William Rowan Hamilton William Edwin Hamilton Elements of Quaternions PDF

H2>
ELEMENTS
OF
QUATERNIONS
BY THE LATE
SIR
WILLIAM ROWAN HAMILTON,

D. C. L.
CANTAB.
LL. D., M, R.
A.,
FELLOW OK THE AMERICAN SOCIEIT OF ARTS AlTD SCir.NCKS;

OF THE SOCIETY OF ARTS FOR SCOTIAND OF THE ROYAL ASTRONOMICAL ROCIKTT OF LONDON; AND OF THB
KOTAL NORTHKRN SOCIETY OF ANTIQ'IAUIES AT COPKNHAGE>f :
CORRESPONDING MKMBER OF THE INSTITUTK OF FRANCE
HONORARY OR CORRESPONDING MEMBER OF THE UrPERTAI- OR ROYAL ACADEMIES OF ST. PETERSBURGK,
BERLIN, AND TURIN OF THE ROYAL SOCIETIES OF EDINBDRGH AND DUBLIN;
OFTHR NATIONAL ACADEMY f)F THE ITNITED STATES;
OF THE CAMBRIDGE PHILOSOPHICAL SOCIETY THE NEW YORK HISTORICAL SOCIETY ;
THE SOCIETY OF NATURAL SCIENCES AT LAUSANNE THE PHILOSOPHICAL SOCIETY OF VENICE
OF
AND
OTHF^ SCIENTIFIC SOCIETIES IN BRITISH AND FORKIGN COUNTRIES
ANDREWS' PROFESSOR OF ASTRONOMY IN THE UNIVERSITY OF DUBLIN
;
AND ROYAL ASTRONOMER OF IRELAND.
EDITED BY HIS SOX,
WILLIAM EDWIN HAMILTON, A.B.T.C.D., C.E.
LONDON
LONGMANS, GREEN,
18G6.
& CO.
DUBLIN:
^rCntett at
t})t
QnibersitB ^reais,
BY M. H.
GILL.
\i^
QA
TO THE
RIGHT HONOEABLE WILLIAM EAEL OF ROSSE,

CHANCELLOR OF THE UNIVERSITY OF DUBLIN,
IS,
BY PERMISSION, DEDICATED,
BY
THE EDITOR.
In
my
late father's
Will no instructions were
to the publication of his Writings,

"
that of the
for his late
nor specially as to
Elements of Quaternions," which, but

fatal illness, would have been before now,
in all their completeness, in the
My
left as
hands of the Public.
brother, the Rev. A. H. Hamilton,
who was
much engaged
in his cle-
named Executor, being

rical duties to
too
undertake the publication, deputed this
task to me.
It
fulfil
was then
for
me
to consider
how
my triple duty in this matter First,
to the
dead
could best
and
chiefly,
secondly, to the present public
and,
came to the conclusion that my duty was to publish the work as I found
it, adding merely proof sheets, partially corrected by
my late father and from which I removed a few typographical errors, and editing only in the literal sense
thirdly, to succeeding generations.
of giving forth.
Shortly before
versations with
my
father's death, I
had several con-
him on the subject of the
"
Elements."
In these he spoke of anticipated applications of Quaternions to Electricity, and to all questions in which
the idea of Polarity is involved
applications which
he never in his own lifetime expected to be able fully
to develope,
bows
to be reserved for the
hands of
another Ulysses.
He also discussed a good deal the
nature of his own forthcoming Preface
and I may
;
intimate, that after dealing with

topics,
its
more important
he intended to advert to the great labour which
the writing of the
"
vi
Elements" had
both mental and mechanical;
as,
cost
him
labour
besides a mass of
subsidiary and unprinted calculations, he wrote out

all the manuscript, and corrected the proof sheets,
without assistance.
And
here I must gratefully acknowledge the generous act of the Board of Trinity College, Dublin, in
relieving us of the remaining pecuniary liability,
and
thus incurring the main expense, of the publication of

volume. The announcement of their intention to
this
gratifying as it was, surprised me the less, when

I remembered that they had, after the publication of
do
so,
former book, " Lectures on Quaternions,"

defrayed its entire cost an extension of their liberality
beyond what was recorded by him at the end of his
my
father's
Preface to the " Lectures," which doubtless he would
have acknowledged, had he lived to complete the Pre"

face of the
Elements."
He intended also, I know, to express his sense of

the care bestowed upon the typographical correctness
of this volume by Mr. M. H. Gill of the
University
and upon the delineation of the figures by the
Engraver, Mr. Oldham.
I annex the commencement of a
Preface, left in maPress,
nuscript by my father, and which he might possibly

have modified or rewritten.
Believing that I have
thus best fulfilled

lished
*'
my
part as trustee of the unpubElements," I now place them in the hands of
the scientific public.
William Edwin Hamilton.

January
1 5^,
1866.
PREFACE
[1.] The volume now submitted to the public is founded on

the same principles as the " Lectures, "^^^ which were published on the same subject about ten years ago
but the plan
:
entirely new, and the present work can in no sense

adopted
be considered as a second edition of that former one. The
is
Table of Contents^ by collecting into one view the headings of

the various Chapters and Sections, may suffice to give, to
readers already acquainted with the subject, a notion of the
course pursued
ductory
sition,
but
it
seems proper to
offer
remarks, especially as regards the
which
it
here a few intro-
method of expo-
has been thought convenient on this occasion
to adopt.
[2.]
designed
The present
to
illustrate it
Three Books, each
develope one guiding conception or view, and to

a sufficient but not excessive number of exam-
by
ples or applications.
iion
treatise is divided into
The
First
Book
relates to the Concept
of a Vector^ considered as a directed right
three dimensions.
line^
The Second Book introduces
in space of
a First Con-
ception of a Quaternion^ considered as the Quotient of two such

And the Third Book treats of Products and Powers
Vectors.
of Vectors^ regarded as constituting a Second Principal

of the Conception of Quaternions in Geometry.
Form
This fragment, by the Author, was found in one of his manuscript books
by the Editor,
TABLE OF CONTENTS,
BOOK
I.
Page
ON VECTORS, CONSIDERED WITHOUT REFERENCE TO

ANGLES, OR TO ROTATIONS,
.
CHAPTER*
1-102
1-11
I.
FUNDAMENTAL PEINCIPLES EESPECTING VECTORS,
On the Conception of a Vector and on EquaOn Differences and Sums of Vectors, taken two
1.
Section!
1-3
lity of Vectors,
Section
2.
3-5
by two,
Section
3.
Section
4.
On Sums of Three more Vectors, ....

On
of Vectors,
5-7
or
8-1
Coefficients
This short First Chapter should be read with care by a beginner

any misconception of the meaning of the word "Vector" being fatal
The Chapter contains explanato progress in the Quaternions.
tions also of the connected, but not all equally important, words
"
or phrases, " re vector," "provector,"
transvector," "actual and
null vectors," " opposite and successive vectors," " origin and term of
a vector," " equal and unequal vectors," "addition and subtraction
of vectors," "multiples and fractions of vectors," &c.
with the notation B - A, for the Vector (or directed right line) ab and a deduction
;
of the result, essential but not pecuUarX to quaternions, that (what

is here called) the vector-sum^ of two co- initial sides of a parallelo"
gram, is the intermediate and co-initial Aiagonal. The term Scalar"
is also introduced, in connexion with coefficients of vectors.
* This
Chapter may be referred to, as I. i. ; the next as I.
Second Book, as II. i. and similarly for the rest.
ter of the
t This Section
ii.
the
first
Chap-
may
be referred
to, as I.
i.
the next, as
Section of the second Chapter of the Third Book, as III.

X Compare the second Note to page 203.
ii.
I.
;
i.
and so
the sixth
on.
CONTENTS.
ii
Pages.
CHAPTER
ArrLICATIOIs^S TO POINTS
II.
AND LINES IN A GIVEN PLANE,
On Linear Equations connecting two

Vectors,
On Linear Equations between three
Section
Section
1.
Co-initial
2.
Co-initial
11-49
11-12
12-20
Vectors,
After reading these two first Sections of the second Chapter, and
perhaps the three first Articles (31-33, pages 20-23) of the following
Section, a student to whom the subject is new may find it convenient
to pass at once, in his first perusal, to the third Chapter of the present
Book; and to read only the two first Articles (62, 63, pages 49-51)
first Section of that Chapter, respecting Vectors in Space, before
proceeding to the Second Book (pages 103, &c.), which treats of Qua-
of the
ternions as Quotients of Vectors.
On Plane Geometrical
......
On Anharmonic Co-ordinates and Equations
of Points and Lines in one Plane,
On Plane Geometrical
Section
resumed,
On Anhaimonic Equations and Vector ExSection
Section
3.
Section
4.
Iêts,
5.
iN'ets,
20-24
24-32
32-35
6.
pressions, for
Among
35-49
Curves in a given Plane,
other results of this Chapter, a theorem is given in page 43,

to ofifer a new geometrical generation of (plane or spheri-
which seems
cal) curves of the third order.
The anharmonic
co-ordinates
and equa-
tions employed, for the plane and for space, were suggested to the
writer by some of his own vector forms; but their geometrical interThe geometrical nets were first discussed
pretations are assigned.
by
Professor Mobius, in his Barycentric Calculus (Note B), but they are
treated in the present work by an entirely new analysis and, at least
for space, their theory has been thereby much extended in the Chapter
:
to
which we next proceed.
CHAPTER
III.
applications of vectors to space,
Section
1.
.49-102
On Linear Equations between Vectors not Com49-56
planar,
been recommended to the student to read the first
two Articles of this Section, even in his first perusal of the Volume
and then to pass to the Second Book.
It has already
Section 2
Space,
On Quinary Symbols
for Points
and Planes in
57_62
CONTENTS.
Section
3.
Section
4.
HI
On Anhannonic Co-ordinates in Space,

On Geometrical
in Space,
Pages.
.
^^Tets
On Barycentres of Systems of Points and on

and
Complex Means of Yectors,
Simple
On Anhannonic Equations, and Yector ExSection 6,
Section 5
62-67
67-85
85-89
pressions, of Surfaces and Curves in Space,
Section
An
7. On
....
Differentials of Yectors,
90-97
98-102
application oifinite differences^ to a question connected with baThe anharmonic generation of a ruled hy-
rycentres^ occurs in p. 87.
perboloid (or paraboloid) is employed to illustrate anharmonic equations and (among other examples) certain cones^ of the second and third
;
orders,
have their vector equations assigned. In the last Section, a defirand scalars) is proposed, which is
nition of differentials {pi vectors
afterwards extended to differentials of quaternions, and which is independent of developments and of infinitesimals, but involves the
Vectors of Velocity and Acceleration are
conception of limits.
tioned ; and a hint of Sodographs is given.
BOOK
men-
11.
ON QUATERNIONS, CONSIDERED AS QUOTIENTS OF

VECTORS, AND AS INVOLVING ANGULAR RELA103-300
TIONS,
CHAPTEE
I.
FUNDAMENTAL PRINCIPLES RESPECTING QUOTIENTS OF VECTORS, 103-239

any, of this Chapter II. i., should be omitted, even
perusal since it contains the most essential conceptions
and notations of the Calculus of Quaternions, at least so far as quotients of vectors are concerned, with numerous geometrical illustra-
Very
in a
little, if
first
tions.
Still
there are a few investigations respecting circumscribed

and ellipsoids, in the thirteenth Sec-
cones, imaginary intersections,
which a student may pass over, and which will be indicated in

the proper place in this Table.
tion,
First Principles
1
Introductory Remarks
103-106
adopted from Algebra,
Section 2. First Motive for naming the Quotient of two
Yectors a Quaternion,
106-110
Section
Sections. Additional
It is
Illustrations,
shown, by consideration of an angle on a desk, or inclined

the complex relation of one vector to another, in length and
l)lane, that
110-112
CONTENTS.
iv
in direction, involves generally a system oifour numerical elements.

"
Many other motives, leading to the adoption of the name, Quaternion," for the subject of the present Calculus, from its fundamental
connexion with the number " Four," are foimd to present themselves
in the course of the work.
On Equality of Quaternions
Sectiost 4
and on the Plane
112-117
of a Quaternion,
On the Axis and Angle of a Quaternion and

117-120
on the Index of a Eight Quotient,
Quaternion,
On the Eeciprocal, Conjugate, Opposite, and
Section
120-129
lôrm of a Quaternion; and on
Quaternions,
On Eadial Quotients and on the Square of a
Section
129-133
Quaternion,
On the Yersor of a Quaternion, of a YecSection
Section
5.
or
6.
]S"ulI
7.
or
8.
tor
and on some General Formulae of Transformation, 133-142
In the five foregoing Sections it is shown, among other things,

that the plane of a quaternion is generally an essential element of its
constitution, so that diplanar quaternions are unequal; but that the
tquare of every right radial (or right versor)
whatever
its
plane
may
be.
The Symbol V -
terpretation, in this as in several other
as real, it
account
equal to negative
unitt/j
admits then of a real in-
systems
but when thus treated
in the present Calculus too vague to be useful on which

is foimd convenient to retain the old signification of that
is
it
is
symbol, as denoting the (uninterpreted) Imaginary of Algebra, or

called the scalar imaginary, in investigations re-
what may here be
specting non-real intersections, or non-real contacts, in geometry.
On "Vector- Arcs, and Yector- Angles, considered as Eepresentatives of Yersors of Quaternions ;

and on the Multiplication and Division of any one such
Section 9
Yersor by another,
142-157
is important, on accoxmt of its constructions of muland division which show that the product of two diplanar
and therefore of two such quaternions, is not independent of
This Section
tiplication
versors,
the order of the factors.
On a System of Three Eight Yersors, in

Three Eectangular Planes and on the Laws of the
Section 10.
Symbols,
157-162
ijk,
The student ought

which are
all
to make hxm&eM familiar with these laws,

included in the Fundamental Formula,
'
i2=p = h'i =
ijk^--'[.
(A)
CONTENTS.
Pages.
In
form,
q=
in "which
while
t,
iv
+ ix+jy + kz,
k are
y,
we
since
(B)
four scalars, or ordinary algebraic quantities,

new symbols, obeying the laws contained in the
tv, x, y, z are
tliree
formula (A), and therefore not

bra
symbolically defined to be a Quadrino-
Quaternion may be
fact, a
niial JSxpresston of tlie
y = + A-,
Section
11
nion
usual rules of alge-
subject to all the
have, for instance,
but ji = -k;
and
iy^k^
=- {ifk^.
On the Tensor of a Vector,
or of a Quater-
and on the Product or Quotient of any two Qua-
162-174
ternions,
Section 12
On
ternions
ternion,
Section 13.
the
Sum
and on the Scalar
On
Quaternion
any two QuaScalar Part) of a Qua-
or Difference of
(or
175-190
the Right Part (or Vector Part) of a
and on the Distributive Property of the
190-233
Multiplication of Quaternions,
Section 14.
On the Reduction of the General Quaternion
to a Standard
Quadrinomial
Form
with a Pirst Proof
of the Associative Principle of Multiplication of Qua-
233-239
ternions,
Articles
omitted at
213-220 (with their sub-articles), in pp. 214-233, may be
first
reading.
CHAPTER
II.
ON COMPLANAE QUATERNIONS, OR QUOTIENTS OF VECTOES IN

ONE PLANE AND ON POWEES, EOOTS, AND LOGAEITHMS OP
;
240-285
QUATEENIONS,
The
in a
first six
first
Sections of this Chapter (II.
ii.)
may be
passed over
perusal.
Section
1.
On Complanar Proportion of Vectors;
Fourth
Third Proportional to Two,

Mean Proportional, Square Root General Reduction
of a Quaternion in a given Plane, to a Standard Binomial Form,
240-246
Proportional to
Three,
Section
tors
On Continued Proportion of Four
more Vecwhole Powers and Roots of Quaternions and
2.
Roots of Unity,
or
246-251
CONTENTS.
VI
Pages.
Section
3. On
the Amplitudes of Quaternions in a given
Plane; and on Trigonometrical Expressions for such

251-257
Quaternions, and for their Powers,
Section 4. On the Ponential and Logarithm of a Quater-
nion
and on Powers of Quaternions, with Quaternions
for their
Section
5.
Exponents,
On Finite (or Polynomial) Equations of Alge-
braic Form, involving Complanar Quaternions

the Existence of n Eeal Quaternion Eoots, of
257-264
and on
any such
265-275
Equation of the w''^ Degree,
Section 6. On the n^ - n Imaginary (or Symbolical)
Roots of a Quaternion Equation of the n*^ Degree, with
Coefficients of the kind considered in the foregoing
275-279
Section,
Section
7.
On the
Reciprocal of a Vector, and on Harof Vectors

with Remarks on the Anhar-
monic Means
monic Quaternion of a Group of Pour Points, and on
;
279-285
Conditions of Concircularity,
In
this last Section (II.
ii.
7) the short first Article 258,
and the
following Art. 259, as far as the formula YIII. in p. 280, should be

read, as a preparation for the Third Book, to which the Student may
next proceed.
CHAPTER
III.
ON DIPLA-NAK QUATERNIONS, OE QUOTIENTS OF VECTORS IN

SPACE AND ESPECIALLY ON THE ASSOCIATIVE PRINCIPLE
:
OF MULTIPLICATION OF SUCH QUATERNIONS,

This Chapter
Section
1.
may
be omitted, in a
286-300
first perusal.
On some Enunciations of the Associative Pro-
perty, or Principle, of Multiplication of Diplanar
Qua-
ternions,
Section
2.
On some Geometrical Proofs of the Associative
286-293
Property of Multiplication of Quaternions, which are
independent of the Distributive Principle,

Section 3. On some Additional FormuloB,
....
....
293-297
297-300
CONTENTS.
BOOK
Vll
III.
ON QUATERNIONS, CONSIDERED AS PRODUCTS OR

POWERS OF VECTORS; AND ON SOME APPLICA-
Pages.
301 to the end.
TIONS OF QUATERNIONS,
CHAPTER
I.
ox THE lUTEEPEETATION OF A PEODUCT OP VECT0E8, OE POWER

301-390
OF A YECTOE, AS A QTJATEENIOX,
.
The
first
first six
Sections of this Chapter ought to be read, even in a
perusal of the work.
Section
of
On a
First Method of Interpreting a Product

Yectors as a Quaternion,
301-303
On some Consequences of the foregoing Inter-
Two
Section
2.
303-308
pretation,
This^r*^ interpretation treats the product

qmtietit
(II.
ii.
where
equal to the
(i.a, as
the previously defined Reciprocal

7) of the vector a, namely a second vector^ which has an m/3
a-i
a~^ (or
Ro)
is
and an opposite direction. Multiplication of Vectors is

thus proved to he (like that of Quaternions) a Distributive, but not
generally a Commutative Operation. The Square of a Vector is shown
verse length,
to be always a Negative Scalar, namely the negative of the square of

the tensor of that vector, or of the number which expresses its length ;
and some geometrical applications of

&c., are given.
The Index of the
initial Vectors, oa, ob, is
this fertile principle, to spheres^
RigJit
Part of a Product of Two Co-
proved to be a right
line,
perpendicular
to
Plane of the Triangle cab, and representing by its length the

Double Area of that triangle while the Rotation round this Index,
the
the Multiplier to the Multiplicand, is positive.
from
This right part,
or vector part, Va/3, of the product vanishes, when the factors are
parallel (to one common line) ; and the scalar part, Sa/3, when they
are rectangular.
Section
3.
On a Second Method of arriving the same

of a Binary Product of Yectors,
308-310
On the Symbolical
a Right
at
Interpretation,
Section
4.
Identification of
Quaternion with
its
own Index
and on the Construc-
tion of a Product of Two Rectangular Lines,
by a Third
310-313
some Simplifications of JNotation, or of
Expression, resulting from this Identification and on
the Concei)tion of an Unit-Line as a Right Tersor,
313-316
Line, rectangular to both,
Section
5.
On
CONTENTS.
viii
Pages.
In
sults
found to agree in all its rebut is better adapted to an extension of the theory,
this second interpretation,
with the
first,
which
is
as in the following Sections, to ternary products of vectors, a product

of the two riffht quaternions, of
of two vectors is treated as the product
which those vectors are the indices (II. i. 5). It is shown that, on
the same plan, the Sum of a Scalar and a Vector is a Quaternion.
Section
or
6.
On the
more Vectors
Interpretation of a Product of Three
316-330
as a Quaternion,
by the substitution, as in recent

oi order of
Sections, of Bight Quaternions for Vectors, without change
the factors. Multiplication of Vectors, like that of Quaternions, is
This interpretation
is eflfected
thus proved to be an Associative Operation.

reduced to the Standard Trinomial Form,
p
= ix +ji/ +
A vector,
generally, is
kz ;
(C)
are the peculiar symbols already considered (II. i.

Vector- TTnits,
10), but are regarded now as denoting Three Rectangular
while the three scalars x, y, z are simply rectangular co-ordinates ; from
in
which
i,
j,
Jc
known theory of which last, illustrations of results are derived.

The Scalar of the Product of Three coinitial Vectors, oa, ob, oc, is found
the
to represent, with a sign depending
Volume of
nishes
the Parallelepiped
when they
direction of a rotation, the
so that
it
va-
are complanar. Constructions are given also for pro-
ducts of successive sides of triangles,
in
on the
under those three lines
and other closedpolygons, inscribed
or in spheres ; for example, a characteristic property of the

contained in the theorem, that the product of the four suc-
circles,
circle is
cessive sides of
an
inscribed quadrilateral is a scalar
and an equally
characteristic (but less obvious) property of the sphere is included in
an
drawn from
this other theorem, that the product oii\iefive successive sides of
inscribed gauche pentagon is equal to a tangential vector,
the point at which the pentagon begins (or ends). Some general Formula of Transformation of Vector Expressions are given, with which
a student ought to render laimsQli very familiar, as they are of continual occurrence in the practice of this Calculus
formulae (pp. 316, 317)
especially the four
V.yV/3a=aS/3y-/3Sya;
= aS^y-/3S7a+ySa/3;
(D)
Vy/3a
(E)
+ /3Syap + ySa/3p
pSai3y=Y/3ySap+VyaS^p + Va/3Syp;
pSajSy
in
which
a,
j8,
y, p are
aS/3yp
any four
(F)
vectors,
while S and
(G)
V are
signs of
the operations of taking separately the scalar and vector parts of a quaOn the whole, this Section (III. i. 6) must be considered
ternion.
to be (as regards the present exposition)
an important one and if

have been read with care, after a perusal of the portions previously
indicated, no difficulty will be experienced in passing to any subse;
it
quent applications of Quaternions, in the present or any other work.
CONTENTS.
Section
7.
IX
On the Fourth Proportional to Three Diplanar

331-349
Vectors,
Section
On an Equivalent Interpretation of the Fourth
8..
Proportional to Three Diplanar Yectors, deduced from

349-361
the Principles of the Second Book,
Section
9.
On a Third Method of interpreting a Product
or Function of Vectors as a Quaternion
and on the
Consistency of the Results of the Interpretation so obtained, with those which have been deduced from the
two preceding Methods of the present Book,
361-364
These three Sections may be passed over, in a first reading. They

contain, however, theorems respecting composition of successive rotations (pp. 334, 335, see also p. 340); expressions for the ^m^-arm of a
an arbitrary pyramid^ as
the angle of a quaternion product^ with an extension, by limits, to the
semiai-ea of a spherical figure bounded by a closed curve^ or to half the
spherical polygon, or for half the opening of
opening of an arbitrary cone (pp. 340, 341)
a construction (pp.
358-
360), for a series of spherical parallelograms, so called from a partial

analogy to parallelograms in o, plane ; a theorem (p. 361), connecting
a certain system of such (spherical) parallelograms with the foci of a

and the concepspherical conic, inscribed in a certain quadrilateral
;
361) of a Fourth Unit in Space (u, or + I), which is of

a scalar rather than a vector character, as admitting merely of change
of sign, through reversal of an order of rotation, although it presents
tion (pp. 353,
theory as the Fourth Proportional (Jj'^k)

angular Vector Units.
itself in this
to
Three Sect-
\l/Section 10. On the Interpretation of a Power of a Vector

as a Quaternion,
364-384
It
may
weU
be
to read this Section (III.
i.
10), especially for
the Exponential Connexions which it establishes, between Quaternions

and Spherical Trigonometry, or rather Polygonometry, by a species of
extension of Moivr^s theorem,
from the plane to space, or to the sphere.

For example, there is given (in p. 381) an equation of six terms,
which holds good for every spherical pentagon, and is deduced in this
way from an extended exponential formula. The calculations in the
sub-articles to Art. 312 (pp. 875-379) may however be passed over ;
and pei'haps Art. 315, with its sub-articles (pp. 383, 384). But Art.
314, and its sub-articles, pp. 381-383, should be read, on account of

the exponential forms which they contain, of equations of the circle,
ellipse,
logarithmic spirals (circular and elliptic), helix,
^^Ql.
screw sur-
face.
Section 11.
On Powers and Logarithms of Diplanar Quawith some Additional Formulae, ....
ternions
384-390
CONTENTS.
Pages.
read Art. 316, and its first eleven sub-articles,

In this Section, the adopted Logarithm^ Iq, of a Quapp. 384-386.
ternion q, is the simplest root, q\ of the transcendental equation,
It
and
may
its
suffice to
expression
is
found to
1^
in -which
T and
be,
= 1T^ + Z^.UV?,
U are the signs
(H)
of tensor and versor, while
Lq\B the
and tt. Such logarithms

q, supposed usually to be between
are found to be often useful in this Calcuhis, although they do not generally possess the elementary property, that the sum of the logarithms
angle of
of two quaternions is equal to the logarithm oi iliQix product : this apparent paradox, or at least deviation from ordinary algebraic rules,
arising necessarily from the corresponding property of quaternimi
not generally a commultiplication, which has been already seen to be
mutative operation {q'q" not - q"q', unless q' and q" be complanar).
here, perhaps, a student might consider his first perusal of this
And
work
as closed.*
CHAPTER
II.
ON DIFFERENTIALS AND DEVELOPMENTS OF FUNCTIONS OF QUATERNIONS AND ON SOME APPLICATIONS OF QUATERNIONS

TO GEOMETRICAL AND PHYSICAL QUESTIONS,
391-495
;
been already said, that this Chapter

perusal of the work.
It has
first
Section
1.
On the
may
be omitted in a
Definition of Simultaneous Differen-
tials,
391-393
* If he should choose to
proceed to the Differential Calculus of Quaternions in
the next Chapter (III. ii.), and to the Geometrical and other Applications in the
third Chapter (III. iii.) of the present Book, it might be useful to read at this
stage the last Section (I. iii, 7) of the First Book, which treats of Differentials of
Vectors (pp. 98-102); and perhaps the omitted parts of the Section II. i. 13,
namely
Articles 213-220, with their subarticles (pp. 214-233),
wbich
relate,
other things, to a Construction of the Ellipsoid, suggested by the present

Calculus. But the writer will now abstain from making any further suggestions
of this kind, after having indicated as above what appeared to him a minimum
among
course of study, amounting to rather less than 200 pages (or parts of pages)
of this Volume, which will be recapitulated for the convenience of the student
at the
end of the present Table.
CONTENTS.
Section
2.
Elementary
XI
Pages.
of the Definition,
Illustrations
from Algebra and Geometry,
394-398
In the view here adopted (comp. I. iii. 7), differentials are not neBut it is shown at a later stage
cessarily, nor even generally, small.
(Art. 401, pp. 626-630), that the principles of this Calculus
^^toz^j
us,
whenever any advantage may be thereby gained, to treat differentials

as infinitesimals ; and so to abridge calculation^ at least in many applications.
Section
3.
On some general Consequences of the Defini398-409
tion,
Partial differentials and derivatives are introduced
tials oi functions offunctions.
Section
4.
Examples of Quaternion
and
differen-
Differentiation,
409-419
One of the most important rules is, to differentiate tjie /actors of a

quaternion product^ in situ ; thus (by p. 405),
dL.qq'
The formula
(p. 399),
= ^q.q' + qAq'.
= -q-^dq.q-\
di.q-^
(I)
(J)
for the differential of the reciprocal of a quaternion (or vector), is also

very often useful ; and so are the equations (p. 413),
dT^
(K)
and
(p.
d.a
411),
a+id^
(L)
q being any quaternion, and a any constant vector-unit, while Hs a

variable scalar.
It is important to remember (comp. III. i. 11), that
we have
not in quaternions the usual equation,
dl^
unless q
if
and d^ be complanar ; and therefore that we have not generally^
p be a variable vector
although
scarcely less simple equation,
which
we
is
have^ in this Calculus, the

useful in questions respecting
orbital motion^
if
a be any constant vector, and
if the
plane of a and p be given (or
constant).
Section
5.
On Successive
Differentials
of Functions of Quaternions,
and Developments,
420-435
CONTENTS.
Xil
Pages.
In
this Section principles are established (pp.
ing qnateraion functions which vanish together
423-426), respectand a form of deve-
lopment (pp. 427, 428) is assigned, analogous'^ to Taylor's Series^

and like it capable of being concisely expressed by the sytnboUcal
As an example of partial and succeseqtmtion, 1 + A = i^ (-p. 432).
sive differentiation, the expression (pp. 432, 433),
= rk*j^kj-^k~^y
may represent ang vector, is operated on ; and an application

made, by means of definite integration (pp. 434, 435), to deduce the
known area and volume of a sphere, or of portions thereof together
which
is
with the theorem, that the vector sum of the directed elements of a
spheric segment is zero : each element of surface being represented by an
inward normal, proportional to the elementary area, and corresponding in hydrostatics to the pressure of a fluid on that element.
Section
6.
On the
Differentiation of Implicit Functions
and on the General Inversion of a Linear Tunction, of a Yector or a Quaternion

with
some connected Investigations,
435-495
of Quaternions
In this Section it is shown, among other things, that a Linear

and Vector Symbol, (p, of Operation on a Vector, p, satisfies (p. 443) a
Symbolic and Cubic Equation, of the form,
= w m'<{> + m"(p^ - <p^

= m' m'<p + <p^=\p,
(N)
whence
m^"i
(N')
another symbol of linear operation, which it is shown how to deduce otherwise from (p, as well as the three scalar constants, m, m', ni'.
The connected
algebraical cubic (pp. 460, 461),
Jf = w + m'c + m"c2 + c' =

is
found to have important applications
and
(0)
0,
it is
provedf (pp. 460,

462) that if SX^p = Sp^\, independently of X and p, in which case
the fimction ^ is said to be self-conjugate, then this last cubic has three
real roots,
Ci, c^, cz
while, in the same case, the vector equation,
Yp<pp
is satisfied
(P)
0,
by a system of Three Heal and Rectangular
namely (compare pp. 468, 469, and the Section
III.
iii.
Directions
7),
those of
the axes of a (biconcyclic) system of surfaces of the second order, represented by the scalar equation,
* At a later
stage (Art. 375, pp. 509, 510), a new Enunciation of Tayloi^s
is given, mth a new proof, but still in
a/orm adapted to quaternions.
Theorem
simplified proof, of
self-conjugation, is
(pp. 698, 699).
some of the chief
results for this important case
given at a later stage, in the few
first
subarticlcs to Art.
of
415
CONTENTS.
Sp<pp
Cp"^
+ C,
in
which C and
Xlll
Pages.
C are constants.
(Q)
and general forms (called cyclic, rectangular,

focal, bifocal, &c., from their chief geometrical uses) are assigned,
one useful pair of
for the vector and scalar functions 0p and Sp^p
such (cyclic) forms being, with real and constant values of g, X, /j,
Cases are discussed
(pp=gp+YXpn,
And
Sp(pp
= ffp^ + S\pfip.
(R)
finally
(pp. 491, 492) that iffg be a linear and quaternion function of a quaternion, q, then the Symbol of Operation, f,
satisfies a certain Symbolic and Biquadratic Equation, analogous to the
shown
it is
cubic equation in 0,
and capable of similar applications.
CHAPTER
III.
ON SOME ADDITIONAL APPLICATIONS OF QIJATEENIONS, WITH

495 to the end.
SOME CONCLUDING kemaee;s,
.
This Chapter, like the one preceding it, may be omitted in a

perusal of the Volume, as has indeed been already remarked.
Section
1.
Remarks
Chapter,
Section 2
first
Introductory to this Concluding
495-496
On Tangents and
iN'onnal Planes to Curves in
Space,
On
and Tangent Planes
On Osculating Planes,
and Absolute lôrmals,
Section
Curves of Double Curvature,
On Geodetic Lines, and Families of
Section
Section
to Surfaces,
JSôrmals
3.
496-501
501-510
4.
to
5.
Surfaces,
511-515
515-531
In these Sections, dp usually denotes a tangent to a curve, and v

to a surface.
Some of the theorems or constructions may
a normal
perhaps be
lels
new
for instance, those connected
with the cone of paral-
(pp. 498, 513, &c.) to the tangents to a curve of double curvature ;
and possibly the theorem

space
(p. 525),
respecting reciprocal curves in
at least, the deductions here given of these results
may
serve
as exemplifications of the Calculus employed. In treating of Families

of Surfaces by quaternions, a sort o^ analogue (pp. 629, 530) to the for-
mation and integration of Partial Differential Equations presents

itself; as indeed it had done, on a similar occasion, in the Lectures
(p. 574).
Section
6.
On Osculating
in Space; with
The
Circles and Spheres, to Curves

some connected Constructions,
531-630
analysis, however condensed, ofthis long Section
(III.
iii.
6),
cannot conveniently be performed otherwise than under the heads of

the respective Articles
(389-401) which compose it: each Article
CONTENTS.
XIV
Pages.
being followed by several subarticles, which form with

Series*
a sort of
it
Article 389. Osculating Circle defined^ as the limit of a circle,

which touches a given curve (plane or of double curvature) at a given
point p, and cuts the curve at a near point q (see Fig. 77, p. 511).
Deduction and interpretation of general expressions for the vector k
of the centre k of the circle so defined. The reciprocal of the radius
KP being called tbe vector of curvature^ we have generally^
Vector of Curvature
and
(o
- k)-i = -p =
Idp
of the curve be
if the arc (s)
made
Vector of Curvature
Y --? = &c.
dp
(S)
dp
the independent variable, then
p" = D,8p
= -^.
(S')
Examples curvatures of helix, ellipse, hyperbola, logarithmic spiral ;

locus of centres of curvature of helix, plane e volute of plane ellipse,
Ajiticle 390
Abridged general calculations; return from (S')
531-535
to (S),
535, 536
Article 391
Centre determined by three scalar equations
Polar Axis, Folar Developable,
F<;côr^^Mai/ow of osculating
Article 392.
Article 393.
j9?ffw^ CMr?;^ to
Intersection (or
which
it
osculates
537
538,539
circle,
intersections) of a circle
with a
539-541
example, hyperbola,
Article 394. Intersection (or intersections) of a spherical curve

with a small circle osculating thereto example, spherical conic ; con;
structions for the spherical centre (or pole') of the circle osculating to
such a curve, and for the point of Mi^^rse<?^w above mentioned, . .
Article 395
541-549
Osculating Sphere, to a curve of double curvature,

defined as the limit of a si^here, which contains the osculating circle to
the curve at a given point p, and cuts the same curve at a near point
Q (comp. Art. 389). The centre s, of the sphere so found, is (as usual)
the point in which the polar axis (Art. 391) touches the cusp-edge of
the polar developable.
Other general construction for the same centre
General expressions for the vector, & = os,

(p. 551, comp. p. 573).
= sf; i?" Ws the spherical curvature (comp. Art.
and for the radius,
Condition of Sphericity
1), for a curve in space.
397).
(S
(as
it is
{S=
and
1),
"When
Coefficient
of Non-sphericity
this last coefficient is positive
for the helix), the curve lies outside the spliere, at least in the
neighbourhood of the point of osculation,
Article 396.
Notations
r, t,
549-553
for D^p, Dg'^p, Sec.
properties
of a curve depending on the square (s^) of its arc, measured from a

=
given point p ; r miit-tangent, t = vector of curvature, r~' = Tr' = curvature (or first curvature, comp. Art. 397), v
tt'
binormal
the
A Table of initial Pages of all the Articles will be elsewhere given, which will
much
facilitate reference.
XV
CONTENTS.
Pages.
three planes^ respectively perpendicular to r,
v, are the
r',
normal
and the osculating plane ; general theory

of emanant lines and planes, vector of rotation, axis of displacement, oscillating screiv surface ; condition of developability of surface of emanants,
plane, the rectifying plane,
654559
Article 397. Properties depending on the cube (s^) of the arc ;

Radius r (denoted here, for distinction, by a roman letter), and Vector
r-'r, of
Second Curvature ; this radius r
be either positive or ne-
may
gative (whereas the radius r of /rs^ curvature is
and
positive),
its reciprocal r"^
may
always treated as
be thus expressed (pp. 663, 559),
d^n
Second Curvature* =r-i
= S ,,,
r-
(T), or,
=S
t"
TT
Vdpd^p
(T')
the independent variable being the arc in (T'), while it is arbitrary in

but quaternions supply a vast variety of other expressions for this
important scalar (see, for instance, the Table in pp. 674, 575).
(T)
We
have
also
(by
p. 560,
comp. Arts. 389, 395, 396),
Vector of Spherical Curvature
= projection
= sp"i=
(p
(r)-i
&c.,
(U)
of vector (r') ot (simple or first) curvature, on radius (li)
denote the linear and angular

of osculating sphere : and if p and
elevations, of the centre (s) of this sphere above the osculating plane,
then (by same page 660),
^ = r tan P= i2 sin P = r'r = rD,r.

Again
(pp. 560, 561), if
=V
=r-ir
T
this line
+ rr' =
X may be
we
(U')
write (comp. Art. 396),
Vector of Second Curvature plus Binormal, (V)
called the Rectifying Vector
inclination (considered first
by Lancret), of this
and if
H denote the
rectifying
Urn (\) to
the tangent (j) to the curve, then
tan
Known
H = /-
tan
P=
>-
'
r.
right cone with rectifying line for its axis,
semiangle,
which
osculates at
p to the developable
(V)
and with
for its
locus of tangents to
the curve (or by p. 568 to the cone of parallels already mentioned)

new right cone, with a new semiangle, C, connected with
by the
;
relation (p. 562),
tanC=^tan^,
(V")
which osculates to the cone of chords, drawn from the given point p
* In this Article, or Series, 397, and indeed also in 396 and

398, several references are given to a very interesting Memoir by M. de Saint-Venant, " Sur
les lignes courbes nx)n planes ." in which, however, that able writer objects to such
known phrases as second curvature, torsion, &c., and proposes in their stead a new
name " cambrure," which it has not been thought necessary here to adopt.
{Journal de T E'eole Polytechnique, Cahier xxx.)
CONTENTS.
XVI
Pagus.
Other osculating cones, cylinders,

Mix, and parabola ; this last being (pp. 562, 566) the parabola which
osculates to the projection of t/ie curve, on its own osculating plajie. De-
to other points
o.
of the given curve.
any near point a, from the osculating circle at p,

decomposed (p. 666) into tivo rectangular deviations, from osculating
helix and parabola. Additional formulae (p. 676), for the general
theory of emanants (Art. 396) case o normally emanant lines, or of
viation of curve, at
tangentialhj
emanant
planes.
General auxiliary spherical curve (pp.
676-578, comp. p. 515) new proof of the second expression (V) for
tan JEf, and of the theorem that if this ratio of curvatures be constant,
the proposed curve is a geodetic on a cylinder : new proof that if each
;
curvature
(r'^, r"i)
be constant, the cylinder
is right,
and therefore
659-578
the curve a helix,
Article 398. Properties of a curve in

fourth and Jifth powers (s^, s^) o its arc (s'),
space, depending
on the
578-612
much
longer than any other in the Volume,

and is supposed to contain so much original matter, that it seems
necessary here to subdivide the analysis iinder several separate heads,
This Series 398
is so
lettered as (a), (b), (c), &c.

(a).
Neglecting
we may
s^,
write (p. 578, comp.
OT?s = ps = p + ST+ |sV + ^s^t" + -i^sir"
or (comp. p. 587),
with expressions
ps
(p.
in terms of r, r,
= p + XsT-\- ysrr' +
588) for the
and
r", r, r',
s.
s^
396),
(W)
Zsrv,
coefficients
If
Ai't.
(or co-ordinates')
("W")
Xs, y, Zs,
be taken into account,
it
be-
comes necessary to add to the expression ("W) the term, x^s^r";

with corresponding additions to the scalar coefficients in (W), introducing
r'"
and
for extending
r"
the laws for forming which additional terms, and

to higher powers of the arc, are assigned in a
them
subsequent Series (399, pp. 612, 617).

(h).
Analogous expressions for r'", v", k", X', </, and p', K, P', "',
which s^ is neglected, are assigned (in p. 579)
to serve in questions in
t"
v',
k, \,
Series 397)
<T,
and p, R, F, H, having been previously expressed (in

while r", v", k", \", <t", &c. enter into investigations
^ the arc s being treated as the independent
which take account of
variable in all these derivations.

(c). One of the chief results of the present Series (398), is the
introduction (p. 581, &c.) of a fiew auxiliary angle, J, analogous in
several respects to the known angle
(397), but belonging to a
higher order of theorems, respecting curves in space : because the neio
angle / depends on \he fourth (and lower) powers of the arc s, while
Lancret's angle
depends only on s^ (including s^ and s"^). In fact,
while tan îs represented by the expressions (V), whereof one is
r'l tan P, tan /admits (with many transformations) of the following
analogous expression
(p. 581),
tan/=ie'-itanP;
(X)
CONTENTS.
xvii
JR' depends* by (b) on *, while / and

depend (397) on no
higher power than s^.
{d). To give a more distinct geometrical meaning to this new angle
/, than can be easily gathered from such a formula as (X), respecting
where
which
it
defined
may
Pages,
be observed, in passing, that /is in general more simply

for its cotangent (pp. 581, 588), than for its
by expressions
tangent, we are to conceive that, at each point p of any proposed

curve of double curvature, there is drawn a tangent plane to the sphere ^
which
and that then the

which envelope (for reasons
osculates (395) to the curve at that point
envelope of all these planes is determined,
afterwards more fully explained) is called here (p. 581) the " Circumscribed Developable :" being a surface nfl/oÔMs to the '''Rectifying
Developable" of Lancret, but belonging (c) to a higher order of ques-
And
tions.
then, as the
Z:;om;w
angle if denotes (2>^1^ ihe inclina-
suitably measured, of the rectifying line (X), which is a. generatrix of the rectifying developable, to the tangent (r) to the curve; so
tion,
the new angle
represents the inclination of a generating line (jp), of

just been called the circumscribed developable, to the same
tangent (r), measured likewise in a defined direction (p. 581), but
in the tangent plane to the sphere. It may be noted as another ana-
what has
logy
(p.
is right
gles
JS"
582), that while
when
and
the curve
/ are
lopables coincide,
which the helix
equal
is
a right angle for a plane curve, so /

For the helix (p. 585), the an-
is spherical.
and the rectifying and circumscribed
with each other and with the right
deve-
cylinder,
on
a geodetic line.
be measured from the given point p, in
(e). If the recent line
a suitable direction (as contrasted with the opposite), and with a suitable length,
it
is
becomes what
may
be called (comp. 396) the Vector of
Rotation of the Tangent Flane (d) to the Oscidating Sphere ; and then
it satisfies, among others, the equations (pp. 579, 581, comp.
(V)),
= V ^,
V
this last being
T0 = i2-i cosec /;
an expression for the
velocity
(X')
of rotation of the plane
just mentioned, or of its normal, namely the spherical radius R, if the

given curve be conceived to be described by a point moving with a con-
* In other
introduces no difierentials
words, the calculation of r' and
higher than the third order ; but that of R' requires \h.Q fourth order of difierentials.
In the language of modern geometry, \kiQ former can be determined by
the consideration oifour consecutive points of the curve, or by that of two consecutive osculating circles ; but the latter requires the consideration of two consecutive osculating spheres,
and therefore of five consecutive ^;o?V^ ^5 of the curve (supOther investigations, in the present and
posed to be one of double curvature).
immediately following Series (398, 399), especially those connected with what
we shall shortly call the Osculating Twisted Cubic, will be found to involve the
consideration of six consecutive points of a curve.
CONTENTS.
xviii
Pages
stantvelocitij,
assumed
= 1. And
denote by v the point in which

nearest to a consecutive radius of the same
if
we
or PS is
the given radius
then this point
kind, or to the radius of a consecutive osculating sphere,
divides the line ps internally, into segments
v
which may (ultimately) be
thus expressed (pp. 580, 581),
Fv = i2sin2/,
But these and other connected
vs=i2cos2/.
results,
(X")
depending on s\ have their
for 7, and r for R), in that earlier theory

analogues (with
introduces only s^ (besides s^ and 6-2) and they are all in(c) which
cluded in the general theory oiemanant Knes and planes (396, 397), of
which some new geometrical illustrations (pp. 582-584) are here
known
given.
(/).
New
auxiliary scalar n {=p-'^RR'
= cot/secP = &e.), = ve-
locity of centre s of osculating sphere, if the velocity of the point
p of
the given curve be taken as unity (e) ; n vanishes with R', cot /, and
(comp. 395) the coefficient S-1 (=wrr~i) of non- sphericity, for the
case of a spherical curve (p. 584). Arcs, first and second curvatures,
lines, of the cusp-edges of the polar and
these can all be expressed without going
and some without using any higher power than s*, or diffe-
and rectifying planes and

rectifying* developables
beyond
s^,
r\ nr, and ri = nr, are the

and second curvature oi the former cusp-edge, ri
rentials of the orders corresponding

scalar radii of first
being positive when that curve turns its concavity at s towards the
determination of the point r, in which the latter
given curve at p
:
cusp-edge
is
touched by the rectifying line \ to the original curve
(pp. 584-587).
(^).
Equation with one arbitrary constant (p. 587), of a cone of

which has its vertex at the given point p, and has
the second order,
with the cone of chords

(397) from that point; equation (p. 590) of a cylinder of the second
order, which has an arbitrary line pe from p as one side, and has
contact of the third order (or four-side contact^
contact of the fourth order {ex five-point contact) with the curve at p ;
the constant above mentioned can be so determined, that the right line
PE shall be a side of the cone also, and therefore apart of the intersect
Hon of cone and cylinder; and then the remaining or curvilinear
part, of the complete intersection of those two surfaces of the second
* The
rectifying plane, of the cusp-edge of the rectifying developable, is the
r', of which the formula LIV. in p. 587 is the equation ; and the
rectifying line rh, of the same cusp-edge, intersects the absolute normal pk to the
plane of X and
given curve, or the radius (r) of first curvature, in the point h in which that
radius is nearest (e) to a consecutive radius of the same kind. But this last theorem, which is here deduced by quaternions, had been previously amved at by
M. de Saint- Venant (comp. the Note to p. xv.), through an entirely different
analysis, confirmed by geometrical considerations.
CONTENTS.
order, is
or -what
XIX
Pages,
known
(by
is briefly
principles) a gawhe curve of the third order,

called* a Twisted Cubic : and this last curve, in
its construction above described, and whatever the asdirection of the auxiliary line pb may be, has contact of the
fourth order (ôr five-point contact) with the given curve of double cur-
virtue of
sumed
vature at p (pp. 587-590, comp. pp. 663, 572).

(Ji). Determination (p. 590) of the constant in. the equation of the
eone {g), so that this cone may have contact of the fourth order (or
with the cone of chords from p the cone thus found

be called the Osculating Oblique Cone (comp. 397), of the second
and the coefficients of its equation inorder, to that cone of chords
five-side contact)
may
volve only r, r, /, r', r', r", but not r'", although this last derivative
is of no higher order than r", since each depends only on s^ (and lower
powers), or introduces only ffth differentials. Again, the cylinder
(J) will have contact of ^q fifth order (or six-point contact) with the
given curve at
p, if the line pe,
which
is
by construction a side of that
cylinder, and has hitherto had an arbitrary direction, be now obliged

to be a side of a certain cubic cone, of which the equation (p. 590) in-
volves as constants not only rrr'rV'r", like that of the osculating cone
just determined, but also r"'. The two cones last mentioned have the
tangent (r) to the given curve for a common side,f but they have also
three other
common
sides,
whereof one
at least is real, since
they are
assigned by a cubic equation (same p. 590) and by taking this side

for the line pe in (y), there results a new cylinder of the second order,
;
which cuts the oscidating oblique cone, partly in that right line pe itself,
and partly in a gauche curve of the third order, which it is proposed to
because
has con-
call
an Osculating Twisted Cubic (comp. again
tact
of the fifth order (or six-point contact) with the given curve at p
{g)),
it
(pp. 590, 591).

(i).
In
sented
and independently of any question of osculation,

passing through the origin o, may be repre-
general,
a Twisted Cubic
(J), if
by any one of the
vector equations (pp. 592, 593),
*
By Dr. Salmon, in his excellent Treatise on Analytic Geometry of Three
Dimensions (Dublin, 1862), which is several times cited in the Notes to this final
Chapter (III. iii.) of these Elements. The gauche curves, above mentioned, have
been studied with much success, of late years, by M. Chasles, Sig. Cremona, and
other geometers but their existence, and some of their leading properties, ap:
pear to have been
first
perceived and published
centric Calculus, Leipzig, 1827, pp.
t This
side,
however, counts as three
intersection (real or
by
Prof.
114-122, especially
imaginary) of these
(p. 614),
tioo cones,
Mobius
(see his
ary-
p. 117).
in the system of the six lines of

which have a common vertex p,
and are respectively of the second dcndi third orders (or degrees). Additional light
in which also
will be thrown on this whole subject, in the following Series (399)
;
shown that
only one osculating twisted cubic, at a given point,

to a given curve of double curvature
and that this cubic curve can be determined,
without resolving any cubic or other equation.
it
will be
there
is
XX
CONTENTS.
+ c)p = a,
Vap + pVy/o + VpVXp/t = 0,
Yap+Ypfp = 0,
or
(Y);
(^
Pages.
(Y')
or
(Y'")
p = (^4-c)-ia, (Y");
in which a, y, \, /* are real and constant vectors, but c is a variable scalar; -vrhile <pp denotes (comp. the Section III. ii. 6, or pp. xii., xiii.) a
or
and vector function, which is Iiere generally not self-conjug ate,

of the variable vector p of the cuMc curve. The number of the scalar
the equation, is
constants, in the form (Y'"), or in any other form of
linear
ten (p. 593), with the foregoing supposition that the curve
it is easy to remove.
passes through the origin, a restriction which
The curve (Y) is cut, as it ought to be, in three points (real or imagi-
found to be
by an arbitrary secant plane ; and its three asymptotes (real or

imaginary) have the directions of the three vector roots (3 (see again
the last cited Section) of the equation (same p. 593),
nary),
y/3^/3
= 0:
(Z)
(P), p. xii., these three asymptotes compose a real and rectangular system, for the case of self-conjugation of the function tp
80 that
by
in (Y).
(y). Deviation of a near point p* of the given curve, from the sphere
(395) which osculates at the given point p this deviation (by p. 593,
;
comp. pp. 553, 584)
is
ns^
R's^
ris^
ultimately equal (p. 595) to the quarter of the deviation (397)

of the same near point Ps from the osculating circle at p, multiplied by
the sine of the small angle spSs, which the small arc sss of the locus of
it is
the spheric centre s (or of the cusp-edge of the polar developable) subtends at the same point p ; and it has an outward or an inward direction, according as this last arc is concave or convex (/) at s, towards the
given curve at p (pp. 585, 595).
It is also ultimately equal (p. 596)
- PsSs,
of the given point p from the near sphere^

which osculates at the near point Ps ; and likewise (p. 597) to the com-
to the deviation ps^
ponent, in the direction of sp, of the deviation of that near point from
the osculating circle at p, measured in a direction parallel to the nor-
mal plane
now expressed to the

has hitherto been considered
at that point, if this last deviation be
accuracy of the /owrÂ order
whereas
it
sufficient to develope this deviation from the osculating circle (397) as

far as the third order (or third dimension of 5) ; and therefore to treat
it as having a direction, tangential to the
osculating sphere (comp.
pp. 566, 594).

Qc). The deviation (Ai) is also equal to the third part (p. 598) of
the deviation of the near point p^ from the given circle (which osculates
at p), if measured in the near normal plane (at p*), and decomposed in
the direction of the radius Ss of the near sphere ; or to the thirdpart

(with direction preserved) of the deviation of the miv near point in
which the given
circle is cut
by the near plane, from the near sphere : or

and still with an unchanged direc-
finally to the third part (as before,
XXI
CONTENTS.
Pages,
from the given sphere, of that other near point

in which the near circle (osculating at Pj) is cut by the given normal
plane (at p), and which is found to satisfy the equation,
tion) of the deviation
c,
sc
3sps
2ip.
(Bi)
Geometrical connexions (p. 599) between these various results {j)

illustrated by a diagram (Fig- 83).
(Jc),
Circle to
(t). The Surface, which is the Locus of the Osculating
a given curve in space, may be represented rigorously by the vector
expression (p. 600),

ws, M
= ps + rsTs sin w + n^ r^' vers u\
(C i)
and u are two independent scalar variables, whereof s is

(as before) the arc pp^ of the given curve, but is not now treated as
small : and u is the (small or large) angle subtended at the centre Ks of
the circle, by the arc of that circle, measured from its point of osculain
which
But the same
tion Vs.
sented also
by the
superficial locus
(comp. 392)
may be
repre-
vector equation (p. 611), involving apparently only
one scalar variable (s),
V-i^+v = 0,
in
which
= r^r/, and
V4
of the surface.
3,
W-ps
w=a>s,M =
(DO
the vector of an arbitrary point

(p. 601), of the Section III. iii.
The general method
shows that the normal
to this surface (Ci), at
thereof, has the direction of
any proposed point

Ws,m <^; that is (p. 600), the direction
which contains the circle through that
of the radius of the sphere,

and has the same point of osculation Ps to the given curve. The
locus of the osculating circle is therefore found, by this little calculation
point,
with quaternions, to be at the same time the Envelope of the Osculating Sphere, as was to be expected from geometrical considerations
(comp. the Note to
p. 600).
curvilinear locus of the point c in (Jc) is one branch of
the section of the surface (J), made by the normal plane to the given
(w).
The
curve at p
new
curve,
dius PS or
and if d be the projection of c on the tangent at p to this

which tangent pd has a ^vcQcXxoTS-pterpendicidar to the ra-
of the osculating sphere at p (see again Fig. 83, in p.

599), while the ordinate DC i^ parallel to that radius, then (attending
only to principal terms, pp. 598, 599) we have the expressions,
ns^
jR*-^
and therefore ultimately
from which
it
(p. 600),
follows that p is a singular point of the section here
considered, but not a cusp of that section, although the curvature

at p is infinite
: the ordinate dc varying

ultimately as tbe power
with exponent f of the abscissa pd.
Contrast (pp. 600, 601), of this
XXU
CONTENTS.
Pages.
with that of the developable Locus of Tangents, made by the

same normal plane at p to the given curve the vectors analogous to
PD and DC are in this case nearly equal to - ^s^t' and \s^x~^v', so
section,
that the latter varies ultimately as the power -| of the former, and the
point p is (as it is known to be) a cusp of this last section.
(n). A given Curve of double curvature is therefore generally a
Singular Line (j^. 601), although, not a cusp-edge, vi]ôn that Surface (f),
which is at once the Locus of its osculating Circle, and the Envelope
of its osculating Sphere : and the new developable surface (d), as being
circumscribed to this superficial locus (or envelope'), so as to touch it
along this singular line (p. 612), may naturally be called, as above,
the Circumscribed Developable (p. 581).

(o). Additional light may be thrown on this whole theory of the
singular line (w), by considering (pp. 601-611) a problem which was
two distinct Sections (xxii. xxvi.) of his wellthe Notes to pp. 602, 603, 609, 610 of these
Elements) ; namely, to determine the envelope of a sphere with varying
radius E, whereof the centre s traverses a given curve in space ; or
discussed
by Monge,
in
known Analyse (comp.
Envelope of a Sphere with One varying Parameter
briefly, to find the
especially for the Case of Coincidence (p. 603, &c.), of

are usually two distinct branches (p. 602) of a certain Charac-
(comp.
what
p.
624)
Curve (or arete de rebroussement), namely the curvilinear enveimaginary) of all the circles, along which the superficial
teristic
lope (real or
envelope of the spheres
touched
is
by those spheres themselves.
Quaternion forms (pp. 603, 604) of the condition of coincidence (o) ; one of these can be at once translated into Monge's equa(j)).
tion of condition (p. 603), or into an equation

slightly more general,
as leaving the independent variable arbitrary ; but a simpler and
more
easily interpretable
form
is
the following (p. 604),
ndr = + JSdi2,
in
which r
envelope
which
is
(o),
the radius of the
while
^'i
is
(Gi)
of contact, of a sphere with its

the radius of (first) curvature of the curve (s),
circle
the locus of the centre s of the sphere.

{q). The singular line into which the two branches of the curviis
linear envelope oxeftised,
an orthogonal
when this condition is
trajectory (p.
that curve,
satisfied, is in general
607) to the osculating planes of the curve
which
is noiv the given one, is therefore

(comp. 391,
607) oith& polar developable, corresponding to
the singular line just mentioned, or to what may be called the curve
In this way there arise
(p), which was formerly the given curve.
(s)
395) the cusp-edge
many
(p.
verifications of formulae
shown
(pp. 607,
608)
for example,
the
with the results of (/).

(r). With the geometrical hints thus gained from interpretation
of quaternion results, there is now no
difficulty in assigning the Complete and General Lntcgral of the Equation of Condition {p), which was
equation (Gi)
presented
is easily
to be consistent
by Monge under the form (comp.
differential equation
of the second order,
p. 603) of a non-littear
involving three variables
CONTENTS.
XXlll
Pages.
considered as functions of a fourth (a), namely the co-ordinates of the centre of the sphere^ regarded as varying with the ra(^,
tt)
ip,
but which does not appear to have been either integrated or
dius,
interpreted
by that
found presents
The general
illustrious analyst.
integral here
iisQlisit^vî'vaQ, quaternion form (p. 609),
translated (^. 610) into the usual
but
is
language of analysis.
easily
less
ge-
neral integral is also assigned, and its geometrical signification exhibited, as answering to a case for which the singular line lately consi-
dered reduces itself to a singular point (pp. 610, 611).

(s). Among the verifications {q) of this whole theory,
(pp. 608, 609) that although,
it is
shown
when the two branches (p)
of the general
curvilinear envelope of the circles of the system are real and distinct^
each branch is a cusp-edge (or arete de rebroussement, as Monge perit to be), upon the superficial envelope of the spheres, yet in the
of fusion (p) this cuspidal character is lost (as was likewise
seen by Monge*) and that then a section of the surface, made by
ceived
case
a normal plane to the singular

pressed by the equation (Fi).
confirmed,
by
calculation and
has precisely the for-m (m), ex-
line,
In
short, the result is in
by geometry,
coincidence (^) is satisfied, the Surface
is,
that
many ways
when the condition of
as in (w), at once the Enve-
of the osculating Sphere and the Locus of the osculating Circle, to

that Singular Line on itself, into which by (g-) the tivo branches (o)
lope
of
general cusp-edge are fused.

Other applications of preceding formulae might be given ;
for instance, the formula for k" enables us to assign general exits
(f).
pressions (p. 611) for the centre and radius of the circle, which osculates at K to the locus of the centre of the osculating
circle, to a given
curve in space
with an elementary
verification, for the case of the
plane evolute of the plane evolute of a plane curve. But it is time to conclude this long analysis, which however could scarcely have been
much abridged, of the results of Series 398, and to pass to a more
brief account of the investigations in the following Series.
Article 399. Additional general investigations, respecting that

gauche curve of the third order (or degree), which has been above
called an Osculating Twisted Cubic (398, (A)), to any proposed curve
of double ciirvature
curve
is
().
a,
with applications to the
case,
where the given

612-621
helix,
In general
(p. 614), the tangent
pt
to the given curve is a
nodal side of the cubic cone 398, (K) one tangent plane to that cone
(C3), along that side, being the osculating plane (P) to the curve, and
therefore touching also, along the same side, the osculating oblique cone
;
(C-i)
of the second order, to the cone of chords (397) from p
while the
other tangent plane to the cubic cone (C3) crosses ihoi first plane
{P),
or the quadric cone (Cg), at an angle of which the trigonometric cotan-
Compare the
first
Note
to p.
609 of these Elements.
CONTENTS.
XXIV
gent (^r')
is
equal to half the differential of the radius (r) of second
And the three

curvature, divided by the differential of the arc (js).
common sides, pe, pe', pe", of these tivo cones, "which remain when the
tangent pt
is excluded,
and of which one at least must be
real, are
the
through the given point p to the three asymptotes (398, (f))

to the gauche curve sought being also sides of three quadric cylin-
parallels
{L"^, which contain those asymptotes as other

and of which each contains the twisted
sides (or generating lines)
cubic sought, and is cut in it by the quadric cone (C2).
Method to the case of a given helix, it
(b). On applying this First
ders,
say (Z2),
(-Z^'2),
is found (p. 614) that the general cubic cone (C3) breaks up into the
system of a new quadric cone, (C2'), and a new plane (P') which latter is the rectifying plane (396) of the helix, or the tangent plane at p
;
whereon that given curve is traced. The two

quadric cones, (Co) ^^d (C2')) touch each other and fhe plane (P) along
the tangent pt, and have no other real common side : whence two of
to the right cylinder,
the sought asymptotes, and two of the corresponding cylinders (a), are
in this case imaginary, although they can still be used in calculation
But the plane (P') cuts the cone (C2), not only
(pp. 614, 615, 617).
in the tangent pt, but also in a second real side pe, to which the real
is parallel (a) ; and which is at the same time a side of a
real quadric cylinder (ia), which has that asymptote for anotlier side
(p. 617), and contains the twisted cubic : this gauche curve being thus
asymptote
the curvilinear part (p. 615) of the intersection of the real cone (C2),
with the real cylinder (2)of thi<; result ; fractional ex(c). Transformations and verifications
pressions (p. 616), for the co-ordinates of the twisted cubic ; expression (p. 615) for the deviation of the Jielixfrom that osculating curve,
which deviation is directed inwards, and is of the sixth order : the

between the tangmt pt and the real asymptote, is a right
line PB, which is cut internally (p. 617) by the axis of the right cylinder (b), in a point a such that pa is to ab as three to seven.
(d). The First Method (a), which had been established in the pre-
least distance,
ceding Series (398), succeeds then for the case of the helix, with a faciwhich arises chiefly from the circumstance (i), that for this case
lity
the general cubic cone (C3) breaks up into two separate loci, whereof
But usually the foregoing method requires, as in
one is a plane (P').
398, (fi)), the solution of a cubic equation : an inconvenience which is
completely avoided, by the employment of a Second General Method,

as follows.
(e). This Second Method consists in taking, for a second locus of the
gauche osculatrix sought, a certain Cubic Surface (^83), of which
every point is the vertex* of a quadric cone, having six-point con-
* It
through
is
known
that the locus of the vertex of a quadric cone,
six given iwints of space, a, b, c, d, e,

f,
which passes
whereof no four are in one
XXV
CONTENTS.
Pages.
tact with the given curve at p
so that this tiew surface is cut hy the
plane at injinity, in the same cubic curve as the cubic cone (C3). It is
found (p. 620) to he a Muled Surface^ with the tangent pt for b. Sin:
gular Line ; and when this right line is set aside^ the remaining (that
is, the curvilinear') part of the intersection of the two loci, (62) and
Twisted Cubic sought
{Si), is the Osculating
thus completely
which gauche osculatrix
generally determined, without any such difficulty

or apparent variety, as might be supposed to attend the solution of a
is
dindi
cubic equation (d),

(p. 621).
Article 400.
().
and with new
verifications for the case of the helix
On Involutes and Evolutes in Space, ....
621626
The usual points of Monge's theory are deduced from the two
fundamental quaternion equations
S((T-p)p' = 0,
(p. 621),
V(<T-p)<r'=0,
(Hi)
which p and <t are corresponding vectors of involute and evolute

together with a theorem of Prof. De Morgan (p. 622), respecting the
in
case
when
(i).
the invohite
An
a spherical curve.
is
involute in space is generally the only real part (p. 624) of
the envelope of a certain variable spJiere (comp. 398), which has its
centre on the evolute, while its radius
is the variable intercept between the two curves but because we have here the relation (p. 622,
comp.
p. 602),
i2'2
+ t'2 =
0,
(Hi')
the circles of contact (398, (0)) reduce themselves each to o, point (or
rather to a pair of imaginary right lines, intersecting in a real point),
and the preceding theory (398), of envelopes of spheres with one
varying parameter, undergoes important modifications in
its results,
the conditions of the application being difierent. In particular, the

involute is indeed, as the equations (Hi) express, an orthogonal trajectory to the tangents of the evolute ; but not to the osculating planes
plane, is generally a Surface, say (S^), of the Fourth Degree : in fact, it is cut by
the plane of the triangle abc in a system of four right lines, whereof three are
the sides of that triangle, and the fourth is the intersection of the two planes,
ABC and DEF. If then we investigate the intersection of this surface (^S^) with
the quadric cone, (a.bcdef), or say {O2), which has a for vertex, and passes
through the five other given points, we might expect to find (in some sense) a
curve of the eighth degree. But when we set aside iihejive right lines, ab, ac, ad,
AE, AF, which are common to the two surfaces here considered, we find that the
(remaining or) curvilinear part of the complete intersection is reduced to a curve

of the third degree, which is precisely the twisted cubic through the six given points.
In applying this general (and perhaps new) method, to the problem of the osculating twisted cubic to a curve, the oaculating plane to that curve
may be excluded,
and then the quartic surface (/S4) is reduced to the

as foreign to the question
cubic surface (S^, above described.
:
CONTENTS.
XXvi
Pages.
of that curve, as the singular line (398, (^)) of the former envelope
was, to those of the curve which was the locus of the centres of the
spheres hefore considered,
when
a certain condition of coincidence (or
oi fusion, 398, {pj) was satisfied.

and
evolute (p. 625) ; if p, Pi, P2,
(c). Curvature of hodograph of
of involute and evolute, and if we
s, Si, 82, . . be corresponding points
draw right
lines sti, st2,
a common length
sp,
of the second order at
Article 401.
in the directions of SiPi, S2P2,
the spherical curve ptiT2
p, tvith the involute
fpiP3
and with
will have contact
(pp. 625, 626).
Calculations abridged, by the treatment of quater-
nion differentials (which have hitherto been finite, comp. p. xi.) as

new deductions of osculating plane, circle, and sphere,
infinitesimals ;*
with the
vector equation (392) of the circle;
cond curvatures o{
a.
On
Section 7
and of the
first
and
^q-
626-630
curve in space,
Surfaces of the Second Order; and on
630-706
Curvatures of Surfaces,
Article 402. References
to some equations of Surfaces, in earlier

parts of the Volume,
Article 403. Quaternion equations of the Sphere (p2 = - 1, &c.),
for norm is employed
In some of these equations, the notation
630,631
631-633
(comp. the Section II. i. 6).

Article 404. Quaternion equations of the Ellipsoid,
One of the simplest of these forms is (pp. 307, 635) the equation,
633-635
....
T(tp + p,c)=K2_,2,
(Ij)
*
Although, for the sake of brevity, and even of clearness, some phrases have
been used in the foregoing analysis of the Series 398 and 399, such 0.% four-side
or five-side contact between cones, and five-point or six-point contact between
curves, or
between a curve and a
surface,
which are borrowed from the doctrine
of consecutive points and lines, and therefore from that of infinitesimals ; with a
few other expressions of modem geometry, such as the plane at infinity, &c.
;
yet the reasonings in the text of these Elements have all been rigorously reduced,
BO far, or are all obviously reducible, to the fundamental conception of Limits :
compare the
definitions of the osculating circle
389, 395. The object of Art. 401

such ultimate reference to limits,
is
to
it is
make
it
and
sphere, assigned in Articles
visible
how, without abandoning
possible to abridge calculation, in several
by treating (at this stage) the differential symbols, dp, d^p, &c., as if
without taking the
they represented infinitely small differences, Ap, A^p, &c.
trouble to write these latter symbols frst, as denoting fnite differences, in the
cases,
rigorous statement of a problem, of which statement it is not always easy to assign

the proper form, for the case of points, &c., Q.t finite distances : and then having
the additional trouble oireducing\hQ complex expressions so found to simplerformsy
in which differentials shall finally appear. In short, it is shown that in Quaternions, as in other parts of Analysis, the rigour of limits
ih.e facility
of infinitesimals.
can be combined with
xxvu
CONTENTS.
Pages,
and k are
in wliich
the
normals.
ci/clic
and constant vectors, in the directions of

This fonn (Ii) is intimately connected with, and
real
indeed served to suggest, that Construction of the Ellipsoid (II. i. 13),

by means of a Biacentric Sphere and a Point (p. 227, comp. Fig. 53,
of the Quap. 226), which was among the earliest geometrical results
The
ternions.
terms of
t,
three scmiaxes, ,
k as follows
c,
*,
are expressed (comp. p. 230) in
= Tt + TK:;
b=^~-T - k)
= Tt-T;
(Ii')
(t
whence
ai"'c
Article 405.
central quadric),
Article 406.
Sp0p=/p = O,
General
= T(i-
(Ii")
k).
Central Surface of the Second Order (or
Sp<pp=fp =
l,
General Cone of the Second Order (or quadric cone),
Article 407.
Bifocal Form of the equation of a central but non-
conical surface of the second order
is
The
bifocal
638-643
with some quaternion formulse,
643-653
relsLting to Confocal Surfaces,

(a).
636-638
form here adopted (comp. the Section
III.
ii.
6)
the equation,
Cfp
in which,
C,
(Jj)
(Ji')
(real) focal unit-lines, common to the whole system of

the (real and positive) scalar I is also constant for that sysbut the scalar e varies, in passing from surface to surface, and
a, a' are
confocals
tem
= (Sp)2 - 2eSapSa'p + (Sa'p)2 + (1 - e^) p2 =

C= (e - 1) (e + Saa') l^.
two
;
be regarded as a parameter, of which the value serves to distinguish one confocal, say (e), from another (pp. 643, 644).
scalar semiaxes (real or ima(J)). The squares (p. 644) of the three
may
ginary), arranged in algebraically descending order, are,
a^^{e^\)P,
whence
and the three
bi^{e+âa')P,
^'=-T~'
c^
= {e-\)P;
(Ki)
^^'^
^"^^^T^'
vector semiaxes corresponding are,
aU(a+a'),
bVYaa',
cU(a-a').
(Mi)
Rectangular, unifocal, and cyclic forms (pp. 644, 648, 650),

of the scalar function fp, to each of which corresponds a form of the
(c).
vector function
<pp
deduction,
by a new
analysis, of several
known
theorems* (pp. 644, 645, 648, 652, 653) respecting confocal surfaces,
* For
example, it is proved by quaternions (pp. 652, 653), that the focal
the focal cone, which has any proposed point p for vertex, and rests on
lines of
the focal hyperbola, are generating lings of the single-sheeted hyperboloid

(of the
given confocal system), which passes through that point and an extension of
this result, to the focal lines of any cone circumscribed to a confocal, is deduced
:
by a similar
But such
analysis, in a subsequent Series (408, p. 656).
theorems respecting confocals can only be alluded to, in these Contents.
known
CONTENTS.
XXVIU
Pages,
and their focal
conies ; the lines a, a' are asymptotes to the focal
hy-
refeperbola (p. 647), whatever the species of the surface may be

rences (in Notes to pp. 648, 649) to the Lectures* for ihe focal ellipse
of the Ellipsoid^ and for several different generations of this last sur:
face.
of the equation
{cT). General Exponential 7}ransformation (p. 651),
of any central quadric ;
= xa + yYa%
,
this auxiliary vector

nent,
(3 is
(Ni), with x^^fa
+ 2/2/UVaa' =
1,
(a-ea)JJYaa
(Ni')
,n.T .v
any one confocal (e) the expoand the coefficients,

however connected
constant, for
in (Ni), is an arbitrary or variable scalar;

y, are two other scalar variables, which are
t,
X and
with each other by the relation
(Ni').
If SiUj fixed value be assigned to t, the equation (Ni) then represents the section made by a plane through a (p. 651), which sec(i).
tion
is
an
ellipse if
the surface be an
ellipsoid,
but an hyperbola for
and the cutting plane makes with the focal plane

of a, a', or with the plane of the focal hyperbola, an angle = it-rr.
(/). If, on the other hand, we allow t to vary, but assign to
X and y any constant values consistent with (Ni'), the equation (Ni)
then represents an ellipse (p. 65 1 ), whatever the species of the surface
either hyperboloid ;
may
be ; x represents the distance of
its
centre
from the
centre o of the
measured along the focal line a; y is the radius of a right

cylinder, with a for its axis, of which the ellipse is a section, or the
surface,
radius of a circle in a plane perpendicular to a, into which that ellipse

can be orthogonally projected : and the angle ^tn is now the excentrie
anomaly. Such elliptic sections of a central quadric may be otherwise

obtained from the unifocal form (c) of the equation of the surface ;
they
are, in
some points of view, almost as interesting as the known

: and it is
proposed (p. 649) to call them Centro-
circular sections
Foeal Ellipses.
And
it is obvious that, by interchanging the two focal lines

a Second Exponential Transformation is obtained, with a
Second System of centro-focal ellipses, whereof the proposed surface is
the locus, as well as of the first system (/), but which have their
(g).
a, a' in
centres
((?),
on the line
a',
and are projected into
circles,
on a plane per-
pendiciilar to this latter line (p. 649).

Qi).
Equation of Confocals
Article 408.
On
(p. 652),
Yv,(pv^
= Nv(l>,v.
(Oi)
Circumscribed Quadric Cones; and on the
Umbilics of a central quadric,
653- G63
* Lectures on
Quaternions (by the present author), Dublin, Hodges and
Smith. 1863.
XXIX
CONTENTS.
Pages.
Equations (p. 653) of Conjuffate Points, and of Conjugate Diwith respect to the surface /p = 1,
(a).
rectians,
=
(P:),and/(p,p') 0;
of the same surface with the riffht
=
/(p,p')
Condition of Contact,
(Pi')
l,
line pp',
(QO
(/(p,p')-i)^=(/p-i)(/p'-i);
form of the equation of the Cone, with vertex at

circumscribed to the same quadric (/p = 1).
this latter is also a

p',
which
is
The condition (Qi) may
(i).
he thus transformed
also
(p. 654),
FYpp' = a^^c^fip p),
F being
(Qi')
a scalar function, connected with

by certain relations of
reciprocity (comp. p. 483) ; and a simple geometrical interpretation
may
be assigned, for this
last equation.
Reciprocal Cone, or Com of Normals
The
(c).
scribed cone (Qi) or
(Q,i')>
may
<t
be represented
at p'.to the circum-
(p.
655) by the very
simple equation,
(Qi")
i?'((T:Sp'(T)=l;
which likewise admits of an extremely simple interpretation.
(<^).
other
given right line
known
(p.
656)
is
touched
by two
confocals,
and
results are easy consequences of the present analysis
example (pp. 658, 659), the cone circumscribed to any sru-face of

the system, from any point of either of the two real focal curves^ is a
cone of revolution (real or imaginary)
but a similar conclusion holds
for
when
focal, and even

more generally (p. 663), when that vertex is any point of the (known
and imaginary) developable envelope of the confocal system..
good,
(e).
the vertex
is
on the third (or imaginary)
central quadric has in general Twelve Umbilics (p. 659),
whereof ovlj four (at most) can be real, and which are its intersections
with the three focal curves : and these twelve points are ranged, three by
three,
on
eight imaginary right lines (p. 662),
at infinity,
and which
it is
proposd to
which
call the
intersect the circle
Eight Umbilicar Ge-
neratrices of the surface.
(/). These (imaginary) umbilicar generatrices of a quadric are
found to possess several interesting properties, especially in relation

to the lines of curvature : and their locus, for a confocal system, is a
developable surface (p. 663),
system.
Article 409.
namely the known
envelope (d) of that
Geodetic Lines on Central Surfaces of the Second

664-667
Order,
One form
of the general differential equation of geodetics

an arbitrary surface being, by III. iii. 5 (p. 515),
(a).
this is
Vvd2p = 0,
shown (p. 664)
integral,
yrhexQ
(Ri),
if
Tdp =
const,
on
(Rj')
to conduct, for central quadrics, to the first
p2i)-2 = Tv2/Udp
P \& ihe perpendicular from
= -^ = const;
the centre
on.
(Si)
ihQ tangent plane,
XXX
Pages,
B is the
and
is
CONTENTS.
parallel
imaginary) semidiameter of the surface, which
(real or
The known equation

anew; this last
the tangent (dp) to the curve.
to
of Joachimstal, P.i)
const,
is
therefore proved
constant, however, being by no means necessarily

be not an ellijJsoid.
Deduction
(b).
(p.
real, if
665) of a theorem of M. Chasles, that the tan-
gents to a geodetic, on
confocal (^,)
any one central quadric [e), touch

and of an integral (p. 666) of the form,
ei sin^ vi
e^
cos^ vi
iii.
Svdpd2p =
(p. 666),
= d. P-2i)-2 = 2Sj/dvdp-i. Svdp-d2p
d/i
by the lately cited Section (III.

equation of the second order,
but,
(Si')
differential of the scalar
be thus decomposed into factors
may
also a conifnon
= e, = const.,
which agrees with one of M. Liouville.

(c). Without the restriction (Hi')? the
in (Si)
the surface
5, p.
(Si")
515), the differential
(Ri")
0,
with an arbitrary scalar variable, represents the geodetic lines on any

surface : the theorem (a) is therefore in this way reproduced.
But we
(d).
or P. J)
but
= h'^,
same time, by (Si"), that the quantity h,

not only for the geodetics on a central quadric,
see, at the
is constant,
a certain other
also for
set
of curves, determined
by the
differen-
equation of ihe first order, Svdi/dp = 0, which will be seen, in the

next Series, to represent the lines of curvature.
tial
Article 410.
lar
on such
(a).
On Lines of Curvature generally;
lines, for the case of a
The
and in particu667-674
Central Quadric,
differential equation (comp. 409, (<?)),
Srdvdp = 0,
(TO
Lines of Curvature, upon an arbitrary surface ;

represents (p. 667)
because it is a limiting form of this other equation,
i\ie
SvAvAp =
0,
(Ti')
which is the condition of intersection (or of parallelism), of the normals

drawn at the extremities of the two vectors p and p + Ap.
(5). The normal vector r, in the equation (Ti), may be multiplied
(pp. 673, 700) by any constant or variable scalar n, without any real
change in that equation but in this whole theory, of the treatment
;
of Curvatures of Surfaces by Quaternions, it is advantageous to consider the expression Svdp as denoting the exact differential of some
scalar function of p
for then (by pp. 486, 487)
we shall have an equa-
tion of the form,
dr = ^dp = a
which usually involves p

(p. 669,
comp. (R),
self-conjugate function of dp,
also.
For
instance,
we may write
(Ui)
generally
p. xiii),
dv=^dp+V\dp;i;
(Ui')
xxxi
CONTENTS.
Pages,
the scalar ^, and the vectors X,
[i
being
and being generally* func-
real,
tions of p, but not inyolving dp.
This being understood, the two\ directions of the tangent dp,

((?).
which satisfy at once the general equation (Ti) of the lines of curva= of the surface, are easilyture, and the differential equation Si^dp
found to be represented by the two vector expressions (p. 669),
UYvX + UVv/i;
(Ti")
they are therefore generally rectangular to each other, as they have

long been
known
to be.
The
surface itself remaining still quite arbitrary, it is found

useful to introduce the conception of an Auxiliary Surface of the Se{d).
cond Order
equation
670), of
(p.
is,
which the variable vector
Sp>p' = ^p'2 + SXp>p' =
more generally =
which the centre is
or
const.
and
it is
is
p',
(W)
1,
proposed
and the
to call this surface, of
Index Surface, partly

diametral section, made by the tangent plane to the given
surface at p, is a certain Index Curve (p. 668), which may be considered to coincide with the known " indicatrice" of
because
at the given point p, the
its
Dupin.
The
().
surface
may
expressions (Ti") show (p. 670), that whatever the given

be, the tangents to the lines of curvature bisect the angles
fonned by the traces of the two cyclic planes of the Index Surface ((f),
on the tangent plane to the given surface these two tangents have
;
was seen by Dupin) the directions of the axes of the Index

Curve (p. 668) and they are distinguished (as he likewise saw) from
also (as
tangents to the given surface, at the given point p, by the

condition that each is perpendicular to its oivn conjugate, with respect
to that indicating curve
the equation of such conjugation, of two
all other
tangents t and
r',
being in the present notation (see again

Sr^r' =
0,
or
Sr^r = 0.
p. 668),
(Ui'")
(/). New proof (p. 669) of another theorem of Dupin, namely

that if a developable be circumscribed to any surface, along any curve
its generating lines are everywhere conjugate, as
tangents to
the surface, to the corresponding tangents to the curve.
(J). Case of a central quadric ; new proof (p. 671) of still another
thereon,
theorem of Dupin, namely that the curve of orthogonal intersection

(p. 645), of two confocal surfaces, is a line of curvature on each.
(A). The system of the eight mnbilicar generatrices (408, (e)), of a
central quadric, is the imaginary envelope of the lines of curvature on
that surface (p. 671)
and each such generatrix
is itself sea.
imaginary
For the case of a central quadric, g, X, /i are constants.

It will soon be
; but in some cases more.
of curvature pass through an umbilic of a quadric.
t Generally two
lines
seen, that three
XXXll
CONTENTS.
Page
of curvature thereon
so that through each of the twelve umbilics

(see again 408, (e)) there pass three lines of curvature (comp. p. 677),
line
whereof however only one, at most, can be real namely two generatrices, and a principal section of the surface. These last results, which
:
are perhaps new, will be illustrated, and otherwise proved, in the

following Series (411).
Article 411. Additional illustrations and confirmations of the
foregoing theory, for the case of a Central* Quadric ; and especially

of the theorem respecting the Three Lines of Curvature through an
TJmhilic,
whereof two are always imaginary and
rectilinear,
....
674 6<
The general equation of condition (Ti'), or Sj/AvA/o = 0, for

the intersection of two finitely distant normals, may be easily transformed for the case of a quadric, so as to express (p. 675), that when
{a).
the normals at p
(i).
and
p' intersect (or are parallel), the
own polar.
Under the same conditions,
pendicular
clwrd pp'
is
per-
to its
if
the point p be given, the locus
usually (p. 676) a quadric cone, say (C) ; and therefore the locus of the point p' is usually a quartic curve, with p for a
double point, whereat two branches of the curve cut each other at right
of the chord pp'
is
and touch the two lines of curvature.

If the point p be one oid, principal section of the given surface,
but not an umbilic, the cone (C) breaks up into a, pair ofplanes, whereof
angles,
(c).
one, say (P), is the plane of the section,
dicular thereto, and
is
and the
other, (P')> i^
not tangential to the surface
perpen-
and thus the
quartic (b) breaks up into a pair of conies through p, whereof one is

the principal section itself, and the other is perpendicular to it.
p be an umbilic, the second plane (P')

(^d). But if the given point
becomes a tangent plane to the surface and the second conic (c) breaks
up, at the same time, into o. pair of imaginary^ right lines, namely
the two umbilicar generatrices through p (pp. 676, 678, 679).
;
(e).
soid, or
It follows that the
normal pn at a real umbilic p (of an
ellip-
not intersected by any other

are in the same principal section ; but
a double-sheeted hyperboloid)
is
which
pn is intersected, in an imaginary sense, by all
which are drawn at points p' of either of the two ima-
real normal, except those
that this real normal

the normals p'n',
ginary generatrices through the real umbilic p
so that each of these
*
Many, indeed most, of the results apply, without modification, to the case of
the Paraboloids ; and the rest can easily be adapted to this latter case, by the consideration of infinitely distant points.
omit the term
central,
We shall therefore often, for conciseness,
and simply speak of
quadrics, or surfaces of the second
order.
t It is well known that the single- sheeted hyperboloid, which (alone of

central quadrics) has real generating lines, has at the same time no real umbilics
(comp. pp. 661, 662).
CONTENTS.
XXxiii
Pages.
imaginary right
lines is
seen
anew to be a line* ofcurmture^ on the sur-
face (comp. 410, (A)), because all the normals p'n', at points of this
one common (JmaginariJ) normal plane (p. 676)

and as before, there are thus three lines of curvature through an umline, are situated in
bilie.
(/). These geometrical results are in various ways deducible from

calculation with quaternions ; for example, a form of the equation of
the lines of curvature on a quadric is seen (p. 677) to become an
an umbilic (v X) while the differential of that equation

breaks up into ivfo factors, whereof one represents the tangent to the
principal section, while the other (SXd^p = 0) assigns the directions of
identity/ at
||
the two generatrices.

(g).
The equation of the
SapA(0
which has already presented
cone (C),
a certain locus of chords

transformations ; for instance (see
itself as
admits of
(b),
p. 675), it
may
many quaternion
be written thus,
Sa'pAp
p being the vector of the vertex p, and p + Ap that of any other point
cone ; while a, a' are still, as in 407, (a), two real focal lines,
p' of the
of which the lengths are here arbitrary, but of which the directions
are constant, as before, for a whole confocal system.
(h). This cone (C), or (Yi), is also the locus (p. 678) of a system
* It
might be natural to suppose, from the known general theory (410, (c))
tivo rectangular directions, that each such
generatrix pp' is crossed perpendicularly, at every one of its non-umbilicar points p', by a second (and distinct,
of the
although imaginary')
2iM^ received result of
lohen a right line
is
line
of curvature. But
it is
an almost equally well known
modem geometry, paradoxical
as
it
must
directed to the circle at infinity, as
ratrices in question are,
then
this
imaginary
at first appear, that
(by 408,
(e)) the
gene-
everywhere perpendicular to
Quaternions are not at all re-
line is
Compare the Notes to pages 459, 672.

sponsible for the introduction of this principle into geometry, but they recognise
a,nd employ it, under the following very simple form: that if a non-evanescent
itself.
vector be directed to the circle at infinity, it is an imaginary value of the symbol Oi

(comp. pp. 300, 459, 662, 671, 672) ; and conversely, that wli)z this last symbol
represents a vector which is not null, the vector thus denoted is an imaginary line,
which cuts that circle. It may be noted here, that such is the case with the reci-
procal polar of every chord of a quadric, connecting any two umbilics which are not
in one principal plane' ; and that thus the quadratic equation (XXI,, in p. 669)
from which the two directions (410, (c)) can usually be derived, becomes an idenas it ought to do, for consistency with
tity for every umbilic, real or imaginary
the foregoing theory of the three lines through that umbilic. And as an additional illustration of the coincidence of directions of the lines of curvature at any
:
non-umbilicar point p' of an umbilicar generatrix,
it
may
be added that the cone
of chords {C), in 411, (5), is found to touch the quadric along that generatrix^
when its vertex is at any such point p'.
XXXIV
CONTENTS.
Pages.
of three rectangular lines ; and if

to a side,
be cut by any plane perpendicular

and not passing through the vertex, the section is an equilait
teral hyperbola.
The same cone (C) has, for three of its sides pp', the normals
677) to the three confocals (p. 644) of a given system which pass
through its vertex p ; and therefore also, by 410, (y), the tangents
(i).
(p.
to the three lines of curvature through that point,
which are the inter-
sections of those three confocals.
(y).
I,
And
because
of 407, (a),
indefinitely
(b),
many
its
we
equation (Vi) does not involve the constant
arrive at the following theorem (p. 678)
common centre o, have

and pass through a common point p,
quadrics. with a
totic cones biconfocal,
their
their normals
at that point have a quadric cone {C)
Article 412
(). If
On
for their locus.

Centres of Curvature of Surfaces,
a be the vector of the centre
section of an arbitrary surface,
:If
asymp-
....
of curvature of a normal
which touches one of the two lines of

we have the two fundamental
curvature thereon, at any given point p,

equations (p. 679),
(T
=p+
i2Uv,
(Wi),
and
ij-^dp
+ dXJj/ =
(Wi')
whence
VdpdUv = 0,
(Wi"),
and
^ + S^ =
0;
(Wi'")
dp
the equation (Wi") being a new form of the general differential equation of the lines of curvature.
(J). Deduction (pp. 680, 681, &c.) of some known theorems from
these equations
and of some which introduce the new and general
conception of the Indeic Swfaee (410, (<?)), as well as^that of the
;
known Index
(c).
in
Curve.
Introducing the auxiliary iscalar (p. 682),
which r (|| dp) is a tangent to a line of curvature, while dv = 0dp,

two values of r, which answer to the two rectangular
as in (III), the
directions (Ti") in 410, (c), are given (p. 680)
= -g-TXiJi.
cos (Z
-^+
A
by the
expression,
(X'l)
).
/x
in which g, X, /* are, for any given point p, the constants in the equation (Ui") of the index surface; the difference of the two curvatures
JB'i
therefore vanishes at an umbilic of the given surface, whatever the
form of that surface may be that is, at a point, where v \ or /i,

and where consequently the index curve is a circle.
(d). At any other point p of the given surface, which is as yet en:
\\
tirely arbitrary, the values of r

*i
may be thus
= ar2,r2=a2-2,
||
expressed (p. 681),

(Xi")
being the scalar semiaxes (real or imaginary) of the index curve

= l, Srp' = 0).
(defined, comp. 410, (d), by the equations Sp'^p'
ai, a2
679-
XXXV
CONTENTS.
Pages.
quadratic equation, of wHcli ri and >'2, or the inverse
squares of the two last semiaxes, are the roots, maybe written (p. 683)
under the symbolical fo7-m,
The
{e).
Sv-i (^
+ r) -ij/ =
which may be developed (same page) into

j-a
+ rSv-ixJ' + Sv-i
(Yi)
this other form,
;|/j/
= 0,
(Y'l)
the linear and vector functions, \// and Xj being derived from the function 0, on the plan of the Section III. ii. 6 (pp. 440, 443).
(/). Hence, generally, the product of the two curvatures of a sur-
face
is
expressed (same p. 683) by the formula,

i2i-iiil2-i
= n r2 Tv -2 = - S
which will be found useful in the following
tP
(Zi)
series (418), in
connexion
with the theory of the Measure of Curvature.
The given
{g).
surface being
still
we
write
it is
easily
quite general, if
(p. 686),
r = Udjo,
and
so that T
r'
r'
= U (vdp),
(A2),
and therefore tt
= Uv,
are unit tangents to the lines of curvature,
proved that
dr'
= rST'dT,
(Bg),
or that
Yrdr'
= 0;
(A'2)
(B'2)
this general parallelism of dr' to r being geometrically explained,
by
observing that a line of curvature on any surface is, at the same time,
a line of curvature on the developable normal surface, which rests upon
and to which
that line,
r'
or vt is normal, if r be tangential to the
line.
(Ji).
If the vector of curvature (389) of a line of curvature be

the normal v to the given surface, the projection
projected on
(p.
686)
face,
is
the vector of curvatiire of the normal section of that sur-
which has the same tangent r
but this
result,
and an analo-
gous one (same page) for the developable normal surface (jg), are
virtually included in Meusnier's theorem, which will be proved by
quaternions in Series 414.
(0- The vector cr of a centre
of curvature of the given surface,

may (by (Wi) and (Xi)) be ex-
answering to a given point p thereon,

pressed by the equation,
<T
which may be regarded
= p + r-iv;
(C2)
also as a general form of the Vector
Equation
of the Surface of Oentres, or of the locus of the centre s

the variable vector (0 of the point p of the given surface being supposed (p. 501)
:
to be expressed as a vector function of two independent

variables, whereof therefore v, r, and a become also
although the two last involve an ambiguous
sign,
and scalar
functions,
on account of the
Ttvo Sheets of the surface of centres.

{j ).
The normal
at
s,
to
what may be called the First Sheet, has

what may (on the same plan) be
the dii-ection of the tangent t to
called the First Line of Curvature at p
and the vector
t;
of the point
CONTENTS.
xxxvi
Pages.
on the corresponding A^< of the Reciprocal (comp.

has (by p. 684) the exprespp. 507, 508) of the Surface of Centres,
corresponding to
s,
sion,
;
which may
also
= r(Spr)-i;
(D2)
(t)) to be a
be considered (comp.
form of the Vector
Equation of that Reciprocal Surface.

The vector v satisfies generally (by same page) the equations
(Jc).
of reciprocity,
Su(T=S<ru
= l,
Su^(7
= 0,
S(rû
(D/)
0,
^(T, Sv denoting any infinitesimal variations of the vectors a and v,

consistent with the equations of the surface of centres audits recipro-
cal, or
any
linear
and
vector elements of those
two
(t).
684, 685),
The equation Sv (w - p) = 0, or more simply,

Suw = 1,
which w
two
surfaces, at
we have also the relations (pp.

=
Spu l, Svw = 0, Svv0y = O.
corresponding points
(Dg")
(E2)
a variable vector, represents (p. 684) the normal plane

to \he first line (y) of curvature at p; or the tangent plane at s to the
or finally, the tangent plane to
first sheet of the surface of centres
in
is
that developable normal surface (^),
which
rests
upon the second line of
curvature, and touches the first sheet 2Xou^ a certain cw?t^, whereof we
And if v be regarded, comp. (t),
shall shortly meet with an example.
as a vector fimction of two scalar variables, the envelope of the variable
plane (Eg) is a sheet of the surface of centres ; or rather, on account of
the ambiguous sign (), it is that surface of centres itself : while, in
manner, the reciprocal surface Q')
like
plan&i
(w).
Saw =
is
the envelope of this other
1.
(E2')
The equations (Wi), (Wi) give (comp. the Note

d(r=diJ.Uv;
to p. 684),
(F2)
combining which with (C2), we
see that the equations (Hi) of p, xxv.

are satisfied, when the derived vectors p' and tr' are changed to the cor-
responding differentials, dp and dc. The known theorem (of Monge),

that each Line of Curvature is generally an involute, with the corre-
sponding Curve of Centres for one of its evolutes (400), is therefore in

and the connected theorem (also of Monge),
this way reproduced
:
that this
evoltite is
a geodetic on
its
own
sheet of the surface of centres,
follows easily from what precedes.

(w). In the foregoing paragi-aphs of this analysis, the given surface has throughout been arbitrary, or general, as stated in {d) and
But if we now consider specially the case of a central quadric,

several less general but interesting results arise, whereof many, but
perhaps not all, are known ; and of which some may be mentioned
(J).
here.
XXX Vll
CONTENTS.
=
=
Supposing, then, that not only dv (pdp, but also i/ 0p, and
Spv =fp = 1, the Index Surface (410, {d)) becomes simply (p. 670) the
from o to p ; whence many
ffiven surface, with its cent7-e transported
Pages.
(o).
simplifications follow.
(p). For example, the semiaxes ai, a2 of the index curve are now
equal (p. 681) to the semiaxes of the diametral section of the given
surface, made by a plane parallel to the tangent plane ; and Tv is, as
in 409, the reciprocal

latter plane
for the
Pi
of the perpendicular ^ from the centre on this
whence (by (Xi) and Xi")) these known expressions
two* curvatures
result
J?ri
= Par*
i?3-'
= Pa2-'.
(Gg)
if a new surface be derived from a given cen-
Hence, by (<?),
any species'), as the
(^).
locus of the extremities of normals^

erected at the centre, to the planes of diametral sections of the given
surface, each such normal (when real) having the length of one of the
tral quadric (of
semiaxes of that section, the equation of this

(p. 683) of being written thus
new surfacef admits
S|t>(0-p-2)-'p
(r).
Under the conditions
(0),
= O.
(H2)
the expression (C2) for a gives (p. 684)
the two converse forms,

<T
whence
= r-i (^ +
r) p,
(I2),
= /'(^ + r)-i(r;
(I2')
(pp. 684, 689),
= y (0 + r)-J ^(T,
and therefore
(p. 689),
by
focal surfaces,
<r =
(^-+r->)v;
(J2')
and by the theory (407) of con-
= ^2-ij/ = 02''^p,
(ri
be formed from
(J2),
(d), (p),
(K2)
by changing the semiaxes abc to azb2C2 it

being understood that the given quadric {abc) is cut by the two confoif 02
cals (aibiCi)
and
<p
{azbzc^),
in the first and second lines of curvature
and that <ri is here the vector of that^rs^

which answers to the^^rs^ line (comp. (/). Of
on the same plan, we have the analogous expression.
through the given point p
centre s of curvature,
course,
Throughout the present Series 412, we attend only (comp. ()) to the curvawhich have the directions of the ttvo
tures of the ttvo normal sections of a surface,
lines of curvature
these being in fact what are always regarded as the two principal curvatures (or simply as the two curvatures) of the surface. But, in a shortly
subsequent Series (414), the more general case will be considered, of the curva:
ture of any section, normal or oblique.

t When the given surface is an ellipsoid^ the derived surface
is
the celebrated
JFave Surface of Fresncl which thus has (H2) for a symbolicalform of its equation.
When the given surface is an hyperboloid, and a semiaxis of a section is
:
imaginary, the (scalar and

is still
to be
made equal
now positive) square, of the (imaginary) normal erected,
to the square of that semiaxis.
CONTENTS.
xxxviii
Pages.
for the vector of the second centre.

(s).
These expressions for
Salmon, namely that the
tn, 02
include
(p.
G89) a theorem of Dr.

given quadric at a
centres of curvature of a
'daa poles of the tangent plane, with respect to the two

confocals through that point ; and either of them may he regarded,
by admission of an ambiguous sign (comp. (t)), as a new Vector Form*
of the Equation of the Surface of Centres, for the case (0) of a given
given point are
central quadric.
(f).
In connexion with the same expressions
observed that
if ri, r^
scalar r in (c),
and
and second
first
for
tri,
it
0-2,
may
be
be the corresponding values of the auxiliary
if r,
/ still
denote the unit tangents
lines of curvature, while abc, aibiC\,
and
(jg)
to the
a^h2C2 retain
their recent significations (r), then (comp. pp. 686, 687, see also p.
652),
n =fT =/IJdp = (^2 - a-f)-^ = &c.,
(L2)
(L2')
r2=/r'=/[Jvdp = (a2-'i2)-i = &c.;
this association of rx and a\ with a%, &;c., and of r^ and a% with i,
&c., arising from the circumstance that the tangents r andr' have respectively the directions of the normals vz and vi, to the two confocal
surfaces, {a^hcz) and (aihic{).
and
here called ^2 is
(ti). By the properties of such surfaces, the scalar
therefore constant, in the whole extent of a first line of curvature;
and the same constancy of ro, or the equation,
d/Ux^dp
may
in various
= 0,
(M2)
ways be proved by quaternions
(p, 687).
(w). Writing simply r and r' for ri and r^, so that r' is constant,
but r variable, for o, first line of curvature, while conversely r is constant and r' variable for a second line, it is found (pp. 684, 685, 686),
that the scalar equation of the surface of centres {i) may be regarded
as the result of the elimination of r-^ between the two equations,
]
= S.(7 (1 +
whereof the
/-V)-20(r,
(X2),
and
= S.(r
(l+>-i^)-302a;
latter is the derivative of the former,
with respect
(IsV)
to the
scalar r'K It follows (comp. p. 688), that the First Sheet of the Sur-
face of Centres is touched by an Auxiliary Quadric (Xo), along a Quartic

Curve (N2) (No'), which curve is the Locus of the Centres of First Curvature, for all the points of a Line of Second Curvature ; the
same
being also touched (see again p. 688), along the same curve, by
the developable normal surface (J), which rests on the same second line :
sheet
with permission to interchange the words, first and

out the whole of this enunciation.
(;).
tors p, a,
The given
second,
through-
surface being still a central quadric (0), the vecas functions of v (comp. {j) (A:) (l)),
V can be expressed
* Dr. Salmon's
result, that this siuface of centres is of the twelfth
be easily deduced from this form.
degree,m&j
XXXlX
CONTENTS,
Pages,
and conversely the

the former
we have,
<T=(l+r-i.^)3
from which
latter
last the
ing Qk) that Sffu
can be expressed as a function of any one of
for example, the reciprocal equations (p. 685),
and
(0,),
0-iy,
= (l +
y-i0)-2 0(t
(Oa')
may be obtained anew, by observHence also, by (r), we can infer the expres-
formula (N2)
= 1.
sions,*
= (^ -1 + r-)
and in fact
it is
= 02 "^
V, (P2),
and v
= ^2 p = V2
(Pz')
easy to see otherwise (comp. p. 645), that vz
\\
||
v,
=1=
Spv, whence V2= i^ as before.

(x). More fully, the two sheets of the reciprocal (J) of the surface
of centres may have their separate vector eqtiations written thus,
and Spv2
f 1 -<p2P=^
and the scalar equationf of this
vo
V2,
0ip
= vi
(P2")
reciprocal surface itself, considered
may (by page 685) be thus written, the

being related as in 408, (J),
as including both sheets,
tions/and
i?'
vi
ftflic-
= (iFv-l)fv,
(Q2)
with several equivalent forms one way of obtaining this equation

being the elimination of r between the two following (same p. 685)
;
Fv + r-'v^
(y).
The two
= 1,
last equations
(Q3')
fv + rv^ = 0.
(Q2")
may also be written thus, for tho^rst
sheet of the reciprocal surface,
= 1,
-F2 wi
in
which (comp.
(R2),
and fUvi
= r,
(R2')
pp. 685, 689),

J?'2V
= Sv02"û = Sw(0-i+r-i)u;
(Its")
and accordingly (comp. pp. 483, 645), we have F2V2-Fv=l, and
f'Uv2=fT = r.
(z).
is
For a
line
of second curvature on the given surfacd", the scalar

and then the two equations (0,2') (Q3"), or
constant, as before;
>
(^2)5 (R's), represent jointly (comp. the slightly different enunciation
in p. 688) a certain quartic curve, in which the quadric reciprocal (E2),

of the second confocal {ao bo Co), intersects the first sheet (ij) of the Eeciprocal Surface (Q2) ; this quartic curve, being at the same time the
intersection of the quadric surface (Q2') or(R2), with the quadric cone
(Q2") or (E2'), which is biconcyclic with the given quadric, fp= 1.
* The
equation
1/
V2,
= the
normal
tually given in the text of Series 412
to the confocal {a^ bi
but
it is
c-i)
at p, is not ac-
easily deduced, as above,
from
the formulae and methods of that Series.
t The equation (Qo)
is one oi'CsxQ fourth degree ; and, when expanded by coagrees perfectly with that which was first assigned by Dr. Booth
(see a Note to p. 685), for the Tangential Equation of the Surface of Centres of a
ordinates,
it
quadric, or for the Cartesian equation of the Reciprocal Surface.
CONTENTS.
xl
Article 413.
On the Measure of Curvature of a Surface,
Pages.
.
G89-G93.
The object of this short Scries 413 is the deduction by quaternions,

somewhat more briefly and perhaps more clearly than in the Lectures^
of the principal results of Gauss (comp. Note to p. 690), respecting
ihe Measure of Curvature of a Surface, and questions therewith connected.
(a).
Let
surface,
and
on the unit
p, Pi, P2be any three near points on a given but arbitrary

r, Ri, R2 the three corresponding points (near to each other)
sphere,
which are determined by th.e parallelism of the radii
OR, ORi, 0R2 to the normals pn, PiNi, P2 N2
then the areas of the two
small triangles thus formed will bear to each other the ultimate ratio
p. 690),
hm.
ARR1R2
V.dUj/Ûi/

_ =_S^1,1
APP1P2
whence, with Oauss's
Yapop
definition of the
1^
measure of curvature, as the
ultimate ratio of corresponding areas on surface and sphere,

the formula (Zi) in 412, (/), hia ftmdamental theorem,
Measure of Curvature
,(S2)
i?i
"'
we have, by
Hz "',
(S2')
Product of the two Principal Curvatures of Sections.

(b). If the vector p of the surface be considered as a function of
two
scalar
variables,
and
tt,
and
if derivations
with respect to these
be denoted by upper and lower accents, this general transformation

results (p. 691),
Measure of Curvature
=S^S^'-(s^^,
V
V
V
\
v = Y/o'p,
in which
(T2)
(T2')
with a verification for the notation pqrst of Monge.

(c). The square of a linear element d*, of the given but arbitrary
surface, may be expressed (p. 691) as follows
:
d52
= (Tdp2 =) ed^2 + 2/d^dM + gdiu"^
c,f,9\
^J',9'\
efg,;
and
the form of this function (p. 692) agreeing, in

*
corresponding expression assigned by Gauss.
(U2)
and with the recent use (J) of accents, the measure (T2)
(same pagef) to be an explicit function of the ten scalai's,
is
e,-2f',^g";
all its details,
proved
(IJ2')
with the
Hence follow
at once (p. 692) two of the most important

mathematician on this subject; namely, that
every Deformation of a Surface, consistent with the conception of it as
an infinitely thin and flexible but ineztensible solid, leaves unaltered.
(<?).
results of that great
* References are
given, in Notes to pp. 690, &c. of the present Scries 413,
to the pages of Gauss's beautiful Memoir, " Disqtiisitiones generates circa Superficies Curvas^^ as reprinted in the Additions to Liouvillc's Monge.
CONTENTS.
xli
Pagea.
the Measure of Curvature at any Pointy and Ilnd, the Total

Curvature of any Area : this last being the area of the corresponding
portion {a) of the unit-sphere.
1st,
By a
(e).
suitable choice of
the expression (U2)
and
u, as certain, yeodetic co-ordinates,
be reduced (p. 692) to the following,
may
ds''=^dt^
+ nHti^;
(U2")
the length of a geodetic arc ap, from a fixed point a to a

variable point p of the surface, and u is the angle bap which this
where
t is
makes with a
variable arc
neighbourhood of
fixed geodetic
we have w = f, and
a,
ab
n'
so that in the immediate
= T>tn = 1.
(/). The general expression (c) for the measure of curvature takes
thus the very simple form (p. 692),
Hi-^Jî-^
and we have (comp.
(i?))
= - n-^n" = - n-^Tit^n
(V2)
the equation (p. 693),
Total Curvature of Area
apq = Aw - J w'dM
(Vg')
bounded by two geodetics, ap and aq, which make with

each other an angle = Aw, and by an arc pa of an arbitrary curve on
the given surface, for which t, and therefore n\ may be conceived to
this area being
be a given function of
u.
pq be itself a geodetic, and if we denote by v the

variable angle which it makes at p with ap prolonged, so that tan v
= ndM dt, it is found that dv = ~ n'du and thus the equation (V2')
(jf).
If this arc
conducts (p. 693) to another very remarkable and general theorem of

Gauss, for an arbitrary surface^ which may be thus expressed.
Total Curvature of a Geodetic Triangle
= what may
abc = a +
b+c 7r,
(V2")
be called the Spheroidal Excess of that triangle, the
total
area (Itt) of the unit-sphere being represented by eight right angles :

with extensions to Geodetic Polygons, and modifications for the case of
what may on the same plan be

tivo
called the Spheroidal Defect,
curvatures of the surface are oppositely directed.

Article 414. On Curvatures of Sections (Normal
of Surfaces
(a).
when
the
and Oblique)
and on Geodetic Curvatures,
694-698
The curvatures considered in the two preceding
Series hav-
ing been those of the principal normal sections of a surface, the present
Series 414 treats briefly the more general case, where the section is
made by an
arbitrary plane, such as the osculating plane at p to an

arbitrary curve upon the surface.
(i). The vector of curvature (389) of any such curve or section
k)-^
being (p
to be (p. 694),
(p
(t)-i
= D,2p,
=
V-
'
its
S -^
normal and tangential components axe found
(p
- (ri)-i
cos2 V
and
(p
(T2)"i
sin2 v,
W3)
(W2')
(p-0-'='-'dp-Svdp-idV;
the former component being the Vector of Normal Curvature of the
CONTENTS.
xlii
Pages.
and the latter

Surface, for the direction of tlie tangent to the curve
being the Vector of Geodetic Curvature of the same Curve (or section).
:
In the foregoing expressions, a and ? are the vectors of the

and x, in -which the axis of the osculating circle to the curve
intersects respectively the normal and the tangent plane to the surface (p. 694) s is also the centre of the sphere, which osculates to
(c).
points s
the surface in the direction dp of the tangent ; a\, a^ are the vectors
of the two centres Si, Sg, of curvature of the surface, considered in Se-
which are at the same time the centres of the two osculating
which the curvatures are (algebraically) the greatest and
and v is the angle at which the curve here considered crosses
ries 412,
spheres, of
least
\h.Q first line
form
of curvature.
The equation (W2) contains
{d).
(p. 695),
i?-
= iJr
cos2 v
a theorem of Euler, under the
+ i?2-i sin2 V
( W2")
contains also Meusnier's theorem (same page), imder the form

(comp. 412, (/)) that the vector of normal curvature (h) of a surface,
it
any given direction, is the projection on the normal v, of the vector

of oblique curvature, whatever the inclination of the plane of the sec-
for
tion to the tangent plane
The expression
(e).
mits
(p.
may
be.
(W2'), for the vector of geodetic curvature, ad-
697) of various transformations, with corresponding expres

T(p ^) of geodetic curvature, which is also the
sions for the radius
radius of plane curvature of the developed curve, when the developable

circumscribed to the given surface along the given curve is unfolded
into a plane
curve
and when this radius
is constant,
so that the developed
or part of one, it is proposed (p. 698) to call the given

curve Q.Bidonia (as in the Lectures'), from its possession of a certain isois
circle,
perimetrical property, which was first considered by M. Delaunay,

and is represented in quaternions by the formula (p. 697),
^JS(Ui^.dp5p) + c^JTdp = 0;
or
c-'dp
by the
(X2)
= V(Uj/.dUdp),
(X'2)
what may be called the Calculus of Variations in Quabeing a constant, which represents generally (p. 698)
rules of
ternions
the radius of the developed circle, and becomes infinite for geodetic
lines, which are thus included as a case of Didonias.
Article
41 5.
Supplementary Remarks,
698-706
Simplified proof (referred to in a Note to p. xii), of the general existence of a system oithree real and rectangular directions, which
{a).
satisfy the vector equation
and
cubic equation Jf =0 (p.
^p^p =
0, (P),
when
is
a linear, vector,
and of a system of three real roots of the

xii), under the same condition (pp. 698-
self-conjugate fimction
700).
(^). It
may happen
(p.
701) that the differential equation,
Srdp =
0,
(Y2)
CONTENTS.
xliii
Pages,
or represents a system of surfaces, tvithout the expression
In this case,
Si/dp being an exact differential, as it was in 410, (i).
there exists some scalar /acôr, n, such that S^?vdp is the exact diffeis integrable,
rential of a scalar function of
tor p
is itself
(pp. 701, 702,
jO,
without the assumption that this vect; and then if we write
function of a scalar variable ^
a,
comp.
p. xxx),
d V =
df = ^dp,
this

manner
tegrability of the
found
it is
equation (Yg)
mula,
(p. 702), that the Condition*
ofln-
expressed by the very simple for-
is
Syv = 0;
(Yz")
in which y is a vector function of p, fiot generally linear, and deduced

from (p on the plan of the Section III.:îi- 6 (p. 442), by the relation,
0dp-0'dp = 2Vydp;
0'
(Yo'")
being the conjugate of <p, but not here equal to it.

(d). Connexions (pp. 702, 703) of the Mixed Transformations in
the last cited Section, with the known Modular and Umbilicar Generations of a surface of the second order.
The equation
{e).
(p. 704),
T(p - V.^Vya) = T(a - V.y V;3p),
(Z^)
which a, /3, y are any three vector constants, represents a central

quadric, and appears to offer a new mode of generation-^ of such a surface, on which there is not room to enter, at this late stage of the
in
work.
(/). The vector of the centre of the quadric, represented by the
equation fp
= m~^\pe
(p.
The equation
(g).
const., with fp
Sp0p, is generally
case of paraboloids, and of cylinders.
28ep
704)
k= ^-'e
(p. 705),
S>qpq'pq"p
+ Sp^p + Syp +
C = 0,
(Z2')
represents the general surface of the third degree, or briefly the General
Cubic Surface ; C being a constant scalar, y a constant vector, and q,
q',
q" three constant quaternions, while

and self-conjugate function of p.
^p
is
here again a linear,
vector,
The General
(A).
Cubic Cone, with its vertex at the origin, is thus
represented in quaternions
* It
comes,
is
shown,
in a
when expanded,
by the monomial equation (same
Note to
the
p.
known
page).
702, that this monomial equation (Y"2) beof six terms, which expresses the con-
eqtiation
dition of integrability of the differential equation pdiX + q^y + rdz - 0.

t In a Note to p. 649 (already mentioned in p, xxviii), the reader will find
references to
t]xQ
Lectures, for several î^QXQTxi generatiotrs of the ellipsoid, derived
from quaternion forms of
its
equation.
CONTENTS.
xliv
Pages.
Sqpq'pg'p
0.
(Z2")
Screw Sections (p. 705)

(t). Screw Surface,
Arch, with illustration by a diagram (Fig. 85,
Skew
Centre of Skew
p. 706).
8.
On a few Specimens of Physical Applications
of Quaternions, with some Concluding Remarks, 707 to the end.
Section
Article 416. On the
Statics of a Rigid
= 2Vai3;
VyS/3
each a
707-709
Body,
Equation of Equilibrium,
(a).
is a vector of application ;
(A3)
the corresponding vector of applied
(3
force ; y an arbitrary vector and this one quaternion formula (A3)

scalar equations
is equivalent to the system of the six usual
:
(X = 0, r=
(A.)
0,
Z= 0, i = 0,
When
M= N= 0).
0,
= 0,
S(2/3.SVa/3)
but wo ^
(B3),
2/3
= 0,
(C3)
resultant = 2/3, which
the appHed forces have an unique

the line whereof (A3) is then the equation, with y
acts along
for its variable
vector.
(c).
When
the condition (C3)
is satisfied,
the forces
compound
themselves generally into one couple, of which the axis = 2 Va/3, whatever may be the position of the sissumed origin o of vectors.
(<?).
2Va/3 =
When
0,
(D3),
with or without (C3),
the forces have no tendency to turn the body round that point o and
when the equation (A3) holds good, as in (a), for an arbitrary vector
y, the forces do not tend to produce a rotation* round any point c,
;
and both the
so that they completely balance each other, as before,

conditions (C3) and (D3) are satisfied.
In the general
(g).
case,
when neither (C3) nor (D3) is
satisfied,
\iq
be an auxiliary quaternion, such that

ff2;3
then
\q
= 2Va/3,
(E3)
the vector perpendicular from the origin, on the central
is
axis of the system ; and if c = S^-, then c2/3 represents, both in quantity and in direction, the axis of the central couple.
(/). If
Q be
another auxiliary quaternion, such that
Q2/3
with T2/3
* It
is
> 0,
then
SQ = c =
(F3)
2a/3,
central
moment divided by
total force ;
easy to prove that the moment of the/ore^ /3, acting at the end of the
o, and estimated with respect to any unit-line 1 from the same ori-
vector a from
gin, or the energy
round that line

St"ia/3
with which the force so acting tends
regarded as ^ fixed axis,

so that when the condition (D3)
t,
to cause the
body
to turn
Siaj3, or
represented by the scalar,
is satisfied, the applied forces have no
is
tendency to produce rotation round any axis through the origin : which origin
"comes an arbitrary point c, when the equation of equilibrium (A3) holds good.
xlv
CONTENTS.
Pages,
and \Q is the vector y of a point c upon the central axis -which does
not vary with the origin o, and which there are reasons for considering
as the Central Foint of the system, or as the general centre of applied
forces
what
is
(7).
in fact, for the case of parallelism, this point c coincides with

usually called the centre of parallel forces.
Conceptions of the Total Iloment 2a/3, regarded as heing ge-
nerally a quaternion
and of the Total Tension,
Sa/3,
considered as
a scalar to which that quaternion with its sign changed reduces itself
for the case of equilibrium (a), and of which the value is in that
case independent of the origin of vectors.
(K). Principle of Virtual Velocities,
^S(3da
Article 417.
(a).
= 0,
(G3)
On the Dynamics of ^ Rigid Body,
709-713
General Equation of Dynamics,
2*S(D2a-?)â = 0;
(H3)
the vector ? representing the accelerating force, or m^ the moving

of which the vector at the time Hs a ;
force, acting on a particle
and da heing any infinitesimal variation of this last vector, geometrically compatible with the connexions between the parts of the
system, which need not here be a rigid one.
(5). For the case of a. free system, we may change each â to
+ Yia,
being any ttco infinitesimal vectors, which do not change in
passing from one particle m to another and thus the general equae
and
tion (H3) furnishes two general vector equations, namely,
and
2mVa (D2a - ?) =
2(D<2a_^) = 0, (I3),
(J3)
which contain respectively the law of the motion of the centre of
gravity, and the law of description of areas.
If a body be supposed to be rigid, and to have sl fixed point
{c).
and we may write,
o, then only the equation (J3) need be retained
;
Dta=Yia,
t
being here
Rotation
o.
(K3)
namely the Vector Axis of Instantaneous

Ut denoting the direction of that axis, and its
finite vector,
its versor
tensor Tt representing the angular velocity of the
time
body about
it,
at the
t.
(d).
When
the forces vanish, or balance each other, or
compound
themselves into a single force acting at the fixed point, as for the case
of a heavy body turning freely about its centre of gravity, then
2;nVa^ = 0,
so that
we
shall
and
if
we
write,
0t=SmaYat,
(M3)
have the equations,
(pDii + YKpiÔ,
whence
(L3)
again denotes a linear, vector, and self-conjugate function,
Sty
(Jfs);
A2
= 0,
'f>i+
(Q3),
= 0,
(O3);
and
St0t=A2j
(p'Dti
= Yiy;
(P3)
(R3)
the vector y being what we may call the Constant of Areas, and the
scalar A' being the Constant of Living Force.
CONTENTS.
xlvi
Pages.
One
of Poinsot's representations of the motion of a body, under

the circumstances last supposed, is thus reproduced under the form,
(e).
that the Ellipsoid of living Force (P3), with

o, rolls without gliding on the fixed plane
point
its ceittre at
(Qti),
which
the fixed
is parallel
Plane of Areas (Sty = 0) ; the variable semidiameter of contact,

being the vector-axis (c) of instantaneous rotation of the body.
to the
I,
i
(/). The Moment of Inertia, with respect to any axis through o,
equal to the living force (A^) divided by the square (Ti^) of the
semidiameter of the ellipsoid (P3), which has the direction of that axis ;
is
and hence may be derived, with the help of the first general construC'
Hon of an ellipsoid, suggested by quaternions, a simple geometrical
representation (p. 711) of the square-root of the moment of inertia
of a body, with respect to any axis ad passing through a given point
= CA, with the help of two other
A, as a certain right line bd, if cd
points B and c,
chosen in more
{g).
which are likewise

ways than one.
A cone of the
fixed in the body, but
may
be
second degree,
Sta/=0,
with
(S3),
J/
= 720t _ A202t,
(T3)
infixed in the body, but rolls in space on that other cone, which is the
locus of the instantaneous axis i
and thus a second representation,
;
proposed by Poinsot, is found for the motion of the body, as the rolling
of one cone on another.
(Ji). Some of Mac Cullagh's results, respecting the motion here
considered, are obtained with equal ease by the same quaternion
analysis
for example, the line y, although fixed in space, describes

an easily assigned cone of the second degree (p. 712), which
in the body
cuts the reciprocal ellipsoid^
= h^,
Sy<p-'7
in a certain sphero-conic
(Us)
and the cone of normals
tioned cone (or the locus of the line
+ /i2y-i)
rolls
to the last
men-
on the plane of areas
= 0).
(Sty
(i).
The Three
(Principal) Axes of Inertia o the body, for the

(p. 712) of the three rectangular and
given point o, have the directions

vector roots (comp. (P), p. xii.,
and the paragraph 415,
(a), p. xlii.)
of the equation
\i(pi=0,
and
(Vz),
because, for each, D<t =
(V3')
denote the three Principal Moments of inertia, correJ, B,

sponding, then the Symbolical Cubic in 
+ ^) (0 + C) =
0.
from moments referred
to those
which correspond
through any other point
of the body.
(W3)
to axes passing
to respectively
CONTENTS.
Article 418.
xlvii
On the motions of a System of Bodies,
as free particles m,
m,
Pages.
considered
whicli attract each other according to the
713-717
law of the Inverse Square

(a). Equation of motion of the system,
SwSD2aâ + ^P=
a
is
P= Swm'T (a - a')"' (Js)

P is the poof the mass or particle m
if
(Xa),
0,
the vector, at the time
tcntial {ox force-function)
t,
and the infinitesimal variations â are ar-
bitrary.
(V).
Extension of the notation of derivatives^
(c).
The
dP=='ES(J)aP.da).
m,
differential equations of
become thus,
?nD<2a
+ DaP=0,
and the laws of the centre of gravity, of

are obtained under the forms,
= (3,
^ml>ta
(A4)
and of living
areas,
force,
2wVaD<a = y;
(B4);
T=- ^^m (Dia)2= P+
and
(3,
(Zs)
motion of the separate masses
y being two vector constants, and
(C4)
(d/)
_ff;
H a scalar constant.
(d). "Writing,
P= r (P+ T) df,
(E4),
and
F= r
2 Tdt
F may
= F+
tH,
(F4)
be called the Principal* Function, and

the Characteristic
Function, of the motion of the system ; each depending on the final
vectors of position, a, a',
and on the
F depending also (explicitly)

depends instead on the
final
and
on the
cottstant
initial vectors
S oi
initial vectors, uq, a'o,
time,
while
t,
but
V (= the Action')
living force, in addition to those
the masses m, m',
being supposed to be
known, or constant.
(e).
We
are led thus to equations of the forms,
mBia + DaP= 0,
(G4)
-mDâ + I>a^F= 0,
(D,P) = -JT,
(H4)
(I,)
whereof the system (Gi) contains what may be called the Intermediate Integrals, while the system (H4) contains the Final Integrals,
of the differential Equations of Motion (A4).
(/).
In
like
manner we
find equations of the forms,
(K4); Di^F=<; (L4)

the intermediate integrals {e) being here the result of the elimination
Dar=-mT)ta,..
(J4);
D^F=mDo,
" 0 General
* References are
given to two Essays by the present writer,
in Dynamics," in the Philosophical Transactions for 1834 and 1835, in
which t\iQ Action (Y), and a certain other function (S), which is here denoted by P,
Method
But the ana^

called, as above, the Characteristic and Principal Functions.
here used, as being foimded on the Calculus of Quaternions^ is altogether
unlike the analysis which was employed in those former Essays.
were
lysis
CONTENTS.
xlviii
Pages.
of H, between the system ( J4) and the equation (L4) and the final
integrals, of the same system of differential equations (A4), being now
;
(theoretically) obtained,
by eliminating the same constant
R between
(K4) and (L4).
The functions F and
(^).
V are
obliged to satisfy certain Partial

which those relative to the
Differential liquations in Quaternions, of

final vectors a, a',
are the following,
(D,JF')-iSm-i(D,i^)2=P, (M4); |2w-i(Dr)2 +
and they are subject
P+J? = 0;
to certain geometrical conditions,
(N4)
from which
can be deduced, in a new way, and as new verifications, the law of mo-
and the law of description of areas.
tion of the centre of gravity,
General approximate expressions
(Ji).
iând F, and
for their derivatives fi'and
(p.
t,
717) for the functions

mo-
for the case of a sAor^
tion of the system.
Article 419.
on the
On the Relative Motion of a Binary System
and
Law
of the Circular Hodograph,

of one body from the other being a, and the dis(fl). The vector
tance being r (=Ta), while the sum of the masses is M, the differen-
equation of the relative motion
tial
square,
Da =
is,
with the law of the inverse
ilfa-ir-i
(O4)
being here used as a characteristic of derivation, with respect to the

time t.
As a
(h).
which holds good
first integral,
of central force,
VaDa = j8 =
which
also for
any
other law
we have
a constant vector
(P4)
includes the two usual laws, of the constant plane (-i-
of the constant ureal velocity
/3),
and
triT,).
Writing r = Da = vector of relative velocitij, and conceiving this

new vector r to be drawn from that one of the two bodies which is
(c).
here selected for the origin

T
is
and
a
o,
the locus of the extremities of the vector
(by earlier definitions) the Hodograph of the Relative Motion

this
hodograph
is
prove* to be, for the
Law of the Inverse Square,
Circle.
(d). In fact, it isshown (p. 720), that for any /azi' of central force,
the radius of curvature of the hodograph is equal to the force, multiplied into the square of the distance, and divided by the doubled areal
velocity
or
by the
constant parallelogram
c, under the vectors (a

or of the orbit and the hodograph.
It follows then, conversely, that the law of the inverse
square
;
and r) oi position and

(e).
velocity,
the only law which renders the hodograph

generally a circle ; so
that the law of nature may be characterized, as the Law of the Circular
is
Hodograph : from which latter law, however, it is easy

the/onw of the Orbit, as a conic section with &, focus at o.
to
deduce
717-733
xlix
CONTENTS.
Pages.
p,
(/). If the semiparameter of this orUt be denoted, as usual,

if h be the radius of the Jwdograph, then (p. 719),
by
and
h
((jr).
city, in
The
= Mc-^=-cp-^ = {Mp-^)^.
the sense that eh
graph, from the point

(/).
The
hodographic
an
circle (e
< 1)
(e>
(^=
a parabola,
it is
1)
in
and
all
of the hodo-
here treated as the centre of force.

is interior to the
when the point
is
ellipse,
And
1).
the distance of the centre
is
which
orbit is
ference of that circle

terior point
(Q4)
orbital excentncity e is also the hodographic excentri-
it is
when
these cases, if
<?=^(l-e)-i
is
on the circtim-
an hyperbola, when o
we
is
an ex-
write
= tf/ri(l-e2)-',
(R4)
the constant a will have its usual signification, relatively to the

orbit.
(e). The quantity Mr-^ being here called the Potential, and denoted by P, geometrical constructions for this quantity P are assigned,
with the help of the hodograph (p. 723) and for the harmonic mean,
2M(r + r')-i, between the two potentials, and P', which answer to
;
the extremities
of any proposed chord of that circle all

constructions are illustrated by a now diagram (Fig. 86).
x, x'
which
ij). If u be the pole of the chord xx' m, m' the points in which
the line oii cuts the circle l the middle point, and n the pole, of the
;
new chord mm', one secant from which

u' the intersection of this secant
last pole is
thus the line nxx'
with the chord mm', or the harmonic
conjugate of the point u, with respect to the same chord ; and nt,x/
any near secant from n, while v, (on the line ou) is the pole of the
near chord
TTI
then the two small
i,x
arcs,
and
I'x/, of the
hodo-
graph, intercepted between these two secants, are proved to be xHiandP'; or to the two
matelj proportional to the two potentials,
the perpendiculars let fall from x and x', on

here be called the hodographic axis ln. Also, the harmonic
ordinates xv,
xV, namely
what may
mean between these two ordinates
the line u'l; while ux, ux', and
hodograph, so that
points, X, x'
and
x,,
is
obviously (by the construction)

u,x/ are four tangents to the
u,x,,
this circle is cut ortJwgonalhj, in
x/,
by two
other circles,
the
ttvo
pairs of
which have the two near
points u, u^ for their centres (pp. 724, 725).

In general, for any motion ofa^mnt (absolute or relative, in
(Ji).
one plane or in space, for example, in the motion of the centre of the
moon about that of the

attractions of the
earth, under the perturbations produced by the

sun and planets), with a for the variable vector (418)
opposition of the point, the time -dt which corresponds to any vectorelement dDa of the hodograph, or what may be called the time of hO'
dographically describing that element, is the quotient obtained by

dividing the same element of the hodograph, by the vector of acceleration
Da
in the orbit ; because
we may
Avrite generally (p. 724),
CONTENTS.
Pages.
(0- For the law of the inverse square (comp. (a) and
measure of the/ore^ is,
TD2a = Mr-^ = M'^P^

the times
arcs T,T
the
())?
(T4)
df, of hodographically describing the small circular

the
t't/ of the hodograph, being found by multiplying
dt,
and
lengths 0") of those two arcs by the mass, and dividing each product
by the square of the potential corresponding, are therefore inversely
as those tivo potentials^ P, P', or directly as the distances^ r, r', in the
orbit
so that
we have
the proportion,
dt:df:dt + dit'=r:r':r + r'.
(Ui)
suppose that the mass, Jf, and the Jive points o, l, m,

the chord mm' are given, or constant, but that the raTj, v^ upon
dius, h, of the hodograph, or the position of the centre h on the hodo-
we
(m). If
graphic axis ln, is altered, it is found in this way (p. 725) that
although the two elements of time, d^, d^', separately vary, yet their
sum remains unchanged : from which it foUows, that even if the tioo
(y) between
two secants from the pole n of tho fixed chord mm', the sum (say, A< +
A^') of the two times is independent of the radius, h.
circular arcs, tt, t't/, be not small, but still intercepted
(w). And hence may be deduced (p. 726), by supposing one secant
become a tangent, this Theorem of JEodographic Isochronism, which
was communicated without demonstration, several years ago, to the
Royal Irish Academy,* and has since been treated as a subject of
to
investigation
by
several able writers
common chord, which passes

through, or tends towards, a common centre offorce, he cut perpendicularly by a third circle, the times of hodographically describing the inter-
If two
circular hodographs, having a
cepted arcs will be equal.

(0).
This common time can easily be expressed
form of the
(p. 726),
under the
definite integral,
_.
Time of tmt
2MC^
= -^
ff^
Jo (1
dw
- e'; cos wy
^
(V4)
^
2g being the length oi the fixed chord mm';
which reduces itself to
when
is at m',
e' the
quotient lo lm,
that is for the case of a pa:
1 for an ellipse, and outside those limits

an hyperbola, but being, in all these cases, constant ; wiule <; is a
certain auxiliary angle, of which the sine = iit ul (p- 727), or
rabolic orbit ; e lying between
for
denote the length pp' of the chord of the orbit, corresponding to the chord tt' of the hodograph ; and w varies from to tt,
when the yfh(ê periodic time 27rw"i for a closed orbit is to be computed
-s{r\ r')"i,
if s
with the
verification, that the integral (V4) gives, in this last case,
Jf =
a^n^, as usual.
* See the
It
Proceedings of the 16th of March, 1847.
of force is occupied by a common mass, M.
centre
common
(W4)
is
imderstood that the
CONTENTS.
li
Pages.
the general composition of
By examining
(p).
tlie definite
inte-
by more purely geometrical considerations, -which are

by Fig. 87, it is found that, with the law of the inverse
gral (V4), or
illustrated
t of
describing an are pp' of the orbit (closed or unfunction (p. 729) of the three ratios^
square, the time

closed) is
Q.
and therefore simply a function of the chord (s, or pp') of the orbit,
and a are
and of the sum of the distances (r + r'y or op + op') when
given which is a form of the Theorem of Lambert,
{q). The same important theorem may be otherwise deduced,
through a quite different analysis, by an employment oi partial deri-
and oi partial
vatives,
differential equations in quaternions,
which
is
analogous to that used in a recent investigation (418), respecting the

motions of an attracting system of any number of bodies, m, m, &c.
now (comp. p. xlvii) the following expression for the
(r). "Writing
relative living force, or for
the mass
(M- m + m'^,
multiplied into the
square of the relative velocity (TDa),

2
T= - JfDa2 = 2 (P+ E) = 31 (2r- - -i)
introducing the two new integrals
J'=f'(P+T)d^
(Y4)
(p. 729),
and
(Z4),
(A5)
V=['^1Tdit=^F+tH,
which have thus (comp. (E4) and (F4)) the same forms as before, but
different (although analogous') significations, and may still be
with
called the Principal
denoting by a,
sition^ or of the
orbit,
two corresponding
when
and Characteristic Functions of the motion
a' (instead of ao,
while
r, r'
and
a) the initial Sind final vectors opoare the two distances, and r, r' the
vectors of velocity, or of the hodograph
it is
found
F may be treated as
a function of a, a', i, or
as a function of a, a, a, or of r, /, s, and JT; and
of r, r, s, t, and
that their partial derivatives, in the first view of these two functions,
that
is
given,
are (p. 729),
Dai^=DF=r,
(l)t)F=-H,
J)a'F=Ba'V=^-r'
(Bs);
and
(Dg);
while, in the second view of the
D^F= Dr=^;
same functions, they
(C5)
(E5)
satisfy the two
'partial differential equations (p. 730),
BrF^Dr'F,
(F5),
and
D,.F=D/r;
(G5)
along with two other equations of the same kind, but of the second
degree, for each of the functions here considered,
mentioned in p. xlviii.
which are analogous
to those
(s).
The equations (F5) (G5)
express, that the two distances, r
and r', enter into each of the two functions only by their sum ; so that,
if
M be
still
treated as given,
F may be regarded as a
function of the
CONTENTS.
lii
Pages.
three quantities, r + r\ a, and t ; while V, and therefore also t by
be a function of the three scalars,
(Eo), is found in like manner to
which last result respecting the time agrees with

s, and a
and furnishes a new proof of Lamberfs Theorem.
+ r',
(^),
(t).
three partial differential equations (r) in
The
F conduct, by
for the three partial

merely algebraical combinations, to expressions
and thus, with the help
derivatives, DrF, Hr V (~IirV), and D^F;
of (E5), to ttvo new definite integrals* (p. 731), which express respecin the relative motion of a Unary
tively the Action and the Time,
system here considered, namely, the two following
}.s\r+r+8
t
whereof the
= l\
4:(t)
^ds;
latter is not to
(Is)
be extended, without modification, be-
yond the limits within which the radical is finite.

Article 420. On the determination of the Distance of a Comet,
new
Planet, from the Earth,

(a). The masses of earth and comet being neglected, and the mass
of the sun being denoted by M, let r and w denote the distances of
or
733, 734
earth and comet from sun, and z their distance from each otber, while
a is the heliocentric vector of the earth (Ta = r), known by the theory
of the sun, and p is the unit- vector, determined by observation, which
Then it is easily proved by
directed from the earth to the comet.
is
quaternions, that
we have
the equation (p. 734),
SpDpDV_^|^_ ^\
z
SpDpJJa
with
tv"^
r^
\r^
z^
w^
- 2zSap
(Ja)
j'
;
(K5)
eliminating w
between these two formulae, clearing of fractions, and

dividing by 2, we are therefore conducted in this way to an algebraical equation of the seventh degree, whereof one root is the sought distance, z.
(5).
The final
and by the
equation, thus obtained, differs only by its notation,,

from that assigned for the same
facility of its deduction,
purpose in the Mecanique Celeste; and the rwZtJofLaplace there given,

by inspection of a celestial globe, which of the two
for determining,
* References are
given to the First Essay, &c., by the present writer (comp.
the Note to p. xlvii.), in which were assigned integrals, substantially equivalent
to (H5) and (I5), but deduced by a quite different analysis.
It has recently been
remarked to him, by his friend Professor Tait of Edinburgh, that while the area
described, with Newton's Law, about ihQ full focus of an orbit, has long been
known to be
proportional to the time corresponding, so the area aboxit th^ empty
focus represents (or is proportional to) the action.
CONTENTS.
liii
Pages,
bodies (earth and comet)
the formula (J5).
is
the nearer
sight from
to the stm, results at
the
Article 421. On the Development of the Disturbing Force of

Sun on the Moon or of one Planet on another, which is nearer
;
the Sun,
Let a, <T be the geocentric vectors of moon and sun r (= Ta),
(rt).
and s(=T<t), their geocentric distances Mthe sum of the masses of
than
itself to
earth and
moon
the mass of the sun
mark of derivation with respect
the
and
7}
J7,
(L5),
D
:
moon about
equation of the disturbed motion of the
D2a=i{fâ+
and
to the time
if
Vector of Disturbing Force
(as in recent Series)
then the
differential
the earth
is,
= <p(a) = a-^Tar\
= S{(pa f (o- a))
<l>a
(M5)
(N5)
denoting here a vector function, but not a linear one.

(b). If we neglect ;, the equation (L5) reduces itself to the form
D2a = M(pa
elliptic
which contains (comp. (O4)) the laws of undisturbed
motion.
(c). If we develope the disturbing vector 77, according to ascending powers of the quotient r s, of the distances of moon and sun from
the earth, we obtain an infinite series of terms, each representing a
:
group oi partial disturbing forces, which
may be thus
= Jji + 72 + 73 + &c.
= ;i,l+>?l,2, ?2 = 72,1+?72j2+?2,3, &C.
?l
finite
denoted,
(O5)
(P5)
these partial forces increasing in number, but diminishing in intensity,

and being conin the passage from any one group to the following
;
nected with each other, within any such group, by simple numerical
ratios
and angular relations.

For example, the two forces
r)uii
(d).
rigorously, proportional to the
arc,
numbers
J7i,2
1
and
of the first group

3
the three forces
of the second gxow^ are as the numbers 1, 2, 5; and

the /o?* forces of the jfAi'rt? group are proportional to 6, 9, 15, 35:
while the separate intensities of the first forces, in these three first
72,2, ?2)3
72, i>
groups, have the expressions,
Sr
T7i=-;
(e).
ZSr'^
T^2,i=-g^;
T,;3,x
5S;-3
-^^.
(Qa)
All \kGSQ partial forces are conceived to act at the moon; but
may be represented by the respectively parallel unit-
their directions
lines JJt]\,i, &c.,

circle
unity),
and
drawn /rom
]),
which passes through

of the sun and moon in
of
its
radius equal to
the geocentric (or apparent)
^;(?ac^5,
the heavens.
J) of ^woowfrom sun
(/). Denoting then the geocentric elongation
by -f 9 and by 0i, 02, and Di, 1)2,
what may be called two fictitious suns, and three fictitious moons,
(in the plane of the three bodies)

])3,
and terminating on a great
the earth,
of the celestial sphere (supposed here to have
which the corresponding elongations from 0,
in the same great
734-736
CONTENTS.
liv
circle,
are
20,
and -0, +
20,
30,
Pages,
by Fig. 88
30, as illustrated
(p. 735); it is found that the directions of the two forces oiihejirst
group are represented by the two radii of this unit-circle^ which terminate in D and J)\ those of the three forces of the second group, by the
three' radii to 0i, 0, and 0o; and those of the /owr forces of the
;
with facilities for exthird group, by the radii to D2, D, Di, and
;
tending all these results (with the requisite modifications), to the
fourth and subsequent groups, by the same quaternion analysis.
(gy And it is important to observe, that no supposition is here
made
respecting any smallness of excentricities or inclinations (p. 736) ;

so that all theformulce apply, with the necessary changes oi geocentric to heliocentric vectors, &c., to the perturbations of the motion of a
comet about the sim, produced by the attraction of a planet,
(at the time) more distant than the comet from the sun.
Article 422
(a). If p and
On
Fresnel's
which
is
736-756
Wave,
be two corresponding vectors, of ray-velocity and

wave-slowness, or briefly Bay and Index, in a biaxal crystal, the veloand if ^p and ^{i be any inficity of light in a vacuum being unity
fi
nitesimal variations of these
two
vectors, consistent
tions (supposed to be as yet unknown), of the
Wave
with the equa-
(or wave- surface^,
and its reciprocal, the Index- Surface (ox surface of wave- slowness) we
have then first the fundamental Equations of Reciprocity (comp. p.
:
417),
Syup=-1,
(II5);
S/i^p
= 0,
(S5);
S/)^/f
which are independent of any hypothesis respecting the
= 0,
(T5)
vibrations of
the ether.
(b). If 5p be next regarded as a displacement (or vibration), tangential to the wave, and if de denote the elastic force resulting, there
exists then, on Fresnel's principles, a relation between these two small
vectors
which relation may (with our notations) be expressed by

two following equations,
either of the
= (!>-^dp,
(U5),
or
Sp
= ^Ss',
(Vs)
being of that linear, vector, and self- conjugate kind,

which has been frequently employed in these Elements.
the function
(c). The fundamental connexion, between the functional symbol

and the optical constants abc of the crystal, is expressed (p. 741,
comp. the formula (W3) in p. xlvi) by the symbolic and cubic equa-
^,
tion,
(^
+ a-2)
(<p
5-2)
(0
+ c-2) =
(NYs)
which an extensive use is made in the present Series.

The normal component, p.-'^Sfids, of the elastic force de, is in((?).
effective in Fresnel's theory, on account of the supposed incompressiof
hility
of the ether; and the tangential component, ^-^Sp- fx-^Sfide, is

same theory, and with present notations) to be equated to
(in the
Iv
conte:nts.
Pages.
fi-^cp, for
the propagation of a rectilinear vibration (p. 737)
"we ob-
tain then thus, for such a vibration or tangential displacement, dp, the
expression,
= (0-1 _ /-2)-i/A-i Bfids
dp
(Xs)
and therefore by (Sg) the equation,

O
= S)tt-^(0-i-/t*-2)-V-S
which is a Symbolical Form of the

face, and may be thus transformed,
= S;t.(,*2_0)-,^.
(^).
surface
(Y5)
scalar Equation of the Index- Sur-
(Z5)
The Wave- Surface,

(^d), is
as being the reciprocal (a) of the indexeasily found (p. 738) to be represented by this other
Symbolical Equation,
0=^Sp-i(i>-p-^y^p-^;
or
(Ae)
(Be)
Sp(|o2-0-i)-ip.
(/). In such transitions, from one of these reciprocal surfaces to

the other, it is found convenient to introduce two auxiliary vectors,
V and (i)(= 0u), namely the lines oix and ow of Fig. 89 ; both drawn
from the common centre o of the two surfaces; but v terminating (p.
738) on the tangent plane to the wave, and hem^ parallel to the direction of the elastic force 6e ; whereas w terminates (p. 739) on the tangent plane to the index- surface, and is parallel to the displacement dp.
{g).
Besides the relation,

(o
or
^v,
t;
= 0"'a;,
(Ce)
connecting the two new vectors (/) with each other, they are connected with p and fi by the equations (pp. 738, 739),
= -1,
(Be);
Spv = 0;
Spu)=-1,
(Fe);
S/iw
S^t;
(Eg)
= 0;
(Ge)
and generally (p. 739), the following Eule of the Interchanges holds
good In any formula involving p, n, v, w, and 0, or some of them,
it is permitted to
with 0'i proexchange p with fi, v with w, and
vided that we at the same time interchange dp with de, but not generally* dfi with 5p, when these variations, or any of them occur.
(A). We have also the relations (pp. 739, 740),
:
- p-i = u- iVv jw = )u + v-i

/a-i = (o'^Ywp = p + w'
(He)
(le)
* This
apparent exception arises (pp. 739, 740) from the circumstance, that
dp and Se have their directions generally fixed, in this whole investigation
(although subject to a common reversal by +), when p and p. are given ; whereas
dp. continues to be used, as in (), to denote any infinitesimal vector, tangential to
the index-surface at the
end of
p..
CONTENTS.
Ivi
rage=
with others easily deduced, which may all be
illustrated
by the above-
cited Fig. 89.

(^).
Among
such deductions, the following equations
(p.
740)
be mentioned,
may
(Yv(pvy + Sv(pv =
(Vw0-iw)2 xSw^-ia) = 0;
(Je);
0,
(Ke)
which show that the Locus of each of the two Auxiliary/ Foints, v and
w, wherein the two vectors v and w terminate (/), is a Surface of
the Fourth Degree, or briefly, a Quartic Surface; of which two lociihe
constructions mQ.j
two reciprocal
he connected (as stated in
p.
741) with those of the
ellipsoids,
Sp(pp=l,
(Lg),
and
8,0^-^
= 1;
(Me)
an arbitrary semidiamefer.
however, a much more interesting use of these two
p denoting, for each,
Q'). It is,
of which (by (W5), &c.) the scalar semiaxes are a,
the^rs^, and a"i, J"i, c-i for the second, to observe that they
ellipsoids,
employed
(pp. 738,
b, c
for
may
be
739) for the Constructions of the JFave and the
Index- Surface, respectively, by a very simple rule, which (at least for
the first of these two reciprocal surfaces (a)) was assigned by Fresnel himself.
In fact, on comparing the symbolical form (As) of the equa(A;).

tion of the Wave, with the form (H2) in p. xxxvii, or with the equation 412, XLI., in p. 683, we derive at once FresneVs Construction :
namely,
through
that if the ellipsoid

its
centre,
and
{abc')
be
by an arbitrary plane
that plane be erected at
cut,
if perpendiculars to
that central point, which shall have the lengths of the semiaxes of
the section, then the locus of the extremities, of the perpendiculars so
erected, will be the sought
Wave-Surface.
A precisely
similar construction applies, to the derivation of

and thus the two
the Indêx- Surface from the ellipsoid (^"'^"'c-i)
(J).
auxiliary surfaces, (Lg) and (Me), may be briefly called the Generating Ellipsoid, and the Reciprocal Ellipsoid.
(m).
The cubic (W5)
inverse function
ing
to -p-2, &c.,
in
<p
enables us easily to express (p. 741) the
e is any scalar
and thus, by changnew forms of the equation (Ae) of the wave are
(jp -f e)-i,
obtained, whereof one
where
is,
= (^-ip)2 + (p2 + a2 + J? +
c^)
Sp0-'p
- amt^
(Ne)
with an analogous equation in p. (comp. the rule in (jf)), to represent

the index-surface : so that each of these two surfaces is of the fourth
degree, as indeed is otherwise
known.
(n). If either Sp^-'p or p2 be treated as constant in (Ne), the
degree of that equation is depressed
and therefore the Wave

Sp0-ip
is cut,
= AS
from the fourth
by each of the
(Oe),
p2
to the second;
tivo concentric qtiadrics,
r2
= 0,
in a (real or imaginary) curve of thefourth degree
(Pg)
of which two quar-
Ivii
CONTENTS.
Pages.
answering
tic curves,
wave
is
(o).
to all scalar values of the constants
h and
the
r,
the common locus.
The new
and
ellipsoid (Og) is similar to the ellipsoid (Me),
similarly placed, while the sphere (Pe) has r for radius ;
and every
because it is, by the

quartic of the second system (n) is a sphero-conic,
with
equation (Ae) of the wave, the intersection of that sphere (Pe)
the concentric and quadric cone,
O=Sp(0 + r-2)-ip;
or,
(Qe)
*
by (Be), with this other concentric quadric,
-l = Sp(^-i + r2)-ip,
(Re)
whereof the conjugate (obtained by changing equation) has
to
in the last
(Se)
<i2-r, ^2-^2, t;2-r2,

for the squares of its scalar semiaxes,
and
is
therefore confocal with
the generating ellipsoid (Le).
{py For any point p of the wave, or
at the
end of any ray
the
p,
tangents to the two curves (w) have the directions of w and fiut ; so
that these two quartics cross each other at right angles, and each is a
common orthogonal in all the curves of the other system.

be parallel
(j2). But the vibration dp is easily proved to
to
hence the curves of the^;-^ system (w) are Lines of Vibration of the
Wave : and the curves of the second system are the Orthogonal Trajectoriesf to those lines.
(r). In general, the vibration Sp has (on Fresnel's principles) the
direction of the projection of the ray p on the tangent plane to the
wave
and the
elastic force ds
projection of the index-vector
surface
has in like manner the direction of the

p,
on the tangent plane
to the index-
so that the ray is ihMs, perpendicular to the elastic force
Abticle 423.Mac Cullagh's Theorem of the Polar Plane,
757-762
* For real curves of the second

system (n), this new quadric (Re) is an hywith one sheet or with two, according as the constant r lies between a
perboloid,
and
b,
or
between
sheets or one, in the
and c and, of course, the conjugate hyperboloid

same two cases respectively.
;
(o)
has two
t In a different theory of light (comp. the next Series, 423), these spheroon the wave are themselves the lines of vibration.
conies
Iviii
CONTENTS.
Table* of Initial Pages of Articles.
Art.
CONTENTS.
Table of Initial Tagbs continued.

[
Art.
lix
ELEMENTS OF QUATERNIONS.
BOOK
I.
ON VECTORS, CONSIDERED WITHOUT REFERENCE TO ANGLES,

OR TO ROTATIONS.
CHAPTER
I.
FUNDAMENTAL PRINCIPLES RESPECTING VECTORS.

Section I. On
Art.
A right
the Conception of
a Vector; and on Equdity
of Vectors.
having not only length,

be a Vector. Its initial point a
is said to be its origin; and its final point b is said to be its
term,
vector ab is conceived to be (or to construct) the
but
line ab, considered as
also direction, is said to
differencô? its two extreme points ; or, more fully, to be the

result of the subtraction of its own origin from its own term ;
and, in conformity with this conception, it is also denoted by

the symbol b - a a notation which will be found to be exten:
sively useful, on account of the analogies which it serves

to express between geometrical and algebraical operations.
When the extreme points a and b are distinct, the vector ab

or b - A is said to be an actual (or an effective) vector but
;
when
two points are conceived to coincide,

the vector aa or a - a, which then results, is said to be null.
vector.
Opposite vectors, such as ab and ba,
b-a
or B - A and a - b, are sometimes
(as a limit) those
called vector
and
sive vectors,
such as ab and bc, or
Succes-
revector,
_
^_2
<
^^s- 1B - a and c - B, are occasionally said

to be vector and provector: the line ac, or c - a, which
TO
Revector.
is
drawn from the
[book
I.
origin a of the first to the term c of the second,
being then said to be the trans-
At
vector.
a later stage,
we
shall
have to consider vector-arcs and

vector-angles ; but at present, our
only vectors are (as above) right A^
>
'b
I'ig- 2.
lines.
2. Two vectors are said to be equal to each other, or the

= CD, or b - a = d - c, is said to hold
equation ab
good, when
and
the
term
of
the
one
can
be brought
when)
origin
(and only
to coincide respectively with the corresponding points of the

It
other, by transports (or by translations) without rotation.
follows that all null vectors are equal,
and may therefore be
denoted by a common symbol, such as that used for zero ; so that
we may
write,
-B
0;
but that two actual vectors, ab and cd, are not (in the present
full sense) equal to each other, unless they have not merely
If then they do not
equal lengths, but also similar directions.
happen to be parts of one common line, they must be opposite
sides
abdc
of a parallelogram,
the two lines ad, bc
becoming thus the two
dia-
gonals of such a figure, and

consequently bisecting each
some point e.
Conversely, if the two equaother,
in
tions,
D- e=E-
A,
and
-E=e-
B,
two lines
AD and bc are commedial, or have
a common middle point e, then even
are satisfied, so that the
they be parts of one right line,

the equation D-c = B-Ais satisfied.
Two radii, ab, ac, of any
one circle (or sphere), can never be
if
directions differ.
Fig. 4.
equal vectors
because their
CHAP.
FUNDAMENTAL PRINCIPLES
I.J
VECTORS.
3. An equation between vectors^ considered as an equidifference of points, admits of inversion and

^
^
alternation ; or in symbols, if
J
D-C=B -
A,
then
C - D
Fig. 5.
=A-B,
and
D- B = C-A.
Two
vectors,
cd and ef, which are
equal to the same third vector, ab,

are also equal to each other ; and
these three equal vectors are, in

general, the three parallel edges of
a prism.
Section
2.
On Differences and Sums of Vectors taken two

by two,
4.
In order to be able to write, as in algebra,

^
= c - B, if c' - a' = c - A,
(c'
a')
a)
(b
we next define,
when a first vector ab is subtracted from

ac
a second vector
which is co-initial with it, or from a third
is
vector a'c' which
equal to that second vector, the remainder
is that fourth vector bc, which is drawn from the term b of the
first to the term c of the second vector
so that if a vector be
subtracted from a transvector (Art. 1), the remainder is the
that
It is evident that this geometrical

provector corresponding.
subtraction of vectors answers to a decomposition of vections (or
o motions) ; and that, by such a decomposition of a null vection into
two
opposite vections,
(b
so that, if an actual vector
A A, the remainder
we have
the formula,
= (a - a) - (b - a) = A - b
a)
is
ab be subtracted from a
the revector ba.
If then
null vector
we
agree to
- a to the
abridge, generally, an expression of the form
=
ab ba a and - a
shorter form,
a, we may write briefly,
being thus symbols of opposite vectors, while a and (- a) are,
;
[bOOK
same reason, symbols of one common vector
for the
we may
I.
so that
write, as in algebra, the identity,
5, Aiming still at agreement with algebra, and adopting

on that account the formula of relation between the two signs,
+ and -,
(b -a) + a==b,
in
which we
sum is
their
shall say as usual that

b,
summands, we
b-
is
added to
a,
and that
while relatively to it they may be jointly called

shall have the two following consequences
:
I. If a vector, ab or b - a, be added to its own

origin
the sum is its term b (Art. 1) and
II. If a provector bc be added to a vector ab, the sum
the transvector ac or in symbols,
a,
is
I.
(b
- a) + A = B
and
11.
(c
b) + (b
a)
= c - a.
first equation is an immediate consequence of the

formula
which, as above, connects the signs + and -,
general
when combined with the conception (Art. 1) of a vector as a dif-
In
the
fact,
; and the second is a result of the same

the definition of the geometrical subwith
combined
formula,
traction of one such vector from another, which was assigned
in Art. 4, and according to which we have (as in algebra) for
ference of two points
any three
points^ a, b, c, the identity,

(c
a)
(b
a)
=c-
B.
clear that this geometrical addition of successive vectors

corresponds (comp. Art. 4) to a composition of successive vec-
It
is
tions, or
two
opposite
revector)
and that the sum of

vectors (or of vector and
motions
is
a null
BA + AB =
0,
line ; so
or (a
that
b) + (b a) =
0.
It follows also that the sums of equal

pairs of successive vectors are equal;
or more fully that
if b'
a'
=b-
A,
and
c'
b'
=c-
b,
then
c'
CHAP.
I.]
VECTORS.
being in general the two oppo^

Art.
prism (comp.
3).
Again, in order to have, as in algebra,
the two triangles, abc and
a'b'c',
site faces oô.
6.
(c'
b')
+ (b - a) = c - A, if
c'
- b' = c - B,
be two successive vectors, ab, bc,

b'c'
be
vector
and
equal to the second, but not successive to the first, the sum obtained by adding the third to the
we
shall define that if there

if a third
that fourth vector, ac, which is

A of the first to the term c of the se-
first is
It follows that the
cond.
drawn from the origin
sum of any
two co-initial sides ^ ab, ac, of any /?rZlelogram abdc, is the intermediate and
CO- initial diagonal
(c
a) + (b
ad
= a) D
or, in
a, if
symbols,
d -c = b - a
Fig. 8.
we have then (by 3) c-a = d-b.

The sum of any two given vectors has
because
7.
independent of their order
is
thus a value which
or, in symbols, a
-)-
/3
= /3 +
a.
If equal vectors be added to equal vectors, the sums are equal

vectors, even if the summands be not given as successive
(comp. 5) ; and if a null vector be added to an actual vector,
sum is that actual vector or, in symbols, + a = a. If

then we agree to abridge generally (comp. 4) the expression
the
+ to +
and
if
still
(- a) = +
8.
a,
a, and + (+ a),
and we have, as
denote a vector^ then +
&c., are other symbols for the

in algebra, the identities,
same vector
+ (- a) = - (+ a) = -
a,
(+ a) + (- a)
0,
&c.
On Sums of three or more Vectors.
Section
3.
The sum
of three given vectors, a,
j3,
y, is
next defined
to be that fourth vector,
S=7 +
(j3
a),
or briefly,
S=7 + j3 + a,
obtained by adding the third to the sum of the first

and second and in like manner the sum of any number of
which
is
vectors
is
formed by adding the
last to the
sum
of
all
that
precede it: also, for any four vectors, a,

S + (7 + /3 + a) is denoted simply by S + 7 +
rentheses, and so
on
for
[bOOK
/3,
7, B, the
I.
sum
+ a, without paany number of summands.

/3
The sum
of any number of successive vectors, ab, bc,

CD, is thus the line ad, which is
drawn from the origin a of the first,
9.
to the
term d of the
last
and be-
cause, when there are three such vectors, we can draw (as in Fig. 9) the
two diagonals ac, bd of the (plane
or gauche) quadrilateral abcd, and
may
"^
p.
then at pleasure regard ad, either as the sum of ab, bd,

sum of ac, cd, we are allowed to establish the follow-
or as the
ing general formula of association, for the case oi any three
summand
lines, a, /3,
(7 +
/3)
+a=7+
(/3
a)=7 +
j3
+ a;
by combining which with the formula of commutation (Art.

namely, with the equation,
a +
/3
j3
7),
a,
which had been previously established for the case of any two
such summands, it is easy to conclude that the Addition of
Vectors is always both an Associative and a Commutative OpeIn other words, the sum of any number of given vectors
ration,
has a value which is independent of their order, and of the
mode of grouping them so that if the lengths and directions of
the summands be preserved, the length and direction of the
sum will also remain unchanged except that this last direction
;
may be regarded as indeterminate, \f\ieJi the length of the sumline happens to vanish, as in the case
which we are about to consider.
1 0. When
any n summand-lines,
AB, BC, CA, or AB, BC, CD, DA, &C.,
arranged in any one order, are the n
successive sides of a triangle abc, or of
a quadrilateral abcd, or of any other

closed polygon, their sum is a null line, aa
Fig. 10.
and conversely,
CHAP.
I.]
when
the
VECTORS.
sum of any given system ofn vectors is thus equal

may be made (in any order, by transports without
to zero, they
n successive
sides of a closed polygon (plane or

be
gauche). Hence,
îven any such polygon (p), supa
it
is
pose
pentagon abcde,
possible to construct another
closed polygon (p'), such as a'b'c'd'e', with an arbitrary initial
rotation) the
if there
a', but with the same number of sides, a'b',

e'a', which
new sides shall be equal (as vectors) to the old sides ab,
ea,
point
taken in any arbitrary order.
For example,
if
we draw^wr
successive vectors, as follows,
AB =
CD,
B C = AB,
CD
EA,
D E = BC,
and then complete the new pentagon by drawing the line e a',
this closing side of the second figure (p') will be equal to the
remaining side de of the^r^^ figure (p).
is still a closed one, when
11. Since a closed figure abc
all its points o^re projected on any Sissumed plane, by any system
.
of parallel ordinates (although the

aiêa of the projected figure a'b'c'
may happen to vanish), \t follows that
.
if
number of given
be zero, and if we
on any one plane by
the sum of any
vectors a,
j3,
project them
parallel lines
y,
all
mities,
drawn from their extrethe sum of the projected vec-
tors a,
/3',
y'i
will likewise be null;
-^
Fig. 11.
so that these latter vectors, like the
former, can be so placed as to become the successive sides of a

closed polygon, even if they be not already such. (In Fig. 1 1
a"b"c" is considered as such a polygon, namely, as a triangle
,
with evanescent area; and
we have
the equation,
a"b" + b"c" + c"a"
0,
as well as
a'b'
b'c'
c'a'
0,
and ab + bc + ca =
0.)
Section
12.
or
by
The
4.
On
[bOOK
I.
of Vectors,
Coefficients
simple or single vector, a, is also denoted by la,

by (+ 1 ) a ; and in like manner, the double vector,
.a, or
a^â^ is denoted by 2a, or 2 a, or (-1- 2) a, &c. ; the rule being,

that for any algebraical integer, m, regarded as a coefficient by
which the vector a is multiplied, we have always,
.
la + ?wa = (1
+m)
the symbol 1 + m being here interpreted as in algebra. Thus,

Oa = 0, the zero on the one side denoting a null coefficient, and
the zero on the other side denoting a null vector; because by
the rule,
la4 Oa =
(l
+ 0)a = la =
a,
and
.'.(ia
a-a = 0.
=
Again, because (I) a + (- 1) a (1 1) a = Oa = 0, we have
- a = -a =
=
-(la); in like manner, since (l)a + (-2)a
(- l)a
= (1 - 2)a = (- 1) a = - a, we infer that (-2)a = -a-a = - (2a)
and generally, (- rn) a = - (ma), whatever whole number m
may be so that we may, without danger of confusion, omit
the parentheses in these last symbols, and write simply, - 1 a,
- 2a, -ma.
;
13. It follows that ivhatever tivo whole numbers (positive or

and n, and tvhatnegative, or null) may be represented by
Fig. 12.
ever
tivo
vectors
may
be denoted by a and
j3,
we have
as in algebra, the formulae,
nama=
(n m)
and (compare Fig.
a,
n (ma) = (nm) a = nma,
12),
(j3
a) = w/3 + ma
always,
CHAP.
I.J
VECTORS.
SO that the multiplication of vectors by coefficients is ii doubly

distributive operation, at least if the multipliers be whole
numbers; a restriction which, however, will soon be re-
moved.
14. If
ma = /3,
the coefficient
m being still whole, the vector
said to be a multiple of a ; and conversely (at least if the

/3
be different from zero), the vector a is said to be a
integer
is
sub-multiple of j3.
multiple of a sub-multiple of a vector is
said to be q. fraction of that vector ; thus, if /3 = ma, and y = wa,
n
then y is a fraction of /3, which is denoted as follows, y =
i3 ;
also j3
and y
said to
is
is
two
by
th|||practional coefficient
It fol-
and y be any two fractions (positive or negative

Avhole numbers being included), and if a and j3 be any
if
v>ectors,
a?
then
yaxa = {yx)a,
results
multipjlied
said to be the product of this multiplication.
lows that
or null,
be
y{xa) = (yx)a = yxa,
x{(5
a) =
xj^
Xa
which include those of Art. 1 3, and may be extended

where x and y are i7icommensur able coefficients, con-
to the case
sidered as limits
oi' fractional
ones.
For any actual vector
a, and for any coefficient a^, of

any of the foregoing kinds, theproduct xa, interpreted as above,
represents always a vector (5, which has the same direction as
the multiplicand-line a, if ro 0, but has the opposite direction
if a; < 0, becoming null if ^ = 0.
Conversely, if a and /3 be any
15.
two actual vectors, with directions either similar or opposite, in

each of which two cases we shall say that they are parallel
vectors,
and
shall write j3
||
a (because both are then parallel,
in the usual sense of the word, to one
always
find, or
common
conceive as found, a coefficient
line),
we can
xÔ, which
shall
= Xa; or, as we shall also write it,

satisfy the equation (5
=
and
the
P ax;
positive or negative number x, so found, Avill
bear to 1 the same ratio, as that which the lenyth of the line
bears to the length of
a.
10
Hence
16.
quotient which
natural to say that this coefficient x is the

from the division of the vector j3, hy the
and to write, accordingly,
it is
a;
we
I.
results,
parallel vector a
SO that
[bOOK
or x = j5:
j3 -r- a,
a,
or
a;
shall have, identically, as in algebra, at least if the
be an actual vector, and
divisor-line a
if the dividend-line
j3
he
parallel thereto, the equations,
(j3
which
a)
.a= a = j3,
will afterwards
non-parallel vectors.
Xa:a=
and
" extended, by
We
may
= x;
definition, to the case
conditions, d = , and may say that the vector a

X
of the division of the other vector
we
shall
have these other
17.
The
(5
the quotient
is
by the number x
so that
identities,
.x = (ax=)By
X
of
write also, under the same
'
X =
and
a.
positive or negative quotient, x--=^,
which is thus
obtained by the division of one of two parallel vectors by another, including zero as a limit, may also be called a Scalar
;
because
structed,
(or axis)
it
can always be found, and in a certain sense con-
by the compainson ofpositions upon one common

;
scale
or can be put under the form,

c -A
AC
where the three points, a, b, c, are collinear (as

Such scalar s are, there- ^
annexed).
^
simply the Reals (or real quanin combinatities) oi Algebra; but,
fore,
in the
figure
'
'
"Z,
Vectors above considered, they

main
elements
of the System, or Calculus, to
form one of the
tion with the not less real
CHAP.
POINTS AND LINES IN A GIVEN PLANE.
II.]
11
which the present work relates. In fact it will be shown, at

a later stage, that there is an important sense in which we can
conceive a scalar to be added to a vector ; and that the sum
so obtained, or the combination,
^^
is
Scalar plus Vector ^^
Quaternion.
CHAPTER
II.
APPLICATIONS TO POINTS AND LINES IN A GIVEN PLANE.
Section
On Linear
Equations connecting two Co-initial

Vectors.
When
18.
one
several vectors, oa, ob,
common
System
we
that all
are
all
drawn from
In the present and future secalways suppose, if the contrary be not expressed,
the vectors a, j3,
which we may have occasion to
the vector
tions
said to be the Origin of the

point o, that point
and each particular vector, such as oa, is said to be
is
of its own term,
a.
shall
consider, are thus
drawn from one common
origin.
But
be
if it
term-
desired to change that origin o, without changing the

points a, . we shall only have to subtract, from each of their
old vectors a, . one common vector w, namely, the old vector
- a>,
- w,
oo' of the new
j3
origin o'; since the remainders, a
will be the new vectors a',
of the old points a, b, . . . For
.
j3',
example,
we
shall
a = o'a =
have
a-o'=
(a
19. If two vectors a,
and
= OA - oo'=
o) -(o'- o)
|3,
or oa, ob, be thus
if their
o
given origin o,
directions be either similar or
w.
drawn from a
b
'
opposite,
a-
'
7*^
so that the three
on one right line (as in the figure

points^ o, a, b, are situated
ELEMENTS Of QUATERNIONS.
12
is
annexed), then (by 16, 17) their quotient a
I.
[bOOK
some
positive or
and conversely, the equation

negative scalar, such as x
= xa, interpreted with this reference to an origin, expresses
(5
the condition qfcollinearity, of the points o, a, b the particu;
lar values,
tions,
o and
20.
The
0,
to the particular po^z-
x=\, corresponding
a,
of the variable point b^ whereof the indefinite
is
the locus,
oa
right line
j3,
a;
linear equation, connecting the two vectors a
more symmetric ^rw, when we write
acquires a
a +
^/3
The
important.
and
thus
however only the ratio

two points
wliere a and h are tioo scalars, of which

is
it
condition of coincidence, of the
a and B, answering above to
a?
is
now
-7-
or,
more
symmetrically,
a^h=0.
when
Accordingly,
a = -b, the linear equation becomes
b{(5-a)-~=0,
since
we do not suppose
=
or
/3-a =
0,
that both the coefficients vanish
and
ob = oa, requires that thepointB should

j3
with the point a a case which may also be conve-
the equation
coincide
a, or
niently expressed by the formula,
B = a;
coincident points being thus treated (in notation at least) as
In general, the linear equation gives,
equal.
OA +
Section
2.
OB =
0,
and therefore
6=
bo
oa.
On Linear Equations betiveen three co-initial

Vectors.
21. If
two
(actual
and
co-initial) vectors, a, )3,
be not con-
= 0, with ???/ tivo

a?ii/ equation of the form aa-^ b[5
scalar coefficients a and b whatever, their directions can neither
nected by
be similar nor opposite to each other ; they therefore determine
CHAP.
II.]
13
a plane aob, in which the (now actual) vector, represented by

the sum aa + ^/3, is situated. For if, for the sake of symmetry,
we denote
sum by the
where c is some
third scalar, and 7 = 00 is
some third vector, so that the
symbol
this
cy,
three co-initial vectors, a,

7, are
j3,
connected by the linear
equation,
aa-^ bj5 + Cy = 0;
and
if
we make
OA =
-aa
then the two auxiliary points, a' and b', will be situated (by
19) on the two indefinite right lines, oa, ob, respectively
:
and we
shall
have the equation,
OC = OA'-fOB',
so that the figure a'ob'c
is
(by 6) a parallelogram, and conse-
quently plane.
22. Conversely, if c be any point in the plane aob,
we can
draw from it the ordinates, ca' and cb', to the lines oa and ob,
and can determine the ratios of the three scalars, a, b, c, so as
to satisfy the two equations.
OA
oa
OB'
OB
which we shall have the recent expressions for oa', ob',

with the relation oc = oa' + ob' as before ; and shall thus be
= 0, which
+
brought back to the linear equation oa +
after
5/3
C'y
therefore be said to express the condition ofcomplanarity of the four points, o, a, b, c. And if we write it under
equation
may
the form,
xa+
7/j3
+ 2y
and consider the vectors a and
/3
=^
0,
as given, but
y as a variable
vector, while x, y, z are variable scalars, the locus of the variable point c will
then be the given plane, oab.
ELEMENTS OF QUATERNIONS,
14
23. It
that the point c is situated on the right

here considered as a given one. In that
may happen
which
line AB,
is
case (comp. Art. 17, Fig. 13), the quotient

to
some
[book
scalar,
suppose
so that
we
shall
AC
must be equal
have an equation of
the form,
or 7 = a + ^(i3-a),
y^^Û
by comparing which
last
or (1
-Oa + ^/3 - 7 =
form
with the linear equation of Art.

21,
of
we
see that the condition
of the
collinearity
is
plane oab,
three
in the given
points A, B, c,
expressed by the
formula,
fl
+ 6+
This condition
c=
Fig. 16.
0.
may also be
~a -b
1
thus written,
OA OB
+
OA OB
or
and under this last form it expresses a geometrical

which is otherwise known to exist.
24.
relation,
When we
have thus the two equations,

a + 6/3 + cy = 0, and + ^ + c =
0,
so that the three co-initial vectors a, /3, 7 terminate on one

right line, and may on that account be said to be termino-collinear, if
we
eliminate,
the three scalars a,
b, c,
successively and separately, each of

are conducted to these three other
we
equations, expressing certain ratios of segments

6 (|3
a)
+ c(7
=
a)
0,
c(y
fi)
+ a(a -
(5)
0,
a(a-y) + b(j5-y) = 0;
or
= i.AB + C.AC = C.BC + a.BA = a.CA + 5.CB.
Hence
follows this proportion, between coefficients and seg-
ments,
tt
:6:c= Bc CA ab.
:
CHAP.
We
II.]
might
also
15
have observed that the proposed equations
give,
a =
^0 + ^7
AB
)3-a
C'y
+aa
_aa + b[5
whence
25. If
we
still
treat a
a+h
and
- as variable, the
equation
as given^ but regard
(5
7 and
a;
^+y
will express that the variable point c is situated someivhere
on the indefinite right line ab, or that it has this line for its
locus
while
it
divides theJinite line
the variable quotient
CB
Let
c'
ab into segments of which

,
is,
be another point on the 5a?we /me, and

,
Xa +
2/'/3
let its
vector be,
16
26.
When the
[bOOK
anharmonic quotient becomes equal to nega-
unity, the group becomes (as

If then we have the two equations,
tive
y' =
rca
+ 2/i3
x+
well
is
xa- pQ
y = ^
,
'
1/
known) harmonic.
-y
the two points c and c' are harmonically conjugate to each other,
with respect to the two given points, a and b and when they
vary together, in consequence of the variation of the value of
;
-, they
form
(in a
well-known
sense),
on the indefinite right
line AB, divisions in involution; the double points {ov foci) of
namely, the points of which each is its own

As a verificonjugate, being the points a and b themselves.
cation, if we denote by /x the vector of the middle point m of
this involution,
the given interval ab, so that
M
^
'
=
/3-/i=iU-a, or/i i(a+/3),
we
'
"
'
easily find that
P-/U
y+ x
y'-fx
so that the rectangle under the distances mc, mc', of the two
variable but conjugate points^ c, c', from the centre m of the
involution, is equal to the constant square of half the interval
between the two double
points, a, b.
More
generally, if
we
write
xa-^y^
''
<2?
'
?/
_lxa + my (5
^
Ix-vmy
XIX
^^^'xs any
where the anharmonic OMO\jiexi\>
^
^ constant scalar.
m xy
then in another known and modern* phraseology, the points
c and c' will form, on the indefinite line ab, two homographic
More
divisions, of which a and b are still the double points.
generally
still, if
we
establish the
* Seethe Gtometrie
Supe'rieure of
1^1.
two equations,
Chasles, p. 107.
(Paris, 1852.)
CHAP.
II.]
y=
being
I
and
^^,
constant, but
still
and y =
^^,
y^jIx + my
x^y
fft
y
-
17
'
variable, while a
oa',
j3'
ob',
the two given lines, ab and a'b', are then homographically divided, by the two variable points, c and c', not
oc',
now supposed
move along one common
to
27. When the linear equation aa +

between
without the relation a + 6 + c =
line.
5/3 +
cy
subsists,
its coefficients,
then
^, y are still complanar, but they

no longer terminate on one right line ; their term-points a, b, c
being now tho corners of a triangle.
the three co-initial vectors
In
tors
more general case, we may propose

y of the three points,
this
a', j3',
a'
= oabc,
C'=
that
is
a,
to find the vec-
b'=ob-ca,
OCAB
to say, of the points in

lines drawn from the
which the
origin o to the three corners of

the triangle intersect the three
respectively opposite sides. The three collineations oaa', &e.,
give (by 19) three expressions of the forms,
where x, y, z are tlirce scalars, which it is required to determine by means of the three other collineations, a bc, &c., with
the help of relations derived from the principle of Art. 23.
Substituting therefore for a its value x'â, in the given linear
equation, and equating to zero the sum of the coefficients of
the new linear equation which results, namely,
x-âa + hji-\^cy;
from the
and eliminating similarly

ginal equation
X=
we
each in its turn,

/3, 7,
find the values,
-a
T
y=
-b
c\a
z=
- c
;
ori-
18
whence the sought vectors are expressed

following ways
[bOOK
in either of the
I.
two
i...a=rr
f3
6 + c
=
c
+a
,;
y=
a^
'
or
,
11.
we
= 6)3
-T
+ Cy
+ c
^ aa
= cy
J
c + a
-.,
ffa
+ ij3
7 =^ a +
J--'
by one of these expressions for a', that a' is on

the line oa and by the other expression for the same vector
As another veria, that the same point a' is on the line bc.
In
fact
see,
;
fication, we may observe that the last expressions for a, /3', y\

coincide with those which were found in Art. 24, for a, j3, y
themselves, on the particular supposition that the three points

A, B, c were coUinear.
We
may next propose to determine the ratios of the

sides of the triangle abc, made by the points
the
of
segments
For
this
c'.
a', b',
purpose, we may write the last equations
28.
for a,
j3',
y'
under the form,
= 6(a-i3)-c(y-a') = c(i3'-7)-(a-/3') = (7'-a)

-Hfi-y')',
and we
see that they then give the required ratios, as follows
ba'_
cb'
a'c
b'
b'a
whence we obtain
as the condition
at once the
Ac'_ b
_a
c'
c'b
known
ba'
cb'
ac'
AC
b'a
c'b
'
equation of six segments,
'
of concurrence of the three right
lines aa', bb',
common point, such as o. It is easy also to infer, from

the same ratios of segments, the following proportion of coefficc', in
cients
and areas,
a:b:c= obc oca gab,

:
in
which we must,
in general, attend to algebraic signs ; a triangle being conceived to pass {through zero) from positive to
negative, or vice versa, as compared with any given triangle in
CHAP.
its
11.]
19
plane, when (in the course of any continuous change)

vertex crosses its base.
It may be observed that with this
own
its
convention (which
ment
Qt{
is,
in fact, a necessary one, for the establish-
general formulae)
equation
we
have, for any three points, the
ABC + BAC =
we had
exactly as
(in Art. 5) for
0,
any two points, the equa-
tion
AB+ BA =
More
fully,
we
0.
have, on this plan, the formulae,
ABC = - BAG = BCA = - CBA = CAB = - ACB
and any two complanar triangles, abc, a'b'c', bear to each other
^positive or a negative ratio, according as the two rotations,
which may be conceived to be denoted by the same symbols
ABC,
a'b'c',
29. If
are similarly or oppositely directed.

and b' bisect respectively the sides bc and ca,
a'
then
a =b=
and
bisects
c'
ab
c,
whence the known theorem
follows, that
of the sides of a triangle concur, in a point

often called the centre of gravity, but which we pre-
the three bisectors
which
is
fer to call the
origin o.
for a,
/3',
At
7'
mean point of the triangle, and which is here the

the same time, the first expressions in Art. 27
become,
^^'2'
"'^"^
whence
^="2'
known theorem results,
this other
that the three bisec-
tors trisect each other.
30. The linear equation between a,

the case last considered, to the form,
a +
/3
the three vectors
+ 7=
0,
j3,
7 reduces
or oa + OB + oc =
itself, in
or oa, ob, oc, are therefore, in this

Art.
case, adapted (by
10) to become the successive sides of a
a, /3, 7,
20
transports loithout rotation
by
triangle,
we complete
[book
and accordingly,
I.
if
(as in Fig. 19) the
parallelogram aobd, the triangle

OAD will have the property in
It follows (by 11)
question.
that if we project XhQfour points
o, A, B, c, by any system o^ parallel ordinates, into four other
points,
o^, A^,
on any
as-
sumed pZawe, the sum of the three

projected vectors, a^, j3^, y^, or
o A &c., will be null; so that
,
Fig. 19.
we
shall
have the new linear
equation.
/
i3,
7.
^>
or.
+ o
oÂ^
and
in fact
B^
o^c^
evident (see
it is
Fig. 20) that the projected

wâ/?oiwô^ willbethe mean
point of the projected triangle,
A^,
Fig. 20.
We shall have also the equation,
(a-a) + (ft-0)
(y,-7) = O;
where
= o\^ - QA = (o A + AA
a^- a
(oo^
+ oâ) = AA^ -
00^
hence
OO^
= J (aA^
-I-
BB^ + CC
),
or the ordinate of the mean point of a triangle

the ordinate s of the three corners.
Section
31.
3.
sides bc, ca, ab, of the triangle
seek to assign the vectors

tion (comp. 27),
ease of Art. 27, in
which
supposed to be unequal, we may next

points a", b", c" do the lines b'c', c'a', a'b'
b, c are
what
meet respectively the
inquire, in
mean of
the
On Plane Geometrical Nets.
Resuming the more general
the coefficients a,
is
a", /3",
or
may
y" of the points of intersec-
chap.
points and lines in a given plane.
ii.]
a" = b'c''BC,
The
first
c"-a'b'ab.
b"=c'a''Ca,
expressions in Art. 27 for
21
j3',
give the equa-
7',
tions,
Fig. 21.
(c + a)/3' + ^j3
=
(a + h)y' + C7
0,
whence
bbut (by 25) one
(a + b)
{c
member
is
-^^
a)
the vector of a point on bc, and

is a value for the
the other of a point on b'c' each therefore

vector a" of a", and similarly for /3 and 7".
;
"
We
may
aa
bl5
there-
fore write,
bj3-cy
/3"
- aa
c- a
= Cy
-b
and by comparing these expressions with the second set of

in Art. 27, we see (by 26) that the points
j3', 7'
a", b", c" are, respectively, the harmonic conjugates (as they
are indeed known to be) of the points a', b', c', with respect
to the three pairs of points, b, c
so that, in the
c, a
a, b
notation of Art. 25, we have the equations,
values of a,
(ba'ca")
And
(cb'ab")
(ac'bc")
=-
I.
because the expressions for a", j3", y" conduct to the

lowing linear equation between those three vectors,
fol-
22
[boOK
I.
{b-c)a"+{c-a)l5"+{a-b)j"=0,
with the relation
between
{b-c)+{c-a) + {a-b) =
coefficients, we arrive (by 23) at this
its
other known
theorem, that the three points a", b", c" are coUinear, as indicated by one of the dotted lines in the recent Fig. 2 1
.
The
32.
sal,
line a"b 'c'
may represent any

cutting the sides of a triangle abc and because
rectilinear transver-
we have
AC
AC'
CB'
, .,
,
Avhile-7- = -, and -t-
BA
CB
- a"
-j
I)
as beiore,
we
arrive at this other
equation of six segments, for any triangle cut by a right line
(comp. 28),
ba" cb' ac'
a"c
b'a
c'b
Avhich again agrees with known results.

33. Eliminating j3 and 7 between either set of expressions
(27) for j3' and 7', with the help of the given linear equation,
we
a,
arrive at this other equation, connecting the three vectors
/3',
7'
O = - a +
(c
+ ) j3'
+ (a +
^) 7'.
on the same plan as the given equation between

Treating
find
that if (as in Fig. 21) we make,
a, j3, 7) we
b'" = OB
C"' = OC a'b',
a'" = OA BC',
CV,
this
the vectors of these three new points of intersection may be exwhereof the first
pressed in either of the two following ways,
shorter,
is
but the second is, for some purposes (comp. 34, 36)
more convenient
T
'"
2a^b^c'
b^
^,^
^
Ib^c-â
cy
^ ,^ 2c a
+ + b'
or
,
_ 2a + 6j3 + C7
2 + 6 + c
r.m
+ a a^b^
, _ 2cy
^
'2c + a + b
_ 2^/3 + C7 + flg
26 + +
<?
CHAP.
And
II.]
the three equations, of which the following
is
(b- c)a"- (2b+c + a)(5"'+ {2c- a + b)y"
23
one,
0,
with the relations between their coefficients w^hich are evident

on inspection, show (by 23) that we have the three additional
coUineations, a"b"'c'", b"c"'a'", c"a"'b'", as indicated by three of
the dotted lines in the figure.
Also, because Ave have the two

^
expressions,
,
(a +
b)y'-i- (c
(a +
Z>)
+ a)j5'
(c + a)
(a + b) y
(c
a) j5'
'
( + Z>)-(c + )
'
-\^
see (by 26) that the two points a", a'" are harmonically co?ijugate with respect to e' and c' ; and similarly for the two
we
other pairs of points, b",
and with
31),
a', b':
we may
b'",
and
c", c'",
compared with c', a',

employed (25,
so that, in a notation already
write,
(bVc'a") = (c'b'Vb") =
(a'c'"b'c")
=-
1.
we
beyin^ as above, with ^wj four complanar points,

no three are collinear, we can (as in Fig.
of
which
o, A, B, c,
what
be
called a First Construction, derive from
18), by
may
34. If
them six lines, connecting them two by two, and intersecting

each other in three new points, a', b', c' and then by a Second
;
Construction (represented in Fig. 21), we may connect these

by three new lines, which will give, by their intersections with
We
the former lines, six new points, a", . d".

might proceed to connect these with each other, and with the given
.
points,
by
sixteen new" lines, or lines of a Third Construction,
namely, the four dotted lines of Fig. 21, and twelve other
lines, whereof three should be drawn from each of the four
given points
and these would be found to determine eighty-
new
points of intersection, of which some may be seen,

are not marked, in the figure.
they
although
But however far these processes of linear construction
four
may
be continued, so as to form w^hat has been called* a plane

*
By
Prof. A. F.
Mobh'S,
trische Calcul, Leipzig, 1827).
in
page 274 of his Bai-ycentric Calculus (dcr barycen-
24
[bOOK
I.
geometrical net^ the vectors of the points thus determined have

all one common property :
namely, that each can be represented
by an expression of the form,

xaa + yhfi
xa
-\-yh
+-
zcy
'
+ zc
In fact
coefficients x, y, z are some whole numbers.
see (by 27, 31, 33) that such expressions can be assigned
for the ni?ie derived vectors, a', .
y", which alone have been
where the
we
hitherto considered
and
it is
not
difficult to perceive,
from
the nature of the calculations employed, that a similar result

must hold good, for every vector subsequently deduced. But
and other connected results will become more completely

evident, and their geometrical signification will be better unthis
somewhat closer consideration of anharmonic

and the introduction of a certain system of anharmonic co-ordinates, for points and lines in one plane, to which
derstood, after a
quotients,
we
shall
next proceed
reserving, for a subsequent Chapter,

to space.
any applications of the same theory
Section
4.
On Anharmonic
Co-ordinates and Equations of
Points and Lines in one Plane,

35. If we compare the last equations of Art. 33 with the
corresponding equations of Art. 31, we see that the harmowc
group ba'ca", on the side bc of the triangle abc in Fig. 21,
has been simply reflected into another such group, bV'c'a", on

b'c', by a harmonic pencil of four rays, all passing
through the point o ; and similarly for the other groups.
the line
More
ob, oc, od, or briefly
generally, let oa,
o.abcd, be
any pencil, with the point o for vertex ; and let the new ray
OD be cut, as in Fig. 22, by the three sides of the triangle
ABC, in the three points
Ai, Bj, Ci
oAi =
so that (by 25)
we
shall
(ba'cai)
ai
let also
+
= yh^ zcy
^^
yo + zc
'-,
have the anharmonic quotients,

=
y
-,
(ca'sai")
z
;
CHAP.
II.]
2-5
us seek to express the two other vectors of intersection, j3i and ^71, with a view to
determining the anharmonic ra-
and
let
tios
of the groups on the two
other sides.
The given
equation
(27),
shows us at once that these two

vectors are,
{y-z)c + ya
{z-y)b
^-^zaa
T
.
_
=
oc -^
oci
yi
\z-y)h + za
whence we derive (by 25) these two other anharmonics,

(cb'aBi)
so that
we have
(BCACi) =
2-- W
-,
the relations,
= (bc'aCi) + (ba'cAj) = I.
(cb'aBi) + (ca'bai)
But
not
in general, for
difficult to
any four coUinear points

AB
BC
whence by the
a, b, c, d, it is
prove that
CD+
AC
CB
BD=
da;
definition (25) of the signification of the
bol (abcd), the following identity
is
sym-
derived,
(abcd) + (acbd)=
I.
Comparing this, then, Avith the recently found relations,

have, for Fig. 22, the following anharmonic equations :
(cab'bi) = (ca'bai) =
we
y
(bAC'Ci)
= (ba'cAi) = y
and we see that (as was to be expected from known
princi-
KLEMKNTS OF QUATERNIONS.
26
[bOOK
I.
pies) the anharmonic of the group does not change, when we

pass from one side of the triangle, considered as a transversal
of the pencil, to another such side, or transversal.
may
We
therefore speak (as usual) of such an anharmonic of a group^

as being at the same time the Anharmonic of a Pencil ; and,
with attention to the order of the rays, and to the definition

(25), may denote the two last anharmonics by the two following
reciprocal expressions
(o.cabd) = -;
(o.bacd) = ^;
^
with other resulting values, when the order of the rays

changed it being understood that
is
(o
if
cabd) =
(c'a'b'd'),
the rays oc, oa, ob, od be cut, in the points
by any one right

36.
The
c', a', b',
d\
line.
expression (34),
xaa + yb^ + zcy
xa^yh^zc
may represent the
vector o^ any point p in the given plane by a

^
suitable choice of the coefficients ic, y, x, or simply of their ratios.

For since (by 22) the three complanar vectors pa, pb,
PC must be connected by some

a'
pa +
6'
linear equation, of the
pb +
c'
PC =
0,
form
CHAP.
11.]
Pi
we
rOlNTS AND LINKS IN A GIVEN PLANE.
= PABC,
P2
= PB-CA,
27
= PCAB,
P3
with the same coefficients xyz, the following
shall have,
the vectors oPi, 0P2,

expressions for
of these three points
or
0P3,
pi, |02, /03,
of intersection,
Pi, P2} P3
_ zcy + xaa
yh^ + zcy
^
yb+zc
xaa + ybj3
xa + yb
^^
which give
at
once the following anharmonics of pencils, or of
groups,
(A
(B
V
BOCP) = (ba CPi) =
COAP) = (cBAPa) = X
(C
AOBP) = (AC'BPa) =
whereof we see that the product
is unity.
Any two of these
three pencils suffice to determine the position of the point p,
when the triangle abc, and the origin o are given ; and there-
appears that the three coefficients x, y, z, or any scalar s

proportional to them, of which the 5'?<o^2>wâ- thus represent the
fore
it
anharmonics of those pencils,
may be
conveniently called the
An HARMONIC
Co-ordinates of that pointy p, with respect to

the given triangle and origin : while the point p itself may be
denoted by the Symbol,
v = (x,y, z).
With
this notation, the thirteen points of Fig. 2 1
come
to be
thus symbolized:
a =(1,0,0),
B =(0,1,0),
a'
B'
=(0,0,1),
=(0,1,1),
=(1,0,1),
'=(1,1,0);
a" = (0, 1,-1), B" = (-1,0, 1), c"=(l,-l,0);
a"'=(2,
1,
1),
b"'= (1,2,1),
'"=(1,1,2).
(1,1,1);
28
two points
37- If Pi and p^ be any

Pi
and
if t
= (^1,
!/l,
P2
Zl),
and u be any two
[bOOK
I.
in the given plane,
G'^2J
^2, Z2),
scalar coefficients, then the following
third point,
p=
(tXi
uXiy
tj/i
+ uy2>
tzx
+ uz^,
with the two former points, or (in other words)

For, if we make
is collinear
is
situated on the right line PiPg.
x=
tXi-\-
UX2,
y=
tifx
+ w^2)
z^tz^^
UZ2,
and
xâa +
Xitt
xâa +
xâ +
xaa +
xa +
these vectors of the three points P1P2P are connected by the

linear equation,
{xâ +
.)
^1 +
w {xâ +
.)/02
(ara
.) |0
which (comp. 23), the sum of the coefficients is zero. Conversely, the point p cannot be collinear with Pi, P2, unless its
in
co-ordinates admit of being thus expressed in terms of theirs.

It follows that if a variable point p be obliged to move along a
given right line P1P2, or if it have such a line (in the given
co-ordinates xyz must satisfy a homogeneous equation of the first degree^ with constant coefficients ;
which, in the known notation of determinants, may be thus
plane) for
its locus, its
written,
or,
more
X,
g,
î>
yij
Zi
Xz,
^2?
2n
fully,
=x
(giZ2
z.y^)
+y
{zix^
^1^2)
+ z
{x.y^
yiX2)
or briefly,
where
l.v
my
+ nzy
Z, m, n are three constant scalars, whereof the quotients

determine the position of the right line A, which is thus the
locus of the point p.
It is natural to call the equation^ which
CHAP.
II.]
29
thus connects the co-ordinates of the j)oint p, the Anharmonic

Equation of the Line A; and we shall find it convenient also
to speak of the coefficients
the Anharmonic Co-ordinates

also
in that equation, as being

of that Line : which line may
tw, w,
/,
be denoted by the Symbol,
A=
m, w].
[/,
For example, the three
38.
sides bc, ca,
ab of the given
triangle have thus for their equations,

a;
y=
0,
2-0,
0,
and for their symbols,
The
[0,0,1].
[0,1,0],
[1,0,0],
three additional lines oa, ob, oc, of Fig. 18, have, in like
manner, for their equations and symbols,
z-x = 0,
y-z = 0,
[0,1,-1],
The
x-y=0,
[-1,0,1],
[1,-1,0].
lines b'c'a", c'a'b", ab'c", of Fig. 21, are
y-\-z-x =
zn-x-y =
0,
0,
x + y -z=0,
or
[-1,1,1],
[1,-1,1],
[1,1,-1];
a"b"'c"', b"c"'a'", c"a"'b''', of the same figure,

manner represented by the equations and symbols,
the lines
y+
z-3x
==
z+
0,
[-3,1,1],
and the
x-3y=0,
[1,-3,1],
are in like
x-\^y-3z =
0,
[1,1,-3];
line a"b "c" is
-\-
y + z ==
0,
or
[1, 1, 1].
Finally, we may remark that on the same plan, the equation

and the symbol of what is often called the line at infinity, or
of the locus of all the irifinitely distantpoints in the given plane,
are respectively,
ax \ by \ cz =
0,
and
\a, b, c]
30
[bOOK
I.
because the linear function^ ax^hy-\- cz^ of the co-ordinates

-c, z/, z of a point p in the plane, is the denominator of the expression (34, 36) for the vector p of that point: so that the
point p is at an infinite distance from the origin o, when, and
only when, this linear function vanishes.
These anharmonic co-ordinates of a
39.
line, although
above interpreted (37) with reference to the equation of that
considered as connecting the co-ordinates of a variable
line,
point thereof, are capable of receiving an independent geometrical interpretation.

For the three points l, m, n, in which
& + my + nz = 0, intersects the three
the line A, or
[/, tw,
sides BC, CA,
ab of the given
x=
lines
w], or
z=0
i/=0,
36) be thus denoted
0,
L = (0,
n,-m)
But we had
A"
evidently (on the plan of
may
m=
(- w, 0,
/)
n = (m, -
/,
0)..
also (by 36),
= (0,1,-1);
whence
(38),
triangle abc, or the three given
it is
c"
b"=(- 1,0,1);
= (l,-l,0);
easy to infer, on the principles of recent articles,
that
= (ba"cl)
m
- = (cb"am)
^
^
= (ac'bn)
^
with the resulting relation,

(ba"cl)
(cb"am)
(ac"bn) =
1.
40. Conversely, this last equation is easily proved, with

the help of the known and general relation between segments
(32), applied to any two transversals, a"b"c" and lmn, of any
triangle abc.
In
fact,
we have
thus the two equations,
bl cm an
LC MA NB
ba" cb" ac"

a"c
'
b"a c"b
on dividing the former of which by the

of the last article results.
have been
led, without
We
latter,
'
the last formula
might therefore
in this
way
any consideration of a variable point
p.
CHAP.
II.]
to introduce three auxiliary scalars,

71
their quotients
TYl
-,
^^QQ^?!
w,
/,
w, defined as
31
having
respectively, as in 39, to the
three anharmonics of groups,
(cb"am),
(ba"cl),
(ac"bn)
and then it would have been evident that these three scalars,
m, n (or any others proportional thereto), are suflScient to
/,
determine the position of the right line A, or lmn, considered

as a transversal of the given triangle abc
so that they might
:
naturally have been called, on this account, as above, the anharmonic co-ordinates of that line. But although the anharmonic co-ordinates of a point and of a line may thus be inde-
pendently defined^ yet the geometrical utility of such definitions

will be found to depend mainly on their combination : or on the
formula
nz =
of 37, which may at pleasure be considered as expressing, either that the variable point {x, y, z) is
situated somewhere upon the given right line [/, m, ri] ; or else
Ix +
my +
in some direction^ through

[/, m, 7i\ passes,
the given point (ar, y, z).
41. If Ai and Ag be any two right lines in the given plane,
that the variable line
Ai =
then any third right line

through the intersection
curs with
them
A2 =
mi, ?2j,
[Zi,
[/2) '^2,
W2],
same plane, which passes

(in other words) which con^
in the
ArA2, or
(at a finite or infinite distance),
may be
repre-
sented (comp. 37) by a symbol of the form,
A=
\tli
+UI2, tmi + w?W2, tni +
U7I2'],
where t and u are scalar coefficients. Or, what comes to the

same thing, if Z, m, n be the anharmonic co-ordinates of the
line A,
then (comp. again 37), the equation
0= Z(wiW2-
=
^1^2) + &c.
/,
m,
/i,
mi, Ui
4, ^2, W2
must be
satisfied
because, if (X, F, Z) be the supposed point
commo?i to the three lines, the three equations
;
32
IX+mY+nZÔ,
I.
LX ^ m,Y^n,Z =(i,
m,Y+n^Z =0,
l,X +
[bOOK
must co-exist. Conversely, this coexistence will be possible,

and the three lines will have a common point (which may be
infinitely distant), if the recent condition
For example, because [a^h^
tisfied.
c^
of concurrence be
sa-
has been seen (in 38)
to be the symbol of the line at infinity (at least if we still retain the same significations of the scalars a, 6, c as in articles
27, &c.),
it
follows that
[/,
m,
lines,
and
A'
two parallel
[/
+ w,
m + ub,
w +
uc']
because they concur at infiproblems respecting intersections of right

collineations of points, &c., in the given plane, when
are symbols of
nity.
ri]
In general,
lines
all
by this anharmonic method^ conduct to easy eliminabetween linear equations (of the scalar kind), on which
treated
tions
we need not here delay the mechanism of such calculations

being for the most part the same as in the known method of
trilinear co-ordinates: although (as we have seen) \hQ geome:
trical interpretations are altogether different.
Section
42. If we
5.
On Plane Geometrical Nets, resumed.
now
resume, for a moment, the consideration of
those plane geometrical nets, which were mentioned in Art. 34

and agree to call those points and lines, in the given plane, ra;
tional points
and rational
lines, respectively,
which have their
anharmonic co-ordinates equal (or proportional) to whole numbers ; because then the anharmonic quotients, which were discussed in the last Section, are rational ; but to say that a point
or line is irrational, or that it is irrationally related to the
given system of four
initial points o, a, b, c,
monic co-ordinates are not thus
all
when
its
anhar-
equal (or proportional) to
it is clear that whatever four points we may assume

and however far the construction of the net may be
the net-points and net-lines which result will all be ra-
integers ;
as initial,
carried,
tional, in the sense just
now
such; and the subsequent
defined.
cZe we waêow^
In
fact,
we
begin with
(41) can never after-
CHAP.
33
PLANE GEOMETRICAL NETS.
II.]
wards conduct to any, that are of the contrary kind the right
which connects two rational points being always a rational
and the point of intersection of two rational lines being
line
The assertion made in Art. 34
necessarily a rational point.
:
line
therefore fully justified.

43. Conversely, every rational point of the given plane,
with respect to the four assumed initial points oabc, is a point
of the net which those four points determine. To prove this,
is
it is
Ai =
show that every rational point

on any one side bc of the given triangle abc, can
evidently sufficient to
(0, y, ^),
be so constructed.
Bi
we have (by
c'Bi
= OAi-CA,
as in Fig. 22,
and
= oai*ab,
Ci
35, 36) the expressions,
Bi
from which
Making,
it is
BC =
(y,
0,y-z),
Ci=(2:, 2r-?/, 0);
easy to infer (by 36, 37), that

(0,
y,z- y),
b'Ci
bc =
(0,
z/
z, z)
and thus we can reduce the linear construction of the rational

point (0, y, z), in which the two whole numbers 7/ and z may
be supposed to be prime to each other, to depend on that of
the point (0, 1, 1), which has already been constructed as a'.
It follows that although no irrational point
q of the plane ca?^
be a net-point, yet every such point can be indefinitely approached

to, by continuing the linear construction;
so that
it
can be included within a quadrilaeven within a tri-
teral interstice PiPgPaPi, or
angular interstice P1P2P3, which interstice of

the net can be made as small as we may de-
Analogous remarks apply to irrational

the plane, which can never coincide
with net-lines, but may always be indefinitely approximated io
by such.
44. If p, Pi, P2 be any three colUnear points of the
net, so
that the formulae of 37 apply, and if p'be
any ^wrM net-point
{x\ y\ z) upon the same line, then writing
sire.
lines in
xâ + yj) + z^c =
xâ + yJb + z^c =
ri,
Vg,
34
we
shall
I.
have two expressions of the forms,

^
tVipi
tVi
in
[bOOK
which the
+ UVzpz
'
+ UVz
t'Vipi
fvi
+ W V2/O2
'
u'vi
coefficients tut'u' are rational, because the co-or-
dinates xyz, &c., are such, whatever the constants abc may be.
have therefore (by 25) the following rational expression
We
for the anharmonic of this net-groiip
\ *
"
1
"2"
^^
'f
tu
{xyi
r-
yxi) {x y^-yx^)
and similarly for every other group of the same kind. Hence
eiêry group of four collinear net-points, and consequently also
every pencil of four concurrent net-lines, has a rational value for

anharmonic function ; which value depends only on the processes of linear construction employed, in arriving at that group
its
or pencil, and
is
quite independent of the configuration or arinitial points : because the three initial
rangement of the four

constants, ,
It
was thus
disappear from the expression which results.

the nine pencils, which had the
5, c,
that, in Fig. 2 1
d" for their vertices, were all harmonic pencils, in whatever manner the four points o, a, b, c
might be arranged. In general, it may be said that plane
nine derived points
a'
geometrical nets are all homographic figures-* and conversely,

in any two such plane^^wre^, corresponding points may be con-
sidered as either coinciding, or at least (by 43) as indefinitely

approaching to coincidence, with similarly constructed points
of two plane nets
that is, with points of which (in their reanharmonic co-ordinates (36) are equal
the
spective systems)
:
integers.
45.
Without entering heref on any general theory of trans-
formation of anharmonic co-ordinates,

if
we
select
any four
we may
already see that

which no three
net-points Oi, Ai, Bi, Ci, of
are collinear, every other point p of the same net

related (42) to these ; because (by 44) the three
*
Compare the Geometrie
Siiperieure of
M.
f See Note A, on Anharmonic Co-ordinates.
is
rationally
new
Chasles, p. 362.
anhar-
CHAP.
CURVES
II.]
monies of pencils,
(Ai
IN A
GIVEN PLANE.
BiOiCiP)
35
&e., are rational
and
therefore (comp. 36) the new co-ordinates Xi, yi, z^ of the point
p, as well its old co-ordinates xyz, are equal or proportional to
wliole numbers.
It follows (by 43) that every point p of the
net can be linearly constructed, if any four such points be

given (no three being collinear, as above) or, in other words,
;
that the whole net can be reconstructed,* if any one of
its
qua-
drilaterals (such as the interstice in Fig. 24) be known.
As
an example, we may suppose that the four points oa'b'c' in

Fig. 21 are given, and that it is required to recovj^r from them
the three points abc, which had previously been among the
data of the construction. For this purpose, it is only necessary to determine first the three auxiliary points a"', b'", d", as
the intersections oa'
b'c',
&c.
liary points a",
mulae, A = b'b"
c'c", &c., will
c", as b'c'
",
and next the three other auxiafter which the forb"'c"', &c.
;
enable us to return, as required,
to the points a, b, c, as intersections of
Section
6.
On
known
right lines.
Anharmonic Equations, and Vector Expres-
sions,
for Curves in a given Plane,
When, in the expressions 34 or 36 for a variable vec=

tor p
OP, the three variable scalars (or anharmonic co-ordinates) X, y, z are connected by any given algebraic equation,
46.
such as
fp{x,y, z)
0,
supposed to be rational and integral, and homogeneous of the

degree, then the locus of the term v (Art. 1) of that vector
p^
is
a plane curve of the
jo^''
order; because (comp. 37)
it is
cut
* This theorem
(45) of the possible reconstruction of a plane net, from any one
quadrilaterals, and the theorem (43) respecting the possibility of indefiapproaching by net-lines to the points above called irrational (42), without
ever reaching such points by any processes of linear construction of the kind here
of
its
nitely
considered, have been taken, as regards their substance (although investigated
by a
from that highly original treatise of Mobius, which was
referred to in a former note (p. 23). Compare Note B, upon the Barycentric Calculus ; and the remarks in the
following Chapter, upon nets in space.
totally different analysis),
36
in
p points
any given right
lx-^my + nz==
line^
For example,
if
we
^
where
t,
and
(distinct or coincident,
v are three
w,
we
are
by
write
_ faa + u'b^ + uVy
a^v^hTv^'
new
^=
= ^%
real or imaginary),
I.
0, in the given plane.
variable scalars, of
suppose that the sum is zero, then,

tween the four equations,
JC
[boOK
<
shall
eliminating these be-
by
2=172,
M^,
which we
+ M+
?;
0,
conducted to the following equation of the second
degree,
^ ^^^ ^ ^^
^f^z'-
2yz
22a;
2^^
= 2, and the locus of p is a conic section. In fact,

/>
the conic which touches the sides of the given triangle abc,
so that here
it is
at the points above called a',

tions with the side bc, by
quadratic with
that there
is
b', c'
for if
making
a;
we
seek
(38),
its intersec-
we
obtain a
namely, {y-zy = Q\ which shows
e^'waZ roots,
contact with this side at the point (0,
1, I),
or a'
(36) : and similarly for the two other sides.

47. If the point o, in which the three right lines a a', bb^,
cc' concur, be (as in Fig. 18, &c.) interior to the triangle abc,
the sides of that triangle are then all cut internally, by the
points a', b', c' of contact with the conic so that in this case
;
(by 28) the ratios of the constants , b, c are all positive, and
the denominator of the recent expression (46) for p cannot va^
nish, for any real values of the va-
and conset, u, v
can
render
such
no
values
quently
riable scalars
',
The
infinite that vector p.
conic is
therefore generally in this case, as in

Fig. 25, an inscribed ellipse ; which
becomes however the inscribed

cle,
cir-
when
a'M 6"M
c'^
-a
-b
-c
^^'
^'^*
a, b, c denoting here the lengths of
the sides of the triangle, and s being their semi-sum.
CHAP.
ANHARMONIC EQUATIONS OF PLANE CURVES.
II.J
But
48.
if
37
the point of concourse o be exterior to the
tri-
angle of tally ents abc, so that two of its sides are cut externally,
then tico of the three ratios oiseyments (28) are neyative ; and
therefore one of the three constants a,
<
0,
but each of the two others as >
6,
0.
c may be treated as
Thus if we suppose
that
oO,
i>0,
a' will
a+oO,
a + 5>0,
fl<0,
be a point on the side b itself but the points b',

on the lines ac, ab, aa' prolonyed, as in Fig. 26
then the conic

ellipse
circle),
o
and
c',
will be.
be an
a'b'c' will
(including the case of a

or a parabola, or an hy-
perbola^ according as the roots of
jj
the quadratic,
{a
^c)f +
2ctu + (6 +
c)?/2
0,
b^
obtained by equating the denominator (46) of the vector p to

zero, are either, 1st, imayinary
Ilird, real
and unequal: that

+ ab>
bc-\-ca
-!
For example,
circle,
the
6"^
c-i
if the conic
known
<
5"^
or
c"^
or <
=0,
we have
;
here negative), according as

or
0,
be what
ratios of
a"^
is
according as
is,
0,
or (because the product abc
or Ilnd, real and equal) or
or > 0.
=0,
is
often called the exscribed
segments give the proportion,
=-
-c
-b
and
-s + s-c + s-b<0.
49. More generally, if c^ be (as in Fig. 26) a point upon
the side ab, or on that side prolonged, such that cc^ is parallel
to the chord b'c', then
c'
Ac'
= cb' ab' = -
:
c,
and
ab
ac'
=a+^ 6
:
writing then the condition (48) of ellipticity (or circularity)
38
under the form,
< r, we
[bOOK
see that the conic
is
an
I.
ellipse,
< or = or > ab ; the

parabola, or hyperbola, according as c^c'
in
other
still,
respects, that which is reprearrangement being
to
sented in Fig. 26.
Or,
express the same thing more symmetrically, if we complete the parallelogram cabd, then according as the point d falls, 1st, beyond the chord 'b'c', with
respect to the point a; or llnd, on that chord; or Ilird,

imthin the triangle ab'c', the general arrangement of the same
Figure being retained, the curve is elliptic^ or parabolic, or
which
hyperbolic. In that other arrangement or configuration,
a + 5+c<0,
answers to the system of inequalities, 6>0,
oO,
upon the side bc itself, but o is on the

a and then the inequality,
through
prolonged
the point
a'a
a' is still
line
(6 +
c)
shows that the conic
is
+ 6c < -
(^>2
+ 5c +
c")
<
0,
necessarily an hyperbola
whereof it
is
is touched by the side bc at. a,

touched in b' and c', by the sides
easily seen that one branch
while the other branch
is
CA and BA prolonged through
a.
The curve
is
either + 6 or a + c be negative, while b

tive as before.
lic, if
50.
When
the quadratic (48) has
equal, so that the conic
is
its
also
and
hyperbo-
c are posi-
roots real
and un-
an hyperbola, then the directions of
the asymptotes may be found, by substituting those roots,

or the values of t, u, v which correspond to them (or any
scalars proportional thereto), in the numerator of the expression (46) for |0
and similarly we can find the direction of the
axis of the parabola, for the case when the roots are real but
;
equal
for
we
which a right
shall thus obtain the directions, or direction, in
line
conic at infinity.
op must be drawn from o, so as to meet the

And the same conditions as before, for dis-
tinguishing the species of the conic, maybe otherwise obtained

by combining the anharmonic equation,
(46), of that
/=
conic, with the corresponding equation ax + by + cz=0 (38) of

the line at infinity ; so as to inquire (on known principles of
modern geometry) whether that line meets that curve in tivo
CHAP.
II.]
DIFFERENTIALS
TANGENTS POLARS.
39
it, or cuts it, in points which (alhere to be considered as real,

are
though infinitely distant)
51. In general, if/"(a;, y, z)-(i be the anharmonic equa-
imaginary points, or touches
tion (46) of any plane curve^ considered as the locus of a variable point p ; and if the differential* of this equation be thus
denoted,
X^x + Ydy + Zdr
d/(rc, y, z)
then because, by the supposed homogeneity (46) of the function/,
we have
the relation
Xx-^Yy
we
shall
have
also this other
+ Zz=^
0,
but analogous
relation,
if
X that
X',y'
\z - z = diXidiy\^Z]
-y
(by the principles of Art. 37),
is
point upon
the tangent
the
to
if p'=(a7',
y\
drawn
curve,
2')
be any
at the point
p=
and regarded as the limit of a secant. The sym(ic, y, 2;),

bol (37) of this tangent at p may therefore be thus written,
[X, F, Z\,
where
or
known
Da, d^, d^ are
[d,/ d,/, d,/]

characteristics
ofpartial deriva-
tion.
52. For example, when /has the form assigned in 46, as answering to the conic lately considered, we have Da-/= 2{x-y-z)y
&c. ; whence the tangent at any point {x, y, z) of this curve
may be denoted by the symbol,
\_x-y-z,
y-z-x,
z-x-y']\
in which, as usual, the co-ordinates of the line
by
any others proportional to them.
which
Thus
may be replaced
at the point
a',
or
evidently (by the form of/) a

the
the
curve,
point upon
tangent is the line [- 2, 0, 0], or
that
is
side bc of the given triangle, as
the
;
[1, 0, 0]
(by 38),
(by 36) at (0, 1,1),
is
* In the
theory of quaternions, as distinguished from (although including) that
it will be found necessary to introduce a new definition
of differentials, on
account of the non- commutative property o^ quaternion-multiplication : but, for the
of vectors,
present, the usual significations of the signs
d and d are
sufficient.
40
[bOOK
I.
was otherwise found before (46). And in general it is easy to

see that the recent symbol denotes the right line, which is (in
a well
known
spect to the
resense) the polar of the point (a;, z/, z), with

or that the line [X', Y\ Z'~\ is
same given conic
the polar of the point {x\ y\ z)
because the equation
Xa!-\-Yy' + Zz=^0,
which
may be
for a conic
X'x + Y'y + Z'z =
written as
0,
expresses (by 51) the condition requisite, in order that a point

(.r, 2/, z) of the curve* should belong to a tangent which passes
through the point {x\ y\ zf). Conversely, the point (x, y, z)

is (in the same well-known sense) the poZ^ of the line [X, Y, Z'\
so that the centre of the conic,
which
ihe pole of the line at infinity (38),

the conditions a~^X = b-^Y=c~^Z \
K = (6 +
sent conic, the point

vector
aa +
OK
bj5
is
the point which
satisfies
therefore, for the pre-
it is
c,
(by known principles)
is
a+
of which the
b),
easily reduced, by the help of the linear equation,

+ cy = (27), to the form,
is
â +
+ C-y
ca-^ ab)
^2j3
2{bc+
'
with the verification that the denominator vanishes, by 48,

when the conic is a parabola. In the more general case, when
this
denominator
is
different
from zero,
it
can be shown that
every chord of the curve, which is drawn through the extremity

K of the vector k, is bisected at that point k which point
:
would therefore
way be
seen again to be the centre.

53. Instead of the inscribed conic (46), which has been the
in this
subject of recent articles, we may, as another example, consider that exscribed (or circumscribed) conic, which passes
through the three corners a, b, c of the given triangle, and
touches there the lines a a", bb", cc" of Fig. 21. The anharis easily seen to be,
monic equation of this new conic

yz-^
* If the curve
/=
zx^ xy =
were of a degree
tions above written would represent
the line-polar, of the point
(x',
y\
what
Jiigher
0',
than the second^ then the two equa-
are called
2'), witli
the, first
polar,
and the
respect to the given curve.
last or
CHAP.
41
VECTOR OF A CUBIC CURVE.
II.]
the vector of a variable point p of the curve
may
therefore be
expressed as follows,
t'âa + u~^bf5
with the condition

centre k'
is
u^v =
+ v^cy
The
0, as before.
vector of
its
found to be,
^2
j^
^2
_ 2hc - lea - lab
c2
'
and it is an ellipse, a parabola, or an hyperbola, according as

the denominator of this last expression is negative, or null, or
And because these two recent vectors^ jc, k, bear a
positive.
scalar ratio to each other,
it follows
(by 19) that the three
are collinear ; or in other words, that the line
of centres kk', of the two conies here considered, passes through
the point of concourse o of the three lines aa', bb', cc'.
More
= [/, m, w]
generally, if l be the pole of any given right line A
points o, K, k'
(37), with respect to the inscribed conic (46), and if l' be the
pole of the same line A with respect to the exscribed conic of
the present article, it can be shown that the vectors ol, ol', or
of these two poles are of the forms,
X, X',
X = ^ {laa + mb^ +
where k and
k'
X'
wc'y),
are scalar s
k'
(laa + mb[5 + wcy),
the three points o,
l, l'
are there-
fore ranged on one right line.

54. As an example of a vector-expression for a curve of
order higher than the second^ the following
OP=p =
with
+ w +r=
find here
tâa + u^b^
0, as before.
by elimination of f,
tâ
is
an
+ v^cy
u% + v^c
Making
a:
= <^ y =
uî
z = r^
we
w, v the anharmonic equation,
{x^y\ zfthe locus of the point p
may
be taken
2*Jxyz
therefore, in this example, a curve
the third order, or briefly a cubic curve.
The mechanism
of
(41)
42
[book
of calculations with anharmonic co-ordinates
is
so
much
I.
the
same as that of the known trilinear method, that it may suffice

to remark briefly here that the sides of the given triangle abc
are the three (real) tangents of injiexion ; the points ofinflexion
being those which are marked as a", b", c" in Fig. 21 ; and the
*
If fl=6 =c, in which
origin of vectors o being a conjugate point
in
o
becomes
case (by 29) this origin
(as
Fig. 19) the mean
point of the triangle, the chord of
inflexion
then the
infinity^
curve
a"b"c"
is
line
at
and the
takes
the
form represented
in Fig. 27;
ing three
hav-
Fig. 27.
irifinite
branches^ inscribed within the angles vertically opposite to

those of the given triangle abc, of which the sides are the
three asymptotes.
55. It would be improper to enter here into any details of

discussion of such cubic curves, for which the reader will na-
But
turally turn to other works. f
it
may be
remarked, in
passing, that because the general cubic may be represented, on

the present plan, by combining the general expression of Art.
34 or 36 for the vector
p,
with the scalar equation
= 27 kxyz,
s^
where
=^+
7/
2*
k denoting an arbitrary constant, which becomes equal to

unity, when the origin is (as in 54) a conjugate point; it fol-
lows that
if
p=
(x, y, z)
of the curve ^ and
if
and
we make
p'
(x', y',
s' =ôc
z) be any two points

shall have the
+y +z, we
relation,
xyzs
*
Answering
= xyz s^,
to the values
cube-roots of unity
xs
or
t,
zs
sy'' sz'~
^=1, u=9, =
which values of
ys
.
;
, v
give
02^
where
is
one of the imaginary
x = y=^z, and p =
0.
t Especially the excellent Treatise on Higher Plane Curves, by the Rev. George
Salmon, F.T.C.D., (Src. Dublin, 1852.
CHAP.
ANHARMONIC PROPERTY OF CUBIC CURVES.
II.]
in -vvliich it is
not
difficult to
= (a'.pbpb');
sx
'
43
prove that
= (c.papa);
sz
^=(b.pcpc);
^
^
sy'
the notation (35) of anharmonics of pencils being retained.

"
obtain therefore thus the following Theorem :
If the
We
of any given plane* triangle abc he cut (as in Fig. 2\)hy

any given rectilinear transversal a"b"c", and if any two points
p and p' in its plane he such as to satisfy the anharmonic rela-
sides
tion
(a'.pbp'b")
(b". pcp'c")
(c". pap'a")
1,
then these tioo points p, p' are on one common cuhic curve, which
has the three collinear points a", b", c" for its three real points
of inflexion, and has the sides bc, ca, ab of the triangle for its
three tangents at those points ;" a result which seems to offer
a new geometrical generation for curves of the third order,
5Q, Whatever the order of a plane curve may be, or whatever may be the degree p of ihQfunctionfm 46, we saw in 51
that the tangent to the curve at any point p = {x,
?/,
z)
is
the
right line
if
w, w],
[/,
Z= D^/,
m = Dyf
n = d^^;
by the supposed homogeneity o^f give the

\my^nz^ 0, and therefore enable us to establish
expressions which,
relation, Ix
the system of the two following diiferential equations,
m^y + wds = 0,
lAx +
x^l +
y^m + z^n z^
we arrive at a new
y, z
which render the
If then, by elimination of the ratios of re, y,

homogeneous equation of the form,
as one that
function /=
is
true for
all
values of
ic,
0.
may require to be cleared offactors,

elimination), we shall have the equation
(although it
introduced by this
f(/,
* This Theorem
may be
m, n) =
0,
extended, with scarcely any modification, from plane to
spherical curves, of the third order.
44
[bOOK
I.
must be satisfied by the tangent A to the

curve, in all the positions which can he assumed by that right line.
as a condition that
And, by
connparing the
dr (/,
we
:
n)
we may
see that
X y z=
772,
DiF
diF
two
differential equations,
icd/
0,
write the
df,
+ ydm + 2d7i =
0,
pjôportioji,
and the symbol
p = (d/f,
d,f, df),
above, the point of contact p of the variable

in
m, w],
any one of its positions, with the curve which
if (x, y, z) be, as
line
\_l,
Hence we can
pass (or return) from the tan=

f
0, of a curve considered as the envelope of
gential equation
a right line A, to the local equation f= 0, of the same curve
is its
envelope.
considered (as in 46) as the locus of a point p : since, if we obtain, by elimination of the ratios of /, m, n, an equation of the
form
0=/(DiF,
D,F,
df),
be necessary, of foreign factors) as a conseof

the
quence
homogeneous equation f = 0, we have only to
substitute for these partial derivatives, D/F, &c., the anhar(cleared, if it
monic co-ordinates
And when
x, y, z,
to
which they are proportional.
the functions /"and f are not only homogeneous (as
always suppose them to be), but also rational and

integral (which it is sometimes convenient not to assume them
as being), then, while the degree of the function^ or of the
wê shall
local equation, marks (as before) the order of the curve, the
degree of the other homogeneous function f, or of the tangential
= 0,
equation f
is
easily seen to denote, in this anharmonic
from the analogy of other and older methods, it
might have been expected to do), the class of the curve to
which that equation belongs or the number of tangents (dis-
method
(as,
and real or imaginary), which can be drawn

that curve, from an arbitrary point in its plane.
tinct or coincident,
to
57. As an example (comp. 52), if we eliminate x, y, z between the equations,

m = y-z- X, n^z-x-y^ Ix^my -^ nz=^,
l = x-y-z,
where /, m, n are the co-ordinates of the tangent
to the inscribed
CHAP.
45
LOCAL AND TANGENTIAL EQUATIONS.
II.]
we are conducted to the following tangenof

that
conic, or cvrve of the second class,
equation
conic of Art. 46,

tial
=
F(l,m^n) = mni-nl+lm 0;
with the verification that the sides [1, 0, 0], &c. (38), of the
this equation.
triangle abc are among the lines which satisfy
Conversely, if this tangential equation were given, we might
(by 6Q) derive from it expressions for the co-ordinates of contact X, y, z, as follows
X = BiF =
m + n,
y=
z=l-m;
n-vl,
[1, 0, 0] touches the conic,

in
the
envelope,
point (0, 1, 1), or a', as
with the verification that the side

considered
before
now as an
and then, by eliminating
back to the local equation, f=
I,
m,
n,
In
of 46.
0,
we should be brought
like
manner, from
the local equation /=yz + zx + xy = of the exscribed conic (53),

we can derive by diiferentiation the tangential co-ordinates,'^
I
iijcf=
y+
rn
z,
n=X+
z-i X,
y,
and so obtain by elimination the tangential equation, namely,

f(/,
m, n) =
l''
+ m^ + n^-
2mn -
2nl
from which we could in turn deduce the

(comp. 40), the very simple formula
Ix
-2lm =
0',
local equation.
And
my +nz = 0,
which we have so often had occasion to employ, as connecting

two sets of anharmonic co-ordinates, may not only be considered (as in 37) as the local equation of a given right line A,
along which a point p moves, but also as the tangential equation of a given point, round which a right line turns :
according
we suppose the set I, m, n, or the set x, z/, z, to be given.

Thus, while the right line a"b"c", or [1, 1, 1], of Fig. 21, was
as
* This name of "

tangential
by Dr. Booth
appears to have been first introduced

which the author of the present Elebut the systeyn of Dr. Booth was entirely
co-ordinates'''
in a Tract published in 1840, to
ments cannot now more particularly refer

different from his own.
See the reference in Salmon's Higher Plane Curves^ note to
page 16.
:
46
[bOOK
I.
represented in 38 by the equation x + y + z = Q, the point o of

the same figure, or the point (1, 1, 1), may be represented by
the analogous equation^
l+m + n =
because the co-ordinates
I,
must
7n,
0;
n of every
line,
which passes
satisfy this equation of the first de-
through
this
gree, as
may be
we may
write the local equation of the inscribed conic (46) as
point o,
seen exemplified, in the same Art. 38, by the
lines OA, ob, oc.
58. To give an instance or two of the use of forms, which,
although homogeneous, are yet not rational and integral (56),
follows
xh
+ yh +
zl
= 0;
and then (suppressing the common numerical
factor i), the
partial derivatives are

/
m = y^,
= x%
n=z'h\
so that a form of the tangential equation for this conic

/-^
+ nr^ + w"^ =
is,
Avhich evidently, when cleared of fractions, agrees with the first

with the verification (48), that
form of the last Article
when the curve is a parabola ; that is, when
a~^ + h~^ + c-^ =
:
it is
touched (50) by the line at infinity (38). For the exwe may write the local equation thus,
scribed conic (53),
.27"^
whence
it is
+ ^^ +
allowed to write
Z=a;"%
jt^
also,
rn^y"^,
n=
2r-,
and
lh
m^+nh=0
;'
a form of the tangential equation which, when cleared of radiAnd it is evident that we could
cals, agrees again with 57.
return, with equal ease, from these tangential to these local
equations.
59 For the cubic curve with a conjugate point (54), the

local equation may be thus written,*
*
Compare Salmon's Higher Plane Curves, page 172.
CHAP.
LOCAL AND TANGENTIAL EQUATIONS.
II.]
och
we may
+ yh + 2^ = 0;
therefore assume for
expressions,
I
and a form of
its
m=
x't^
47
its
tangential co-ordinates the
n=
?/-!,
tangential equation
z'l
thus found to be,
is
H + mh + nh = 0.
Conversely, if this tangential form were given,
turn to the local equation, by making
X=
/-f ,
?/
z=
= wrf,
we might
re-
n~i,
which would give x^ + i/î- z^ = 0, as before. The tangential

equation just now found becomes, when it is cleared of radicals ^
= /-2 + m-2 +
when
or,
it is
w-2
- 2m-' nr' -
2n-' t'
- 21' m'
also cleared o^fractions,
= F = w}n^ + nH^ + Pm^ - 2nPm -
2lrren
- 2mnH
of which the biquadratic form shows (by 56) that this cubic
a curve of the fourth class, as indeed it is known to be.
is
The
this
inflexional character (54) of the. points a", b", c"
curve
we make
upon
here recognised by the circumstance, that when

= 0, in order to find the four tangents from
is
m~n
a" =(0, 1,- 1) (36), the resulting biquadratic, Q -m^- 4lm^, has
three equal roots ; so that the line [1, 0, 0], or the side bc,
counts as three, and is therefore a tangent of inflexion : the fourth

tangent from a" being the line [1, 4, 4], which touches the
cubic at the point (- 8, 1, 1).
60. In general, the
two equations
nnjcf- lDzf=
may be
0,
(56),
nDyf- mDzf=
0,
considered as expressing that the homogeneous equa-
tion,
f{nx,ny, -lx-my) =
which
is
0,
obtained by eliminating z with the help of the rela-
tion Ix + my-{
nz=
0,
from f{x, y, z) =
0,
and which we may
48
[bOOK
I.
= 0, has two equal roots a;

{x, y)
y, if /, /w, n be
the co-ordinates of a tangent to the curve /*; an
equality
which obviously corresponds to the coincidence oftwo intersec-
denote by
still
tions of that line with that curve.

Conversely, if we seek by
the usual methods the condition of equality of two roots
x:y of
the homogeneous equation of the p*^ degree,
(a;,
by eliminating the
y) =f{nx, ny,
-Ix- my),
y between the two derived homo= Dj/0, we shall in general be

Dj.0,
ratio x
geneous equations,
conducted to a result of the dimension 2p(pand of the Jbrm,
= wP^P-i) F (/, m, n)
1) in
Z,
m,
w,
and
by
so,
by the
rejection of the foreign factor
we
this elimination,*
n:P^P-^\
introduced
shall obtain the tangential equation
F=0, which will be in general of the degree p(p-
such being
generally the known class (56) of the curve of which the

order (46) is denoted by p with (of course) a similar mode of
passing, reciprocally, from a tangential to a local equation.
:
61.
As an
assigned in 54,
example, when the function /has the cubic form

we are thus led to investigate the condition for
the existence of two equal roots in the cubic equation,
= (l>{x, y) ==
l(n
I)
X + (m - l)y]^ + 27n^xy(lx + my),
by eliminating x y between two derived and quadratic equa; and the result presents itself, in the first instance, as of
:
tions
the
tivelfth
but
it is
dimension in the tangential co-ordinates
found to be
effected, it is
is
well
by
n^,
and when
/,
m, n
this division is
reduced to the sixth degree, thus appearing to
imply that the curve

cubic
divisible
is
known
of the sixth class, as in fact the general

A. further reduction is however
to be.
possible in the present case, on account of the conjugate point

o (54), which introduces (comp. 57) the quadratic factor,
*
Compare the method emjjloyed
in
Salmon's Higher Plane Curves, page 98, to
find the equation of the reciprocal of a given curve, with respect to the
conic, x2
-I-
y8
-I-
jZ
= 0,
In general,
then (xyz)0^ &ndf(xi/z)
if
imaginary
the function f be deduced from /as above,
are equations of ttco reciprocal curves.
CHAP.
VECTORS OF POINTS IN SPACE,
III.]
(/
m + 71)^ =
49
and when this factor also is set aside, the tangential equation
found to be reduced to the biquadratic form* already assigned
is
59
in
the algebraic division, last performed, corresponding

geometric depression of a cubic curve with a
to the
known
But it is time
double point, from the sixth to the fourth class.
to close this Section on Plane Curves ; and to proceed, as in
the next Chapter we propose to do, to the consideration and
comparison of vectors ofpoints
in space^
CHAPTER
III.
APPLICATIONS OF VECTORS TO SPACE.
Section
I.
On Linear Equations between Vectors not Com"

planar.
When
three given and actual vectors oa, ob, oc, or

are
not
contained in any common plane, and when
j jSj 7,
the three scalars a, b, c do not /Z vanish, then (by 21, 22)
62.
the expression aa + b[5 + cy cannot become equal to zero; it

must therefore represent 50?we actual vector (1), which we may,
for the sake of
the
new
* If
we
(actual) vector S, or od,
we multiply
that form f
obtain a biquadratic equation in
= 1^ (?,
and
we
symmetry, denote by the symbol dS
if
ni)
we then
(59) by
Z
is
z^,
and then change nz
to-lx- my,
m, namely,
=(l- w)2 Qx + wz t/)2 + 2hn {I + m)
eliminate
where
not contained in any one
{Ix
+ my) z + ^2^,^222
between the two derived cubics,
are conducted to the following equation of the twelfth degree,
= Dî//, = D,ni^,
= x'^y'^z^fix, y, 2),
where /has the same cubic form as in 64.
"We are therefore thus brought hack

(comp. 59) from the tangential to the local equation of the cubic curve (54) complicated, however, as we see, with the /acôr x^y'^z'^^ which corresponds to the sys;
tem of the three
real tangents of inflexion to
that curve, each tangent being taken
The reason why we have not here been obliged to reject alio the foreign
factor, 22^ as by the general theory (60) we might have expected to be, is that we
multiplied the biquadratic function f only by z\ and not by z^.
three times.
50
[book
I.
of the three given and distinct planes, boc, coa, aob, unless
some one, at least, of the three given coefficients , h, c, vanishes and where the new scalar^ d^ is either
greater or less
than zero.
shall thus have a linear
equation between four
;
We
vectors^
aa +
which
~aa
OA
where
-aa
d
_b^
'
b(5
+ cy + dB = 0;
will give
OB
-b3
'
oc
-cy
d
'
-cy
or
CD = OA + OB + oc
or
are the
vectors of the three points

a', b', c', into which the
point D
is projected^ on the
three given lines oa, ob, oc,
by planes drawn parallel to

the three given planes, boc,
&c. so that they are the
;
three co-initial edges of a

parallelepiped, whereof the
and
given
lines,
by
and then
sum,
od or
^,
is
the internal
Or we may
project d on
lines da", db", dc" parallel to the three
co-initial diagonal
the three planes,
Fig. 28.
(comp.
shall
6).
have oa" = ob' + oc' =
bfi
+ cy
,
&c.,
and
g = OD = oa' + oa" = ob' + OB = oc' + oc".
"
And
evident that this construction will apply to any fifth

D
point
of space, if \kiQfour points oabc be still supposed to be
and
not complanar : but that some at least of the three
given,
ratios of the four scalars a, b, c, d (which last letter is not
it is
here used as a
sition
mark of differ entiatio'ii)
will vary with the
of the point d, or with the value of
example,
we
63.
We
vector
^.
po-
For
0, if d be situated in the plane boc

two other given planes through o.
may inquire (comp. 23), what relation between
coefficients must exist, in order that the point d
and similarly
these scalar
have a =
its
shall
for the
CHAP.
VECTORS OF POINTS IN SPACE.
III.]
may be
situated in the fourth given plane
condition
abc
51
;
or
what
of complanarity oi the four points, a, b, c, d.
is
the
Since
the three vectors da, db, dc are now supposed to be. complanar,
they must (by 22) be connected by a linear equation, of the
form
comparing which with the recent and more general form (62),
we
see that the required condition
is,
a +^-fc + c?=0.
This equation
may be
written (comp. again 23) as
-c
-a -b
-3- +-3- + 7=*
d
+=
OB
oc
OA
ob'
oa'
or
1,
oc'
1;
and, under this last form, it expresses a known geometrical

property of a plane abcd, referred to three co-ordinate axes
OA, oB, oc, which are drawn from any common origin o, and
terminate upon the plane.
have also, in this case of com-
We
planarity (comp. 28), the following proportion of coefficients

and areas :
a:b
or,
:- d = i)Bc: dca
dab abc
:
more symmetrically, with attention
to signs of areas,
aibic: d = bcd - cda dab - abc

:
where Fig. 18 may serve for illustration, if we conceive o in

that Figure to be replaced by d.
64. When we have thus at once the tioo equations,
aa +
b(5
+ cy + dd =
0t
and
so that the four co-initial vectors a,
on one common plane, and
/3,
a +b+c+
d=0,
7, S terminate (as above)
therefore be said (comp. 24) to

evident that the tivo right lines,
may
be termino-complanar, it is
DA and BC, which connect two pairs oi the four complanar
points, must intersect each other in some point a' of the plane,
at a finite or infinite distance.
And
there
no
difificulty
in
perceiving, on the plan of 31, that the vectors of the three
52
points
of intersection, which thus result, are the
a', b', c'
lowing
[book
I.
fol-
for
= bc*da,
a'
a^ d
b-\-c
cy + aa
c + a
for b'=
ca'db,
/3'
for c'=
ab'DC,
y' =
,
+ d^
bj3
b + d
aa + b3 Cy + dd
a+b
c + d
1
expressions which are independent of the position of the arbitrary origin o, and which accordingly coincide with the cor-
responding expressions in 27, when we place that origin in the

=
Indeed, these last results hold good
point D, or make 8 0.
when
vectors a, j3, y, S, or ihe^ve
even
the^wr
31),
(comp.
For, although there
points o, A, B, c, D, are all complanar.
then exist tivo linear equations between those four vectors,
which may
in general
a a + h^ +
c'y
be written thus,
+ d'^ =
without the relations,

coefficients,
if
yet
a'
a"a +
0,
+ &c. =
we form from
6"j3
cy
+ &c. =
0, a"
-f
d"^ = 0,
between the
0,
these another linear equation,
of the form,
(a"
+ ta)a +
and determine
(5"
tb')f5
by the
t
+ (c" + tc)y + (d" + td')S = 0,
condition.
b"+c"+d"
a" +
a +b' + c
we
have to make
shall only
tions written at the
'
-\-
d'
a=a" + ta\ &c., and
the two equa-
commencement of the present

and
article will
conduct to the expressions

for the three vectors of intersection which
above,
assigned
vectors may thus be found, without its being necessari/ to em-
then both be
satisfied;
will
ploy those processes of scalar elimination, which were treated

of in the foregoing Chapter.
As an Example,
let
aa +
the two given equations be (comp. 27, 33),
6/3
+ cy =
0,
(2a
+ c)o"'- aa =
CHAP.
and
let it
53
III.]
be required to determine the vectors of the intersections of the three pairs
of lines bc, aa'"
CA, ba'"
(2a +
and determining
by
and ab,
we have
1,
6/3
whereof the
first
= a',
6/3
+ cy) =
0,
the condition,
(2o + 6
which gives
Fonning the combination,
ca'".
+ c)a"' - aa + t{aa +
by
27.
- a + <(a + 6 +
c)
= 0,
for the three sought vectors the expressions,
+ cy
+c
c)
cy
'
+ 2aa
c+2a
'jaa
'
2a
6/3
'
+h
Accordingly, in Fig. 21, the line
A a'"
intersects
bc
in
and although the two other points of intersection here considered,

which belong to what has been called (in 34) a Third Construction^ are not marked
in that Figure, yet their anharmonie symbols (36), namely,
(2, 0, 1) and (2, 1, 0),
the point a'
might have been otherwise found by combining the equations y = and x=2z for the
two lines ca, ba'" and by combining z = 0, a: = 2y for the remaining pair of lines.
;
case, when i\iQ four given points

in
common
are
not
A, B, c, D,
pla7ie, let e be any ffih given
a.nj
on
of
not
situated
point
any one oi the four faces of the
space,
nor
on
given pyramid abcd,
any such face prolonged and let
In the more general
Qo.
its
vector oe =
Then the/owr co-initial vectors^
e.
ea, eb, eg,
ED, whereof (by supposition) no three are complanar, and which

do not terminate upon one plane, must be (by 62) connected
by some equation of the form,

a.EA + 5.EB + C.EO +
C?.ED
= 0;
where the^wr scalar s, nt, Z>, c, d, and their sum, which we shall
denote by - e, are all different from zero.
Hence, because
EA = a - , &c., we may establish the following linear equation
between five co-initial vectors, a,
/3,
7, S,
e,
whereof wo ^wr are
termino-complanar (64),
aa+bfi + cy + dd + ea = 0;
a+ b + c+ d+ e = 0, between the five scalars

whereof no one now separately vanishes. Hence
= {aa + 6j3 + cy + d^) ( + 6 + c + ^, &c.
with the
relation,
a, b, c, d,
also,
^^.
e,
Under
these conditions, if
Di
= DE'ABC,
we denote by
and
we
write
0Di =
i,
the vector of the point Di in which

the right line de intersects the plane abc, we shall have
that
is,
if
54
^
^1
In
aa +
bj5
7
+ cy dS +
= -3
[bOOK
I.
ee
.
these two expressions are equivalent, or represent one

vector, in virtue of the given equations; but the first
fact,
common
shows (by 63) that this vector ^1 terminates on the plane abc,
and the second shows (by 25) that it terminates on the line
DE ; its extremity Dj must therefore be, as required, the intersection of this line with that plane. We have therefore the two
equations,
..a(a-g0 + ^(i3-î) + c(7-S0 = 0;
I.
II..
,d(^-S,) + e(e-^,) = 0;
wlience (by 28 and 24) follow the two proportions,

T.
a:b:c= DiBC DiCa DiAB

d:e= EDii DiD
ir.
the arrangement of the points, in the

annexed Fig. 29, answering to the case
where
all
the four coefficients a,b,c,d
are positive (or have one common si^n),

and when therefore the remaining coefficient e is negative (or
67.
tion,
^^*
'
has the opposite sign).

triangles, in the
For the three complanar
we may
substitute
first proporwhich
three
volumes,
pyramidal
any
rest upon those triangles as their bases, and which have one
common vertex, such as D or e and because the collineation
;
DEDi gives DDiBC - EDiBC = DEBC, &c., wc may write

proportion,
I".
aib:c = DEBC deca deab.

:
Again, the same collineation gives

EDi DDi = EABC
:
we have
therefore,
DABC
by IP., the proportion,
11".
dl
-e = EABC DABC.
;
But
DEBC + DECA + DEAB + EABC = DABC,
and
this other
CHAP.
III.]
a-\-
we may
c-i-
d = -e;
therefore establish the following fuller formula of
proportion, between coefficients and volumes
HI.
a:b:c:d: -e = DEBC deca: deab eabc dabc;
55
the ratios of all these five pyramids to each other being considered as positive, for the particular arrangement of the points
which is represented in the recent figure.
The formula III. may however be regarded as perfectly general, if we agree to say that pyramidal volume changes
68.
2>,
sign^ or rather that it changes its algebraical character, as po^

sitive or negative, in comparison with a give?i pyramid, and
with a given arrangement of points, in passing through zero

(comp. 28) namely when, in the course of any continuous
;
change, ang one of its vertices crosses the corresponding base.

With this convention* we shall have, generally,
DABC = - ADBC = ABDC = - ABCD, DEBC = BODE, DECA = CDEA
the proportion III. may therefore be expressed in the following more symmetric, but equally general form :
Iir.
a:b:c:d:e = bode
odea: deab eabc

:
abcd
the sum of these five pyramids being always equal to zero,

when signs (as above) are attended to.
69.
We saw
(in 24) that the
aa +
5/3
+ c^ =
two equations,
+6+c=
0,
0,
gave the proportion of segments,

a
c= Bc CA
:
ab,
whatever might be the position of the origin o. In like manner we saw (in 63) that the two other equations,
*
Among the consequences of this convention respecting signs of volumes, which
has already been adopted by some modern geometers, and which indeed is necessary
(comp. 28) for the estabUshment of general formula:, one is that any two pyramids,
ABCD,
a'b'c'd', bear to
rotations,
BCD and
each other a positive or a negative ratio, according as the two

supposed to be seen respectively from the points A and a',
b'c'd',
have similar or opposite directions, as right-handed or left-handed.
56
aa +
b[5
a+b+c +
Cy+dS=0,
[bOOK
I.
d=0,
gave the proportion of areas,
a:b:c:d= bcd - cda dab - abc

:
where again the origin
is
arbitrary.
And we have just deduced
of volumes,
(in 68) a corresponding proportion
from the two
analogous equations (65),
aa +
bj5
-\^
cy + d^ +
ee
a+b+
0,
c+d+e=0,
with an equally arbitrary origin. If then we conceive these

segments, areas, and volumes to be replaced by the scalars to
which they are thus proportional, we may establish the three

general formulcB
OA.BC + OB.CA+ OC.AB =

II. OA.BCD -OB.CDA-f OC.DAB -OD.ABC =
III. OA.BCDE + OB.CDEA+ OC.DEAB + OD EABC+ OE ABCD =
I.
where
in I.,
in II.,
and in
a, b, c are a72i/ three collinear points ;
A, B, c,
III.,
a, b, c,
D are any four complanar points ;

d, e are any five points of space ;
an entirely arbitrary
remembered, that the additions
point.
and subtractions are supposed to be performed according to the
rules of vectors, as stated in the First Chapter of the present
Book the segments, or areas, or volumes, which the equations
while o
is,
in each of the three formulas,
It must, however, be
indicate, being treated as coefficients of those vectors.
We
might still further abridge the notations, while retaining the

meaning of these formulas, by omitting the symbol of the arbitrary origin o and by thus writing,*
;
A.BC + B.CA +
r.
CAB
0,
any three collinear points with corresponding formulae II'.

and Iir., for any four complanar points, and for any five points
for
of space.
We should
thus have some of the notations of the Bart/centric Calculus (see
Note B), but employed here with
different interpretations.
CHAP.
III.]
QUINARY SYMBOLS FOR POINTS
Section 2.
On Quinary
IN SPACE.
57
Symbols for Points and Planes in

Space.
The equations
of Art. ^6 being still supposed to hold good,

the vector p of any point P of space may, in indefinitely many ways,
be expressed (comp. 36) under the form:
70.
_.
I.
in
which the
..0P=/> =
ratios
xaa + vhQ + zc^ + wdh +
differences of
of the
vee
^-4
the^ve
coefficients^
xyzwv, de-
In fact, because the four points

termine the position of the point.
ABCD are not in any common plane, there necessarily exists (comp.
65) a determined linear relation between the four vectors drawn to
them from the point
which may be written thus,

x'a PA + y'b PB + z'c PC + w'd pd = 0,
?,
giving the expression,
_ x'aa + y'b^ +
x'a + y'b-\in
which the
ratios of
z'cf^
z'g
+ w'dh
*
+ w'd
the/owr scalars x'y'z'w\ depend upon, and
conversely determine, the position of p ; writing, then,
x = tx' -^-v,
z = tz' +
y-ty'-vv^
ty
V,
= tw' + v,
where i and v are two new and arbitrary scalars, and remembering
that aa + + 66 = 0, and +
+ e = (65), we are conducted to the
form for /?, assigned above.
71. When the vector p is thus expressed, the point p may be
. .
. .
denoted by the Quinary Symbol {x,
y, z^
w, v)
and we may write
the equation^
p
But we
see that the
oHhe same
{x, y, z,
w,
same point p may
v).
also
be denoted by this
other
provided that the followbetween

ing ;?ropor<iow
differences of coefficients (70) holds good:
x' -v' :y' -v^'.z' ~v'\w' -v' = x-v:y-v:z-v:w -V.
symbol,
Under
kind,
this condition,
we
(a;', 3/',
z\ w',
v'),
shall therefore write the following /ormw/a
of congruence,
E {x,
{x', y', z', w', v')
y,
z,
w, v\
to express that these two

quinary symbols, although not identical in
composition, have yet the
common point.
And we
{x',
y,
sam^ geometrical
signification, or
denote one
shall reserve the symbolic equatiott,

z',
w', v')
I
{x, y, z,
w,
r),
58
'
ar,
equal to
= V.
v'
I.
v\ of the one symbol, are

coefficients, x'
the corresponding coefficients of the other,
to express that the Jive

separately
[bOOK
72. Writing also, generally,
= t(x, y, z, w, v),
=
v' + v)
{x\
v), &c.,
r') + (a;,
(te, ty, tz, tw, tv)
(x'
+ re,
. .
and abridging the particular symbol*
(Q0
(Q),
{x\
in
. .
v')j
(0,
and
(a?,
v),
For example,
coefficients.
while
{U\
"^^y briefly denote the quinary symbols

we may thus establish the congruence (71),
and u are arbitrary

=
0,0, 0, 1)
(1,1, 1,1,0),
which
(1, 1, 1, 1, 1) to
(0, 0, 0, 1,
1)E(1,
0,0);
1, 1,
each symbol of the first pair denoting (65) the given point E; and
each symbol of the second pair denoting (66) the derived point Dj.
When
may
the coefficients are so simple as in these last expressions, we

occasionally omit the commas, and thus write, still more briefly,
E (11 110);
(00001)
(00011)
(11100).
/>,
p'
p'\ expressed each under the first
be
termino-coUinear
form (70),
(24) and if we denote their denomimust then (23) be
x"a +
tors, xa +
^ xa +
, by m, m, m", they
connected by a linear equation, with a null sum of coefficients, which
73. If three vectors,
may be
written thus
tmp +
We have,
t
t'm'p'
+ t'^m"p" =
0.
+ vee) + V {x'aa + + v'ee) + V {x"aa + + v"e^ =

t{xa^. + ve) + V [x'a ^- .\v'e)\ t" {x"a + + v^'e) =
{xaa +
where
tm + t'm' + t"m" +
therefore, the two equations of condition,
t,
. .
V, i" are three
new
scalar s, while the five vectors a
e,
and
the five scalars a..e, are subject only to the two equations (65):
but these equations of condition are satisfied by supposing that
tx
+ Vx' + V'x" =
= î? + t'v* + t"v" = -u,
where u is some new scalar, and they cannot be satisfied otherwise.

Hence the condition of collinearity of the three points p, p', p", in
which the three vectors p, p', p" terminate, and of which the quinary symbols are (Q), (Q'),
{Q>"),
may
briefly be expressed
by the
equation,
* This
quinary symbol
( {/) denotes
no determined point, since
r/by 70, 71) to the indeterminate vector p

;h other quinar}'
symbols, as above.
=-
but
it
it
corresponds
admits of useful combinations
CHAP.
QUINARY SYMBOLS OF PLANES.
III.]
so that if any four scalars,
t\ t'\ u,
t^
59
can he found^ which
satisfy this
then, but not in any other case, those three

For example, the three
points vv'v" are ranged on one right line.
points D, E, Di, which are denoted (72) by the quinary symbols,
last symbolic equation,
(00010), (00001), (11100),

And
three symbols is ( U).
SLve collinear ;
we have
if
because the sum of these
the equation,
(Q'0 = (Q) + <'(QO + ^(^),

where
<,
on
p'',
f,
u are any three scalars^ then {Q!') is a symbol for a point

For example, the symbol (0, 0, 0, t^ t') may
the right line pp'.
denote any point on
the line de.
By reasonings precisely similar it may be proved, that if

(QO CQ'O (Q^'O be quinary symbols for a.ny four points pp^p^^p'^/
74.
(Q)
any common plane,
so that the four vectors pp'p^p'" are terminoan

then
complanar (64),
equation, of the form
in
t{Q)^t'{a')^t"{Q")+t"'{Q:'^) = -u{U),
must hold good; and conversely, that \i the fourth symbol can be
{Q'") =
t{a)U\Q')-vt^'{Q')û{U\
with any scalar values of

ated in the plane
ppV
if,
t", u,
t',
then the fourth point v'" is situFor example, the four
of the other three.
points,
(1 0000),
(01000),
(11100),
(00100),
or A, B, c, Di {^^^^ are complanar; and the symbol
may represent any point
When
{t,
t\ t", 0, 0)
in the plane abc.
a point p
is thus complanar with three given points,

therefore expressions of the following forms, for
the five coefficients x, ..v of its quinary symbol, in terms of the^teen given coefficients of their symbols, and olfour new and arbitrary
75.
Po, Pi, P2>
scalars
we have
X-
t^XQ
And hence, by
t^Xy
+ Uj)C2 + w
v= toVo + tiVi + tzV.^ + u.
elimination of these four scalars,
<o
">
we
are con-
ducted to a linear equation of the form
=
l(x-v) + m{y-v) + n (z -v) + r{w-v)
which may be
called the
Quinary Equation of
the
0,
Plane
P0P1P2,
or of
the supposed locus of the point p: because it expresses a common

property of all the points of that locus; and because the three ratios
of the four new coefficients ly m, w, r, determine the position of the plaiie
60
in space.
It
I.
however, more symmetrical, to write the quinary
is,
equation of a plane
as follows,
my + nz + rw + sv = 0,
Ix +
where the fifth
[bOOK
coefficient^ s, is
connected with the others by the rela-
tion,
+ m-\-n + r+s=^0;
and then we may say that [/, m, n, r, 5] is (comp. 37) the Quinary
Symbol of the Plane 11, and may write the equation^
n=
For example, the
m,
n, r, 5].
coefficients of the
symbol
[_l,
ABC may be thus expressed (comp. 74)

z = t^ +
a? = fo + ,
y = ti + u,
for a point
p in the plane
v=m;
w-u,
u,
between which the only relation, independent of the four arbitrary

scalars to.,u, is w-v=zO', this therefore is the equation of the plane
ABC, and the symbol of that plane is [0, 0, 0, 1,-1]; which may
(comp. 72) be sometimes written more briefly, without commas, as
[000 11]. It is evident that, in any such symbol, the coefficients may
be multiplied by any common factor,
all
76.
The symbol of the plane
we may next
plane
is
PqPiPz having been thus determined,

propose to find a symbol for the pointy P, in which that
intersected
by a given
line P3P4:
or to determine the
coefficients
or at least the ratios of their differences (70), in the quinary

symbol of that point,
a;
v,
(x, y, z,
w,
= p = PoPiP,-P3P4.
i;)
Combining, for this purpose, the expressions,

X = t^X^ + tiXi + U^,
V= f^Vs + t^v^ +
.
u'f
(which are included in the symbolical equation (73),
and express the
coUinearity PP3P4,) with the equations (75),

â;
+ ..+5V=0,
^+.. +
we
(which express the complanarity PP0P1P2,)
= 0,
are conducted to the
formula,
<3
{IX^
SVs)
<4
{Ix^
4-
+ sv^ =
which determines the ratio t^ t^, and contains the solution of the
problem. For example, if p be a point on the line de, then (comp.
:
73),
CHAP.
but
III.]
if it
therefore
QUINARY TYPES OF POINTS AND PLANES.
be also a point in
^3-^ = 0; hence
the
plane abc, then
iv-vÔ
61
(75),
and
(Q) = ^3(00011)+'(lllll), or (Q)E(OOOll);

which last symbol had accordingly been found (72) to represent the
intersection (fiQ), Di
= abc de.
77. When the five coefficients, xyzwv, of any given quinary

are
symbol (Q) for a point p, or those of any congruent symbol (71),
whole numbers (positive or negative, or zero), we shall say
any
to the Jive given pointsj

(comp. 42) that the point p is rationally related
A E; or briefly, that it is a Rational Point of the System, which
And in like manner, when the five
those five points determine.
.
quinary symbol (75) of a plane 11 are either

to
or
integers, we shall say that the plane is a Raproportional
equal
tional Plane of the same System; or that it is rationally related to the
same five points. On the contrary, when the quinary symbol of a
coefficients, Imnrs, of the
point, or of a plane, has not thus already whole coefficients, and cannot be transformed (comp. 72) so as to have them, we shall say that
the point or plane is irrationally related to the given points; or

it is irrational.
right line which connects two rational
briefly, that
the intersection of two rational planes, may be called, on

the same plan, a Rational Line ; and lines which cannot in either
points, or
is
of these two
ways be constructed, may be said by contrast to be

It is evident from the nature of the eliminations
employed (comp. again 42), that a plane, which is determined as containing three rational points, is necessarily a rational plane; and in
like manner, that a point, which is determined as the common inter-
Irrational Lines.
of three rational planes, is always a rational point : as is also

every point which is obtained by the intersection of a rational line
with a rational plane ; or oi tivo rational lines with each other (when
section
they happen to be complanar).

78. Finally, when two points, or two planes, differ only by the arrangement (or order) of the coefficients in their quinary symbols, those
points or planes
to be syntypical.
be said to have one common type; or briefly

For example, Xhefive given points, a,
e, are thus
may
syntypical, as being represented
. .
by the quinary symbols (10000),

and the ten planes, obtained by taking all the
(00001)
ternary
combinations of those five points, have in like manner one common
.
type.
Thus, the quinary symbol of the plane abc has been seen
(75) to be [OOOlT]; and the analogous symbol [iToOO] represents

the plane cde, &c.
Other examples will present themselves, in a
62
[bOOK
I.
But
shortly subsequent Section, on the subject of Nets in Space.
seems proper to say here a few words, respecting those Anhar-
it
monic Co-ordinates, Equations^ Symbols, and Types, for Space, which

are obtained from the theory and expressions of the present Section,
by reducing (as we are allowed to do) the number of the coefficients,
in each
symbol or equation, ivovafioe
Section
79.
3.
On Anharmonic Co-ordinates
When we
we may)
to four.
in Space.
adopt the second form (70) for />, or suppose (as

we get this
coefficient in the ^r5^ form vanishes,
that the fifth
other general expression (comp. 34, 36), for the vector
of a point
in
space:
ov = p=
xaa + yh3 + zc^ + wdb

-J',
xa +^^
yo+zc + wd
and may then write the symbolic equation (comp. 36, 71),
7^{x,y,z, w),
and
we
the Quaternary Symbol of the Point p: although

shall soon see cause for calling it also the Anharmonic Symbol of
call this last
Meanwhile we may remark, that the only congruent

symbols (71), of this lastform^ are those which differ merely by the
introduction of 2i common factor : the three ratios of the /owr coeffi-
that point.
X .w, being all required, in order to determine the position of

whereof those four coefficients may accordingly be said
(comp. 36) to be the Anharmonic Coordiriates in Space.
.
cients,
the point;
80.
When we
the point
p,
of a plane
IT,
thus suppose that v = Q, in the quinary symbol of
we may suppress
lx+ .+sv =
.
the fifth term sv, in
(75)
the quinary equation
and therefore may suppress
also (as
here unnecessary) the fifth coefficient, s, in the quinary symbol of that

plane, which is thus reduced to the quaternary form,
n = p,
This last
the Plane, of
m,
n,
r~\.
hethe Anliar'inomc Symbol of

which the Anharmonic Equation is
Lc + my + nz + riy =
may
also be said (37, 79), to
the /owr
coefficients,
Imnr, which we
shall call also (comp. again 37)
the Anharmonic Co-ordinates of that Plane n, being not connected

themselves by any general relation (such as /+
+ 5 = 0): since
their three ratios (comp. 79) are all in general necessary, in order to
among
determine the position of the plane in space.

81. If
we suppose
that the fourth
coefficient,
w, also vanishes, in
CHAP.
ANHAHMONIC CO-ORDINATES
III.]
63
IN SPACE.
is in theplane abc ; and maythen be sufficiently represented (as in 36) by the Ternary Symbol
And if we attend only to the points in which an arbitrary
{x, y, z).
the recent symbol of a point, thsit point P
the given plane abc, we may suppress its fourth cobeing for sucli points unnecessary. In this manner,
then, we are reconducted to the equation^ Ix+my + nz= 0, and to the
= [^, m, w], for a right line {37) in the plane abc, consideied
symbol, A
here as the tiâce, on that plane, of an arbitrary plane H in space. If
plane
intersects
efficienty r,
as
this plane 11 be given
the ternary symbol for

coefficients, r and s.
82. In the
to the plane
its
by
quinary symbol (75),
its trace
we thus
obtain
A, by simply suppressing the two
last
case, when the point p is not confined

we denote (comp. 72) its quaternary symbol by
more general
abc,
if
(Q), the lately established formulae of collineation and complanarity

(73, 74) will still hold good: provided that we now suppress the
symbol
U), or suppose
its coefficient to
be
zero.
Thus, the formula,
(Q)='(QO+''(a'o+^'''(Qn
expresses that the point p is in the plane p'p'^p^'^ ; and if the coefficient
vanish, the equation which then remains, namely,
thus complanar with the two given points v', v",

and with an arbitrary third point; or, in other words, that it is on
the right line v'v"', whence (comp. 76) problems of intersections of
lines with planes can easily be resolved.
In like manner, if we designifies that p is
note briefly by [i2] the quaternary symbol

n, the formula
expresses that the plane
planes,
Ti',
II",
W"
and
\l,
m, n, r] for a plane
passes through the intersection of the three

we suppose t"^ = 0, so that
if
the formula thus found denotes that the plane II passes through
the point of intersection of the two planes, 11', 11^'', with any third
plane; or (comp. 41), that this plane 11 contains the line of intersection of n', n'' ; in which case the three planes, 11, 11', 11'',
may be
said to be collinear.
Hence
it
appears that either of the two expres-
sions,
I.
may be used
.t'{ QO
as a
consider that line
1" ( Q"),
II.
Symbol of a Right Line
either,
1st,
t'
in
[7^']
V' [i?"],
Space : according as we
two giien poirits, or
as connecting
64
[bOOK
I.
Ilnd, as being the intersection of two given planes. The remarks (77)
on rational and irrational points, planes, and lines require no modification here; and those on t^pes (78) adapt themselves as easily to
quaternary as to quinary symbols.

83.
From
it
the foregoing general formulse of collineation and comfollows that the point p^ in which the line ab inter-
planarity,
sects the plane cdp through
of space,
may
CD and any proposed point p = (xyzw)
be denoted thus
p'
for example, e
= AB CDP = {xyOO)
and c' = ab cde =
(1
1 1
1),
(1 100).
In general,
if
ABCDEF be any six points of space, the four collinear planes (82), abc,
abd, ABE, ABF, are said to form o, pencil through ab; and if this be
cut by any rectilinear transversal, in four points, c, D, e, f, then
(comp. 3o) the anharmonic function of this group of points (25) is
Anharmonic of the Pencil of Planes: which may be
called also the
thus denoted,
(ab cdef) =
.
(c'd'e f').
Hence (comp. again 25, 35), by what has just been shown
ing c' and p', we may establish the important formula:
X
(cd aebp) = (ac'bpO = .
respect-
so that this ratio of coefficients, in the
symbol (xyzw)
for a variable
point p (79), represents the anharmonic of a pencil of planes, of which

the variable plane cdp is one; the three other planes of this pencil
being given.
In like manner,
(ad
becp) = -,
^
and
(bd ceap) = - ;
X
.
so that (comp. 36) the product of these three last anharmonics is

unity.
On
the same plan
(bC
AEDP)^ =
^
.
we have
also,
1J
(CA BEDP) = ,
.
(aB
CEDP)
=-
so that the three ratios, of the three first coefficients xyz to the
fourth coefficient w, suffice to determine the three planes, bcp, cap,
ABP, whereof the ôm^ p is the common intersection, by means of the
anharmonics of three pencils of planes, to which the three planes reAnd thus we see a motive (besides that of analogy
spectively belong.
to expressions already used iov points in a given plane), for calling
t\iQ four coefficients, xyzw, in the quaternary symbol (79) for
point in
21.
space, the
Anharmonic Co-ordinates of that Point.
84. In general, if there be any four collinear points,
Pq,
P3,
so
CHAP.
ANHARMONIC CO-ORDINATES
III.]
65
IN SPACE.
that (comp. 82) their symbols are connected hy two linear equations,
such as the following,
=
(Qa) t'{Qo) + w'(QO,
(Qi) KQo) + u{Q^),
then the anharmonic of their group may be expressed (comp. 25, 44)
as follows:
(PoPi>2P3)
as appears
AB
ut'
,;
by considering the pencil (cd PoPiPsPs), and the transversal

in like manner, if we have (comp. again 82) the two
.
And
(83).
other symbolic equations, connecting four collinear planes Hq

[i?i]
n^,
=
[^3] '[i?o] + t*li?J,
= <[i?o] + w[i?J.
the anharmonic of their pencil (83)
is
expressed by the precisely
similar formula,
(nn,n,n3)
as
may be proved by supposing
-;
the pencil to be cut by the same
transversal line ab.

85. It follows that if f{xyzw) dindfi{xyzw) be
geneous and linear functions of

collinear planes IIo
re,
w; and
y, z,
if
any two homo-
we determine four
by the four equations,

/, = 0,
A = kf,
f=f,
Xlg (82),
/=0,
where k is any scalar; we shall have the following value of the anharmonic function, of the pencil of planes thus determined
:
(nonin2n3)
Hence we derive
this Theorem,
= ^' =
which
is
4
important in the application
of the present system of co-ordinates to space:
" The
Quotient of any two given homogeneous and linear Functions,
the
anharmonic Co-ordinates (79) of a variable Point p in space, may
of
be expressed as the
Anharmonic (rioninalls) of a Pencil of Planes ;
whereof
three are given,
point
and through a given right line
p,
while the fourth passes through the variable
which
is
common
to the three
former planes.''^
86.
And
Theorem:
in like
manner may be proved
this other
but analogous
" The
Quotient of any two given homogeneous and linear Functions,
of the anharmonic Co-ordinates (80) of a variable Plane n, may be expressed as the Anharmonic (P0P1P2P3) of a Group of Points; whereof
three are given
of their
and
and the fourth is the intersection,

given right line A, with the variable plane
collinear,
common and
'
IT."
fl,
^Q
if
[bOOK
I.
More fully, if the two given functions of Imnr be f and f,, and
we determine three points P0P1P2 by the equations (comp. 57)
= 0, Fi = F, Fi = 0, and denote
by P3 the intersection of their com-
mon
line
with n, we shall have the quotient,
^
For example,
we suppose
if
A2=(I001),
A'2
(PoPiP.,P3).
that
B2=(010l),
= (1001),
C2=(0011),
= (010T),
b'2
c'2
(001T),
so that
A2 = DA*BCE,
and
(da2Aa'2)=- 1, &c.,
n
to r, in the symbol IT =
Z, m,
\^lmnr'],
may be expressed (comp. 39) under the form of anharmonics of
we
&c.,
find that the three ratios of
groups, as follows:
-=
where
q, r, s
(DA'gAQ)
-=
(DB^gBR)
- =
(cCgCS)
denote the intersections of the plane n with the three

And thus v/e have a 7notive (comp.
given right lines, da, db, dc.
83) besides that of analogy to
lines in
a given plane (37), for calling
ihe/bur coefficients Z, m, ?i, r, in the quaternary symbol (80)

& plane n, the Anharmonic Co-ordinates of that Plane in Space.
(as above)
for
87. It may be added, that if we denote by l, m, n the points in

which the same plane 11 is cut by the three given lines bc, ca, ab,
and retain the notations k'\ "b", c^' for those other points on the same
three lines which were so
now
write (comp. 36)

A^'- (0110),
marked before
B^^
= (T010),
(in 31, &c.), so that
c'^=:
we may
(llOO),
we shall have (comp.
39, 83) these three other anharmonics of groups,

with their product equal to unity:
= (ca'^bl)
n
-=
I
(ab'^cm)
= (bo'ân)
on
and ihQsix given points, A.", b'\ &\ A'2, b'o, c'g, are all in one given plane
the equation and symbol are:
[e], of which
=
= 0',
x-\-y + z + w
[e] [11111].
The
six groups of points, of
which the anharmonic functions thus
represent the six ratios of the four anharmonic co-ordinates, Imnr,

of a variable plane 11, are therefore situated on the six edges of the
given pyramid, adcd
tivo
points in each group being cornei^s of that
CHAP.
GEOMETRICAL NETS
III.]
IN SPACE.
67
intersections of the edge with

pyramid, and the two others being the
and
n.
the tivo planes, [e]
Finally, the plane [e] is (in a known
modern sense) the plane of homology,"^" and the point e is the centre
of homology^ of the given pyramid abcd, and of an inscribed pyramid
A,BA^b where
Ai
fication {QQ, 76),
Ai
= ea'BCD, &c.; so that Di retains its recent

and we may write the anharmonic symbols,
= (OIll),
And
Bi
= (1011),
Di
Ci-:(1101),
signi-
= (IllO).
the harmonic conjugates to these

last points, with respect to the lines ea, eb, eg, ed, so that
if
we denote by a!{b!xQ'xT>\
(eaiAA'i)
we have
= (ed,dd'i) = -
the corresponding symbols,
B^ = (1211),
A^=(2111),
c'i
d^ = (1112).
= (1121)
Many other relations of position exist, between these various points,

lines, and planes, of which some will come naturally to be noticed,
in that theory of nets in space to which in the following Section we
shall proceed.
Section
When we
88.
4.
On Geometrical Nets
have
in Space.
(as in Q5>) five given points
e,
whereof no
four are complanar, we can connect any two of them by a right line^
and the three others by a plane, and determine the point in which
these last intersect one another: deriving thijs a system oHen lines Aj,
ten planes IIi,
Po,
and
ten points Px,
by what may be
from the given system o^ five points
called (comp, 34) a First Construction.
We may
next propose to determine all the new and distinct lines, Ag, and
planes, 112, which connect the ten derived points p^ with the five
given points Pq, and with each other; and may then inquire what
new and
ivith
distinct points P3 arise (at this stage) as intersections of lines
planes, or of lines in one plane with each other : all such
new lines,
and points being said (comp. again 34) to belong to a Second

And then we might proceed to a Third Construction
Construction.
of the same kind, and so on for ever
building up thus what has
planes,
been calledf a Geometrical Net in Space.
To
express this
geome-
(71, 75, 82) appoints, planes, and

lines, and by quinary types (78), so far at least as to the end of the
second construction, will be found to be an useful exercise in the
trical process
by quinary symbols
* See Poncelet's Traite des

Propritte's Projectives (Paris, 1822).
t By Mobius,
ia p.
291 of his already
cited Barycentric Calculus.
68
application of principles lately established

mately in that Method of Vectors, which
[BOOK
and therefore
I.
ulti-
the subject of the

present Book. And the quinary form will here be more convenient
than the quaternary, because it will exhibit more clearly the geomeis
dependence of the derived points and planes on the five given

points, and will thereby enable us, through a principle of symmetry,
trical
to reduce the
number of distinct
Of the
89.
types.
five given points, Pq, the quinary type has
been seen
(78) to be (10000); while of the ten derived points p,, of first construction, the corresponding type may be taken as (00011); in fact,
considered as symbols, these two represent the points a and d^. The
nine other points Pi are AVc'AiBiCiAoBaCz ; and we have now (comp.
83, 87, 86) the symbols,
Ai = EA BCD = (10001),
=
=
DA
BCE
A2
(10010);
A'= BC ADE = (01 100),
also, in
to
any symbol or equation of the present form,

a, b, c to b, c, a,
change
the third,
first,
and second
second, and third: thus, b'
provided that
it is
permitted
we at the same time write
co-ej6&cients, in
the places of the
= ca bde = (lOlOO),
&c.
first,
The symbol
(xyOOO) represents an arbitrary point on the line ab; and the symbol [OOwrs], with 71 + r + 5 = 0, represents an arbitrary plane through
that line : each therefore may be regarded (comp. 82) as a symbol also
of the
ab
line
itself,
and
at the
same time
as a type of the ten lines
A^; while the symbol [000 ll], of the plane abc (75),
(78) as a type of the ten planes
BCDE,
and the
CADE,
IIi.
may betaken
Finally, the five pyramids^
ABDE,
ABCD,
ABCE,
such as abc, whereof each is a common face of

two such pyramids, may be called pyramids Bi, and triangles Ti, of
ten triangles,
the First Construction.

90. Proceeding to a Second Construction (88), we soon find that
may be arranged in two distinct groups; one group con-
the lines A3
sisting of fifteen lines A2,

nects two points Pj,
section of
1,
such as the line* aa'd^, whereof each con
and passes
two planes
IIi
through that point, as here of abc, ade
while the other group consists of
thirty lines A2,
2,
such as
b'c',
connecting two points Pi, but not passing through any point
being one of the thirty edges of five new pyramids R^, namely,
C'b'AzAi,
*
A'c'B^Bi,
also through one point Pq, being the inter-
BVC2C1,
A2B2C2D1,
Po,
AiBiCiDj:
AB1C3, ABoCi, DAAi, ea'aj. are other lines of this group.
each
and
CHAP.
GEOMETRICAL NETS
111.]
IN SPACE.
69
which pyramids R^ may be said (comp. 87) to be inscribed homothe centres of homology for these
logues of the five former pyramids i?i,
five pairs
ofpyramids being the five given points A

five planes
of homology being
[a]., [e],
e ; and
whereof the
last
t\iQ planes
has been
a third conalready mentioned (87), but which belong properly to

struction (88).
The planes ITj, oi second construction, form in like
o^ fifteen planes Ila, i, such as the
plane of the five points^ AB1B2C1C2, whereof each passes through one
point Po, and ihioM^ four points Pi, and contains two lines Ag,!, as
manner two groups; one consisting
here the lines AB1C2, ACiBg, besides containing /owr lines A2,2, as here
BiB^, &c. ; while the other group is composed of twenty planes Ti.^,2i
such as AiBiCi, namely, the twenty faces of the five recent pyramids i?2
whereof each contains
three points Pi,
not pass through any point
It is
Pq.
and
three lines A2,2,
now required
but does
to express these
geometrical conceptions* of the forty-five lines A2; the thirty five planes
lis; and the five planes of homology of pyramids, [a]
[e], by qui.
nary symbols and
types,
before proceeding to determine the points Pg
of second construction.
91. An arbitrary joozrz^ on the right line aa'Di (90) may be represented by the symbol (ifwwOO); and an arbitrary ji?Za/ze through
that line by this other symbol, [OwTwrr], where m and r are written
(to save
commas) instead
of-wand-r;
hence these two symbols
(comp. 82) denote the line aa'Di itself, and may be used as
types (78) to represent the group of lines Ag, j. The particular sym-
may
also
bol [01111], of the last form,

through the last-mentioned line,
of the
112,1.
same group
The
and may serve
line b'c',
represents that particular plane

also the line AB1C2
which contains
as a type for the
and the group
A2,2,
may
group of planes
be represented by
and s = -s; while
=
agreef to write s t + Uf
the plane b'c'a^, and the group 112,2, i^^y he denoted by [11112].
Finally, the plane [e] has for its symbol [11114]; and the four
(stuOO) and [tttus^, if
we
other planes [a], &c., of homology of pyramids (90), have this last
for their
92.
common
The points
type.
Y^, of second construction (88), are
more nume-
Miibius (in his Barycentric Calculus, p. 284, &c.) has very clearly pointed
out the existence and chief properties of the foregoing lines and planes ; but besides
that his analysis is altogether different from ours, he does not appear to have aimed
at enumerating, or
even at classifying,
(88) the second construction, as
all i\\Q
we propose
points of what has been above called
shortly to do.
f With this convention, the line ab, and the group Ai,
the plane symbol [OOtus']
their
point-symbol being (^mOOO).
may
be denoted by
70
[bOOK
I.
lines Ag 2i,i[idi planes II2 of that construction: yet with

the help of types, as above, it is not difficult to classify and to
enumerate them. It will be sufficient here to write down these
rous than the
which are found to be eight, and to offer some remarks respecting, them in doing which we shall avail ourselves of the eight
following typical points, whereof the two first have already occurred,
and which are all situated in the plane of abc
types,
= (21100);
A"" = (12100);
= (0lT00);
A^' = (02100);
A^^^
A''
a'^
=(21100);
a^
a^"'
= (32100);
a'"=:(23T00);
=(02100);
the second and third of these having (lOOll) and (30011) for conIt is easy to see that these eight types repregruent symbols (71).
sent, respectively, ten, thirty, thirty, twenty, twenty, sixty, sixty,
and sixty
mark
distinct points, belonging to eight groups,
which we
shall
number
of the points Pg is 290.

If then we consent (88) to close the present inquiry, at the end of
what we have above defined to be the Second Construction, the total
as
Po,
1,
P28;
so that the total
number of the nt points, p^ P2, which are thus derived by lines

and planes from the five given points Po, is found to be exactly three
hundred: while the joiiit number of the net-lines, A^, A2, and of the
(1.)
To
has been seen to be one hundred, so
111, Yl^,
net-planes,
far.
the type P2,i belong the ten points,

A"b"c",
a'iB'iC'iD'i,
A'2B'2C'2,
with the quinary symbols,
a"= (OllOO),
A'2= (lOoTo),
a'i
= (10001),
which are the harmonic conjugates of the ten points

A2B2C2,
A'b'c',
d'i
(00011),
Pi, namely, of
AiBxCiDi,
with respect to the ten lines Ai,on which those points are situated so that we have
ten harmonic equations, (ba'ca") = 1, &c., as already seen (31, 86, 87). Each point
;
P2,
is
the
common
intersection of a line
blish the four following /ormM?
mulae)
Ai with
three lines
A 2, 2
thus
we may esta-
of concurrence (equivalent, by 89, to ten such
for-
a" =BC*b'c' 'Bid '6202;

a'i
Each point
P2,
group
112, 1
is
= EA DiA2" b'Ci
A'2
C'Bi
d'i
= DA*DiAi'B'C2'C'B2;
= DE A1A2 BiB2'CiC2.
'
also situated in three planes ITi
and in six planes
112, 2
for
in three other planes, of the
example, a"
is
a point
common
to tlie
twelve planes,
ABC, BCD, BCE;

b'c'Ai,
Each
line,
BiCiAi,
AB1C2C1B2,
B2C2A2,
Ai or A 0,2, contains one point
Db'BiC'Ci,
b'c'A2,
?>, i;
plane, 11 1 or 112,2, contains f^ree such points;
Eb'B2C'C2
BiCjDi,
B2C2D1.
but no line A2, 1 contains any. Each

and each plane 112,1 contains two.
CHAP.
GEOMETRICAL NETS
III.]
IN SPACE.
71
which are the intersections of opposite sides of a quadrilateral Q3

whereof the diagonals intersect in a point Po
of the quadrilateral B]B2C2Ci,
each other in the point
which
is
in
that plane,
example, the diagonals B1C2, B2C1

(by 90) in one of the planes 112,1, intersect'
:
for
while the opposite sides CiBi, B2C2 intersect in a"
and
the two other opposite sides, B1B2, C2Ci have the point d'i for their intersection.
The ten points P2, 1 are also ranged, three by three^ on ten lines of third construction
on the axes of homology^
A3, namely,
A"b'iC'i,
of ten pairs
of
triangles
which the centres
is
also
a'iA'2D'i,
homology are the ten points Pi

is the axis of homology of the two
:
inscribed in the former, with the point
ranged six by
namely
six,
A"b"c",
which are situated in the ten planes
in Fig. 29), to represent their centre of
platies Us,
a"b'2C'2,
of
a"b"c", in Fig. 21,
the latter
Ti, 72,
lli,
in that figure (replaced
The same ten points
homology.
and of
example, the dotted line

triangles, abc, a'b'c', whereof
for
by Di
P2,i are
and the ten
in the
last lines A3 are ranged four by four, in five

planes of homology of five pairs of pyramids, i?i, H^y
already mentioned (90) for example, the plane [e] contains (87) the six points
a"b"c"a'2b'2c'2, and the four right lines,
:
a"b'2C'2,
b"c'2A'2,
latter are the intersections of the four faces,
of the
pyramid abcd, with the corresponding
DCB,
DiCiBi,
DBA,
DAC,
DiAid,
of its inscribed homologue AiBiCiDi
A"b"c"4
c"a'2B'2,
which
ABC,
faces,
DiBiAi,
AiBiCi,
and are contained,
besides, in the four other
planes,
A2B'c',
B2C'a',
A2B2C2
C2A'b',
the three triangles, abc, AiBiCi, A2B2C2, for instance, being all homologous, although
in different planes, and having the line a"b"c" for their common axis of homology.
We may
also say, that this line
a"b"c"
is
the
common
trace (81) of two planes flojo,
namely of AiBiCi and A2B2C2, on the plane abc and in like manner, that the point
a" is the common trace, on that plane IIi, of îô lines A2,2, namely of BiCi and B2C2
;
being also the
common
trace of the
two
lines b'ic'i
and
b'2c'2,
which belong
to the
third construction.
(2.) On the whole, these ten points, of second construction, a". . ., may be
considered to be already well known to geometers, in connexion with the theory
of transversal^ lines and planes in space
but it is important here to observe,
:
with what simplicity and clearness their geometrical relations are expressed (88),
For example, the colby the quinary symbols and quinary types employed.
linearity (82) of the four planes, ABC, AiBiCi, A2B2C2, and [e], becomes evident
from mere inspection of their four symbols,
Compare the Note to page 68.

t The coUinear, complauar, and harmonic relations between the ten points,
which we have above marked as P2, 1, and which have been considered by Mobius
also, in
connexion with his theorv of nets in space, appear to have been

in a Memoir upon transversals.
by Carnot,
first
noticed
72
[OOOllJ,
[III2I],
[bOOK
I.
[11114],
[11112],
which represent (75) the four quinary equations,
w-v=0, x + + z-2w v=0,
x+i/ +
i/
w-2v=0,
with this additional consequence, that the ternary
+ y + z + M; -4i; =
symbol (81) of the common trace,
a?
of the three latter on the former, is [111]: so that this trace is (by 38) the line
A"B"c"of Fig. 21, as above. And if we briefly denote the quinary symbols of the
four planes, taken in the same form and order as above, by [i?o] [-î] [^2] [-^3]) we
see that they are connected
IR{\
whence
=-
by the two
[7?o]
[i22]
relations,
[i?3]
= 2 [i?o]
we
4- [
K2]
denote the planes themselves by IIi, 112, n'2, Ha,

the following value for the anharmonic of their pencil,
if
(UiUÛ'Ûs)
a result which can be very simply verified,
ramid, and E (comp. 29) is its mean point
=-2
for the case

:
we have (comp. 84)
the plane
when abcd
lis,
or [e],
is
a regular py-
becoming
in this
case (comp. 38) the plane at infinity, while the three other planes, abc, AiBiCi,
A2B2C2, are parallel ; the second being intermediate between the other two, but twice
as near to the third as to
\ihQ first.
We must be a little more concise in
(3.)
our remarks on the seven other types of
points P2, which indeed, if not so well

quite so interesting
in space
although
it
known,* are perhaps also, on the whole, not

seems that some circumstances of their arrangement
deserve to be noted here, especially as affording an additional exercise
may
(88), in the present system of symbols and types. The type P2, 2 represents, then, Sl group
oi thirty points, of which a"', in Fig. 21, is an example; each being the intersection
of a line A2,
with a line A2,2, as a'"
is
the point in which aa' intersects b'c'
but
each belonging to no other line, among those which have been hitherto considered.
But without aiming to describe here aZZ the lines, planes, and points, of what we have
called the third construction,
we may
already see that they must be expected to be
and that the planes 1X3, and the lines A3, of that construction, as well as
the pyramids i?2, and the triangles 2^2, of the second construction, above noticed, can
numerous
only be regarded as specimens, which in a closer study of the subject, it becomes neT2,x.
Accordingly it is
cessary to mark more fully, on the present plan, as lis, 1,
found that not only is each point P2, 2 one of the corners of a triangle Tz, 1 of third
.
is of a"'b"'c"' in Fig. 21), the sides of which new triangle are

passing each through one point P2, 1 and through two points P2, 2 (like
the dotted line a"b"'c"' of Fig. 21) ; but also each such point P2,2 is the intersection
construction (as a'"
lines A3,
of two
2,
new
lines of third construction,
A 3, 3, whereof each
connects a point pq with a
* It does not
appear that any of these other types, or groups, of points
hitherto been noticed, in connexion with the net in space, except the one
have ranked as the fifth,

struction
5,
and which represents two points on each
have
line Ai, as
has been seen to represent one point on each of those ten lines of first conbut ihai fifth group, which may be exemplified by the intersections of the
the type Pg,
line
P2,
P2,
which we
DE with
page 290 of
the two planes AiBiCi
his already cited work),
sult of a different analysis.
and A2B2C2, has been indicated by Mobius
(in
although with a different notation, and as the re-
CHAP.
GEOMETRICAL NETS
III.]
For example, the point
point P3,
V.
two new
lines, da'i,
congruent symbols,
ea's
We may
73
the common trace (on the plane abc) of the

we adopt for this point a"' the second of its two
a"' is
because, if
we have (comp.
A"'= (loOll)
IN SPACK.
73, 82) the expressions,
= (d) - (A'l) =
(e)
(A'2).
of concurrence (comp. the
therefore establish ih^ formula
A"' = aa' b'c' da'i
E a'o
first
sub-article)
Avhich represents a system of thirty such formulae.

(4.) It has been
remarked that the point a!"
represented, not only
may be
quinary symbol (21100), but also by the congruent symbol, (1001 1);
if
by the
then
we
write,
Ao = (11100),
new points AoBqCo,
Bo = (iriOO),
Co
= (11100),
must be considered to be syntypicaU

in the quinary sense (78), with the three points
a"'b"'c"', or to belong to the same
group P3, 2, although they have (comp. 88) a diflferent ternary type. It is easy to
see that, while the triangle a"'b"'c"' is (comp. again
Fig. 21) an inscribed homologue Ti,\ of the triangle a'b'c', which is itself (comp. sub-article- 1) an inscribed
these three
in the plane of abc,
horaologue To, 1 of a triangle Ti, namely of abc, with a"b"c" for their common axis
of homology, the new triangle AqBoCo is on the
contrary an exscrihed homologue
But from the syntyT-i, 2, with the same axis A3, 1, of the same given triangle Ti.
pical relation, existing as above for space between the points a'" and Ao, Ave may
expect to find that these two points P2, 3 admit of being similarly constructed, when
the Jive points Pq are treated as entering symmetrically
(or similarly), as geojnetrical elements, into the constructions.
The point Aq must therefore be situated, not
only on a line A2, 1, namely, on aa', but also on a line A2,2j which is easily found to
be A1A2, and on two lines A3, 3, each connecting a point Pq with a point P2,i which
latter lines are soon seen to be bb" and cc".
We may therefore establish the formula
;
of concurrence (comp. the last sub-article)
Ao = aa' A1A2 bb" cc"
and may consider the three points
Ao, Bq, Cq as the traces of the three lines AiAo,
while the three new lines aa'', bb", cc", which coincide in position
B1B2, C1C2
with the sides of the exscribed triangle AoBqCo, are the traces A3, 3 of three planes
112, 1, such as AB1C2B2C1, which pass through the three given points A, b, c, but do
:
not contain the Hnes A2,i whereon the six points P2,2 in their plane IIi are situated.
Every other plane IIi contains, in like manner, six points P2 of the present group ;
them and every plane 1X3,2 contains three ; each

through two such points, but each line A2,2 only through one.
But besides being (as above) the intersection of two lines A3, each point of this group
P3,2 is common to two planes Ui, four planes 112,1, and two planes 1X2,2; while
every plane
112,
contains eight of
line A2,i passing
is also a common corner of two different

triangles of
#AiVd construction, of the lately mentioned kinds Ts,! and 23,2, situated
respectively
in the two planes oi first construction which contain the
It may be
point itself.
added that each of the two points Po, 2, on a line A2, 1, is the harmonic conjvgate of
each of these thirty points
one of the two points pi, with respect to the point Pq, and to the other point Pi ou
thus we have here the two harmonic equations,
that line
(aa'dia'")
by which the
positions of the
two points
(adia'ao)
a'"
=-
1,
and A might be determined,
74
[bOOK
I.
third group, PojS, o{ second construction, consists (like the preceding group)
(5.)
of thirty points, ranged two hy two on the fifteen lines A 2,1, and six by six on the
ten planes ITi, but so that each is common to two such planes ; each is also situated
and on one line ^3,1 in which (by sub -art. 1)

these two last planes intersect each other, and two of the five planes ITs, 1 ; each
plane 112,1 contains /oMr such points, and each plane 112,2 contains three of them ;
in two planes 112,1, in two planes 112,2,
but no point of this group is on any line Ai, or A2,2- The six points P2,3, which
are in the plane abc, are represented (like the corresponding points of the last
group) by two ternary types, namely by (211) and (311) and may be exemplified
;
by the two following
which these
points, of
last ai'e the ternary
symbols
A'^ = A a' a"b"c" = AA' AiBiCi A2B2C2

Ai'^ = aa' d'ia'2A 1 = aa';b'ciC2-c'biB2.
three points of the first sub-group a'^
The
are coUinear
new
of the second sub-group are the corners of a

to the triangle abc,
and
but the three points Ai'^
triangle, Tz,z,
to all the other triangles in its plane
which
. .
homologous
which have been hitherto
and A0B2C2; the

common axis of homology and the sides
considered, as well as to the two triangles AiBiCi
former points being their
is
line of the three
of the
new
trian-
being the traces of the three planes (comp. 90) of homology of pyramids, [a], [b], [c] ; as (comp. sub-art. 2) the line a'^'b'^c"^ or a"b"c" is the common trace of the two other planes of the same group Ha, 1, namely of [d] and [e]. We
gle, Ai'^Bi'^Ci^^,
may
also say that the point A\^^ is the trace of the line A'iA'2
b'co, c'bo are the traces of the
may
two planes
112,2 in
and because the
which that point
is
lines
contained,
we
write the fonnula of concurrence,

Ai''
= AA'
a'ia'2 b'Co c'bo.
may be also remarked, that each of the two points P2, 3, on any line A2, 1, is
the harmonic conjugate of a point P2,2, with respect to the point Pq, and to one of
(6.) It
the two points Pi on that line
being also the harmonic conjugate of this
last point,
with respect to the same point Po, and the other point Pa, 2 thus, on the line aa'di,
we have the four harmonic equations, which are not however all independent, since
:
two
of
them can be deduced from the two
equations of the fourth sub-article
others,
with the help of the two analogous
(aa"'aV^) = (AA'AoA'') = (AAoDiAi'^) = (aDiA"'Ai'^) =:
1.
And
the three pairs of derived points Pi, P2,2, P2,3, on any such line A2, i, will
be found (comp. 26) to compose an involution, with the given point Pq on the line for
one of
its
two double points (or foci')
a point P3 of third construction
the other double point of this involution being
namely, the point in which the line A2,
one of the five planes of homology
ITa, 1,
which corresponds (comp. 90)
meets that
to the par-
Thus, in the present example, if we denote by A the

point in which the line aa' meets the plane [a], of which (by 81, 91) the trace on
ABC is the line [ill], and therefore is (as has been stated) the side Bi''^Ci'^ of the
lately mentioned triangle T3, 3, so that
ticular point Po as centre.
A''
we
shall
= (1 22) = aa'
BC'" Cb'"
Bi'^Ci"',
have the three harmonic equations,

(aa'a'^di)
which express that
this
new
= (aa"'aÂo) = (aa"'aÂi''') = -
point A^
is
the
common harmonic conjugate
of the given
CHAP.
GEOxMETRICAL NETS IN SPACE.
III.]
75
and
point A, with respect to the three pairs of points, a'di, a"'ao, a'Âi"';
that these three pairs form (as has been said) an involution, with
therefore
A and A* for
its
two
double points.
found that
(7.) It will be
we have now exhausted
all
the types of points of
there being only four

second construction, which are situated upon lines A2, i
But there are still to be considered two new groups
such points on each such line.
;
of points P2 on lines Ai, and three others on lines A2,2-
twenty points, situated
any
line
A2
common
Attending first
observe that each of the two new types,
we may
set of lines,
ttvo
Po,
4,
former
to the
ro,
5,
represents
by two on the ten lines oi first construction, but not on

s-ix in the ten planes IIi, each point however being
and therefore six by
to three such planes
also each point P2,4 is
common
to three planes 112,2,
one such plane while each of these last planes

If we attend only to points in the
contains three points P2, 4, but only one point P2, 5.
plane abc, we can represent these two new groups by the two ternary types, (021)
and each point
P2, 5 is situated in
and (021), which as symbols denote the two typical

A^
we have
It
Bi^'',
= BC*c'AiA2'DiAiBi*DiA2B2
points,
A^'= BC c'BiB2 = BC
c'bq
also the concurrence,
may
which
be noted that A^
A^'
= BC
is
the harmonic conjugate of
on the same trace
last point is
c'Ao DiC" Ab'".

c',
with respect
c'Aq, of the plane c'aiA2
to
Ao and
and that
a"^' is
with respect to c' and Bq, on the trace of the plane

c'biB2, where Bi^ denotes (by an analogy which will soon become more evident) the
so that we have the two equations,
intersection of that trace with the line ca
harmonically conjugate to
Bi^",
(aoc'bi'â^)
(8.)
Each
= (boBi^C V') = -
1.
contains thus two points P2, of each of the two last
line Ai,
new
groups, besides the point Po, 1, the point Pi, and the two points Po, which had been
previously considered it contains therefore eight points in all, if we still abstain (88)
:
from proceeding beyond the Second Construction. And it is easy to prove that these
distinct modes, be so arranged as to form
eight points can, in two
(comp. sub-art, 6)
an involution, with two of them for the two double points thereof. Thus, if we attend
only to points on the line bc, and represent them by ternary symbols, we may write,
o=(001),
B=:(010),
a^'
Av=(021),
(021),
a'=(011),
Aiv
= (012),
A"=(On);
Ar' = (0r2);
and the resulting harmonic equations

I.
II.
will
then
suffice to
= (bA^CA'Ô = (BArCAi^') = I,
(a'ba'c) = (aVa"ai^) = (a'a^'a"ai^') = -
(ba'ca")
show
1st.,
I,
that the two points Pq, on any line Ai, are the double
points of an involution, in which the points Pi, Piriform one pair of conjugates,
while the two other pairs are of the common form, P2,4, P2,5
and Ilnd., that the
;
two points Pi and P2, 1, on any such line Ai, are the double points of a second invoAlso
lution, obtained by pairing the two points cf each of the three other groups.
each of the two points Pq, on a line Ai, is the harmonic conjugate of one of the
two points
Po, j
to the point Pi
on that
line,
on the same
with respect
line
thus.
to the other point of the
same group, and
76
(ba'ai''a^)= (ca'a"ai")
three other
(9.) It remains to consider briefly
=-
I.
[bOOK
1.
groups of points
P2,
each group
containing sixty points, -which are situated, two by two, on the thirty lines A2,2, and
six by six in the ten planes TIi.
Confining our attention to those which are in the
plane abc, and denoting them by their ternary symbols, we have thus, on the line
new
b'c', the three
with which
same
may
line b'c'
typical points, of the three remaining groups, P2,6, P2,7, r2,8:
A = (23T);
A^"' = (321);
A"' = (121);
be combined these three others, of the same three types, and on the
Ai^"
= (112);
= (312);
Ai^"'
Ai = (213).
Considered as intersections of a line A2,2 with lines A3 in the same plane Hi, or with
planes 112 (in which latter character alone they belong to the second construction),
the three points a^", &c., may be thus denoted :
= b'c'*bb"*cb"''Aa"^'=b'c' BC1A2A1C2
b'c' DiCiAi D1C2A2
g'(,' j5^b" ab"'av
A** = b'c' a'CqBi'^Ci^B^' BA'^Bi^^Bi'^^ = b'c' a'CiCs
A^"
^vni _
"
with the harmonic equation,
=-
(coa'ciâ'^)
1,
and with analogous expressions for the three other points, Ai^", &c. The line b'c' thus
it
intersects one plane ITo, i (or its trace bb" on the plane abc), in the point a'^^
intersects two planes 112,2 (or their common trace Dib") in A"^"" ; and one other plane
;
n2,2 (or
its
trace a'co) in a'^
Each plane
three groups.
112,
and similarly
i
for the other points, Ai"^", &c., of the
contains twelve points P2,6, eight points
P2,
7,
same
and eight
points P2,8; while every plane 112,2 contains six points P2,6, twelve points P2,7,
and nine points Pojs- Each point P2,6 is contained in one plane IIi ; in three
planes 112,1; and in two planes n2,2planes
112,1,
and
in four planes 112,2.
in two planes 112,1,
(10.)
each such
The
and
Each point
And
Po,? is in one plane IIi, in
each point P2,8
is
two
situated in one plane ITi,
in three planes 112,2.
points of the three last groups are situated onlg on lines A2,2
hut,
on
two points of each of those three groups are situated ; which, along
with one point of each of the two former groups, P2, 1 and P2,2j and with the two
points Pi, whereby the line itself is determined, make up a system of ten points upon
that
line.
line,
For example, the
line b'c' contains, besides the six points
the last sub -article, the four others
b'=(101);
Of
mentioned in
a" = (011);
c'=(110);
these ten points, the two last mentioned,
a"'=(211).
namely the points
P2,
and
P2, 2
upon the
are the double points (comp. sub-art. 8) oi Sinew involution, in which the two
points of each of the four other groups compose a conjugate pair, as is expressed by
line A2,
2,
the harmonic equations,
(a"b'a"'c')
And
(a"aâ"'ai^")
(a'V-"'a"'ai^"')
= (a"a'a"'ai) =-l.
the analogous equations,
(bV'c'a'")
show that the two
= (b'a^'"c'a^''") =
points Pi on
any line
A2,
(b'ai^"c'ai^'")
=- 1,
are the double points of of another invo-
lution (comp. again sub-art. 8), whereof the
two points
P2,i, P2,2
on that
line
form
CHAP.
GEOMETRICAL NETS
III.]
one conjugate
pair,
IN SPACE.
while each of the two points P2,6
In
is
fact, the
77
paired with one of the points
eight-rayed pencil (a c'b'a"'a"aÂÂxÂi^")

coincides in position with the pencil (a . bca Wa^'AiÂiÔ> ^^d may be said to be
a pencil in double involution ; the third and fourth, the fifth and sixth, and the seventh and eighth rays forming one involution, whereof the first and second are the
P2,
as its conjugate.
sixth
; while the first and second, the fifth and seventh, and the
and eighth rays compose another involution, whereof the double rays are the third
and fourth of the pencil.
two double* rays
(11.) If
we
proceeded to connect systematically the points P2 among themselves,

of
Po, we should find many remarkable lines and planes
and with the points Pi and
third construction (88), besides those which have been incidentally noticed above
example,
we should have a group
1X3,2 of
for
twenty new planes^ exemplified by the
two following,
[E]
= [11103],
[dJ=
[11130],
which have the same common trace A3, 1, namely the line a"b"c", on the plane abc,
as the two planes AiBiCi, A2B0C2, and the two planes [d], [e], of the groups 1X3,2 and
lis,
1)
1X3, 2
which have been considered
would be found
three points P2,
in former sub-articles
and each of these new planes

1, six points P2, 2, and
to contain one point Po, three points P2,
might be proved also that these twenty new planes are the
new pyramids B3, which are the exscribed homologues of the five
It
3.
tioenty faces oifive
old pyramids Ri (89), with the five given points Pq for the corresponding centres of
homology. But it would lead us beyond the proposed limits, to pursue this dis-
cussion further
although a few additional remarks
may
be useful, as serving to
establish the completeness of the enumeration above given, of the lines, planes,
and
points of second construction.
93. In general, if there be any n given points^ whereof no four

are situated in any common plane, the number JV of the derived
which are immediately obtained from them, as intersections

drawn through two of the
and
each
three
plane through
given points,
others), or the number of
points of x\iQform ab'CDE, is easily seen to be,
points,
A n
of line with plane (each line being
_
^"^'^^"
so that
n{n-\){n-2){n-^){n-A)
2.2.3
10, as before,
formula to the case
n=
when n = 6.
15,
we should
But
if
we were
to
find, for that case,
apply this
the value,
i\r=/(i5)=15.l4. 13.11 = 30030;

and thus fifieen given and independent points of space would conduct,
by what might (relatively to them) be called a First Construction
(comp. 88), to a system of more than thirty thousand points. Yet it
has been lately stated (92), that from the fifteen points above called
Po> Pi,
there can be derived, in this way, only two hundred and ninety
*
Compare page 172
of the
Geom.
Snpe'rieure of
M.
Chasles.
78
points
than
P3,
as intersections of the
That
three hundred.
points, at the
[bOOK
I.
form* A'll; and therefore /ezcer

number of derived
this reduction of the
end of what has been called (88) the Second Construcfrom the dependence of the ten points
tion for the net in space, arising

Pi
on theîê points
Pq,
would be found
to
be
so considerable,
might
not perhaps have been anticipated; and although the foregoing examination proves that all the eight types (92) do really represent
points P2, it may appear possible, at this stage, that some other tyiê
of such points has been omitted.
study of the manner in which
the types of points result, from those of the lines and planes of which
they are the intersections, would indeed decide this question ; and
was, in fact, in that way that the eight types, or groups, Pg, 1, .P2,8)
of points of second construction for space, were investigated, and
found to be sufficient: yet it may be useful (compare the last sub-
it
art.) to verify, as
(1.)
below, the completeness of the foregoing enumeration.
Th^ fifteen points,
Po, Pi,
admit of 105 binary^ and of 455 ternary combina-
from determining so many distinct lines and planes. In fact,

those 15 points are connected by 25 collineations^ represented by the 25 lines Ai,
Az,!*, which lines therefore count as 75, among the 105 binary combinations of
tions; but these are far
points
and there remain only 30 combinations of
which are constructed
this sort,
by the 30 o^êr lines, A2,2- Again, there are 25 ternary combinations of points,
which are represented (as above) by lines, and therefore do not (determine any plane.
- 6) triangles T\, T2, because
Also, in each of the ten planes IIi, there are 29 (=35
each of those planes contains 7 points Pq, Pi, connected by 6 relations of collinearity.
In like manner, each o{ the fifteen planes 112,1 contains 8 (=10 2) other triangles
To, because
it
contains 6 points Po, Pi, connected by two collineations.
There re-
20 (=455 25 290-120) ternary combinations of points to

The completebe accounted for; and these are represented by the 20 planes 112,2.
ness of the enumeration of the lities and planes of the second construction is therefore
main
therefore only
verified; and it only remains to verify that the 305 points, Pq, Pi, P2, above considered, represent all the intersections A 11, of the 55 lines Ai, Ao, with the 45 planes
Hi, Ha.
(2.)
Each plane
Aojzj each plane
ITo,
IIi
1
contains three lines of each of the three groups, Ai, A2,
contains two lines
A 2,1, and
1,
four lines A2,2; and each plane
Hence (or because each line Ai is contained in three

two planes ITi, and in two planes ITo,!; and each
two planes 172, 1, and in two planes 112, 2), it follows that,
rro,2 contains three lines A2,2
planes IIi
each line
-A2,i
jine A2, 2 in one plane IIi, in
in
without going beyond the second construction, there are 240 (= 30
The
planar
definition (88) of the points P2 admits, indeed, intersections
lines,
when they
are not already points Po or Pi
are also points of the form A* IT
so that no generality
to this last form, as in the present discussion
we
but
is lost,
all
30
+ 30 + 30
A'A of corn-
such intersections
by confining ourselves
propose to do.
CHAP.
+ 60 + 60)
cases
intersection
(3.)
is
GEOMETRICAL NETS
III.]
is
o/ coincidence of
line
reduced, hereby, from
IN SPACE.
and plane;
79
so that the
65.45 = 2475,
to
number
2235 (= 2475
of cases
- 240).
of
Each point Tq represents twelve intersections of the form Ai-ITi because it

to four lines Ai, and to six planes Hi, each plane containing two of those
;
common
four lines, but being intersected
ABC, for example,

point Po
is
common
is
by the two
A by
intersected in
to three planes 1X2,
1,
others in that point Pq
the two lines,
ad and
ae.
as the plane
Again, each
no one of which contains any of the four
Ai through that point ; it represents therefore a system of twelve other intersections, of the form Ai IIo, 1.
Again, each point Pq is common to three lines Ao, 1,
lines
contained in two of the six planes ITi, but intersects the four others
in that point Po
which therefore counts as twelve intersections, of the form A2, 1 ITi.
Finally, each of the points Po represents three intersections, A2,i'n2, 1; and it reeach of Avhich
is
presents no other intersection, of the form
within the limits of the present
IT,
Thus, each of the five given points is to be considered as representing, or

inquiry.
constructing, thirty-nine (= 12 + 12 + 12+3) intersections of line with plane; and
2040 (=2235 195) other cases of such intersection A- IT, to be

accounted for (in the present verification) by the 300 derived points, Pi, P2.
(4.) For this purpose, the nine columns, headed as I. to IX. in the following
there remain only
Table, contain the numbers of such intersections
which belong respectively
to the
nine forms,
Ai'iii,
Ai'iia,!,
Ai-n2,2;
A2,i'ni,
A2,i'n2,i,
A2,i"n2,2;
A2,3*TIi,
A2,2*n2,
A2,2*n2,2,
1,
each of the nine typical derived points, a'

A'^, of the nine groups Pi, P2,i,
Column X. contains, for each point, the sum of the nine numbers, thus tabuPo.sfor
lated in the preceding columns
and expresses therefore the
entire
number
of inter-
which anyone such jsotn^ represents. Column XI. states the number of the
points for each type ; and column XII. contains ihQ product of the two last numbers, or
the number of intersections A 11 which are represented (or constructed) by t\\Q group.
sections,
Finally, the stim of the
numbers
in each of the
and because the 300 derived points,
two last columns
is
written at
its foot
and second constructions, are thus found

to represent the 2040 intersections which were to be accounted for, the verification is
seen to be complete : and no new type, of points P2, remains to be discovered.
(5.)
of first
80
(G.) It
is
to be
[bOOK
remembered that we have not admitted^ by our
Siny points "which can only he determined
as belonging to the second construction
by
intersections
nor have
we
I.
definition (88),
of three planes
counted, as lines
A2
IIi,
02,
of that
any lines which can only be found as intersections of two such planes.
For example, we do not regard the traces aa", &c., of certain pZanes A2, 1 considered
in recent sub-articles, as among the lines of second construction, although they would
construction,
And any
present themselves early in an enumeration of the lines A3 of the third.
point in the plane abc, which can only be determined (at the present stage) as the
intersection of two such traces,
it
however
not regarded as a point P2.

A student might find
an exercise, to investigate the expressions for such
is
to be not useless, as
and for that reason it may be noted here, that the ternary types (comp.
81) of the forty-four traces of planes ITi, 112, on the plane abc, which are found to
compose a system of only twenty-two distinct lines in that plane, whereof nine are
intersections
lines Ai, A2, are the seven following
[100],
[Oil],
(comp. 38)
[Oil],
[111],
["iuj,
which, as ternary symbols, represent the seven

BC,
aa',
[211],
[211];
lines,
aa",
a"b"c",
b'c',
Dia'',
a'co.
Again, on the same principle, and with reference to the same definition, that
point, say f, which may be denoted by either of the two congruent quinary
(7.)
new
symbols (71),
(43210)
(01234),
and which, as a quinary type (78), represents a neio group of sixty points of space
(and of no more, on account of this last congruence, whereas a quinary tj-pe, with all
five coefficients unequal, represents generally a group of 120 distinct points), is
not regarded by us as a point P2; although this new point f is easily seen to be the
its,
intersection of three planes
which
all
of second construction, namely, of the three foUoTAnng,
belong to the group
lis,
[oiITi],
or aa'diCiB2, cc'diBiA2, eb'b2c'c2.
each plane
common
94.
points
112,
(as
is
[iioli],
It
[iiTio],
may, however, be remarked
contains twelve points P3 of this
new group
in passing, that
every such point being

evident from what has been shown) to three such planes.
1
From
Po,
the foregoing discussion it appears that the Jive given

three hundred derived points Pj, Pg, are arranged in
and the
upon the fifty -five lines A^, A^, and in 1\\q forty-five planes Hi,
Each line Ai contains eight of the 305 points, forming
on it what may be called (see the sub-article (8.) to 9'^) a double inEach line Ao, contains seven points, whereof one, namely
volution.
space,
Ilo,
as follows.
the given point, Po, has been seen (in the earlier sub-art. (6.)) to be
a double point oi another involution, to which the three derived pairs
And each line A^,^ conPi, p.^, on the same line belong.
tains ten points, forming on it a new involution; while eight of these
ten points, with a different oj^der of succession, compose still another
of points,
CHAP.
GKOMETIIICAL NETS IN SPACE.
III.]
81
involution* (92, (10.))Again, each plane 11, contains ^^^-;i6'o

three
given points, four points ofjirst, and 45 points
points^ namely
Each plane IT..,! conta.ins forty-seven points,
o[ second construction.
v/hereof o
is
a given point, four are points
and 42 are points
Pi,
* These theorems
respecting the relations of involution, of given and derived
points on lines oi first and second constructions, for a net in space, are perhaps
new
although some of the harmonic relations, above mentioned, have been noticed under
other forms by Mobius
to wliom, indeed, as has been stated, the conception of such
:
a net
is
due.
Tiius, if
we
consider (compare the Note to page 72) the two intersec-
tions.
El
we
easily find that they
= DE
may
El
AiBiCi ,
= DE
'
A2B2C2,
be denoted by the quinary symbols.
E2= (00021);
(00012),
they are, therefore, by Art. 92, the
by the theorem
E2
two points
P2, 5
on the line de
and consequently,
stated at the end of sub-art. 8, the
harmonic conjugate of each, taken

must be one of the two points d, e on
with respect to the other and to the point Di,

line.
Accordingly, we soon derive, by comparison of the symbols of these ^t>e
points, DED1E1E2, the two following harmonic equations, which belong to the same
that
type as
tlie
two last of that sub-art. 8
-1
(D1DE2E1) =
(DiEEiEi)
=-
but these two equations have been assigned (with notations slightly different) in the
formerly cited page 290 of the Barycentric Calculus.
(Corap. again the recent Note
The geometrical meaning
to page 72.)
conceiving that
abcd
is
of the last equation

a regular pyramid, and that E is its
be illustrated, by
mean point; for then
base abc DiD is the altitude
(comp. 92, sub-art. (2.) ), Di is the mean point of the

of the pyramid ; and the three segments DiE, DiEi, DiEo
may
;
are, respectively, the
quar-
the third part, and the half of that altitude they compose therefore (as the formula expresses) a. ha7-monic progression ; or Di and Ei are conjugate points, with
ter,
But in order to exemplify the double involution of the same

E2would be necessary to consider three other points P2, on the same line
de; whereof one, above called d'i, belongs to a, known group P2,i (92, (2,)); but
the two others are of the group Po,!, and do not seem to have been previously noticed.
As an example of an involution on a line of third construction, it may be remarked
respect to
e and
sub-art. (8.),
it
that on each line of the group A3, 3, or on each of the sides of any one of the ten triangles T3, 2, in addition to one given point po, and one derived point P2, 1, there are
two points
P2, 2, and two points P2, e ; and that the two first points are the double
points of an involution, to which the two last pairs belong : thus, on the side
Aqbco of the exscribed triangle AqBoCo, or on the trace of the plane BC1A2A1C3, we
have the two harmonic equations,

(baob"co)
= (ba"'b"ci^") = -
1.
Again, on the trace a'co of the plane a'ciCj, (which latter trace is a line not passing
through any one of the given points,) Oo and ei'^ are the double points of an involution, wherein a' is conjugate to cr and a'* to b^'.
But it would be tedious to
luultij-ly
such instances.
82
[book
I.
P2: of which last, 38 are situated on the six lines A2 in the plane, but
four are intersections of that plane 11^,, 1 with/owr other lines of second
construction.
Finally, each plane 112,2 passes through no given
point,
but cont^m^ forty-three derived
of second construction.
And
points,
whereof 40 are points
because the planes oi first construc-
tion alone contain specimens of all the ten groups of points, Pq, Pi,
^218?
P2i>
given or derived, and o^ all the three groups of lines, A^,
â i> ^2i 2j at the close of that second construction (since the
types
Ai are not represented by any points or lines in any plane
nor are the types Pq, A^, A2,i represented in a plane 112,2), it
P214J P2>5j
ITg,!,
has been thought convenient to prepare the annexed diagram (Fig.

30), which may serve to illustrate, by some selected instances, the
arrangement oi the fifty -two points Pq, Pi, P2 in a plane Ilj, namely, in
the plane abc ; as well as the arrangement of the nine lines A,, A,
in that plane,
and the
traces
A3 of other planes upon
it.
View of the Arrangement of the Principal Points and Lines
in
a Plane
of First Construction.
In this Figure, the triangle abc is suppposed, for simplicity, to be the equilateral
hose of a regular pyramid abcd (comp. sub-art. (2.) to 92) ; and Di, again replaced
inscribed triangle, a'b'c',
by o, is supposed to be its mean point (29). The first
therefore, bisects the three sides
finity (38)
the number
1,
and the axis of homology a''b"c"
on the line
c'b'
is
the line at in-
prolonged, being designed to suggest that
GEOMETRICAL NETS
CHAP, m.]
IN SPACE.
83
the point a", to which that Hue tends, is of the type P2,i, or belongs to the ^rs^
group of poiuts of second construction. A second inscribed triangle, a"'b"'c"', for
which Fig. 21 may be consulted, is only indicated by the number 2 placed at the
middle of the side
group of points
b'c',
P2.
to suggest that this bisecting point a'" belongs to the second
The same number
2,
but with an accent,
2', is
placed near the
corner Aq of the exscribed triangle AoBoCq, to remind us that this corner also belongs
(by a syutypical relation in space) to the group P2, 2- The point A'^, which
by the number 3, on the dotted line at the top
infinitely distant, is indicated
the
same number with an
accent, lower
Finally, the ten other numbers,

8,
8',
And
mentioned,
may
now
while
down, marks the position of the point Ai".
unaccented or accented,
denote the places of the ten points, a^, Ai%
A'*, Ai"^.
is
;
4, 4',
5, 5',
6, 6',
7, 7',
a", a^"*
the principal harmonic relations, and relations of involution, above
be verified by inspection of this Diagram.
a'', Ai^', a^", Ai^",
95. Hovêver far the series of construction of the net in space

continued, we may now regard it as evident, at least on com-
may be
parison with the analogous property (42) of the plane net, that every
which such constructions can conduct, must
point, line, or plane, to
necessarily be rational (77); or that it must be iâtionalhj related to

the system o^ th^ five given points : because the a/iAarmom'c co-ordinates (79, 80) of every net-point, and of every net-plane, are equal or
proportional to whole numhei'S.

Conversely (comp. 43) every point,
or plane, in space, which is thus rationally related io the system of
line,
points ABODE, is a point, line, or plane of the net, which those five points
determine. Hence (comp. again 43), every irrational point, line, or
plane (77),
is
indeed incapable of being rigorously constructed, by any

but it admits of being inde;
processes of the kind above described
approximated to, by points, lines, or planes of the net. Every

anharmonic ratio, whether of a group of net-points, or of a. pejicil of
finitely
or of net-planes, has a rational value (comp. 44), which depends only on the processes of linear construction employed, in the
generation of that group or pencil, and is entirely independent of the
net-lines,
arrangement, or configuration, of the five given points in space. Also,

of collineation, and of complanarity, are p>reserved, in the
all relations
passage from one net to another, by a change of the given system of

points: so that it may be briefly said (comp. again 44) that all geometrical nets in space are homographic figures. Finally, any
five points*
of such a net, of which no four are in one plane, are sufficient
(comp.
These general properties (95) of the space-net are in substance taken from
Mobius, although (as has been remarked before) the analysis here employed
appears
new
most of the theorems above given, respecting thepoints of second

construction (92), at least after we pass beyond the Jirst group po, i of ten such
points
which (as already stated) have been known comparatively long.
to be
as do also
84
45) for the determination of the
[boOK
I.
ivlwle net: or for the linear constrnc-
tion of all its points, including the five given ones.

let the five points AiBiCiDi and e be now supposed to be
be required to derive the four points abcd, by linear constructions,
from these new data.
In other words, we are now required to exscribe a pyramid
As an Example,
(1.)
given
and
let it
ABCD
to a given pyramid AiBjCiDi, so that

point E for their given centre of homology.
may be homologous thereto, with the

An obvious process is (comp. 45) to irt.
it
scribe another homologous pyramid, A3B3C3D3,, so as to have A3
and then
AaBsCa
= eai'BiCiDi,
&c.
to determine the intersections o^ corresponding faces, such as AiBiCi
and
these/owr lines of intersection will be in the common plane [e], of homology

and will be the traces on that plane of the/owr sought planes,
for
of the three pyramids,
ABC,
(fee,
drawn through the
construct one corner
a"
called
of
four given points Dj, &c.
exscribed pyramid,
tlie
If
it
we might
were only required to

above
find the point
connnon intersection of three planes, as follows,
as the
a" = AiBiCi AiUiE
and then should have
this other
A3B3C3
formula of intersection,
A = EArwiA'^.
Or
A might be determined by
the point
the anharmonic equation,
(EAA1A3)
which
for
(2.)
a regular pyramid
As
easily verified.
is
regards the general passage from one net in space to another,
symbols Pi
= 3,
{xi
are complanar
v{),
and
P5
let the
denote any Jive given points, whereof no four

and u be six coefficients, of which the five ratios are
(ajg
let a'b'c'd'e
v^)
such as to satisfy the symbolical equation (comp. 71, 72),

a'(Pi)
+ b'(Fd +
or the five ordinary equations

a'xi
Let
p'
(p')
by
it
e'xs
= a;a'(Pi) + y&
precisely the
the point
from
d'(Pi)
+ e'(P5) = - '( U)
includes, namely,
.
= a'vi +
+ evt = -
'.
(P2)+ ?c'(p3) + wd'ivC) + ve{y-^û\
found that this
will be
same
last point p'
{xyzwv)
ABCDE
from A
thus
;
is
can be derived ixora. the
constructions, as those
rived from the five points ABCDE.
is
it
be any sixth point of space, of which the quinary symbol satisfies the equa-
tion,
then
c'(P3)
which
P5
= {xyzwv) is deif ' = a;4-y + z+ tr 3r, then
by which the
As an example,
U)
five points Pi
point
derived from AiBjCiDjE, by the same constructions as (xyzwv)

itself may be constructed from Ai . E, as the point p =
(30001)
.
which would conduct anew
to the
anharmonic equation of the
last
sub-article.
(3.) It
may
be briefly added here, that instead of anharmonic ratios, as con-
nected with a net in space, or indeed generally in relation to spatial problems, we

are permitted (comp. 68) to substitute products (or quotients) oi quotients of volumes
of pyramids; as a specimen of
harmonic
tion,
which
substitution,
it
may
be remarked, that the an-
admits of being replaced by the following equainvolving one such quotient of pyramids, but introducing no auxiliary point :
relation, just referred to,
CHAP.
MEANS OF VECTORS.
III.]
KA
In general,
space,
if
AiA
3eBiCiDi AiBiCiDi.
:
xi/zw be (as in 79, 83) the anharmonic co-ordinates of a point p in
we may
write,
pbcd ebcd
1/
pcda'ecda'
with other equations of the same type, on which
Section
85
5.
On
we cannot
here delay.
Barycentres of Systems of Points
Simple and Complex Means of
when
96. In general,
sum 2a of any number
the
and on
Vectors,
of co-initial
vectors,
ai
is
= OAi,
Ow
0A,,
divided (16) by their number^ m, the resulting vector,
a = OM = 2a = 2oA,
m
of those m
m
is
in
be the Simple Alean
said to
which
mean
this
(comp. 18)
vectors
and the point
m,
and of which the position
vector terminates,
easily seen to be independent of the position of the
is
common
origin o, is said to be the Mean Point (comp. 29), of the

It is evident that we have the equaA;;,.
system of the
points, Ai,
tion,
= ("i - y") +
or that the
sum of the
+ (a, -
/t)
= 2 ( - /*) = 2ma
m vectors, dravvn/7'07?z the
mean point
to
1,1,
the
points A of the system, is equal to zero. And hence (comp. 10, 11, 30),
it follows, 1st., that these
vectors are equal to the m successive
sides of a closed
polygon
Ilnd., that if the system
and
mean
its
point be projected, by any parallel ordinates, on any assumed plane

(or line), the projection m^, of the mean point m, is the mean point of
the projected system
point,
is
the
mean
and
Illrd., that the ordinate mm', of the
of all the
otJier ordinates,
mean point
lows, also, that if n be the
AjA'i,
a^â'^.
of another system,
mmn
It folBj, ..
b;
and if s be the mean point of the total system, Ai

b, of the m + n
= s points obtained
by combining the two former, considered as partial systems ; while V and a
may denote the vectors, on and os, of
.
these two last
mean
points
mfi=â,
wi/
m{aso that the general
iii)
mean
then we shall have the equations,
= 2^,
= n{vpoint,
5CT
71 J/,
7/i.MS =?z.SN;
a),
s,
= 2a + 2/3 = 771/4 +
is
situated on the right line mn,
which connects the two partial mean points,
M and
n; and divides
86
ms and
that line (internally), into two segments
As an Example,
more fully, let
(I.)
point
two whole numbers^
proportional to the
let
abcd be a gauche
[boOK
which are
sn,
and
I.
inversely
n.
quadrilateral, and let
E be
its
mean
or
OE = ^ (oA + OB
-f-
oc + od),
or
that
is to
say, let
a=6
=c=
rf,
= H + i3 + y +
lately used, for certain derived points di, &c., if
= ai = i(/3 + y +
OA3=a2 = |(a + ^),
OA' = a'=|(/3 + y),
OAi
we
have seven
shall
And
Then, with notations
write the vector formulcB^
î=i(a + /3+y),
= Kr + ^).
y2
r'=K+/3),
we
mean
vector, e
namely, the follow-
= ^(a + 3ai) = ..= i(5+3î)

= K'+ 2)=. .=i(r' + r2).
these conduct to the seven equations between segments,
AE =
a'e
which prove (what

AAi,
5),
.
different expressions for the
ing:
5);
in the equations of Art. 65.
is
3eai,
= eA2,
otherwise
DE = 3edi
known)
c'e=ec2;
that the four right lines, here denoted by

pyramid abcd with the mean
DDi, whereof each connects a corner of the
point of the opposite face, intersect and quadrisect each other, in one common
e and that the three common bisectors a'ao, b'bo, c'ca, of pairs of opposite
edges, such as BC and da, intersect and bisect each other, in the same mean point
point,
so that the /our middle points,
c', a', C2,
A2, of the four successive sides
ab, &c,, of
common plane, which

AC and bd.
bisects also
the gauche quadrilateral abcd, are situated in one

the
common
bisector, b'B2, ofthe^zô diagonals,
(2.) In this
D being ybwr, the number

example, the number s of the points A
which thus cross each other in their general mean point E is seen
.
of the derived lines,

to be seven
namely,
and the number
in the notation lately
of the derived planes through that point is nine :

used for the net in space, four lines Ai, three lines A2, 1,
Of these nine planes, the six former may (in

IIi, and three planes 112,1.
the present connexion) be called triple planes, because each contains three lines (as
six planes
the plane abe, for instance, contains the lines aai, bbi, c'c2), all passing through the
point e; and the three latter may be said, by contrast, to h^non triple planes,
mean
because each contains only two lines through that point, determined on the foregoing
principles.
(3.)
In general,
let
() denote the
point s of a total system of

posed into partial systems
s
;
number of the lines, through the êera/ mean
given points, which is thus, in all possible ways, decomlet/(s) denote the number of the triple planes, obtained
by grouping the given points into three such partial systems let (s) denote the
number of non-triple planes, each determined by grouping those s points in two dif//
ways into two partial systems and let f() =f(s) + ^ ()

number of distinct planes through the point s so that
ferent
represent the entire
0(4) -7,
/(4) =
6,
//(4)
3,
f(4)
9.
CHAP.
Then
new
lines,
we
if
according as
partial systems,
namely cs
87
we introduce a new point c, each old line mn furwe group the new point with one or other of the
(M) aud (A') and that there is, besides, one other new line,
easy to perceive that
it is
nishes two
two old
MEAN POINTS OF SYSTEMS.
111.]
have, therefore, the equation infinite differences,
which, Avith the particular value above assigned for 0(4), or even with the simpler
and more obvious value, 0(2)= 1, conducts to the general expression,
<p(s)
= 2^-^-1.
(Af) (iV) (P) be any three partial systems, which jointly make
(4.) Again,
up the old or given total system (^S) and if, by grouping a new point c with each
of these in turn, we form three new partial systems, (M') (iV) (P); then each
if
old triple plane such as mnp, will furnish three new triple planes,
MNP,
while each old
line,
We
plane be obtained in any other way.

ferences
/( +
l)
0(sf
we
then
CKL
nor can any new triple
have, therefore, this
new equation
in dif-
But we have seen that

if
mnp
mn'p,
kl, will give one new triple plane,
write, for a
l)
= 20(s) +
l;
moment,
/(s)
we have
= 3/ + 0(*).
+ 0()=x(),
this other equation in finite differences,
=
x(* + l)
3xW +
Also,
/(3)-l,
0(3)
therefore,
2x()
and
2/(s)
(5.) Finally,
it is
clear that
l.
x(3) = 4:
3,
3--l,
= 3--2*+l.
we have
the relation,
3/W + ^(*)=i0W. (0(0-
l)
= (2-'-l)
(2-2-1);
because the triple planes, each treated as three, and the won-triple planes, each treated
as one, must jointly represent all the binary combinations oHhe lines, drawn through
the
point s of the whole system.
mean
24/()
Hence,
= 22-2+
3.2S-I- 3-
and
=
f(s) 22-3+2*-2-3->;
so that
f(s+1)-4f(s) = 3-1-2i,
and
^(+1)-4;//()
which
last equation in finite difierences
= 3/(*);
admits of an independent geometrical inter-
pretation.
(6.)
For instance, these general expressions give,
0(5)=15;
80 that if
/(5) = 25;
we assume a gauche
^(5) = 30;
pentagon, or a system
f(5) = 55;
offve points
in space,
e.
88
[boOK
I.
this system, there will in general be a set oi fif-
and determine the inean point F of
teen lines, of the kind above considered, all passing through this sixth point f
and
these will be arranged generally in fifty-five distinct planes, vf'hereo{ tiventy-five will
be what we have called triple, the thirty others being of the non-triple kind.
More
97.
and
co-initial vectors,
lars (17),
if "i
generally,
and
then that new
., be, as before, a
,
from these by the formula,

a,a +
+ a,,,a,,,
.
B=
a,
system of
given
a^he any system of m given scaco-initial vector ^, or ob, which is deduced
.
if a^,
2aa
2a
a.
or OB
2aoA
2a
or by the equation
2a(a-/3) = 0,
or
2aBA =
0,
m given vectors , or
with
that system of given
as
considered
OA,
affected (or combined)
scalars, a, as coefficients^ or as multipliers (12, 14). It may also be said
may
be said to be the Complex
Mean
of those
that the derived point b, of which (comp. 96) the position is indeo, is the Barycentre (or centre ofgravity)
pendent of that of the origin
of the given system of points Aj
.,
considered as loaded with the
intersections of lines and planes

;
given weights ai
of
these
from
the
complex means, or harycentre^, of
arise,
comparison
.
and theorems of
which are entirely analogous to those lately

considered (96), for simple means of vectors and of points.
partial and total systems,
(1.)
As an Example,
in the case of Art. 24, the point
is
the barj-centre of the
system of the two points, A and B, with the weights a and 6; while, under the conditions of 27, the origin
weights a,
b, c
and
given points A, b,
c,
we
when
if
is
the barycentre of the three points A, b, c, with the three
use the formula for p, assigned in 34 or 36, the same three

loaded with xa, yh, zc as weights, have the point p in
Again, with the equations of 65, e is the barycensystem of the /bwr given points, A, b, c, d, with the weights a, b, c, d
the expression of 79 for the vector op be adopted, then xa, yb, zc, wd are
their plane for their barycentre.

tre of the
and
if
equal (or proportional) to the weights with which the same four points a
In
be loaded, in order that the point p of space may be their barycentre.
.
d must
all
these
thus proportional (by 69) to certain segments, or areas, or

and what we have called the
volumes, of kinds which have been already considered
anharmonic co-ordinates of a variable point p, in a plane (36), or in space (79),
cases, the weights are
be said, on the same plan, to be quotients of quotients of weights.

The circumstance that the position of a barycentre (97), like that of a simple mean point (96), is independent of the position of the assumed origin of vectors,
might induce us (comp. 69) to suppress the symbol o of that arbitrary and foreign
may
(2.)
point; and therefore to write' simply, under the lately supposed conditions,
*
lus
We should thus
have some of the principal notations of the Barycentric CalcuCompare the Note to page o6.
but used mainly with a reference to vectors.
CHAP.
BARYCENTRES OF SYSTEMS OF POINTS.
III.]
or
bB=âA,
if
89
= a.
It is easy to prove (comp. 96), by principles already established, that the ordinate of the harycentre of any given system of points is the complex mean (in
the sense above defined, and with the same system of weights), of the ordinates of
the points of that system, with reference to any given plane : and that the projection
of the bary centre, on any such plane, is the harycentre of the projected system.
(3. ) Without any reference to ordinates, or to any foreign origin, the barycentric
notation B
be
Da may
interpreted,
by means of our fundamental convention
A, considered as
(Art. 1) respecting the geometrical signification of the symbol b
denoting the vector from A to B together with the rules for multiplying such vectors by scalars (14, 17), and for taking the sums (6, 7, 8, 9) of those (generally
:
new)
vectors,
which are (15) the products of such multiplications. For we have only
to write the formula as follows,
in order to perceive that
it
2a(A-B) = 0,
be considered as signifying, that the system of the
may
vectors from the barycentre B, to the system of the given points Ai, A2,
by the scalars
tiplied respectively
comes (generally) a new system of vectors with a null sum

these last vectors, ai bai, ao. BA3,
.
when mul-
(or coefficients) of the given system ai, a^,
in such a
be-
manner that
can be made (10) the successive sides ofa closed
polygon, by transports without rotation.

(4.) Thus if we meet the formula,
we may
indeed interpret
it
= Âi -f A2),
as an abridged form of the equation,
OB = |(oAi 4 OA2)
o be any arbitrary point, and if o'be the point which completes

(comp. 6) the parallelogram A1OA2O', then b is the point which bisects the diagonal
00', and therefore also the given line A1A2, which is here the other diagonal. But we
which implies that
may
if
also regard the formula as a
mere symbolical transformation
of the equation,
(a3-b) + (ai-b) = 0;
which (by the earliest principles of the present Book) expresses that the two vectors,
from B to the two given points Ai and A2, have a null sum; or that they are
equal in
length, but opposite in direction : which can only be, by B bisecting A1A2, as before.
(5.) Again, the formula, Bi
= i(Ai +
ment of the equation,

OBi
A2 + A3),
may
be interpreted as an abridg-
= J (OAi + OA2 + OA3),
which expresses that the point B trisects the diagonal 00' of the
parallelepiped
(comp. 62), which has OAi, 0A2, 0A3 for three co-initial edges. But the same formula niay also be considered to express, in full consistency with the
foregoing interpretation, tliat the sum of the three vectors, from b to the three points Ai, A2,
vaA3,
nishes
A1A2A3.
which
And
is
the characteristic property (30) of the
similarly in
mean point of the
triangle
more complex cases: the legitimacy of such

transformaa consequence of the original interpretation
of the
tions bt-'ing here regarded as
(1)
symbol b - a, and of the rules for operations on vectors, so far as as they have been
hitherto established.
KLEMENTS OF QUATERNIONS.
90
Section 6
[bOOK
I.
On Anharmonic
sions,
Equations, and Vector-Expresand

Curves in Space.
Surfaces
of
When, in the expression 79 for the vector /> of a variable

P
of
point
space, the four variable scalars, or anharmonic co-ordinates, xyzw, are connected (comp. 46) by a given algebraic equation,
98.
w) = 0, or briefly /= 0,
supposed to be rational and integral, and homogeneous of the j^''*
dimension, then the point p has for its locus a. surface of the p*^ order
whereof /= may be said (comp. 56) to be the local equation. For
fp{x, y,
z,
if
we
substitute instead of the co-ordinates x
.w, expressions of the
forms,
X=
tXo
to indicate (82) that p
+ UXly
is
collinear
W= tWQ + Wlî,
with two given points,
Pq, Pi,
the
resulting algebraic equation \at\u is of the p*^ degree ; so that (according to a received modern mode of speaking), the surface may be
said to be cut in
p poitits
(distinct or coincident,
and
real or imagi-
And in like manner, when

nary*), by any arbitrary right line, PyPi.
the four anharmonic co-ordinates Imnr of a variable plane 11 (80) are
connected by an algebraical equation, of the form,
'Fq{l,
m, n, r) =0, or briefly F
= 0,
where f denotes a rational and integral function, supposed

mogeneous of the q*^ dimension, then this plane 11 has for
lops (comp. 6Q) a surface of the q*^ class, with
equation: because
we make
I = tlo + wî, ...
q*^ degree,
for its tangential
= ?ro +
we
uri,
n passes through a given
are conducted to an algebraical equation of the

(real or imaginary) values for the ratio t:u,
which gives q
and thereby assigns q

*
be ho-
its enve-
if
to express (comp. 82) that the variable plane

right line rio'Ili,
f=
to
(real or
imaginary!) tangent planes
to the sur-
be observed, that no interpretation is here proposed, ior imaginary intersectio7is of this kind, such as those of a sphere \vith a right line, which is who/It/
external thereto.
The language of modem geometry requires that svch imaginary
It is to
intersections should be spoken of,
and even that they should be enumerated : exactly

we should count what are called the imagi-
as the language of algebra requires that
nary roots of an equation.
But
it
would be an
error to confound geometrical imagi-
of negatives, for which it will suon be seen

that the Calculus of Quaternions supplies, from the outset, a d-finite and real innaries, of this sort, with those square roots
terpretation.
+ As regards the unijiterpreted character of such imaginary contacts in geometry,

the preceding Note to the present Article, resp icting imaginary intersections, may be
consulted.
CHAP.
fcuce^
ANHARMONIC EQUATIONS OF SURFACES.
III.]
drawn through any such given but arbitrary right
line.
91
We
51, 56), that if the functions /and f be only homogeneous (without necessarily being rational and integral)^ then
may add (comp.
[d./, D,/, D,/,
D,/]
the anharmonic symbol (80) of the tangent plane to the surface

/= 0, at the point (xyzw) ; and that
is
(d^f, d,f, df,
in like manner, a
is
[/mnr], with its enveloped surface^ f

teristics of partial derivation.
(1.)
d,f)
of contact of the plane
for the point
symbol
d^,
As an Example, the surface of the second
order,
d^,
being charac-
which passes through the
nine points called lately

A,
has for
its local
B,
c',
C,
a',
C2,
D,
A2,
E,
equation,
0=f=xz-ywt,
which
gives,
by
differentiation,
I
= T>xf= 2
rn.=:
n=i>zf=x;
10yf=
w,
=D,/=-y:
so that
[z,
is
- w,
-y]
X,
a symbol for the tangent plane, at the point (x, y, z, to).

(2.) In fact, the surface here considered is the ruled (or hyperbolic) hyperboloid,
on which the gauche quadrilateral abcd
And
through the point E,
P=
et
namely, the lines
two
pairs,
we
is
superscribed,
R=
(OyzO),
= (OOzw),
annexed Figure 31),

drawn through p to intersect the
(see the
ab, cd, and bc, da, of opposite sides
of that quadrilateral abcd, are the
two generating
hues, or generatrices, through that point

their plane, qrst, is the tangent
face, at the point p.
If,
we
then,
plane
so that
to the sur-
denote that tan-
we have
gent plane by the symbol [Imnr'],
the
equations of condition,
= lx-r- my my + nz=-nz + rw = rw + lx;

whence
follo\v3 the proportion,
l:7n:n:r = x-^
or,
y~^
2~'
u?
i
;
because xz = yw.
I
m:n:
r--
w:x:-y,
as before.
(3.)
At
the
same time we
and which passes
write
Q = (oryOO),
(xyziv),
then Qs and
if
see that
(ac'bq)
(dcocs)
also
92
so that the variable generatrix QS divides

(as
AB and DC homographicallg*
versely, if it
were proposed
ad, bc, and
13
c'c-z
to find the locus
known)
of the right
two given
lines,
and ad, bc,
I.
the two fixed generatrices
being three of
which thus divides
we might
c'c2 (with the recent
Con-
its positions.
line QS,
homographically (com p. 26) two given right lines in space,

for those
[bOOK
take
AB and dc
meanings of the
letters)
and then should have, for the two variable but corresponding (or homologous^ points Q, s themselves, and for any arbitrary
point p eollinear with them, anharmonic symbols of the forms,
for three given positions of the variable line
Q=
because,
by
we
82,
(s, M, 0, 0),
= (0, 0, , ),
we write p = (x, y,
so that the
locus.,
p = (s*,
tu^ tt, s)
should have, between these three symbols, a relation of the form,

(p)
if then
z,
= <(Q) +
we have
w),
whether of the
f(s):
the anharmonic equation xz
line qs, or of the
is
point p,
= yw^
as before
known) a
is
(as
ruled
surface of the second order.

(4.)
As
observe that
regards the
if
we
known
write, in like
T = (<00),
R=(0^0),
we
shall
it
may
suffice to
(p)=tt(R) + s(T),
have again the expression,
p=
as before
double generation of that surface,
manner,
so that the
{st, tu,
xz = i/w,
giving
uv, vs),
same hyperboloid is
divides the other pair of opposite sides bc,
also the locus of that other line bt,
ad
which
same gauche quadrilateral abcd

and the lines a'ao,
its positions
of the
;
ba, cd, and a'a2 being three of
supposed to intersect each other in the given point e.
(5.) The symbol of an arbitrary point on the variable line rt is (by sub-art. 2)
of the form, t(0, y, 2, 0) + u{x, 0, 0, w'), or (wic, fy, tz, uw) while the symbol of an
homographically
c'c2 being
still
arbitrary point on the given line c'c3

sent one
common point (comp.

p'
Hence the known theorem
And
these
two symbols repre-
Fig. 31),
when we suppose
t
is (f, t\ ', ').
= RT-c'c2=(y,y,2,2),
^
=y, u -z, t=\,
y
=-=-,
results, that a variable generatrix, rt,
tersects three fixed lines, BC,
ad,
which are generatrices of
c'co,
of one system,
Conversely, by the same comparison of symbols, for points on the two lines
o'c2,
we
section
line,
rt and
should be conducted to the equation xz =yw, as the condition for their interand thus should obtain this other known theorem, that the locus of a right
which intersects three given right
with those three lines
lines in space, is generally
for generatrices.
p"
= QS
similar analysis
a'a2
an hyperboloid
shows that QS
which may be thus denoted
a'a2, in a point (comp. again Fig. 31)
(6.)
in-
the other system.
intersects
= (xyyx).
As another example of the treatment of surfaces by their anharmonic and

we may remark that the recent symbols for p' and p'', combined with
local equations,
Compare
p.
298 of the Geometrie Super ieure.
CHAP.
III.]
AKHAllMONIC EQUATIONS OF SURFACES.
the symbols of 83, 86

s, t; with
and with the equation xz = yw, give the expressions
93
for c', a', Ca, A2,
those of sub-art. 2 for p, q, r,
k;
= (q) + (s)=Cr) + (t);
(p)
(e)
whence
it
= (c')
+ (C2) =
(a')
+ (A2)
follows (84) that the
two points
ABCD, divide the four generating

ratios
(P')
= y(c')+^(^'2)=(R) +
(p")
= y (A-) +
and the
p', p',
through p and e
lines
x (as)
^CT);
= (q) + (s)
f
sides of the quadrilateral
in the following
anharmonic
(cECaP')
(a'eA2P")
so that (as again is
= - = (ba'cr) = (aa2Dt)
(qp"sp)
= y- =
= (rp'tp)
known) the
(bc'Aq)
= (CC2DS)
variable generatrices, as well as the Jlxed ones, of the
hyperboloid, are all divided homographically

of the present surface is easily found, by the expres(7.) The tangential equation
sions in sub-art. 1 for the co-ordinates Imnr of the tangent plane, to be the follow.
ing:
= F = Zn mr
which
may
second
be interpreted as expressing, that this hyperboloid
class,
is
the surface of the
which touches the nine planes,
[1000], [0100], [0010], [0001], [1100], [0110], [0011], [1001], [1111]

or with the literal symbols lately
BCD,
CDA,
DAB,
ABC,
employed (comp. 86, 87),

CDc",
DAa",
ABC'2,
BCA'2,
and
[e].*
Or we may interpret the same tangential equation f = as expressing (comp. again

86, 87, where q, l, n are now replaced by t, r, q), that the surface is the envelope of
a plane qrst, which
the two connected conditions of
satisfies either of
(bc'aq)
(ca'br)
'
^
=
n
=
n
= (cc-iDs)
=
r
(DA2AT)
homogrnphy
And as rea double generation of the hyperboloid thus showing itself in a new way.
gards the passage (or return), from the tangential to the local equation (comp. 66),
we have in the present example the formulae
:
X = DiF n;
y = d,p = -
= df =
w = DrF = m
whence
xz yw
= 0,
as before.
and therefore also

(8.) More generally, when the surface is of the second order,
and f, when presented under rational
of the second class, so that the two functions
and integral forms, are both homogeneous of the second dimension, then whether we
derive I .r from x .why the formulae,
.
In the anharmonic symbol of Art. 87, for the plane of homology [e], the co-
ffficient 1 occurred,
through inadvertence, five times.
94
or X.
from
by the converse
I.
formulae,
Z = J>nF,
y = D,F,
X=I>iF,
[bOOK
W = DrF,
the point p = (xyztv) is, relatively to that surface, what is usually called (comp. 52)
the pole of the plane 11 = IJmnr] ; and conversely, the plane 11 is the polar of the
point p ; wherever in space the point P and plane IT, thus related to each other,
And
be situated.
may
because the centre of a surface of the second order
is
known
to be (comp. again 62) the pole of (what is called) the plane at infinity ; while (comp.
38) the equation and the symbol of
this last
ax-^hy+cz^dw 0,
if
the four constants abed have
still
with reference to the system of the
and
same
the
five
plane
are, respectively,
[a, 6, c, d],
significations as in 05, 70,
given points
abcde
79, &c.,
follows that
it
we may
denote this centre by the symbol,
K = (DoFo,
where Fq denotes,
DcFo, DjFo)
DftFo,
Fâbcd), and
for abridgment, the function
is still
a scalar con-
stant.
we have
(9.) In the recent example,
Fo
K = (c, - d,
we assume (comp.
if
Accordingly
p=
where
i'
if
also
we
= bt+ cv,
p'
= cu-\- ds,
-ac- hd,
have the symbolic
shall then
p" = {x"y"z"w"^ be that
x"
and
= 2e'st cw',
new
is
a new point, while
u =dv
w'
-{
'
at,
= ast + btu + cuv + dvs
= as + 6m
point K, and
e'(p)-(p') = (p"),
point, of which the co-ordinates are,
z" =^2e'uv aw\
ax" + by" + cz" + dw" =
to"
= 2e'v3 -\-bw',
0.
any chord of the hyperboloid, which passes through the fixed

p" be the harmonic conjugate of that fixed point, with respect to that
is to say, if
if
pp' be
variable chord, so that (pkp'p")
distant plane \abcd^^
which pass through
With
it,
=-
1,
then this conjugate point p" is on the infinitely

K bisects all the chords pp'
or in other words, the fixed point
and
is
therefore (as above asserted) the centre of the surface.
same meanings
(65, 79) of the constants a, b,

point (96) of the quadrilateral abcd, or of the system of its comers,
(10.)
relations,
y" = 1e'tu\ dw,
therefore,
That
- 6).
= (s't', t'u\ uv\ = v's'),
e'(p)+(p')=''(K),
if
a,
write, for abridgment,

e'
we
t'
and the anharmonic symbol
sub-arts. 3, 4),
(st, tu, MP, vs),
v are any four scalars, and p'
, t, u,
= ac bd;
becomes thus,
for the centre of the hyperboloid
the
c,
d,
may
the
mean
be denoted
by the svmbol,
M=(a-i, 6",
if
then this
mean
c-i, rf-i);
point be on the svrface, so that
ac-bd=0,
the centre
is
on the plane
[a, b,
c, d']
or in other words,
it is
infinitely distant:
so
CHAP.
VECTORS OF SURFACES AND CURVES IN SPACE.
III.]
95
In genethat the surface becomes, in this case, a ruled (or hyperbolic) paraboloid.
ral (comp. sub-art. 8), if Fo = 0, the surface of the second order is a paraboloid of
some kind, because
its
centre
then at infinity^ in virtue of the equation
is
(aDa +
+ cT>c + dud) Fo =
hl>b
or because (comp. 50, 68) the plane [abed'] at infinity
is
then one of
its
tangent
= 0.
planes, as satisfying its tangential equation, F
(11.) It is evident that a curve in space may be represented
by a system of two
anharmonic and local equations ; because it may be regarded as the intersection nf
two surfaces. And then its order, or the number of points (real or imaginary*), in
which
cut by
it is
two surfaces
an arbitrary plane^
is
obviously the product of the orders of those
or the product of the degrees of their
be rational and integral.

(12.) A curve of double curvature
may
also
two
local equations,
supposed to
be considered
sion (or arete de rebroussement) of a developable surface,
fts the edge of regresnamely of the locus of the
; and this surface may be supposed to be circumscribed at once

two given surfaces^ which are envelopes of variable planes (98), and are repreIn this view, a curve of double cursented, as such, by their tangential equations.
tangents to the curve

to
by a system of two anharmonic and tangential eqaa.the class of such a curve be defined to be the number of its osculating
planes, which pass through an arbitrary point of space, then this class is the product
of the classes of the two curved surfaces just now mentioned: or (what comes to the
vature
may
tions
and
itself he represented
if
same thing)
details
two tangential equations, by

But we cannot enter further into these
the product of the dimensions of the
it is
which the curve
(on this plan) symbolized.
is
the mechanism of calculation respecting which would indeed be found to be
the same, as that employed in the
known method (comp. 41)
of quadriplanar co-or-
dinates.
99. Instead of
other system of
anharmonic co-ordinates, we may consider any

scalars, Xy,
x,i, which enter into the ex-
w variable
pression of a variable vector, p; for example, into an expression of

the form (comp. 96, 97),
p=
And
if
then,
xai+ x_a^ +
= Ixa.
those n sc'dars x be oW functions of one independent
variable scalar,
t,
tion of that single
we may regard this vector p

scalar; and may write,
and
as being itself a. func-
I.,.p=(p{t).
But
if
the n scalars
variables,
and
u,
be functions of
tivo independent and scalar

then p becomes a function of those two scalars,
a;
and we may write accordingly,

II.
= <^{t,
u).
In the 1st case, the term P (comp. 1) of the variable vector p has
*
Compare the Notes
to
page 90.
96
[bOOK
I.
which may be plane or of

become
a
even
right line, according to the
may
form of the vector-function cp and p may be said to be the vector of
this line; or curve.
In the Ilnd case, p is the vector of a surface, plane
generally for its locus a curve in space,
double curvature, or
or curved, according to the form of <p (^, u) ; or to the manner in which

this vector p depends on the two independent scalars that enter into
its
expression.
(1.)
As Examples (comp.
'
25, 63), the expressions,
Il...p=____.
l...p=.-;
signify, 1st, that p is the vector of a variable point p on the right line ab ; or that
it is the vector of that line itself, considered as the locus of a point; and Ilnd, that
the vector of the plane abc, considered in like
is
manner
as the locus of an arbitrary
point P thereon.
(2.)
The equations,
I.
= ara + y/3,
II. ..
= a;a + y/3 + zy,
wth
a; 4-
y2 = 1 for the
1st,
and
x"^
+ y'^ -^ z^ =
I for
the Ilnd,
signify 1st, that p is the vector of an ellipse, and Ilud, that it is the vector of an
for their common centre, and with OA, ob, or OA, ob,
ellipsoid, with the origin o
DC, for conjugate semi-diameters.

(3.)
The equation (comp. 46),

p
expresses that p
centre),
which
is
= t'â^^ 2/3 + (< + uy y,
the vector of a cone of the second order with o for

^
touched by the three planes OBC, OCA,
is
cab
its
vertex (or
the section of this cone^
made by the plane abc, being an ellipse (comp. Fig. 25), which is inscribed in
abc and the middle points a', b', c', oii\ie sides of that triangle, being
triangle
the
the
points of contact of those sides with that conic.

(4.)
The equation (comp. 53),

= râ + "' /3 +
p
expresses that p
is
vertex, but with OA, ob,
plane
abc
is
a new
y'l
y,
with
+ m + r = 0,
the vector of another cone of the second order, with
ellipse,
oc
The
for three of its sides

(or rays).
still
for
by
the
section
circumscribed to the triangle abc, and having
its
tangents
at the corners of that triangle respectively parallel to the opposite sides thereof.
(5 J
The equation (comp.

p
signifies that
the origin
its
54),
= f3a + u^^ + v^y,
with
\-u-\-v
0,
the vector of a cone of the third order, of which the vertex is still
section (comp. Fig. 27) by the plane abc I)eing a cubic curve, whereof
is
the sides of the triangle
abc
are at once the asymptotes, and the three (real) tangents
of inflexion; while the mean point (say o') of that <na^Ze is

the curve; and therefore the right line oo', from the vertex o
may
be said to be a conjugate ray of the cone.
(Jo.)
The equation (comp.
98, sub-art. (3.)),
a.
conjugate point oi
to that
mean
point,
CHAP.
VECTORS OF SURFACES AND CURVES
III.]
staa
IN SPACE.
97
+ tub^ + uvcy + vsdd

+ <t6 + uvc + vsd
sta
-and-
which
in
u
scalars,
are two variable scalars, while
and
a,
presses that p
are
rf
still
four constant
but p is still a variable vector, exy, d are four constant vectors,

the vector of a ruled (or single-sheeted) hi/perboloid, on wldch the
abcd is superscribed, and which passes through the given point
/3,
is
gauche quadrilateral
E, whereof the vector
assigned in 65.
e is
we make (comp.
(7.) If
a,h,c^
"
98, sub-art (9.)),
s'taa -
s't'a
t'u'hfi
t'u'b
+ u'v'cy - v's'dd
+ u'v'c v's'd
where
s'=zbt
then p'
it is
it
= op'
+ cv,
t'
= cu + ds,
=dv + at,
v'
= as+ bu,
the vector of another point f' on the same hyperboloid ; and because
last vectors is constant,
is
found that the suvi of these two
follows that k
passes through
is
it
the vector oi
z.
fixed point k, which bisects every chord pp' that

K is the centre of
or in other words (comp. 52), that this point
the surface.
(8.)
The
three vectors,
^^^
+y
rt
'
'
are termino-collinear (24) ; if then a gauche quadrilateral abcd be superscribed on

a ruled hyperboloid, the common bisector ofthe two diagonals, AC, bd, passes through
the centre k.
(9.)
When
ac
bd, or
when we have the

sta
st
or simply,
p
p
is
art.
= sta^-
tu[3
+ uvy
equation,
+ uvy + vsS
tuji
tu
-f
vsd,
-T
uv
'
-\-
vs
with
+u = t + v = l,
then the vector of a ruled paraboloid, of which the centre (comp. 52, and 98, subis
distant, but upon which the quadrilateral abcd is still super(10.)
scribed.
),
And
infinitely
this surface passes through the
of the system of the four given points
mean point
;
of that quadrilateral, or
when
because,
= t = u = v = ^,
tho
variable vector p takes the value (comp. 96, sub-art. (1.)),

/*
(10.) In general,
it
this paraboloid is the locus
tides
AB and dc of
site sides,
= i(a + /3 +
is easy to prove,
of a right
from
line,
y+
which divides similarly the two opposile
the same gauche quadrilateral
bc and ad.
c).
the last vector-expression for p, that
abcd
or the other pair of

opp^-
98
Section
On
The equation
100.
(99,
[bOOK
Differentials of Vectors.
I.)>
= op is generally the vector of a point p of a. curve

/>
gives evidently, for the vector oq of another point
same curve, an expression of the form
which
in
PQ
I.
.,
i7i
space,
Q of the
P + Ap=(p{ti-At);
Bo that the chord p% regarded as being
itself a vector, comes thus to be repre-
sented (4) by the Jinite difference,
Ta=Ap = A(p(t) = (p(t + At)-(p{t).

Suppose now that the
other finite dif-
new
ference, A^, is the n*^ part of a

scalar,
u ; and that the chord
Fig. 32), the
r^^
or pq,
A/),
part of a new vector,
is
ff,
in like
or pr
manner (comp.
so that
we may
write,
nAt =
u,
and
?iA/>
= w pq =
.
<r
= pr.
Then, if we treat the two scalars, t and u, as constant, but the number n as variable (the/orm of the vector-function (f), and the origin o,
being given), the vector p and the point p will he fixed: but the two
points Q
n^P, or
be
and
ff,
r,
the two differences Ai and
A/),
will (in general) vary together.
indefinitely increased, or
made
to tend
and the multiple
And
if this
to infinity,
vector
number n
then each of the
two diflferences A^, A/> will in general tend to zero ; such being the
common limit, of n-û, and of ^{t + n~û) - ^(t): so that the variable
will te?id to coincide with the fixed point p. But
point Q, of the curve
although the chord pq will thus be indefinitely shortened, its n*^ mulPR or <r, will tend (generally) to definite limit,* depending on
the supposed continuity of the function 4>(^); namely, to a certain
or o-, or (say) t, which vector pt will evidently
definite vector, pt,
tiple,
to the curve: or, in other words, the variable

(in general) tangential
to a fixed position t, on thetangent to that curve at p.
tend
will
R
point
be
We
shall thus
T
have a
limiting equation, of the
= pT = lim.
PR = ra5 =
lim. 7iA0(i), if
n=
and u being,
as above, two given

*
form
nAt=u;
00
and {generdXXy) finite
Compare Newton's
Principia.
scalars.
And
CHAP.
if
DIFFERENTIALS OF VECTORS.
III.]
we then agree
ferential of the
to call the second of these
and
first,
to denote it
99
two given
by the symbol
scalars the dif-
d,
we
shall de-
^/?ethat the vector-limit, r or o-, is the (corresponding) differential of

the vector />, and shall denote it by the corresponding symbol^ d/>; so
as to have,
under the supposed conditions,

u = dt, and t = dp.
Or, eliminating the two symbols u and t, and not necessarily supposing that p is appoint of a curve, we may express our Definition"^ of the
Differential of a Vector />, considered as a Function (p of a Scalar t^
by the following General Formida

d/)
= d<f>(O =
lim.nJ0(if+
-)-<(>(m,
71
n = 00
/
in
which
and
nerally finite
on those
is,
and independent scalars, both genew and finite vector, depending
dt are two arbitrary
and dp
two
in general, a
according to a law expressed by the formula,
scalars,
and derived from that given law, whereby the old oi former
or
(p
(t),
(1.)
depends upon
As an example,
vector,
the single scalar, t
let the
given vector-function have the form,
p = ^(t') = ^tâ,
where a
a given vector.
is
Then, making A^
= -,
where
is
any given
scalar^
and n
is
a variable whole number,
we have
a
CTn
and
finally,
m\2
ff
= nAp = au{t-\-\;
writing dt and dp for u and
dp =
u \
aul
(Ta.
atu
era,,
= dl
d(p(t)
-^\
= atdt.
(2.) In general, let (p(t) =af(t), where a is still a given or constant vector, and
/(<) denotes a scalar function of the scalar variable, t. Then because a is a common
factor within the brackets { } of the recent general formula (100) for dp, we may
write,
dp
provided that
we now
= d0(O = d.a/(O = ad/(O;
scalar function of two independent scalars,

similar formula
of a scalar function, fif), is a new

and dt, determined by the precisely
define that the differential

t
d/(()
i'f.-l-^c+f)-^'
Compare
the Note to page 39.
100
which can
easily be proved to agree^ in all its consequences,
dijêrentiating functions
[bOOK
I.
with the usual rules for
of one variable.
(3.) For example, if we write d^ = nh, where A is a new variable scalar., namely,
the n*^ part of the given and (generally) finite differential^ d^, we shall thus have
the equation,
3^ =
lim
d^
in
which the
first
member
is
/.=o
/(^ + ^)-/(0
h
here considered as the actual quotient of two finite sca-
"We may, however, as

df(t) d^, and not merely as a differential coefficient.
usual, consider this quotient, from the expression of which the differential d< disappears, as a derived function of the former variable, t and may denote it, as such, by
lars,
either of the two usual symbols,
f'(t) and Dtf{t).

(4.)
In
the formtda
these
two
manner we may
like
of a vector-function,*
write, for the derivative
<p(f),
last
forms denoting that actual and finite vector, p or tp' {t), which is
by dividing (comp. 16) the not less actual (or finite) vector,
obtained, or derv)ed,
dp or d^
by the finite scalar, dt. And if again we denote the n" part of this
h, we shall thus have the equally general formula :
(t),
last scalar
by
vtp
= Dt(p (0 = hm.
!-^
A=
7-^-^
with the equations,
dp
exactly as
(5.)
if
The
= Dtp
d*
p'dt
<10(O
= vt<p (0
dt
(b'(t)
dt,
the vector-function, p or 0, were a scalar function,

= 1, gives thus dp = p'; so that the derived vector
particular value, dt
(with our definitions) a particular but important case of the differential of a

vector.
In applications to mechanics, if t denote the time, and if the term p of
p' is
the variable vector p be considered as a moving point, this derived vector

p' may be
called the Fector of Velocity : because its length represents the amount, and its di-
And if, by setting off vectors ov = p' (comp,

is the direction of the velocity.
again Fig. 82) from one origin, to represent thus the velocities of a point moving in
space according to any supposed law, expressed by the equation p = 0(<), we conrection
struct a
p'
curve PQ.
vw . of which the corresponding equation may be written as

then this new curve has been defined to be the HoDOGEAPH,f as the old
new curve
= ^'(0)
may
be called the orbit of the motion, or of the moving
point.
In the theory of Differentials of Functions of Quaternions, a definition of the

which is expressed by an equation of precisely
will be proposed,
differential dtp (ji)
but it will be found

the same form as those above assigned, for d/(), and for d0 (t)
that, for quaternions, the quotiefit d0(g): dq is not generally independent ofdq
and consequently that it cannot properly be called a derived function, such as 0'(9)
of the quaternion q alone. (Compare again the Note

f The subject of the Hodograph will be resumed,
work.
In fact,
what appears
it
to
page 39.)
at a subsequent stage of this
almost requires the assistance of Quaternions, to connect

mode, with Newton's Law of Gravitation.
to be the best
it.,
in
CHAP.
DIFFERENTIALS OF VECTORS.
III.]
We may
(6.)
a vector-function twice (or oftener), and so obtain
differentiate
For example,
sticcessive differentials.
101
if
we
differentiate the derived vector p',
its
we
obtain a result of tbe form,
= |[)"d^,
dp'
where p" = T>tp' ^''Ot'^p,
by an obvious extension of notation

dd^ or
of the scalar
d^^,
t is
d<2, as usual,
zero^ then the
= ddp = d
d2p
where
and if we suppose that the second
p'd<
differential,
second differential of the vector p
= dp'
6.t
p". dt^
is,
denotes (dty ; and where it is important to observe that, with

is as finite a vector as dp, or as p itself. In applications
the definitions adopted, d^p
to motion, if < denote the time,
We may
(7.)
may
p"
be said to be the Vector of Acceleration.

of the Vector
also say that, in mechanics, \h& finite differential dp,
of Position p, represents, in length and in direction, the right line (suppose pt in

Fig. 32) which woidd have been described, by a freely moving point P, in the finite
interval of time dt, 'imvci%(imiQ\y following the time t, ifaX the end of this time t all
had ceased
to act.
(8.) In geometry, if p
= ^(f)
foreign forces
be the equation of a curve of double curvature, re) of a developable surface, then the
garded as the edge of regression (comp. 98, (12.)
equation of that surface itself, considered as the locus of the tangents to the curve,
may be thus written (comp 99, II.)
:
p
if it
= ^ (0 + ^' (0
01^
simply, p
= (0 + d^ (t),
be remembered that u, or At, may be any arbitrary scalar.

(9.) If any other curved surface (comp. again 99, II.) be represented
tion of the form, p
dent
and scalar
= ^(x, y),
variables,
where
x and
y,
by an equa-
now denotes a vector -function of two indepenwe may then differentiate this equation, or this
<f)
expression for p, with respect to either variable separately, and so obtain what may
be called two partial (hut finite) differentials, d^p, dyp, and two partial derivatives,
Dj.p, T>yp,
whereof the former are connected with the latter, and with the two arbitrary
(hut fnite) scalars, dx, dy, by the relations,
dxp
And
these
two
= Dxp
da;
d^p
= Dyp
6y.
differentials (or derivatives) of the vector
p of the surface denote two
tangential vectors, or at least two vectors parallel to two tangents to that surface at
the point p
so that their plane is (or is parallel to) the tangent plane at that point.
(10.) The mechanism of all such differentiations of vector-functions is, at the
:
present stage, precisely the
same
as in the usual processes of the Differential Calcu-
lus; because the most general form of such a vector-function,
which has been consi-
dered in the present Book, is that of a sum of products

(comp. 99) of the form xa,
where a is a constant vector, and a; is a variable scalar: so that we have only to
operate on these scalar coefficients x ., by the tisual rules of the calculus, the vectors a.
being treated as constant factors (comp. sub-art. 2). But when we shall
.
come
products of vectors, or generally those new functions of

which can only be expressed (in our system) by Quaternions, then some few
new rules of differentiation become necessary, although deduced from the same
(or
nearly the same) definitions, as those which have been established in the present
to consider quotients or
vectors
Section.
As
is
well illustrated bv Atwood's machine.
102
[bOOK
I.
of a vector
(11.) As an example of partial differentiation (comp. sub-art. 9),
" vector"
being here used as an adjective) of two scalar variables^
function (the word
let us take the equation,
p = 0(a:,y)=|{a:2a+y2/3 + (x + y)2y};
which p (comp. 99, (3.) ) is the vector of a certain cone of the second order; or
more precisely, the vector of owe sheet of such a cone, if x and y be supposed to be
in
Here, the two partial derivatives of p are the following
real scalars.
%p = y/3 + (a; + y) y
= xa + (x + 7/)y;
i>:,p
and
therefore,
2p
= XD:rp + y^yp
so that the three vectors, p, D^p, Dyp, if
drawn (18) from one common origin, are con-
tained (22) in one common plane; which implies that the tangent plane to the surand thereby verifies the conical
face, at any point p, passes through the origin o
character of the locus of ihaX point p, in which the variable vector p, or op, termi:
nates.
(12.)
If,
in the
same example, we make
whence
it
a:
follows that the middle point, say
the values,
c',
of the right line ab,
is
one of the
and that (comp. again the sub-art. 3 to Art. 99, and the
9) the right lines oa and ob are parallel to two of the tangents to the
points of the conical locus
recent sub-art.
= 1, y = - 1, we have
Dyp = -^;
D^p = a,
p = i(a + ^),
touched by the plane aob, along

be proved, that the same cone is
touched by the two other planes, boc and COA, at the middle points a' and b' of the
two other lines BC and CA and therefore along the two other sides (or rays), oa'
surface at that point
so that the cone in question
And
the side (or ray) oc'.
in like
manner
it
is
may
and ob'
which again agrees with former results.

(13.) It will be found that a vector function of the sum of two scalar variables,
and df, may generally be developed, by an extension of Tai/lor's Series, under the
:
form,
0(< +
dO = 0(O + d^(O + i(lV(O+^d3^(O + -d2
=(i + d + - +
it
d^(pt
= 0,
&c.
where a
d3
+ ..)^(<) = ^V>(0;
being supposed that d'^t=
(as in sub-art. 1),
0,
is
dH 0, &c. (comp. sub-art. 6).

a constant vector, we have d(pt
Thus,
if
atdt,
d^(pt=adt'i,
<pt= ~at^,
and
(pit+
dt)
= |a (< -h dty = \at^ + atdt + âdfi,
rigorously, without
(14.)
= ^{t +
any supposition that dt is small,

we thus suppose At = dt, and develop e
When
0(i),
dt')
the finite difference,
A^{t)
the first term of the development so obtained, or the term of first
dimension relatively to
d^, is
hence (by a theorem, which holds good
tions, as well as for scalar functions) the first differential d<pt
for vector-func-
of the function
but
we do
not choose to define that this Differential is (or means') that first term : because the Formula (100), which we prefer, does not postulate the /Jossi6i7tVy, nor even
suppose the conception, of any such development.
appear more
clear,
when we
shall
theory of Quaternions, to which
come
to conyiect
we next
proceed.
recent remarks will perhaps

them, at a later stage, with that
Many
BOOK
II.
ON QUATERNIONS, CONSIDERED AS QUOTIENTS OF VECTORS,

AND AS INVOLVING ANGULAR RELATIONS.
CHAPTER
I.
FUNDAMENTAL PRINCIPLES RESPECTING QUOTIENTS OF VECTORS.
Section
1.
Introductory Remarks
First Principles adopted
from Algebra,
Art.
101.
The only angular
relations^ considered in the fore-
going Book, have been those of parallelism between vectors

(Art. 2, &c.) and the only quotient s^hXthQvto employed, have
been of the three following kinds
;
I.
tion
Scalar quotients of scalar s^ such as the arithmetical fracin Art. 14;
II. Vector quotients^
of vectors divided by scalars, as
in Art. 16;
III. Scalar quotients

lar or opposite, as
- =
ic
of vectors, with directions either simi-
in the last cited Article.
But we now
propose to treat of other geometric Quotients (or geometric

Fractions, as we shall also call them), such as
=for
each of which
q,
tlie
with jSwo^ll a (comp. 15);

Divisor (or denominator), a or oa, and
or ob, shall not only both be
/3
the Dividend (or numerator),
104
[bOOK
an Angle,
Vectors, but shall also be inclined to each other at
from zero, and from
distinct (In general)
102. In introducing this
new
II.
two'^ right angles.
conception, of a General
Quo-
of Vectors, with Angular Relations in a given plane, or

in space, it will obviously be necessary to employ some properties of circles and spheres, which were not wanted for the purtient
But, on the other hand, it will be

to
and
useful
possible
suppose a much less degree of acquaintance with many important theoriesf of modern geometry, than
pose of the former Book.
that of which the possession was assumed, in several of the

foregoing Sections. Indeed it is hoped that a very moderate
amount of geometrical,
ration will be found
algebraical, and trigonometrical prepasufficient to render the present Book, as
well as the early parts of the preceding one, fully and easily
any attentive reader.
intelligible to
proper to premise a few general principles

respecting quotients of vectors, which are m(^QQdi suggestedhy
And 1st, it is
algebra, but are here adopted by definition.
103. It
may be
evident that the supposed operation of division (whatever its

full geometrical import may afterwards be found to be), by
which we here conceive ourselves to pass from a given divisor-
and from a given dividend-line
line a,
(provisionally)
their geometric
j3,
to
what we have
quotient, q,
called
(or rather
may
must) be conceived to correspond to some converse act (as yet

not fully known) oIl geometrical multiplication: in which new
act the former quotient, q, becomes a Factor, and operates on
the line-u, so as to produce (or generate) the line
therefore lorite, as in algebra,
/3
5'.
a,
or simply,
j3
= qa, when
/3
j3.
a=
We
shall
$- ;
More generally speaking, from every even multiple of a right angle.

t Such as homology, homography, involution, and generally whatever depends
on anharmonic ratio : although all that is needful to be knoAvn respecting such
ratio, for
the applications subsequently made, may be learned, without reference to

It was,
treatise, from the definitions incidentally given, in Art. 25, &c.
any other
perhaps, not strictly necessary to introduce any of these modern geometrical theories,
any part of the present work ; but it was thought that it might interest one class,
at least, of students, to see how they could be combined with that fundamental con-
in
ception, of the
Vectoe, which
the First
Book was designed
to develope.
CFIAP.
even
I.]
FROM ALGEBRA.
FIRST PRINCIPLES ADOPTED
if the
two
lines a
and
or
/3,
105
oa and ob, be supposed
And
be inclined to each other, as in Fig. 33.
to
this very sim-
ple and nvLinvoS. notation (comp. 16) will then allow us to treat
as identities the two following formulae
:
\13
13
qa =
q;
,,
a=)â=(5;
a
although we shall, foj' the present, abstain from writing also

such formulae* as the following:
* = j3,
-a =
q,
where a, j3 still denote tivo vectors, and q denotes their geometrical quotient : because we have not yet even begun to consider the multiplication of one vector by another, or the division
of a quotient by a line,
104. As a Ilnd general
we
shall
next lay
=-,
3'
i3
principle, suggested
down, that
it
and
by
algebra,
if
a = a,
then
j3'
= /3
*
;
or in words, and under a slightly varied form, that unequal

The
vectors, divided by equal vectors, give unequal quotients.
importance of this very natural and obvious assumption will

soon be seen in its applications.
105. As a Ilird principle, which indeed may be considered to pervade the whole of mathematical language, and
without adopting which we could not usefully speak, in any
case, of
EQUALITY
quotients,
we
between any two geometrical

next assume that two such quotients can
as existing
shall
never be equal to the same third^ quotient, without being at the

same time equal to each other: or in symbols, that
if
*
q,
and
q"
q,
then
q"
= q.
be seen, however, at a later stage, that these two formulae are permitted,
in the development of the Quaternion System.
scarcely necessary to add, what is indeed included in this Ilird
It will
and even required,
It is
in virtue of the identity q
= q^
that
if q'
= q,
principle,
then q
q'
or in words, that
we
shall
never admit that any two geometrical quotients, q and q\ are equal to each other in
one order, without at the same time admitting that they are equal, in the opposite
order also.
1.
CHAP.
I.]
QUOTIENT OF TWO VECTORS A QUATERNION.
107
cannot generally be equal to any

of tlie (so-called) reals of algebra, whether of the positive or of
the negative kind. For let x denote any such (actual*) scalar,
LAR
or in other words,
it
let a denote any (actual) vector; then we have seen (15)

that the product xa denotes another (actual) vector, say j3',
which is either similar or opposite in direction to a, according
as the scalar coefficient, ov factor, x, is positive or negative ;
and
in neither case, then, can it represent any vector, such as j3,

is inclined to a, at any actual angle, whether acute, or
which
right, or obtuse
=
j5'
/3,
or Xa =
or in other words (comp. 2), the equation

impossible, under the conditions here sup-
j3, is
posed.
But we have agreed
= x;
we must,
xci
(16, 103) to write, as in algebra,
therefore (by the Ilnd principle of the fore-
going Section, stated in Art. 104), abstain from writing also
= X, under
the same conditions
denoting a scalar.
still
Whatever else a quotient of two inclined vectors

to be, it is thus, at least, a Non-Scalar.
may be found
forming the conception of the scalar itself,

two
parallel^ vectors (17), we took into acof
count not only relative length, or ratio of the usual kind, but
also relative direction, under the form o^ similarity or opposition.
109.
Now,
in
as the quotient
In passing from a to Xa, we a Zêrec? generally (15) the length of

the line a, in the ratio of + a; to 1 and we preserved or reversed
;
the direction of that line, according as the scalar coefficient x

was positive or negative. And in like manner, in proceeding to
form, more definitely than we have yet done, the conception of
the non-scalar quotient (108), $' = j3 a = Ob oa, optica inclined
:
vectors,
*
is
By
which
for simplicity
may
an actual scalar, as by an actual vector (comp.
different from
zero.
be supposed (18) to be co1),
we mean here one that
An
actual vector, multiplied by a null scalar, has for product

therefore unnecessary to prove that the quotient of two actual
(15) a null vector ; it is

vectors cannot be a null scalar, or zero.
It is to be
remembered that we have proposed (15) to extend the use of this

two vectors which are (in the usual sense of the
word)
terra parallel, to the case of
parallel to one
right line.
common
line,
even when they happen to be parts of one and the same
ELEMENTS OF QUATERNIOiNS.
108
[bOOK
II.
we have 5^7/ to take account both of the relative length,

and of the relative direction, of the two lines compared. But
initial^
while the former element of the complex relation here considered, between these two lines or vectors, is still represented
by a simple Ratio (of the kind commonly considered in geometry), or by a number* expressing that ratio the latter element of the same complex relation is now represented by an
Angle, aob and not simply (as it was before) by an alge;
braical sign, + or -.
110. Again in estimating this angle, for the purpose of
distinguishing one quotient of vectors from another, we must
consider not only
its magnitude (or quantity), but also its

since otherwise, in violation of the principle stated
in Art. 104, we should have ob': oa = ob : oa, if ob and ob'
Plane:
were
tivo distinct
rays or sides of a cone of revolution, with oa
which case (by 2) they would necessarily be

unequal vectors. For a similar reason, we must attend also to
the contrast between two opposite angles, of equal magnitudes,
for its axis; in
and
in
one common plane.
In short, for the purpose of know-
mg fully the relative direction of two co-initial lines oa, ob in

space, we ought to know not only how many degrees, or other
parts of some angular unit, the angle
aob contains; but
the direction of the rotation from oa to

OB including a knowledge of the plane,
:
in
also (comp. Fig. 33)
which the rotation
is
performed
^^""^
o^'S- 33.
and
of the hand (as right or left, when viewed from a known side of
the plane), towards which the rotation is directed.
111. Or, if we agree to select some one fixed hand (suppose
the right ^ hand), and to call all rotations positive when they
* This
number, -which we shall presently call the tensor of the quotient, may be
whole or fractional, or even incommensurable with unity but it may always be
equated, in calculation, to a positive scalar : although it might perhaps more pro;
perli/ be said to be a signless
lengths, without
any
number, as being derived solely from comparison of
reference to directions.
If right-handed rotation be thus considered as positive, then the positive axis
of the rotation aob, in Fig. 33, must be conceived to be directed downward, or below
the plane of the paper.
CHAP.
I.]
QUOTIENT OF TWO VECTORS A QUATERNION.
109
are directed towards this selected hand, but all rotations weâtive when they are directed towards the (dher hand, then, for
any given angle aob, supposed for simplicity to be less than two
right angles, and considered as representing a rotation in a given
plane from oa to ob, we may speak q{ one "perpendicular oc to
that plane aob as being the positive axis of that rotation and
of the opposite perpendicular oc' to the same plane as being the
;
negative axis thereof: the rotation round the positive axis being
zYê//* positive, and vice versa. And then the rotation aob may
be considered to be entirely known,
if we
know,
1st, its quantity,
and llnd, the

but not without a knowledge
of these two things, or of some data equivalent to them. But
whether we consider the direction of an Axis, or the aspect of
or the ratio which
it
bears to a right rotation
direction of its positive axis,
a Plane, wê find (as indeed
oc
is
well
known) that the determi-
nation of such a direction, or of such an aspect, depends on

polar co-ordinates* i or other angular elements.
two
112. It appears, then, from the foregoing discussion, that

for the complete determination, of what we have called the geometrical Quotient of two co-initial Vectors, a System of Four
Elemerits, admitting each separately of numerical expression,
is
generally required. Of these four elements, one serves (109)
to determine the relative length of the two lines compared ;
and the other three are in general necessary, in order to determine^//^ their relative direction. Again, of these three latter
elements, one represents the mutual inclination, or elongation,
of the two lines
between them
or the magnitude (or quantity) of the angle

while the two others serve to determine the
direction of the axis, perpendicular to their common plane,

round which a rotation through that angle is to be performed,
one (or towards

a fixed and previously selected hand), for the purpose of passing (in the simplest way, and therefore in the plane of the two
Vines) from the direction of the divisor-line, to the direction of
in a sense previously selected as the positive
The
actual (or at least the frequent)
7ise
of such co -ordinates
is
foreign to the spirit
but the mention of them here seems likely to assist a student,

by suggesting an appeal to results, with which his previous reading can scarcely fail
to have rendered him familiar.
of the present
System
110
the dividend-line.
And
[book
II.
no more than four numerical elements
are necessary, for our present purpose: because the relative

length of two lines is not changed, when their two lengths are
altered proportionally^ nor
when
is
their relative direction changed,
the angle which they form
is merely turned about, in its

of
this essential connexion of
On account, then,
plane.
that complex relation (109) between two lines, which is com-
own
pounded of a relation of lengths, and of a relation of directions^

and to which we have given (by an extension from the theory
of scalars) the name of a geometrical quotient, with a System
o/'FouR numerical Elements, we have already a motive* for
" the
Quotient of two Vectors
saying, that
is
generally a Qita-
ternion."
Section
3.
Additional
Illustrations.
Some
additional light may be thrown, on this first concepby the annexed Figure 34. In that Figure,
the letters cdefg are
113.
tion of a Quaternion^
designed to indicate
corners of a prismatic desk^ resting
upon
table.
horizontal
The
hcd
(supposed to be one of
angle
thirty degrees) repre-
sents a (left-handed)
rotation^
whereby the
horizontal ledge cd of
the desk is conceived
to be
elongated
Fig. 34.
(or
removed) from a given horizontal line oh, which may be imagined to

be an edge of the table. The angle gcf (supposed here to contain
forty degrees) represents the slope\ of the desk, or the amount of its
inclination to the table. On the/ace cdef of the desk are drav/n two 5^-
milar and similarly turned triangles^ aob and a'g'b^ which are supposed
to be halves of two equilateral triangles ; in such a manner that each
'
Several other reasons for thus speaking will offer themselves, in the course of the
present work.
t These two angles, hcd and gcf, may thus be considered to correspond to longitude of node, and inclination of orbit, of a planet or comet in astronomy.
CHAP.
ADDITIONAL ILLUSTRATIONS.
I.]
Ill
aob or a'o'b' is one of sixty degrees, and is directed towards

common hand (namely the right hand in the Figure) while if
rotation^
one
the side ob is to the side OA, in one trilengths alone be attended to,
is to the side o'a', in XhQ other; or as the numside
o'b'
as
the
angle,
ber two to one.
114.
Under
these conditions of construction,
we
consider the two
quotients, or the two geometric fractions,
OBiOA and ob':oa,
or
OB
OA
o^b'
-
and
o'a'
we regard the two lines, oa and

relative
the
same
length, and the same relative direction,
having
And we consider and speak of
as the two other lines, o'a^ and o'b^
as being equal to each other; because
OB, as
each Quotient, or Fraction, as a Quaternion: because its complete construction (or determination) depends, for all that is essential to its
conception,
and requisite to
distinguish it
from others, on a system of
four numerical elements (comp. 112); which
are, in this
Example, the
four numbers,
2,
Of these
115.
30,
60,
and
four elements (to recapitulate
40.
what has been above supposed), th
o'b', is
expresses that the length of the dividend-line^ OB or

double of the length of the divisor-line, OA or o'a'.
The Ilnd numerical
element,
namely
namely the number
1st,
2,
60, expresses here that the angle
grees; while the corresponding rotation, from
wards a known hand (in
OA
aob
or a'o'b',
is
one oi sixty de-
to ob, or from o'a' to o'b', is to-
hand, as seen by a person looking at the

the same for both of these two equal angles.
this case the right
face CDEF of the desk), which hand
is
The Ilird element, namely 30, expresses that the horizontal ledge CD of the desk
makes an angle of thirty degrees with a known horizontal line oh, being removed
it, by that angular quantity, in a known direction (which in this case happens
be towards the left hand, as seen from above).
Finally, the IVth element,
namely 40, expresses here that the desk has an elevation oi forty degrees as before.
from
to
Now
an alteration in any one of these Four Elements, such as an alterawould make (in the view here taken) an essential change in the Quaternion, which is (in the same view) the Quotient
of the two
lines compared
although (as the Figure is in part designed to suggest) no such
116.
tion of the slope or aspect of the desk,
when the triangle AOB

own plane, without being turned over (comp. Fig. 36)
change
its
is
conceived to take place,
is
;
merely turned about, in
or
when
the sides of that
triangle are lengthened or shortened proportionally, so as to preserve the ratio (in the
old sense of that word), of any one to any other of those sides.
may then briefly
refesay, in this mode oi illustrating the notion of a Quaternion*
geometry,
We
* As to the mere
word. Quaternion, it signifies primarily (as is well known), like
Latin original, " Quaternio," or the Greek noim reTpaKTvg, a Set
of Four : but
is
obviously used here, and elsewhere in the present work, in a technical sense.
its
it
by
112
rence to an angle on a desk, that the Four Elements which
ing
it
[book
II.
involves are the follow-
Ratio,
Angle,
Ledge, and Slope;
although the two latter elements are in fact themselves angles also, but are not im7nediately obtained as such, from the simple comparison of the tioo lines, of which the
the Quotient.
is
Quaternion
On
Section 4
117. It
is
Equality of Qiiaternions ; and on the Plane

of a Qy,aternion.
an immediate consequence of the foregoing con-
that two quaternions^ or tiuo quotients

ception of a Quaternion,
for simplicity to be all co-initial
vectors^
supposed
of
(18), are
as being equal to each other, or that the equation,
regarded
CD
|3
oc
a'
OB
oa'
by us considered and defined to hold good, when the two triare similar and similarly turned, and in
angles, AOB and COD,
one common plane, as represented in the j,
is
annexed Fig. 35
the relative
length
relative direction
(109), and" the
two
the
of
lines, oa, ob, being
(110),
respects the same as the relative length and the relative direction
then in
all
of the two other
oc, od.
lines,
Fig. 35.
Under the same conditions, we
118.
shall write the following
A AOB
formula of
COD
oc
direct similitude,
reserving this other formula,
A
which we
AOB
oc'
A a'ob
aob', or
shall call
o.
oc'
a'ob',
inverse simili-
formula of
two triangles, aob and
tude, to denote that the

aob', or a'ob
and
a'ob',
although otherwise simi-
lar (and even, in this case, equal,* on account

of their having a common side, oa or oa'), are
*
That
signs (28).
is
to say, equal in absolute amourit of area,
but with opposite algebraic
OB OA, and ob' OA, although not equal (110), will

be conjugate quaternions.
Under the same conditions, we shall
The two
soon be defined to
Fig. 36.
quotients
write also the formula,
A Aon'a'coD.
CHAP.
I.]
CONDITIONS OF EQUALITY OF QUATERNIONS.
113
oppositely turned (comp. Fig. 36), as if one were the reflexion

of the other in a mirror or as if the one triangle were derived
;
(or generated) from the other, by a rotation of its plane through

two right angles.
may therefore write,
We
OB OD
=
OA OC
When
119.
mon
.,.
,
II
A AOB
the vectors are thus
OC
COD.
all
drawn from one com-
ihQ plane aob oi any two of them maybe called

the Plane of the Quaternion (or of the Quotient), ob: oa; and
origin o,
of course also the plane of the inverse (or reciprocal) quaternion (or of the inverse quotient), oa; ob.
And any two quacommon
which
have
a
ternions^
plane (through o), may be said
to be
Complanar* Quaternions^ or complanar quotients, or

fractions
but any two quaternions (or quotients), which have
;
different planes {intersecting therefore in a right line
through
the origin), may be said, by contrast, to be Diplanar.

120. Any tioo quaternions, considered as geometric fractions (101),
can be reduced
a common denominator without
to
Let
chanoje
of the valuej- of either of them, as follows.
'
be the two given fractions, or quaternions
and
if
OB
and
OA
they be
complanar (119), let oe be any line in their common plane; but

if they be diplanar (see
again 1 19), then let oe be any assumed
part of the
line
of intersection of the two planes
each case, the line oe
is
also in the plane cod.

lines, OF,
so that, in
situated at once in the plane aob, and

can then always conceive two other
We
og, to be determined so as to satisfy the two condi-
tions of direct similitude (118),
A EOF a
*
It
is,
Aeogoccod;
AOB,
however, convenient to extend the use of this word, complanar, so as to
include the case of quaternions represented
by angles
in parallel planes.
Indeed, as
which have equal lengths, and similar directions, are equal (,2), so the
quaternion, which is a quotient of two such vectors, ought not to be considered as
all
vectors
undergoing any change, when either vector

port without rotation.
t That
is to say,
the
new
is
merel}^
changed
in position,
hy
a.
trans-
or transformed quaternions will be

respectively equal to
the old or given ones.
14
and therefore
[bOOK
II.
between quotients (117,
also the tioo equations
118),
OF
OE
OB
OG
OD
Oa'
OE
OC
and thus the required reduction
is
mon denominator sought, while
of,
numerators.
It
may
plane aob, such that
OE
OH
oe being the com-
effected,
new or reduced
h be a new point in the
aob, we shall have also,
og
be added that
A hoe oc
'
OB
are the
if
OF
~0A~ OE
'
and therefore, by 106, 107,

OD OB OG + OF
OD OB OG
~
~
OE
OC OA
OC*OA~Of'
OD OB
'
OC OA
OG
OH
'
whatever two geometric quotients (complanar or diplanar)

be represented by ob oa and od oc.
:
may
121. If now the
two triangles aob, cod are not only

but
similar
planar
directly
(118), so that A aob oc cod, we
a
eof
have
A
eog; so that we may write of
evidently
=
tioo
neio lines of, og (or the two
F
the
G, by 20),
(or
comshall
= og
new
The general construction

points f, g) in this case coinciding.
for
the
reduction
to
a
common
denominator, gives there(120),
fore here only one new triangle, eof, and one new quotient,
of OK, to which in this case each (comp. 105) of the two given
:
equal and complanar quotients, ob oa and od oc, is equal.

122. But if these two latter symbols (or the fractional
:
forms corresponding) denote two diplanar* quotients, then the

two new numerator -lines, of and og, have different directions,
as being situated in two different planes, drawn through the new
denominator-line oe, without having either the direction of that
line itself or the direction opposite thereto
they are therefore
;
(by 2) unequal vectors, even

equally long;
whence
if
it follow^s
they should happen to be
(by 104) that the
tivo
new
and therefore
also (by 105) that the two old or given

quotients,
are
quotients,
unequal, as a consequence of their diplanarity.
* And therefore non
scalar (108)
for a scalar, considered as a quotient (17),
has no determined plane, but must be considered as complanar with every geometric
;
quotient; since
it
may
be represented (or constructed) by the quotient of two simiany proposed plane whatever.
larly or oppositely directed lines, in
CHAP.
PLANE OF A QUATERNION.
I.]
115
from this analysis, that diplanar quotients of

and therefore that Diplanar Quaternions (119), are
always unequal ; a new and comparatively technical process
thus confirming the conclusion, to which we had arrived by
general considerations, and in (what might be called) a popular
way before, and which we had sought to illustrate (comp. Fig.
It results, then,
vectors,
34) by the consideration of angles on a desk : namely, that a

Quaternion, considered as the quotient of two mutually inclined
lines in space, involves generally a Plane, as an essential part
(comp. 110) of its constitution, and as necessary to the completeness of its conception.
123.
propose to use the
We
mark
III
as a Sign of Complanarity, whether of lines or of quotients

thus we shall write the formula,
7lll.^>
to express that the three vectors, a,
be made)
/3,
y,
supposed to be (or to
plane ; and the
co-initial (18), are situated in one
analogous formula,
orîir^,
a
y
q\\\q,
to express that the two quaternions, denoted here by q and q',

and therefore that ihefour vectors a, /3, y, 8, are complanar
And because we have just found (1*22) that diplanar
(119).
^
quotients are unequal,
we
includes two complanarities
may
see that one equation of quaternions

of vectors; in such a manner that we
write,
7|||a,/3,
and
the equation of quotients,

the four lines
from o be
in
|||
a, /3,
=
one
if
-=^;
a
y
,
being impossible, unless
common plane.
We
all
shall also
employ the notation

to express that the vector
the quaternion q.
is in
(or parallel to) the plane
of
116
124.
With
[book
II
we may
the same notation for complanarity,
write generally,
xa|||a,/3;
and
j3 being any two vectors^ and x being any scalar ; because, if a = OA and j3 = ob as before, then (by 15, \1) xa = oa',
where a' is some point on the indefinite right line through the
so that the plane aob contains the line oa'.
points o and a
For a
similar reason,
we have
generally the ibllowing formula
o^ complanarity of quotients,
whatever two scalars x and y may be

any two vectors.
125. It
if
whence
it is
/3
still
denoting
evident (comp. Fig. 35) that
is
A AOB a
a and
and Aaocc/bod;
then Aboaocdoc,
COD,
easy to infer that for quaternions, as well as for
ordinary or algebraic quotients,

jS
if
^ .,
o
CL
a
,
-, then, inversely, ts
.
-4,
and
h
C
,.
alternately^
being permitted now to establish the converse of the

mula of 1 18, or to say that
it
II
OB OD
OA
Under the same
oc
condition,
then
A aob a
^
7s'->
last for-
cod.
by combining inversion with
nation, wê have also this other equation,

7
alter-
= ^.
o
126. If the sides, oa, ob, of a triangle aob, or those sides

either
way
prolonged, be cut (as in
Fig. 37) by &UJ parallel, a'b' or a"b",

to the base ab, we have evidently the
relations of direct similarity (118),
a'ob' oc
aob,
a"ob"
cx
aob
whence (comp. Art. 13 and Fig.

it
follows that
we may write,
12)
for qua-
ternions as in
algebra, the general ^'

equation, or identity,
Fig. 37.
CHAP.
AXIS
I.]
AND ANGLE OF A QUATERNION.
xa
117
'
again ani/ scalar, and a, j3 are ani/ two vectors. It

is
easy also to see, that for any quaternion q, and any scalar x,
we have the product (comp. 107),
where x
is
Xq
^
=-'
p a
=z
=
a
X~â

x â
qj;
SO that, in the multiplication of a quaternion by a scalar (as in

a scalar, 15), the order of the
the multiplication of a vector
factors
is
by
indifferent.
Section 5
the
On
the
Axis and Angle of a Quaternion
and on
Index of a Right Quotient^ or Quaternion.
has been already said (111, 112), we are

that the Axis, or more fully that the
to
define
led
naturally
ob : oa,
positive axis, of any quaternion (or geometric quotient)
127.
is
From what
a right line perpendicidar to the plane aob of that quaternion

is such that the rotation round this axis, from the divisor;
and
:
or (as we shall
the
directed towards
right-hand,* like the
line OA, to the dividend-line ob, is positive
henceforth assume)
motion of the hands of a watch.
To
more
definite this conception of the

add, 1st, that the rotation, here
spoken of, is supposed (112) to be the simplest possible, and
therefore to be in the plane of the two lines (or of the quater-
128.
render
axis of a quaternion,
still
we may
nion), being also generally less than a semi-revolution in that

plane; Ilnd, that the axis shall be usually supposed to be a
ox drawn /rom
the assumed origin o

and llird, that the
to
of
this
line
shall
be
be
given, or fixed, and
supposed
length
so that the term
to be equal to some assumed unit of length
line
X, of this axis ox, is situated (by its construction) on a given

spheric surface described about the origin o as centre, which
surface
we may
call
the surface of the unit-sphere.
129. In this manner, for every given non-scalar quotient

*
This
is,
of course, merely conventional,
stitute the /e/V-hand throughout.
and the reader may
(if
he pleases) sub-
118
[boOK
II.
(108), or for every given quaternion q which does not reduce

degenerate) to a mere positive or negative number, the
itself (or
axis will be an entirely definite vector, which
may be called an
on
its
account
of
assumed
UNIT-VECTOR,
length, and which we
shall denote*, for the present, by the symbol Ax q. Employing then the usual sign of perpendicularity, J_, we may now
write, for any two vectors a, j3, the formula
.
Ax.â;
a
Ax.2j3;
a
Ax.^
a
or briefly,
The Angle of a
R
[<^
quaternion, such as ob oa, shall

simply be, with us, the angle aob between the two lines, of
130.
which the quaternion
is the quotient
this angle
being supto
one
of
usual
kind
here
be
the
posed
(such as are considered
;
by Euclid) and therefore being acute, or right, or obtuse (but

not of any class distinct from these), when the quaternion is a
shall denote this angle of a quaternion
non-scalar (108).
:
We
q,
by the symbol, L q
inequalities^ following
and thus
Z<7>0;
where
tt is
131.
shall have, generally, the
two
Lq<Tr;
used as a symbol for two right angles.
When
the general quaternion,
q^
degenerates into a
scalar, x, then the axis (like the planeX) becomes entirely indeterminate in its direction; and the angle takes, at the same
time, either zero or two right angles for its value, according as
the scalar is positive or negative. Denoting then, as above, any
such scalar by
x, wê
have
* At a later
stage, reasons will be assigned for denoting this axis, Ax .5, of a
quaternion q, by the less arbitrary (or more systematic) symbol^ UVgr ; but for the
present, the notation in the text may suffice.
t In some investigations respecting complanar quaternions^ and powers or roots
it is convenient to consider negative angles^ and angles greater than
two right angles: but these may then be called amplitudes and the word "An"
Ratio," may thus be restricted, at least for the present, to its
gle," like the word
of quaternions,
ordinary geometrical sense.

X
Compare the Note to page 114. The angle,

when the quaternion reduces itself to
determinate,
as well as the
zero
unless
axw, becomes
we happen
to
in-
know
a law, according to which the dividend-line tends to become null, in the transition
from
to -.
CHAP.
I.]
CASE OF A RIGHT QUOTIENT, OR QUATERNION.
Ax
Z
a:
= an indeterminate unit-vector
Z re = tt, if a? < 0.
0, if X >
a-
119
132. Oi' non-scalar quaternions, the most im- B

as in
portant are those of which the angle is right,
the annexed Figure 38

q=
OA
OB
and when we have thus,
,
,
and OB _L oa,
or
""
Lq = -,
^
.
>
Right
then be said to be a
the quaternion q may

Quotient ;* or sometimes, a Eight Quaternion.
(1.) If then
but p
a = OA and
p-op, where o and a
OA
are two given {or fixed) points,
a variable pointy the equation
is
expresses that the locus of this point p

line
Fig. 38.
(2.)
the plane through o, perpendicular to the
is
ofperpendicularity p j_ a (129).
/3= ob, b being any third given point, the equation,
for it is equivalent to the ybrmttZa
More
generally, if
P
a
one sheet of a cone of revolution, with o for vertex,

and OA for axis, and passing through the point b because it implies that the angles
AOB and AOP are equal in amount, but not necessarily in one common plane.
expresses that the locus of p
is
(3.)
The equation (comp. 128, 129),
Ax.^ = Ax.?,
a
a
expresses that the locus of the variable point
indefinite half-plane,
which contains
is
the given plane aob ; or rather the

p that are at once complanar
the points
all
with the three given points o, A, b, and are also at the same side of the indefinite
right line OA, as the point b.
(4.)
The system
of the two equations,
Ax.^ = Ax.^,
l'-=L^-,
a
expresses that the point p
is situated,
either
prolonged through /, but not through o
on the^wiVe right
^j^^
line OA, or
so that the locus of
may
on that
line
in this case be
said to be the indefinite half-line, or ray, which sets out from o in the direction of the
understood to be a scaand we may write p = x/3, ic>
vector OB or /3
(cc being
;
lar), instead of the equations assigned above.
* Reasons will afterwards be

assigned, for equating such a quotient, or quatera Vector; namely to the line which will presently (133) be called the Index
nion, to
of the Right Quotient.
120
Ax.e=-Ax.^,
a
a
ê=^_,^,
a
a
expresses that the locus'oi p
tor
BO
is
the opposite ray from o
situated on the prolongation
is
(1)
or that p =aj/3,
= x(3\ x>0,
(Comp.
II.
This other system of two equations,
(5. )
or that
[boOK
a:
<
if /8'
of the revec-
or that
y'
-pX
= ob' = - /3.
Fig- 33, bis.
Fig. 33, bis.)
(6.)
Other notations, for representing these and other geometric
loci, will
be found
by the Calculus of Quaternions ; but it seemed

the present stage, as serving already to show that even
to be supplied, in great abundance,
proper to point out these, at

the two symbols of the present Section, Ax. and Z., when considered as Characteristics of Operation on quotients of vectors, enable us to express, very simply and concisely, several useful
geometrical conceptions.
line, oi, be drawn in the direction of the

ox of such a right quotient (and therefore perpendicular,
133. If a third
axis
by 127, 129, to each of the tioo given rectangular lines, oa,

ob) and if the length of this new line oi bear to the length
of that axis ox (and therefore also, by 128, to the assumed
unit of length) the same ratio, which the length ofthe dividend;
line,
OB, bears to the length of the divisor-line, oa ; then the

determined, is said to be the Index of the Right
line 01, thus
And it is evident, from this definition of such an

with our general definition (117, 118) of
combined
Index,
Quotient.
Equality heticeen Quaternions, that two right quotients are

equal or unequal to each other, according as their two indexlines (or indices) are equal or unequal vectors.
Section 6
On
the Reciprocal, Conjugate, Opposite,
and Norm
of a Quaternion; and on Null Quaternions.

134.
The Keciprocal
quaternion, such as
=
^^
is
(or the Inverse,
comp. 119) of a
that other quaternion,
which
is
formed by interchanging the divisor- Hjw and the diviand in thus passing from any non-scalar quater-
dend-line;
nion to
its reciprocal, it is
evident that the angle (as lately
CHAP.
RECIPROCAL OF A QUATERNION.
I.]
121
defined in 130) remains unchanged^ but that the axis (127,

128) is reversed in direction: so that we may write generally,
pa
The
product of two reciprocal quaternions is always

to
equal
positive unit?/; and each is equal to the quotient of
divided
by the other; because we have, by 106, 107,
unitt/
135.
It
is
new or peculiar
mutual relation existing between a
reciprocal; since, \f one be denoted by the
therefore unnecessary to introduce any
notation, to express the
quaternion and its

q, the other
symbol
may
(in the present
System, as in Alge-
bra) be denoted by the connected symbol,*
q,
have thus the two general formulae (comp. 134)

1
or -.
We
A
A
= -Ax.^7.
Ax.-
L-=Lq\
'
yet entering on the general \X\QOxy of multiof quaternions, beyond what has been
and
division
plication
done in Art. 120, it may be here remarked that if any two
quaternions q and cq be (as in 134) reciprocal Xo each other, so
136.
Without
(by 135), and \f q be any third quaternion, then

(as in algebra), we have the general formula,
that
q-q^
q :q
q .q
=q
'-\
if (by 120) we reduce q and q" to a common denominaand denote the new numerators by /3 and 7, we shall have
because
tor a,
(by the definitions in 106, 107),
7/3770
137.
When two
complanar triangles aob, aob', with a com-
* The
symbol q'^, for the reciprocal of a quaternion /, is also permitted in the
present Calculus; but we defer the use of it, until its legitimacy shall have been
established, in connexion with a general theory of powers of Quaternions,
'
122
man
[bOOK
II.
side oa, are (as in Fig. 36) inversely similar {I IS), so that
the formula
OB
Quotients.*
^
OA
aob' a' aob holds good, then the hvo unequal
and
OA
OB'
r-\
/>
Conjugate Uuater-
are said to be
and if they^r^^ of them be still denoted by q, then the

is thus the
which
second,
conjugate of that j^r^^, or of any other
which
is
quaternion
equal thereto, is denoted by the new symNiONS
K^
bol,
which the
in
K may be said
letter
to be the Charac-
teristic of Conjugation.
Thus, with the construction above
supposed (comp. again Fig. 36), we may write,
OB
OA
1st,
138.
From
that
if
ob'
^
xôb
OA
t^
OA
-*
this definition of conjugate quaternions, it follows,
the equation
^
= K holdgood, then
OA
OA
the
/meon'mavbe
-^
called (118) the rejkxion of the line ob (and conversely, Xkvo^latter line
the reflexion of the/ormer), with respect to the line oa Ilnd, that, under
;
the same condition, the line OA (prolonged

pendicularly the line bb', in
and
some point
if
necessary)
bisects
per-
a' (as represented in Fig. 36)
any two conjugate quaternions (like any Uvo reciprocal

quaternions, comp. 134, 135) have equal angles, but opposite axes:
Ilird, that
so that
we may
write, generally,
K^ = L
q;
Ax K^ = - Ax
.
and therefore! (by 135),
zKa^Z.-;
139.
-, or
x''^,
The
Ax.Ko = Ax.-.
reciprocal of a scalar, x,
having the same algebraic
is
simply another scalar,
sign,
and
in all other re-
But the conjugate K^, of a

speCts related to x as in algebra.
scalar x, considered as a limit of a quaternion, is equal to that
scalar x itself; as may be seen l)y supposing the two equalhut
opposite angles, aob and aob', in Fig. 36, to tend together to
*
Compare the Note
It will
to page 112.
soon be seen that these two last equations (138) express, that the con-
jugate and the reciprocal, of any proposed quaternion

although they have in general unequal tensors.
q,
have always equal
versors,
CHAP.
I.]
CONJUGATE AND NULL QUATERNIONS.
123
We may therefore
write, gene-
two right
zero, or to
rally,
angles.
Kx = x,
if
a;
be any scalar
and conversely*,
5-
=a
if
scalar,
K^*
because then (by 104) we must have ob = ob', bb'=0; and

therefore each of the two (now coincident) points, b, b', must
be situated somewhere on the indefinite right line oa.

140. In general, by the construction represented in the
same Figure, the sum (comp. 6) of the tivo numerators (or dividend-lines,
ob and
ob'), of the tioo conjugate fractions (or
quoand K^' (137), is equal to the double

whence (by 106), the sum of those two conju-
tients, or quaternions), q
of the line oa'
gate quaternions themselves
^ +
Kq
is,
= 0+
^
^^ =
Kg
2oa'
oa
this sum is therefore always scalar, being positive if the angle

Z ^ be acute, but negative if that angle be obtuse.
141. In the intermediate case, when the angle aob is right,
the interval oa' between the origin o and the line bb' vanishes ;
mentioned numerators, ob, ob', become two

of
which the sum is null (5). Now, in geneopposite vectors^
will be found useful, or rather
and
is
natural,
ral, it
necessary
and the two
lately
admit that a null

an actual vector, gives always a Null Quaternion as the quotient; and to denote this null quotient by
In fact, we have (by 106) the
the usual symbol for Zero.
(for consistency yf\i\\ former definitions), to
vector, divided by
equation,
a
the zero in the numerator of the Ze/J^-hand fraction representing here a null line (or a null vector, 1,2); but the zero on the
rz^^^hand side
quaternion).
"of
And
* Somewhat later
as
9 = 0; and
it
the equation denoting a null quotient (or

thus we are entitled to infer that the sum,
will be seen that the equation
that this last
is
degenerates (131) into a scalar.
another
mode
^q = q may
also be written
of expressing that the

quaternion, q,
124
K^* +
[bOOK
II.
or q +
Kq, of a right-angled quaternion^ or right quo(132), and of its conjugate, is always equal to zero.
142.
have, therefore, the three following formulae,
whereof the second exhibits a continuity in the transition from
5^,
tient
We
the^r^^ to the third
IT
!...</ + K^' >
0,
if
ll...q + Kq =
0,
if
+ K^- <
0,
if
III.
5'
<
^ q
lq=l;
^q > ^'
id
And
because a quaternion, or geometric quotient, with an actual and Jinite divisor-line (as here oa), cannot become equal to
zero unless
dividend-line vanishes, because (by 104) the
its
equation
/3
L
=
= - requires the equation
a
if a
the
j3
0,
be any actual and finite vector, we may infer, conversely, that

sum q + K^' cannot vanish, without the line oa' also vanish-
that is, without the lines ob, ob' becoming opposite vectors,
and therefore the quaternion q becoming a right quotient {1^2),
ing
We are therefore entitled
to establish the three following cow-
verse formulae (which indeed result
from the three former)
+ K^' > 0,
then
Lq
= 0,
^ + K(7
then
Lq=-\
.\i q-V l^q < 0,
then
Lq>-.
r.
if
II'.
if
Iir.
9-
<--,
IT
143.
When
two opposite vectors (1), as
j3
and-j3, are both
divided by one common (and actual) vector, a, we shall say that

the two quotients, thus obtained are Opposite Quaternions;
so that Ihe opposite of any quaternion q, or of any quotient
j3
a,
may be denoted
-P
as follows (comp. 4)
0-^0/3
^=0-^=-^;
'
CHAP.
OPPOSITE QUATERNIONS.
I.]
125
while the quaternion q itself maj, on the same plan, be denoted

The sum of any two
(comp. 7) by the symbol + q, or + q.
opposite quaternions is zero, and their quotient is negative
unHy; so that we may write, as in algebra (comp. again 7),
(-^):^ = -l;
=
=
(-^) + ^ (+^) + (-^) 0;
because, by 106 and 141,
a
The
site
-^ = (-1)^;
reciprocals of opposite quaternions are themselves oppoor in symbols (comp. 126),
11,
= --,
a
-a
a
= -q- = -75-75
-(5
15
j5
because
-q
Opposite quaternions have opposite axes, and supplementary

angles (comp. Fig. 33, bis) ; so that we may establish (comp.
132, (5.)
the two following general formulae,
L{-
q)
= Tr- Lq'i
Ax.(-
q)
= - Ax.q.
We
144.
may also now write, in full consistency with the
recent formulae II. and IF. of 142, the equation,
IF..
Kq = -q,
and conversely* (comp. 138),

=IF'.
if
.
K^'
Z^ =
if
then
q,
TT
Kq ^Lq = ^.
In words, the conjugate of a right quotient, or of a right-angled

(or right) quaternion (132), is the right quotient opposite
thereto ; and conversely, if an actual quaternion (that is, one
which
is
not null) he opposite to
its
own
conjugate,
it
must be
a right quotient,
(1.)
If then
we meet
the equation,
Ke = -^,
a
>ve shall
know
that p
-i-
or
and therefore
^
a
(if
+ K^ = 0,
a
= OA, and
= op,
as before), that the
It will be seen at a later stage, that the

equation K^ = 5, or g + K7 =
be transformed to this other equation, Sg =
and that, under this last form, it
expresses that the scalar part of the quaternion 7 vanishes : or that this quaternion
is a right
quotient (132).
may
126
locus of the point
On
(2.)
is
[book
the plane through o, perpendicular to the line
OA
II.
(as in 132,
the other hand, the equation,
K'
^-K^ = 0,
or
expresses (by 139) that the quotient p a is a scalar ; and therefore (by 131) that
or tt ; so that in this case, the locus of p is the indefiits mtgle L
(p a) is either
:
nite right line
145.
through the two points O and A.
As
the opposite of the opposite, or the reciprocal of the red-
procal^ so also the conjugate ofthe conjugate, of any quaternion, is that
or in symbols,
quaternion
itself;
so that,
abstracting from the subject
by
briefly,
of the operation, we
may write
K2=KK=1.
is
easy also to prove, that the conjugates of opposite quaternions are
themselves opposite quaternions ; and that the conjugates of reciprocals
are reciprocal: or in symbols, that
It
I...K(-g) = -K(?,
K^+K(-^) = 0;
or
and
or
1:%,
11.
K^.K-=1.
9
Kq is included (comp. 143) in this more general

(1.) The equation K(q') =
formula, 'K.(^xq) = xKq, where x is any scalar; and this last equation (comp. 126)
may be proved, by simply conceiving that the two lines ob, ob', in Fig. 36, are
multiplied by any common scalar or that they are both cut by any parallel to the
;
line bb'.
(2.)
To prove
that conjugates of reci.
procals are reciprocal, or that
we may
K^
K - = 1,
conceive that, as in the annexed /

bis, while we have still the
Figure 36,
relation of inverse similitude,
A aob' a' AOB
(] 18, 137),
as in the former Figure 36, a

is
new point c
OA itself,
determined, either on the line
or on that line prolonged through A, so as

Fig. 36
to satisfy either of the two following connected conditions of direct similitude

:
A bog a
or simply, as a relation
aob' ;
b'oc
between the /our points
A BOG
oc
aob
o, A, b, c, the formula,
a' AOB.
127
GEOMETRICAL EXAMPLES.
CHAP.
I.]
For then
vfe shall
have the transformations,
OA
OB
= , = .
=
OB
OC
1
OA
1
K- = K = K
Ob'
OC
OB
Jvg
The two quotients, ob OA, and ob oc, that is to say, the quaternion q
and the conjugate of its reciprocal, or* the reciprocal of its conjugate, have
the same angle, and the same axis ; we may therefore write, generally,
:
(3.)
itself,
Ax.K- = Ax.5.
L^-=lq;
(4.) Since
OA ob and OA ob' have thus been proved (by sub-art. 2) to be

we can now infer this theorem, that ang two geo:
a pair of conjugate quotients,
and
metric fractions,
which have a common numerator
ternions, if the denominator
a, are conjugate
of the second be the reflexion of the denominator
/3'
common numerator
qua(5
of
138, I.); whereas it had

only been previously assumed, as a definition (137), that such conjugation exists,
under the same geometrical condition, between the two other (or inverse) fractions,
ihQ first, with respect to that
and
the three vectors a,
(5.) Conversely,
if
we
being supposed to be
j3, j3'
know
all co-initial
(18).
meet, in any investigation, the formula
OA ob' = K (oA
we
(^covnp.
ob),
the reflexion of the point b, with respect to the

li?ie OA ; or that this line, OA, prolonged if necessary in either of two opposite direcline bb', in some point a', as in either of the two
tions, bisects at right angles the
shall
that the point b'
is
Figures 36 (comp. 138, II.).

conditions of construction,
(6.) Under the recent
mentary
it
follows from the
most
ele-
which passes through the three points

and that this line is, in length, a mean
principles of geometry, that the circle,
A, B, c, is touched at B,
hythe right line OB
Let then od be such a geometric mean, and

proportional between the lines OA, oc.
let it be set off from o in the common direction of the two last mentioned lines, so
that the point
symbols, y, d;
d falls between A and c also let the vectors

we shall then have expressions of the forms,
;
= aa,
positive scalar, a
>
(comp. sub-art. 2) with this scalar
a,
where a
is
some
= K OA
OB
OB
(7.) Conversely, if
we
still
verse similitude of triangles,
or
= K OB
^^
OA
OB
A boc
<x'
aob
and
K - and -
q
Ng
Kq
may
it
of
/3
a,
it is
to the quaternion itseU divided
will
be connected
Pa
this last
formula expresses the in-
expresses nothing more: or in other
common
value of these two equal quater-
be represented by either of the two
or in words, that
by the formula,
aâ ^ R
-
= K -.
or
= aâ,
suppose that y
It will bo seen afterwards, that the
nions,
or q
oD be denoted by the
= aâ,
and the vector
and with the vector
OC
,^
oc
y
;
oc,
new
symbols, Ug-
Tq,
equal to the versor divided by the tensor ; and also
by the norm.
128
words,
by the vector
it is satisfied
(3
of every point B,
II.
[bOOK
which gives that inverse
simili-
purpose it is only requisite that the length of ob should be (as

above) a geometric mean between the lengths of OA, oc or that the two lines, ob,
OD (sub-art. 6), should be equally long: or finally, that b should be situated some-
But
tude.
for this
where on the surface of a sphere, which is described so as to pass through the point D
for its centre.
(in Fig. 36, bis), and to have the origin o
we meet an
(8). If then
equation of the form,
pa
= K-,
a'â
in
- OA,
which a
of the point P
aa
for
is
a radius
-K-=a-,
a
a
or
p = op, and a is a scalar, as before, we shall know that the locus

a spheric surface, with its centre at the point O, and with the vector
and
this spheric locus of
we determine a
common orthogonal to
also that if
is
by the equation oc = a'â,

apc, which can be
A and c because every radius
point c
all the circles
described, so as to pass through the two fixed points,
OP
ob
of the sphere
is
to
ABC
is
a tangent, at the variable point
P, to the circle
apc, exactly as
in the recent Figure.
(9.) In the
same
Fig. 3G, his, the similar triangles
of
ples) that the length
in the simple ratio of
bc
OD
to that of
is
to
OA
ab
in the
or as the scalar a to 1.
we
search, the recent equation in p (sub-art. 8),
length of(^p
show (by elementary
sub-duplicate ratio of oc to
aâ) = a x
shall
length
If then
know
we
meet, in
princi-
OA
or
any re-
that
of(p a);
while the recent interpretation of the same equation gives this other relation of the
same kind
length of p
a subsequent stage,
(10.)
supplies Rules
of
or
x length of a,
shown that the Calculus of Quaternions

of Transfoi-mation, by which we can pass from any one to any other
these last equations respecting p, without (at the time) constructing any Figure,
(immediately) appealing to Geometry: but it was thought useful to point out,
At
already,
it
will be
how much geometrical meaning*
is
contained in so simple a formula, as that
of the last sub-art. 8.
(11.)
The product of two conjugate quaternions
NoRM,t and
is
denoted thus
is
said to be their
common
qKq = N5.
*
student of ancient geometry may recognise, in the two equations of sub-art.
a sort oi translation, into the language of vectors, of a celebrated local theorem of
Apollonius of Perga, which has been preserved through a citation made by his early
9,
commentator, Eutocius, and may be thus enunciated Given any two points (as here
A and c) in a plane, and any ratio of inequality (as here that of 1 to a), it is possible
:
to construct a circle in the plane (as here the circle bdb'), such that the (lengths of
here ab and cb, or ap and cp), which are inflected from the
the) two right lines (as
two given points to any common point (as B or p) of the circumference, shall be to
ratio. (Auo do9svT(ov arjfxiiiov, k. t. \. Page 11 of Halley's
Edition of Apollo ni us, Oxford, mdccx.)
f This name, Norm, and the corresponding characteristic, N, are here adopted,
as suggestions from the Theory of Numbers ; but, in the present work, the\' will not
each other in the given
CHAP.
RADIAL QUOTIENTS, RIGHT RADIALS.
I.]
129
It follows that NKg = Nj ; and that the norm of a quaternion is generally a positive
scalar: namely, the square of the quotient of the lengths of the two lines, of which
In fact we have, by sub-art.
(asvectors) the quaternion itself is the quotient (112).
6,
and by the definition of a norm, the transformations
OB _
OB
OC OB
OA
OA
OB'
As a
N0 =
limit,
we may say
_ OC OB _ OC
OB OA
OA
that the
\lengthofa
norm of a
&:'
OA
null quaternion is zero; or iu symbols,
0.
(12.) "With this notation, the equation of the spheric locus (sub-art. 8), which
has the point o for its centre, and the vector aa for one of its radii, assumes the
shorter form
N? = a3:
Section
7.
or
N-^=l.
On Radial Quotients;
and on
the Square of a
Quaternion.
was early seen (comp. Art. 2, aud Fig. 4) that any

AB, AC, of any one circle, or sphere, are necessarily
146. It
two radii,
unequal vectors ; because their directions differ. On the other

hand, when we are attending only to relative direction (110),
we may suppose that all the vectors compared are not merely
co-initial (18), but are also
equally long; so that if their common length be taken for the unit, they are all radii, oa, ob,
.
of what
we have called the
Unit- Sphere (128), described round
the origin as centre; and may all be

said to be Unit-Vectors (129).
And
then the quaternion, which is the

quotient of any one such vector divi-
ded by any other, or generally the

^'^' ^^"
quotient of any two equally long vectors, may be called a Radial Quotient; or sometimes simply a Radial. (Compare the annexed Figure 39.)
be often wanted, although it may occasionally be convenient to
employ them.
we shall soon introduce the conception, and the
of the
For
Tensor, Ty, of
a quaternion, which is of greater geometrical
utility than the Norm, but of which it
will be proved that this norm is
simply the square,
characteristic,
qKq = 'Sq = (Tqy,

inpare the Note to sub-art. 3.
130
[book
II.
The two
Unit- Scalar s^ namely, Positive and Negaconsidered as limiting cases of radial quobe
Unity may
and tt, of the
tients^ corresponding to the two extreme values^
147.
tive
In the intermediate
angle aob, orZ5'(131).
when aob
case^
is
a right angle^ or
Lq =
-^
as in Fig. 40, the resulting quotient, or quaternion, may be called (comp. 132) a Right
Radial Quotient; or simply, a Right Radial.
Fig. 40.
The consideration of such right radials
will be found to be of great importance, in the whole theory

and practice of Quaternions.
148. The most important general property of the quotients
mentioned is the following that the Square ofevery Right

Radial is equal to Negative Unity ; it being understood that
last
we
write generally, as in algebra,
q.q=qq = q\
and
product of two equal quaternions the square of
call this
as in Fig. 41, we
describe a semicircle aba', with o for cen-
For if,
each of them.
and with ob for the bisecting radius,

then the two right quotients, ob oa,
tre^
and
oa'
ob, are equal (comp. 117); and
therefore their
107)
\\iQ
common square
is
(comp.
product.
OB
OA'
OaJ
OB OA
OB
OA
OA
where oa and ob may represent any

ttvo equally long, but
mutually rect-
More
generally, the
angular
lines.
Fig. 41, his.
of every Right Quotient

to a Negative Scalar; namely, to the negative of
the square of the number, which represents the ratio of the
or to zero
lines compared
lengths'^ of the two rectangular
Square
(132)
is
equal
Hence, by 145, (11.),
g^^-Ny,
if
Lq =
CHAP.I.]
GEOMETRICALSQUAREROOTSOFNEGATIVE UNITY. 131
minus the square of the number which denotes (comp. 133) the*
as appears from
length of the Index of that Right Quotient
in
and not (as beob
is
an
which
ordinate,
41,
bis,
only
Fig.
:
fore) a radius, of the semicircle aba'
oa'
=-
(length
pA/
OA
\lengta oj oaJ
Thus
everi/
obV
o/obV
^,.
for
^.
we have
...
,
if
OB
thus,
OA.
Right Radial is, in the present System,

Roots
of
of Negative Unity ; and may thereSquare
the
one
Values of the Symbol ij -\\ which
fore be said to be
of
149.
the
07ie
celebrated symbol has thus a certain degree of vagueness, or at

least
of indetermination, of meaning in this theory, on account

we shall not often employ it. For although it thus
of which
admits of 2i perfectly clear and geometrically real Interpretation,

as denoting
what has been above
called a
Right Radial Quo-
yet the Plane of that Quotient is arbitrary; and therefore

the symbol itself must be considered to have (in the present
system) indefinitely many values; or in other words the Equatient,
tion,
has (in the Calculus of Quaternions) indefinitely many Roots,*

which are all Geometrical Reals : besides any other roots, of
a purely symbolical character, which the same equation may be
conceived to possess, and which may be called Geometrical
Conversely,
Imaginaries.'t
the
q be any real quaternion, which
* It will be
subsequently shown, that
is unity, so that
if
ic,
y, z be ani/ three scalars,
of which
sum of the squares
a;2
and
if
if i,j,
+ y2+22 =
i.
h be ani/ three right radials, in three mutually rectangular planes ; then
the expression,
= ix \jy + kz,
denotes another right radial, which satisfies {as such, and by symbolical laws to be
assigned) the equation q^ =l- and is therefore one of the geometrically real values
of the symbol V 1.
f Such imaginaries
tertiions (or rather
will be found to offer themselves, in the treatment
by what
of ideal contacts, in geometry; but

metrical reals alone.
by Quaand
will be called Biquaternions'), of ideal

intersections,
we
confine our attention, for the

present, to geoto page 90.
Compare the Notes
132
the equation q^ = - 1,
as in Fig. 42, we suppose that
satisfies
we
shall
it
II.
must be a right radial; for if,

boc,
A aob a
have
oc OB
^
'
OB
\OA
and
[book
OC
OA
OA
square of q cannot become equal to

- oa,
negative unity^ except by oc being =
or = oa' in Fig. 4 1 ; that is, by the line ob
this
being at right angles to the line oa, and

being j\,t the same time equally long, as in O
Fig. 40.
(1.) If then
we meet
the equation,
1,
(;)
= oa,
Fig. 42.
= op,
know that the locus of the point p is

the circumference of a circle^ with o for its centre, and with a radius which has the
same length as the line oa ; while the plane of the circle is* perpendicular to that
where a
given
centre
line.
is
and p
we
as before,
shall
In other words, the locus of p
the origin
and the given point
is
a great
A, on the
circle,
on a sphere of which the

surface, is one of the
same spheric
poles of that circle.

= a^, where a is any (real) scalar, requires
(2.) In general, the equation q^
that the quaternion q (if real) should be some right quotient (132)
the number a
;
denoting the length of the index (133), of that right quotient or quaternion (comp.
Art. 148,
and
Fig. 41, bis).
But the plane
of q
is still
entirely arbitrary
and
therefore the equation
like the equation
g'=
1,
which
it
includes,
must be considered
to
have
(in the
present system) indefinitely many geometrically real roots.

(3.) Hence the equation,
(;)=-
which we may suppose that a>0, expresses that the locus of the point p is a
(new) circular circumference, with the line OA for its axis,* and with a radius of
in
which the length
=ax
the length of oa.
150. It may be added that the ijidex (133), and the axis (128),
of a right radial (147), are the same; and that its reciprocal (i3i\ its
conjugate (137), and its opposite (143), are all equal to each other. Conversely, if the reciprocal of a given quatermon q he equal to the opposite
*
It
being understood, that the axis of a circle
the plane of that circle, and passing through
its
is
centre.
a right line perpendicular to
CHAP.
RADIAL QUOTIENTS CONSIDERED AS VERSORS.
I.]
of that quaternion, then q
is
right radial;
because
133
its square, q^,
then equal (comp, 136) to the quaternion itself, divided hy its opBut the conjugate
posite; and therefore (by 143) to negative unity.
is
quotient is equal to the reciprocal of Âai quotient ;

in Fig. 36, we conceive that the three lines oa, ob, ob' are
equally long, or if, in Fig. 39, we prolong the arc ba, by an equal arc
of every radial
because
if,
we have
ab',
the equation,
Kg'
And
oa
==
OA ob
ob'
-.
5-
conversely,*
if
then the quaternion
5'
K5'=
is
-,
or
if
qKq-
1,
a radial quotient.
Section 8. On the Versor of a Quaternion, or of a Vector

and on some General Formulce of Transformation.
151.
When
a quaternion 5' = j3 a is thus a radial quotient

the lengths of the two lines a and /3 are equal,
:
(146), or
when
the
of this quaternion
effect
q,
considered as a
Factor
(103),
in the equation qa = j3, is simply the turning of the multiplicand-li7ie a, in the plane of g (119), and towards the hand de-
termined by the direction of the positive axis Ax q (129),

through the angle denoted hj Lq (130) so as to bring that
line a (or a revolving line which had coincided thercAvith) into
.
a new direction : namely, into that of the product-line

with reference to this conceived operation of turning,
j3.
And
we
shall
now
say that every Radial Quotient is a Versor.

152.
Versor has thus, in general, q> plane, an axis, and
an angle ; namely, those of the Radial (146) to which it cor-
responds, or is equal : the only difference between them being

a difference in the points of view'\ from which they are respectively regarded
namely, the radial as the quotient, q, in the
;
Hence, in the notation of norms (145, (11.)), if Ng'= 1, then 5 is a radial ;

and conversely, the norrn of a radial quotient is always equal to positive unity.
t In a slightly metaphysical mode of expression it may be said, that the radial
quotient is the result of an analysis, wherein two radii of one sphere (or circle) are
compared, as regards their relative direction ; and that the equal versor
mnt of a corresponding synthesis, wherein owe radius is conceived to he
is
the imtru-
generated, by
a certain rotation, from the other.
134
[bOOK
II,
formula, 5' = j3 a ; and the versor as the {Qqvxdl) factor, q, in

the converse formula, (3 = q.a; where it is still supposed that
:
the two vectors, a and
A versor^
153.
/3,
are equally long.
like a radial (147),
cannot degenerate into ascalary
except by
acquiring ona or other of the two limit-values^
and TT. In the first case, it becomes positive unity ; and in the second
its angle
case, it becomes negative unity : each of these two unit-scalars (147)

being here regarded as ti factor (or coefficient, comp. 12), which operates on a line, to preserve or to reverse its direction.
In this view, we
say that
may
is
sor with an angle

line
on which
an Inversor
=-
is
it operates^
(comp. Fig. 41).

right versor (like
b.
and that every Bight Versor (or ver-
Semi-inversor :* because
or turns
it
it
Aa//*-mv6rf5 the
through haJf of two right angles
For the same reason, we are led to consider every

every right radial, 149, from which indeed we have
just seen, in 152, that it differs only as factor differs from quotient),
as being one of the square-roots of negative unity : or as one of the values
of the symbol y'

154. In fact
we may observe
that the
sidered as operating on a line (in its
own
of a right versor, conplane), is to turn that line,

effect
towards a given hand, through a iHght angle. If then q be such a verand if 5'a=/3, we shall have also (comp. Fig. 41), q^ = -u; so
sor,
that, if a
be any
line in the plane of a right versor q,
equation,
whence
we have
the
q.qa = -a',
natural to write, under the same condition,
it is
On
the other hand, no versor, which is not right-angled,

^/ -\; or can satisfy the equation qâ --a, as Fig.
For it is included in the meaning of this
serve to illustrate.
as in 149.
can he a value of
42 may
as applied to the theory of versors, that a rotation

or
2
through the double of the angle of q itself, is equiLq,
through
last equation,
* This word^ " semi -inversor," will not be often used ; but the introduction of it
seems adapted to throw light on the view taken, in the present work,
here, in passing,
1, when regarded as denoting a certain important class (149) of

symbol V
There are uses of that symbol, to denote Geometrical ImagiReals in Geometry.
navies (comp. again Art. 149, and the Notes to page 90), considered as connected
with ideal intersections, and with ideal contacts ; but ivith such uses of V - 1 ^6
of the
have, at present, nothing to do.
CHAP.
I.]
VERSOR OF A QUATERNION, OR OF A VECTOR.
135
valent to an inversion o^ direction; and therefore to a rotation through
two right angles.
155. In general, if o be any vector^ and if a be used as a

temporary* symbol for the number expressing its length; so
that a
is
here a positive scalar, which bears to positive unity,

1, the same ratio as that which the length of
or to the scalar +
the line a bears to the assumed unit of length (comp. 128);

then the quotient a a denotes generally (comp. 16) a new vector, which has the same direction as the proposed vector a, but
:
has
1
length equal to that assumed unit

the
Unit- Vector in the direction of a.
46)
its
unit-vector
by the symbol,
Ua =
Ua
-,
if
and so
so that
it is
(comp.
We shall denote
this
shall write, generally,
a = length of a
a be, as above supposed, the number

(commensurable or incommensurable, but positive) which represents that length, with reference to some selected standard.
156. Suppose now that 5^ = j3 a is (as at first) 2^ general
that
is,
more
fully, if
quaternion, or the quotient ofany two vectors, a and j3, whether

Such a Quaternion will not (geneequal or unequal in length.
rally) be a Versor (or at least not simply such), according to the
definition lately given
because
its effect,
when
operating as a
factor (103) on a, will not in general be simply to turn that

line (151)
but will (generally) alter the length,^ as well as the
direction.
But if we reduce the two proposed vectors, a and /3,
to the two unit-vectors Ua and Uj3 (155), and iovm. the, quotient
:
of these, we shall then have taken account of relative direction

:
and the result will therefore be a versor, in the sense
alone
lately defined (151).
We propose to call the quotient, or the
versor, thus obtained, the versor-element, or briefly, the Versor, of the Quaternion q ; and shall find it convenient to em*
We
shall soon propose a general notation for representing the lengths of vectors,
according to which the symbol Ta will denote what has been above called a ; but
are unwilling to introduce more than one new characteristic of operation, such as K,
or T, or U, &c., at one time.
By what we
shall soon call call
an act of tension, which
consideration of the tensor of a quaternion.
will lead us to the
136
[bOOK
II.
ploy the same* Characteristic, U, to denote the operation of

taking the versor of a quaternion, as that employed above to
denote the operation (155) o^ reducing a vector to the unit of

On this plan, the
length, without any change of its direction.
symbol
Uq
ofq ; and the foregoing

enable us to establish the General Formula :
will denote the versor
finitions will
de-
Va
which the two unit-vectors, Ua and Uj3, may be called, by

analogy, and for other reasons which will afterwards appear,
the versors-f of the vectors, a and /3.
157. In thus passing from a given qiiaternion, q, to its ver-
in
we have
only changed (in general) the lengths of the

two lines compared, namely, by reducing each to the assumed
sor, \]q,
unit of length (155, 156), without
Hence
making any change
in their
plane (119), the axis (127, 128), and

the angle (130), of the quaternion, remain unaltered in this
passage so that we may establish the two following general
formulae
directions.
i\iQ
L\]q = Lq',
More
generally
we may
Ax U^ = Ax
.
q.
write,
* For the moment, this double use of the characteristic U, to assist in

denoting
Ua derived from a given line a, and also the versor JJq derived
both the unit-vector
from a quaternion q^ may be regarded as established here by arbitrary definition;

but as permitted, because the difference of the symbols, as here a and q, which serve
for the present to denote vectors
and quaternions, considered as the subjects of these
two operations U, will prevent such double use of that characteristic from giving
to
any confusion.
But we
rise
shall further find that several important analogies are
anticipation expressed, or at least suggested,
when
the proposed notation
is
by
employed.
Thus it will be found (comp. the Note to page 119), that every vector a may usefully
be equated to that right quotient, of which it is (133) the index ; and that then the
unit-vector \] a may be, on the same plan, equated to that right radial (147), which
is
(in
the sense lately defined) the versor of that right quotient.
We shall
also find
ourselves led to regard every unit-vector as the axis of a quadrantal (or right') rotation, in a plane perpendicular to that axis; which will supply another inducement,
On the whole, it appears that there will

be no inconvenience, but rather a prospective advantage, in our already reading the
^^
versor of a;'" just as we may read the analogous symbol Vq, as
symbol Ua as
to speak of every such vector as a versor.
*^
versor ofq."
t Compare the Note immediately preceding.
CHAP.
I.]
EQUAL AND RECIPROCAL VERSORS, REVERSORS. 137
Lq = Lq, and Ax
<][
Ax
q,
if
U^' = JJq
the versor of a quaternion depending solely on, but conversely

being sufficient to determine, the relative direction (156) of the
two
of which (as vectors) the quaternion itself is the quo(112); or the axis and angle of the rotation, in the plane
lines,
tient
of those two lines, from the divisor to the dividend
(1
28)
so
two quaternions, which have equal versors, must also

have equal angles, and equal (or coincident) axes, as is expressed by the last written formula.
Conversely, from this
that any
dependence of the versor \]q on relative direction'^ alone, it

follows that any two quaternions, of which the angles and the
axes are equal, have also equal versors; or in symbols, that
if
\Jq'=\]q,
Lq'=Lq,
and
Ax.^-'
Ax.'.
For example, we saw (in 138) that the conjugate and the reciprocal of any quaternion have thus their angles and their
axes the same ;
it
follows, therefore, that the versor
of the
conjugate is always equal to the versor of the reciprocal; so

that we are permitted to establish the following general for-
mula, f
158.
it
Again, because
follows that the versor of the reciprocal oi ^nj quaternion is,

same time, the reciprocal of the versor ; so that we may
at the
write,
* The unit-vector
Ua, which we have recently proposed (156) to call the versor
manner on the direction of that vector alone; which
exclusive reference^ in each of these two cases, to Direction, may serve as an addi-
of the vector a, depends in like
tional motive for employing, as
we have
and one common
U, to
characteristic,
lately done, one
common name, Versor,
assist in describing or
denoting both the Unit-
= U/3 Ua all
danger of confusion being sufficiently guarded against (comp. the Note to Art. 156),
by the difference of the two symbols, a and q, employed to denote the vector and the
Vector
Ua
itself
and
the Quotient of two such Unit- Vectors, TJq
quaternion, which are respectively the subjects of the two operations
two operations agree
in this essential point, that each
titative element, of absolute or relative length.
Compare
the Note to Art. 138.
while those
serves to eliminate the

quan-
138
U- = :^;
q
or
II.
[bOOK
U^.U-=1.
^
q
\Jq
Hence, by the recent result (157), we have
UK^ = =i-;
or,
also, generally,
U^.UK^ =
1.
Also, because the versor \Jq is always a radial quotient (151,

152), it is (by 150) the conjugate of its own reciprocal ; and
therefore at the
same time (comp. 145), the
reciprocal of its
oion conjugate; so that the j5roc?wc^ of tioo conjugate versor s^

or what we have called (145, (H.) ) their common Norm, is
always equal to positive unity ; or in symbols (comp. 150),
NU^ = U^.KU^=1.
For the same
reason, the conjugate of the versor of any quaequal to the reciprocal of that versor^ or (by what
has just been seen) to the versor of the reciprocal of that quaternion; and therefore also (by 157), to the versor of the con-
ternion
is
jugate; so that we may write generally, as a

cent results, the formula
summary of re-
each of these four symbols denoting a new versor, which has

the same plane, and the same angle, as the old or given versor
\]q, but has an opposite axis, or an opposite direction of rotation : so that,
with respect to that given Versor,

Re versor.
it
may
na-
turally be called a
159. As regards the versor itself, whether of a vector or of

a quaternion, the definition (155) of Ua gives,
\]xa = + Ua,
or
= - Ua,
according as
rc> or <
because (by 15) the scalar coefficient x preserves, in the first

case, but reverses, in the second case, the direction of the vector a; whence also, by the definition (156) o^JJq, we have
generally (comp. 126, 143),
=
= - Uq,
or
\]xq + U^',
The
according as
x>
or < 0.
versor of a scalar, regarded as the limit of a

quaternion
(131, 139), is equal to positive or negative unity (comp. 147,
CHAP.
I.J
GENERAL TRANSFORMATIONS OF A VERSOR.
139
153), according as the scalar itself is positive or negative
or
in symbols,
\Jx = +
or = -
1,
1,
according as
aj
> or <
the plane and axis of each of these two unit scalar s (147), considered as versors (153), being (as we have already seen) indeterminate.
The versor of a null quaternion (141) must be re-
garded as wholly arbitrary, unless we happen to know a laio,'^

according to which the quaternion tends to zero, before actually
reaching that limit in which latter case, the plane, the axis,
and the angle of the versor] UO may all become determined, as
The versor of a right quotient
limits deduced from that law.
;
(132), or of a right-angled quaternion (141), is always a right

radial (147), or a jnght versor (153) ; and therefore is, as such,
one of the square roots of negative unity (149), or one of the
- 1 ; while
values of the symbol
(by 150) the axis and the
index of such a versor coincide ; and in like manner its reciprocal, its conjugate, and its opposite are all equal to each other.
160. It is evident that if a proposed quaternion q be already

a versor (151), in the sense of being a radial (146), the operation of taking its versor (156) produces no change; and in
manner that, if a given vector a be already an unit-vector ^

remains the same vector, when it is divided (155) by its oivn
length ; that is, in this case, by the number one. For example,
like
it
we have assumed
is
(128, 129), that the axis oi every quaternion

may therefore write, generally, in the no-
an unit-vector ; we
tation of 155, the equation,
U(Ax.|7)
-Ax
5'.
A second operation U leaves thus the result of the^r^^ opera*

U unchanged, whether the subject of such successive ope-
tion
rations be a line, or a quaternion;

*
Compare the Note
When
we have
therefore the
two
to Art. 131.
the zero in this symbol^ UO,
is
considered as denoting a null vector

(2),
the symbol itself denotes generally, by the foregoing principles, an indeterminate

unit-vector; although the direction of this unit- vector may, in certain questions, he-
come determined, as a
limit resulting
from a law.
140
[boOK
II.
following general formulas, differing only in the symbols of

that subject
:
UUa = Ua;
JJ\Jq
= \Jq;
whence, by abstracting (comp. 145) from the subject of the

operation, we may write, briefly and symbolically,
161. Hence, with the help of 145, 158, 159, we easily deduce
the following (among other) transformations of the versor of a quaternion
Kq
'Uqf= Uxq^
We may
Uj
q
if
a;
>
=-
\Jxq, if
^ ^
a;
<
0.
also write, generally,
the parentheses being here unnecessary, because (as will soon be more
fully seen) the st/mbol Vq^ denotes one common versor, whether we
interpret it as denoting the square of the versor, or as the versor of
The present Calculus will be found to abound in

General Transformations of this sort ; Avhich all (or nearly all), like
the foregoing, depend ultimately on very simple geometrical conceptions ; but which, notwithstanding (or rather, perhaps, on account
the square, of q.
of) this
extreme
simplicity of their origin, are often useful,
a,s
elements
new kind
oi Symbolical Language in Geometry: and generally,

as instruments of expression, in all those mathematical or
physical
of a
researches to which the Calculus of Quaternions can be applied.

It
is, however, by no means necessary that a student of the subject,
make himself familiar with all the

recent transformations of U^'; although it may be well that he
should satisfy himself of their correctness, in doing which the fol-
at the present stage, should
lowing remarks will perhaps be found to

(1.)
To give a geometrical
recent equation,
illustration^ -wliich
q:Kq = (Vqy-,
assist.
may
also serve
&.m proof, of the
CHAP.
we may employ
Fig. 36, bis; in which,
OA
Kq
As regards the
(2.)
141
GEOMETRICAL ILLUSTRATIONS.
1.]
Ob'
Ob'
equation,
by 146,
(2.),
\ODJ
=
11(52)
ê have
OA
(JJqy, we have only
to conceive that the
three lines oa, ob, oc, of Fig. 42, are cut (as in Fig. 42, bis) in
new
three
points, a', b', c',
by an
unit-circle (or
a radius equal to the unit of length), which

their
then
is
by a
circle
with
described about
origin o as centre, and in their common plane ;

these three lines be called a, p, y, the three new lines oa',
the three unit- vectors denoted by the symoc'are
for
common
if
(by 155)
ob',
Uo,
bols,
Uy; and we have the transformations
U/3,
(comp. 148,
149),
As
regards other recent transformations (161), although

seen (135) that it is not necessary to invent any new or
peculiar symbol, to represent the reciprocal of a quaternion, yet
of present convenience, and as a merely temporary
if, for the sake
(3.)
we have
notation,
we
for a
moment, the
(comp. 145, 158)
letter
of the operation of taking the reciprocal,
we
as a characteristic of reciprocation, or
shall then
have the symbolical eqtiations
R2 = K2
we have
Fig. 42, bis.
write
employing thus,
bat
A'
RK = KR;
1;
also (by 160),
U2= U
whence
RU = UR = KU=UK;
it
easily follows that
U = RUR = RKU = RUK = KUR = KRU = KUK

= URK = UKR = UKUR = UKRU = (UK)2 = &c.
(4.)
The equation
U ^ = U 2,
a
expresses that the locus of the point

(4.)),
which
tion
because
is
drawn /rom o
it is
On the
is
the indefinite right
line,
or ray (comp. 132,
in the direction ofoiR,* but not in the opposite direc-
equivalent to
.|=i,
(5.)
Up = Vp,
or simply,
other
or
I-
or p
= x(3, x>0.
hand the equation,
U^ = -.U^, orUp=-U/3,
a
a
expresses (comp. 132, (5.)) that the locus of p is the opposite ray from o or that
it is the
because it may be transformed to
indefinite prolongation of the revector bo
;
In 132,
(4.), p. 119,
OA and a ought
to
have been ob and
b.
142
U^ = -l;
(6.) If a,
(3,
or
z.^
= 7r;
or
p= x/3,
a;
[boOK
< 0.
in sub-art 2) the three lines oa, ob,
y denote (as
of Fig. 42, bis), so that (bv 149)
we have
equation,
the equation
II.
y
-
3Y
= -
\a
oc
of Fig. 42 (or
then this other
/^pV^^ya'
a)
expresses ^cweraWy that the locus of p

the whole indefinite right
or that it is
is the system of the two last loci
both ways prolonged, through the two points o and B
line,
(comp. 144, (2.)).

(7.)
bis},
But
if it
happen that the
line y, or oc, like oa' in Fig.
has the direction opposite to that of
a, or of
41 (or in Fig. 41,
oa, so that the last equation takes
the particular form,

r)\2
1,
("Îthen
U-
must be (by 154) a
right versor
and
reciprocally, every right versor, with
a plane containing a, will be (by 153) a value satisfying the equation.

therefore, the Zooms
of the point P
In this
case,
(as in 132, (1.), or in 144, (1.)) the plane

through O, perpendicular to the line OA ; and the recent equation itself, if supposed
by a real* vector
to be satified
equivalent /orwis
p,
is
may
be put under either of these two earlier but
TT
On Vector- Arcs, and Vector-Angles, considered

9.
as Representatives of Vers or s of Quaternions ; and on the
Multiplication and Division of any one such Versor by
another,
Section
162. Since every unit-vector oa (129), drawn from the

in some point a on the surface of what we
origin o, terminates
have called the unit-sphere (128), that term a (1) may be

considered as a Representative Point, of which the position on
that surface determines, and may be said to represent, the
direction
(12,
of the
line
17) by any
which
which
is
is
oa
in space
positive scalar.
or of that line multiplied

then the Quaternion
And
the quotient (112) of any two such unit-vectors, and

in one view a Radial (146), and in another view a
Versor (151), may be said to have the arc of a great circle,

AB, upon the unit sphere, which connects the terms of the two
*
to
Compare 149,
page 90.
(2.)
also the second
Note
to the
same Article
and the Notes
CHAP.
I.]
REPRESENTATIVE AND VECTOR ARCS.
vectors, for its Representative Arc,
arc a
Vector Arc, on account
We may
143
also call this
of its having a definite direcindicated (for example) by a
tion (comp. Art. 1), such as is

curved arrow in Fig. 39 and as being thus contrasted with
its own opposite, or with what may be called by analogy the
Revector Arc ba (comp. again 1) this latter arc represent;
at once the reciprocal (134),
on the present plan,
ing,
conjugate (137), of the former versor; because
it
and the
represents
the corresponding Reversor (158).

163. This mode o^ representation^ of versors of quaternions
vector
arcs, would obviously be very imperfect, unless
by
We
shall therefore
equals were to be represented by equals.
that
a
as
it
to
vector
is
natural
otherwise
arc, ab,
do,
define,
upon the unit sphere, is equal to every other vector arc cd
which can be derived from
it, by simply causing (or conceivown great circle, without any change of
In fact, the two isosceles and
length, or reversal of direction.
plane triangles aob, cod, which have the origin o for their
common vector, and rest upon the chords of these two arcs as
bases, are thus complanar, similar, and similarly turned so
ing)
it
to slide* in
its
that (by 117, 118)
we may
A aob
qc
here write,
cod,
OB CD
=
OA oc
the condition of the equality of the quotients (that is, here, of

the versors), represented by the two arcs, being thus satisfied.
We
arcs,
sometimes denote this sort of equality of two vector

ab and cd, by the formula,
shall
ab = ^ CD;
and then
it is clear
(comp. 125, and the earArt. 3) that we shall also have, by what
may be called inversion and alternation,
lier
these two other formulae of arcual equality,
'^BA=^DC;
ÂC = -BD.
/
o^-----
-'A.
Fig. 35, if..
(Compare the annexed Figure 35, bis^

* Some aid
34
to the conception
may
here be derived from the inspection of Fig

on the surface of one com-
in -which two equal angles are

supposed to be traced
144
[bOOK
II.
164. Conversely, unequal versors ought to be represented

(on the present plan) by unequal vector arcs ; and accordingly,
we purpose to regard any two such arcs, as being, for the present purpose, unequal (comp. 2), even when they agree in
quantity, or contain the same number of degrees^ provided that
differ in direction
they
which
principal ways, as follows.
a versor, ob
ba
two
and so may represent (162)
oa, and the corresponding reversor, oa: ob, reOr, Ilnd, the two arcs may belong to different
ab and bc in Fig. 43
radial quotients
two
they represent
great
in either of
For, 1st, they may be opposite

for example, a vector arc ab, and
arcs of one great circle; as,

the corresponding revector arc
spectively.
may happen
circles, like
in
which
latter case,
(146) in different planes ; or (comp.

119) two diplanar versors, ob oa,
:
and 00 OB
:
but
it
has been shown
generally (122), that diplanar quaternions are always unequal: we

consider therefore, here again the
arcs,
ab and
bc, themselves, to be
been said) unequal vectoi^s.

165. In this manner, then, we may be led
(comp. 122) to
the
a
or
of
i\\Q
conception of plane,
regard
position of a great
circle on the unit sphere, as entering,
essentially, in general,*
(as has
into the conception ofa vector-arc, considered as the

representative of a versor (162).
But even without expressly referring
to versors, we may see that if, in Fig. 43, we
suppose that b
the middle point of an arc aa' of a great circle, so that in a

recent notation (163) we may establish the arcual
is
equation,
-^
we ought
ba',
then (comp. 105) not to write also,

^
mon
AB = ^
AB =
'^
BC;
Or
the four lines oa, ob, oc, od, of Fig, 35,

may now be conceived to
be equalli/ long; or to be cut by a circle with o for centre, as in the modification of
that Figure, which is given in Article 163, a little lower down.
desk.
We
say, in general; for
it
will soon be seen that there is
great semicircles^ considered as vector arcs,
may
a sense in which all

be said to be equal to each other.
CHAT.
145
ARCUAL EQUATIONS, CO-ARCUALITY.
I.]
because the two co'initial arcs, ba' and bc, which terminate
to be, as vector-arcs^
differently, must be coiiLudered (comp. 2)
On the other hand, if we should refuse to admit (as

unequal.
in 163) that any two complanar arcs, if equally long, and simiab and cd in the recent
could not usefully speak of
Fig. 35, his, are equal vectors, we
as existing under any circumequality between vector-arcs
like
larly (not oppositely) directed,
stances.
We are then thus led again
to include, generally, the
circle as distinguished
conception of a plane, or of one great
from another, as an element in the conception of a Vector-Arc,
And hence
an equation between two such arcs must in general

be conceived to include two relations of co-arcuality. For
^ ab =
cd, of Art. 163, includes geneexample, the equation
of
its
signification, the assertion (comp. 123)
rally, as apart
'^
that the four points a, b, c, d belong to or\Q common great ciror that each of the two points, c and D,
cle of the unit-sphere
;
is
co-arcual with the two other points, a and b.
166. There is, however, a remarkable case of exception, in which

two vector arcs may be said to be equal, although situated in diffe-
when they
rent planes: namely,
are both great semicircles.
In
fact,
upon the present plan, every great semicircle, aa', considered as a

vector arc, represents an inversor (153); or it represents negative
unity (oa'
oA = -a
a=-
1),
considered as one limit of a versor;
but we have seen (159) that such a versor has in general an indeterminate plane.
Accordingly, whereas the initial and final points, or
(comp. 1) the
origiri
A and the term
b, of
a vector arc ab, are in ge-
neral sufficient to determine the plane of that arc, considered as the
most direct path (comp. 112, 128) from the one point
on the sphere; in the particular caê when one of the
shortest or the
to the other
two given points
is diametrically opposite to the other, as a' to A,

the direction of this path becomes, on the contrary, indeterminate.
If then we only attend to the effect produced, in the way of change
of position of a point, by a conceived vection {or motion) upon the

we are permitted to say that all great semicircles are equal
vector arcs; each serving simply, in the present view, to transport a
sphere^
point from one position to the opposite; and thereby to reverse (like
the factor - 1, of which it is here the representative) the direction of
the radius which is drawn to that point of the unit sphere.
146
(1.)
The
equation,
rk
aa'
=n
[bOOK
II.
bb',
and b' to b, satisfies evidently

the general conditions of co-arcuality (165); because the /ottr points aba'b' are all
on one great circle. It is evident that the same arcual equation admits (as in 163)
in
which
it is
here supposed that
of inversion and alternation
n
(2.)
effect
We may
a'a
a' is opposite to a,
so that
r\
n AB =
and
b'b,
'-^
a'b'.
also say (comp. 2) that all null arcs are equal, as producing
on the position of a point upon the sphere
n AA =
r>
bb
and thus may write
no
generally,
= 0,
with the alternate equation, or identity, <^ ab = r> ab.

on the present plan, of the
(3.) Every such null vector arc AA is a representative,
other unit scalar, namely positive unity, considered as another limit of a versor (153) ;
its plane is again indeterminate (159), unless some law be given, according to
which the arcual vection may be conceived to begin, from a given point A, to an in-
and
definitely
near point B upon the sphere.
The
167.
theory,
is
principal use of Vector Arcs^ in the present

to assist in representing ^ and (so to speak) in con-
by means of a Spherical Triangle^ the Multiplication and Division of any two Diplanar Versor s (comp. 119,
In fact, any two such versors of quaternions (156),
164).
structing^
considered as radial quotients (152), can easily be reduced (by

the general process of Art. 120) to the forms,
5-- j3 :a
= OB
OA,
5^'
7 j3= oc
:
ob,
where a, b, c are corners of such a triangle on the unit sphere;

and then (by 107), the former quotient multiplied by the latter will give for product
^'.5'
'y
a= OC
OA.
If then (on the plan of Art. 1) any two successive arcs, as ab

and Bc in Fig. 43, be called (in relation to each other) vector
W[id provector ; while that third arc ac, which is drawn from
the initial point of the first to the final point of the second,
be called (on the same plan) the transvector : we may now

in the multiplication of any one versor (of a quaterthat
say
nion) by any other, if the multiplicand* q be represented {IQ2)
by a vector-arc ab, and if the multiplier q be in like manner
shall
*
Here, as in 107, and elsewhere, we write the symbol of the multiplier towards
the left-hand, and that of the multiplicand towards the right.
CHAP.
I.]
CONSTRUCTION OF MULTIPLICATION OF VEIISORS. 147
represented by
tion
is
?i
mode
provector-arc bc, which
of representa-
always possible, by what has been already shown, then
the product q. q, or q'q, is represented, at the same time,

the transvector-arc ac corresponding.
One
168.
by
of the most remarkable consequences of this con-
of the multiplication ofversors is the following that

the value of the product of two diplanar versors (164) depends
upon the order of the factors ; or that c[q and qq are unequal,
struction
For let aa' and cc' be

unless q be complanar (119) with q.
in
of
two
arcs
great circles,
different planes, bisecting each
any
other in the point b, as Fig. 43 is designed to suggest; so
that we have the two arcual equations (163),
^
AB =
and
ba',
bc =
'^
c'b
then one or other of the two following alternatives will hold

Either, 1st, the two mutually bisecting arcs will both
good.
be semicircles, in which case the two new arcs, ac and
c'a',
will
indeed both belong to one great circle, namely to that of which

B is a pole, but will have opposite directions therein because,
in this case, a' and c' will be diametrically opposite to a and c,
;
and therefore (by 166,
(1.)
^
the equation
AC =
'^
a'c',
but not the equation

'>
will
be
satisfied.
AC =
'^
c'a',
Or, Ilnd, the arcs aa and
cc',
which are
supposed to bisect each other in b, will not both be semicircles,

even if one of them happen to be such ; and in this case, the
arcs AC, c'a' will belong to two distinct great circles, so that
they
will be diplanar, and therefore unequal, when considered as
vectors.
(Compare the
1st
and Ilnd cases of Art. 164.)
In
each case, therefore, ac and c'a' are unequal vector arcs; but the
former has been seen (167) to represe?it the product q'q; and
the latter represents, in like manner, the other product, qq', of
the same two versors taken in the opposite order, because it is
the Jiew transvector arc,
cated by
when
and ba' (= ab)
c'b
(=bc)
is
treated as the
new
as the neiv provector arc, as is indithe curved arrows in Fig. 43.

The two products^
vector arc,
148
qq and
'</,
[BOOK
II.
are therefore themselves unequal, as above asserted,
under the supposed condition of diplanarity,

169. On the other hand, when the two
factors, q and q\
are complanar versors^ it is easy to prove, in several different
ways, that their products, q'q and qq\ are equal, as in algebra.
Thus we may conceive

turn round
that the arc cc', in Fig. 43,
is
made
to
middle point b, until the spherical angle cb a'

vanishes; and then the two neiv transvector-arcs, ac and cV,
will evidently become not only complanar but equal, in the
its
sense of Art. 163, as being
still equally
long, and being now
Or, in Fig. 35, bis, of the last cited Articonceive a point e, bisecting the arc bc, and there-
similarly directed.
cle,
we may
which
fore also the arc ad,
is
commedial therewith (comp.
Art. 2, and the second Figure 3 of that Article) ; and then,

if we represent the one versor q by either of the two
equal
arcs, AE, ED, we may at the same time represent the other
so that
q' by either of the two other equal arcs, ec, be
the one product, q'q, will be represented by the arc ac, and
the other product, qq', by the equal arc bd.
Or, without re-
versor
ference to vector arcs,
we may suppose
that the
two factors
are,
5^
=
=j3: a OB
OA,
-'
=7
a= oc
OA,
any three complanar and equally long right

Us) for thus we have only to determine a fourth line, S or od, of the same length, and in the same
=
plane, which shall satisfy the equation S:'y j3:a (117), and
therefore also (by 125) the alternate equation, ^:j3 = 7:a;
OA, OB, 00 being
lines (see again Fig. 35,
and
it
will then immediately follow* (by 107), that
i3
We may therefore infer,
for any two versor s of quaternions, q

and q\ the two following reciprocal relations
:
It is evident that, in this last process of reasoning, we make no use of the supposed equality of lengths of the four lines compared so that we might prove, in exactly the same way, that qqqq '\iq' \\\q (123), without assuming that these two
;
complanar factors, or quaternions, q and
q',
are versors.
CHAP.
I.]
MULTIPLICATION OF RIGHT VERSORS.
l,.,qq^qq\
ll...l^q'q=qq\
if
149
/||k(123);
then
q'\\\q{\Q^)',
in the product)
convertibility offactors (as regards their places
2^
oi
a
and
thus
at
once
complanarity.
proof
being
consequence
170. In the 1st case of Art. 168, \kvQ factors q and q* are both
rigU versors (153) and because we have seen that then their two
products^ q'q and qcf are versors represented hy Qq^xoMy long but oppositely directed arcs of one great circle, as in the 1st case of 164, it
;
follows (comp. 162), that these
two products are
at once reciprocal
(134), and conjugate (137), to each other; or that they are related
as versor and reversor ( 58).
We may therefore write, generally,
1
I.
and
qq=Kq'q^
11. ..
qq'=-j--,
if q and q be any two right versors; because the

multiplication of
any two such versors, in two opposite orders, may always be represented or constructed by a Figure such as that lately numbered
43, in which the bisecting arcs aa' and cc' are semicircles. The Ilnd
formula may also be thus written (comp. 135, 154):

III. .. if
and under
cause
it
this
2'2
= -l, and q'^=~\^ then
form
it
q'q
.qq-+
\\
evidently agrees with ordinary algebra, be-
expresses that, under the supposed conditions^
q'q.qq'^4Kf',
but
be found that this
last equation is not an identity., in the

of
general theory
quaternions,
171. If the two bisecting semicircles cross each other at right
it
angles.,
will
the conjugate products are represented by two quadrants^

one great circle. It follows that if two right
oppositely turned^ of
two mutually rectangular planes, he multiplied together in two

opposite orders, the two resulting products will be two ojjposite right
versors, in a third plane, rectangular to the two former; or in symbols,
versors, in
that
if 5^
=-
1,
q^=^-l, and Ax. q
-u
Ax.
q,
then
(?'?)^=(^??=-i,
q'q
= -q^\
and
Ax.
In
q^q J~
this case, therefore,
Ax.
q,
Ax.
q'q
J-
Ax.
we have what would be
namely the equation,
{qqy = -g'\q\
g'.
in algebra a paradox,
150
if
q and
q'
[bOOK
II.
be any two right versors, in two rectangular planes ; but
we
see that this result
is
not more paradoxical, in appearance, than the
equation
which
under
exists,
the
same
And when we come
conditions.
to ex-
amine what,
in the last analysis, may be said to be the meaning of this

last equation, we find it to be simply this : that any two quadrantal or
right rotations, in planes perpendicular to each other,
a third
selves into
right rotation, as their resultant, in
compound them-
a plane perpendi-
to each of them : and that this third or resultant rotation has

one or other of two opposite directions, according to the order in which
cular
the
tivo
component rotations are taken, so that one shall be successive
to the other.
We
172.
propose to return, in the next Section, to the
consideration of such a System of Right Versors, as that which
we have here briefly touched upon but desire at present to
remark (comp. 167) that a spherical triangle ABcm?ij ser veto
:
construct,
by means of represeiitative arcs (162), not only the
multiplication, but also the division, of any one of two diplanar

versors (or radial quotients) by the other.
In fact, we have
only to conceive (comp. Fig. 43) that the vector arc ab represents a given divisor, say 5', or j3 a, and that the transvector
arc AC (167) represents a given dividend, suppose 7", or y a;
:
for then the provector arc
on the same
q"
'.q,
or 7
bc (comp. again 167)
plan, the quotient
of these two
will represent,
versors,
(106), or the versor lately called q
/3
have generally, by 106, 107, 120,

bra, the two identities
= q"\
namely
since
we
for quaternions, as in alge-
qq'q q''
{q"-.q).q
173. It is however to be observed that, for reasons already assigned, we must not employ, for diplanar versors, such an equation
as q. {q": q)
sors, the
fact
q^'
because
ordinary algebraic
by 169,
we may now
I.
11.
we have found (168)
that, for such ver-
identity, qq'^q'q-, ceases to he true.
establish the
if
q{q'''q)=q".
iiq{q'''.q)
In
two converse formulae:
= q", then
$'M!I?(123);
q"
\\\
q.
Accordingly, in Fig. 43, if q, q', q" be still represented by the

arcs AB, BC, AC, the product q {q" q), or qq', is not represented by
:
CHAP.
REPRESENTATIVE AND VECTOR ANGLES.
I.]
151
but by the different arc c'a' (168), which as a vector arc has been
seen to be unequal thereto: although it is true that these two last
and therefore subtend
arcs, AC and c'a', are always equally long,
AC,
of the unit sphere; so that we may write,

equal angles at the centre o
two quaternions)*
versors
two
for
(or indeed for any
any
generally,
q and q" , the formula,
Lq{q'':q)==Lq".
174. Another mode of Representation of Versors, or rather two
such new modes, although intimately connected with each other,
be briefly noticed here.

We may consider the angle aob, at the centre o of the unitconceived to have not only a definite quantity, but also
when
sphere,
may
1st.
a determined pZane (110), and a given direction therein (as indicated

arrow in Fig. 33),
by one of the curved arrows in Fig. .39, or by the
as being
what may be
called
by analogy
a Vector- Angle
and may
the Representative Angle of, the

of the unit- sphere.
are
radii
ob
Vei'sor OB oa, where oa,
this
rectilinear
Ilnd. Or we
angle aob at the centre,
replace
say that
it represents,
or that
it is
may
by the equal Spherical Angle ac'b,
at
what may be
called the Positive Pole of the representative arc
^î^
ab;
and o^b are quadrants; and the rotation,

from the first of these two quadrants
to the second (as seen from a point outside the
sphere), has the direction which has been selected
so that c'a
at this pole c^
(111, 127) for the positive one, as indicated in the

annexed Figure 44: and then we may consider this
spherical angle as a
or ob
new Angular
W
C'
Fig. 44.
Representative of the same versor
q,
OA, as before.
now that after employing Q,frst spherical trianto

construct
gle ABC,
(as in 167) the multiplication of any one given
versor q, by any other given versor q', we form a second or polar
175. Conceive
which the corners a', b', c' shall be respectively (in the
sense just stated) the positive poles of the three successive sides, bc,
CA, AB, of the former triangle ; and that then we pass to a third triangle a'b'^c', as part of the same lune b'b'' with the second, by tak-
triangle, of
ing for b" the point diametrically opposite to
b';
so that b'' shall
be
It will soon be seen that several of the formulje of the

present Section, respecting the multiplication and division of versors, considered as radial quotients (151),
require little or no modification, in the passage to the corresponding operations on
quaternions, considered as general quotients of vectors (112).
ELEMENTS OP QUATERNIONS.
152
[book
II.
the negative pole of the arc ca, or the positive pole of what was lately
called (167) the transvector-arc ac: also let
c^' be,
in like
manner, the point opposite
to c' on the unit sphere.
Then we may not
only write (comp. 129),
Ax
q = oc'.
but shall
also
Ax ^ = oa',
Ax
q'q
= ob'^
have the equations,
Z^ = b''c'a^ Z^'=c'aV,
Lq'q =
Q."v>"A'\
namely the two

and a\ and the external
these three spherical angles^
base-angles at c^
angle at b"^ of the new or third

triangle a'b'V/, will therefore represent^ respectively, on the plan of 174, II., the mul-
vertical
Fig. 45.
and the pro(Compare the annexed Figure 45.)
tiplicand^ q, the multiplier^ q\
duct, q'q.
Without expressly referring to the former triangle abc,

connect this last construction of multiplication ofversors (175)
176.
we can
with the general formula (107), as follows.

Let a and /3 be now conceived to be two unit-tangents*
sphere at
the two
to the
perpendicular respectively to
arcs c
and c'a^ and drawn toc',
wards the same sides of those arcs as the

points a' and b' respectively and let two
other unit-tangents, equal to these, and
denoted by the same letters, be drawn (as
;
in the annexed Figure 45, his) at the points

b'^ and A^ so as to be normal there to the
same arcs c'b'' and
c'a',
and to
fall
towards
the same sides of them as before. Let also
two other unit- tangents, equal

other, and each denoted
by
7,
to each b'
Fig. 45,
be drawn at
bis.
the two last points b" and a', so as to be both perpendicular to the
Then
a'b''', and to fall towards the same side of it as the point c'.
(comp. 174,11) the two quotients, /3 a and 7 /3, will be equal to the
arc
two versors, q and q\ which were lately represented

(in Fig. 45) by the
*
By an
unit tangent is here meant

simply an unit line {or unit vector, 129) so
as to be tangential to the unit-sphere, and to have its
origin, ox its initial
point (1), on the surface of that sphere, and not (as we have usually supposed) at
the centre thereof.
drawn
PRODUCT ON ORDER OF FACTORS. 153
CHAP.T.] r3EPENDENCE OF
two base angles, at
c'
and
a',
of the spherical triangle a'b'^c' ; the pro-
two versors, is therefore (by 107) equal to the third

a
and consequently it is represented, as before, by the
quotient, 7
external vertical angle c'^b^'a' of the same triangle, which is evidently
equal in quantity to the an^le of this third quotient, and has the same
duct^ q^q, of these
:
oW, and the same direction of rotation,

may assist to show.
axis
his,
177. In each of the
two
last Figures,
as the
arrows in Fig. 45,
the internal vertical angle
thus equal to the Supplement, ir - l q'q, of the angle of the

product; and it is important to observe that the corresponding roat a"
is
tation at the vertex b'^,
the side b'â' to the side b'^c', or (as
from
ôm
may
we
the point a' to the point c', is positive; a
briefly express it)

result which is easily seen to be a general one,
the foregoing Article.*
by the reasoning of
then infer, generally, that when
We may
of any two versors is constructed by a spherical trianof which the two base angles represent (as in the two last Articles)
ih.Q factors, while the external vertical angle represents i\iQ product,
then the rotation round the axis (ob") of that product q'q, from the
the multiplication
gle,
axis (oa') of the multiplier q', to the axis (oc') of the multiplicand q, is
whence it follows that the rotation round the axis Ax. q'
positive:
of the multiplier,
from the axis Ax. q of the multiplicand, to the
Ax.
axis
same
Or, to express the

q'q of the product, is also positive.
rotations
hitherto
since
the
considered
more
only
thing
fully,
have hQQM plane ones (as in 128, &c.), we may say that if the two
latter axes be projected on a plane perpendicular to the former, so as
to
still
from
have a common origin
the projection of
Ax. q
o,
then the rotation round Ax.
to the projection of
Ax.
q'q, will
5',
be di-
rected (with our conventions) towards the right hand.
178.
We have therefore thus 2(,new mode of geometrically
exhibiting the inequality of the two products^ q'q and^g-', of two

diplanar versors (168), when taken as factors in two different
orders.
For
this purpose, let
Ax.g=0P, Ax.5'=OQ, Ax.qq = OR\

and prolong to some point s the arc pr of a great circle on the
unit spliere.
Then, for the spherical triangle pqr, by prin* If a
person be supposed to stand on the sphere at b", and to look towards the
it would appear to him to have a right-handed direction, which is the one
arc a'o',
here adopted as positive (127).
154
[book
II.
we shall have (comp. 175) the followof

internal
base angles at p and q, and of
the
two
values
ing
the external vertical angle at ii
ciples lately established,
PQR = Lq
RPQ = Lq\
SRQ = Lqq
and the rotation at q, from the side qp to the side qr

Let fall an arcual perpendiright-handed.
cular, RT, from the vertex r on the base pq,
and prolong this perpendicular
a manner as to have
/^
RT =
also prolong pr' to

'
'^
tr'
to
r',
will
be
in such
some point
s'.
We shall
new
triangle pqr', which will

be a sort of reflexion (comp. 138) of the old
one with respect to their common base pq
and this new triangle will serve to construct
then have a
Fig. 46.
the new product^ qq. For the rotation at p
from PQ to pr' will be right-handed,

we shall have the equations,
QPR' = Zg;
r'qp = Z^';
as
it
ought to be
qrV = z$'^';
and
or' = Ax.^^';
so that the new external and spherical angle^ qr's', will represent the neio versor, qg', as the old angle srq represented the
old versor, q'q, obtained from a different order of the factors.
And
although, no doubt, these
always
equal in quantity, so that
the general formula,
tivo angles,
at
r and
are
r',
we may establish (comp.
73)
Lqq = Lqq,
yet as vector angles (174), and therefore as representatives of

versor s, they must be considered to be unequal: because they
the tangent planes to the sphere
different planes, namely,
at the two vertices R and r'; or the two planes respectively
the centre o.
parallel to these, which are drawn through
have
179. Division of Versors (comp. 172) can be constructed by

means oi Representative Angles (174), as well as by representative arcs
Thus to divide q" by q, or rather to represent such division
(162).
similar to that last employed for
geometrically^ on a plan entirely
CHAP.
CONICAL ROTATION OF AXIS OF VERSOR.
I.]
multiplication,
in Fig. 46,
we have only
determine the two points P and
r,
by the two conditions,

OP = Ax.
and then
to
155
to find a third point
RPCl
=Lq^
0R = Ax.<^'^
5-,
q by the two angular equations,

QRP = TT - Z
2'^
the rotation round p from PR towards pci being positive; after which
we
shall have,
Ax.
(^'^
^)
= oq;
(^'^
^)
pcir.
which connects
(1.) Instead of conceiving, in Fig. 46, that the dotted line rtr',
the vertices of the two triangles, with
pq
common base (178),

base, we may imagine it
for their
a great circle, perpendicularly bisected by that

of a small circle, described with the point P for
is
an arc of
an arc
to be
its positive pole (comp. 174, II.).

passage (comp. \lo) from the versor 9", or q'q, to
the unequal versor q (5" 9), or qq\ is geometrically performed by a Conical Rotation of the Axis Ax. 5", round the axis Ax. q, through an angle = 2 Z. ^, without
And
then
we may say
that
i\iQ
{(quantitative') change of the angle Lq"

formula (comp. again 173),
L q (9" :q)
any
so that
',
we
=L
we have,
as before, the general
q".
employ the construction of multiplication and division by

(2.)
representative arcs, which Fig. 43 was designed to illustrate, and conceive that a
ncAV point c" is determined in that Figure by the condition ^ a'c" = ^ c'a', we may
Or
if
prefer to
then say that in the passage from the versor 9", which
is
represented
by AC,
to the
versor 9 (9" 9), represented by c'a' or by a'c'', the representative arc of 9" is made
to move, without change of length, so as to preserve a constant inclination * to the
:
representative arc
in passing
AB ofq, while tfs
from A to
initial point describes the double
of that arc Ab,
a'.
(3.) It maybe seen, by these few Examples, that if, even independently of some
new characteristics of operation, such as K and U, new combinations of old symbols,
such as 9(9": 9), occur in the present Calculus, which are not wanted in Algebra,
they admit for the most part of geometrical interpretations, of an easy and interesting kind and in fact represent conceptions, which cannot well be dispensed with,
;
and which
it is
useful to be able to express, with so
(Compare the remarks
in Art.
161
much
and the sub-articles
simplicity
and
conciseness.
to 132, 145.)
180. In connexion with the construction indicated
two Figures 45,
it
by the
here remarked, that if abc be any

a', b', c' be (as in 175) the positive
may be
spherical triangle, and if

poles of its three successive sides, bc, ca, ab, then the rotation
(comp. 177, 179) round a! from b' to c', or that round b' from
* In a manner
analogous to the motion of the equator on the ecliptic, by luni-
Vw
precession, in astronomy.
ELExMENTS OF QUATERNIONS.
156
[bOOK
II.
to a', &g., is positive.

The easiest way, perhaps, of seeing
the truth of this assertion, is to conceive that if the rotation
round a from b to c be not already positive, we make it such,
c'
by passing to the diametrically opposite triangle on the sphere,

which will not change the poles a', b', c'. Assuming then that
these poles are thus the near ones to the corresponding corners
of the given triangle, we arrive without any difficulty at the
conclusion stated above which has been virtually employed

in our construction of multiplication (and division) of versors,
:
by means of Representative Angles
(1 75,
176)
and which may
be otherwise justified (as before), by the consideration of the

unit-tangents of Fig. 45,
bis.
a, 13, y be ant/ three given unit vectors, such that the rotation
from the second to the third, is positive (in the sense of Art. 177)
(1.) Let then
round the
and
let a',
first,
(3',
a'
y' be three other unit vectors, derived
= Ax.
/3'=Ax.(a:y),
(y:/3),
then the rotation round
from
a',
j8'
to
y',
from these by the equations,

y'
= Ak.Q3
will be positive also
a)
and we
shall
have
the converse formulae,
a = Ax.(y':/3'),
/3
= Ax.
round a from
(a': y'),
= Ax
(/3'
a').
y were given to be negative, a', j8', y'

being still deduced from those three vectors by the same three equations as before,
then the signs of a, (3, y would all require to be changed, in the three last (or reci(2.) If the rotation
j3
to
procal^ formulae ; but the rotatitm round a', from /3' to y', would still be positive.
(3.) Before closing this Section, it may be briefly noticed, that it is sometimes
convenient, from motives of analogy (comp. Art. 5), to speak of the Transvector-
Arc (167), which has been
Arcual Sum
of the
seen to represent
two successive
the factors; Provector being
still
o,
vector-arcs,
said to be
product of two versors, as being the

which represent (on the same plan)
added
to Vector
but the Order of such
Addition of Diplanar Arcs being not now indifferent (168), as the corresponding
order had been early found (in 7) to be, when the vectors to be added were right
lines.
(4.)
We may
also
speak occasionally, by an extension of the same analogy, of
the External Vertical Angle of a spherical triangle, as being the Spherical Sum of
the two Base Angles of that triangle, taken in a suitable order of summation (comp.
Fig. 46); the Angle which represents (174) the Multiplier being then said to be
added (as a
sort of
Angular Provector^ to that other Vector-Angle which represents

what is here called the sum of these two angles (and is,
the Multiplicand; whilst
with respect to them, a species of Transvector- Angle) represents, as has been proved.
the Product.
(5.) This conception of angular transvection
when (on the plan
of 174, 1.)
we assume the
becomes perhaps a little more clear,

o as the common vertex of three
centre
angles aob, boc, aoc, situated generally in three different planes.
For then we may
CHAP.
SYSTEM OF THREE RIG^T VERSORS.
I.]
157
conceive a revolving radius to be either carried by two successive angular motions,
from OA
to OB,
and thence
to
oc
or to be transported immediately,
by one such
motion, from the^rs^ to the third position.

(6.) Finally, as regards the construction indicated
by Fig. 45,
bis, in
which tan-
be well to remark distinctly hero, that

a'b"c', in that Figure, may be ang given spherical triangle, for which the rotation
round b" from a' to c' is positive (177) and that then, if the two factors, g and q',
gents instead of radii were employed,
may
it
be defined to be the two versors, of which the internal angles at c' and a' are (in the
sense of 174, II.) the representatives, the reasonings of Art. 176 will prove, without
necessarily referring, even in thought, to any other triangle (such as abc), that the
external angle at b" is (in the same sense) the representative of the product, q'q, as
before.
Section
10.
On a System of Three B.ight Versors, in
Rectangular Planes; and on

I
the
Laws of
the
Three
Symbols,
k.
J,
ok
are any three given and coinitial but rectangular unit-lines, the rotation round the first
from the second to the third being positive; and let oi', oj,
181. Suppose that oi, oj,
ok' be the three unit-vectors respectively opposite to these, so

that
= -or,
ok'=-ok.
oj'^-oj,
Let the three new symbols i,j, k denote a system (comp. 172)
oi'
of three right versors, in three mutually reHangular planes,

with the three given lines for their respective axes ; so that
Ax.^ = OI, Ax./ = oj, Ax.^ = 0K,
and
2
as
= ok:oj,
J=oi:ok,
Figure 47
may
A=oj:oi,
We shall then have these other expressions for the

I
same three versors
=
K'
Fig. 47.
while the three respectively opposite versors

pressed
may
be thus ex-
= OJ
=
-J OK
- ^ = 01
z
OK =
01
OJ = oj'
ok'
oi'
oj = oj' ok' = ok
ok = ok' oi' = oi
oi = oi'
oj' = oj
:
oj'
ok' ;
oi'.
element;s of quaternions.
158
[book
ii.
And
from the comparison of these different expressions seveimportant symbolical consequences follow, which it will be
worth while to enunciate separately here, although some of
ral
them
are virtually included in the results of former Sections.

182. In the j^r5^ place, since
i^
= (o j' ok) (OK oj) =

.
oj'
we deduce (comp. 148) the following

squares of the new symbols
L..^^ = -l;
/ = -l;
oj, &c.,
equal values for the
as
^^
= -1;
might indeed have been at once inferred (154), from the
circumstance that the three radial quotients (146), denoted here
by
z,
versors
J, k, are all right
(181).
In the second place, since

= (oj ok') (ok' oi) = oj
ij
.
we have
oi, &c.,
the following values for the products of the
same three
symbols, or versors, when taken Uoo hy two, and in a certain

order of succession (comp. 168, 171)
:
II.
But
jkî;
ij=k)
ki=j.
in the third place (comp. again 171), since
we have
(01
ok)
(ok
oj) = oi
oj, &c.,
these other and contrasted formulae, for the binary
products of the same three right versors,

tors with an opposite order :
III.
,ji==-k;
kj
when taken
= -i;
ik
as fac-
= -j.
Hence, while the square of each of the three right versors, denoted by these three neiv symbols, ijk, is equal (154) to negative unity,
the product of any
tioo
of them
is
equal either to the third itself, or to the opposite (171) of that third versor, according as
the multiplier precedes or follows the multiplicand, in the cyclical succession,
j\
k,
i?
which the annexed Figure 47,

remembering.
/^
/
\l
Fig. 47, his.
his,
may give some
help towards
CHAP.
LAWS OF THE SYMBOLS,
I.]
(L) To connect such

arcs (162),
multiplications of
j,
159
J, K.
k with the theory of representative

may regard any one of the four
and of representative angles (174), we
quadrantal arcs, jk,

presenting the versor
opposite to
I,
k'j, in Fig. 47, or any one of the four spherical right

which those arcs subtend at their common pole i, as re-
k.i', j'k',
angles, jik, kij', j'ik', k'ij,
i'
I,
which
The squaring
and similarly for^ and
is
A,
with the introduction of the point
to be conceived as being at the
back of the Figure.
= -l,
comes thus to be geometrically

constructed by tbe doubling (comp. Arts. 148, 154, and Figs. 41, 42) of an arc, or of
an angle. Thus, we may conceive the quadrant kj' to be added to the equal arc jk,
(2.)
their
sum
of
i,
or the equation i^
being the great semicircle sj' which (by 166) represents an inuersor (153),
a.
Or we may add the right angle kij' to the
factor.
,
or negative unity considered as
equal angle jiK, and so obtain a rotation through two right angles at the pole
I,
or
o; which rotation is equivalent (comp. 154, 174) to an inversion of

direction, or to a passage from the radius OJ, to the opposite radius oj'.
at the centre
= k, may in like manner

(3.) The midtiplication of j ly i, or the equation ij
be arcually constructed, by the addition of k'j, as a provector-arc (167), to ik' as
a vector-arc (162), giving ij, which is a representative of A, as the transvector-arc,
Or the same multiplication may he angularly conor arcual'Sum (180, (3.)).
with the help of the spherical triangle ijK ; in which the base-angles at I
structed,
and J represent respectively the multiplier, i, and the multiplicand, j, the rotation
round i from j to K being positive : while their spherical sum (180, (4.)), or the ex~
ternal vertical angle at
(comp. 175, 176), represents the same product,
k,
as
before.
i
(4.) The contrasted multiplication of hyj, or of^ into* i, may in like manner
be constructed, or geometrically represented, either by the addition of the arc ki, as
a new provector, to the arc jK as a new vector, which new process gives Ji (instead
of ij) as the new transvector ; or with the aid of the new triangle ijk' (comp. Figs.
46, 47), in which the rotation round i from j to the new vertex k' is negative, so
that the angle at i represents now the multiplicand, and the resulting angle at the
neio pole k' represents the
new and opposite product,
ji
= k.
we have thus ji = - ij (as we had qq = - qq in

see
that
the laws of combination of the neio symbols,
171),
in
not
are
all
ifj, k,
respects the same as the corresponding
183. Since
we
laws in algebra; since the Commutative Property of Multiplication, or the convertibility (169) of the places of the^côr^
without change of value of the product^ does not here hold

good: which arises (168) from the circumstance, that the
factors to be combined are here diplanar versors (181).
It is
therefore important to observe; that there
is
a respect in which
*
multiplicand is said to be multiplied by the multij)lier while, on the other
hand, a multiplier is said to be multiplied into the multiplicand a distinction of this
sort between the two factors being necessary, as we have seen, for
quaternions,
;
although
it is
not needed for algebra.
160
[bOOK
II.
the laws of i, J, k agree with usual and algebraic laws
:
namely,
Property of Multiplication; or in the pro-
in the Associative
perty that the
mula (comp.
new symbols always obey
the associative for-
9),
whichever of them
may be substituted for t, for k, and for X

which equality of values we may omit the point, in
any such symbol of a ternary product (whether of equal or of
unequal factors), and write it simply as lkX. In particular
wê have thus,
;
in virtue of
i.jk
2.2
2*2
= _
ij
.k
= k.k = k^ = -\
or briefly,
ifk
We may,
Formula
therefore,
by
= -\.
182, establish the following important
P=f = k' = ijk^-l

to which
we
shall occasionally refer, as to
(A)
" Formula
A," and
which we
shall find to contain (virtually) all the laws

of the
to
a
and
therefore
be
symbols ijk,
sufficient symbolical basis
for the
whole Calculus of Quaternions
shown that every quaternion can

mial Form,
:*
because
it
wîll
be
be reduced to the Qiiadrino-
q=w + ix ^-jy +
kz,
where w, x, y, z compose a system offour

same three right versors as above.
scalars, while
i,
j, k
are the
(1.)
direct proof of the equation, ijk
of the symbols in Art. 181.

tions
were seen
In
=-
1,
may be
we have only
fact,
to
derived from the definitions
remember that those
defini-
to give,
This formula (A) was accordingly made the basis of that Calculus in the first
communication on the subject, by the present writer, to the Royal Irish Academy in
1843
and the
letters,
i,
j, A,
continued to be, for some time, the only peculiar sym-
bols of the calculus in question.
But
it
was gradually found to be useful to incorK and U, &c.), for representing
porate with these a few other notations (such as
Operations on Quaternions.
It
was
also thought to be instructive to establish the
principles of that Calculus, on a more geometrical (or less exclusive)}' symbolical)

foundation than at first which was accordingly afterwards done, in the volume en;
Lectures on Quaternions (Dublin, 1853) and is again attempted in the present work, although with many differences in the adopted plan of exposition, and in
titled
the applications brought forward, or suppressed.
CHAP.
I.]
and
may
I,
= Oj'
= oi'
j = OK
OK,
Ol',
161
J, K.
OJ
by the general formula of multiplication (107), whatever four
to observe that,
lines
LAWS OF THE SYMBOLS,
be denoted by a,
y'
/3,
y' a
we have always,
y, ^,
y/3a'
jS'a
r briefly, as in algebra,
/3
the point being thus omitted without danger of confusion
= oj'
ijk
we have
Similarly,
On
in like
oj) (oj
(oj
oi) (oi
of association
and
1,
=+
jik
is still
obeyed
= ik = -j = iy = a j,
i .ji = i.>-k = -ik=j= ki =
ij
- 1 = - 1 = ^J = if-Ji
i/=
.
= -j;
= +j
w=-
ij
iij
.
f,
iji
with others deducible from
these,
by mere
plan illustrated by Fig. 47, his.

(4.) In general, if the Associative
symbols whatever of a given
of lines
1.
182 give also these other ternary products, in which the
in
offactors
i
tion
ok) (ok oj) =oj oj = +
ikj= +
law
= ok' ok = - 1
oj') (oj' ok)
=-1
ok') (ok' oi) = oi' oi
:
manner,
The equations
(3.)
1,
the other hand,

kji
and
oi) (oi
so that
as before.
two other ternary products
these
jki = (ok'
= (oi'
kij
(2.)
OJ =
class,
and
cyclical permutation of the letters,
Law
of Combination exist
for a given
in' Art. 9, or for multiplication
mode
for
on the
any three
of combination, as for addi-
in the present Article, the saine
of ijk
law
any^?<r (or more) symbols of the same class, and combinations of the same
For example, if each of the four letters i, k, X, fi denote some one of the three
exists for
kind.
symbols
j,
i,
k (but not necessarily the same one),

I
k\/li
X/i
= IK
\;[t
tK
we have
/i
the formula,
= tfcX n = ikXh.
(5.) Hence, any multiple (or complex) product of the symbols ijk, in any manner
repeated, but taken in one given order, may be interpreted, with one definite result^
by any mode of association, or of reduction to partialfactors, which can be performed

without commutation, or change of place of the given factors.
For example, the
symbol ijkkji may be interpreted in either of the two following (among other) ways ;
ij
kk.ji
= ij. -ji = i -j2

.
=u=-
ijk
.kji=-l.l=-l.
The formula (a) of 183 includes obviously the three equaTo show that it includes also the six other
equations, (II.), (III.)> of the last cited Article, we may observe that
184.
tions (I.) of 182.
it
with the help of the associative principle of multiplicatioa

may be suggested to the memory by the absenceof the jpoi/i^
gives,
(which
iu the
symbol
ijk),
162
~-ij .lck = - ijk ^ = +

= -k;
ji =j .jk=fk
=
= }'2 = _ i ;
kj ij j
.
ij
^'
[bOOK
jk = -i,ijk = -\-i\
ih
ki
= i ij = Pj = -j
= - Jc'j = -ji^ = +j.
.
And then it is easy to prove, without any reference to geometry^

foregoing laws of the symbols be admitted, that we have also,
jki = kij =
kji
II.
=jik =
ikj
=+
if
the
and geometrically shown in recent sub-articles. It may

be added that the mere inspection of the formula (a) is sufficient to
show that the three"^' square roots of negative unity, denoted in it by
as otherwise
I, ^',
h,
cannot he subject
to all the
ordinary rules of algebra: because
that formula gives, at sight,
2V'F = (-iy=-l=-(z}%-V;
the non-commutative cha7'acter(\ S3), of the multiplication ofsuch roots
among themselves, being thus put in evidence.
Section 11.
On the Tensor of a Vector, or of a Quaternion
and on the Product or Quotient of any two Quaternions.
185.
Having now
sufficiently availed ourselves, in the
two
of the conceptions (alluded to, so early as in the

First Article of these Elements) of a vector-arc (162), and of
last Sections,
a vector-angle (174), in illustration\ of the laws o^ multiplication and division of versors of quaternions
we propose to return to that use of the word, Vector, with which alone the
;
First Book, and the first eight Sections of this First Chapter
of the Second Book, have been concerned and shall therefore
:
henceforth
mean
word
again, exclusively, by that
''
vector," a
Directed Right Line (as in 1). And because we have already

considered and expressed the Direction of any such line, by
*
It is evident that
symbol V
- i, j, h
are also^ on the
same
principles, values of the
1;
because they also are right versors (153); or because (- qy=q\

More generally (comp. a Note to page 131), \ix, y, z be any three scalars which saz'i =
1, it will be proved, at a later stage, that
tisfy the condition x^ -\- y'i
-\-
t One of the chief uses of such
vectors, in connexion with those laws, has been
to illustrate the non-com'v.utative property
hibiting a corresponding property of
operation of the
a sphere.
(168) oi multiplication of versors, by ex-
what has been
same kind on linear vectors
Compare 180,
(3.), (4.).
called,
by analogy
(5), the addition of arcs
to the earlier
and angle$ on
CHAP.
TENSOR OF A VECTOR.
I.J
163
introducing the conception and notation (155) of the UnitVector^ Ua, which has the same direction with the line a, and
which we have proposed (156)

tor,
to call the Versor
of that Vecand express the Length of

by introducing the new name Tensor, and the
we now propose
to consider
the same line
a,
neio symbol,'^
Ta which
;
Tensor of the Vector a
latter
and
symbol we
shall read, as the
shall define it to be, or to denote,
Number (comp. again 155) which represents the Length of

that line a, by expressing the Ratio which that length bears
to some assumed standard, or Unit (128).
186. To connect more closely these two conceptions, of
the versor and the tensor of a vector, we may remember that
when we employed (in 155) the letter a as a temporary symbol for the number which thus expresses the length of the line
the
a,
we had
the equation,
Ua = a
, as
one form of the
defini-
denoted by Ua.
We might therefore
have written also these two other forms of equation (comp. 15,
tion of the unit-vector
16),
a=.Ua,
a = a:Ua,
to express the dependence of the vector, a, and of the scalar,

a, on each other, and on what has been called (156) the versor,
Ua. For example, with the construction of Fig. 42,

161, (2.)
),
fl
we may
= OA
oa',
his
(comp.
write the three equations,

6 = ob:ob',
= oc:oc',
if a, b, c be thus the three positive scalars, which denote the

lengths of the three lines, oa, ob, oc; and these three scalars
may then be considered as factors, or as coefficients (12), by
which the three
unit-vectors Ua, Uj3, Uy, or oa', ob', oc' (in

the cited Figure), are to be respectively multiplied (15), in
order to change them into the three other vectors a, /3, y, or
OA, oB, oc, by altering their lengths, without any change in

But such an exclusive Operation, on the
their directions.
Length (or on the extension^ of aline, may be said to be an Act

of Tension ;t as an operation on direction alone may be called
have then thus a motive
(comp. 151) an act of version.
We
Compare the Note to Art. 155.

t Compare the Note to Art. 156,
in
page 135.
164
for the introduction of the
positive number which
And when
line.
for such a tensor,
(as above) represents the length of a
see that
Ta
we
(instead of a) is employed
may write generally, /(?r awj^
vector a, the equations (compare again 15, 16)
Ua = a Ta
:
Ta = a
II.
name. Tensor, as applied to the
the notation
we
[bOOK
XJa
a=
Ta Ua = Ua Ta.
.
= a (160),
For example, if a be an unit- vector, so that Ua

Td = 1 and therefore, generally, whatever vector may
be denoted by a, we have always,
then
TUa=l.
For the same
q,
we
reason, whatever quaternion may be denoted

have always (comp. again 160) the equation,
(1.)
Hence the equation
T(Ax.^) =
where
p = op,
Tp =
by
l.
l,
is
is
the spheric surface
expresses that the locus of the variable point
the surface of the
unit sphere (128).

(2.)
The equation Tp
= Ta
with o for centre, which passes through the point a.

(3.)
On
the other hand, for the sphere through o, which has
its
centre at a, wê
have the equation,
T(p-a) = Ta;
which expresses that the lengths of the two
(4.)
More
lines,
ap, ao, are equal.
generally, the equation,
T(p-a) = T(/3-a),
expresses that the locus of
is
the spheric surface through b, which has
its
centre
at a.
(5.) The equation of the Apollonian* Locus, 145,
under either of the two following forms
(8.),
(9.),
may
be written
T(p-a2a) = aT(|0-a);
from each of which we shall
T|0=aTa;
find ourselves able to pass to the other, at a later stage,
by general Rules of Transformation, without appealing

(6.) The equation,
T((0
a)
to
geometry (cova^. 145, (10.)).
= T(,o-a),
expresses that the locus of p is the plane through o, perpendicular to the line oa
because it expresses that if oa' = - oa, then the point p is equally distant from the
;
two points A and
a'.
It represents therefore the
Compare the
first
Note
same locus as the equation,
to
page 128.
CHAP.
1.1
Z^=^,
2
a
6o
of 132,(1.);
or as the equation,
+ k!^=0,
(3
= -1,
or as the simple geometrical formula, p
by General Rules of
mulas into any other of them
possible,
of 144, (1.)
of 161, (7.)
J-a
(129).
And
in fact
it
will
be found
these ê fortransform any one of
this Calculus, to
or into this sixth form^
a
which expresses that the scalar part* of the quaternion -
is zero,
and therefore that
this quaternion is a right quotient (132).
(7.) In like
manner, the equation
T(p-/3)=T(p-a)
the plane which perpendicularly bisects the line ab ;
equally distant from the two points A and b.
tensor, Ta, being generally a positive scalar, but vanishing (as a limit)
because
it
expresses that p
The
(8.)
we
with a,
is
is
have,
Txa = + xTa,
thus, in particular,
(9.)
x>
according as
T (- a) = Ta
and
<
or
TOa = TO =
0.
That
T(/3
+ a) = T/3+Ta,
U^ = Ua,
if
but not otherwise (a and /S being any two actual vectors), will be seen, at a later
stage, to be a symbolical consequence from the rules of the present Calculus ; but in
mean time it may be geometrically proved, by conceiving that while a = oa, as

- OA = AC (4) for thus we shall
usual, we make /3 -f a = oc, and therefore jS = oc
the
see that while, in general, the three points o. A,
c are corners of a triangle, and there-
oc is less than the sum of the lengths of the two other

AC, the former length becomes, on the contrary, eg^waZ to the latter sum,
fore the length of the side

sides
OA and
in the particular case
line
oc
equation
in
which
triangle vanishes,
AC, or a and
jS,
by the point A falling on

have one common
the finite
direction, as the
Ua = U/3 implies.
(10.) If a and
when the
OA and
case,
j3
be any actual vectors, and
if their
versors be unequal
(Ua
not
U/S), then
T(/3 + a)<T,3 + Ta;
an inequality which results at once from the consideration of the recent triangle OAC
but which (as it will be found) may also be symbolically proved, by rules of the
;
calculus of quaternions.
Compare the Note
Chapter.
to
page 125
and the following Section of the present
166
(11.) If U/3
= - Ua,
+ a) = +
then T(/3
(T/3
- Ta),
II.
[bOOK
according as T/3
>
or
< Ta
but
T(l3+a)>{T(5-Ta),
The
187.
quotient,
Uj3
U/3o?=-Ua.
Ua, of the versors of the two vec-
has been called (in 156) the Versor of the Quoor quaternion, q = [5 a and has been denoted, as such,
tors, a
tient,
and
if
j3,
On
by the symbol, U^'.

call the quotient, Tj3
the same plan, we propose now to

Ta, of the tensors of the same two vec-
tors, the Tensor* of the Quaternion q, or f5: a, and to denote

it by the corresponding symbol, l^q.
And then, as we have
the
called the letter
characteristic
of the operation
(in 156)
of taking the versor, so

racteristic
sor,
we may now speak of T
as the
Cha-
of the (corresponding) Operation of taking the Ten-
whether of a Vector,
a,
We shall
or of a Quaternion, q.
thus have, generally,
T(j3
a)
= T/3 Ta,
as
we had U(i3
a)
U|3
Ua
(156)
and may say that as the versor U^ depended solely on, but
conversely was sufficient to determine, the relative direction
(157), so the tensor Tq depends on and determines the relative
length] (109), of the two vectors, a and j3, of which the quaternion q
(1.)
is
the quotient (112).
Hence the equation
presses that the locus of p
T- = l,
is
like
Tp = Ta,
the sphere with
to
which
for centre,
it
is
equivalent, ex-
which passes through the
point A.
Compare
the Note to Art. 109, in page 108
and that to Art. 156,
in page
135.
It
has been shown, in Art. 112, and in the Additional Illustrations of the
third Section of the present Chapter (113-116), thsit Relative Leriffth, as well as
an essential element into the very Conception of a QuaAccordingly, in Art. 117, an agreement of relative lengths (as well as an
agreement of relative directions) was made one of the conditions of equality between
relative direction, enters as
ternion.
any two quaternions, considered
as quotients of vectors
so that
we may now
say,
that the tensors (as well as the versors) of equal quaternions are equal.
Compare
the first Note to page 137, as regards what was there called the quantitative element,
of absolute or relative length,
the characteristic
U;
which was eliminated from
serves on the contrary to retain that element alone,

called
a, or
from
q,
by means
of
whereas the new characteristic, T, of the present Section,
and
by contrast the qualitative element, of absolute or
to eliminate
what may be
relative direction.
CHAP.
(2.)
TENSOR OF A QUATERNION.
I.]
The equation comp. 186,
1G7
(6.)),
T^â =
l,
p-
expresses that the locus of p is the plane through o, perpendicular to the line OA.
from the sub-arti(3.) Other examples of the same sort may easily be derived
cles to 186,
(4.)
(5.)
for the tensor
by introducing the notation (187)
T(/3
The
a) >, =, or
<
of a
according as T/3 >, =, or
1,
quotient, or qua-
of a vector (185).
ternion, as additional to that for the tensor
tensor of a right quotient (132)
is
always equal
< Ta.
to the tensor of its in-
dex (133).
(6.)
nerally,
and
The
tensor
of a radial (146)
by 156,
in particular,
is
always positive unity
TUg =
by 181,
T (- 9) = Tg^,
Ta;
(8.)
have, ge-
= T; = T/f=l.
Ti
thus, in particular,
we
thus
l;
Hxq = + xHq, according
(7.)
as c> or
<
or the tensors of opposite quaternions are equal.
= + X,
according as
a;
>
<
or
thus, the tensor of a scalar is that scalar taken positively.
(9.) Hence,
so that,
may
TTa = Ta,
by abstracting from the
TT^=Tg;
subject of the operation
(comp. 145, 160), we
establish the symbolical equation,
T^
= TT = T,
(10.) Because the tensor of a quaternion

tensor
is its
own conjugate (139)
positive unity
its
generally a positive scalar, such a

zero (131) ; and its versor (159) is
is
is
angle
or in symbols,
KTq^Tq;
/LTq
0',
VTq=l.
T(l:5) = T(a:/3) = Ta:T/3 = l:T9;
(11.)
or in words, the tensor
of the reciprocal of a quaternion
is
equal to the reciprocal of
the tensor.
(12.) Again, since the two Imes,

(137) of a conjugate gives
ob and
ob', in Fig. 36, are equally long, the de-
finition
TKq = Tq;
or in words, the tensors of conjugate quaternions are
equal.
(13.) It is scarcely necessary to remark, that any two quaternions which have
equal tensors, and equal versors, are themselves equal : or in symbols, that
q'
188. Since
= q,
we
if
Tq'
= Tq,
and
Vq'='Uq.
have, generally,
'
U0 T/3
U^Û0
T
Ta
la 'Tt
\Ja
Ua =Ti^-T^
Ua la
= Ti3
/J_
= T/3.Ui3 ^Tj3
/3
we may
T
tT
Ta.Ua
ia.Ua
establish the
Ui3
,
,^,
,,
126, 186),
(comp.
^
^
^
,
two following general formulae of decom-
168
[bOOK
II.
position of a quaternion into two factors^ of the tensor 2CDidiver'

sor kinds :
1.
q=Tq
\Jq
W.
= \]q
Tq
which are exactly analogous to the formulae (186)
for the cor-
responding decomposition of a vector, mio factors of the same

two kinds namely,
:
r.
To
a =
Ta.Ua;
11.
a =
Ua.Ta.
decomposition of a quaternion, q, or
OB oA, into factors, we may conceive that aa' and bb' are two
concentric and circular, but oppositely directed arcs, which
illustrate this last
terminate respectively on the two

lines OB and oa, or rather on the
longer of those two lines itself, and
on the shorter of them prolonged,
as in the
annexed Figure 48
so
that oa' has the length of oa, but

the direction of ob, while ob', on the
contrary, has the length of ob, but
the direction of oa and that therefore
we may
write,
by what
has been defined respecting versors and tensors oi vectors

156, 185, 186),
OA' =
Ta.U/3;
{\
65)
0B'=T/3.Ua.
Then, by the definitions in 156, 187, of the versor and tensor

of a quaternion,
TJq
T^-
= U(oB oa) = oa' oa = ob ob'

= T (ob oa) = ob' oa = ob oa'
:
whence, by the general formula of multiplication of quotients

(107),
I.
5^
= OB
OA = (ob
oa')
^^
= ob
OA = (ob
ob')
(oa'
(ob'
oa)
Tq
Uq;
and
II.
oa) = IJq Tq,

.
as above.
189. In words, if
tor
we wish
or from the line
oa
to pass
from the vector a
to the line ob,
we
to the vec-
are at liberty either,

1st, to begin by turning^ from oa to oa', and then to end
by stretching.
/3,
CHAP.
TENSOR OF A QUATERNION.
I.]
169
from oa' to ob, as Fig. 48 may serve to illustrate; or, Ilnd, to begin
by stretching, from oa to ob', and e7id by turning, from ob' to ob.
The act of multiplication of a line a by a quaternion q^ considered as
z,
factor (103), which affects hoth length and direction (109), may
thus be decomposed into two distinct and partial acts^ of the kinds
which we have called Version and Tension ; and these two acts may
be performed, at pleasure, in either of two orders of succession. And
although, if we attended merely to lengths, we might be led to say
that the tensor of a quaternion was a signless number,^ expressive of
a geometrical ratio of magnitudes, yet when the recent construction
(Fig. 48) is adopted, we see, by either of the two resulting expressions (188) for Tg, that there is 2k propriety in treating this tensor
as
21.
positive scalar, as
we have
and propose systemati-
lately done,
cally to do.
TKg = T^,
190. Since
we may
by 187, (12.), and UK^=1:U^, by 158,

any quaternion and its conjugate, the
write, generally, for
two connected expressions
,q = Tq.Vq',
whence, by multiplication and
III.
? K^ = (T^)2
I.
Kg = T^:U^;
II. ..
division,
lY. .q-A^q = {\Jq)\

;
formula had occurred before; and we saw (161) that in it
ihQ parentheses might be omitted, because (U<^)^ = U(^2).
In like
manner (comp. 161, (2.) ), we have also
.
This
last
{Tqy=T{f)=Tq\
parentheses being again omitted ; or in words, the tensor of the square
of a quaternion is always equal to the square of the tensor: as
appears (among other ways) from inspection of Fig. 42, his, in which
the lengths of oa, ob, oc form a geometrical progression
T.f Y=T
OA
\OAy
At
the same time,
gate quaternions,
we
^'^^
T.OA
whence
^^'^^
\T.OA,)=-!
see again that the product
which has been
called (145,
qKq of two conju(H.) ) their common
Norm, and denoted by the symbol
'^q, represents geometrically the

square of the quotietit of the lengths of the two lines, of which (when
considered as vectors) the quaternion q is itself the quotient (112).
>ye
may
V.
therefore write generally,!

.
qYiq
*
= 1q' = l^q',
VI.
Compare the Note
in
t Compare the Note
in
Tq= y/^q^ V(^K^).
page 108,
page 129.
to Art. 109.
170
(1.)
We
have
by
also,
tensor of a quatemion
VII.
II.,
[bOOK
the followhig other general transformations for the
.Tq = Kq.JJq]
VIII.
Tg = U? Kg
of which the geometrical significations might easily be exhibited

of
which the validity

(2.) Also
(3.)
now be
The
is suflSiciently
proved by what precedes.
reciprocal of a quaternion, and the conjugate* of that reciprocal,
may
may
q~Tq2~-^q~ Tq
"We
by a diagram, but
(comp. 158),
thus expressed
(4.)
II.
Tq' Vq'
Vq' Tq
Kq
Tq
Tg2
also write, generally,
IX.
.Kq=Tq. KUq = ^q
q.
'.
191. In general, let ani/ two quaternions, q and^', be considered as multiplicand and multiplier, and let them be reduced (by 120) to the forms /3 a and 7 j3 then the tensor
:
and versor of
theit
third quaternion, y.a, which
be thus expressed
their product qq,
is
(by 107)
may
L..Tqq^T(y:a) = Ty:Ta={Ty:Tp).(T^:Ta) = Tq\Tq;
II...U^V = U(7:a) = U7:Ua=(U7:U^).(Ui3:Uo) = U^'.U^;
where Tqq and Uqq are written, for simplicity, instead of
T(q'.q) and
Hence, in any such multiplication, the

($'.').
tensor of the product is th.Q product of the tensor; and the versor of the product is the product of the versor s ; the order of
the factors being generally retained for the latter {Q,(yay^. 168,
&c.), although it may be varied for the former^ on account of
In like manner, for the diviany one quaternion q\ by any other q, we have the
analogous formula
= JJq'lJJq;
=
IV.
III.
(q'
q)
Tq
the scalar character of a tensor.

sion of
T{q
q)
Tq'
or in words, the tensor of the quotient of any tAvo quateris equal to the quotient of the tensors ; and similarly, the
versor of the quotient is equal to the quotient of the versors.
nions
And
because multiplication and division of tensors are per-
formed according to the rules of algebra, or rather of anMz^*
Compare
Art. 145, and the Note to page 127.
CHAP.
I.]
PRODUCT OR QUOTIENT OF TWO QUATERNIONS.
17
(a tensor being always, by what precedes, a positive number)^ we see that the difficulty (whatever it may be) of the
tic
general multiplication and division of quaternions is thus reduced to that of the corresponding operations on versors : for
which latter operations geometrical constructions have been

assigned, in the ninth Section of the present Chapter.
(1.)
two
The two
and qq\ of any two quaternions taken as factors in

two factors are compla-
products, q'q
different orders, are equal or unequal, according as those
nar or diplanar
because such equality (169), or inequality (168), has been already

is unity
but we have always
proved to exist, for the case* when each tensor

(comp. 178),
Tq'q
= Tqq,
TT
(2.) If
Lq = Lq ->
and
= Kq'q
then qq'
Lqq= L
(170)
qq.
two right
SO that the products of
quotients, or right quaternions (132), taken in opposite orders, are
always conju-
gate quaternions.
(3.) If
Lqq'=lq'q = ^,
80 that the product
Ax. 5 -i-Ax.y,
and
Lq-= Lq' "= t
Ax
9'?
-i-
Ax
then
Ax.
^,
qq'
J-
q'q
= q'qi
A^
q {171)
of two right quaternions, in two rectangular planes,
is
a third
to both ; and is changed to its own opposite,

right quaternion, in a plane rectangular
when the order of the factors is reversed : as we had ij=k=-ji (182).
(4.) In general,
if
q and q' be any two diplanar quaternions, the rotation round
g to Ax q'q, is positive (177).

is a quaternion with the same tensor,
(5.) Under the same condition, q[{q' 5)
and same angle, as q', but with a different axis ; and this new axis, Ax g{q' g),
Ax
q',
from
Ax
may
be derived (179, (1.)
positive direction)
(6.)
round
The product
from the old
Ax
q,
through an
or quotient of
Ax q', by a
angle = 2 Lq.
axis,
conical rotation (in the
two complanar quaternions
is,
in general, a third
they be both scalar, or both right, then this

into a scalar.
product or quotient degenerates (131)
Whether q and q' be complanar or diplanar, we have always as in algebra
quaternion complanar with both
but
if
(7.)
(comp. 106, 107, 136) the two identical equations:
V.
(8.) Also,
by 190,
(5'
9)
g-
= g'
and 191,
V.,
VII.
norm of the product
or in words, the
192. Let
5'
= j3
a,
and
VI.
I.,
we have
(5'
g)
q'.
this other general
formula
N5'5=N9'.Ng;
is
equal to the product of the norms.
q'
= y:
(5,
as before
then
= (1 ^) (1
1
/^)
t)
:^');
gV= 1 (7
of
two
the
the
so that
any
quaternions is
product
reciprocal of
:
= (
) = 7
Compare the Notes
to
pages 148, 151.
172
[bOOK
II.
equal to the product of the reciprocals, taken in an inverted

order : or briefly,
I.
if
Rq'q
Rg
R^',
be again used (as in 161, (3.)) as a (temporary) charac-
of reciprocation. And because we have then (by the

= UR, or in
sub-article) the symbolical equation,
words, the conjugate of the versor of any quaternion q is equal
teristic
KU
same
(158) to the versor of the reciprocal of that quaternion ; while

the versor of a product is equal (191) to the product of the
versors
we
see that
UR^ UR^' = KU^
KU^'^ = UR^'^ =
KUg'.
But
%= Tq
by 191
190, IX.
KU^, by
we
and Tq'q =
Tq Tq = Tq,Tq\
,
arrive then thus at the following other important
and general formula
11.
.Kq'q==Kq.Kq';
or in words, the conjugate of the product of any two quaternions is equal to the product of the conjugates, taken (still)
in
an inverted order.
(1.) These
two
results, I., II.,
may
be illustrated, for versors (Tq
the consideration of a spherical triangle
AB and Bc (comp. 167) may
abc (comp.
represent q and
q',
= Tq' = 1), by
in which the sides

Fig. 43)
the arc Ac then representing q'q;
For then the ncAV multiplier Eg- = 'Kq (158) is represented (162) by ba, and the new
multiplicand Kg' = Kg' by CB whence the new product, B-q.^q =^q.^q\ is represented by the inverse arc CA, and is therefore at once the reciprocal "Rq'q, and the
;
conjugate Kq'q, of the old product q'q.

(2.) If q and q' be right quatei-nions, then
Kq = - g,
Kg' =
g'
the recent formula II. becomes, Kq'q = qq% as in 170.

(3.) In general, that formula II. (of 192) may be thus written
III.
a,
/3,
and
|8
may
B^
y^'j
verse simi'/iYwde of triangles (118),
Aaoboc'bog,
ABOEa'DOB;
and therefore (by 137) the two equations,
/3
K^ = K^.K^;
denote ang three vectors.

then
that, as in the annexed
Suppose
(4.)
Fig. 49, we have the two following relations of in-
where
(by 144)
a'
/3'
"-^^
^/a/^-"""^
i^^-""'^^
^
1^^^--^^
^"^-.
\\
'
IiAt)
K\
Fig. 49.
CHAP.
we
TWO RIGHT QUATERNIONS.
CASE OF
I.]
shall have,
by
ye
K
III.,
4;
so that this third
A DOC a aoe
or
-,
formula of inverse similitude
is
173
a consequence from the other two.
any two circles, whether in one plane or in space,

touch one another at a point b
and if from any point o, on the common tangent bo,
two secants OAC, oed be drawn, to these two circles the four points of section^
(5.) If then (comp. 145, (6.) )
;
A, c, D, E, will be on one
common
circle
for
quence (through equal angles, &c.), from the
such concircularity
an easy conse-
is
last inverse similitude.
(6.) The same conclusion (respecting concircularity, &c.) may be otherwise and
geometrically drawn, from the equality of the two rectangles, AOC and doe, each
being equal to the square of the tajigent on ; which may serve as an instructive
and as an example of the consistency of the

which calculations with quaternions conduct.
(7.) It may be noticed that the construction would in general give three circles,
although only owe is drawn in the Figure but that if the two triangles abc and
verification of the recent formula III.,
results, to
then these three
T>Ê,\)Q &\i\x3iiQdi in different planes,
points abode, are situated on one
common
circles,
and
of course i\iQ five
sphere.
An important application of the foregoing general

of
theory
Multiplication and Division, is to the case of Right
Quaternions (132), taken in connexion with ihoiv Index- Vec193.
tor s^ or Indices (133).

Considering division first,
and employing the general forand y be each _L a and let j3' and y be the
=
respective indices of the two right quotients, q [5 a, and
=
q'
y:a. We shall thus have the two complanarities, /3' j3, y,
and y j3, 7 (comp. 123), because the four lines /3, y, j3', y
and within their common plane it
are all perpendicular to a
mula of
06, let
/3
|||
III
easy to see, from definitions already given, that these four

lines form a proportion of vectors^ in the same sense in which
is
"
i3?
75 S did so, in the fourth Section of the present
so that
we may
7':/3'
In
fact, Ave
Chapter
write the equation of quotients,
t:^.
have (by 133, 185, 187) the following relations of
length,
T/3'
T/3
Ta,
T7'
T7
Ta, and
.-.
T (7'
/3')
T (7
while the relation of directions, expressed by the formula,
U(y:|3>U(7:^),
is
easily established
or
U7
UjS'
U7
by means of the equations,
UjS,
/3)
174
Z(y:y) =
Ax .(y' :y) = Ax
^((3':/3)
|;
[boOK
.
(/3' :/3)
II.
= Ua.
We
arrive, then, at this general Theorem (comp. again 133):

that ^^the Quotient of any two Right Quaternions is equal to
the Quotient of their Indices y*'
(1.) For example (corap. 150, 159, 181), the indices of the right versors i^j, Jc
are the axes of those three versors, namely, the lines oi, oj, ok ; and we have the
equal quotients,
= OI
(2.) In like manner, the indices

i
j = oj'
(3.) In general the quotient
their axes
oj'
= A = OJ
oi,
of - i, j, k
:
oi'
= A = OI
&c.
are
Oj',
ok'
oi', oj',
and
&c.
of ant/ two right versors
equal to the quotient of
is
as the theory of representative arcs, and of their poles,
may
easily
As regards the multiplication of two right quaternions,

connexion with their indices, it may here suffice to observe
194.
in
by 106 and 107, \\iQ product y a = (y j3) (j3 a) is equal

(comp. 136) to the quotient, (7 P) (" i^) whence it is easy
to infer that 'Hhe Product^ q'q, of any two Right Quaternions,
that,
equal to the Quotient of the Index of the Multiplier, q\ divided by the Index of the Reciprocal of the Multiplicand, q^
It follows that the plane, whether of the product or of the
is
quotient of two right quaternions, coincides with the plane of

; and therefore also with the plane of their axes ;
their indices
because-
we
have, generally, by principles already established,
the transformation,
if
We
5'
= -, then Index of q = 'Yq
Ax
q,
have thus a new point of agreement, or of connexion, between right qua-
their index-vectors, tending to justify the ultimate assumption (not yet

made), of equality between the former and the latter. In fact, we shall soon prove
that the index of the sum (or difference), of any two right quotients (132), is equal to
ternions,
and
the sum (or difference) of their indices ; and shall find it convenient subsequently to
interpret the product f3a of any two vectors, as being the quaternion-product (194)
of the two right quaternions, of which those two lines are the indices (133):
after]
which, the above-mentioned assumption of equality will appear natural, and be found
to be useful.
(Compare the Notes to pages 119, 186.)
CHAP.
SUM OF TWO QUATERNIONS.
I.]
Section I2. O71

nions ; and on
Sum
the
175
or difference of any two Quater-
the Scalar {or Scalar Part)
of a Quater-
nion,
195. The Addition of any given quaternion q, considered

other given
a
as
geometrical quotient or fraction (101), to any
can
a
as
also
fraction,
always be acquaternion ^, considered
complished by the first general formula of Art. 106, when these

two fractions have a common denominator ; and if they be not
already given as having such, they can always be reduced so as
And because the adto have one, by the process of Art. 120.
dition of any two lines was early seen to be a commutative operation (7, 9), so that we have always 7 + j3 = j3 + 7, it follows
(by 106) that the addition of any two quaternions is likewise a
commutative operation, or in symbols, that

I.
.q + q=^q+q',
Sum
of any two* Quaternions has a Value, which

their Order : and which (by what precedes)
of
independent
must be considered to be given, or at least known, or definite,
when the two summand quaternions are given. It is easy also
so that the
is
to see that the conjugate of any such

of the conjugates^ or in symbols, that
II.
(1.)
is
sum
is
equal to the sum
..K(^+y)=K^' + K^.
The important formula
last -written
becomes geometrically evident, when
it
Let obdc be any parallelogram, and let oa

drawn from one comer of it, but not generally in its plane. Let
presented under the following form.
be any right
line,
the three other corners, b,
c, d,
new
to that line oa, into tbree
be reflected (in the sense of 145, (5.) ) with respect

or let the three lines ob, oc, od be
b', c', d'
points,
138) with respect
reflected (in the sense of
to the
same
at right angles the three joining hnes, bb', cc', dd', as
line
it
OA; which thus
does bb' in Fig. 36.
bisects
Then
each of the lines ob, oc, od, and therefore also the whole jaZanej^^wre obdc, may be
considered to have simply revolved round the line oa as an axis^ by a conical rota-
through two right angles ; and consequently the new figure ob'd'c', like that old
one OBDC, must be a. parallelogram. Thus (comp. 106, 137), we have
tion
= oc' + ob',
^'=:y'
and the recent formula IL
is justified.
od'
+ j8',
5':
= (y':a)+
(]8':
a);
* It will be found that this result admits of

being extended to the case of three
(or more^ quaternions
but, for the
moment, we content ourselves with
two.
176
(2.) Simple as this last reasoning
any new Diagram
to illustrate
it,
and unnecessary as
is,
the reader's attention
[bOOK
it
II.
appears to be to draw
once more invited to
may be
the great simplicity of expression, with which
many important geometrical concepspace of three dimensions, are stated in the present Calculus: and
are thereby kept ready for future application, and for easy combination with other
tions, respecting
Compare the remarks already made
same kind.
results of the
in 132, (6.)
145,
192,(6.); and some of the shortly following sub-articles

196, respecting properties of an oblique cone with circular base.
(10.); 161; 179,(3.);
One
196.
to
of the most important cases oi addition ^ is that

it has been
of two conjugate summands, q and K^ of which

seen (in 140) that the sum is always a scalar.
;
now
sum by
to denote the ^^Z/" of this
We
propose
the symbol,
Sq;
thus writing generally,
I.
or defining the
II.
q +
Kq = Kq + q =
new symbol Sq by
..S^ = i(^ + K^);
2Sq;
the formula,
or briefly,
11'.
S=i
(1
+ K).
will soon more fully appear, we shall also

quantity, ^q, the scalar part, or simply the Scalar, of the Quaternion, q ; and shall therefore call the letter
S, thus used, the Characteristic of the Operation of taking the
For reasons which

call this
new
^caZar of a quaternion. (Comp. 132, (6.) ; 137; 156; 187.)

It follows that not only equal quaternions, but also conjugate
quaternions, have equal scalars ; or in symbols,
III.
8*7'
= S^,
if
q^q-,
or briefly,
IV'.
and
IV.
SK5 = S^
.SK = S.
And because we have seen that K^ = + 5', if 5' be a scalar (139),

but that lLq = -q, if ^^ be a right quotient (144), we find that
the scalar of a scalar (considered as a degenerate quaternion,
131) is equal to that scalar itself, but that the scalar of a right
quaternion
V.
and
Saj
is
= X,
zero.
if
a;
We may
be a scalar
VII.
now write (comp. 160):

VI. .SS^ = ^q, 8^ = 88 = 8;
therefore
.8^ =
0,
Again, because oa' in Fig. 36

multiplied thereby,
we may
is
if
z^ = |.
multiplied by x,
write, generally,
when ob
is
CHAP.
SCALAR OF A QUATERNION.
I.]
VIII.
and therefore
Sxq = xSq,
in particular
IX.
Also because
..
S^
if
177
X be any scalar;
(by 188),
S(T^.U^)=T^.SU^.
IV., while KU'5' =
SKq=Sq, by
U-, by
158,
we have the general equation,

X.
,SUq = SV-;
X'.
or
SU@ = SU ^
whence, by IX.,
XI.
.S^ = T^.Sui;
XI.
or
and therefore
XII.
also,
S^ =
by 190, (V.),
since
To^si = %.S-;
results of 142,
S^ = T^. SU^
'
T^.T- =
xir...
1,
"
S^ = N^.S
a
$'
The
combined with the recent
/3
II.,
definition
I.
or
enable us to extend the recent formula VII., by WTiting,
XIII.
S^^ >, =,
or < 0, according as Z y <, =, or >
and conversely,
XIV.
5-
<, =,
or > -, according as Sg >, =, or < q.
we compare that definition I. with the formula of

and
with Fig. 36, we see at once that because, in that
140,
In
fact, if
Figure,
S (oB oa) = oa' oa,

:
we may
XV.
write, generally,
.
S9 = T^.cosZ^;
or
XVI.
SU^ = cosz^;
equations which will be found of great importance, as serving

to connect quaternions with trigonometry ; and which show
that
XVII. .,Lq=Lq,
if
SU^' = S U^,
the angle Lq being still taken (as in 130), so as not to

and tt whence also,
outside the limits
;
2 a
fail
178
XVlll. ..Lq'=Lq,
if
S^'
= S^,
[bOOK
and
II.
T^'=Ty,
the anple of a quaternion being thus given, when the scalar

and the tensor of that quaternion are given, or known. Finally because, in the
same Figure 36 (comp.
15, 103), the
line,
oa' = (oa'
=
oa) oa oa
.
S (ob
oa),
may be said to be the projection of ob on oa, since a is the

foot of the perpendicular let fall from the point b upon this
latter line oa, we may establish this other general formula
:
XIX.
a result which
aS = S a = projection off3 on a
a
a
Avill
be found to be of great
gations respecting geometrical

written thus
loci,
utility, in investi-
and which may be
also
XX.
=
Projection o/ j3 ow a
Ua
Tj3
SU ^
a
with other transformations deducible from principles stated

It is scarcely necessary to remark that, on account
above.
of the scalar character of
we
Sj',
have, generally,
by 159, and
187, (8.), the expressions,
XXL
while, for the
US^ = 1;
XXII.
TS^ = +S^;
same reason, we have always, by 139, the equa-
tion (comp. IV.),
XXIII.
and,
by
S^
or
XXIir.
which
:*
S^^
0,
or =
TT,
The
unless
Lq = -;
187.
equation,
8-^
now
KS = S
by 159, but TS^ vanishing, by 186,
is
0, by VII., and therefore L Sq is inde\]Sq becoming at the same time indeterminate,
last case S^^
terminate
(1.)
131,
XXIV.
in
KS^ =
= 0,
seen to be equivalent to the formula, p J-
Compare the Note
in
page 118,
and therefore
to Art. 131.
to denote the
J
^
CHAP.
I.]
same plane locus
which
for p, as that
equations of 186, (6.)
or
is
l,
p-a
The
of the four other
of 187, (2.).
equation,
sP::^=o,
se=s?,
a
a
or
expresses that
bp
j_
oa
or that the locus of p

(3.)
any one
by the equation,
T^^ =
(2.)
represented by
179
The
is
or that the points b and p have the same projection on

the plane through b, perpendicular to the line OA.
;
oa
equation,
sue=su5,
a
a
expresses (comp. 132, (2.) ) that
and oa
on one sheet of a cone of revolution, with o
is
and passing through the point B.

(4.) The other sheet of the same cone is represented by
vertex,
for
for axis,
this other equation,
and both sheets jointly by the equation,
(6.)
The
equation,
S^ =
a
it
self; or that the angle
(6.)
the plane through A, perpendicular to the line OA j

expresses (comp. XIX. j that the projection of op on oa is the line oa it^
because
SU^=T-,
a
or
l,
On
the other
oap
is
is
right
or that
= 0.
hand the equation,
S^=l,
or
SU? = T^,
)3'
expresses that the projection of ob on op is op itself or that the angle opb is right ;
or that the locus of p is that spheric surface, which has the line ob for a diameter.
;
(7.)
Hence the system of the two equations,
S^ = l,
S^=l,
represents the circle, in which the sphere (6.), with

plane (5.), with OA for the perpendicular let fall on
(8.)
And
therefore this
new
for a diameter, is cut
last, represents
the Cyclic* Cone (or cone of the
Historically speaking, the oblique cone with circular base
named
by the
from o.
equation,
obtained by multiplying the two
ob
it
may
deserve to be
the Apollonian Cone, from Apollonius of Perga, in whose great
work on Co-
180
[bOOK
II.
second order, but not generally of revolution), which rests on this last circle (7.) as
and has the point o for its vertex. In fact, the equation (8.) is evidently
its base,
satisfied,
when
the two equations (7.) are so; and therefore every point of the circuby those two equations, must be a point of the locus, re-
lar circumference, denoted
But the
presented by the equation (8.).
fore,
when
latter equation
remains unchanged, at
least
changed to xp, x being any scalar ; the locus (8.) is, theresome conical surface, with its vertex at the origin, O and consequently it can
essentially,
is
be none other than that particular cojie (both ways prolonged), which
above) on the given circular base (7.).
(9.)
The system
rests (as
which the point may be omitted,) represents a conic section ;

in which the cone (8.) is cut by the new plane, which has oc
(in writing the first of
namely that
section,
for the perpendicular let fall
upon
it,
from the origin of vectors
o.
(10.) Conversely, every plane ellipse (or other conic section) in space, of which
the plane does not pass through the origin, may be represented by a system of two
because the cone which rests on any such conic as

equations, of this last form (9.)
and has its vertex at any given point o, is known to be a cyclic cone.
;
its base,
(11.) The curve (or rather the pair of curves), in which an oblique but cyclic
cone (8.) is cut by a concentric sphere (that is to say, a cone resting on a circular
base by a sphere which has its centre at the vertex of that cone), has come, in mo-
dern times, to be called a Spherical Conic. And any snch conic may, on the foregoing plan, be represented by the system of the two equations,
S^ S^=l,
a
Tp=l;
the length of the radius of the sphere being here, for simplicity, supposed to be the
unit of Lngth.
But, by writing Tp a, where a may denote any constant and positive scalar,
we can
venient to do
(12.)
at once
The equation
191, VII.)
remove
this last restriction, if it be
thought useful or con-
so.
(8.)
may
be written, by XII. or XII'., under the form (comp.
p
or briefly,
/3
"Mil
sJs?
nics (icojvtKuiv), already referred to in a
Note
to page 128, the properties of such a
cone appear to have been first treated systematically although the cone of revoluBut the designation ^^ cyclic cone" is shorter; and
tion had been studied by Euclid.
;
it
seems more natural, in geometry, to speak of the above-mentioned oblique cone

marking its connexion with the circle, than to call it, as is
thus, for the purpose of
now
usually done, a cone of the second order, or of the second degree
these phrases also have their advantages.
although
CHAP.
I.]
if a'
= Ta.U/3,
/3T^
so that
a and
which
on the sphere
is
=1
/3'
(6.),
aT^=T^.Ua
and Fig. 48.
by the plane
(5.) in the circle (7,),
but also by the (generally) new plane, S
which
new
this
S^,=
a
by the system
1,
new
in the
sphere,
two equations,
of the
S^'=l.
l,
when
(14.) In the particular case
-=
plane cuts the (generally)
or in the circle which is represented
which must be
ob', of Art. 188,
(8.) is cut, not only
(generally) neto circle, in
and
are here the lines oa'
/3'
Hence the cone
(13.)
and
181
/3
a (15), so that the quotient
|1
/3
is
a sca-
and greater than unity, in order that the plane (5.)

(really) cut the sphere (6.), and therefore that the circle (7.) and the cone (8.)
lar,
be real, we
write
may
^ = a'â,
and the
may
may
positive
T(/3:a) = a2,
a>\,
= a,
= f3;
f3'
(13.) coincides with the circle (7.).

(15.) In the same case, the cone is one of revolution; every point p of its circular base (that is, of the circumference thereof) being at one constant distance from
circle
namely at a distance
the vertex o,
(7.) give,
= aTa.
For, in the case supposed, the equations
by XII.,
N^ = S-:S-=l:S- = a2:S^=a2;
a
(Compare 145,
(12.),
and 186,
To = aTa.
or
p
(5.). )
(16.) Conversely, if the cone be one of revolution, the equations (7.) must conduct to a result of the form,
a=^=N^ =
a
S^S?=S^:SV
a
p
p
p
which can only be by the

(14.)
since othftrwMse
the base
we should
would be situated
actual cone)
line
aâ
f3
in one
s'^Î^Ô;
p
vanishing, or
have, by XIV., p
-i-
by our having
aâ,
(S
and
plane passing through the vertex
j3
= a'â,
as in
all the points
o,
which
(for
of
any
would be absurd.
(17.) Supposing, then, that

as in (14.), nor even a'
tion (or
(comp.(2.)),
^
what
is
\\
(3'
\\
we have not (H a, and

we see that the cone
\\
therefore not a'
(8.)
(3,
often called a right cone)
but that
it is,
And we
(or scalene) cone, although still a ct/clic one.
is
= a,
/3'
= (3,
not a cone of revolu-
on the contrary, an oblique
see that such a cone is cut in
two distinct series* of circular sections, by planes parallel to the two distinct (and
mutually non-parallel) planes, (5.) and (13.) or to two new planes, drawn through
;
the vertex o, which have been calledf the two Cyclic Planes of the
cone, namely, the
two following
1\\QS,Q
two series o^ subcontrary (or antiparallel) hut circular sections of a

have been first discovered by Apollonius see the Fifth Propo-
cyclic cone, appear to

sition of his First
Book,
in
which he
(page 22 of Halley's Edition).
By M.
Chasles.
says, KuXeiaOuj î
i)
roiavTr} rofxri
vwivavria
182
a
while the
two lines from the vertex,
planes respectively,
may
(18.) Of these two

the sphere (6.)> which
fi
OA and
ob, which
Sive
perpendicular to these two
be said to be the two Cyclic Normals.
a and
lines,
may
/3,
the second has been seen to be a diameter of
be said to be circumscribed to the cone
when
(8.),
that
considered as having the circle (7.) for its base ; the second cyclic plane (17.)
therefore the tangent plane at the vertex of the cone, to i\xdX first circumscribed
cone
is
II.
[bOOK
is
sphere (6.).
(19.) The sphere (13.) may in like manner be said to be circumscribed to the
cone, if the latter be considered as resting on the new circle (13.), or as terminated by
that circle as
its
new base ; and the diameter of this new sphere is the line ob', or
direction of the line a, or of the^rs^ cyclic normal (17.)
which has by (12.) the
that (comp. (18.)) t\\Q first cyclic plane

second circumscribed sphere (13.).
(20.)
Any
and which
is
other sphere through the vertex, which touches the first cyclic plane,
where
b' is
some
scalar
by the equation,
S^ = l,
S2' = i;
^
p
or
therefore cuts the cone in a circle, of which (by (12.)
S^, =
a
\\
/3
plane of section of sphere and cone is parallel

(21.) In like manner any sphere, such as
h
the equation of the plane
is
l,
S-^,=
b a
or
6',
80 that the perpendicular from the vertex is b'a

this
so
the tangent plane at the vertex, to the
therefore has its diameter from the vertex =b'(S\
co-eflScient, is represented
it
/3',
(comp. (5.) ) and consequently

second cyclic plane (17.).
;
to the
1,
where
6 is
any
scalar,
P
which touches the second
cle,
and
cyclic
plane at the vertex, intersects the cone (8.) in a
cir-
of which the plane has for equation,
is
therefore /)ara//eZ to the first cyclic plane.
(22.)
The equation
of the cone (by IX., X.,
SU^.SU^ = T^;
or,
cos
XVI.) may
Z^ .cos
Z-g
also
be thus written
= T^;
of the inclinations, of any vaof an oblique cyclic cone, to two fixed lines (a and /3), namely to the
two cyclic normals (17.), is constant ; or that the product of the sines of the inclinaof the cone, to two fixed planes, namely to
tions, of the same variable side (or ray, p)
it
expresses, therefore, that the product of the cosines
viable side (p)
the two cyclic planes,

(23.)
is
thus a constant quantity.

in which the concentric sphere
The two great circles,
cyclic planes,
have been
called the two Cyclic
Bv M.
Tp =
1 is cut
by the two
Arcs* of the Spherical Conic
Chasles.
(11.), in
CHAP.
SCALAR OF A SUM OR DIFFERENCE.
I.J
which that sphere

sines
is
It follows
cut by the cone.
of the (arcual) perpendiculars,

on
conic,
its
two cyclic arcs,
is
let fall
(by (22.)
from
) that the
ani/ point
183
product of the
v of a given spherical
constant.
are not put for(24.) These properties of cyclic cones, and of spherical conies,
ward as new; but they are of importance enough, and have been here deduced with
sufficient facility, to
show that we are already
own Rules* of Transformation,
in possession of a Calculus, with its

yrhereby one enunciation of a geometrical theorem, or
problem, or construction, can be translated into several others, of which some

be clearer, or simpler, or more elegant, than the one first proposed.
197.
and
let
then, if
may
Let a, j3, 7 be any three co-initial vectors, oa, &c.,

0D = S = 'y+/3, so that obdc is a parallelogram (6);
we
write
y:a = q\
(5:a^q,
^: a = q" =
and
q'
+ q (106),
and suppose that b', c', d' are the feet of perpendiculars let
from the points b, c, d on the line oa, we shall have, by
fall
196,
XIX., the
expressions,
= aSq,
^ = aSq" = aS (q' 4- q).
y = aSq,
But also OB = CD, and therefore ob' = c'd', the similar projections of equal lines being equal ; hence (comp. 11) the sum of
(ob'
=.) |3'
must be equal to the projec-
the projections of the lines /3, y

tion of the sum, or in symbols,
od'
= oc'+ob',
g'
S':
y+j3',
Hence, generally, for any

the formula
tivo
a = (7': a)+(j3': a).
quaternions, q
and q, we have
L.
S(^'+9)
S^'
+ S^;
or in words, the scalar of the sum is equal to the sum of the

scalar s. It is easy to extend this result to the case of any three
(or more) quaternions, with their respective scalars
q be a third arbitrary quaternion, we may write
thus, if
S(S"+ (2+ ?)) = 8/+ S (j'+ ?) = S/ + (S</+ Sq);

where, on account of the scalar character of the summands, the
last
parentheses
may be
omitted.
We
may
therefore write,
generally,
II.
where S
is
SS^ = 2Sg,
or briefly,
used as a sign of Summation

*
Comp. 146,
(10.), &c.
SS = 2S
:
and may say that
184
the Operation
[bOOK
the Scalar of a Quaternion
of taking
As
tributive Operation
is
II.
a Dis-
to the general Subtrac-
(comp. 13).
of any one quaternion from any other, there is no difficulty
in reducing it, by the method of Art. 120, to the second genetion
nor in proving that the Scalar oftheDiffealways equal to the Difference of the Scalars. In
formula of 106
ral
rence^
is
symbols,
III.
.S(^'-^) = S^'-S^;
or briefly,
SA=AS;
IV.. .SA^-ASg,
when A
is
used as the characteristic of the operation of taking
a difference, by subtracting one quaternion, or one
scalar,
from
another.
any number
an associative and a commutative operation (comp. 9).
(1.) It has not yet been proved (comp. 195), that the Addition of
of Quaternions, 9,
But we
g',
y",
is
that the scalar of the sum of any such set of quaternions has
independent oftlieir order, and of the mode oi grouping them.
(2.) If the summands be all right quaternions (132), the scalar of each separately
wherefore the scalar of their sum vanishes also, and that
vanishes, by 196, VII.
see, already,
a value, which
is
sum
consequently itself, by 196, XIV., a right quaternion a result which it is

-^
-L- a and
-L a, because a is then
a, then y + /3
easy to verify. In fact, if j3
y
perhence, by 106, the sum of any two right quapendicular to the plane of j3 and y
is
ternions
is
a right quaternion, and therefore also the
sum
of
any number of such qua-
ternions.
(3.)
Whatever two quaternions q and
the two identities (comp. 191, (7.) )
q'
y..^{q-q) + q = q\
198.
may
be,
we have always,
as in algebra,
Yl.
(q'
+ q) -q
= q.
Without yet entering on the general theory
o^ scalars of
products or quotients of quaternions, we may observe here that because, by 196, XV., the scalar of a quaternion depends only on the
tensor
and
the angle^
and
is
independent of the axis,
to write generally (comp. 173, 178,

I.
Sqq'
= Sq'q;
II.
we
are at liberty
and 191,
(1.), (5.)),
S ^
(^': ^)
S^';
the two products, qq' and q'q, having thus always equal scalars,
although they have been seen to have unequal axes, for the general
It may also be noticed, that in vircase of diplanarity (168, 191).
tue of what was shown in 193, respecting the quotient, and in 194
Examples have already occurred
in 196, (2.), (5.), (16.).
CHAP.
I.]
SCALAR OF A PRODUCT, QUOTIENT, OR SQUARE.
185
respecting the product, of any two right quaternions (132), in con-
nexion with their indices (133),

such quaternions, the formulae
we may now
establish, for
any
III.
IV.
^q^q
= S(V:I^) = T(?':j).cosZ(Ax.5':Ax.^);
.S(^':<?)
= S
= S f !$': I i = -
(?'
Tg^V- cos L (Ax.
^':
Ax. q)\
where the new symbol Iq is used, as a temporary abridgment, to

denote the Index of the quaternion q^ supposed here (as above) to be
a right one.
With the same supposition, we have therefore also
these other and shorter formulae
V.
VI.
SU(g'':^)=+ cos z ( Ax. 5-' Ax. 5^)

^^q'q = cos Z. (Ax. q' Ax. q) ;
:
which may, by 196, XVI., be interpreted as expressing

the same condition of rectangulariiy of q and q',
VII.
VIII.
that,
under
.L{q:q)-^L (Ax. ?': Ax. ^) ;

Lq'q = 7r-L (Ax. q' Ax. q).
.
In words,
to the
the Angle of the Quotient of two Right Quaternions is equal

Angle of their Axes; but the Angle of the Product^ of two such
equal to the Supplement of the Angle of the Axes.

proving these results otherwise, by constructions such as that employed in Art. 193; nor in illustrating
quaternions,
There
is
no
is
difficulty in
them by the consideration of isosceles quadrantal
triangles,
upon the
surface of a sphere.
199. Another important case of the scalar of m product, is

the case of the scalar of the square of a quaternion. On referring to Art. 149, and to Fig. 42, we see that while we have
always T (^2) ^ (Xg)2, as in 190, and U(^^) = U(^)2, as in 161,

we have also,
I.
but,
Z(g)^
= 2z^,
by the adopted
and
Ax.
= Ax.
(^^^
definitions of
^,
if
Z^<^;
z^(130), and of Ax.g-
(127, 128),
II.
..z(/) = 2(7r-Z5),
Ax.
In each case, however, by 196,

III.
(,7^)
= - Ax.
XVI., we may
SU(^2) = cos Z
2 B
{q^)
if
g,
z^>|.
write,
= cos 2 Z
^
186
when
a formula which holds even
z.
or -, or
is 0,
[bOOK
tt,
II.
and
which gives,
IV... SU(^0 = 2(SU^)2-1.

Hence, generally, the scalar of
the two following forms
V.
S(^2)
q-
may be put under
either of
T5^cos2z
VI.
^;
=2
(Sg)^ T^^
((72)
where we see that it would not be safe to omit the parentheses,

without some convention previously made, and to write simply
this last symbol shall be
Sg^, without first deciding whether
understood to signify the scalar of the square, or i\\Q square of
the scalar of q: these two things being generally unequaL
The
them, however, occurring rather oftener than the

appears convenient to fix on it as that which is to
latter of
it
former,
be understood by S^^, while the other may occasionally be

written with a point thus, S ^^ and then, with these conven.
we may write
tions respecting notation*
VII.
= (S^)^
Sq'
But the square of the

II.) the
IX.
(1.)
g^
= S
{q%
formula:
K {q^) = {KqY = T^2
u^2.
A qnaternion, like
square roots
site
conjugate of any quaternion is easily seen

; so that we have generally
K^^ =
of the squqre
to be the conjugate
(coAp. 190,
VIII.
a positive scalar, may be said to have in general two oppobecause the squares of opposite quaternions are always equal
(comp. (3.) ). But of these two roots the principal (or simpler) one, and that which
we shall denote by the symbol Vg, or Vg, and shall call by eminence the Square Root
of 5,
is
that which has
angle acute, and not obtuse.
its
We
shall therefore write,
generally,
X.
l\'q=^ Lq;
Ax. \/q= Ax. q
As, in the Differential Calculus, it is usual to write dx^ instead of (dx)But as d^x denotes a seco7id differential,
while d(x2) is sometimes written as d.x^.
so it seems safest not to denote the square of Sq by the symbol S^q, which properly
;
or Sj, as in 196,
signifies SSg-,
(9.),
VI.
the second scalar (like the second tensor, 187,

the^r*^ Still every calculator will
or the second versor, 160) being equal to
of course use his

as cos *x
is
own
discretion
often written for (cos
and the employment
of the notation S^^ for (87)^,
xy, may sometimes cause a saving of space.
CHAP.
TENSOR AND NORM OF
I.]
A SUM.
with the reservation that, when I q = 0, or = tt, this

comes (by 131, 149) an indeterminate unit-line.
187
common
axis of q and
Vg
be-
(2.) Hence,
XI.
..SV5>0,
if
Z5'<7r;
while this scalar of the square root of a quaternion may, by VI., be thus trans-
formed
..SV9 = VU(Tg + S9)};

Lq = 7r.
XII.
a formula which holds good, even at the limit

(3.)
The
principle* (1.), that in quaternions, as in algebra, the equation,
XIII... (-9)2 = 92,

an
identity, may be illustrated by conceiving that, in Fig. 42, a point b' is determined by the equation ob' =bo for then we shall have (comp. Fig. 33, bis),
is
=
(qy
"^
^
V<^^
=
OA
because
<7^,
-^
a b'oc.
aob'
200. Another useful connexion between scalars and tensors (or

norms) of quaternions may be derived as follows. In any plane tri-
we havef
angle AOB,
the relation,
(T. ab)2= (T. oa)2 2(T. oa) (T. ob) cos aob + (T. ob)^;
which the symbols T. oa, &c., denote (by 185, 186) the lengths of
the sides oa, &c. but if we still write q = OB: oa, we have
q-\
= ab OA dividing therefore by (T oa)^, the formula becomes
(by
.
in
196, &c.),
or
II.
But q
is
change
its sign^
III.
And
.N(^-1)=%-2S^+1.
here a perfectly general quaternion;
T (1
since
and write,
+ g)^ = 1 + 2S^ +
easy to
it is
IV.
T^^
+
i)
whatever two quaternions q and
we
S ^\
+ ^) =
%-
?=?'+?,
be, while
may
q'
S q'Kq = S qKq^,
.
easily infer this other general formula,
VII.
which
gives, if
((?'
x be any
VIII.
Compare the
By
g)
=%' + 2S
qKq' f
%;
scalar,
.
first
the Second
%+
(<^
.t)
Note
to
page 162.
Book
of Euclid, or
therefore
+ 2S^ +
prove (by 106, 107) that
v...(|'
VI.
we may
2a;S^ + x\
by plane trigonometry.
%.
188
We
(1.)
[bOOK
II.
now prepared to effect, by rules* of transformation, some other powmode of expression to another, of the kind which has been alluded to,
are
sages from one
and partly exemplified,
in former sub-articles.
T^-^=l,
p~a
Take, for example, the fonnula,
of 187, (2.);
or the equivalent formula,
T(p + a) = T(p-a),
which has been
plane through
which
is
perpendicular to the line
o,
To pass now from
locus,
and therefor the same
namely the
locus as that
of 196, (1.)-
ahy
q,
and
by
we have
calculation,
only
to observe that the first or second form, as just
becomes then,
T(g +
or finally,
OA
the former equations to the latter,
to denote the quotient p:

cited,
by the equation,
represented
S- = 0,
now
of 186, (6.)
seen, on geometrical grounds, to represent a certain
by
1)
= T(5-1);
or
N(9 +
1)
=N
(9
-1)
and IV.,
II.
8^=0,
which gives the third form of equation, as required.
from S -
(2.) Conversely,
and IV.,
to the equation
= t[
\a
or to
T (p
can return, by the same general formulae
^- 1 Uni ^+ 1 \
N"[
-
= 0, we
a)
= T (p + a),
or to
or
by
- =
and
III. to
as above
^-
p-a
I.
1,
t(
II.
and gene-
rally,
Sq =
gives
T (q-l) = T(q+l),
or
T^=l;
while the latter equations, in turn, involve, as has been seen, the former.
(3.) Again, if
of the
we take
two forms 186,
the Apollonian Locus, 145, (8.), (9.), and employ thejirst
(5.) of its equation, namely,
T
where a
is
(p
- aâ) = aT (p -
T(7-a2) = aT(5-l),
or
by VIII.,
or, after
suppressing
%
which
a),
a given positive scalar different from unity, we
last is the
2a^Sq +
may
= a2;
or,
a* = a2 (Nj
Np = a2Na;
(5.),
and
any immediate appeal
write
N(g-a2) = a2N(9-
- 28^ +
1)
is
or,
it
as
1);
2a^Sq, transposing, and dividing by a^
second form 186,
lation alone, without
or as
1^
Tp = aTa;
thus deduced from the first,
bi/
calcu-
to geometry, or the construction of
any
diagram.
Compare
145, (10.)
and several subsequent
sub-articles.
CHAP.
I.]
(4.) Conversely if
we take
189
the equation,
N^ = a2,
of
U5,
(12.),
which was there seen to represent the same locus, considered as a spheric surface,
with o for centre, and aa for one of its radii, and write it as ISq = a\ we can then
by calculation return
form
to the
N(g-a2) = a2j^(5-l),
or finally,
T (9-a2) = aT(g-l),
or
T (p - a2a) = aT (p - a),
or
Tp=aTa,
as in 186, (5.);
deduced /rom the second^ namely from
this ^r^/orm of that sub-article being thus
T^=a.
a
(5.) It is far from being the intention of the foregoing remarks, to discourage
of expression, and
attention to the geometrical interpretation of the various forms
general rules of transformation, which thus offer themselves in working with quaternions ; on the contrary, one main object of the present Chapter has been to es-
and
tablish a firm geometrical basis, for all such forms
dation has once been
laid, it
is,
as
we
see,
rules.
not necessary that
But when such
we should
a.
foun-
continually
recur to the examination of it, in building up the superstructure.

That each of the
two forms, in 186, (5.), involves the other, may hQ proved, as above, by calculation ;
but it is interesting to inquire what is the meaning of this result and in seeking to
:
interpret
(6.)
we should be led anew to the theorem

The result (4.) of calculation, that
it,
N (9 - a2) = a2N {q - 1),

may
if
N^ = a2,
be expressed under the form of an identity, as follows
IX.
in
of the Apollonian Locus.
which q
N(g-%) = Ng N (^ - 1)
.
be any quaternion.
(7.) Or, by 191, VIL, because it will soon be seen that
may
q{q
we may
write
it
'^) = q'^ q,
as this other identity
X.
T (qr -
..N(^-%) = N(52-g).
then S - = - and conversely, the former equation follows

I
q
from the latter; because each may be put under the form (comp. 196, XII.),
(8.) If
1)
as in algebra,
1,
%=
(9.) Hence,
196, (6.)
),
if
T (p -
a) = Ta,
2Sg.
la
then S
and
1,
reciprocally.
each of these two equations expresses that the locus of p
which passes through
o,
and has
its
centre at
or
which has on
meter.
(10.)
In fact (comp.
By changing
if
q to 5
Tg =
1,
f-
1 in (8),
then
we
find that
0,
and
reciprocally.
is
the sphere
= 2a
for a dia-
190
Hence
(11.)
if
then
Tp=Ta,
S^
^ =
0,
[boOK
and reciprocally
II.
because (bv
^ - 106)
a.
p-^
P_
Each
(12.)
+a
of these
sphere through a, which has
nitio%
by
semicircle
its
/\
\a
two equations
(11.) expresses that the locus of
centre at
and
their proved
pis the
agreement is a recog-
quaternions, of the elementary geometrical theorem, that the angle in a

is
a right angle.
Section
13.
ternion
cation
On
Right Part (or Vector Part) of a Quathe Distributive Property of the Multipli-
the
and on
of Quaternions.
A given vector ob
can always be decomposed, in one

but in only one way, into two component vectors, of which it
is the sum (6)
and of which one^ as ob' in Fig. 50, is parallel
201.
(15) to another given vector oa, while

the other^ as ob" in the same Figure, is
perpendicular to that given line oa
fall
the perpendicu-
namely, by letting
on oa, and drawing ob" = b'b, so
that ob'bb" shall be a rectangle.
In
lar bb'
other words, if a and
j3
pj
ta
be any two given,
and co-initial vectors, it is always possible to deduce

from them, in one definite way, two other co-initial vectors,
and j3", which need not however both be actual (1); and
j3'
actual,
which
shall satisfy
(comp.
6, 15,
129) the conditions,
when /3 _L a and /3" being null, when j3 a;

]3' vanishing,
but both being (what we may call) determinate vector-funcAnd of these two functions, it is evident
tions of a and j3.
;
that
/3'
that /3"
o,
is
||
the orthographic projection of j5 on the line a ; and

corresponding j^ro/ecêow of [5 on the plane through
is the
which
202.
is
perpendicular to
Hence
it is
a.
easy to infer, that there
is
always one,
of
a
one
but only
way,
decomposing
given quaternion,
q
ifito
= 0B
0A = j3
a,
two parts or summands (195), of which one shall be, as in
CHAP.
UIGIITPART OF A QUATERNION.
I.]
191
196, a scalar^ while the other shall be a right quotient (132).

parts, ihQ former has been already called (196)
Of these two
the scalar part, or simply the Scalar of the Quaternion, and

has been denoted by the symbol Sq so that, with reference
;
we have
to the recent Figure 50,

I.
S3'
= S(oB
oa)=
OA
ob':
or,
(j3
a)
j3': a.
And we now propose to call the Z^êr part the Eight Part*
of the same quaternion, and to denote it by the new symbol
writing thus, in connexion with the same Figure,
II.
.Yq = V(oB
oa)
= ob"
oa
Y([5 a) =
or,
jS"
a.
The System of
Notations, peculiar to the present Calculus,

will thus have been completed
and we shall have the follow;
ing general Formula of Decomposition of a Quaternion into two

Summands (comp. 188), of the Scalar and Right kinds
:
III..
or, briefly
.^=s^
v^ y^
and symbolically,
IV.. 1 = S +
.
(1.) In connexion with the
same
+ s^,
V = V+S.
Fig. 50,
we may
write also,
V (ob: oa) = b'b: OA,

because,
by
construction, b'b
C2.) In like
= ob".
manner, for Fig. 36, we have the equation,
Y (oB
(3.)
Under the
V(/3':a)
(4.) In general,
V.
(5.)
More
= a'b
oa)
OA.
recent conditions,
g=
0,
and
S(,3":a)
and
Vj =
= 0.
evident that
it is
if
0,
Sg'
0,
and
reciprocally.
generally,
VI. ..^''=5',
if
= Sg',
Sg'
VII...V9 = 0,
(6.) Also
YIII.
or
and
V5r'
= V5'
Zg = 0,
V(/3:a) = 0, if
if
with the converse.
or
/3
||
=-7r;
the right part of a scalar being zero.
*
the
tify
This Right Part, Yq, will come to be also called the Vector Part, or
simply
of the Quaternion
because it will be found possible and useful to iden-
Vectob,
such part with
136, 174.
its
own Index- Vector
(133).
Compare
the Notes to pages 119,
192
On
(7.)
II.
may now
be
the other hand,
lX...Yq = q,
a right quaternion being
203.
[bOOK
We had
its
oum
(196,
if
lq='î
right part.
XIX.)
a formula which
written thus,
I.
ob'= S(oB
or
oa). OA,
j3'
= Si-a,
a
to express the projection of ob on oa, or of the vector /3 on a ;

and we have evidently, by the definition of the new symbol
Yq, the analogous formula,

II.
.ob" = Y(ob:oa).oa,
or
/3"
V^.a,
a
to express the projection of (5 on the plane (through o), which

drawn so as to be perpendicular to a ; and which has been
is
considered in several former sub-articles (comp. 186,

It follows (by 186, &c.) that
196, (1.) ).
III.
Tf^"
= TY'Ta= perpendicular
distance
(6.),
and
of b from oa;
being here considered with reference to its

characteristic T of the tensor implies.
the
It
length alone, as
this perpendicular
is
V a
to be observed that because thefactor^
in
the recent
II. for the projection /3", is not a scalar, we must write

that factor as a multiplier, and iiot as a multiplicand ; although
formula
we were
at liberty, in consequence of a general convention

(15), respecting the multiplication of vectors and scalars, to
denote the other projection j3' under the form,
r. ..i3'
(1.)
The
= aS2(196,XIX.).
equation,
v=o.
(2.)
The
is
the indefinite right line oa.
equation,
v^::^=o,
a
or
ve = v^,
a
CHAP.
drawn through the point

(3.) The equation
is
the indefinite right line bb", in Fig. 50, which
S^ = S^,
or
0,
OA
it
we combine
ofl96, (2.),
has been seen to express that the locus of

that the point
is
b, parallel to the line oa.
=
S^^
a
to the line
1^
I.]
is
the plane through b, perpendicular
with the recent equation (2.), we shall express

two last mentioned loci ; or that
if
then
is
situated at the intersection of the
it
coincides with the point B.
we take
(4.) Accordingly, whether
forms
(2.), (3.),
stZ^ = o,
ve^=0,
a
we can
v.,
0,
whence
jo
The
two
last of these recent
142
p=
and
we have
so that
se = S^,
in the first case
consequence from each system, the equality p
(5.)
or the
- /3= 0, by
through 202, VI., that
the coincidence,
first
V^ = V^,
or
infer this position of the point
p-3- =
that -
the two
namely,
by
inferring,
in the second case
through 202,
by
inferring,
in each case (comp. 104), or as a
= /3,
or
op = ob
or finally (comp. 20)
b.
^^ p ^
equation,
expresses that the locus of the point
^^
a
the cylindric surface of revolution, which

passes through the point b, and has the line oa for its axis ; for it expresses, by III.,
that the perpendicular distances of p and B, from this latter line, are equal.
(6.)
The system
is
TV^ = TV^,
a
made by
S^ = 0,
y
the (generally) elliptic section of the cylinder

(5.),
the plane through o, which is perpendicular to the line oc.
(7.) If
we employ an
analogous decomposition of p, by supposing that
= P'+P">
the three rectilinear or plane

briefly written
is
loci,
p"-^,
P'll>
(1.),
(2.), (3.),
may have
their equations thus
p"
= 0;
p"
= /3";
p'
= ^':
while the combination of the two last of these gives p = /3, as in (4.).
(8.) The equation of the cylindric locus, (5.), takes at the same time the form,
Tp" = T/3";
which
last equation expresses that the projection
p" of the point p, on the plane through
o perpendicular to oa, falls somewhere on the circumference of a circle, with o for

and this circle may "accordingly be considered as the base
centre, and ob" for radius
:
of the right
204.
cylinder, in the sub-article last cited.
From
the mere circumstance that
right quotient (132),
whenceUV*/
2 c
is
Yq
is
always a
a right versor (153), of
194
which the plane
[boOK
II.
and the axis (127), coincide with those

we have
(1 19),
of q^ several general consequences easily follow. Thus
I.
principles ah'eady established, the relations
by
generally,
..zV^ = ^;
III.
IV.
11.
Ax.V^=Ax.UV^ = Ax.y;
KVq = -Yq,
..SV^
V.
KV = - V (144)
or
or
0,
SV=0(196,
VII.);
.(UV^)2 = -1 (153,159);
(V^)^ = (TV^)^ =
and therefore,
VI.
N V^,*
because, by the general decomposition (188) of a quaternion

into ^côr^, we have
VII.
We have also
.
and
V^ = TV^ UV^.
.
(comp. 196, VI.),

= 0, or
VIII.
IX.
VS^
VV^ = V^,
X.
..VK^
VS =
(202, VII.)
V V = V (202,
VK = -V,
or
V^ =
or
= -V^,
IX.)
because conjugate quaternions have opposite right parts, by the

definitions in 137, 202, and by the construction of Fig. 36.
For the same
we have
reason,
XL
K^
this other general formula,
or
S^-V^,
K = S-V;
but we had
y
= S^ + V^,
or
1=S + V,
hence not only, by addition,

or
1 +
^ + K^=2S^,
but also, by subtraction,
whence the
Characteristic,
III.,
K = 2S, asin
..^-K^ = 2V^,
XII.
by 202,
or
196,
IV.
I.,
1-K = 2V;
V, of the Operation of taking
the
RightPartofa Quaternion
196),
may
equations
XIIL
be defned hj
(comip. 132, (6.); 137; 156; 187;

either of the two following symbolical
V=
XIV
- S
(202, IV.);
whereof the former connects

*
it
V = i(l -K);
with the characteristic
Compare the Note
to
page 130.
S,
and
CHAP.
PROPERTIES OF THE RIGHT (OR VECTOR) PART.
I.]
the latter with the characteristic
K on S and V
195
while the dependence of
expressed by the recent formula XI. ; and

that of S on
by 196, 11'. Again, if the line ob, in Fig. 50,
be multiplied (15) by any scalar coefficient, the perpendicular
bb' is evidently multiplied by the same hence, generally,
is
XV.
Yxq
xYg,
if
x be any scalar
and therefore, by 188, 191,
XVI.
V^ = T^ VU^,
and XVII.
But the consideration of the right-angled

same Figure, shows that
XVIII.
because,
by 202,
TV^ = Tq .TVU^.
triangle, ob'b, in the
II.,
.TYq^Tq.sinLq,
we have
.
TV^ = T(ob":oa) = T.ob":T.oa,

and
T.ob"= T.ob
we
sin
aob
arrive then thus at the following general
and useful
mula, connecting quaternions with trigonometry anew

XIX. .TVU^ = sinZy;
for-
by combining which with the formula,
SUg = co3Zg(196, XVI.),

we
arrive at the general relation
XX.
(SU^)2 + (TVU|7)2 = I;
..
which may also (by XVII., and by 196, IX.) be written thus
XXI. ..(S^)^+(TV^)^=(T^)'-;
and might have been immediately deduced, without sines and
cosines, from the right-angled triangle, by the property of the
square of the hypotenuse, under the form,
+ (T b'b)2 = (T ob)2.
The same important relation may be expressed in various other
(T
ways
ob')2
we may write,
XXII.
^q = Tq^=^q^-Yq\
for example,
where
it is
assumed,
as
an abridgment oi notation (comp. 199,
VII., VIII.), that
XXIII.
V^'-
= (Yq)\
but that
XXIV.
V.
j^
Y{q%
V96
[bOOK
II.
the import of this last symbol remaining to be examined.

And because, by the definition of a norm, and by the properties of S^ and Yq^
XXV.
we may
NS^ =
XXVL
but
S^^
NV^ = - Yq\
write also,
XXVII.
a result which
% = N(S^ + V^)
NS^ + NV^
indeed included in the formula 200, VIII.,
is
since that equation gives, generally,
XXVIII.
.N(^ + a;) =
X being, as usual, any scalar.

(by 106, 143)
we
%
It
+ Naj,
may
^^ = ^;
if
be added that because
have, as in algebra, the identity,
= -^'-^,
XXIX. ..-(^'+^)
the opposite of the sum of any two quaternions being thus equal
to the sum of the opposites, we may (by XL) establish this
other general formula
XXX. ..-K^ = V^-S^;

the opposite of the conjugate of any quaternion q having thus
the same riff ht part as that quaternion, but an opposite scalar
part,
(1.)
From
the last formula

if
and
q=-Kg,
it
may
then
be inferred, that
Yq' = + Yq,
but
Sq' =
-Sg]
therefore that
and
Tq'=Tq^
Ax. 5'= Ax.
9,
but
Lq'
= Tr- Lq\
might have been deduced from 138 and 143, without the
introduction of the characteristics S and V.
which two
(2.)
last relations
The
equation,
(v^Y=fv^Y,
a ]
\
a]
or (by
XXVL),
NV^ = NV^,
like the equation of 203, (5.), expresses that the locus of
is
the right cylinder, or
the point B.
cylinder of revolution, with oa for its axis, which passes through
(3.) The system of the two equations,
(^^T=(-fI'
-;-
system in 203, (6.), represents generally an elliptic section of

but if it happen that y a, the section then becomes cir-^
the same right cyhnder
like the corresponding
cular.
||
CHAP.
I.]
The system of the two
(4.)
s^
x^yTa
(1
equations,
x>-l,
with
fv-]=a;2-l,
represents the circle,* in
197
x<I,
which the cylinder of revolution, with oa
for radius, is perpendicularly cut
by a plane
for axis,
at a distance
and with
= + xTa from
the vector of the centre of this circular section being xa.
While the scalar x increases (algebraically) from 1 to 0, and thence to

to 1, and then decreases
the connected scalar V(l ar^) at first increases from
from 1 to ; the radius of the circle (4.) at the same time enlarging from zero to a
(5.)
1,
maximum = Ta,
and then again diminishing
to zero
while the position of the centre
of the circle varies continuously, in one constant direction, from ajirst limit-point a',
if
oa' =
a,
to the point A, as a second limit.
The locus of all such circles is the sphere, with aa' for a diameter, and therewith o for centre namely, the sphere which has ali-eady been represented by the
(6.)
fore
equation
Tp = Ta
of 186, (2.)
or
by
T^ = 1,
a
S^^^ = 0,
p
but which
now
+a
presents itself under the
of 187, (1.)
or
by
of200, (11.);
new form,
[îh(^^h'
obtained
by eliminating x between the two
(7.) It
is
recent equations (4).
easy, however, to return from the last
to the first, or to the third,
by
neral relations between the symbols used.

written,
form to the second, and thence
rules of calculation already established, or
In
fact, the last
by the ge-
equation (6.)
may
be
by XXII., under the form,
Ne=ii
a
whence
T^=l,
a
and therefore
also
Tp = Ta, by
187, and
by 190, VI.;
S
^^ =
p
-\-
0,
by 200,
(11.).
a.
through a, with o for centre, might already have

definition and property of a norm, stated in
145, (11.), to ad-
(8.) Conversely, the sphere
been seen, by the
first
mit (comp. 145, (12.) ) of being represented by the equation

fore,
by XXII., under the
variable scalar
those equations
S -,
:
recent form (6.)
as in the
first
of the
and thus might be
in which
two equations
if
N-=1
a
we
(4.),
we
write
and thereto denote the
recover the second of
led to consider, as in
(6.), the sphere in question
the word " circle," in these pages, is usually meant a

circumference, and
"
"
"
; and in like manner, the words
sphere,"
cylinder,"
cone," &c., are
usually here employed to denote surfaces, and not volumes.
By
not an area
198
[bOOK
II.
as the locus of a variable circle, which
is
(as above) the intersection of a variable
plane perpendicular to its axis.
(9.) The same sphere may also, by XXVII., have its equation written thus,
cylinder, with a variable
= l;
Nfs^ Ve^
(10.)
If,
or
Tfs^ V^]
= l.
in each variable plane represented^by the first equation (4.), we conceive

circle, or that of the variable cylinder, to be multiplied by any con-
the radius of the
stant and positive scalar a, the centre of the circle and the axis of the cylinder re -
maining unchanged, we
new system
shall pass thus to
a new system of circles, represented by
this
of equations,
(v^Y = ^^-l.
8^ = 0.,
(11.) The locus of these new circles will evidently be a Spheroid of Revolution ;
the centre of this new surface being the centre o, and the axis of the same surface
being the diameter aa', of the sphere lately considered which sphere is therefore
either inscribed or circumscribed to the spheroid, according as the constant a > or
:
<
because the radii of the
cond case
of the spheroid
(12.)
new
circles are in the first case greater,
than the radii of the old
less,
circles
respectively the sphere
= Ta.
of the two co-axal cylinders of revolution,
which envelope
and spheroid (or are circumscribed thereto) are
v5Y=-i,
aa
NV^
NV^=1,
TV^ = a.
TV^=1,
a
The system
but in the se-
or because the radius of the equator
while the radius of the sphere
=aTa,
The equations
(13.)
of the tAvo equations,
^t=x,
(v-^|=a;2-l,
represents (comp. (3.) ) a variable ellipse,
if
with
the scalar
/3
M0<
II
x be
a,
still
treated as a va-
riable.
The
(14.)
this
new
result of the elimination of
x between the two
equation,
or
NS^ + NV^=1,
a
by XXV., XXVI.
or
Nfs^ + vM=l,
by XXVII.;
or finally,
t[s^ v|'J=i,
byl90,VL,
last equations,
namely
CHAP.
I.]
QUATERNION EQUATION OF THE ELLIPSOID.
199
of all such ellipses (13.), and will be found to be an adequate

representation, through quaternions, of the general Ellipsoid (-with three unequal
axes) that celebrated surface being here referred to its centre, as the origin o of
represents the locvs
vectors to
its
points
and the
six scalar (or algebraic) constants,
which enter into
the usual algebraic equation (by co-ordinates) of such a central ellipsoid, being here
virtually included in the two independent vectors, a and (3, which may be called its
two Vector- Constants*
(15.)
The equation (comp.
(12.)),
or
(ST-.
TV^
NV|=1,
= 1,
represents a cylinder of revolution, circumscribed to the ellipsoid,
along the ellipse which answers to the value a? = 0, in (13.)

this ellipse of contact is represented by the equation,
and touching
it
so that the plane of
a
the normal to this plane being thus (comp. 196, (17.) ) the vector a, or
the axis of the lately mentioned enveloping cylinder is (3, or on,
(16.) Postponing
ellipsoid (14.), it
oa
while
any further discussion of the recent quaternion equation of the

be noted here that we have generally, by XXII., the two fol-
may
lowing useful transformations
Yq, of any quaternion q
XXXI.
for the squares, of the scalar Sq,
and of the right part
Sg2
= T52
f V52
XXXII.
V52 = S53
_ TqK
(17.) In referring briefly to these, and to the connected formula XXII., upon
occasion, it may be somewhat safer to write,
(S)2
+ (Y)2,
(T)2
than S2t=T2 + V^,
(V)2
= (S)2 -
(T)2,
(T)2
= (S)2-(V)2,
because these last forms of notation, S2, &c., have been

otherwise interpreted already, in analogy to the known Functional Notation, or Notation of the Calctdus of Functions, or of Operations (comp. 187,
(9.); 196, VI.
-&c.
and 204, IX.).

(18.) In pursuance of the same analogy, any scalar
may be denoted by the gene-
ral symbol,
V-'O;
because scalars are the only quaternions of which the right parts vanish.
(19.) In like manner, a right quaternion, generally, may be denoted by the
sym-
bol,
S-i
and since
this includes
(comp. 204, I.) the right part of any quaternion,

establish this general symbolic transformation of a Quaternion :
we may
^=v-io + s-io.
(20.)
With
this
form of notation, we should have generally, at
least for realf
quaternions, the inequalities.
It will be found,
ellipsoid,
may
however, that other pairs of vector-constants, for the central
occasionally be used with advantage.
t Compare Art. 149; and the Notes
to
pages 90, 134.
200
[bOOK
II.
(S-'0)2<0;
(V-i0)2>0;
so that a (geometrically real) Quaternion
is
generally of the form
Square-root of a Positive, plus Square-root of a Negative.
The equations 196, XVI. and 204, XIX.
(21.)
ternions
and
trigonometry/, the formula
XXXIII.
= TVUg SU^ = TYq
tan Z g
new link between qua-
give, as a
S^.
not be entirely in accordance with the theory of that Functional

(or Operational) Notation, to which allusion has lately been made, but it will be
found to be convenient in practice, to write this last result under one or other of the
(22.) It
may
abridged forms :*
XXXIV.
TV
tan Z 9
= -
XXXIV.
or
tan Z g
= (TV
S) q ;
which have the advantage oi saving the
when
repetition of the symbol of the quaternion ,

that symbol happens to be a complex expression, and not, as here, a single let-
ter, q.
The transformation
(23.)
by
II., for
any quaternion
XXXV.
so that
we may
q,
194, for the index of a right quotient, gives generally,
the formulae
IVg = TV^ Ax. g

.
(24.)
lUVg = Ax.
lUV = Ax.
- Ax. \q, by 135, and therefore = Ax. q,
because Ax. (1 V^) =

write also, by XXXV.,
And
we may
XXXVI.
establish generally the symbolicalf equation,
XXXVI'.
II.,
XXXV.
I (1
V^)
= - Ax.
by
TYq.
205. If any parallelogram obdc (comp. 197) be projected

on the plane through o, which is perpendicular to oa, the projected figure ob"d"c" (comp. 11) is
that
od" = oc" + ob" (6), or
still
S"
a parallelogram; so
= y" +
/3"
and therefore, by 106,
g":a=(7":a)+(j3":a).
Hence, by 120, 202, for any two quaternions^ q and q\ we ha\e
I...
V(^'+^) = V^'+V^;

t At a later stage it will be found
possible (comp. the
to write, generally,
IUV9 = UV9;
YVq^Yq,
and then (comp. the Note
XXXVr.,
in
Note to page 174, &c.),
page 118 to Art. 129) the recent equations,
will take these shorter forms
Ax. q =
VYq
Ax. = UV.
XXXVI.,
CHAP.
RIGHT PART OF A SUM OR DIFFERENCE.
I.]
with which
it is
easy to connect this other,
..Y(q'-q) = Yq'-Yq.
11.
Hence
also, for
Y[q"-v{q' +
201
any three quaternions,
q\
q, q,
q)}^Yq"+Y{q+q)=Yq"^{Yq' + Yq)^
for any greater number of sumraands

write generally (comp. 197, II.),
and similarly
we may
III.
Y^q = S V(7,
or briefly
III'.
so that
VS = SY
while the formula II. (comp. 197, IV.) may, in like manner,
be thus written,
IV. ..VA^ = AV^,
IV.
or
..VA = AV;
the order of the terms added, and the mode o^ grouping them,
in III., being as yet supposed to remain unaltered, although
both those restrictions will soon be removed.
We conclude
then, that the charactetnstic V, of the operation oi taking the

right part (202, 204) of a quaternion, like the characteristic S
o^ taking the scalar (196, 197), and the characteristic

of
is
a
Distributive
or
the
conjugate (137, 195*),
taking
Symbol,
represents a distributive operation: whereas the characteristics. Ax., Z, N, U, T, of the operations of taking respectively
theaa;w(128, 129), the7?^/e(130), thenorm{\A5, (11.) ), the

versor (156), and the tensor (187), are not thus distributive
symbols (comp. 186, (10.), and 200, VII.)
upon a whole
(or sum),
by operating on
or do not operate
its
parts (or sum-
mands),
(1.)
We may now recover the symbolical
(comp. 196, VI.
V.
(2.) In like
other,
202, IV.
equation
.K2 = (S-V)2 =
VI.
S2
V2
1 (145),
under the form
S2-SV-VS + V2 = S + V=1.
manner we can recover each
of the expressions for S^,
under the forms (comp. again 202, IV.)
VII.
K^ =
and 204, IV. VIII. IX. XI.):
V^ from the
= (1-V)2 = 1-2V + V2 = 1-V = S, as in 196, VI.;

= (1-S)3=1-2S + S2 = 1-S = V, asia204, IX.;
or thus (comp. 196, II'.,
and 204, XIV.), from the expressions
for
S and V in terms
ofK:
Indeed,
for the case of
it
has only been proved as yet (comp. 195, (1.)), that

summands but this result will soon be extended.
two
2 D
KEq-'EKq,
202
VIII.
IX.
[bOOK
II.
.S2=:i(l+K)' = i(l + 2K + K2) = i(l + K) = S;

V2 = 4(l-K)2 = i(l-2K + K2) = |(l-K) = V.
(3.) Similarly,
X..
XL
and
206.
SV = i(l + K)(l-K) = i(l-K2)=0, asin204, IV.;

YS = i(l-K) (1 + K) = i(l-K2) = 0, as in 204, VIII.
As
regards the addition (or subtraction) of such right

parts, Yq, V^-', or generally of any two right quaternions
(132), we may connect it with the addition (or subtraction) of
Let obdc be again any paraloa be now an unit-vector (129)
their indices (133), as follows.
lelogram (197, 205), but

perpendicular to
its
plane
let
;
so that
Z(/3:a) = /(y:a) = Z(g:a)=^,
Ta=l,
8=7 +
/3.
Let ob'd'c' be another parallelogram in the same plane, obtained by a positive rotation of the former, through a right
angle, round oa as an axis; so that
Ax.
(j3'
P)
= Ax,
(y 7) = Ax.
:
Then
(g'
g)
a.
the three right quotients, /3 a, 7 a, and S a, may represent any two right quaternions, q^ q, and their sum, q + q,
Avhich is always (by 197, (2.) ) itself di right quaternion; and
:
the indices of these three right quotients are (comp. 133, 193)
so that we may write, under the fore/3', 7', S',
the three lines
going conditions of construction,
^'=I(i3:a),
7=I(7:a),
g'
= I(g:a).
But this third index is (by the second parallelogram) the sum
of the two former indices, or in symbols, ^' = 7' + /3' ; we may
therefore write,
I.
(^'
+ ^) = Ig + I^,
if
Z ^= z
Sum* of any two Right QuaterSum of their J?idices, Hence, generally,

any two quaternions, q and q', we have the formula,
IL..IV(^' + ^) = IY^'+IV^,
or in words the Index of the

nions
for
^'
is
equal to the
Compare
CHAP.
I.]
GENERAL ADDITION OF QUATERNIONS.
203
because V^', V^-' are always right quotients (202, 204), and
{q + q) is always their sum (205, I.) ; so that the index of

the right part of the sum of any two quaternions is the sum of
the indices of the right parts. In like manner, there is no difficulty in proving that
m...l{q-q)^lq-lq,
if
Z^'=Z^ = ^;
and generally, that

lY. ,.lY{q-q) = lY(^-lYq',
the Index of the Difference of any two right quotients, or of
the right parts of any two quaternions, being thus equal to the
Difference of the Lidices*
may then reduce the addition
We
or subtraction of
any two such quotients, or parts,
tion or subtraction of their indices
to the addi-
a right quaternion being

when
its index is given, or
determined,
always (by 133)
;
known.
We
207.
see, then, that as
the Multiplication of any
Q_uaternions was
(in 191) reduced to (1st) the aynthmetical

of
operation
multiplying their tensors^ and (Ilnd) the geometrittoo
cal operation of multiplying their versors^ which latter was constructed by a certain composition of rotations, and was represented (in either of two distinct but connected ways, 167, 175)
by sides or angles of a spherical triangle: so the Addition of
any two Quaternions maybe reduced (by 197, 1., and 206, II.JI
to, 1st, the algebraical addition oftheir scalar parts considered
as two positive or negative numbers
(16) and, Ilnd, the geo^
metrical addition of the indices of their right parts, considered

as certain vectors (1)
this latter Addition of Lines h^mg performed according to the Rule of the Parallelogram (6.).t In
:
Compare again the Note
It does not fall
the subject
but
to
page 174.
within the plan of these Notes to allude often to the history of

ought to be distinctly stated that this celebrated Rule, for what
it
be called Geometrical Addition of right lines, considered as

analogous to composition of motions
(or of forces), had occurred to several writers, before the invention
of the quaternions
although the method adopted, in the present and in a former
may
work, of deducing that rule, by algebraical analogies, from the symbol

for the line ab,
may possibly not have been anticipated. The reader
pare the Notes to the Preface to the author's
(Dublin, 1853).
Volume
b-a
(1)
may com-
of Lectures on Quaternions
204
[bOOK
II.
manner, as the general Division of Quaternions was seen (in

191) to admit of being reduced to an arithmetical division of
like
geometrical division ofversors, so we may now

(by 197, III., and 206, lY.) reduce, generally, the Suhtrac-
tensors,
tion
larsy
and
?i>
of Quaternions to (1st) an algebraical subtraction ofscaand (Ilnd) geometrical subtraction of vectors : this last
?i,
operation being again constructed by a parallelogram, or even

by a plane triangle (comp. Art. 4, and Fig. 2). And because
the sum of any given set of vectors was early seen to have a
independent of their order, and of the mode

of grouping them, we may now infer that the Sum of any number ofgiven Quaternions has, in like manner, a Value (comp.
value (9), which
is
is independent
of the Order, and of the
Grouping of the Summands: or in other words, that the general
A ddition of Quaternions is a Commutative^ and an Associative
197 J (l.))> which
Operation.
(1.)
The formula,
V2^=2Vg,
of 205, III.,
now
seen to hold good, for any number of quaternions, independently of the arrangement of the terms in each of the two sums, and of the manner in which they may be
is
associated.
(2.)
We can infer anew
K {q
that
+ g-) = Kg' +
under the form of the equation or
S (g' + g) (3.)
More
generally,
it
Kg, as in 195,
V (g' + g) = (Sg' - Vg') +
may
II.,
identity,
(Sg
- Vg).
be proved, in the same way, that
K2g = 2Kg,
or briefly,
whatever the number of the summands
may
KS = SK,
be.
208. As regards the quotient or product of the right parts, Yq and

Yq', of any two quaternions, let t and f denote the tensors of those
two parts, and let x denote the angle of their indices, or of their axes,
or the mutual inclination of the axes, or of the planes,]^ of the two
quaternions q and q' themselves^ so that (by 204, XVIII.),
*
t Two planes, of course, make with each
other, in general, two unequal and supplementary angles ; but we here suppose that these are mutually distinguished, by
taking account of the aspect of each plane, as distinguished from the opposite aspect :
which
is
most easily done (lll.)> by considering the axes as above.
CHAP.
QUOTIENT OR PRODUCT OF RIGHT PARTS.
I.]
= TV(? = T^
sin
t'
Z?,
= TV2'= T^'
sin
205
Z$^
and
x = L {lYq' IV?) = L (Ax. q' Ax. q).

Then, by 193, 194, and by 204, XXXV., XXXV'.,
I.
V?^V2 = IV2':IV? = + (TV2':TV?).(Ax.5':Ax.^);
:
11.
V^'. Yq
= IV^'
^=
(T^?' TV?) (Ax.
?'
Ax.
?)
and therefore (comp. 198), with the temporary abridgments proposed above,
S {Yq' Yq) = ft' cos x ;

IV.
S (V?'. Yq)=-t't cos ;r;
VI.
VII. ,,L(Yq':Yq) = X',
VIII.
III.
V.
We have also generally (comp.

IX.
TV {Yq' Yq) =
sin
XL TV {Yq' Yq) = Vt sin
.
and
t't-'
a;
SU (Yq' V?) = + cos x

SU (Yq' V?) = - cos a;;
:
(V?^ V?) =
zl
tt
- a;.
204, XVIII., XIX.),
X.
a; ;
XII.
T VU {Yq' Yq) = sin a;
TVU {Yq'
Yq) = sin a:
in particular,
XIII.
V {Yq'
Yq) = 0,
and
if?'
III?
XIV.
V?^ V?) = 0,
(123);
because (comp. 191, (6.), and 204, VI.) the quotient or product of
the right parts of two complanar quaternions (supposed here to be
both non-scalar ( 1 08), so that t andit' are each > 0) degenerates (131)
into a scalar, which
XV.
may be
V?': V? = + i'rS
thus expressed
and
XVI.
Yq'.Yq = -t%
if a;=
V?'.V? = +
ifa;
0;
but
XVII.
the
V?':V? = -'r^ and XVIII.
..
first
case being that of coincident,
..
i'^,
= ^;
and the second case that of
In the more general case oi diplanarity (119), if we

the unit-line which is perpendicular to hath their axes,
opposite axes.
denote by
and therefore common

intersect,
and which
Ax. ?
Ax.
II.
to
to their
two planes, or in which those planes
so directed that the rotation
is
q' is positive
round
it
from
(comp. 127, 128), the recent formula)
I.,
give easily,
XIX.
Ax. (V?': V?) =+
XX.
5;
Ax. (V?'. V?) =- ^;
and therefore (by IX., XI., and by 204, XXXV.), the indices of the
and product of the right parts of any two
right parts, of the quotient
may be expressed
diplanar quaternions,
XXI.
XXII.
as follows:
I V ( V?' V?) = + a ft-' sin x

.lV(V?'.V?) = -5.nsina;.
:
206
[bOOK
II.
ABC be any triangle upon the unit-sphere (128), of which the spheriand the corners may be denoted by the same letters A, b, c, while the sides
as usual be denoted by a^h^ c\ and let it be supposed that the rotation
(comp.
(1.) Let
cal angles
shall
177) round
A from c
Then
as in Fig. 43.
we
and therefore that round b from A
to b,
we have
writing, as
(3:
and
a,
q'
=y
'
to c, &c., is positive,
often done,
where
ft,
= oa,
&c.,
easily obtain the the following expressions for the three scalars
the vector d
= smc;
we have
(2.) In fact
x,
t',
and
for
= sin o
*'
and
a;
= tt b
= - ft.
here,
Tq = Tq' =
whence
t,
Lq = c,
l,
Also
are as just stated.
= a;
Lq'
be (as in 175) the positive poles

of the three successive sides bc, ca, ab, of the given triangle, and therefore the points
A, b,
t'
If a', b', c'
c the negative poles (comp. 180, (2.)
Ax. q =
the
) of
Ax.
OC',
q'
new
= oa'
arcs b'c', c'a', a'b', then

;
but X and d are the angle and the axis of the quotient of these two axes, or of the
therefore x is, as above
quaternion which is represented (162) by the arc c'a'
;
stated, the supplemejit of the angle b,
which
and S
directed to the point
is
upon the
sphere,
point b.
(3.) Hence, by III. V. Vil. VIII. IX. XI., for any triangle abc on the unitsphere, with a = OA, &c., we have the formulae:
is
diametrically opposite to
XXIII.
V ^ V - = - sin a cosec c cos b
sf
XXIV.
i\i&
ft
= + sinasinccosB;
V^.V-
s(
aj
XXVI....(vl.V^)
XXV....(vI:vf]=.-B;
XXVn. Tvf V^:v5 = + sinacoseccsinB;
.
XXVIIL
= B,
ft
V^.V^ = + sina sine sin B.
TV(
ft
by XIX. XX. XXI. XXIL,
(4.) Also,
still
if
the rotation round b from
to c be
positive,
XXIX..
Ax.fv^:V^'j
XXXL.
XXXIL
(5.) If,
IVI
IV(
= -/3;
XXX..
Ax.
moment
ai
= a,
v|. V^^
+ ZB;
=-/3sinacoseccsinB
V^:V^
V^.V- = -|-/3sinasincsinB.
)
on the other hand, the rotation round b from
writing for a
fti
= ft,
yi
= y, we
to c were negative, then
should have a
new and oppo-
which the rotation round Bi from Ai to Ci would be positive,

but the angle at bi equal in magnitude to that at b so that by treating (as usual)
all the angles of a spherical triangle as positive, we should have bj = b, as well as
site triangle, AiBiCi, in
ci
= c,
and ai
=a
and
therefore, for example,
by XXXI.
CHAr.
COLLINEAR QUATERNIONS.
I.J
IV V P"
Pi
V
or
IV
\
V
ai
= - /3i
207
sin ai cosec ci sin bi,
= + /3sinacoseccsinB:
V^:V-l
a)
P
the four formulae of (4.) would therefore

direction of rotation in the given triangle,
still
subsist,
we were
to
new
provided that, for this
change the sign of (3, in the
second member of each.
lYq: Sq to (IV: 8)5-, as TV5: Sq was abridged, in

(TV: S)g, we have by (5.), and by XXIV., XXXII., this other
(6.) Abridging, generally
204,
XXXIV'.,
to
general formula, for any three unit- vectors a,

at the corners of a spherical triangle abc
(3,
considered
y,
still
as terminating
XXXIII.
(IV:S)[ V^.V^'j
= i/5tan
the upper or the lower sign being taken, according as the rotation round b from
c,
or that
and which
round
(3
from a to
7,
which might perhaps be denoted by the symbol
to
ajSy,
in quantity is equal to the spherical angle b, is positive or negative.
When
the planes of any three quaternions q^ q\ q'\ consipassing through the origin
(119), contain any coinmon
line, those three may then be said to be Collinear^" Quaternions ; and
because the axis of each is then perpendicular to that line, it follows
209.
dered as
all
that the Axes of Collinear Quaternions are Complanar : while conversely, the complanarity of the axes insures the colUnearity of the
quaternions, because the peipendicvlar to the plane of the axes
to the planes of the quaternions.
is
a line
common
collinear ; but the converse proposition

(1.) Complanar quaternions are always
does not hold good, collinear quaternions being not necessarily complanar.
(2.) Collinear quaternions, considered as fractions (101), can always be reduced
to a
common denominator (120)
so reduced, as to appear
those quaternions
and conversely,
if
three or
more quaternions can be
common denominator e,
then common to all their
under the form of fractions with a
must be
collinear
because the line
c is
planes.
(3.)
Anj/ two quaternions are collinear with any scalar
the plane of a scalar
being indeterminate^ (I'^l)(4.) Hence the scalar and right parts, S5, Sg', V^, V5', of any two quaternions,
are always collinear with each other.
(5.)
The conjugates
of collinear quaternions are themselves collinear.
Quaternions of which the planes are parallel to any

Compare the first Note to page 113.
said to be collinear.
t Compare the Note
to
page 114.
common
line
may
also be
208
210. Let
common
(comp. 120) three other lines
/3,
7, 6,
a'
and thus may conclude that
such that
^
a'
^'
(as in algebra),
l..,{q' + q)q'' =
by
II.
q" be any three collinear quaternions; and let a

Then we may determine
to their planes.
q, q',
denote a line
because,
[bOOK
qY^qq".
106, 107,
In like manner, at
may be proved
J8
a 8
a 8
under the same condition of collinearity,*
least
it
that
II.
{q'-q)q''
Operating by the characteristic

attending to 192, II.,
III.
IV.
II.,
upon these two
we
equations, and
find that
Kq" .{Kq' +^q) = Kq" .Kq' + Yiq'' .Kq-,

^q" .{Kq' -Kq)^Yiq" ..q' -Yiq".l^q\
and 195,
= q'q''-qq'\
where (by 209, (5.) ) the three conjugates of arbitrary collinears,

K5, Yiq\ Kg-'', may represent any three collinear quaternions. We
have, therefore, with the same degree of generality as before,
V.
q"
{q'
g)
= q"q' + q"q
-q)= q'^q' - q'^q.

any four collinear quaternions^ we may estaVI.
q"
{q'
If, then, 5-, q', q'\ q'" be

blish the formula (again agreeing with algebra):
VII.
{q"'
+ q")
(?'
5)
= ql"q' + q^'q; +
q;"q + q^'q ;
and similarly for any greater number, so that we may write

VIII.
.. 'S.q^
briefly,
.^q^^q'q,
where
2g' = ^1 +
?2
+ ?m,
2^^ = ^\ +
^'2
+ q'n,
and
^q'q = q\q, +
q/q^^ \-q'2qi
of Collinear] Quaternions
+ q',,q^,
In words (comp. 13),
and n being any positive whole numbers.
the Multiplication
is
a Doubly Distributive
Operation.
* It will soon be
seen, however, that this condition is unnecessary.
distributive property ofmultiplication will soon be found (compare the last
t This
Note)
near.
to
extend
to the
more general
case,
in
which the quaternions are not
collie
CHAP.
DISTRIBUTIVE MULTIPLICATION OF COLLINEARS. 209
I.]
by 209,
(1.) Hence,
for the
and 202,
(4.),
product of ant/ two quaternions
IX.
Hence
(2.)
(comp. 126
199, VII.
q'q
also, for the
this general transformation,
= Sq'. Sq + Yq'. Sq + Sq'.Yq + Yq'.Yq.

we have
square of any quaternion^
and 204, XXIII.)
we have
III.,
X..
the transformation
q^=Sq^ + 2Sq.Yq + Yq^,
we
(3.) Separating the scalar and right parts of this last expression,
other general formulae
find these
XL
whence
also,
XIII.
(4.)
8.92
= 892 +
SU(92)
XIL
792;
= (SU9)2 + (YUqy
By supposing
= K9, in IX.,
q'
Y.q^ = 2Sq.Yq;
we have
dividing by T92,
XIV.
VUC^^) = 2SU9 .YUq.
and therefore Sq' = S9, Yq'=-- V9, and trans-
posing the two conjugate and therefore complanar factors (comp. 191, (1.) ), we obtensor (comp. 190,
of^â
tain this general transformation for a norm, or for the square
V.
202, III.
and 204,
XV.
T92
XL)
= N9 = qKq =
:
(S9 + Yq) {Sq
- Yq) = Sq^ - Yq^
which had indeed presented itself before (in 204, XXII.) but is now obtained in a
new way, and without any employment of sines, or cosines, or even of the well-known
theorem respecting
tlie
square of the hypotenuse.

by XV., from XI., and dividing by T^^^
we
(5.) Eliminating Yq^,
XVI.
S 92 = 2S92 - T92
XVII.
find that
SU(92) = 2 (SU9)2
agreeing with 199, VI. and IV., but obtained here without any use of the
formula for the eoUne of the double of an angle.
(6.)
Taking the
general expressions
scalar
and right parts of the expression IX., we obtain these other
XVIII.
XIX.
in the latter of
Sq.
We may
(7.)
which we
S9'9
= Sq. S9 + S (Yq'. Yq)
Yq'q = V9'. S^
126) transpose the two
may (by
+ Yq. Sq + Y(Yq'. Yq)
factors, V9', Sq, or
(by 206, 207) write, instead of XIX., this other formula
also
Yq,
= IV9'. S9 + IV9. Sq + IV(V9'. V9).

If we suppose, in VII., that 9" = K9, 9"' = K9', and transpose
(comp. (4.) )
complanar (because conjugate) factors, q' -\-q and K(9' + 9), we obtain the
XIX'.
the two
known
IV9'9
following general expression for the
(9
+ 9)
or briefly,
XX.
norm of a sum
K (9' + 9) = 9'K9' + 9K9' + 9'K9 + 9K9
N(9' + 9) = N9' + 2S.9K9' + N9,
as in 200,
'
VIL
because
9'K9
(8.)
Kq + X,
= K.9K9',
by 192,
By
changing
where x is any
XXL
9' to
and (1 + K).9K9'= 2S.9K9', by 196,
II.,
scalar,
in
we
XX.,
or
by forming the product of 9
find that
N (9 + ) = N9 + 2a;S9 +
a;2,
as in 200, VIII,
whence, in particular,
XXr.
N(9 -
I)
= N9 - 2S9 +
2 E
1,
as in 200, II.
II'.
+ a;
and
210
Changing q to ft: a, and multiplying by the square

a and [3, the formula,
(9.)
two
of
Ta, we
[bOOK
II.
gef, for
any
vectors,
XXII.
T(/3
- a)2 =
T/32
2T/3
Ta SU ^ + Ta^,
.
which Ta2 denotes* (Ta)2; because (by 190, and by 196, IX.),
in
K(^-,Wt^=(I^Y,
Ta
a
\a
\
and
S^-=fsvi
Ta
a
a
(10.) In any plane triangle, abc, with sides of which the lotgths are as usual
denoted by a, 6, c, let the vertex c be taken as the origin o of vectors ; then
= CA,
we
/3-a =
^ = CB,
Ta =
AB,
6,
T(3
= a,
T(/3-a) =
c,
SU- = cosc;
a
recover therefore, from XXII., the Jnndamental formula of plane trigonometry,
under the form,
XXIII.
(11.) It is
= a2 -
c2
important to observe that
2ah cos c +
we have not
62.
here been arguing in a circle
because although, in Art. 200, we assumed, for the convenience of the student, a previous knowledge of the last written formula, in order to arrive more rapidly at certain
applications, yet in these recent deductions
from
multiplication of (at least) collinear quaternions,
i\\Q
distributive property
we have founded
YIW. of
nothing on the re-
and have made no use of any properties of oblique-ansults of that former Article
gled triangles, or even of right-angled ones, since the theorem of the sqiiare of the
hypotenuse has been virtually proved anew in (4.) nor is it necessary to the argument, that any properties of trigonometric functions should be known, beyond the
;
mQXQ
definition oi
a.
cosine, as a cexidJin projecting factor,
from which the formula
XVI. was derived, and which jusrifies us in writing cose in

The geometrical Examples, in the sub-articles to 200, may
(10.).
196,
again, and their validity be seen anew, without
being
now
any appeal
to even
the last equation

therefore be read
plane trigonometry
supposed.
= Tq^ + V52, as in 204, XXXI. ; and we know

that V^^j as being generally the square of a right quaternion, is equal to a negative
scalar (comp. 204, VI.), so that
(12.)
The formula XV. gives S52
XXIV
in each of
which two cases
Vg2 <
unless
0,
V9 = 0, by
Lq = 0,
or
= tt,
202, (0.), and therefore
its
square vanishes
II6IICC
XXV.
(SVqy <
Sg2 <Tg2,
1,
inevery other case.
We
Ta^ as denoting also T(a-) ;

assigned any meaning to the square of a vector, or generally
In the Third Book of these Elements it will be shown,
to the product of two vectors.
that such a square or product can be interpreted as being a quaternion : and then it
because
are not yet at liberty to interpret the symbol
we have not yet
will be found (comp. 190), that
T(a2)
whatever vector a
may
be.
(Ta)2
Ta2,
CHAP.
APPLICATIONS TO SPHERICAL TRIGONOMETRY.
I.]
211
(13.) It might therefore have been thus proved, -without any use of the transforSUq = cos Lq (196, XVI.), that (for any real quaternion q) ws have the in-
mation
equalities,
XXVI.
SU9<+1,
..
and
SVq>-l,
Sq<+Tq,
Sq>-Tq,
happen that Z 9 = 0, or = tt ; SU^ being = + 1, and Sq = + Tq, in the first

case
whereas SVq =1, and 85- = Tq, in the second case,
= ^q, and Tq.Tq :=T.qKq' =:T q'Kq = Bq.T (g' q), while
(14.) Since Tq^
unless
it
S qKq'
.
=S
XXVII.
if
= Ng S (^'
q'Kq
9), the
XX.
by XXVI.,
gives,
(Tq'+Tqy-T(q' + qy = 2(T-S)qKq'=21^q.(T-S)
we adopt the abridged
XXVIIL
XXIX.
which excepted
q'
each
case,
(15.) Writing q=:
a,
(5
member
g'= y
unless
q'
a,
in which latter case, but not in
any
and
x>0;
= (1 + x)Tq.
and multiplying by Ta, the formula XXIX.
of this last inequality
T7 + Tj3>T(y + /3),
> 0,
hence,
= xq,
becomes,
XXX.
(?)
T7 - Sg = (T - S) ?,
not a positive scalar
is
T9' + T9>T(9'+5),
(<?'
notation,
and suppose that the quotient
in
formula
other,
unless
becomes
= x[3,
we have Uy = U/3
x>0',
We
(155).
therefore
anew
at the results of 186, (9.), (10.), but without its having been necessary
to consider any triangle, as was done in those former sub-articles.
arrive
(16.)
On
the other hand, with a corresponding abridgment of notation,
we have,
by XXVI.,
XXXI.
also,
by XX.,
XXXII.
Tg + S5=(T+S)^>0,
unless
Lq = 7r;
&c.,
.
Tiq + qy (Tq' -Tqy=:2(T + S)qKq' =:2^q.(T+S) (q:q);
hence,
XXXIII.
where either sign
(17.)
And
T (g'
may
+ 9)
> + (Tq - Tg),
hence, on the plan of (15.), for
XXXIV.
unless
but, on the
x>
y,
/3,
unless
Uy = -
if
Uy = -U/3,
Ui3,
cctntrary,
XXXV. ..T(y + /3) = +CTy- T/3),
the upper or the lower sign being taken, according as
what was
= - a;^,
any two vectors
T(y + /3) > (Ty - T/3),
whichever sign be adopted
agrees with
5'
be taken.
inferred, in 186, (11.),
Ty >
or
<
T/3
all
which
from ^reome^ricaZ considerations alone,
In fact, if we make /3 = on, y = oc, and - y

then one' will be in general a plane triangle, in which the length of the side
Bc' exceeds the difference of the lengths of the two other sides ; but if it happen that
combined with the definition of Ta.
= oc',
the directions of the
oc have
two
lines ob, oc' coincide, or in other
words that the
lines
ob,
opposite directions, then the difi'erence of lengths of these two lines becomes
equal to the length of the line bc'.

(18.) With the representations of q and q\ assigned in 208, (1.), by two sides of
a spherical triangle ABC, we have the values,
^q
cos
c,
Sg'
= cos a,
^q'q
= S (y ) = cos
:
212
XVIII. gives
the equation
by 208, XXIV., the fundamental formula of
therefore,
spherical trigonometry (comp. (10.)
XXXVI,
COS h
II.
[bOOK
),
as follows
cos a cos c
sin
a sin
c cos b.
(19.) To interpret, with reference to the same spherical triangle, the connected
equation XIX., or XIX'., let
round b from c to A
arc CA,
a'
if b'
= oa',
be
&c.,
still,
it
now
so that
as in 208,
(2.),
supposed, as in 208, (5.), that the rotation
b and
b' are situated at the
same
side of the
Then writing
the positive pole of that arc.
we have
IVg'
= y'sinc;
lVg''
IV (V5'. Yq) =
and
be
is positive,
= a'sina
sin
j8
a sin
IVg^'^r
c sin
= /3' sin J
b (comp, 208,
(5.) ),
with the recent values (18.), for Sq and Sq'; thus the formula XIX'. becomes, by
transposition of the two terms last written:
XXXVII.
j3
sin
sin c sin
b=a
a cos
sin
+ /3' sin & + y' sin
c cos a.
= OP be any unit- vector; then, dividing each term of the last equa(20.) Let jO
tion by jo, and taking the scalar of each of the four quotients, we have, by 196, XVI.,
this
new
equation
XXXVIII,
sin
sin c sin
+
where
still,
b cos pb
sin c cos
= sin a
cos c cos pa'
a cos Pc'
of the sphere.
of the arcual perpendiculars, let
when p
is,
sin b cos pb'
abc, and a', b', c' are

is an arbitrary point,
a, h, c are as usual the sides of the spherical triangle
as in 208, (2.), the positive poles of those sides
upon the surface
but p
Also cos pa', cos pb', cos pc', are evidently the sines
from that point upon those sides being positive
fall
relatively to them, in the
same hemispheres as the opposite corners
triangle, but negative in the contrary case
of the
so that cos aa', &c., are positive,
and
are the sines of the three altitudes of the triangle.
(21.) If
we
place
at b,
two of these perpendiculars vanish, and the
last
formula
becomes, by 208, XXVIII.,
XXXIX.
such then
tiplied
is
sin 6 cos bb'
= sin a sin c sin B = TV V^.V- ];

\ (3
a)
(
the quaternion expression for the product of the sine of the side ca, mul-
by the
sine of the perpendicular let fall
upon that
side,
from the opposite ver-
tex B.
(22.) Placing p at A, dividing by sin a cos c, and then interchanging b and c,
get this other fundamental formula of spherical trigonometry,
XL.
and we
see that this is
cos
aa'= sin
c sin
= sin 6 sin c
we
included in the interpretation of the quaternion equation
XIX., or XIX'., as the formula XXXVI. was seen

XVIII.
in (18.) to be the interpretation
of the connected equation
By assigning other positions to p, other formulie of spherical trigonometry

Thus if we suppose p to cobe deduced, from the recent equation XXXVIII.
incide with b', and observe that (by the supplementary* triangle),
(23.)
may
* No
previous knowledge of spherical trigonometry^ properly so called, is here
supposed ; the supplementary relations of two polar triangles to each other forming
rather a part, and a very elementary one, 0^ spherical geometry
.
CHAP.
GENERAL DISTRIBUTIVE PROPERTY.
I.]
b'c'
cV = tt b,
=71 a,
a'b'=
213
c,
tt
while
cos bb'
we
= sin a sin c = sin c sin A, by
XL.,
easily deduce the formula,
XLI.
sin
sin c sin
A sin b sin c =
which obviously agrees, at the plane
sin
cos c cos c sin A - cos a cos A sin c
with the elementary
limit,
relation,
A + B + C = TT.
(24.) Again,
by placing p
XLII.
sin
at a', the general equation becomes,
a cos
= sin 6 cos c + sin c cos o cos b
with the verification that, at the plane
limit,
a= 6 COSC+ ccos b.
But we cannot here delay on such
to be
worth while
deductions, or verifications
although it appeared
whole of spherical trigonometry may thus be
to point out, that the
developed, from the fundamental equation of multiplication of quaternions (107), when

that equation is operated on by the two characteristics S and V, and the results
interpreted as above.
211. It
may next be proved,
as follows, that the distributive for-
of the last Article holds good, when the three quaternions,

which enter into it, without being now necessarily colliq, q\ q"^
7iea7\ are right; in which case their 7'eciprocals (135), and their sums
mula
I.
(197, (2.)
),
Let then
will be right also.
/Lq
qq = \;
Lq=Lq''=^,
and therefore,
We shall then
have,
by 106, 194, 206,
W+q')q = l{q^'+q'Ylq,
= {lq:''.lq?^-V{lq':lq)^q'q + q'q',
and the distributive property in question
(1.)
By
taking conjugates, as in 210,
tributive formula, 210, V., holds
good
for
q(s"^q') = qq" + qq\

(2.)
it is
is
proved.
easy hence to
infer,
that the of/ter dis-
any three right quaternions; or that

if
Lq = Lq'= Lq" = ~.
For any three quaternions, we have therefore the two equations
{Yq" + Yq') V^ = V^" V^ + V^' V^

Yq ( V5" + Yq') = Yq.Yq"+Yq. Yq'.
.
(3.)
The quaternions
7, 7',
q" being
still
arbitrary,
we have
thus,
by 210, IX.,
214
[bOOK
II.
(?" +q')l = (Sq" + Sq') Sq + (V^" + Yq') Sg + Vg (Sq" + Sq') + (Yq" + Yq') Yq
= (Sg". S^ + V^". S? +Yq,Sq" + Yq".Yq)^{^q.^q + Yq'. Sq + Yq.Sq' + Yq' .Yq)
= q"q + q'q
.
SO that the formula 210,

is
and therefore
I.,
also
(by conjugates) the formula 210,
"V.,
valid generally.
The General*
212.
Multiplication of Quater?iions
is
there-
fore (comp. 13,210) 2i Doubly Distributive Operation; so that

we may extend^ to quaternions generally ^ the formula
(comp,
210, VIII.),
I.
^q. '^q='2g'q:
summands of each set may
.
however many
the
be,
and whe-
ther they be, or be not, colUnear (209), or right (211).

(1.) Hence, as
an extension of 210, XX., we have now,

II.
where the second sign of

quaternions
q\
g,
N27 = 2% + 2 SS gKg'
summation
refers to all possible
binary combinations of the
And, as an extension of 210, XXIX., we have the
(2.)
III. ..
inequality,
ST^>T2g,
bear scalar and positive ratios to each other, in

g, q,
which case the two members of this iuequaUty become equal so that the sum of the
unless all the quateniions
tensors, of
set of quaternions, is greater
any
than the tensor of the sum^ in every
other case.
an extension
(3.) In general, as
IV.
(4.)
The
(5.)
We have
i^Tqy -
formulae, 210, XVIII.,
of 210,
XXVII.,
(T2gr)2
= 22(T - S) qKq.
XIX., admit
easily of analogous extensions.
168) the general equation,
also (comp.
V...(2y)2-S(g2) = 2(^g' +
in which,
by 210,
VI.
because,
5'g);
IX!,
qq'+ q'q=2{Sq.Sq' + Yq.Sq +Yq'.Sq
-^^
S(Yg'.Yq)y,
by 208, we have generally

VII.
VIII.
or
(Comp. 191,
(2.),
V(V/.V5) = -V(V^.V5');
V^'^ =
- Yqq,
if
^ g
=Z
g'
= -.
and 204, X.)
213. Besides the advantage which the Calculus of Quaternions

gains, from the general establishment (212) oHh.Q Distributive Principle,
or Distrihiitive Property of Multiplication^

*
Compare
the Notes to page 208.
by
being,
so
far,
CHAP.
INTERSECTIONS OF RIGHT LINES AND SPHERES. 215
I.]
assimilated to Algebra^ in processes which are of continual occurrence, this principle or property will be found to be of great im-
portance, in applications of that calculus to Geometry; and especially

in questions respecting the (real or ideal*) intersections of right
lines with spheres^ or other surfaces of the second order, including
contacts (real or ideal), as limits of
such intersections.
The
follow-
ing Examples may serve to give some notion, how the general disin
tributive principle admits of being applied to such questions
some of which however the less general principle (210), respecting
:
the multiplication of collinear quaternions (209), would be sufficient.

And first we shall take the case of chords of a sphere^ drawn from a
given point upon
its surface.
(1.) From a point a, of a sphere -with o for centre,

chord AP, which shall be parallel to a given
line
OB
more
or
OP,
let it
be required to draw a
fully, to assign the vector,
of the extremity of the chord so drawn,
as a function of the two given vectors, a = OA,

= OB ; or rather of a and Uj(3, since it
j3
and
is
evident that the length of the line
affect the result of the construction,
51
/3
cannot
which Fig.
may
ap ob, or p a
begin by writing the expression,
(2.) Since
||
/o
||
/3,
we may
= a + ar/3(15),
which may be considered (comp. 23, 99) as a form of the equation of the right line
AP and in which it remains to determine the scalar coefficient x, so as to satisfy the
;
equation of the sphere,
Tp=Ta
we
In short,
T(a-f
by some
(186, (2.)).
are to seek to satisfy the equation,
scalar
x which
= Ta,
C)8)
shall be (in general) different
stitute this scalar in the expression

p
= a + a;/3,
from zero
and then to sub -
in order to determine the required
vector p.
(3.)
For
and
square,
this purpose,
to
an obvious process
is,
after dividing
both sides by T/3, to
employ the formula 210, XXI., which had indeed occurred
before, as
200, VIII., but not then as a consequence of the distributive property of multiplication.
In
division
this
by
x,
manner we
are conducted to a quadratic equation,
and gives then,
-2S
= a-2/3S
/3'
/3'
Compare the Notes
to
page 90, &c.
which admits of
216
[bOOK
the problem (1.) being thus resolved, with the verification that
/3
may
II.
be replaced
U/3, in the resulting expression for p.
by
As a mere
(4.)
exercise of calculation,
dividing the last equation (2.)
we may vary
the last process (3.),
and then going on as
of T/3,
by Ta, instead
by
before.
This last procedure gives,
and
therefore,
a;
= -2S-:N^ = -2s5(by
a
(5.) In general,
by 196,
196, XII'.), as before.
(3
11'.,
1-2S = -K;
hence,
and
by
(3.),
finally,
new
expression for p, in which it is not permitted generally, as

* instead of the multiplicand.
f3 as the multiplier,
it
was
in (3.), to
treat the vector

(6.) It is
now
easy to see that the second equation of (2.)
is satisfied
for the
expression (5.) for p gives (by 186, 187, &c.),
Tp =
T^.T/3
= Ta,
as was required.
To
(7.)
interpret the solution (3.), let
chord AP, and
let
expression (3.) for p gives,
and accordingly, ocad
To
is-
pa
-f = K-,
we have only
to observe
the supplementary angle
let fall
/3s|=OD;
a parallelogram.
interpret the expression (5.),
/3
51 be the middle point of the

from A on ob then the
in Fig.
by 196, XIX.,
CA=|(a-p) =
(8.)
be the foot of the perpendicular
which gives
OA
=K
OB
op'
or
.^
if
ob'
)3
OP =PO,
(comp. 138) that the angle aop'
aop
externally,
by the
is
bisected internally, or
indefinite right line
ob
(see again
Fig. 51).
(9.) Conversely, the geometrical considerations
and
(8.) to interpret or to verify the
which have thus served in
two forms of solution
(3.), (5.),
(7.)
might have
been employed to deduce those two forms, if we had not seen how to obtain them,
rules of calculation^ from the proposed conditions of the question.
(Comp. 145,
by
(10.), &c.)
(10.) It
is
evident, from the nature of that question, that
Compare the Note
to
page 159.
a ought
to be deduci-
CHAP.
blc
from
IMAGINARY INTERSECTIONS.
I.]
(3
and
duce p from
by exactly the same
p,
and
/3
processes as those which have served us to de-
Accordingly, the form (3.) of p gives,
a.
S| -S|
and the form (5.)
gives,
-K|.,.
since the first form can be recovered
back
to the parallelism,
(11.)
The
from the second, we see that each leads us
p-a\\(3 (2.).
x shows that
solution (3.) for
x=
And
p-2/3S^;
p+2/3S^
K|=-|
And
2lT
= a,
0,
= A,
if
S(a:/3) =
or if
0,
(3
-i^
a.
namely, that a right line

drawn at the extremity of a radius oa of a sphere, so as to be perpendicular to that
its second point
radius, does not (in strictness) intersect the sphere, but touches it
the geometrical
meaning of
this result is obvious
of meeting the surface coinciding, in this case, as a limit, with the first.
Hence we may
(12.)
infer that the plane represented
or
s^::-"=o,
a
is
by the
equation,
sP=i,
a
the tangent plane (comp. 196, (5.)) to the sphere here considered, at the point A.
(13.) Since /3 may be replaced by any vector parallel thereto, we may substitute
for it
rt,
if
= oc
be the vector of any given point c upon the chord ap, whether
(as in Fig. 51) the middle point, or not;
therefore write,
-^ .(y -^=-K
a
-a
by
(3.)
and
(5.),
a),
In the Examples of the foregoing Article, there was no
214.
room
= a-2(y-a)S
we may
for the occurrence of imaginary roots of an equation, or for
ideal intersections of line
and surface.
such imaginary intersections
may
To give now a case in which

we shall proceed to con-
occur,
drawing a secant to a sphere, in a given direcfrom a given external point ; the recent Figure 51 still serving
sider the question of

tion,
us for illustration.
is the vector of any given point b, through which it is
(1.) Suppose then that a
draw a chord or secant epoPi, parallel to the same given line (3 as before.
required to
We have
now,
if
po = opo,
po
= +a;oA
xo2
being a
new
p.
scalar
and
= . + x,/3.
Ta = Tpo = T( + xoi3),
2oSi+Ni-N^
similarly, if pi
.,
= 0,
= oPi,
-si^{(T-;)%(vi)'},
2 F
218
by transformations* which
will easily occur to
[boOK
any one who has read recent
II.
articles
And
with attention.
the points Po, Pi wall be together real, or together

imaginarr/^
according as the quantity under the radical sign is positive or negative ; that is, ac-
cording as
(2.)
we have one
The equation
two following
or other of the
(conip. 203, (5.)
represents a cylinder of revolution, with
of
its base.
If
E be a point
sign in (1.) vanishes
this case, then, the line
as
inequalities,
),
ob
for its axis,
and with
Ta
for the radius
of this cylindric surface, the quantity under the radical
and the two roots

through
which
e,
xq, xi of the quadratic

is
become equal.
In
parallel to on, touches the given sphere
otherwise evident geometrically, since the cylinder envelopes the sphere (comp.
If E be internal to the cylinder,
204, (12.) ), and the line is one of its generatrices.
is
the intersections
Pq, Pi are real ;
but
if
e be external
to the
same
surface, those in-
tersections are ideal^ or imaginary.

(3.) In this last case, if
and
we make,
for
abridgment,
being thus two given and real scalars,

A'o
where V
is
= s-tV -1;
xi
we may write,
= 3+t\' -1;
the old and ordinary imaginary symbol of Algebra, and
vested here with any sort of Geometrical Interpretation.^
We
is
not in-
merely express thus
the /acf of calculation, that (with these meanings of the symbols a, /3, c, s and <)
the formula Ta = T( \-x^'), (1,), when treated by the rules of quaternions, conducts
to the
quadratic equation,
(X
sj-
+ <2 = 0,
the reason being that the right line through E is, in the
present case, wholly external to the sphere, and therefore does not really intersect it
at all ; although, for the sake oi generalization of language, we may agree to say,
which has no real root
as usual, that the line intersects the sphere in two imaginary points.
We
must however agree, then, for consistency of symbolical expression^ to

(4.)
consider these two ideal points as having determinate but imaginary vectors, namely,
the two following
po
in which
e'
it is
= c + /3-</3V-l;
easy to prove,
of the perpendicular
(as above) parallel to
let fall
on
pi
= + /3 + #^V-l;
that the real part s + s(3 is the vector e' of the foot
from the centre o on the line through E which is drawn
1st,
and Ilnd, that the real tensor
tT/S
of the
coefficient
of
It does not seem to be necessary, at the present stage, to supply so many references to former Articles, or Sub-articles, as it has hitherto been thought useful to
give
but such
may
still,
from time to time, be given.
t Compare again the Notes
to
page 90, and Art. 14&.
CHAP.
V-
in the vnaginary
e'e" to the sphere,
e'e
which proves the
we
part of each expression, represents the length of a tangent
drawn from that external
(5.) In fact, if
And
we
write oe'
point, or foot, e'.
= t' = e + s/3, we
shall
= 6 - 1' = - fi/3 = /3S = projection
have
whether the points
1st assertion (4.),
oe on ob
of
be real or imaginary.
Po, Pi
because
have, for the case of imaginary intersections,

<T/3
and the Ilnd assertion (4.)
in
219
CIRCUMSCRIBED CONES.
I.]
(6.)
An
which
(3
and y are two real
di
vectors, while
V-
1 is the (scalar)
imaginary of al-
geometrically real right versor (149, 153),
may
be said
a BivECTOR.
(7.)
and
T. E'E",
expression of the form (4.), or of the following,
gebra, and not a symbol for

to be
= V(T'2- Ta2) =
is justified.
In
like
manner, an expression of the form
are two real scalars, but
V-
(3.), or
1 is still the ordinary
be said by analogy to be a BisCALAr.
a;'
=s+ <V
1,
where
imaginary of algebra,
may
Imaginary roots of algebraic equations are
thus, in general, biscalars.

(8.)
And
if
a bivector (6.) be divided by a (real) vector, the quotient, such as
a
which go and
two real quaternions, but

be said to be a Biquaternion. *
in
qi are
V-
1 is,
as before, imaginary,
may
The same
distributive principle (212) may be employed in

investigations respecting circumscribed cones, and the tangents (real
or ideal), which can be drawn to a given sphere from a given point.
215.
(1.) Instead of conceiving that o, a, b are three given points, and that limits of
E are sought, as in 214, (2.), which shall alloAvthe points of in-
position of the point
tersection Po, Pi to be real,
we may
suppose that o, a, e (which
which shall permit the same reality
epoPi being
still
may
be assumed to
now the
ob are to be assigned,
drawn parallel to ob, as in
be collinear, without loss of generality, since a enters only by

data of the question and that limits of direction of the line
its
tensor) are
214, (1.).
(2.)
Dividing the equation
Ta = T( +
x(3)
by Te, and squaring, we have
Compare the second Note
to
page 131.
220
x may
the quadratic in
and
roots are real
its
and unequal, or
according as
is,
II.
therefore be thus written,
sin
EOB <
or
and
real
TVU^<
e
that
[bOOK
equal, or imaginary, according as
or= or>T-;
or
> T.oa
T.oe.
then Te < Ta, T(a c) > 1 ; but TVU^ can

(3.) If E be interior to the sphere,
never exceed unity (by 204, XIX., or by 210, XV., &c.) ; we have, therefore, in
this case, the first of the three recent alternatives, and the two roots of the quadratic
:
are necessarily real and unequal, whatever the direction of
/3
may be.
Accordingly
evident, geometrically, that every indefinite right line, drawn through an internal point, must cut the spheric surface in two distinct and real points.
E be SM/jer/aa?, so that Te = Ta, T (a :) !, then the first
(4.) If the point
alternative (2.) still exists, except at the limit for which /3 -^ f , and therefore
it is
T VU (/3
= 1,
we have
One root of the quaroot, namely

ar = 2S(:/3), is likewise always real, but vanishes for the case when the angle
EOB is right. In short, we have here the same system of chords and of tangents,
:
dratic in
f)
in -which case
now =0,
is
surface, as in
from a point upon the

write
213
and the other
the only difference being, that
we now
for a, or t for a.
(5.)
then
the second alternative.
for every direction of /3
But
e be an external
finally, if
TVU(j8
t)
may
point, so that
that any one of the three alternatives (2.)
ing direction of
(6.)
To
Tt
> Ta,
and
T(a
t)
< 1,
it
either fall short of this last tensor, or equal, or exceed
may come
to exist, according to the
so
vary-
/3.
illustrate geometrically
the law of passage from one such
we may
alternative to another,
ob-
serve that the equation.
TVU" = T-,
sin
EOP = T OA T. OE,
.
represents (when e
thus external)
a real cone of revolution, with its
is
vertex at the centre o of the sphere
and according
as the line
side this cone, or on
the
first
it,
on
lies in-
or outside
it,
or the second or the third of
the three alternatives (2.) is to be

adopted ; or in other words, the line
through
E,
drawn
^^S* ^2*
parallel (as before) to on, either cuts the sphere, or touches

it at all.
(Compare the annexed Fig. 52.)
does not (really) meet
it,
or
CHAP.
cone, with its
an external
still
to the sphere
it
221
point, the cone of tangents which can be drawn

real; and the equation of this enveloping or circumscribed
vertex at e, may be obtained from that of the recent cone (6.), by
IfKbe
(7.)
from
POLAR PLANES, CONJUGATE POINTS.
I.]
is
simply changing p to p
TVU^^=T-;
c
c
sinoEP = T.OA
or
any quaternion, and x any
(8.) In general, if q be
the recent equation (7.)
or at least one form of
it is, therefore,
may
it is,
T.OE.
scalar,
therefore be thus written
'
p-
or
T. p'p T. EP
:
be the foot of the perpendicular let
if p'
tient is evidently
(9.)
We may
=T. OA
fall
T. oe,
from p on oe ; and
in fact the first
quo-
= sin oep.
also write,
or
TV^ T^.t[^-i'J;
(^S^y-N^
N^-^N^-
2Se +
j;
or
as another form of the equation of the circumscribed cone.
(10.) If then
we make
also
N^ = l,
or
to express that the point
cone,
we
is
on the enveloped sphere, as well as on the enveloping
plane of contact, or of what

with respect to the given sphere :
find the following equation of the
polar plane of the point
e,
(S^-N^J
while the fact that
ponent
N^ = N-,
2,
(11.)
or by
The
its
it is
or
0;
a plane of contact*
is
equation entering through
is
called the
by the occurrence
of the ex-
S^-N^
exhibited
its
square.
vector,
f'=S^ = 6N- = OE',

6
is
right line
oe
and we
was
to be expected
ofE
cuts perpendicularly the
see that
T.Tt'
as
that of the point e' in which the polar plane (10.)
= Ta2,
or
T.oe.T.oe'= (T.oa)2,
from elementary theorems, of spherical or even of plane geo-
metry.
* In
fact a
modern geometer would
say, that
we have
here a case of two coinci-
dent planes of intersection, merged into a single plane of contact.
222
The equation
(12.)
formed
(10.), of the polar piano of e,
II.
easily be thus trans-
V"
=
si=(sP.Ni
P J
P
\
it
may
[bOOK
continues therefore to hold good,
Si-N^
or
when
0:
and p are interchanged.
If
then
we
take,
the vertex of a new enveloping cone, any point c external to the sphere, and
situated on the polar plane ff' . . of the former external point e, the new plane of
a.s
of the new point c, will pass through the former

a geometrical relation of reciprocity, or of conjugation, between the two
points c and e, which is indeed well-known, but which it appeared useful for our purpose to prove by quaternions* anew.
contact, or the polar plane dd'
vertex e
(13.) In general, each of the two connected equations,
which
may
P'
also be thus written,

i
iJst'L.-st^Y'-^''^
a
a
a
Y ap
= S. ?!!',
a
may be said to be a form of the Equation of Conjugation between any two points p and
because it exp' (not those so marked in Fig. 52), of which the vectors satisfy it
:
presses that those
two points
respect to the given sphere,
a well-known sense, conjugate to each other, with

= Ta.
are, in
Tp
(14.) If one of the two points, as p', be given by its vector p', while the other
point p and vector p are variable, the equation then represents a plane locus;
namely, what is still called the polar plane of the given point, whether that point be
external or internal, or on the surface of the sphere.

(15.) Let P, p' be thus two conjugate points; and let it be proposed to find the
in which the right line pp' intersects the sphere.
s, s',
Assuming (comp. 25)
points
that
OS
and attending
= =
(T
a;jO
+ yp',
x-\-y
l,
to the equation of conj ugation (13.),
TcT
we
= Ta,
have,
by 210, XX,,
or
by
200, VII., the following quadratic equation iny.x,
(a
+ y)2 = N[a;^+y^'V^^N- + 2a;y+i/2N^;

\
which
gives,
it is evident that, if the points of intersection s, s' are to be real, one

(16.) Hence
of the two points p, p' must be interior, and the other must be exterior to the sphere ;
norms here occurring, one must be greater and the other less than
because the two roots of the quadratic, or the two values o^y.x, differ
because, of the two

unity.
And
* In
fact, it will easily
be seen that the investigations in recent sub-articles are
put forward, almost entirely, as exercises in the Language and Calculus of Quaternions,
and not as offering any geometrical novelty of result.
CHAP.
EQUATION OF ELLIPSOID, RESUMED.
I.]
223
divided
sigyis, it follows (by 2G) that the right line pp' is harmonically
it is well known to
s' at which it meets the sphere
at
the
indeed
two
s,
points
be),
(as
or that in a notation already several times employed (25, 31, &c.), we have the haronly by their
monic formula,
(Fsp's')
From a
(17.)
drawn
as being
of
we can still speak of a cone of tangents,

we must say that those tangents are ideal,
real but internal point p,
to the sphere
or ijnaginary ;* and
contact
= -1.
but
if so,
as terminating on an imaginary circle
must consider them
which the real but wholly external plane
dern geometry, recognised as being (comp. (14.)
is,
by
quaternions, as
of
by mo-
the polar plane of the supposed
internal point.
Some
216.
readers
ing, to see here a
may
find
it
useful, or at least interest-
few examples of the application of the General
Distributive Principle (2 1 2) of multiplication to the Ellipsoid,
of which some forms of the Quaternion Equation were latelyassigned (in 204, (14,) ); especially as those forms have been
found to conduct! to a Geometrical Construction, previously
unknown,
for that celebrated
and important Surface or raIn what follows, it will

:
ther to several such constructions.
be supposed that any such reader has made himself already

sufficiently familiar with the chief formulae of the preceding
Articles
and therefore comparatively few references J

upon the present subject.
will
be
given, at least
(1.)
To
prove,
first,
that the locus of the variable ellipse,
=
I...S^
which locus
is
a:,
(v^y=a;2-l,
204,(13.)
represented by the equation,
the two constant vectors a, /3 being supposed to be real, and to be inclined to each
other at some acute or obtuse (but not right) angle, is a surface of the second order ^
Compare again the second Note to page 90, and others formerly referred
t See the Proceedings of the Royal Irish Academy, for the year 1846.
X Compare the Note
to.
to page 218.
4-a, the system I. represents (not an ellipse but) a pair of right lines,
real or ideal, in which the cylinder of revolution, denoted by the second
equation of
that system, is cut by a plane parallel to its axis, and represented
by the first equa-
If
tion.
/3
224
in the sense that
it is
cut
by an
\ = ol,
fi,
= om,
II.
arbitrary rectilinear transversal in two (real or ima-
ginary) points, and in no more than two,

vectors
[bOOK
as given
and
let
us assume two points l, m, or their

us seek to determine the points p (real or
let
imaginary), in which the indefinite right line lm intersects the locus II. ; or rather
the number of such intersections, which will be sufficient for the present purpose.
Making then p =
(2.)
(26),
we
have, for
z,
the following quadratic
equation,
III...(yS^+.S^J-(3,V^
zVey
= (j, + .).;
without proceeding to resolve which, we see already, by its mere degree, that the number sought is two ; and therefore that the locus II. is, as above stated, a surface of
the second order.
p is substituted for p
(3.) The equation II. remains unchanged, when
surface has therefore a centre, and this centre is at the origin o of vectors.
(4.) It has been seen that the equation of the surface
IV.
it
..Tfs^ +
may
the
also bo thus written
V^'\=l;
gives therefore, for the reciprocal of the radius vector
204,(14.)
from the centre, the expres
-i-l4^-?)^
has a real value, which never vanishes,* whatever real value may
be assigned to the versor Up, that is, whatever direction may be assigned to p
the
and
this expression
surface
is
therefore closed,
(5.) Introducing
and Jinite.
two new constant and auxiliary
expressions.
p-rCL
we have
vectors, determined
by the two
CHAP.
(6.)
CIRCULAR SECTIONS, CYCLIC PLANES.
I.]
Under the same
conditions, for
any arbitrary vector
225
we have
p,
the trana-
formaUons,
.p
VIII
= iIp +
IX.
=i/e_eV
p\
..^ + K^=S^ + V^;

a
the equation IV. of the surface
may
X.
..xf^+K^'j^l;
therefore
he thus written
or thus,
X'.
t/
|
+K^^=
meaning of which new forms will soon be seen.

The system of the two planes through the origin, which are respectively
perpendicular to the now vectors y and d, is represented by the equation,
the geometrical
(7.)
0,
or
XI...S^S^
combining which with the equation
XIII...l =
II.
we
XII...fs?Y = (s^
H)'{
get
fs^y-fv^y=N^;
or,
XIV.
Tp =
T/3.
These two diametral planes therefore cut the surface in two circular sections, with T/3
common radius and the normals y and d, to the same two planes, may be
for their
called (comp. 196, (17.) ) the ci/clic normals of the surface;

selves may be called its ci/elic planes.
while the planes them-
if we seek the intersection of the surface with the concentric

(8.) Conversely,
is T/3, we are conducted to the
equation XII. of
sphere XIV., of which the radius
the system of the two cyclic planes, and therefore to the two circular sections (7.) ;
so that every radius vector of the surface, which is not drawn in one or other of these
either greater or less than the radius T/3 of the sphere.
planes, has a length
all these marks, it is clear that the locus II., or 204, (14.), is (as above
(9.) By
its centre being at the origin
(3.), and its mean semiaxis
asserted) an Ellipsoid;
while Uf3 has, by 204, (15.), the direction of the axis of a circum^
two
= Ti3;
being
scribed cylinder of revolution, of which cyhnder the radius is T/3 ; and a

last cited sub-article, perpendicular to the plane of the ellipse of contact.
is,
by the
who are familiar with modern geometry, and who have caught the
(10.) Those
notations of quaternions, will easily see that this ellipsoid II., or IV., is a deformation of what may be called the mean sphere XIV., and is homologous thereto ; the
in the direction of /3
infinitely distant point
being a centre of homology, and either
two planes XI. or XII. being a plane of homology corresponding.
of the
X. or X'., of the quaternion equaadmits of being interpreted^ in such a way

as to conduct (comp. 216) to a simple construction of that surface
which we shall first investigate by calculation, and then
217.
The
recent form,
tion of the ellipsoid,
illustrate
by geometry.
2 G
226
(1.) Carrying on the

ing that (by 190, fee),
Roman numerals from
^=K^.N^,
and
[book
II.
the sub-articles to 216, and observ-
:?
= -.Ne,
the equation X. takes the form,
XV.
T^2
^-'^H]-^
XVI.
if
we make
=-
XVII.
T^2
when
and
<2
i- =
Ty2
<2'
and k are two new constant vectors, and < is a new constant scalar, which we
which the value may be chosen at pleasure.
shall suppose to be positive, but of
(2.) The comparison of the forms X. and X'. shows that y and d may be interchanged, or that they enter symmetrically into the equation of the ellipsoid, although
they may not at first seem to do so it is therefore allowed to assume that
;
XVIII.
for the supposition
T(j3
which
latter case
We
(.3.)
Ty > T^, and therefore that

Ty = T5 would give, by VI.,
XVIII'.
a)
= T(/3-a),
was excluded
and
.'.
Tt
(by 186, (6.) &c.)
in 216, (1.).
have thus,
XIX. ..Ut =
XX.
Ti
m;
T(J
XXI.
(4.)
Tk2
Tl2
Let ABC be a plane triangle,
such that
XXII. ..CB = I,
CA=/c;
let also
AE =
(0.
a sphere, which we shall call the

diacentric sphere^ be described round the
Then
if
point c as centre, with a radius
T/e,
and
therefore so as to pass through the centre
(here written instead of o) of the ellipand if D be the point in which the
soid,
line
AE meets
this sphere again,
we
shall
have, by 213, (6.), (13.),
XXIII.
CD = -K-.p,
Fig. 53.
and therefore
xxiir. ..DB = t+K-.p;

p
> T/c
a,
CHAP.
CONSTRUCTION OF THE ELLIPSOID.
I.]
XVI. becomes,
so that the equation
XXIV.
The
227
= T.AE.T.DB.
<2
external to the diacentric sphere (4.), by the assumption (2.) ;

a real tangent (or rather cone of tangents) to this sphere can therefore be drawn from
that point ; and if we select the length of such a tangent as the value (1.) of the sca(5.)
lar
that
*,
point
is to
is
say, if
we make each member
of the formula
XXI.
equal to unity^
sphere, as in
and denote by d' the second intersection of the right line bd with the
Fig. 53, we shall have (by Euclid III.) the elementary relation,
XXV.
whence follows
this Geometrical
<2
= T.db.T.bd';
Equation of the Ellipsoid,
XXVI. ..T.AE=T.BD';
somewhat more
or in a
familiar notation,
XXVII.
where ae denotes the length of the
AE = BD^;
and similarly for bd'.

The following very simple Rule of Construction (comp. the
line ae,
recent Fig. 53)

thereforeTrom our quaternion analysis
From a fixed point A, on the surface of a given sphere, draw any chord ad let
d' he the second point of intersection
of the same spheric surface with the secant ZD^
drawn from a fixed external* point B and take a radius vector ae, equal in
(6.)
results
and
length to the line bd',
AD
the locus of the point
in direction either coincident with, or opposite
will he
an
ellipsoid,
with
A.
for
its centre,
to,
the
chord
and with Bfor
a point of its surface.
Or thus:
(7.)
If,
length
of a plane but variable quadrilateral Abed', of which one side ab
and
if their intersection
given in
and
upon the surface of a given sphere, whereof

a chord, then the opposite side be is a chord of
be always situated
the side ad' of the quadrilateral is
a given
is
in position, the two diagonals AE, bd' be equal to each other in length,
ellipsoid.
218.
From
either of the
two foregomg statements, of the
Rule of Construction for the Ellipsoid to which quaternions

have conducted, many geometrical consequences can easily be
inferred, a few of which may be mentioned here, with their
annexed
proofs by calculation
course,
(1.)
in fact
still
the present Calculus being, of
employed.
That the corner b, of what may be called the Generating Triangle abc, is
a point of the generated surface, with the construction 217, (6.), may be
merely to fix the conceptions, that the point b is here supposed to be exterthe calculations and the construction would be almost the same, if we assumed B to be an internal point, or Tt < Tk, Ty < T^.
nal
It is
(f>.)
ELEMENTS OF
228
proved,
position
QUATERNIO:,'S.
[bOOK
11.
by conceiving the variable chord ad of the given diacentric sphere to take the
AG; where G is the second intersection of the line ab with that spheric sur-
face.
A (instead of g), and therefore x> to g

AE (or of ad) then tends to become tangential to the
of ae (or of bd') tends, by the construction, to become
(2.) If
be conceived to approach to
(instead of a), the direction of
sphere at A, while the length

equal to the length of bg the surface has therefore a diametral and circular section,
;
which touches the diacentric sphere at
in a plane
a,
and with a radius
(3.) Conceive a circular section of the sphere through A,
dicular to
G,
bc
and the
radius of a
d move along this
= bg.
made by a
plane perpen
move along a parallel circle through

such then is the
length of bd', or that of ae, will again be equal to bg
second diametral and circular section of the ellipsoid, made by the lately
if
circle,
d' will
mentioned plane.
(4.) The construction gives us thus two cyclic planes through A ; the perpendiculars to which planes, or the tivo cyclic normals (216, (7.)) of the ellipsoid, are
seen to have the directions of the two sides, CA, cb, of the generating triangle XBC
(1.).
(5.) Again, since the rectangle
BA bg = BD bd' = bd ae =
.
we have
double area of triangle
abe
sin
bdb,
the equation,
XXVIII.
perpendicular distance of
e from ab = bg
sin
bde
the third side, ab, of the generating triangle (1.), is therefore the axis of revolution
of a cylinder, which envelopes the ellipsoid, and of which the radius has the same
length, bg, as the radius of each of the
(6.)
For the points
two diametral and
of contact of ellipsoid
circular sections.
and cylinder, we have the geometrical
relation,
XXIX.
bde =
a right angle
or
XXIX'.
adb = a
right angle
the point d is therefore situated on a second spheric surface, which has the line ab
for a diameter, and intersects the diacentric sphere in a circle, whereof the plane passes
through A, and cuts the enveloping cylinder in an ellipse of contact (comp. 204,
(15.), and 216, (9.) ), of that cylinder with the elHpsoid.
in f, and let bf meet it
again in f'
(7.) Let AC meet the diacentric sphere again
(as in Fig. 63) ; the common plane of the last-mentioned circle and elhpse (6.) can
then be easily proved to cut perpendicularly the plane of the generating triangle abc
in the line af'
so that the line f'b
is
normal
to this
plane of contact; and there-
(by conjugate diameters, &c.) to the ellipsoid, at b.

the construction (217), to which many
(8.) These geometrical consequences of
others might be added, can all be shown to be consistent with, and confirmed by, the
fore also
quaternion analysis from which that construction itself was derived. Thus, the two
and their two cycircular sections (2.) (3.) had presented themselves
216, (7.)
clic normals (4.), or the sides CA, cb of the triangle, being (by 217, (4.) ) the two
vectors k,
i,
have (by 217,
(1 .) or (3.) ) the directions of the
which again agrees with 216,

(9.) Again,
it
/3;
y, ^
two former vectors
y, 5
(7.).
will be found that the
constant vectors, a,
and
i,
k,
assumed relations between the three pairs of
any one of which /;ai>s
is
sufficient to deter-
CHAP.
CONSEQUENCES OF THE CONSTRUCTION.
1.]
229
mine the
left to
ellipsoid, conduct to the following expressions (of which the investigation

the student, as an exercise) :
XXX...a =
^^,=^^J=^^Û(.
XXXI...,3 = ^4-y=j4-^ =
is
+ .)=K'B;
^^0(.-)
= BO,
f', g referring here to Fig. 63, while ajSyd retain their former meanings (21G), and are not interpreted as vectors of the points abcd in that Figure.
Hence the recent geometrical inferences, that AB (or bg) is the axis of revolution of
the letters b,
an enveloping cylinder
and that f'b
(6.),
normal
is
to the plane of the ellipse of con-
tact (7.), agree with the former conclusions (216,
such an axis, and that a

(10.) It
is
that
),
/3 is
easy to prove, generally, that
is
g+1^ Ng-1
1
g9-l_g (g-l)(K9+l) ^ Ng-
or 204, (15.)
(9.),
such a normal.
(9
+ l)(Kg+l)
N(*7
q-1
l)'
N(5-l)'
whence
^
~, T (TT^^'
T(, + ;c)2'
whatever two vectors t and k may be. But we have here,
<2 = Tt2 - T/c2, by 2 17,
XXXIII.
;
XXXII.
s^
(5.)
the recent expressions (9.) for a and
XXXIV.
The
last
written
= + (i + /c)
1+
/3
form 204, (14.), of the equation of the
1+
which the sign of the right part
may
ellipsoid,
S
I
fc)
"
^-V-^:S
S-^:S
K
K
+K
l
may
therefore be
now
thus
for
we
see that the last equation
XXXVI.
a value of p which evidently
(11.)
From
= AB =
II.,
thus
we
is
verify
by calcu-
a point of the sur-
when we suppose
i-k: =
^:S-;
a
the form 216, IV.
satisfies also
the form 216,
And
namely that b
is satisfied,
1=1
be changed.
lation the recent result (1.) of the construction,
face
= - (i-
XXXV. ..Tl
in
become, therefore,
(3
combined with the value
XXXIV.
of a,
it is
easy
to infer that the plane,
XXXVII.
which corresponds
..
s^ = i,
or
to the value
xxxvir.
S-^K = S^^^,
1 in 216,
I., touches the ellipaoid at the point b,

of -which the vector p has been thus determined (10)
the normal to the surface^ at
a:
that point, has therefore the direction of
that the last geometrical inference (7.)
is
or of a, that
so
is, of fb, or of f'b
thus confirmed, by calculation with
quater-
-f
k-,
nions.
219.
few other consequences of the construction (217)

may
be here noted
eiêcially as regards the geometrical determination
230
II.
[bOOK
of the three principal semiaxes of the ellipsoid, and the major and
minor semiaxes of any elliptic and diametral section ; together with
the assigning of a certain system of spherical conies, of which the
considered to be the locus.
may be
surface
denote the lengths of the greatest, the mean, and the least semithen if the side bc of the generating triangle cut
the diacentric sphere in the points h and h', the former lying (as in Fig. 53) between
{!.) Let a,
axes of the
6, c
ellipsoid, respectively
b and
the points
c,
we have
the values,
XXXVIII. ..a = B?;
so that the lengths of the sides of the triangle

of these semiaxes,
XXXIX.
and we
may
..
bc =
If,
\-
ca=Tk=
c
;
ab = T(i-k) =
ac
in the respective directions of the
BH, the points
we
L, N, thus obtained, will
And
of the surface.
if
we
two supplementary chords ah, ah' of the

an, with the lengths of eh',
be respectively a major and a minor summit

a of that surface, a peipendicular am
erect, at the centre
and
circular sections,
mean summit
c=T-T/c.
set off lines al,
to the plane of the triangle, with a length
two
expressed, in terms
a c
Ti2
Tk2
h= - - -;
r (i
k)
a = Tt+TK;
sphere, or in the opposite directions,
to the
= bh;
write,
XL.
(2.)
=
a
Ti
= bg;
abc may be thus
= bg,
will be situated
the point M (which will be common,

on the enveloping cylinder) will be a
thereof.
and ellipsoid are both cut by a plane through a, on

(3.) Conceive that the sphere
which the points b' and c' shall be supposed to be the projections of b and c ; then c'
will be the centre of the circular section of the sphere
and
if
the line b'c' cut this
which Di may be supposed to be the nearer to b',

the two supplementary chords adi, ado of the circle have the directions oithQ major
and minor semiaxes of the elliptic section of the ellipsoid while the lengths of those
new
circle in the points Di, DO) of
semiaxes
ba.bg:
are, respectively,
bdj, and
ba.bg: bd2
or bd'i
and
bd'2, if the
secants bdi and bdo meet the sphere again in Di' and D2'.
(4.) If these
note by
two semiaxes of the section be called
we have
the tangent from b to the sphere,
XLI.
= <2
BDi
a^
= aca-'^
a,
BD2 =
and
and
c,,
if
we
still
de-
thus,
<2
= acc~^
we
denote by p\ and /?2 the inclinations of the plane of the section to the two
cyclic planes of the ellipsoid, whereto CA and CB are perpendicular, so that the projections of these two sides of the triangle are
but
if
(c'a
XLII.
'c'b'
we have
XLIII.
= CA
= CB
= ^ (a c) sin/)i,
= i(a + c) sin/?2,
sin/j2
sin/)i
= b'd22 - b'di^ = 4b'c'
BDa^ - BDi^
c'a
= (a2 - C-) sin p^ sin p^
whence follows the important formula,
XLIY.
c,-2
a/2
(c-2
a-2) sin/î
unpz
CHAP.
231
SEMIAXES, SPHERICAL CONICS.
I.]
or in words, the
known and
" the
difference of the inverse
useful theorem, that
diametral section of an ellipsoid, varies as

squares of the semiaxes, of a plane and
the product of the sines of the inclinations of the cutting plane, to the two planes of
circular section.
As
(5.)
verifications, if the plane
be that of the generating triangle abc,
we
have
and
/'i=f2=->
but
if
P2 =
a^
= a,
the plane be perpendicular to either of the
0,
and
c^
two
= ei
ca, cb, then either pi or
sides,
= a^.
c^
(6.) If the ellipsoid be cut
by any concentric
sphere, distinct from the
mean
sphere XIV., so that
XLV.
AE = Tp = r ^
where
6,
r is a given positive scalar
then
XLVI.
so that the locus of
BD =
what may be
and which
is still
is,
^ba;
called the guide-point D, through which,
by the
ae
of the ellipsoid (or one of its prolongations)

at a constant distance from the given external point b, is
construction, the variable semidiameter

passes,
that
<2r-iâc6-i,
now
again a circle of the diacentric sphere, but one of which the j)Zae does not pass
(as it did in 218, (3.) ) through the centre A of the ellipsoid. The point e has therefore here, for one locus, the cyclic cone which has a for vertex, and rests on the last-
mentioned
circle as its
base ; and since
it is
must be on one or other of the two spherical

cone and sphere last mentioned intersect.
on the concentric sphere XLV., it

which (comp. 196, (11.)) the
also
conies, in
The intersection of an ellipsoid with a concentric sphere is therefore, genesystem of two such conies, varying with the value of the radius r, and becoming, as a limit, the system of the two circular sections, for the particular value
(7.)
rally, a
=5
and the
ellipsoid itself
nies, including those
two
may
be considered as the locus of
all
such spherical co-
circles.
(8.) And we see, by (6.), that the two cyclic planes (comp. 196, (17.), &c.) of
any one of the concentric cones, Avhich rest on any such conic, coincide with the two
cyclic planes of the ellipsoid : all this resulting, with the greatest ease, from the con-
struction (217) to
(9.)
With
which quaternions had conducted.

which was designed
respect to the Figure 53,
plained.
But
as regards the other letters
minor summit of the

chord af', of what
surface, so that an'
we may
to illustrate that con-
abcdd'eff'ghh'ln has been already ex-
struction, the signification of the letters
we may
= na
here add,
Ilnd, that
1st,
is
that n'
is
a second
a point in which the
here call the diacentric circle agf, intersects what
may
be called the principal ellipse,* or the section nblen' of the ellipsoid, made by the
plane of the greatest and least axes, that is by the plane of the generating triangle
ABC, so that the lengths of
ak and bf
are equal
this ellipse at this point, is parallel to the side
* In the
plane of what
bola of the ellipsoid.
is called,
Ilird, that the tangent, vkv', to
ab of the
triangle, or to the axis
of
by many modem geometers, the focal hyper-
232
revolution
[bOOK
II.
(5.) being in fact one side (or generatrix)

are thus two conjugate semidiameters of the
of the enveloping cylinder 218,
of that cylinder
IV th,
ab
that ak,
ellipse, and therefore the tangent tbt', at the point b of that ellipse, is parallel to
the line akf', or perpendicular to the line bff' ; Vth, that this latter line is thus the
normal (comp. 218, (7.), (11.) ) to the same elliptic section, and therefore also to the
ellipsoid, at b
Vlth, that the least distance kk' between the parallels ab, kv, being
= the radius b of the cylinder, is equal in length to the line bg, and also to each of
;
the two semidiameters, as, as', of the ellipse, which are radii of the two circular
sections of the ellipsoid, in planes perpendicular to the plane of the Figure
that AS touches the circle at
that
drawn
circle,
which
220.
The reader
is
and Vlllth, that the point
at right angles to the side
bc
Vllth,
on the chord ai of
s' is
of the triangle.
will easily conceive that the quaternion equa-
tion of the ellipsoid admits of being put under several other forms
among which, however,
here suffice to mention one, and to
may
it
assign its geometrical interpretation.

For any three vectors,
( I.)
we have
k, p^
i,
XLVII. ..Nf- +
\p
the transformations,
+ N-+2S-K-VNp
p
p p
p
j
= N-JSr-+N-N- +
K
VP
whence follows
(2.) If then
by the
we
introduce two
new
T-T"
\p
'I
= T['u.Ti + K HiJ^ p\
.
auxiliary and constant vectors,
and
k',
de-
equations,
XLIX.
which
K-.p^
-p p
"
2S
this other general transformation
XLVIII. ..Tf +
fined
i'
= -U/c.T/,
K'
= -Ut.T<c,
give,
L.
we may
Ti'
= Tt,
Tk'
write the equation
cisely similar
form
T (i' - /c') = T (i - k),
= Tk,
XVI,
(in
217) of the
ellipsoid
Tt'2
- Tk'2 = ^2,
under the following pre-
U...- =
T^.-+K-.<,
in
which
i'
and k have simply taken the
places of
and
k.
Retaining then the centre A of the ellipsoid, construct a new diacentric

sphere, with a new centre c', and a new generating triangle ab'c', where b' is a new
fixed external point, but the lengths of the sides are the same, by the conditions,
(3.)
LII.
Ac' =
K,
c'b'
draw any secant b'd"d"' (instead
=+
I,
of bdd'),
and
and
therefore
ab'
set off a line
ae
= - k'
t'
in the direction of
CHAP.
STANDARD QUADRINOMIAL FORM.
I.]
233
ad", or in the opposite direction, with a length equal to that of bd'";

the point E will be the same ellipsoid as before.
the locus
of
which we shall here* draw from this new construction

known) a second enveloping cylinder of revolution, and that
but that the radius of this second
its axis is the side ab' of the new triangle ab'c'
cylinder is equal to that of the first, namely to the mean semiaxis, b, of the ellipsoid
(4.)
The only
inference
that there exists (as
is,
is
and that the major semiaxis,
a,
or the line
AL
in Fig. 53,
bisects the angle bab',
between the two axes of revolution of these two circumscribed cylinders: the plane
of the new ellipse of contact being geometrically determined by a process exactly
similar to that employed in 218, (7.); and being perpendicular to the new vector,
i'
k',
as the old plane of contact
was (by 218, (11.))
to
t -f
k.
Section 14. On the Reduction of the General Quaternion

to a Standard Quadrinomial Form ; with a First Proof of
the Associative Principle of Multiplication of Quaternions.
'
221. Ketaining the significations (181) of the three rectangular unit-lines oi, oj, ok, as the axes, and therefore also
the indices (159), of three given right versors i,j, k, in three

mutually rectangular planes, we can express the index OQ of
any other right quaternion, such as V^^, under the tiinomial

form (com p. 62),
=
=
I.
IV5' OQ aj.on-y.oj + Z.OK;
.
where xyz are some three
scalar coefRcients, namely, the three
rectangular co-ordinates of the extremity q of the index, with

Hence we may write
respect to the three axes 01, oj, ok.
by 206 and 126,
also generally,
II.
Yq = xi
-\-
yj-^zk- ix -YJy + kz
and
this last form, ix +jy + kz^ may be said to be a Standard

Trinomial Form, to which every right quaternion, or the right
part Yq of any proposed quaternion q, can be (as above) re-
we denote by w the scalar part, Sq, of the same

we shall have, by 202, the following
General Reduction of a Quaternion to a Standard Quadri-
duced.
If then
general quaternion q,
nomial
Form
(183)
* If
room shall allow, a few additional remarks may be made, on the relations
to the ellipsoid, and on some other constructions of
t, k, &c.,
of the constant vectors
that surface, when, in the following Book, its equation shall
new
form,
T(tp+pK)=K2_i2.
2 H
come
to be
put under the
"234
III.
q^
(Sq + V^^ =^)w +
ix
[bOOK
^jy + kz
II.
which the four scalars, icxyz^ may be said to be the Four

Constituents of the Quaternion.
And it is evident (comp. 202,
(5.), and 133), that if we write in like manner,
in
IV.
where
q'
w' ^ ix \jy'
+ kz^
denote the same three given right versors (181) as
ijk
before, then the equation
V...?' =
?,
between these two quaternions, q and'', includes the four following scalar equations between the constituents :
VI.
which
is
iv'=
x=x,
IV,
==
new justification (comp. 112,

we have done throughout
z'
3/>
2/'
of naming, as
=^
z;
116) of the pi' opriety

the present Chapter,
the General Quotient of two Vectors (101) a Quaternion.

222. When the Standard Quadrinomial Form (221)
we have then not
adopted,
1.
as before,
but
only
and
V^
ix
^jy +
kz,
by 204, XI.,
also,
II.
And
^q = w,
is
K^ =
(S^
V^- =) w
ix
-jy
- kz.
because the distributive property of multiplication of qua-
ternions (212), combined vvith the laws of of the symbols ijk

(182), or with the General and Fundamental Formula of this
whole Calculus (183), namely with the formula,

2=J3
^2
Z/A
=-
(A)
1,
gives the transformation,

III.
we
{ix
+jy + kzf = -{x^ +
have, by 204, &c., the following
IV.
VII.
VIII.
IX.
+ z^),
expressions
NVg=(TV^)2=- V^2â;2 + ?/2-f 22.

TYq=^(x' + y' + z^);
U V^ = {ix +jy +kz): ^ (rc^ + f^ + z^)
V.
VI.
new
y^
U(/
=
.
=
(?f7
= Sq' - V^2 = w^+x'' + y' + z'

T^'*
.Tg= v'(^' + a;' + 3/' + 2');
+ ix +jy + kz):
yj {w"^
x"^
+y^ + ^0
'>
CHAP.
LAW OF THE NORMS.
I.]
X.
XI.
SU^ =
VU^
10'.
^J
+jy
(ix
23^5
{w'^-^x^ +
y''
-vkz)', y/ (m;2
-^ x'^
z')
y'^^- z"")
-A^^
XII...TVU,=
^
-v^^ +^
(1.)
To prove the
we may arrange
recent formula III.,
the multiplication (comp. again 182)
"^-y
as follows the steps of
Yq = ix +jy +
kz,
Y q = ix +jt/ + kz
= - a;2 -f kxy -jxz

=
+ iyz,
jy.Yq -y^- kyx
kz .Yq = z^
+jzx izy
V52 = Vg ? = - a;2 - y2 _22.
ix . V^
(2.)
"We have, therefore,
Xm...{ix+jy + kzy = -l,

a result to which
we have
if
already alluded,* in
x^+y^+z^ =
l,
connexion with the partial indeter-
minateness of signification, in the present calculus, of the symbol V 1, when considered as denoting a right radial
(149), or a right versor (153), of which the plane
or the axis is arbitrary,
If q"
(3.)
= q'q,
= w' + &c., q"=w" -^

5'
then
8ic.,
Ng"=Ng'.Ng, by
191,
but
(8.);
w" =
XIV
"
r"
= (i^j'z +
^"''^ "^ '^'"'^ "^ ^^'"^
we ought
" '^'^-^'
z'w')
+ {xy yx")
to,
and may be
.w" ^ ix" ->tjy" + kz" = (w' + ix' +jy' + kz) (w + ix +jy + kz)
&c.,
the quaternion formula,
therefore,
XVI.
=w+
and conversely these four scalar equations are jointly equivalent

in,
WW - (x'x + y'y + z'z),
''"
[:
summed up
XV.
if
then
W"2
under these conditions XIV., to have the equation,
+ a:"2 + y"2 + z"3 = (,'2 + x'^ + y'2 + 2'2)
(^j^2
+ x^ + y^ +
z^)
can in fact be verified by so easy an algebraical calculation, that its truth

be said to be obvious upon mere inspection, at least when the terms in the four
Avliich
may
quadrinomial expressions w"
2" are arrangedf as above.
Compare the first Note to page 131 and that to page 162.
t From having somewhat otherwise arranged those terms, the author had some
;
little
trouble at
first,
in verifying that the twenty-four double products, in the ex-
pansion of w'"^ + &c., destroy each other, leaving only the sixteen /jrocfwc^* ofsquares^,
or that XVI. follows from XIV,, when he was led to anticipate that result through
He believes, however, that the algebraic theorem

XVI., as distinguished from the quaternion formula XV., with which it is here con-
quaternions, in the year 1843.
nected,
had been discovered by the celebrated Eulbr.
236
The
223.
we
principal use which
[bOOK
shall here
11.
make of the
the genew^hich
ral associative property of multiplication of quaternions
can now with great ease be done, the distributive^ property
standard quadrinomial form (221),
is
to prove
by
it
(212) of such multiplication having been already proved. lu

we write, as in 222, (3.),
fact, if
I.
=w
q
y'
<j
i
q"
ix
^ jy + kz^
= z^' +
= w" +
ix
-\-
ix"
jy' + kz\
+ jy" + kz'\
that the relation q" = q'q', or any other

between the three quaternions ^, q\ q\ and
inquire whether it be true that the associative formula
without
now assuming
relation, exists
II.
holds good,
we
see,
qq^q^q^qq,
principle, that
by the distributive
only to try whether this last formula
is
valid
when
we have
the three
quaternion factors ^, ^', q are replaced, in any one common

order on both sides of the equation, and with or without repe-
by the three given right versors ijk but this has alarrive then, thus, at the
been
proved, in Art. 183.
ready
the
that
conclusion,
^General Multiplication of Quaimportant
ternions is an Associative Operation, as it had been previously
tition,
We
seen (212) to be a Distributive one: although we had also

found (168, 183, 191) that such Multiplication is not (in general) Commutative : or that the two products^ q'q and qq, are
generally unequal.
We
therefore omit the point (as in

II. by the
may
183), and may denote each member of the equation
symbol
q'q'q-
(1.) Let
ternions,
and
= Vg,
Kv'y
Let this
v'
therefore,
= u',
= Yq\
by 191,
v" =
Yq"
(2.),
Su't)
so that
= I (w't? + ry'),
last right quaternion be called
shall then
w,
v\ v" are
any three right qua-
and 196, 204,
u^,
and
Ww =('
let ^v'v
= s
rr/).
so that v'v
= s^ +
v^
we
have the equations.
* At a later
stage, a sketch Avill be given of at least one proof of this Associative
Frinciple of Multiplication, which will not presuppose the Distributive Principle.
CHAP.
ASSOCIATIVE PRINCIPLE OF MULTIPLICATION.
I.]
2 Vv"y
v'v^
vv"
v"s^
sy"
237
whence, hy addition,
= v". v'v v'v v"

= (v"v' + v'v")v - v'(y"v +
= 2vSv'v" 2v'Sv"v
2 Vr'tf^
vv"')
and therefore generally,
v" be
if c, ',
IIL
right, as above,
still
= vSv'v" v'Sv'v
v"yv'v
a formula with which the student ought to make himself completely familiar, on account of its extensive utility.
(2.)
With
the recent notations,
V
we have
v"%v'v =Yv''8^
= v"s^ = v"8vv';
therefore this other very useful formula,
IV.
where the point

in Avhich it is
still
v"v'v
= vSv'v"-
member may
in the first
'St'"y
+ v"8vv\
often for simplicity be dispensed with
and
supposed that
lv = iv = Lv =-.
(3.)
The formula
III. gives
V.
IV
(by 206),
v'Yv'v
= Iv
Sv'v"
Iv'.
Sv"v
hence this last vector, which is evidently complanar with the two indices Iv and It'',
is at the same time
(by 208) perpendicular to the third index lv'\ and therefore (by
(1.) ) complanar with the third quaternion q".
(4.)
With
the recent notations, the vector,
VI.
is
lv,= IYv'v = lY(Vq'.Yq),
(by 208, XXII.) a line perpendicular to both Iv and Iv'; or common to the planes
being also such that the rotation round it from Iv' to iv is positive :
of q and q'
while
its length,
TIv,,
or
or
Ty,,
or
TY.v'v,
TYCYq'.Yq),
bears to the unit of length the same ratio, as that which the parallelogram under the
indices, Iv and Iv', bears to the unit of area.
(6.)
To
interpret (comp. IV.) the scalar expression,
VII.
(because Sv"^= 0),
Sv'v'v
we may employ
--=
^v"v,
= S.t>"Vw'v,
the formula 208, V.
which gives the the trans-
formation,
VIII.
%v"v'v
where Tv" denotes the length of the
Tv".
Tw
cos
(tt-x);
and Tr, represents by (4.) the area

while x is (by 208), the
(positively taken) of the parallelogram under Iv' and Iv
This angle will be obtuse, and therefore the
angle between the two indices Iv", Iv,.
cosine of its supplement will he positive, and equal to the sitie
of the inclination of
the line Iv" to the plane oflv and Iv, if the rotation round Iv" from Iv' to Iv be
negative, that is, if the rotation round Iv from Iv to Iv" be positive ; but that cosine
line Iv",
will be equal the negative of this sine, if the direction of this rotation
We
have therefore the important interpretation

IX.
^v"v'v
=^
be reversed.
+ volume of parallelepiped under
Iv,
Iv, Iv"
238
II.
[boOK
the upper or the lower sign being taken, according as the rotation round Ir, from
I' to Iw", is positively or negatively directed.
For example, we saw that the ternary products
(6.)
lues,
namelv,
ijk
= -i,
kji
= +l, by
and
ijk
kji
have scalar va-
183, (1,), (2.);
of indices becomes, in this case, an unit-cube ;

while the rotation round the index of i, from that ofj to that of k, is positive (181).
a-ccovdrngly the parallelepiped
SLV\d
In general, for any three
(7.)
X.
and when the three
7-ight
quaternions
Svv'v"
we have
vv'v'\
= - Sv"v'v
the formula,
indices are complanar, so that the volume mentioned in IX. va-
two
nishes, then each of these
last scalars
new Formula of Complanarity
XL
becomes zero
so that
we may wnte,
as a
^v"v'v
= 0,
if
Ir"
1
Iv (123)
1',
while, on the other hand, this scalar cannot vanish in any other case, if the quaternions (or their indices) be still supposed to be actual (1, 144); because it then represents an actual volume.
Hence
(8.)
also
three qtiatcrnions
we may
XII. .. S (Vq" Yq. Yq)

.
that
is,
by 209,
if
if
= 0,
if
lYq"
lYq',
lYq
q" have any common line.

we employ the standard trinomial form 221,
the planes of g,
(9.) In general,
v
Formula of Collinearity,
establish the following
for
any
q',
= Yq = ix -\-jy + kz,
v'
= ix' + &c.
w" = ix"
11.,
+ &c.
namely,
the laws (182, 183) of the symbols i,j, k give the transformation,
XIII,
and accordingly
Sy"t'
of the parallelepiped,
edges,
if
= x" {z'y y'z) + y" {x'z -
this is the
known
z'x)
which has the three
z" {jj'x
volume
expression for the
lines op, op',
(Avith
op"
the rectangular co-ordinates* of the four comers, o,
z'y)
a suitable sign)
for three co-initial
p, p',
p" be 000, xyz,
x'y'z\ x"y"z".
(10.) Again, as another important consequence of the general associative property of multiplication, it may be here observed, that although products of more than
two quaternions have not generally equal scalars, for nil possible permutations of the
factorsj since
we have
just seen a case X. in
produces a change of sign in the
which such a change of arrangement

permutation is permitted, under
result, yet cyclical
or in symbols, that for any three quaternions (and the result is easily extended to any greater number of such factors) the following formula holds good
the sign S
Xiy.
In
fact, to
prove this equality,
^q'q'q =
we have only
Sqq"q'.
to write it thus,
Xm..S(9'V-9) = S(9.9'V),
and
to
remember that the
tered (198, I.),
when
* This result
may
scalar of the product of
any two quaternions remains unal-
the order of those two factors
is
serve as an example of the
changed.
manner
in
which quaternions,
although not based on any usual doctrine of co-ordinates, may yet be employed to
deduce, or to recover, and often with great ease, important co-ordinate expressions.
CHAP.
COMPLANAR QUATERNIONS.
I.]
239
it may be inferred that

= K (g", q'q) = Kq'q Kq" = Kq Kq
(11.) In like manner, by 192, II.,
XV.
K."gq'q
Kq",
with a corresponding result for any greater number of factors; whence by 192, I.,
if
Ilg and U'q denote the products of any one set of quaternions taken in two opposite orders,
we may
XVI.
But
(12.)
XVIII.
if
write,
.
KUq = n'Kq
XVII.
RUq = H'Rg.
Kv = -v, by 144
XX.
Snc = + Sn'v
r be right, as above, then
KUv = n'v;
XIX.
number
upper or lower signs being taken, according as the
hence,
.
Vnv = + YWv
of the right factors
is
even or odd; and under the same conditions,
XXI.
as
was
SHi; = h(Uv n'v)
lately exemplified (1.), for the c&se
(13.) For the case where that
XXIII.
XXIV.
results
XXII.
number
VHi; = ^ (Hv + U'v)
where the number

is three,
is tico.
the four last formulsa give,
= - Svv'v" = ~(v"v'v vv'v'")

Yv'v'v = -f Yvv'v" = I (y"v'v + vv'v")
Sv"v'v
which obviously agree with X. and IV.
224. For tbe case of Complanar Quaternions (119), the power of

reducing each (120) to the form of a fraction (101) which shall have,
at pleasure, for its
denominator or for
line in the given plane, furnishes
its
numerator, any arbitrary-
some peculiar
facilities for
proving
the commutative and associative properties o^ Addition (207), and the

distributive and associative properties oi Midtiplication (212, 223);
while, for this case of multiplication of quaternions, we have already
seen (191, (I-)) ^^^^ *^ commutative property also holds good, as
it
does in algebraic multiplication.
It
may
therefore be not irrele-
vant nor useless to insert here a short Second Chapter on the subject
oisuchcomplanars: in treating briefly of which, while assuming as
proved the existence of all the foregoing properties, we shall have an

opportunity to say something of Powers and Roots and Logarithms ;
and of the connexion of Quaternions with Plane Trigonometry, and
with Algebraical Equations. After v^hich, in the Third and last
Chapter of this Second Book, we propose to resume, for a short time,

the consideration oi Diplanar Quaternions; and especially to show
how the Associative Principle of Multiplication can be established,
for them, without* employing the Distributive Principle,
*
Compare the Note
to
page 236.
240
CHAPTER
[bOOK
II.
II.
ON COMPLANAR QUATERNIONS, OR QUOTIENTS OF VECTORS IN

ONE PLANE; AND ON POWERS, ROOTS, AND
LOGARITHMS OF QUATERNIONS.
Section
1.
On
Corn-planar Proportion of Vectors;
Fourth
Three, Third Proportional to Two, Mean

Proportional
Proportional, Square Root; General Reduction of a Quaternion in a given Plane, to a Standard Binomial Form.
to
The Quaternions
of the present Chapter shall all be

be
complanar (119); their common plane being
supposed
assumed to coincide with that of the given right versor z ( 1 8 1 ).
225.
to
And all the lines, or vectors, such as
a, j3, 7, &C., or ao, oi, as,

be
conceived to be in that
shall
be
here
to
&c,,
employed,
write
of
that
so
we
z;
(by 123), for the purmay
given plane
poses of this Chapter, the forrnulcB of complanarity :
Under
226.
these conditions,
interpret any symbol of the form

S in the given plane; which line
we can always (by

:
(/3
0)7,
103,
17)
as denoting a line
may also be denoted (125)

not*
but
the
la)
(comp. 103) by either of the
by
symbol (7
.^^
two apparently equivalent symbols,
that
we may
(j3.
7)
a,
{y.j5):a\ so
write,
and may say that
this line ^ is the
* In fact the
symbols
Fourth Proportional to the
us
/3. y, y /3, or /3y, y/3, have not as yet received with
and even when they shall come to be interpreted as representing certain quaternions, it will be found (comp. 168) that the two combinations,
any
interpretation
y and
have generally
different significations.
CHAP.
COMPLANAR PROPORTION OF VECTORS.
II.]
three lines
a, |3,
or to the three lines a, 7,
241
/3
the two
/3
y, of any such Complanar Proportion of Four
Vectors, admitting thus of being interchanged, as in algebra.
and
Means,
Under the same
conditions
we may
write also (by 125),
so that (still as in algebra) the tioo Extremes, a
and
^,
of any
such proportion of four lines a, j3, 7, , may likewise change

while we may also make the means
places among themselves
:
we at the same time change the extremes to means. More generally, if a, j3, 7, S, e
be any
odd number of vectors in the given plane, we can always find
become the extremes^
if
another vector p in that plane, which shall satisfy the equation,
"I
M-"'
"^'
Mr'-'
and when such a formula holds good, for any one arrangement
of the numerator-lines a, 7, e,
and of the denominator-lines
.
can easily be proved to hold good also for any

^
/o, j3,
other arrangement of the numerators, and any other arrange.
ment
it
For example, whatever four (commay be denoted by /378, we have the trans-
of the denominators.
planar) vectors
formations,
the
two numerators being thus interchanged.
so that the
Again,
two denominators
also may change places.

interesting case of such proportion (226) is that
in which the means coincide; so that only three distinct
lines,
such as a, /3, 7, are involved and that we have
Art.
227.
An
149,
(comp.
and Fig. 42) an equation of the form,

I.
..7 = ^)3,
a
2
or
a=^/3,
'
242
but not* y =
/3/3
a,
nor a =
/3j3
[bOOK
II.
7. In this case, it is said that
the three lines ajSy form a Continued Proportion; of which a

and y are now the Extremes^ and /3 is the Mean : this line (5
being also said to be t Mean Proportional between the two

others, a and y ; while y is the Third Proportional to the tAvo
lines a
and
tional to
and d
j3
y and
j3.
SO that this mean,
portional (226)
We have also
is,
at the
j3,
betAveen a and 7, is also i\iQ fourth proas first, and to those two other lines.
to itself,
(comp. again 149),
\aj
whence
it is
same time, the third proporconditions, we have
Under the same
\y
natural to write,
l3
_ fy\h
a
/3
_ fa\h
and therefore (by 103),
although we are not here to write j3 = (ya)i, nor j3 = (ay)^.

But because we have always, as in algebra (comp. 199, (3.) ),
the equation or identity, (-
qy = g'^, we
are equally well enti-
tled to write,
the symbol qi denoting thus, in general, either of two opposite

quaternions, whereof however one, namely that one of which
the an^le
which
is
shall
acute, has been already selected in 199, (1.), as that

be called by us the Square Root of the quaternion
* Compare the Note to the foregoing Article.

We say, a mean proportional ; because we
j3,
shall shortly see that the opposite
same sense another mean; although a rule will presently be given,

for distinguishing between them, and for selecting one, as that which may be called,
by eminence, the mean proportional.
line,
is in
the
CHAP.
II.]
CONTINUED PROPORTION, MEAN PROPORTIONAL. 243
and denoted by
q,
Wc may
\/ q,
therefore establish the for-
mula,
as above, a continued proportion ; the upper

when (as in Fig. 42) the angle aoc, between
taken
signs being
the extreme lines a, 7, is bisected by the line ob, or /3, itself;
if a, j3,
7 form,
but the lower signs, when that angle is bisected by the opposite
or when j3 bisects the verticallg opposite angle (comp.
line,
jS,
again 199, (3.) ): but the proportion of tensors,
VIII. ..T7:T/3 = Tj3:Ta,
and the resulting formulae,

= Ta .T7,
IX.
T/3^
T/3
v^
(Ta .T7),
in each case holding good.

And when we shall speak simply
of the Mean Proportional between two vectors, a and 7, Avhich
make any
acute, or right, or obtuse angle with each other, we

always henceforth understand the former of these two
bisectors
namely, the bisector ob of that angle aoc itself, and
shall
not that of the opposite angle
thus taking upper signs, in the
recent formula VII.

At
(1.)
the limit
when
each of these two unit-lines
if
a be any positive
At the
(2.)
VI
aoc vanishes, so that Uy = Ua, then U/3 =

mean proportional /3 has the same common
two given extremes.
direction as each of the
X.
the angle
and the
= + 1,
This comes
and generally,
X'.
our agreeing to write,
to
V(a2) = -f
a,
scalar.
other limit,
when Aoc = 7r,
Uy =- Ua,
or
the length of the
mean
determined by IX., as the geometric mean (in the usual sense)

between the lengths of the two given extremes (comp. the two Figures 41) ; but,
even with the supposed restriction (225) on the plane in which all the lines are
proportional
situated,
is still
/3
an ambiguiti/
arises in this case,
perpendiculars at o, to the line AOC,
To remove
tor.
of
(to
lar),
which axis
from the
supposition
is
we
from the doubt which of the two opposite
to be taken as the direction of the
shall suppose that the rotation
all the lines considered in this
first line
oa
are,
vec-
by 225, perpendicuwhich
to the second line ob, is in this case positive ;
V-
=+
and
XI'.
V(- a^) = ia,
if
> 0.
be carefully observed that this square root of negative unity is not in

imaginary, nor even ambiguous, in its geometrical interpretation, but
It is to
sense,
Chapter
mean
round the axis
equivalent to writing, for present purposes,
XI.*
any
this ambiguity,
is
denotes a real
and given
right versor (181).
244
And
thus the
mean proportional between two
in all cases, determined
II.
[bOOK
vectors {in the given plane) becomes,
at least if their order (as first
and third) be given.
(3.) If the restriction (225) on the common plane of the lines, were removed, we
might then, on the recent plan (227), fix definitely the direction, as well as the
length, of the meon OB, in every case hut one: this excepted case being that in
which, as in (2.), the two given extremes, OA, oc, have exactly opposite directions so
that the angle (aoc = tt) between them has no one definite bisector. In this case, the
;
sought point b would have no one determined position, but only a loctts : namely the
circumference of a circle, with o for centre, and with a radius equal to the geometric
mean between oa,
line
AOC.
oc, while
its
plane would be perpendicular to the given right
(Comp. again the Figures 41
and the remarks in 148, 149, 163, 154,
on the square of a right radial, or versor, and on the

racter of the square root of a negative scalar,
when
partially indeterminate cha-
interpreted,
ou the plan of this
Calculus, as a real in geometry.)
228. The quotient of any two complanar and right quaternions has been seen (191, (6.) ) to be a scalar ; since then we
here suppose (225) that q\\\ii we are at liberty to write,
l...Sq = x;
Yq:i=y;
Yq=-yi=iy;
establish the following Reduction of a

in
the
Quaternion
given Plane (of 2*) to a Standard Binomial
and consequently may
Form* (comp. 221):

11.
X and y being some
if
q^x+iy,
tioo scalarsy
q\\\i;
which may be called the
tivo
(comp. again 221) of this binomial. And then an

equation between two quaternions, considered as binomials of
this form, such as the equation,
co7istituents
III.
q'=
or
q,
breaks up (by 202, (5.)
III'.
x +
iy=x + iy,
into two scalar equations between
their respective constituents^ namely,
IV.
x=x,
y=yi
notwithstanding the geometrical reality of the right versor,
i.
marked as III.,
(1.) On comparing the recent equations II., III., IV., with those
v., VI., in 221, we see that, in thus passing from general to co?n/>Zanar quaternions,
we have merely suppressed the coefficients ofj and k, as being for our present purpose,
ntdl
and have then written x and
It ia permitted,
form X
-i-
iy is shorter,
by 227, XL,
and perhaps
y, instead of
w and
x.
to write this expression as

less
x + yV
1; but the
Uable to any ambiguity of interpretation.
CHAP.
STANDARD BINOMIAL FORM, COUPLE.
II.]
As
245
word "binomial" has other meanings in algebra, it may be conveCouple and the two constituent scalars x and y, of
which the values serve to distinguish one such couple from another, may not unnaturally be said to be the Co-ordinates of that Couple, for a reason which it may be
(2.)
the
nient to call the form II. a
useful to state.
(3.) Conceive, then, that the plane of Fig.
round Ax.t
positive rotation
which
may
axis of
i is
50 coincides with that of
in that Figure, directed
is,
t,
and that
towards the left-hand;
be reconciled with our general convention (127), by imagining that this

o towards the back of the Figure or below* it, if horizon-
directed from
This being assumed, and perpendiculars bb', bb" being let fall (as in the Figure) on the indefinite line OA itself, and on a normal to that line at o, which normal we may call oa', and may suppose it to have a length equal to that of oa, with
tal.
a left-handed rotation aoa', so that
V.
while
/8'
= ob',
oa'
and
then, on whichever
= I. OA,
V.
or briefly,
/3"= ob", as in 201, and g
oa
side of the indefinite right line
a'
= ia,
= (3:
a,
the point
as in
202
b may be
situated,
a comparison of the quaternion q with the binomial form II. will give the two equations,
VI.
(=
ir
S^y)
so that these two scalars,
= /3'
x and
y (= V^ i=
:
/3"
ia) =/3"
y, are precisely the two rectangidar co-ordinates
of
the point B, referred to the two lines OA and oa', as two rectangular unit-axes, of
And since every other quaternion, q z=x' -{- iy\
the ordinary (or Cartesian) kind.
can be reduced to the form y a, or oc OA, where c is a point

which can be projected into o' and c" in the same way (comp. 197,
that the two new scalars, or constituents, x' and y', are simply (for
in the given plane,

in that plane,
205), we see
the same reason) the co-ordinates of the
axes.
new point
c,
referred to the
same
pair of
the principles of the foregoing

(4.) It is evident (from
Chapter), that if we thus
express as couples (2.) any two complanar quaternions, q and q', we shall have the
following general transformations for their sum, difference, and product :
VII..
VIII.
(5.) Again,
Vg = iy,
for
q'q = (ix'x) + i(^y'y);
= {x'x - y'y) + i (x'y + y'x).
q'.q
any one such couple,
q,
we have (comp. 222) not only S*; = x, and
as above, but also,
lX...Kq = x-iy;
XII.
U, =
X.
-^,
of
(6.) Hence, for the quotient
[q'
XIV...
J5
|_
l^q
x'
^T^
XL
XIII.
any two such
+ iy'
x"
= x^ +y^
x"
+ iy"
-^m7^'
= x'x + y'y,
couples,
^^
Tg =V(a;2 +y2);
4^^
we
.
&c.
have,
,,
^ ^'^ =^^5'
=
y" y'x - x'y.
to
page 108.
246
(7.)
The law of the norms (191, (8.) ), or the

by the well-known
pressed here (comp. 222, (3.))
XV.
in
which xyx'y'
Section
tors
..{af^
may
+ y'^)
[bOOK
formula, Ng'g-
= Ng^
Nj,
is
ex-
algebraic equation, or identity,
+ y'^) = (x'a; - y'yf + (x'y + y'x^
(a;2
II.
be any four scalars.
On Continued Proportion
of Four or more VecWhole Powers and Roots of Quaternions ; and Roots
2.
of Unity.
The
229.
conception of continued proportion (227)
may
easily be extended from the case oi three to that oi four or

more (complanar) vectors ; and thus a theory may be formed
o^ cubes and higher whole poioers of quaternions^ with a corre-
spondingly extended theory of roots of quaternions, including

roots of scalars^ and in particular of unity.
Thus if we suppose that the four vectors ajSyS form a continued proportion,
expressed by the formulae,
I..J
^
p
= ^,
whence
11.
.U-l^- J^-a
(by an obvious extension of usual algebraic notation,) we may

say that the quaternion S a is the cube, or the third poicer, of
and that the latter quaternion is, conversely, a cube/3 a ;
root (or third root) of the former
which last relation may na:
turally be denoted
III.
..^
by writing,
or
^^Y,
Iir.
..i3=(^Ya(comp.227,IV.,V.).
230. But it is important to observe that as the equation

=
Q, in which q is Sk sought and Q is a given quaternion,
qwas found to be satisfied by two opposite quaternions q, of the
form V Q (comp. 227, VII.), so the sliglitly less simple
equation q^= Q is satisfied by three distinct and real quaternions, if Q be actual and real whereof each, divided by either
;
of the other two, gives for quotient a real quaternion, which

is equal to one of the cube-roots of positive unity.
In fact, if
the
annexed Fig. 54) that j3' and j3" are
we conceive (comp.
two other but equally long vectors
in the given plane, ob-
CHAP.
CUBE-ROOTS OF A QUATEBNION, AND OF UNITY.
II.J
tained from
j3
by two successive and
247
positive rotations, each
through the third part of a circumference,

so that
or
"^
'13'
15"
i3'
and therefore
v-(IT=(f)-=..-v-f-(ITf=(f
we
shall
have
we are equally entitled, at this stage, to write, instead

of III. or Iir., these other equations
so that
vii...a'=f?Y,
^'=^-^v
vir...ff = f^>,
/3"=
or
231.
(real
have three
and actual) quaternion
Q may
thus be said
; of which
however only one can have an anÎe less than sixty degrees ;
while none can have an angle equal to sixty degrees, unless the
proposed quaternion Q degenerates into a negative scalar. In
to
(real, actual, and) distinct cube-roots
every other case, one of the three cube-roots of Q, or one of the

three values of the symbol Qa, may be considered as simpler
than either of the other two, because it has a smaller angle

(comp. 199, (10)5 and if we, for the present, denote this one,
which we
shall call the Principal
Q, by the symbol
^ Q, we
Cube-Root of the quaternion

be enabled to establish
shall thus
the formula of inequality,
VIIL
232.
At
the limit,
scalar, one of its
..Z^Q<^,
when Q
cube-roots
is
if
zQ<7r.
degenerates, as above, into a negative

itself a
negative scalar, and has there-
248
fore its angle =7r; while eac^ of the
In
this case,
these
among
[bOOK
two other roots has
its
II.
angle
two roots of which the angles are
equal to each other, and are less than that of the third, we shall
consider as simpler, and therefore as principal, the one which an-
swers (comp. 227, (2.) ) to a positive rotation through sixty degrees ;

and so shall be led to write,
IX...y-l = lp^^
and
X...Z^-1=|;
using thus thQ positive sign for the radical ^3, by which i is multiplied in the expression IX. for 2y/- 1 ; with the connected for-
mula,
IX^..y(-a3) = |(l+eV3),
it
although
might
>0;
if
have seemed more natural to adopt as
at first
principal the scalar value, and to write thus,
3/-l=-l;
which
latter is in fact one value of the symbol, (-
We have,
(1.^
The
(2.)
however, on the present plan, as in arithmetic,
..^1 = 1;
XI.
1)*.
and
Xr. ..
^(a3)
= a,
if
a>0.
equations,
can be verified in calculation^ by actual cubing, exactly as in algebra

ference being, as regards the conception of the subject, that although
equation
i^
= 1,
it is
the only dif-
satisfies
the
regarded here as altogether real; namely, as a real right ver-
sor* (181).
233. There
no difficulty in conceiving how the same general

be extended (comp. 229) to a continued proportion
is
principles may
of n 4- 1 complanar vectors,
I.
a, ai, ag,
a,
This conception differs fundamentally from one which had occurred to seve-
before the invention of the quaternions ; and according to which

the symbols 1 and V 1 were interpreted as representing a pair of equally long and
mutually rectangular right lines, in a given plane. In Quaternions, no line is repreral able writers,
sented
that
by the number, One, except
we
as regards its length
the reason being, mainly,
require, in the present Calculus, to be able to deal with all possible planes
and that no one right
line is
common
to all such.
CHAP.
FRACTIONALPOWERS, GENERAL ROOTS OF UNITY. 249
II.]
when n
is
a whole
number
greater than three ; nor in interpreting,
in connexion therewith, the equations,
\a
\a
Denoting, for the moment, what we shall call the principal n*^ root
of a quaternion Q by the symbol
y^Q, we have, on this plan (comp.
231, VIII.),
TT
V.
YI...L (/this last condition,
/.yQ<
1)
ZQ<7r;
if
TT
Vir.
^;
n'
namely that there
V(/- l)u>0;
shall be a positive (scalar) co-
of
z, in the binomial (or

efficient y
couple) form x-\-iy (228), for the
thus
to
quaternion;;/-],
serving
complete the determination of
thsit
principal n^^ root of negative unity
lar, since
formulas.
write,
may be changed
And
to
-a,
or of any other negative scain each of the two last
if
a>0,
as to the general n^^ root

of a quaternion^
we may
on the same principles,
VIIL.. Q^=l^.j;/Q;
where the
factor 1, representing the general n*^ root of positive

has
n
unity,
different values, depending on the division of the circumference of a circle into n equal parts, in the way lately illustrated, for the case 7i = 3, by Figure 54 ; and only differing from
ordinary algebra by the reality here attributed to i. In fact, each

of these n*^ roots
of unity is with us a real versor; namely the quotient of two radii
of a circle, which make with each other an angle^
equal to the
n^^
part of some whole number of circumferences.

I
We propose,
however, to interpret the particular symbol

noting the principal value of the n<^ root of i thus writing,
(1.)
t",
as always de-
IX.
whence
in
= -v/i
when this root is expressed under the form of a couple

(228), the two constituents x andy shall both be positive, and the quotient y: x
shall have a smaller value than for
any other couple x + iy (with constituents thus
it
will follow that
which the n^^ power equals i.

(2 ) For example, although the equation
positive), of
52
satisfied
by the two
values,
(1
= (x-|-iy)2 = t,
^2,
we
2k
shall write definitely,
250
(3.)
And
ia satisfied
[BOOK
II.
although the equation,
by the
three distinct and real couples, (
+ V3)
2,
and -
i,
we shall adopt
only the one value,
XI.
we
(4.) In general,
have the expression,
shall thus
XII.
= V =
i4
= cos - + 1 sin 2n
which we
shall occasionally abridge to the following
Xir.
and
2'
= CIS
2n'
thus interpreted, denotes a versor, which fwrns any line on which it

an angle equal to the n*^ part of a right angle, in the positive direction of rotation, round the given axis of t.
this root, i",
operates, through
and n be
ani/ two positive whole numbers, and q

the
definition
contained in the formula 233,
any quaternion,
of
whole
the
us to write, as in algebra,
enables
II.,
poiver, q^,
234. If
the two equations

I-
^'"^"
q^"'^
II.
(^")'"
= y"^"
and we propose to extend the former to the case of

negative whole exponents, writing therefore,
III.
and
^''=
and
IV.
mill
q^ri-n^gm.gn
in particular,
V.
We shall also
^"^
=-=
^^
reciprocal''^
extend the formula
VI.
II.,
(^")"
(134) of q.
by writing
q~',
whether m be positive or negative so that this last symbol,

if ?w and n be still whole numbers, whereof n
may be supposed
to be positive, has as many distinct values as there are units in
;
the denominator of its fractional exponent,

*
Compare the Note
to
when reduced
page 121.
to its
CHAP.
AMPLITUDE OF A QUATERNION.
II.]
251
m
terms; among which values of q^,
consider as the principal one, that which
least
the principal
n*^
we
is
shall naturally
the m*^ power of
root (233) of q.
For example, the symbol qi denotes, on
this plan, the square of any cubehas therefore three distinct values, namely, the three values of the cuberoot of the square of the same quaternion q ; but among these we regard as principal,
(1.)
root of 9
it
the square of the principal cube-root (231) of that proposed quaternion.

(2.) Again, the symbol q^ is interpreted, on the same plan, as denoting the
= + Ij this square has only
square of any fourth root of q but because (1*)^ = 1^
two distinct values, namely those of the square root q^, the fractional exponent |
being thus reduced to its least terms; and among these the principal value is the
;
square of the principal fourth root, which square
is,
square root (199, (1.), or 227) of the quaternion
q.
(3.)
The symbol
same
q~i denotes, as in algebra, the reciprocal of
while 5"2 denotes the reciprocal of the square, &c.

(4.) If the exponent t,ma symbol of the form
incommensurable),
at the
we may
a variable fraction tends
q^,
time, the principal
a square-root of q ;
be stiU a scalar, but a surd (or

as a limit, towards which
t,
consider this surd exponent,
and the symbol itself may then be interpreted as the
corre-
sponding limit oi a fractional power of a quaternion, which has however (in this case)
until a selection
indefinitely many values, and can therefore be of little or no use,
shall have been made, of one value of this surd power as principal, according to a law
which will be best understood by the introduction of the conception of the amplitude
of a quaternion, to which in the next Section
we
shall proceed.
Meanwhile (comp. 233), (4.) ), we may already definitely interpret the symV as denoting a versor, which turns any line in the given plane, through t right
(5.)
bol
angles,
ponent,
round
t,
Ax
i,
in the positive or negative direction, according as this scalar eX"
whether rational or
irrational, is itsehf positive or negative
and thus
may
establish the formula,
^^^
..
VII.
t*
= cos
f
*7r
'
sm
tir
;
^j
or briefly (comp. 238, XII'.),
VIII. ..i' = cis^.

2
Section
3.
On
the Amplitudes of Quaternions in a given
and on Trigonometric Expressionsfor such Quaterand for their Powers,
Plajie;
nions,
235. Using the binomial or couple form (228) for a quaternion in the plane of i (225), if we introduce two new and
real scalars, r and z, whereof the former shall be supposed to
be positive, and which are connected with the two former scalars X and y by the equations,
I.
^=
r cos
2;,
?/
= r sin
2,
r > 0,
252
we
[bOOK
have the formulae (comp. 228,

.Tq = T(x + ii/) = r;
shall then evidently

11.
III.
JJq
(re
= cos z +
iy)
(5.)
II.
i8mz;
last expression may be conveniently abridged (comp.

233, Xir., and 234, VIII.) to the following
which
IV.
U^
= cis
2: ;
Y.
so that
g^rclsz.
the arcual or angular quantity, z, maybe called the Amplitude^ of the quaternion q this name being here preferred
And
'^
by us to
Angle'' because we have already appropriated
the latter name, and the corresponding symbol L q, to denote
(130) an angle of the Euclidean kind, or at least one not ex-
and tt ; whereas the

ceeding, in either direction, the limits
as
considered
obliged only to satisfy the equaamplitude, z,
tions
I.,
may have any
We shall denote
real arid scalar value.
this amplitude, at least for the present, by iho, symbol,^ am.^,

or simply, am q ; and thus shall have the following formula,
of connexion between amplitude and angle,
VI.
{z=)Qm.q = 2mr Lq',
*
t The symbol V was spoken of, in 202,
peculiar to the present Calculus
and
as completing the system of notations
in fact, besides the three letters,
i,
j, h, of
which
the laws are expressed by ihQ fundamental formula (A) of Art. 183, and which were
originally (namely in the year 1843, and in the two following years) the only peculiar
symbols of quaternions (see Note to page 160), that Calculus does not habian}'- more than the_^ye characteristics of
tually employ, with peculiar significations,
operation,
part)
K,
S, T,
U, V, for conjugate, scalar, tensor, versor, and vector (or right

mark N for norm, which in the present work has been
although perhaps the
adopted from the Theory of Numbers, will gradually come more into use than
As to the marks, I, Ax., I, R,
it has yet done, in connexion with quaternions also.
and now am (or am^), for angle, axis, index, reciprocal, and amplitude, they are to
be considered as chiefly available for the present exposition of the system, and as not
and the same remark
often wanted, nor employed, in the subsequent ;?racfce thereof
.
applies to the recent abridgment
cis,
for cos
sin
to
some notations
in the present
Section for powers and roots, serving to express the conception of one n*^ root, &c.,
and to the characteristic P, of what we shall call in the
as distinguished from another
;
next section the ponential of a quaternion, though not requiring that notation afterwards. No apology need be made for employing the purely geometrical signs, -i-,
for perpendicularity, parallelism, and complanarity : although the last of
j|, 111,
them was perhaps

useful.
first
introduced by the present writer,
who has found
it
frequently
CHAP.
ADDITION AND SUBTRACTION OF AMPLITUDES.
II.]
253
the upper or the lower sign being taken, according as Ax. q

= Ax. i ; and n
being any whole number, positive or negative
We
or null.
may
then write also (for any quaternion ^
the general transformations following
VII.
(1.)
tity,
U5'
Writing q
to the line
cis
= ^:
/3
am
5'
VIII.
5'
= Tj
.cis
am q,
amplitude am. gr, or am (/3 a), is thus a scalar quanproper sign) the amount of rotation, round Ax. i, from the
and admitting, in general, of being increased or diminished by
expressing (with
line
|||
a, the
its
any whole number of circumferences, or oi entire revolutions, when only the

tions of the two lines, a and /3, in the given plane of i, are given.
direc-
(2.) But the particular quaternion, or right versor, i itself, shall be considered
as having definitely, for its amplitude, one right angle; so that we shall establish the
particular formula,
IX.
am.i
z. 1
= -.
When, for any other given quaternion q, the generally arhitrary integer
VI. receives any one determined value, the corresponding value of the amplitude may be denoted by either of the two following temporary symbols,* which we
(3.)
in
here treat as equivalent to each other,
am .q,
so that (with the
same
X.
may
Z 9
rule of signs as before)
mula than VI., the equation
and
or
we may write,
as a
more
definite for-
am,! .q
= inq = 2mr Lq;
say that this last quantity is the n^^ value of the amplitude of q while the
amo q, may be called the principal amplitude (or the principal value of
;
zero-value,
the amplitude).
(4.)
With
and with the convention, amo ( 1)
these notations,
= + tt, we may
write,
XI.
XII.
am a
amo q = loq = iq]

am,j 1
z,i 1
= 2ra7r,
if
>
and
XIII.
if
a be
still
am (- a) = am(- 1) =
Z (- 1) = (2 + 1) tt,
a positive scalar,
236. From the foregoing definition of amplitude, and from

the formerly established connexion of multiplication of versor^
with composition of rotations (207), it is obvious that (within
the given plane, and with abstraction made of
tensors) multiplication
*
and
division
Compare
of quaternions answer respectively to
the recent Note, respecting the notations

employed.
254
[bOOK
II.
and subtraction of amplitudes : so that,

the symbol am 5' be interpreted in the general (or indefinite)
sense of the equation 235, VI., we may write
(algebraical) addition
if
I.
am {q'. q) = am q + am q\
member
am {q'l q) = am q'- am q
that, in each formula, ojie of the values of the
implying hereby
first
II.
among the values of the second member ; but
is
not here specifying ivhich value. With the same generality

of signification, it follows evidently that, for a product of any
number of (com planar) quaternions, and for a lohole power of any
we have
one quaternion,
III.
the analogous formulae
amn2' = Samg'
IV.
2im.q^
^p.oxaq
where the exponent p may be any positive or negative integer,

or zero.
was proved,
in 191, II., that for any two quaternions, the formula Vq'q
a result which, by the associative principle of multiplication
(223), is easily extended to any number of quaternion factors (complanar or diplaso that we may write, generally,
nar), with an analogous result for tensors
(1.) It
= U^'. \]q
holds good
V.
uiig = nuj
VI.
Tn^ = nxg.
(2.) Confining ourselves to the first of these two equations, and combining
III., and with 235, VII., we arrive at the important formula :
VII.
whence
n cis am 2 (= nUj = Ullg' = cis am Hq) = cis 2 am g
at least if the exponent p be
In these
2,
IX.
z
.
&c.
(cis
am q)p cis(p am '),

.
any whole number.

the amplitudes am. 5, am. 9',
we may
n cis z = cis Sz
and
sines of
X.
the
if
number
thus,
IX'.
(cis z)P
known and
&c.,
may represent any
thus,
= c\spz
useful theorems, respecting
sums and multiples of arcs.
For example,
whence
them
therefore write
including thus, under abridged forms, some
(4.)
still
last formulae,
angular quantities,
cosines
with
in particular (comp. IV.),
VIII.
(3.)
it
cis z'
cis
form
of factors of the
= cis (z' + z)
X'.
cis
(cis z)^
z be two,
= cis 2z
cos (z' + z) = S (cis z' cis z) = cos 2' cos z - sin 2' sin z
sin (z' + z) = i-i V (cis z' cis z) = cos z' sin z + sin z' cos z
sin 22 = 2 cos sin z
cos 2z = (cos zy (sin z)^
.
we have
with similar results for more factors than two.

(6.)
Without expressly introducing the conception, or at least the notation of

we may derive the recent formulae IX. and X., from the consideration of,
amplitude,
the power
i<
(234), as follows.
That power
oft, with a scalar exponent,
t,
has be
CHAP.
255
POWERS WITH SCALAR EXPONENTS.
II.]
symbol satis^nng an equation which
interpreted in 234, (6.), as a
may
be written
thus:
XI.
We
be any scalar.
same
or
any other
t*
= cis z,
see then at once,
= îr
if
which turns a
or geometrically as a versor,
line
right angles, where
through
may
this interpretation, that if i' be either the
from
scalar, the formula,
XII.
V V = iM',
And
must hold good, as in algebra.
XIII.
or
t<
= ^*,
because the number of the factors
seen to be arbitrary in this last formula,
XIV.
we may
is easily
write also,
.(i^> = iP^
p be any whole* number. But the two last formulae may be changed by XI., to
the equations IX. and X., which are therefore thus again obtained ; although the
later /orms, namely XIII. and XIV., are perhaps somewhat simpler: having in-
if
deed the appearance of being mere algebraical identities, although we see that their
geometrical interpretations, as given above, are important.
(6.) In connexion with the same intei^retation XI. of the same useful symbol i\
it
mav
be noticed here that
XV. ..K.t* = t-';

and that
therefore,
XVI.
cos
= S.
t-rr
XVII.
sm
i-1
i'
= i(i< + i-0
V. i* = i e-i
(i*
- i'O-
(7.) Hence, by raising the double of each member of XVI. to any positive whole
power p, halving, and substituting z for ^tir, we get the equation,
XVIII.
2P-1 (cos z)P = i (i
+ i-*)P = I (iP + t'Ô + Ip (i^P-^)* + i(2-^)0 + &c
= cos joz+p cos(/) - 2)z

with the usual rule for halving the
+^-^
coefficient of cos Oz, if
cos (p
- 4) 2+ &c.,
be an even integer ; and
with analogous processes for obtaining the known expansions of 2-P"' (sin z)p, for any
positive whole value, even or odd, of p ; and many other known results of the same
kind.
237. lip be
still
a whole number,
we have thus
the transforma-
tion,
I.
in
jp
= (r cis zy = rP
cis pz
which (comp. 190, 161) the two

may be thus written
kinds,
= i^qY cis {p am
.
factors, of the tensor
and versor
II.
T {qy = {Tqy = T^^
III.
U (q^) = (U^)^ = Uj"
and any value (235) of the amplitude am.q
g') ;
may be
taken, since all
be seen that there is a sense, although one not quite so definite, in

formula holds good, even when the exponent p is fractional, or surd ;
namely, that the second member is then one of the values of the first.
which
It will soon
this
256
will conduct to one

for
I.,
we
common value
of this whole power
IV.
n',
(qP)n
= T/
cis
^=
(p am q\ with
.
n being whole numbers whereof the
to the second, so that the exponent
prime
And
q^.
if,
substitute this slightly different formula (comp. 235,
(3.)),
m',
II.
[bOOK
supposed to be
first is
here
is
n^>0,
2,
fraction in
its
with a
positive denominator n\ while the factor Tq^ is

a.
as
positive scalar (of which the positive or negative
interpreted
logarithm, in any given system, is equal to ^ x the logarithm of Tq)y
least terms,
then the expression in the second member admits of n' distinct vaanswering to different values of n ; which are precisely the n'
lues,
values (comp. 234) of the fractional power q^, on principles already

established the principal value of that power corresponding to the
:
n=0.
value
For
(1.)
anj/ value of the integer n,
by the formula
recurring periodically, when/?
Abridging (1p) to
(2.)
V.
a
restriction
(4.)
(-
l)p,
as above,
we have
a,
that the symbol (qP)n, defined

such values, however,
power qP
thus, generally^
shall soon
Up
by 235, XII.,
be any fraction,
remove
and
in particular,
Principal value
1*0=+!,
Iin=cisn7r,
27r
Hn = CIS -
.
,.
1*0
= ,1,
,,
Ih
fraction.
oflP= IPq =
making successively ;> = |, p = i, we have
.
^.
.
1^,,,
which however we
(3.) Thus,
_.^
VIII.
is,
lP = cis 2pmr,
VI.
Vn.
we may say
IV., represents the n*^ value of the
1.
lî=-l,
-l + tV3
^.
1*2
1*2
= +!,
-l-tV3
=
r
&c.;
= 1,
.
^.
,
1*3
&c.
Denoting in like manner the n'* value of (- 1)p by the abridged symbol
have, on the same plan (comp. 235, XIII.), for any fractional* value
we
ofp,
IX.
(j-iyn = cisp(2n+ l)7r; whence (comp. 232),
X...(-l)io = cis|= +
.-,
(-l)ii
cis^
= -i,
(-l)i2
= + t,
&c.;
and
xi...(-i)^=l:fi^
these three values of
(5.)
(-
XII.
n*'
(-,)^=L^%,,
l)i recurring periodically.
The formula IV.
so that the
(-1).,=-,,
value of qP
gives, generally,
.
is
=
(qp)n
by V., the transformation,
(qPô cis
2pn7r
lPn(qP)o
equal to the principal value of that power of
* As
before, this restriction
is
q,
multi-
only a temporary one.
CHAP.
PONENTIAL OF A QUATERNION.
II.]
257
plied by the corresponding value of the same power of positive unity } and it may be
remarked, that if the base a be any positive scalar^ the principal jp'^ power ^ (<^)o,
is simply, by our definitions, the arithmetical value of aP.
n^^ value of the p^^ power of any negative scalar, a, is in like man(G.) The
ner equal to the arithmetical p^'* power of the positive opposite, +a, multiplied by
tlie
corresponding value of the same power of negative unity; or in symbols,
XIII.
(7.)
The formula
(- a)Pn = (- 1)P (aP)o
= CaP)o cis p (2n + 1) tt.
IV., with its consequences V. VI. IX. XII. XIII.,
extended so as to include, as a
when
limit, the case
the exponent p being
may
still
be
scalar,
becomes incommensurable, or surd; and although the number of values of the power
qp becomes thus unlimited (comp. 234, (4.)), yet we can still consider one of them
as the principal value of this (now) surd power : namely the value,
XIV.
which answers
We may
238.
to the principal
(g'P)o
= TqP
cis
(p amo
9),
amplitude (235, (3.)) of the proposed quaternion
q.
therefore consider the symbol,
^^
in
which the
any quaternion, while the exponent, p,

as
now
any scalar^
being
fully interpreted; but no interpretation has been as yet assigned to this other
symbol of the
q, is
base,
is
same kind,
in
qq'^
which both the base
q,
and the exponent q\ are supposed
to be (generally) quaternions, although for the purposes of this

Chapter complanar (225). To do this, in a way which shall
be completely consistent with the foregoing conventions and

conclusions, or rather which shall include and reproduce them,
where the new quaternion exponent, q, degenerates

a scalar, will be one main object of the following
for the case
(131) into
Section which however will also contain a theory of logarithms of quaternions, and of the connexion of both logarithms
and powers with the properties of a certain function, which
:
we
shall call the ponential
of a quaternion, and to consider
which we next proceed.
Section
4.
On
ternion;
the Ponential and Logarithm

of a Quaterand on Powers of Quaternions, with Quaternions
for their Exponents.

239. If
we
consider the polynomial function,

I.
P(^,
w)=lt^, + ^2 +
2 L
..<7,,
258
in which q
while
it is
any quaternion, and
is
is
any positive whole number,
n
'^^
'
(^
r(m+l)/
1.2.3..wiV
it is
II.
supposed (for conciseness) that

TT
then
[bOOK
not difficult to prove that however great, hut finite and

Tq may be, a Ji72ite number m can be assigned, for
given^ the tensor
which the inequality

III.
T(P(^, m-\-n)-V{q,m))<a,
if
shall be satisfied, however large the (positive whole)
be,
is
>0,
number n may
and however small the (positive)
given.
In other words,
lY,
.q = x +
a finite value of the
if
scalar a, provided that this last

write (comp. 228),
we
V{q, m) = X, + iY^,
iy,
number
can always be assigned, such that the
ra
following inequality,
V.
shall hold good,
(X,,,-X,)' + (r,.-
(but given and >0) the

each of the two scalar
VI.
converges ultimately
to
may
x,=
\-vx,
a fixed and finite
and the
Ave
must consider
following
series
this
limit,
Y^,
whereof the one may be

and of which each
x and
y.
Writing then
Q = Xoo+iYcK>=X+iY,
quaternion Q (namely the limit
-r,, ..
a;+"'^^V.
latter Foo, or F,
a certain function of the two scalars,
VIII.
and however small
T, = y-^xy,..
Y,=y,
called Xoo, or simply JT,
n,
It follows evidently that
be.
or succession of scalar functions,
series,
x,=
..Xo=i,
scalar a
VII...ro = 0,
is
7j2<aS
however large the number
to
which the
of quaternions^
IX...P(^,0)=1, P(^,l)=l+^, P(^,2)=l + g + |',.. P(?,m),...

converges ultimately) as being in like
we
manner a ce,xtQ.\nfunctioji, which
shall call the ponential function, or simply the Ponential of q, in
consequence of its possessing certain exponential properties

which may be denoted by any one of the three symbols,
P (?.
We have therefore
X.
Qo),
or
P {q),
or simply
P^.
the equation,
Ponential
ofq=Q-'Pq=\-\-qx +
Avith the signification II. of the
term q^.
q2-\-..+q:g>i
and
CHAP.
EXPONENTIAL PROPERTY.
II.]
259
(1.) In connexion with the convergence of this ponential series, oi" with the in= Tq, and r, = Tq,i
it may be remarked that if we write
(conip. 235) r
equality III.,
we shaU have, by 212,

XI.
it
is
sufficient
terms, r,n+h
shall
(2.),
m + n)-P(g,m))<P(r, m + n)-P(r, 771);
T(P(^,
then to prove that this last difference, or the sum of the n positive
Now if we take a number p>2r -1, we
rm+n, can be made <.
have rî <|rp, rp<.2<
be assigned, such that
XII.
P (r, w +
the asserted inequality
is
Ao
a finite number
m>p>2r -
can
and then,
- P(r, m) < a(2-i + 2-2 +
n)
.'+ 2-'0
<a
therefore proved to exist.
an ascending
(2.) In general, if
XIII.
fr^+i, &c., so that
<a
r,/,
+ Aig +
A2g2
series
4-
with positive
where
&c.,
coefficients,
Ao >
be convergent when q is changed to a positive scalar,

when 5' is a quaternion.
Ai > o,
o,
such as
&c.,
will ^fortiori converge,
it
240. Let q and q' be any two complanar quaternions, and

be their sum, so that
I...q^^ = q^-^q,
let
q"
q'^\\\q'\\\qi
with the signification 239, 11. of q^^ and with

corresponding significations of q^^^ and q'^î we have
then, as in algebra,
where
like
5-0
9'o
manner
= l- Hence, writing again r = Tq, 7\ = Tq,, and

= Tq', r" = Tq'\ &c., the two differences,
in
r'
III.
P {i\ m) P (r, m) - P {r"

.
m),
and
IV.
P(r", 2m) -P(r', m).P(r, m),
can be expanded as sums of positive terms of the form r'p,.rp (one

sum containing ^m{m-\- 1), and the other containing m{rn+ 1) such
terms); but, by 239, HI-, the sum of these two positive differences
can be made less than any given small positive scalar a, since
V.
provided that the
P (r^^ 2m) - P {r'\ m) < a,

number
is
therefore, separately tends to 0, as
must
exist
a fortiori, when the
quaternions, q,
Exponential
q',
q".
a>
if
taken large enough
tends to oo
tensors, r, r',
The function Vq
is
0,
each difference,
a tendency which
r", are replaced by the
;
therefore subject to the
Law,
\\,..V{q'^q) = Vq'.Vq=Vq.Vq',
if
q'
\\\
q.
260
we
(1.) If
VII.
write (comp. 237, (5.)
PI =
VIII.
then
c,
II.
),
.
Pa;
= ()o = arithmetical value
off' ;
base of the natural system of logarithms, and x is any scalar.

shall henceforth write simply t^ to denote this principal (or arithmetical) value of
where
We
[bOOK
e is
known
the
the x^^ power of
c,
and so
shall
have the simplified equation,

VIII'.
(2.)
Already we have thus a motive

IX.
Pa;;=*.
for writing, generally,
P9 = ;
here to be considered merely as a definition of the sense in which

we interpret this exponential symbol, e9 ; namely as what we have lately called the
ponential function, Vq, considered as the sum of the infinite but converging series,
but this formula
is
239, X.
however be soon seen
It will
(comp. 238) of the symbol

(3.)
For any scalar
X.
to be included in
we have by VIII.
x,
a more general
definition
q^'.
x = \Px natural
the transformation
logarithm of ponential of x.
241. The exponential law (240) gives the following general decomposition of a ponential into factors,
I.
in
Vq = V{x + iy) = Tx,Viy',
which we have just seen that the
The other
factor Pic
is
a positive scalar.
proved to be a versor, and therefore

to be the versor ofFq, while Pa; is the tensor of the same ponential; because we have in general,
11.
since
is
factor, Viy,
easily
.Pg.P(-^) = PO=l,
IV.
(K$)"' =K{q'")
and
..P% = KPy,
III.
= {s2iy) Kq"" (comp.
199, IX.);
and therefore, in particular (comp. 150, 158),
V.
P2> = P(-
We may therefore
VII.
IX.
iy)
= KFiy,
or
VI.
NFiy =
write (comp. 240, IX., X.),
1.
VIII,
TP^ = PS^ = Pa;=e-;
a;=S^=lTP^;
=
=
=
e^y = cis
P
UP^
V^ Fiy
y (comp. 235, IV.)
.
this last transformation being obtained
X.
XL
from the two
=
SFiy=l-'^ + &c. cosy;
^-'VP^>=?/-|^
+ &c.=siny.
Hence the ponential Fq may be thus transformed

XII.
Fq=F{x\ iy) =
e''
cis y.
series,
CHAP.
CONNEXION WITH TRIGONOMETRY.
II.]
(1) If
we had
lidity
not chosen to assume as known the series for cosine and sine, nor
as that known one on which their va-
any one unit of angle, such

depends, we might then have proceeded
to select (at first)
xiii.
we should
Pi>=/y +
have, by
XV.
\}a!&
%,
as follows.
Writing
/(- y) ^+fy,
<}>(ry)
= - 'Pv^
the exponential law (240),
XIV.
and then
261
/(y + y') = S (Viy
Viy')
.f{y-y)=
^fy .fy' - <\>y
<^y'
fy-fy'^r^y^W-,
functional equation, which results, namely,
XVI.
/(y + y') +f(y -y') = 2/y .//,
would show that
XVII.
whatever unit of angle
by the
^=
cos
-X a
right angle \
be adopted, provided that
may
we determine
the constant c
condition,
XVIII.
XVIII'.
= least positive
of the equation Jy(= SPiy)
root
or nearly,
1-5708, as the study of the series* would show.
motive would thus arise for representing a right angle by this numerical
(2.)
constant, c; or for so selecting the angular unit, as to have the equation (tt still denoting two right angles),
XIX.
TT
= 2c = least positive
giving nearly,
XIX'.
7r
root
= 3-14159,
of the equation fy=
we
should reduce XVII. to the simpler form,
(3.)
As
to the function (py, since
XXI.
it is
sines or
(/y)2
= cosy,
.fy
(0y)2
= Pi>.P(-ty) = l,
y and it is easy to prove that the upper sign is to be

In fact, it can be shown (without supposing any previous knowledge of cosines) that 0c is positive, and therefore that
evident that
taken.
0y = +
as usual;
for thus
XX.
sin
XXII.
0c
= + 1,
or
XXIII.
Pic =
whence
XXIV.
0y = S.i-iPty
= SPt(y-c)=:/(y-c),
and
XXV. ..Pzy=/y + i7(y-c).

If then
we
* In
replace c
fact,
by -,
we have
the value of the constant c
may
be obtained to this degree of

accuracy,
by simple interpolation between the two approximate values of the function/,

/(1-5)
and of course there are
=+
0-070737,
artifices,
/(I -6) =
- 0029200
not necessary to be mentioned here, by which a far
more accurate value can be found.
262
XXVI.
y-
^y = cos
-cos^y-l")
'
(4.)
The
= sin ^
XXVII.
and
Viy =
cis y, as in
II,
IX.
for cosine and sine might thus be deduced, instead of being

and since we have the limiting value,
X. XI.
series
assumed as known
[bQOK
XXIX.
lim.y-i siny
= lim.y-it-i VPiy = 1,
follows that the nnit of angle, which thus gives Piy = cisy, is (as usual) the angle
subtended at the centre by the arc equal to radius ; or that the number w (ôr 2c) is
to 1, as the circumference is to the diameter of a circle.
it
(5.) If
any other angular unit had been,
for
any reason, chosen, then a right
angle would of course be represented by a different number, and not by 1'5708 nearly
but we should still have the transformation,
XXX.
though not the same
242.
Pzy =
cis
-Xa
series as before, for cos
The usual
I.
right angle
y and
unit being retained,
2m7r =
and
II.
j,
sin y.
we
see,
P {q
-\-
by 241, XII., that
2m7r)
= Vq,
n be any whole number it follows, then, that the inverse ponenor what we may call the Imponential, of a given
quaternion q, has indefinitely many values, which may all be repre-
if
tial function, V'^q,
sented by the formula,

III.
and of which each
..P-'^ = lT^ + tam^;
satisfies
the equation,
IV...PP,r'^ = ^;
while the one which corresponds to n-0 may be called the PrinciIt will be found that when the exponent p is any
pal Imponential.
scalar, the definition already given (237, IV., XII.) for the n^^ value
o the p^^ power of q enables us to establish the formula,
V...
(?"),.=
P(pP-'?);
and we now propose to extend this last formula, by a new definition,

to the more general case (238), when the exponent is a quaternion q':
thus writing generally, for any two complanar quaternions, q and q',
the General Exponential Formula,
VI...(e'% = P(2'P.-2);
the principal value of q'^' being still conceived to correspond to n = 0,
or to the principal amplitude of q (comp. 235, (3.) ).
CHAP.
LOGARITHM OF
II.]
For example,
(1.)
VII.
(9)o
= F(qo-h) = Fg,
we
the ponential Tg, which
now
263
QUATERNION.
because
Po-
= Ic = 1
agi-eed, in 240, (2.), to denote simply
by
9, is
therefore
seen to be in fact, by our general definition, the principal value of that power,
or exponential.
(2.)
With
VIII.
these
the same notations,

'V
two last only
= cis y,
cos
diflfering
y = ^ (> + f-'V),
IX,
and we
may
y =
from the usual imaginary expressions
by the geometrical reality* of the versor i.

(3.) The cosine and sine of a guaternion
by the equations
sin
(in the given plane)
('V
-V)
for cosine
and sine,
may now be defined
cos 5
= ^ ( + -9)
X.
sin
(' - -^)
write (comp. 241, IX.),
XI.
cis
= *9 = Fig.
of cis g, the exponential properties, 236, IX., X.,

(4.) "With this interpretation
and we may write,
continue to hold good
;
XII.
(g9')n
= P C^'IT?) P OV amn g) = (Tg)o' cis(g' amn 5)

.
a formula which evidently includes the corresponding one, 237, IV., for the n<^ value
of the p*^
(0.)
power of g, when p is scalar.

definitions III. and VI., combined with 235, XII., give generally,
The
XIII.
l,;^'
(I5')n
=P
(6.)
The same
Po-'i
equation agrees with a
last
2imrg'
XIV.
iq^% = irfi'. (q^')o
definitions give,
XV.
which
formula 237, XII.
this last equation including the
known
XVI.
(tOo
= "F
interpretation of the symbol,

^v-i
considered as denoting iu algebra a real quantity.

extended to the case where the exponent g' is
(7.) The formula VI. may even be
a gvaternion, which is not in the given plane ofi, and therefore not complanar with
the base g
thus
we may
write,
XVII. ..(i;)o=P(iPo-^0
but
it
p(-^y-^;
would be foreign (225) to the plan of this Chapter
to enter into
any further de-
on the subject of the interpretation of the exponential symbol gi', for this case
of diplanar guaternions, though we see that there would be no difficulty in
treating
tails,
it,
after
what has been shown respecting complanars.

*
Compare 232,
(2.),
and the Notes to pages 243, 248.
264
As
243.
i\iQ
regards
given plane),
[bOOK
II.
^ of a quaternion q (in the

which satisfies the
quaternion
any
general logarithm
we may regard
it as
equation,
I.
= Vq' = q;
e'^'
simply the Imponential P"^g, of which the n*^

the
formula 242, III. But the principal impoexpressed by
= 0, may be said to be the prinanswers
which
nential^
(as above) to n
the
or
Logarithm, of the quaternion q, and may
simply
cipal logarithm^
and in
value
this
view
it is
is
be denoted by the symbol,

so that
we may
or
more simply,
write,
I.
still
II.
because
ITUgÛ = 0,
1^
l^
(llq
U^) = IT^ + lU^,
WJq =
anio q.
have thus the two general equations,

=
V.
IV.
V1^ = 1U^;
S1^ 1T^;
which YVq
in
= ITg- + 2 amo ^ ;
and therefore,
III.
We
Po"'2'
the scalar and natural logarithm of the positive
is still
scalar T^'.
(1.)
As examples (comp.
VI.
(2.)
235, (2.) and (4.)
It
= |z7r
The general logarithm

log
of q
this last denoting the n'^ value
q,
;
VIII.
(3.)
VIL
may
or log
1(- 1)
= iV.
be denoted by any one of the symbols,

q,
or (log q)n,
and then we
.
),
(log g)
shall have,
= lj + 2imr.
The formula,
IX.
log
g'g= log q + log
holds good, in the sense that every value of the
5,
first
if
\\\
member
q,
one of the values of
is
the second (comp. 236).
and one value of log

value of qi' = 9'i3
(4.) Principal
of two general logarithms,
The
quotient
(5.)
qi'
= q\q,
X...(Iog,),.Go..).J0||^.
be said to be the ^'eweraZ logarithm of the quaternion, q', to the complanar quaand we see that its expression involves* two arbitrary and indepenternion base, q
while its principal value may be defined to be Iq \q.
dent
may
integers,
As
the corresponding expression in algebra, according to Graves and
Ohm.
CHAP.
EQUATIONS OF ALGEBRAIC FORM.
II.]
Section
5.
On
265
Finite"^ (or
Polynomial) Equations of AlgeComplanar Quaternions ; and on the

Existence ofn Real Quaternion Roots, of any such Equation of the n"* Degree,
braic Fornif involving
244.
We have
seen (233) that an equation of the form,

I.
..^-Q = 0,
is any
given positive integer, and Q is anyt given,
and actual quaternion (144), has always n real, actual,
where n
real,
and unequal quaternion
roots, q,
complanar with
namely,
the n distinct and real values of the symbol Q" (233, VIII.),
determined on a plan lately laid down. This result is, however, included in a much more general Theorem, respecting
Quaternion Equations of A Igebraic Form ; namely, that if

,
qi, qo,
qn be any n given, real, and complanar quaternions,
.
then the equation,

II.
^" +
5'i5'"-i
9-2^"-=
+ 5 =
0,
has always n real quaternion roots, q, q',

and no more
5-^"),
in the given plane; of which roots it is possible however that
.
some, or
all
may become
equal, in consequence of certain
between the n given coefficients.

another statement of the same Theorem,
relations existing
245.
As
if
we
write,
I.
Fnq
= q'' +
qiq'''^+
..+
qn,
the coefficients qi
qn being as before, we may say that every
such polynomial function, Fnq, is equal to a product ofn real,
.
complanar, and linear (or binomial) factors, of the form q-q'',

or that an equation of the form,
11. ..
can hQ proved in
*
By
infinitely
many
?,a.y'mg finite
cases to exist: although
equations,
we merely
we may
not be
intend to exclude here equations with
which has been seen (242) to have infinitely

= 2tn7r, where n may be any whole
represented by the expression q
many
roots,
Tnq=^{q-q){q-q)..{q-q^-^),
all
terms, such as
P^= 1,
number.
It is true that
we have supposed Q
other case, from substituting for
1||
the versor
(225")
U V Q,
but nothing hinders
us, in
and then proceeding as
any
before.
266
able, with
II.
our present methods, to assign expressions for the
terms of the coefficients ^'i,

qnsay that there is always a certain system
real quaternions, q\ &c., ||| i, which satisfies the system of
roots^ q\
246.
ofn
[bOOK
q^^\ in
Or we may
equations, of known algebraic form,
+ 5^^"^ =-qi
= + 52
qq" + q'q" + q'q" +
=
^VV"+.. -^3; &c.
/ + ^" +
III
<j
the difference Tnq - ^nq is divisible by

q q\ as in algebra, under the supposed conditions of complanarity (224), it is sufficient to say that at least one real quater247.
Or because
nion q always exists (whether

satisfies the equation,
we can
IV. ..F^' =
assign
or not), ichich
it
0,
with the foregoing form (245, 1.) of the polynomial function f.

248. Or finally, because the theorem is evidently true for
the case 72=
the case
9'n
while the case 244,
1,
is
satisfied
1.,
has been considered, and
by the supposition
5'
0,
we may,
without essential loss of generality, reduce the enunciation to

the following:
Every equation of the form,*
h-^q(q-q){q-q")-^{q-q'^-'^)-Q,
in
which q, q,
and Q are any n real and given quaternions

whereof at least Q and q may be supposed
satisfied by at least one real, actual, and com.
in the given plane,
actual (144),
is
planar quaternion,
q.
* The
con-esponding /orm, of the algebraical equation of the n^h degree, was proposed by Mourey, in his very ingenious and original
little
work, entitled
La
vraie
theorie des Quantites Negatives, et des Quantites pretendues Imaginaires (Paris,

Suggestions also, towards the geometrical proof of the theorem in the text
1828).
have been taken from the same work
in which, however, the curve here called (in
251) an oval is not perhaps defined with sufficient precision the inequality, here
numbered as 251, XII., being not employed. It is to be observed that Mourey's
book contains no hint of the present calculus, being confined, like the Double Alge:
De Morgan (London, 1849), and like the earlier Avork of Mr. Warren
(Cambridge, 1828), to questions within the plane : whereas the very conception of the
Qttaternion involves, as we have seen, a reference to Tridimensional Space.
bra of Prof.
CHAP.
II.]
GEOMETRICAL EXISTENCE OF REAL ROOTS.
267
249. Supposing that the m-\ last of the n-l given quaterfirst of them are actual,
q^ .q^"-'^^ vanish, but that the
where 7n may be any whole number from 1 to n - 1, and introduc-
nions
n-m
new
ing a
satisfies
real,
known, complanar, and actual quaternion
which
qo,
the condition,
Q
qq
we may
"*>
..
q'^'*
write thus the recent equation
and may (by 187, 159, 235) decompose

IV. ..T/g=l; and Y. .Vfq=],
I.,
it
into the
or
VI.
two following:
.
am/^ = 2;?7r;
some whole number (negatives and zero included).

A, be
give a more geometrical form to the equation, let
and
let it be supposed that a, ^,
line
or
assumed
given
\\\i,
any
and op, os, are n-m + 2 other lines in the
and p, 0-, or OA, ob,
same planes, and that <[>/> is a known scalar function of /a, such that
in
which p
250.
is
To
VII.
a = q'\
13
- q'%
p=
q\
o-
= qoK
and
^^
-^^
\(tI
\osl
/3
OA OB
the theorem to be proved may then be said to be, that whatever sysand s, in a given plane, and whatever
tem of real points, o, A, b,
.
number w, may be assumed, or given, theiê is always at

one real point p, in the same plane, which satisfies the two condi-
positive whole
least
tions:
IX.
T^/t)
X.
am
^p
= 2p7r.
251. Whatever value i\\\i we may assume for the versor (or
unit- vector) U/>, there always exists at least one value of the tensor
Tp, which satisfies the condition IX.
because the function T(pp va-
when T/3 oo, having varied

continuously (although perhaps with fluctuations) in the interval.
Attending then only to the least value (if there be more than one)
nishes with
T/3,
and becomes
infinite
which thus renders T(pp equal to unity, we can conceive a real,

unambiguous, and scalar function Y^t, which shall have the two fol-
of Tp,
lowing properties
XI, ..T,p(ifc) = l;
And
in this
way
Xll.,.T<p(xiyln)<l,
if
a;>0,
the equation, or system of equations,
<
1.
268
XIII. ,,p
= if I,
or
XIV.
Tp =
.Vp=i,
[bOOK
II.
ft,
may be conceived to determine a real, finite, and plane closed curve,

which we shall call generally an Oval, and which shall have the two
following properties
1st, every right line,
or
7âi/,
drawn/rom
ike ori-
gin
any arbitrary direction within the plane, meets the curve
other given points
once, but once only; and Ilnd, no one of the
o, in
n-m
because 0a = 0,3 = = 0.
252. This being laid down, let us conceive a point p to perform
one circuit of the oval, moving in the positive direction relatively to the
A, B,
on
is
the oval,
given interior point O; so that, whatever the given direction of the

line OS may be, the amplitude am (/j iff), if supposed to vary conti-
have increased by four right angles, or by 27r, in the

course of this one positive circuit ; and consequently, the amplitude
of the left-hand factor (pia)"", of 0p, will have increased, at the same
nuously,'^ will
Then, if the point A be also interior to the oval, so

must be prolonged to meet that curve, the ray ap will
have likewise made one positive revolution, and the amplitude of the
But if a be an exterior
factor (^p - a) a will have increased by 27r.
oa
the
the
in a point M, and
that
line
intersects
so
curve
finite
point,
time,
by
2m7r.
that the line oa
therefore never meets
it
again
if
prolonged, although the prolonga-
ao must meet it once in some point n, then

while the point p performs first what we may call the positive halfcircuit from M to N, and afterwards the other positive half-circuit
from N to M again, the ray ap has only oscillated about its initial and
tion of the opposite line
final direction,
namely that of the
opposite direction
if still
line ao,
without ever attaining the
in this case, therefore, the amplitude
am(AP:
oa),
supposed to vary continuously, has only fiuctuated mils x&\Me,
and has (upon the whole) undergone no change at all. And since
precisely similar remarks apply to the other given points, b, &c.,
follows that the amplitude, am 0P, of the product (VIII.) of all
these factors, has (by 236) received a total increment =2{m-Yt)'7r, if
t be the number
(perhaps zero) of given internal points, a, b,
;
it
while the number
forms (as above)

at least
form
2p7r;
satisfied.
times,
is
(by 249)
at leasts
one. positive circuit,
and therefore
a value of the
X. has been at least once
at least once, through
and consequently the
But the
Thus, while p perthe amplitude am <pp has passed

1.
conditio?i
other condition, IX., is satisfied throughout,
by the
* That is, so as not to receive any sudden

increment, or decrement, of one
more whole circumferences (comp. 235, (!))
or
CHAP.
II.]
GEOMETRICAL ILLUSTIIATIONS, QUADRATICS.
supposed construction of the oval
there
is
269
therefore at least one real
upon that curve, for which 0p or fq = 1 ; so that, /or this

of
that point, the equation 249, III., and therefore also the
position
equation 248, I., is satisfied. The theorem of Art. 248, and conse-
position P,
quently
also,
245 and 246,
by 247, the theorem of 244, with its transformations

is therefore in this manner proved.
253. This conclusion
is
so important, that
ful to illustrate the general reasoning,
it
may
by applying
be use-
it
to the
case of a quadratic equation, of the form,
P P
or 11.
I.
'
9o\q
We have now to prove (comp. 250, VIII.) that a

exists,
OP AP
OS OA
(real) point p
which renders the fourth
proportional (226) to the three

lines OA, op, ap equal to a
given line os, or ab, if this latFig. 55.

ter be drawn = os; or which
satisfies the following condition of
similarity of triangles
(118),
III.
which
A AOP a PAB
includes the equation of rectangles,
I V.
OP.AP = OA-AB.
(Compare the annexed Figures, 55, and

55y bis.)
Conceive, then, that a contiFig. 65, his.
described as a locus (or
as part of the locus) of p, by means of this
equality IV., with
the additional condition
nuous curve*
when
is
necessary,
that
shall be luithin it; in
a manner that
when
such
(as in
Fig. 56) a right line from

o meets the general or total
Fig. 56.
locus in several points, m.

*
This curve of the fourth degree is the well-known

Cassinian; but when it
breaks up, as in Fig. 56, into two separate
ovals, we here retain, as the oval of the
proof, only the one round o, rejecting for the present that round A.
270
[bOOK
II.
but the point m which is nearest to o> as not

1
25
belonging (comp.
XII.) to the oval here considered. Then
while p moves upon that oval, in the positive direction relatively to o, from M to n, and. from n to m again, so that the
we
m', n',
reject all
ray op performs one positive revolution, and the amplitude of

the factor op os increases continuously by 27r, the ray ap
:
performs in like manner one positive revolution, or (on the

whole) does not revolve at all, and the amplitude of the factor
AP OA increases by
:
27r
or
by
according as the point a is inIn the one case, therefore, the
0,
terior or exterior to the oval.
amplitude amcpp of the pj'oduct increases by47r (as in Fig. 55,

bis)
and in the other
so that in each case,
case, it increases
(as in Fig.
by 27r
56)
passes at least once through a value of

the form ^pir, whatever its initial value may have been. Hence,
for at least one real position, p, upon the oval, we have
V.
am 0p =
it
and therefore
..T0^ =
VII.
but
VI.
U^jO
1,
throughout, by the construction, or by the equation of the locus

the geometrical condition 0/o = 1 (II.) is therefore satisfied
IV.
by
and consequently the quadratic

by at least one real quaternion
=
But
the recent form I. has the same
(250,
VII.).
q p:X
at least one real vector p
equation fq
root^
(I.) is satisfied
generality as the earlier form,
VIII.
where
q^
and
quaternions
Foq
(f
+ q^q +
q^Ô (comp.
245),
q^ are any two given, real, actual, and complanar

thus there is always a real quaternion q[ in the
given plane, which
satisfies
the equation,
= q^ +
= (comp. 247) ;
qxq +^2
and
dividing by q-q\ as in algebra
subtracting, therefore,
Vlir.
(comp. 224),
we
F2^'
obtain the following depressed or linear equa-
tion q,
IX.
5'
+ 5''+
^1
0,
or
The quadratic VIII. has
IX.
5'
/ = - 5''-5'i (comp. 246).
therefore a second real quatertiion root,
q, related in this manner to ê first ; and because the quadratic function T2q (comp. again 245) is thus decomposable
into fivo linear factors, or can be put under the form,
CHAP.
RELATIONS BETWEEN THE ROOTS.
II.]
X.
271
T^q=(q-^) (q-q"),
cannot vanish for ant/ third real qtmternion, q ; so that

(comp. 244) the quadratic equation has no more than two such
it
real roots.
(1.)
The
cubic equation
X.
Vzq
may
therefore be put under the ybrm (comp. 248),
= q^-^q\q^ + ?2? + ?3 = 9 (? -
O (9-9") + 93=
the general proof (252"), which has been
has therefore one real
root, say q\ hy
above illustrated by the case of the quadratic equation; subtracting therefore (compare 247) the equation F^q^ = 0, and dividing by 9 q\ we can depress the cubic to
a quadratic, which wHl have two new real roots, q" and 5'" ; and thus the cubic
it
be put under the form,
may
function
XI.
which cannot vanish

fore
for
F39 = (9
any fourth
-9
(9
9") (9
real value of 9
-9
"),
the cubic equation X. has there-
no more than three real quaternion roots (comp. 244)
and similarly
for
equa-
tions of higher degrees.
(2.) The existence of two real roots q of the quadratic I., or of two real vectors,
p and p\ which satisfy the equation II., might have been geometrically anticipated^
from the recently proved increase = 47r of amplitude 0p, in the course of one circuit,
for the case of Fig. 55, bis, in
p and
p', ore
milarity III.
lighter^ oval,
consequence of which there must be two real positions,

which each satisfies the condition of si-
the one oval of that Figure, of
and
for the case of Fig. 56,
which
in this case exists,
from the consideration that the second (or
although not employed above,
is
related to
exactly as the^Vs^ (or dark) oval of the Figure is related to o so that, to the real
position p on the first, there must correspond another real position p', upon the se;
cond.
As regards
(3.)
the law of this correspondence,
if
the equation II. be put under
the form,
and
if
we now
write
XIII. ..p
= 5a,
we may
write
XIV. ..91 =
q2
1,
= (J:a,
and then the recent relation IX'. (or 246) becomparison with the form VIII.
tween the two roots will take the form of the following relation between vectors,
for
XV.
+ p' = a
or
XV'.
op'
= p' = a-
= PA
so that the point p' completes (as in the cited Figures) the parallelogram opap',
and
by the middle point c of OA. Accordingly, with this position
we have (comp. III.) the similarity, and (comp. II. and 226) the equation,
the line pp'

of
p',
is
bisected
XVI.
(4.)
The
Aop' a p'ab
XVII.
other relation between the
two
<pp'
= ^ (a p) = ^p = 1.
roots of the quadratic VIII.,
(comp. 246),
XVIII
q'q"
92,
gives
XIX.
p'
=-
(T
namely
272
and accordingly, the line ct, or

and Ap, or a, p, and p'.
The actual
(5.)
planar quaternions,
a fourth proportional
os, is
[bOOK
II.
to the three lines oa, op,
by calculation, of the quadratic equationYIlI. in cornperformed exactly as in algebra ; the formula being,
solution,
is
XX. ..9=-5iVa9i2-52),
in which, however, the square root
ciples already laid
to
is
be interpreted as a real quaternion, on prin-
down.
(6.) Cubic and biquadratic equations, with quaternion coeflScients of the kind
considered in 244, are in like manner resolved by the known ybrmwZir of algebra;
but we have now (as has been proved) three real (quaternion) roots for the former,
and four such
real roots for the latter.
The
254.
following
is
another
mode
of presenting the geometri-
cal reasonings of the foregoing Article,
without expressly intro-
The equation
ducing the notation or conception of amplitude.

= 1 of 253 being written as follows,
0)0
I.
<T
we may
= x/)=^(p-a),
a
or II.. .T(r=Tx/>, and III.
thus regard the vector
or the point
s as
o-
as a
Uo-=Ux/J,
known function of the
vector p,
a function of the point p; in the sense that, while o
and A arefxed, P and
s vary together : although it may (and does) hapto a former position without p having similarly
return
may
Now the essential property of the oval (253) may be said
pen, that s
returned.
to be this: that it
tensor
of
OS,
is the
locus
of the points p nearest
Txp has a given value, say b namely the

when the point s, like o and a, is given.
;
to o,
for which the
given value of To-, or

If then we conceive
the point p to move, as before, along the oval, and the point s also to
move, according to the law expressed by the recent formula I., this
latter point must move (by II.) on the circumference of a given circle
(comp. again Fig. 6^), with the given origin o for centre ; and the
theorem
is,
that in so moving, s will pass, at least once, through every
position on that circle, while p performs one circuit of the oval.

this
And
observing that (by III.) the angular motion of

equal to the sum of the angular motions of the two rays,
may be proved by
the radius
os
is
OP and ap; but this latter sum amounts to eight right angles for the
case of Fig. 55, his,, and to four right angles for the case of Fig. 56;
the radius os, and the point s, must therefore have revolved twice in
first case, and once in the second case, which proves the theorem
the
in question.
(1.) In the first of these
the three angular velocities
two
is
cases,
namely when a is an interior point, each

and the mean angular velocity
positive throughout,
oj
CHAP.
CASSINIAN OVALS, LEMNISCATA.
II.]
273
the radius os is double of that of each of thefuô rays op, ap. But in the second case,
when A is exterior, the mean angular velocity of the ray ap is zero; and we might
for a moment doubt, whether the sometimes negative velocity of that ray might not,
for parts of the circuit, exceed the
always positive velocity of the ray op, and so

for a while.
This cannot be, however ; for
move backwards^
cause the radius os to
we conceive p to describe, like p', a circuit of the other (or lighter) oval^ in Fig. 56,
the point s (if still dependent on it by the law I.) would again traverse the whole of
the same circumference as before ; if then it could ev&c fluctuate in its motion, it
if
would pass more than twice through some given series of real positions on that circle,
during the successive description of the two ovals by p ; and thus, within certain
limiting values of the coefficients, the quadratic equation would have more than two
real roots : a result which has been proved to be impossible.
(2.) While s thus describes a circle round o, we may conceive the connected point
to describe an equal circle round A and in the case at least of Fig. 56, it is easy
;
from the constant equality (253, IV.) of the rectangles op. ap

and OA. AB, that these two circles (with t'u and t'u' as diameters), and the two ovals
(with MN and m'n' as axes), have two common tangents, parallel to the line oa,
to prove geometrically,
which connects what we
new
or third circle,
call the
may
which
two given foci (or focal points), o and
OA
described on this focal interval
is
the
as diameter, passing
through the four points of contact on the ovals, as the Figure may senê to exhibit.
(3.) To prove the same things by quaternions, we shall find it convenient to
change the origin (18), for the sake of symmetry, to the central point o; and thus
now cp by p, and ca by a, writing also ca = Ta = tr, and representing still
the radius of each of the two equal circles by b.
We shall then have, as ihQ joint
to denote
equation of the system of the two ovals, the following
IV.
=
T(,o + a).T(p-a)
2a*;
or
V.
But because we
VI.
T(92_i)=2c,
ha.ye generally
S ^2 = 2S52
.
VII.
whereof the
=^
and
-.
(by 199, 204, &c.) the transformations,
may (by
210, (8.)
% - 2NV^,
be written under either of the two
(% - 1)2 + 4NV5 = 4c2
first
VIII.
(N5 +
1)2
_ 4NS9 = 4c2
shows that the maximum value of
happens
for this case of Fig.
Tg =
or
1,
- T52 = X52 + 2V52 = 2NSg - N^ =
the square of the equation V.

following forms
if
Tp = a
56; and that this
results which,
when
TVg is c, at least if 2c < 1 as

maximum corresponds to the value
interpreted, reproduce those of the preceding
sub-article.
When
(4.)
then
2c
>
1,
it is
permitted to suppose Sq
= 0, NV9 = N^ = 2c - 1
we have only one continuous
though
>
in that Figure),
before) a
and
oval, as in the case of Fig. 55, bis; but if c

1
~, there exists a certain undulation in the form of the curve (not represented
TV^- being a minimum for
maximum when Tq =
two summits
ai,
<
N,
1,
Sq=
0,
or for p 4- a, but
becoming (as
and vanishing when Sg2
where the oval meets the
= 2c -f
1,
namely at the
axis.
case, when 2c = I, the Cassinian curve IV. becomes

(as
of
which the quaternion equation
a
lemniscata;
known)
may, by V., be written
(comp. 200, (8.) ) under any one of the following forms;
(5.)
In the intermediate
is
2 N
274
IX.
.T(q^-1)
1;
or
X.
.N92 = 28.^2
XI.
[book
.
Tr/
= 2SU
II.
g2
or finally,
XII.
which
when
last,
T|o2
= 2Ta2
cos 2 ^
written as
Xir.
cp2= 2cA*.
cos 2acp,
agrees evidently with known results.

(6.) This corresponds to the case when
and
XIII.
XIV.
=P
+
|,
iu253, XII.,
that quadratic equation having thus its roots equal; and in general, for all degrees,
cases of equal roots answer to some interesting peculiarities ofform of the ovals, on
which we cannot here delay.

(7.) It
may, however, he remarked,
in passing, that if
we remove
the restriction
that the vector p, or cp, shall be in a given plane (225), drawn through the line
which connects the two foci, o and a, the recent equation V. will then represent the
surface (or surfaces') generated by the revolution of the oval (or ovals), or lemniscata,
about that line oa as an axis.
we look back, for a moment, on the formula oi similarity,

we shall see that it involves not merely an equality of rect-
255. If
253,
III.,
but also an equality of angles, aop and pab; so that

the angle oab represents (in the Figures 55) a given difference of the
base angles aop, pao of the triangle gap: but to construct a triangle,
angles, 253, IV.,
of such a given difference, combined with a given base, and

a given rectangle of sides, is a known problem of elementary geomeTo solve it briefly, as an exercise, by quaternions, let the given
try.
by means
base be the line aa', with o for its middle point, as in the annexed
Figure 57 ; let baa' represent the given difference of base angles, paa' - aa'p ; and let oa ab
.
be equal to the given rectangle of sides, ap-jTf.

We shall then have the similarity and equation,
I...Aoa'p<xpab;
II...i-"=!^;
a
a
p
whence
it
follows
by the simplest calculations,
that
III.
or that ^
mean
(^1
-1) +
proportional (227) between a and y3.

Draw,
which shall be in length a geometric mean between the two given lines, oa, ob, and shall also bisect their angle
is
therefore, a line op,
CHAP.
AOB
its
IMAGINARY QUATERNION ROOTS.
II.]
extremity will be the required
of the sought triwhich geometrical syn-
vertex, p,
a result of the quaternion analysis,

thesis* easily confirms.
angle aa'p
275
The equation III. is however satisfied also (comp. 227) by the opposite \ecop' = po, or p' = - p ; and because /3 = (p a) p, we have
(1.)
tor,
+^
IV.
'
p-\^
=^=^=:t
IV'
or
'
a''
= .
..==
p'a
oa op oa'
'
'
so that the /or followiog triangles are similar (the two first of
equal)
V.
as geometry again
the
a'op'
The angles
sum
of the angles in the triangle
a AGP
oc
POB aAP'B
would confirm.
(2.)
ap'b, bpa, are therefore supplementary^ their
oap
whence
it
sum being
equal to
follows that the four points

a,
are concireular:f or in other words, the quadrilateral APbp' is inscriptible

which (we may add) passes through the centre c of the circle oab (see
P, B, p'
in
them indeed being
circle,
Fig. 57), because the angle
agam
aob
is
double of the angle af'b^ by what has been
already proved.
(3.) Quadratic equations in quaternions
may
also be
employed
in the solution
other geometrical problems ; for example, to decompose a given vector into

two others, which shall have a given geometrical mean, &c.
of
many
On the n^-n Imaginary (or Symbolical) Roots

Section 6
a
Quaternion Equation of the n^^ Degree, ivith Coeffiof
cients
of the kind considered
in the
The polynomial function Fnq
256.
foregoing Section,
(245), like the quaternions
which it depends, may always be reduced to the form of

a couple (228) and thus we may establish the transformation (comp.
q, ^1,
q^ on
239),
I.
F,,q
= Fn(x + iy) = X + iY,,=
Gn(co,y)
iB'n{x,y},
Gn and Kn, being two known, real, finite, and scalar

functions of the two sought scalars, x and y\ which functions, rela-
X and
In
r, or
fact, the
two triangles
I.
are similar, as required, because their angles at
and p are equal, and the sides about them are proportional.
t Geometrically, the construction gives at once the similarity,
A AOP a
and
if
POB,
whence
we complete the parallelogram
A oa'p a op'b,
bpa = opa + pad = po a'
apa'p', the
gives
L ap'b
new
similarity,
= oa'p + a'po = aop
thus the opposite angles bpa, ap'b are supplementary, and the quadrilateral apbp' is
It will be shoAvn, in a shortly subsequent Section, tliat these four
inscriptible.
points, a, r, b, p',
form a harmonic group upon their
common
circle.
276
[bOOK
II.
tively to them, are each of the n^^ dimension, but which involve also,
though only in the^r^^ dimension, the 2n given and real scalars,
And
since the one quaternion (or couple) equation^

equivalent (by 228, IV.) to the system of the two scalar
3/1,... Xny yn-
a?!,
F^q = 0,
is
equations,
II.
X =
r, =
0,
or
0,
III.
Gn(x. y)
= 0,
H,,{x, y)
= 0,
we see (by what has been stated in 244, and proved in 252) that
such a system, of two equations of the n*^ dimension, can always be
by n systems (or pairs) of real scalars, and by not more than
satisfied
such as
n,
IV. ..a;^y;
rr^y';..
a;W,
j/C');
although
may happen that two or more of these systems shall coincide with (or become equal to) each other.
it
(1.) If
II. or III.
rem
is,
and y be treated as co-ordinates (comp. 228,
a;
(3.)
represent a system of two curves^ in the given plane
),
the two equations
and then the theo-
that these two curves intersect each other (generally*) in n real points, and
:
although two or more of these n points may happen to coincide with
more
in no
each other.
(2.) Let h denote, as a temporary abridgment, the old or ordinary imaginary,
o( algebra, considered as an uninterpreted symbol, and as not equal to any
V 1,
real versor, such as
(comp. 181, and 214, (3.)
),
but as following the rules of sca-
lars, especially as regards the comi(a<it7e/)roper<y of multiplication (126)
V.
A2
(3.) Let q denote
meaning
just
now
and
=0,
still
VI.
hi
VIII.
form
let
any
[9']
VII.
but
h not
so that
= i.
a real quaternion, or real couple, x + iy

and with the
h, let [] denote the connected but imaginary alge;
proposed of
braic quantity, or bi-scalar (214, (7.)),
and
= ih,
q^x + iy,
x + hy
IX.
but
biquaternion (214), (8.), or (as
so that
[q^
we may
=x+
hy
here call
bi-couple, of the
it)
[g"], be said to be complanar with i; with the old notation (123) of
complanarity.
for the polynomial equation in real and complanar quaternions,
(244, 245), we may be led to substitute the following connected algebraical
equation, of the same degree, n, and involving real scalars similarly :
(4.) Then,
Fnq =
X.
[Fnq-]
= M" +
[?l]
[gy-'
'-
* Cases of
equal roots may cause points of intersection, which are generally imaginary, to become real, hut coincident with each other, and with, former real roots:
for instance the hyperbola, x^
by the pair of right
lines
two pairs of lines, x^
xy
y- =
y^ = a,
= 0,
if
is
intersected in two real
the scalar a
and xy
0,
may
>
or
<
and
distinct points,
but for the case a
= 0,
the
be considered to have four coincident
intersections at the origin.
CHAP.
NEW SYMBOLICAL ROOTS
II.]
OF UNITY.
277
-1
which, after the reductions depending on the substitution V. of

the form,
Xn and Vn
where
are the
XI... [Fnq'\=Xn + hrn = 0;

same real and scalar functions as
But we have seen
(5.)
nish together,
by
in I.
can be made to va-
in II., that these two real functions
selecting ant/ one
ofn
for A', receives
real pairs IV. of scalar values,
.v
and y
the
General Algebraical Equation X., of the n^^ Degree, has therefore n Real or Imagi 1 ; and it has no more than n such roots.
nary Roots,* of the Form a: + y V
Elimination of y, between the two equations
(6.)
to
braic roots (5.), real
and scalar
to
roots, x',
XII.
and which are
degree 2
conducts generally
may
x, of the
;
correspond ; and n (
values of y, which
II. or III.,
which equation has therefore n^ algeor imaginary

namely, by what has been lately proved, n real
y('0 (comp. IV.) of y
x^^\ with real and scalar values y\
an algebraic equation in
1) other roots, with the same
number
of corresponding
be thus denoted,
.
[a;(-i),
[;r("')]
either themselves
XIII.
[y(+i)],
[>(2)]
imaginary (or bi-scalar, 214, (7.)), or at least cor-
respond, by the supposed elimination, to imaginary or bi-scalar values ofy; since if

a;('+i) and y ('+'), for example, could lothh& real, the quaternion equation Fnq=0
would then have an (n + l)st real root, of the form, jO'+i) = ;("+J) H-iy^w+O, contrary
to
what has been proved (252).
257-
On
it results that the equation F,^q = {) in

of the w'* degree, with real coefficients,
the whole, then,
complanar quaternions,
while it admits of only n real quaternion
I.
q\ q'\
j(^
roots,
(244, &c.),
is symbolically satisfied also (comp. 214, (3.)) by n(n- 1) imaginary

quaternion roots, or hy n^ -n bi-quaternions (214, (8.) ), or bi-couples
(256, (3.)
),
which may be thus denoted,
and of which the

III.
where
[^^"^1)]
first, for
example, has the/onw,
= [a;(^)] + zXi/^""')] = a;/^) + Aa?//^"^') + i{y,^''''^ + %//("^'0 ;
a;/"*'\ X/}''"^^\ y/'*'^^\
and
^//^"""'^
the imaginary of algebra (256, (2.)

(1.)
There must, for instance, be n(
are four real scalars, but h is
).
1) imaginary n^^ roots
of unity,
in the
given plane off (comp. 256, (3.)), besides the n reaZ roofs already determined (233,
* This celebrated Theorem

of Algebra has long been known, and has been proved
ways; but it seemed necessary, or at least useful, for the pui-poseof the pre-
in other
sent work, to prove
it
anew, in connexion with Quaternions
the theorem (244, 252), to

which
in the present Calculus
it
or rather to establish
corresponds.
Compare
278
237)
ofl
and accordingly in the case n

two real and two imaginary
\\\i,
IV.
by 256,
for,
(2.),
= 2, we
II.
have the four followmg square-roots
-1;
..+1,
[bOOK
-hi;
+hi,
we have
V.
= ^'' = (- 1)
( ^0'
(-
1)
= + 1.
the two imaginary roots of the quadratic equation J'g^ = 0, which

generality
exist, at least as symbols (214, (3.) ), may be obtained by multiplying the squareroot in the formula 253, XX. by At ; so that in the particular
case^ when that radi-
And
cal vanishes, the four roots of the equation
become real and equal
zero having thus
only itself {or a square-root.

{2.) Again,
we write (comp.
if
VI.
237, (3.)
),
-l + iV3
..9=1*1=
g^
so that 1, q, q^ are the three real cube-roots
and
if
we
80 that
roots
= ll2 = -l-tV3
of positive unity,
in the given plane
write also,
and
6^ are (as usual) the two ordinary (or algebraicat)

imaginary cube-
of unity ; then the nine cube-roots of\(\\\i) are the following

VIII.
9, 92
0, 02
Oq, e^q
9%
O^q^
the two next are real couples, or quaternions

\\\i
the two following are imaginary scalars, or biscalars; and the four that remain are
whereof the
a real scalar
first is
imaginary couples, or bi-couples, or biqnaternions.

(3.)
The
sixteen fourth roots of unity
IX.
..+1;
i;
(||| i)
h;
are:
i(lA)(l0;
hi;
the three ambiguous signs in the last expression being all independent of each other.
(4.)
Imaginary
roots, of this sort, are
calculations respecting ideal intersections,*

in
sometimes useful, or rather necessary, in
and ideal
what remains of the present Volume, we
shall
contacts, in geometry : although
have
little
or no occasion to
employ
them.
(5.)
"We may, however, here observe, that when the
restriction
/?7ane of the quaternion q is removed, the General Quaternion
(225) on the
Equation of the
n*^
Degree admits, by the foregoing principles, no fewer than n* Hoots, real or imaginary : because, when that general equation is reduced, by 221, to the Standard
Quadrinomial Form,
X...Fnq= Wn + iXn +j Vn + kZ,, = 0,

it
breaks up (comp. 221, VI.) into a System of Four Scalar Equations, each (gene-
rally) of the n*^ dimension, in w, x, y, z
XI.
and
..r =
o,
namely,
:^=o,
r=o,
z=o;
z be eliminated between these four, the result is (generally) a scalar

(or
algebraical) equation of the degree n\ relatively to the remaining constituent, w ;
if X, y,
Comp. Art. 214, and the Notes
there referred
to.
CHAP.
RECIPROCAL OF A VECTOR.
II.]
which therefore has n* (algebraical) values,

three other constituents, x, y,
real or
imaginary
and
similarly for the
of the sought quaternion q.
even happen, when no plane
may
(6.) It
z,
279
is
given, that the
become
solutions) of a ^ni7e* equation in quaternions shall
number of roots (or
infinite; as has been
=1
seen to be the case for the equation q^

(149, 154), even when we confine ourwhat we have considered as real roots. If imaginary roots be admitted,
selves to
we may
write, still
more generally, besides the two biscalar
sion,
XII.
..
(-l)i
values,
^v=^v' = ^vv'=.0,
= r+Ar',
h,
the expres-
Nw-Nr'=l;
being thus any two real and right quaternions, in rectangular planes, pro
vided that the norm of i}xQ first exceeds that of the secondhj unity.
and
v'
And
(7.)
in like
manner, besides the two real and scalar values,
this general symbolical expression for a square root of positive unity,
the diflference of the norms reversed
XIII.
li
we have
1,
with merely
Nr' - NtJ =
Su = Sv' = Svv =0,
= + hv',
1.
Section 7.
On the Reciprocal of a Vector, mid on Harmonic Means of Vectors; with Remarks on the Anharmonic
Quaternion of a Group of Four Points, and on Conditions
of Concircularity,
When two vectors, a and a',

Ua Ta, and therefore
258.
I.
a =-
are so related that

II.
a=-
Ua Taf
:
or that
III.
we
Ta
Ta' =
and
1,
shall say that each of these
of the other
Ud + Ua
0,
the Reciprocal]
and shall (at least for the present) denote this
is
and every right quotient v,

VIIL
Râ = RRa = a;
so that for every vector a,
VII.
259.
procals
between them, by writing

V. ..a' = Ra, or VI. ..a = Ra';
relation
and
IX.
IV.
two vectors
is
Ra = -Ua:Ta;
RIt; = IR?; (comp. 161, (3.),
One of the most important
and 204, XXXV'.).

properties of such reci-
contained in the following theorem
Compare the Note to page 265,

t Accordingly, under these conditions, we shall afterwards denote this reciprobut we postpone the use of this notation, until
cal of a vector a by the symbol a"'
we shall be prepared to connect it with a general theory of products and powers of
;
Compare 234, V., and the Note to page 121. And as regards the temporary use of the characteristic R, compare the second Note to page 252.
vectors.
280
If any two vectors oa, ob, have

cals, then (comp. Fig. 58) the right
is
[bOOK
II,
oa', ob' for their reciproline a'b'
parallel to the tangent od, at the origin o,
to the
circle
oab
and the
tioo
triangles,
Or
ob'a', are inversely similar (118).
OAB,
in symbols,
I.
if
ob' =
and
oa =R.0A,
E.ob,
then
A oab a' ob'a.

Of course, under the same
(1.)
conditions, the tangent at
to the circle oa'b' is
parallel to the line ab.
The angles bao and ob'a' or bod being equal, the fourth proportional (226)
and ob, or to ba, oa, and ob, has the direction of od, or the direction op-
(2.)
to AB, AO,
its length is easily proved to be the reciprocal (or inverse)

posite to that of a'b' ; and
of the length of the same line a'b', because the similar triangles give,
II.
it
(oa
ba) ob
.
(ob'
a'b').
ob =
a'b',
being remembered that

III.
we may
OA
oa' =
OB ob' =
therefore write,
IV.
(oa ba) OB
.
whatever two vectors a and
Changing a and
(3.)
/3
= R. a'b',
may
V.
or
(3
= 'R(R(3- Ra),
be.
their reciprocals, the last formula becomes,
/3 to
Ra
VI.
(4.)
.R(/3-a) = rT^.Ri3;
Ra Rp
The
inverse similarity
I.
VII.
or
,^
IX.
^|3
.
by
195,
II., or
(oa'; b'a').ob'
= R. ab.
R/3
207, (2.),
+a
î3
1 = K'-^^,
K-+
a
a^
gives also, generally, the relation,
a
(6.) Since^ then,
we ^have
'
^
X.
Ra+R/3 =
-^
R(3
Ra
:
R(/3a)'
the lower signs agreeing with VI.

the reciprocals of opposite vectors are themselves opposite
(6.) In general,
or
in symbols,
XI.
(7.)
More
XII.
R(-a)=-Ra.
generally,
.
Rxa = x-^ Ra,
X be any scalar.
in X., changing a to y, diNiding, and taking conjugate*,
(8.) Taking lower signs
we find for any three vectors a, /?, y (complanar or diplanar) the formula
if
-^^^^^Ra-R/B
if
a = oa,
=
(5
OB, and y
VR(i3-y)'
= oo,
as usual.
^-a'-y-An'
CHAP. II.] ANHARMONIC AND EVOLUTIONARY QUATERNIONS. 281

we
(9.) If then
extend, to
any four points ofspace^ the notation (25),

^
'
BC DA
and defining that
interpreting QSLch oi ihQ^Q iyfo factor-quotients as a quaternion,
anharmonic quaternion function, or simply the

Anharmonic, of the Group offour points A, B, c, D, or of the (plane or gauche) Qua'
drilateral ABCD, we shall have the following general and useful ^rmwZa of transfortheir
product (in
mation
this order^ is the
where oa',
(10.)
ob', ob' are
With
supposed to be reciprocals of oa, ob, oc.

XIV., we have generally, and not merely for coUinear
this notation
groups (35), the relations
XVI.
(abcd) + (acbd)
XVII.
(abcd). (adcb) =
(11.) Let o, A, b, c, D be any five points, and oa',

OD we shall then have, by XV.,
1.
od' the reciprocals of oa,
XVIII.
^=K
bo'
and
= K (oadc)
^
^
therefore,
XIX.
if
^
DA
(ocba),
^
^
we agree
K (a'b'c'd') = (oadc)
to write generally, for
XX.
any six
(12.) If then the five points
AB CD EP
.
BC DE FA
(abcdef)
= - (oadcba),
points, the formula,*

-
(ocba)
be complanar (225),
we
have,
by 226, and
by XIV.,
XXI.
K (a'b'c'd') = (abcd),
XXI'.
or
(a'b'c'd')
= K (abcd)
the anharmonic quaternion (abcd) being thus changed to its conjugate,

four rays OA, . . od are changed to their reciprocals.
when
the
260. Another very important consequence from the defifrom the recent theo-
nition (258) of reciprocals of vectors, or

rem (259), may be expressed as follows
If any three coinitial vectors oa, ob, oc, he chords of one

common circle, then (see again Fig. 58) their three coinitial re^
* There is a convenience in
calling, generally, t\A^ product of three quotients,
(abcdef), the evolutionary quaternion, or simply the Evolutionary, of the Group
A . . F, or (if they be not collinear) of the plane or gauche Hexagon

because the equation,
of Six Points,
abcdef
(abca'b'c')
=-
1,
expresses either 1st, that the three pairs of points, aa', bb', cc', form a collinear involution (26) of a
well-known kind
or Ilnd, that those three pairs, or the three cor-
responding diagonals of the hexagon, compose a complanar or a homospheric Involution, of a new kind suggested by quaternions (comp. 261, (11.) ).
2o
282
[bOOK
II.
ciprocals, oa', ob', oc', are termino-collinear (24) : or, in other

words, \^ \hQ four points o. A, b, c be concircular, then the three
points
are situated on one right line.

conversely, if three coinitial vectors^ oa, ob', oc', thus
c'
a', b',
And
terminate on one right line, then their three coinitial reciprocals, OA, OB, oc, are chords of one circle; the tangent to which
circle, at the origin, is parallel to the right line; while the
anharmonic function (259, (9.) ), oi \hQ inscribed quadrilateral

OABC, reduces itself to a scalar quotient of segments of that line
(which therefore is
I.
its
(oABc)
own
conjugate,
b'c'
bV = (oo a'b'c') = (o
by 139)
namely,
OABc),
the symbol oo be used here to denote the point at infinity on

the right line a'b'c' and if, in thus employing the notation
(35) for the anharmonic of a plane pencil, we consider the null
if
chord, 00, as having the direction^ of the tangent, od.

(1.) If jO =OP be the variable vector of a point
ternion equation of that circle may be thus written
II.
the coefficient x
Ep =
bemg
E/3
a;(Ea
Or we may
where
III.
thus a variable scalar (comp. 99,
variable position of the point

(2.)
- R/3),
p upon the
circle
oab, the qua-
.x = (oabp)
I.),
which depends on the
p on the circumference.
write,
^
as another form of the equation of the
'
t+u
same
circle
oab
with which
may usefully
be
contrasted the earlier form (comp. 25), of the equation of the line ab,
V.
(3.) Or, dividing the second
we have
p
^
= ta+u(3
.
t+u
member
of IV.
by the
first,
and taking conjugates,
for the circle,
+ ;
VL..^+^=.
a
while
YII.
/^ ^"^ = t + u,
PR
(5
for the right line.
(4.)
Or we may
this latter symbol,
write,
by 204,
Compare the remarks
by
II.,
(18.), denoting
in the second
awy
Note
scalar.
to
page 139, respecting the
determinateness of signification of the symbol UO,

which vanishes according to a laiv.
when
possible
the zero denotes a
line,
CHAP.
CIRCULAR AND HARMONIC GROUP.
II.]
Or
(5.)
etlU
more
IX.
briefly,
V(OABP) =
(6.) It the four points o, a, b,

in their plane, while POi,
we have
the relation,
the ybwr
new points Oi
X.
283
o be
IX'.
still
= V->
(oabp)
coneircular,
pci are the reciprocals of po,
(OiAiBiCi)
or
and
.
if
0.
p be any fifth point
PC, then
= K(OABC) = (OABC) = "
XXL,
by 259,
Ci are therefore generally concircular.
(7.) If, however, the point p be again placed on the circle oabc, those four new
points are (by the present Article) collinear; being the intersections of the pencil
p. OABC with a parallel to the tangent at p.
In this
therefore, we have the
case,
equation,
XI.
(p.
oabc)
(oiAiBiCi)
= (oabc)
anharmonic of the pencil (35) is thus seen to be equal

we have defined (259, (9.) ) to be the anharmonic of the grovp.
so that the constant
And
(8.)
187, (8.) ) to
write, for
because the anharmonic of a circular group

its
own
Q.
(oabc)
= T T (oabc) = + (Ua
falls,
according as the quadrilateral
is,
Hence
(9.)
it is
equal (by
we may therefore
Bc) (ab co),
or the lower sign being

between the points a' and c'
an uncrossed or a crossed one.
of rectangles of opposite sides; the upper
quotient
taken, according as the point b'
that
a scalar,
taken
what
any inscribed quadrilateral oabc, the formula,

XII.
= '^
is
tensor, either positively or negatively
to
it is
or does not
oabc
is
fall,
easy to infer that /or any circular group o, A, b,
c,
we have
the
equation,
XIII...U^ = ~
+U^;
AB
the upper sign being taken
CB
the succession
oabo
occurs, that the
it
therefore
by
a direct one, that
is,
when
is
lateral is crossed: while conversely this equation
ever
is
uncrossed; and the lower sign, in the contrary case,

the succession is (what may be called) indirect, or when the quadri-
the quadrilateral
namely, when
OABC
when
anharmonic (oabc)
(5.) that the
grovp
is
XIII.
is sufficient
to prove,
when-
a negative or a positive scalar, and

not linear), as above.
is
circular (if
E be any five homospheric points (or points upon the surface

of one sphere), and if o be any sixth point of space, while oa',
oe' are the reciprocals
of OA,
OE, then the five new points a'. e' are generally homospheric (with each
(10.) If A, B, c, D,
o happens to be on the sphere abode, then a' . e' are complanar,

plane being parallel to the tangent plane to the given sphere at O
with resulting anharmonic relations, on which we cannot here delay.
other)
their
but
if
common
26 1
An
interesting case of the foregoing theory is that

generally scalar anharmonic of a circular group becomes equal to negative unity : in which case (comp. 26), the
.
when the
cidar
is said to be harmonic.
few remarks upon such cir^
and harmonic groups may here be briefly made the stu-
group
284
dent being
now
to
left to
fill
him an easy
up
hints for himself, as
(oABc) =
or
III.
and under
monic
we
this condition,
Mean between
II.
what must be
exercise of calculation.
(1.) For such a group (comp. again Fig. 58),

I.
[bOOK
we have
and therefore
II.
thus the equation,

a'b'
= b'c'
..R/3=i(Ra + Ry);
shall say (comp. 216, (5.) that the Vector
the two vectors,
a and
is
(2.) Dividing, and taking conjugates (comp. 260, (3.), and 216, (5.)
obtain the equation,
Iy...2-^2=2;
a^y
'
the
Har-
y.
),
we thus
= -â;
y.../3=-^y
+a
+ a^
or
|.y
or
YL..(3 = -y = la,
if
VII. ..
= |(y + a);
thus denoting here the vector oe (Fig. 58) of the middle point of the chord Ac.
then say that the harmonic mean between any two lines is (as in algebra)
We may
to their
the^wrÂ proportional
VIII.
semisum^ and to themselves.

thus the similar triangles,
we have
(3.) Geometrically,
.
A AOB a EOC
are equal,
we may
infer
the four points o, a, b,

(4.)
We have
c,
VIII'.
A AOE a BOC
or because the angles
oac and obc
(comp. 260, (5.) ) that, when the equation

if not coUinear, are concircular.
I. is satisfied,
whence, either because the angles
oba and oca,
also the similarities,
and
and
IX'.
A oea a aeb
r...t/=^%
at
CHAP.
INVOLUTION IN A PLANE, OR IN SPACE.
II.]
285
diagonals; which geometrical relation answers to either of the two anharmonic

equations (comp. 259, (10.))
XIII.
(oBAc) = + 2
(8.) Hence, or in other ways,
it
XIII'.
may be
conjugate chords of the circle to which they belong

through the pole of the other, and that thus the line
Ac prolonged
the point d, in which the chord
Under the same
(9.)
conditions,
it is
is
BC and ba
in the sense that each passes
db
is
the second tangent from
easy to prove, either by quaternions or by
we have the harmonic equations

1
XrV.
(abco) = (bcoa) = (coab) =
the harmonic mean between ab and Ao bo is such a mean between
:
AC
= + ^.
intersects the tangent at o.
geometry, that
so that
(ocab)
inferred that these diagonals, ob, ao, are
and ca between co and cb.
(10.) In any such group, any two opposite points (or opposite comers of the quadrilateral), as for example o and b, may be said to be harmonically conjugate to each
two other points^ A and c
other, with respect to the
A and c
and we
see that
o (whether
in space) there always corresponds di fourth point b, which is
points
are given, then to every third point
when
these two
in a given plane, or
in this sense conju-
always complanar with the three

points A, c, o, and being even concircular with them, unless they happen to be collinear with each other; in which extreme (or limiting) case, the fourth point b is still
gate to that third point
this fourth point being
is now collinear with the others

(as in 26, &c.).
When, after thus selecting two* points, A and c, or treating them as given
or fixed, we determine (10.) the harmonic conjugates b, b', b", with respect to fêw,
oi any three assumed points, o, o', o", then the three pairs of points O, b
o', b';
determined, but
(11.)
o", b",
be said to form an Involution,^ either on the right line AC, (in which
will only be one of an already well-known kind), or in a plane through that
may
case
it
line,
or even generally in space
and the two points
two Double Points (or Foci) of
said to be the
A,
c may in
opened, for geometrical investigation by Quaternions,
is far
all
these cases be
But the
this Involution.
field
thus
too extensive to be
more
than mentioned here.

(12.)
We
shall therefore only at present add, that the conception of the
mean between two
vectors
not be limited to the plane
n arbitrary vectors ai,
XV.
may
:
a,
we may
since
when
R7 =
(Rai +
define
it satisfies
+ Kan)
(13.) Finally, as regards the notation Ra,

cal of a vector,
it
harmonic
any number of such, and need

that r] is the harmonic mean of the
easily be extended to
the equation,
or
XVI.
and the
nRtj
definition
may be observed that if we had chosen to
= 2Ea.
(258) of the recipro-
define reciprocal vectors as
having similar (instead of opposite) directions, we should indeed have had the positive sign in the equation 258, VII. ; but should have been obliged to write, instead of
much
258, IX., the
less
simple formula,
RI=-IRr.
* There
even
when
is
the
Superieure of
a sense in which the geometrical process here spoken of can be applied,

two fixed points, or foci, are imaginary. Compare the GeomCtrie
M.
Chasles, page 136.
t Compare the Note
to 259, (11.).
286
CHAPTER
[bOOK
II.
III.
ON DIPLANAR QUATERNIONS, OR QUOTIENTS OF VECTORS IN

SPACE AND ESPECIALLY ON THE ASSOCIATIVE PRINCIPLE OF MULTIPLICATION OF SUCH QUATERNIONS.
:
Section
1.
On
some Enunciations of the Associative Pro-
perty, or Principle, of Multiplication of Diplanar Quaternions.
262. In the preceding Chapter v^ê have confined ourselves

almost entirely, as had been proposed (224, 225), to the consideration of quaternions in a îven plane (that of i) ; alkiding
only, in some instances, to possible extensions* of results so
obtained.
But we must now return to consider, as in the
First Chapter of this Second Book, the subject of General

Quotients of Vectors : and especially their Associative Multi-
which has hitherto been only proved in connexion with the Distributive Principle (212), and with the
plication (223),
Laics of the Symbols iyj, k (183). And first we shall give a

^QW geometrical enunciations of that associative principle, which
shall be independent of the distributive one, and in which it
will be sufficient to consider (comp. 191) the multiplication of
versors; because the multiplication of tensors i^ evidently an
associative operation, as corresponding simply to arithmetical
multiplication, or to the composition of ratios in geometry .f

shall therefore suppose, throughout the present Chapter,
We
that g^
are
r, s
given and
some
three given but arbitrary versors, in three

and our object will be to throw
distinct planes ;X
* As in
227, (3.); 242, (7.); 254, (7.); 257,
;
260, (10.); and 2G1, (11.) and (12.).
(6.)
and
(7.)
259, (8.), (9.),
(10.), (11.)
Or,
more
generally, for
any three
pairs of magnitudes, each pair separately
being homogeneous.
X If the factors
g, r, s
were coinplanar, we could always (by 120) put tbera
CHAP.
III.]
ASSOCIATIVE PRINCIPLE, SYSTEM OF SIX PLANES. 287
some additional light, by new enunciations in this Section,

and by new demonstrations in the next, on the very important, although very simple, Associative Formula (223, II.),
which may be written thus
:
I.
or thus,
more
II.
q',
sr.q-s.rq;
fully,
g'g
= t,
if
g'
= sr,
and
rq^
and t being here three new and derived

and derived planes.
s\
neio
ss'
versors, in three
263. Already we may see that this Associative Theorem

of Multiplication, in all its forms, has an essential reference to
a System of Six Planes, namely the planes of these six versorSf
lY.
'
q, r, 5, rq, sr, srq,
or
IV.
q, r, 5, s, q', f;
on the judicious selection and arrangement of which, the clearness and elegance of every geometrical statement or proof of
the theorem must very much depend while the versor character of the factors (in the only part of the theorem for which
:
proof is required) suggests a reference to a Sphere, namely to

what we have called the unit-sphere (128). And the three
following arrangements of the six planes appear to be the most
natural and simple that can be considered namely, 1st, the
:
which the planes all pass through the centre of

arrangement
the sphere; Ilnd, that in which they all touch its surface;
and Ilird, that in which they are the ^\k faces of an inscribed
in
solid.
We
proceed to consider successively these three ar-
rangements.
264. "When thQ first arrangement (263)
to
employ arcs of great
is
adopted,
it is
natural
circles^ as representatives of the versorSj
on the
under the forms,

(i
/3
and then should have (comp. 183,
(1.) ) the two equal ternary productB,
d y
SB = -d ^-^
= s.rq;
pa a y a
sr.q :=-!so that in this case (comp.

difficulty.
224) the associative property would be proved without any
288
[BOOK
II.
plan of Art. 162. Representing thus the factor q by the arc ab,
and r by the successive arc bc, we represent (167) their product rq^
or s\ by AC; or by any equal arc (165), such as de, in Fig. 59, may
be supposed to be. Again, representing s by ef, we shall have df
as the representative of the ternary
^
y^
product s.rqy or ss', or ^, taken in

one order of association. To rep resent the other ternary product,
we may
sr. q^ or q'q^
new
three
first
points, g, h,
i,
STTI
"~>Lg
''/\
<^
determine
^\^^
by arcual
"\
//^^Nv
J^
*ji^^:^^^5^
y
^^"^
-^"^
c*^^^^^^-^^^^:^-
equations (165), between gh, bc,

and between hi, ef, so that bc, ef
Fig. 59.
intersect in h, as the arcs representing 5' and s had intersected in e;

and then, after thus finding an arc Gi which represents sr^ or q^, may
determine three other points, k, l, m, by equations between kl, ab,

and between lm, gi, so that these two new arcs, kl, lm, represent q
and q', and that ab, gi intersect in l for in this way we shall have
;
an
q^q as required. And the theoequal to the former arc df, in the
namely km, which represents
arc,
rem then
that this last arc
is,
km
is
full sense of Art. 165; or that when (as under the foregoing conditions of construction) the five arcual equations,
I. ..
exist,
r>
AB =
<^
KL,
riBC
= '^GH,
<->
EF =
then this sixth equation of the

II.
'^
f^
'ÂC = ODE, nGI = '^LM,
HI,
same kind
is satisfied also,
DF = '> km:
K and m, being both on the same great circle as the two

determined
points, d and f ; or d and m being on the
previously
f
and
k: and the two arcs, df and km, of that
circle
through
great
the two points,
great circle, or the two dotted arcs, dk,

equally long,
and
fm
in the Figure, being
similarly directed (165).
nine points a . . i so as to satisfy the three middle

(1.) Or, after determining the
the three other points, k, l, m, without any other
I., we might determine
equations
arcual equations, as intersections of the three pairs of arcs ab, df ; ab, gi ; df, gi ;
and then the theorem would be, that (if these three last points be suitably distin-
extreme equations I., and

guished from their own opposites upon the sphere) the two
the equation II., are satisfied.
The same geometrical theorem may also be thus enunciated If the first,
:
(2.)
third,
and
and fifth
sides (kl,
GH, ed) of a spherical hexagon
klghed
be respectioely
and third sides (ab, bc, ca) of a. spheand sixth sides (lg, he, dk) of the same
arcunlly equal (165) to the first, second,
rical triangle Abc, then the second, fourth,
hexagon are equal

angle, MIF.
to the three successive sides (mi, if,
fm) of another spherical
tri-
CHAP.
FIRST AND SECOND
III.]
(3.) It
may
ARRANGEMENTS OF PLANES. ^89
also be said, that iffive successive sides (kl,
ed) of one spherical
hexagon be respectively and arcually equal to thence successive diagonals (ab, mi,
of theirs*
BC, IF, ca) of another such hexagon (ambicf), then the sixth side (dk)
is
equal to the sixth diagonal (fm) of the second.
we adopt the conception mentioned in 180, (3.), of an arcualsum, and

sum by inserting + between the symbols of the two summands, that of
added arc being written to the left-hand, we may state the theorem, in connexion
(4.) Or, if
denote such a
the
with the recent Fig. 59, by the formula
'^BA=n EF+
III.,. '^DF +
n BC,
'^
if
DA = n EC
where b and f
may denote any two points upon the sphere.

may also express* the same principle, although somewhat
We
(5.)
as follows (see again Fig. 59, and compare sub-art. (2.)
IV.
(6.)
If,
for
if
'^
ED + n GH +
'>
a moment, we agree
V.
we may then express the

VI.
(7 )
..ifD-E
Or
KL =
still
more simply,
if
<->,
f^',
or in words, that Addition

(8.) Conversely, if
(o"+'^')
^"
little
dk + he +
'^
lg =
0.
more lucidly thus
be supposed to denote any three dipla-
to the ride (180, (3.) )
'^
k-d+e-h + g-l = 0.
then
0,
nar arcs, which are to be added according

the theorem may be said to be, that
VII.
simply
AB = B - A,
recent statement IV. a
H-G + L-K =
less
to write (comp. Art. 1),

-^
then
0,
^"
of ArcK on a Sphere
above referred
to,
+ ('^'+");
is
an Associative Operation.
any independent demonstration be given,
of the truth of
any
one of the foregoing statements, considered as expressing a theorem of spherical geometry, f a new proof will thereby be furnished, of the associative property of multiplication
of quaternions.
265. In the second arrangement (263) of the six planes, instead

of representing the three given versors, and their partial or total
products, by arcs, it is natural to represent them (174, II.) by anConceive then that the two versors, q and r,
gles on the sphere.
are represented, in Fig. 60, by the two spherical angles, eab and
ABE; and therefore (175) that their product, rq or s\ is represented
by the external vertical angle at e, of the triangle abe. Let the
* Some of these formul and
figures, in
connexion with the associative principle,
are taken, though for the most part with modifications, from the author's Sixth Lecture on Quaternions, in
which that whole subject
is
very fiUly treated.
Comp. the
Note to page 160.
t Such a demonstration, namely a deduction of the equation

equations
I.,
by known
II.
from the five
properties of spherical covics, will be briefly given in the en-
suing Section.
2 p
290
[book
II.
second versor r be also represented by the angle fbc, and the third
versor s by bcf; then the
other binary product, sr or
q\ will be represented by
the external angle at f, of
the new triangle bcf. Again,
to represent the Jirst ternary
product, t ss^ s.rq, we have

only to take the external angle at
D of the triangle ecd,

60.
Fig.
D be a point determined
by the two conditions, that the angle ecd shall be
and DEC supplementary/ to bea. On the other hand,
if
equal to bcf,
we conceive
if
a point d' determined by the conditions that d'af shall be equal toEAB,
and afd' supplementary to cfb, then the external angle at d', of the
triangle afd', will represent the second ternary product, q^q =
which (by the associative principle) must be equal to the
Conceiving then that ed is prolonged to g, and fd' to

two spherical angles, gdc and ad'h, must be equal in all respects
vertices
d and
d' coinciding,
and the
rotations (174, 177)
sr. q,
first.
h,
;
the
their
which they
represent being not only equal in amount^ but also similarly directed.
Or, to express the same thing otherwise, we may enunciate (262) the
by saying, that when the three angular equations^

..ABE = fbc, bcf = ecd, DEC = 7r-BEA,
Associative Principle
I.
are
satisfied, then these three other equations^
II.
DAF = EAB,
FDA = CDE,
AFD =
TT
- CFB,
are satisfied also. For not only is this theorem of spherical geometry a
consequence of the associative principle o^ multiplication of quaternions y
but conversely any independent demonstration* of the theorem

at the same time, a proof of the principle.
266.
The
the six planes
is,
third airangement (263) of

may be illustrated by con-
ceiving a gauche hexagon, ab'ca'bc', to be

inscribed in a sphere, in such a
manner that
the intersection d of the three planes, c'ab',

b'ca^ a'bc', is on the surface; and therefore that the three small
circles,
denoted by
these three last triliteral symbols, concur

* Such as
we
-p.
shall sketch, in the following Section,
properties of the spherical conies.
Compare the Note
with the help of the known
to the foregoing Article.
CHAP.
in
THIRD ARRANGEMENT, SPHERICAL HEXAGON. 291
III.]
one point d
while the second intersection of the two other small
circles, ab'c, ca'b,
Let
Fig. 61.
may be
denoted by the letter
be also for simplicity at
it
the annexed
d', as in
supposed, that (as in
first
the Figure) the^z;e circular successions^
ab'cd^
c'ab'd,
ca'bd',
b'ca'd,
a'bc'd,
or that the Jive inscribed quadrilaterals^ denoted by

;
these symbols I., are all uncrossed ones.
Then (by 260, (9.) ) it is
allowed to introduce three versors^ q, r, 5, each having two expres-
are
all direct
sions, as follows
__
II.
__b'd
=+
q = JJ
DC'
ab'
V-;
AC'
^^
ca'
^^da'- =
r=U
+ U
B'D
CB
^;
CD'
BD'
ca'
a'b
although (by the cited sub-article) the last

formulae should receive the negative sign,
fourth of the successions
were to become
sponding quadrilaterals were crossed ones.
I.
members of these three

if
the
first,
indirect,
We
third,
and
or if the corre-
have thus (by 191)
the derived expressions,

III.
5'
= ro' =
=U
DC'
Q^
BC'
=U
= sr=JJ
CB'
AB'
whereof, however, the two versors in the first formula would differ
in their signs, if the fifth succession I. were indirect ; and those in
the second formula,
IV
and
be
since,
.t
= ss'
by the
if
the second succession were such.
=s.r(7=\J
^
bc'
q'q
^
^
= sr.q
= \j
^
associative principle, these
two
AC'
Hence,
;
last versors are to
under the supposed conditions of construcc', a, d', compose a circular and direct suc-
equal, it follows that,
tion, i\iQ four points, b,
cession
or that the quadrilateral, bc'ad',
is
plane, inscriptible,* and
uncrossed.
267.
It is easy,
by
suitable changes of sign, to adapt
the recent reasoning to the case where some or all of the successions I. are indirect and thus to infer, from the associa;
tive principle, this theorem
of spherical geometry
T/'ab'ca'bc'
Of course, since the four points bc'ad' are known to be homospheric (comp.
260. (10.)), the inscriptihility of the quadrilateral in a circle would follow from its
being plane, if the latter were otherwise proved : but it is here deduced from the
equality of the two versors IV., on the plan of 260, (9.J.
292
[bOOK
II.
he a spherical hexagon^ such that the three small circles c'ab',

b'ca', a'bc' concur in one point d, then, 1st, the three other small
bc'a, concur in another point, d';
circles, ab'c, ca'b,
of the
six circular successions, 266,
of those ivhich are indirect
and
I.,
bc'ad',
and Ilnd,
the number
always even (including zero).

And conversely, any independent demonstration* of this geometrical theorem will be a new proof o^ the associative prinis
ciple.
268.
The same
fertile principle of associative multiplication
may
be enunciated in other ways, without limiting the factors to be verThus we

sors, and without introducing the conception of a sphere.
may
say (comp. 264, (2.)), that
if
o.abcdef (comp. 35) be any
pencil of six rays in space, and o.a'b'c' any pencil of three rays, and
if the three angles aob, cod, eof of the first pencil be respectively
equal to the angles b'oc', c'oa', a'ob' of the second, then another
pencil oi three rays, o.a''b''ô'', can be assigned, such that the three
other angles boc, doe, foa of the first pencil shall be equal to the
angles b'ôc'', c^'oa''', a'ôb'^ of the third: equality of angles (with

one vertex) being here understood (comp. 165) to include complanarity,
and
similarity of direction of rotations.
(1.) Again (comp. 264, (4.)), we may establish the following formula, in M'hich
the four vectors a/3y5 form a complanar proportion (226), but c and Z, are any two
lines in space
T
I.
for,
under this
last condition,
we have (comp.
we determine (120)
six
new
[0 =
>?
I
IV.
-,
T)
/3
proof,
and
whence
the following
kX/z, so that
OK
=
-,
e
and
<{
An elementary
lowing Section.
vectors, r]6t,
is
1^4;
aye
then
125),
Another enunciation of the associative principle
ni...it'JA
a
for if
^_^.
~j
7
^?=? ^ = i i^f
'ye ayt a' d e'
II
(2.)
^^_î3
,f
,11
a
y
^-
f -
i?
by stereographic projection,
will be proposed in the fol-
PROOFS BY SPHERICAL CONICS.
CHAP.
III.]
we
hare the transformations,
shall
293
VI. is true,
(3.) Conversely, the assertion that this last equation or proportion
are connected by the five proportions IV., is a
fi
whenever the twelve vectors a
form of enunciation of the associative principle
for it conducts
(comp. IV. and V.)
to the equation,
even with
but,
this last restriction, the three factor-quotients in
VII.
may
represent
ant/ three quaternions.
On
some Geometrical Proofs of the Associative

Property of Multiplication of Quaternions^ which are inde-
Section
2.
pendent of the Distributive* Principle.
We propose,
269.
in this Section, to furnisli three
geome-
Demonstrations of the Associative Principle, in connexion with the three Figures (59-61) which were employed
and with the three arin the last Section for its Enunciation
trical
rangements oi six planes, which were described in Art. 263.

The two first of these proofs will suppose the knowledge of a
few properties oi spherical conies (196, (11.)); but the third
will only
employ the doctrine of stereographic projection, and

more strictly elementary character. The
will therefore be of a
Principle itself is, however, of such great importance in this

Calculus, that its nature and its evidence can scarcely be put
in too many different points of view.
270.
The only
properties of a spherical conic, which we shall in

as known,^ are the three following 1st, that
assume
this Article
through any three given points on a given sphere, which are not on a
great circle, a conic can be described (consisting generally of two oppo-
which shall have a given great circle for one of its two cyclic
arcs; Ilnd, that if a transversal arc cut both these arcs, and the conic,
the intercepts (suitably measured) on this transversal are equal; and
site ovals),
Ilird, that
while
if
its legs
the vertex of a spherical angle move along the conic,

pass always through two fixed points thereof, those legs
Compare 224 and 262 and the Note to page 236.

t The reader may consult the Translation (Dublin, 1841, pp. 46, 50, 55) by the
present Dean Graves, of two Memoirs by M. Chasles, on Cones of the Second Degree, and Spherical Conies.
;
294
[bOOK
II.
intercept a constant interval^ upon each cyclic arc, separately taken.

Admitting these three properties, we see that if, in Fig. 59, we conceive a spherical conic to be described, so ^s to pass through the
three points b, f, h, and to have the great circle daec for one cyclic
arc, the second and third equations I. of 264 will prove that the arc
GLIM
the other cyclic arc for this conic; the
is
first
I.
equation
proves
next that the conic passes through k and if the arcual chord fk be
drawn and prolonged, the two remaining equations prove that it
;
meets the cyclic arcs in d and m; after which, the equation 11. of
the same Art. 264 immediately results, at least with the arrangeadopted in the Figure.
ment"^*
The
(1.)
1st property
is
easily seen to correspond to the possibility of circum-
scribing a circle about a given plane triangle,
namely that of which the corners are
the intersections of a plane parallel to the plane of the given cyclic arc, with the
three radii drawn to the three given points upon the sphere
but it may be worth
:
an
Ilnd property by quaternions.

Take then the equation of a cyclic cone, 196, (8.), which may (by 196,
while, as
(2.)
exercise, to prove here the
XII.) be written thus
and
let
II.
I...S^Se=N|,
S^
n|',
being thus two rays (or sides} of the cone, which may also be considered to
be the vectors of two points p and p' of a spherical conic, by supposing that their
p and
p'
Let r and
lengths are each unity.
be the vectors of the two points t and t' on
r'
the two cyclic arcs, in which the arcual chord pp' of the conic cuts them
in.
..S- = 0,
a
The theorem may then be

Y.
if
or that this expression
Now, by
III.
and
S^=0,
IV.
Tr =
Tr'
so that
1.
(3
stated thus
p=XT-\- XT',
VL
that
then
VI.
satisfies II., if the
p'
= x'r + xr
equations
I.
III.
IV. V. be
satisfied.
V. VI., we have
(3
/3a?'/3
members of I. and II. are equal, and it only remains

to prove that their second members are equal also, or that T|o' = Tjo, if Tr' = Tr.
Accordingly we have, by V. and VI.,
whence
it
follows that the
VIIL
first
^7?.^^^ = S-iO,''
by
^-y^ = X
+X T+T
(O' + p
and the property
200,'^
(IL),
^' and 204,''^
(19.):
>"
in question is proved.
* Modifications of that
arrangement
be easy to adapt the reasoning.
may
be conceived, to which however it would
CHAP.
III.]
PROOF BY STEREOGRAPHIC PROJECTION.
295
271- To prove the associative principle, with the help of Fig. 60,
three other properties of a spherical conic shall be supposed known :*
1st, that for every such curve two focal poi7its exist, ipossessing seve-
important relations
and one tangent arc he
ral
to
it,
one of which
is,
that if these two foci
given, the conic can
be constructed; Ilnd,
that if, from any point upon the sphere, two tangents be drawn to the
conic, and also two arcs to the foci, then one focal arc makes with one
tangent the same angle as the other focal arc with the other tangent ;
and Ilird, that if a spherical quadrilateral be circumscribed to such
a conic (supposed here for simplicity to be a spherical ellipse, or the
opposite ellipse being neglected), opposite sides subtend supplementary
the two (interior) foci. Admitting these known

and supposing the arrangement to be as in Fig. 60, we
may conceive a conic described, which shall have e and f for its two
focal points, and shall touch the arc bc ; and then the two first of the
angles, at either of
properties,
equations I., in 265, will prove that it touches also the arcs ab and
CD, while the third of those equations proves that it touches ad, so
ABCD is a circumscribedf quadrilateral: after which the three

equations IL, of the same article, are consequences of the same properties of the curve.
that
272. Finally, to prove the same important Principle in a

more completely elementary way, by means of the arrangement
represented in Fig. 61, or to prove the theorem of spherical
geometry enunciated in Art. 267, we may assume the point d
as thejfJoZ^ of a
stereograpjhic projectio7i, in which the three
small circles through that point shall be represented by
right
the three other shy circles
li7ies,hut
being in one common plane. And

then (interchanging accents) the
all
theorem comes to be thus stated
be any three points

(comp. Fig. 62) on the sides bc,
If A,
b', c'
CA, AB of any plane triangle, or on

those sides prolonged, then, 1st,
""
^Fig, 62?
the three circles,

* The reader
In
may again
consult pages
strictness, there are of course
46 and 50 of the Translation

lately
cited.
four foci, opposite two by two.
t The writer has elsewhere proposed the notation, ef(. .;

abcd, to denote the
f to this circumscribed quadrilateral.
relation of the focal points e,
296
I.
will meet in one point d and Ilnd, an even

of the six (linear or circular) successions,
number
II.
and
ab'c, bc'a, ca'b,
II.
c'ab', a'bc', b'ca',
[bOOK
11'.
(if
any)
c'ab'd, a'bc'd, b'ca'd,
be direct; an even number therefore also (if any) being

indirect.
But, under ihi^form,* the theorem can be proved
will
by very elementary
considerations, and
still
without any em-
of the distributive principle (224, 262).
ployment
of the theorem, as thus stated, is evident from the Third Book

but to prove both parts together, it may be useful to proceed as follows,
The first part
(1.)
of Euclid
admitting the conceptiou (235) oi amplitudes, or of angles as representing roto<io*,

which may have any values, positive or negative, and are to be added with attention
to their signs.
We may
(2.)
thus write the three equations^

III.
ab'c
= mr^
bc'a
= nV,
= n"7r,
ca'b
to express the three collineations, ab'c, &c. of Fig. 62
the integer, n, being odd or
on a prolongation of that
even, according as the point b' is on the finite line AC, or

line
or in other words, according as the
and similarly
for the
(3.) Again, if
two other
coefficients, '
opqr be any four

.
succession II.
POQ
-j-
QOR = FOR +
V.
and
ab'd
-1-
db'c
may
= n;r,
direct or indirect
we may
establish the for-
2m7r,
with the same conception of addition of amplitudes
plane of the triangle abc, we
is
and n".
points in one plane,
mula,
IV.
first
if
then
d be any point
in the
write,
bc'd + dc'a
= nV,
ca'd
+ da'b = n'V
therefore,
VI.
(ab'd + dc'a) + (bc'd + da'b) + (ca'd
(4.) Again,
we have
if
any
four points
-f
db'c)
= (w +
n'
n")
tt.
be not merely complunar but concircidar,
VII.
the integer p being odd
opqr
The
CPQ + QRO=/J7r,
or even, according as the succession
Associative Principle of Multiplication
was
opqb
is
direct or indi-
stated nearly under this^r7n,
same simple diagram, in paragraph XXII. of a communication by the present author, which was entitled Letters on Quaternions, and has
been printed in the First and Second Editions of the late Dr. Nichol's Cyclopcedia of
The same commuthe Physical Sciences (London and Glasgow, 1857 and 18G0).
and was
illustrated
by
the
nication contained other illustrations and consequences of the same principle, which it
has not been thought necessary here to reproduce (compare however Note C) and
others may be found in the Sixth of the author's already cited Lectures on Quaternions (Dublin, 1853), from which (as already observed) some of the formulas and
;
figures of this Chapter
have been taken.
CHAP.
III.]
red
then
if
whereof c'
is
ADDITIONAL FORMULA, NORM OF A VECTOR.
we
a,
denote by
the second intersection of the
first intersection,
VIII.
we
first
circles I.,
have
shall
ab'd + dc'a
p and p' being odd, when the two
and second
first
297
BCD -f da'b =/)'7r,
=jtJ7r,
successions
are direct, but even in the con-
11'.
trary case.
by VI., we have,
(5.) Hence,
IX.
ca'd
db'c p'tt,
-1-
the third succession
through the intersection

say,
p"
is
odd
is
II',
of the two first
or even, according as the
previously considered,
is itself
number of indirect successions,

ing zero), or odd.
(6.)
where X.
In every
even or
among
if q, r, s
(7.)
sr.
q=:
p'
/j -f
+ p" =
ji
re'
if
and
";
I.
passes
direct or indirect, that is to
is
number
even (includ-
is
of snccessions of each kind
is
even,
the importance of the second part of it

any, of the six successions II. IT.) arising from
:
we have always,
+ s.7'q,
it
the five previously considered,
case, therefore, the total
(respecting the even partition,
XI
number of even coefficients, among ih^five

odd ; or in other words, according as the
and both parts of the theorem are proved

the necessity of proving that
therefore always circular, or the third circie
and newer
as in algebra,
XIl.
sr
g=^
s.rq,
be any three actual quaternions.
The
associative principle of multiplication
distributive principle,
also be proved, without the
may
by certain considerations of rotations of a system, on which we
cannot enter here.
On some Additional Formulae.
Section 3
273. Before concluding the Second Book, a few additional remarks may be made, as regards some of the notations and transformations which have already occurred, or others analogous to them.
And
first as
Book the
to notation, although
interpretation of
we have
reserved for the Third
such expressions as â, or
a^,
we have
yet
agreed, in 210, (9.) to abridge the frequently occurring symbol (Ta)^

to Ta^ ; and we now propose to abridge it still further to Na, and to
square of the tensor (or of the length) of a vector^ a, the Norm

of that Vector: as we had (in 190, &c.),' the equation Tq^ = 'Nq, and
called N^* the norm of the quaternion q (in 145, (11.) ).
shall
call this
We
therefore
now
write generally, for any vector

I.
(1.)
The equations (comp.

II.
..Np =
l;
186, (1.) (2.) (3.) (4.)
the formula,
.(Ta)2 = Ta2 = Na.
III.
V.
a,
),
IV.
..Np = Na;
N(p-a) = N(/3-a),
represent, respectively, the unit-sphere ; the sphere

the sphere through o, with a for centre ; and the
N(p - a) = Na
through A, with o for centre
sphere through b, with the sa"^

"' *^^
centre a.
2q
298
The equations (comp. 186,
(2.)
(6.) (7.)
..N(p + a) = N(p-a);
VI.
[bOOK
II.
),
VII.
N(p-/3) = N(p-
a),
oa
represent, respectively, the plane through o, perpendicular to the line
and the
plane which perpendicularly bisects the line ab.
274. As regards transformations, the few following may here be

added, which relate partly to the quaternion forms (204, 216, &c.)
of the Equation"^ of the Ellipsoid.
Changing K(fc
(1.)
p) to
Ep
R/c,
by 259, VIII.,
in the equation 217,
XVI.
of the ellipsoid, and observing that the three vectors p, Rp, and Rk are complanar,
while 1 Tp = TRp by 258, that equation becomes, when divided by TRp, and when
:
the value 217, (5.) for
<2 is
I.
taken, and the notation 273
Rp
surface,
we
member will soon be

second member as k^ i^.
with the
(2.) If, in connexion
same
employed
..Tf^+^VNt-N/c;
Rc
of which the first
and the
is
introduce a
seen to admit of being writtenf as T(tp +
earlier
new auxiliary
p^:),
forms (204, 216) of the equation of the

vector^ a or os, such that (comp. 2 1 6,
VIII.)
the equation may,
by 204,
(14.), be reduced to the following extremely simple

III.
T(T
fonn
= Ti3;
neio auxiliary point s is what we have called

while the line PS, or a p, which connects any two
which expresses that the locus of the
mean
the
sphere, 216,
XIV.
corresponding points^ p and s, on the ellipsoid and sphere, is seen to be parallel

the fixed line j3; which is one element of the homology, mentioned in 216, (10.).
(3.) It
and
if p'
easy to prove that
is
IV.
o-'
to
.S?c = S^
a
S^,
and therefore
V.
^':
87c = 8^ S?,
be the vectors of two new but corresponding points,
p'
and
s',
on the
and sphere; whence it is easy to infer this other element of the homology,
that any two corresponding chords, pp' and ss', of the two surfaces, intersect each
ellipsoid
other on the cyclic plane which has ^ for

fact,
its cyclic
normal (comp. 216,
they intersect in the point t of which the vector

,^^
VI.
xp +
=
x'p'
x+xr-
xa +
xâ'
x^x r-,
if
p'
x=S^,
d
* In the verification 216,

(2.) of the equation 216,
senting a surface of the second order,
=toad of
hasj
V-
and
V-
V and V^
nions (Dul
thi
figures of
in
and
a;'
=- 8^
(1.), considered as repre-
ought to have been printed,
but this does not affect the reasoning.
others iTJompare the Note to page 233.
(7.) )
is,
in
CHAT.
and
HOMOLOGIES OF ELLIPSOID AND SPIIEUE.
III.]
this point is
299
ou the plane just mculioued (comp. 216, XL), because
VIL.
(T
.T(T = T/3,
we had assumed
if
= f- S^ + V^^î3 = p-2/3S^
which would have given again, as

IX.
S^=0.
would have followed,
(4.) Quite similar results
VIIL
in IIL,
X.
but with
..S-=-S^ S^;
the other cyclic plane, with y instead of d for its normal, might therefore have been
taken (as asserted in 216, (.10.) ), as another plane of homology of ellipsoid and
namely, i\iQ point at infinity on
sphere, with the same centre of homology as before
:
the line
/3,
or on the axis (204, (15.)
two circumscribed cylinders of
of one of the
revolution (comp. 220, (4.)).
The same ellipsoid is, in two other ways, homologous to the same mean
same two cyclic planes asp/anes of homology, but with a new centre
(5.)
sphere, with the
which
of homology,
is
the infinitely distant point on the axis of the second circum-
scribed cylinder (or on the line ab' of the sub-article last cited).
Although not
(6.)
transformations
specially connected with the ellipsoid, the following general
be noted here (comp. 199, XIL, and 204,
may
XL..TVV? = V{KT7-S5)};
XIII.
r
the
tanU? = (TV: S) Vg =
The equations 204, XVI. and XXXV., give
(7.)
XIL
UV5 = UVU5
XIV.
XIV'.
XXXIV.)
UIV^ = Ax.
Î|^^.
easily,
XV.
TI Vg
= TVg
more symbolical forms,
Xlir.
UVU = UV
UIV = Ax.
XV'.
TIV = TV
and the identity 200, IX. becomes more evident, when we observe that
XVL .g-Nj=g(l-Kg).
.
(8.)
We
have
also generally
XVII
(9.)
),
'"g + 1
(g
+ l)(Kg +
l)
Ng+ 1 + 285'
The formula,*
XVIIL
in
(comp. 200, (10.) and 218, (10.)
i^ = (g-l)(Kg-M) ^ Ng-l + 2Vg
U(r5 + K5r) = U(Sr.Sg + Vr.Vg) = r-i(r252)^"S
which q and r may be any two quaternions,
in itself,
is not perhaps of
any great importance
but will be found to furnish a student with several useful exercises in trans-
formation.
When it was said, in 257, (1.), that zero had only itself iov a square-root,
meaning was (comp. 225), that no binomial expression of the form x-\- 1^(228)
(10.)
the
could satisfy the equation,
XIX.
= 92 = (x + iyy =
(,t2
_ y) -f 2ixy,
This formula was given, but in like manner without proof, in page 587 of the
author's Lectures on Quaternions.
300
II.
[bOOK
for any real or imaginary' values of the two scalar coefficients x and y, diflferent
from zero ;* for if hiquaternions (214, (8.) ) be admitted, and if h again denote, as
in 256, (2.), the
imaginary of algehray then (comp. 257, (6.) and (7.)) we
may
= 0, the imaginary expression,

Sv = Sv' = StV = Nr' - N =
write, generally, besides the real value 0^
XX.
r 4 hv%
if
being thus any two real right quaternions, with equal norms (or with equal
tensors), in planes perperpendicular to each other.
c
and
v'
(11.) For example, by 256, (2.)

formations,
XXI. ..(i +
hj)2
andby the laws (183)
= is -f +
(ij
+ji)
oft/A,
AO
we have
the trans-
one of the imaginary values of the symbol 0^.

(12.) In general, when hi-quaternions are admitted into calculation, not only the
square of one, but the product of two such factors may vanish, without efMer of them
separately vanishing a circumstance which may throw some light on the existence
so that the bi-quaternion
hj
is
of those imaginary (or symbolical) roots of equations, which were treated of in 257.
(13.) For example, although the equation
XXII.
has no real roots except
any other real

ternionf v
1,
..
g2_i = (^_l)
(5 4-1)
and therefore cannot be
scalar, or real quaternion, for q, yet if
hv',
verified
we
by the
substitution of
substitute for q the bi-qua-
with the conditions 257, XIIL, this equation XXII.
is verified.
(14.) It will be found, however, that when two imaginary but non-evanescetit
factors give thus a null product, the norm of each is zero; provided that we agree
to extend to bi-quaternions the formula N5 = Sg2 Vg^ (204, XXII.)
or to define
that the iVorwi of a Biquaternion (like that of an ordinary or real quaternion) is
;
equal to the Square of the Scalar Fart, minus the Square of the Right Part
of these
each
two parts being generally imaginary, and the^rwer being what we have
called a Bi-scalar.
(15.) With this definition, if q and q' be any two real quaternions, and
as above, the ordinary imaginary of algebra, we may establish the formula
if
be,
XXIII.
or (comp. 200, VII.,
(16.)
and 210, XX.),
XXIV.
As
N(9 + hq') = (Sq + hSqy - (Yq + hYq'^
regards the
N (g + hq) = Ng - Ng* + 2AS
norm of the sum
of
gKg'.
any two real quaternions,
or real vec-
tors (273), the following transformations are occasionally useful (comp, 220,
(2.) ):
N(g' + 5) = N(Tg'.Ug + Tg.Ug');

.N(/3 + a)=N(T/3.Ua + Ta.U/3);
in each of which it is permitted to change the norms to the tensors of which they are
the squares, or to write T for N.
XXV.
XXVI.
'

t This includes the expression + hi,
positive unity.
Other such roots are
-.^
of 257, (1.), for a symbolical square-root oi

hj, and + hk.
BOOK
III.
ON QUATERNIONS, CONSIDERED AS PRODUCTS OR POWERS OF

VECTORS AND ON SOME APPLICATIONS OF QUATERNIONS.
;
CHAPTER
I.
ON THE INTERPRETATION OF A PRODUCT OF VECTORS, OR

POWER OF A VECTOR, AS A QUATERNION.
Section
1.
On a First Method of interpreting a Product of

Tico Vectors as a Quaternion.
Art. 275. In the
First
Book
of these Elements
we
inter-
any two directed right lines in

space (4) Ilnd, the sum of two or more such lines (5-9) Ilird,
the product of one such line, multiplied by or into a positive
or negative number (15)
IVth, the quotient of such a line,
preted, 1st, the difference of
divided
Scalar (17)
generally a
such
we have called
and Vth, the sum of a system of
by such a number (16), or by what
lines,
each affected (97) wdth a scalar
coefficient (99), as
being in each case zYseZ/* (generally) Directed Line* in

or what we have called a Vector (I).
2i>
Space
276. In the Second Book, the fundamental principle or

pervading conception has been, that the Quotient of tivo such
Vectors
is,
generally, a
Quaternion (112, 116). It is howwe have included under this ge-
ever to be remembered, that
which usually relates to what may be called

an Oblique Quotient, or the quotient of two lines in space
making either an acute or an obtuse angle with each other
neral conception,
* The Fourth
Proportional to any three complanar lines has also been since interpreted (226), as being another line in the sayne plane.
302
[bOOK
III.
(130), the three following particular cases: 1st, the limiting

case, when the angle becomes null, or when the two lines are
similarly directed, in which case the quotient degenerates (131)
into 2L positive scalar ; Ilnd, the other limiting case, when the
equal to two right angles, or
is
angle
when
the lines are oppo-
and when in consequence the quotient again

but
now into a negative scalar and Ilird, the
degenerates,
intermediate case, when the angle is right, or when the two
sitely directed,
lines are perpendicular (132), instead
of being parallel (15),
and when therefore their quotient becomes what we have

called (132) a Right Quotient, or a Right Quaternion
which has been seen to be a case not less important than the
two former ones.
:
But no
277.
Interpretation has been assigned, in either of
the two foregoing Books, for a Product of two or more Vectors ; or for the Square, or other Power ofa Vector : so that
the Symbols,
I.
/3a,
7j3a5
and
II.
a^
a',
a'\
aS
which a, j3, 7
denote vectors, but t denotes a scalar, remain as yet entirely uninterpreted ; and we are therefore /ree
in
any meanings to these neio symbols, or

new combinations of symbols, which shall not contradict each
other, and shall appear to be consistent with convenience and
And to do so will be the chief object of this First
analogy.
of
the Third (and last) Book oit\iQ^Q Elements: which
Chapter
to assign, at this stage,
is
designed to be a
going.
278.
As
much
shorter one than either of the fore-
commencement oi ^Moh Interpretation we
shall
here define, that a vector a is multiplied by another vector j3,

or that the latter vector is multiplied into* the former, or
that the product j3a is obtained, when the multiplier-line /3
is divided
by the reciprocal Ba (258) of the multiplicand-line a ;
we had proved (136) that one quaternion is midtiplied into

when it is divided by the reciprocal thereof. In symbols, we shall therefore write, as a first definition, the foras
another,
mula:
*
Compare
the Notes to pages 146, 159,
CHAP.
I.
INTERPRETATION OF A PRODUCTOF TWOVECTORS. 303
I.]
where
./3a=i3:Ra;
And we
II.
Ra = - Ua Ta
(258, VII.).
proceed to consider, in the following Section,
some of
the general consequences of this definition, or interpretation, of

a Product of two Vectors, as being egual to a certain Quotient^
or Quaternion.
Section
2.
07i some Consequences of the foregoing Interpretation.
279.
I.
it
The
.
definition (278) gives the formula
i3a
:^
Ka
and
I'.
similarly,
aB =
^nr,
-tvp
by 259, VIII., the general relation,

= Kaf5
11. ..0)3 = KjSa.
or
/3a
gives therefore,
II.
The Products
two opposite orders, are

and the Multiplication of
oftioo Vectors^ taken in
therefore Conjugate Quaternions;
Vectors, like that of Quaternions (168),

Commutative Operation,
(1.) It follows
from
II.
(by 196, comp. 223, (1.)
III.
is
),
(generally) a
Non-
that
= i(/3a + a/3).
Si8a = + Sa/3
(2.) It follows also (by 204, corap. again 223, (1.) ), that
IV.
Vj3a
= - Va/3 = \{â-
a/3).
280. Again, by the same general formula 259, VIIL,

have the transformations,
R(a+a')
it
follows, then,
Kj3
also,
Ra
Ra"
from the definition (278), that

II.
j3(a-fa')=/3a + i3a';
.
whence
RjS
RjS
we
by taking conjugates (279), we have
this other
general equation,
III.
(a + a
) i3
aj3
-f
a'jS.
Multiplication of Vectors is, therefore, like that of Quaternions

(212), a Doubly Distributive Operation.
As we have not yet
assigned any signification for a

as 7/80, we are not yet presuch
ternary product of vectors,
281.
304
[bOOK
III
pared to pronounce, whether the Associative Principle (223)

oi Multiplication of Quaternions does or does not extend tc
But we can already
Vector-Multiplication.
derive several other
consequences from the definition (278) of o. binary product, j3a

among which, attention may be called to the Scalar character
of a Product of two Parallel Vectors; and to the Right cha-
racter of a Product
of two Perpendicular
lines at right angles
with each other.
The
(1.)
definition
I.
it
may
(278)
Vectors, or of
two
be thus written,
../3a = -T|3.Ta.U(/3:a);
gives, therefore,
III..
U...T(3a=T(3.Ta;
= -U(/3
U/3a
a)
= U/3.Ua
the tensor and versor of the product of two vectors being thus equal (as for quaternions, 191) to the product of the tensors, and to the product of the versors, respectively.
(2.) "Writing for abridgment (comp. 208),
IV.
we have
..a = Ta,
= T/3,
= Ax.(/3:a),
a;
= Z.(/3:a),
thus,
V.
VII.
T(3a
SU/3a
- 6a cos a;
= ba;
VI.
S/3a = Sa|3 =
= SUai3 = -cosx;
VIII.
Lâ=7r-X',
.
so that (comp. 198) the angle
of the product of any two vectors
is
the supplement of
the angle of the quotient.
We have next the
(3.)
IX.
TV/3a =
XL
XII.
transformations (comp. again 208),
TVa/3 = 6a sm
a;
= - yba
IV/3a
sin
X.
..IUV/3a=Ax./3a = -y;
so that the rotation
XI'.
TVU/3a = TVUaj3 = sin x

.
= + ya6 sin
IUVa/3 = Ax. a/S = + y
a;
IVa/3
XII'. ..
round the axis of a product of two
vectors,
from the multiplier
to
the multiplicand, is positive.
by IX., that the tensor of the right part of such a product,

equal to the parallelogram under the factors; or to the double of the area of
the triangle OAB, whereof those two factors a, /3, or OA, ob, are two coinitial sides :
(4.) It follows also,
(3a, is
so that if
we denote
here this last-mentioned area by the symbol
A OAB,
we may
write the equation,
XIII.
TY(3a = parallelogram under
a,
(3,
= 2A OAB
and the index, lYj3a, is a right line perpendicular to the plane oi this parallelogram,
of which line the length represents its area, in the sense that tliey bear equal ratios
to their respective units (of length
(5.)
XIV.
(6.)
and of
area).
Hence, by 279, IV.,

.
For any two
T((3a
vectors, a,
a(3)
(3,
=2 x
parallelogram
= 4 A gab.
CHAP.
PRODUCTS OF PARALLELS AND PERPENDICULARS. 305
I.]
XV. ..S/3a = -Na.S(/3:a);

or briefly,*
XVI.
V/3a
XVII. ../3a = -Na.(/3:
= -Na
V(/3
a);
a),
with the signification (273) of Nor, as denoting (Ta)2.

(7.) If the two factor-lines be perpendicular to each other, so that a; is a right
angle, then the parallelogram (4.) becomes a rectangle, and the product (ia becomes
a right quaternion (132); so that we
XVIII.
S/3a
may write,
= Sa)3 = 0, if 13 -^
and
a,
reciprocally.
Under the same condition of perpendicularity,
(8.)
XIX.
^/3a=Iai3 = |;
XX.
I/3a
=-
76a
XXI.
Ia/3
=+
yo&.
(9.) On the other hand, if the two factor-lines he parallel, (he right part of their
product vanishes, or that product reduces itself to a scalar, which is negative or positive according as the two vectors multiplied have similar or opposite directions ; for
we may
establish the formula,
XXII.
and, under the
if /3
same condition
XXIII.
II
then
a,
V/3a
= 0,
/3a
We may
also write
(by 279, (1.) and (2.)
&nd formula ofparallelism
pendicularity,
XXIV.
XXV.
(11.) If a,
/3,
= 0/3 = S/Ja = Sa/3 = + ba,
the upper or the lower sign being taken, according as

(10.)
Va/3
oi parallelism,
then
if /3 -I- a,
.. if j3\\ a,
then
y be any three
a;
= 0,
or
tt.
the following ^rmZa ofper-
(3a
=- al3,
and reciprocally
with the converse.
f3a=^-\- afS,
unit-lines, considered as vectors of the
comers
c of a spherical triangle, with sides equal to three new positive scalars, a, b, c,

then because, by XVII., (3a=- (3: a, and y/3 = - y /3, the sub-articles to 208 allow
A, B,
us to write,
XXVI.
S (Vyj3 V/3a) =
.
sin
sin c cos
XXVII.
IV(Vy/3.V/3a) = + /3sinasincsinB;
XXVIII.
(IV: S) (Vy/3.V/3a) = /3 tan b
.
upper or lower signs being taken, in the two last formulae, according as the rotation
round (3 from a to y, or that round b from a to c, is positive or negative.
(12.)
The equation 274,
XXIX.
T(p
I.,
of the Ellipsoid,
+ px) = Ti2-Tk:2;
or
may now
XXX.
be written thus
T(tp
+ pjc) = Nt-Nic.
Under the general head of a product of two parallel

two interesting cases occur, which furnish two first
examples of Powers of Vectors : namely, 1st, the case when
282.
vectors,
* All the
consequences of the interpretation (278), of the product (3a of two vectors, might be deduced from this formula XVII.
which, however, it would not have
been so natural to have assumed for a definition of that symbol, as it was to assume
;
the formula 278,
I.
2 R
306
[bOOK
III.
the two factors are equal, which gives this remarkable result,
that the Square of a Vector is always equal to a Negative Sca-
and Ilnd, the case when the factors are (in the sense
already defined, 258) reciprocal to each other, in which case
it follows from the definition (278) that their product is
equal
to Positive Unity : so that each may, in this case, be consi-
lar;
dered as equal to unity divided by the other, or to the Power

of that other which has Negative Unity for its Exponent.
When (3 = a, the product /3a reduces itself to what we may call the square
may denote by a^; and thus we may write, as a particular but important
(1.)
of a, and
case of 281, XXIII., the formula (comp. 273),

I.
a2=-a2 = -(Ta)2 = -Na;
is equal to the negative of the norm (273) of that

vector; or to the negative of the square of the number Ta, which expresses (185)
the length of the same vector.
so that the square
of any vector a
More immediately, the definition (278) gives,

II. .. a2 = aa = a Ra = (Ta)* = Na,
(2.)
as before.
Hence (compare the notations 161, 190, 199, 204),
(3.)
III.
S.a2=-Na;
IV.
..V.a2=0;
and
V.
T.a2 = T(a2) = + Na
= (Ta)2 = Ta2;
the omission of the parentheses^ or of the pointy in this last symbol of a tensor, * for
the square of a vector, as well as for the square of a quaternion (190), being thus
justified
and
in like
manner we may
VI.
write,
U.a2 = U(a2) = -l = (Ua)2 = Ua2;
the square of an unit-vector (129) being always equal to negative unity, and parentheses (or points) being again omitted.
(4.)
The equation
VII.
it
p2
= a2,
represents therefore,
VII'.
gives
Np = Na,
(2.), the sphere
by 186,
or
VII".
Tp = Ta
with o for centre, which passes
through the point a.

(5.)
The more general
equation,
VIII.
represents the sphere with
(p
for centre,
a)2
= (/3 - a),
(comp.f 186,
which passes through the point
(4.),
b.
(6.) For example, the equation,
IX.
represents the sphere with
(p
- a)2 = a2,
for centre,
Compare the Note
t Compare
(comp. 186,
which passes through the origin
to
(3.),
o.
page 210.
also the sub-articles to 273.
CHAP.
SQUARE AND RECIPROCAL OF A VECTOR.
I.]
The equations (comp. 18G,
(7.)
X.
(p
lepi'esent, respectively, the
(6.), (7.)),
= (p-a)2;
a)2
307
XL
(p ~/3)2
= (p- a)2,
OA
o, perpendicular to the line
plane through
and the
plane which perpendicularly bisects the line ab.

(8.) IhQ distributive principle of vector-multiplication (280), and the formula
279, III., enable us to establish generally (comp. 210, (9.) ) the formula,
XII.
IX'.
The
a)2 = ^2 + 2Sl3a + a2
(/3
and X. may
the recent equations IX.
(9.)
p'i=2Sap
and
X'.
Sap =
0.
equations,
XIII.
.pHa2 = 0;
o
represent the spheres with
for centre,
so that this very simple formula, p'
of the Unit-Sphere (128), and
tion
therefore be thus transformed
XIV.
which have a and
= 0,
is,
.. p2
is
+ l = 0,
1 for their respective radii
(comp. 186, (1.)
a form of the Equa-
as such, of great importance iu the present
Calculus.
(10.)
The
equation,
XV.
may
p-2Sap + c = 0,
be transformed to the following,
XVI. ..N(p-a) = -(p-a)2 = c-a2 = c4-Na;

XVI'.
T(p-a) = V(c-a2) = V(c + Na);
or
it
represents therefore a (real or imaginary) sphere, with
for centre,
and with
this
last radical (if real) for radius.
(11.) This sphere

scalar constant,
but
it
therefore necessarily real, if c be a positive scalar
is
becomes imaginary,
(12.)
if c
The radical plane

XVII.
or
if this
of the two spheres,
- 2Sap + c = 0,
p2
p2
XVIII. ..2S(a'-a)p
therefore always real,
Na
+ Na < 0.
has for equation,
it is
though negative, be (algebraically) greater than a^, or than-
c,
if
- 2Sa'p +
c'
0,
= c'-c;
the given vectors a, a and the given scalars
c,
be
such, even if one or both of the spheres themselves be imaginary.
(13.)
The equation 281, XXIX.,
ellipsoid -svith its centre
or
plified,* as follows:
XIX.
(14.)
The
it is
may now be still further sim-
.T(tp +
pK:)
jc-ia.
definition (278) gives also,
XX.
whence
of the Central Ellipsoid (ox of the
XXX.,
taken for the origin of vectors),
aRa = a a = 1
:
or
XX'.
Ra a = Ra Ra =
.
natural to write, f

t Compare the second Note to page 279.
308
XXI.
if
we
[bOOK
III.
Ra = l:i = a-i,
a
80 far anticipate here the general theory of powers of vectors^
above alluded
to
last symbol to denote the quotient, of unity divided by the vector

(277), as to use this
a ; so as to have identically, or for every vector, the equation,
XXII.
(16.) It follows,
XXIII.
(16.) If
by
.
a.a-i
= a-'.a=l.
258, VII., that
a-i
= - Ua Ta
:
XXIV.
and
/3a
= /3 aK
:
the equation XXIII. as a definition'^ ofthesymbol a'\

might have been used, as a formula of interpretation for
we had adopted
then the formula
XXIV.
But we proceed to consider an entirely different method, of arriving

the symbol (3a.
or of a Binary
at the same (or an equivalent) Interpretation of this latter symbol
:
Product of Vectors, considered as equal
Section
3.
to
a Quaternion.
On a Second Method of arriving at
the
same
In-
terpretation^ of a Binary Product of Vectors.
283. It cannot
to have been observed
fail
reader of the Second Book,
by any attentive
how close and intimate a connexion^
has been found to exist, between a Right Quaternion (132), and

Index^ or Index-Vector (133). Thus, if u and v denote (as
in 223, (1.), &c.) any two right quaternions, andif Iv, Iv de-
its
note, as usual, their indices,

I.
It;'
111.
to
already seen that
and conversely (133);

=
I(i;'v) Iv'It;(206);
if
Ii;,
11.
we have
v'
v,
Ii;':I=v':tj(193);
which may be added the more recent formula,

IV. .RIt;=IRv(258, IX.).
.
284. It could not therefore have appeared strange, if we

this new formula of the same kind,
had proposed to establish

I.
\v' .lv
= v.v =
v'v,
as a definition (supposing that the recent definition 278 had

not occurred to us), whereby to interpret the product of any two
of right quaternions, as being equal to the product oj

those tivo quaternions themselves.
And then, to interpret thi
two
of
product j3a,
given vectors, taken in a given order
any
indices
Compare
t Compare the Note
to
page 174.
CHAP.
SECOND INTERPRETATION OF A PRODUCT.
I.]
309
to conceive (as we always may), that

proposed yacôr^, a and (5, are the indices of two right
we should only have had

the two
quaternions, v and v\ and to multiply these latter, in the same

order. For thus we should have been led to establish the for-
mula,
II.
or
/3a
we should have
if
v'v,
this slightly
III. ..j3a
in
and
Iv,
j3
Iv'
more symbolical equation,
= /3.a = Pi3.r';
which the symbols,

I'a and
are understood to denote the
the
two
To
(1.)
and
lines a
establish
are the indices.
j3
now the
T'/B,
two right quaternions, whereof
substantial icZewîfy of these two interpretations,
278 and
284, of a binary product of vectors [3a, notwithstanding the difference ofform of

the definitional equations by which they have been expressed, we have only to observe that it has been found, as a theorem (194), that
IV. ..v'v
= Iw':
I (1
= Iv': IRr
r)
but the definition (258) of Ra gave us the lately cited equation, RIy
therefore, by the recent formula II., the equation,
Y.
as in 278,
I.
It;My
a and
/3
= I':RI;
still
VI.
or
/3.a
= /3
^ IRv
we have
Ra,
The two
denoting any two vectors.
interpretations
have been obtained by

ferent processes, or suggestions, and are expressed by two different formula.
therefore coincide, at least in their results, although they
dif-
(2.) The result 279, II., respecting conjugate products of vectors, corresponds
thus to the result 191, (2.), or to the first formula of 223,
(1.)(3.) The two formulae of 279, (1.) and (2.), respecting the scalar and right
parts of the product
(3a,
answer to the two other formulae of the same sub-article,
223, (1.), respecting the corresponding parts of

(4.)
The doubly
r'v.
distributive property (280), oi vector-multiplication, is
on this
plan seen to be included in the corresponding but more general property (212), of
multiplication of quaternions.
(5.)
of Art.
By changing YVq, lYq, t, t', and o, to a, (3, a, b, and y, in those formulae

208 which are previous to its sub-articles, we should obtain, with the recent
definition (or interpretation) II. of j3a, several of the consequences lately given (in
sub-arts, to 281), as resulting
from the former
definition,
278,
I.
Thus, the equa-
tions,
VI., VII., VIII,, IX., X.,

of 281, correspond
to,
XL,
and may (with our
XII., XXII., and XXIII.,
last definition)
be deduced from, the
for-
mulae,
v., VI., VIII.,
of 208.
(Some
XL,
XII.,
XXIL, XX.,
XIV., and
of the consequences from

the^jsub- articles to
considered, in 281, (11.))
XVL, XVIIL,
208 have been already
310
(6.)
The geometrical
properties of the line
[bOOK
lYâ, deduced from ihejirat
(278) of (3a in 281, (3.) and (4.), (namely, the positive rotation round
tion
III.
defini-
tha.t line,
/3 to a ; its perpendicularity to their plane ; and the representation by the same

of the paralellogram under those two factors^ regard being had to units of length
of area^) might also have been deduced from 223,
(4.), by means of the second
from
Ime
and
definition (284), of the
Section
4.
On
ternion with
same product^
(5a.
the Symbolical Identification of a Right Quaown Index: and on the Consti'uction of a
its
Product of Tioo Rectangular Lines^ by a Third Line,

angular
rect-
to both.
285. It has been seen, then, that the recent formula 284,
may replace the formula 278, 1., as a second definition
II. or III.,
of a product of two vectors^ which conducts to the same consequences, and therefore ultimately to the same interpretation
of such a product, as the first.
in the second formula,
Now,
have interpreted that product, /3a, by changing the two factor-lines, a and j3, to the two right quaternions, v and v, or
wê
I~'a
and
I"^j3,
of which they are the indices; and by then de-
fining that the sought product )3a is equal to the product v'v,
of those two right quaternions. It becomes, therefore, important to inquire, at this stage, how far such substitution, of T'a
for a, or of v for lu, together with the converse substitution, is
permitted in this Calculus, consistently with principles already
established.
For
it is
evident that
if
such substitutions can
be sliown to be generally legitimate, or allowable, we shall

thereby be enabled to enlarge greatly the existing field oi interpretation:
and to
treat, in aZZ cases. Functions
of Vectors, as
being, at the same time. Functions of Right Quaternions.

286.
We have first,
I.
by 133 (comp. 283, I.), the

..ri/3 = I-ia, if ^=a.
In the next place, by 206 (comp. 283,
II.),
we have
equality ^
the formula of
addition or subtraction^
II.
..I-i(â) = I-'/3ria;
with these more general results of the same kind (comp. 207
and 99),
III.
r^2 = 2l-'fl
IV.
I'lxa = ^xV'a.
.
CHAP.
RIGHT QUATERNION EQUAL TO
I.]
In the third place, by 193 (comp. 283,
311
ITS INDEX.
III.),
we
have, for division,
the formula,
V... r>^:I-'a = ^:a;

while the second definition (284) oi multiplication of vectors, which has
been proved to be consistent with ih^ first definition (278), has given
us the analogous equation,
VI.
r'^.râ = /3.tt = â.
would seem, then, that we might at once proceed to define^ for the
purpose of interpreting any proposed Function of Vectors as a Quater-
It
ternion, that the following general
VII.
or
still
more
sulject of
.râ = a;
briefly
the operation I
VIII.
or
and
Equation exists
.
I??
symholically, if it
is
if
v,
1;
=
^;
be understood that the
always a right quaternion,

IX. ..1=1.
But, before finally adopting this conclusion, there is a case (or rather
a class of cases), which it is necessary to examine, in order to be certain that no contradiction to former results can ever
be thereby caused.
The most
general form of a vector function, or of a vector

of other vectors and of scalars, which was
a
function
as
regarded
considered in the First Book, was the form (99, comp. 275),
287.
I.
/3
= 2ica;
and we have seen that if we change,
in this form, each vector a to the
corresponding right quaternion Tâ, and then take the index of the
new right quaternion which results, we shall thus be conducted to
precisely the same vector p, as that
tained before; or in symbols, that
II.
2aâ
ll^xlr'a
which had been otherwise ob-
(comp. 286, IV.).
But
another form of a vectorfunction has been considered in the Second Book ; namely, the form,
III.
..^=...Jâ(226,
in
which
And
a,
before
of 286,
we
y3,
7, ^, e
III.);
are any odd number of complanar vectors.
we
accept, as general, the equation VII. or VIII. or IX.

must inquire whether we are at liberty to write, under
the same conditions of complanarity, and with

of the vector p, the equation,
i\iQ
same
signification
312
288.
planar
To examine
vec/or^,
[bOOK
III.
there be at first only three given com^; in which case there will always be (by
in the same plane, which will
/>,
represent or
this, let
7|||a,
226) 2L fourth vector

construct the function (7:^). a; namely, t\iQ fourth proportional to
/3, 7, a.
Taking then what we may call the Inverse Index-Functions,
or operating on these four vectors a, y3, 7, p by the characteristic I'S
obtain /owr collinear and right quaternions (209), which
may be
denoted by v, v'^ v'\ v'" \ and we shall have the equation,
we
V.
v"'\v={p\ a=^\^=:)v"\v'\
VI.
or
v''^
{v'':v').v\
which proves what was required.
Or, more symbolically,
^=tîj:h.
a
VII
I'a
VIIL..^.a
/3
1-1/3'
= ^ = I(I =
l(I;;i.I-ia).
And it is so easy to extend this reasoning to the case of any greater

odd number of given vectors in one plane, that we may now consider the recent formula lY. as proved.
289.
We
shall therefore adopt, as general, tlie symholical
equations VII. VIII. IX. of 286 ; and shall thus be enabled,

in a shortly subsequent Section, to interpret ternary (and other)
products of vectors, as well as powers and other Functions of
Vectors, as hQing generally Quaternions; although they
may,
in particular cases, degenerate (131) into scalars, or may become right quaternions ( 1 32) in which latter event they may,
:
to, their
own
symbols,
we
y,
same
principle, be represented by, and equated

indices (133), and so be treated as vectors.
In
shall write generally, for any set of vectors a, (5,
in virtue of the
and any function X, the equation,

= /a"^,
I.
)
f{a, /3, 7>
.
1% r^y,
...)
q,
while in the particular case

q being some quaternion;
this quaternion
is
right, or
when
when
=
q v= 8-^0= I>,
'
CHAP.
we
I.]
PRODUCT OF TWO RECTANGULAR LINES ALINE.
313
and usually by preference (for that
case),
shall write also^
the formula,
II.
./(a,
r,
I/(l-', I-'/3, 1-'7, ...)=/,

vector,
a
|0 being
290. For example, instead of saying (as in 281) that the
Product of any two Rectangular Vectors is a Right Quaternion,
/3,
.)
with certain properties of its Index, already pointed out (284,

(6.) ), we may now say that such a product is equal to that in-
And hence will follow the important consequence, that

Product of any Two Rectangular Lines in Space is equal
(or may be constructed by) a Third Line, rectangular to
dex.
the
to
both; the Rotation round this Product -Line, from the Multiplier-Line to the Multiplicand' Line, being Positive : and the
Length of the Product being equal to the Product of the

Lengths of the Factors^ or representing (with a suitable reference to units) the Area of the Rectangle under them.
And
generally
we may now,
pression, identify'^ a
Section 5
and
own Index.
for all purposes of calculation
Right Quaternion with
its
ex-
On
some Simplifications of Notation, or of Expression, resulting from this Identification ; and on the Conception of an Unit-Line as a Right Versor,
291.
An
immediate consequence of the symbolical equa-
tion 286, IX., is that

I, of the
which it
we may now
suppress the Characteristic
Index of a Right Quaternion, in all the formulae into

has entered
and so may simplify the Notation. Thus,
;
instead of writing,
Ax. q = IJJVq,
Ax. q = UIV^,
or
or
Ax. = lUV,
Ax. = UIV,
as in 204, (23.),
as in 274, (7.),
we may now write simply f,
I...Ax.q=\JYq',
or
II.
..Ax. = UV.
The Characteristic Ax., of the Operation of taking the Axis of

a Quaternion (132, (6.) ), may therefore henceforth be
replaced
Compare the Notes to pages 119, 136, 174, 191, 200.

t Compare the first Note to page 118, and the second Note
2s
to
page 200.
314
[bOOK
III.
whenever we may think fit to dispense with it, by this combination of two other characteristics, U and V, which are of greater
and more êw^7-/ utility, and indeed cannot* be dispensed with,
in the practice of the present Calculus.
are now enabled also to diminish, to
292.
We
some extent,
the number of technical terms, which have been employed in

the foregoing Book.
Thus, whereas we defined, in 202, that
the right quaternion Yq was the Right Part of the Quaternion
we may now, by
of the sum Sq + Yq,
q, or
290, identify
that part with its own index-vector TVq, and so may be led to
call it the vector part, or simply Âe Vector,! of that Quater-
nion
q, without henceforth speaking of the right part: although

the plan of exposition, adopted in the Second Book, required
that we should do so for some time.
And thus an enuncia-
tion, which was put forward at an early stage of the present

work, namely, at the end of the First Chapter of the First
Book, or the assertion (17) that
" Scalar
becomes entirely
when
and acquires a perfectly definite

are in this manner led to conceive a
intelligible,
For we
signification.
Number
plus Vector equals Quaternion,'*
(positive or negative) as being added to a Li7ie,X

added (according to rules already established) to
it is
that right quotient (132), of which the line is the Index,

symbols, we are thus led to establish the formula,
I.
9^
=a+
a,
when
\1.
a-\- I'*a
In
Of course, any one who chooses may invent new symbols, to denote the same
operations on quaternions, as those which are denoted in these Elements, and in the
elsewhere cited Lectures, by the letters
and "V ; but, under some form, such sym-
appears to have been hitherto thought expedient, by other

in
writers, not hastily to innovate on notations which have been already employed
As to the
several published researches, and have been found to answer their purpose.
bols
must be used : and
it
type used for these, and for the analogous characteristics K, S, T, that must evidently
be a mere affair of taste and convenience and in fact they have all been printed
as small italic capitals, in some examination-papers by the author.
:
t Compare the Note to page 191.

be the st/m of a
J On account of this possibility of conceiving a quaternion to
number and a line, it was at one time suggested by the present author, that a Quaternion might also be called a
Grammarithm, by a combination
of the
two Greek
words, y pa fifi-^ and dpidnog, which signify respectively a Line and a Number.
CHAP.
CONCEPTION OF ANUNIT-LINE ASA RIGHTVERSOR. 315
I.]
and whatever
lohatever scalar^
and
vector,
may be denoted by a
And
because either of these two parts, or summands,

vanish separately, we are entitled to say, that both Sca-
a.
may
Numbers and Lines, are included in the

a
Qiiaternion, as now enlarged or modified.
Conception of
293. Again, the same symbolical identification of \v with
lar s
and
Vectors, or
VIII.) leads to the forming of a new conception of an

Unit- Line, or Unit-Vector (129), as being also a Riff ht Versor
(153) or an Operator, of which the effect is to tuvji a line, in
V (286,
through a positive quadrant of

rotation : and thereby to oblige the Operand-Line to take a
neiv direction, at right angles to its old direction, but without
a plane perpendicular to
itself,
any change of length. And then the remarks (154) on the

=
equation q'^ ~\, where q was a right versor in i\\Q former
sense (which is still a permitted one) of its being a right radial quotient (147), or the quotient oftico equally long but mutually rectangular lines, become immediately applicable to the
interpretation
of the equation,
= -l,
p2
where p
(1.)
we have
is still
^2
(282,
XIV.)
an unit-vector.
Thus (comp.
Fig. 41), if
a be any
line perpendicular to stick
a vector p,
the equations,
..pa = /3;
I.
(B
or
II.
pâ
= p(5 = a' = -a;
being another line perpendicular to p, which is, at the same tinie, at right angles
and of the same length with it and from which a third line a\ or a, oppo-
to a,
the line a, but
equally long, is formed by a repetition of the operation,

denoted by (what wê may here call) the characteristic p or having that unit-vector p for the operator, or instrument employed, as a sort of handle, or axis*
of rosite to
still
tation.
C2.) More generally (comp. 290), if a, (3, y be any three lines at right angles to
each other, and if the length of y be numerically equal to the product of the lengths
of a and [3, then (by what precedes) the line y represents, or constructs, or is
equal
to,
the product
of
the two other lines, at least if a certain order of the

factors
we may write the equation (comp. 281,
(comp. 279) be observed: so that

III.
..a/3
= y,
if
lY.
(3
-i-
XXI.),
a,
Compare the
-i-
first
a,
Note
-i-
(3,
and
to page 136.
V.
Ta. T/3 = Ty,
316
[bOOK
provided that the rotation round a, from /3 to y, or that round y from a to

has the direction taken as the positive one.
(3.)
In
more general
this
we may
case,
(3,
III.
&c.,
conceive that the multiplier-line
still
a has operated on
diict-line
the multiplicand-line (3, so as to produce (or generate) the prohut not now by an operation of version alone, since the tensor of /3 is
(generally) multiplied by that of a, in order to form, by V., the tensor f the pro-
duct y.
(4.)
And
if
(comp. Fig. 41,
his, in
which a was
first
we repeat this compound operation, of tension and

if we multiply again hy a, we obtain a. fourth line
a')
or
with a direction opposite to that of

the
line,
VI.
(5.)
The operator
gative scalar,
multiplier (15)
/3,
and with a length generally
ay = aa^ = a2/3 =/3' = aT-ji,
see,
and then
)3,
to
different
namely
a = Ta.
e^c? on
/3,
to the ne-
now with a
completely consistent, in
6.
if
a^, or aa, is therefore equivalent, in its
be aâtn deduced^ but
Section
to
Na, considered as a coefficient, or as a (scalar)

(Ta)2, or
whence the equation,
a^, or
VII. ..a2 = -Na (282,
may
changed
version combined (comp. 189),

/3', in the plane of /3, y, but
On
all its
ne?:?
I.),
interpretation,
which
consequences, with the one
the Interpretation
is,
first
however, as we
proposed (282).
of a Product of Three or
more Vectors^ as a Quaternion,

294. There
is
now no
difficulty in interpreting a ternary
product of vectors (comp. 277, I.), or a product o^ more vectors than three ^ taken always in some given order ;
namely, as
the result (289, I.) of the substitution of the corresponding
right quaternions in that product: which result is generally
what we have lately called (276) an Oblique Qtwtient, or a
Quaternion with either an acute or an obtuse angle (130) ; but
degenerate (131) into a scalar, or may become itself a.

right quaternion (132), and so be constructed (289, II.) by a
new vector. It follows (comp. 28 1), that Multiplication of Vec-
maj
tors, like that
see that
we may
it is
of Quaternions (223), in which indeed

included, is
we now
an Associative Operation: or that
write generally (comp. 223,

the Formula,
II.), for
ang three
tors^ a, /3, y,
I.
(1.)
The formulaB 223,
III.
II.
III.
7j3.a
and IV.,
are
= 7.j3a.
now
V.yV/3a = aS/3y
Vy/3a = aS/3y
replaced by the following
-/3Sya;
- /3Sya +
ySa/3
vec-
CHAP.
in
TERNARY PRODUCTS OF VECTORS.
I.]
which Yy(3a
is
written, for simplicity, instead of V(7j3a), or V. y/3a;
which, as with the earlier equations referred

it useful to render himself vevT/ familiar.
form of the equation
(2.) An(ther useful
IV.
The equations IX. X.
(3.)
vectors, the formula
V.
317
to,
and with
a student of this Calculus will find
II. is the following:
V(Va/3.y) = aS/3y ^/3Sya.

XI"V^. of
223 enable us now
to write, for
any three
SyjSa
= - Sa/3y = Say (3 = - S/3ya = S(3ay = - Sya/3
=+
=+
volume of parallelepiped under a,

X volume of pyramid oabc
y,
(3,
upper or lower signs being taken, according as the rotation round a from
j8
to
is
or in other words, the scalar SyjSa, of the ternary product of

vectors y/3a, being positive in the first case, but negative in the second.
positive or negative
(4.)
pressed
The condition of complanarity of

by the equation (comp. 223, XI.)
VI.
(5.) If a,
j8,
Sy/3a
y be any three
y, is therefore
/3,
ex-
complanar with a and
VI'.
Sa/3y
&c.
complanar or diplanar, the expression,
..5 = aSi3y-/3Sya,
Sy^=
represents therefore (comp. II.
to y, but
or
vectors,
VII.
VIII.
gives
it
three vectors, a,
and
0,
and IV.)
(3
a,
IX.
Srt/35
fourth vector
= 0;
d,
which
'i3
perpendicular
or in symbols,
X. ..^j.y,
and
XL
..5
a,
111
/3.
(Compare the notations 123, 129.)

(6.)
For any four
vectors,
XII.
XIII.
we have by
and IV. the transformations,
II.
YiYa(3 .Vy^) = dSa(3y - ySa[3d

V (Va/3 Vy^) = aS^yd- (SSayd
.
and each of these three equivalent expressions represents & fifth vector e, which is at
once complanar with a, (3, and with y, ^; or a line oe, which is in the intersection
of the two planes, oab and ocd.
(7.) Comparing them, we see that any arbitrary vector p may be expressed as
a linear function of any three given diplanar vectors, a, (3, y, by the formula
:
XIV.
which
is
(8.)
pSaPy = aS(3yp + (3Syap
found to be one of extensive
-\-
pSa/3y
utility.
Another very useful formula, of the same kind,
XV.
ySa(3p
= V/3y. Sap + Vya
S(3p
is
the following:
+ Va/3. Syp
member of which, the points may be omitted.

(9.) One mode of proving the correctness of this last formula XV., is to
on both members of it, by the three symbols, or characteristics of operation,
in the second
XVI. ..S. a,
the
common
results
S./3,
operate
S.y;
on both sides being respectively the three scalar products,
XVII.
where again the points
Sap
may
8a(3y,
be omitted.
S(3p Sa(3y,
Syp Sa/3y
.
318
We
(10.)
[bOOK
here employ the principle, that if the three vectors a,
/3,
y he actual
and diplanar, then no actual vector \ can satisfy at once the three scalar
XVIII. .. Sa\ = 0, S/3\ = 0, SyXÔ;
because
equations^
cannot be perpendicular at once to those three diplanar vectors.

then, in any investigation with quaternions, we meet a system of this
it
(11.)
III.
If,
form XVIII., we can at once infer that
XIX. ..\=0,
while, conversely, if
= 0,
or Sa/3y
(12.)
X he an actual
XX.
if
vector, then a,
Sa/3y
y must be complanar vectors,
/?,
as in VI'.
Hence
under the same condition XX., the three scalar equations,
also,
XXI.
SaX = Sa^, S^X = S/3/i,

XXII,
X = ix.
.
give
SyX = Sy/x,
Operating (comp. (9.)) on the equation XV. by the symbol, or characS.^, in which 8 is any new vector, we find a result which may be written
thus (with or without the points)
(13.)
teristic,
XXIII.
where
a,
/3,
y, d,
= Sap
S(3yd
S(3p
Sy â +
- So>
Sa/3y
SyjO SSaf3
.
denote any five vectors.
may
that the equation XV. holds

(14.) In drawing this last inference, we assume
:
are
which in fact must be true,
three
vectors
the
when
even
a,
complanar
/3,
y
good,
as a limit, since the equation has been proved, by (9.) and (12.), to be valid,
little out of the plane of a and /3.
if
y be
ever so
We
(15.)
have therefore
this
new formula
XXIV. ..V/SySap + Vya S,3p + YafSSyp = 0,

in
which p
of a,
may
denote any fourth vector, whether
if
Sa(3y
or out of, the
in,
= 0;
common plane
y.
(3,
to that plane, the last formula

(16.) If p be perpendicular
evidently true, each
is
member vanishing separately, by 281, (7.) and if we change p to

a vector d in the plane of a, /3, y, we are conducted to the following equation, as an
interpretation of the same formula XXIV., which expresses a known theorem of
term of the
first
plane trigonometry, including several others under
XXV.
for
sin
Boc
aod +
cos
sin
coa
cos
it
bod +
sin
aob
cos
any four complanar and co-initial lines, OA, ob, oc, od.
a line perpendicular thereto, but
(17.) By passing from od to
plane,
we have
XXVI.
this other
.
sin
which, like the former,
known*
BOC
sin
equation
aod +
admits of
sin
many
XXIV.
obtained as above
by
sin
aob
transformations, but
when we
is
sin
cod =
seek to interpret the formula
by S.^, and changing p
to
e,
we have
Compare page 20
only mentioned here
quaternions.
(18.) Operating on that formula
equation
common
in their
coA sin bod +
as offering itself naturally to our notice,
cod = 0,
of the Geometrie Superieure of
M.
Chasles.
this
new
CHAP.
ELIMINATION OF A VECTOR.
I.]
XXVII.
= SatSpyd -f
Si3tSyad
if
SyeSapd,
-f
319
=
Sa(3y
which might indeed have been at once deduced from XXIII.
/?,
(19.) The equation XIV., as well as XV., must hold good at the
y are complanar ; hence
XXVIII.
+ ySa/3p =
aS(Byp + pSyap
if
0,
Saf5y
limit,
when
a,
= 0.
(20.) This last formula is evidently true, by (4.), if p be in the common plane
of the three other vectors ; and if we suppose it to be perpendicular to that plane,
so that
XXIX.
and
by 281,
therefore,
XXX.
we may
Yya Ya^,
= 0,
S^yp = S(V/3y.p) = V/3y.,o,
..p\\ V/3y
since
(9.),
.
II
\\
S(S/3y.p)
&c.,
divide each term by p, and so obtain this other formula,
XXXI.
aVj3y
+ /3Vya + y Vaj3 = 0,
if
Sa/3y
= 0.
(21.) In general, the vector (2^2") of this last expression vanishes
expression
own
therefore equal to its
is
XXXII.
whatever three vectors
may
aY(3y
we have
XXXIII.
and the formula
by
a,
II.
the
3Sa/3y,
y.
/3,
we suppose
if
by
write,
that the three vectors are
the proportion,
.
V/3y
Vya
XXXI, becomes
XXXIV.
may
+ (3Yya + yVa/3=
be denoted
(22.) For the case of complanar ity,
equally long,
and we
scalar,
OA
Va/3 = sin boc
coA
sin
sin
aob
thus,
BOC + ob
sin
sin
coA + oc
aob =
sin
where OA, ob, oc are any three radii of one circle, and the equation
is
interpreted as
in Articles 10, 11, &c.
(23.)
XV., by
The equation XXIII. might have been deduced from XIV.,

operating with S.^, and then interchanging 5 and p.
instead of
first
(24.)
A vector
in general be considered
may
(221) as depending on three scacannot then be determined hy fewer than three

be eliminated between /ezrer than four.
lars (the co-ordinates of its term^
scalar equations
(25.)
nor can
As an example
three given
it
of such determination of a vector, let a,
and diplanar vectors
XXXV.
in Avhich a, b, c are
p has for
its
(26.)
same
S/3p
let
S(3yp
XXXIX.
As an example
XL.
y be again any
= 0,
Syp =
Then the sought
vector
Syap =
= a,
e,
we
6',
S(3p
of a
b,
Sa/3p
have,
p = e-i (â'a + 6'/3 +
of elimination
Sap = a,
XXXVII.
if
.. e
Sa/3y.
the three equations be,
signification of the scalar
equations,
j3,
the three given equations be,
= c-i(aVi3y + 6Vya + cVa/3),
XXXVIII.
(27.)
= a,
Sap
let
by Xy.,
As another example,
then, with the
and
supposed to denote three given scalars.
expression,
XXXVI.
it
c' ;
by XIV.,
c'y).
vector, let there be the
Syp =
r,
S(^p
d;
four scalar
320
then,
by XXIII., we have
only the /our vectors, a
XLI.
which p does not enter, but
this resulting equation, into
and the ^bur
.d,
a S/3y5 -
(28.) This last equation
a..d:
scalars,
SySa +
III.
[bOOK
SSaf5-d. Sa(5y=^0.
therefore be considered as the condition of concur-
may
rence of the four planes, represented by the four scalar equations XL., in one common point; for, although it has not been expressly stated before, it follows evidently
from the definition 278 of a binary product of
combined with 196,
vectors,
that every scalar equation of the linear form (comp. 282,
XLII.
= OA,
which a
in
vector of the foot
Sap =
or
a,
Spa
= a,
= op,
as usual, represents a plane locus of the point P

of the perpendicular on that plane from the origin, being
and p
s,
XLIIL
OS
= (T=aRa = aa-i
(5.),
XVIIL),
(282,
the
XXL).
we
conceive a pyramidal volume (68) as having an algebraical (or scalar^ character, so as to be capable of bearing either a positive or a negative ratio to
the volume of a given pyramid, with a given order of its points, we may then omit
(29.) If
the ambiguous sign, in the last expression (3.) for the scalar ofa ternary product of
vectors : and so may write, generally, oabc denoting such a volume, the formula,
XLIV.
= a positive or a
Sa/3y
oabc,
negative scalar, according as the rotation round
OA from ob
to
oc
is
negative or positive.
(30.) More generally, changing
the formula :
XLV.
which
it
DABC = S(a -
^) (/3
to d,
- ^)
(y
and oa or a to a -
- ^) = Sa/3y -
d,
SjSyS
&c we have thus

,
+ Syda - SSa(3
may be observed, that the expression is changed to its own opposite, or

1, when any two ofthe four vectors, a, /3, y, d, or when
negative, or is multiplied by
any two of the four points. A, B, c, D, change places with each other; and therefore
in
is
restored to
former value, by a second such binary interchange.
its
(31.) Denoting then the
XLV.,
XLVI.
and may then write the
EBCD = BECD = &C.)
origin of a,
which
(32.)
A, b, c, d,
And an
XL VIII.
result
by
E,
we have first, by XLIV.,
BCDE i CDEA + DEAB
In
-f
EABC + ABCD =
e may denote any five points of space.
analogous formula (69, III.) of the First Book, for any six points
65, 70),
OA BCDE + OB CDEA + OC DEAB + OD EABC + OE ABCD = 0,

.
being treated as scalar

results.
(comp. 68) under the more symmetric form (because
which the additions are performed according
and
y, d
DABC = EABC EBCD + ECDA EDAB
oabcde, namely the equation (comp.
in
(3,
XLVII.
in
new
the equation,
fact,
XLIX.
to the rules of vectors, the
by
.
or, substituting a, (3, y,
volumes
recovered from the foregoing principles

XLVII., this last formula may be written as
coefficients, is easily
ED. EABC
= EA EBCD + EB ECAD
.
d for ea, eb, eo, ed, as
+ EC, EABD
CHAP.
STANDARD TRINOMIAL FORM FOR
I.]
L.
^Sa/3y
aS/3y5
A VECTOR.
321
+ i3Sya5+ySrt/35;
Mliicli is only another form of XIV., and ought to he familiar to the student.
(33.) The formula 69, II. may be deduced from XXXL, by observing that, when
the three vectors a, (3, y are complanar, we have the proportion,
LI.
Vj3y
Yya
Va/3
V (/3y +
ya +
a/3)
= OBC OCA OAB

:
abc,
signs (ôr algebraic ov scalar ratios) of areas be attended to (28, 63); and the
formula 69, I., for the case of three collinear points a, b, c, may now be written as
if
follows
= 2V(/3y + y + a/3)
= 2V(/?~a)(y-a) = 0,
LII...(/3-y) + /3(y-a)+y(a-/3)
if
the three coinitial vectors a,
The case when four
(34.)
or
to
/3,
y be termino- collinear (24).
coinitial vectors a,
/3,
y,
d are termino-complanar
(^64:)
when they terminate in /owr complanar points A, b, c, d, is expressed by equating

zero the second or the third member of the formula XLV.
(35
Finally, for ternary products
LIII.
The
295.
of vectors in general,
wê
have the formula
a2/32y2 + (Sa/3y)2 ^ (Va/3y)2

(S/3y /3Sya + ySa/3)2
a2 (S/3y)2 + /33 (Sya)2 + y2 (Sa/3)3 - 2S/3y Sy Sa/3.
identity (290) of a right quaternion with its inunit-line as a right versor,
dex, and the conception (293) of an
now
allow us
to treat the three importnnt versors,
?,
J, k, as
constructed by, and even as (in our present view) identical

ivith, their own axes ; or Avith the three lines ox, oj, ok of 181,
considered as being each a certain instrument, or operator^ or

agent in a right rotation (293, (1.) ), which causes any line, in
a plane perpendicular to
itself, to
turn in that plane, through
With
^positive quadrant, without any change of its length.
this conception, or construction, the Laws of the Symbols ijk
are
still
included in the Fundamental Formula of 183, namely,
22=/ = A2=//^=-l;
(A)
and if we now, in conformity with the same conception, transfer
the Standard Trinomial Form (221) from Right Quaternions
an expression of the form,
to Vectors, so as to write generally
=
or
I'.
I.
a = ia +jb + kc, &c.,
p ix +jy + kz,
where xyz and abc are scalars (namely, rectangular co-ordi.
nates),
we
and can,
As
can recover
if
we
think
many
fit,
of the foregoing results with ease

them with co-ordinates,
connect
Fundamental Formula A, the law

be interpreted on the plan of 293, (1.), as representing the reversal which results from two successive quadrantal rotations.
(1.)
i^
=- 1,
to the laws (182), included in the
&c.,
may
2 T
322
The two contrasted
(2.)
[boOK
III.
laws, or formulae,
= + A,
ji
=-
(182, II. and III.)
*,
may now
be interpreted as expressing, that although a positive rotation through a

right angle, round the line i as an axis, brings a revolving line from the position j to
the position k, or + k, yet, on the contrary, a positive quadrantal rotation round the
new axis, brings a new revolving line from a new initial position, i, to a
new final position, denoted by k, or opposite* to the old final position, + k.
= 1 (183) may be interpreted by conceiving, that we
(3.) Finally, the law ijk
linej, as a
operate on a line a, which has at
tions of
i,
first
the direction of
-\-j,
by the three
lines, k,j,
i,
which gives three new but equally long lines, j3, y, d, in the direc+ h, j, and so conducts at last to a line a, which has a direction op-
in succession
posite to the initial one.
The foregoing laws of ijk, which are all (as has been
Formula A, when combined with the recent expression
(4.)
in the
) for
222, (1.)
the square of that vector the value

II.
this
Tp
square of the line p
is
p2
= (ix ^jy + kzy = -
said) included (184)

for p, give (comp.
I.
(.r2
+ 2,5 + z2)
therefore equal to the negative
norm Np
(185), or to the negative of its
of the square of its length

(273), which agrees with the former
resultt 282, (1.) or (2.).

(5.) The condition of perpendicularity of the two lines p and a, when they afe
represented by the two trinomials I. and I'., may be expressed (281, XVIII.) by the
formula,
III.
= Sap = - (oa; +
6y
cz)
which agrees with a well-known theorem of rectangular co-ordinates.

(6.)
The condition of complanarity
oi three lines, p, p, p", represented
by the
trinomial forms,
IV.
is
= ix +jy + kz,
p'
- ix' +
&c.,
p" = ix" +
&c.,
(by 294, VI.) expressed by the formula (comp. 223, XIIL),
V.
= Sp"p'p = x"{z'y - y'z) \ y"(x'z - z'x) + z\y'x -
agreeing again Avith

(7.)
When
known
x'y)
results.
the three hues p, p' p", or op, op', op", are not in one plane, the
by 294, (3.), the volume of the parallelepiped
,
recent expression for 8p"p'p gives,
In the Lectures, the three rectangular unit-lines,
i,
j, k,
were supposed
(in
order to fix the conceptions, and with a reference to northern latitudes) to be directed,
respectively, towards the south, the west, and the zenith ; and then the contrast of
= k, came to be illustrated by conceiving, that we at

ij ==+ k,ji
a moveable line, which is at first directed westward, round an oms
the south, with a right-handed (or screwing) motion,
(or handle^ directed towards
through a right angle, which causes the line to take an upward position, as its ^na'
the two formulae,
one time
one
tur-n
and that at another time we operate, in a precisely similar manner, on a line

first southward, with an axis directed to the west, which obliges this nevf
directed at
line to take finally
t Compare
a downward (instead
of,
as before, an
upward)
direction.
also 222, IV.
CHAP.
PRODUCT OF ANY NUMBER OF VECTORS.
I.]
323
(comp. 223, (9.) ) of which they are edges ; and this volume, thus expressed, is a
to p" is itself
positive or a negative scalar, according as the rotation round p from p'
positive or negative : that is, according as it has the same direction as that round
+ X from +y to +z
(8.) It may be
vectors, then
(or round
from j
VI.
Sa(3y
the important
VIII.
we may
write for
if a, /3,
= + Vy/3a = |(ai3y +
(12.)
y be ang three
),
yi8a).
since a vector, considered as represent-
always (by 144) the opposite ofits own conjugate, so

*
formula,
ing a right quaternion (290),
we have
that
),
= - SyjSa = | (a/Sy - y/3a)
Va/3y
More generally (comp. 223,
(9.)
to i), or the direction opposite thereto.
and V.) we have
III.
(by 294,
VII.
that
noticed here (comp. 223, (13.)
is
Ka = - a,
and
IX.
therefore
KUa = Il'a,
ang number of vectors, the transformations,

X.
XL
sna=+sn'a = Knan'a),
vna = + vn'a = |(na
+n'a),
upper or lower signs being taken, according as that number is even or odd
understood that
XII. . . n'a = ...yi3a, if na = a/3y...
(10.)
The
result at once
follows
it
being
relations of rectangularity,
XIII.
which
Ax.
i -I-
Ax. j
Ax.y -i- Ax. k
Ax. k J- Ax.
may now be
from the definitions (181),
t,
written more briefly, as
XIV. ..i-i-j;
and similarly
in other cases,
kJ-i',
j-i-k,
where the axes, or the planes, of any two right quater-
nions are at right angles to each other.

(11.) But, with the notations of the Second Book,
we might also have writtten,

by 123, 181, such formulas of complanaritg as the following, Ax.^' \\\i, to express
(comp. 225) that the axis oij was a line in XhQ'plane of i and it might cause some
;
we were now to abridge that formula to j [||

that we should not henceforth employ the sign
confusion, if
convenient
either symbols
of three
lines,
considered
still
as
i.
\\\,
complanar; or
In general,
it
seems
except as connecting
else
symbols of three
right quaternions, considered as being collinear (209), because their indices (or axes^
are compla7iar : or finally, any two complanar quaternions (123).
(12.) On the other hand, no inconvenience will result, if we now insert the sign of
parallelism, between the symbols of two right quaternions which are, in the former
sense (123), complanar
for
example,
we may
XV...xi\\i,
if
xyz be any three
If,
in
write,
yjy,
on our present plan,
zk\\k,
scalars.
Uke manner, we
as equivalent to Ul'â,
interpret,
on our present plan, the symbols Ua, Ta, Na

are reconducted (compare the Notes to
Tl'a, Nl'â, we
page 136) to the same significations of those symbols as before (155, 185, 273)
it ia evident that on the same plan we have now,
Sa =
0,
Va = a.
and
324
[bOOK
III.
296. There are a few particular but remarkable cases^ of ternary

and oihQv products of vectors, which it may be well to mention here,
and of w^hich some may be worth a student's while to remember:
especially as regards the products
inscribed in circles, or in spheres.
(1.) If A, B, c,
of successive
sides
of closed polygons,
we know, by
be any four concircular points,
the sub- articles to
260, that their anharmonic function (abcd), as defined in 259, (9.),

also positive or negative, according to a
which has been already
is
scalar; being
law of arrangement of those four
points,
stated.
(2.) But, by that definition, and by the scalar (thoMgh negative^ character of the
square of a vector (282), we have generally, for ang plane or gauche quadrilateral
ABCD, the formula
:
I.
iu
e2(ABCD) = AB.BC.CD.DA=<Ae continued product of the four sides;
which the
or of
two
coefficient e^ is a positive scalar,
positive squares, as follows

II.
(3.) If then
abcd be
e2
namely the product
of
two negative
= BC2
Da2
= BC2. DA^ > 0.
a plane and inscribed quadrilateral,
we have, by
260, (8.),
the formula,
III.
AB BC CD DA = a positive
.
according as this quadrilateral in
or negative scalar,
circle is a crossed or
an uncrossed one.
The product
a(3y of ang three complanar vectors is a vector, because ita

scalar part Sa/3y vanishes, by 294, (3.) and (4.); and if the factors be three sue
cessive sides AB, bc, cd of a quadrilateral thus inscribed in a circle, their product has
(4.)
either the direction of the fourth successive side, DA, or else the opposite direction,
or in symbols,
IV.
according as the quadrilateral

(5.)
By
ab.bc.cd da >
or
abcd
is
< 0,
an uncrossed or a crossed one.
conceiving the fourth point d to approach, continuously and indefinitely,

we find that the product of the
to the first poi7it A,
three successive sides of
ang plane triangle, abc,

given by an equation of^ the form
y^
is
V.
AT being a
line
AB BC CA = AT
.
-^L^:^^
(comp. Fig. 63) which touches the
circle,
or
presents the initial direction of motion, along the circumference, from A through 'B to C: while the length
of
tliis
tangential product-line, AT,
is
represents, with the usual reference to
of the three
sides, of the
same inscribed
equal
an
U7iit
to,
(more fully) which touches

or re~
the segment ABC of that circle, at the point a
circumscribed
^^\c
^^,.,~--^
//''
ÔNj^^;^^^^
-^^S*
"^
or
of length, the product of the lengths
triangle abc.
theorem respecting the product of the sides of an inscribed

triangle be supposed to have been otherwise proved, and if it be remembered, then
(6.) Conversely, if this
since
it
will give in like
manner the equation.
CHAP.
PRODUCTS OF SIDES OF INSCRIBED POLYGONS.
I.]
VI.
i( T>
AC.CD.DA=AU,
325
he ant/ fourth point, concircular with A, B,
Au
while
c,
is,
as in the annexed
Figures 63, a tangent to the new segment ACD, we can

recover easily the theorem (3.), respecting the product
the sides of an inscribed quadrilateral ; and thence
of
can return to the corresponding theorem (260, (8.) ),

respecting the anharmonic function of any such figure
ABCD
we
for
shall thus have,
by V. and
VI., the
equation,
AB BC CD DA = (aT AU) (CA Ac),

in which the divisor cA. AC or N. ac, or AC^ is always
positive (282, (1.) ), but the dividend at.au is negaVII.
fT
Fig. 63,
bis.
foi^ the case of an uncrossed quadrilateral (Fig. 63), being on the

contrary positive for the other case of a crossed one (Fig. 63, bis').
a given point A, which touches at a
(7.) If p be afljrpoint on the circle through
given origin o a given line ox = r, as represented in Fig. 64, we shall then have by
tive (281, (9.))
(5.) an equation of the form,
VIII.
in
which x
is
oA.AP.PO=a;.OT,
some scalar
coeflScient,
OA =
a,
and op =
IX.
IX'.
IX'
p, as usual,
a (p
p-'
.
a) jO
':
Vrp-1
which
Making then
varies with the position of p.
we
have
shall
= XTy
XT
a'ipi,
Vra-i
64.
Fig.
and any one of these may be considered as a
form of the equation of the circle, determined by the given conditions.
(8.) Geometrically, the last formula IX." expresses, that the line p"i-a-i, or
Rp - Ra, or a'p' (see again Fig. 64), if oa' = a"' = Ra = R. OA, and op' = p-^ = R. op,
is
parallel to the given tangent t at
which agrees with Fig. 58, and with Art.
260.
(9.) If B be the point opposite to o upon the circle, then the diameter ob, or
as being 4- r, so that tjS'^ is a vector, is given by the formula,
X.
in
.. Ti3-i
= Vra-i;
which the tangent r admits, as
or
X'.
/3
=-r
Vra"!;
ought to do, of being multiplied by any

without the value of j3 being changed.
(10.) As another verification, the last formula gives,
XI.
it
OB = T(3 = Ta T VUra' = 5a
:
(3,
sin
scalar,
aot.
(11.) If a quadrilateral oabc be not inscriptible in a circle, then, whether

plane or gauche, we can always circumscribe (as in Fig. 65) tico circles, oab and
it
be
obc
about the two triangles, formed by drawing the diagonal ob and then, on the plan
of (6.), we can draw two tangents ox, ou, to the two segments cab, obc, so as to
represent the two ternary products,
;
326
and
OA.AB.BO,
after
which we
XII.
where the
shall
[bOOK
III
OB.BC.co;
have the quaternary product)
OA.AB.BC.CO = OT.OU:OB4;
bo ob,
divisor, oB^, or
or
is
ob,
positive scalar, but the dividend or. ov, and therefore also the quotient in the
product in thejirst member,
second member, or the

is a quaternion.
(12.) The axis of this quaternion is perpendicular to the plane TOU of the two tangents ; and
therefore
OABC,
if
to the
then the axis
is
at the point o
is
ou,
itself of the quadrilateral
plane
that be a plane figure
but
if it
be gauche,
normal to the circumscribed sphere

being also in
such, that the rotation round
all cases
it,
from ox
to
positive.
(13.)
The angle
of the
same quaternion
tween the two tangents above mentioned
is
it is
the supplement '^Pthe angle
tou
be-
therefore equal to the angle u'ot,
if
ou' touch the new segment OCB, or proceed in a new and opposite direction from o
it may therefore be said to be the
(see again Fig. 65)
angle between the two arcs,
;
OAB and OCB, along which a point should

circumferences, to the opposite corner
move, in order to go from o, on the two

B of the quadrilateral oabo, through the two
A and c, respectively or the angle between the arcs ocb, oab.

(14.) These results, respecting the axis a.nd angle of the product of the four suc-
other corners,
cessive sides, of
any quadrilateral oabc, or abcd, apply without any modification
to
the anharmonic quaternion (259, (9.)) of the same quadrilateral; and although,
for the case of a quadrilateral in a circle, the axis becomes indeterminate, because
the quaternary product and the anharmonic function degenerate together into scabecause the figure may then be conceived to be inscribed in indefinitely many
lars, or
be determined by the same rule as in the general

and uncrossed quadrilateral (Fig. 63)
for the inscribed and crossed one (Fig. 63, bis).
spheres, yet the angle
case
but
may
this angle being -=
= 0,
ir,
still
for the inscribed
(15.) For the gauche quadrilateral oabc, which

inscribed in
a determined sphere, we
may
say,
by
be conceived to be
may always
(13.), that the angle
of the qua-
ternion product, Z. (oA.AB.BC.Co), is equal to the angle of the lunule, bounded

the two arcs of small circles oab, ocb ; with the same construction
(generally) by
for the equal angle of the anharmonic^
L (oabc),
L (oA ab. bc
or
co).
that the general principle 223, (10.), of the permissibility of

(16.) It is evident
of quaternion factors under the sign S, must hold good for
cyclical permutation
the ease
when
those quaternions degenerate (294) into vectors
obvious, that every permutation of factors

cyclical permutation
is
XVI.) under
the sign
Hence
generally, for
any four
XIII.
Sa/3y5
= S/Syâ
XV.
z:
and
it is still
allowed, under the sign
again allowed, under this other sign
also (comp. 196,
(17.)
is
SU
and consequently
C
vectors,
XIV.
a/3y5
we have
.
the three equations,
SUa/3yo
^ySa
more
whence
= SU/Syâ
CHAP.
and
PENTAGON
I.]
in particular, for
we have
IN A SPHERE.
327
the guccessive sides of any plane or gauche quadrilateral abcd,
the four equal angles ^
XVI.
L (ab
bc CD da)
L (bc CD
.
da ab) = &c.
with the corresponding equality of the angles of the four anharmonics,
XVII.
Z.
=L
(abcd)
(bcda)
Z.
=Z
(cdab)
(dabc)
or of those of the four reciprocal anharmonics (259, XVII.),
XVII'.
I (adcb)
/.
(badc)
Z.
(cbad)
=Z
(dcba).
(18.) Interpreting now, by (13.) and (15.), these last equations,

them the following theorem, for the plane, or for space :
Let abcd be any four points, connected hy four

passing through three of the points
we
derive from
each
circles,
then, not only is the angle
at a, between the arcs abc, adc, equal to the angle at c, between cda and cba, but also it is equal (comp. Fig. 66) to the
angle at b, between the two other arcs bcd and bad, and to
the angle at d, between the arcs dab, dcb.
(19.) Again, let abcde be smy pentagon, inscribed in a

; and conceive that the two diagonals AC, ad are drawn.
sphere
"We
have three equations, of the forms,
shall then
XVIII.
ab.bc.ca = at; ac.cd.da = au

AD DE EA = ay
where at, au, av are three tangents
XIX.
to the sphere at A, so that their

product is
But the equations XVIII. give
fourth tangent at that point.
= (at AU Av) (^C^

aP^)
which touches the sphere at A.
AB.BC CD DE EA
.
= AW = a new
vector,
have therefore this Theorem, which includes several others under it

" The
product of the five successive sides, of any {generally gauche) pentagon
inscribed in a sphere, is equal to a
tangential vector, drawn from, the point at which
"VVe
and
the pentagon begins
ends."
p be a point on the sphere which passes

through
three given points a, b, c
we shall have the equation,
(20.) Let then
o,
and through
XX.
..
= S(oA.AB.BC.CP.Po) =
Sa(/3-a)(y-^) (p_y)(_p)
= a2S/3yp + (S^Sy ap + y^SajSp p^SaySy.
(21.)
Comparing with 294, XIV., we
tial vectors a, /3,

y,
their
common
see that the condition for the four co-ini-
p thus terminating on one spheric surface, which passes through

may be thus expressed
origin o,
XXL
if
p = xa + y/3 + zy,
then
p2
= xa* + y(3^ + zy^.
(22.) If then we project (comp. 62) the variable point p into points a',
b', c' on
the three given chords OA, OB, oc, by three planes
through that point p, respectively
parallel to the planes boc, coa, aob, we shall have the equation :
XXII.
p,
op2
= oa.oa' +
+ oc.oc'.
ob.ob'
(23.) That the equation XX. does in fact represent a spheric locus for the
point
is evident from its mere
form (comp. 282, (10.) ) and that this sphere passes
;
328
through the four given points^ o, A,

tion
when we change p
is satisfied,
may be proved by
B, c,
any one
to
OD
(24.) Introducing an auxiliary vector^
XXIII.
or
by the system
^Sa/3y
XX.
is
= a2Y)Sy+/32Vya + yVaj3,
),
= S^y,
Sâ, /32 = S^/3,
y2
Sâ-' = S^/S'i = S^y-i = 1
of the sphere becomes simply,
XXV.
D
p2
XXV'.
or
s^p,
S^p-
the point of the sphere opposite to o, and ^
is
a diameter (comp. 282,
IX'.; and 196, (6.)).

(25.) The formula XXIII., which determines this diameter,
this other
wav
may
be written in
XXVI.
XXVI'.
or
y.
j3,
determined by the equation,
d,
a2
XXIV.
or
so that
or
III.
observing that the equa-
of the four vectors, 0, a,
of the three scalar equations (comp. 294, (25.)
XXIV.
the equation
[bOOK
^Sa/5y =
Va03-a) (y-/3)y;
= -V(0A.AB.BC.C0)
6.0ABC.0D
where the symbol oabc, considered as a
coefficient^ is interpreted as in
namely, as denoting the volume of the pyramid oabc, which
is
XLIV.
294,
here an inscribed
one.
(26.) This result of calculation, so far as
it
regards the direction of the axis of
and may be used to confirm, the theorem

(12.), respecting i\\Qproduct of the successive sides of a gauche quadrilateral, oabc
including the rule of rotation, which distinguishes that axis from its opposite.
the quaternion
oa.ab.bc.co, agrees
with,
(27.)
The formula XXIII.
XXVII.
for the diameter d
o\ Sa-i/3-i y-i
may
also be thus written
= V (/3-1 yi 4 y' a-i +
a-i /J-i)
= V(/3-i-a-i)(y-i-a-i);
and the equation XX.
of the sphere
XXVIII.
which expresses (by 294,
be transformed to the following
may
= S (/3-1 -
a-i) (y-i
a-i)
(pi
- a'')
comp. 260, (10.)), that the four reciprocal vec-
(34.),
tors,
XXIX.
OA'=a'=a-i,
OB'=/3'
= /3-^
oc'=y' =
y-i,
op'
= |o'=p-^
the plane a'b'cV, in which they all terminate, being

because the perpendicular let fall
parallel to the tangent plane to the sphere at o
are termino-complanar (64)
on
this plane
from o
is
XXX.
.b'
= d-\
as appears from the three scalar equations,
XXXI.
(28.) In general,
then
if
d be
XXXII.
because this expression
Sa'5
= S/3'^ = Sy'S = 1.
the foot of the perpendicular from o, on the plane abc,

.. d
XXXIII.
= Sa/3y
V(/3y
+ ya + a/3)
and may be deduced from, the three equations,
satisfies,
.
Sa5-i
= 8/3^-1 = Sy^-i =
1.
As
a verification, the formula shows that the length TS, of this perpendicular, or
altitude, OD, is equal to the sextuple volume of the pyramid oabc, divided by the double area
of the triangular base ABC.
(Compare 281,
(4.),
and 294,
(3.), (33.).)
CHAP.
EQUATION OF HOMOSPHERICITY.
I.]
329
have been obtained by the

(29.) The equation XX., of the sphere oabc, might
. and
elimination of the vector d, between the four scalar equations
XXV., on
XXIY
the plan of 294, (27.).

to the deve(30.) And another form of equation of the same sphere, answering
lopment of XXVIII., may be obtained by the analogous ehmination of the same vector
between the four other equations
^,
XXIV.
and XXV'.
is generally a quabut the product of the successive

sides of a hexagon abcdef, or any other even-sided figure^ inscribed in a circle, is
a scalar : because by drawing diagonals AC, ad, ae from the first (or last) point a
The product
(31.)
any even number
of
of
complanar vectors
ternion with an axis perpendicular to their plane
of the polygon, we find as in (6.) that it differs only by a scalar coefficient, or divisor,
from the product of an even number of tangents, at the first point.
(32.) On the other hand, the product of any odd number of complanar vectors is
always a line, in the same plane; and in particular (comp. (19.)), the product of
the successive sides of a pentagon, or heptagon, &c., inscribed in a circle, is equal to
a tangential vector, drawn from the first point of that inscribed and odd-sided poly-
gon
because
it difiers
only by a scalar coefficient from the product of an odd num-
ber of such tangents.

of any number of lines in space is generally a quaternion
they be the successive sides of a hexagon, or other even- sided polygon,
inscribed in a sphere, the axis of this quaternion (comp. (12.) ) is normal to that
The product
(33.)
(289)
and
if
sphere, at the initial (or final) point of the polygon.
(34.) But the product of the successive sides of a heptagon, or other odd-sided
polygon in a sphere, is equal (comp. (19.) ) to a vector, which touches the sphere at
the initial or final point
because it bears a scalar ratio to the product of an odd
;
number
of vectors, in the tangent
(35.)
The equation XX.,
or
plane at that point.
its
XXVIII., may be
transformation
called the con-
dition or equation
c,
of homosphericity (comp. 260, (10.)) of the five points o, A, B,

and the analogous equation for the five points abode, with vectors apydt
from any arbitrary origin
may be written thus

= S(a-/3) (iS-y) (y- d) {S- e) {e XXXV. .. = aa* + fc/32 + cy2 + dd^ + ea^,
o,
XXXIV...
or thus,
six times the second
member
of this last formula being found to be equal to the se-
cond member of the one preceding
XXXVI.
or
more
= BCDE,
it,
if
= CDEA,
= DEAB,
d = EABC,
= ABCD,
fully,
XXXVII.
so that,
a);
6a = S (y
- /3)
(5
- /3)
(6
- /3) = S (yê -
by 294, XLVIII. and XLVII., we have
XXXVIII.
with the relation between the
.0 = aa +
or condition
&c.
+ cy + dd+ee,
+b+
which allows (as above) the origin of vectors
The equation
+ ejSy - fSyS),
also (comp. 65, 70) the equation,
coefficients,
XXXIX.
(36.)
b(3
Se(3
+ d + e,
to be arbitrary.
XXXV. may
be obtained as the result of an
elimination (294, (27.) ), of a vector k, and of a scalar g, between

ûe scalar equations of the form 282, (10.) namely the five
following,
2u
330
XL.
a2-2S/ca + ^ =
0,
K being the vector of the centre

written as
/32
p2_2Sx:p + 5r
III.
of which the equation
Abcd,
and on which, by
2_2S/c + ^
</
of the sphere
XLL.
g being some scalar constant
point E is situated.
- 2Skj3 + = 0,
[bOOK
may
be
= 0,
condition referred
tlie
to,
the Jifth
(37.) By treating this fifth point, or its vector e, as arbitrary, we recover the
condition or equation of concircularily (3.), oi lh.Q four points A, b, c, d; or the
formula,
XLIL
(38.)
= V(a-
The equation of the
in general
be written thus
circle
/3)
03-7) (y-5) (5-a).
ABC, and the equation of the sphere abcd,
may
XLm...0 = V(a-/3)(/3-r)(r-p)(p-a);
XLIV. ..0 = S(a-/3)(/3-y)(y-^)(^-p)(p-a);
other locus.
p being as usual the vector of a variable point p, on the one or the
to the circle abc, and of the tangent plane
(39.) The equations of the tangent
abcd, at the point
to the sphere
XLV.
XLVL
and
(40.) Accordingly, whether

or
XLIV. and XLVI., we
XLVIL
a, are respectively,
0=V(a-/3) (^-y) (y-) (p-a),

= S(a-/3) {(3-y) (y-^) (5-a) (p-a).
we combine
the two equations XLIII. and
XLV.,
find in each case the equation,
- a)2 =
(p
gi\ing
0,
or
=A
(20)
the four
being supposed that the three points A, b, c are not collinear, and that
points, A, B, c, D are not complanar.
it
of the sphere
(41.) If the centre
XLVIII.
abcd be taken
a2=|32 = y2=^2 = _r2,
or
for the origin o, so that
XLIX.
Ta = T/3 = Ty = T5 = r,
the positive scalar r denoting the radius, then after some reductions
we
obtain the
transformation,
..V(a-i8)(^-y)(y-^)(^-a) = 2aS(/3-a)(y-a)(^-a).
if k be, as in (36.), the centre of the sphere, we have the
(42.) Hence, generally,
L.
equation (comp. XXVI'.),

LI.
V(ab.bc.cd.da) = 12ka.abcd.
enunciate this theorem :

(43.) "We may therefore
" The vector
successive
the
sides, of a gauche quadrilaproduct offour

part of
teral inscribed in a sphere, is equal to the diameter drawn to the initial point of the
volume of the pyramid, which its four points depolygon, multiplied by the sextuple
termine.''^
reductions (41.), the following general for mulce of trans(44.) In effecting the
may be useful on other occasions
formation havp been employed, which

LII.
where a
may
aq + qa=^1 {aSq + Sqa)
be any
vector,
LII'.
aqa = a^Kq + 2aSqa
and q may be any quaternion.
CHAP.
FOURTH PROPORTIONAL TO DIPLANAR VECTORS. 331
I.]
Section
On
7-
the Fourth Proportional to Three
Diplanar
Vectors.
297. In general, when any fovr quaternions,
the equation of quotients,
I.
q"'\ q''
q^ q',
q^\ q"\ satisfy
q'\q,
or the equivalent formula,
ll...q"'={q':q),q" = qYq",
we
form a Proportion ; and that the fourth,

the Fourth Proportional to the^rî, second^ and third
shall say that they
namely
q'", is
quaternions, namely to ', q', and q", taken in this given order.
This definition will include (by 288) the one which was assigned in
226, for the fourth proportional to three complanar vectors, a, yS, 7,
fourth vector in the same plane, ^= /3a"^7, which has been
already considered; and it will enable us to interpret (comp. 289)
the symbol
namely
i\i2i\,
III.
Pa^<î,
when 7
not
\\\
a, /3,
new case,
as denoting not indeed a Vector, in this
hut at least a Quawhich

be
the
called
ternion,
may
present general plan) the Fourth
(on
Three
to
these
Proportional
Diplanar Vectors, a, /3, 7. Such fourth
some
interesting
properties, especially with reproportionals possess
ference to their vector parts, which it will be useful briefly to consider,
and to illustrate by showing their connexion with spherical

and generally with spherical geometry.
trigonometry,
be (as in 208, (1.), &c.) the vectors of the corners of a triangle

/3, y
the unit-sphere, whereof the sides are a, 6, c ; and let us write,
(1.) Let a,
ABC on
a
IV.
(n
where
it is
being also at
three cosines,
less
=
=
cos 6
cose
=
=
= -S/3y,
= - Sya,
Si3a-i
Sa^;
understood that
V.
it
Say'i
=cosa=Sy/5-i
/,
a2 = /32
= y2=-l,
or
VI.
Ta = T/3 = Ty =
supposed, for the sake of fixing the conceptions, that each of these
m, n, is greater than zero, or that each side of the triangle abc is
first
than a quadrant.
(2.) Then, introducing three
VII.
new
vectors, 5,
= Yy^-â =
f , ^,
Va/3-'y
defined
= ny
-\-
la
by the equations,
mfi^
332
we
VIII.
IX.
so that
(3.) This
have the
shall
than unity ; for
less
r, is
length,
Sa/37
if
we
+ r2=N/3a-V =
e2
(4.) Dividing the three lines ^,
l;
the corners are determined
(5.)
The
.OD =
m=
? by their length,
new
versors (155, 156); and so obtain a
write,
= S/3a-V = e,
scalar e is different from zero, because the vectors a,
XII.
length, say r, because
relation,
XL
and the
common
common
one
N^=N = N? = ;2 + ^2 + 2_2Zmn = r2;

T^ = T = T? = r = V(Z2 + mS + 7i2 _ 2/mw).
all
III.
namely,
X,
we
have
find that these three derived vectors
they have one common norm
[book
new
three
r-l^;
OE = Ue =
y are diplanar.
to their
def, on the unit-sphere, of which
triangle,
by the
/3,
we change them
r,
unit-vectors,
r-i;
sides opposite to d, e, f, in this
new
or de-
rived triangle, are bisected, as in Fig. 67, by the corners

A, B, c of the old OT given triangle ; because we have the d^
three equations,
XIII.
+ ? = 2Za;
show
&c.), the equations XIII.
XIV.
cos
d+E = 2ny.
+ ^=2î3;
(6.) Denoting the halves of the
a=
by a', 6', c' (so that the arc ef = 2<

by IV. and IX., that
new
also,
sides
= r cos 6',
cos b
r cos a',
cos c
=r
cos
c'
the cosines of the half-sides of the new (or bisected^ triangle, def, are therefore ^r
portional to the cosines of the sides of the old (or bisecting) triangle ABC.
(7.)
The equations IV.
XV.
we have
.. 2Z
by
therefore,
by 279,
give,
may
2n
= - (a^ + j3a)
VII., the three following equations between quaternions,
XYI...as =
which
(1.),
2m = -(ya + ay),
= -(^y + -y/3),
Ka,
(3K
= S(3,
yd = ay;
= (5d,
dy = y,
XVI'.
= aK,
?/3
and express in a new way the relations of bisection (5.).

(8.) We have therefore the equations between vectors,
XVII.
XVir.
or
(9.)
Hence
also,
= aZa'\
=
XVIir.
= /3^/3-i,
^ = /3^/3-i,
aa-i,
by V., or because
XYIII.
or
.. 2
a,
y are
(3,
= yy-
unit-vectors,
= -aKa,
= -I3^(3,
.Z=- am,
= - /3?/3,
.s
f=y^y-i.
= -ysy;
= - y^y.
(10.) In general, whatever the length of the vector a mag be, the first equation
line e is (comp. 138) the reflexion of the line c,, with respect
XVII. expresses that the

to that vector a
because
;
XIX.
(11.) Another
mode
it
may
be put (comp. 279) under the form,
Kfa-
or
of arriving at the
XIX'.
..
fa-i=Kâ-
same inteipretation
of the equatio
CHAP.
= aZa-\
rallel
is to conceive ^ decomposed into two summand vectors,

and the other perpendicular to a, in such a manner that
XX.
for
we
then
333
EXPRESSIONS FOR CONICAL ROTATION.
I.]
shall have,
XXI.
..?=r+r,
by 281,
riia,
and
C one pa-
r-j-;
(10.), the transformations,
= aK'a-^ + aK"a-^ = âa-i - K"aa-i = ^' _ ^"
the parallel part of ^ being thus preserved, but the perpendicular part being reversed,
hy the operation a
)a-^.
= aâ'^ to the form ea = a?, that is, to the first
(12.) Or we may return from
and then this equation between quaternions will show, as suggested
equation XVI'.
in (7.), that whatever may be the length of a, we must have,
(^
XXII. ..T = T^,

so that the two lines
that from
to ^
Ax.*a = Ax.a^,
Lea=LaKi
^ are equally long, and the rotation from
to
is
equal to
common
these two rotations being similarly directed, and in one
plane.
(13.) "We
XVII. XVII'. under
also write the equations
may
XXIII.
= a-iâ,
&c.
XXIII'.
the forms,
..Zâ'ha,
&c.
(14.) Substituting this last expression for ? in the second equation XVII'.,
derive this
new
XXIV.
that
more
is,
XXV.
An
(15.)
= i3a-if a/3-1
XXIV.
or
= a/3-i^/3a-
briefly,
we
equation,
= 9g-i,
XXV'.
and
= 5-1^^,
if
XXVI.
5 =â-J.
namely one with such a symbol as
expression of this form,
XXVII.
9 (
) 9-1
an operator, occurred before, in 179, (1.), and in 191, (5.) and was seen to indicate a conical rotation of the axis of the operand quaternion (of which the symbol
is to be conceived as being written within the parentheses), round the axis of
q,
for
through an angle =2 Lq, Avithout any change of the angle, or of the tensor, of that
operand; so that a vector must remain a vector, after any operation of this sort, as
being
sttll
a right-angled quaternion (290)
XXVIII.
or (comp. 223, (10.) ) because
= Sg-i^p = Sp = 0.
S9P5-1
we
conceive two opposite points, p' and p, to be determined on the

unit-sphere, by the conditions of being respectively the positive poles of the two op(16.) If then
posite arcs,
AB and
XXIX.
we can
infer
op'
from
ba, so that
= Ax. /3a-i = Ax.
XXIV.
nical rotation round the line Op' as

the angle
aob
(17.)
It
because
And
(if
o be
in like
still
g,
that the line
an
and
od may
employing
a)3-i
= Ax.
5-1,
axis, through
an angle equal
to the double
of
the cenire of the sphere).
manner we can
from XXIV'., that the
infer
line
OB admits
in 291, that this characteristic Ax. can be dispensed with,

but there may still be a convenience in
admits of being replaced by
was remarked
it
op = p'o = Ax.
be derived from the line OE, by a co-
UV
it
occasionally.
334
[book
III.
of being derived from od, by an equal but opposite conical rotation, round the line
as a new positive axis, through an angle equal to twice the angle boa.
OP
To
(18.)
drawn
and other connected
illustrate these
which p
in
results, the
annexed Figure 68
is
represents, as above,
the positive pole of the arc ba, and arcs are
drawn from
circle
it
through
(The other
to d, e, f, meeting the great
a and b
in the points r, s, t.
letters in the Figiu*e are not, for
the moment, required, but their significations will soon
be explained.)
(19.) This being understood, we see,

first, that because the arcs ef and fd are
bisected (5.) at
A and
b, the
thi-ee
arcual
perpendiculars, ES, ft, dr, let fall from E,

F, D, on the great circle through A and b,
are equally long; and that therefore the
P is the interior pole of the small cir-
point
cle def', if f' be the point diametricallv
Fig. 68.
op-
posite ioF: so that a conical rotation round
pole p, or round the axis op, would in fact bring the point D, or the line od,
is one part of the theorem (17.)-
this
to
the position E, or OE, which
(20.) Again, the quantity of this conical rotation, is evidently measured by the
arc RS of the great circle with p for pole ; but the bisections above mentioned give
(comp. 165) the two arcual equations,
XXX.
RB= '^ BT,
f^
TA = '> AS
whence
XXXI.
'^
rs
= 2 ba,
and the other part of the same theorem (17.) is proved.

(21.) The point p may be said to be the reflexion, on the sphere, of the point D,
with respect to the point b, which bisects the interval between them and thus we
;
say that two successive reflexions of an arbitrary point upon a sphere (as here
froraD to F, and then from f to e), with respect to two given points (b and A) of a
may
given great circle, are jointly equivalent to one conical rotation, round the pole (p) of
that great circle ; or to the description of an arc of a small circle, round that/)o/e, or
parallel to that great circle
is
and that the angidar quantity (dpe)
of this rotation
by the arc (ba) connecting the two given points ; or is

angle (bpa), which that given arc subtends, at the same pole (p)_
double of that represented
the double of the

(22.) There is, as
rotation
but
it
is
we see, no diflSculty in geometrically proving this theorem of

remarkable how simply quaternions express it : namely by the
formula,
XXXII.
in
which
a,
(5,
may
a.(3-^pl3.a-^=ai3-\p.(3a-\
denote any three vectors
and which, as we
see
by the points^
involves essentially the associative principle of multiplication.

(23.) Instead of conceiving that the point D, or the
,- 'JF/
has been reflected into the position f, or of,

with respect to the point b, or to the line ob, with a similine OD,
lar successive reflexion
from F
to e,
we may
a point has moved along a small semicircle, with b for

from D to f, as indicated in Fig. 69, and then along
pole,
R B
SU
conceive that
Fig. 69.
CHAP.
I.]
CONSTRUCTION OF A FOURTH PROPORTIONAL.
another small semicircle, with A for pole, from f to e

of these two successive and semicircitlar motions
an arc de of a third small
circle,
which
is
and we
effect,
is
335
see that the result, or
equivalent to a motion along
parallel (as before) to the great circle

as before) is double of
(still
through B and A, and has a projection us thereon, which

the given arc ba.
(24.) And instead of thus conceiving two successive arcual motions of a point D
upon a sphere, or two successive conical rotations of a radius OD, considered as compounding themselves into one resultant motion oi ih-At point or rotation of that ra,
we may
conceive an analogous composition of two successive rotations of a

solid body (or rigid system), round axes passing through a point o, which \s fixed in
space (and in the body) and so obtain a theorem respecting S7ich rotation, which
dius,
from what precedes, and on which we may perhaps return.

draw some additional consequences from the equations VII., &c., and
easily suggests itself
(25.) But to
from the recent Fig. 68, especially as regards the Construction of the Fourth Proportional to three diplanar vectors, let us first remark, generally, that when we have
(as in 62) a linear equation, of the
form
connecting /o?^r co-initial vectors a

fifth vector,
is
= aa+
evidently complanar (22) with a,
S,
0,
whereof no three are complanar, then this
bfi=
j3,
d^=
aa^^b^^^ cy
cy
dS,
and also with
y, S
(comp. 294, (6.)
it is
of intersection of the plane aob, cod, of these
therefore part of the indefinite line
two pairs of vectors.

(26.)
And
if
we
divide this fifth vector
lars,
a+
and
6,
by the two (generally unequal) sca-
d,
the two (generally unequal) vectors,
(aa +
i/3)
and
(a + b),
(cy + d^)
(c
d),
which are obtained as the quotients of these two divisions, are (comp. 25, 64) the
vectors of two (generally distinct) points of intersection, oi lines withf/aMe, namely
the
two following
AB OCD,
cd oab.
and
(27.) When the two lines, ab and CD, happen to intersect each other, the two
last-mentioned points coincide ; and thus we recover, in a new way, the condition
(63), for the complanarity of t\iQ four points o, A, B, c, or for the termino-complanarity of the four vectors a, (3, y, 8 ; namely the equation
which may be compared with 294,

(28.) Resuming
now
XLV. and
L.
the recent equations VII., and introducing the
XXXIII.
\=
Za
new
vector,
= i(6 - 5),
7/3
which gives,
XXXIV.
we
see that the
two
Sy\ =
0,
and
XXXV.
TX = \/(r^ -n*)=r
sin
c',
arcs ba, de, prolonged, meet in a point l (comp. Fig. 68), for
which OL = U\, and which
is
distant by a quadrant from c
a result which
may be
confirmed by elementary conaiderations, because (by a well-knolv^n theorem respect-
336
iug transversal arcs) the

the third side in a point
common
ba
bisector
sinDL=
(29.) To prove by quaternions this
common value, we have only to observe
XXXVI.
in which,
of the
two
sides,
de and
ef,
must meet
which
for
i>,
III.
[bOOK
sin el.
of sines, and to assign their
last equality
that
by XXXIII.,
V5X = Vf\ = ^V5;
TA = T\=r2 8inc',
TVê =
and
r2 sin 2c'
the sines in question are therefore (by 204, XIX.),
XXXVr.
On
(30.)
TVU^X = TVUA = ir2 sin 2c'
similar principles,
XXXVII.
in
we may
V/3X
interpret the
= ZV/3a,
= cos
r^ sin c'
c'.
two vector-equationsy
VaX = wVjSa,
which
XXXVIII.
an equivalent
(31.) Accordingly,
_-
we
if
and
bisects RS,
XL.
nd
tion (18.),
c'
sin c
= tan c':
is
will be
tan
CD
cosBC
cos
tan
AB
sin
bl
sin
let fall
tan
c,
'^
LM = '^ AB =
quadrants
ac
al
the perpendicular CQ on ab (see again Fig.
we determine two new
if
tions,
the arcs mr,
sin
to the trigonometric equations,
^^
so that
TX TV/3o = r
the pole of dr, and
'^
n by
points m,
r.
QR,
LN = n CD,
and because the angle at
dm
is
6i
the arcual equi
a quadrant; whence
is
right by construe
is
the pole of
and the angle lnm is right conceiving then that the arcs CA and cb are drawn,
have three triangles, right-angled at q and n, which show, by elementary principle
:
that the three trigonometric quotients in

namely cos cq, or cos l.
(32.)
To prove
Mary points m, n,
this last result
q, r,
we have
XLI...co,L =
XLII.
XXXIX.
have in
fact a
common
valu
by quaternions, and without employing the auJ
= T^.S^=T^;
SU^=Sn'^"
VJj
V/3a
\pa
y\
yX
= ny-X,
UV^f = UyX,
V^f = 2yX,
=y+X,
and
XLIII...S^ = ?^=-Sâ-yX-'=.-S^X-'=l,
yX
it
(yX)-
being remembered that X -L
y,
VyX = yX = -Xy,
whence
(yX)2 =
y2X
= X2,
SyX'l
= 0.
At the same time we see that if p be (as before) the positive pole of ba,
k, k' be the negative and positive poles of de, while l' is the negative (as l
(33.)
and
is
if
the positive) pole of cq, whereb}' all the letters in Fig. 68 have their significations
determined,
we may write,
OP = UVâ
XLl V.
while
ok'
= yUX
OL = + UX,
ok = - yUX
as before.
ol'
= - UX
CHAP.
ANGLE OF A FOURTH PROPORTIONAL.
I.J
337
(34.) Writing also,
= - y\, or X = y/c, and /A = /3a-'X,

XLV. OK = Uk, and om = U/t,
= pi\-i.XK- =/iK-i
XLVI.
/3a-i.y
XLV.
we have
/c
so that
this fourth proportional, to the three equally long hut
therefore
tive
diplanar vectors, a,
/3,
y,
is
which the representative arc (162) is km, and the representakdm, or l'dr, or edp and we may write for this versor, or qua-
versor, of
is
angle (174)
ternion, the expression
XLVII.
/3a"iy
= cos l'dr + od
sin l'dk.
(35.) The double of this representative angle is the sum of the two base-angles of
the isosceles triangle dpe ; and because the two other triangles, epf', f'pd, are also
isosceles (19.), the lune ff' shows that this sum is what remains, when we subtract
the vertical angle F, of the triangle def, from the
d and e
base-angles
same
gles of the
of that triangle
triangle
or
when we
sum of the supplements of the two

subtract the sum of the three an-
ôm four right angles. We have therefore this very simple
expression for the Angle of the Fourth Proportional
XL VIII.
we
if
(36.) Or,
introduce the area, or the spherical excess^ say 2, of the triangle
DKF, writing thus
XLIX.
we have
these other expressions
L.
2 = D +E +F -
TT,
Lâ-^y = \ir-l'2.;
= l'dr = tp - |(d + e + f).

j3a-'y
LL
/3a-"y
= sinfS
r-^ cos
^2
because
OD = U^=r-i^, by XIL
(37.) Having thus expressed /3a-i y, we require no new appeal to the Figure, in
order to express this of/jer fourth proportional, ya'^/3, which is the negative of its
conjugate, or has an opposite scalar, but an equal vector part (comp. 204, (1.), and
295, (9.)
the geometrical difference being merely this, that because the rotation
round a from
to
must
(3
(38.)
be,
to y has been supposed to be negative, the rotation round a from y

on the contrary, positive.
/3
We may thus
LII.
and we
a,
/3,
write, at once,
ya-i/3
= - K/3a-i y = - sin |2 + r-^d cos |2

new fourth
have, for the angle of this
y, of
which the second and third have merely changed places with each
the formula
(39.)
other,
LIIL
we have
proportional, to the sarne three vectors
But the common

therefore,
vector part of these ieô fourth proportionals
is 6,
by VII
by XI.,
LIV.
the upper sign
= KDL = i(D + E + F) = i7r + ^2.
Z.ya-i|3
= cos|2;
being taken, when the
= + sin|2;
rotation round
a from
the sides 2a',
2b', 2c',
/3
to y
is
negative, as
above supposed.
(40.) It follows
bv
(6.) that
when
2x
of a spherical
triangle
338
DEF, of which the area

triangle, of
[bOOK
III.
2, are bisected by the corners A, n, c of another spherical

a, 6, c, then
is
which the sides* are
LV.
cos a
cos a'
= cos b
cos
b'
= cos
cos
c'
= cos ^S.
from what has been recently shown, that the angle rdk, or
in Fig. 68, represents the semi-area of the bisected triangle def ;
(41.) It follows also,
MDN,
mn
or the arc
whence, by the right-angled triangle lmn, we can infer that the sine oi this semi-area.
is equal to the sine of a side of the bisecting triangle abc, multiplied into the sine
of
the perpendicular, let fall
gle
upon that
side
from the opposite corner of the
latter trian-
we have
because
LVI.
(42.) The same
12 =
sin
sin
mx = sin lm
conclusion can be
sin
= sin ab
sin CQ.
drawn immediately, by quaternions, from the
expression,
LVII.
in
which one
sini2
= e = Sa|8y = S(V/3a.y-i) = TV/3a.SU(V/3a:y);

and the other factor
factor is the sine of ab,
is
the cosine of cp, or the
sine of cq.
Under the same
(43.)
conditions, since
LVIII.
we may
= i^"'( + 0,
a = U( +
&c.,
write also,
LIX.
in which,
.sinA2=SU(+0(^+^)
{S
e)
= SdeK:4:lmn;
by IV. and XIII.,
LX. ..4Zmn = -S(^ +

(44.)
Hence
LXI.
tan 12
^
LXII
and under
(e
= ^'--S(f^ + ?^+^0-
by LIV.,
also,
cos
i2 = r = (r3 r3_rS^^ +
we have
this last form,
rS {eZ + Kd
^^'^
^^=
45+^0
Se) )
4lmn
__5JM___
- tiUeK-SV^d
-SXJSe
'
general expression for the tangent of half the

spherical opening at o, of any triangular pyramid odef, whatever the lengths Td,
Te, T^ of the edges at o may be.
(45.)
As a
verification,
a.
we have
(4/mn)3 = -i(e + ?)3 (? 4 Sy Cd + e^

= 2 (r2 - SeK) (r2 - S?^) (r^ - S^f)
LXIII.
but the elimination of
LXIV.
we ought then
LXV.
if
52
^2 between LIX. LXI.
gives,
(ilmny = (SdeKy +
(r3
r{8sK
+ S^^ + SSe)
)2
to find that
.
= 2 = ^3 = _
{S5eKy = r^
r2
and
- r^ {(SeKy + (8^^)24-
(S^f)^}
in fact this equality results
^SeKSKSSSe,
immediately from the general
formula 294, LTII.

(46.)
Under the same
condition, respecting the equal lengths of d,
e,
?,
we have
also the formula.
* These sides abc, of the

bisecting triangle ABC, have been hitherto supposed for
it will be found that the
for-
simplicity (1.) to be each less than a quadrant, but
mula LV. holds good, without any such
restriction.
CHAP.
CONNEXION WITH SPHERICAL AREA.
I.]
LXVI. ..-Y(d + e)(6 +

whence other
(47.) If
verifications
(?
= 2J (r2 -
^)
SeK
- S?^ -
339
S6t)
= Sbnnd
be derived.
may
denote the area* of the bisecting triangle abc, the general principle
infer that
<T
LXII. enables us to
^^^
LXVII.
tan -
'
'^
- S/3y - Sya - Sa/3
sin c sin
p denote
= sin
sinp = sin b
+m+
cq from c on ab,
the perpendicular
l-t-/
+ cos a + cos 6 + cos c'
1
if
Sa/3y
sin c sin
by (IX.) and (XL),

LXVIII.
e^ + O. + 1 +
so that
A = &c. (comp.
210, (21.)
).
(48.) But,
m + ny =2(1 + 1) (1+m) (1+n)
(abc
4 cos - cos - cos 2
hence the cosine and sine of the new semi-area are,

<r
,,,,,
LXIX.
cos
-i-
cos a + cos 6 4 cose
abc
a
.
sin
sui
LXX.
4cos - cos - cos 2

2
2
sin
sm c
&c.
cos(49.) Returning to the bisected triangle, def, the last formula gives,
x^^rr
LXXI.
1^ =
sm ~S
.
sin a' sin i' sin
cose
= smp sm c
.
sec c
denote the perpendicular from f on the bisecting arc ab, or ft in Fig. 68;
if />'
but cos
^2 =
cos c sec
by LV.
c',
LXXII.
Accordingly, in Fig. 68,
sin
ft
we
tan
hence
|S =
= sin ft
sinp' tan c
tan ab.
have, by spherical trigonometry,
= sin es = sin le
sin
= cos ln sin mn cosec lm = tan mn cot ab.
(50.) The arc mn, which thus represents
in quantity the semiarea of def, has its

be considered as the representative arc (162) of a certain
new quaternion, Q, or of its versor, of which the axis is the radius OD, or \Jd ; and
pole at the point d, and
may
new quaternion may be thus expressed

LXXIII.
Q = dya^ = -
this
its
6^
+ d8a(3y
= r2 +
e^
tensor and versor being, respectively,
LXXIV.
(51.)
lows
is
An
TQ = r = cos|2;
LXXV.
UQ= cos ^S
important transformation of this last versor
+ on.
maybe
sin
12.
obtained as fol-
* The
reader will observe that the more usual symbol 2, for this area of
here employed (36.) to denote the area of the cxscribed
triangle def.
abc
340
III.
[bOOK
so that
LXXVII. ..l^ =
LQ=L 5ya(i = l (56-)i (fÔ^ (?5-i)i
these powers of quaternions, with exponents each
= |,
being interpreted as square

roots (199, (1.) ), or as equivalent to the symbols V(5-J), &c.
(52.) The conjugate (or reciprocal) versor, UQ'i, which has nm for its representative arc,
may
UQ by simply interchanging
be deduced from
the corresponding quaternion
LXXVIII.
/3
and
y, or
and
is,
.
Q'
= KQ = S(3ay = r -
e^
and we have
LXXIX.
UQ' = cos|2-OD.sini2 = (5ri)i
(^-1)^ ( ^'>
the rotation round d, from e to f, being still supposed to be negative.

= j^ ; also
(63.) Let H be any other point upon the sphere, and let oh
the area of the
new spherical
LXXX.
if
the rotation round
cos IS'
dfh
let 2'
be
then the same reasoning shows that
+OD.sin ^2'=
(5:-i)i (?;-i)i (t^^'Os
H be negative and therefore, by multiplication of

LXXVI. and LXXX., we have by LXXV. the analogous
d from f
the two co-axal versors,
formula
triangle,
to
LXXXL
cosKS +
2')
+ oD.sin i(2 +
2')
= (5-i)^
(fÔ^' (S^/'O^ (ô-^^;
where 2-1-2' denotes the area of the spherical quadrilateral, defh.

(54.) It is easy to extend this result to the area of anj/ spherical polygon, or to
the spherical opening (44.) of any pyramid; and we may even conceive an extension of it, as a limit, to the area of any closed curve upon the sphere, considered as
decomposed into an
mon
indefinite
number of indefinitely small triangles, with some comon the spheric surface, and with indefinitely small
vertex, such as the point d,
arcs EF, FH,
of any
of the
curve, for their respective bases
cone, expressed thus as the
or to the spherical opening
Angle of a Quaternion, which
product of indefinitely many factors, each equal

which differs indefinitely little from unity.
is
to the square-root
the limit* of the
of a quaternion,
(55.) To assist the recollection of this result, it may be stated as follows (conip.
180, (3.) for the definition of an arcual sum):
" The Arcual Sum
of the Halves of the successive Sides, of any Spherical Polygon, is equal to an arc of a Great Circle, which has the Initial {or Final) Point of
This Limit
is
we may say
closely analogous to a definite integral, of the ordinary kind
or
a Definite Integral, but one of a new kind, which could

not easily have been introduced without Quaternions. In fact, if we did not employ
the non-commutative property (168) oi quaternion multiplication, the Products here
rather,
that
it is
considered would evidently become each equal to unity
: so that
they would furnish no expressions for spherical or other areas, and in short, it would be useless to
speak of them. On the contrary, when that property or principle of multiplication
introduced, these expressions of product-form are found, as above, to have extremely useful significations in spherical geometry ; and it will be seen that they suggest and embody a remarkable theorem, respecting the resultant of rotations of a sys-
is
tem, round
any number of successive axes,
other req?ects succeeding each other with
all
passing through one fixed point, but
in
any gradual or sudden changes.
CHAP.
AREA OF POLYGON OR CURVE ON SPHERE.
I.]
341
its Pole, and represents the Semi-area of the Figure;'' it being understood that this resultant arc is reversed in direction, when the half-sides are (ar-
the Polygon for
cually)
added
(56.)
in
an opposite order.
As regards the order thus
referred to,
it
may be observed that
in the arcual
which corresponds to the quaternion multiplication in LXXVI., we cona point to move, first, from b to f, through half the arc df which half-side
addition,
ceive
right-hand factor, or square-root, (^^~i)i. We

then conceive the same point to move next from f to a, through half the arc fe,
which answers to the factor placed immediately to the left of the former ; having
def answers
of the triangle
to the
thus moved, on the whole, so far, through the resultant arc
ba
(as a transvec-
And
180, (3.)), or through any equal arc (163), such as ml in Fig. 68.
half the arc ed, or through any arc equal to
finally, we conceive a motion through
tor,
ln
that half, such as the arc
same Figure,
in the
to correspond to the
extreme
left-
the final resultant (or total transvector arc), which

answers to the product of the three square- roots, as arranged in the formula, being
thus represented by the ^naZ arc mn, which has the point d for its positive pole, and
hand factor
in the formula
the half-area, |S, for the angle (51.) of the quaternion (or versor) product which
it
represents.
(57.)
Now
the direction of positive rotation on the sphere has been supposed to

e ; and therefore along the perimeter, in the order dfe,
be that round d, from f to

as seen* from
any point of the surface within the triangle that is, in the order in
which the successive sides df, fe, ed have been taken, before adding (or compound:
And accordingly, in the conjugate (or reciprocal) formula

took the opposite order, def, in proceeding as usual from right-hand
to left-hand factors, whereof the former are supposed to be multiplied bgf the latter;
ing)
their halves.
LXXIX., we
while the result was, as

the area
(58.)
we saw
in (52.), a
new
was simply changed
of the triangle
To give an example
versor, in the expression for which,
to its
own
negative.
of the reduction of the area to zero,
we have only
to
conceive that the three points d, e, f are co-arcual (165), or situated on one great
circle ; or that the three lines d, e, ^ are complanar.
For this case, by the laws*
we have
of complanar quaternions,
LXXXII.
thus cos 12
1,
and 2
the formula,
(de-^)h (^-1)^ (?5-i)i
1,
if
SdtK =
is
imagined
In this and other cases of the
the spectator
sort,
point of the sphere, round which the rotation on the surface
formed
ple,
his
0.
body being outside the sphere.
that the rotation round the
line,
And
to stand
on the
conceived to be per-
is
when we say, for examfrom the line OB to the line OC,

Figures, we mean that such would apsimilarly
or radius, OA,
is negative (or left-handed), as in the recent

pear to be the direction of that rotation, to a person standing thus with hisfeet on
A, and with his hody in the direction of OA prolonged : or else standing on the centre
Compare 174, II. 177; and the Note
(or origin) o, with hia head at the point A.
;
to
page 153.
t Compare the Notes to pages 1-16, 159.

X Compare the Second Chapter of the Second Book.
342
(59.) Again, in (53.) let the point
h be co-arcual
[bOOK
d and
with
f, or let SS^t]
III.
then, because
LXXXir.
(?,- )i (17^-1)^
if
(^^-0'.
S5?j;
0,
the product of four factors LXXXI. reduces itself to the product of three factors
LXXVI. ; the geometrical reason being evidently that in this case the added area
2' vanishes ; so that the quadrilateral defh has only the same area as the triangle
DEF.
But
(60.)
ple
when
the
added area (53.) may even have a negative* effect^ as for examh falls on the old side de. Accordingly, if we write
this
new
point
LXXXIII.
and denote the product
LXXXI,
formation,
LXXXIV.
which shows (comp. (15.)
Qi
= (f^-i)i
(^,,-i)i (^-i)J,
of four square-roots
Qo = idt~
'
)i
Qi
by
^- 1 )^
we shall have the
Q2,
S^ ;
if
trans-
that in this case the angle of the quaternary product Qi
that of the ternary product Qi, or the half-area of the triangle efh (= def dhf),
although the axis of Q2 is transferred from the position of the axis of Qi, by a rotation round the pole of the arc ed, which brings it from oe to od.
is
(GL) From
formula
this
LXXXL
example, it may be considered to be sufficiently evident, how the

be applied and extended, so as to represent (comp. (54.) ) the
may
area of any closed figure on the sphere, with any assumed point D on the surface as
a sort of spherical origin; even when this auxiliary point is not situated on the perimeter, but is either external or internal thereto.
(62 ) A new quaternion Qo, with the same axis od as the quaternion Q of (50.),
but with a double angle, and with a tensor equal to unity, may be formed by simply
squaring the versor \] Q and although this squaring cannot be effected by removing
;
the fractional exponents,^ in the formula

in other ways.
tlie
For example we have,
transformations
LXXVI.,
l)y
yet
it
can
easil}^
be accomplished
LXXIII. LXXIV., and by VII. IX. X,
:J
LXXXV.
Qo = UQ2=r-2(^ya/3)2=-
=-
(ya/3)2
= - (e - ;j)2 = r2 -
S-^
ya(3d dyajS
.
+ 2e^
= r. od, by XII., the trigonometric values LIV. for r and e

fact, because S
enable us to write this last result under the form,
and in
LXXXVI.
(63.)
To show
its
Qo
= - (yafSy = cos S + od
geometrical signification,
let
sin S.
us conceive that
abc and lmn
* In some
investigations respecting areas on a sphere, it may be convenient to
distinguish (comp. 28, 63) between the two symbols def and dfe, and to consider
them as denoting two opposite triangles, of which the ? is zero. But for the pre-
we
sent,
are content to express this distinction,
by means
of the
two conjugate qua-
and (52.).
t Compare the Note to (54.).
ternion products, (51.)
and
The equation dyafi = y aj3d is
in general,
qp
= pq,
if p
||
Yq.
not valid generally
butwehave Aere S= \yaf.
CHAP.
CASE OF SIDES GREATER THAN QUADRANTS.
I.]
343
have the same meanings in the new Fig. 70, as in Fig. 68 and that AiBiMi are
three new points, determined by the three arcual equations (163),
;
LXXXVII. nAC = '^CAi,

MN = o NMi
'^BC='^CBi,
r.
which easily conduct to this fourth equation of

the same kind,
LXXXVir.
nLMi =
-^BiAi.
This new arc lmi represents thus (comp. 167, and

Fig. 43) the product aiy-^.yf3r^=ya-^.(3y-^
while the old arc ml, or
which has
MN
cause
its
equal
ba
(31.), represents a/?-^
whence the arc mmi,
pole at D, and is numerically equal to the whole area 2 of def (bewas seen to be equal (50.) to half that area), represents the product
its
ya-J/3y-i.a^-i, or
and
interpreted,
(ya/3)2, or Qq.
LXXXVI.
The formula
has therefore been
be said to have been proved anew, by these simple geometri-
may
cal considerations.
We
(64.)
see, at
the
same time, how to
LXXXVIII.
interpret the symbol,
Qo=^^^;
a
y
namely as denoting a
versor, of
which the axis
of a certain auxiliary spherical triangle def,

to D, E, F, are bisected (5.)
a from
to
j8
is
each
less
(3
directed
whereof the
by the given points A,
or from, the corner
to,
according as the rotation round
B, c,
of that auxiliary triangle
sides, respectively ojo/JosiYe
negative or positive; and of which the angle represents, or
cally equal to, the area
we have
is
at least
if
we
still
is
numeri-
suppose, as
hitherto for simplicity done (1.), that the sides of the given triangle
abc are
than a quadrant.
298.
The
case
when
instead of being all
the sides of the given triangle are all greater,

than quadrants, may deserve next to be
less,
(although more briefly) considered; the case when they are all
equal to quadrants, being reserved for a short subsequent Article:
and
other cases being easily referred to

from a given line to its opposite,
(1.) Supposing
now
by limits, or by passing
that
I.
or that
these,
II.
m < 0,
< 0,
..a>
6>
< 0,
n
,
c>
2'
we may still retain the recent equations IV. to XI. XIII. and XV. to XXVI., of
297 but we must change the sign of the radical, r, in the equations XII. and XIV.,
;
and
also the signs of the versors
U^, Uc,
U^
in XII., if
we
desire that the sides of
be bisected (as in Figures 67,

68) by the corners of the given triangle abc, of which the sides a, b, c are now each greater than
a quadrant. Thus, r being still the common tensor of 5, e, ^, and therefore
being still
supposed to be itself >0, we must write now, under these new conditions I. or II.
the auxiliary triangle, def,
the
new
equations.
may
still
344
III.
OD = -U^ = -r-^;
IV.
cos a
OE = -U6 = -r->;
r cos a',
cos 6
The equations IV. and VIII.
(2.)
V.
2r cos a cos
h'
of
= r cos 6',
297
still
OF = cos
U^ = -r- ^
c= r cos c.
holding good,
= cos a'2 + cos 6'2 +
cos c
[bOOK
;
we may now
cos c'2 -
III.
write,
1,
we adopt positive values (297), or negative values (298), for the com, n of the sides of the bisecting triangle ; the value of r being still supposed
according as
sines
?,
to be positive.
(3.) It
is
VI.
not
difficult to
the recent formula V.
VII.
prove (comp. 297, LIV., LXIX.), that
= 4:Cos|2,
may
according as
/>0,
or
&c.,
Z<0, &c.
therefore be written unambiguously as follows
2 cos a' cos
h' cos c'
cos
^S = cos a- + cos b"^ -f cos c'^
and the formula 297, LV. continues to hold good.

(4.) In like manner, we may write, without an ambiguous sign (comp. 297, LI.)
the following expression for the fourth proportional fSa'^y to three unit-vectors a, /3,
round the
y, the rotation
first
VIII.
from the second to the third being negative
= sin ^2 +
/3a-iy
where the scalar part changes sign,
when
od. cos
the rotation
is
iS
reversed.
(5.) It is, however, to be observed, that although this /ormttZa VIII. holds good,
not only in the cases of the last article and of the present, but also in that which has
been reserved for the next, namely when Z= 0, &c ; yet because, in the prese7it case
(298) we have the area 2 > tt, the radius od is no longer the (positive) axis US of
the fourth proportional fSa-^y ; nor is ^tt ^2 any longer, as in 297, L., the (posiOn the contrary we have now, for this axis and angle,
tive) angle of that versor.
the expressions
IX.
(6.)
To
Ax.
= DO = -OD;
|3rt-iy
illustrate these results
by a
X.
construction,
Z /3a-iy = 1(2
we may remark
tt).
that
if,
in Fig.
ab be supposed each greater than a quadrant, and if

a new Figure, analogous to 68, the perpendicular cq will
7, the bisecting arcs bc, ca,
we
proceed to form from
also exceed a quadrant,

also
and r
will fall
it
and the
poles
p and
on the arcs lq and
ql,'
will fall between the points
or the angle kdm, or l'dr, or edp, may still

present the versor (3a-^y, yet the con-esponding rotation round the point
a negative
(7.)
at D,
is
now o'
common
but we have now, as in Fig. 71, a new arrangement, in virtue of which this angle is to be
found by halving what remains, when the sum of

tlie supplements of the angles atD and e, in thetri-
\ ^^v/^^
c
jrw 71.
now, in agreement with the recent expression X.,

.
l(3a-iy
= x(p
~//^^
rrJ-^'
^"-^11^^
angle def, is subtracted /rom the angle at r, instead

of our subtracting (as in 297, (35.) ) the latter angle from the former
XL
'
^'''"^V^v.
^n^^ /')
D^;Z~
tlierefore
as regards the quantity of this rotation, or the

magnitude of the angle
again, as in Fig. 68, a base-angle of one
of three isosceles triangles, with p for their
vertex
character.
And
it is
c and q
prolonged: and although the arc km,

be considered, as in 297, (34.), to re-
r)-7r.
sum
it is
CHAP.
MODIFICATIONS OF THE CONSTRUCTION.
I.]
The negative
(8.)
of the conjugate of the formula VIII. gives,
XII.
= -sin^S + OD.cosiS
ya-'|3
and by taking the negative of the square of

following
TT
we
are conducted to the
.^?
= cosS + OD.sinS;
|=-(ya-i/3)2
which had only been proved before (comp. 297, (62.), (64.)) for the case
and in which it is still supposed that the rotation round a from /3 to y is
result
S<
this equation,
XIII.
345
negative.
With
(9.)
we have
the same direction of rotation,
also the conjugate or recipro-
cal formula,
XIV.
-^- = -(/3a-W)2 = cos2-OD.sin2.

a/3y
happened that only one side, as ab, of the given triangle abc, was
while each of the two others was less than a quadrant, or that we had Z > 0,
(10.) If
greater,
it
m > 0,
but n
means
of the foregoing constructions
<
and
if
we wished
opposite to c, or to change
to y'
to represent the fourth proportional to a,
we
=y
j8,
y by
should only have to introduce the point c'

for thus the new triangle abc' would have
each side greater than a quadrant, and so would fall under the case of the present
Article
after employing the construction for which, we should only have to change
;
the resulting versor to
its
negative.
manner, if we had I and m negative, but n positive, we might

again substitute for c its opposite point c', and so fall back on the construction of
Art. 297: and similarly in other cases.
(11.)
And
in like
(12.) In general, if we begin with the equations 297, XII., attributing any arbitrary (but positive) value to the common tensor^ r, of the three co-initial vectors
5, f, Z, of which the versors, or the unit-vectors U^, &c., terminate at the corners of
a given or assumed triangle def, with sides = 2o', 26', 2c', we may then suppose
(comp. Fig. 67) that another triangle abc, with sides denoted by a, 6, c, and with
by I, m, n, is derived from this one, by the condition of

and therefore by the equations (comp. 297, LVIII.),
their cosines denoted
ing
its sides ;
XV.
OB = /3=U(^+^),
.OA=a = U( + 0,
with the relations 297, IV. V. VI., as before
or
oc = y =
bisect-
U(5 + ),
these other equations (comp.
by
297, XIII. XIV.),
XVI.
When
= 2ra
cos
a',
+ 5 = 2rj3 cos 6',
54-= 2ry
cose'.
we have at once (comp. 297,

(13.)
LX.), by merely taking scalars of products of vectors, and without any reference to
ireas (compare however 297, LXIX., and
298, VII.), the equations,
XVII.
this simple construction is

adopted,
4 cos a cos
= -r-2S(^+
= 4 cos b cos c cos a' = 4 cos c cos a' cos b'

+ c) = &c. = 1 + cos 2a' cos 26' 4- cos 2c'
b' cos c'
3) (5
-1-
3r
XVIII
^^
cos a'
^Q^^
cos
6'
cose
cos a'2
cose'
vhich can indeed be otherwise deduced,
by
lometry.
2 Y
-f
cos b'^+ cos
2 cos a' cos 6' cos
the
known
c'2
1
'
c'
formulse of spherical trigo-
346
[booK
III.
We
see, then, that according as the sum of the squares of the cosines of
(14.)
the half-sides, of a given or assumed spherical triangle, def, is greater than unity,
or equal to unity, or less than unity, the sides of the inscribed and bisecting triangle,
ABC, are together
less
than quadrants, or together equal
to
quadrants, or together
greater than quadrants.

(15.) Conversely, ly the sides of a given spherical triangle abc be thus all less,
or all greater than quadrants, a triangle def, but only one* such triangle, can be
exscribed to
it,
so as to
have
new
as above
its sides bisected,
to let fall a perpendicular, such as
CQ
the simplest process being
from c on ab, &c.
in Fig. 68,
and then to draw
perpendicular to these perpendiculars, and therefore coinciding in position with the sought sides de, &c., of def.
(16.) The trigonometrical results of recent sub-articles, especially as regards the
arcs,
through
c, &c.,
triangle, are probably all well known, as certainly some of them

but they are here brought forward only in connexion with quaternion formulce;
and as one of that class, which is not irrelevant to the present subject, and includes
areaf of a spherical
are
the formula 294, LIIL, the following
XIX
(a/3y)2
be mentioned, wherein a,
may
three vectors, but the order of the factors
is
important
(3,
y denote any
= 2a2/32y2 + a^(/37)Hi82(ar)2 + yKi3/-4aySai3S/37.
And if, as in 297, (1.), &c., we suppose that a, (3, y are three ttnit-vecOA, OB, oc, and denote, as in 297, (47.), by (t the area of the triangle abc,
the principle expressed by the recent formula XIII. may be stated under this appa(17.)
tors,
rently different, but essentially equivalent form
^^
XX.
which admits of several
a + /?y + a/3+y
- /3 + y a + /3
y+a
.
= cos
+ a sm
0-
it
as follows (comp. 297,
LXVII.)
(y + ) ^ -2e+2a(l + ? + m + n)
+2e+ 2a(l + /+m + n)
K(a+/3)(i3 + y) (y + a)
-( + /^)(/3 + y)
_l + l + m + nj-ea _
f
>
+ a tan -
cos
+ l + m + n-ea
a tan
ff
cos -
Article,
istence of indefinitely
+ a
we
many
Sin
0-\2
cos
(t
-+ a
sin
a"
a
cos
<t
a sm -
+ a sin
* In the next
cr
verifications.
(18.) "We may, for instance, transform
XXI '"
c, as above.
shall consider a case of indeterminateness, or of the ex-
exscribed triangles
def
namely, when the sides of ab(
are all equal to quadrants.
t This opportunity may be taken
of referring to
an interesting Note, to page
96, 97 of Luby's Trigonometry (Dublin, 1852); in which an elegant construction

connected with the area of a spherical triangle, is acknowledged as having been men
tioned to Dr. Luby,
F.T.C.D.
297, was suggested
Sadleir,
by a
since deceased
and lamented
A construction nearly
to the present writer
by
friend, the
Rev. William Digb;
the same, described in the sub- articles
quaternions, several years ago.
11
CHAP.
CASE OF SIDES EQUAL TO QUADRANTS.
I.]
347
that if a, j8,
(19.) This seems to be a natural place for observing (comp. (16.) ),
any four vectors, the lately cited equation 294, LIIL, and the square of the
y, 5 be
equation 294, XV., with 5 written in
very general and symmetric formula
XXII.
it
instead of p, conduct easily to the followmg
a2/32y222 4.^S/3ySa^)2+
(SyaS|35)U (Sa/3Sy^)2
+ 2^2Sa/3S/3ySya
= 2SyaSa/3S/3^Sy^+ 2Sa/3S/3ySy^Sa5+ 2S/3ySyaSa^S/35
+ ^'^ y2 (Sa^)2 + y2a2 (8,35)2 a2^2 (Sy 5)2
+ a*52(S|3y)2 + /3252(Sya)2 + yH^{â^)K
2a2Si3yS;3^Sy5
+ 2/3SyaSy^Sa5 +
2y2Sa/3Srt^S/35
-J.
we take any
(20.) If then
XXIIL
treating a,
/3,
Z'
spherical quadrilateral abcd, and write
= cos AD = - SUa5,
m' = cosBD
of the arcs bc, ca, ab, as in 297, (1.),
XXIV.
= - SU/35,
as the unit-vectors of the points A, b, c, and
we have
n'
/,
= coscD = &c.,
m, n as the cosines
the equation,
+ /2Z'2 + m'^m'-^ + ri^n"^ + 2lmn' + 2mnT + Inl'm^ 2lmn

= 2mnm'n' + 2nln'l' + 2lml'm'
+ Z24TO2+n2 + Z'2+TO'2+n'2;
which can be confirmed by elementary considerations,* but

an interpretation of the quaternion formula XXII.
is
here given merely as
In squaring the lately cited equation 294, XV., we have used the two
XXII., and 210, XVIII.), in
(21.)
following formulas of transformation (comp. 204,
which o,
XXV.
may be any thi-ee vectors, and which are often found to be useful
= (Sai8)2 - a2/32
XXVI. S (V/3y Vy a) = y^Sa^ - S/3ySy a.
(Va/3)2
/3,
.
The two cases, for which the three sides a, b, c, of the given
are all less, or all greater, than quadrants, having been
aec,
triangle
considered in the two foregoing Articles, with a reduction, in 298,
299-
(10.)
and
(11.), of certain other cases to these, it
only remains to
consider that t/m^d principal case, for which the sides of that given
triangle are all equal to quadrants : or to inquire what is, on our
general principles, the Fourth Proportional to Three Rectatigular
Vectors.
And we shall find, not only that this fourth proportional
is not itself a Vector , but that it does not even contain
any vector
part (292) different from zero although, as being found to be equal

to a Scalar, it is still included
(131, 276) in the general conception
of a Quaternion.
:
(1.) In fact, if
I.
we
Z
is
TO
= 0, n =
0,
or that
II.
= 6 = c = -.
formula equivalent to this last equation of seventeen terms,

connecting the
which join, two by two, the corners of a spherical quadrilateral

given at page 407 of Carnot's Geometrie de Position (Paris, 1803).
cosines of the arcs

r>,
suppose, in 297, (1.), that
= 0,
348
or
III.
.S^y=Sya = Sa/3 = 0,
VIL give,
V. ..^ = 0,
IV.
while
Ta =
III.
[bOOK
T^^Ty = 1,
the formulae 297,
?=0;
0,
but these are the vector parts of the three pairs oi fourth proportionals to the three
rectangular unit-lines^ a, /3, y, taken in all possible orders; and the same evanescence of vector parts must evidently take place,
the three given lines be only at
if
right angles to each other, without being equally long.

that they are unit lines, and
(2.) Continuing, however, for simplicity, to suppose
that the rotation round
r=0, and e=l,
a from
/3
in 297, (3.);
to y is negative, as before, we see that we have now

and that thus the six fourth proportionals reduce
In
themselves to their scalar parts, namely (here) to positive or negative unity.

manner we find, under the supposed conditions, the values
this
VI.
/3a-'y
(3.)
VIL
In
= r^"â = r~'/^ = +l;
Vr.
ya-ii3=a/3-iy
For example (comp. 295) we have, by the laws (182) of

.
ij-'^k=jk-H=^ki-^j = +
Vir.
the two fourth proportionals, ij-^k and
fact,
(4.)
To connect
we may
t,
j, A, the values,
lj-H = ik-\j=ji-^k = ~l.
kj-î,
two ternary products, ijk and kji, and therefore to

cluded in the Fundamental Formula A (183).
ticles,
= /3y-'a = -l.
are respectively equal to the
and
1,
by the laws
this important result with the constructions of the
observe that
when we
two
last
in-
Ar-
on the general plan of 298, (15.),
seek,
to
exscribe a spherical triangle, def, to a given tri-quadrantal (or tri-rectangular)

triangle, ABC, as for instance to the triangle ijk (or jik) of 181, in such a manner
that the sides of the
problem
new
triangle shall be bisected
sumed, in the interior of the

be taken, with respect to the three sides
TV),
PF
PE,
by the corners
of the old, the
many solutions. Any point p may be asgiven triangle abc and then, if its reflexions D, E, f
found to admit of indefinitely
is
are perpendicularly
a, b, c,
so that (comp. Fig. 72) the arcs
by those
bisected
three sides, the three other arcs ef, fd, de will be

because
bisected by the points a, b, c, as required
:
the arcs ae,
af have each
the same length as ap,
and the angles subtended at A by pe and pf are together equal to two right angles, &c.
of the auxiliary points, D, e,
(5.) The positions
F,
are therefore, in the present case, indeterminate,
or variable
but the sum of the angles at those three
and equal to four right angles ;

the six isosceles triangles on pd, pe, pf as bases, that sum of the
three angles d, e, f is equal to the sum of the angles subtended by the sides of the
given triangle abc, at the assumed interior point p. The spherical excess of the
points
is
constant,
because,
by
triangle
def
is
therefore equal to two right angles,
and
its
area
2 = tt
as
may
otherwise seen from the same Figure 72, and might have been inferred from the
mula 297, LV., or LVI.

(6.) The radius OD,
in the formula 297,
XLVIL,
be
for-
for the fourth proportional
becomes therefore, in the present case, indeterminate ; but because the angle
- r), in the same equation, vatiishes, the formula becomes simply
l'dr, or ^ (tt
/3a~^y,
CHAP.
I.J
OTHER VIEW OF A FOURTH PROPORTIONAL.
= 1, as in the recent equations VI.

/3a-iy
class here considered.
and similarly
in
349
other examples, of the
(7.) The conclusion, that the Fourth Proportional to Three Rectangular Lines
a Scalar, may in several other ways be deduced, from the principles of the present
Book.
For example, with the recent suppositions, we may write,
is
VIII. . .)3a-i = -y,

VIII'. .. ya-J= + /3,
y/3-i
= -a,
a)8-'=
+ y,
oy->=-/3;
= + a;
/3y-^
the three fourth proportionals VI. are therefore equal, respectively, to
to
a^,
(8.)
Or (comp.
(3.)
we may
XXI.*),
IX.
in
y^, -
and consequently to + 1 while the corresponding expressions VI'. are equal

- 1.
/3, + y2, + a2, and therefore to
/32,
which the factor a-^
is
|3a-iy
of rectangularity
= a-2.|3ay,
a-2=
if
a,
always a scalar, whatever vector a
vector part of the ternary product

ditions
write generally the transformation (comp. 282,
ây vanishes, by
294,
III.,
may
when
be
while the
the recent con-
III. are satisfied.
product jSay, and this fourth proportional /3a~'y,

can never reduce themselves to scalars, unless the three vectors a, (3,
y (supposed to
be all actual (Art. 1) ) are perpendicular each to each.
(9.) Conversely, this ternary
On
an equivalent Interpretation of the Fourth

Three Diplanar Vectors, deduced from the
Proportional
the
Second Book.
Principles of
Section
8.
to
300. In the foregoing Section, we naturally employed the results

of preceding Sections of the present Book, to assist ourselves in attaching a definite signification to the Fourth Proportional
(297)
to
Three Diplanar Vectors ; and thus, in order to
interpret the
sym-
we
availed ourselves of the interpretations previously obtained, in this Third Book, of a'^ as a line, and of a^, a/37 as quaternions.
But it may be interesting, and not uninstructive, to inquire
hoi /3a'^7,
how
the equivalent symbol^

I.
or
(^: a). 7,
7,
with 7 not
\\\
a, yS,
might have been interpreted, on the principles of the Second Booky without at first assuming as known, or even seeking to discover,
in-
any
terpretation of the three lately

II.
It will
mentioned symbols,
a~",
a^,
al3'y.
be found that the inquiry conducts to an expression of the
form,
* The formula
here referred to should have been printed as
Ra =
= o"*.
350
III.
where
S is the
same
vector^
sub-articles to 297; while
[bOOK
(/3:a).7=^ + ew;
and e is the same scalar,
III.
is
employed
as a
as in the recent
temporary symbol,
to
denote a certain Fourth Proportional to Three Rectangular Unit

Lines, namely, to the three lines oq, ol', and op in Fig. 68; so
that, with reference to the construction represented by that Figure,
we should be
equation
by the
led,
principles of the Second Book, to write the
IV.
. .
And when we
oa).oc=od.cos12 +
(ob:
(ol': oq). op. sin -^2.
proceed to consider what signification should be at-
tached, on the principles of the same Second Book, to that particular

fourth proportional, which is here the coefficient of sin ^-2, and has
been provisionally denoted by
we
w,
find that although
it
may be
regarded as being in one sense a Line, or at least homogeneous with a

line, yet it must not he equated to any Vector: being rdiih^v analogous^
in Geometry, to the Scalar Unit
of Algebra, so that
and conveniently denoted by the usual symbol
1,
it
or
may be
+
1,
But when we thus write u=\,
to Positive Unity.
naturally
or be equated
the last term
of the formula III. or IV., of the present Article, becomes simply

, or sin ^2 ; and while this term (or part) of the result comes to be
considered as a species of Geometrical Scalar, the complete Expression for the General Fourth Proportional to Three Diplanar Vectors
takes the
297,
Form
XL VII.,
of a Geometrical Quaternion: and thus th^ formula
or 298, VIII.,
is
reproduced, at least if
we
substitute
for the present, (/3: a).7 for /Sa''^, to avoid the

necessity of
interpreting here the recent symbols II.
in
it,
(1.)
The
construction of Fig. 68 being retained, but no principles peculiar to the
Third Book being employed, we

as before,
V.
write,
with the same significations of
OB OA = OR OQ = cos c + (ol' oq) sin c

VI.
oc = OQ cos ja + OP sin /?
:
(2.)
may
Admitting then, as
p, &c.,
is
c,
natural, for the purposes of the sought interpretation,
that distributive property which has hQ&a. proved (212) to hold good for the multiplication of quaternions (siS \i Aoqs for multiplication in algebra); and writing for
abridgment,
VII.
we have
u = (ol'
the quadrinomial expression
VIII.
oq) OP;
.
(oB oa). oc = ol'.
+ OP
in
cos c sin/)
sin c C0SJ9
+M
+ OQ
sin c sin p
cos ccosjo
which it may be observed that the sum of the squares of thefour
coefficients ofth
CHAP.
SCALAR UNIT IN SPACE.
I.]
three rectangular unit-vectors, OQ,
op,
oi.',
351
and of their fourth
proportional, u,
is
equal to unity.
But the
(3.)
which may be regarded as a
coefficient of this fourth proportional,
species oi fourth unit, is
IX.
we must
sin c sin
p = sin mn =
sin
12 =
therefore expect to tind that the three other coefficients in VIII.,
when
di-
^S, or by r, give quotients which are the cosines of the arcual distances
or that a point
of some point x upon the unit-sphere, from the three points l', q, p
X can be assigned, for which
vided
by
cos
X.
sin c
= r cos l'x
cosp
(4.) Accordingly
substitute
XI.
d
.
for
01.'.
it is
and
cos c
cosp
= r cos Qx
we have
therefore that
3inccos/)
p=
cos c sin
found that these three last equations are
r cos px.
satisfied,
when we
the transformation,
+ OQ.cosccos/j + OP.coscsinp = OD cosiS = ^,

.
whence follow the equations IV. and III. and it only remains to study and interpret
the fourth unit, , which enters as a factor into the remaining part of the quadrino;
mial expression VIII., without employing
Book
tt',
any
principles except those of the
and therefore without using the Interpretations 278, 284, of
(3a,
301. In general, when two sets of three vectors,

7', are connected by the relation,
Second
&c
a, /3, 7,
and
)3',
I.
..--=
a
7'
it is
or
1,
II...
=
a ^,
-7y
a'
natural to write this other equation,

TTT
III.
^ = ^'
7 7
ft
'
ft
these two fourth proportionals (297), to a, (3, 7, and

are equal to each other: whatever the /w^Z signification of
each of these two last symbols III., supposed for the moment to be
and to say that

to
a', ^', 7',
may be afterwards found to be. In short,

make
it a condition of the sought
to
may propose
Interpretation,
the principles of the Second Book, of the phrase,
not yet fully known,
we
on
^''Fourth Proportional to Three Vectors y'
and of either of the two equivalent Symbols 300,

Equation III. shall /oZZow from I. or II.; just as,
I.,
that the recent
at the
commence-
ment
of that Second Book, and before concluding (112) that the

general Geometric Quotient jSia of any two lines in space is a Quaternion,
we made it a conditioji (103) of the interpretation of such a quotient,
that the equation {^fi: a).a ^ should be satisfied.

302. There are however two tests (comp. 287), to which the recent equation III.
must be submittedj before
its final
adoption; in
352
order that
we may be
sure of
III.
[bOOK
with Xhe previous
its consistency, 1st,
Fourth Proportional to Three Complanar

their common plane; and Ilnd, with the gene-
interpretation (226) of a
Line in
Vectors, as a
ral principle of all mathematical language (105), that things equal to

And it
the same thing, are to be considered as equal to each other.
is
found, on
trial,
that hoth these tests are
home :
so that they
no objection to our adopting the equation 301, III., as true hy

tion^ whenever the preceding equation II., or I., is satisfied.
(1.) It
d, as in
may happen
226
that the
namely, when
a,
first
member
of that equation III.
In
y are complanar.
(3,
is
this case,
form
defini-
equal to SiUne
we have by
II.
the equation,
7y
are also complanar (among themselves), and the line d is their

/3', y'
and the equation III. is satisfied, both members being
fourth proportional likewise
so that a,
symbols for one common line, ^, which is in general situated in the intersection of
the two planesy afSy and a'ji'y'
although those planes may happen to coincide,
;
without disturbing the truth of the equation.

(2.) Again, for the more general case oi diplanarity of a, /3, y, we
ceive that the equation* II. co-exists with this other of the same form,
V.
X = 21
a y
if
which gives
the definition 301 be adopted.
principles of equality,
we ought
when
give,
con-
2 y =Cr",
a
a
If then that definition be consistent with general

to find,
= ^'/';
^7'
a
a
III.
by
tween two fourth proportionals holds good

VII.
VI.
may
and VI., that
this third equation be-
or that
VIII.
2, 1,
a y
= ^,
a
And accordingly, those two equations

the equations II. and V. are satisfied.
principles of the Second Book, respecting quaternions considered
by the general
as quotients of vectors, the transformation,
= Î.Z = ^Jl = ^',ae

^%
ay y' ay" a'
y
required.
a'
then permitted to interpret the equation 301, III., on

the principles of the Second Book, as being simply a transformation
or I.;
(as it is in algebra) of the immediately preceding equation II.,
303. It
is
and therefore to write, generally,

l...q^{ = q'^\
if
II. ..
5(7:7') = ^';
* In this and other cases of

reference, the numeral cited
is
always supposed
to be
the one which (with the same number) has last occurred before, although perhaps
it may have been in connexion with a shortly preceding Article.
Compare 217, (1.).
CHAP.
where
which
FOUIITH PROPORTIONAL RESUMED.
I.]
7, 7'
are any two vectors^ and
quaternions^
we have (by
9,
Now,
satisfy this last condition.
353
are any two quaternions^
and v be any two
if v
right
193, comp. 283) the equation,
III.
Iv.lv'
= v:v'
==
vv'^
or
IV.
v~^
(Iv
Iv^)
v'-^
whence
V.
v-Mv = v'-^
It;',
by the principle which has just been enunciated. It follows, then,

that '''if a right Line (Iv) le multiplied by the Reciprocal {v^) of the
Eight Quaternion (v), of which it is the Index, the Product {v^lv) is
independent of the Lengthy and of the Direction, of the Line thus operated on f' or, in other words, that this Product has one common Value,
for
Lines (a) in Space: which common or constant
all possible
may be regarded as a kind of neiv Geometrical Unit, and is equal

what we have lately denoted, in 300, III., and VII., by the tem-
value
to
porary symbol w; because, in the last cited formula, the line op is

the index of the right quotient oa: ol'.
Retaining, then, for the
a in space, considered
this
we
line
have, for every
moment,
symbol, u,
as the index of a right quaternion, v, the four equations:
Yl...v-â = u;
Yll.
IX.
in
which
a=vu;
=w a
V"'
understood that a =
it is
Iv,
VIII.
a: w;
and the three
last are
here re-
garded
merely transformations of the first, which is deduced
and interpreted as above. And hence it is easy to infer, that for
any given system of three rectangular lines a, p, 7, we have the general
as being
expression
X.
(/8
where the scalar
a)
7 = xu,
co-efficient, x,
the equation,
XI. .a; = (T/3:Ta).T7,
.
This coefiicient x
if
is
a _L
/3,
-L 7,
of the new unit, u,
according as
XII.
is
j_
a;
determined by
U7 = + Ax. (a: ^).
therefore always equal, in magnitude (or absolute
quantity), to i\\Q fourth proportional to the lengths of the three

lines 0^87
rotation
but
round
it is positively
the third line 7,
given
or negatively taken, according as the
from
the second line
/3,
to the first line
itseM positive or negative: or in other words, according as the

rotation round the first line, from the second to the third, is on the
a, is
contrary negative ov positive (compare 294, (3.)

(1.)
).
In illustration of the constancy of that fourth proportional which has been,

denoted by m, while the system of the three rectangular unit-lines
for the present,
2 z
354
from which
it is
conceived to be derived
is
in
[bOOK
III.
any manner turned about, we may ob-
serve that the three equations, or proportions,
XIII.
?f
y:a=a:-y;
y=/3: a;
/3:
-y = y:/3,
conduct immediately to this fourth equation of the same kind,
XIV.
M:a =
= (y:/3).a;
or*
y:/3,
we admit that this new quantity, or symbol, m, is to be operated on at all^ or combined with other symbols, according to the general rules of vectors and quaternions.
if
to change the three letters a, j3, y, by a cyclical per(2.) It is, then, permitted
mutation, to the three other letters /3, y, a (considered again as representing unitwithout altering the value of the fourth proportional, m; or in other words, it
lines'),
is allowed to make the system of the three rectangular lines revolve, through the third
part of four right angles, round the interior and co-initial diagonal of the unit-cube,
of which they are three co-initial edges.
no such change of value will take place, if

(3.) And it is still more evident, that
we merely cause the system of the two first lines to revolve, through any angle, in
since thus we shall merely substitute,
its own plane, round the third line as an axis
;
But by combining these two last

another factor equal thereto.
modes of rotation, we can represent any rotation whatever, round an origin supposed
for the factor
/3
a,
to be fixed.
And
(4.)
/3'
a'
y', to
as regards the scalar ratio of
any
cient to suggest that, without
pose
it to
be so
any real change

prepared, that we shall have
XY...a' = a;
X being some
any one fourth proportional, such as

a y, or u, it is suffi/3
other, of the kind here considered, such as
scalar coefficient,
/3'
in the former,
we
are allowed to sup-
Y = xy;
= /3;
and representing the
i-atio
required.
304. In the more general case, when the three given lines are
not rectangular, nor unit-lines, we may on similar principles determine their fourth proportional, without referring to Fig. 68, as
Without any real loss of generality, we may suppose that
follows.
the planes of
a, /3
and
a,
7 are perpendicular to each other; since
this
comes merely
y3: a,
another quotient equal thereto.
I.
..
Ax.
(/8:
a)
J-
to substituting, if necessary, for the quotient
Ax.
(7: a),
where /3' and 7' are parallel

to it, and to each other; so
to a,
that,
Having thus
= /3' + ^'', 7 = 7' +
/3
II.
let
but
j3"
by 203,
and
I.
7",
are perpendicular
and II., we shall have
7''
the expressions,
III.
.^^=S^.a,
* In
equations of this form, the parentheses
clearness they are here retained.
y-S^.a,
may
be omitted, though for greate
CHAP.
SPHERICAL PARALLELOGRAM.
1.]
and
IV.
We may
./3"=V-.a,
a
r/^
=Y^
355
a.
then deduce, by the distributive principle (300,
(2.)
),
the
transformations,
a
ad
where
VI.
The
.. a
f^/
/3
/3
fv
= ^S-+7"^-=7S- + )S''SA and VII.
latter part, xu^ is
Q'l
ojm
= ^-^ 7^
a
what we have called (300) the (geometrically)
scalar part, of the sought fourth proportional ; while the former part
B may (still) be called its vector part: and we see that this
part is
represented by a line^ which is at once in the two plajies^ of ^, 7'', and

of 7, ^'^ ; or in two planes which may be generally constructed as follows, without
now assuming
lar, as in I.
Let
7'
that the planes a/3 and 07 are rectangube the projection of the line 7 on the plane of
and operate on this projection by the quotient /3:a as a multithe plane which is drawn through the line )3 a 7' so obtained,
plier
a, /?,
at right angles to the plane ay3, is 07ie locus for the
and the plane through

is
which
7,
And
another locus for that line.
or vector part
x%
it is
is
sought line d
perpendicular to the plane 77',
as regards the length of this
line,
and the magnitude
d,
(or quantity) of the scalar part
easy to prove that
YIII.
T^ = ^coss,
and
IX.
x = tsms,
where
X..
.?
= T/3:Ta.T7,
and
XI.
sin 5
= sin c
denote the angle between the two given lines
if c
sin
a, 13,
j9,
and
the
inclination of the third given line 7 to their plane: the sign of the
scalar coefficient, x, being positive or negative, according as the rotation
round a from
(1.)
dition
13
to
is
negative or positive.
Comparing the recent construction with Fig.
I. is satisfied,
the four unit-lines
four radii oc, OQ, or, od,
(2.) It
may
U/3,
68,
Ud
we
see that
the con-
may
be called
CQRD on
the sphere.
be remarked that the area of this quadrilateral
of the triangle
when
take the directions of the
which terminate at the four comers of what
a tri-rectangular quadrilateral
//the area 2
Uy, Ua,
def
which may be
is
inferred, either
exactly equal to
from the circum-
356
stance that
MDN
its
spherical excess (over four right angles)
is
[bOOK
III.
constructed by the angle
dbr and eas
being together equal to the triangle abf,

so that the area of desk is S, and therefore that of cqrd is ^S, as before.
(3.) The two sides CQ, qr of this quadrilateral, which are remote from the obtuse
or from the triangles
angle at d, being still called p and c, and the side cd which is opposite to c being
also let
still denoted by c', let the side dr which is opposite to p be now called p'
;
be denoted by d and d' ; and let s denote the spherical excess

or the area of the quadrilateral.
"We shall then have the relations,
^tt),
the diagonals CR,
(CDR
QD
d = cosp cos
p tan
tan c'= cos

cos
!cos
of which
s=
cosp
= cosp cos c'

= cos c tan p
seep' = cos c sec c' = cos d sec cT
some have virtually occurred
cos d'
tan p'
and
before,
all are easily
proved by right-an-
gled triangles, arcs being when necessary prolonged.

(4.) If we take now two points, a and b, on the side qr, which satisfy the arcual
equation (comp. 297, XL., and Fig. 68),
XIII.
and
is
if
we then
n AB
join AC, and let fall on this
= OQE
new
easy to prove that the projection b'd' of the side
arc,
and that the angle dbb'
XIV.
is
right
b'd'
and the new quadrilateral bb'd'd

(5.)
Hence
the point
d may
so that
= o AC
arc the perpendiculars bb', dd'
bd on
we have
XV.
the arc
the two
dbb'
AC
new
= 7r
is
it
equal to that
equations,
also tri-rectangular.
is
he derived from the three points abc, by any two of
the four following conditions: 1st, the equality XIII. of the arcs ab,
qr
Ilnd, the
Ilird, the tri-rectangular characcorresponding equality XIV. of the arcs Ac, b'd'
ter of the quadrilateral CQRD ; IVth, the corresponding character of bb'd'd.
;
(6.) In other words, this derived point D is the common intersection of the four
perpendiculars^ to the four arcs ab, ac, cq, Bb', erected at the four points R, d', C, b ;
CQ, bb' being still the perpendiculars from c and b, on ab and AC and R and d'
being deduced from q and b', by equal arcs, as above.
;
305. These consequences of the construction employed in 297,

mentioned merely in connexion with that theory of
&c., are here
fourth proportionals to vectors, which they have thus served to illustrate; but they are perhaps numerous and interesting enough, to
Spherical Parallelogram^''^* for
justify us in suggesting the namey
'*'
the quadrilateral cabd, or bacd, in Fig. 68 (or 67) ; and in proposing

to say that d is the Fourth Point, which completes such &parcdlelogram,
when the
three points c, a, b, or b, a, c, are given
asjirst^ second,
and
that the analogy

*
By
the
third.
to the
It
plane
is
sphere,
carefully observed,
here thus far impet'fect, that in the
same analogy, the quadrilateral cqrd,
Spherical Rectangle.
upon the
must however be
in Fig. 68,
may
be called
CHAP.
SEKIES OF SPHERICAL PARALLELOGRAMS.
I.J
357
general case, when the three given points are not co-arcual, but on the
contrary are corners of a, spherical triangle abc, then if we take c, d, b,
or B, D, o, for the th-ee first points of a new spherical parallelogram, of

the kind here considered, the new fourth point, say Ai, will not coincide with the old second point A; although it will very nearly do so,
if the sides of the triangle abc be small: the deviation aa,
being in
fact
found to be small of the third
order, if those sides of the given
triangle be supposed to be small of the first order; and being always

directed towards the foot of the perpendicular, let fall from a on bc.
To
(1.)
investigate the law of this deviation, let
/3, y be still any two given

making with each other an angle equal to a, of which the cop or op be any third vector. Then, if we write,
unit-vectors, ob, oc,

sine is
and
let
I.
..pi
= ^(p) = iNp.f-7 +
OQ=Up,
OQi
= Upi,
^/3j,
the
new
or derived vector, <pp or pi, or oPi, will be the
common
vector part of the
two fourth proportionals, to p, (3, y, and to p, y, (3, multiplied by the square

of the
andsQCQi will be what we have lately called a spherical parallelogram.
length ofp
;
We
have the transformation (compare 297, (2.)),
shall also
II...pi = 0p
/3S^+yS|-pS|;
and the distributive symbol of operation

III.
..0P
IV.
but
111/3, y,
.
(2.) This being understood,
V.
so that
p',
- Ip,
0p =
..p = p' + p";
will
<p
be such that
0V = p, ^ PJllAy;
and
if
Ax. (y
11
/3).
let
0p'
= p'i;
p'll|/3,y;
p"|iAx.(y:^);
or op', is the projection of p on the plane of /3y
and p", or op", is the

Then we shall have
part (or component) of p, which is perpendicular to that plane.
an indefinite series of derived vectors,
p\, pa, ps,
ceeding each other alternately, as follows
yj^
= ^p = p'l -
fpi
'
or rather two such series, suc-
Ip"
Pi = ^2p
\p3 = ^ V = p'l - ^'P"
P4
= p' + pp"
= 0V = P' + ^*P'
&C.
the two series of derived points. Pi, P2, P3, Pi,

.
being thus ranged, alternately,
on the two perpendiculars, pp' and Pip'i, which are let fall ^rom the points p and Pi,
on the given plane Boc ; and the intervals, PP2, P1P3, P2P4, forming a geometri.
cal progression, in which each
is
equal to the one before
it,
multiplied by the con-
by the negative of the
cosine o{ the given angle boc.

and tt, and also
supposed to be distinct from
in general from the intermediate value Jrr, we shall have the two
limiting values,
stant factor
/,
or
(3.) If then this angle be
VII.
still
p2
= p',
or in words, the derived points Po, P4,
other derived points, Pi, P3,
p2n+l
.
= p'i,
if
= Qo;
of even orders, tend to the point
of odd orders, tend to the other point
p'l,
p',
and the
as limiting
358
[bOOK
III.
positions: these two limit points being the feet of the two (rectilinear) perpendiculars, let fall (as above) from p and p' on the plane bog.
But even the^rs^ deviation ppg, is small of the third order, if the length Tp
op be considered as neither large nor small, and if the sides of the spheri-
(4.)
of the line
BQC be small
cal triangle
For
of the first order.
we have by
VI. the following ex-
pressions for that deviation,
VIII.
pP2 = |02-p
= ('*-l)p"=-sina2.sin/>a.T|O.U/o";
where /> denotes the inclination of the line p to the plane (3y or the arcual perpendicular from the point q on the side bc, or a, of the triangle. The statements lately
;
made (305)
are therefore proved to have been correct.
And
(5.)
if
we now resume and extend
the spherical construction, and conceive
deduced from baiC, as A\ was from bdc, or d from bac ; while A2 may
be supposed to be deduced by the same rule from bdiC, and D2 from BA2C, &c.,
through an indefinite series of spherical parallelograms, in which the fourth point
that Di
of
is
any one
is
treated as the second point of the next, while the^rs^ and third points
:
we see that the points Ai, A2, . . are all situated on the arcual
remain constant
perpendicular
let fall
from A on bc
and that
all
We
see also that the ultimate positions, Ac
of those two perpendiculars
and
manner the
in like
situated on that other arcual perpendicular,
are
which
it
would perhaps be
by any other method than that of the Quaternions,

simple as those which have been employed above.
cult to prove,
lations so
(6.) It
may be remarked that
from d on bc.
Doo, coincide precisely with the feet
a remarkable theorem, which
points Di, D2,
is let fall
diflS-
at least with calcu-
the construction of Fig. 68 might have been other-
wise suggested (comp. 223, IV.), by the principles of the Second Book, if we had
sought to assign the /oMri^/t/?ro/>or<ionaZ (297) to three right quaternions; for example, to three right versors, v, v', v", whereof the unit lines a, /3, y should be supFor the result would be in general a quaternion v'v'^v", with
posed to be the axes.
e for its scalar part, and with d for the itidex of its right part : e and d denoting
the same scalar, and the
same
306. Quaternions
struction^
which
shall
mination of the two
may
also
be employed to furnish a new con-
complete (comp. 305, (5.)) the ^rajo/c<?Z deter-
series of
I.
when the
vector, as in the sub-articles to 297.
derived points^
D, Ai,
A2, Do, &C.,
c are given upon the unit-sphere; and

thus shall render visible (so to speak), with the help of anew Figure,
the tendencies of those derived points to approach, alternately and
three points a, b,
d'
indefinitely, to the feet, say
let fall
from the two
and
a',
of the two arcual perpendiculars
opposite corners^
d and
a, of
the/r5^ spherical
parallelogram, bacd, on its given diagonal bc which diagonal

have seen) is common to all the successive parallelograms.
;
The given triangle ABC being supposed for simplicity to have its sides ab^
than quadrants, as in 297, so that their cosines Imn are positive, let a', b', c'
(1.)
less
(as wej
CHAP.
CONSTRUCTION OF THE SERIES.
I.]
the feet of the perpendiculars
also let
II.
an and bc
SO that the arcs
on these three sides from the points A,
let frJl
and n be two auxiliary
points, determined
BM = n
ri
359
AM = rk MN
'^
MC,
by the
bisect each other in m.
Let
b,
equations,
;
fall
from n a perpendicular
nd' on BC, so that

III.
-^
bd'= n
a'c
c" be two other auxiliary points, on the sides h and

prolonged, which satisfy these two other equations,
and
let b",
IV.
b" and
points^
b'b"
r\
Then the perpendiculars
(2.)
r^
r\
AC,
c'c"
to these last sides,
=n
bacd
and
or on those sides
AB.
CA and ab,
c", will intersect each other in the point D,
c,
erected at these last
which completes
(^BOb) the
foot of the perpendicular from this point d,

on the third side bc of the given triangle, will coincide (comp. 305, (2.) ) with the
foot d' of the perpendicular on the same side from N so that this last perpendicular
kd' is one locus of the point D.
spherical parallelogram
i\\Q
To obtain another
(3.)
locus for that point, adapted to our present purpose, let
B denote now* that new point

other
in
which the two diagonals, ad and bc,

) we have the expression,
V. . OD u(rw)3 +
write (comp. 297, (25.), and (30.) ),
.
we may
VI.
intersect each
then because (comp. 297, (2.)
OE
= u (mj3 + ny),
whence VII.
sin
ny-?a),
be
sin
ec
= m = cos ba'
:
cos a'c
ad
thus dividing the arc bc into segments, of which the sines are proportional to the cosines of the adjacent sides of the given triangle, or to the cosines
of their projections ba' and a'c on bc
so that the greater segment is adjacent to the
the diagonal
lesser side,
(4.)
and the middle point
The
intersection
is
of
bc (!)
therefore a
lies
known
between the points a' and E.

and the great circle through
point,
A and E is a second known locus for

D which point may therefore be
;
found, as the intersection of the arc

AE prolonged, with the perpendicular
nd' from N (1.).

And because e lies
(3.) beyond the middle point m of bc,
with respect to the foot a' of the perpendicular on
is
falls
BC
bc from
a,
but (as
it
easy to prove) not so far beyond

as the point d', or in other words
between
is,
M and d'
(when the arc

a
as above supposed, less than
quadrant), the prolonged arc ae cuts

nd' between n and d'; or in other
words, the perpendicular distance of
the sought fourth point d,
given diagonal
gram,
is less
bc of
from
the
the parallelo-
than the distance of the
Fig. 73.
given second point A, from the same given

diagonal. (Compare the annexed Fig.
73.)
It will
be observed that m, n, e have not here the same

significations as
360
(5.) Proceeding
III.
[boOK
next (305) to derive a new point Ai from b, d, c, as d has been

we see that we have only to determine a new* auxiliary point
derived from b, a, c,
F,
by the equation,
VIII.
and then
draw df, and prolong
to
'ÊM=n MF;
it till it
meets aa' in the required point Ai, which

bdoAi, with BC (as before) for a given
will thus complete the second parallelogram^
diagonal.
to complete (comp. 305, (5.) ), the third parallelogram,
(6.) In like manner,
BAiCDi, with the same given diagonal bc, we have only to draw the arc AiE, and
prolong it till it cuts nd' in Di after which we should find the point A2 of a fourth
;
by drawing
successive parallelogram BD1CA2,
The constant and
(7.)
limit-point d',
limit-point A',
diF,
and so on
for ever,
indefinite tendency^ of the derived points D, Di,
and of the other (or alternate') derived points Ai, A2,

becomes therefore evident from this new construction
limiting) results of which,
we may
to the
to the other
the final (or

express by these two equations (comp. again
;
305,(5.)),
IX. ..Dco=D';
Aao=A.
smallness (305) of tliQfirat deviation AAi, when the sides of the

(8.) But the
given triangle abc are small, becomes at the same time evident, by means of the
same construction, with the help of the formula VII. which shows that the interval^
;
mf
small of the third order, when the sides of the

given triangle are supposed to be small of the^r** order: agreeing thus with the
equation 305, VIII.
EM, or the equal interval
The theory
(5.), is
some interesting apconnexion with spherical conies ; on which however we cannot enter here, beyond the mere enunciation of a Theorem, % of which (comp. 271)
(9.)
of such spherical parallelograms admits of
plications, especially in
the proof by quaternions
Fig. 68
is
easy
and that the present
letters c'
and c" correspond to q and r in that Fi-
gure.
* This new
and the
point,
intersection of the perpendiculars of the given trian-
new Figure 73, as the points denoted by the

former Figure 68
although the four points a, b, c," d
are conceived to bear to each other the same relations in the two Figures, and indeed
not the same in the
gle, are evidently
same
letters,
f and
in Fig. 67 also
p, in the
bacd
what we have proposed
being, in that Figure also,
to call a
Compare the Note to (3.).

spherical parallelogram.
t The formula VII. gives easily the relation,
VII'.
tan
EM = tan ma'
tan j
hence the interval

generally, if
interval
em
X This
<
em
is
small of the third order, in the case (8.) here supposed
-, as in (1.), while 6
is less
and
c are unequal, the formula
than ma', or than d'm, so that e
Theorem was communicated
to the
falls
Royal
as a consequence of the principles of Quaternions.
(Vol. III., page 109).
between
Irish
m and d',
Academy
and
shows that this

as in (4.
in June, 1846J
See the Proceedings of that dat
CHAP.
toe
INTERPRETATION OF A PRODUCT*
TttIRt>
I.]
"
//"kumn he any spherical quadrilateral,
complete the spherical parallelograms,
and
X.
RUNG,
then these
KILA,
LIMB,
any point on the sphere ; ifalso
NIKD,
E and F of the diagonals km and ln of the quadrilateral i

two poles are the foci* of a spherical conic, inscribed in the derived quadri-
and determine
lateral
361
the poles
ABCD, or touching
its
four sides."
(10.) Hence, in a notationf elsewhere proposed,

ditions of construction, the
XI.
EF
formula
ABCD
(. .)
we
shall have,
under these con-
or
XI'.
EF
(. .)
BCDA
&C.
(11.) Before closing this Article and Section, it seems not irrelevant to remark,
that the projection y' of the unit-vector y, on the plane of a and /3, is given by the
formula,
^_.
XII.
and that therefore the point
abc
diculars of the triangle
XIII.
(12.) It
may
= a sin
,
.
which
p, in
a cos B
+ /3
smc
=a
A
-
(see again Fig. 73) the three arcual perpen-
on the vector,
intersect, is
.
sin 5 cos
tan
A+
jS
tan
b + y tan c.
be added, as regards the construction in 305,
(2.), that the right
lines J
XIV.
however
and
in ttvo points s
t,
PPi,
OS
and which thus become two fixed

in space
is
Section
P2P3,
P3P4,
may
XV...
P1P2,
be continued, intersect the given plane BOC, alternately

of which the vectors are,
far their series
= eii^',
1+Z
'
ox =
^î;
l+l
poirits in the plane,
when
the position of the point
given, or assumed.
On a
9.^
Third Method of interpreting a Product or
Function of Vectors as a Quaternion ; and on the Consistency of the Results of the Interpretation so obtained, with
those which have been deduced fiom the two
preceding
thods of the present Book.
307.
The
Me-
Conception of the Fourth Proportional to Three
Rectangular Unit- Lines, as being itselfa species of iôM?'^^ Unit

in Geometry, is eminently characteristic of the present Calcu-^
lus
and
offers a
Third Method of interpreting a Product of

is however found to be
two Vectors as a Quaternion: which

* In the
language of
modem geometry, the conic in question

four real, namely the sides AB, BC, CD, da
ginary, namely two from each of the focal points, e and F.
t Compare the Second Note to page 295.
touch eight given arcs
3 A
may
;
be said to
and /o?<r ima-
362
III.
[bOOK
with the two former methods (278,

admits of being easily extended
and
284) of the present
in space and generally to
or
more
lines
three
to products of
Functions of Vectors (289). In fact we have only to conceive*
consistent^ in all its results^
Book
* It was in a somewhat
analogous
that
way
Des Cartes showed,
in his
Geome-
Amsterdam, 1659), that all products and powers of lines,

their lengths alone, and without any reference to their direc-
<r?a (^Schooten's Edition,
considered relatively to
by the suitable introduction of a line taken for

however high the dimension of the product or power might be. Thus (at
page 3 of the cited work) the followmg remark occurs
" Ubi notandum
est, quod per a^ Vel b^, sirailesve, communiter, uon nisi lineas
tions, could be interpreted as lines,

uniti/,
onmino simplices concipiam,
licet illas, ut
nominibus in Algebra
usitatis utar,
Qua-
drata aut Cubos, &c. appellem."
But
it
was much more
directed lines in space
difficult to accomplish the corresponding multiplication

on account of the non-existence of any such line, which
symmetrically related to all other lines, or common to all possible planes (comp. tl
The Unit of Vector-Multiplication cannot properly be itself
Vector, if the conception of the Symmetry of Space is to be retained, and duly cora-
Note to page 248).
bined with the other elements of the question.

at least in theory,
when we come
This difficulty however disappea:

new Unit, of a scalar kind (300
to consider that
which has been above denoted by the temporary symbol
w,
and has been obtain
in the foregoing Section, as a certain Fourth Proportional to

Unit-Lines, such as the three co-initial edges, AB, AC,
Unit- Cube
ad
of
Three Rectanguh
what we have
called
by the proposed conception of it, uuderg(

any manner moved, or turned ; and therefoi
for this fourth proportional,
no change, when the cube abcd
may be considered
is
in
to be symmetrically related to all directions
of lines
in space, or
(or translations) of a point, or body. In fact, we conceive its

termination, and the distinction of it (as + u) from the opposite unit oi the same ki:
( w), to depend 07ily on the usual assumption of an wiit of length, combined wi
all possible vections
the selection of a hand (as, for example, the i-ight hand), rotation towards whi
hand shall be considered to be positive, and contrasted (as such) -with rotation b
wards the other hand, round the same arbitrary axis. Now in whatever manner the
supposed cube
may
be thrown about in space, the conceived rotation round the edge
AB, from AC to AD, will have the same character, as right-handed or left-handed, at
If then the fowth proportional to these
the end as at the beginning of the motion.
by + m, or simply by + 1, at one stage of

that arbitrary motion, it may (on the plan here considered) be denoted by the same
symbol, at every other stage : while the opposite character of the (conceived) rota-
three edges, taken in this order, be denoted
tion,
round the same edge ar, from ad to AC, leads us to iiegard the fourth proporAC as being on the contrary equal to , or to 1. It is true that
tional to AB, AD,
this conception of a
directions therein,
new
may
unit for space, symmetrically related (as above) to all linear

appear somewhat abstract and metaphysical ; but readers
who think it such can of course confine their attention to the rules of calculatiov
which have been above derived from it, and from other connected considerations aiu.
which have (it is hoped) been stated and exemplified, in this and in a former "V
lume, with sufficient clearness and fullness.
:
CHAP.
CONCEPTION OF THE FOURTH UNIT.
I.]
363
that each proposed vector â^ is divided hj the neio or fourth

unit, w, above alluded to ; and that the quotient so obtained,
which is always (by 303, VIII.) the right quaternion I'â,
whereol' the vector a

tor
is
the index^
venient, midtiplied into the
we shall merely reproduce

now as a consequence of a
tinct
is
substituted for that vec-
the resulting quaternion being finally, if Ave think
same fourth
For
unit.
it
con-
in this
way
the process of 284, or 289, although

different train of thought, or of a dis-
but Consistent Interpretation : which thus conducts, by a

to the same Rules of Calculation as before.
new Method,
(1.)
The equation of the
unit-sphere, p2
+1=
XI Y.), may
(282,
ceived to be an abridgment of the following fuller equation
thus be con-
the quotient p : w being considered as equal (by 303) to the right quaternion, I'^p,
which must here be a right versor (154), because its square is negative unity.
(2.)
The equation of the
ellipsoid,
T(tp +
may
pfc)
= k2 - i2
(282, XIX.),
be supposed, in like manner, to be abridged from this other equation
\u u
and similarly
(3.)
uu
\u
\u
in other cases.
We might also write these equations,
other, but connected forms
of the sphere
and
ellipsoid,
under these
= -;
III...^p
u
with intepretations which easily
IV...Tfip+^.U-.-ie;
u J
u
u
\u
offer themselves,
on the principles of the foregoing
Section.
(4.) It
this
is,
however, to be distinctly luiderstood, that we do not propose to adopt
Form of Notation,
suggest
it,
in the practice of the present Calculus
in passing, as one
which
may
and that we merely
serve to throw
Conception, introduced in this Third Book, of
some additional light on the

a Product of two Vectors as a Qua-
ternion.
(5.) In general, the Notation of Products, which has been employed throughout
the greater part of the present Book and Chapter, appears to be much more convenient, for actual use in calculation, than any Notation of Quotients : either such as
has been just now suggested for the sake of illustration, or such as was employed in
the Second Book, in connexion with that First Conception of a Quaternion
(112),
which that Book mainly related, as the Quotient of two Vectors (or of two directed lines in space). The notations of the two Books are,
however, intimately conto
nected,
and the former was judged
to be an useful preparation for the

latter, even as
364
regarded the quotient-forms of many of the expressions used

tics of Operation^ such as
S,
[bOOK
III.
while the Characteris-
V, T, U, K, N,
are employed according to exactly the
same laws
In short, a reader of the
in both.
Second Book has nothing to unlearn in the Third although he may be supposed to
have become prepared for the use of somewhat shorter and more convenient pro;
cesses,
than those before employed.
Section 10
On
of a Power of a Vector
the Interpretation
as a Quaternion.
308. The only symbols, of the kinds mentioned in 277,

which we have not yet interpreted, are the cube a% and the
general power a*, of an arbitrary vector base, a, with an arbifor we have already assigned intertrary scalar exponent, t
pretations (282, (1.), (14.), and 299, (8.)) fox t\iQ particular
symbols a^, a"^ a"S which are included in this last^rm. And
;
we
shall preserve those particular interpretations if
define, in full consistency
we now
with the principles of the present and
preceding Books, that this Power a* is generally a Qy,aternion,

which may be decomposed into two factors, of the tensor and
versor kinds, as follows
Ta' denoting the arithmetical value of the t^^ power ofthe/?onumber Ta, which represents (as usual) the length of the
sltive
base-line a
and
p, perpendicular
\Ja*
denoting a versor, which causes any
to that line a, to revolve
round
it
line
as an axis,
right angles^ or quadrants, and in a positive or negative direction,

according as the scalar exponent, t, is itself a
through
positive or negative
number (comp. 234,
(5.)
).
(1.) As regards the omission of parentheses in the formula I., we may obser\'e
that the recent definition, or interpretation, of the symbol a', enables us to write
(comp. 237,
II.
II.
IIL),
.
T(aO = (Tay = Ta*;
(2.) The axis and angle of the power

determined by the two following formulae
IV.
Ax.
= Ua
III.
a*,
U (a*) = (Ua)'= Ua'.
considered as a quaternion, are generally
V.
a*
2n7r
+ ^tTT
the signs accompanying each other, and the (positive or negative or null) integer, n,
and tt.
being so chosen as to bring the angle within the usual limits,
CHAP.
POWER OF A VECTOR A QUATERNION.
I.]
365
of the (positive or negative) product

(3.) In general (comp. 235), we may speak
as being the amplitude of the sa.me power, with reference to the line a as an
Itir,
axis of rotation
and may write accordingly,
VI.
We may
(4.)
VII.
Ua* = cos
am. a* = |<7r.
write also (comp. 234, VII. VIII.),
+ Ua
sin
VIII.
or briefly,
Ua* = cas
--.
iS
(5.) In particular,
IX.
Ua= ens nTT = +
IX'.
Ua2+i=
Ua
number n (supposed to be whole)

upper or lower signs being taken, according as the
For example, we have thus the cubes,
is even or odd.
..Ua3 = -Ua;
X.
(6.)
The eonjvgate and norm
remembered that
to turn a line J-
ing that line through
XI.
of the
round
a through -
- a)
a3
= -aNa.
be thus expressed
may
round
-tir
+ a, is
XII.
Na* = Ta
being
= a.
parentheses being unnecessary, because (by 295, VIII.) Ka
are given
of
and
the
same
The
vector,
power
scalar,
reciprocal
(7.)
mulae
(it
equivalent to turn-
Ka' = Ta< Ua"' = (- ay
+ ItTr
X'.
power a'
by the
for-
XIII.
XV.
S.a<
.
= Ta<.cos~;
V.a = TrtÛa.
= Ta-.Ua- = a-*= Ka^ Na*
a*
XIV.
sin
(comp. 190, (3.)
^;
).
vector p into parts p' and p", which are respectively

have the general transformation :*
parallel and pex'pendicular to a, we
(8.) If
we decompose any
XVI.
a*pa-^
= aîp + p") a-' = p' + Ua^^.
p",
= the new vector

tity
obtained hy causing p to revolve conically through an angular quanexpressed by tir, round the line a as an axis (comp. 297, (15.)).
(9.) More generally (comp. 191, (5.) ), if q be any quaternion, and if
XYIL..afqa-f = q,
the new quaternion q
Ax.
q,
through
formed from q by such a conical rotation of its own axis

round a, without any change of its angle L q, or of its tensor T5.
tir,
(10.) Treating
ijTt
is
as three rectangular unit-lines (295), the

symbol, or expres-
sion,
XVIII.
in
rj^kj-^k'\
or
XIX.
= rk^ji^k^-t,
which
XX...r>0,
may
tion^- to k
being srr
s>0,
s2l,
<>0,
f<2,
the length or tensor of this line

p being r ; its inclinaand the angle through which the variable joZuwe kp
may be
represent any vector
the shortly following sub-article (11.).

conceive (compare the first Note to page 322) that the two Hues i and
7
are directed respectively towards the south and west
points of the horizon, while the
third line k is directed towards the zenith, then sir is the zenith-distance of
p; and
tTT is the azimuth of the same
line, measured /row south to west, and thence
(if necessary) through north and east, to south again.
Compare
If
we
366
conceived to have revolved, frem the initial position
wards the position kj\ being trr.

(11.) In accomplishing the transformation
XVIII.
pression
ki,
with an
XVL, and
111.
[bOOK
initial direction to-
from the ex-
in passing
symmetric but equivalent expression XIX., we employ
to the less
the principle that
XXI.
kj-^
= S-^ = - K (kj-s) =jsk
which
easily admits of extension, and

VII. or VIII.
(12.) It
oi the
power
be confirmed by such transformations as
may
scarcely necessary to remark, that the definition or interpretation
is
a* oi
I.^
vector a, gives (as in algebra) the exponential property,
any
XXII. ..a*a*=a+,
whatever scalars
two
may
be denoted by s and
and similarly when there are more
As verifications of the expression XVIII.,

we may observe that it gives,
(13.)
torJ
XXIIL..p=- Kp;
(14.)
More
generally,
it
considered as representing a vec
XXIV.
p2
= _ r2.
* be any
scalar,
we have
tl
XXV. .p^=
fact,
and
will be found that if
eminently simple transformation
In
tha
factors of this form.
(rk*fkj-^k-^
= rA<jsA'y-A-.
the two last expressions denote generally two equal quaternions, becai
= r Ilnd, equal angles^ each = I (A") ; an(
1st, equal tensors, each
they have,
Ilird, equal (or coincident) axes, each formed from
two successive
rotations, one through stt
khj
one common system

t7r round
round j, and the other through
309. An?/ quaternion, q, which is not simply a scalar, ma]

a*, by a suitable choice of the base, a J
be brought to the foj^m
andof the exponent, t; which latter may moreover be supposec

to fall between the limits
and 2 ; since for this purpose w<
have only to write,
.#=?^;
I..
II.
TT
..Ta = T/;
III.
.Ua = Ax.^;'
and thus the general dependence of a Quaternion, on a Scalai

and a Vector Element, presents itself in a new way (comp. 17,
207, 292).
*
The employment
note
ber.
we
it
the proposed quaternion is a versor,
of this letter , to denote
what we
called, in the
Tq
to
be denoted at
all)
by the usual symbol,
1,
1,
In gene-
a.
(when necessary
two preced-
fourth unit, &c., was stated to be a merely temporary one.

shall henceforth simply equate that scalar unit to the number one
ing Sections,
ral,
When
and
for that
</e-
num-
CHAP.
EXPRESSIONS FOR VERSORS AS POWERS.
I.]
we have
Ta =
thus
367
or in other words, the base a, of the
an
unit- line.
equivalent potrer a\
Conversely, every versor
be
considered
as
an
a
miit-line^ with a scalar exmay
power of
is
ponent^
which
t,
than two
less
maybe supiposed to he in general positive ^^nd

we may write generally,
so that
..% = a^
IV.
V. ..a = Ax.^ = T-4,
with
VI.
and
>
t<2\
0,
1, the exponent
degenerate into 1 or
or 2, and the base a has an indeterminate or art becomes
And from such transformations ofversors
bitrary direction.
if this versor
although
new methods may be deduced,
Conceive that p, q, k, in Fig. 46, are replaced by a, b, c, with unit-vecand 2, determined

y as usual and let a;, y, z be three scalars between
(1.)
tors a,
for treating questions of sphe-
and generally of spherical geometry,
rical trigonometry,
/3,
by the three equations,
VIL
where
A, B,
ajTT
= 2a,
= 2B,
^7r
c denote the angles of the spherical
z7r
= 2c;
The
triangle.
three versors, indi-
cated
by the three arrows in the upper part of the Figure, come then
noted
VIII.
we have
so that
follow
by easy
last,
/3ya^= y2-2
divisions
q^^V;
g'g
= y2-z;
or
and
X.
y==/3s'a*
=- 1
multiplications, these
two others immediately
X'.
a^y^^y
the rotation round a from

(2.) In X.
to
/3
XI.
is
=-
X".
jSâ^y^
y being again supposed
=-
to be negative.
we may write (by 308, VIII.),

.
a*
= caSA
and then the formula becomes,

tation
= a^
the equation,
IX.
from which
to be thus de-
as above
for
^v = Cj8sB
any spherical
y*
= cysc
triangle, in
which the order of ro-
XII.
cysc
c/3sB
casA
or (comp. IX.),
XIII.
(3.)
S/3a= -
cos c + y
Taldng the
cos
c,
we
sin
= (cos b + j3 sin b)
on both sides of
scalars
(cos a
this last equation,
sin a).
and remembering that
thus immediately derive one form of the fundamental equation of

; namely, the equation,
spherical trigonometry
XIV.
(4.) Taking the vectors,
XV.
which
is
sin
-f
cos
we have
A cos B = cos c sin a sin b.
this other
formula
= a sin A cos B + |8 sin b cos A +
easily seen to agree
with the equation 210,
cos c
with 306, XII., and
XXXVII.
may
Yf3a sin a
sin
also be usefully
compared
368
The
.{5.)
XV. may
result
[bOOK
III.
be enunciated in the form of a Theorein, as follows
"
If there be any spherical triangle ABC, and three lines he drawn from the
centre o of the sphere^ one towards the point a, with a length =sin A cos B
another
towards the point b, with a length = sin B cos A and the third perpendicular to the
plane AOB, and towards the same side of it as the point c, with a length = sin c sin A
;
sin
and
if,
with these three lines as edges, we construct a parallelepiped : the

from o will he directed towards c, and will have a length
intermediate diagonal
= sin c."
(6.) Dividing both
we
members
same equation XV. by
of the
cular let
positive
fall
when c and p
common
are situated at one
and taking
side of that arc (or in one
A and b
hemisphere, of the two into which the great circle through
no
p,
scalars,
p be any fourth point on the sphere, and Q the /oo< of the perpendifrom this point on the arc AB, this perpendicular pq being considered as
find that if
surface),
XVI.
sin
we have
c cos pc
common
then,
= sin a cos b
cos
pa +
sin
cos
cos
pb
-f
sin
A sin b
sin c sin
pq
a formula which might have been derived from the equation 210, XXXVIII., by fir
cyclically changing ahcARC to fccaBCA, and then passing from the former triangle
its
polar, or supplementary
and from which
many
less
general equations maj'
deduced, by assigning particular positions to p.

(7.) For example, if we conceive the point p to be the centre ofthe circumscrihe
small circle abc, and denote by R the arcual radius of that circle, and by s tl
scmisum of the three angles, so that 2s=A + B + c=7r

297, (47.), the area* of the triangle ABC, whence
XVII.
the formula
XVI.
PA = PB = PC =
i?,
and
sin
pq =
if(T
<T,
again denote, as
ii
R sin (s c),
sin
gives easily,
XVIII.
2 cot i? sin
= sin A sin b
sin c
a relation between radius and area, which agrees with knov\n

we may, by 297, LXX., &:c., deduce the known equation
;
'
results,
and from whic
XIX.
c tan i2
= 4 sin - sin - sin 2
in
which we have
still,
2'
as in 297, (47.), &c.,
XX.
(Sa/3y =) sin a sin 6 sin c
= t&c.
manner we might have supposed, in the corresponding general equation 210, XXXVIII., that p was placed at the centre of the inscribed small circle,
and that the arcual radius of that circle was r, the semisum of the sides being *
(8.) In like
and thus should have with ease deduced
this other
known
relation,
which
is
a sort
of polar reciprocal of XVIII.,
XXI ...
But
to
2 tan r
sin s
= e.
these results are mentioned here, only to exemplify the fertility of the formulae,
which the present calculus conducts, and from which the theorem in
early seen to be a consequence.
Compare the Note
to the cited sub-article.
'
divides the splie-j
(5.) was,
CHAP.
I.]
(9.)
We
EXPONENTIAL FORMULA FOR THE SPHERE.

might develope the ternary product
loped the binary product XIII.
New
S p'.
the latter, by the symbol
forms, wonld thus arise, of which

suggested the investigation.
(10.)
in the equation XII., as
compare scalar and vector parts
we deve-
and operate on
general theorems^ or at least
may
it
369
new
general
be sufficient in this place to have merely
regards the order of rotation (1.) (2.), it is clear, from a mere inspecformula XV., that the rotation round y from /3 to a, or that round c from
As
tion of the
B to A, must be positive, when that equation
XV.
holds good; at least
if
the angles
abc, be (as usual) treated as positive : because the rotation

round the line V/3a from /3 to a is always positive (by 281, (3.) ).
(11.) If, then, for any given spherical triangle, abc, with angles still supposed
A, B, c, of the triangle
round c from b to a should happen to be (on the conwhich could be

should be obliged to modify the formula XV.
done, for example, so as to restore its correctness, by interchanging a with (3, and at
the same time a with b.
to be positive, the rotation
trary) negative,
(12.) There
still
remaining
we
however, a sense in which the formula might be considered as

without any change in the mode of writing it namely, if we were
is,
true,
symbols A, B, c as denoting negative angles, for the case last supAccordingly, if we take the reciprocal of the equation X., we get this
to interpret the
posed (11.)-
other equation,
XXII.
where
x, y, z are positive, as before,
are negative,
(13.)
if
On
X, y, z,
the whole, then,
which
form VII.
if a,
[3,
|8
to
y be
iVse//'
a from
/3
to
- y,
z,
negative, as in (1.).
y be any given system of three co-initial and

assign a system of three scalars,
we can always
shall satisfy the exponential equation X.,
the rotation round
As
=-
and therefore the new exponents, x,
to the spherical angles A, B,
80 as to fall between the limits
(14.)
a-^/3-2' y-^
the rotation round a from
diplanar unit-lines, OA, OB, oc,

the
2, will
is
and
shall
have relations of
c; but these three scalars,
if
determined
be all positive, or all negative, according as
negative, as in (1.), or positive, as in (11.).
regards the limits just mentioned, or the inequalities,
XXIII. ..x<2,
y<2,
z<2;
x>-2, y>-2,
z>-2,
they are introduced with a view to render the problem of finding the exponents
in the formula X. determinate ; for since we have, by 308,
XXIV.
.a4
= |3^ = y4=+l,
if
Ta = T/3 = Ty =
xyz
l,
we might otherwise add any multiple (positive or negative) of the number four, to
the value of the exponent of any unit-line, and the value of the resulting j?ower would
not be altered.
(15.) If
we admitted exponents = + 2, we might
render the problem of satisfy-
ing the equation X. indeterminate in another way ; for it would then be sufficient to
suppose that any one of the three exponents was thus equal to + 2, or 2, and that
the two others were each
(16) When
X.,
if
it
was
limited as above,
or else that all three were of the form
+ 2.
lately said (13.), that the exponents, x, y, z, in the formula
would have one common
sign, the case
was
tacitly excluded,
which those exponents, or some of them, when multiplied each by a quadrant,

give angles not equal to those of the spherical triangle abc, whether positively or
for
3 B
370
[bOOK
III.
negatively taken but equal to the supplements of those angles, or to the negatives
of those supplements.
;
- 1), that the equation

a^ = (3'^ = yi
(17.) In fact, it is evident (because
X., or
the reciprocal equation XXII., if it be satisfied by any one system of values of xi/z^
Avill still be satisfied, when we divide or multiply any two of the three exponential
by the squares of the two unit-vectors, of which those factors are supposed to
he powers: or in other words, if we subtract or add the number two, in each of two
factors^
exponents.
(18.)
We may,
for
XXV.
=a2-*/3^yy-
example, derive from XXII. this other equation
which, when the rotation
is
XXVI.
or
a2-^/32-y= y-^
as supposed in (1.), so that
xyz aie
positive,
maybe
in-
terpreted as follows.
A and b on its two bounding semicircles,

a from b to c or, what comes to the same thing,
with a positive rotation round A from b to c'. Then, on the plan illustrated by Fi b, of the angles a and b in the
gures 45 and 46, the supplements tt-a, tt
triangle
ABC, or the angles at the same points A and b in the co-lunar triangle abc', will
(19.) Conceive a lune cc', with points
and with a negative
rotation round
represent two versors, a multiplier, and a multiplicand, which are precisely those
denoted, in XXVI., by the two factors, a^-* and jS^-y-, and the product of these two
factors,
and
is
taken in this order,
at that point c'
at c, or to the angle
unit-line oc', or
fore,
is
that third versor, which has
its
axis directed to
o',
same general plan (177), by the external angle ofthe lune,

which, in quantity, is equal to the external angle of the same lune
represented, on the
by
y,
tt
c.
This product
which has
its
exponent
(tt
TT
-z
c) = 2
power of the
we have
there-
this construction, the equation,
XXVII.
which (by 308,
310.
a2-^|3-J/
= (-y)2-*;
(6.) ) agrees with the recent formula
The
2c
results
XXVI.
equation,
I.
which
therefore equal to ihoi
is
from 309,
2b
2a
7'^j3'^ d"^
(1.),
= -l,
and in which
a, /3,
7 are the
unit-vectors oa, ob, oc of any three points on the unit-sphere

while the three scalars a, b, c, in the exponents of the three
;
factors, represent generally the angular quantities of rotation,

unit-lines, or radii, a, /3, 7, from the plane
Aoc to the plane aob, from boa to bog, and from cob to coa,
and are positive or negative according as these rotations of
or negative must be regarded
planes are themselves positive
round those three
as an important formula, in the applications of the present

It includes, for example, the whole doctrine of
Calculus.
Spherical Triangles; not merely because
it
conducts, as we
CHAP.
SPHERICAL SUM OF ANGLES.
I.]
have seen (309,
(3.)
to one form of the
),
371
fundamental scalar
equation of spherical trigonometry^ namely to the equation,

II.
cos c + cos A cos B = COS c sin A sin b
gives a vector equation (309, (4.) ), which

serves to connect the angles, or the rotations, a, b, c, with the
but also because
it
of the radii, a, j3, 7, or oa, ob, oc, for any system

It may, thereof three diverging right lines from one origin.
additional
make
a
to
here
few
be
not
remarks,
fore,
improper
directions*
respecting the nature, evidence and extension of the recent
formula
I.
members
(1.) Multiplying both

20
'f
we have
of the equation
the transformation (comp. 309, (1.) )

21
2b
IIL
/3^
2c
by the inverse exponential
I.,
2 (tt
a^ = -y"'r =y
c)
-^
2a
(2.) Again, multiplying both
members
of
I.
inta\ a
mula:
IV.
2c
2b
y'r
jS'^
2a
=-a~'^
we
t,
obtain this other for-
2(ff a)
=a
"
2a
IV. hy a"" ^
(3.) Multiplying this last equation
we
derive these other forms
* This
may
2o
and the equation
III. into y'f,
be considered to be another instance of that habitual reference to
direction, as distinguished
from mere quantity (or magnitude), although combined
and is eminently characteristic of

whereas Des Cartes, on the contrary, had aimed to reduce all problems of geometry to the determination of the lengths of right lines : although (as all who use
his co-ordinates are of course well aware) a certain reference to direction is even in
therewith, which pervades the present Calculus,
it
his theory inevitable, in
connexion with the interpretation of negative roofs (by him
called inverse or false roots) of equations.
Thus
in the first sentence of Schooten's
recently cited translation (1659) of the Geometry of
" Omian Geometriae Problemata
facile
Des
Cartes,
we
find it said:
ad hujusmoditerminosreduci possunt, ut
deinde ad illorum constructionem, opus tantura sit rectarum quarundam longitudinem

cognoscere."
The very
makes
it
different view of geometry, to which the present writer has been led,
the more proper to express here the profound admiration with which he re-
gards the cited Treatise of Des Cartes
containing as
it
does the germs of so large a
portion of all that has since been done in mathematical science, even as concerns
imaginary roots of equations, considered as marks of geometrical impossibility.

f For the distinction between multiplying a quaternion i7ito and by a factor, see
the Notes to pages 146, 159.
ELEMENTS OF QU ATP:RNI0NS.
372
2a
V.
a"
2b
2c
80 that cyclical permutation
2b
=_!
y'r ^77
VI.
of the letters, a,
/3,
/S'f
2a
III.
2c
= -1
a'T y'r
and
y,
[bOOK
allowed
is
A, B, c,
the
as indeed was to be expected, from the nature of the theorem which

equation I.
that equation expresses.
;
(4.)
From
either
V. or VI. we can deduce the formula:

2a
VII.
by comparing which with

is
yT = -/3
and IV
III.
2b
2c
a'T
we
B)
2(7r
= /3
see that cyclical
permutation of
letters
permitted, in these equations also.
(5.) Taking the reciprocal (or conjugate) of the equation

pare 309, XXII.) this other equation
I.,
we
obtain
(com-
2a
VIII.
2
or
IX.
- A)
(ff
which cyclical permutation of
in
III.)
we can
2b
The equation X. may

2(7r
XI.
at the
And
2(7r-c )
2(w.-B)
/3
letters is
=+
again allowed, and from which (or from
2a
(7.)
7""^=-!;
at once derive the formula,
X.
(6.)
2c
a^^"^
all
a ^
2C
-n
=-ylt,
also be thus written (comp. 309,
A)
2 (TT
''
B)
2(7r
-^
c)
y~
XXVII
2(7r-c)
TT
=(_y)
may be
interpreted {comT^. 309, (19.) ), and

a reference to that general construction (177) for the
the foregoing equations
same i\mQ proved, by
2b
multiplication ofversors, which the Figures 45 and 46 were designed to illustrate; if

we bear in mind that a power a*, of an unit-line a, with a scalar exponent, t, is (by
308, 309) a versor, which has the effect of turning a line

round a as an axis of rotation.
-I-
a,
through
right an-
gles,
(8.) The principle expressed by the equation I., from which all the subsequent
equations have been deduced, may be stated in the following manner, if we adopt the
definition proposed in an earHer part of this work (180, (4.) ), for the spherical sum
of
two angles on a spheric surface

^^
For any spherical triangle, the Spherical Sum of the three angles, if taken
:
suitable Order, is equal to

(9.) In fact,
add the angle b
when
Two Right
the rotation round
a from b
to the angle a, the spherical
sum
to
c, the
angle c
itself,
we
get
sl
is
negative,
if
we
spherically
so obtained is (by the definition
referred to) equal to the external angle at c; if then
ment of
in a
Angles."
we add
sum, or suppleexactly equal to
to this
final or total sum, which
is
addition of spherical angles at one vertex, and therefore in one plane, being acTT
complished in the usual manner; but the spherical summation of angles with different vertices being performed according to those neio rides, which were deduced in the
;
Ninth Section of Book II., Chapter I. j and were connected (180, (5.) ) with the
conception of angular transvection, or of the composition of angular motions, in dif'
ferent
and
successive planes.
CHAP.
ADDITION OF ARCS ON A SPHERE.
I.]
373
(10.) Without pretending to attach importance to the following notation,

just propose
ception here referred
Using a plus
to.
XII.
the symbol of an
which
to
it is
in parentheses, as a
symbol or characteristic
(4.) )
be abridged as follows:
to the left of the
symbol of the multiplicand,
plier to the left* of the
may
symbol of the angle
to
because such addition corresponds (as above)

have agreed to write the symbol of the multi-
we
a multiplication ofversors, and
I.
.c(+)B(+)A=7r;
added angle being written
added (comp. 264,
may-
and represent the con-
of angles, the formula
of such spherical addition
we
to recall
which may serve
in passing, as one
it
in
every multiplication of qua-
ternions.
311. There
is, however, another view of the important equation

according to which it is connected rather with addition of
arcs (180, (3.) ), than with addition of angles (180, (4.)); and may
be interpreted^ and proved anew, with the help of the supplementary
310,
I.,
or polar triangle, k'b'c', as follows.

rotation round a from b to c being still supposed to be negative, let
be (as in 175) the positive poles of the sides bc, ca, ab ; and let a', /3', y'
be their unit- vectors.
Then, because the rotation round a from y' to /3' is positive
(1.)
The
a', b', c'
(by 180, (2.)

duct
),
and
is
in quantity the
supplement of the spherical angle a, ihe pro-
will be (by 281, (2.), (3.) ) a versor, of which a is the axis,

angle; with similar results for the two other products, a'y', ^'a.
y'fi'
(2.) If then
we
write (comp. 291),

I.
a'
supposing that
11.
we
shall
..Ta
UV/3y,
IV. ..a
V.
..
/3'
= T/3 = Ty =
have (comp, again 180, (2.)
and
= UVya,
l,
and
y'
III.
= UVa/3,
.
Saj3y
> 0,
),
UVy'/3',
i3
A=/ly'/3',
= UVaV, y = UV/3'a',
= Z.a'y', c = Z./3'a';
whence (by 308 or 309) we have the following exponential expressions

three
la^&i
2a
.
y'/3'=a";
2b
a'y'=/3'r;
(3.) Multiplying these three expressions, in
the
new product
2o
^'al^y^.
an inverted order, we have, there-
2c
VII.
and the equation 310,

(4.)
for these
products of unit-lines,
VI.
fore,
and a the
2b
2a
^^ a = /3'a'.
'^
y-^
a'y'.
y%V
= y'2/3'2a'2 = -
proved anew.
And because, instead of VI., we might have written,
I. is
in this -Ây
Compare the Note
to
page 146.
374
^
VIII. ..
'
/3
we
see that the equation to be proved
first
along the side
the successive side
from
c'a',
a' to b', it will
b'c', of
from
/^'
may
be reduced to the form of the identity
a'
/3'
and may be interpreted as expressing, what
move
12.
y^ = --
a
/8'r=-^;
ttT=-|;;
is evident,
that
if
a point be supposed to
then along
the polar triangle a'b'c', from b' to c'
c' to a'
III.
[bOOK
and
finally along the
remaining side
thus have returned to the position from which
it set out,
a'b',
or will
on the whole have not changed place at all.

(5.) In this view, then, we perform what we have elsewhere called an addition of
arcs (instead of angles as in 310) ; and in a notation already used (264, (4.) ), we
may
express the result by the formula,
X.
rt
a'b'
+n
c'a'
+ o
b'c'
each of the the two left-hand symbols denoting an arc, which is conceived to be added
(as a successive vector-arc, 180, (3.) ), to the arc whose symbol immediately /oWomj*
it,
or is written next
it,
but towards the right-hand.
The expressions VI.
or VIIL, for the exponential factors in 310, I., show

new way the necessity of attending to the order of those factors, in that formula
for if we should invert that order, without altering (as in 310, VIII.) the exponents,
we may now see that we should obtain this new product :
(6.)
in a
2a
XL
.a^
2b
(i^
2o
'
'
/a'
= + (y'^V)2;
y7r=_I. ^
p y a
which, on account of the diplanarity of the lines a, /3', y', is not equal to negative
unity, but to a certain other versor ; the properties of which may be inferred from
what was shown
in 297, (64.),
and in 298,
(8.),
but upon which
we cannot
here
delay.
312. In general (comp. 221), an equation, such as
1...?'=?,
between two quaternions, includes a system of/our* scalar equations, such as the following :
II.
Sq^Sq;
Saq'
= Saq
Sj5q'
= Sj5q
Syq'
= Sy^
where a, j3, y may be any three actual and diplanar vectors :

and conversely, if a, /3, 7 be any three such vectors, then the
four scalar equations II. reproduce, and are sufficiently re-
* The
propriety, which such results as this establish, for the use of
Quaternions, as applied to this whole Calculus, on account of its

with the number Four, does not require to be again insisted on.
the-
name,
essential connexion
CHAP.
I.]
QUATERNION EQUATION INCLUDES FOUR.
375
placed by, the 07ie quaternion equation I. But an equation

between two vectors is equivalent only to a system of three scalar equations^ such as the three last equations II. ; for example, in
294, (12.), the one vector equationX.Xll. is equivalent

under the immediately
to the three scalar equations XXI.,

preceding condition of diplanarity
XX.
In like manner, an
equation between two versors of quaternions,* such as the equation
..U/=U^,
III.
includes generally a system of three, but of not more than

because the versor \Jq depends genethree, scalar equations
on
a
rally (comp. 157)
system of three scalars, namely the two
;
which determine
Ax. q, and the
axis
its
one which determines
or because the versor equation III. requires to

angle Lq
combined
be
with the tensor equation,
its
',
IV.
in order to
Tq=Tq,
compare 187
reproduce the quaternion equation
I.
(13.),
Now the
re-
evidently o^ this versorform III., if

a J i3, 7 be still supposed to be unit-lines. If then we met that
equation, or if one of li^form had occurred to us, without any
cent equation, 310,
I., is
knowledge of its geometrical signification, we might propose to

resolve
three
it,
with respect to the three scalars
unknown
quantities.
problem thus proposed,
A, b, c, treated as
The few following
may
remarks, on the
be not out of place, nor unin-
structive, here.
(1.) Writing for abridgment,
V.
and
VI.
cot B = ,
cot c = V,
= cosec a cosec b cosec c,
cot A=-t,
.
the equation to be resolved becomes (by 308, VII., or 309, XII.),
VII. ..(u + y)(tt-t-j8)(*
+ a) = s;
in
which the tensors on both
or
* An
equation, U|o'= Up, or UV9' = UVg-, between two versors of vectors (156),
between the axes of two quaternions (291), is equivalent only to a system oitwo
scalar equations
tem
of two
sides are already equal, because
because the direction of an axis, or of a vector, depends on a sys-
angular elements (111)-
376
VIII.
(y2
1) (2
1)
(<2
[bOOK
111.
1).
a, and dividing the re(2.) Multiplying the equation VII. by t-\- a, and into t
sult by <2 + I, we have this new equation of the same ybrm, but differing by cyclical
permutation (comp. 310, (3.)
and
in like
(<4-a)(+y)(M+/3) = 5;
IX.
X.
.(u + /S)(<+a)(f y)
manner,
Taking the half

279, IV., and 294, II.)
XI
two
difference of the
(3.)
- y^) = V. ^Vay =
]KM
"\iC/3a-a/3) = Vâ,
ySa/3
|(/3y
we
arrive at this
new
equation, of vector form
XII.
which
is
= s.
last equations,
= vY(3n +
tY(3y
ySa(3
aS/3y
S./3 (comp. 294, (9.)
it
we
it
gives
and
for if we oj
by S.a, we get (comp. 298, XXVI.),

XIII.
if
).
(4.) It enables us, however, to determine the two scalars,
and
aS(3y,
= V/3y,
y/3)
equivalent only to a system of two scalar equations, because
when operated on by
rate on
and observing that (by
tSaf3y
a2S|3y
- S(3a Say = S (V/3a
operate on the same equation XII. by S
XIV.
t;Sa/3y
y,
we
.Yay);
get in like manner,
= y2Sa/3 - SaySy/3 = S(Vay.Vy/3).
(5.) Processes quite similar give the analogous result,
XV.
and thus the problem
ttSa/3y
is
= (S^'Sya -
Sy/3S/3a
= S(Vy^ .V^Sa)
resolved, in the sense that expressions
have been found
the three sought scalars t, u, v, or for the cotangents V. of the three sought angle
whence the fourth scalar, s, in the quaternion equation VII., can easily be
a,'b, c
:
deduced, as follows.
(6.) Since (by 294, (6.), changing S to a,
we
have, for any three vectors a,
XVI.
13,
y,
= V(V/3a
aSa/3y
ySa/3y
the expressions XIII.
XV. XIV.
and afterwards
cyclically permuting)
the general transformations,
.Vay),
(5Sa(3y
= V(Vy/3 .Yf3a),
= V(ay.Vy/3),
give,
Ut +a)Saf3y = Y(3a.Vay;
XVII.
((
whence, by VII
XVIII.
(7.)
And
and thus the remaining
XVII.
+ /3) Sa/3y = Vy/3 .YjSa

+ y)Sa/3y = Vay.Vy/3;
s(Sa^y)3
scalar,
s, is
the equation VIII.
give,
((<*
XIX.
XIII.
V
(8.) The equations
with repressions
:
= (Vy/3)2
(V/3a)2 (Vay)2;
also entirely determined.
may
be
verified,
by observing that the expressions
+l)(Sa/3y)2 = (V/3a)2(Vay)2;
(2
1) (Sa/3y)2
(t;2
1) (Sa/3y)2
XIV. XV. XVI.
= (Vy^)^
= (Vay)2
give,
(V/3a)2
(Vy/3)2.
by elimination of Sa/3y,
the
CHAP.
SOLUTION OF THE EXPONENTIAL EQUATION.
I.]
XX.
a<-
= (V
S) (Y(3a .Vay)
/3-
= (V
377
S) (Vy)3 Y(3a)
.
yt;-i=(V:S)(Vay.Vy/3);
by comparing which with the formula 281, XXVIII.,
characteristic
I,
we
find that the three scalars,
u, c,
t,
after suppressing (291) the

are either 1st, the cotangent*
of the angles opposite to the sides a, b^ c, of the spherical triangle in which the three
given unit-lines o, (3, y terminate, or Ilnd, the negatives of those cotangents, the
angles themselves of that triangle being as usual supposed to he positive (309, (10.) ),
according as the rotation round a from /3 to y is negative or positive : that is (294,
or finally, by XVIII., according as thej^wrÂ
as Sa(3y>or <
;
), according
scalar, s, is negative or positive, because the second member of that equation XVIII.
(3.)
is
always negative, as being the product of three squares of vectors (282, 292).
which is that of 309, (1.), we see then anew, by V. and VI.,
(9.) In the 1st case,
that
we
are permitted to interpret the scalars A, B, c, in the exponential formula
as equal to the angles of the spherical triangle (8.), which are usually deBut we see also, that we may add any even multiples of
noted by the same letters.
TT to those three angles, without disturbing the exponential equation ; or any one
310,
I.,
even,
and two odd multiples of
tt,
of cosecants, because s is, for this

(10.) In the Ilnd case, which
in
any
order, so as to preserve a positive
case, negative in VI.,

is
product
by (8.).
we may,
that of 309, (ll.)i

for similar reasons,
1., as equal to the negatives of the
interpret the scalars A, B, c, in the formula 310,
angles of the triangle
and as thus having, what VI. now
positive (8.), a negative product
of cosecants, while
requires, because s is
their cotangents
now
have the values
But we may also add, as in (9.), any multiples of ir, to the scalars thus
required.
found for the formula, provided that the number of the odd multiples, so added, is
itself
even (0 or 2).
The conclusions of 309, or 310, respecting the interpretation of the expO'
(11.)
neniial formula, are therefere confirmed,
present
new analysis
in conducting
and might have been
which
it
is
evident that
anticipated,
by the
we have been
dealing
with real scalars, and with real vectors, only.

(12.) If this last restriction were removed, and imaginary values admitted, iu
we might have begun by operating, sa

on that equation, by theûr characteristics,
the solution of the quaternion equation VII.,

in II.,
XXI.
S,
a,
/3,
and
S y
.
which would have given, with the significations 297, (1.), (3.), of /, m,
and therefore with the following relation between those ^wr scalar data,
XXII.
e2
= 1 _ Z2 _ ^3 ._
n,
and
e,
2lmn,
a system offotir scalar equations, involving the four sought scalars, s, t, u, v; from
which it might have been required to deduce the (real or imaginary) values of those
four scalars,
(13.)
by the ordinary processes of algebra.
The
four scalar equations, so obtained, are the following
= e + lt + mu -\-nv tuv + s
=
XXIII.
ei 4-
mtu + ntv -\-uv
I;
= - CM + Itu+tvi nuv + m-2ln;

= ev + tu + Itv + mvv n
;
3 c
378
eliminating uv and u between the three last of which,

the determinant,
l
1, m<, ntv + et
XXIV.
m,
t,
It
n,
and analogous eliminations
+ l)(ei,
of XXII.,
_n + Zm);
give,
= e(t2+l)
XXV.
= (^^ + 1) {e2t, _ (m - nV)
.
XXVI.
and
we find, with the help
= c(<2
Itv + ev-n
e, tv+ m 2ln
III.
[bOOK
{eu-7n + nl\
(14.) Rejecting then the factor
i^
we
(n
- Ini) + (1 - Z^)
find,
(g*
_ ; + ,) }
as the only real solution of the
problem (12.), the following system of values:
XXVII.
et = l mn;
eu = m nl;
ev = nlm;
XXVIII. ..eh = -Q- l^) (1 - m^) (I - n^)
.
and
which correspond precisely
to those otherwise found before, in (4.) (5.) (6.), and might

therefore serve to reproduce the interpretation of the exponential formula (310).
(15.) But on the purely algebraic side, it is found, by a similar analysis, that
the four equations XXIII. are satisfied also by a system oifour imaginary solutions,
represented by the following formulas
:
=
XXIX. ..f'+ 1=^5 r2+i 0;
= tuv It mu n?> e =
\5
which
be
have mentioned in passing, since they do not appear to

have any such geometrical interest, as to deserve to be dwelt on here though, as
regards the consistency of the different processes employed, it may be remembered
that in passing (2.) from the equation VII. to IX., after certain preliminary multi-
may
it
sufficient to
plications,
we divided by
+ 1,
t"^
we were
as
was supposed
solutions, because t
entitled to do,
when seeking only for real
to be a scalar.
(16.) This seems to be a natural occasion for remarking that the following general transformation exists, whatever three vectors may be denoted by a, /3, y :
XXX.
which proves
Vya
in
to Va/3, is
rotation round
S(V/3y .Vya .Va/3)
=-
(Sa/3y)2
180), that the rotation round the line Y^y,from

or is directed in the same sense (281, (3.) ), as the
always positive ;
from a to )3, &c.
Ya^
(17.) In like
and
manner we have
XXXI.
XXXII.
BO that the rotation
ment the
new way (comp.
round
generally,
S (Va/3 .Vya .V/3y) = + (Sa/3y)2,

S (Vy/3 Vay V/3a) = + (Saj8y)2
.
Yy^ from Vay
three diplanar vectors a,
/3,
V/3a is negative, whatever arrangehave among themselves.
to
may
(18.) If then a", b", c" be the negative poles of the three successive sides, BC, CA,
AB, of any spherical triangle, the rotation round a" from b" to c" is negative: which
is
entirely consistent with the opposite result (180), respecting the system of the
hree positive poles

^19.)
ôr
(8.)
a', b', c'.
quantitative interpretation' of the equation
we may infer from
The bC
it
(by 281,
(4.),
XXX. may also
and 294,
(3.) ) that
-f normals oa', ob', oc' to the three faces bog, COA,
witn orpressions:
be easily
z/oabc
be
Aob have
as-
any
their
-ouq/ to the areas of those faces (as bearing the same ratios to
CHAP.
EXTENSION TO SPHERICAL POLYOONS.
I.]
379
then (with a similar reference to units) the volume of the new pyramid^
of the square of the volume of the old pyramid,
units, &c,),
Oa'b'c', will he three quarters
OABC.
313. But an allusion was made, in 310, to an extension oi

the exponential formula which has lately been under discusand in fact, that formula admits of being easily extended,
sion
;
from triangles to polygons upon the sphere
we may
for
write,
generally,
2A
I.
if A1A2
An-i
2A.i
2Ai
2A2
a~ an.~
02^
~= (- 1)",
ai
and
A be any spherical polygon,
if
the scalars
Ai, A2,
exponents denote the positive or negative
that
of
polygon, considered as the rotations aAiA2,
angles
A1A2A3,
namely those from AiA to A1A2, &c. while n is any
number* > 2.
whole
positive
.
in the
Let oc = y
(1.) One mode of proving this extended formula is the following.
be the unit-vector of an arbitrary point c on the spheric surface and conceive that
arcs of great circles are drawn from this point c to the n successive comers of the
;
polygon.
We
shall thus
have a system of n spherical
the polygon will (generally) be decomposed into

angles,
which may be thus denoted

II.
III.
two
triangles,
and each angle of
(positive or negative) partial
CAiA2 =
AAiC
CA2A3
Ai',
= Ai",
= A2',
= A2",
.
AiAoC
SO tbat, with attention to signs of angles in the additions,
rV.
Also
Ai
= Ai' + Ai",
A2 = A2' + A2", &c.
let
V.
AoCAi = Ci,
A3CA2
= C2,
&c.
and therefore
VI.
Ci
C2 +
+ Crt =
an even multiple of
tt,
which reduces itself to 27r in the simple case of a polygon with no re-entrant angles,
and with the point c in its interior.
(2.) Then, for the triangle CA1A2, of
by 310,
which the angles Are
2A2"
VII.
and in
Ci, Ai', A2",
III., the equation,
like
manner,
a2 ^
2Ai'
ai
for the triangle CA2A3,
"T
2Cx
=y
w
;
we have
* The formula admits of interpretation, even
for the case
2.
we
have,
380
2A3
VIII.
we
2A3"
IX.
we have
so proceeding,
2A2
'T
as
at last,
2Ai"
X.
which reduces
310, (3.)
ai
itself to
is
&c.
IV., the product,
by
2(Ci + C2)
2Ai'
=+
by VI., a product of the form,

2A3
2A,j
a,t'^
when
I.,
The theorem
).
'^
jt
obtain,
ai T
'^
fl[2
III.
2C2
=-
ao T
when we multiply VII. by VIII.,
But,
and
2A2'
as ^
[bOOK
it is
2Ai'
= (-
ai ^
a%'^
multiplied
bt/
!)
and into a "
(comp.
therefore proved.
"
the spherical sum of the successive angles of

(3.) In words (comp. 310, (8.) ),
any spherical polygon^ if taken in a suitable order, is equal to a multiple of two right
angles, which is odd or even, according as the number of the sides (or corners) of the
polygon is itself odd or even" the definition formerly given (180, (4.) ), of a Sphe:
And the reasoning may be briefly

being of course retained.
When an arbitrary point c is taken on the spherical surface, as in (1.),
stated thus.
the spherical sum of the two partial angles, at the ends of any one side, is the supplement of the angle which that side subtends, at the point c but the sum of all such
rical
Sum of Angles,
subtended angles
fore the
sum
is
of their supplements can differ
n be the number of the

(4..)
tions
I.,
therernr, if
sides.
Whatever that number may
in the formula
some whole multiple thereof:

only by some such multiple from
either four right angles, or
we have
be, if
we denote by j9,j
for every vector p,
and
the exponential product
for every quaternion q, the equa-
XI.
PnPPn^ = p
XII.
pnqpn^
=q
whereof the former may (by 308, (8.), be thus interpreted:

"
If any line op, drawn from the centre o of a sphere, be made to revolve coniOA, as n successive axes of rotation, through ancally round any n radii, OAi,
.
gles equal respectively to the doubles
of the angles of the spherical polygon Ai
.a,,,
the line will be brought back to its initial position, by the composition of these n rotations.^^
(5.) Another way of proving the extended formula I., for any spherical juoZy^-on,
analogous to that which was employed in 311 for the case of a triangle on a sphere,
and may be stated as follows. Let Ai', A2', ... a/ be the positive poles of the arcs
is
AiAo, A2A3,
AAi
Then the point Ai
is
and
let
a/, a^,
a,/ be the unit- vectors of those
the positive pole of the
polygon at that point
is
new
arc
Ai'a,,',
n
and the angle Ai
poles.
of the
measured by the supplement of that arc with similar reThus we have the system of expressions
;
sults for other corners of the polygon.
(comp. 311, VI.):

2Ai
XIII.
so that the
unit-lines,
ai w
2A^
ai'a';
a^
'^
= a^'aVi
product of powers in I. is equal to the following product ofn squares

and therefore to the n'^ poiver of negative unity,
CHAP.
XIV.
a'na'n-x
and thus the extended theorem

(6.) This latter process
which
381
FORMULA FOR A SPHERICAL QUADRILATERAL.
I.]
it is possible
that
a'rt-i
a'n-z
a'sa'i
a\a'n = (- 1)"
proved anew.
be translated into another theorem of rotation, on
and last Chapter of
briefly return,* in the Second
is
may
we may
Third Book, but upon which we cannot here delay.

may be remarked however here (comp. 309, XII.), that the extended
exponential formula I. may be thus written:
this
(7.) It
XV.
cas A
For example,
(8.)
if
ca-iSA_i
(suitably measured) are denoted
gative rotation from
ad
ca2S A2
cais Ai
= (-
I)".
by
ab, &c., while a,
to
/3,
y, ^ are the unit vectors of its cor-
then
ners,
XVI.
XVII.
(cos c
therefore,
XVIII.
in fact,
c^s
D cys c
.
B cas a = +
.
c/3s
Hence (comp. 309, XIII.), we may write
(9.)
and
spherical quadrilateral, of which the angles

a, . d, so that a represents the positive or ne-
abcd be any
-y
sin c) (cos
by taking
scalars
cos c cos D +
sin
d - ^ sin d) =
on both
c sin
sides,
cos
each member of this last formula
the angle aeb, or ced, if

(10.) Let p
= OP
face; then operating

r
'
'
'
(cos
b + /? sin b)
and changing
cd =
1.
also,
cos
(cos
a+ a
sin
a)
signs,
b cos a +
sin
b sin a cos ba
equal (by 309, XIV.) to the cosine of

the opposite sides ad, bc of the quadrilateral intersect in e.
is
be the unit vector of any fifth point, p, upon the spheric

by S p on XVII., we obtain this other general formula,
sur-
= sin a cos B cos ap + sin b cos a cos bp + sm A sin B sm ab sm pq

sin c cos D cos cp + sin d cos c cos dp + sin c sin d sin cd sin pr
-I-
which the sines of the sides AB, CD are treated as always positive ; but the sines
of the perpendiculars pq and PR, on those two sides, are regarded as positive or neA to b and from c to d, are negative
gative, according as the rotations round p, from
in
or positive
and hence, by assigning particular positions to
p, several other
but less
general equations of spherical tetragonometry can be derived.

of the opposite sides
(11.) For example, if we place p at the intersection, say F,
AB, CD, the two last perpendiculars will vanish, and two of the six terms will disappear, from the general formula XIX. ; and a similar reduction to four terms will
occur, if we make the arbitrary point p the pole of a side, or of a diagonal.
314. The definition of the power a\ which was assigned in 308,

enables us to form some useful expressions, by quaternions, for circular, elliptic, and spii'al loci, in a given plane, or in space, a few of
which may be mentioned here.

(1.) Let
cular to
it
a be any given unit- vector oa, and

then,
by the
(3
any other given
we
write,
Compare 297,
(24.).
definition (308), if
line ob, perpendi-
382
I.
OB for
we
(2.) If
in a
Ta =
OP = /o=a'/3,
Ta =
of a
= 0,
circle,
with O for centre, and
to OA.
plane perpendicular
retain the condition
Sa/3
l,
will be the circumference
but not the condition Sa/3
1,
0,
be in general a quaternion, and not merely a vector ; but

vector-part (292), we can form this new vector-expression,
product
its
and
radium,
the locus of the point
a*j3 will
II.
where
III.
and now the locus
OP = p = V. a*/3 = /3 cos a; + y sin ar,

= t-K, and IV. . y = oc = Vaj3
2x
a plane
for its major and minor semiaxes
called the excentric anomaly.
(3.)
of
is
If we write, under the
V. .. ob'=^' = V/3a: a
so thatB'
and
p' are
b and P
then,
by
= j3' cos a; +
of this projected point
p' is
then the
if
we take
and with ob and oc

DC
is
what
is
often
en
II.,
sin
a;
circle,
p'
= Vpa a = aVpa,
:
a plane drawn through o,
o/i
the equation,
= a*/3'
with ob' and oc for two rectan-
radii.
(4.)
Under the same
conditions, the elliptic locus (2.), of the point
section of the right cylinder
(compare 203,
VIII.
made by
(5.)
itself, is
the
),
TVap = TVa/3 = Ty,
the plane,
IX.
= Sy /3p,
as a confirmation of
or
IX'.
. .
= Sa/3S/3p
/32Sap
which last form we have, by
X.
Sap = Saj3
(5.) If we
we may again write
cos x,
retain the condition Sa/3
II.
S/3p
(comp. 298,
XXVI.)
and IV.,
= (3^
cos x.
(1.), but not now the condition Ta= 1,

but the locus ofv will now be a loga-
the equation I. for p ;

rithmic spiral, with o for its pole, in the plane perpendicular to oa ; because equal
angular motions, of the turning line op, correspond now to equal multiplications of
the length of that line p.
(G.)
For example, when the scalar exponent
volving unit line,
t is
increased
by
4, so that the re-
XL ..Up = UaÛj3
which it had before the increase oft

returns (comp. 309, XXIV.)
was made, the length Tp of the turning line p itself, or of the radius vector of the
Ta* ; which constant and positive scalar is not now equal to
locus, is multiplied by
to the direction
unity.
(7.) If
we
reject both the conditions (I.),
Ta=l,
and
Sa/3
= 0,
power a', is now neither an unit-line, nor perto the line on which that power operates, as a factor, we
pendicular to (B, namely
must again take vector parts, but we have now this new expression
so that the line a, or the base of the
XII.
in
which we have written,
OP
= p = V.
a'/3
for abridgment.
aj
(2.),
VI. .. op' =
and
we have
p'
centre at o,
its
same conditions
the projections (203) of
VII.
gular
with
while the angular quantity, x,
= a-iy,
right angles to the unit-line OA,
60 that the locus
ellipse,
:
III.
[boOK
= a<(/3 cos X +
sin
a?)
CHAP.
383
EXPRESSIONS FOR CERTAIN SPIRALS.
I.]
..a = Ta,
XIII.
y=V(Ua./3).
more complex case, the locus of p is still a plane curve, and may be
(8.) In this
said to be now an elliptic* logarithmic spiral; for if we suppress the scalar factor,
an ellipse as the locus: but when
a*, we fall back on the form II., and have again
we take
centric
accotmt of
anomaly
tiplications
we
find
ex(comp. (2.)) that equal increments of

mul-
of the length (Tp), of the vector of the new
We may
(9.)
tha.t fsLCiox,
to equal
(x), in the auxiliary ellipse so determined, correspond
also project
through o perpendicular
auxiliary vector y
to
b and
spiral.
p, as in (3.), into points b'
and
p',
still
and then, since it is

becomes (with
Ta > or < 1),

= a'(/3' cos cc + y sin x) = a'j8'
tion of the projected spiral
XIV.
are brought back to the case (5.), and the projected curve
logarithmic spiral, of the known and ordinary kind.
we
so that
on the plane
contains the extremity c of the

= Ua . j8', the equaeasily proved that y
oa, which plane
is
seen to be a
are easily represented, on the same ge(10.) Several spirals of double curvature
by merely introducing a vector-term proportional to t, combined or not
neral plan,
with a constant vector-term, in each of the expressions above given, for the variable
For example, the equation,
vector p.
XV.
.. p
= cta + a%
Ta = 1,
with
and
Sa/3
= 0,
constant scalar different from zero, represents a helix, on the right

circular cylinder VIII.
while c
is a7iy
And
(11.)
hand
terra,
we
if
and so
introduce a
new and
write,
XVI.
we
variable scalar, u, as
a,
factor in the right-
= cta + ua%
have an expression for a variable vector p, considered as depending on two

variable scalars (t and w), which thus becomes (99) the expression for a vector of a
shall
surface
of that important
namely
diculars, let fall
Screw Surface, which
from the various points of a given
of revolution, on wliich that helix, or spiral curve,
315.
Without
it
helix,
is
is the locus
of the perpenon the axis of the cylinder
traced.
at present pursuing farther the study of these loci

may be remarked that the definition (308) of the
by quaternions,
power a*, especially
when Ta= 1, combined with the

and with the identification (295) of those three
important right versors with their own indices, enables us to establish the following among other transfoi'mations, which will be found
useful on several occasions.
laws (182) of
(1.) Let
I.
for the case
i,j, k,
a be any
unit-vector,
The wwanogarithmic
a circular logarithmic spiral.

projection of
what
is
and
let t
II..
..S.a-^=S.a';
be any scalar ; then,
S. a'-i
=S
a^^^
=- S
a<-;
might perhaps be called, by contrast to this one,

Compare the following sub-article (9.), respecting the
spiral
here called an elliptic logarithmic spiral.
384
III.
a*=S.a' + aS.a^-J;
V.
(2.)
Let a and
be
VI..
(S
any two
(5.)
by
also,
X.
XI.
jY.
it
(6.)
(7.)
a-'= S. a<- aS.
a<-i
= aâ-i== 1.
be
let t
still
V. a<
= aS.
=S
a+i.
a^-i
any scalar ; then

a-i
= AV.
= -Y.kt;
;&
=S
jY. = V. t* ;
kY.ji
= Y.j*
(10.), for
a<+J.
and laws,
= -Y.it',
kY. t
Y. kt^_ Y.jt.
an arbitrary vector
p,
may be put
under the
it
= rY. A2+i + rki^Y. {2*.
be expanded as follows
may
XIII.
We
III.
j\ k,
XII. .. p
And
and
the same principles
The expression 308,
following form
i,
iY. it =jY.jt
a*-i)2
roiit-vectors^
.iY. jt =.Y.kt;
by the laws of
"We have
IV.
+ (S
S.a'=:S.i*;
VIII.
aV. a< = a2S
IX.
(4.)
a02
VII.
(3.) Hence,
[bOOK
= r{( cos<7r+Jsin tir) sin stt
shall return, briefly, in the
k cosstt}.
-j-
Second Chapter of
this
Book, on some of
these last expressions, in connexion with differentials and derivatives of powers of

vectors ; but, for the purposes of the present Section, they may suffice.
Section
11
On Powers and Logarithms ofDiplanar Qua-
ternions; with some Additional Formulce.

316.
We
shall conclude the present Chapter with a short
Sup-
plementary Section, in which the recent definition (308) of a power

of a vector, with a scalar exponent, shall be extended so as to include
the general case, of a Power of a Quaternion, with a Quaternion Exponent, even when the two quaternions so combined are diplanar:
and a connected definition shall be given (consistent with the less
general one of the same kind, which was assigned in the Second
Chapter of the Second Book), for the Logarithm of a Quaternion in
an arhitrary Plane :* together with a few additional Formulae, which

could not be so conveniently introduced before.
(1.)
We propose,
then, to write, generally,
known base of the natural (or

q being any qttaternion, and e being the real and
real and positive scalars
so that (as usual),
pierian) system of logarithms, of
Na
* The
quaternions considered, in the Chapter referred to, were all supposed
But see the Second Note to page 265.
i.
in the plane of the right versor
to
CHAP.
LOGARITHMS OF DIPLANAR QUATERNIONS.
I.J
II.
(Compare 240,
(2.)
we
We
(1.)
and
= i=l-f-+ J--+&c. =
III.
shall
2-718 28...
(2.).)
shall also write, for
any quaternion
shall call its principal logarithm, or
and thus
385
9,
its
the following expression for
Logarithm
what
lq=^\Tq+ /.q.TJYqi
simply
have (comp. 243) the equation,

IV.
When
= q.
cW
q
any actual quaternion (144), which does not degenerate (131)
(3.)
into a negative scalar, the formula III. assigns a definite value for the logarithm^
Avhich is such (comp. again 243) that
Iq
is
Y.
VII.
.S\q =
lTq',
UVI9 = JJYq
YI. ..Ylq = /.q.VYq;

VIII. . . TVI9 = /.q',
the scalar part of the logarithm being thus the (natural) logarithm of the tensor ;
and the vector part of the same logarithm \q being constructed by a line in the direction of the axis Ax, q, of which the length bears, to the assumed unit of length, the
same ratio as that which the angle L q bears, to the usual unit of angle (comp. 241,
(2.), (4.)).
were merely required
(4.) If it
to satisfy the equation,
= q,
IX. ..'
in
q is supposed to be a given
which
any negative scalar
(3.),
and actual quaternion, which
we might do
X.
g'
this
and in this view, what

what may be considered
the exponential equation IX., and may, as such, be thus
is
we have
called the logarithm, \q, of the quaternion q, is only
denoted
not equal to
= (log q)n = Ig + 2n7rUVj,
where n
any whole number, positive or negative or null
as the simplest solution of
is
by writing (compare again 243),
XI. ..15
= (log 5)0.
The excepted
case (3.), where g is a negative scalar, becomes on this plan

a case of indetermination, but not of impossibility : since we have, for example, by
(5.)
the definition III,, the following expression for the logarithm of negative unity,
XII. .. l(-l)
which
in its
signifying
of
TT
to 1
form agrees
any
Ivith old
unit-vector, of
= 7rV-l;
and well-known
results,
which the length bears
but
is
here interpreted as
to the unit of length the ratio
(comp. 243, VII.).
We propose also to write,
generally, for any two quaternions, q and q', even

ifdiplanar, the following expi-ession (comp. 243, (4.) ) for what may be called the
principal value of the power, or simply the Power, in which the former qxiaternion q
(6.)
is
the base, while the latter quaternion q'
XIII.
is
the exponent
g9'
{9'l3;
quaternion power receives, in general, with the help of the definitions I.

a perfectly definite signification.
a rector, p, \ta angle becomes a right angle ; the
(7.) When the base, q, becomes
and thus
nnd
this
III.,
utfuiition III. gives therefore, for this case,
3 D
ELEMKNTS OF QUATERNIONS.
386
XIV.
and
this is the
quaternion which
.lp=lTp + Ûp;
be multipled by
is to
XV.
same
for the
When,
(8.)
last
formula becomes
XVII.
we
see that the
as one
which
we may
is
write,
I.,
p9'
9',
in the expression,
= 9'lp.
vector-base, the exponent q' becomes a scalar,
p*
= f'lP = Tp. *Up,

=-
the relation (Up)^
= cos a; -f Up
e^^P
t,
the
XVI.
and because, by
III.
[bOOK
2a;
of the
I.,
= fTT
XVI
or briefly,
sin x,
former definition, 308,
if
1 gives,
power
I'.
a*, is
e^^^
in this
= cpsa;,
ay reproduced,
^^
included \n the more general definition XIII., of the power

by the last mentioned definition,
XVIII.
=
(Up)* = ÛP cps
qi'
for
y (comp. 234, VIII.),
with the recent values XVI. and XVII., of x and tÛp.

of diplanar quaternions, we cannot expect to find
(9.) In the present theory
sum of the logarithms of any two proposed /actors, shall hQ generally equal
that the
to the logarithm of the product
but for the simpler and earlier case of cowjp/anar

may be considered to exist, with due modifica-
quaternions, that algebraic property

tions for multiplicity
of value*
to establish generally the very

(10^) The definition III. enables us, however,
simple formula (comp. 243, IT. III.)
:
XIX.. .\q=]{Tq.Vq) = lTq^\Vq;

in
which (comp. f3.)
),
XX...\Vq==Lq. VYq = Y\q
We
(11.)
have
quaternion base,
q,
also generally,
in
XXI.
TlUg = Z
by XTIL,
iov
XXII.
UIU9 = VYq.
any scalar exponent,
t,
and any
the power,
XXlil.
or briefly,
9<
= i3 = (Tqy.
(cos
Lq+ UVg
XXIW. ..q*=TqK CVS tlq,
which the parentheses about Tq
XXIV.
may
if
sin * ^ 9)
v=VYq;
be omitted, because
T (9O = (Tqy=Tqf
(comp. 237,
II.).
When the
p',
base and exponent of a power are two rectangular vectors, p and

(12.)
the product pip is, by XIV., a vector; but
then, whatever their lengths may be,
fa
is
always a versor,
XXV.
we have
"
= cos Ta + Ua sin Ta,
if
a be any vector
therefore,
1!
might have been observed, that every value o^ each member of
the formula IX., there given, is one of the values of the other member; and a similar
remark applies to the formulae I. and II. of 236.
* In 243, (3.),
it
CHAP.
POWERS AND FUNCTIONS OF QUATERNIONS.
I.]
XXVI.
or in words, the
power
pp' is
if
T.pP'-l,
under
versor,
387
8.pp'=0;
this condition of rectangularity.
For example (comp. 242, (7.))* and the shortly following funnula
(13.)
XXVIII.),
XXVII.
.1^= tm = - A
/ = ii\J = + A
the base be an unit-line, and the exponent a line of any length, but
perpendicular to the base, the axis of the power is a line perpendicular to both ; unless the direction of that axis becomes indeterminate, by the power reducing itself to a
and generally,
scalar,
which
if
in certain cases
may
(14.) Thus, whatever scalar c
XXVIII.
this
power, then,
is
i<'J
happen.
may be, we may write,
= 6'yi* =
a versor (12.), and
-^c*''
its
= cos
axis
is
-*
sin
+k
generally the line
but in the
any whole and even number, this versor degenerates into positive or negative unity (153), and the axis becomes indeterminate (131).
case
when
(15.)
XXIX.
c is
any real quaternion
If, for
.
VYq = V,
q,
and therefore
we
write again,
XXX.
.vq =
qv,
XXXI.
and
.v^
= -l,
the process of 239 will hold good, when we change i to v, the series, denoted in I.
by 3, is therefore always at last convergent,^ however great (but finite) the tensor
1q may be ; and in like manner the two following other series, derived from it, which
represent (comp. 242, (3.) ) what we shall call, generally, by analogy to known expressions, the cosine and sine of the quaternion q, are always ultimately convergent :
XXXII.
XXXIII.
We
..
.. cos5
= i(e"5+
-j?^+ j-|l-^-&c.,
f- j^+ ,-^|^.-&c.
shall also dejine that the secant, cosecant, tangent,

still
XXXIV.
XXXV.
to be real, are the functions
sec g
tan g
and thus shall have the usual
We
sin, = i(."^-.') =
(16.)
a quaternion, supposed
(17.)
"?)
2
cosec g
= ^
cot q
relations, sec g
and cotangent of
2y
-^
;
cos g, &c.
shall also have,
XXXVI.
e"*
= cos g + V sin g,
6""^
= cos g - w
sin g
* In the
theory of complanar quaternions, it was found convenient to admit a
of value for a power, when the exponent was not a whole number; and therefore a notation for the principal value of a power was employed, with
certain multiplicity
which the conventions
of the present Section enable
us
now
to dispense.
convergence exists, even when the quaternion is imaginary, or when it is replaced by a biquaternion (214, (8.) )
but we
have no occasion here to consider any but real quaternions.
t In
fact, it
can be proved that this
final
388
and
therefore, as in trigonometry
XXXVII.
whatever quaternion q
(18.)
more
And all
may
[bOOK
III.
(comp. 315, (l.))?
(cos g)2
+ (^sin
5)2
= c''^
-*'
= = 1
be.
theformulee of trigonometry, for cosines and sines of sums of two or
arcs, &c., will thus hold
good for quaternions
common plane ;
nions to be combined are in any
XXXVIII.
cos (9'
9)
= cos q
for
- sin
cos q
provided that the quater-
also,
example,
if
q' sin q,
9'
q.
(19.) This condition of complanarity is here a necessary one; because (comp,

between sums and
(9. ) ) it is necessary for the establishment of the exponential relation
powers.
(20.) Thus,
we may
indeed write,
XXXIX.
69'-3
= 9'.
but, f general, the developments of these
XL.
9'+ 9
9' 9
= il_^L^ 4. terms
XLI.
and
if
9,
\\
',
two expressions give the

of third
difference,
and higher dimensions
i(iqq'-q'q)=Y(Yq.Yq'),
an expression which does not vanish, when the quaternions q and q' are diplanar.
(21.) A few supplementary formulas, cormected with the present Chapter, may be
appended here, as was mentioned at the commencement of this Article (316). And
may be remarked, as connected with the theory of powers of vectors, that if
first it
a,
j8,
y be any three
unit-lines, OA, ob, oc,
triangle abc, then the formula 298,
and
XX. may
if
a denote the area of the spherical
be thus written:
2<r
ALU.
a ^ + r_ T
a
=a
+ (3, y +
/3+y a + jS y+a
..
the exponent being here a scalar.

for any three vec(22.) The immediately preceding formula, 298, XIX., gives
the relation :
tors,
XLIII.
(Ua/3y)2
(Uj3y)2
+ (Uay)^ + (Ua/3)2 + 4Uay
SUa/3. SU/3y =
-2
y be made equal to i, J, k, the first member of this equation be= -2.

comes, 1-1-1-1 +
much more complex identity, involving as it does not
(23.) The following is a
but also /our arbitrary scalars, a, b, c, and r
vectors
three
arbitrary
/3, y,
a,
only
for example, if a,
(3,
has some geometrical applications, and a student would find it a good exercise
in transformations, to investigate a proof of it for himself. To abridge notation, the
but
it
three' vectors a,
/3,
y,
and the three
scalars a, b,
c,
are considered as each composing
a cycle, with respect to which are formed sums 2, and products

may be thus exemplified
IT,
on a plan which
XLIV.
SaV/Sy
= aV/3y + 6Vya +
This being understood, the formula to be proved
cVa/3
is
Ua^ = a^b^c^.
the following
- r2(2a (/3 - y) y
(Sa/3y)2 + (SaV^y)^' + r2(SV/3y)2
+ 2n (r2 + S)3y + 6c) = 2n (r^ + a2) + 2 ElaS
+ S(r2 + a2 + a){(Vi3y) + 26c(r2 + S/3y)-r2(^-y)};
XLV.
the sign of summation in the last line governing all that follows
it.
CHAP.
I.]
ADDITIONAL FORMULA, CONTACTS ON A SPHERE. 389
(24.) For example, by making the four scalars a,

any three vectors a, /3, y, the relation,
6, c, r
each =0, this formula
gives, for
XLVI.
(Sai3y)2
+ 2nS|8y = 2na2
-f
a2(V/3y)2
which agrees with the very useful equation 294, LIII., because
XLVII.
a? ( V/3y)^
a^
(S^y)^
/32y2 }
= (aS/3y)2 - Haa.
(25.) Let a, /3, y be the vectors of three points A, B, c, which are exterior to a
given sphere^ of which the radius is r, and the equation is,
XLVIII.
p2
ri8
(comp. 282, XIII.)
denote the lengths of the tangents to that sphere, which are drawn
from those three points respectively. We shall then have the relations
and
let a, 6, c
XLIX.
+ a2 =
.. a2
/32
4-62=,^2^c2 = -r2;
a2= -a2, &c., and the second member of the formula XLV. vanishes the
member of that formula is therefore also equal to zero, for these significations of
thus r'+
first
the letters
and thus a theorem
in the investigation
is
obtained,
which
is
found to be extremely useful,
of the system of the eight (real or imaginary)

small circles^ which touch a given set of three small circles on a sphere.
by quaternions
We
cannot enter upon that investigation here; but may remark that be(26.)
cause the vector p of the foot p, of the perpendicular op let fall the origin o on the
right line ab, is given by the expression,
as
may
be proved in various ways, the condition of contact of that right line
the sphere
XLVIII.
LI.
or
is
TV/3a
ab with
expressed by the equation,
= 7-T (a-p;);
or
LIL
(V|3a)2
= r2 (a - ^)2
by another easy transformation, with the help of XLIX.,

LIII.
(r2
Sa)3)2
= (r2 +
a^) (r^
+ /32) = a262.
(27.) This last equation evidently admits of decomposition into two factors, representing two alternative conditions, namely,
LIV.
and
we
r2
+ Sa/3 - a6 =
LY.
.. r^
Saj3
+ ah =
',
consider the tangents a and b (25.) as positive, it is easy to prove, in

several difierent ways, that the first or the second factor is to be
selected, according
as the point p, at which the line ab touches the sphere, does or does not fall between
if
still
A and b or in other words, according as the length of that line

sum, or to the difference, of those two tangents.
(28.) In fact we have, for the first case,
the points
is
equal
to the
LVL..T(/3-a) = 6 + a,
in virtue of the relations
LVn. ..T(/3-a)=+(6-a),
and it may be remarked, that we
oi {he two co7iditions
{21 .),
LII. and LI.
= (/3-a)2 + (6+
or
XLIX.
a)3
= -2 (r3+ Sa/3-a6),
but, for the second case,
or
might
= (/3 -a)2 + (6-)3 = -2(r2+ Sa^S +

in this
way have been
and thence the equation
a6)
led to find the

system
LIII,, or its transformations
390
We may
(29.)
[bOOK
III.
conceive a cone of tangents from A, circumscribing tl;e sphere

it along a small circle, of which the pZane, or the /?oZar/)/ane
XLVIII., and touching

of the point A,
is easily
LVIII.
and
manner the
in like
found to have
Sap +
r2
for its equation,
(comp. 294, (28.), and 215, (10.));
equation,
LIX.
S/3p
+ r2 = 0,
represents the polar plane of the point b, which plane cuts the sphere in a second
small circle : and these two circles touch each other, when either of the two conditions (27.)
is
satisfied;
such contact being external for the caseLIV., hwtintemal
LV.
The condition of contact
for the case
have been
(30.)
(26.), of the line and sjoAere, might
otherwise found, as the condition of equality of roots in the quadratic equation
(comp. 216, (2.)),
= (^xa + y/3)2 + (a: + y^ r^,

= ar2 (r^ + a^) + 2xy (r^ + Sa/3) + y2 (r2 + ^2)
LX.
LXI.
or
the contact being thus considered here &s a case of coincidence of intersections.
(31.) The equation of conjugation (comp. 215, (13.)), which expresses that
each of the two points a and b is in the polar plane of the other, is (with the present
notations),
LXIL
.r2-fS/3 = 0;
LXI., which then exist if the line cuts the sphere,

answering here to the well-known harmonic division of the secant line ab (comp.
215, (16.) ), which thus connects two conjugate points.
the equal but opposite roots of
(32.) In like manner, from the quadratic equation* 216, III.,
we get
this analo-
gous equation,
a
connecting the vectors X,
the ellipsoid 216, II.
of
/*
and
we
if
any two points
l,
place the point
fi
(3 j
m, which
L on
ai*e
conjugate relatively to
the surface, the equation
will represent the tangent plane at that point l, considered as the locus
gate point
But
all
whence
it is
LXIII.
of the conju-
easy to deduce the normal, at any point of the
ellipsoid.
researches respecting normals to surfaces can be better conducted, in con-
nexion with the Differential Calculus of Quaternions, to which we shall next proceed.
(33.) It may however be added here, as regards Powers of Quaternions with

scalar exponents (11.), that the symbol q^rq-^ represents a quaternion formed from r,
by a conical rotation of its axis round that of q, through an angle = 2t Lq-, and that
both members of the equation,
LXIV.
are symbols of one
common
(qrq-^y = qriq-\
quaternion.
CoiTected as in the
first
Note
to
page 298.
CHAP.
DEFINITION OF DIFFERENTIALS.
II. 1
CHAPTER
391
II.
ON DIFFERENTIALS AND DEVELOPMENTS OF FUNCTIONS OF

QUATERNIONS; AND ON SOME APPLICATIONS OF QUATERNIONS, TO GEOMETRICAL AND PHYSICAL QUESTIONS.
Section
On
the Definition
of Simultaneous
Differentials.
317. In the foregoing Chapter of the present Book, and in

several parts of the Book preceding it, we have taken occasion
we went along, a considerable variety of Examof the Geometrical Application of Quaternions : but these
to exhibit, as
ples^
have been given, chiefly as assisting to impress on the reader

the meanings of new notations ^ or oi new combinations of sym-
when such presented themselves in turn to our notice.

In this concluding Chapter, we desire to offer a few additional
examples^ of the same geometrical kind, but dealing, more
bols^
than before, with tangents and normals to curves and

surfaces ; and to give at least some specimens, of the applica-
freely
tion of quaternions to Physical Inquiries,
cessary that
we should
first
establish here
some Notations, respecting Differentials

of their Functions, generally.
318. The usual definitions,
o1^
But
it
seems ne-
some Principles, and

of Quaternions, and
differential coefficients,
and
of derived functions, are found to be inapplicable generally to

the present Calculus, on account of the (generally) non-commutative character of quaternion-multiplication ( 1 68, 191). It
becomes, therefore, necessary to have recourse to a new Definition of Differentials, which yet ought to be so framed, as to
be consistent with, and to include, the usual Rules of Different^ atio7i: because scalar s (131), as well as vectors
(292),
have been seen to be included, under the general Conception
f Quaternions.
319. In seeking for such a
new
definition,
it is
natural to
392
[bOOK
III.
go back to the first principles of the whole subject of Diffeand to consider how the great Inventor of Fluxions
be
supposed to have dealt with the question, if he had
might
been deprived of that powerful resource oi common calculation^
which is supplied by the commutative property of algebraic
rentials
multiplication
or
by the familiar equation,

xy = yx,
considered as a general one, or as subsisting for every pair of

factors^ X and y while limits should still be allowed^ but in;
Jlnitesimals be
still
and indeed the Jiuxions them-
excluded:
be regarded as generally finite^* according to

what seems to have been the ultimate view of Newton.
selves should
The answer
320.
which a study of the

contained in the following
believe to be a perfectly general one, as
to this question,
Principia appears to suggest,

Definition^
which we
is
we propose
regards the older Calculus, and which

for Quaternions
to adopt
"
Simultaneous Differentials (or Corresponding Fluxions)

are Limits of Equimultiple s\ of Simultaneous and Decreasing
Differencesr
*
Compare the remarks annexed
to the
Principia (Third Edition, London, 1726)
Lemma of the
Second
Second Book of the
and especially the following passage (page
244):
"Neque enim
spectatur in hoc
nascentium proportio.
Eodem
Lemmate magnitudo momentorum,

momentorum usurpentur vel
recidit si loco
incrementorum ac decrementorum (quas

titatum nominare
etiani motus,
licet) vel finitse quaevis
sed prima
velocitates
mutationes et fluxiones quan-
quantitates velocitatibus hisce proportion-
ales."
f As regards the notion

4ties
which
dent, it
may
especially
(p. 31),
of multiplying such differences, or generally any quantidiminish together^ in order to render their ultimate relations more evibe suggested by various parts of the Principia of Sir Isaac Newton but
all
by the First Section of the First Book.
See for example the Seventh

"
occur
under which such expressions as the following
AB et AD ad puncta longinqua
h ^t
produci,"
intelligantur semper
ideoque rectaa semper finitas
:
"
.
Lemma
."
The direction, " ad puncta longinqua produci," is repeated in conAh, Ad,
nexion with the Eighth and Ninth Lemmas of the same Book and Section; while
under the former of those two Lemmas we meet the expression, " triangula semper
.
finita," applied to
the magnified representations of three triangles, which all diminish

:
and under the latter Lemma the words occur, " manente longi-
indefinitely together
tudine Je," where
ing multiplication
Ae
is
a.
finite
and constant
line,
obtained
of a constantly diminishing line
AE
by a constantly
increas-
(page 33 of the edition
cited).
CHAP, n.] LIMITS OF EQUIMULTIPLES OF DIFFERENCES.
393
And
conversely, whenever any simultaneous differences^ of

any system of variables, all tend to vanish together^ according
to any law, or system of laws
then, if any equimultiples of
;
those decreasing differences all tend together to any system of

finite limits, those Limits are said to be Simultaneous lîffe-
System ; and are de-
rentials of the related Variables of the
noted, as such,
by prefixing the
a characteristic of
letter d, as
differentiation, to the
Symbol of each such variable.
More
let
321.
fully
and symbolically,
I.
..
.q,r,s,.
denote any system of connected variables (quaternions or others); and

let
II.
Ag-,
Ar, A5,
denote, as usual, a system of their connected (or simultaneous) diffe'

rences ; in such a manner that the sums,
III.
shall
2 + Ag,
+ Ar,
+ As,
be a 7iew system of variables^ satisfying the same laws of conwhatever they may be, as those which are satisfied by the old
nexion,
Then, in returning gradually from the new system to the

system I.
old one, or in proceeding gradually from the old to the new, the
II. can all be made (in general) to approach
since it is evident that they may all vanish together.
if while the differences themselves are thus supposed to decrease''^
simultaneous differences
together to zero,
But
we
indefinitely together,
multiply them all
by some one common but
increasing number, n, the system of their equimultiples,
IV.
may
nl\q,
nAs, ...
nAr,
tend to become equal to some determined system offinite
And when
by
this
happens, as in
a suitable adjustment
&c., the limits
all
o|[^tlie
ordinary cases
of the related variables,
q, r, s\
increase of
thus obtained are said
to
it
limits.
may be made to do,
to the decrease of Ag-,
be simultaneous
and are denoted,
as such,
differentials
by the sym-
bols,
V.
quaternion
crease indefinitely,
may
when
^q,
dr,
d5,
be said to decrease, when

that tensor tends to zero.
3e
its
tensor decreases
and
to de-
394
Section
2.
[bOOK
III.
Elementary
Illustrations of the Definition^ from

Algebra and Geometry.
To leave no possible doubt, or obscurity, on the im~

of
the
port
foregoing Definition^ we shall here apply it to determine the differential of a square^ in algebra, and that of a
322.
rectangle, in geometry; in doing which we shall show, that

while for such cases the old rules are reproduced, the differentials
tion,
treated oineed not be small; and that
and not a
were introduced into their expressions,
would be a
vitia-
for the purpose of ren-
all the differentials equal to the
dering
it
correction, of the results, if any additional terms
corresponding
diffe-
though some of them may be assumed to be so,

in
the first Example, one, and in the second Examnamely,
rences
ple, two,
(1.) In Algebra, then, let us consider the equation,
I.
..y = ^2,
which gives,
and
therefore, as usual,*
III.
er
what comes
to the
Ay = 2a;Aa;+
Aa;2;
same thing,
IV.
wAy = 2xnAx + "'(nAa;)2,
an arbitrary multiplier, which may be supposed, for simplicity, to be a

positive whole number.
(2.) Conceive now that while the differences Ax and Ay, remaining always con-
where n
is
nected with each other and with

to zero, the
number n
x by the equation
III., decrease,
increases, in the transformed equation IV.,
and tend together
and tends
to infi-
such a manner that the product, or multiple, nAx, tends to some Jinite limit
which may happen, for example, by our obliging Ax to satisfy always the con-
nity, in
dition,
V.
after
Aa;
= n-â,
nAx = a,
a previous selection of some given and finite value for
We write here,
as usual,
a.
as is common, Aa;' to denote (Aa;)2; while A.a;2 would be

same known plan, for A {x^), or Ay. In like manner we shall wnte
for (da;)2
and shall denote d(a;') by d.x^. Compare the notations
written, on the
da;2j
or
S32, S.52, and Yq^,
Y.q^
in
199 and 204.
CHAP.
We
(3.)
by IV.,
ILLUSTRATION FROM ALGEBRA.
II.]
395
shall then have, ^vith this last condition V., the following expressicm
for the equimultiple
VI.
But because
Ay, of the other
and therefore
Ay:
difference,
nAy = 2xa + n-â^ = b + n-â^,
= 2xa.
if
given and
a,
finite, (2.), while the number n increases indefinitely, the term n-â^, in this expression VI. for
nAy, indefinitely tends
to zero, and its limit is
Hence the two finite quantities, a and b
rigorously null.
a2, is
(since x is supposed to be finite), are two simultaneous limits, to which, under the
supposed conditions, the two equimultiples, nAx and wAy, tend;* they are, therefore,
by the definition (320), simultaneous differentials of x and y and we may write ac:
cordingly (321),
VII.
= a,
'fla;
dy
= 6 = 2xa
or, as usual, after elimination of a,
VIII.
(4.)
And
it
dy = d..r2
would not improve, but
vitiate,
2a;da;.
according to the adopted definition
(320), this usual expression for the differential of the square of a variable x in algebra, if we were to add to it the term dx^, in imitation of the formula III. for the
difference
For this would come
A.x^.
supposing that, for a given and finite
to
value, a, of dx, or of Aa;, the term n'â^, or n'^dx^, in the expression
could fail to tend to zero, while the member, n,
by which the square
VI. fornAy,
of dx
is
divided^
increases without limit, or tejids (as above) to iyifinity.

(5.)
As an arithmetical example,
let there
and
be the given values,
y = x2 = 4,
IX. ..ar = 2,
da;
= 1000;
be required to compute, as a consequence of the definition (320), the arithrithmetical value of the simultaneous differential, dy.
have now the following
let it
We
equimultiples of simultaneous differences,
X.
.nAa;
= da; = 1000;
Ay = 4000+ 1000000->
but the limit of the n^^ part of a million (or of any greater, but given scad finite number) is exactly zero, if n increase without limit ; the required value of dy is, therefore,
rigorously, in this example,
XI.
(6.)
And we
see that these
dy
= 4000.
two simultaneous
XII. ..
da;
= 1000,
differentials,
dy
= 4000,
are not, in this example, even approximately equal to the

rences,
XIII.
= da; =
value n = 1
Aa;
Ay = 10022 -
1000,
22
two simultaneous
diffe-
1004000,
although, no doubt, from the very conception of

simultaneous differentials, as embodied in the definition (320), they must admit of
having such equisubmultiples of themselves taken,
which answer to the
XIV.
* In
but
it is
n^dx
this case, indeed, the multiple
nAx
and
n'dy,
has by V. a constant value,

namely a
found convenient to extend the use of the word,
case of constants
or to say, generally, that a constant
limit, so as to
is its
own
limit.
include the
396
number
as to be nearly equal, for large values of the
n,
to
[bOOK
III.
some system of simulta-
neous and decreasing differences,
XV.
and
Aa;
Ay;
and more and more nearly equal to such a system, even in the way of ratio,
all become smaller and smaller together, and tend together to vanish.
as they
(7.) For example, while the differentials themselves retain the constant values
XII., their millioyith parts are, respectively,
XVI.
= 0-001,
.n-ida;
n-My = 0004,
and
and the same value of the number n
gives,
by
if
n= 1000000;
X., the equally rigorous values of two
simultaneous differences, as follows,
XVII. ..
Aa;
= 0001,
Ay = 0-004001;
and
so that these values of the decreasing differences
XV. may
already be considered to
be nearly equal to the two equisubmultiples, XIV. or XVI., of the two simMZâneowa
And it is evident that this approximation would be improved^
differentials, XII.
by taking higher values of the number, n, without the rigorous and constant values
XII., of da; and 6y, being at all affected thereby.
(8.) It
is,
however, evident
we might have assumed any
also, that after
assuming y = x^, and
a;
= 2,
as in IX.,
other finite value fi)r the differential dx, instead of the
and should then have deduced a different (but still ^ni7e) value for the
other differential, dy, and not the formerly deduced value, 4000: but there would
always exist, in this example, or for this form of the function, y, and for this value
value 1000
between the two simultaneous
of the variable, x, the rigorous relation
dx and
XVIII.
which
differentials,
dy,
is
.dy =
obviously a case of the equation VIII.
4da;,
and can be proved by similar
rea-
sonings.
323. Proceeding to the jiTom'ised
we
shall agai7i see that differences
Example from Geometry
and
diflferentials are
(322),
not in gene-
and that the latter (like the

shall also see that the differentials
ral to be confounded with each other,
former) need not he small.

{like the differences)^
But we
which enter into a statement of relation, or
into
the enunciation of a proposition, respecting quantities which vary together, according to any law or laws, need not even he homogeneous
among
themselves
homogeneous with
is
the
differential,
it
being sufficient that each separately should be

which it corresponds, and of which it
the variahle to
as line of line, or area of ayêa.
It will also be seen
that the definition (320) enables us to construct the differential of a

rectangle, as the sum of two other {finite) rectangles, without any reference to units of length, or of area, and without even the thought of
employing any numerical calculation whatever.
CHAP.
ILLUSTRATION FROM GEOMKTRY.
II.]
annexed Figure 74, abcd be any given rectangle, and
(1.) Let, then, as in the
BE and dg be any
let
increments of
its
397
arbitrary but given and finite

Complete the
sides, ab and ad.
G.
increased rectangle gaef, or briefly af, which will
thus exceed the given rectangle ac, or ca, by the
cg
of the three partial rectangles, ce, cf,
sum
or
by
what we may call the gnomon,* cbefgdc. On the
diagonal cf take a point i, so that the line ci may
;
be any arbitrarily selected submultiple of that diago-
and draw through i, as in the Figure, lines hm,

;
KL, parallel to the sides ad, ab and therefore in-
nal
Fig. 74.
on the
tercepting,
sides ab,
ad
dk
prolonged, equisubmultiples bh,
two given
of the
increments, be, dg, of those two given sides.

(2) Conceive
we
now that, in this construction, ihepoint i approaches to c, or that

new points i, on the given diagonal cr, nearer and nearer to the
take a series of
c, by taking the line ci successively a smaller and smaller part of that

Then the two new linear intervals, bh, dk, and the new gnomon, cbhikdc,
sum of the three new partial rectangles, CH, ci, CK, will all indefinitely de-
given point
diagonal.
or the
and
crease,
will tend to vanish together
remaining, however, always a system of
three simultaneous differences (or increments), of the
two given
AB, AD, and
sides,
of the given area, or rectangle, Ac.

(3.) But the given increments, be and DG, of the two given sides, are always
(by the construction') equimultiples of the two first of the three new and decreasing
differences; they may, therefore, by the definition (320), \>Q arbitrarily taken a.^ two
simultaneous differentials of the two sides,
AB and ad, provided
as the corresponding or simidtaneous differential
equimultiple of the
rectangles,
the
of
AC and
new gnomon
Ai,
(2.), or of the
whereof the^rs^
is
that
we
of the rectangle AC, the
then treat,
limit
ofthe
decreasing difference between the two
given.
(4.)
We are then, first, to increase this new gnomon, or the difference of AC, ai,
sum
(2.) of the three partial rectangles, ch, ci, ck, in the ratio of
DG
to
DK
limit will,
and secondly,
by the
be
to
or
bh, or
of the area so increased. For this last

and rigorously equal to the sought dift/the given and finite increments, be and dg, be as-
to seek the limit
definition (320), be exactly
ferential of the rectangle
turned (as by (3.) they
AC
may)
to
be the differentials of the sides, ab, ad.
Now when we
thus increase the two new partial rectangles, ch and ck, we

get precisely the two old partial rectangles, ce and CG which, as being given and
constant, must be considered to be their own limits, f But when we increase, in the
(5.)
same
ratio, the other
new partial rectangle
ci,
we do
not recover the old partial
rectangle cf, corresponding to it ; but obtain the new rectangle CL, or the equal
rectangle cm, which is not constant, but diminishes indefinitely as the point i ap-
proaches to c
or CM,
is
in such a
manner that the
limit
of the area, of
this
new rectangle CL
rigorously nulL
* The
word, gnomon, is here used with a slightly more extended signification
than in the Second Book of Euclid,
t Compare the Note
to
page 395.
398
(6.)
tials
then, the given increments, be, dg,
//*,
sides AB,
of the given
be
still
[bOOK
III.
assumed
AD (an assumption which has been
to be the differenseen to he permitted),
the differential of the given area, or rectangle, AC, is proved (not assumed) to be, as
a necessary consequence of the definition (320), exactly and rigorously equal to the
sum of the two partial rectangles CE and CG because such is the limit (5.) of the
;
multiple of the
And
(7.)
new gnomon
(2.), in the construction.
any one were
if
to suppose that he could
improve this known value for
the differential of a rectangle, by adding to it the rectangle CF, as a new term, or

part, so as to make it equal to the old or given gnomon (1.), he would (the definition
being granted) commit a geometrical error, equivalent to that of supposing that the
two similar rectangles ci and CF, bear to each other the simple ratio, instead of bearing (as they do) the duplicate ratio, of their homologous sides.
Section
324.
3.
On some general Consequences of
the Definition.
Let there be any proposed equation of the form,

I. ..Q = i^(y, r,...);
be any assumed (but generally /wzYe) and

let d^-, dr,
whether
simultaneous differentials of the variables^ q^ r,
scalars, or vectors, or quaternions, on which Q is supposed to
and
I.
Then the corresponding (or simuloi
ihîr
function, Q, is equal (by the detaneous) differential
finition 320, compare 321) to the following limit:
depend,
II.
by the equation
dQ = lim.
w[2^(5' + w-M5',
n=
r + w'^dr,
..
,)-F(q,r,
...)};
00
any whole number (or other positive* scalar) which,

is conceived to become
indefinitely
and
so to tend to infinity.
And if, in
greater and greater,
where n
is
as the formula expresses,
particular, we consider the function

variable q, so that
III.
as involving only one
..Q =/(?)=/?,
then
lY.
.dQ = efq = lim.nlfiq

n=
+n-^dq)-fq];
00
formula for the differential of a single explicit function of a

which agrees perfectly with those given, near
single variable^
Si
the end of the First Book, for the differentials of a vector, and
of a scalar, considered each as a function (100) o^ a single sca*
Except
in
tion, this scalar
some rare casQs of discontinuity, not

n
may
at present under our considera-
as well be conceived to tend to negative infinity.
CHAP.
CONSEQUENCES OF THE DEFINITION.
II.]
lar variable,
but which
is
now
399
extended, as a consequence
when the connected
of the general definition (320), to the case
Q, and their differentials, Aq, dQ, are quaternions :

with an analogous application, of the still more general Formula of Differentiation II., to Functions of several Quatervariables, q,
nions.
As an example
(1.)
of the use of the formula IV., let the function of 5 be
its
square, so that
Then, by the formula,
VI.
AQ = Afq = \\m.n{{q + nÂqy
n=
= lim.
n=
(g'
-q"'}
CD
dj + dj 5 +
.
n"^ ^5^),
oD
Avhere dg- signifies* the square of
VII.
dy
or without the pointsf between q
Vir.
that
is,
d.g2=g.d^+dg.5;
and dg,
.
d.52
= g,dg + dg
q;
an expression for the differential of the square of a quaternion^ which does not in general admit of any fm-ther reduction : because q and dq are not generally commutative,
as factors in multiplication.
= dq.q,
it happens, as in
algebra, that q.dq
being complanar, the expression VII. then
When, however,
by the two quaternions q and
dg-
evidently reproduces the usual form, 322, VIII., or becomes,
VIII. ..d.5'
(2.)
As another example,
= 2gdg,
X.
dq\\\q{12S).
Q = fq=q-\
IX...
Then, because
if
the function be the reciprocal,
let
./(g + n-idg')-/g = (g + n-idg)-i-g-i

= (gr + n-i dg)-! 9 - (g + n-idg) } g-i
= n-^{q-\-n~^dqy^.dq.q-^,
{
of which,
when
multiplied
by
n, the limit is
pression for the differential of the reciprocal
XL
- q'^dq.q'^, we have
the following ex-
of a quaternion^
..d.g-i=-g-i.dg.g-;

t The point between d and q^, in
distinguish the differential
the
first
member
of VII.,
is
of the square from the square of the
indispensable, to
differential.
But
denoted briefly by dq^, so the products, q dq and dq

q,
may be written as qdq and dq q the symbol, dq, being thus treated as a whole one,
or as if it were a single letter.
Yet, for greater clearness of expression, we shall re-
just as this latter square
is
tain the point between q

mulae, leaving
it
and dq,
in several (though not in all) of the subsequent forat his pleasure.

it,
to the student to omit
400
[BOOK
III.
or without the points* in the second member, 6q being treated (as in VII'.) as a
whole symholy
XI'.
&.q'^
= q-^(}iq
j"!
an expression which does not generally admit of being any farther reduced^ but becomes, as in the ordinary calculus,
XII...d.9-l = -g-2dy,
that
is,
for the case
if
dÎIlg,
of complanarity of the quaternion and

,
its differentiaL
325. Other Examples of Quaternion Differentiation will be given

but the two foregoing may serve sufficiently
;
in the following Section
and to show the analogy of

those of the older Calculus, while exemplifying also
to exhibit the nature of the operation,

its results to
And we shall
the distinction which generally exists between them.
here proceed to explain a notation^ which (at least in the statement of
the present theory of differentials) appears to possess some advantages
and will enable us
nition^
(1.)
to offer a still
more brief
symbolical dejir
of the differential of a function fq^ than before.
We
have defined (320, 324), that
if dry
be called the differential of a (qua-
ternion or other) variable, 9, then the limit of the multiple,
L..n{/(9 + n-id5)-/5},
of an indefinitely decreasing difference of the function fq, of that (single) variable g,
when taken relatively to an indefinite increase of the multiplying number, n, is the
corresponding or simultaneous differential of that function, and

by the symbol dfq.
(2.)
But
before
we thus
plier, n, is stiU considered
is
denoted, as such,
pass to the limit, relatively to n, and while that multitreated sls finite, the multiple I. is evidently a func-
and
tion of that number, n, as well as of the two independent variables, q
we
and
d^.
And
propose to denote (at least for the present) this new function of the three variables,
II.
n, q,
and dq,
of which the form depends, according to the law expressed

form of the given function, f, by the new symbol,
by the formula
I.,
on the
m...f(q,dqy,
in such a
manner
as to write, for
any two
equation,
IV.
which
may
.fnCq, q')
and q, and any number,
n, the
= n{fCq + n-^q')-fq};
obviously be also written thus,
V.
and
variables, q
f(q
+ n-1 q) =fq + -!/ (q,
q'),
here regarded as rigorously exact, in virtue of the definitions, and without

anything whatever being neglected, as small.
is
Compare
the Note immediately preceding.
CHAP,
(3.)
DISTRIBUTIVE PROPERTY.
ir.]
For example,
and from 324,
And
/(9,
little
calculation in 324, (1.), that,
q) = qq + q'q + n"! q^,
iffq
= q^
(2.), that,
VII.
(4.)
appears from the
it
VI.
401
fn(q, q')
the definition of 6fq
= - (g + n"! g')"! q'q-\

may now
iffq
= ?-'
be briefly thus expressed
VIII.
or, if
..d/5=/j9,d9);
understood, we may write, still more
the sub-index ^ be
simply,
IX...6fq=f(q,dq);
this last expression, /(g-, dq), ox f{q, 5'), denoting thus A function
dent variables^ q and q, of which the form
is
of two independerived* or deduced (comp. (2.) ), from
the^'icen ox proposed form of the function y^ of a single variable, q, according to a

law which it is one of the main objects of the Differential Calculus Cat least as re-
gards Quaternions) to study.
One
326.
of the most important êwera/j9rcper^/e5, of the
functions of this class f(q, q'), is that they are all distributive
with respect to the second independent variable, q\ which is introduced in the foregoing process of what we have called de-
from some given function fq, of a single variable, q:

may be proved as follows, whether the two
rivation,'\
a theorem which
independent variables be, or be not, quaternions.

(1.) Let q" be
any third independent
variable,
and
let
n be any number ; then
the formula 325, V. gives the three following equations, resulting from the law of derivation of/(g,
q)
fxoxa fq
..f(q + -l q") =fq + n-% {q, q") ;

.f(q + n-iq" + n-iq')=f(q + nlq") + nYniq + n-ig\
III.
.f{g + n-lq+n-^q)=fq+n-lfn(q, q+q")',
I.
II.
q')
* It
was remarked, or hinted, in 318, that the usiial definition oi a, derived funcnamely, that given by Lagrange in the Calcul des Fonctions, cannot be taken
as a foundation for a diiFerential calculus of quaternions : although such derived
themselves occasionally in the applications of that calfunctions of scalars present
and (4.), and in some analogous but more general cases, which
culus, as in 100, (3.)
tion,
will be noticed soon.
kind, since
it
Law of Derivation is of an entirely different

from a given function of one variable, to a derived
The present
conducts, as
we
see,
function of two variables, which are in general independent of each other.

function/,, (9,
9')'
of the three variables,
re,
9, 9',
it
The
be called a derived
func-
deduced, hy the Jixed law IV., from the same given function fq,
has in general a less simple form than its own limit, f^ (9, 9'), or
tion, since it is
although
may also
t Compare the Note immediately preceding.
3 F
402
by comparing which we
the^rm
and
of the original function, fq, and the values oi the four variables,
remaining altogether arbitrary : except that n
we
(2.) For example, if
may
(or
may
V.
.
is
q, q', q",
supposed to be a number, or
not^ be quaternions.
take the particular function /g'
form 325, VI. of the derived function /,i($',
VI.
III.
see at once that
at least a scalar, while q, q, q"
and
[bOOK
g'2,
which gives the
we have
q'),
+ q"q + n-lq"i',
= q{q'+q')\{q' + q"\q+n-y{q^q'y;
.fn{q, q")=qgf'
./(^, q'^q")
therefore
VII.
,fn{q,
q+q")-fn(q, q") = qq'+ qq + n-Kq'''+

= {q + w-i q") q' + q'Cq + n^q") + "' q'^
q'q"
+ q"q')
=Mq + n-iq",q'),
as required
by the formula IV.
then that formula as proved, for all values of the number n, we

(3.) Admitting
have only to conceive that number (or scalar) to tend to infinity, in order to deduce
this limiting
form
of the equation
VIII.
f^(q, q'+q") =/^(^,
q')
+f^(q, q")
or simply, with the abridged notation of 325, (4.),
IX.
./(^,
^+^")=/(?,
ql+fCq.q")',
which contains the expression of the functional property, above asserted to exist
(4.) For example, by what has been already shown (comp. 325, (3.) and (4.)),
X...iffj= q%
in each of
then
f(q,
q')
= qq + c[q
andXI. ..if/^ = j-S then /(^, ?')=-r^?V';

which instances we see that the derived function f(jq, q')
relatively to
q',
although
it is
only in the first of them that
butive with respect to q also.

(5.) It follows at once from the formula IX. that
it
is distributive
happens to be
distri-
we have generally*
XII.../(ry, 0) = 0;
and
it is
not difficult to prove, as a result including this, that
XIII.
As a
(6.)
f(^q, q')
may
.f{q, xq')
= xf(q,
q'), if
confirmation of this last result,
x he any
we may
scalar.
observe that the definition of
be expressed by the following formula (comp. 324, IV., and 325, IX.):
XIV.
.fiq,
q')
= \im.n{f(iq+n-iq')-fq};
71= CD
we have
therefore, if
x be any
XV.
.fiq,
finite scalar,
and
m = x-^n,
= x.hm.m{f(q+
xq')
m-iq')-fq}
a transformation which gives the recent property XIII., since

letter
m may
We
be written instead of n, in the formula of definition
abstract here from
it is
evident that the
XIV.
some exceptional cases of discontinuity ^ &c.
chap.ii.]differentialquotients,derivedquaternions.403
327.
Eesuming then
the general expression 325, IX., or
writing anew,
l...&fq^f{q,Aq),
we
d^',
see (by 326, IX.) that this derived function^ dfq, of q and
is always (as in the examples 324, VII. and XL) distribu-
tive with respect to that differential dq,
considered as an inde-
pendent variable, whatever
may
be.
the^rw of the given function fq

see also (by 326, XIII.), that if the differential
We
dq of the variable, q, be multiplied by any scalar, x, the differential dfq, of the function fq, comes to be multiplied, at the
same time, by the same scalar, or that
=
II.
.f{q, xdq) xf(q, dq), if x be any scalar.
And in fact it is evident, from the very conception and
.
defini-
tion (320) of simultaneous differentials, that every system of
must admit of being all changed together to

of
any system
equimultiples, or equisubmultiples, of themselves,
without ceasing to be simultaneous diiferentials or more genesuch diiferentials
rally, that
it is
permitted
system, by any common
to multiply all the differentials^
of a
scalar.
(1.) It follows that the quotient,
Ul...dfq:dq=f(q,dq):dq,
and d*^, does not change when the differendq is thus multiplied by any scalar ; and consequently that this quotient III, is
of the two simultaneous differentials., dfq

tial
independent of the tensor Tdg-, although it is not generally independent of the versor
\Jdq, if q and dq be quaternions : except that it remains in general unchanged, when
we merely change
this
that versor to
its
own
opposite (or negative),or to

is a scalar.
comes to multiplying dg- by 1, which

(2.) For example, the quotient,
lY.
d.q^ :dq=:q-^ dq .q.dq-'^
Ud^, because
= q + 'Udq.q.JJdq-\
in which d^'*' and XJdq-^ denote the reciprocals of d^' and Vdq, is very far from
being
independent of dq, or at least of JJdq since it represents, as we see, the sum of the
given quaternion q, and of a certain other quaternion, which latter, in its geometrical
;
interpretation (comp. 191, (5.)),
conical rotation of
may
be considered as being derived from
Ax.q round Ax.dq, through an angle = 2ldq
q,
by a
so that both the
axis and the quantity of this rotation depend on the versor Udg', and vary with that
versor.
we may, if we please, say that the

we ought not to call it a Differential
(3.) In general
Quotient; but
quotient III.
is
a Differential
Coefficient (corap. 318), beof decomposition into two factors^ whereof one
cause dfq does not generally admit

shall be the differential
dq, and the other
di
function ofq alone.
404
[bOOK
III.
(4.) And for the same reason, we ought not to call that Quotient a Derived
Function (com p. again 318), unless in so speaking we understund a, Function of Two*
independent Variables, namely of q and Vdq, as before.

(6.)
riable,
When, however, a quaternion, q, is considered

so that we have an equation of the form,
V.
it is
as
a,
function of a scalar va-
t,
q=ft, where
denotes a scalar,
then permitted (comp. 100, (3.) and (4.)) to write,
= lim.i-i{/((+A)-î}
/i=0
and
to call this limit, as usual, a derived function oft, because
tion
of that scalar variable,
t,
and
alone,
is
it is
(in fact) a funcindependent of the scalar differential,
dt.
(6.)
We may
also write, under these circumstances, the differential equation,
VII.
and ma,y
dg
Dtq.
VIII.
or
dt,
dfq =f't.dt,
call the derived quaternion, Dtq, or ft, as usual, a

differential coefficient in
because the scalar differential, dt, is (in fact) multiplied by

expression thus found for the quaternion differential, dq or dft.
(7.) But as regards the loffic of the question (comp. again 100, (3.)),
this formula,
it,
in the
it is
im-
portant to remember that we regard this derived function, or differential coefficient,
IX.
or
ft,
X.
or
Btft,
as being an actual quotient VI., obtained
Dtq,
by dividing an actual quaternion,
dft,
or
d^,
by an actual scalar, dt, of which the value is altogether arbitrary, and may (if we
choose) be supposed to be large (comp. 322); while the dividend quaternion X. depends, for its value, on the values of the two independent scalars, t and dt, and on
the form of the function ft, according to the law which is expressed by the genera/
formula 324, IV.,
328. It
is
for the differentiation
of explicit functions of any single variable.
easy to conceive that similar remarks apply to

variables than one; and that when
quaternion fund ions of more
the differential of such a function

under the form,
I.
dQ = dF(q,
r, ^,
.)
is
expressed (comp. 324,
= F(g,
r, s,
dq, dr, d^,
II.)
.),
is always distributive, with

the new function
respect to each
the
of
dr,
d*,
differentials, dq^
being also homoseparately
dimension
with
the
first
(comp. 327),
respect to all
geneous of
.
those differentials, considered as a system

*
Compare the Note
to 325, (4.).
in such a
manner
CHAP.
that,
tion,
405
PARTIAL AND TOTAL DIFFERENTIALS.
II.]
whatever may be the form of the gwe7i quaternion funcQ, or F, the derived* function F, or the third member of
I., must possess this genQmlfimctional property

and 327, II.).
XIII.,
326,
(comp.
the formula
II.
F (Qi
r, Sy
xdg, xdr, xds
.)
=^xF{q,r,
s,
d^',
dr,
d.9,
.),
so that products, as well as
where x may be any scalar:
of q, r, &c. considered
squares, of the differentials dq, dr, &c.,
as so many variables on which
depends, are excluded i'rom the
expanded expression of the differential dQ of the function Q.
For example,
(1.)
the function to be diflferentiated be a product of two qua-
if
ternions,
III...
then
n-^.dq.dr
function
is null,
when
the
number n increases without
limit) the differential of the
is,
IV.
-with
Q = F<iq,r)=qr,
the limit of
easily found from the general formula 324, II., that (because
it is
analogous
dQ = d.^r = dF(?,r) = F(^,
results, for differentials of
(2.) Again, if
we take
r,
dj, dr)=r g. dr
+ d^.r;
products of more than two quaternions.
this other function,
v..
Q = F(g,r) = g-ir,
same general formula 324, II., and observing that we have, for
the number (or other scalar), n, and of the four quaternions, 5, r, 5', r',
then, applying the

all values of
the identical transformation (comp. 324, (2.)
YI.
=
we
find, as
the required
the function
VII.
which
is
n{(9-f
),
n-^q')-^ (r
+ n-h')-q-^r}
-(q + n-^q')~^ q'q-^(r+ rrV),

limit, when n tends to infinity, the following
q-^r'
differential
of
dQ= d.q-'^r-diF(^q,r)==F{q,r,
dq, dr)
again, like the expression IV., distributive
= q-^
dr
- q'^ dq q-W;
.
with respect to each of the dif-
and does not involve the product of those two

q, r,
although these two differential expressions, IV. and VII., are both entirely rigorous, and are not in any way dependent on any supposition that the tensors of dq and dr are small (comp. again 322).
ferentials dq, dr, of the variables

differentials
329. In thus differentiating a function of more variables

we are led to consider what may be called Partial
than one,
of Functions of two or more Q_uaternions ; which

be thus denoted^
Differentials
may
Compare
the Note last referred
to.
406
I.
[bOOK
III.
.d,Q, drQ, d,Q,...
Q be
a function, as above, of q, r, s,
which is here supdifferentiated
with
to
be
each
to
variable sepaposed
respect
others
constant.
as
the
were
And
then, if dQ deif
rately,
If
note, as before, what may be called, by contrast, the Total

Differential of the function Q, we shall have the General For-
mula^
II.
.dQ = d5Q + d,(i + d,Q + ...;
and symbolically,
III.
d = dq + dr + ds +
or, briefly
.,
denote the quaternion variables on which the

function
depends, of which the total differential is
quaternion
if q, r, s,
to be taken
whether those variables be
connected with each other,

(1.)
by any
For example (comp. 328, (1.)

IV.
Q=qr,
if
and the sum of these two partial
independent, or be
all
relation or relations.
),
dgQ = d5'.r, and drQ = g.dr;
then
differentials of
Q makes
up its
total differential d Q,
as otherwise found above.

(2.)
Again (comp. 328,
(2.)
Q = q-^r,
dqQ + drQ =the same dQ
V.
and
if
),
d,.Q
thendgQ^-g-idgr.gr-ir;
5-idr;
which was otherwise found before,
as that
for this
form
of the function Q.
(3.)
and
To exemplify the
r, let
have
possibility of a relation existing
between the variables q

supposed equal to each other in V. we shall then
and accordingly we have here dqQ = q~^dq = drQ.
those variables be
Q= 1
dQ=
now
= c = any constant quaternion; we

and may infer that
in IV., let 5r
(4.) Again,
have
0=dQ=dqQ + drQ
VI.
dr
= - ^-i.dg.r,
if
gr
= c = const.
shall then again
a result which evidently agrees with, and includes, the expression 324,
differential
(5.)
A quaternion,
scalar variables,
its
t,
u,
may happen
q,
.
and then
to
it
case write,
VII.
lus),
we
.dtg
be expressed as
a,
function of two or more
= Dtq d^
.
because,
by the present Article,

by 327, VII., we may in this
duq
= D,,^. dw,
coefficients are
independent of the differentials (as in the ordinary calcushall have (by II.) an expression for the total differential dq, of the form,
VIII.
and may at pleasure

ternions,
for the
will have, as such,
But
partial differentials, dtq, duq, &c.
where the
XL,
of a reciprocal.
dq = dtq
+ duq+
=J)tq.dt+Duq.du-\-
say, under the conditions here supposed, that the derived qua-
CHAP.
ELIMINATION OF DIFFERENTIALS.
II.]
IX.
Dnq,
Dtq,
407
are either the Partial Derivatives, or the Partial Differential Coefficients^ of the
Quaternion Function,
X...9 = F(^
when
with analogous remarks for the case,

289) into a vector, p.
330. In general,
may be
it
,...);
the quaternion, g, degenerates (corap.
considered as evident, from the

is zero ;
definition in 320, that the differential of a constant

be changed to any constant quaternion,
so that if
equation 324,
I.,
dQ
then
c,
in the
to be replaced by 0, in the
is
And
rentiated equation, 324, II.
diffjê-
there be given any .jy^êm
if
of equations, connecting the quaternion variables, q, r, s, ,.

we may treat the corresponding system of diffê rentiated equations, as holding good, for the system of simultaneous differen,
and may therefore, legitimately in

tials, Aq, dr, d^, .
;
theory, whenever in practice it shall be found to be possible,
eliminate any one or more of those differentials, between the
.
equations of this system.,

(1.)
As an example,
there be the
let
I.
where
c denotes
gr
= c,
two equations,
and
II.
= r^,
Then (comp. 328,
a constant quaternion.
(1.),
and 324, (L)
we
have the two differentiated equations corresponding,

III.
in
Qf.dr
d<7. r
IV.
which the points^ might be omitted.

V.
and when we substitute

tial dr,
and obtain
this
new
(2.)
I.
dr
= -^-id^.r,
ds
= r. dr + dr. r
we
gives,
as in 329,
this value in the latter,
VI,
The equation
The former
VL
thereby eliminate the differen-
differential equation,
.
= rq-^.dq.r q-'^.dq.r^.
6s
gives also the expression,
VII. ..r=5->c;
the equation II, gives therefore this other expression,
VIII.
.s = (g-ic)2=
by elimination before differentiation.

expressions VII. and VIII, for r and
And
s,
we
if,
5-1 c^-'c,
in the formula VI.,
we
substitute the
get this other diflferential equation.
to 324, (1.).
408
IX.
=-
d.(9-lc)2
X.
331.
d.(g-'c)2
III.
q-^cq'K6q.q'^c- q-Kdq.q'^ cq'^c;
which might have been otherwise obtained

the form,
[bOOK
(conip, again 324, (1.)
= g-ic.d
and (2.)
),
under
iq-^c)-\-d(q-^c).q-^c.
No
special rules are required, for the differentiation of

functions offunctions of quaternions but it may be instructive
;
how
to show, briefly,
the consideration of such differentiation
conducts (comp. 326) to

class
f{q,
2k
general property offunctions of the
and how that property can be otherwise esta-
q');
blished.
(1.) Let/,
and
(p,
\p
denote any functional operators, such that
I...4;q = <p(fq);
then writing
II.
and
=fq,
III.
whence V.
That is, we may (as usual)
an independent variable, r
.s
= (pr, we
.ds
differentiate the
;
and
have IV.
= -ij/q;
= 6ipq- d^r.
compound function,
<p(fq^,
as iffq were
then^ in the expression so found, replace the diffe-
by differentiating the simple function, fq. For this

comes virtually to the elimination of the differential dr, or of the symbol dfq, in a
way which we have seen to be permitted (330).
rential dfq by its value, obtained
(2.) But, by the definition* of d/5 mdfn(q, q'\ we saw (325, VIII. IX.) that
the differential d/g' might generally be denoted hyf^(^q, dq), or briefly by f(q, dq)\
d<pr and d\pq may also, by an extension of the same notation, be represented
by the analogous symbols, 0^(r, dr) and i/'(<7, d^), or simply by 0(r, dr) and
whence
(3.)
"We ought,
VI.
therefore, to find that
dq)
(q,
= ^ (fq, / (9,
dq)),
if
^Pq
= <p{fq)
or briefly that
VII. ..x^g,
for any two quaternions,
9') =</>(/?,
q,
iire
tions/,, (9, 5'), ^n(9, ?'), ^n{q^ q')
law
the
to
expressed
according
<pq, \pq,
limits to
if
/(9,0)
q\ and any two functions,
P9=0/?,
^,
provided that the func-
deduced (or derived) from the functions /j,

by the formula 325, IV. and that then the
;
which these derived functions fn{q,
q'),
&c. tend,
when the number n
to infinity, are denoted by these other functional symbols, f^q, q'), &c.
this otherwise, or to establish this general property VII.,
(4.) To prove
tends
offuncwithout any use of differentials, we may observe that the

of this class f{q, q'),
of the formula 325, IV. by which the
general and rigorous transformation 325, V.,
are defined, gives for all values of n the equation
tions
functions/,, (5,
q)
VIII.
<l>f{q
n-^q')
= <l>ifq + n-^f(q,
= <Pfq + "' 0n(/9, fn{q,

but also, by the same general transformation.
?'))
q))
CHAP.
DIFFERENTIAL OF A PRODUCT OR POWER.
II.]
IX.
.xp(g
+ n-^ q') =ypq + n~^ rpn (q,
hQUce generally, for all values of the number

independent quaternions,
write,
X.
and
q, q,
for all
n, as well as for all values
forms of the two functions,
= <Pn (fq, fn {q, <?'))>

= oo, is
limiting form, for n
q)
rPu (q,
q')
H = ^/Q
^^
409
of the two
<p,
we may-
'
an equation of which the

(with the notations used) the
equation VII. which was to be proved.
(5.) It is scarcely worth while to verify the general formula X., by any particular example yet, merely as an exercise, it may be remarked that if we take the
:
forms,
of
Xl...fg = q^
which the two
XII.
H = q\
= q\
<pq
by 325, VI., the common derived form,
first give,
/(<?,
q)
= 0 (g,
= qq'
q')
-\-
q'q
"' q"
the formula X. becomes,
XIII.
Mq, q') = Mq', qq+q'q+n-'q')
= 5* (qq + q'q + n^ q'^) + (qq + q'q + n'^q'^) ^2 + -l (^qq' + q'q + -! ^'2)2
.
which agrees with the value deduced immediately from the function
definition 325, IV.,
\pq or 5*,
by the
namely,
XIV.
M9, 0={(? + -NT-9^}
= n{(52 + n-i(^qq'+ ^'^ + -l5'2))2_ (^2)2}.
.
(6.)
In general, the theorem, or
& junction
rule, for differentiating as in (1.)
a function, of a quaternion or other variable,

pressed
and
if
by the formula,
we
did not otherwise
may
of
be briefly and symbolically ex-
XV...d(^/)g = dK/g);
know it, a proof oi its correctness would
be supplied, by
the recent proof of the correctness of the equivalent formula VII.
Section 4
Examples of Quaternion Differentiation,
now be
easy and useful to give a short collection of

Examples of Differentiation of Quaternion Functions and Equations^
additional to and inclusive of those which have incidentally occurred
332. It will
of the subject.
already, in treating of ihQ principles
in 330), then
(1.) If c be any constant quaternion (as
..dc = 0;
I.
III.
d. c/j
II.
= cdfq
= d/^;
= 6fq
d(/^. c)
d(/^ +
IV.
c)
c.
(2.) In general,
V.
if
d(/9
+ ^g +
= 4/? + d^3 +
^"^
briefly,
VI.
dS =
2d,
be used as a mark of summation.

(3.) Also,
VII.
=
d(/g ^?) d/j ^? + /g
.
^q
and similarly for a product of more functions than two the rule being simply, to
in its own place, or without disturbing the order
differentiate each factor separately,
:
3 G
410
of the factors (conip. 318, 319)
and then
add
to
[bOOK
III.
together the partial results (comp.
329).
(4.)
In particular,
VIII.
if
be any positive whole number,
= g-' dq + 5-2 dq.q.

<7'
qdq
.-\-
q''^
+ dq jwl
.
and because we have seen (324, (2.)) that

IX.
d.q-^=
q-^.dq.q-^,
we have
this analogous expression for the differential of a power of a

quaternion, with
a negative but whole exponent,
X.
= -q-^dq. q-^ - q-^ dq

(5.)
To
differentiate
XI.
if
we
root,
q^.q^
= - 5-d
q^-"*
a square
.d.qi
qî
q-^>*
.
we
q*n
q^-'
dq q'^ -q-^dq. q'K

.
are to resolve the linear equation,*
= dq',
or
XI'.
rr'+
r'r
= q',
XII.
= q\
q'=dq,
(6.) Writing
XIII.
whence (by 190, 196)
XIV.
(140, 145),
we
rs
sum
the product and
= d.qi = dr.
r'
also, for this purpose,
have,
it
= K/-=K.^,
will follow that
= Nr = T/-2=T9,
and XV.
= 2Sr= 28.?^,
of these two conjugate quaternions, r and
by
s,
being thus scalars
XI'.,
XVI.
r'^q's
= r's
-{
sr
whence, by addition,
XVII.
and
q'-^r-'^q's
= {r+s)r +
r'
(r
s)
= 2r' (r + )
finally,
r^ ^' + ^"'V
XVIII...
2(r
an expression
+ )
,
'
or
^
XIX...d.5i='-^^^^li-S
4S ji
'
for the diflferential of the square -root of
found to admit of
many
a quaternion, which will be
transformations, not needful to be considered here.
(7.) In the three last sub-articles, as in the three preceding them,
posed, for the sake of generality, that q

if in
any application they happen, on the contrary,
d .^ = m5"-i dq ;
XXII. ..d.q'^ =
and
lq-idq,
XXL
if
is
easily
problem presents
rentiation
enough accomplished in the present
difficulties,
d.q-mz=~ mq-*-idq
XXIII. .. dg ||1 ^ (123);
.
Although such solution of a linear equation, or equation
quaternions,
has been sup; but
to be complanar, the expressions
are then simplified, and take usual, or algebraic forms, as follows
XX.
it
and d^ are two diplanar quaternions
of the first degree, in
instance, yet in general the
without the consideration of which the theory of
of implicit functions of quaternions would be
entirely incomplete.
diffe-
But a
general method, for the solution of all such equations, will be sketched in a subse-
quent Section.
CHAP.
II.]
DIFFERENTIALSOFTRANSCENDENTAL FUNCTIONS. 411
because,
when q
pression
XVIIL,
More
(8.)
complanar with
is
and therefore with
g-,
the numerator of that expression

if
generally,
nary calculus, but
x be any scalar exponent, we may
s).
write, as in the ordi-
under the condition of complanarity XXIII,,
still
XXIV.
or with r, in the ex-
g^,
be written as r'^q (r+
may
d g* = rcg*- J dg
XXV.
or
gd g^
.
= irg= da-.
333. 'IhQ functions of quaternions, whicb have been lately differentiated, may be said to be of algebraic form ; the following are a
few examples of differentials of what

transceiidental functions
may be called, by contrast,

the condition oi complanarity
of quaternions
being however here supposed to be satisfied, in order that
the expressions may not become too complex. In fact, with this sim(dg-
III
2')
plification,
they will be found to assume, for the most part, the known
and usual forms, of the ordinary

(1.)
we have
Admitting the definitions in 316, and supposing throughout that dg

the usual expressions for the differentials of <z and Ig, namely,
I.
(2.)
differential calculus.
We have
III.
(3.) Also, if r
d.
the
by
also,
= 2dg
II.
same system of
d sin g = cos gdg
IV.
dlg
d.
= g-Mg.
= - sin gdg
d cos g
dg, then,
= d. -i = g^-d.rlg = g'-(lgdr +
g*
g,
definitions (316),
and dr be complanar with q and
IV'.
|]|
&c.
by 316,
g- Vdg)
or in the notation of partial differentials (329),
V.
(4.)
In particular,
if
VII.
may
If then the base
d^
.g"
= g'lgdr.
the base g be a given or constant vector, cr, and if the exthe value 316, XIV. of Ip) the recent forthen
t,
ponent r be a variable scalar,

mula IV. becomes.
(5.)
VI.
and
d2.g'-= rg'-idg,
(by
d.a
[.a* =
{
(
lTa
a be a given unit
+ |Ua Va^dt.
1
line,
so that
ITa =
0,
and
Ta =
1.
Ua = a, we
write simply,
VIII.
a^
=-
if
a<+i d#,
da
= 0,
useful formula, for the differential of a

(6.) This
with a variable scalar exponent,
may
and
power of a constant unit
line,
be obtained more rapidly from the equation
308, VII., which gives,
IX.
a*
= cos +
sin
if
evident that the differential of this expression
since
it is
itself
multiplied
(7.)
=by |7rad^, because a2
The formula VIII. admits
nected with the rotation through
also of
Ta =
is
equal to the expression
1.
a simple geometrical interpretation, con-
right angles, in a plane perpendicular to a, of
412
which
rotation, or version, the
power
a',
or the versor Ua*,
is
[bOOK
III.
considered (308) to be
the instrument* or agent, or operator (comp. 293).
334. Besides algebraical and transcendental forms, there are other

of operation on a quaternion, q^ or on a function thereof,
results
as forming a new class (or kind) of funcof

the
out
tions, arising
principles and rules of the Quaternion Calculus itself: namely those which we have denoted in former Chapters
which may be regarded
the symbols,
by
I.
or
K^, S^, Yq, N^, T^, Vq,
by symbols formed through combinations of the same
operation,
II.
And
it is
si^ns of
such as
essential that
SU^, YUq, VYq, &c.
we should know how
to differentiate expres-
sions of these forms, which can be done in the following manner,

with the help of the principles of the present and former Chapters,
and without now assuming the complanarity, ^q\\\q.

(1.) In general, let /represent, for a
any two quaternions, q and
q',
we
HI.
and therefore alsof (comp. 326,
IV.
..
moment, any distributive symbol,

have the equation,
so that for
shall
/(?+?') =/?+/?';
(5.)),
f(xq)
= xfq,
if
x be any
scalar.
(2.) Then, with the notation 325, IV., we shall have
V.
and
./n(?,
q')^n{f{q^n-^q')-fq} =fq';
we shall have
by 325, VIII., for any such function /$,
therefore,
the differential
expression,
VI. ..d/^=/d5r.
(3.)
But
S,
V,
have been seen
to
be distributive symbols (197, 207)
we can
therefore infer at once that
VII.
dK} = Kdg
VIII.
dSg = Sdg
IX.
or in words, that the differentials of the conjugate, the scalar,
dV^ = Vd^
and
the vector of a
quaternion are, respectively, the conjugate, the scalar, and the vector of the differential
of that quaternion.
To find the differential of the norm, Ny, or

dN^, we have (by VII. and 145) the equation.
(4.)
Compare
the second
Note
to
to
deduce an expression for
page 133.
t In
quaternions the equation III. is not a necessary consequence of IV., although the latter is so of the former for example, the equation IV., but not the
equation III., will be satisfied, if we assume fq = qcq-^c'q, Avhere c and c' are any
;
two constant quaternions, which do not degenerate
into scalars.
CHAP.
DIFFERENTIALS OF TENSOR AND VERSOR.
II.]
d^g = d.qKq=dq.Kq + q.Kdq;

qKq = K q'Yiq, by 145, and 192, II.
+ K) j'K^ = 2S qKq = 2S(K^ q), by 196, II., and
XI.
dNg = 2S(K<y dq).
X.
but
and
(1
therefore
413
198,
I.
(5.) Or we might have deduced this expression XI. for dN^', more immediately,
by thQ general formula 324, IV., from the earlier expression 200, VII., or 210, XX.,
for the norm of a sum, under the form,
XI'.
dN^ =lim n{N(g + - dq) - N^}
= Um
n = oo
{ 2S(Kg
dq)
+ n-i Ndg }
= 2S(K5.d^),
as before.
(6.)
and
Ny
The
tensor, Tq, is the square-root
differential
of the tensor of a quaternion, the expression (comp. 158),

XII.
a result which
is
more
easily
dTg =
^^
= S(KU5.d5) =
The
q divided by
versor
its
Uq
tensor
is
Tq
= s^.
f?
Tq
q
equal (by 188) to the quotient, q Tq, of the quaternion

hence the differential of the versor is,
:
XIV...dU5^ = d;f =
Tq
whence follows at once
S:^,
remembered, under the form,
XIII...
(7.)
(190) of the norm, Nj; and because Tq

applied which gives, for the
XXII. may be
are scalars, the formula 332,
this formula,
[^-SÛ=V^.U,;
q JTq
\q
q
analogous to XIII.
and
like
it
easily
remem-
bered,
= V^.
XY...^*
q
Vq
(8.)
q^Tq.
We might also have observed that because (by 188), we have generally
U^, therefore (by 332, (3.)) we have also,
XVI.
d^ = dT^
U^ + T^
dU^',
and
xvii...?i
q
if
then
deduce
= ^^ ^
+ l^^
T?
Ug
'
we have in any manner established the equation XIII., we can immediately

XV. and conversely, the former equation would follow at once from the
;
latter.
as remarkable, that
(9.) It may be considered
for any two quaternions, q and dq, the formula :*
we should
thus have generally, or
* When the connexion of the theory of normals to surfaces, with the

differential
calculus of quaternions, shall have been (even briefly) explained in a subsequent
Section, the student will perhaps be able to perceive, in this formula XVIII., a recognition,
though not a very direct one, of the geometrical principle, that the radii
of a sphere are its normals.
414
XVIII.
S (dU^
Uy) =
lows.
158
..dUq:Vg = Si
XVIII'.
or
but this vector character of the quotient dUj-
[boOK
;
Ug- can easily be confinned, as fol-
Taking the cojijugate of that quotient, we have, by VII. (comp. 192,

and 324, XI.),
XIX.
K (dU^
= KUr 1 dKU^ = U^
U^-i)
III.
d (U^-i)
= - dU^ Ug"'
.
II.
Avhence
XX.
(1 + K) (dU? U^-i) =
which agrees (by 196, II.) with XVIII.

(10.) The scalar character of the tensor^ Tq, enables us always
to write, as in
the ordinary calculus,
XXI.
butlTg = Sl^, by
= dTg
dlT^
Tg
316, V.; the recent formula XIII.
may
therefore
by VIII. be
thus written,
XXII.
(11.)
Sdl7
= dSI? = dT^:d7 = S(d^:
When
d^'
XV.
takes the form,
And
in fact
5, this last difference
Ij
XXIII.
we have
generally,
XXII'. .. dl^-
or
g);
vanishes,
by 333,
II.
9-1
d^ =
S-i 0.
and the equation
dWq = Ydlg = dY\q.
= 'S[\q, by
\\Jq
316, XX., although the differentials

members of the re-
of these two equal expressions do not separately coincide with the
when
cent formula XV.,
We may however write generally
and dg are diplanar.
(comp. XXII.),
XXIV.
335.
dlUg - dU^
\]q
- dg
V(dlg
= dl^-dg
g)
q.
We have now differentiated the six simple functions 334,
which are formed by the operation of the

K,
S,
I.,
six characteristics,
U;
V, N, T,
and as regards the differentiation of the compound functions 334, II.,

which are formed by combinations of those former operations, it is
easy on the same principles to determine them, as
the few following examples.
(1.)
The axis Ax.
UVg
334, IX. and XIV., be thus expressed

.
seen in
of a quaternion has been seen (291) to admit of being rethe differential of this axis may therefore, by
;
g-
presented by the combination
I.
may be
d (Ax. q)
= dUV^ = V (Vdg Vg) UVg

.
whence
"
The
or
differential
s\t\\a.i&d
of the axis
in the plane
d(Ax.g)
Ax. ^
is therefore,
(2.)
Vg
is
'
generally, a line perpendicular to that axis,

it
when the plane (and

the quaternion and its
vanishes,
constant ; or
when
IV.
111
complanar.
Hence,
III.
and conversely
Vdg
UVg
of the quaternion; but
therefore the axis) of that quaternion

differential are
_dUVg^
this
dUVg = 0,
complanarity IV.
if
may
dg
be expressed by the equation III.
CHAP.
DIFFERENTIALS OF AXIS AND ANGLE.
II.]
(3.) It
is
415
easy to prove, ou similar principles, that
and
VL
(4.)
But
dSU^ = SdU^ = S f V
in general, for
?^. U7\
any two quaternions, q and q\ we have (comp. 223,
(o.) ) the transformations,
VII.
S(V5' ^) = S (Yq.Yq)=^,q'Yq
and when we thus suppress the characteristic V before dg 5, and insert it before
Jjq, under the sign S in the last expression VI., we may replace the new factor VU5'
:
- TVU^ UV^
(188), or by TVU^ UV<? (274, XIII.), or by
(204, v.), where the scalar factor TVUg^ may be taken outside (by 196, VIIL);
by TVUj.
UVU?
also for 5-1
UVg we may
the formula VI.
may
substitute 1
VIII.
Now
(UVg'
^), or 1
qUYq,
because
UV5
dSUg = - S
-^
qijYq
XIX.),
TVUg.
IX.
&\Jq
= cos
q,
TVUg = sin L
also, in 316, these expressions for the angle of
X.
we may
the earliest connexions of quater-
be remembered, that
it may
among
(5.)
ternions with trigonometry^ the folloAving formulae occurred (196, XVI.,
we had
Jl|
therefore be thus written,
and 204,
a quaternion,
Zg=TVl5 = TlU^;
therefore establish the following expression for the differential
of the angle
of a quaternion,
XL
(6.)
The following
is
d Z q=6.TY\q
another
way
= dTlU^ = S
-^.
of arriving at the
same
differentiation of the sine instead of the cosine of the angle, or
tion of
that
For this purpose, it is only necessary to remark

of dSUg-.
have, by 334, XII. XIV., and by some easy transformations of the kind
dTVUg-, instead
we
lately
result, through the

through the calcula-
employed
in (4.), the formula,
dividing which by SUiy, and attending to IX. and X., we arrive again at the exXL, for the differential of the angle of a quaternion.
pression
(7.) Eliminating
S (dg gUVj) between VIII. and XII., we obtain the

:
tial equation,
XIII.
SU^ dSUg + TVU^ dTVU^ =

.
differen-
of which, on account of the scalar character of the differentiated variables, the inte-
gral
is
evidently of the /brm,
XIV.
(SU^)2
+ (TVUg)2 = const.;
and accordingly we saw, in 204, XX., that the sum in the

tion is constantly equal to positive unity.
first
member
of this equa-
416
The formula XI. may
(8.)
dUV^ =
by
quotient,
Vl^
HI.
XV...d/.q = S (Y{dq
with the verification,
[bOOK
q) UV^);
when we suppose d^|||^, as
:
in IV., and therefore

the expression under the sign S becomes the differential of the
VYg, and therefore, by 316, VI., of the angle Z. q itself.
that
III.,
:
336. An important application of tHe foregoing principles and

rules consists in the differentiation of scalar functions of vectors, when
those functions are defined and expressed according to the laws and
notations of quaternions.
be found, in
It will
that such diffe-
fact,
rentiations play a very extensive part, in the applications of quaternions to geometry ; but, for the moment, we shall treat them here, as
The
merely exercises of calculation.
following are a few exam-
ples.
(1.) Let p denote, in these sub-articles, a variable vector;
and
let
the following
equation be proposed,
..r2
I.
so that r
279, III.,
we
have,
in
Vr =
which
0,
VIII., and 324, VII., the equation,
by 322,
II.
In
+ p2 = 0,
a (generally variable) scalar. Differentiating, and observing that, by

= 2Spp', if p' be any second vector, such as we suppose dp to be,
pp' + p'p
is
fact, if r
..rdr
+ Spdp = 0;
be supposed positive,
or
III.
it is here,
dr
= - r-iSpdr = rSp-idp.
by 282,
II.,
the tensor of p
last expression III. for dr is included in the general formula, 334,
(2.) If this tensor,
IV.
r,
be constant, the differential equation
Spdp =
- p2 = const,
if
0,
or
V.
so that
we have
r2
= T(ip + pK),
and k are here two constant

(by 334, XI., because Kip
VI.
with
XIX.),
= 0,
dic
= 0,
&c.),
2r3dr = idN(tp
+ pK) = S(pt + icp)(tdp+dpK)

= Ci2 + /cO Spdp + 2SKptdp
;
VII. .. 2r-idr
if
V be an auxiliart/
vector, determined
VIII.
which admits
(4.)
.r*v
= Svdp,
by the equation,
= (i2 + k2) p + 2Vicpt
of several transformations.
For example we
may
IX.
write,
.
by 295, VII.,
= (i2 + ^2) p + Kpi + ipK
r*j/
= t(tp + pK) + K(pt+Kp);

or,
by 294,
III.,
and 282, XII.,

X.
becomes simply,
Then, squaring and differentiating,
vectors.
= pi,
di
r4i/
= (i2 +
- k)2
(t
k2) p
+2
(ffStp
2 (iS/cp
so that this
dTp = 0.
if
(3.) Again, let the proposed equation be (comp. 282,
II.
XIII.
- pSiK +
fcStp)
&c.
iSkp)
CHAP.
SCALAR FUNCTIONS OF VECTORS.
II.]
The equation V. gives
(5.)
tion,
XI.
r*
when squared without
(coinp. 190, V.),
= N (ip + pk) = (tp + pK)
= (t^ +
417
(pi
differentia-
+ Kp)
+ cpicp + pKpi
= (t2 + K2)pa + 2Stpicp
k:2)p*
= (t k)2 p2 + 4Sp Skp = &c.,

by transformations of the same kind as before we have therefore, by the recent extwo variable
pressions for r*v, the following remarkably simple relation between the
vectors, p and v,
or
Xir. .Spv= 1.
XII.
Srp = 1
;
When
(6.)
the scalar,
tion,
XIII.
Svdp =
r, is
constant,
whence
we
have, by VII., the differential equa-
XIV.
also
Spdv =
by XII.
0,
of reciprocity thus existing, between the two vectors p and

geometrical signification will soon be seen.
a relation
(7.)
v,
of which the
Meanwhile, supposing r again to vary, we see that the
for
2r^dr
two
last
may
last expression VI.

be otherwise obtained, by taking half the differential of either of the
expanded expressions XI.
for r*
it being remembered, in all these little
permutation of factors, under the sign S, is permitted
(223, (10.)), even if those factors be quaternions, and whatever their number may
be and that if they be vectors, and if their number be odd, it is then permitted,
;
calculations, that cyclical
under the sign V, to invert their order (295, (9.)), and so to write, for instance,
YipK instead of Vicpt, in the formula VIII.
a vector, let
(8.) As another example of a scalar function of
mity (or nearness) of a variable point p to the origin o so that
p denote the
proxi-
XV.
j9
= (- p2)-i = Tp->,
XV'.
or
p-2
+ p2 =
0.
Then,
XVI.
d/?
= Si/dp,
XVII.
if
v=jp3p=/>2u^.
V being here a new auxiliary vector, distinct from the one lately considered (VIII.)
and having (as we
see) the
same
versor (or the
same
direction') as the vector
it-
tensor equal to the square of the proximity of v too; or equal

to the inverse square of the distance, of one of those twa points from the other.
but having
self,
337.
On
its
the other hand,
we have
tions, to consider vectors as functions
often occasion, in the applicaof scalars^ as in 99, but now
with forms arising out of operations on quaternions^ and therefore

such as had not been considered in the First Book. And whenever
we have thus an expression such

or
I. ../i = 0(O,
as either of the
II.
/)
0(5, 0,
for the variable vector of a curve^ or of a surface
and
being two variable scalarSy and
<P(t)
two following,
and
(comp. again 99),
^(5,^) denoting ani/
functions of vector form^ whereof the latter is here supposed to be entirely independent* of the /ormer, we may then employ (comp. 100,
*
We
Articles,
are therefore not employing here the temporary notation of
according to which we should have had, d(pq
3h
<p(q, dq).
some recent
418
[bOOK
III.
and the more recent sub-articles, 327, (5.), (6.), and

notation of derivatives^ total or partial; and so may
the
329, (5.) )
write, as the differentiated equations^ resulting from the forms I. and
(4.)
and
(9.)
TT. respectively,
the following:
III.
IV.
dp = (p't.dt = p'dt = -D,p.dt;

6p = d^ + dtp = D^.ds + Dtp
d^
of which the geometrical significations have been already partially

seen, in the sub-articles to 100, and will soon be more fully developed.
(1.) Thus, for the circular locus, 314, (1.), for which
V.
we
= a%
= D,p =
VI. ..p'
(2.)
And
Sa(S
1,
= 0,
|a'^'/3=|ap.
for the elliptic locus, 314, (2.), for
VII.
we
Ta =
have, by 333, VIII., the following derived vector,
=V
Ta =
a/3,
which
but not Sa/3
1,
= 0,
have, in like manner, this other derived vector,
VIII.
(3.)
As an example
..p'=D,p = |v.aî/3.
of a vector-function of
more scalars than
one, let us
resume
the expression (308, XVIII.),
IX.
in
which we
now suppose
shall
p =rk*j^kj-sk-i',
that the tensor r
is
given, so that p
is
the variable
upon a given spheric surface, of which the radius is r, and the cenat the origin while s and t are two independent scalar variables, with respect
vector of a point
tre is
to which the two partial derivatives ofthe vector p are to be determined.

(4.) The derivation relatively to t is easy ; for, since ijk are vector-units (295),
and since we have generally, by 333, VIII.,
X.
if
Ta =
1,
and therefore
d.a*=-a'îda;,
and
if
x be any
scalar function of
XII.
D,p =
^ {kp
t,
XI.
we may
pk)
D( a* =
.
a^' Dfa?,
write, at once,
= 7rVA;p
by 279,
IV.,
and we see that

XIII. ..SpDip
= 0,
a result which was to be expected, on account of the equation,
XIV.
which
.p2
+ r2=0,
by 308, XXIV., from the recent expression IX. for p.

C5.) To form an expression of about the same degree of simplicity,
follows,
partial derivative of p,
(its scalar
vanishing)
we may
hence
observe thatys+U^"-*
is
equal to
its
for the other
own
vector part
VECTOR FUNCTIONS, DERIVED VECTORS.
CHAl'. II.]
XV.
by
D,p
-nk^jkip
XVI.
or
thus the equation,
XVII.
which
is
XIV. with
The
(6.)
must be perpendicular
two
other fo7tns for the

.
0,
s.
Dsp must be a vector
by XV.
hence,
or
XVI., p
under
by the expression 315, XII. for p. In fact that

and (4.), and by the recent values XII. XVI., these
easily confirmed
is
SpD,p =
we have
is zero,
to the vector k*jk'^, or k^^j, QXJk-^^\ a result which,
expression gives, by 315, (3.)
XVIII.
the scalar of jk-*^
respect to the variable scalar
partial derivative
the last form,
= irk'^^jp = Trjk'itp,
D^p
XIIL, and might have been otherwise obtained, by taking the
analogous to
derivative of
And because
the transformation 308, (11.).
419
partial derivatives of
= Trrk^ty.f-^',
Dip
XIX.
which have been above considered
jo,
D,p
= 7rr(A;2V.i2*î -Y.k^);
which might have been immediately obtained, by partial derivations, from the expression 315, XII. itself, and of which both are vector-forms.
And
(7.)
XIII.,
these
new forms for, the
partial derivatives of
= 7rr (j cos trc - i sin ttr^ sin sir

=
sin <7r) cos sir k sin sn
7rr
D,p
{ (i cos tir +j
}
XX.
XXI.
by derivating the expanded expression 315,
or immediately
hence,
we obtain
D(p
We may
add that not only is the variable vector p perpendicular to each of

(8.)
the two derived vectors, Dsp and D<p, but also they are perpendicular to each other ;
for
we may
by XII. and XVI.,
write,
XXII..
S(Dsp.Dp) = -5r2S.A;2'yp2A
A vector may
(9.)
bles,
such as
vector of
r, s,
may
or rather
= 7r2r2S.A2î=0;
be drawn from the expressions XX. and XXI.

be considered as a function of three independent scalar varia-
and the same conclusion
it
must be
an arbitrary point of space
so considered, if
and then
it
will
it is
have a
to
admit of being the
total differential
(329)
of the trinomial form,
XXIII.
and
dp
= drp + dsp4-dtp = Drp.dr + D*p.ds+D<p.d<;
have three* partial derivatives.

(10.) For example, when p has the expression IX., we have
will thus
derivative,
XXIV..
which may
D,p = r-ip = Up,
be thus more fully written (comp. again 315, XIIL),
also
XXV.
and we
this third partial
Drp =
k^j^kj'^k'^
= (i cos tTT -\-j sin tir') sin stt +
A-
cos
tt
see that the three derived vectors,
XXVI.
Drp,
D^,
D,p,
compose here a rectangular system.
* That
is to
say, three
of the first order ;
sider successive differentials,
for
we
of functions of one or
shall soon
more
have occasion
variables,
and so
to
con-
shall be
conducted to the consideration of orders of differentials and derivatives, higher than

the first.
420
Section
5,
On
Successive Differentials
[bOOK
III.
and Developments,
of Functions of Quaternions.
338. There will
now be no
difficulty in the successive dif-
ferentiation, total or partial, of functions of one or more quaternions ; and such differentiation will be found to be useful,
as in the ordinary calculus, in connexion with developments of
functions : besides that it is necessary for many of those geo-
and physical applications of differentials of quaterwhich we have not entered yet. A few examples of
successive differentiation may serve to show, more easily than
any general precepts, the nature and effects of the operation
and we shall begin, for simplicity, with explicit functions ofone
metrical
nions, on
quaternion variable.
(1.) Take then the square, q^, of a quaternion, as a function /7, which is to be
twice differentiated.
saw, in 324, VII., that a first differentiation gave the
We
equation,
I.
we
but
are
now
of the function
d/9
= d.j3 = 5.dg + dg.5;
to differentiate again, in order to
form the second differential d^fq
treating the differential of the variable q as
q\
still
equal to d^, and
a new arbitrary quaternion, of which the

And thus we get, in general, tiiis
tensor, Tds^, need not be small (corap. 322).
twice differentiated expression, or differential of the second order,
in general writing dcl^
II.
(2.)
324,
d^y,
where
dflfq
= d2. 52 = g,a^q +
XL),
III.
(0
2d^2
d^g. q.
differential of the reciprocal of a quaternion
The second
(3.) If
di-q is
d^.q-^
is
generally (comp.
= 2{q-idqYq--^-q-^d^q.q~K
be a variable vector, then (comp. 336, (1.)) we have, for the
second differentials of its square, the expressions
IV.
d p2
.
= 2Spdp
V.
d2. p^
2Spd'2p
dfp =2Svdp, when v
is
in

(5.)
may be
= 2Svd2p +
d2/p
briefly,
(6.)
As an example
dVpdp = Vpd2p
because
Sdvdp, instead of S(di/.dp).

will be found useful, in the
motions, performed under the influence of central forces

.
(comp. again
expressed as follows
2Sdi'dp
The following very simple equation

VIII.
if
another variable vector,
then the second differential of the same function

VII.
and
+ 2dp2.
any other scalar function of a variable vector p, and

the sub-articles to 336) its first differential be put under the form,
(4.) If /p be
VI.
first
theorj' of
V. dp^ =
0.
of the second difterential of a quaternion, considered as a
CHAP.
SUCCESSIVE DIFFERENTIALS, DEVELOPMENTS.
II.]
421
fnnction of a scalar variable (comp. 333, VIII., and 337, (l.))j the following may
be assigned, in which a denotes a given unit line, so that a^ = 1, da = 0, but x
a variable scalar
is
IX.
d2.a'=df
j= |aîd3a;-(|
|a*^da;
(7.) Tlae second dilTerential oi i\xQ
product
be expressed as follows (comp. II.)
may
X.
The second
339.
= d2/gr
d* (/g 09)
oi
]a*da;2.
any two functions oi
a.
quaternion
+ Ufq
<l>q
^q +fq
d^q.
differential, d^q, of the variable
quaternion
q^
enters generally (as has been seen) into the expression of the second
differential d^fq, of the function /', as a new and arbitrary quaternion: but, for that very reason, it \q permitted^ and it is frequently
found
be convenient, to assume that this second differential d^^
to
equal to zero
ferential
And
constant.
is
dq
what comes
or,
I.
=
=
d^q 0,
d^ constant, or V.
of
course more simple,
dYq become
.
the expressions for
is
same thing, that the fast difwhen we make this new supposition,
to the
as in the
following examples.
With
(1.)
tials,
this last supposition,
of the square
IL
(2.)
d2.g2
Agam,
= 2d52;
m.
we suppose
if
that
V.
dfq
= codq
ciqc2
in writing which, the points*
The first
(3.)
and 325,
(2.)
),
we now suppose
I.
/^ =
or
may
any three constant quaternions, and
Coqciqcz,
its first
VI.
and second
dfifq
differentials are,
= 2codq
cidg-
C2
be omitted.
remaining still entirely ariiVrary (comp. 322, (8.),

no supposition is made that its tensor Tdq is small, although
differential, dq,
so that
this differential
an equation which
dq
may
to be constant (I.)
in like
we have
rigorously,
+ d^)2=j2 + d.j2 + ^d2.52;
be also written thus,
VIII. ..(^+dg)2
And
the following second differen:
= 2(5-id5)25-i = 29-Kd?.^-0'.
that
1'.,
+ coqcidq. c^
VII. ..(^
(4.)
we have
are
co, ci, C2
IV.
under the same condition
find,
I'.,
d2.9-i
take the function,
we
or
I.
and the reciprocal of a quaternion
manner we
shall have,
(l
+ d + ^d2).92.
more generally, under the same condi-
tion of constancy of dq, the equation,
IX.
if
the function /^ be the
sum
../(^+dg) = (l4-d+id2)/5,
of
any number of monomes, each separately of the ybrm
Compare
the second Note to page 399.
422
[bOOK
111.
IV., and therefore each rational^ integral, and homogeneous of the second dimension,
with respect to the variable quaternion, q or of such monomes, combined with others
of the first dimension, and with constant terms : that is, if oq, bo, bi, b'o, 6'i, . . and
;
Co, ci, C2, c'o, c'l, c'l,
be ani/ constant quaternions, and
X.
fq = ao+ ^boqbi + ^coqciqcz.
is easy to carry on the operation of differentiating, to the

and higher orders ; remembering only that if, in any former
we have denoted t\\Q first differentials of q, d^',
by d^", d^^,
340. It
third
stage,
we then
continue so to denote them, in every subsequent stage of the

successive differentiation : and that if we find it convenient to treat
any one
we must then
differential as constant^
treat all its successive
few examples may be given, chiefly

with a view to the extension of the recent formula 339, IX., for the
differentials as vanishing.
function f(q + dq) of a sum^ of any two quaternions, q and dq, to polynomial forms, of dimensions higher than the second.
The
(1.)
third differential of a square

I.
More
(2.)
tiojis
d3 g2
More
(3.)
=^
d^j
+ d'g
is
gr
generally (comp. 338, II.),
+ Z(Aq
generally, the *Atrc differential oi
(comp. 338, X,)

II.
may
be thus expressed
d3(/g.05) = d3/g.0g
generally
A^q
di^q
+ 3d2/g.d0g + 3d/9.d2^5+/5.d^0g.
the n'^ differential of a product
still,
dg).
product of two quaternion func-
a.
as in the ordinary
is,
calculus,
III.
d'(/5.^g)
= d/?.^5 + d'-/9.d^g + 3d-2/5.d205+

n(-l)(n-2)
n(n-l)
the only thing peculiar to quaternions being, that

rally) the order
in each
of the factors,
we
IV.
Hence
d.r<7
= dr.5 +
+
V...(^l+d
(l + d +
^d^+J3d3
|d4-^d3+..
(6.) Hence,
by 339, (4),
the equation (comp. 316,
it is
if
nd-ir.dg,
under this condition that dq

(5.)
stant quaternion, we have the transformation,
also,
+ ... +
is
+fq.6'*(pq,
are obliged to retain (gene-
term of this expansion
the function
(4.) Hence, in particular, denoting briefly
<pq to q,
III.
r,
and changing
0.
if
c be
any other con-
fq by
d^^
constant,
--^^dnyr,c
+ ^g>'
2.3...(-l/ ''"'y-^g
easy to infer that if we interpret
=
'^
^"'
= ^-
the symbol
t'^
by
I.),
VI.
-1
d + ^d2
+ -i- d5+
2 o
&c.,
that
is,
if
we
interpret this other
symbol
i'fq, as concisely
denoting the series which
CHAP.
FUNCTIONS WHICH VANISH TOGETHER.
II.]
423
and if the
is formed from fq, by operating on it with this symbolic development
function fq, thus operated on, be any finite polynome, involving (like the expression
339, X.) no fractional nor negative exponents ; we may then write, as an extension
;
of a recent equation (339, IX.), the formula
if
VIL..d/g=/(5+d9),
which
is
d2g
= 0;
here a perfectly rigorous one, all the terms of this expansion for
2i
function
of a sum of two quaternions, q and dg^, becoming separately equal to zero, as soon as
the symbolic exponent of d becomes greater than the dimension of the polynome.
(7.)
We shall
soon see that there is a sense, in which this exponential transform

be extended, to other functional forms which are not composed as
and that thus an analogue of Taylor's Theorem can be established for Qua-
mationYll.
above
ternions.
may
Meanwhile
it
may
VIII.
which
t^fq
be observed that by changing dg to

write the formula as follows
we may
expansion obtained as above,
=f(q + Aq)
=t
+ A)fq,
(1
Ag', in i\\Q finite
or briefly,
IX.
<1
= 1+ A
be operated on, or transformed, as in the usual

calculus of differences and differentials.
For instance, it being understood that we
treat A^q as well as d^q as vanishing, we have thus (for any positive and whole exlast symbolical equation
may
ponent m), the two following transformations of IX.,
X.
A*"
= (d _
and
!)",
XI.
d"
= (log(l + A)>;
the results of operating, with the symbols thus equated, on
any polynomial function

always j^wife expansions, which are rigorously
fq, of the kind above described, being
equal to each other.
341. Let
Fx and ^x
be
I.
Then
loith that variable ; so
iÔ
two functions of a scalar va-
a7iy
riable, of which both vanish

satisfy the two conditions,
0,
00 =
that they
0.
the three simultaneous values^

II.
X,
Fx,
(px,
of the variable and the two functions, are at the same time
(comp. 320, 321) three simultaneous differences, as compared
with this other system of three simultaneous values,

III.
If,
0,
2^0,
00.
then, any equimultiples,
IV.
nx,
nFx,
ncpx,
can be made, by any suitable ina'case

combined with a decrease of the variable, x,
of the three values
of the number, n,
II.,
to tend together to any system of limits, those limits must (by

the definition in 320, compare again 321) admit of being con-
sidered as a system of simultaneous differentials^
424
V.
dFx,
dx,
[boOK
III.
dcjiX,
answering to the system of initial values III. ; and must be

proportional to the ultimate values of the connected system of
derivatives,
VI.
F'Xy
1,
We may therefore write,

lues of the
VII.
two
tends to zero.
as expressions for those ultimate va-
last derived functions,
FO = lim. nF-,
^'0
7i==o
And
when x
(ji'xy
= lim. w^ -,
n
n==D
iÔ
if
= 00 =
0.
even if these last values vanish,- ox if the îô new condi-
tions
VIII.
i^'O
0,
0,
are satisfied, so that x, F'x, and <^x are now (comp. II.) awîf?
system of simultaneous differences, we may 5^27/ establish the
following equation of limits of quotients, which

of these last conditions VIII.,
lyi.
if
.\\m{Fx:(px)=\\m{Fx'.^'x),
=
x
a:=o
is
independent
jR) =
0O = O;
being understood that, in certain cases, these tivo quotients

may 6oÂ vanish with x ; or may tend together to in/lnity, when
it
tends, as before, ô zero.
confirm it by a
(1.) This theorem is so important, that it will not be useless to
geometrical illustration, which may at the same time serve for a geometrical proof;
at least for the extensive case where both the functions fx and ^x are of scalar forms,
and consequently may be represented, or constructed, by the corresponding ordinates, XY and XZ (or ordinates answering to one common abscissa OX), of two
curves OyY and OzZ, which are in one plane, and set out from (or pass through)
We shall afterwards see that
one common origin O, as in the annexed Figure 75.
the result, so obtained, can be extended to quaternion functions.
(2.) Suppose then, first, that the ordinates of these two curves are proportional,
or that they bear to each other one fixed and constant ratio-, so that the equation,
X.
is satisfied
sponding ordinates
first curve, to
secant tX,
XY XZ = xy xz,
OX and Ox, however great or
:
for evert/ pair of abscisses,
we
may
be.
meet the axis of

see at once
Prolonging then
abscissae in
(if
some point
t,
and so
to
Yy
of the
determine a sub-
(by similar triangles) that the corresponding chord Zz of

same axis in the same point, t and therefore that
thes econd curve will meet the

it will
small the corre-
necessary) the chord
determine (rigorously) the same subsecant, tX.
CHAP.
GEOMETRICAL ILLUSTRATIONS.
II.]
425
if the point x be conceived to

approach to X, so that the secant Xyt
curve tends to coincide with the tangent YT to that curve at the pohit Y,
the secant Zzt of the second curve must tend to
(3.) Hence,
of
\^^Q
first
comcide with the line ZT, which line therefore

must be the tangent to that second curve or in
:
other words, corresponding suhtangents coincide,

and of course are equal, under the supposed coti-
ditionX., of a constant proportionality ofordinates.
(4.) Suppose next that corresponding ordi-
nates only tend to bear a given or constant ratio

to each other ; or that their (now) variable ratio
when
tends to a given or fixed limit,
mon
i\iQ
abscissa
point
is
the
com-
indefinitely diminished, or
when
tends to
and
let
be
still
^^-
the
variable point in which the tangent to the first curve at

line
TX
curve at
some
is still th.Q
Avill
Then the corresponding tangent
first suhtangent.
so that the
to the second
not in general pass through the point T, but will meet the axis in
But the ratio of the two corresponding suhtangents,
and
TX
different point U.
UX, which had

was
''^*
Y meets the axiis,
been a ratio
c\i
equality,
now
satisfied rigorously, will
when
the condition oi proportionality
we shall
at least tend to such a ratio; so that
X.
have,
under this new condition, of tendency to proportionality of ordinates, the limiting

equation,
XL
whence the equation IX.
XII.
(5.)
"We might
also
lim (tan
(TX UX) =
lim
under the geometrical form,
results,
XTY
have observed
tan
that,
XUZ) = lim (XY

when
XZ).
the proportion X.
is
rigorous, cor-
responding areas* (such as xXYy and xXZz) of the two curves are then exactly in
the given ratio of the ordinates ; so that this other equation, or proportion,
XIIL
is
Hence
then also rigorous.
OXYyO OXZ2O = XY
:
if
we
XZ,
only suppose, as in (4.), that the ordinates tend
some fixed limiting ratio, the areas must tend to the same ; so that ty* the second
member of the equation IX. have any definite value, as a limit, the first member
must have the same : whereas the recent proof, hy suhtangents, served rather to
show that if the first (or left hand) limit in IX. existed, then the second limit in
to
that equation existed also, and

(6.)
If the function
Fx
was equal to the

be a quaternion, we
XIV.
where W, X, Y,
Z are four
..
(by 221) express
it
as follows,
Fx=JV+ iX + jF+kZ,
scalar functions oi x, of which each separately can be
Compare the Fourth Lemma
cially its Corollary, in
first,
may
of the First
Book
of the Principia
which the reasoning of the present sub -article
ticipated.
3t
is
and see
espe-
virtually an-
426
constructed, as the ordinate of a plane curve
quaternion function reduces

(7.)
And
if
and the recent geometrical* reasoning

Fx: which
and therefore
will thus apply to each of them,
to their linear combination
W= 0.
a vector function of x, when
itself to
were another quaternion or vector function,
y\jx
we might
sub-
first
Fx, and then eliyninate the scalar function 0.r so that a limiting equation of the /orm IX. may thus be proved to hold good, when hoik the functions com-
stitute it for
pared are vectors, or quaternions, supposed
The general
(8.)
to vanish with x.
still
considerations, however, on Avhich the equation IX.
more simple and
established, appear to be
direct
and
in like manner, this other but analogous equation, in which

derivatives,
and the conditions VIII. are now supposed
XV.
lira
x=0
And
nue
so
we might
{F'x
= lim {F"x
<p'x)
lately
F"x and ^"x
are second
to be satisfied:
F'O
if
<p"x),
was
evident that they give,
it is
= 0,
<p'0
= 0.
x=0
pi-oceed, as long as successive derivatives,
of higher orders, conti-
to vanish together.
(9.) Hence, in particular, if
we take
this scalar form,
X"'
XVI..
2
which evidently gives the values,
XVII.
and
if
we suppose
<J)0
= 0,
= 0,
0'O
XVIII.
while F(")0 has any
FO = 0,
in
= 0,
F"0 =
we may then
.
lim
x=o
0('- 1)0
0WO =
= 0,
1,
such that
is
= 0,
-F'O
finite value,
XIX.
f'6
Fx
that the function
(Fx
0,
Fi^- DO
= 0,
establish this limiting equation
(px)
= FWQ
which the function Fx, and the vahie
quaternions
although they
may
F^"*)0, are here supposed to be generally

happen, in particular cases, to reduce themselves
(292) to vectors, or to scalars.
* Instead of the
equation IX., it has become usual, in modem works on the Difone of the following form (deduced from principles of La-
ferential Calculus, to give
grange)
F(x)
F'COx)
0(a;)
(pXOxy
.,
V y
rv
and 1. And a geometrical

9 denoting some proper fraction, or quantity between
which is also a geometrical proof, when the functions Fx and (px can be
illustration,
constructed (or conceived to be constructed) as the ordinates of two plane curves, is

sometimes derived from the axiom (or geometrical intuition), that the chord of any
finite and /)/ane arc must be parallel to the tangent, drawn at some point of that
finite arc.
But
this parallelism
no longer
when Hhe functions
are quaternions
a vector.
exists, in general,
and accordingly the equation
of double curvature
or even
in this
when one
when the curve
Note
of
is
them
is
one
not generally true,
is
a quaternion, or
CHAP.
Taylor's series extended to quaternions. 427
II.]
342. It will
now be
easy to extend the Exponential Transis a sense in
formation 340, VII. and to show that there

which that very important Formula,
;
which is, in fact, a known* mode of expressing the Series or

Theorem of Taylor, holds good for Quaternion Functions generally, and not merely for those functions q^finite and j)oly-
and
iiomial form, with, positive
was
which it
For let j^ and/(5' + d^)
of which neither is infinite, of
lohole exponents, for
lately deduced, in 340, (6.).
denote any two
states, or values,
any function of a quaternion; and of the

II.
&fq,
&fii,
in
d"/^,
first differentials,
which
d^-
const.,
be supposed that no one is iifinite, and that the last of

different from zero ; while all that precede it, and the
let it
them
is
functions^ and/(^ + d^') themselves, may or may not happen

to vanish.
Let the first m terms, of the exponential development of the symbol {^^ - \)fq, be denoted briefly by qx, q^,
.
qm
and
series,
may be called the remainder of the

or the correction which must be conceived to be added
sum of these
to the
denote what
let rj
m terms, in order to produce the exact value
of the difference,
III.
in
Afq ^f(g + Ag) -fq =f{q + dj) -fq
such a manner that
shall
have rigorously, by the nota-
employed, the equation,
tions
I^-
we
f{q
^q)=fq-^qi +
q2
"
+ qm +
r,,,,
where
q,
=
^.j
term qm being different from zero, but no one of the terms

Then we
infinite, by what has been above supposed.
shall prove, as a Theorem, that
this
being
*
Lacroix, for instance, in page 168 of the First Volume of his larger Treatise
on the Differential and Integral Calculus (Paris, 1810), presents the Theorem of
Taylor under the form,
dw
d^M
6?u
d*
where u denotes the value which the function u receives, when the variable x
ceives the arbitrary increment 6.x (I'accroisseraent quelconque du;).
re-
428
V.
Td^ =
remainder may
lim (Tr, Tq,) = 0,
lim.
if
III.
[bOOK
;
be made to
or in words, that the tensor of the
as
we
a
ratio
the
last term
as
small
to
the
tensor
bear
please,
of
retained, by diminishing the tensor, loithout changing the versor, of the differential (or difference) ^q. And this very general result, which will soon be seen to extend to functions of
several quaternions, is in the present Calculus that analogue

of Taylor's theorem to which we lately alluded (in 340, (7.) )
;
and it may be called, for the sake of reference,

Theorem adapted to Quaternions^
VI. .. iâ;=/(^
shall
a:dg)-/g-^d/5--^d2/^-..-^-^
F'x = d^fCq + xdq) - dVg -
F('.
because,
^ 3
= d- '/(g + xdq) - d*"" /S'

= d*f(q + xdq)
^^^_3^
d--J/5,
x,
d-i/^
*^<^ finally,
by 327, VI., and 324, IV.,
..
Df (q + xdq) = l\m. n{f(q + xdq +n-\dq)-f(q + xdq)}
in like manner,
the
mark
IX.
We
X.
have
.
of derivation
and not here on
(2.)
i)a;
Fi^')x
and
alone,
have the following successive derivatives with respect to
TIL
VIII.
Taylor
Writing
(1.)
we
^^
ar,
D3/(^
4-
= d^f(q + xdq), &C
xdq)
=df{q-\- xdq),
referring to the scalar variable x, while d operates on q
nor on dq.
by VI. and VII., the values,
therefore,
FO = 0,
F'O =
0,
we have
whence, by 341, XIX.,
F"0 =
0,
F('
1)0
FMO = d*"/?
= 0,
this limiting equation,
or
XII...lim(fx:+x) =
l,
if
x=o
(3.)
But
these
two
functions,
^r =
{;~^\
m
\z.6
.
Fx and
i//ar,
are fonned by IV. from

q,n
+ r^ aud
Qm* by changing dq to xdq; and instead of thus multiplying dq by a decreasing scatensor Tdq, without changing its versor
lar, X, we may diminish its
Vdq. We may
therefore say that,
when
this other quotient r,

tion
V. holds good.
this is done, the quotient
(q^ + rm)
q,n to zero, as its limit; or in
qm tends
to unitj/, or
other words, the limiting equa-
CHAP.
EXAMPLES OF QUATERNION DEVELOPMENT.
II.]
As an example,
(4.)
let
the function fq be the reciprocal, 5-'
= const.),
=
=
d/g
d".g-i
2.3...7n.(7-i(-r>,
429
then (comp. 339,
III.) its m" differential is (for d<7
XIIL
and
it is
XIV.
.(q
we have
= dq.q-î
rj)-!
=5-1(1
4.
r)-l
= j-{l - r + r2-. + (- r)*" + (- r)+l

.
(1
r)- };
therefore here
XV...q,n = q-'(-r)'n,
and
if
easy to prove, without differentials, that
r,
=-
q^r (1
+ ry\
this last tensor indefinitely diminishes
T(r,
= Tr
9^)
T(l +r)-l;
with Tdq, the quaternion q being sup-
posed to have some given value different from zero.
we
(5.) In general, if
XVL
.f(q
establish the following equation,
+ n-Mq)=fq + n-^dfq +
"
"'
"^d'-fq
^ 3
d">-l/y
fl-m
as a definitional extension of the equation 325,
the function /5
itself,
nor any one of
V.
and
if
we suppose
that neither
as far as dr^'^fq is infinite ; the

then be expressed by the formula,
its differentials
result contained in the limiting equation
XI.
may
XVIL ../(J)(g, d?) = d-/5;

value m = 1, if we suppress the upper
which for the particular

index, coincides with
the form 325, VIII. of the definition
dfx, but for higher values of m contains a theo-
rem: namely (when d"'/;^ is supposed neither to vanish, nor to become

what we have called Taylor's Theorem adapted to Quaternions.
infinite),
343. That very important theorem may be applied to cases, in

which a quaternion (as in 327, (o.) ), or a vector (as in 337), is expressed as Q. function of a scalar ; also to transcendental forms (333),
whenever the differentiations can be effected; and to those new
forms (334), which result from the peculiar operations of the present
Calculus itself.
A few such applications may here be given.
(1.)
I.
we
first this
Taking
.
qr=ft
transcendental and quaternion function of a variable scalar,
= a%
Ta =
with
l,
da =
0,
d<
= const.,
have, by 333, VIII., the general term,
"
d.a*
TT
dividing then t^
and convergent
in.
TTT
a<
'-=2:3^;:=
.
a'
by
and
a:a
(xa)r
7radt\^
i)
we obtain an
namely (comp. 308,
.ad<=l +
(2.) Correct
a^,
2-8^7;;
infinite series,
(4.)
"
which
is
^^-''^*'
found to be correct,
),
(xay^^
=
^-^ + ..4-2^y^-f..=* cos
+
Trdt
Trdt
asin
S(g + dq), V(^ + d^), K{q + d^), and

t^ on Sy, Yq, Kg-, and Ng; for exconstant), the third and higher differentials of Ng vanish by
finite expansions, for
^{q-Vdq'), are obtained
ample {dq being still

334, XL, and we have
aHxay^
=2:1::^
when we operate with
430
IV.
e'^-Nq
an expression
VII.
To
(3.)
+ d + id2)N<7 = Ng + 28(Kq
(1
dq) + Nd*?
and
this quotient
VI.... = Vr =
V^ = ^.
q,
V... = S.=
S^=^,
q
Then writing
of which
still
m
1]
also,
and
notation o{ factorials,
m
r"+> ==
= -S.(s-f-)2=-(s2+t;2);
ds
We have
treated as constant, the equation,
= d'wr = [d(s + v)
[- 1]
easy to separate the scalar and vector parts
it is
IX.
by
= (-!). (-2). (-3).... (-m),
by 342, XIII., dq being
VIII.
known
abridgment, as in a
also, for
VII. .. [-!]
shall have,
^q
i-q
(4.)
6q)
versor of a sum, let us again
and denote the scalar and vector parts of

by 334, XIII. and XV.,
so that,
(5.)
= N(<7 +
III.
of a sum, which agrees with 210, XX., and with 200,
for ihe 7iorin
develope, on like principles, the tensor
write r for dq
we
[bOOK
dv
for
u)'+i,
example,
= -Y.(s + vy=-2sv.
by V. and VI.,
d'"iTfl
dw'Ta
XI...'=( +
d)-^'
= ..=(.+
d>l;
the notation being such that we have, for instance, by IX.,
XII.
XIII.
(6.)
(5
(u
+ d)l=s;
+ d)l = tj;
XV.
dividing and
XVI..
s
d)2l
= ( + d)s = s'^+d* = -j;2.

= u2 + d=u2_2s.
gives, therefore,
I.,
+ dy)=dT^ = *+dl.Tg;
U(^ + d9) = t'iU^ = t'+Jl.U5;
T((7
substituting,
.
T(H-s + ) = f+di.
and V being here a scalar and a
other
The exponential formula 342,
XIV.
or,
(s
(+d)21=(t; + d)r
but of which, in the applications,
order that the series
may
XVIL
U(H-5 + t;) = i''^Jl;
which are entirely independent of each

the tensors must not be taken too large, in
vector,
converge.
(7.) Tlie symbolical expressions,
XVI. and XVII.,
two
for those
series,
may
be
developed by (4.) and (5.) thus, if we only write down the terms which do not exceed
the second dimension, with respect to s and v, we have by XII. and XIII. the deve;
lopment,
XVIII.
XIX.
JJ(1
T(l + + t;)=H-s-ir3+.
+ s + v')=l + v + Qv^ - sv-) +
..,
.;
which accordingly the product is 1 + s + v, to the same order of approximation.

a swm of tAvo quaternions can sometimes be developed, with(8.) A /wjiciftow o/
out differentials, by processes of a more algebraical character ; and when this hapof
pens,
we may compare
to quaternions in 342,
function; for example,

series 342,
XIV.,
the result with the /orm given by Taglor^s Series, as adapted

and so deduce the values of the successive differentials of the
we can
infer the expression 342, XIII. for d". q \ from the
for the reciprocal
of a sum.
CHAP.
(9.)
SUCCESSIVE DIFFERENCES AND DERIVATIVES.
II.]
And
not only
may we
verify the recent developments,
431
XVIII. and XIX., by
comparing them with the more algebraical fonns,
XX.
XXI.
but
if
also,
which are
XXII.
the
s + f)=(l + +v)i (l + s-w>,

U(l+5+t>) = (l + s-|-r)Kl + -H
.T(l +
.
of these for example (when expanded by ordinary processes^

have given us, without difi'erentials,
first
in this case applicable)
T{q f
r?')
= (1 + -
|w2
we can then
infer the values of

quaternion, as follojvs :
XXIII.
..
where
.)Tj,
tYie first
and second
dT<y-S^.T^;
^^Tq
whereof the
it,
first
= ^qq-\
and
= Yq'q-\
differentials of the tensor of a
-[^'1>'
agrees with 334, XII. or XIII., and the second can be deduced from
under the form,
(10.) In general, if we can only develope a function /(gr + q') as far as the term
which are of the^rs^ dimension relatively to ', we shall still obtain thus
or terms
an expression for the^rs^ differential dfq, by merely writing dq in the place of q'.
But we have not chosen (comp. 100, (14.) ) to regard this property of the differential of a function as the fundamental one, or to adopt it as the definition of dfq
be;
we have not
chosen to postulate the general possibility of such developments of

functions of quaternion sums, of which in fact it is in many cases difficult to discover
the laws, or even to prove the existence, except in some such way as that above excause
plained.
(11.) This opportunity may be taken to observe, that (with recent notations)
have, by VIII., the symbolical expression,
XXV.
*+'^^+n
= l + s+;
XXVI.
or
f+d
we
+ r.
344. Successive differentials are also connected with successive dif-
by laws which it is easy

words need here be said.
ferences^
a few
(1.)
We
to investigate,
and on which only
can easily prove, from the definition 324, IV. of dfq^ that
stant,
I.
dVi?
= lim 2{/(?+ 2-'dg) - 2/(^+

.
n= 00
if
dq be con-
-'dg)+/g};
with analogous expressions for differentials of higher orders.

(2.) Hence we may say (comp. 340, X.) that the successive differentials,
n...dfq,
are limits to
III.
all
simultaneously tend,
least tends to
(3.)
d^fq,
d^fq,..
for
d2^=0,
which the following multiples of successive
And
A/g,
when
become equal
hence
n2A2/^,
n3A3/5,..
the multiple
n^q
thereto, while the
we might
prove, in a
is either
number n
new way,
differences,
for
A2^ = 0,
constantly equal to dq, or at

increases indefinitely.
that if the function
f(q + d^)
432
[bOOK
III.
can be developed, in a series proceeding according io ascending and whole dimenbions

with respect to dq, the parts of this series, which are of those successive dimensions,
must follow the law expressed by Taylor's Theorem* adapted to Quaternions
(342).
345. It
may be
easy to conceive that the foregoing results
is
ex-
tended (comp. 338), to the successive differentiations of functions of

several quaternions; and that thus there arises, in each such case, a
system of successive differentials^ total and partial : as also a system of
partial derivatives^ of orders higher than the first, when a quaternion,
or a vector,
is
regarded (comp. 337) as a function of several scalars.
The general expression
(1.)
involves d^q, d2r,
but
it is
for the
second total differential,
often convenient to suppose that all these second dif-
ferentials vanish, or that the first differentials dq, dr,
d'Q, or d'"jP(^,
r,
.),
becomes a
the m^^ dimension, of those
rational, integral,
first differentials
dq, dr,
are constant ; and then
and homogeneous function of

which may (comp. 329,
.
III.) be thus denoted,

II.
d'Q
(dj
+ d;. +
.)'"
or briefly.
Q;
III.
d"*
= (dg +
dr
.)',
which symbolical power, the multinomial theorem o{ algebra may be

because we have generally, for quaternions as in the ordinary calculus,
in developing
employed
IV.
(2.)
For example,
V.
Q = rqr,
And
in general,
F{q,
r),
.
VI.
then
dAq =
drdq Q
VII.
and suppose
r',
drQ = r'qr-{
= dqdr Q = r'q'r + rqr
each of the two equated symbols IV. gives, by
jF(^, r
which the numbers n and n are supposed
We may
operation on
its
I.),
^ n-^dr)-F{jq\ n-^dq,
r)
+ F(q,
r)};
to tend to infinity.
also write, for functions of several quaternions,
X...
or briefly,
rqr
the limit of this other function, or product (comp. 344,
nn'{F(^ + n-ld^, r + n'-'dr)
(3.)
d,d,.
dqQ^rq'r;
'
in
we denote dq and dr by q and
VIII.
and
IX.
if
Q+AQ = F(q + dq,r+dr,..)=eV^^'-F(q,r)i

XL
+ A = e^-'^r'' ' =
e'^;
with interpretations and transformations analogous to those which have occurred

already, for functions of a single quaternion,
as an example of successive and partial derivation, if we resume
(4.) Finally,
the vector expression 308, XVIII. (comp. 315, XII. and XIII.), namely,
XII.
p= rktfhj'^k't,
* Some remarks on the

adaptation and proof of this important theorem will be
found in the Lectures, pages 589, &c.
CHAP.
SCALAR AND VECTOR INTEGRALS.
II.]
which has been seen
to be capable of representing the vector of
we may observe that
any point of space,
principle mentioned in
by the
gives, without trigonometry^
it
433
308, (11-)? and by the sub-articles to 315, not only the form,
XIII.
but
also, if
a be any vector
XIV.
XV.
whence
We
(5.)
rA-'/-''A;i-,
as in 308,
XIX.,
= rk*^^j^^k't = rkf(kS.a^ + iS.a*'-^).k-*;

= rV. ^25+1 + rk^fy. i2, as in 315, XII.
unit,
have therefore the following new expressions (compare the sub-articles

two partial derivatives odhe first order, of this variable vector p,
to 337), for the
taken with respect to
and
XVI.
Dsp='n-rktjHj-^k-t
= -irpk*jk'*,
with the verification, that
XVII.
pDsp =
Trr^
k^j^kph*
k'jHj-^k-t
= Trr^k^jk-i
= Trrk^tY.j^' = rrrk'^^jS a2-i = r-ipD^p S. a2-i,

XIX.
D,p Dtp = rr^rpS
pD<p = rD^p S a^'-\ and XX.
= r-ip = ktj^kj-oJc-i, as in 337, XXV.
XXI.
D,.p
and
whence
whUe
XVIII.
Dtp
we have
so that
a'î
the following ternary product of these derived vectors of the
order,
XXII.
Drp.Dsp.Dtp=
first
= -7rr2DS.a2*;
7r2p2S.a2-i
which product depends (comp. 299, (9.)) on the circumthat the vectors thus multiplied compose (337, (10.) ) a rectangular system.
the scalar character of

stance,
(6.) It
easy then to
is
XXIII.
XXIV.
(7.)
Dr2p
D,2p
The
same
three scalar variables,
r, s, t,
the expressions
= D<D,p = r-'D^p ;
D,.D<p
DrDsp = DsDrp = r-iDsp
- 7rrA2<V. i2.
=
=
D<2p =
D,D<p D<D,p Tr^rk^^Y.f^^^
;
= - 7r2p
the six partial derivatives of p, of the second
infer, for
order, taken with respect to the
three partial differentials of the first order, of the
same variable vector
p, are the following:
XXV.
d,.p
= r-ipd;-
dp
with the products,
XXVI.
XXVII.
.
d,p
drp
d<p
dsp
= Dp
6s
= - TrrpdS
dtp
7rr2dr
d^p
= Dtp
a^K dt
dS
a^'
6t
dt.
These differential vectors, d,p, d^p, dtp, are (in the present theory) generally finite ; drp, like Drp, being a line in the direction of p, or of the radius of this
(8.)
sphere round the origin, at least if dr, like r, be positive ; while d^p, like D^p, is
(comp. 100, (9.) ) a tangent to the meridian of that spheric surface, for which r
and t are constant ; but dtp, like Dtp, is on the contrary a tangent to the small circle
(or parallel^,
on the same sphere,
for
which r and
s are constant.
Treating only the radius r as constant, and writing p = op, if we pass from
the point P, or (*, <), to another point q, or (s + As, t), on the same meridian, the
(9.)
chord PQ
is
represented
from p to a point R, or
by the finite
(s, t
difference.
At), on the
Asp
sented by the other partial and finite difference, Atp
may
be denoted by
and
in like
manner,
if
same parallel, the new chord pr

;
while the point (s
we
As,
is
pass
repre-
t+ At)
s.
(10.) If now the two points Q and r be conceived to approach to p, and to come
very near it, the chords PQ and PR will very nearly coincide with the two cor-
to be
3 K
434
responding arcs of meridian and parallel
drawn
[bOOK
III.
or with the tangents to the same two cir-
have the lengths of those two arcs or finally with the differential and tangential vectors^ dsp and d^p, if we suppose (as we may, comp. 322)
that the two arbitrary and scalar differentials, d and df, are so assumed as to be
cles at P, so
as to
As and
constantly equal to the two differences,

with them.
Whether the
(11.)
like the product
A^, and consequently to diminish
and d^ be large
differentials ds
or small, the product d,p.d<p,
Dsp D^p, represents rigorously a normal vector

.
(as in
XXVI. and
which the length bears to the unit of length (comp. 281) the same ratio, as
that which the rectangle under the two perpendicular tangents, d^p and d<p, to the
XX.)
of
sphere, bears to the unit
Hence, with the recent suppositions (10.),
of area.
we
regard this product dsp . d^jO as representing, with a continually and indefinitely
increasing accuracy, even in the way of ratio, what we may call the directed element
may
of spheric surface, pqrs, considered as thus represented (or constructed) by a normal at P and the tensor of the same product, namely (by XXVI.),
;
XXVIII.
T(d,,o
dtp)
= - 7rr2dS
a^Kdt,
1, while
in which the negative sign is retained, because S . a~^ decreases from + 1 to
to 1, is an expression on the same plan for what we may call by cons increases from
trast the undirected element of spheric area, or that element considered with reference
merely to quantity, and not with reference to direction.
from < = to
(12.) Integrating, then, this last differential expression XXVIII.,
<
= 2,
and from
PQRS between
= o
these
to s
= s\,
that
is,
taking the limit of the sum of all the elements

find the following equation
bounding values, we
XXIX.
Area of Spheric Zone =
27rr2S (a^ô
- a^h)
whence
XXX.
and
Area of Spheric Cap
(s)
= 27rr2(l -S.a^) = 47rr2 (TV. a*)^
finally,
XXXI.
Area of Sphere =47rr', as
usual.
the expression XXVII., with its sign changed (on account

(13.) In like manner
of the decrease of S a^*, as in (11.) ), represents the element of volume ; and thus,
= to 5 = 1, and from t = Q to t = 1,
=
by integrating from r = ro to r ry, from s
.
we
obtain ancAV the
known
XXXII.
values
Volume of Spheric Shell
{n^
- ro^)
and
47rr3
XXXIII.
Volume of Sphere (r)=
are however only specimens of

(14.) These
tion,
although connected with quaternion forms
as usual.
what may be
;
called Scalar Integra-
audit will be more characteristic
of the present Calculus, if we apply it briefly to take the Vector Integral, or the limit
oi i\xQ directed elements (11.), of a portion of a spheric surface:
of the vector-sum
a problem which corresponds,
in hydrostatics, to calculating the resZ<an<
of the pres-
sures on that surface, each pressure having a normal direction, and a quantity prothe element of area.
portional to
For this purpose, we may employ the expression XXVI. with its sign
(15.)
in order to denote an inward normal, or a. pressure acting /row tvithout ;
changed,
and
if
we then
substitute for p its value
XV., and observe that
CHAP.
DIFFERENTIALS OF liMrLIClT FUNCTIONS.
II.]
XXXIV.
k'^'dt
aud remember that V.^^^+i
XXXV.
XXXVI.
we
a^*,
k^=~l,
easily deduce the expressions:
Sum of Directed Elements of Elementary Zone=7rr^kd (S. a2*)^

Sum of Directed Elements of Spheric Cap () = - irr-Jc (1 - (S.a-^y)
=^S
because
0,
['
435
= Trr^k (V. a2*)2 = n'^k (D<p)2 = nk (YkpY.

(16.)
But the radius
responding,
is
of (he
TVArp, so that
its
plane and circular base, of the spheric segment corarea is in quantity =-tt (Ykp^ ; and the common
inward normals is that of +

hence if we still represent the
by normals thus drawn inwards, we have this new expression
direction of all its
rected elements
i?;
di-
XXXVII.
Sum of Directed Elements of Circular Base = -7rk(Ykpy

XXXVI., we arrive at the formula,
comparing which with
XXXVIII.
a result which
may
Sum of Directed Elements of Spheric Segment = Zero ;
be greatly extended, and which evidently answers to a
known
case of equilibrium in hydrostatics.
(17.) These few examples may serve to show already, that Differentials of Quaternions (or of Vectors) may be applied to various geometrical and physical questions ;
and
that,
when
so applied,
it is
permitted to treat them as small,
venience be gained thereby, as in cases of integration there always
is.
if
any conBut we must
now
pass to an important investigation of another kind, with which differentials will

be found to have only a sort of indirect or suggestive connexion.
Section
6.
On
the Differentiation
of Implicit Functions of
Quaternions ; and on the General Inversion of a Linear

Function, ofa Vector or a Quaternion : with some connected
Investigations.
We
346.
saw, when difFerentiatlng the square-root of a
quaternion (332, (5.) and (6.) ), that it was necessary for that
purpose to resolve a linear equation,* or an equation of the
frst degree; namely the equation,

I
rr'
r'r
q\
and q represented two given quaternions, q^ and

r
while
represented a sought quaternion, namely dr or d q^.
d^',
And generally, from the linear or distributive form (327), of
in
which
the quaternion differential

11.
..dQ = d/^=/(^,d^),
of any given and explicit function fq, when considered as depending on the differential d<7 of the quaternion variable q, we
see that the return
*
from the former

Compare
the
Note
to
differential to the latter,
page 110,
436
[bOOK
III.
dQ
from
to dq, or the differentiation of the inverse or

implicit function f-^Q, requires for its accomplishment the So-
that
is
of an Equation of the First Degree : or what may be

called the Inversion of a Linear Function of a Quaternion.
are therefore led to consider here that general Problem
lution
We
to which accordingly, and to investigations connected Avith

which, we shall devote the present Section, dismissing however now the special consideration of the Differentials above
mentioned, or treating them only as Quaternions, sought or

given, of which the relations to each other are to be studied.
347.
Whatever the
tributive function^ fq,

I.
particular form of the given linear or disbe, we can always decompose it as follows:
may
./?=/(S? + V2)=/Sj+/V? = Sy./l +/V2;
taking then separately scalars and vectors, or operating with S and
V on the proposed linear equation^

../^ = r,
11.
where r
eliminate
and
is
and q a sought one, we can
a given quaternion,
Sg',
and
so reduce the
problem
in general
to the solution of a linear
vector equation^ of the form,

III.
where
o-
is
a given vector^ but p{=
<pp=
<r\
Yq)
is
a sought one, and ^
is
used
as the characteristic o^ a given linear and vector function of a vector,

which function we shall throughout suppose to be a real one, or to
involve no imaginary constants in its composition. But, to every such
function
linear
0/),
there always corresponds what
and vector function
<P'p,
may be called
connected with
it
by
a conjugate
the following
Equation of Conjugation^
IV.
S\<pp=Spf\',
where a and p are any two vectors. Assuming then, as we may, that
V are two auxiliary vectors, so chosen as to satisfy the equa/i and
tion,
and therefore
also,
VI.
SXff
SXyuv,
S/i<T
0,
Si'o'=0,
where \ is a fAtrt? auxiliary and arbitrary vector, we may (comp. 312)

replace the one vector equation III. by the three scalar equations,
CHAP.
INVERSION OF LINEAR AND VECTOR FUNCTIONS. 437
II.]
VII.
And
these give,
S/J^'A
Spf'fJL
SA/tl/,
=r-
0,
S/J^'
l^
by principles with which the reader
0.
is
supposed to
be already familiar,^ the expression,

VIII.
mp =
Yr<r,
IX.
or
p = <p-^(T =
m~^ylr<r^
an auxiliary linear and vector function,

yjr
of which the value and the form are determined by the two following
be a vector-constant^ and
if OT
equations
X.
WiSA/Ai^
XL
= S(0a.0>.0V);
Y-(VH = V(?'>.0V);
..
or briefly,
X'.
mSX/Ai/=S.^'X0>^V,
and
XI'. .. YrV/ii/
= V.^>^V.
And
thus the proposed Problem of Inversion^ of the linear and vector

function 0, may be considered to be, in all its generality, resolved;
it is
always possible so to prepare the second members of the
equations X. and XL, that they shall take the/orms indicated in the
first members of those equations.
because
(1.) For example, if we assume any three diplanar vectors

from them three other vectors /3o, /3'o, /3"o, by the equations,
XII.
/3oSaa'a"
= Ya'a",
(3'oSaa'a"
= Ya'a,
a,
a, a", and deduce
j8"oSaa'a"
= Yaa',
then ani/ vector p may, by 294, XV., be expressed as follows,
XIII.
if
we
then
XIV.
we
write,
.
/3
= (3oSap + /3'oSa'p + ^"oâ'p
= <p(3o,
i3'
= (p(3o\
(i"
= î3"o,
have the following Geyieral Expression, or Standard Trinomial Form,

a Linear and Vector Function of a Vector,
shall
XV.
containing, as
we
0p
= /3Sap + /3'Sa> + /3"Sa"p
see, three vector constants,
jS,
j3',
/3",
for
or nine scalar constants,
such as
XVI.
Sa/3, Sa'iS, Sa"/3;
Sa^S', Sa'/3',
Sa"/3'
Sa/3",
Sa(5", Sa"/3";
(and generally will) all vary, in passing from one linear and vectorfunction <Pp to another such function ; but which are all supposed to be real, and given,
which
may
for QSi.ch particular
(2.) Passing to
XV.
gives,
form of that function.

what we have called the conjugate
linear function
(p'p,
the form
by IV., the expression.
student might find
it useful,
at this stage, to read again the Sixth Section of
the preceding Chapter; or at least the early sub-articles to Art. 294, & familiar ac-
quaintance Avith which
is
presumed in the present Section.
438
XVII.
(p'p=^
aSfSp
a'S/3'p
[bOOK
a"S/3'V
III.
but
V. (aS/3/t
+ a'S/3V) = Vaa'S /3' (vSj3/i - /tS/Sv)

= Va'S ./3T. /3V|a/ = Vaa'S ./3'/3V;itv
a'S/3V) (aS/3v
and
therefore the transformation XI. succeeds,
gives,
XVIII. ..^pp=- yaa"Sf3"(3'p + Ya"aS(3^"p + YaaS(3'(3p,

as an expression for the auxiliary function
written,
XIX.
so that
is
-tp
..^P'p
changed
to
when
is
of
which the conjugate may be thus
+ V^"/3Saa"p + Yl3(3'Saap
V/3'/3"Sa'-a'p
\p',
-^
to
changed
by interchanging each
0',
of the
three alphas with the corresponding beta.

(3.) If
we
write, as in this
XX.
whole investigation we propose
= Yfiv,
X'
= YvX,
fi'
to do,
= VX/Lt,
the formulae XI. and X. become,
XXI.
i//X'
= V.
XXII.
and
(p'ix(p'v,
XXII.
of Saa'a" in this last expression
S/3XS/3"/3'X'
Sy3'XS/3/3"X'+ S/3"XS/3'/3X'
we
= S/3"/3'/3SXX',
find the expression,
XXIir.
coefficient
..m = Saa'a"S/3"/3'/3;
XXIII.
so that
and because the
by XVII. XVIII.,
is
the division by SXX', or by SX/iV, succeeds, and
which may
= S. 0'Xi|/X',
thSXX'
with the same sort of abridgment of notation as in XI'.
does not change
when we
m = S/3/3'/3"Sa"a'a,
pass from
<p
to
on which account we
<p',
may
write also,
XXIV.
because,
by
(2.),
niSXX'
XXV.
(4.)
We ought
its
.. xP'X'
and accordingly,
mSX/tv
S.cpXtpfxfpv,
= Y.<pn<pp.
= ^p = )8Sap + /3'Sa'p + /3"Sa"p,
(T
solution expressed (comp. VIII.)
XXVII.
the conjugate expression,
then to find that the linear equation,
XXVI.
has
XXI V.
or
S.ipX-sp'X',
we can deduce from XI.
by the formula,
pSaa'a"S/3"/3'/3=Va'a"S/3"/3V+ Va"aS/3/3"<r + Vaa'S/3'/3(7;

if
we
operate on the expression
XXVI.
for
<r
with the three sym-
bols,
XXVIII.
we
(3"(3\
/3/3",
(S'I3,
obtain the three scalar equations,
XXIX.
S/3"/3V = S/3"/3'i8Sa|0, &c.,
from which the equation XXVII. follows immediately, without any introduction of
the auxiliary vectors X,
/u,
v,
although these are useful in the theory generally.

XXVII. were given^ and the value of <t sought,
(5.) Conversely, if the equation
we might
operate with the three symbols,
XXX. ..S. a,
and
for
so obtain the three scalar equations
a would follow.
S./3,
S.y,
XXIX., from which
the expression
XXVI.
CHAP.
STANDARD TRINOMIAL FORM.
II.]
an
(6.) It will be found
on formula} of
useful check
439
this sort, to consider
each beta^
what we have called the Standard Form (1.) of 0|O, as being oHheJirst dimension ;
for then we may say that
and <p' are also of i]\& first dimension, but ;// and ;//' of
in
the second, and m of the third; and every formula, into which these symbols enter,
will thus be homogeneous: a, a',
a", and X, /x, v, p, being not counted, in this mode
a being treated as of the /rs< dimension, when
of estimating rfwe7i07i*, but
taken as representing
And although the trinomial form XV. has been

we choose to take the more expanded form,
(7.)
yet
ral,
it is
(pp.
seen to be svfficientli/ gene-
if
XXXI.
.^p
XXXII.
which gives
2/3Sap,
f p = 2S/3|0,
any number of terms of ^p, such as /SSap, (3'Sa'p, &c being now included in the
sum 2, there is no diflSculty in proving that the equations VIII. and IX. are satis,
fied,
when we
Avrite,
XXXIII.
.^p = 2 Vaa'S/3'/3p,
XXXIV.
with
;|/'p
= 2V/3/3'Sa'ap,
and
XXXV.
(8.)
own
2Saa'a"S/3"/3'/3=2S]3i3'/3"Sa"a'a.
The important property
when
(2.), that the auxiliary function \p is
changed
to its
be proved without any reference to

the form 2/3Sap of 0p, by means of the definitions IV. and XI., of
0' and ;//, as folconjugate
xj^
(p
is
Whatever four vectors
lows.
XXXVI.
v, fxi,
//,
X'l = Yixivi,
to 0',
changed
may
and vi
may be,
XXXVII.
and
if
.
we
write
rP'Yfxv
= V. 0ju0i/,
adopting here this last equation as a definition of the function i//', we may proceed to
prove that it is conjugate to ip, by observing that we have the transformations,
XXXVIII.
S\'ii//'\'=S(V/iivi.V.0/x0v)=S./ii(V.i/iV.0/i0v)
S/ii0/i Sjî^v
Sfltp'fXi Spfp'vi
= B.fi (V. j/V. (p'fii<p'vi) = S
(V/f v.V.
= S/ii0v
= S/Kp'Vi
Sj/10/i
Sv^'fii
which establish the relation in question, between
xjj
and
(p'fXKp'vi)
= SX'xl^X'i
xp'.
(9.) And the not less important property (3.), that m remains unchanged when
we pass from to 0', may in like manner be proved, without reference to the /orjn
XV. or XXXI. of 0p, by observing that we have by XXXVII., &c. the transfor-
mations,
XXXIX.
S 0X0/Lt0v = S
.
^W\' = SX'^^X = mSX'X = mBXfiv,
because the equations III. and VIII. give,
XL.
=mp, whatever
xp(pp
so that the value of this scalar constant
linear function 0, exactly as
348. It
is
I.
as li^ formula
identities,
may
be
be derived from the original

was in X. or X'. from the conjugate function
it
0'.
found, then, that the linear and vector equation,

.
0p =
cr,
gives
II.
mp^tfja,
of solution; with i\\Q general method, ohove ex-
plained, of deducing
two
vector p
m may now
and
\p
from
(p.
We have therefore the
440
III.
772(T
[bOOK
mp = -ifj^p
^;//(T,
III.
and symbolically,
or briefly
Iir.
7W
0;//
= ;//0
we may connect
with which, by what has been shown,
these
conjugate equations ^
Iir.
7W
Changing then successively
=
0';//'
^'^'.
and v to
fx
and
-^'ix
equation of definition of the auxiliary function

formula,
îMfiv
we
in the
t//'v,
;//,
or in the
347, Xr.,
V.^>0'v,
get these two other equations,
IV.
-^Y,
= mV.iLL(l>'v
v^P'ti
V.
xpY.xp'fixP'v
= m^Yfxv
which the points may be omitted, while in

may be exchanged with unaccented
and
we see that the law of homogeneity
of
symbols
operation
in the former of
each of them accented

:
(347,
preserved. And many other transformations of
the same sort may be made, of which the following are a few
(6.) ) is
examples.
(1.) Operating on V.
by
;p-i,
VI.
or
.
by m-i0, we get
V.;|/'/xipV
this
= m^Vjuv
new
formula,
comparing which with the lately cited definition of ;//, we see that we may change
to i//, if we at the same time change \p to m(p^ and therefore also m to w^
<p' being
;
then changed to
VII.
fi
and
<//'
For example, we
(2.)
in
;//',
which we
to m<p'.
thus pass from IV. and V. to the formulas,
may
^Yv<p'ij,=Yfi\p'v,
and
VIII.
(pY.^'jx(p'v
=mYfii'
law of homogeneity is still observed.

(3.) The equation VII. might have been otherwise obtained, by interchanging
and V in IV., and operating with - m-'^, or with - ;//-! ; and the formula VIII.
see that the lately cited
be at once deduced from the equation of definition of -^j by operating on it

In fact, our rule of inversion, of the linear function <p,
0.
may be said to be
contained in the formula,
may
with
IX.
where
(4.)
is
^'iV/ii'
= m-iV.^'/i0V;
a scalar constant, as above.
By
similar operations
X.
XI.
XII.
XIII.
substitutions,
(p^Y.(p'n(p'v
?n0V.
and
= m(pYinv = Y.\f/'ij^'v
= mWnv = '>pY.\p'i^\l/'v
m2V. (py^'v =TO^-^V/i V

V.
(l>'iJ,^'v
<p'^fi(p'î^
= -il/W. ^'^\}/'v
= xpY. (b'fx(p'v = -^^Yîv
&c.
CHAP.
(5.)
SECOND FUNCTIONS, QUATERNION CONSTANTS. 441
II.]
But we have
also,
XIV.
90 that the second functions
XIII.,
formed from
\p' is
S X^2p
.
and
<p^
as
02^
=s
may
to
p0'^\,
;//
to 0,
we change
at the
(S
= SX/ii/S,.
fpk<pn(f>v)^
^\y\;n-^v
same time
0' to
to the equation,
other analogous formulae.
many
349.
conjugate (compare 347, IV.) ; hence, by

and generally it will be found, that if
from
;
remarked that the changes (1.) conduct
also be
XV.
and
=S
^'2 are
n be ant/ whole number, and if we change

to i//", ^' to ;//'", and m to m".
0'", ^
(6.) It
(pp<p'\
The
expressions,
X 0X +
!Ji(l>ix
+ v'^v,
with the significations 347,
X^//X
XX.
of
X',
ju'i/'ju
v'\pv
y, and others of the
ju',
same type, are easily proved to vanish when X, ju, v are complanar, and therefore to be divisible by SX/xv, since each such
expression involves each of the three auxiliary vectors X, û, v
in the^r^^ degree only
the quotients of such divisions being
;
therefore certain constant quaternions, independent of X, /i, v,

and depending only on the particular form of 0, or on the
but
(scalar or vector,
real) constants,
composition of that given function.

I.
and
II.
which enter into the
Writing, then,
= (X'^X + ^ 0/u + v<pv) SX/zr,

= (X';/^X + /x'\//,tt + v'^v) SX^v,
^2
:
^1
.
we
shall find it useful to consider

separately the scalar and
vector parts of these two quaternion constants, q^ and q^'t
which constants are, respectively, of i\iQ Jirst and second dimensions, in a sense lately explained.
(1.) Since
and qi change
hand,
VX'0\ = /[tSv0\- i^SX^'yu, &c., it follows that the

signs, when f is changed to <p\ and therefore i// to
we may change
the arbitrary vectors X,
/x,
v to
X',
/*',
v,
vector parts of
^'.
if
we
On
at the
time change X' to V/iV, or to XSX/xi^, &c., and SX/xi', or SXX', to -(SX/ij^)2
SX/iv, we find these new expressions,
viding then by
III.
IV.
operating on which by S,
q\
5^3
we
= (kcpX + fupii + i^0v')

= (Xt//X' + jin///i' + v^v')
^'i
the other
same
;
di-
SX/iv,
:
SXjuv
return to the scalars of the expressions
I.
and
II.,
with
changed to 0' and ;//'.

(2.) Hence the conjugate quaternion constants, Kqi and Kqo, are obtained by
passing to the conjugate linear functions ; and thus we may write,
and
;//
3 L
442
or,
V.
VI.
Kg, = (\y\ + /i>'/i + v'ip'v) SXfiv

Kqz = (X'^p'\ + fi'4/'fi + v'rp'v) SX/iv
:
where
= Yfiv,
X'
X', &c., in the dividends,
.Kqi = (X0'X' +
VII.
VIII.
interchanging \ with
III.
[bOOK
Kq2=
fKp'fi
+ v<pV)
SX/iv
(Xi//>'+ fi^'fi'+ v^p'v)
SX^v
&c., as before.
V.p on Vgi, and observing
(3.) Operating with
while
&c.,
+ /iS/i'p + vSv'p) = 0pSX/i v,

+ fiSfKp'p + v'Sv^'p = 0'pSX/u v,
(XSX'p
X = 0(XSX'p)
X'SX^'p
with similar transformations
V. PV92, we find that
for
IX.
and
X.
V. pYqi =^<pp- 0'p ;

V. pYq^ = i//p ^'p.
(4.) Accordingly, since
- 0'p) = - Sp(^p - f p) = 0,
Sp(^p
^p ^'p,
the vector
do not vanish, must be a
if it
fore of the ^rm,
XI.
in
which y
some constant vector
is
XII.
<^p
^'p
so that
= 2Vyp,
we may
= 0op + V-yp,
^p
and there-
line perpendicular to p,
where the function 0oP is ' own conjugate^ or

0p and <p'p namely the part,
-write,
0010
ip'p
is
the
- Vyp,
common
self- conjugate
part of
XIII.
And we
see that,
XIV.
<Pop
= h('pp-^<l>'p)'
with this signification of y,
V(X>X +
/i'0/i
+ v>i/) = - 2ySX/iV,
or
XIV'.
V91 = - 2y
XV'.
V52 = -
while we.have, in like manner,
XV.
V(X'i//X
if
As a
0X
+ f^^f/f^ + v'4>v) = - 25SX/i/, or

= 2V^p.
XVI.
//p-i//'p
confirmation, the part 0o of
to
VyX,
(XSyX' +
juSy/i'
<p
has by (1.) no
member
&c., in the first
of
eficct
XIV., we have
on Yq\
and
C4.),
XVII.
we change
or
XIX.
and
XX.
(1.),
= - 2yS\fjiv,
fitpfi' + v(pv')
+ fi^fi' + v^j/v') = - 2SS\fiv,
= + 2ySX/iv,
Y(X(p'X' + fKp'fi' + j/^V)
Y(X'^'X fi'\l/'fi + v'^'v) = + 2dSXfiv,
.
this relation
if
+ vSyv') - yS (XX' + fiix' 4- vv) = ySXfiv ~ 3ySX/i v.

- ^VX^'X', &c., by 348, VII., while we may write, by
XVIII.
we have
and
thus,
=
(5.) Since VX';//'\
(2.),
2^,
Y(\<pX'
-{-
V(Xi|/X'
-\-
between the two new vector constants,
XXI.
= -fy = -<p'y=-<poy;
and ô have all the same effect, on this particular vector, y.

(6.) "We may add that the vector constant y is of the^rs^ dimension, and that
S is of the second dimension, with respect to the betas of the standard form ; in fact,
for 0, 0',
with that ^rm, 347, XV,, of fp,
we have
the expressions,
CHAP.
SYMBOLIC AND CUBIC EQUATION.
II.]
y = i V(/3a + fi'a + ^"a"%

= iV(V/3'i3".Va'a"+ V/3"/3.Va"a + V;3/3'.Vaa').
XXII.
XXIII.
and
(7.) If
we
we
443
denote by
and
i//o
wiq,
what
i^
w become when
and
is
changed
to 0o,
easily find that
XXIV.
i//p
= //op - ySyp + V^p
XXV.
so that the se{/*-con/â?e /7or* of ^p contains
vector y, but only in the second degree;
XXVI.
<p
to 0', or from,
to
the
in like manner,
m = Jo+Sy5 = wio-Sy0y;
m remains unchanged, when we pass
y.
we have
(8.) It is evident that
XXVII.
and
an even degree^ because
y again entering only in
from
and
= i//op - ySyp - V^p

ySyp, which involves
..//>
a term,
the relations,
.
Wo =
^0*^0
that, in a sense already explained, ô, V'o,
'/'oô
and mo are of the^ri^, second^ and
third dimensions, respectively.
350. After thus considering the vector parts of the two

quaternion constants, q^ and qz, we proceed to consider their
; which will introduce tioo new scalar constants,
m" and m\ and will lead to the employment of two new conjualso will result
gate auxiliary functions, xp ^^^ XP whence
scalar parts
>
the establishment of a certain Symbolic

I.
m-
m'(p
and Cubic Equation,

-
m"(p'^
<^^
is satisfied by the Linear Symbol of Operation, tp, and

of great importance in this whole Theory of Linear Func-
which
is
tions.
(1.) Writing, then,

II.
we
w" = Sî,
and
III.
m'
85-2,
two new constants changes value, when we pass from

because, in such a passage, it has been seen that we only
see first that neither of these
to <p', or from y to
y
change qi and 73 to Kqi and Kqz. Accordingly, if we denote by w'q and m"o what
m' and m" become, when
is changed to <po, we easily find the expressions,
(p
IV.
(2.) It
may
m" = m"o
be noted that m", or
V.
and
to"o,
is
m' =
m'o-
w'o are of the second, with respect to the standard form of

that form
we
have,
VII.
(3.)
If
we
introduce two
new
IX.
linear functions,
.
xp and
x'p, such that
X Vjuv = V(;xf r i^f ^),

-
x'V/ui'= V(/i0v
t'ffi),
and
and accordingly, with
VIII.
and
m" = Sa(3 + Sa'/3' + Sa"/3"

VI.
m'= S (Va'a". V/3"/3' + Va"a.V/3|3"+ Vaa'. V/3'/3).
.
and
y^.
of the^rs^ dimension, but that m'
444
[bOOK
III.
easily proved that these functions are conjugate to each other, and that each is
of the first dimension; in fact, with the standard form of <pp, we have the expres-
it is
sions,
X.
XI.
and S
xp = V(aV/3p + a'V/3> + a"V/3,

= V(/3Vap 4- (i'Ya'p + /S" Va"p)
x'P
XaY (3p = S p(BYa\

.
&c.
xo be formed from
if
Also,
(po,
as
x from
^,
it -will
be found that
XII.
where xo
is
XP = Xop- VyP,
..
of the
first
and
XIII.
x'P
= XoP + Yyp
dimension.
(4.) Since
= S.\(/i0V r^'/t) = S(/i'0'/u + v'(p'v},
SXxX'
the expression II. gives, by 349, V., the equation,
XIV.
X and
m"S\X' = S. X(0 + x)X',
being two arbitrary and independent vectors
X'
having the functional
which can only
be,
by our
relation,
XV.
00
or briefly and symbolically,
XVI.
+ xp =^ '"p
x + = m".
Accordingly it is evident that the relation XV. is verified, by the form X. of xp,
combined with the standard form of ^p, and with the expression VI. for tlie constant m".
The formula XVI.
(5.)
gives,
XVII.
and accordingly the
changing ft and v to
i//'/a
V. xp'ftxl/'v =
we
for thus
x0 = ^" and remembering that
Yp V = - <pYv(p'fi
and
;//
have,
XVIII.
x^V/*!^
= Y(fi\f/'v - v^'p) = 0V(/A0'v -
vtp'p)
ipxYpv,
as required.
(6.) Since, then,
S X0x^' = S
.
the value III. of m' gives,
XIX.
X and
X' being
\(p,\p'v
v-^'p)
S(p'\p'p,
+ v'yp'v),
by 349, VI., the equation,

.
w'SXX' =S.\(\p+ <px)
independent vectors
XX.
X',
hence,
.ypp^<pxp
= rn'p,
or briefly,
XXI.
And
in fact, with the standard
XXII.
which
i//
+ 0x =^
fonn of 0p,
wê
0XP = X0P = V(V/3'/3". Vp Va'a" +
verifies the
equation XX., when
with the expression 347, XVIII.
for
it is
(pxp
=ip= m,
we
V/3"/3
.YpYa"a +
V/3/3'. VpVaa')
combined with the value VII. of m', and
<//,o.
(7.) Eliminating the symbol x, between the
remembering that
^''
have
two equations XVI. and
find the symbolic expression,
XXL,
and
CHAP.
II.]
BINOMIAL FORM, FIXED LINES AND PLANES.

XXIII.
m^-1
= ^ = m'- m"(p + <p^
and thus the symbolic and cubic equation
I. is
(8.) And because the coefficients, m, m',

vemain unaltered, in the passage from ' +
proved.
m"0'2
write also this conjugate
^'3.
m'^xp^, and reducing

(9.) Multiplying symbolically the equation I. by
= m, we eliminate the symbol <p, and obtain this cubic in i^,
xj/if)
XXV.
in
445
which ^'
fore
= 7n2 - mm"xp + m'xp^ -
be substituted for
may
(10.) In general, it
to 7n0, as before,
i|/
;p.
be remarked, that
may
we change
t|/3
by
not only
when we change ^
m to
m"^,
to
but also m' to
ip,
and there-
mm", and m" to
m' while x is at the same time changed to (pXi o^" to x'Pi ^^^ the quaternion 91 is
changed to 52- Accordingly, we may thus pass from the relation XVI. to XXI.
and conversely, from the latter to the former.
;
And
(11.)
if
new
the two
auxiliary functions,
by the equations VIII. and IX.,
x and
x',
be considered as defined
their conjugate relation (3.) to each other
may be
proved, without any reference to the standard form of <pp, by reasonings similar to
those which were employed in 347, (8.), to establish the corresponding conjugation
of the functions
and
\p
^'.
(12.) It may be added that the relations between <p, (p', x, x'l ^^^ "*" give the
following additional transformations, which are occasionally useful
:
XXVI.
XXVII.
= Y^nxv + v(pfi) - - V(j^xi^ + J0^)

^'Vjuv
= - V(vxV + H'^''^)
0V/iv = V(/ixV + v^'\i)
;
>
with others on which we cannot here delay.
35 1
The
cubic in
I,
(f)
may be
thus written
mp - m'cpp
m'<l)^p
cji^p
where p is an arbitrary vector. If then it happen that for some

particular but actual vector, p, the linear function ^p vanishes,
so that (pp = 0, 0^|o = 0, (p^p
zero ; or in symbols,
II.
if <^p
0,
&c., the constant
0,
T^ >
and
then
0,
must be
m = 0.
Hence, by the expression 347, XXIII. for m, when the

standard form for (pp is adopted, we must have either
III.
Saa'a" = 0,
or else
IV.
S/3"/3'/3
so that, in each case, that generally trinomial form^ 347,
must admit of being reduced
we have thus
to
a binomial. Conversely,
a function of the partictdar form^
XV.,
when
446
we have
III.
[boOK
then,
VI.
(JiYaa
so that if a and a be actual and non-parallel lines, the real and

actual vector Yaa will be a value of |0, which will satisfy the
=
but no other real and actual value of p, exequation ^p
=
cept p xYaa'i will satisfy that equation, if j3 and /3' be actual,
;
and
In this case V., the operation (p reduces

non-parallel.
every other vector to the Jixed plane of j3, j3', which plane is
therefore the locus of (^p ; and since we have also,
VIL .0> = aSj3|O+a'Si3>,

.
we
see that the locus of the functionally conjugate vector, <p'p,

another Jixed plane, namely that of a, a'. Also, the normal
to the latter plane is the line which is destroyed hj the former
is
operation,
is
in like
operation,
namely by
manner the
(}>',
since
we
while the normal to the former plane

line, which is annihilated by the latter
;
have,
VIII. ,.(^'Vi3i3' =
but not
0,
any actual p, in any direction except that

which may however, for the present
In this case we have
purpose, be regarded as the same.*.
also monomial forms for "^/o and ifjp, namely
tp'p
of VjSjS'j or
IX.
0, for
its opposite,
..^f^p
= Yaa'SP'(5p,
so that the operation
plane (of
j3,
j3'),
xp
and
X.
destroys every
i//>
Vj3/3'Sa'ap
line in the first fixed
and the conjugate operation
\p'
annihilates
every line in the second fixed plane (of a, a). On the other
hand, the operation \p reduces every line, which is out of the
first plane, to the fixed direction of the normal to the second
plane; and the operation \p' reduces every line which is out of
the second plane, to that other fixed direction, which is normal
to
the^r*^ plane.
operate first with
then with
*
\p
And thus it
\p,
or first
comes to pass, that whether we

and then with
or first with (p, and
with \p' and then with cp' or first with <^',
;
Accordingly, in the present investigation, whenever we shall speak of a fixed

or the " direction of a given line," &c., we are always to be understood
" or the
of that direction."
^''
direction^''''
as meaning,
opposite
CHAP.
II.]
EQUAL ROOTS, RECTANGULAR LINES AND PLANES. 447
and then with
we
in all these cases,
ip'
arrive at last at a null
Ihie, in conformity with the symbolic equations,
XI.
<^\p
-ip^=
=
\fj(^'
<p'\p'
=w=
0,
which belong to the case here considered.

(L) Without recurring
to the
standard form of
the equation 348, VI.,
<pg.
namely Y.rf/'fiyp'v = m<l>Yfiv, and the analogous equation V. i|//i^v = ni^'Vjiti',

might have enabled us to foresee that ^'p and i//p, if they do not both constantly va-
must (ifm = 0) hare each a Jixed direction ; and therefore that each must be
by a monome, as above the fixed direction of ^f/p being that of a line
which is annihilated by the operation (p, and similarly for \^'p and <p'.
nish,
expressible
(2.)
And
by 347, XI. and XXV., we have
because,
^Yfiv = V. (p'fKp'v,
and
t//'V/iv
= V. <pfi(pv,
so that the line ^'/i, if actual, is perpendicular to \pYfiv,

lar to \l/'Y[j,v,
we
present case) a plane locus
functions, <pp
and
and the
line
(pfi
perpendicu-
of the two lines, ^'p and and ^p, must have (in the
whence the binomial forms of the two conjugate vector
see that each
0'p,
might have been foreseen
i//p
and
yj/'p
being here supposed to
be actual vectors.
lines (or directions'), to the
(3.) The relations of rectangularitt/, of the two fixed
two fixed planes, might also have been thus deduced, through the two conjugate binomial forms, V. and VII., without the previous establishment of the more general
trinomial (or standard) form of 0p.

a plane locus for 0p, and of another for 0'p, for the case
(4.) The existence of
= 0, might also have been foreseen from the equations,
when
<p\<l>fx<pv
=S
<p'X(l>'fi^'v
= mSXfiv
and the same equations might have enabled us to foresee, that the scalar constant
m mast be zero, if for any one actual vector, such as X, either (pX or p'X becomes
null.
(5.)
condition
And
the reducihility of the trinomial to the binomial form,
is satisfied,
when
might have been anticipated, without any reference
position of the constant
this last
to the
com-
m, from the simple consideration (comp. 294, (10.)), that
no actual vector p can be perpendicular, at once, to three diplanar
lines.
may happen, that besides the recent reduction

the
of
linear function cpp to a binomial form, when the
(351)
352. It
relation
I.
m=
between the cow5tow^5 of that function, in which case the

symbolic and cubic equation 350, I. reduces itself to the form,
- i
II.
+
0^
exists
thus losing
its
0^
"^2
jji'fp
absolute term, or having
o7ie
root equal to zero,
448
this equation
roots
equal to each other;
becoming
.
III.
a further reduction, by two of its

namely either by our
may undergo
having
III.
[bOOK
77i'
lY.
and
0,
<^''{^-m) =
{)',
by the existence of these other equations,

and VI.
^m'y = 0.
((^
In each of these two cases, we shall find that certain neiv geo-
or in another way,
V.
m"^ -
4m =0,
metrical relations arise, which
it
be interesting briefly to
may
investigate and of which the principal is the mutual rectangularity o? two Jixed planes, which are the loci (comp. 351) of
;
certain derived, ^ndi functionally conjugate vectors : namely, in

the case III. IV., the loci of ^p and ^'p and in the case V.
;
VI., the
loci
VII.
of
^p and
- im",
^=
if
<I'/o,
VIII.
and
60 that, in this last case, the
0)'
0'
\m",
satisfies this neiv cubic,
symbol
= ^2 (^ ^. Lm!')
while ^' satisfies at the same time a cubic equation with the
same coefficients (comp. 350, (8.)), namely
IX.
We saw

whenm =
the line i//'/o hsi?, generally ajixed

perpendicular ; and that in like manner tlie
If then the
line i//p has then another fixed direction, to which tp'p is perpendicular.
plane loci of 0p and ^'p be at right angles to each other, we must also have the
(1.)
direction, to
Jixed
lines
ill
351, (1.), (2.), that
which that of the
line <pp \s
^'\ and ^fi rectangular, or

XI.
S.^'X^lx
independently of the directions of \ and
XII.
since
is
..0 = t//V,
p,
S\^P^n,
whence
XIII. .,
or
4/2
= 0,
an arbitrary vector.
Now in general, by the functional relation 350, XXI. combined with

= w, we have the transformation,
= ;//(m' 0j^) = m'i// m;)^
XIV.
i//'
must satisfy the depressed or quadratic equathen m 0, as in I., the symbol
(2.).
i//0
if
;//
tion,
XV.
which
is
0=m';//-//2;
accordingly a factor of the cubic equation,
XVI.
whereto the general equation 350,
nishing.
= TO-.//2-;//3,
XXV.
is
reduced, by this supposition of
va-
CHAP.
CASE OF DEPRESSED EQUATION.
11.]
we have
(3.) If then
condition XIII.
not only
by XV.
is satisfied,
m = 0,
;
as in
m'
\ivX also
I.,
449
= 0,
and the two planes, above
as in III., the
referred to, are ge-
nerally rectangular.
(4.)
We might
indeed propose to satisfy that condition XIII., by supposing that
we had always,
XVII.
for every direction of
= 0,
..;//
that
XVII'.
is,
i/'p
= 0,
but in this case, the quaternion constant qî would vanish (by
m\ as being its scalar part (by 350, III.),
349, II.); and therefore the constant
would
still
(5.)
be equal to zero.
The
particular supposition
metrical character oi thQ qnQiiion

directions of the lines <pp and
lines
(j)'p
XVII. would however
alter completely the t/ô-
would imply (comp. 351, (2.)) that the
for it
-,
(when not evanescent^ arejixed, instead
(6.) On the side of calculation, we should thus have, for the

functions, fo and ^'p, monomial expressions of the forms,
XVIII.
whence, by 347,
.(Pp
= (iSap,
(p'p
we
350, VII.,
= aS/3p
two
corijnyate
should recover the equations, T^p=
= 0.
and m'
We
(7.)
should have also, in this particular case,
XIX.
fp now
so that
in like
XVIIL, and
of those
for their loci, as before.
having only certain planes
manner
= 0,
0p
(p'p
if
XX.
and
Ma,
.fp =
if
0,
-i-
(3;
and
p be any line in the Jixed plane perpendicular to a
a null line, if p be in that other Jixed plane, which is at right
vanishes,
is
angles to the other given
if
line, 13.
These two planes, or their normals a and (3, or the fixed directions of the
two lines (p'p and <pp, will be rectangular (comp. (1.) ), if we have this new equa(8.)
tion,
XXI.
for
every direction of p
ip^p
02
= 0,
XXI'.
or
02p
= 0,
and accordingly the expression XVIII. gives
= Sa/3. 0p = 0,
if
(3
-^
a,
and
reciprocally.
a and (3, the equation 350, XXIII. shows

(9.) Without expressly introducing
satisfies (compthat when i|/ = 0, and therefore also m' = 0, as in (4.), the symbol
or
new
the
equation,
depressed
quadratic
(2.) )
XXII.
which
cible,
is
accordingly
unless
(10.)
a.
_ ^"^
02
factor of the cubic IV., but to which that cubic

= 0, as well as m' = 0.
;/x
is
not redu-
we have thus
The
condition, then, of the existence
which we have respectively (pp =

useless
nishing (a case which it would be
equations should subsist
XXIII.
.. 7n
or that the cubic IV.,
and rectangularity of the two planes
and
(7.), for
0'p
without 0p generally va-
0,
to consider), is that the four following
= 0,
and
to'=0,
m" =
0,
XVII.
and
i//
quadratic factor XXII., should reduce themselves to
its
the very simple forms,
XXIV.
the cubic in
having thus
its
03
= 0,
and
XXV.
three roots equal,
and
02
null,
= 0;
and
\pp vanishing.
450
We may
(11.)
that
I. is zero,
is
when even one root of the general cubic 350,
also observe that as,
m = 0,
when
the vector equation
0p =
by one real direction of
XXVI.
was seen
m' =
0,
351) to be
(in
or
when
satisfied
the cubic in
XXVII.
common
by one
p, so
when we have
0jo
= 0,
^pp
also
= 0,
direction of the real and actual line p
by 350, XVII. and XX., the general

\pp
And
has two null roots, or takes the form IV., then the
tp
two vector equations,

are satisfied
III.
[bOOK
because
we have,
relation,
= m'p -
x^P-
by 350, XV., we have
also the relation xp

"p
0p, it folthe three roots of the cubic all vanish, or when the three scalar
equations XXIII. are satisfied, then the three vector equations,
(12.)
lows that
because,
when
XXVIIL
^p
^p = 0,
0,
have a common (real and actual) vector root
XP =
0,
or are all satisfied
by one common
direction of p.
(13.) Since
m" 
the cubic IV.
lowing forms,
XXIX.
in
which accented
signification
may
may
be written under any one of the
fol-
= 02^ = ^x0 = x^2 = ^.0j^ = &c.,
may be substituted for unaccented symbols and its geometrical

be illustrated by a reference to certain ^xed /me*, and fixed planesj
:
as follows.
(14.) Suppose
that the operation
that
first
so that the cubic in

late every vector p, so that (comp. (4.) (5.) (6.) )
^^ to
^p and
<p'p
x1>i
XXIV.
zero,
;
and
does not annihi-
have no< the
directions
o{ two fixed lines, hnt have only (comp. (1.) and (3.)) two fixed and rectangular
and 11', as their loci ; and let the normals to these two planes be denoted
planes,
by \ and
X',
in the planes
so that these
13'
and
two rectangular
lines,
\ and
\',
are situated respectively
II.
(15.) Then it is easily shown (comp. 351) that the operation ^ destroys the line \'
while it reduces* every other line (that is, every line which is not of the form
xX, with Yx = 0) to the plane 11 -^ X ; and that it reduces every line in that
itself,
plane to a fixed direction, /x, in the same plane, which is thus the common direction
And the symbolical equaof all the lines ^^p, whatever the direction of p may be.
tion, X 02= 0, expresses that this fixed direction p, of p^p may also be denoted by
.
X~^0
or that
we have
the equation,
XXX.
= x/* = ^' p - <PPj
which can accordingly be otherwise proved
if
A*
= 0^P
with similar results
for the conjugate
symbols, 0' and x'*
* "We
a line,
propose to include the case where an operation of this sort destroys
it to zero, under the case when the same operation reduces a line to a fixed
or reduces
direction, or to a fixed plane.
CHAP.
CASE OF MONOMIAL FOBM.
II.]
4^1
(16.) For example, we may represent the conditions of the present case by the
following system of equations (comp. 351, V. VII. IX. X., and 350, VI. VII. X.
XI):
= /3Sap + /3'Sa'p, (p'p = aSf3p + a'Sfi'p,

= m' = S (Vaa'. V/3'/3) = Sa/3 Sa'^' - Sa/3'Sa'i3,
{<pp
XXXI.
'.
(m" = Sa/3 +
Sa'/3';
= V(aV/3p + a'V/3'p) = m"p - <pp,

X>=V(/3Vap + /3'Va = m"p-^'p,
f X9
XXXII
and
may
XXXIII "
A ==
\i
which we
after
x =
i//'p
V/3/3'Saa'p
0'x'p
[
then write (not here supposing \' = V/ir, &c.),
V = Vaa',
V/3/5''
0/3
II
0/3',
/*'
S\\'=0,
= 0'a'
= S\>' =
SX/*
0'a,
II
easily find that
XXXIV
i>M=^'>i
?^VII/^'
/^^'"=^'
X/*
= 0;
(:
we have thus
(17.) Since
we have
0'2p,
xV = ^ where
is
/t'
XXXV. ..0 = Spxy=S/t'xp,

the locus of
the
is
î
line in the fixed direction of
to a plane,
which
= 0, 02.^ =
^ 0x
.
or
xp-^"/*';
therefore a plane perpendicular to the line
is
xp
normal
equations,
also the equation,
reduces every line in this
0,
is
may be interpreted
new plane
of
xp
XXXII.,
because <pX
(18.)
n,
/i'
the fixed direction of Yaa'y or of X'
in like
manner,
symbolical
and
with analogous
results,
Accordingly we
see,
and that
<p
<Pxp
by
= 0,
= 0.
We
see also, that the operation ^x? o^
which the operation
to
and
-A-nd the
as expressing, that the operation
are interchanged with unaccented symbols.
0XP has
that
to the Jixed direction of ^-'O, or of X'
that the operation 0^ destroys every line in this plane -L
when accented
/x'
the locus of the line x'p'
reduces every line
X0> destroys every line in the plane
and that thus the symbolical equa-
= 0, may
be interpreted.
be remarked that theâjcd direction X', of <pxp
or x0P> ought to be that of the line of intersection of the tvro ^xed planes of (pp and
tions, <pX'<P
(19.)
As
0,
x0
a verification,
it
may
Xp; and accordingly it is perpendicular by XXXIII. to their two normals, X and

with similar remarks respecting the fixed direction X, of <p'x'p or
which
xW)
fjt,'
is
perpendicular to X' and to
p,.
(20.) Let us next suppose, that besides

that
m"
sion for
0p reduces
from
itself to
zero.
In this
m = 0,
case, it
and m'
0, we have
i//
= 0,
but
has been seen (6.) that the expres-
the monomial form, fSSap; and therefore that the operaa,

fixed plane (-L a), while it reduces every other line to
destroys every line in
tion
a.
different
is still
fixed direction Q\
now
Sa/3
/3),
is
not contained in that plane, because
we have by
(21.) In this case
XXXVI.
i^'
(
XXII.
in 0,
" ^^"^'
(16.), equating
we have not
^'^
" "^^^'
(3'
to 0, the expressions,
< ^'
=
V.
x>
/3Vap = (m" 0') p,
and the depressed cubic, or the

quabe written under the very simple form.
XVIII. are reproduced
may
a or
""" " ^"^
=
XP = V. aV/3p (m" 0) p,
so that the equations
dratic
which
= 0.
452
[bOOK
III.
XXXVII. ..o = ^x=x^

here reduces an arbi(22.) Accordingly (comp. (5.) and (7.) ), the operation ^
trary line to the fixed direction of /3, while x destroys every line in that direction ;
and conversely, the operation x reduces an arbitrary line to the fixed plane perpen-
But because we do not

destroys every line in that fixed plane.
the Jixed direction of 0p is not contained in the Jixed plane
dicular to a, and
m" = 0,
here suppose that

of
xp; and (comp.
(8.)
and (10.)) the directions of
<pp
and ^'p are
yiot
rectangular
to each other.
(23.)
On
nish, or that
tion }pp
the other hand, if we suppose that the three roots of the

we have m=0, m' = 0, and to" = 0, as in XXIII., but
not satisfied for all directions of
is
p,
cubic in
<}>
va-
that the equa-
then the binomial forms
XXXI.
of ^p and 0'p reappear, but with these two equations of condition between their rector
constants, whereof only one had occurred before
:
XXXVIII.
(24.)
We
= Sa/3Sa'/3' - Sa/3'Sa'/3,
= Sa/S +
Sa'/3'.
have also now the expressions,
XXXIX.
xp=-<pp,
x'p =-1>'pi
and the cubic in ^ becomes simply (p^ = 0, as in XXIV. but it is important to observe that we have not here (comp. (9.) ) the depressed or quadratic equation 02 = Q,
since we hare now on the contrary the two conjugate expressions,
;
XL.
0V = ^/p = Vaa'S/3'/3p,
f^p = ^'p =
V/3/3'Sa'ap,
is now interpreted^
the equation 0^ =
by obreduces
serving that 02 here reduces every line to the fixed direction of 0"iO ; while
an arbitrary vector to that jftxed plane, all lines in which are destroyed by 02,
are equal,
(25.) In this last case (23.), in which all the roots of the cubic in
And
which do not generally vanish.
and
common
are null, the theorem (12.), of the existence of a
XLI.
= 0,
(pYaa
vector root of the three
be verified by observing that we have now,
may
equations XXVIII.,
-.pYaa'
x^' = ^
= 0,
the third of which would not have here held good, unless we had supposed i"=
(26.) This last condition allows us to write, by (16.),
XLI I.
the lines
planes
fx'
and
and
p,
11'
new
vector
which were
i'
0>' =
0,
0,
= 0,
Yfi'X
J/
= VAX'
11
V/x/i',
so that
Sîv
0,
will be a line in the intersection of those
lately seen (14.) to be the loci of the lines
(comp. (17.)
the loci of
xp and
compose a rectangular system.

(27.) In general, it is easy
XLIV.
in
YfiX =
0,
S/i/i'
= 0,
thus coinciding in direction with the normals \ and

if then we write,
XLIII.
this
0/A
.
'
'
x'p
'y
and the three
S/x'v
X',
to the
= 0,
two rectangular planes^
0p and 0'p, and are now

v (or X', X, i/) will
ft, /*',
lines
to prove that the expressions,
= /î + ^f^'h /3' = a'/3i + fe'/3'i,

//^
\ai=aa + a'a, a\ = ba + b'a',
which
lars,
0.
a, /3, a', /3' may be ang four vectors, and a, b, a', b' maybe any four scaconduct to the following transformations (in which p may be
any vector)
:
CHAP.
453
VECTOR AND QUATERNION INVARIANTS.
II.]
XLV.
Saii3i+Sa'i/3'i
= Sa|3 + Sa'/3';
= /3Sap + /3'Sa'p;
XLVI.
|6iSaip + /3'iSa'ip
XLVII.
Vaia'i.Vi3'ii3i=Vaa'.V/3'/3
.
so that the scalar, Sa/3 + Sa'/3' ; the vector, (3Sap + /3'Sa'p ; and the quaternion*
a ffiven system oi four
Yaa.Yf3'(3, remain unaltered in value, when we pass from
to another system of four vectors ai/3ia'i/3'i, by expressions of the
vectors
af3a'(3',
forms XLIV.
With the help of this general principle (27.), and of the remarks in (26.),
be shown, without difficulty, that in the case (23.) the vector constants of
the binomial expression (3Sap + (3'Sa'p for <})p may, without any real loss of genera(28.)
it
may
be supposed subject to the four following conditions,
lity,
XLVIII.
which evidently conduct
XLIX.
and thus put
= 0,
cubic 0^
= Sa/3=Sa')3 = S)6/3' = Sa')S';
to these other expressions,
in evidence, in
^V
=/3Sa/3'Sa'p,
03p
a very simple manner, the general non-depression of the
to the quadratic,
= 0.
<p^
not only i = 0, m' = 0, m" = 0, but

a depressed form of <j)^ 0, by the linear function
= between its con0p reducing itself to the monomial /3Sap, with the relation Saj3
and thus the consequences of the
stants, has been already considered (in (10.))
(29.)
also
t|/
The
= 0,
case, or sub-case,
and therefore
<p^
when we have
0, as
supposition III., that there are (at least) two equal but null roots of the cubic in 0,
have been perhaps

(30.)
0,
sufficiently discussed.
As regards
the other principal case oi equal roots, of the cubic equation in
namely that in which the vector constants are connected by the
relation V., or
by
the equation of condition,
L.
it
may
= m"2 - Am' = (Sa/3 + Sa'/3')2 - 4S (Vaa'. V/3'/3)

= (Sa|8 - â'^y + 4Sa/3'Sa'|3,
remark that
suffice to
it
LI.
and that thus
its
X02
as given in (14.),
353.
and
= 0*2^
interpretation
by VL,
conducts,
bolical equation,
if
,j,
or
by VII. and IX.,
= ^ _ Irn"
to the
sym-
precisely similar to that of the analogous equation,
is
= O,
where
x="-0>
XXIX.,
in the following sub-articles.
When we
m = 0,
have
but not
0,
nor m"'^=Am\
the three roots of the cubic in ^ are all unequal, while one of
them is still null, as before and the two roots of the quadratic
;
and scalar equation, with real coefficients (347),

= c2+ m'c^m'.
I.
.
*
have, in these transformations, examples of
nion Invariants.
We
what may be
called Quater-
454
[bOOK
III.
formed from the cubic by changing ^ to - c, and then

dividing by c, are also necessarily unequal, whether they be
which
is
We
shall find that when these two scalar

real or imaginary.
are
there
are then two real directions, pi and
real,
rootSf Ci, C2,
P2> in that Jixed
plane
which
is
the locus (351, 352) of the
line ^p, possessing the property that for each of them the homogeneous and vector equation of the second degree,
II.
Npt^p =
0,
or
(i>p
II
p,
without p vanishing ; namely by our having, for the

Jirst of these two directions, the equation
is satisfied,
III.
and
0ioi
= -cipi,
or
if
0ipi=O,
01
+ Ci;
them the analogous equation,

=
= +
or
if
C2P2,
02p2 O,
02
for the second of
J.Y.
=(f)p2
^2:
but that no other direction of the real and actual vector
p, sa-
the equation V., except that third which has already

been considered (351), as satisfying the linear and vector equatisfies
tion,
V.
It will also be
0p =
0,
with
>
T/o
shown that these two
0.
directions, pi,
/02,
are not
but rectangular, to each other and to the third
realy
only
direction p, when the linear function (pp

or when the condition
(4.) ),
VI.
'
(p'p
or
(pp,
Vr.
is
self conjugate (349,
S\(pp
= Sp0X,
satisfied by the given form of 0, or by the constants which

enter into the composition of that linear symbol; but that when
this condition of self-conjugation is wo^ satisfied, the roots of the
is
quadratic I. may happen to be imaginary : and that in this

case there exists no real direction of p, for which the vector
equation II. of the second degree is satisfied, by actual values

of p, except that one direction which has been seen before to
satisfy the linear equation
(1.)
is to
The most obvious mode
V.
of seeking to satisfy II., otherwise than through V.,
= x^-\= /3Sap + jS'Sap,
assume an expression of the form, p
the equation, (0
c)
p=
0,
with
<pp
a;'/3',
and
by
to seek thereby to satisfy
satisfying separately the two
scalar equations,
VII.
= X (c +
Sa/3)
xSa(S',
= x(c +
Sa'|3')
+ xBa'f3,
CHAP.
which
CASE OF UNEQUAL REAL ROOTS.
II.]
by elimination ofV:
give,
VIII.
which
and
IX.
X.
we may
XI.
in like
write
form of
p2
it
c0i3
= (to" +
0P2
"We
XIV.
/3, )3'
In
and IV.
fact,
the expression IX. gives

or
îpi
0,
- ^/3 = - C2/3 - 0i3 = -
C2)(3
may
= - ci/3 - ^/3 = - ^1/3 = - 0/3 - m'c2-^(5y
0/3
C20/3
= -0^-iV^/3;
^2/3
be thus written,
m'j3
= - C2P,
or
02/O2
= 0.
also write,
may
XIII.
Denoting then, as above, by cj

we have
I.
thus,
and gives,
(2.)
c,
= Sa^'Sa%
Sa'/S')
= ci<J>l3 + m'l3 = -cipi,
= (m" +
pi
supposed for the present to be real,
I.,
manner, the expression X.
XII.
and
Sa|3) (c
for p, in the plane 11 of
satisfy the equations III.
because
the following quadratic in
x,
= i3(ci + Sa'/3') - /3'Sa'^ = ci(3 + Va'V|3'/3

pi
= /3 (C2 + Sa'jS') - i8'Sa'/3 = c^^ + Va'V/3'/3
p2
i>pi
and
(c
the roots of that quadratic
two real directions
which
easily seen to be only another
is
Co,
these
455
= /3'(ci + Si3) - )3Sa/3' = ci/3' + VaV/3/3'=- 02/3^

= /3'(c2 + Sa[S) - iSSa/3' = cajS' + VaV^/3' = - 0i/3'
p'2
p'l
||
||
pi
p2
have the equations,
shall then
XV.
0ip'i
0,
<p2p2
be the same by VIII. as those of pi and

so will furnish no new solution of the problem just resolved.
but the directions of
we have
(3.) Since
p'2 will
pg,
02
%
0p
02/3'
II
02/3
fixed direction of 0i'iO

in the
and
thus,
follows that the operation 02 reduces every line in the fixed plane of
XVL
it
and
p'l
II
pi
and
and
01-10,
II
that, in like
XVI'.
0i/3'
1|
0i/3
||
p2
||
to the
manner, the operation 0i reduces every line,
same fixed plane of 0p, to the other fixed direction of

Hence we may write the symbolic equations,
02-'O.
(4.)
XVII.
in which the points
XVIII.
may
.
(5.) If
1//
is
we propose
to
therefore,
XX.
îp
and
from
//i
010
= 0,
we have
ci)
0i,
(0
by the same general
rule (347, XI.)
by
= V. (p'lfKp'w = V. (0'/i + cifi) (0V + civ),
= i/zp + cixp
XXII.
if
02-
in fact
definition 350, VIII. of %,

ci^p,
and in like manner,

even
= 0,
we have
\p{Vfiv
by the
= 0201 = (0 +
form
formed from 0,
XIX.
and
01 020
+ c-i) = 02 - m"(p + m'=^,

0102.0= 0201-0 = '//0 = TO =0.
0102
so that
which
be omitted
be different from zero, and
(6.) Accordingly, without
or
XXI.
1//1
1//
+ cix
+ ci2
^2 = yp + C2X+cz'^,
if ci, C2
assuming that
be arbitrary scalars.
ra vanishes,
if
we
operate on
i/zjp
with
456
^1, or symbolically
multiply the expression
product,
XXIII.
where m\
is
XXIV.
as appears
what the
scalar
<l>i\Pi
XXI.
for -î
when ^
becomes,
by
the definitions of
î,
= {(p + Ci) (t// + cix + Ci2)

= <pyp+ cv{<px + ^) + ci(^ +
m+ am' + ci^m" + ci^ = twi,
changed to
is
= S.^'iXfi/ifiv = S.(^'X+ci\)
niSX^j/
by
[bOOK
xO,
(p',
X) "*
^%
"^"i
we
get the symbolic
x) +
ci3
î, or is
+ ci/i)
(<p'n
and by the
III.
such that
((p'v
civ);
relations
between
those symbols which have been established in recent

Articles, or in the sub-articles
appended to them.
m = 0,
now again that

we have by XXIII.,
(7.) Supposing
tic I. in c,
XXV.
0i;//i
= mi =
and
and that
in like
be formed from mi, by changing ci to

(8.) Comparing XXV. with XVII., we
if i2
intimate connexion existing between

tor to the fixed direction of
tical, because,
xpi
and
^rÔ, or of pi
by XXI., and by the known
are the roots of the quadra-
ci, ca
XXVI.
manner
^2'>p2
= i2 = 0,
C2.
may be
led to suspect the existence of
an
(p2^,
since each reduces an arbitrary vec-
and
in fact these
relations
two operations are iden-
between the symbols, we have
XXVII.
i//i
= + cix + Ci2 = (m' - m'> + (p^) + a {m" - 0) + Ci*

= 02 - (m"+ ci) = 02 4 C20 = 002
i//
and
XXVIII.
similarly,
while
-^
(9.)
= 0102,
i^2
= 02 +
ci0
= 0^^
as before.
We have
thus the new symbolic equation,
XXIX.
00102 =
0,
which the three symbolic factors 0, 0i, 02 may be in any manner grouped and
and in which the subject
transposed, so that it includes the two equations XVII.
in
Its interpretation has been already partly

an arbitrary vector p.
reduces every vector to the fixed plane IT,
given but we may add, that while
01 reduces every line to another fixed plane, Hi, and 02 reduces to a third plane,
of operation
is
thus 0102, or 020i, while it destroys two lines pi, p2, and therefore every line in
IT, reduces an arbitrary line to t\\Q fixed direction of the intersection of the
two planes 111112, which intersection must thus have the direction of 0-'O ; and in
II2
the plane
like
manner, the fixed direction pi of 0r'O, as being that to which an arbitrary vec-
tor is reduced (3.) bj'the
compound operation 020, or 002, must be that of the interand pii or 02~'O, has the direction of the intersection of
section of the planes 11112;
while on the other hand 002 destroys every line in
and 00i every
line in
niTi
II2
so that these three planes, with their three lines of intersection, are the chief
IIi,
elements in the geometrical interpretation of the equation 00i02

(10.)
= 0.
The conjugate equation,
XXX.
0'0'i0'2
= 0,
may be interpreted in a similar way, and so conducts to the consideration of a conjugate system of planes and lines ; namely the planes 11', FI'i, n'2, which are the
loci of 0'p, 0'ip, 0'2p, while the operations 0'i0'2, 0'20'i, and 0'0'i destroy all lines
CHAP.
CASE OF IMAGINARY ROOTS AND DIRECTIONS.
II.]
457
planes respectively, and reduce arbitrary lines to the fixed directions

of 0'->O, (p'xÔ^
II'II'i, which are also those
in these three
of the intersections, n'in'2, ir'all',
(11 J It
important to observe that these three last lines are the normals to the
is
three first planes,
that, in like manner, the three
and
IT, II";
11,
To prove
9xe perpendicular to the three latter planes.
this,
lines
/ormer
to ob-
is sufficient
it
serve that
XXXI.
and
Sp>p = Sp^y =
(12.) Instead of eliminating x'
have eliminated c
x\
and x^
x\
XXXIV.
X2 be the two values of x'
xix^
(xix'z
xi(3
+ Xix'i)
+ x\(3',
x'lx'z
P2
252/3
II
= - Sa/3'
and consequently it is satisfied, if the given function

cause we have then the relation,
XXXVI.
binomial form of
in fact the
XXXVII.
.<p'p-(pp
^'-'0
x, then
+ X2l3',
- Sa'/3')
(Sa/3
pi, p2,
V/3a
Sa'/3
or 0i"'O, ^2"'0,
= - /32Sa/3' + S/3/3'(Sai3 - Sfl'/3') 4- iS'zSa'jS = S
4-
this other quadratic^
hence the condition of rectangularity of the two lines

pressed by the equation,
XXXV.
tpp
= a;2Sa'/3 + a:.r'(Sa'/3'-Sa/3)-a;'2Sai3';
XXXIII. ..pi\\
and
or that
0,
between the two equations VII., we might
a;
which would have given
XXXII.
also, if
f p'=
if
0,
J- 0-'O, &c.
similarly, 0'p
/3i3T(/3a
is
ex-
(i'a)
be self-conjngate (VI.), be-
+ V/3'a' =
gives (comp. 349, XXII.),
= (aS/3p - /3Sap) +
which cannot vanish independently of
(a'S/3'p -/3'Sa'p)
= V.pV(/3a +(5'a'),
unless the constants satisfy the condition
p,
XXXVI.
(13.)
With
this condition then, of self-conjvgation of ^,
we have
the relation of
rectangularity,
XXXVIIT.
at least if these directions p\
The
as follows.
XXXIX.
condition
=S
hence
and
Spip2=
and pg be
XXXVI.
aa'V(/3a
{a^ Sa')3
or
0,
0rÔ
-i-
(pfÔ
which they can
real,
easily be
proved to
be,
gives,
+ /3'a') =
a'2Sa|8')2
a^Sa'/J
= (Saa')^
Saa'(Sa'/3'
(Sa/3
- Sa/3) - a'^Sa^S'
Sa'/3')2,
- 4m') = aa'2 { (Sa/3 - Sa'/3')2 +

a2a'2(m"
4Sa/3'Sa'/3}
= (a2a'2 - (Saa')2) (aj3 - Sa'|3')' + (a2Sa'i3+ a'2Sa/3')2 > 0,
- Sa'/3')2 + 4Sai3'Sa'/3 = i"2 - 4m' >
XL.
(Sa/3
.
two quadratics, I. (or VIII.), and XXXII., has real and unequal
roots : a conclusion which may also be otherwise derived, from the expressions
= aa + ba% ^' = ba + a'a\ which the condition allows us to substitute for f3 and /3'.
|3
the ybwr vectors ajSa'fi' are com(14.) The same condition XXXVI. shows that
so that each of the
planar, or that
we have
XLI.
hence
Yaa\
the function
or 0''O is
<p
is
the relations,
Sai3/3'
0,
now normal
Sa'/3/3'
to the
= 0,
plane
Y(Yaa'.Y(S'(3)
and therefore by
(13.), when
self-conjugate (VI.), the three directions,
3n
458
XLII.
or
p, pi, p2,
[bOOK
III.
0-10, (prÔ, 02"^ 0,
compose a real and rectangidar system.

(15.) In i]\Q present series of subarticles
roots of the cubic in
(to 353), we suppose that the three

= 0) having
are all unequal, the cases of equal roots (with
<p
been discussed in a preceding series (352); but it may be remarked in passing, that
when a self-conjugate function 0p is reducible to the monomial form l3Sap, we must
have the relation Y(3a =

(comp. 352, (5.) and (6.)
and that thus the
pendicular to the fixed plane (352,
line
(3,
to
thefxed
direction of which
an arbitrary vector, is perevery line in which is destroyed by that
) the operation
tlien reduces
(7.) ),
operation 0.
be thus self- conjugate, it is evident that the three planes

(16.) In general, if
which are (comp. (10.) ) the loci of fp'p, (pip, fzp, coincide with the
TV, n'l, n'2,
planes
n,
which are the
lli, 112,
When
(17.)
is
not
loci of (pp, <pip, <p2p.
self- conjugate,
and
so that cpp
(p'p
are not generally equal,
has been remarked that the scalar quadratic I., and therefore also the symbolical
cubic in 0, may have imaginary roots; and that, in this case, the vector equation If.
it
of the second degree cannot be satisfied
which
satisfies the linear
and
real
(pp itself to vanish, while p remains

of such imaginary scalars, as roots of I., and of
As an example
actual.
real direction of p, except that one
by any
equation V., or causes
what may be called imaginary directions, or imaginary vectors (comp. 214, (4.)),
which correspond to those scalars, and are themselves imaginary roots oi II., we may
take the very simple expressions (comp. 349, XII.),
XLIII.
which y denotes some
in
the condition VI., the function

self-conjugate part 0o
its
0p
= Vyp,
and given
real
0'p
vector,
= Vyp;
and which evidently do not
satisfy
own conjugate,
so that
being here the negative of its
We have thus,
zero (comp. 349, XIII.).
is
XLIV. ..mo=0,
to'o
= 0,
0o =
m"o=0,
O,
;/'o
= 0,
xo = 0,
and consequently, by the sub-articles to 349 and 350,
XLV.
the quadratic
m = 0,
and
I.,
m'
= -y^,
roots
its
ci, cg,
XLVI. ..c2-y2 =
m" =0,
become
Ci
0,
i//p
=-
= + yri.Ty,
where
imaginary of algebra (comp. 214,
and XXII.) we have now
XLVII.
1 is the
;//i<T
=-
ySyff
ySyp,
xP-~ ^79
therefore,
C2
= -v/r7.Ty,
(3.) )
- CiYya + ciâ = (y - ci)Vy(T,
thus by
^^(T
XX.
or
XXL,
= (y - C3)Vy(r
hence
= yif/i(T,
Vy^îff
= 0,
Sy\|/i(T
&C.,
and
XLVIII.
and
in like
we
= (0 +
01X^10-
manner
= (y +
Ci)4/i(T
XLVIII'.
02^//2<^
- Ci)Vy(T = (y2 (y
Ci)
Ci^)Yy(T
take an arbitrary vector c, and derive (or rather conceive as derived) from
if
then
it
two (imaginary) vectors pi and pg by the {imaginary) operations
shall
have (comp.
III.
XLIX.
and
= 0,
L.
p2
xpi
and IV.) the equations,
pi
i//i(T,
^^cr,
0ipi
02p2
= O,
0,
= -cipi, Vpi0pi=O,
= - C2p2, Vp20p2 = 0,
0pi
0p2
and
1//2,
we
CHAP.
as ones
VECTOR AND QUADRATIC EQUATION.
II.]
which are
We
at least symbolically true.
459
find then that the two
imaginary
or as far as calculation
directions, pi
po, satisjy (at least in a symbolical sense,
is concerned) the vector equation II., or that pi and pz are two imaginary vector roots
and
Vp^p =
but that, because the scalar quadratic I. has here imaginary roots,
;
one in the
equation II. has (as above stated) no real vector root p, except
direction of the given and real vector y, which satisfies the linear equation V., or
of
this vector
= 0.
gives <pp
have been more simply treated,

(18.) This particular example might
wish
to satisfy the equation,
"We
follows.
as
method,
general
LI.
which gives, when we operate on

LII.
if
then
we wish
it
by V. y and V. p,
= Vyp
to avoid supposing <pp
LIII.
we can
Syp
satisfy these
- Vyp = 0, we must
= 0,
by any
when we assume the
is satisfied,
LI V.
seek to satisfy the two
p2=0;
(real or
imaginary)
Now
the vector equation LI.

(really or symbolically)
satisfied
LIII.
if
these others,
= p^-Vyp
Syp,
scalar equations,
and conversely,
less
= V.pVyp = pSyp-p2y;
by a
p,
we
shall
have
the first equation
expression,
= (c + y)Vy = Vy
cr
(T
(c
- y ),
an arbitrary vector, and c is any scalar, or symbol subject to the laws of

where a
scalars ; and this expression LI V. for p, with its transformation just assigned, gives
is
LV.
the quadratic
XLVI.
is
p2=(c2-y2)(Vy(T)2 =
therefore reproduced,
and imaginary directions, as

(19.)
0,
and
c2
if
Ave
- y2 = Q
have the same imaginary
roots,
before.
Geometrically, the imaginary character oi the recent problem, of satisfying

pVyp = by any direction of p except that of the given line y, is
the equation V.
apparent from the circumstance that 0p, or Vyp, is here a yector perpendicular to p,
and that therefore the one cannot be also parallel to the
both be actual lines
if
other, so long as both are
cles
real*
354. In the three preceding Articles, and in the sub-artiannexed, we have supposed throughout that the absolute
=
wantingy or that the condition w
in which case we have seen (351) that it is always
term of the cubic in

is satisfied
(p
is
= 0, by at least o?ie
possible to satisfy the linear equation ^p
real and actual value of p (with an arbitrary scalar coefficient)
;
or
by
*
at least one real direction.
Accordingly the two imaginary
be those which in
It will
directioiis,
modern geometry are
be easy
above found
called the directions
now
to show,
for p, are easily seen to
oUines drawn in a given
plane (perpendicular here to the given line y), to the circular points at infinity
which supposed directions the imaginary character may be said to be precisely
that each
is
its
(in the given plane)
own perpendicular.
of
this,
460
[bOOK
III.
that although conversely (comp. 351, (4.) ) the function (j)p

cannot vanish for ani/ actual vector p, unless we have thus
m=0, yet there is always at least one real direction for which
the vector equation of the second degree,
I.
Yp<pp
0,
in combination with
which has already been considered (353)

the condition m = 0, is satisfied and that
if
the function ^ be
a self-conjugate one, then this equation I. is always satisfied

by at least three real and rectangular directions, but not generally
by more
directions than three ; although, in this case of
self-conjugation,
II.
namely when
=
or
ir.
(f)'p
0p,
for all values of the vectors p
pen
to
become
and
S\(pp
X, the
= Sp<pX,
equation
true, for o7ie real direction of p,
direction perpendicular thereto
I.
may hap-
and
for every
or even for all possible direc-
tions, according to the particular
system of constants, which
We
shall
enter into the composition of ihQ function (j)p.
also that the scalar (or algebraic) and cubic equation,
III.
show
m + m'c + m'c^ + c^
formed from the symbolic and cubic equation 350, I.,

- c, enters
importantly into this whole
by changing ^ to
if
that
it have one real and two imaginary roots,
and
theory
the quadratic and vector equation I. is satisfied by only one
real direction of p but that it may then be said (comp. 353,
(17.) ) to be satisfied also by two imaginary directions, or to
have two imaginary and vector roots : so that this equation
which
is
I.
may be
lines,
said to represent generally a system of three right

least must be real.
For the case II., the
whereof one at
scalar roots of III. will be proved to be ahvays real; so that

and m"o be formed (as in sub-articles to 349 and 350)
if mo, m'o,
from the self-conjugate part (pop of any linear and vector function (J>p, as m, m', and m" are formed from that function (pp itself,
then the neic cubic,
ly
which thus
results,
= Wo +
Wi'oC I W2"oC-
+ C%
can never have imaginary
roots.
CHAP.
(1.) If
v.
CASE OF SELF-CONJUGATION.
II.]
we
write,
*p = 0p
Avhere c
is
+ cp,
$'p
an arbitrary
^'p
if
..* =
VI.
and
VII.
VIII.
with
Hence
may
it
i//
.
we denote by
<]>
+c
*",
*=
^', and
-f
*' = 0'
c,
M what
i//,
^',
c,
and
or $, the calculations in 353, (5.), (6.),
+ ex + c*,
*''
= i//'+cx'+c2,
M= m + m'c + rn'c^ + c^,
M= *^ =
V.
or briefly,
cp,
and
scalar,
w become, by this change of ^ to

show that we have the expressions,
(2.)
461
^
x'l
^^^ the constants
m',
m" become,
IX.
X = Dc^ = X+2c, X'= Dc^' = X' + 2c,

r Af' = DcM= m' + 2m
+ 3c2,
"c
'
'
\ Af"= |Do2iH = m" +
3c
with the verifications,
XI.
as
we
*+X=
*'
*X + ^ = *'X' + ^' = iW',
+ X' = M",
had, by the sub- articles to 350,
^
(3.)
The new
+ X = ^' + X' = "*"' 0X + 4' = fx' + = w'.

* must satisfy the new cubic,
'/''
linear symbol
XII.
= M- M'^ +
3/"$2 _ $3
which accordingly can be at once derived from the old cubic 350,
XIII.
(4.)
Now
it is
under the form,
always possible to satisfy the condition,
XIV.
M=0,
by substituting
for c a real root of the scalar cubic III,
new symbolical
cubic XII. to the form,
XV.
which
I.,
- *)^
(c
= m 4 '(c - *) +^"(<^ - *)2 +
and thereby
to
reduce the
0=$3_i\f "$34.i\f'$;
precisely similar to the form,
is
= (p^-m"<pi-\-m'<l>,
and conducts
to analogous consequences,
since they can easily be supplied
352,11.,
which need not here be developed
by any one who
in detail,
will take the trouble to read again
the few recent series of sub-articles.

it happen that *'p constantly vanishes, in which case
not identically null) takes a monomial form, which is reduced to
zero (comp. 352, (7.) ) for every direction of p in a given plane, the operation
reduces (comp. 351) an arbitrary vector to a given direction; and the operation *
For example, unless
(5.)
M' = 0, and $p
(if
destroys every line in that direction
way
the equation
(6.)
And
see that
so that, in every case, there is at least one real
*p -
0,
and therefore
also (as
above asserted)
without causing p itself to vanish.

since that equation I. may be thus written,
I.,
XVI.
we
of satisfying the vector equation
it
can be
yp'i>p
= 0,
or
*p
II
p,
satisfied without 4>p vanishing, if this new; scalar
equation,
XVII.
= C2 +
M"C^ M',
and quadratic
comp. 353,
1.,
462
have real and unequal roots
Co
Ci,
$p wiU generally have
the line
and
<f>i
and $2
and
"^îp
Ci,
III.
write,
*2=<J+C2,
and there
will be two real
which $ipi = 0. while

and these are precisely the
in that plane, for one of
each
for the other, so that
$2^2=
fixed directions of
= *+
for its locus a given plane,
and p2
distinct directions pi
we then
for if
XVIII. .. *i
[bOOK
^2iOj if
XVI., or I.
formed from
satisfies
î and
*"2 be
by changing
io
respectively.
(7.) Cases of equal and of imaginary roots need not be dwelt on here ; but it may
be remarked in passing, that if the function ^p have the particular form (g being
scalar constant),
any
XIX.
XIV.
XX.
then
(pp=gp,
the cubic
(^-^)3 = 0,
or III. having thus all
its
and
XXL
M= (^ + c)3
roots equal, and the equation
by every direction of p, in this particular case.

The general existence of a real and rectangular system
satisfying L, when the condition II. is satisfied, may be proved
I.
being sa-
tisfied
of three directions
(8.)
as in 353, (14.);
unnecessary to dwell on the case where, by two roots of the cubic becoming
equal, all lines in a given plane, and also the normal to that plane, are vector roots
and
it is
of I., with the
And
(9.)
same condition
II.
because the quadratic,
+ ni'c + m' (353, I.), has been proved to

= <Pp, the analogous quadratic XVII.
c'^
have always real roots (353, (13.)) when

must liliewise then have real roots, Ci, Cg
<p'p
whence it immediately follows (comp.

XII. and XIII.), that (under the same condition of self-conjugation) the cubic III.
has three real roots, c, c + Ci, c + C2 and therefore that (as above stated) the other
;
cubic IV., which
formed from the self-conjugate part <po of the general linear and
and which may on that account be thus denoted,
is
vector function 0,
XXII.
If
(10.)
m = TOo -
we
Mo= 0,
?'
On
c,
= m'o-y2,
which
XXV.
whence
roots always real.
c,
the equation
recover the relations,

to"
= 7n"o, as in 350,
= y.\pfx^v
a similar plan, the equation m(p'yfiv
and v are arbitrary
jx
we
and
XXIV.
in
its
Syôy (349, XXVI., comp. 349, XXI.) becomes,

XXIII.
J!f=Mo-Sy4>oy;
whence, by comparing powers of
(11.)
has
denote in like manner by $0 the symbol ^0
(m +
m'c
follow these
+ c^)
m"c^
new
becomes,
M^'V^v=Y.^fi^v,
vectors,
(<p'
(1.).
and
c is
comp. 348,
an arbitrary scalar ; or more
+ c)Yfiv - V. (i///i +
cxfi
C-fi) (^)pv
(1.),
fully,
+ c^v + c^v)
equations,
XXVI.
(w + m'<p')Ynv = V(x///i .xv-'^v. \ju),
- vxpft + xA* X*')?
XXVII.
(m' -f m"<p')Yiiv = Y{fi-^v
XXVIII.
(m" + (J>')Yfxv = Y(nxv ~ vxn),
.
which can
&c.)
all
we can
be otherwise proved, and from the last of which (by changing
infer this other of the
XXIX.
(12.)
As an example
(w'
xp')Y fxv
if,
= Yd-i^x"*^ ~ J^^XM)'
of the existence of a real
directions (8.), represented jointly
to
same kind,
by an equation
and rectangular system of three

form I., and of a system of
of the
CHAP.
REAL AND RECTANGULAR SYSTEM.
II.]
three real roots of the scalar cubic IIL,
when
take the form,
XXX.
<pp=gp
the condition IL
+ YXpix =
g being here any real and given scalar, and X,

vectors
to -which/ornj, indeed,
We have
0op can be brought.
XXXL
fi
is
satisfied, let
us
<l>'p,
any
and non-parallel given
real
shall soon find that every self-conjugate function
now
(after
some reductions),
^p=Y\pfiS\fi-YXfiS\pn-g(X8fip-\- nSXp) + g^p,

XXXIL xp = - {XSup + fiSXp) + 2gp,
.7n=(^g- SX/i) (^2 _ X2^2)^ rn' = - X'^fi'^ - 2g8X^ + 3^2,
m" = - SXfi + 3g
XXXin.
and
we
463
where the part of
which
xpp
is
independent of^'
may be put under several other forms,
such as the following,
XXXIV.
V(Xp/*SX/i
= X(pSX/tt
and *,
*",
&c.
M" may
be formed from 0, rp, Xi ^j ^'1 *"j by simply

The equation 3f= has therefore here three real and unequal
X, M, M',
changing g to c+g.
roots,
- XfxSXpfi) = XpfiSXfi - XfiSXpfi

= ^X(Xjup + pX/i)/i = X(XSnp + fiSXp - Xpfi)fi,
SXfxp)fi
-t-
namely the three following,
.c =
-^ + SX/z, e+Ci = -g + TXfi, c+ C3 = -(7-TX)u;

the corresponding forms of ^p are foimd to be,
XXXVL ^p = VX/iSX/xp, îp= - (XT/i + /xTX; S p(XTfi + /iTX),
XXXV.
and
i//2p=- (XT/i-/iTX)S.p(XT/x-/iTX).
Thus
""Fp,
îp, and *'2p have in fact the three fixed and rectangular directions of
+ juTX, and XT/* fxTX, namely of the normal to the given plane of X,
VX^t, XTfj,
fi,
and the bisectors of the angles made by those two given
lines
and
these are ac-
cordingly the onlg directions which satisfy the vector equation of the second degree,
XXXVIL
(Vp0p=V.pVXp^=)VpXS/ip + VpjuSXp = O;
so that this last equation represents (as
was expected) a system of three right
lines,
in these three respective directions.
(13.) In general, if
cubic equation
M=
0,
XXXVIII.
if
the corrresponding values of
XXXIX.
also
we have
..
>I'l
denote the three roots (real or imaginary) of the
ci, C2, C3
and
we
write,
<S?i=<p
Ci,
$2 =
+ C2,
*3 =
C3,
will be (comp. VI,),
= ;^+CiX+Ci2,
*2=^ + C2X+C22,
^3 = ^+C3X+<'32;
the relations,
- m" = - C2 + C3 =
X,
+ CbCi + C1C2 = + to' = 0x +
0;//
ici
C1C2C3 = m =
4-
C2C3
'I'j
whence
it is
easy to infer the expressions,
XLI.
*i
= (C2 - C3)-'
*3
which enable us
- ^2), *2 = (C3 - ci)-i (^1 ^3),

- C2)-l (^2 - ^1)
(^3
(Cl
to express the functions $ip,
*2p, *3P as binomials (comp. 351,

&c.), when îp, ^2p ^3p have been expressed as monomes, and to assign the
planes (real or imaginary), which are the loci of the lines *ip, *2p, *3p.
464
(14.) Accordingly, the three operations, *,
by which
ti, 42,
lines in the three
lately determined directions (12.) are destroyed, or reduced to zero,

first
111.
[bOOK
and which
at
present themselves under the forms,
XLII.
$p = XS/^p + /iSXp,
$ip
*2 = VXp/i - pTX/u,
= VXp/i + pTX/z,
are found to admit of the transformations,
VTTTT
XLIII.
A = ^2p-îpL;
$p
'
^2P-^P
= .
g,
^
<>Ip
'^
where
'*',
î,
*2 have
*2p=T
'f^
2TX/i
TX/i+SX/i'
the recent forms
^p-îp
TX/tt-SX/i'
XXXVI., and the
locioi *p, *ip, $2p
com-
pose a system of three rectangular planes.

(15.) In general, the relations (13.) give also (comp. 353, (8.)),
XLIV.
whence
<>iî
XLVI.
also,
^2 = *3$i, ^3 = *i$2,
= $2'5'2 = *3^3 = <Ei*2*3 = 0,
^1^2 = '^2^z - î^\ = 0,
^1 = $2*3,
XLV.
and
the symbols (in any one system of this sort) admitting of being transposed and grouped
at pleasure; if then the roots of
be real and unequal, there arises a system
of three real and distinct planes, which are connected with the interpretation of the
M=
symbolical equation,
$i<E>33
= 0,
exactly as the three iânes in 353, (9.) were con-
nected with the analogous equation, 001^2 = 0.

(16.) And when the cubic has two imaginary roots, it may then be said that there
is one real plane (such as the plane -L y in 353, (18.), (19.) ), containing the two
imaginary directions which then satisfy the equation I. and two imaginary planes,
;
which respectively contain those two

line (such as the line
the same equation
355.
in the
directions,
example
cited),
and
intersect each other in one real
namely the one real vector root
of
I.
Some
be thrown upon that vector

by considering the system of the two
additional light
equation of the second degree,
may
scalar equations,
I.
SAP0/3
and
0,
II.
Sa/j
0,
and investigating the condition of the reality of the two* directions,

are generally satisfied, and for each of
px and p2^ by which they
which the plane of p and ^/s contains generally the given line x in
I.,
or
is
normal to the plane locus
two directions are always real and

become indeterminate), when the
gate
and that
then, if
of />.
We shall find that these
A be a root po of the vector equation,

III.
II.
rectangular (except that they may

linear function ^ is its own conju-
Yp^ppÔ,
Geometrically, the equation I. represents a cone of the second order, with X

and with the three lines p which satisfy III. for three other sides ; and
for one side,
II. represents
a plane through the vertex, perpendicular to the side X. The two direc-
tions sought are thus the two sides, in
which
this plane cuts the cone.
CHAP.
NEW PROOF OF EXISTENCE OF THE SYSTEM.
II.]
which has been already otherwise discussed, the

of that equation
also roots
three real
when
<p'p
XXII.) from the

tor function
(1.) Let \,
/I,
if =0 was formed from
= yfi + zv,
and
and
the equation II. is satisfied,
the roots of
i,j, k.
symbols
which quadratic
VII.
will be real
and unequal,
if
(Sv(pv-Sfi^fiy + ASiJi<pvSv<pn>0;
= S (yi/i + ziv)
{yin
VIII.
Introducing
(2.)
= Sii(}>v,
Sv(pii
now
/*
the expression,
and
v.
0p = 0op + VyjO (349, XII.), the condi-
and VIII. take the forms,
tions VI.
.
(Sv(pov
which are both
S/i0o/*)2
satisfied
+ 4S ()u0ov)2 > 4 (Sy/ivy,
generally
being, that the quadratic
when y = 0,
V. may happen
to
or
and
= ^' =
become an
^0
X.
;
identity^
Syfiv
the only exception
by
all its coefficients
but the opposite inequality (to VI. and IX.) can never hold good, that
to say, the roots of that quadratic can never be imaginary, when
is thus
self-
vanishing
is
+ z-iv)
if
= - (yiyz + 2122)
if
at least for this particular pair of vectors,
IX
= y^/i + z^v,
becomes,
and the corresponding directions of p will be rectangular,
is,
<pp
zfi,
= y2Si;0/A + yz(Sj/0i'-S/i0/i)-2S/i0i';
VI.
that
units, following in all
Writing then,
\p = yv
therefore,
I.
and vec-
<p.
V be a system oi three rectangular vector
V.
M= 0, for this
and therefore of the necessary reality of

cuhic^ Mq = 0, which is formed (354, IV. or
;
respects the laws (182, 183), of the
IV.
this equation III.
self-conjugate part ^0 of the general linear
as
|,
p^ are
also will follow a
reality of the scalar roots of the cubic
case of self-conjugation of
the roots of that other
and
the general existence (354) of a system of
and rectangular directions^ which satisfy

= 0/3, being thus proved anew : whence
new proof of the
lines p\
465
conjugate.
On the other hand, when y is actual, or ^'p not generally =0p, the condiX. of rectangularity can only accidentally be satisfied, namely by the given or
fixed line y happening to be in the assumed plane of p, v; and when the two direc(3.)
tion
when
tions of p are thus not rectangular, or
have only to suppose that the square of
the scalar
this scalar
Sypv
does not vanish,
we
becomes large enough, in order to
render (by IX.) those directions coincident, or imaginary.

= <p, or y = 0, we may take /x and v for the two rectangular direc(4.) When <l>'
tions of p, or may reduce the quadratic to the very simple form yz =
;
but, for this
purpose,
we must
establish the relations,
XI.
(5.)
shall
And
if,
at the
same
time,
Sp(pv
= Sv^/i = 0.
satisfies
have these other scalar equations,
3 o
the equation
III., so
that <p\
||
\,
we
466
XII.
whence
= S/i0\ = Sv0\ = SX^/i = S\0j/
Yv\
||
XIII.
or,
X,
^/x
and
fi,
<pv
YXfi
\\
||
v,
= VX0\ = Yfi(pfi = Yv^v
\\
111.
[bOOK
V thus forming (as above stated) a system, of <Aree real and rectangular roofs
;/,
of that vector equation III.

(6.) But in general, if III. be satisfied by even two real and distinct directions
of p, the scalar and cubic equation
can have no imaginary root; for if those
M=
two
^p:
tient
then
p,
the third root
is
common
give one
ci
and
also real
real
and
will take the form,
XIV.
is
easily
shown from
have
= (j)p
and
an imaginary
so
the
see, at
same
time,
deduced from the symbolical cubic 350,
Mq =
established.
M=
has all
which
p,
by
roots real, if
its
(as above stated) can never have
how
i\\Q
scalar cubic
M=
might have been
or from the equation 351,
I.,
if
^p
= - cp,
supposition,
then
(p^p
= c^p,
is diff'erent
<p^p
from
zero.
the quadratic Y. fails to assign any definite values to
2, or
in the given plane to p, this case is evidently distinguished

.
as the con-
of indetermination, above alluded
(9.) Finallj', as regards the case*
XV.
I.,
by any actual p namely by ob= - c^p, &c., and therefore Mp = 0, in
dition for the vector equation III. being satisfied
serving that
it
root.
And we
(8.)
and pa
at least two equal roots, since
and formulae already
therefore that the equation
pi
0p = ap,
that 0p must be of
- ci + Spip2i3),
(c
(7.) It is then proved anew, that the equation
^'p
for the quo'
C2,
for that quotient, then
the plane of pi, p2

will
= (c - ci)2
principles
and
on the other hand, the two directions
if,
and scalar value, such as ci,
= (0 + <;i)p = 0, for every Ziwe in

cip + j3Spip2P, and the cubic
or $ip
values, ci
real roots of the cubic, of which therefore
C2 are <mjo
the form,
as
and scalar
directions give two unequal but real
S/i^/x
any
by the
to,
when
definite directions
condition,
= Sr0v,
in combination with the equations XI.
356.
which
by the
linear
Syinbolic
and
and Cubic Equation
(350),
vector symbol 0, suggests a Theo'
of Geometrical Deformation, which may be thus enunciated
rem'\
of
The existence of the
is satisfied
It will
be found that this case corresponds to the circular sections of a surface

while the less particular case in which <p'p = <pp, but not
the second order;
= Sv^v, so that the two directions of p are determined, real,

Sfi^fi
corresponds to the axes of a non-circular section of such a surface.
f This theorem was stated,
and
rectangular,
nearly in the same way, in page 568 of the Lectures;
and the problem of inversion of a
linear
and vector function was
treated, in the few
preceding pages (559, &c.), though with somewhat less of completeness and perhaps
of simplicity than in the present Section, and with a slightly different notation. The
general form of such a function which was there adopted may now be thus expressed
0p = 2/3Sap + Yrp,
the resulting value of
was found
r being a given quaternion
to be
(page 561),
CHAP.
II.]
THEOREM OF SUCCESSIVELY DERIVED
LINES.
467
*'
yy, hy any given Mode, or LaWy of Linear Derivation, of the

kind above denoted by the symbol ^, we pass from any assumed Vector p to a Series of Successively Derived Vectors, /î, /jg, pz,
or ^'/),
>
^^d i/i ^y constructing a Parallelepiped^ we decompose any

0V> ^V>
Line of this Series, such as p^, into three partial or component lines
'>
mp, -m'pi, m"p2^ in the Directions of the three which precede it, as here
of p} Ply Pi th^^ t^^ Three Scalar Coefficients, m, - m', m", or the Three
'>
Ratios which these three Components of the Fourth Line p^ hear to the
Three Preceding Lines of the Series, will depend only on the given Mode
Law
of Derivation, and will he entirely iiidependent of the assumed

and
Direction of the Initial Vector.^^
Length
or
(1.)
As an Example
L
which answers
of such successive Derivation, let us take the law,
= 0p = - Yftpy,
pi
p3
= 02p = _ V/3piy,
to the construction in 305, (1.), &c.,
Treating them at
are unit-lines.
as
first
&c.,
when we suppose that
any two given
vectors, our general
and y
method
/3
conducts to the equation,

II.
with the following values of the
TIL
as
may
in 354,
coefiicients,
.m = -/32y2s/3y,
new
and -
be seen, without any
XXXIII.,
= mp - m'pi + m"p2,
j03
to 0,
ft,
m" = S/3y
m'=-|32y2,
by merely changing
calculation,
g, X,
and
p,,
y.
for comparison with 305, that

(2.) Supposing next,
IV.
so that
j8,
y are unit
values III. become,
and the equation

VI.
II.,
.
lines,
.i82=y2=-l,
and
I is
and
S/3y=-Z,
the cosine of their inclination to each other, the
Y...m = l,
w" =
m' = -l,
/;
connecting /or successive lines of the
= ?p + pi-^p2,
p3
or
VII.
series,
takes the form,
p3-pi = -K>2-p);
m = 2:Saa'a"Si3"/3'/3 + 2S {rY aa .Y ^' ^) + SrSSa/3r - SSarS/Sr +

and the auxiliary function which we now denote by
m^-i<T
where the sum
A student
sions
= ;//<T=
of the
might
by the
SVaa'S/3'/3(T
two
find it
last
i/'c
SrTr2
was,
+ 2V. aV(V/3(r.r) +
terms of
an useful
(VcrrSr
- VrSerr)
might have been written as
(xrSr
- rSor.
exercise, to prove the correctness of these expres-
Section.
principles of the present
One way
of doing so would be, to
to ôP + Vyp and

2/3Sop and r as respectively equal
transform m and i^ff, as above written, into the following,
treat
+
-aro-S(yt)Cô + 0(y + 0, andô<T-(y + )S(y
and â of the
that is, into the new values which the
takes the new value, $p = (ô + c) p + V(y + ) p.
)(T
c;
which would
+ V(7(0o + c)(y+O;
Section assume,
when $p
468
a
result
which agrees with 305,
interval P1P3
(2.),
since
we
III.
[bOOK
there found that
if
= op,
&c., the
= Z x PP2.
was
function (347), or the

(3.) And as regards the inversion of a linear and vector
return from any one line pi of such a series to the line p which precedes it, our general method gives, for the example I., by 354, (12.),
VIII. ..;|/pi=
1/3 (/3ypi
+ pi^y)y.
and
a result which
it is
easy to verify and to interpret, on principles already explained.
We
are now prepared to assign some new and gene357.

Forms, to which the Linear and Vector Function (with real
constants) of a variable vector can be brought, without assum-
ral
ing its self-conjugation ; one of the simplest of which forms

the following,
I.
qo
(j)p
= Yqop + VXp/i,
with
r.
qo
is
g + y;
being here a real and constant quaternion, and X, p, tivo real

vectors, which can all be definitely assigned^ when
and constant
is
given : except that A and ju may be
295,
VII.), and that either may be multiplied
interchanged (by
It will
if
other be divided by the same.
the
by any scalar,
the particular form of ^
follow that the scalar, quadratic, and homogeneous function of

a vector, denoted by Spcpp, can always be thus expressed
:
II.
Sp(pp=gp'^ + SXpiup
or thus,
ir.
Sp<j)p
=gp^+ 2SXpSpp,
if
g'
SA/x
a general and (as above remarked) definite transformation,

which is found to be one of great utility in the theory of >Swr-
faces* of the Second Order.

(1.) Attending first to the case of seZ/"-co7i;Môfe functions 0o|O, from which we
can pass to the general case by merely adding the term Vyp, and supposing (in virtue of what precedes) that aia2az are three real and rectangular vector-units, and
ciczcz three real scalars (the roots of the cubic
Mq
0),
such that
* In the
theory of such surfaces, the two constant and real vectors, \ and
called the cyclic normals.
have the directions of what are
/i,
CHAP.
in.
RECTANGULAR AND CYCLIC TRANSFORMATIONS. 469
II.]
= Cô + ci)ai=0,
îai
we may
02a2
(0o
<?2)a2
= O,
âaa
= (ô +
C3) 03
0,
write
IV.
and therefore
V,
so that
VI.
azSazp + aaSaap),
= ciaiSaip + C2a2Sa2p + czozSazp
= {c2 - ci) a2Sa2p + (C3 - ci) asSaap,

(îp
=
icz-Ci)azSazp + ici-C2')aiSaip,
}(p2p
- C3) aiSaip + (c2 - C3) a2Sa2p,
=
<pop
= - (aiSaip +
^3P
(ci
the binomial forms of ^1, 02, ^3 being thus put in evidence.

have thus the general but scalar expressions:
(2.)
We
- p2 = (Saip)2 + (Sa2p)2 + (Sa3p)2

= Sp56op = ci(Saip)2+ C2(Sa2p)2+ csCSasp)^
= - cip2 + (C2 - ci) (Sa2p)2 + (C3 - ci) (Sasp)^
= _ C2p2 - (C2 - Ci) (Saip)2 + (C3 - Cg) (Sasp)^
- (C3- C2) (Sa2p)2
- ci)
=VII.
VIIL
Sp<pp
C3p
which
iu
and
in general permitted to
IX.
< C2 < C3,
ci
(Saip)2
(ca
it is
assume that
X.
or that
C2
- ci = 2e2,
C3
- C2 = 2e'2,
introduced for a
being real scalars, and the numerical coefficients being
e'
which
tive of convenience
Comparing the
(3.)
mo-
will presently appear.
but one of the expressions VIII. with
last
II'.,
we
see that
we may bring Sp^p to the proposed form II., by assuming,

= - eai + e'as, 5^ = SX/i - C2 = - 1 (ci + C3),
XL \ = eni + e'a3,
.
because SX/x
But
(4.)
;li
C3
in general
(comp. 349, (4.) )
XII.
that
| (ci + cz).
= e2 _ e'2 =
.
Sp0p = Sp0'p,
we
from
II. that 0op

5'P
thus the transformation
and
0'
p,
be equal
I. is proved to be possible, and real.

with the values XI. of X, /u, g, the expression,
(5.) Accordingly,
XIV.
XIII.
0op=<^V
; we can therefore infer
+ V\p/u, because VXp/z = V/^pX = its own conjugate; and
unless the self-conjugate parts of ^
is,
cannot have, for all values of
unless
0op
=5'|0
+ VXp/t = p{g- SXjw) + XS/ip +
/itSXp,
becomes,
XV.
0op
=-
C2P
+ (e'a3 4
eai) S(e'az-eai)p +{e'az
= C2P 2e2aiSaip + 2e'2a3Sa3p

which agrees, by X., with VI.
(6.) Conversely if ^r, X, and
0oVXju
roots
= p'VXjLt, where g' =g-
one of the three directions of

is
that of ai, az,
XVI.
we
are,
by VI.,
0,
ai, 02,
and the normal to the plane of X, fx must have

03 if then we assume, on trial, that this plane
;
and write accordingly,

.
X = aai +
a'az,
p.
= ba\ + b'az,
to seek for scalars aa'bh'
XVII.
but these give
S(e'az+eai)p
eai)
be constants such that 0op=5'P + VXp/*, then

hence - g' must be one of the three
fi
SX/z, as before
Mq
of the cubic
ci, Co, cz
2a6
XV in.
ci
C2,
{2aby =
(p%p
which shall
2a'6'
(26a')2
C3
satisfy the three conditions,
- cj,
= (c3 -
XSjwp + /iiSXp,
a6'
c^) (cz
6a'
- cj),
470
SO that if the tranaformation is to be a real one,
C3
c-z
we must suppose
are either both positive, as iu IX., or else both negative
we must
so
arrange the three real roots of the cubic, that
intermediate in value between the other two.

the values X.,
we
satisfy the conditions
XIX.
and are thus
/*,
[bOOK
led
a'
cz
ci
and
or in other words*
be (algebraically)
may
Adopting then the order IX., with
XVII. by supposing
= 6' = c',
that c^
III,
=6=e
that
back from XVI. to the expressions XI., as the only real ones for X,
possible the transformations I. and II. ; except that X and /*
and g which render
be interchanged, &c.
may
as before.
We
see, however, that in an imaginary sense there exist two other solutions
(7.)
of the problem, to transform (pp and Sip<pp as above ; for if we retain the order IX., and
ci or C3, we may in each case conceive the corresponding
equate 5^' in II'. to either
sum of two squares

tors
VIII. as heing the product of two imaginary but linear fac-
in
the planes of the two imaginary pairs of vectors which result being real, and
perpendicular respectively to ai and a^.

(8.) And if the real expression XIV. for (p^p be given, and it be required to pass
from it to the expression V., with the order of inequality IX., the investigation in
354, (12.) enables us at once to establish the formulae
XX.
XXL
ci
= -5r-TX/i,
C2=-^+SX/f,
= U(XT/i - /iTX),
ai
az = UVX/t,
C3
03
= -^+TX/i;
=
+ /xTX)
U(XT/x
fl
;
in which however
it is permitted to change the

sign of any one of the three vector
Accordingly the expressions XI. give,
units.
+ SX/A = 2e2 = c2-ci,

TX = T/ii, X-/i = 2eai,
TX/i
(9.)
We have also the two
XXII.
XXIII.
= 2e'2 = c3-C2,
=
= + 2ee'a2,
2ee'a3ai
VX/i
TX/i-SX/i
X+
for
any
identical transformations,
and may (among other ways)

p., p,
XX. and XXI., from II. and VIII.
three vectors, X,
duced, through the expressions
(10.) Finally, as regards the expressions VI. for 0ip, &c.,
responding forms of
.//p
by
i/zip, S:c.,
we have (comp.
if
354, (15.)
which are as usual (comp. 351, &c.) oi monomial form
= (pz(pzp = (c2 - ci)

= ^30ip = (cz cz)
}
\(/zp
= 0i^2p = (ci - C3)
4'3p
I
XXIV.
\pip
and Avhich
= ^ + C2;
= 2e'a3.
/4
- SX/i)-i,
SXp/ip = p2TX/i + { (SX/ip)2 + (SXpT/i + S)upTX)2 } (TX/*
= - p2TX^ - { (SX/ip)^ + (SXpT/4 - S/ipTX)2} (TX/i + SX/i)-,
SXp/ip
which hold good
pressions,
SX/i
verify the relations 354, XLI.,
we denote
:
ci)
a2Sa2p
- Cz) asSasp
(C3
(c2
the cor-
these other ex-
- C3) aiSaip
(ci
be de-
and several other parts of the whole
fore-
going theory.
358. The general linear and vector function ^p 0^ a. vector has been
seen (347, (!)) * contain, at least implicitly, ?ii?ie scalar constants ;
and accordingly the expression 357, 1, involves that number, namely
four in the term
V^'o/a, on account of the constant quaternion q^^ and

in
other
term
the
and \]fi,
Y^pfJi, each of the two unit-vectors,
five
as
two
and
the
tensor
as
one
more.
But
a selfscalars,
TX/a
counting
UX
'S
CHAP.
471
FOCAL TRANSFORMATIONS.
II.]
and vector function, or the self-conjugate part <pop of

the general function <pp, involves only six scalar constants ; either because three disappear with the term Y^p of (pp ; or because the conconjugate linear
2Vâ = 27 =
dition of self-conjugation,
(comp. 349, XXII. and 353,
form 2^S/3 (347,

equivalent to a system of three scalar equations, connecting the nine constants. And for the same reason the general quadratic but scalar function, S/j^/j, involves in like manner only six scalar
XXXVI.), which arises when we take

XXXI.),
for (pp the
is
constants. Accordingly there enter only six such constants into the
357, II., IK, V., VIII., XIV.; Ci, Cg, C3, for instance,
expressions
being three such, and the rectangular unit system
SpfP and
(pop,
ai, a^, a^
answer-
The
ing to three others.
following other general transformations of

not quite so simple as 357, II. and XIV., in-
although
volve the same number (six) of scalar constants, and deserve to be

briefly considered: namely the forms,
I.
II.
which
in
We
tors.
shall
two
a, h are
= a{Vapy + h(S^py;
= - aaVap + bjSSlBp
Sp<pp
0oP
and
real scalars,
down
set
merely
a, y3
two
are
real unit-vec-
the leading formulse, leaving the
reader to supply the analysis, which at this stage he cannot find

difficult.
(1.) In accomplishing the reduction of the expressions,
Sp^p = ci(Saip)2+ C2(Sa2p)2+ C3(Sa3p)2,

and
= ciaiSaip -f
ôp
to these
new forms
I.
and
cia-^âzQ
found that,
II., it is
must be that root of the scalar cubic iVfo=
if
4-
357, VIII.
357, V.,
0303803^,
the result
is
to be a real one,
the reciprocal of which is algebraically

It is therefore convenient
intermediate^ between the reciprocals of the other two.
here to assume this new condition, respecting the order of the inequalities,
III.
which
C3,
cri>C2->C3-i;
with the arrangement 357, IX., if the three roots

be all positive, but will be incompatible with it in every other case.
will indeed coincide
(2.) This being laid
down
the (real) values of a, 6, a,
(or even,
may
/3
IV.
V.
in
0,
if
we choose, the opposite order being
be thus expressed
a = C2, 6 = ci C2 + C3
= xai + zas, /3 = x'a\ + nâz
.
which
VI.
c\
VII.
if
^=
C3
h{xx'
cr'
zz')
C3
= - fcSa/3 = (say)
h'
ci, C2,
taken),
472
VIII.
IX.
arH y2 = x"^ +y' = 1
= c-iXz
- a6 CSai3)2
cix-i + C322 = C1C2-IC3 = 6-'6'2 = 6 (Sa/3)2,
C1C3 =
= -6/3Sa/3 = Cia;ai + c^zaz; &c.
XII.
6'j3
6'2
C1C2-IC36
= ci2a;2 + C3222
X.
XI.
And
(3.)
hx'z'
HI.
XIII.
[bOOK
there result the transformations
=(ci C2)aiSaip +
02P
= -C2(a;ai+za3)
S(a;ai
C2) 0380310
(cs
+ 203)?
(a;ciai
+ ^6303)
S(a:ciai
20303)05
C1C3
XIV.
ôp
= ciaiSaip + C2a2Sa2p + 030380310
= C2 (aoi + 203) V(a;ai + 203) p +

XV.
Sp^p
=- C2(V(arai
-}-
(ajcioi
C1C3
203)p)2+
+ ZC3O3) 8 (a;cioi + zcza^p
(S(a;ciai
-f
2C303)p)2;
C1C3
which
last, (/"cics
XVI.
a;2
and
he positive, gives this ofêr real form,
Sp^p =
N { S(a;ciai + 2C303) p + (ciCs)^ V(xai + Z03) p }
z^ being determined
by the expressions VI.
Those expressions allow us to change the sign oi z: x, and thereby to deter(4.)

mine a second pair of real unit lines, a and (S', which may be substituted for o and (5
the order of inequalities III. (or the opposite order), and the
in the forms I. and II.
;
values IV. of a and
formations
XVII.
XVIII.
6,
remaining unchanged.
We have therefore
the double trans-
- C2(Vop)2 +
8p^p =
0op
(Ci
-C2 +
= czaYap + (ci - cg +
C3) (S/3p)2
C3)/3S/3p
= - C2(Vo2
+ (ci-c2+C3)(S,8'p)2;
= cgo'Va'p + (ci -C2 + C3)i3'S/3'p.
(5.) If either of the two connected forms I. and II. had been given, we might
have proposed to deduce from it the values of C1C0C3, and of 010203, by the general
method of this Section. We should thus have had the cubic,
XIX.
= Jlfo = (c +
..
and because the quadratic
XX.
(c
o){c2+(a-&)c-a&CSoi8)2};
a)-iiVfo=
may be
thus written,
(c-i+a-i)2 (Soi8)2-(c-i4 o-i) (a-iS.(o/3)2
+6
i)
a-2(Vo/3)2
0,
gives two real values of c'^ + a-^, one positive and the other negative ; if then we
in the order III., we have the
arrange the reciprocals of the three roots of
it
Mo=
expressions,
^^^-
f<'^
\f3
= K^-'') + l^^'^(.o-'''+â-î'^S.(al3y + b-^)',

=
(6
- a) - lab
V(o- + 2a-i6-iS. (o/3)2 + 62)
C2
= -a;
the signs of the radical being determined by the condition that (ci- C3) a6(So/3)2
= cj-i _ C3-1 > 0. Accordingly these expressions for the roots agree evidently with
:
the former results, IV. and XI., because S (afS)^ = 2 (Soj3)2 - 1.

for
(6.) The roots ci, C2, C3 being thus known, the same general method gives
.
the directions of 01, 02, 03 the versors of the following expressions (or of their negatives)
CHAP.
SECOND FOCAL TRANSFORMATIONS.
II.]
XXII.
= ac3-
i//ip
'
(csa
473
+ 6/3Sa/3) S (câ + b0Sa(3) p
U2p = a6Va/3S/3ap;
= acri (cio + 6/3Sa/3) S (cia + 5/3Sa/3) p
ipsp
of which the monomial forms
XXII'.
ai
= + U(c3a +
aa =
6^Sa/3),
UVa/3,
(7.) Accordingly the expresssions in (2.), give (if
XXIII.
again be noted, and which give,
may
csa + J/3Sa/3 = (03 -
as
we suppose
Va/3 = (x'z -
ci)xau
= U(cia + 5/3Sa/3).
azai
= + 02),
cia
xz') az,
6/3Sa/3
{ci
Ci)zaz
and as an additional verification of the consistency of the various parts of this whole
theory, it may be observed (comp. 357, XXIV.), that
XXIV.
- C2)
(c3
(co
XXV.
(C3
XXVI,
we have
Ci) a'
(C3
a),
= 5/3a/3 - aa, (cg - ci) /3' = aaj3a = (6â/3 - aa)2 = (/? - bfif
ci)2
true for
is
may
made by
any vector
any two unit
p,
a, a',
(c3-ci)a'
a = ci +
a+
At the same time we have
b(3,
;
62)
and any two
/3,
scalars a, 6.
C3),
a'
(ci -cj)/?'
2xai,
a'
(ci
= Izaz
and as the external
C3)/3- 2aaSa/3;
angle, of the lines a, a'.
these other expressions,
- ci) (/3 + 18') = 2 (ci^ - aaSa/3),

(ci
which can
easy to
write,
.
so that a\ bisects the internal angle,
lines a,
= (c3 + ci)a + 26/3Sa/3,
XXIX.
(10.)
Ca).
whence, by XXIII., we may
XXX.
evident from (4.), that ai, 03 must be the bisectors of the

and also of those made by /3, (3' and the expressions XXV.
be thus written (because 6
XXVIII.
is
it is
(9.) Accordingly
angles
it
the following equation,
XXVII.
(a(Vap)2+6(Sj3/>)) (a2+ 2a5S.(a/3)2 +
= a (V(6/3a/3 - aa)py + 6(S(aa/3a - 6/3)p)^
which
C3) (C3
transformations, XVII. and XVIII.,
(8.) As regards the second

prove that we may write,
so that
- cg), a3(Va/3)2
(ci
- aci-\cia + bl38a(3y = (ci -
- ac3-i(c3a + h^â^y = (cg - ci)
(C3
- ci)
(/3
jS')
= 2 (cg/S - aaSa/3)
easily be reduced to the simple forms,
XXXI.
.^-{^^^
Ix'ai,
(i-(i'=2z'az,
with the recent meanings of the coefficients x' and z'.

sake of obtaining real transformations,
(11.) And although, for the
we have
supposed (comp. III.) that
XXXII.
(ci-i
- C2-O
(C2-'
C3-1)
>
0,
= a;ai + za3
between the three unit vectors aa\az,

whereof the two latter are rectangular, gives 332 + 22= 1, as in IX., so that each of
the two expressions VI. involves the other, and their comparison gives the ratio,
because the assumed relation o
XXXIII.
x2
22
= (ci-i 3 P
C2-1)
(c2-i
- C3-'),
474
yet
we
see that, without this inequality
XXXII.
[bOOK
III.
existing, the foregoing transforma-
good in an imaginary (or merely symbolical) sense : so that we may say,

in general, that the functions Sp^p and 0op can be brought to the /orms I. and II.
in gix distinct ways, whereof two are real, and the /owr others are imaginary.
tions hold
(12.) It may be added that the

the following,
XXXIV.
first
equation
XXII. admits
of being replaced
by
;//ip=-Jcr'(tfii3-aaSa/3)S(<Ji/3-aaSa/3)/3,
with a corresponding form
for
;//3jo;
and that
thus, instead of XXII'.,
we
are at
liberty to write the expressions,
XXXV.
U (ci^ - aaâp),
ai
for the rectangular unit system,
359. If
we
I.
and
II.
we
call, as
= UVa/3,
I.
ag
= U (c3/3 - aaSa/3),
or II.
naturally may, the expressions
+ Câ^â^p + CâSo'zp^
= Ci(Sairt' + 02(802/9)2 + C3(Sa3/))2,
(pop
8/30/j
az
deduced from
CittiStti/a
357, V.,
357, VIII.,
the Rectangular Transformations of the Functions (pop and Sp(pp,

then by another geometrical analogs/, which will be seen when we
come to speak briefly of the theory of Surfaces of the Second Order
^
we may
call the expressions,
III.
and
IV.
0o/>
Spi>p
=gpi- y^p^,
= gp^ + S\pfip,
357, XIV.,
357,
the Cyclic* Transformations of the same two functions; and

say that the two other and more recent expressions,
V.
and
VI.
0o/>
= - auYap + b(3S^p,
= a(Vapy + b(S^py,
Sp(pp
are Focal j Transformations of the same.
how
how
(358)
(357)
to
We have
II.,
may
358,
358,
II.,
I.,
already shown
ones; and also
exchange rectangular forms with cyclic

from rectangular expressions to /oca/ ones, and
but it may be worth while to consider briefly the mu-
to pass
:
reciprocally
tual relations which exist, between cyclic and /ocaZ expressions, and
the modes of passing from either to the other.
first
(1.) To pass from IV. to VI., or from the cyclic to the/ocaZ form, we may
accomplish the rectangular transformation II., with the values 357, XX., and XXL,
of ci, C2, C3, and of ai, a^, crs, the order of inequality being assumed to be
Compare the Note
to Art. 357.
of 358, are the two real focal

f It will be found that the two real vectors a, a',
lines of the real or imaginary cone, which is asymptotic to the surface of the second
order,
Sp0p
= const.
PASSAGE FBOM CYCLIC TO FOCAL FORMS.
CHAP.
II.]
and then
shall
VIL
475
as in 357, IX.
C3>C2>ci,
have (comp. 358, XV.) the following expressions
= { S p (ci^<U\ - U/i) + caÛX + U/)) } 2

- {V. p(cii(U\ + U/i) + caÛX - U,i))}2
*
4Sp^p = { S p ((- ci)i (U\ U/i) + (- C3)i (U\ + U/t)) }
+ {V.p((- cOi (UX + U/i) + (- C3)i (UX U/i))}2
- co)2 Sp0p = { V. p (cai VX/i + (- C2)i (XT/i + /iTX)) }
(C3
+ { S p ((- C2)i VX/i - Cgi (XT/i + /iTX)) } 2
(c3-ci)2Sp0p=-{V.p((-ci)iVX/i + C2KXT/i-/iTX))}2
- {S.p(- C2iVX/i + (- ci)i (XT/i -/iTX))}2;
VIIL
ASpcpp
Vlir.
IX.
X.
in
which
XL
= -$f-TX/i,
ci
remembered that (by 357, XX.),
to be
it is
C2
= -5' + SX/i,
C3
= -g-T\ii;
+ SXp/tp
Sp^p, if ^, X, /i, p be real. And in Mts symbolical sense, although they have
been written down as /owr, they only count as three distinct ybcaZ transformations,
and of which
aZZ are symholically true, or give (as in
IV.) the real value gp^
for
of a given
and real
cij
die form
because the expression VIII'.
is
an immediate con-
and other formulae IX'. and X'. might in like manner be at once
sequence of VIII.
derived from IX. and X.
;
But
(2.)
if
we wish
to confine ourselves to real focal forms, there are then
cases to be considered, in each of
IX. X.
to be adopted, to the exclusion of the other three.
is
XIL
if
the /orw VIII.
XIIL
XIV.
If
C3
C3
if
if all
and therefore
0,
the only real one.
ci"
When
same condition of
/3
fa
in
VI.
may
or in the case
when VIII.
be thus expressed
l(^]\uX-U/i)
XVII'.
U/i);
be written (comp.
4Sp0p = N { (C3i + cii) p .UX + (C3i cii) U/i p } ;

4Sp^p =N {(C3i+ cii) UX.p + (cai ciJ)p.U/i};
.
XVIII.
we
may
10
= i (C3* +
cii)
UX,
co
= 1 {c^^ - c^) U/i,
shall have, briefly,
XIX.
the real fo-
|^^y(UX+U/i);
(4.) In the same case VIII., the expressions for 4Sp^p

358, XVI.) under either of these two other real forms :
XVII.
is
+
+
|/3=*(^'y(UX-U/i) ^(^fy(UX
6 = ci - ci + C3 as before (358, IV.).
with
be
reality.
all the roots c axQ positive,
XVI.
we
0,
If
cal form, the unit lines a,
so that if
Thus,
> cf'^ > C3-1 >
>
adopted, under the

(3.)
> C2 > ci >
> ci, C2" > C3-1 > > ci'^, then X. is the real form.
C2 >
> > C3 > ci, C3-1 > > ci-i > C2-1, the only real form is IX.
XV.
> C3 > C2 > ci,
> ci'i > ci-^ > 03"
be negative, then VIII'. is the form to
the roots of the cubic Mq =
.
Finally
C3
is
that
is,
^bwr
which some one of the /owr equations VIII. VIII'.
Sp0p = N(eop + p-o)=N(pio+K:op)-
476
and
= ^ (ci-i + <?3-i) UX,
shall then
which
III.
Or we may make
(5.)
XX.
[bOOK
= i (ci'i -
-whence
tf3-i)U/i,
k^
t^
= cf^ c^'^
have the transformation,
be compared with the equation 281, XXIX. of the ellipsoid, and for the
which form, or of its two vector constants, i, k, it is necessary that the
may
reality of
roots c of the cubic should all be positive as above.

(6.) It was lately shown (in 358, (8.), &c.), how to pass from a. given and real
focal form to a second of the same kind, with its new real unit lines a', j8' in the
same plane as the two old or given lines, a, (3 but we have not yet shown how to
pass from a focal form to a ci/clic one, although the converse passage has been re;
two
that the
given values
We
(7.)
now
Let us then
cently discussed.
scalar constants a, b,
suppose that the form VI. is real and given, or

and the two unit vectors a, /3, have real and
and let us seek to reduce this expression VI. to the earlier form IV.
might, for this purpose, begin by assuming that
XXII.
> C2-1 > C3-\
ci-i
as in 358, III.
which would give the expressions 358, XXI. and XXII., for C1C2C3 and aiaaaa, and
so would supply the rectangular transformation, from which we could pass, as before, to
(8.)
form
is
the
CT/clic one.
to vary a little the analysis, let us now suppose that the given focal
some one of the four following (comp. (1.) )
But
XXIIL
XXIV.
= (S/3op)2 - (Vaop)2; XXIIF.

Sp<l>p
= (S/3op)2 + (Vaop)2
XXI V.
Spipp
Sp^p = (Vaop)^ (S/3op)2

Sp^p = (Vaop)2 (S/3op)2
which ao and (3o are conceived to be given and real vectors, but not generally unit lines; and which are in fact the four cases included under the general
form, a(Vap)3 f 6(S/3p)2, according as the scalars a and 6 are positive or negative.
in each of
It will be sufficient to consider the
two others
two
cases,
XXIII. and XXIV., from which the
will follow at once.
For the case XXIII. we easily derive the real
(9.)
XXV.
Sp^p
= ^p2 +
XXVI.
where
X = j3o +
and the equations 357,
cyclic transformation,
= (S/3op)2 - (Saop)2 + aoV^

= S(/3o+ ao)p.S(/3o- ao) p + oV
SXpfip
ao,
ft
^{g- SX/z)p2 + 2SX/iS/xp,
= ()8o - ao),
5^
IW +
/So')
(9.) enable us to pass thence to the two imaginary cyclic
forms.
(10.) For example,
XXVII.
we may
the proposed function be (comp. XIX.),
if
Sp^p =N(top + pKo) = (S(io+
-(V(to
fco)p)2,
write
ao
= <o-'co,
/3o
XXVIII.
To
treat the case
city the subindices
0,
X=2to,
to+Ko,
and the required transformation
(11.)
ico)p)2
is
fi=Ko,
g=
io'^
+ Ko^;
(comp, 336, XI,),
N((op4
p/fo)
XXIV, by
= ('o2+/co2)p2 + 2StoP'cop.
our general method,
we may omit for
and write simply (comp. V. and VI.) the expressions.
simpli-
CHAP.
XXIX.
in
PASSAGE FROM FOCAL TO CYCLIC.
II.]
= - aYap + /3S/3p,
0p
which however
sions
a and
to be observed that
it is
Sp^p
= (Vap)2 + (S/3p)2
though real
(3,
aVap = aSap a^p, we
Hence because
unit lines (8.).
XXX.
and
477
vectors, are not
easily
now
form the expres-
XXXI. ..mâ' (Sai3)2, m' = a^ (a^ - /32) - {â^f m" = /S^ - 2a2
= Ya(3S(3ap - a^(âYap + /3V/3p) + a^p
(i//p
= Vap/3Sa/3 + a (a^ - /S^J Sap,
XXXII.
j
XP = - (aSap + /3S/3p) + (/32 -a'^)p;
and therefore
XXXIII.
M=(c- a^) (c2 + (/32_a)c-(Sa/3)2),
and XXXIV.
^p = Vap/3Sa/3 + (/32 - a2) (cp - aSap) - c (aSap + iSS^Sp) + c2p
= (a(a2- ^2 _ c) + /3Sa/3)Sap + (aSajS - c/3)Si3p + (c2 + (i32- a2) c - (Sa/3)2)p.
,
(12.) Introducing then a real and positive scalar constant,
r,
such that
= (a2 - /32)2 + 4 (Sa^)2 = {a^ + |82)2 + 4 (YajSy

= a* + (ai3)2 + (|3a)2 + /S* = a4 + 2S. (a|3)2 + /34
= a-2 (a3 + /3a/3)2 = /3-2 (/33 + ai3a)2 = &c.,
XXXV.
in
r*
which (by 199, &c.),
= (Sa/3)2 +
(a/^)3
the roots of iJ/=
(Va|3)2
= 2 (Sa/3)3 - a2/32 = 2 (Va/3)2 + a2|32,
admit of being expressed as follows
XXXVI.
ci
- /32 + r),
(a2
cg
and when they are thus arranged, we have the
XXXVIL
Ci
> > C3 > C2,
may
of
^p
C3
= i (^ - i^^ -
be variously transformed and
r)
inequalities,
> > C2-' >
ci-i
are the three
verified,
<?3-'.
monomial
" C3-Xac3 + /3Sa/3) S (âc^ + ^Sa/3) p,

/'^'^
= cri(aci + j3Sa/3) S (aci + /3Sap) p
I'/'ap
XXXVIII
which
The corresponding forms
(13.)
= a3,
expressions,
^2P
= Ya(3S(3ap,
and give the three following
rect-
angular vector units,
XXXIX.
in
ai
= U(ac3 + /3Sa/3),
connexion with which
+ /3Sa/3)
easy to prove that
T (ac3 +
(3Safi)
TYap =
I
03 = U(aci
= (- 03)* (ci - 2)^ (ci - c^^ = r(ci- 2)^ (- Ci)^,

- 02^ (cs - ca)*
(ci
- C2)i (ci - ^3)^ = r =
+
( T(aci
j3Sa/3)
ai^cs
(cs
02)^ Ci^
XL.
it is
a2 = VYa(3,
the radicals being all real,
We
(14.)
have thus,
by XXXVII.
for the given focal form
XXX.,
the rectangular transfor-
mation,
XLI.
Sp0p
= (Vap)2 + (S/3p)2
= g^(S(g3+i3Sa/3)p)2
g2(Sa/3p)2
f3(S(aci
+ /3Sa/3)p)2
'
-C3(Cl-(;2)r2
{Ci-Co){Cz-CoS)
C\{cz-C2^r'^
or briefly,
XLII.
Sp0p
= (Yapf +
(S/3p)2
+
in
which the
first
term
is positive,
= ci (S. pU(a<?3 + )3Saj3)p)2

+ C3(S .pU(a(;i + /5Sa/3))2
a2(S. pUVa/3)
but the two others are negative, and
the roots of the quadratic,
XLIIL
= c2+(,S2-a2)c-(Sa/3)2.
c\, C3
are
478
We have
(15.)
XLIV.
acz
+ /3Sa/3
It
- aSa/3,
= (Sa/3)2
aci
jSci
C1C3
may
XLV.
111.
also the parallelisms,
because
and
[bOOK
+ jSSajS
||
I3c<i
- aSa(3,
therefore write,
.
Sp(pp
= (Vap)2 + (S/3p)2 = ci(S

+
a2 (S
pU(/?ci
aSa/3))2
pU Va/3)2 +
f?3
(S pU(/3c3
.
- aSa/3))2
while
XLVI.
T(/3ci-aSa/3)
r = (ci
and
<?3')i,
with
new
XLVII.
T(/3c3- aSa/3) =r(-c;3)i ((?3-^2>,
real radicals as before.
(16.) Multiplying then

tain this
= rcii(ci-C2>,
by r2(TVa/3)2,
or
by
(ci
Co)
(ci-cz) (C3-C2),
we
ob-
equation,
(ci
- cs) {(TVa/3)2 ((Vap)2 + (S/3p)2) - a2 (Sa/Sp^}

= (^3 - a2) (ciS)3p - Sa/3Sap)2 - (ci - a2) (^gS/Sp -
aSa/3)2
only another way of expressing the same rectangular transformation as bebut has the advantage of being freed from divisors.
which
fore,
is
(17.) Developing the second

obtain this new transformation
member
of
XLVIL, and
we
dividing by C1-C3,
XLVIII.
in
= - (Ya(Sy ({Yapy + (S/3p)2)

= a2 (Sa/3p)2 - (Sai3)2 (Sap)* + 2aSai3SapS|3p +
(TVai3)2 Sp^p
which we have written
for
XLIX.
(18.)
The expressions
and accordingly, when
C= C1C3 - a2 (ci + C3).
XXXVI.
L.
becomes an
C(S/3p)2 ;
abridgment,
for ci, C3 give thus,
C = -a*-(Va/3)2;
this value is substituted for
identity, or holds
good
in
XLVIII., that equation
for all values of the three vectors, a,
(S,
as
be proved* in various ways.
may
(19.) Admitting this result, we see that for the mere establishment of the equaXLVII., it is not necessary that ci and C2 should be roots of the parîcw/arjua-
tion
dratic XLIII.
It is sufficient, for this purpose, that they should be roots of any qua-
dratic,
LL
c2+^c + P=0,
between
its coefficients.
with the
reZaîora
LII.
But when we combine with
â2
+^+ a4 + (Va/3)2 = 0,
this the condition ofrectangu-
-^ ai, or
larity, as
LIII.
we
=S
(ci/3
- aSa/3)
(C3/3
- aSa/3) = A (Sa/3)2 +
S(i^
a(Sa/3)2,
obtain thus a second relation, which gives definitely, for the two coefficients, the
values,
LIV.
and
so conducts, in a
*
left,
Many
new way,
^ = /32 - a8, B=- (Sa/3)2

to the equation
XLIII.
such proofs, or verifications, as the one here alluded
at this stage, as exercises, to the student.
to,
are purposely
CHAP.
AND IMAGINARY.
IDENTITIES, REAL
II.]
479
(20.) In this manner, then, we might have been led to perceive the truth of the
rectangular transformation XLVII., with the quadratic equation XLIII. of which
ci and C3 are roots, without having previously found the cubic XXXIII., of which
the quadratic
a factor, and of which the other root
is
is C2
= a^.
But
if
we had not
employed the general method of the present Section, which conducted us to formers*
that cubic equation, there would have been nothing to suggest the particular form
XLVII., which could thus have only been by some sort of chance arrived
(21.) The values of aia2a3 give also (comp. 357, VII.),
LV.
that
is,
p2
= (S
pV((3ci
- aSa/3))2 +
(S pUVa/3)2
.
+ (S pVQScz - aSalS^y
.
by XL. and XLVL,
LVI.
ciC3(ci
- C3)
(p2(Va/3)2
- (SaiSp)^) = 03(03 - a^) (ciS/3p - Sa(3Sapy

- a (ci - a2) {czS(3p - Sa(3Sap)
and accordingly the values XXXVI. of ci,

this last equation under the common form,
Comparing
(22.)
by 357,
see,
(6.),
(6.) at present sought, the plane of \,
We
as in 357, (3.), &c.).
the equations
formation
03
/i
enable us to express each
ciCz(ci C3)
the recent inequalities
rangement 357, IX., we
(aS(3p
member
of
/3Sap)2.
ci>C3>C2 (XXXVII.) with
the ar-
that for the real cyclic transformation
is to
be perpendicular to 03 (and not to a2,

- Sa(3Sapy between
are therefore to eliminate (c^SfSp
XLVII. and LVI., which
gives (after a few reductions) the real trans-
LVIL
=
which
at.
- Ci/32) ((Vap)2 4 (S/3p)2) - ((?i - a2) (Sa/3)2p2

= (ciS/3p - Saj3Sap)2 - ci (Sa/3p)2
- aSa/3 - CiiVa/3)
p (cif3 aSa/3 + CiiVajS) S p (ci/3
.
((Sa/3)2
of the kind required.
is
(23.) Accordingly
LVIII.
it
will be
found that the following equation,
((Sa/3)2-c/32) (Vap)2 + (c-a2) (c(Si3p)2
- p2S(a/3)2)
= (cS/3p - Sa/3Sap)2 - c(Sa/3p)2,
is
an
identity, or that
a, p, p;
since,
it holds good for all values of the scalar

c, and of the vectors
by addition of c(Va/3)2p3 on both sides, it takes this obviously iden-
tical form,
LIX.
((Sfl/3)2
- c/32)
(Sap)2
c(c
- a2)
(S^p)2
= (cS|8p -
Sa;3Sap)2
-c(a8(3p-(3Sapy-;
- a2) = (Saf3y so that if ci be either root of the quadratic XLIIL, or if

ci(ci
ci/32,
the transformation LVIL is at least symbolically valid : but we must take, as
above,
the positive root of that quadratic for ci, if we wish that transformation to be a real
which it employs. And if we had happened (comp.
(20.)) to perceive this identity LIX., and to see its transformation LVIII., we
might have been in that way led to form the quadratic XLIIL, without having
previously formed the cubic XXXIII.
one, as regards the constants
(24.) Already, then, we see how to obtain one of the two imaginary cyclic transformations ot the given focal form XXX., namely by changing ci to C3 in LVIL ;
and the other imaginary transformation is had, on principles before
explained, by
eliminating (Sa,3p)2 between

the equation,
XLVII. and LVI.
a process which easily conducts to
480
LX.
(Yapy +
+ a2p2 = (ci - cs)"' {ci-i(cS^8p -
(S/3p)2
III.
[bOOK
Saj3Sap)2
-C3-Kc3Si3/t>-Sa/3Sap)2},
where the second member
is
the
sum of two squares
(ci
>
being
but
0,
C3
< 0),
as the
second expression LVII. would also become, if ci were replaced by C3. Accordingly,
each member of LX. is equal to (Sap)2 + (S^p)3, if ci, C3 be the roots of any quadratic LI.,
with only the one condition,
LXL
.ciC3 =
5 = -(Sa/3)2;
which however, when combined with the condition of rectangularity LIIL, suffices
to give also J. = /32 a', as in LIV., and so to lead us back to the quadratic XLIII.,
which had been deduced by the general method, as a/acôrofthe cubic equation
XXXIIL
And
(25.)
ber of
LX.
since the values
XXXVI.
to the simple form (Sap)2
ploying the roots ci, C3 at

cyclic transformation
all,
of
ci, C3
reduce, as above, the second
+ (S/3p)2, we may
mem-
thus, or even without
em-
deduce the following expression for the last imaginary
LXII.
Sp^p = (Vap)2 + (S/3p)2 =
a2p2
+ S (a +
\/^^) p
where
S (a
- -v/^/3) p,
v/- 1 is the imagmary of algebra (comp. 2 14, (6.) ) while the real scalar
XXXV. may at the same time receive the connected imaginary form,
r* of
LXIII.
H = (a2 - /32)2 + 4 {SajSy = (a+ ^/~l^f (a - \/~l/3)2.
(26.) Finally, as regards the passage from the given

real focal form (comp. 358, (4.)
LXIV.
in
which
and
a'
with them,
it
be
(Vap)2+ (Si3p)2= (Va'p)2 + (Si3'p)2,
sufficient
LXV.
r2a'
with the same real value of
to a second
form XXX.,
or the transformation,
+a
are real vectors, distinct from
/3'
may
),
(comp. 358, (8.)
= - (a' + jSajS),
r as before
),
r2/3'
so that
and +/3, but in the same plane
to write
down
the formulae
= - (^ + aâ),
(by
XXXV.,
&c,)
we have
the
relations,
= T/3, Sa'^' = Sa^

Ta' = Ta,
T/3'
+ ') = (^2 - 2 + ^2) - 2/3Sai3 = - 2 {ac^ + /SSa^S) ai,
\r2(a aO = a (r2 + a2 /32) + 2/3Sai3 = 2 (aci + /3Sa/3) as
+ ^') = /3(r2 + a2 - /32) - 2aSa/3 = 2 (^ci - aSa^) ai,
/'"'^^
^r2(/3_^') = /3(r2_a2 + ^2)^.2aSa/3 = -2(/3c3-aSaj8)||a3.
LXVI.
XVIT
||
Z'''^^"
|1
XVTTT
||
the identity,
(27.) "We have then
LXIX.
(V(a3
+ )3a/3)p)2 +
(S(i33
-[-
a/3a)p)2
= (a4+2S.(a/3)^ + i3*)
with which
LXX.
may
- (V(a3 - /3a/3)p)2 +
which enables us
two new
its
that
we can
to pass
(-(Vap)2 +
(S/3p)2),
from the focal form XXIII., to a second real focal form,
lines in the
pass from
- a/3a)p)2
- 2S. (a/3)2 + /3*)
(a4
(S(/33
with
((Vap)2 + (S/3p)2)
be combined this other of the same kind,
same plane
LXIX.
to
as the
two old ones
LXX,, by changing a
to
and
a^/-
it
1.
may
be noted
CHAP.
BIFOCAL AND MIXED TRANSFORMATIONS.
II.]
481
360. Besides the rectangular, cyclic, and focal transformations

there are others, alS/>0/>, which have been already considered,
of
though perhaps of less importance: but we shall here mention only

two of them, as specimens, whereof one may be called the Bifocal^
and the other the Mixed Transforination.
(1.)
The two
LXV., being called /ocaZ lines* an expression

them both may be called on that accoiuit a bifocal transforma-
lines a, a', of 359,
shall introduce
which
tion.
(2.) Retaining then the value 359,
which
liary constant e,
I.
j82
- a2 = rê,
III.
so that
the
first
and
therefore,
and
4e2
of
therefore
II.
(Sa^^^ (1 - c^)
r*,
and introducing a new auxi-
4 (Sa/3)2 =
(^2
r* (1
e2),
_ a2)3,
gives,
- a') = 2/3Sa,3,
r2 (ea
LXV.
equation 359,
IV.
XXXV.
shall satisfy the equation,
V.
r^
(eSp
- Sap) =
28a(SS(3p
XXX. of Spfp,
VL (1 - e2) Sp<pp = (1 -c2) ((Vap)2 + (S/3p)2)
= (1 - c2) (Vap)2 + (eSap - Sa'p)2
= (e2 - 1) a2p2 + (Sap)2 - 2eSapSa'p + (Sa'p)2
with the form 859,

.
in
which a2
= a'2, by
359,
LXVL,
a and a may be considered
so that
which
metrically into this last transformation,
is
to enter
of the bifocal kind above
sym-
men-
tioned.
found for Sp0p involves again

(3.) For the same reason, the expression last
(comp. 358) six scalar constants; namely, e, Ta(=Ta'), and the four involved in
the two unit lines, Ua, Ua'.
scalar and quadratic function Sp^p
(4.) In all the foregoing transformations, the
has been evidently homogeneous^ or has been seen to involve no terms below the se-
cond degree in
mixed form,
We may however
p.
VIL
in
which g\
X,
Sp0p =^'(p
have the same
p.
VIII.
IX.
and
in order that the expression

.
(5.) Other general

real,
)2
employ
+ 2SX(p -
this
apparently heterogeneous or
^) S/x (p
significations as in 357, but
e, ,
^ are three new
two conditions of homogeneity,
constants, subject to the
X.
also
,g'e
g't^
+ XSpK + pSX^ = 0,
+ 2S\KSpK + e = 0,
VIL may
admit of reduction to the form,
Sp^p'=pV2 ^ 2SXpS/ip,
as in 357,
11'.
homogeneous transformations of Sp0p, which are themselves
although connected with, imaginary^ cyclic forms (comp. 357, (7.)), because
Compare the Note
t Xi +
y/
pi,
to Art. 359.
and X3
v/-
/"3,
may
here be said to be two pairs of ima-
ginary cyclic normals, of that real surface of the second order, of which the equation is, as before, Sp^p = const.
Compare the Notes to pages 468, 474.
3 Q
482
[bOOK
a sum of two squares of linear and scalar functions is, in an imaginary

duct of two such functions, are the two following (comp. 357, (9.) )
sense, a
III.
pro-
XL
in
+ S\p^p=5fip2+(S\ip) + (S/iip)2;
= ffp'' + SXpfip = g^p^ - (SXzpy - (S/i3p)2
Sp0p=5'p2
XII.
Sp<pp
which (comp. 357, (2.) and (8.)
.gi=ff-^T\fl = Ci, 5-3 =5r TX/i = C3,

= YXfi (TX/it - SX/i)-^ fii = (XT/x 4- /iTX) (TXfi - SX/i)-J,
= (XT^ - /iTX) (TX/i + SX/t)-i
X3 = VX/i (TX/z + SXfiyi,
fiz
XIII.
XIV.
XV.
and
and
/tti,
We have
(6.)
XVI.
XVII.
\i
so that gi, Xi,
),
5-,
X,
fi
be such.
new mixed transformations following
therefore the two
Sp0p=iî(p-6i)2 + (SXi(p-î))*+(S/ii(p-?i))2 +
^3, X3, /U3 are rea/, if
Spcpp
=gzip-
t^y
- (SXa (p -
â))^
- (S/xsCp -
^3))^
ci;
es
with these two new pairs of equations, as conditions of homogeneity,
+ iiSWi = 0,
+ (SîXO + (S?i;ii)2 + ei = 0,
- XaS^sXg juaS^s/â = 0,
XX.
^33
- (8^3X3)2 XXI.
pgfaS
(8/13^3)2 + 3=0.
XVIII.
XIX.
and
g,n + XiSîXi
361.
îi2
We
saw, in the sub-articles to 336, that the diffeof a scalar function of a vector, may in general be
rential, d/]b,
expressed under the form,

I.
where
tor
/o,
i/
is
= wSvdjO,
dfp
a derived vector function^ of the same variable vecAnd we now propose to

is a scalar
coefficient.
and w
show, that
if
II,
,fp
= Spcpp,
denoting the linear and vector function which has been

considered in the present Section, and of which ô/o is still the
(pp still
we
self-conjugate part,
have the equation
shall
I.
with the va-
lues,
III.
so that the part
(pop
may
w = 2,
v = 0o/o;
thus be deduced from
<pp
by
operat-
ing with idS.p, and seeking the coefficient of dô under the
sign S. in the result: while there exist certain general rela-
of reciprocity (comp. 336, (6.)), between the two vectors

and
1/, which are in this way connected, as linear functions of
p
tions
each other.
(1.)
We have
here,
by the supposed
(comp. 334, VI.),

IV.
linear form of^p, the differential equation
d^p =
(pdp
CHAP.
RECIPROCITY OF FORMS.
II.]
also
and
S(d|O.0p)=S(0|O.dp),
heuce,
by 349,
XIII., we
V.
As an example
(2.)
from 00,
let
8(p.<pdp)
483
= S((p'p.dp)',
have, as asserted,
dSp^p
= S{(pp+ 'p) dp = 2S
ôpdp.
of the employment of this formula, in the deduction of 0op
us take the expression,
VI.
VII.
= SjSSap,
347,
XXXI.,
= Sp0p = 2SapS/3p,
./p
0p
which gives,
and therefore
VIII.
d/p = 2 (/3Sap +
aS/3p)dp.
mparingthis with the general formula,

IX.
we
find that the
|d/p = Svdp
form VI. of 0p has
X.
S. ôpdp,
for its self-conjugate part^
1/= 0op
(/3Sap
+ aS/3p)
and in fact we saw (347, XXXII.) that this form gives, as

pression,
XL
conjugate,
2aS/3p.
is given, or
made, selfnecessary) the semisum of itself and its own conjugate funcinstead of 0o, and shall thus have, simply,
by taking
we may
XII.
whence
.(}>'p
Supposing now, for simplicity, that the function
(8.)
tion,
conjugate, the ex-
its
write
(if
tp
..v = (!>p,
XIII.
also (comp. 348,
XV.
I.
/p
= Si/p,
XIV.
= 2Svdp
d/p
II.),
0-'v
= m-^xpv,
and
XVI.
Svdp = Spdv.
(4.) Writing, then,
XVII.
^ve shall
XVIII.
so that
Fv=Sv<p-^p = m-^Bv4^v,
have the equations,
mag
Fv =fp,
XIX.
be deduced from Fv, as
.dFv = 2Spdv =
above stated, there exists a perfect reciprocity of relations,

V, and also between their scalar functions, fp and Fv.
(5.)
may
As regards
in fact, these last
^-ii^da/
and generally, as
between the vectors p and
',
the deduction, or derivation, of v from/p, and of p from Fv,

it thus
occasionally be convenient to denote
XX.
2S.
v was deduced from fp
V=
may
(S dp)-id/p
.
it
XXI.
KS
dvy^dFv
be considered as only symbolical transformations of the ex-
pressions,
XXII.
d/p
2S(dp.i/),
dFv = 2S(dj/.p),
which follow immediately from XIV. and XIX.

(6.) As an example of the passage from an expression such as fp, to an equal
expression of the reciprocal form Fv, let us resume the cyclic form 357, 11. writing
,
thus,
XXIII.
/p
= S/.0P = ^p2 + SXp/ip,
and supposing that g^ \, and p. are real. Here, by what has been already shown (in
sub-articles to 354 and 357), if 0p be supposed self- conjugate, as in
(3.), we have,
484
XXIV.
(l>p=gp + YXpfx
XXV. ..m = (g- SX/x)

XXVI.
(^2
[bOOK
III.
_ X2^2) = _ eiCoC3
^l/v=YXvfiS\n-Y\fiSXvfi-g(\Sfiv-\-fi^\v)ig^v',
and therefore
XXVII.
7nFv
Sv^l^v
= SXv^vSXfi + (SXvfiy - 2gSXvSfiv + g^v^

= (^2 _ \2^2) ^2 ^ X3(;S/il/)2 + ,i2(S\v)2 - 2^S\vS/i/
which
cal form:
of
when compared with
last,
360, VI.,
is
seen to be
what we have
called a hifo-
/ocaZ lines a, a' (360, (1.)) having here the directions of \, ju, that is
be called the cyclic lines* of the /orm XXIII.
The cyclic and bifocal
its
what may
transformations are therefore reciprocals of each other.

(7.) As another example of this reciprocal relation between cyclic and focal lines,
in the passage from fp to Fv^ or conversely from the latter to the former, let us now
begin with the focal form,
XXVIII.
in
which a and
/3
f^
XYTX
and
/p
Sp(pp
are supposed to be given
= (Yapy + (S/3p)2,
and
= <PP = -ciYap +
real vectors.
(3Sf3p,
359,
XXX.,
We have now, by 359,
m = 2(Sai8)2,
therefore,
XXX.
mFv = a2 (Sa/3)2 Fv = S)'i//j/

= Savf3vSaf3 + (a - /32) (Sai.)2
= - v2(Sa/3)2 + Sai/((a2- /32)Sajy +
= - v2(Sa/3)* + SarS(a3 + /3a/3)j/,
2Sa/SS/35/)
an expression which is of cyc^/cybrm; one cyclic line oi Fv being the given focal
and the other cyclic line of Fv having the direction of + (a^ + /3a/3),
line a of fp
and consequently (by 359, LXV.) of + a', where a is the second real and focal line
;
of /p.
(8.)
And
to verify the equation
XVIII., or
to
show by an example that the two
Fv
are equal in value, although they are (generally) different in

sufficient to substitute in XXX. the value XXIX, of v
which, after a
functions fp and
form, it is
few reductions, will exhibit the asserted equality.
362. It
is
often convenient to introduce a certain scalar
and sym-
metric function of two independent vectors^ p and /a', v^hich. is linear

with respect to each of them, and is deduced from the linear and
self-conjugate vector function 0/?, of a single vector p, as follows:
I.
With
./(/>, P')
this notation,
=
=fip', P)
S/>'0/>
S/.0/)'.
we have
*
They are in fact (compare the Note to page 468) the cyclic normals, or the
normals to the cyclic planes, of that surface of the second order, which has for its
equation /p
const.
while they are, as above, the focal lines of that other or re-
ciprocal surface, of which v
is
the variable vector, and the equation
is
Fv =
const.
CHAP.
485
DEBIVED LINEAR FORMS.
II.]
II... /(/>+/)=/^+2/(/>,p0 +//>';
IV.
VI
and
../(/>,/ +
III.
/(^, P) =fp
/(a7/>, y/>)
=/(/>,/>')+/(/>. p'O
V.
= xyf{p, p%
d/> = 2/(/>, dp)

if
Va; = V?/ =
as a verification,
Yll.., f(xp)^x'fp,
a result
new
which might have been obtained, without introducing
(1.)
It appears to be unnecessary, at this stage, to write do\vn proofs of the fore-
going consequences,
that
this
function I,
we
VI., of the definition I.; but it may be worth remarking,

in the formula V., from a notation (325) which was
II. to
here depart a
little,
used in some early Articles of the present Chapter, although avowedly only as a
temporart/ one, and adopted merely for convenience of exposition o{ the principles of
Quaternion Difi'erentials.
(2.) In that provisional notation (comp. 325, IX.)
we should have
had, for the
differentiation of the recent function /^j (361, II.), the formulae,
d/p=/(|0,dp),
/(p,p')=2SpVp;
them
the numerical coefiicient being thus transferred from one of
compared with the recent equations, I. and V. But there

adoptmg these last equations V. and I., namely,
= 2/(iO,
A/P
dp),
f(p,
p')
is
= Sp'<pp
to the other, as
a convenience
now
in
because this /wnc/ion Sp'^p, or Sp^p', occurs frequently in the applications of quaternions to surfaces of the second order, and not always with the coefficient 2.
and treating dp as constant, or d^p as

offp gives, by IV. and V., the formulae,
(3.) Retaining then the recent notations,

null, successive differentiation
VIII.
so that the theorem 342,
I. is
d2/p
= 2/(dp)
d3/p
&c.
here verified, under the form,
IX...eyp = (l + d+id2)/p
=/P + 2/(p,dp)+/dp;
X.
or briefly,
s<ifp
=f(p +
dp),
an equation which by II. is rigorously exact (comp. 339, (4.)), without any supposition whatever being made, respecting any smallness of the tensor Tdp.
^
36.3. Linear and vector functions of vectors, such as those considered in the present Section, although not generally satisfying the
condition of self-conjugation^ present themselves generally in the difIn fact, if
ferentiation of non-linear but vector functions of vectors.
we denote
for the
moment
such a non-linear function by
tx){p),
or
simply by wp^ the general distributive property (326) of differential

expressions allows us to write,
I.
da'(/j)= 0(d/3),
or briefly,
I'.
daj/>
= 0d/j;
486
III.
[bOOK
has all the properties hitherto employed, including that of

where
not being generally self-conjugate, as has been just observed. There
is,
we
however, as
shall soon see, an extensive
and important
ccwe,
in which the property of self-conjugation exists^ for such a function

0; namely when the differentiated function, wp, is itself the result v
of the differentiation of a scalar function fp of the variable vector /?,
although not necessarily/ a function of the second dimension^ such as
has been recently considered (361); or more fully, when it is the

coefficient of d/o, under the sign S., in the differential (361, I.) of
that scalar function//?, whether it be multiplied or not
lar constant (such as w, in the formula last referred to).
sca-
by any
And
generally (comp. 346), the inversion of the linear and vector function
in I. corresponds to the differentiation of the inverse (or implicit) function
u}-^
manner that the equation
in such a
or V,
I.
may
be writ-
ten under this other form,

II.
(1.)
As a very
dw-V = 0"'do- = m"'-^do-,
III.
This
= o)(p) =^ pap,
VI.
and
To
is
us take
a constant vector.
= ^dp = dwp = p'ap + pap' = 2Vpflp'

S\0p' = 2SXpap' = Sp'<p'X
= 2 VXpa = 2 VapX, ^'p' = 2 Vapp'
^'\
.^p'
V.
(2.)
where a
let
dp = p',
IV.
so that 0p'
a=u3p.
simple example of a non-linear but vector function,
the form,
gives, if
if
0'p' are unequal,
and the
find its self-conjugate pari
linear function (pp' is not self-conjugate.
<pop',
by the method
of Art. 361,
we
are to
form the scalar expression,
VII...|/p' = |SpV = P''Sap;

of
which the
VIII.
differential,
=
|d/p'
taken with respect to
On
giving
and accordingly this is equal

0p' and its conjugate.
(3.)
p', is
S ôp'dp' = 2SapSp'dp',
to the
IX.
0op'
= 2p'Snp
semisum of the two expressions, IV. and VI.,
for
the other hand, as an example o/the self-conjvgation of the linear and
vector function,
X.
even
if
dv = dwp = ^dp,
the scalar function
last function
XL
Here
XIII.
fp
%vhen
X'.
d/p
= 2S
wpdp,
have the form,

.
= ^9P9P9''Pi It l'i 9" being three constant quaternions.

'fp
XII. . v = wp = h^{qpq'pq" + q'pq'pq + q'pqpq')
.
2Sj^dp
be of a higher dimension than the second, let this
dv= 0dp = <pp'=^y(qp'qpq" +
q'pq'p'q)
+ îqp'q"pq + q"pqp'q')
+ ^(q"p'qpq'+qp9p'q")-
CHAP.
LINEAR FUNCTION OF A QUATERNION.
II.]
XI V.
and
so that
(p'
=
SX^p' = ^S gpq'Q^qP + P'?^) + &c. Sp'^X
.
we may
as asserted.
<p,
(4.) In general, if d be used as a second

tion,
487
and independent symbol of differentia-
write (comp. 345, IV.),
XY. .,Sdfq = ddfq,

where fq
denote any function of a quaternion
may
principles of the present
Chapter (comp. 344,
I.,
in fact, each member is, by the

and 345, IX.), an expression for
the limit y*
XVI.
(5.)
hm. nn'{f(q + n-^dq +
As another statement
ferentiation o(fq,
n'-^Sq)
-f(q + n-^dq) -f(q + n-^Sq) +fq }
same theorem, we may remark that a
of the
XVII.
that
is,
=fiq, dq), dfq =f{q, dq)

we differentiate the first of these with
dfq
if
d,
operating only on q with each, and not on dq nor on dq,

equal results, of these other forms,
d,
XVIII.
For example,
common
.fq
= q(^q,
se-
obtain
= ddfq.
c is a constant quaternion,
ddfq=:ddfq
is,
= dq.c.dq + dq.c,dq.
= 2Siopdp,
XXI.
dfp
d/p = 2Swpdp,
= (pdp, with dwp = 0dp, as before,
XXII.
dojp
.
we have the general
equation,
XXIII.
ip,
we
Writing then, by X.,
and
in
where
value of these last expressions
XX.
(6.)
ddfq =f(q, dq, dq) =f{q. dq, dq)
and the
if
XIX.
the
first dif-
with each symbol separately taken, gives results of the forms,
and then the assertion

cond with
S(dp.^^p)
= S(^p.0dp),
which dp and dp may represent ajiy two vectors ; the linear and vector function,
which is thus derived from a scalar function fp by differentiation, is therefore (as
above asserted and exemplified) always self-conjugate.

(7.) The equation XXIII. may be thus briefly written,
XXIV.
and
it
..
Sdp^v = S^pdv;
will be found to be virtually equivalent to the following system of three
known
equations, in the calculus of partial differential coefficients,
XXV.
D^Dj,
= T>yD^,
DyDz = D,Dy,
D.D;,
= D;,Dz.
364. At the commencement of the present Section, we

reduced (In 347) the problem of the inversion (346) of a linear
(or distributive) quaternion function of a quaternion^ to the
*
of
We may
a:'yi,
and
y,
also say that each of the two symbols XV. represents the coefficient
development oif[q + xdq-\-ydq) according to ascending powers of a:
when such development is possible.
in the
488
[bOOK
III.
corresponding problem for vectors; and, under this reduced

or simplified form, have resolved it.
Yet it may be interesting,
and
it
now be
will
easy, to resume the linear
and quater-
nion equation,
I.
/^ =
with
r,
II.
.f{q + q) =fq +fq',
and to assign a quaternion expression
for the solution of that
equation, or for the Inverse quaternion function,
IIL..y=/-V,
with the aid of notations already employed, and of results
al-
ready established.
The conjugate of
(1.)
the linear and quaternion function /j being defined (comp.
347, IV.) by the equation,

in
which
p and
q are arbitrary quaternions,
if
we
set
the ybnn,
V.
in
which
s, s',
more than
and
t,
/y =tqs +
we
sufficiently general,
form^
YI.
whence
VII.
t'qs'
XXXI.)
with
^tqs,
are arbitrary but constant quaternions, and which
..
t',
out (conip. 347,
.f'p
/I =
shall have (comp. 347,
XXXII.)
= spt-\-s'pt' + ... = Espt;

and VIII.
/I = 2sf
.
îs,
is
the conjugate
then possible, for each given particular form of the linear function /9, to assign
one scalar constant e, and two vector constants, f, t', such that
it is
IX.
and then we
./l
= e + f, fl=e+e';
shall have the general transformations (comp. 347, I.)
= S.qfl=eSq + St'qi
Sfq
XI. ..Yfq = tSq + Y.fYq = tSq + (jiYq
XII. ../(? = (e +
S? + Sf '5 + (pYq
X.
and
in which Se'q
=S
e'Yq,
and ^Yq or
YfYq
the kind already considered in this Section

of /5,
a Zmear and vector function of V5, of

being also such that, with the form V.
is
;
we have
XIII.
.<pp
-EYtps.
regards the number of independent and scalar constants which enter, at

(2.)
least implicitly, into the composition of the quaternion function /g-, it may in various
As
to be sixteen; and accordingly, in the expression XII., the scalar e

one; the two vectors, e and e% count each as three; and the linear smd vector
function, (pYq, counts as nine (comp. 347, (1.)).
ways be shown
is
(3.)
(p,
we
Since
we may in
XIV.
.
already
know
(347, &c.)
how
to invert
a function of this
XV.
last kind
general write,
.
= Sr + Vr = Sr + 0p,
where
(p-^Yr
= jniipVr
the scalar constant, m, and the auxiliary linear and vector function, ^, being deduced
CHAP.
PROBLEM OF QUATERNION INVERSION.
II.]
from the function

or S^
and Yq,
by methods already explained. It is required then to express

Sr and p, so as to satisfy the linear equation,
(p
+ f ) Sg + Se'q + ^Vq = Sr + 0p
(e
and the form of
e\
e, e,
for this
Assuming
(4.)
q,
in tenns of r, or of
XVI.
the constants
489
^, being given.
purpose the expression,
XVU...q = g'+p,
in
which
q' is
new equation,
.fq'=Sr+ (Pp -fp = S (r
= S (r - t'p) ./-i 1
XIX.
5'
XX.
? = p + S(r-'p)./-il;
new sought
quaternion, ^ve have the
XVIII.
whence
it
(by supposition) a knovm vector^ and S(r I'p) is a known scalar ; so

only remains to determine the unknown but constant quaternion, /"U, or to
which p
that
and
in
is
resolve the particular equation,
XXI.
c
in which
./9o=l,
XXII.
= ^+7=/-^,
50
being a new and sought scalar constant, and y being a new andsouglit vector con-
stant.
Taking scalar and vector parts, the quaternion equation XXI. breaks up
two following (comp. X. and XL)
(5.)
into the
XXIII.
.. 1
= ec 4- Se'y
y)
= S/(c +
XXIV.
= V/(c + y) = fc + ^y
which give the required values of c and y, namely,
XXV.
= (e - Sf'^-'O
XXVII. ../-!=
whence
XXVI.
and
= - c(p'h
Zt
']
e-Ss<p-h'
and accordingly we have, by XII., the equation,
XXVIII.
./(l-0->)=e-StV-> = V
10.
The problem of quaternion inversion is therefore reduced anew to that of

but we can now advance some steps further,
vector inversion, and solved thereby
in the elimination of inverse operations, and in the substitution for them of direct
ones.
Thus, if we observe, that (p'^ =m-ii//, as before, and write for abridgment,
(6.)
XXIX.
so that
is
a new and
XXIX.,
and
= me-St'xpe=f(m-ipe'),
known scalar
constant,
we
XXXI.
XXX. ..mp = )|/Vr;
XXXII.
jnnq = n\l^Yr + (mSr
.
shall have,
n/-il
St'-*//
by XV. XX. XXVII.
=m -
Vr) (m
.
;//
- ^f ),
an expression from which all inverse operations have disappeared, but which still admits of being simplified, through a division by m, as follows.
in the term n.pYr of XXXII., the value me
(7.) Substituting (by XXIX.),
Se'rpe for n,
and changing (by
viously divisible
by m, such a
XXXIII.
vihere
XXXIV.
XXX.)
tpYr to mp, in the terms which are not ob-
di%'ision gives,
.nq = (m- i/ze) Sr + e.//Vr - Se'i/zVr + (t,

c = -pSs'ê + rPeSe'p = Y.t'Vp^Pe.
.
But (by 348, VII., interchanging accents) we have the transformation,
XXXV.
Yp^Pt
= - fYe(pp = - 0' VfVr,

3 R
490
[bOOK
III.
because ^p = Vr, by XIV. or XV. everything inverse therefore again disappears,

with this new elimination of the auxiliary vector p, and we have this final expres;
sion,
XXXVI.
nq=nf-ir = (me-Ssype).f'^r
= (m- xpe) Sr + c;//Vr - St'rpYr - VeV'VtVr,
.
in
which each symbol of operation governs all that follows it, except where a point
and which it appears to be impossible further to reduce, as
indicates the contrary,
the formula of solution of the linear equation

nion function, fq.
I.,
with the form XII. of the quater-
(8.) Such having been the analysis of the problem, the synthesis, by which an
a posteriori proof oi the correctness of the resulting formula is to be given, may be
simplified by using the scalar value XXIX. of/(m--^6); and it is sufficient to
show (denoting Vr by w), that

with the same form XII. of/:
XXXVII.
for every vector
.f{e^Poj-
(9.) Accordingly, that form
Se'i/zw)
the following equation holds good,
-/VcyVfw = (me - S';|/). w.
of/ gives, with
the help of the principle employed
in
XXXV.,
XXXVIII
because
+ mto), -fSt'xpio = - (e + c) St'i/zw,

=
= Y{Yeu).^'t') êSs^/uj - wSe'^/zf,
\-fYe'(p'Yeuj
(pYs'^'Veut
=
&c.
and
thus
the
equation XXXVI. is proved, by actually
Swip'c'
Ss'tpu),
= ^ (Sf
(10.)
with/
As an example,
we have
we take
if
the particular form,
XXXlX...r==fq=pq + qp,
p = a + a a given quaternion,
XL.
which
then,
XLI.
linear
XLII.
XLII I.
therefore,
so that, dividing
XLIV.
or
or
or
by
/I =/'! =2p,
whence by the theory of
and
'i//a>
operating
in
J^/*^"^
2a,
t'
^p = 2ap
2a,
and vector functions,
= 2ap, ^p = Aa^p, m = 8a^,

= 8aâ, 7n - ipf = Sa^ (a - a), n = 1 6a^ {a"^ -
^'p
.xpe
8a, the formula
XXXVl.
a^)
becomes,
= a (a - a) Sr + a2 Vr - aS aYr - aV. aYr,

= aSr+(a + a)Vr-Sar,
a)g'
XLVI.
2p5Sp = S rKp + pYr = r S/) + V (Vp Vr),
XLVII.
4pgSp = 2rSp + (pr rp) =pr + rKp
.
2a (a2
XLV.
a2) 5
2a(a +
or finally,
XLVIII.
=/ V= ml::^ =
4Sp
rVK^
4Sp
Accordingly,
XLIX.
.(j>r-\-
rKp) + {rp + Kp
r)
2r (p
+ Kp) = 4rSp.
as the last, we may with advantage avail our(11.) In so simple an example

selves o{ special methods; for instance (comp. 346), we may use that which was
employed in 332, (6.), to differentiate the square root of a quaternion, and which
conducted there more rapidly to a formula (332, XIX.) agreeing with the recent
XLVIII.
(12.)
We
might
also
have observed, in the same case
XXXIX.,
that
CHAP.
L.
II.]
SYMBOLIC AND BIQUADRATIC EQUATION.
491
..pr-rp=:p'-q-qp^ = 2V(Y(p^).Yq) = 4:Sp.Y(Yp.Yq)=2Sp.(pq-gp);
yfhence pq
- qp, and
pg and
therefore
qp, can be at once deduced, with the
sulting value for q, or for/-ir, as before
and generally it
is
same
re-
possible to differentiate,
on a similar plan, the n*^ root of a quaternion.
We
365.
shall conclude this Section on Linear Functions,
of the kinds above considered, by proving the general existence of a Symbolic and Biquadratic Equation, of the form,
= w - w/+ n'p - n"f^ +/S

which is thus satisfied by the Symbol (/) of Linear and Quaternion Operation on a Quaternion, as the Symbolic and Cubic
I.
Equation,
r.
m - m'(p + W2"02 _ ^3^
350,
I.,
was satisfied by the symbol {(f) of linear and vector operation

on a vector ; the/owr coefficients, n, ri, n', vi", being j^wr scalar constants^ deduced from the function f in this extended or
quaternion theory, as the three scalar coefficients m, m', m"
were constants deduced from 0, in the former or vector theory.
And at the same time we shall see that there exists a System,
of Three Auxiliary Functions, F, G, H, of the Linear and
Quaternion kind, analogous to the two vector functions, \p and
which have been so useful in the foregoing theory of vecand like them connected with each other, and with the
Xi
tors,
given quaternion function j^ by several simple and useful relations.
(1.)
tion
fq
The formula
= r,
of solution, 364, XXXVI., of the linear and

quaternion equabeing denoted briefly as follows,
II.
so that
(comp. 348,
III'.)
we may
nq
nf-^rFr,
write, briefly
and symbolically,
Ul...fF=Ff=^n,
it
may
next be proposed to examine the changes which the scalar n and the function
Fr undergo, when/r is changed to /r + cr,

or/to/i-c, where c is any
to
is, by 364, XII., when e is changed to e+ c, and
+c
m being at the same time changed, according to the laws of the earlier
stant; that
scalar con-
^', ;p,
theory.
(2.) Writing, then,
IV.
and
we may
../c=/+c,
V.
represent the
//c
=
)//
new form
VI.
ec
=e+
+ ex +
c2,
= 4c, fc=f + c,
0c
mc=m + m'c+ m"e^+ c^,
c,
of the equation 364,
ncfc-^r
= Fcr,
or
VII.
XXXVI. as
fFc = c
follows:
and
492
VI II.
where
Fat
and
HI.
= (m^ - i//cf ) Sr + e^c^r ~ Ss'^êVr - Ye<p'cVeYr,

IX,
nc=Ccmc St'i/zcC.
.
(3.) In this manner
XI.
we may
seen that
it is
write,
..
Fc=F+eG+ c^H+ c^
nc
= n + nc + n"c^ + n"'c^ +
X.
and
[bOOK
c*
where F, G, H, are three functional symbols^ such that
{Fr={m- î) Sr + e^Yr - Sf ';//Vr - Vt'^' VtVr

Gr = (m' - xO Sr + (ex + i//)Vr - Se'xVr - Yi'YiYr
I
= (m" - c) Sr + (e + x) Vr - ScV
( ZTr
;
XII.
and
w, ', n", n'" are
/owr scalar constants, namely,
XIII.
= em-
n'
= TO +
St'ê (as in 364,

em' S'x
XXIX)
n"=TO' + em"
St'e
n" = m" + e.
(4.) Developing then the symbolical equation VIL, with the help of X. and XI.,
and comparing powers of c, we obtain these new symbolical equations (comp. 350,
XVI. XXI. XXIII.)
(H=n"~f;
XIV.
and
G = n" -fB:= n" - n"'f+p

F= n' -fG = n- n"f+ n'p -p
;
finally,
XV.
= iiy= '/- n"/2
+ ny
_^4,
only another way of writing the symboUe and biquadratic equation I.

(5.) Other functional relations exist, between these various symbols of operation,
which we cannot here delay to develope but we may remark that, as in the theory
which
is
of linear and rector functions, these usually introduce a mixture of functions with
their conjugates (comp. 347, XI., &c.).
(6.) This seems however to be a proper place for obsei-ving, that if

for any four quaternions, p, q, r, s, the equations,
temporaiy notations,
XVI.
Ipqr-]
= (p5r) 4-
XIX.
and
so that [pql
is
XVII.
[pq}=pq-gp;
XVIII.
(7.) In the
may
(pqrs)
is
\_pqr'\ is
a quaternion, we
[pp-]=0;
(pqr)
if t
be any fifth quaternion, the quaternion equation^
= /) {qrst) +
q (rstp) +r(sfpq)
s(tpqr)
+ t(pqrs'),
XXV.
and which
= S ./>
= - {qpr) = {qrp) = &c., {ppr) =

= - {qpr-\ = [qrp] = &c., [ppr] =
Ipqr-]
= - (qprs) = (qrps) = (qrsp) = &c., (pprs) = 0.
(pqrs)
.
next place,
XXIV.
which
(;>?r)
Sp + [pr] Sg + [qp] Sr ;
S.p [jr],
XX...[pq-]=-lqpl
XXI.
XXII.
XXIII.
write, as
[r^]
a vector, {pqr) and (pqrs) are scalars, and
shall have, in the first place, the relations
and
we
q(prst)=p{qrst) + r(pqst) + s(prqt)+t{prsq),
analogous to the vector equation,
XXVI.
= aS/3y^-i3Sy(â + ySfa/3-^Sa/3.y,
CHAP.
GENERAL QUATERNION TRANSFORMATIONS.
II.]
or to the continually* occurring transformation (comp. 294,
XXVII.
is satisfied
SSa(3y
generally, because
which
q(prst)
XXX.
[rst]
5Sai3y
= r,
this other
Spq
= s,
= t.
general quaternion equation,
[stp]
Srq +
Ssq
[tpr']
[prs'] Stq,
XXIX.
XV.),
= V/3ySa5+VyaS/35+Va/3Sy^;
gives true results,
when
operated on by the ybwr
it is
symbols (comp. 312),
XXXI.
(9.)
q =p,
for the /our distinct suppositions,
analogous to this otherf useful vector formula (comp. 294,
is
distinct
we have
(8.) In the third place,
XXIX.
XIV.),
= aSdjSy + pSady + ySa(3d,
it is satisfied
XXVIII.
493
S.r,
S.JO,
Assuming then any four quaternions, p,
S.s,
S.f.
r, s, t,
which are not connected by
the relation,
XXXII.
and deducing from them /owr
(prsO = 0,
others, p', r, s\
t',
by the
equations,
rXprst)=^-flstp-],
//^'(^^Ô^/C^Ô'
X^Xprst) =f{tpr-],
XXXIII
which /is
t'{prst)
= -f[prsl
supposed to be a symbol of linear and quaternion operation on a

XXIX. allows us to write generally, as an expression for
the function /j, which may here be denoted by q' (because r is now otherwise used) :
in
still
quaternion, the formula
XXXIV.
and
in
its
q'
^fq =p'Spq + r'Srq + s'Ssq + t'Stq
sixteen scalar constants (comp. 364, (2.)) are
its^bwr quaternion constants, p', r,
s',
now
those which are involved
t'.
(10.) Operating on this last equation with the four symbols,
XXXV.
we
..S.[rV<'],
XXXVI
= {P'rs't') Spq
= {s't'pV) 8sq
Xiq't'p'r')
nion inversion
= (r's't'p) Srq
(qW') = (t'p'r's) Stq
(q's't'p')
Spq,
Srq,
XXIX., we have the
Ssq,
Stq,
following new formula of quater-
= (py/t') {prst^f-^q'
XXXVIII.
Cpr's't') (prst)q
+ Istp'] (q's't'p') + [<pr] (q'tp'r) + [prs] (q'p'r's')
.
for the four scalars,
XXXVII.
are substituted in the formula
S.[/>'rV],
/(^'^'s'*')
and when the values thus found
S. [<'//],
S.[sr/>'],
obtain the four following results
[rsf] {q'r's't')
The equations XXVII. and XXX., which had been proved under slightly different forms in the sub-articles to 294, have been in fact freely employed as transformations in the course of the present Chapter, and
the student.
t Compare
ai*e
supposed to hQ familiar to
494
which shows, in a new way, how

put under what we
may
[bOOK
a linear equation
to resolve
III.
when
in quaternions,
(comp. 347, (!)) the Standard Quadrinomial Form,
call
XXXIV.
(11.) Accordingly, if we operate on the formula XXXVIII. with/, attending to
the equations XXXIII., and dividing by (^prst"), we get this new equation,
XXXIX,
(ip'r's'f)fq
- r' {q's't'p) + ' {q't'p'r') - 1' {q'p'r's')

= q', by XXV.
=p\qr'st')
whence
fq
(12.) It has been remarked (9.), that p,
four quaternions, which do not

sume,
XL. ../j=l,
with the laws of 182, &c., for the symbols
XLI.
and then
it
r, s, t,
satisfy the equation
in recent formulae,
XXXII.
sj,
i,
i,j, k,
we may
may
be any
therefore as-
= k,
because those laws give here,
= -2;
(It/A)
be found that the equations XXXIII. give simply,
will
XLII.../=/l,
= -/t,
r'
so that the standard quadrinomial
s'
= -fj,
t'=-fk',
form XXXIV. becomes, with
this selection of
prst,
XLIII. ..fq=fl.Sq -fi. Siq -fj. Sjq -fk

an immediate
and admits of
pressed (comp. 221)
verification,
^liq,
because any quaternion,
q,
may
be ex-
the quadrinomial,
by
XLIV. ..q = Sq- iSiq -jSjq - kSkq.

we set out with the expression,
if
(13.) Conversely,
XLY.
which gives,
= w + ix+jy + kz,
221,111.,
XLVI. ..fq = wfl+ xfi + yfj + zfk,
or briefly,
XLVII.
- aw
-^
hx
-\-
cy
-\-
dz,
abode being here used to denote five known quaternions, while wxyz are
four sought scalars, the problem of quaternion inversion comes to be that of the sethe letters
parate determination (comp. 312) of these four scalars, so as to satisfy the one
and it is resolved (comp. XXV.) by the system of the four folequation XLVII.
;
lowing formulse
XLVIII
fw{abcd) =
'
'
{ebcd)',
\y {abed) = (abed')
x(abed)
= (aecd);
z(abcd)
= (abce)
the notations (6.) being retained.

(14.) Finally it may be shown, as follows, that the biquadratic equation
linear functions of quaternions, includes* the cubic I'., or 350, I., for vectors.
* In like manner
one,
when we
which case
then
^;;^
it
may be
said, that the cubic
I.,
equation includes a quadratic
confine ourselves to the consideration of vectors in one plane
m = 0,
= m'
;//
and
also
= m',
or
-^l^p
= 0^
if
for
Sup-
p be a line in the given plane
for
for
we have
CHAP.
495
ADDITIONAL APPLICATIONS.
III.]
and quaternion function, fq, reduces
that the linear

pose, for this purpose,
itself to
the last term of the general expression 364, XII., or becomes,
XLIX. ../5 = ^Vg,

the coefficients n,
n', w",
LI.
and the biquadratic
I.
so that
n" take
.
n=
n'
0,
=' = 0, /1=/'1
by XIII., the
= m,
n"
= 0;
values,
= m',
n" = m"
= i-m + Tn'f-m"p+f^)f.
also the operation

vector, by XLIX., and it may be any vector, p
now equivalent to that denoted by 0, when the subject of the operation is a vecwe may therefore, in the case here considered, write this last equation LII. under
Bntfq
tor
then,
..c = 0,
becomes,
LIL
/is
L.
is
now a
the form,
LIII.
which agrees with 351,

the operand (p)
is
= (- w + m'^ - m>2 + ^3)p^
and reproduces the symbolical
I.,
cubic,
when
the symbol of
suppressed.
CHAPTEE
III.
ON SOME ADDITIONAL APPLICATIONS OF QUATERNIONS, WITH
SOME CONCLUDING REMARKS.
Section
1.
Remarks Introductory
to this
Concluding
Chapter.
When
the Third Book of the present Elements was

was
begun,
hoped (277) that this Book might be made a
much shorter one, than either of the two preceding. That
purpose it was found impossible to accomplish, without injustice to the subject
but at least an intention was expressed
366.
it
(317), at the commencement of the Second Chapter, of rendering that Chapter the last : while some new Examples of Geotcith this
understanding as
to the
operand.
m = 0),
((p^~m"^
and therefore
(02
In
the cubic gives here (because
fact,
+ m')(pp = 0;
- m"^ +
m")
(t
be already the result of an operation with 0, on any vector p

above supposed, a line in the given plane.
if <T
that
is if it be,
as
^6
metrical Applications,
ones, were promised.
367.
The
TIT.
and some few Specimens of Physical
promise, thus referred
ready in part
[bOOK
redeemed
to,
for instance,
has been perhaps alby the investigations
(315) respecting certain tangents, normals, areas, volumes, and

pressures, which have served to illustrate certain portions of
the theory oi differentials and integrals of quaternions. But it
may be admitted, that the six preceding Sections have treated
chiefly of that Theory of Quaternion Differentials, including
of course
its
scarcely less
tions,
tions
Principles
and Pules; and of the connected and
important theory of Linear or Distributive Func-
of Vectors and Quaternions

Examples and Applicahitherto
a
thus
merely subordinate or illusplayed
having
:
trative part, in the progress of the present
Volume.
368. Such was, indeed, designed from the outset to be,

upon the whole, the result of the present undertaking which
was rather to teach, than to apply, the Calculus of Quaternions.
:
Yet
it still
appears to be possible, without quite exceeding
and accordingly we shall now endeavour, to

condense into a short Third Chapter some Additional Examples, geometrical and physical, of the application of the princisuitable limits,
and rules of that Calculus, supposed to be already knoivn,

and even to have become by this time familiaj^* to the reader.
And then, with a few general remarks, the work may be
ples
brought to
its close.
Section 2
Tangents and formal Planes
On
to
Curves
in
Space.
369. It was
shown (100) towards the
close of the First Book,
that if the equation of a curve in spacCy whether plane or of double

curvature, be given under the form,
I.
where
^ is
P =
(p{t)
a scalar variable, and
is
= 0^,
a functional sign, then the de-
rived vector,
11.
D/)
= D0 =
(P't
= p'=ap
* Accordingly, even references to former Articles
:dt,
will
now
m
be supplied more
sparingly than before.
CHAP.
TANGENTS AND NORMAL PLANES TO CURVES.
III.]
497
represents a line which is, or is parallel to, the tangent to the curve,
drawn at the extremity of the variable vector p. If then we sup-
a point situated upon the tangent thus drawn to a

p and that u is a point in the corresponding normal
and if we denote the vectors
plane, so that the angle tpu is right,
OP, oT, ou by p, T, V, the equations of the tangent line and normal
pose that T
curve
is
PQ, at
may now be
plane at p
thus expressed
V(t-p)p' = 0;
III. ..
IV..
S(i;-p)/=:0;
the vector t being treated as the only variable in III., and in like
manner v as the only variable in IV., when once the curve pq is
and the point p
given^
(1.) It
is
permitted, however, to express these last equations m\diQT other forms ;

may replace p' by dp, and thus write, for the same tangent line and
example, we
normal plane,
for
is selected.
V.
V(r-p)dp =
where the vector differential dp
and
the curve at P,
(2.)
We may
VI.
0;
may
S(v-p)dp = 0;
any
represent
parallel to the tangent to
line,
not necessarily small (compare again 100).

also write, as the equation of the tangent,
is
VII.
= p-\- xp\
where
a scalar variable
a; is
and as the equation of the normal plane,

VIII.
because
i\\\s
dpT(v-p)
partial differential of
IX.
(3.)
T(v p),
XL
is,
dT(v-p)=0,
or of pu,
(1.), or
is
dw = 0;
if
(by 334, XII.,
5;c.),
337, (1.), of which the equation
Ta =
with
and
1,
Sa/3
where
is
new
is,
= 0,
by VII., and by the value 337, VI.
r= p +yctp,
dT(v-p) = S(U(i;-p).dp).
= a%
the equation of the tangent
Vlir.
or
For the circular locus 314,
X.
tlie
= 0,
of
scalar variable
p',
;
perpendicularity of the tangent to the radius being thus put in evidence.

(4.) For the plane but elliptic locus, 314, (2.), or 337, (2.), for which,
XIL
the value 337, VIII. of
= V. a%
p'
Ta =
with
but not
1,
shows that the tangent,
Sa/3
= 0,
at the extremity of any one semi-
diameter
that is, to the

p, is parallel to the conjugate semidiameter of the curve
one obtained by altering the excentric anomaly (314, (2.)), by a quadrant: or to
the value of p which results, when we change t to t-\-\.
(5.)
For the
helix, 314, (10.), of
..p = cta^ a'/3,
XIII.
c being
a scalar constant,
XIV.
XVL
we have
p'=ca + '^
TVa-ip' =
^T^,
which the equation

with
Ta =
is,
and
1,
Sa/3
= 0,
the derived vector,
a<^/3
and
XV.
whence
XVIL
'^
Sa"
V=
c,
(TV: S)a-1p'=i^
2c
3s
'
498
[bOOK
III.
the tangent line (p') to the helix is therefore inclined to the axis (a) of the cylinder
M^hereon that curve is traced, at a constant angle (a), whereof the trigonometrical
and accordingly, the numerator
tangent (tan a) is given by this formula XVII.
;
7rT/3 of that formula represents the semicircumference of the cylindric base
the denominator 2c
measured
spires,
in
an expression
XVIII.
if
while
half the interval between two successive

a direction parallel to the axis.
We may then write,
is
for
= 2c tan a = 2c cot 6,
7rTj3
a thus denote the constant inclination of the
-helix to the axis, while b denotes the
constant and complementary inclination of that curve to the base, or to the circles
which it crosses on the cylinder.
(6.) In general, the parallels p' to the tangents to a curve of double curvattire,
which are drawn from
Si
fixed origin o, have a certain cone for their locus; and for
is given by the formula XVII., or
by
the case of the helix, the equation of this cone
any legitimate transformation
XIX.
it is
thereof, such as the following,

.
SUa-ip'=+coso=4: sin6;
therefore, in this case, a cone
of revolution, with
its
semiangle a.
As an example of the determination of a normal plane to a curve

curvature, we may observe that the equation XIII. of the helix gives,
(7.)
XX.
p2= ,32 -
and
c2<2,
therefore
XXI.
Spp'
of double
-cH',
the equation IV. becomes therefore, for the case of this curve,
XXII.
Sjo'v
en, with the value
XIV.
of
p'.
If then it be required to assign the point u in which the normal plane to the
(8.)
helix meets the axis of the cylinder, we have only to combine this equation XXII.
with the condition v
XXIII.
the line
pu
is
and we
a,
find,
by XIII. and XIV.,
ou = V = - cHa Sap' =
:
XXIV.
eta,
Sa (u - p) =
therefore perpendicular to the axis, being in fact a
normal
to the cy-
linder.
370. Another view of tangents and normal planes may be proposed,

shall connect them in calculation with Taylor"* s /S^mes adapted
which
to quaternions (342), as follows.

(1.) Writing
I.
p<
= po + t^p'o,
or
brieflj',
I'.
= p + ntp',
the coffiecient ut or u will generally be a quaternion, but its limiting value will be
positive unity, when t tends to zero as its limit ; or in symbols,
II.
Ko
= liin. M=
1.
<=o
(2.) Admitting this, which follows either from Taylor's Series, or (in so simple a
case) from the mere definition of the derived vector p', we may conceive that vector
be constructed by some given line pt, without yet supposing it to be known that
is tangential at P to the curve pq, of which the variable vector is OQ = pt,
while op = po = p, so that the line pq = utp' is a vector chord from p, which diminishes
p' to
this line
indefinitely with the scalar variable,
t,
and
is
small,
if t
be small.
CHAP.
CONNEXION WITH TAYLOR*S SERIES.
III.]
next that
(3.) Conceiving
B,
PQ
we may
= OB= the
lo
a perpendicular
let fall
into the two rectangular lines,
vector of
499
some new and
arbitrary' point
qm on the line PR, and so decompose the chord

PM and mq which, when divided by the same
;
chord, give rigorously the two (generally) quaternion quotients,
ITT
the variable
'
^
=
PQ
ttp'(w
V.
'
'
^
^
PQ
- p)
up'iu}
p)
Vp'(oJ
lim.
we have
limits,
^^ = ?^^-^),
PQ
these other rigorous equations,
VI.
'
lim.
by comparing which with 369,
and IV,, we
III.
Vp'(w
p)
0,
Pi(^-p)
see that those
sent respectively, as before stated, the tangent and the

if
M^X
'^ =
PQ
p'(a;-p)
curve at p; because,
enters into
it
1.
Passing then to the
(4.)
jY
'
thus disappearing through the division, except so far as
which tends as above to
w,
- P)
p)
S"P'(^
the chord
pq
two equations repre-
normal plane
to the proposed
by V. or VI.,
tends,
to coin-
both in length and in direction, with its projection pm on the line pr while
on the other hand, if Sp'(te p) = 0, that projection tends to vanish, even as compared
cide,
with the chord pq
which chord tends now to coincide with
Hne PR, erected so as

be a normal to the curve at
its
other projection mq,
or with the perpendicular to the
to reach the point
FR must,
p.
in this last case,
We may also investigate
(5.)
an equation
for the
whence
normal plane, by considering
as the limiting position of the plane which perpendicularly bisects the chord.
If
it
be supposed to be a point of this last plane, then, with the recent notations, the vector
w = OR must
VII.
satisfy the condition,
IX.
which
- p - t<^p')2 = (w - p)2,
VIII.
(a>
=
2Sup'(a}
p)
<(p')2,
T(w-p<) = T(a>-po),
or
or
be noted that up' is a vector (in the direction of the chord, pq), alsuch then is the equation of the
generally a quaternion, as before
bisecting plane, with w for its variable vector, and its limit is,
in
may
it
though M
itself is
X.
(6.)
The
XL
last process
may
S|o'(a>
p) =
0,
also be presented
as before.
under the form,
= lim. <-i{T(w - p^) - T(w - po)} = r>T(a> - pt),
when
= 0;
and thus the equation 369, VIII. may be obtained anew.

(7.) Geometrically, if we set off on p.q a portion rs equal in
length to RP, as in the annexed Figure 76,
we
shall
/T
have the
.^^7iO~~
7\\/
/
limiting equation,
XII.
+ SQ
= (rq Rp) PQ = (ultimately^ cos rpt

P(4
:
which agrees with 369, IX.

(8.)
If then
the
point
//
be taken
plane at p, this limit of the quotient,
out of the normal

divided by pq,
rq rp
XXf\
^^-^^^
J^^'-"'^'^
pig. 76.
negative; and if the chord pq be

called maW of the ^r< order, the difference oi distances of its extremities from R
may then be said to be small of the same (first) order. But if r be taken in the nor-
has
definite value, positive or
mal plane
at
p (and not
coincident with that point p itself), this difference of dis-
500
[bOOK
III.
may then be said to be small, of an order higher than the first: which answers to the evanescence o{ the first differential of the tensor, T(w-p) in XL, or
tances
369, Vlir.
T(v-p)in
371. A curve may occasionally be represented in quaternions, by

an equation which is not of the form, 369, I.i although it must
always be conceived capable of reduction to that form: for instance,
new
this
equation,
I.
is
Yap .Ypa' = {Yaay,
T Vaa' > 0,
with
not immediately of the form p = 0^, but
it is
reducible to that form
as follows,
II.
An
= ia +
r'a'.
equation such as I. may therefore have its differential or its deriwith respect to the scalar variable t on which p is thus
vative taken,
conceived to depend, even if the exact law of such dependence be unknown : and d/>, or />', may then be changed to the tangential vector
a>
- /> to which
order to form an equation of the tanwhich the vector w of a point on that sought
it is parallel, in
gent, or a condition
line
must
satisfy.
To
pass from
(1.)
III,
whence,
and
IV.
t'
I.
to II.,
*a
first
operate with the sign V, which gives,
or simply.
we may
being scalars,
.
we may
= 0,
pSaa'jO
III'.
Nap = i'\aa\
+ t'ci,
Saa'p
Ypa = tYaa',
and the required reduction is effected while the return from

nation of the scalar t, is an even easier operation.
:
write,
tt'
1,
II. to I., or
the elimi-
Under the form II., it is at once seen that p is the vector o^ & plane hyperwith the origin for centre, and the lines a, a' for asymptotes ; and accordingly
the properties of such a curve may be deduced from the expression II., by the
(2.)
bola,
all
rules of the present Calculus.

(3.)
For example, since the derivative of that expression
V.
the tangent
may
io
is,
p'=a-^2a',
its
equation thus written:
{t + x)a +
t-^(Jt-x)a
therefore the lines a, a' in the points of which the vectors are It a, 2t'â'
the intercept, upon the tangent, between the asymptotes,

bisected at the point of contact : and the intercepted area is constant, because
so that (as
is
(comp. 369, VII.) have
VI.
it intersects
V(<a.<-a')
(4.)
is
well
known)
= Vaa',
&c.
But we may
also operate immediately, as
and thus arrive (by substitution ofo>

tion,
VII.
pfordp,
above remarked, on the fi)rm
Vaw .Yâ + Yap Vwa' = 2 (Vart')2,

.
I.
&c.) ht the equation of conjuga-
CHAP.
III.]
NORMALS AND TANGENT PLANES TO SURFACES. 501
= or, as before, then

=
which expresses (conip. 215, (13.), Sec.') that if p OP, and w
R is on the tangent to the curve, at the point p, or at least each of these two
is situated on the polar of the other, with respect to the same hyperbola.
either
pouits
it is frequently convenient to consider a curve as the intersection of

(5.) Again,
two surfaces; and, in connexion with this conception, to represent it by a system of
two scalar equations, not explicitly involving any scalar variable : in which case,
both equations are to be differentiated, or derivated, with reference to such a varia-
and dp or p' deduced, or replaced by w

p as before.
we may substitute, for the equation I., the system of the two follow(6.) Thus
had occurred as III'.)
ing (whereof the first
ble understood,
VIII.
Saa'p =
p2Saa'
0,
and the derivated equations corresponding
IX.
Saa'p'
- SapSa'p = (Yaay
are,
= 0, 2Saa'Spp' - Sap'Sa'p - SapSa'p' =

w p for p', &c.,
=
2Saa'Spw SawSa'p SapSa'w 2 (Yaay
;
with the substitution of
or,
Saa'w
= 0,
the last of which might also have been deduced from VII., by operating with S.
remarked that the two equations VIII. represent respectively
(7.) And it may be
in general a plarie
or II.
I.
and an hyperholoid, of which the intersection (5.)

and an hyperbolic cylinder, if Saa' = 0.
Section
3.
On Normals and Tangent Planes

shown (100,
372. It was early

is
the hyperbola
that
(9.))
when
to Surfaces.
a curved surface
represented by an equation of the form,

I.
in
is
or a plane
which
is
a functional sign,
<p{x,y),
and
x,
y are two independent and
scalar variables, then either the two partial differentials, or the
two
partial derivatives, of X\iq first order,

II.
d^P, dyP,
or
III.
D^p, Dyp,
represent two tangential vectors, or at least vectors parallel to two

tangents to the surface^ drawn at the extremity or term p of />; so
two differential vectors, or of lines parallel

parallel to) the tangent plane at that point: and
the principle has been since exemplified, in 100, (11.) and (12.),
and in the sub-articles to 345, &c. It follows that any vector v,
that the plane of these

to them, is (or
is
which is perpendictdar to both of two such non-parallel differentials,

or derivatives, must (comp. 345, (11.)) ^^ ^ normal vector at p, or at
least
one having the direction of the normal to the surface at that
point; so that each of the
IV.
if actual,
two
vectors,
V.d,.pd,p,
represents such a normal.
V.
V. D.PD,|0,
502
(1.)
As an
[bOOK
IIT.
additional example, let us take the case of the ruled paraboloid, on
which a given gauche quadrilateral abcd is superscribed. The expression for the
vector p of a variable point p of this surface, considered as a function of two independent and scalar variables, x and y,
may be thus written (comp. 99, (9.))

Tl..,p = xya + il-x)y^ + {l-x) {l-y)y +x(l-y)8;
where the supposition y = 1 places the point p on the

on CD and a;= 1, on da.
y= 0,
(2.)
line
ab
x=Q places it on bc
We
have
by
here,
partial derivations,
Dyp = rr(a-5)+(l-x)(/3-y);
VII...D.p=yC-^) + (l-y)(5-r);
these then represent the directions of two distinct tangents to the paraboloid VI., at
what may be called the point (a:, y) ; whence it is easy to deduce the tangent plane
and the normal
at that point,
cept to remark that
through
s, T,
we
by constructions on which we cannot here delay, exwe draw two right lines, QS and et,
(comp. Fig. 31, Art. 98)
if
p, so as to cut the sides
shall
VIII
= ^ +
_|oQ = a;^ + (l-^)A
therefore,
OR=y/3 + (l-y)y,
OT=ya+
(l-a;)y,
(1
y)^,
by VII.,
IX.
so that these
of the quadrilateral in points q, r,
'
\os
and
da
ab, bc, cd,
have by VI. the vectors,
Dxp = RT,
Dyp
= SQ
two tangents are simply the two generating
lines of the surface,
which
pass through the proposed point P.

the tangents thus found are the sides
(3.) For example, at the point (1, 1), or a,
ba, da, and the tangent plane is that of the angle bad, as indeed is evident from
geometrj".
of the screw surface (comp. 314,
(4.) Again, the equation
X,
= c.ra+ya*/3,
with
Ta =
and
l,
XVI.),
Sa/3
= 0,
gives the two tangents,
XI.
whereof the latter
is
Dxp
= ca +
|ya^+i)3,
T)yp
= a^/?,
perpendicular to the former, and to the axis a of the cylinder ;

normal to the surface X. at the point (x^ y) is represented
so that the corresponding
by the product,
XII.
373.
Whenever
V=
D;rp
Dyp =
ca^+i/3
a variable vector
/>
is
^yfi-a.
thus expressed or even

x and
conceived to be expressed, as a function of two scalar variables,
(or s
and
t,
&c.), if
we assume any
&c.), the
three diplanar vectors, such as
scalar expressions^ Sap^ Sy3p, S7p

*^>
7 (or , K,
functions of the same two scalar,
then
be
will
(or Stp, Skp, Sajo, &c.)
variables ; and will therefore be connected with each other by som(
thi'ee
one scalar equation, of the form,

I.
or briefly,
F{Sap, S^p,
87,0)
0,
CHAP.
CONNEXION WITH QUATERNION DIFFERENTIALS. 503
III.]
../P =
II.
of expression
generality
introduced (instead of zero) for greater

and F^ f are used as functional but scalar
If then (comp. 361,
signs.
this scalar function
XIV.) we express the first

..d/p=2Svdp,
and is
a certain derived vector,
is
i'
of
here considered as being
two scalar varia-
a vector function (like p) of the

(at least implicitly)
we
above mentioned,
bles
differential
fp under the form,

III.
which
in
(7;
C is a scalar constant,
where
shall
have the two equations,
two other and corresponding ones,

V.
Si^D,p = 0,
Sj/D/ = 0;
or these
from which
follows (by 372) that v has the direction of the normal to the surface I. or II., at the point p in which the vector p terit
Hence the equation of that normal (with w for

may, under these conditions, be thus written:
VI. ..\v{ic-p) = 0',
minates.
vector)
its
variable
and the corresponding equation of the tangent plane at the same point
p
is
VII. ..Si^(a;-p)=:0.
For example,
(l.J
if
VI 11.
we take the
expression 308, XVIII., or 345, XII.,
p- rj^kj-^Jc~\
treating the scalar r as constant, but *
in
and
which
namely
kj~^ =j^^, &c.,
as variable,
we have
then (comp. 345,
XIV.), the equations, a denoting any unit- vector,

IX.
Sip
= rS
between which s and

(a')^
a2<S
and
I.
XI.
which
last results
(2.)
With
bles, s
we
and
a^*-^ S
(St2 + (S/p)2 +
p2
XIV.
Skp = rS
a^^î,
- r2 =
(SAjo)2
= _ r2 = const.,
or
form XI.
d/p
of fp,
we
a2*+2
XI'.
0,
Tp = r
before.
hsLve the differential expression of the fivst
= 2Si/dp = 2Spdp,
conceive that p
whence
XIII.
=p
as above,
be supposed to be unknoum (or to be forgotten),
if
some vector function of two scalar variais,

although the particular law VIII. of its dependence on them may now
still
t,
= rS
had indeed been otherwise obtained
this
XII.
if
/p
order,
and
S/p
namely (comp. 357, VII., and 308, (10.) and (13.)),
II.,
X.
and
a2+i,
can be eliminated, by simply adding their squares, becaase

by 315, V., if Ta = 1. In this manner then we arrive at equa-
+ (a'-i)2 = 1,
tions of the forms
we may
write also,
= Svdp = Spdp = Sp (d, + d^p = SpD.p

|d/p
then the /i/ncîon/p have (as above) a value,
=-r2, which
ds
is
+ SpD^p
d^
cnvstant, or
is
inde-
504
[bOOK
III.
pendent of both the variables, s and t, while their differentials are arbitrary^ and are
independent of each other, we shall thus have separately (comp. V., and 337, XIII.,
XVIL),
XV.
The radius p
of the sphere XI.
own
the normal at
its
Dsp and Dtp,
at that point;
is
S|oDp
= 0,
SpD^p
therefore in this
extremity, because
it is
= 0.
seen to have the direction of
way
perpendicular to two distinct tangents,
which are indeed, in the present
case, perpendicular to
each other also (337, (8.)).

(3.) Instead of treating the two scalar variables,
x and
y, or s
and
t,
&c., as both
and independent, we may conceive that one is an arbitrary (but

scalar) /wnc^fon of the other; and then the vector v, determined by the equation
III., will be seen anew to be the normal at the extremity p of p, because it is perentirely arbitrary
pendicular to the tangent at P to an arbitrary curve upon the surface, which passes
through that point or (otherwise stated) because it is a line in an arbitrary normal
:
plane at
p, if a
normal plane
to a curve
on a surface be called (as usual) a normal
plane to that surface also.

(4.)
t,
if we conceive that s in YIII.

XIV. will take the form,
For example,
the last expression
XVI.
whence, d< being
XVII.
= ^d/p = S
still
arbitrary,
S p (s'D.p + Dtp)
p (a'D.p
we have
thus an arbitrary function of
is
+ D^p) df,
As
if
= s'At
the one scalar equation,
= 0,
XVIII.
or
and although, on account of the arbitrary
coefficient
equivalent to the system of the two equations XV., yet

XVIII., that the directed radius p, of the sphere XL,
-J-
this
s',
it
s'Dsp 4 Dtp
one equation XVII.
immediately
is
signifies, as in
perpendicular to the arbi-
is
trary tangent, s'Dsp + Dtp or to the tangent to an arbitrary spherical curve through
p, the centre o and tensor Tp (or undirected radius, r) remaining as before.
',
(5.) As regards the logic of the subject, it may be worth while to read again the
jaroof (331), of the validity of the rule for differentiating a function of a function ;
because this rule is virtually employed, when after thus reducing, or conceiving as
fp of a veclor p, to another scalar function such as Ft of

by treating p as equal to some vector function <pt of this last scalar, we
reduced, the scalar function
a scalar
t,
infer that
XIX.
(6.)
And
tions given
difficulty is
dF<
= 6f(l)t =
2S.
if
v6<pt,
d/o
= 2Sj^dp,
as regards the applications of the formula3 VI.
as before.
and VII,, or of the equa-
for the normal and tangent plane to a surface generally, the

only to select, out of a multitude of examples which might be given:
not be useless to add a few such here, the case of the sphere having of
by them
yet it may
course been only taken to illustrate the theory, because the normal property of
radii was manifest, independently of any calculation.
(7.)
Taking then the equation of the
XX.
of
which the
first differential
XXI.
may
ellipsoid,
XXII.
under the form,
TOp + pjc) = /c2 - 12,
282 XIX.,
,
(see the sub-articles to
336) be thus wiitten,
= S-{(t-K)2p + 2(tSKp + KSip)}dp = S'dp,
and introducing an auxiliary vector, on or

.
?,
ON = ^ = - 2 (t -
such that
k)
its
(tS/cp
KSip),
CHAP.
III.]
NORMALS TO SURFACES OF THE SECOND ORDER. 505
we have v\\p ^, and may
normal
write, as the equation of the
at the extremity
of p, the following,
XXIII.
V. (^
- p)
- p) = 0,
(a>
XXIY.
or
u)
= p + x(^ -
p),
a scalar variable (comp. 369, VII.)

making then x=l, we see that
the vector of the point n in which the normal intersects the plane of the two
which
in
is
a; is
fixed lines
t,
supposed to be drawn from the origin, which
k,
is
here the centre of
the ellipsoid.
we look back on the sub-articles to 216 and 217, we shall see that these
K have the directions of the two real cyclic normals, or of the normals to the
two (real) cyclic planes; which planes are now represented by the two equations,
(8.) If
lines
t,
XXV. ..Sep = 0,
XX.
Accordingly the equation

under the cyclic forms,
XXVI.
Sp0p =
Sfcp
(i3
+ K2)p2 +
= 0.
be put (comp. 336, 357, 359)
may
of the ellipsoid
2Stp/cp
= (c - /c)2 p2 + 4StpSKp = (k2 - t2)2 = coust.

hence each of the two diametral planes XXV. cuts the surface in a
mon radius of these two circular sections being
Tt2
XXVII...
circle,
the com-
_ TrS
Tp=,^^_^=,
where 6 denotes, as in 219, (1.), the length of the mean semiaxis of the ellipsoid
and in fact, this value of Tp can be at once obtained from the equation XX., by
;
making
(9.)
either ip
By
= - pi,
or pK
= Kp,
in virtue of
lengths,
XXVIII.
a = Tt
+ T/c,
XXV.
and
the sub-article last cited, the greatest
semiaxes have for their
least
= Tt-TK;
and the cons^rwcfion in 219, (2.) shows (by Fig. 53, annexed to 217, (4.)) that
have the respective directions of the lines,
these three semiaxes a, b, c
XXIX.
all
XXX
1-
^^"^^
(K2-t2)2
iTk- kTi,
YiK,
iTk
kTi
+ pO^
_/T(ip
_/ T(ip+pO
Y
Ti + T/c
k2-i2
S.pU(tTic-icTO Y
V
in
which agrees with the rectangular transformation,
T(t-OS.pUVt<c) V
Tt*-T/e2
"^
"^
S.pU(tT/c+ icTQ \2
Tt-T/c
J'
deducing which (comp. 359, (1.)) from 357, VIII., by means of the formulae
XX. and XXI., we employ the values (comp. XXVL),
357,
XXXI.
.^=i2 +
k2,
X=
2t,
fl
= K.
is therefore also
(7.), of the cyclic normals i and k (8.),
(10.)
the plane of the extreme semiaxes, a and c (9.), or that which may be called per-
The fixed plane
haps the principal plane* of the ellipsoid
* This
plane may also be said
to
namely, the plane of the generating
be the plane of the principal
tri-
elliptic section
(219, (9.)) ; or it may be distinguished (comp. the Note to page 231) as the plane
of the /ocaZ hyperbola, of which important curve we shall soon assign the equation
in quaternions.
3 T
506
[bOOK
III.
angle (218), (l.)), in that construction of the surface (217, (6.) or (7.)) which
tion
of
is
and was deduced as an interpretation of the quaternion equaXX., or of the somewhat less simple form 217, XVI., with the value Ti2- Tk'
illustrated
by
Fig. 53,
<-.
(11.) Let n denote the length of that portion of the normal, which
between the surface and the principal plane (10.), so that, by
XXXII. .. = NP=:T(p-0,
with the value XXII. of
is
intercepted
(7.),
n^=^-{p-l)\
= os
Let c
be the vector of a point s on the surface of

a new or auxiliary sphere, described about the point N as centre, with a radius = n,
and therefore tangential to the ellipsoid at p and let us inquire in what curve or
|.
curves, real or imaginary, does this sphere cut the ellipsoid.
(12.)
The equations (comp. 371,
(5.)) of the sought intersection are the
two fol-
lowing,
XXXIII.
whereof the
+ n2 = 0,
^)2
(cr
expresses that s
first
XXXIV.
and
T(t(T
+ ck) = k2 _ ,2
a point of the sphere, and the second that
is
it is
point of the ellipsoid ; while p or op enters virtually into XXXIII., through | and w,
but is here treated as a constant, the point p being now supposed to be a given one.
shall remove (18) the origin to this point p of the ellipsoid, if we
(13.)
We
write,
XXXV.
(r
= p+(T',
or
XXXV'.
0-'=
(T-p
= ps;
and thus we obtain the new or transformed equations,
XXXVL.
in
(7.),
XXXVIII.
also 210,
comp.
.
XXXIX.
and
= (i .
(14.) Eliminating then
XXXVIL
like
..
=N(i(r'
(t'*:)
2Sj/(t';
XX.),
+ 2 (iS/cp + /cStp) = (i - (c)2 (p - ^),

+ g'k) = {i- nf tr's + 4St(T'S/ca'.
K)2p
(((t'
(t'2,
we
.
XXXVI.
obtain from the two equations
this other,
XL.
which
XXXVIL
0=<7'2+2S(p-O<y',
which (as in
them is of the second degree
Si(7'.Sk(t'
in
and
= 0',
a, but breaks up, as
Sind scalar factors, representing two distinct planes, parallel
we
by
see, into
XXV.
two linear
to the
two
diametral and cyclic planes of the ellipsoid.

The sought intersection consists then
of a pair of (real) circles, upon that given surface
namely, two circular (but not
;
diametral^ sections, which pass through the given point p.

(15.) Conversely, because the equations XXXVIL XXXVIIL
XXXIX.
XL.
XXXVI.
and XXXIII., with the foregoing values of ^ and n, it follows that

these two plane sections of the ellipsoid at p are on one common sphere, namely
give
that which has
for centre,
and n
for radius, as
above
and thus we might have
found, without differentials, that the line pn is the normal at P ; or that this normal
crosses the principal plane (10.), in the point determined by the formula XXII.
(16.) In general, the cyclic form of the equation of any central surface of the
second order, namely the form (comp. 357,
XLL
shows that the two
II.),
Sp0p =/p2 + 2SXpS/ip
circles (real or
imaginary) in
two planes,
XLII.
S\p = /,
= C= const.,
which that surface
S/ip
= m,
is
cut by
CHAP.
RECIPROCAL SURFACES.
III.]
507
drawn parallel respectively to the two real ct/clic planes^ which are
sented (comp. XL., and 216, (7.)) by the one equation,
XLIII.
(17.)
/p2 +
S\pS/ip
= 0,
on that one sphere of which the equation
are homospherical, being both
XLI V.
jointly repre-
But the centre (say n) of
XLV.
2 (/S^p
this
new
+ mSXp) = 2?w +
is,
C.
sphere, has for its vector (say 4),
ON = 4 = - g-\ln + mX)
plane of the two real cyclic normals, \ and p, and if

/ and m in XLV. receive the values XLII., then this new ^ is the vector of intersecbecause it is (comp. 15.))
tion of that plane, with the normal to the surface at p
it is
therefore situated in the
which touches (though
the vector of the centre of a sphere
also cuttinp, in the
two
circular sections) the surface at that point.
(18.)
We
can therefore thus infer (comp. again (15.)), without the differential
calculus, that the line,
XLVL
ffXp
as having the direction of np,
is
-V)= g'p + XSjup + juSXp = ^p,

the normal at
and may be considered as confirming
with,
clusion otherwise obtained
through the
XLVIL
(if
to the surface
XLI.
which agrees
confirmation were required), the con-
differential expression (361),
dSp0p
= 2Svdp = 2S^pdp;
the linear function <pp being here supposed (comp. 361, (3.)) to be self-conjugate.
the equation of the tangent plane to a
(19.) Hence, with the notation 362, I.,
same point
central surface of the second order, at the

written,
in
XL VIII.
which
/(w, p)
C,
if
Sp0p =
may by
C= const.
VII. be thus
remembered, that
to be
it is
p,
XLIX.
=
/(w, p) =/(p, w) Sw0p
= Sp^w.
And if we
choose to interpret this equation XLVIIL, which is only of the

with
respect to each separately of the two vectors, p and w, or op
first degree (362)
and OR, and involves them symmetrically, without requiring that p shall be a point
(20.)
we may then say (comp. 215,
on the surface,
mula
in question is
points
p and
(13.),
and 316, (31.)), that the
for-
an equation of conjugation, which expresses that each of the two
r, is situated in the
(21.) In general,
if
polar plane of the other.
we suppose
that the length and direction of a line v are so
adjusted as to satisfy the two equations (comp. 336, XII. XIII. XIV.),
L.
Svp
Svdp
1,
= 0,
and therefore
also
then, because the equation VII. of the tangent plane to
be thus written,
LII.
it
follows that
Sr(w-v-i) =
represents, in length
that tangent plane at
so that v
and
i^se//"
LI.
Spdv =
any curved surface may now
0,
direction, the
perpendicular from o on
represents the reciprocal of that perpendi-
cular, or what may be called (comp. 336, (8.)) the vector of proximity, of the tanAnd we see, by LI., that the two vectors, p and r, if
gent plane to the origin.
drawn from a common origin, terminate on two surfaceu which are, in a known and
508
[bOOK
III.
important sense (comp. the sub-arts, to 361), reciprocals* of one another: the line
p-\ for instance, being the perpendicular from o on the tangent plane to the second
surface, at the extremity of the vector v.
we have
374. In the two preceding Articles,
treated the symbol

a
curve
on a given surface,
dp
and therefore also to that surface itself; and thus the formula
as representing (rigorously) a tangent to
Spdp = has been considered as expressing that v has the direction of

the normal to that surface, because it is perpendicular to two tangents
(372), and therefore generally to every tangent (373), which can be
drawn at a given point p. But without at present introducing any
oiher\ signification for this symbol dp, we may interpret in another
way, and with a reference to chords rather than to curves, the differential equation,
I.
dfp=2Svdp,
be a rigorous one (in virtue of our definitions of difsupposed

which
do not require that dp should be small); and may
ferentials,
still
to
deduce from it the normal property of the \ector v, but now with
the help of Taylor'^s Series adapted to quaternions (comp. 342, 370).
In fact, that series gives here a differenced equation, of the form,
still
II.
where
is
Afp = 2SvAp +
Pi',
a scalar remainder (comp. again 342), having the pro-
perty that
III.
whence
IV.
lim.
.
(E TAp) = 0,
:
Compare the Note
It is permitted, for
=
TA/?) 2 lim. Si/UA/>,
whatever the ultimate direction of Ap
lim. TA/>
if
lim. (Afp
may
be.
If then
we
conceive that
page 484.
example, by general principles above explained, to treat the
differential dp as denoting a chorda! vector, or to substitute it for Ap, and so to represent the differenced equation of the surface under the form (comp. 342),
to
= A/"p = (6d-l)/p=d/p + id2/p + &c.;

but with this meaning of the symbol 6p, the equation dfp = 0, or SvdjO = 0, is no
longer rigorous, and must (for rigour) be replaced by such an equation as the following.
the remainder
= 2Svdp +
Sdj/dp
vanishing,
when
+ R,
if
d/p
the surface
= 2Svdp,
is
as before
only of the second order (comp.
some investigations, especially

in those which relate to the curvatures of normal sections : but for the present it
362, (3.)).
Accordingly this last /orm
is
useful in
seems to be clearer to adhere to the recent signification of dp, and therefore to

it as still denoting a tangent, which
may or may not be small.
treat
CHAP.
NEW ENUNCIATION OF TAYLOR's THEOREM.
III.]
509
a small and indefinitely decreasing chord pq of the surdrawn from the extremity p of />, so that
face^
^P represents
V.
A/p =/ (/> +
A/>)
-fp =
and
0,
lim.
TAp = 0,
the equation IV. becomes simply,
VI..
lim. Si/UA/3
and thus proves, in a new way, that v
by proving that
proposed point p,
the
it is
when
chords vafroin that point,
is
= 0;
normal
to the
surface at the
ultimately perpendicular to all
those chords become indefinitely
small, or tend indefinitely to vanish.

For example,
(1.)
VII.
= p^,
.fp
if
v=
thus, for every point
of space,
IX.
and
we have
/?
in fact, either of these
two
and
R TAp = - TAp
:
- TAp
const., we have
A/p TAp = 2SpUAp

:
2SpUAp = TAp,
= Ap^,
rigorously, with this form of ^,
Q of the spheric surface, fp
for every point
X.
VIII.
then
p,
XI.
or
=
.
with equal rigour,
pq = 2op.cosopq
of a
and of
last formulas expresses simply, that the projection
diameter of a sphere, on a conterminous chord,

course diminishes with it.
is
equal to that chord
itself,
(2.) Passing then to the limit, or conceiving the point Q of the surface to approach indefinitely to p, we derive the limiting equations,
XII.
lim.
SpUAp =
XIII. .. lim. cos opq
which shows, in a new way, that the radii of a sphere are its normals ;
with the analogous result for other surfaces, that the vector v in I. has a normal dieither of
rection, as before: because its projection
on a cAorc? pq tends indefinitely to diminish
with that chord.

(3.)
II.
and
We may
III., that
also interpret the diflferential equation I. as expressing,
the plane 373, VII.,
which
direction perpendicular to v, is the tangent
jection of the chord
Ap on
the
XIV.
normal v
.
is
through
in a
drawn through the point p
plane to the surface
because the pro-
to that plane, or th& perpendicular distance,
- S (Uj/. Ap) =
|i2.Ti/-i,
of a near point Qfrom the plane thus drawn through p, is small of an order higher
than the first (comp. 370, (8.)), if the chord fq itself be considered as small of the
first order.
375. This occasion may be taken (comp. 374, 1. II. III.), to give
a neiu Enunciation of Taylor'' s Theorem, in a form adapted to Quaternions, which has some advantages over that given (342) in the preshall therefore now express that
ceding Chapter.
important
We
Theorem as follows
"
If none of the
m\\ functions.
510
I.
become
infinite
fq, ^fq,
d'^fq,
III.
d^^ = 0,
which
in
d'-fq,
[bOOK
in the immediate vicinity of a given quaternion
q^
then the
quotient,
II.
a= [/fe+c3^)-/?-d/^-^-J- ^^-&c.
d^fq
d^"
^
'
2.3..W'
2.3..m)
can he made
versor
to
Udg, by
(1.)
tend indefinitely to zero^for any ultimate value of the

indefinitely diminishing the tensor Tdg-."
The ^roo/ of the theorem, as thus enunciated, can
easily be supplied
attentive reader of Articles 341, 342, and their sub-articles; a few hints
ever here be given.

do not
(2.)
We
we
now
only assume that
by an
may how-
suppose, as in 342, that d'"/(^ must be different from zero;

not infinite : and we add^ to the expression 342, VI. for
it is
Fx^ the term,
- ^""^"/g
III
2. 3... TO
(3.)
Hence each of the expressions 342, VII.,
for the successive derivatives oi
Fx, receives an additional term; the last of them thus becoming,

IV.
DFa;
= F(.*>^)x = d'/(5 + xdx) - ^^fq
we have now (comp. 342, X.) the values

V... 7^0 = 0, F'O, i^"0=0,... FC"-J)0 = 0,
so that
(4.)
F(')0
= 0.
Assuming therefore now (comp. 342, XII.) the new auxiliary

a;'"do
^
VI.
i//a;=
with
function,
Tdj>0,
7n
which gives,
VII.
we
find
ipO
(by 341,
= 0,
;//'0
= 0,
-.^"0
= 0,
^//('"-DO
0,
;p(')0
= dj*",
comp. again 342, XII.) that
(8.), (9.),
VIII.
lim.
{Fx ^x) =
:
0.
x=0
(5.)
But
these
two new
the divisor of the quotient
functions,
in II.,
Fx and
;//
j,
are formed from the dividend and
by changing d^
to
xAq
instead of thus multiplying a given quaternion differential dg,

nitely decreasing scalar, x,
we may
and (comp. 342, (3.))

by a small and indefi-
indefinitely diminish the tensor, Tdg, without
changing the versor, Udg-.

(6.)
to zero,
And even (/"Ud^ he changed, while the differential dj is thus made

we can always conceive that it tends to some limit ; wliich limiting
mate value of that versor Udg
may
to tend
or ulti-
then be treated as if it were a constant one, with-
out affecting the limit of the quotient Q.

as above enunciated, is therefore fully proved
and we arc at
(7.) The theorem,
liberty to choose, in any application, between the two forms of statement, 342 and
;
375, of which one
is
more convenient
at one time,
and the other at another.
CHAP.
OSCULATING PLANES TO CURVES IN SPACE.
III.]
511
Ow
Osculating Planes, and Absolute Normals, to

Curves of Double Curvature.
Section 4
variable vector pt of a curve in space may in general

be thus expressed, with the help of Taylor's Series (corap. 370,
The
376.
(1.)):
I.
"
Pt^ P^tp'
-^
with
^fûp",
Wo
u being here abridged symbols for pq^ p'o, p''qi w<; and the
/), /)', p"j
up"
being a vector, although the factor u is generally a quaproduct
And the different terms of this expresternion (comp. 370, (5.)).
sion
may be thus
I.
II.
while
III.
= OP
.
constructed (compare the annexed Figure 77):

tp'
/>i=OQ,
= PT
and
\fup" = TQ
^ ^fup" = ^(^-^
;
tp'
the line tq, or the term ^tûp", being thus what

may be called the deflexion of thecwrrePQR, at q,
from
its
tangent
pt
measured in a
at P,
direction
which depends on the law according to which

varies with t^ and on the distance of q, from
pt
p.
The equation of the plane of the
triangle ptq, is
rigorously (by II.) the following,
with w for
its
variable vector,
IV.. .0=Sw/y(tu-/5);
IV. then touches the curve at
this plane
Q; so that if the point

the resulting formula,
V.
q,
Sp"p'{iv
p,
and (generally)
cuts it at
be conceived to approach indefinitely top,
-p\
or
= S//^(w - p\
that limiting position^ in which
v.
is
the equation of the plane ptq, in
is
called the osculating plane, or is said to osculate to the curve pqr,
it
at the point p.
a,
function ps of a variable
a function of the former scalar variable t, we shall then
(1.) If the variable vector p be immediately given as
scalar,
s,
which
is itself
have (comp. 331) the expressions,

Yl.
p't
= s'DsPsi
p"t
= s"DsPs+s"''Ds^ps,
with
s'
= T>ts,
s"='Dt'^s;
thus the vector p" may change, even in direction, when we change the independent
scalar variable ; but p" will always be a line, either in or parallel to the osculating
plane ; while p will always represent a tangent, whatever scalar variable may be
selected.
(2.)
As an example,
let
us take the equation 814, XV., or 369, XIII., of the
512
With
helix.
the independent variable
of that equation,
[bOOK
III.
we have (comp. 369, XIV.)
the derived expressions,
VII.
= ca +
p'
1 a^-iA
P"
-(fp/3
[^]V-p);
p" has therefore here (comp. 369, (8.)) the direction of the normal to the cylinder
and consequently, the osculating plane to the helix is a normal plane to the cylinder
;
of revolution, on which that curve

soon be greatly extended.
is
traced
a result well known, and which will
into a. plane curve, its oscu(3.) When a curve of double curvature degenerates
The condition of planarity of a
lating plane becomes constant, and reciprocally.
curve in space
therefore be expressed
may
VIII.
or,
by 335,
II.,
UV|o'p"
by the
=+ a
equation,
constant unit line
and 338, VIII.,
IX...0 = V^yp7 = v"^PP'"-
'
X.
or finaUy,
= 0,
Sp'p'p"'
(4.) Accordingly, for a
Vp'p"
or
XI.
plane curve,
if
Vp'p"
.
p"
p',
p".
\ be a given normal
to its plane,
we have
the three equations,
XII
which conduct, by
(5.) For example,
S\p'
294, (11.), to
if
= 0,
SXp" =
0,
SXp'"
X.
we had not otherwise known that the equation 337, (2.)

we might have perceived that it was the equation of some
represented a plane ellipse,
plane curve, because
XIII.
p'
which are complanar

(6.)
And
it
gives the three successive derivatives,
Va'î/3,
lines,
p"=-
(^^y
ya%
p"'
=-
i^^Jâ^^
the third having a direction opposite to the
first.
generally, the formula X. enables us to assign, on
curvature, for which p
is
expressed as a function of
t,
resembles a plane curve, or approaches most closely to
any curve of double

tbe points* at which it most
its
own
osculating plane.
377. An important and characteristic property of the osculating

plane to a curve of double curvature, is that the perpendiculars let
fall on it, from points of the curve near to the point of osculation,
are small of an order higher than the second^ if their distances from,
thatômf be considered as small of XhQ first order,
(1.)
To
exhibit this
by quaternions,
let
us begin by considering an arbitrary
plane,
* Namely, in a modern phraseology, the places of four-point contact with a

The equation, Vp'p" = 0, indicates in like manner the places, if any, at which
plane.
a curve has three-point contact with a right
line.
For curves of double curvature,
these are also called points of simple and double inflexion.
CHAP.
513
CONE OF PARALLELS TO TANGENTS.
III.]
I.
S\(w-p) =
TX = 1,
with
0,
drawn through a point p of the curve. Using the expression 376, I., for the vector
of the same curve, we have, for the perpendicular dispt, of another point q
OQ, or
tance of
Q from the plane

II.
this other rigorous expression,
I.,
.
SX(pt
- p) = tSXp +
^f^SXup"
which represents, in general^ a small quantity of the^r< order y if
be assumed to
be such.
The expression
(2.)
second order,
however, generally^ a small quantity of the
II. represents
the direction of
if
satisfy the condition,
III.
that
the plane
is, if
(3.)
And
if
I.
SX/o'
= 0;
the condition,
IV.
be also satisfied
by
SXp"=0,
X, <Ae, but not otherwise, the expression II. tends to bear
an order higher than the second.

But the combination of the two conditions. III. and IV., conducts
evanescent ratio to
(4.)
touch the curve.
f^,
expression,
V.
378.
is,
to the
.X = UVpy';
comparing which with 376, V., we see that the property above stated
belongs to the osculating plane, and to no other.
curve
an
or is small of
is
one which
Another remarkable property* of the
osculating plane to a
the tangent plane to thecowe of parallels to tangents
which has its vertex at the point of osculation.
that
(369, (6.)),
it is
(1.) In general, if p
= ^jc
be (comp. 369,
I.)
the equation of a curve in space,
the equation of the cone which has its vertex at the origin, and passes through this
curve,
is
of the form,
I... p
in
= y0x;
which x and y are two independent and scalar variables.

(2.) We have thus the two partial derivatives,
II.
Dxp
and the tangent plane along the side

in.
(3.)
.0
= y<p'x,
= S(oj.<px. <}>'x)
Changing then
x, ^, 0',
Dyp
= ^x
(x) has for equation,
to
or briefly.
t,
p',
p'\
v., of the oscxdating plane to the curve 376,
III',
a> - p, we
I., is
= Sw^^'.
see that the equation 376,
also that of the tangent
plane to
the cone ofparallels, &c., as asserted.
379.
Among
all
the normals to a curve^ at any one point, there

;
namely the one which is in
are two which deserve special attention
* The writer does not remember

seeing this property in print ; but of course
an easy consequence from the doctrine of infinitesimals , which doctrine however
has not been thought convenient to adopt, as the basis of the present exposition.
is
3 u
it
it
514
[bOOK
III.
the osculating plane, and is called the absolute (or principal) normal ;
and the one which is perpendicular to that plane, and which it has
been lately proposed to name the hinormal* It

pressions,
(1.)
and
p",
334)
quaternions, for these
by
as follows.
The absolute normal, as being perpendicular to p', but complanar with p'
has a direction expressed by any one of the following formulae (comp. 203,
I.
or
Vp"p'.p'-i;
II.
good
or
dUp';
(2.) There is an extensive classf of cases, for
III.
dUdp.
which the following equations hold
IV.
and
easy to assign ex-
is
two normals,
Tp'
= const.
V.
p'2
= const.
such cases, the expression I. reduces

representative of the absolute normal.
in all
(3.)
For example, in the case of the
VI.
itself to p",
Sp'p"=
which
is
therefore then a
with the equation several times be-
helix^
and accordingly the absolute norwith the normal p" to the cylinder, on which it is traced ;
the locus of the absolute normal being here that screw surface or ffelicoid, which
has been already partially considered (comp. 314, (11.), and 372, (4.)).
fore employed, the conditions (2.) are satisfied
mal
to that curve coincides
(4.)
cause
And
it is
as regards the binormal,
it
may
be sufficient here to remark, that beit has the direction expressed

by
perpendicular to the osculating plane,
one or other of the two symbols (comp. 377, V.),
VII.
(5.) There
Yp'p",
or
Vir.
Vdpd2p.
a system ot three rectangular planes, the osculating

plane being one, which are connected with the system of the three rectangular lines,
the tangent, the absolute normal, and the binormal, and of which any one who has
exists, of course,
studied the Quaternions so far can easily form the expressions.

(6.)
And
a constructionX for the absolute normal
may
be assigned, analogous
to and including that lately given (378) for the osculating plane, as
tion of the expression II. or III., or of the
symbol dUp'or dUdp.
an interpreta-
From any
origin
conceive a system of unit lines (Up' or Udp) to be drawn, in the directions of the
successive tangents to the given curve of double curvature
these lines will terminate
;
*
By M. de Saint-Venant, as being perpendicular at once to two consecutive elements of the curve, in the infinitesimal treatment of this subject. See page 261 of the
very valuable Treatise on Analytic Geometry of Three Dimensions (Hodges and Smith,
Dublin), by the Rev. George Salmon, D. D., which has been published in the present
year (1862), but not till after the printing of these Elements of Quaternions (begun in
1860) had been too far advanced, to allow the writer of them to profit by the study
of it, so much as he would otherwise have sought to do.
t Namely,
those in which the arc of the curve, or that arc multiplied
by a
scalar
taken as the independent variable.

X This construction also has not been met with by the writer in print, so far as
he remembers but it may easily have escaped his notice, even in the books which he
constant,
is
has
seen.
CHAT.
ABSOLUTE NORMALS, GEODETIC LINES.
III.]
on a certain spherical curve; and the tangent, say

8 which corresponds to the point
solute
to this
ss',
515
new curve,
at the point
of the old one, will have the direction of the ab-
normal at that old point.

At the same time, the plane oss' of thQ great
(7.)
circle,
which touches the new
curve upon the unit sphere, being the tangent plane to the cone of parallels (378),
has the direction of the osculating plane to the old curve; and the radius drawn to
its
pole
is
parallel to the hinormal.
As an example
of the auxiliary (or spherical) curve, constructed as in (6.),

take again the helix (369, XIII., &c.) as the given curve of double curvature, and observe that the expression 369, XIV., namely,
(8.)
we may
VIII.
p'
whence Tp'
+^
ca
gives IX.
a+i/3,
= - c2
we have
constant (as in IV.), and
is
p'2
-f
^^-~-
= const,
(comp. (3.))
the equation (comp. 369,
XV.
XIX.),
/
X.
SaUp' = -c[
7r2j32YJ
c2
J-
=-cosa = const.,
a being again the inclination of the helix to the axis of its cylinder ; which shows
new curve is in this case a plane one, namely a certain small circle of
that the
the unit sphere.
In general,
(9.)
if
the given curve be conceived to be an orhit described
by a
which moves with a constant velocity taken for unity, the auxiliary or spherical curve becomes what we have proposed (100, (5.)) to call the hodograph of that
point,
7notion.
(10.)
city,
And
the given curve be supposed to be described with a variable velo-
if
the hodograph
Section
5.
is still
some curve upon the cone of parallels
to tangents.
On Geodetic Lines, and Families of Surfaces,
380. Adopting as the definition of a geodetic line, on any proposed

curved surface, the property that it is one of which the osculating
plane is always a normal plane to that surface, or that the absolute
normal to the curve
is
also the
normal to the surface,
we have two
principal modes of expressing by quaternions this general and characteristic property.

For we may either write,
I.
^vp'p'fÔ,
or
II.
Si/d/5d-/)
= 0,
normal v to the surface (comp. 373) is perpendicular to the hinormal Yp'p'' or Vd/3d^/3 to the curve
(comp, 379,
to express that the
VII. VII'.)
or
III.
else, at
.
pleasure,
Yv{\]p')'
to express that the same
= 0,
or
IV.
normal v has the
YudJJdp =
0,
direction of the absolute
normal
II. III.), to
line.
deduce the known
such
(UpJ or dUd/5 (comp. 379,

And thus it becomes easy to
lines (or curves) to
some
\m-ôii2Jit families
the same geodetic

relations of
of surfaces, on which
516
III.
[bOOK
they can be traced. Accordingly, after beginning for simplicity

with the sphere, we shall proceed in the following sub-articles to de-
termine the geodetic lines on cylindrical and conical surfaces, with
show how the correspondbe investigated, upon developable surfaces, and surfaces
arhitrary hoses; intending afterwards to
ing lines can

of revolution.
On
(1.)
a sphere, with centre at the origin, -we have v
equation IV. admits of an immediate integration ;* for
V.
b)
= VpdUdp = dV/oUd/o,
being some constant vector
whence VI.
the curve
is
it
\ p,
and the
differential
here becomes,
YpVdp = w, and
VII.
a great
therefore in this case
Swp =
circle,
0,
as
being wholly contained in one diametral plane.

(2.) Or we may observe that the equation,
VIII.
Spp'p"= 0,
IX.
or
SpdpdV =
0,
obtained by changing v to p in I. or II., has generally for a first integral (comp.

335, (1.)), whether Tp be constant or variable,
X.
it
UVpp' = UVpdp = w = const.
expresses therefore that p is the vector of
the origin
which curve must consequently be here a great
(3.) Accordingly, as a verification of X., if
XI.
where
we
t is still
shall
some curve (or
some independent
we
x and y being
ajc 4- j8y,
line) in
write
scalar functions of
and
scalar variable,
a plane through
circle^ as before.
a, /3 are
f,
two vector constants,
have the derivatives,

XII.
p'
= aar' + ^y',
p"
= ax" -^ (Sy" \\\ p, p'
',
so that the equation VIII. is satisfied.
(4.)
we may
For an arhitrary
cylinder, with generating lines parallel to a fixed line a,
write,
XIII.
.âv = 0,
XIV.
SadUdp =
XV.
0,
SaUdp = const.
a geodetic on a cylinder crosses therefore the generating lines at a constant angle,

and consequently becomes a right line when the cylinder is unfolded into a plane :
both which known properties are accordingly verified (comp. 369, (5.), and 376,
(2.)) for the case of a cylinder of revolution, in
(5.) For
an arbitrary
which case the geodetic is a helix.

we have the equations,
cone, with vertex at the origin,
XVII.
Svp = 0,
SpdUdp = 0,
XVIII.
dSpUdp = S (dp.Udp) = - Tdp
XVI.
multiplying the last of which equations

= - d p^, we obtain the transformations,
by 2SpUdp, and observing that
- 2Spdp
We here
assume as evident, that the differential of a variable cannot be conand we employ the principle (comp. 338, (5.)),
;
stantly zero (comp. 336, (7.))

- VTdp = 0.
that V. dp Udp
CHAP.
XIX.
III.]
GEODETICS ON SPHERES, CONES AND CYLINDERS. 517
= d{(SpUdp)a + p2} = d. (VpUdp),
the vertex, on
a tangent
XX.
TVpUdp = const.
any one geodetic upon a cone, has
to
the perpendicular from

therefore a constant length; and all such tangents touch also a concentric sphere,*
or one -which has
its
centre at the vertex of the cone.
(6.) Conceive then that at each point
p or
p' of the geodetic
drawn, and that the angles otp, ot'p' are right

been shown,
is
XXI.
and
or = or' = const. = radius
we
pV
a tangent pt or
by what has just
shall have,
of concentric sphere
the cone be developed (or unfolded) into a plane, this constant or

length, of the perpendiculars from o on the tanif
common
gents, will remain unchanged, because the length
OP and
the angle
velopment
opt
are unaltered
by such de-
the geodetic becomes therefore some
plane line, with the same property as before

and although this property would belong, not
;
only to a
but also to a circle with o
rigJit line,
(compare the second part of the annexed Figure 78), yet we have in this result 0^
for centre
merely an
effect of
the foreign factor SpUdp,

in (5.), in order to facili-
which was introduced
Fig. 78.
tate the integration of the differential equation
XVIII., and which (by that very equation) cannot be constantly equal to zero. We
are therefore to exclude the curves in which the cone is cut by spheres concentric
it
and there remain, as the sought geodetic lines, only those of which the developments are rectilinear, as in (4.).
(7.) Another mode of interpreting, and at the same time of integrating, the
with
which can
equation XVIII., is connected with the interpretation of the symbol Tdp
be proved, on the principles of the present Calculus, to represent rigorously the differential ds of the arc (s) of that curve, whatever it may be, of which p is the variable vector ; so that we have the general and rigorous equation,
;
XXII.
Avhether that arc
and whether
its
Tdp
= d*,
some other
we
scalar,
t,
tiple, n-^dt, of
t,
suppose, for the sake of greater generality, that the vector
scalar s are thus both functions, pt
lar variable,
thus denote the arc
if s
be taken as the independent variable;

differential ds be small or large, provided that it be positive.
itself,
(8.) In fact if
and the
or
and
st,
of
p
some one independent and sca-
our principles direct us first to take, or to conceive as taken, a submulthe fnite differential dt, considered as an assumed and arbitrary in-
crement of that independent variable,
to determine
next the vector
pt+iT^dt,
and
the scalar st+n^dt, which correspond to the point Vunht of the curve on which
pt terminates in p^, and of which st is the arc, ^^ti measured to p^ from some fixed point
Po on the
same curve
When
the cone
to take the differences,
is
of the second order, this becomes a case of a
respecting geodetic lines on a surface of the
same second
one of which curves touch also a confocal surface.
known theorem
order, the tangents to
any
518
pt+n~^dt
a^<i
Ph
St^nÂt
III.
[bOOK
St,
which represent respectively the directed chord, and the length, of the arc P<p<+n''d<,
which arc will generally be small, if the number n be large, and will indefinitely diminish when that number tends to infinity ; to multiply these two decreasing
and su by n
and
diffe-
which the products tend,

when n thus tends to od such limits being, by our definitions, the values of the two
sought and simultaneous differentials, dp and ds, which answer to the assumed varences, of pt
finally to seek the limits to
lues of
and
And
d*.
because the small arc, As, and the length, TAp, of
its
small
chord, in the foregoing construction, tend indefinitely to a ratio of equality, such

must be the rigorous ratio of ds and Tdp, which are (comp. 320) the limits of their
equimultiples.
(9.)
Admitting then the exact equality XXII. of TdjO and ds, at least when the
we have only to substitute d for Tdp iu]
latter like the former is taken positively,
the equation
XVIIL, which
XXIII.
then becomes immediately iutegrable, and gives,

s
+ SpUdp = - S (p Udp) = const.

:
where S (p Udp) denotes the projection tp, of the vector p or op, on the tangent to]
the geodetic at p, considered as a positive scalar when p makes an acute anglet
:
with dp, that
is,
when
the distance
or
Tp
op from
denotes, as above, the length of the arc PqP of the
the vertex
same
is
increasing; while
curve, measured from somei
fixed point Vq thereon, and considered as a scalar which changes sign, when the va-j
liable point
passes through the position Pq.
(10.) But the length of tp does not change (comp. (6.)), when the cone
loped, as before ; we have therefore the equations (comp. again Fig. 78),
XXIV.
^^
TP = const.
PoP
/-N
= Pqp' - t'p',
XXV.
r-N
.
pp'
is
deveî
t'p'
- tp,
which must hold good both before and after the supposed development of the conic
surface ; and it is easy to see that this can only be, by the geodetic on the cone b
In
coaling a right line, as before.
tp
the tangent
member
of
in
XXV.
plane be supposed to interse

be conceived to approach to p, the second
bears a limiting ratio of equality to the first member, increased oi
a point
and
t',
fact, if ot' in the
if p'
diminished by tt* which latter line, and therefore also the angle tot* between
perpendiculars on the two near tangents, or the angle between those tangents them^
;
selves, if existing,
must bear an
indefinitely decreasing ratio to the arc PP'
the radius of curvature of the supposed curve
the development
(11.)
is rectilinear
XXVI.
p'
so thi
as before.
may
equation, Tdp = ds (XXII.), conducts to manj

be put under several other forms.
For example, we
write generally,
XXVIII.
also
if
or t' coincides with T,
The important and general
other consequences, and
may
is infinite,
and
s'
be the
(12.)
And
Sp'p"+
if
= 1,
= 0;
XXVII.
(D,p)2 + 1
=
+ (Dis)2 = 0, or XXIX.
p'2 + s'2
TD,p
s's"
o,
and
s, taken with respect to any independei

whence, by continued derivation,
derivatives of p
first
scalar variable, such as
XXX.
(d<p)2
= 0,
XXXI.
Sp'p"'+ p"2 + s's"
+ s"2= 0,
&c.
the arc s be itself iak&a. as the independent variable, then (comp.
379, (2.)) the equations
XXXII.
XXIX.,
.
p'2
&c., become,
0,
Sp'p"
= 0,
Spp"'+
p"2
= 0,
&c.
CHAP.
III.]
DIFFERENTIAL AND SCALAR EQUATION.
519
381. In general, if we conceive (comp. 372, I.) that the vector p

of a given surface is expressed as a given function oitwo scalar varia-
X and y, whereof one, suppose y, is regarded at

known function of the other, so that we have again,
bles,
1.
= (p{x, y),
where the form of
is
but now with
first as
an un-
II.
^ =fx,
that
is
but
known^
of/ sought; we may then
.
regard p as being implicitly a function of the single (or independent)

scalar variable^ x, and may consider the equation,
III.
p = (p(x,fx\
as being that of some curve on the given surface, to be determined by

Denoting then the unknown total derivative
assigned conditions.
the
known partial derivatives of the same first
but
T><p(x,fx) by />',
order by Dj0 and Dy0, with analogous notations for orders higher
than the
IV.
in
p'
first,
(comp. 376, VI.) the expressions,
= D^0 + y'-Dy(p,
which
normal
we have
p" =D/0 +
2y'j>jyy(i)
=
-f'x, y" j)^y-f'x^
y' -T)^y
&c.
+ 3^''d/0 -v y"^y^, &c.
Hence, writing for the
v to the surface the expression,
V.
1/
= V (d^0
Dj,0)
= V. D^0Dj,0,
comp 372,
.
or this vector multiplied by any scalar, the equation 380,

geodetic line takes this new form,
VI.
by
or,
I.
V.,
of a
o = Si^/>y^=s(v.D,0D,0.V/)yO;
a general transformation which has been often employed
already (comp. 352,
VII.
XXXI.,
&c.),
O = S/d,0.S/>''d,0-S/>'d,0.S/X0;
and thus, by substituting the expressions IV. for />' and p'^, we obtain an ordinary (or scalar) differential equation, of the second order,
X and y, which is satisfied hy all the geodetics on the given surface,

and of which the complete integral (when found) expresses, with two
arbitrary and scalar constants, theform of the scalar function f in II.,
in
or the law of the dependence of
y on
x, for
the geodetic curves in
question.
(1.)
As'an example,
let
us take the equation,
VIII.
of a cone with its vertex at the origin
form
of the vector function
//
is
given, that
through which the sides of the cone
(a;,
y) =yy\fx,
comp. 378, I,
which cone becomes a known one, when the

is,
all pass.
when we know a guiding curve p = y^x,
We
have here the partial derivatives
520
IX.
X.
and
D^20
T>x<p=!/Dx^x=î/\p\
= t/r>x^x = y;^",
=
= \p\
Dj,0
d^Dj^^
[bOOK
= -^j
i//a;
Dy^(p
III.
comp. 378, II.;
the expressions IV. become, then,
XI.
p'
= y;p' + ///,
and
since only the direction of the
and
write,
= yr\/" + 2y'//' + y"^
p"
normal
XII. ..
is
important,
divide V.
(2.)
XIII.
=S(V4/^'.Vp'p")
;//3
and
i//'2
for the geodetic*
=S|o"i//Sp'//'-Sp";|/'Sp';//
= (ySi//4/" + 2y'Si//^//' + y"j//2)

- (ySi//'-^" + 2y'//'2 + y"Si//;//')
which
(y//'2
+ y'S;//;//')
(ySi//i//'
are abridged symbols for (;â:)3 and
+ y^^),
(;//'a;)2
but this equation in
X and y may be greatly simplified, by some permitted suppositions.

(3.) Thus, we are allowed to suppose that the guiding curve (1.)
XIV. ..Tâ; =
if
we
^2 = -l,
l,
pendent variable,
x,
we have
then,
XIV.,
With
(4.)
its
^^p'x =1,
4/'2
=-
combined with the
(12.),
= 0,
s^'xp"
1,
taken as the inde-
sî//"
=-
//'2
last equation
1.
0-(y-y")(-y)-(-2y')(-y')=yy"-2y'2-y2;
is
complete integral
found bg ordinary methods to be,
XVII.
in which 6
by 380,
1|/'2
is
these simplifications, the differential equation XIII. becomes,
XVI.
and
SJ/^" +
S-^4/'=0,
further assume that the arc of this spherical curve
XV.
the intersec-
is
with the concentric unit sphere^ so that
tion of the cone
and
y,
equation of the second order,
diflferential
.
by
= V4'//'.
a/
The expressions XI. and XII. give (comp. VI. and VII.)
on the cone VIII., the
in
we may
and
c are
two
y=
6 sec
(a;
c),
arbitrary but scalar constants.
To interpret now this integrated and scalar equation in x and y, of theêoon an arbitrary cone, we may observe that, by the suppositions (3,), y represents the distance, Tp or op, from the vertex o, and x + c represents the angle aop,
in the developed state of cone and curve, from some Jixed line OA in the plane, to the
(5.)
detics
the projection of this new
variable line
op
stant (being
= 6, by
line,
op on
t\iaX,
OA
fixed line
XVII.), and the developed geodetic
is
is
therefore con-
again found to be a right
as before.
382. Let ABODE
ries 0^ points in space.
annexed Figure 79) be any given
(see the
Draw
sive right lines, ab, bc, cd, de,

long them to points b', c', d',
the succes.
and pro-
e',
the
lengths of these prolongations being arbitrary ;
join also b'c', c'd', d'e^
shall thus
. .
We
have a series ofplane triangles,
nerally in different planes
so that
se-
.---""'\s.c'
xp^
^
^
^^^^^g-^^^"^
-^:^
-^
"""ê'
^^^' ^^'
b'bc', c^cb', d'de', ... all ge-
are generally
bcd'cV, cde'd'c',
gauche pentagons^ while bcde'dVb' is a gauche heptagon, &c. But we
;
CHAP.
DEVELOPABLE SURFACE, CUSP-KDGE.
III.]
can conceive the Jirst triangle b'bc' to turn round
521
its side bcc', till it
comes into the -pl&ne of the second triangle, c'cd'; which will transform the Jirst gauche pentagon into a plane one, denoted still by
We can
BC-D''cV.
then conceive
this
plane figure to turn round
its
comes into the plane of the third triangle^ d'de';

the
first gauche heptagon will have become a plane one, dewhereby
noted as before by bcde'd'c'e' and so we can proceed indefinitely.
are conceived
Passing then to the limit, at which the points abode
side odd', till it
to be each indefinitely near to the one which precedes or follows it in

the series, we conclude as usual (comp. 98, (12.)) that the locus of
the tangents to a curve
of double curvature is a developable surface : or

admits of being unfolded (like a cone or cylinder) into a plane,
without any breach of continuity. It is now proposed to translate
that
it
these conceptions into the language of quaternions, and to draw from

especially as regards the determi-
them some of their consequences:

nation of the geodetic
(1.) Let
lines,
or simply
tl/x,
on such a developable surface.
denote the variable vector of a point upon the curve^
i//,
or cusp-edge, or edge of regression of the developable, to which curve the generating

lines of that surface are thus tangents, considered as a. function yp of its arc, x, mea-
A upon
sured from some fixed point

will be of the
L
y being a second
IL
p'
= (1
Y.
The
so that while the equation of the surface
= <p(x,y) = ypx-\-yi''x='-^-\-y>l/\
we
shall
have the relations (comp. 381, XV.),
Sfi//"=0,
= -i//"2 = 2^
S;//';//"'
z=1^".
if
D^^ = ^' + y^", Dy0 =

+ y') ;^' + y^", p" = y";//' + (1 + 2/)^/" + yxjy'"
III.
and
(3.)
= -l,
//'2
Hence
IV.
scalar variable,
.Ti//'.,= 1,
(2.)
it
form (comp. 100, (8.)),
v=
|/'
V;//'^"
= \p'4'"j
multiplied
differential equation of the geodetics
by any
may
scalar.
therefore be thus written
(comp. 381, XIIL),
VL
in which,
by
(1.)
and
S(Vi//'i//".Vp'p";=Sp';//"Sp>'-Sp>"Sp'^';
(2.),
^Sp'^"
VII
= -y*^
ISp";^" =
(1
Sp"xP'=-y" + y2^,
2y') z2
_ y^2',
the equation becomes therefore, after division
VIII.
= - (1 + y')
by z,
= z{(l +yy + Q/zy} +
Sp'^P'
(1 +y') {yz)'-y"yz,
or simply,
IX.
(4.)
that
z,
.2+t;'=0,
or IX'.
To
now
interpret
or T//", or Td;//'
two near tangents
\f/'
and
this
TdJ/'
+ dt;=0,
if
X.
.tan
v=-i!^,=^^.
1+y l+y'
very simple equation IX. or IX'.,
we may
observe
dx, expresses the limiting ratio, which the angle between

//'
+ A\p',
to the cusp-edge (1.), bears to the small arc
3 X
Ax
522
[bOOK
III.
of that curve which
is intercepted between their points of contact ; while v is, by IV.,

that other angle ^ at which such a variable tangent, or generating line of the developable, crosses the geodetic on that surface; and therefore its derivative, v' or dr dar,
:
represents the limiting ratio, which the change A of this last angle, in passing from
one generating line to another, bears to the same small arc Ax of the curve which
those lines touch.

(5.) Referring then to Figure 79, in which, instead of two continuous curves,
there were two gauche polygons, or at least two systems of successive right lines, connected by prolongations of the lines of the first system, we see that the recent formula
IX. or IX'.
equivalent to this limiting equation,
is
^,
XI.
cd'c'-bc'b'
,.
.
lim.
CCD
=-1
but these three angles remain unaltered, in the development of the surface the hent
space becomes therefore ultimately a straight line in the plane, and si:
line b'c'd' for
milarly for all other portions of the original polygon, or twisted

which b'c'd' was a part.
is
line, b'c'd'e'
.,
of
(6.) Returning then to curves and surfaces in space, the quaternion analysis (3.)|
by this simple reasoning,* to conduct to an expression for the known andl
found,
characteristic property of the geodetics on
a developable : namely that they become]

and on cones (380, (6.) and (10.), or]
right lines, as those on cylinders (380, (4.)),
when
381, (5.)), were lately seen to do,

is
the surface on which they are thus

traced]
unfolded into a plane.
known
383. This
result, respecting geodetics on developahles^
may
be very simply verified, by means of a new determination of the absolute] normal (379) to a curve in space, as follows.
(1.)
The
arc s of
write (comp. 376,

I.
p.s
any curve being taken
I.),
we may]
for the independent variable,
Series, the following rigorous expressions,
by Taylor's
= p- sp' + is'û_sp",
po
= p,ps = p+ sp' + |2^p",
with
uo
1,
for the vectors of three near points, p_a, Pq, p, on the curve, whereof the second bisects the arc, 2*, intercepted
(2.) If then
between the
first
and
we conceive the parallelogram
have, for the two diagonals of this

II.
III.
we
shall]
figure these other rigorous expressions,
P-sPs=p,-(0-s=2sp' + ^'2(4-tt_s)p";
new
third.
p_4PoPsRs to be completed,
PoKs
(Os
+ p_s - 2/00 =
ls\Us
+ M_s) p"
',
* In the Lectures
(page 581), nearly the same analysis was employed, for geoon a developable but the interpretation of the result was made to depend on
detics
an equation which, with the recent
significations of \p
and
v,
may
be thus written, as
the integral of IX'., v + jTd;//' = const. ; where jTdi//' represents the finite angle between the extreme tangents to the finite arc jTdxp, or Ax, of the cusp-edge, whenl
that curve
is
t Called
developed into a plane one.

also,
and perhaps more usually, the principal normal.
CHAP.
GEODETICS ON DEVELOPABLES.
III.]
523
which ^ve the limiting equations,

IV.
(3.)
But
lim. s"ip-4Ps
8=0
is
V.
what may be
the length f_sP of
of the double arc, 2s,
= 2p'
= p".
lim. "2poRs
=0
called the long diagonal, or the
ultimately equal to that double arc
we have
chord
therefore,
by
IV., the equation,
VI.
if
Tjo'= 1,
p'
= Dsp,
and
if s
denote the arc,
considered as the scalar variable on which the vector p depends
a result agreeing
with what was otherwise found in 380, (12.).

(4.)
At
the same time, since the ultimate direction of the same long diagonal is
we see anew that the same ^rs< derived vector p'
evidently that of the tangent at Pq,
what may be called the unit-tangent* to the curve at that point.

And because the lengths of the two sides P-sPq and PqPs, considered as chords
represents
(5.)
of the
two successive and equal arcs,
each other,
it
and
are ultimately equal to
s,
follows that the parallelogram (2.)
is
them and
fore that its diagonals are ultimately rectangular; but these diagonals,
v., have ultimately the directions of p' and p"
VII.
=
Sp'p" 0,
if
which had been otherwise deduced
to
ultimately equilateral, and there-
by IV. and
we find therefore anew the
equation,
the arc be the independent variable,

before, in 380, (12.).
But under the same condition, we saw (379, (2.)) that the second derived
p" has the direction of the absolute normal to the curve such then is by V.
(6.)
vector
the ultimate direction of what

in (2.)
or, ultimately,
we may
call the short
diagonal PqRs, constructed as
the direction of the bisector of the (obtuse) angle p.PoP, be-
tween the two near and nearly equal chords from the point Pq
the parallelogram becomes ultimately the osculating plane.
while the plane of
(7.) All this is quite independent of the consideration of any surface, on which
But if we now conceive that this curve
the curve may be conceived to be traced.
is
(comp. Fig. 79), by wrapping round a develophad been drawn, and if the successive porof that line be supposed to have been equal, then because the two
formed /rom a right
line b'c'd'
able surface a plane on which the line

tions b'c', c'd',
c'd' originally made supplementary angles with any other line

c'c in the plane, the two chords c'b' and c'd' of the curve on the developable tend to
make supplementary angles with the generatrix c'c of that surface ; on which ac-
right lines c'b'
and
count the bisector (6.) of their angle b'c'd' tends tohe perpendicular to that generating line c'c, as well as to the chords' J)', or ultimately to the tangent to the curve at
The absolute normal (6.) to the curve
c', when chords and arcs diminish together.
is therefore perpendicular to two distinct tangents to the surface at
c',
consequently (comp. 372) the normal to that surface at that point whence,
the definition (380), the curve is, as before, a geodetic on the developable.
thus formed
and
by
is
(8.)
As regards
b'c'd' to the line c'c,
the asserted rectangularity (7.), of the bisector of the angle

when the angles cc'b' and cc'd' are supposed to be supple
mentary, but not in one plane, a simple proof
Compare the Note
may
to
be given by conceiving that the
page 152.
524
right line b'c'
is
two equally long
[bOOK
III.
prolonged to c", in such a manner that c'c"= c'd' ; for then these
lines from c' make equal angles with the line c'c, so that the one may
be formed from the other by a rotation round that line as an axis; whence c"d',
which
evidently parallel to the bisector of b'c'd', is also perpendicular to c'c.

In quaternions, if a and p be any two vectors, and if t be any scalar,
is
(9.)
have the equation,

VIII.
which
is,
by 308,
S.a(a<pa-'-p)
an expression
(8,),
we
0;
for the geometrical principle last stated.
384. The recent analysis (382) enables us to deduce with ease,

by quaternions, other known and important properties of developable surfaces: for instance, the property that each such surface may
be considered as the envelope of a series of planes^ involving only one
and arbitrary constant {or parameter)
scalar
and that each plane of
in their
common equation;
this series osculates to the cusp-edge of the de-
velopable.
(I.)
The equation
of the developable surface being

I.
its
normal v
is
easily found to
the scalar variable
equation
be, or
(a;,
y)
have as
I.,
whether
be not, the arc of the cusp- edge, of which curv'e the

II.
p=
by 373, VII., the equation

III.
\l/x.
of the tangent plane takes the form,
Sw;//'i//"
= St//i//'i//",
from which the second scalar variable y thus disappears

all
as in 382,
in 382, V., the direction of Yxj/'xp",
is,
(2.) Hence,
still,
+ y^'x =-<\^+ y4>\
=i//a:
this
common
equation, of
the tangent planes to the developable, involves therefore, as above stated, only
one variable and scalar parameter, namely x

the developable surface
and the envelope of
all these /j/anes is
itself.
(3.) The plane III., for any given value of this parameter x, that is, for a.i\y given
point of the cusp-edge, touches the surface along the whole extent of the generating
line,
which
is
the tangent to this last curve.
And by comparing
its equation III. with the formula 376, V., we see at

once that this plane osculates to the same cusp-edge, at the point of contact of that
curve with the same generatrix of the developable.
(4.)
385. If the reciprocals of the perpendiculars, let fall from a given

origin, on the tangent planes to a developable surface, be considered
as being themselves vectors from that origin, they terminate on a
curve, which is connected with the cusp-edge of the developable by
some interesting relations oi reciprocity (comp. 373, (21.)): in such
a manner that if this new curve be made the cusp-edge of a new developable, we can return from it to XhQ former surface, and to its cusp-
edge,
by a similar process of construction.
CHAP.
THEOREM OF RECIPROCAL CURVES IN
III.]
(1.) In general, if i^^
x ^^7
tions,
I.
which
briefly \p and x be two vector functions of a

be deduced from \p by the three scalar equa-
and Xx, or
scalar variable x, such that
in
525
SPACE.
..Sipx
= c,
&^'x =
81//% is written briefly for S(i//a!.x*)
o,
srx = o,
and
c is
any
scalar constant,
we have
then this reciprocal system of three such equations,
lI...Sx^ = c,
Sx'^ = 0,
Sx> = 0;
an intermediate step being the equation,

III.
(2.)
..Sfx' = Sx'f =0.
Hence, generally,
lV...ifx = |^fJ^!,
S'/^'f'y
But
then
V...;p='^^.
Sxxx
p be the variable vector of a curve in space, and p', p" its first and
(3.)
second derivatives with respect to any scalar variable, then, by the equation 376, V.
if
of the osculating plane to the curve,
VI.
so that if
i^
we have
the general expression,
= perpendicular from
and x he considered
origin on osculating plane;
as the vectors of two curves, each vector is c
reciprocal of the perpendicular, thus let fall from a
plane to the other.
common
point,
x the
on the osculating
(4.) We have therefore this Theorem :

If from any assumed point, o, there be drawn lines equal to the reciprocals oj
the perpendiculars from that point, on the osculating planes to a given curve of double curvature, or to those perpendiculars midtiplied by any given and constant scalar; then the locus of the extremities of the lines so drawn will be a second* curvCy
from which we can return to the first curve by a precisely similar process.
386. The theory of developable surfaces, considered as envelopes

of planes with one scalar and variahle parameter (384), may be additionally illustrated by connecting it with Taylor'' s Series , as follows.
(1.) Let at denote
I.
or,
by another
any vector function

a<
of a scalar variable
= flo + ttita'o = a + tua,
with
t,
so that
uq= 1\
step in the expansion,
II.
at
= ao +
â'o
^f^vta'a
= a + ta
-\-
where u and v are generally quaternions, but ua' and
vq
\tiva",
I;
va." are vectors.
* The two curves
may be said to be polar reciprocals, with respect to the (real or

imaginary) sphere, p^ = c, and an analogous relation of reciprocity exists generally
when the points of one curve are the poles of the osculating planes of the other, with
respect to any surface of the second order: corresponding tangents being then reciprocal polars.
Compare the theory oi developables reciprocal to curves, given in
Salmon's Analytical Geometry of Three Dimensions, page 89 see also Chapter XI.
(page 224, &c.), of the same excellent work.
;
526
[bOOK
III.
(2.) Then, as the rigorous equation of the variable plane, the reciprocal of the
- at, we have either,
perpendicular on which from the origin is
III.
1 = Satp = Sap -r
tSua'p,
or
IV.
we adopt
according as
- 1 = Sap +
the expression
I.,
iSa'p
+ ^t^Sva'p,
more rigorous ex-
or the equally but not
pression II., for the variable vector at.

(3.) Hence, by the form III., the line of intersection of the two planes, which
and t of the scalar variable, or parameter, t, is
answer to the two values
rigorously
represented
of the two scalar equations,
by the system
V.
And
(4.)
Sajo
Sua'p
0,
0.
the limiting position of this right line V., which answers to the con-
ceived indefinite approach of the second plane to ihQ first,
by
the equations,
VI.
whereof
it is
respect to
t,
Sap +
= 0,
Sa'p
given with equal rigouri
is
may be formed from ihQ first, by derivating with^

and treating p as constant: although no such rule of calculation had:
seen that the second
been previously laid down, for the comparatively geometrical process which
supposed to be adopted.
is
here]
(5.) The locus of all the lines VI. is evidently some ruled surface; to determine^
the normal v to which, at the extremity of the vector p, we may consider that vector to be a function (372) of two independent and scalar variables, whereof one is
<,j
and the other may be
moment w, and
called for the
thus
we
giving
shall
have the two'
partial derivatives,
VII.
Hence the
(6.)
Sap +
= 0,
SaD^p
a has the
line
SaDM,p
= 0,
||
a.
normal v
direction of the required
the plane
touches the surface (comp. 384, (3.)) along the whole extent of the /miting line VI. ; and the locus of all such lines is the envelope of all the planes, of]
the system recently considered.
1
The
(7.)
line
VI. cuts generally the plane IV., in a point which
is
rigorously de-
termined by the three equations,
VIII.
and the
mined by
this other
Sap +
0,
Sa'p
which
it
may
= 0,
St>a"p
with equal rigour, the point deter-
is,
system of equations,
IX.
in
limiting position of this intersection
Sap
+ 1 = 0,
be remarked (comp.
Sa'p
= 0,
Sa"p
(4.)), that the third
is
the derivative of the
p be treated as constant.
locus of all these points IX. is generally some curve upon the surface
(8.) The
which is the locus of the lines VI., and has been seen to be the envelope (6.) of
(.5.),
second,
if
the planes III. or IV. ; and to find the tangent to this curve, at the point answering
to a given value of t, or to a given line VI., we have by IX. the derived equations,
X. ..Sap'=0,
Sa'p'=0,
comparing which with the equations VI. we

which is thus their common envelope.
whence
p'||
Vaa';
see that the lines
VI. touch the curve^
CHAP.
(9.)
FAMILIES OF SURFACES.
III.]
527
"We see then, in a new way, that the envelope of the planes III., which have
(t) in their common equation^ and may represent any system of
one scalar parameter
is a developable surface : because it is in general

(comp. 382) the locus of the tangents to a curve in space, although this curve may
reduce itself to a pointy as we shall shortly see.
planes subject to this condition,
may add
(10.) "We
that
if
at in III. be considered as the vector of a given curve^
this curve is the locus
of the poles* of the tangent planes to the developable, taken

with respect to the unit sphere; and conversely, that the developable surface is the
envelope of the polar planes of the points of the same given curve, with respect to
the
same
sphere.
(11.) If then it happen that this given curve, with at for vector,
so that we have this new condition,
XI.
S/3a<
+1 =
0,
/3
is
a plane one,
being any constant vector,
namely the vector of the pole of the supposed plane of the given curve, the variable
plane III., or Spat+ 1 = 0, of which the surface (5.) is the envelope, passes constantly through this fxed pole ; so that the developable becomes in this case a cone,
with
(3
for the vector of its vertex
the equations IX. giving
now p =
f3.
(12.) The same degeneration, of a developable into a conical surface, may also
be conceived to take place in another way, by the cusp-edge (or at least some finite
portion thereof) tending to become indefinitely small^ while yet the direction of its
tangents does not tend to become constant.

developable which
written thus
XII.
For example, with recent notations, the

may have its equation
the locus of the tangeuta to the helix
is
= 0(w, y) = c(a;a+-2 tan a. a*U]3)+ya(l + tan a. a*U/3);
which when the quarter -interval, c, between the spires, tends to zero, without their
a being changed, tends to become a cone of revolution
round that axis, with its semiangle = a.
inclination a to the axis
387. So far, then, wê may be said to have considered, in the present Section, and in connexion -with geodetic lines, the four following
First,
families of surfaces (if the first of them may be so called).
spherical SU7faces, of
which the
characteristic joroêri^
is
expressed
by the equation,
I.
Yv'^p-a) =
0, if
a be vector of centre;
second, cylindrical surfaces, with the property,

II.
Si/a=
0, if
a be parallel
to the generating lines;
third, conical surfaces, with the property,

III.
Si^ (/>
a)
= 0,
if a
be vector of vertex
and fourth, developable surfaces, with the distinguishing property

expressed by the more general equation,
*
Compare the Note
to
page 525.
528
IV.
Vi^di/
dp have the direction of a generatrix
if
=0,
[bOOK
III.
V being in each the normal vector to the

surface, so that
V.
8*^(1/3
for all tangential directions of d/>;
0,
and the fourth family including the third, which in its turn includes
the second, A few additional remarks on these equations
may be
here made.
The geometrical signification of the equation I. (as regards the radii) is obbut on the side of calculation it may be useful to rexnark, that elimination of
(1.)
vious
V between
I.
and V.
gives, for spheres,
S(p a)d|0 = 0, or VII.

T(p - a) = const.
The equations II. and V. show that dp, and therefore Ap, may have
VI.
(2.)
given direction of a
for
an arbitrary cylinder, then,
(372),
VIIT.
where
we have
the
the vector equation
= <p{x,y) = xpa: +ya,
an arbitrary vector function of x.

at once infer, that
\px is
From VIII. we can
(3.)
lX...S(3p =
Syp=Sy;|/.,
Si3xP
the scalar equation (373) of a cylindrical surface
(comp. 371,
is
a=
if
V^y;
therefore generally of the /br/n
(6.), (7.)),
X.
(i
and y being two constant
0=F(Sj3p, Syp);
and the generating
vectors,
lines being perpendicular to
both.
The equation
(4.)
XI.
III. may
SvUa = Ta-Svp
;
be thus written,
whence
XII.
Sa/Ua
= 0,
if
Ta = <
the equation for cones includes therefore that for cylinders, as was to be expected,
and reduces itself thereto, when the vertex becomes infinitely distant.
III., when compared with V., shows that dp may have
and therefore that p a may be multiplied by any scalar the
The same equation
(5.)
the direction of p
a,
vector equation of a conical surface
XIII.
(6.)
The
a-'r y\px,
is
4'^
being an arbitrary vector function.
scalar equation of a cone
of a scalar variable
f,
therefore of the form,
may
be said to be the result of the eZmtnah'on
between two equations of the
XIV...S(p-a)x. = 0,
forms',
S(p-a)x'^ = 0,
which express that the cone is the envelope (comp. 386, (1 !)) ^^ ^ variable plane,
which passes through a. fixed point, and involves only one scalar parameter in its
with a new reduction to a cylinder, in a case on which we need not here
equation
:
delay.
(7.)
The equation IV.
tion along
which we
may
implies, that for each point of the s.'irface there is a direc-
move, without changing the tangent plane ; and therefore
that thfe surface is an envelope of planes, &c., as in 386,

developable, in the sense of Art. 382.
(8.)
The
the form,
and consequently that
it is
vector equation of a general developable surface inay be written under
CHAP.
ELIMINATION OF ARBITRARY FUNCTIONS.
III.]
XV.
529
= <p(x,t/)^^^ + i/-UrP':,;
the sign of a oersor being here introduced, for the sake of facilitating the
passage,
at a certain limitj to a cone (comp. 386, (12.)).
And
(9.)
the scalar equation of the
same nrhitrary developable may be repre-
sented as the result of the elimination of
XVI.
in -which xt is
Spx<
between the two equations,
+1=0,
Sf)x'<
an arbitrary vector function of t.
The envelope of a plane with two arbitrary and scalar parameters, t and
generally a curved but undevelopable surface, which may be represented by the
system of the three scalar equations^
(10.)
tt,
is
XVII.
where
Bpxu +
= 0,
SpDix =
0,
SpDx =
denotes the reciprocal of the perpendicular from the origin on the tangent plane to the surface, at what may be called the point (t, u),
388. It remains, on the plan lately stated (380), to consider

of revolution^ and to investigate the geodetic lines, on
briefly surfaces
which the equation, analogous

IV. in 387, for spheres, cylinders, cones,
of the form,
this additional /amz??/ of surfaces; of
to those
marked
I. II. III.
and developables,
is
I.
Sa/3i/
= 0,
if a be a given line in the direction of the axis of revolution,

supposed for simplicity to pass through the origin ; but which may also
be represented by either of these two other equations, not involving
the normal
v,
n.
where
so
T/>=/(Sa/)),
or
III.
i:Yap = F{âp),
F are
used as characteristics of two arbitrary but scabetween which Sap may be conceived to be eliminated,
a third form of the same sort obtained.
/ and
lar functions
and
(1.) In fact, the equation I. expresses that v \\\ a, p, or that the normal to the
surface intersects the axis ; while II. expresses that the distance from a. fixed point
is ^function of its own
projection on the %a.mQ fixed line, or that the
made by planes perpendicular to the axis are circles ; and the same circu-
upon that axis

sections
larity of these sections is otherwise expressed
by
III., since
that equation signifies
that the distance from the axis depends on the position of the cutting plane, and
constant or variable with

in calculation,
it
by means of the general

IV.
(2.)
The equation
tion of i\ie first order,
I. is
and
is
while the two last forms are connected with each other
relation (comp. 204,
(Tap)2
(Sap)2
XXI.),
+ (JYapy.
analogous, in quaternions, to a partial differential equatwo other equations, II. and III., is analogous
either of the
to the integral of that equation, in the usual differential calculus of scalars.
3 Y
530
To accomplish
(3.)
the corresponding integration in quaternions, or to pass from

III. can be deduced by IV., we may observe that the
whence
the form
I.
to II.,
equation
I.
allows us to
Y.
two
so that the
must
III.
[bOOK
scalars
Avrite
(because Svdjo
= xa-\-yp,
Sap and
VI.
= 0),
.
a:Sad|0
+ yS|odp = 0,
T/o are together constant,
or together variable,
and
therefore h^ functions of each other.
from the form
(4.) Conversely, to eliminate the arbitrary function
II,,
quater-
nion differentiation gives,
VII.
so that
VIII.
hence
= S (Up dp) +/' (Sap) Sadp = S (Up + a/'Sap) dp

v Up + af'Sap, and IX.
v
a, p, as before
.
[j
we can
return in this
way
to the equation
I.,
the functional sign
f disappear-
ing.
(5.)
We have
nions* analogous
thus the germs of a Calculus of Partial Differentials in Quaterto that employed by Monge, in his researches respecting ^miVies
of surfaces: but we cannot attempt

(6.)
But as regards the geodetic
to pursue the subject farther here.

lines
which gives at once the

X.
whence, by a
first
(because
S(adp.Udp) =
- SaTdp = 0)
integration, c being a scalar constant,
XI.
of
I.,
= SapdUdp = d. SapUdp
revolution, we have only

by 380, IV., the expression dUdp,
upon a surface of
to substitute for v, in the recent formula
= SapUdp = TVap.SU(Vap. dp).
(7.) The characteristic property of the sought curves is, therefore, that for each
them the perpendicular distance from the axis of revolution varies inversely as the
cosine f of the angle, at which the geodetic crosses a parallel, or circular section of
because, if Ta = 1, the line Yap has the length of the perpendicular let
from a point of the curve on the axis, and has the direction of a tangent to the
the surface
fall
parallel.
* The same remark was made in

page 574 of the Lectures, in which also was
given the elimination of the arbitrary function from an equation of the recent form
was also observed, in page 578, that geodetics furnish a very simple example
what may be called the Calculus of Variations in Quaternions ; since we may
III. It
of
write,
Sjds = ^JTdp = J ^Tdp = -JS(Udp. 5dp)
= - J S(Udp
and therefore dUdp
v, or
d^p)
= - AS (Udp
Vi-dUdp = 0,
^p)
+ J S (dUdp
dp),
as in 380, IV., in order that the expression
may vanish for all variations dp consistent with the

equation of the surface: while the evanescence of the part which is outside that sign
J supplies the equations of limits, or shows that the shortest line between two curves
under the
last integral sign
on a given surface
t Unless
the geodetic
is
is
perpendicidar to both, as usual.
happen that this cosine is constantly

a meridian of tlie surface.
it
zero, in
which case
0,
and
CHAP.
(8.)
OSCULATING CIRCLES TO CURVES IN SPACE.
III.]
The equation XI. may

XII.
and
is
also
be thus written,
= Sapju',
cTp'
531
where
= Dtp
p'
be supposed to denote the time, while the geodetic

conceived to be a curve described by a moving point, then while Tp' evidently reif
= ds
presents the linear velocity of that point, as being
denote the arc (comp.
d#, if s
'
100, (5.), and 380, (7.), (H.))? i' is easy to prove that Sapp represents the double
areal velocity, projected on a plane perpendicular to the axis; the one of these two
velocities varies therefore directly as the other
nics, that
each velocity would be constant,*
subject only to the
and
i(
in fact, it is
known from mecha-
the point were to describe the curve^
normal reaction of the surface, and undisturbed by any other
force.
As regards the
analysis, it is to be observed that the differential equation

not only by the geodetics on the surface of revolution, but also by the
which fact of calculation is connected with the obvious
parallels on that surface
geometrical property, that every normal plane to such a parallel contains the axis of
(9.)
X.
is satisfied,
revolution.
(10.) In fact if we draw the normal plane to any curve on the surface, at a point
it crosses a parallel, this plane will intersect the axis, in the point where that
where
axis
is
met by the normal
this construction _/ai7s to
to the surface,
touch a parallel, at the point where
Section
drawn
determine that normal,

its
at the
if
same point of crossing
but
the curve coincide with, or even
normal plane
is
drawn.
On
Osculating Circles and Spheres, to Curves

Space; with some connected Constructions.
6.
in
389. Resuming the expression 376, I. for pt, and referring again
we see that if a circle pq,d be described, so as to touch a
to Fig. 77,
given curve pqr, or its tangent pt, at a given point p, and to cut the
curve at a near point Q, and if pn be the projection of the chord pq
on the diameter pd, or on the radius cp, then because we have, rigorously,
I.
we have
PQ =
</)'
^fup^',
with
u=l
for
0,
also
II.
.PN = J^Vw/>V:/>',
and
PC
_ 2 _ 2pn
PD
PQ^
Yup'^p'
(^pf
îtu/ypf'
Conceiving then that the near point q approaches indefinitely to the

given point p, in which case the ultimate state or limiting position of
* This remark
rential
is virtually made in page 443 of Professor De

Morgan's Diffeand Integral Calculus (London, 1842), which was alluded to in page 678
of the Lectures
on Quaternions.
532
the
circle
pqd
that point
is
we
p,
[bOOK
III
said to be that of the osculating circle to the curve at

see that while the plane of this last circle is th'e os-
plane (376), the vector k of its centre k, or of the limiting poof the point c, is rigorously expressed by the formula
Ctilating
sition
IV.
/c
/>
which may however be in many ways transformed, by the rules of

the present Calculus.
we may
(1.) Thus,
write, as transformations of the expression IV., the follow-
ing:
p'
Tp'
Tp'
or introducing differentials instead of derivatives, but leaving
still
the independent
variable arbitrary,
VI.
dp
_ +
dp3
_ _
_ Tdp _ _ d*
dUp'"^ dUd^'
vd2pdp-i~^
ic-p-^-^^^2p-P
be the arc of the curve so that the last expression gives this very simple formula, for the reciprocal of the radius of curvature, or for the ultimate value of
if s
CP,
VII.
(p
k)-* = DjiUp',
where
Up'
= Udp,
as before.
To interpret this result, we may employ again that auxiliary and spherical
upon the cone of parallels to tangents, which has already served us to conin 379, (6.) and (7.), the osculating plane, the absolute normal, and the bi-
(2.)
curve,
struct,
normal, to the given curve in space.
PC has ultimately the
And
direction of dUp',
thus
we
while the semidiameter
see, that
and therefore that
of the absolute
normal
(379, II.) at p, the length of the same radius is ultimate!}' equal to the arc pq (or
As) of the given curve, divided by the corresponding arc of the auxiliary curve ; or
that the radius of curvature, or radius of the osculating circle at p, is equal to the
ultimate quotient of the arc pq, divided by the angle between the tangents, pt and
at p, and to its prolongation qr at Q, although these
(say) QU, to that arc pq itself
two tangents are generally in different planes, and have no common point, so long
as PQ remains ^wi7e; because we suppose that the given curve is in general one of
double curvature, although i\iQ formula, and the construction, above given, are applicable to
plane curves
also.
formula IV. gives,

(3.) For the helix, the
and
VIII.
K^cta-
a*(3 cot2 a,
whence
p, p', p",
osculating circle
to the axis
- k = a^(3 cosec^ a,
but otherwise on a co-axal cylinder, of which the radius
T/3, multiplied
by the square
= T(p - k) = T/3
linder,
is
IX.
the locus of the centre of the

therefore in this case a second helix, on the same cylinder, if
a being the inclination of the given helix
a=
by values already assigned for
a, the expression,
and
is
X cosec^
a, so
of the cotangent of o;
the given radius
and the radius of curvature
that this radius always exceeds the radius of the cy-
cut perpendicularly (without being prolonged)
by the
axis.
CHAP.
VECTOR OF CURVATURE, EXAMPLES.
III.]
= const.,
(4.) In general, if Tp'
$33
(2.)), the
and therefore S|o'p"=0 (comp. 379,
expression IV. becomes,*
X.
= p + ^,
XI.
whence,
= p - p'''^,
if
Tp'
= 1,
P
that
Under
the arc be taken as the independent variable (380, (12.)).

formula VII. reduces itself to the following,
if
is,
this
last condition, then, the
XII.
(p
k)-' = p" = j)s^p = ultimate
so that p" (for
Tp"
and
is
Tp'= 1)
a numerical measure for
versor
its
Up"
reciprocal of radius cp;
be called the Vector of Curvature, because
may
what
its
tensor
usually called the curvature^ at the point P,

represents the ultimate direction of the semidiameter pc, of the
is
circle constructed as above.
(5.)
As an example
of the application (2.) of the formula IV. for k, to
a.
plane
curve, let us take the ellipse,
XIII.
Ta =
= Ya%
Sa/3
1,
337, (2.),
^ 0,
The expressions
considered as an oblique section (314, (4.)) of a right cylinder.
376, (5.) for the derivatives of p, combined with the expression XIII. for that vector itself, give here the relations,
XIV.
and therefore comp. (338,
XV.
Ypp' = const. $
0,
Vp'p"'
Yp'p" = const.
^ (3y,
=
-^
Ypp
Va</3
interpret this result,
we may
write
XVII.
= 0;
K = p^- ^, ^]
where
['^Y f3y,
if
= Ya(3
IV.,
= p-
XVI.
To
Vpp" =
we have by
hence for the present curve
(6.)
(5.)),
it
(Va<+i/3)3 (/3y)-'.
as follows,
XVIII.
Ypp.p"^
pi
= - p'= Va<+i/3;
TT
SO that pi is the conjugate semidiameter of the ellipse (comp. 369, (4.)j, and Vpp':p'
recover then,
is the perpendicular from the centre of that curve on the tangent.
We
is
known
radius of curvature of an ellipse

equal to the square of the conjugate semidiameter, divided by the perpendicular.
may also write the equation XVI. under the form,
(7.)
by
this simple analysis, the
result, that the
We
XIX.
= p Pi'
where
XX.
Vppi
* The
expressions X. XI.
may
also be easily
Yppi = /3y =
deduced by
const.
limits,
from the con-
struction in 383, (2.).
f It may be remarked that the quantity z, or Tv//", in the investigation (382)

respecting êorfefic* on a deoelopable, represents thus the curvature of the cusp-edge,
for any proposed value of the arc, x, of that curve.
X These values XV. might have been obtained unthout integrations, but this
seemed to be the readiest way.
534
and may
interpret
it
III.
[bOOK
as expressing, that the radius of curvature
is
equal to the cube
by the constant parallelogram under any two

such conjugates or by the rectangle under the major and minor semiaxes, which
are here the vectors /3 and y (comp. 314, (2,)).
of the conjugate semidiameter, divided
;
XVI. or XIX. for k is easily seen to vanish, as it ought to

(8.) The expression
do, at the limit where the ellipse becomes a circle, by the cylinder being cut perpen= being satisfied ; and accordingly if we write,
dicularly, or by the condition Sa/3
XXI.
we
= SUa/3 = excentricity
or
of ellipse,
XXII.
y2
= (1 _ ^2)^2^
easily find the expressions,
XXIII.
= (5S.at+yS.at-i,
XXIV...p,^ = /3ni-HS.aO'^),
so that the formula
XXV.
thus containing
(9.)
And
XIX.
^=|i
/3-[/3S.a^+^];
becomes,
= e^l(3{8.aty e^ as
= -/3S.a'-i + yS.a;
pi
Zi|^^''\
= e2 (^(s.aO^ -
-(S.a'-i)'),
factor.
may be remarked in passing, that the expression XVI., or its recent

XXV., for k as a function of t, may be considered as being in qua-
it
transformation
ternions the vector equation (comp. 99,
I.,
or 369, I.) of the evolute* of the ellipse,
of centres of curvature of that plane curve and that the

last form gives, by elimination of ^ (comp.f 315, (1.), and 371, (5.)), the following
system of two scalar equations for the same evolute,
or the equation of the locus
XXVI...(s|]n(sp]^
XXVI'.
or
which
= .^
S/3y.
(S^k/ + (Syfc)^ = (e/3)^,

known results.
= 0;
&c.
will be found to agree with
(10.)
As another example
of application to a plane curve,
hyperbola,
XXVII.
with a and
(3
for asymptotes,
rived vectors are,
XXVIII.
XXIX.
whence
and with
p'
p
its
= ta + t-%
we may
consider the
comp. 371,
centre at the origin.
II.,
In this case the de-
= a - ^2^,
Vp"p'
p" = 2t-^f3,
= 2^3 Vj3a = r^Vpp',
and the formula IV. becomes,
XXX. ..K-p =
,=
Ypp :p
the perpendicular from the centre
ov
o on the tangent to the curve at r,

same point p and an
where ov
is
and PT
the portion of that tangent, intercepted between the
is
asymptote (comp. (6.) and 371,
That
is
(3.)).
to say, of the plane evolute; for
we
shall soon
have occasion
to consi-
der briefly those evolutes of double curvature, which have been shown by Monge to
exist, even when the given curve is plane.
t In lately referring (373, (1.)) to the formula 315, V., that formula was inadvertently printed as (a*)-
(a*-^)2=
1,
the sign S. before each power being omitted.
CHAP.
ABRIDGED GENERAL CALCULATION.
III.]
We may
(11.)
also interpret the
tion of the semidiameter
op on
XXX.
denominator in
we have
of the osculating circle,
XXXI.
whereof the
then
is
the foot of
be the sought centre
the geometrical equations,
NP PK = PT2,
if
as denoting the projec-
np where N
the normal, or as the line
the perpendicular from the curve on that normal line
535
XXXII.
Z NTK =
an extremely simple construction for the centre

shall soon find to admit of being extended,
last furnishes evidently
of curvatiire of an hyperbola, which we

with
and
modification, to a spherical conic*
little
The logarithmic
(12.)
spiral vîih
XXXIII.
= a%
its cyclic
arcs.
at the origin,
its pole
Sa(3
0,
Ta
comp. 314, (5.)
^ 1,
be taken as a third example of a plane curve, for the application of the foregoing
may
formulae.
XXXIV.
first
derivation gives,
by 333, VII.,
= (c + y)jO = p(c -
jo'
y), jo'p-i
the constant quaternion quotient, p'
makes with the tangent pt a constant
XXXV.
tan
= (T V
p,
= c + y,
= ITa, and
if
= - Ua;
here showing that the prolonged vector
angle, n,
S) (p' :p)
which
= c- iTy,
is
OP
given by the formula,
or
cot
= - ITa
and a second derivation gives next,
XXXVI.
p"
The formula IV. becomes
XXXVII.
(C
Vp"p'
y)2 p,
= (C2 - y) p2y =
/c
= p+p'y-i = pcy-i = -cy-ip=?^.a<+*/?;

TT 1
the evolute
is
p'2y.
therefore, in this case,
therefore a second spiral, of the
same kind as the
first,
and the radius
of curvature kp subtends a right angle at the common pole. But we cannot longer
here delay on applications within the plane, and must resume the treatment by quaternions of curves of double curvature.
390.
When
the loîc by which the expression 389, IV. was ob-
tained, for the vector k of the centre of the osculating circle, has
once been fully understood, the process may be conveniently and safely
abridged, as follows. Referring still to Fig. 77, we may write briefly,
* It was in fact for the
spherical curve that the geometrical construction alluded
was first perceived by the writer, soon after the invention of the quaternions, and
but it has been thought that a sub-artias a consequence of calculation with them
to
cle or
may
t
two might be devoted,
which
serve at least as a verification.

If r be radius vector,
= 1,
T/3
cot
as above, to the plane case, or hyperbolic limit,
and Q polar angle, and
if
we
suppose for simplicity that
the ordinary polar equation of the spiral becomes r = a^, with a
= la,
as usual.
= Taw,
and
536
which are
as equations
sense which
I.
PT = d/),
by 203,
&c.
or true at the
all ultimcctely true,
supposed to be now distinctly seen
is
[bOOK
in a
limits
PN = (part of pq _L pt =)
TQ = JdV,
III.
whence, ultimately,
_ ^^ _
as before: this last expression,
in
-^^^
which
^f^^
denotes briefly
Vd^dp
being rigorous, and permitting the choice oi any scalar,

And then, by writing,
to be used as the independent variable.
V(d^/3.d/>),
III.
the factor
dt^
d/>
= p'd^t,
dH = 0,
we
disappears, and
dV = p^^dt\
pass at once to the expression,
which had been otherwise found before.
When
(1.)
the arc of the curve
380, (12.), &c.) the expresssion
V.
is
taken for the independent variable, then (comp.
reduces itself to the following,
II.
dp2
(c-p=-
and accordingly the angle ptq in Fig. 77
we may
so that
"^
we have thus
(2.) Suppose then that
in
then ultimately right (comp. 383, (5.)),
Kp = (ult.) PC = Cult.)
^
the expression V. for k
II.,
is
Sd2pdp=0;
at once write, with this choice of the scalar variable,
VI.
us consider
because
p,
how we might
which that variable
pt2
2tq
dp^
-r-, as above.
d-^p
geometrically (and very simply) deduced
for this particular choice of the scalar variable
and
let
thence pass, in calculation, to the more general formula

is left
arbitrary.
For
this purpose,
we may
write,
by
principles already stated,
vn
V
(n
<.P
^'P
.L-v^
;
^^^^^2
Tdp
^P
^dUdp^
Tdp
Tdp
Vd3pdp-i.Udp
Tdp
Vd2pdp-i
_ Vdpd2p
dp
and the required transformation
accomplished.
if s denote the arc of any curve of which
p
(3.) And generally,
vector, we may establish the symbolical equations,
VIII.
dp^
is
D,=^-Ld;
is
the variable
*'
i-.^=t-5^<'tT/=(53^^ T'
For example (comp. 389, XII.), the Vector of Curvature, D^'p, admits of
generally under any one of the five last forms VII.
expressed
being
(4.)
CHAP.
POLAR
III.]
AXIS,
POLAR DEVELOPABLE.
537
391. Instead of determining the vector k of the centre of the osculating circle by one vector expression^ such as 389, IV., or any of
its
transformations,
we may determine
it
by
a system of three scalar
equation^^ such as the following,

I.
S(ic-p)p' =
III.
II.
S{k-p)p'^-p'^ = 0;
..S{k-p)p'p'' = 0,
.
0',
of which
it
first, if
be treated as constant (comp. 386,
a:
be observed that the second
may
is
the derivative of the
(4.))
and of which the
expresses (369, lY.) that the sought centre is in the normal

plane to the curve, while the third expresses (376, V.) that it is in
first
the osculating
absolute
(1.)
plane ; and the second serves to fix
Using
differentials instead of derivatives,
variable arbitrary,
we may
.S(K-p)dp = 0;
IV..
of which the second

(2.) It
is
is
the differential of the
VIII.
For instance, we
first, if
/c
the independent
VI. .. S(fc-p)dpd2p= 0;
be again treated as constant
.d2T(K-p) = 0;
may
and which
I.
and
is
II.,
IX.
dUV(K - p)dp =
interpretations.
say that the centre of the osculating circle is the point,

VI. or IX., is intersected by the aocis of that
III. or
namely, by the right line which
to its plane
still
and 380, (2.)), with the

to write these equations under the forms,
k,
which the osculating plane,
drcle
S(K-p)d2p-dp2 = 0;
them with geometrical
to connect
(3.)
in
but leaving
also permitted (comp. 369, (2.), 376, (3.),
..dT(K-p) = 0;
VII.
position on the
intersect.
establish this equivalent system of three equations,
V.
same supposition respecting
and
its
normal (379), in which those two planes
or
is
by
represented
drawn through
centre, at right angles
its
the two scalar equations,
IV. and V.,
or
VII. and VIII.
And we may
observe (comp. 370, (8.)), that whereas for a point r taken

arbitrarily in the normal plane to a curve at a given point p, we can only say in general, that if a chord pq be called small of ih.Q first order, then the difference of dis(4.)
tances,
RQ IIP,
is
small of an order higher than the first
yet, if the point
B be
taken on the axis (3.) of the osculating circle, then this difference of distances is
small, of an order higher than the second, in virtue of the equations VII. and VIII.
(5.)
The
right line
which may be
through
it,
is
I.
IL, or IV. V., or VII. VIII., as being the locus of points
called poles of the osculating circle,

also called the
on
Polar Axis of the curve
given /joini of osculation.

(6.) And because the equation II.
is
all possible
itself,
spheres passing
corresponding to the
(as above remarked) the derivative of
I.,
the
known theorem
follows (comp. 386), that the locus of all such polar axes is a developable surface, namely that which is called the Polar Developable, or the envelope
of the normal planes
to the given curve;
of tvhich surface
casion to consider briefly the cusp-edge.
3 z
we
shall soon
have oc-
538
[bOOK
III.
The following is an entirely different method of investigatby quaternions, not merely the radius or the centre of the oscu-
392.
ing,
but the
lating circle to a curve in space,
and in a way which
itself:
is
vector equation of that circle

applicable alike, to plane q.\xv\qs^ and to
curves of double curvature.

ot = r is a given tangent
moment taken as the origin and let
(1.) In general, conceive that
point which
the
is for
riable tangent^
drawn
at the
to a circle, at
extremity of the variable chord
the intersection^ gt'Pp', of these two tangents.
combined with the formula 324, XI.
a given
= p' represent a vaOP = p also let u be
pp'
Tlieu the isosceles triangle gup,
for the differential of a reciprocal, gives easily
the equations,
I.
p'
if
pr-ip
II
III.
II.
=-
Vrp-ip'p-i
(Vrp-i)'
as in 296, IX.",
= const. = Vra-i,
Vrp-i
a be the vector oa of any second given point A of the circumference.

(2.) The vector equation of the circle pqd (389) is therefore,
IV.
= V -^ =
V -?^
whence,
- V.
passing to the limit (<
ciccle is at
once found to
(1
Qt-p
= 0, m=
= - V.p"p'-i (1 +
htup"p'-\)-^
1), the analogous equation of the osculating
be,
v...vi!^.-v?;.
<o p
,'
,i..t(
\^~ P ?,>"
.,
with the verification (comp. 296,

VII.
the vector k of the centre
is
(9.)), that
a>
when we
- p = 2 (fc - p)
a.
suppose,
p',
seen to satisfy the equation,
VIIL..-^=-vC
p
K-p
which agrees with recent
i^p"p'-0-^
IX...-i^ + V^^=0;
K-p
dp^,
or
results (389, IV., &c.).
so as to touch a given
(8.) Instead of conceiving that a circle is described (389),
curve (Fig. 77) at p, and to cut it at one near point q, we may conceive that a circle
cuts the curve in the given point p,
and also
in two near points,
Q and
R,
uncon-
nected by any given law, but both tending together to coincidence with p and may
inquire what is the limiting position (if any) of the circle pqk, which thus intersects
:
the curve in three near points, whereof one (p) is given.

if a, /3, p be three co-initial chords, OA, ob, op, of any one
(4.) In general,
cle, their
reciprocals a"', /3"S p'l,
plying which principle,

tîf!.!
we
if still co-initial,
are termino-collinear (260)
cir-
ap-
are led to investigate the condition for the three co-ini-
vectors
X.
with Mo
1,
(w
p)-i,
(sp'
+ |s2,p")-i,
4-
(<p'
thus ultimately terminating on one right line
itûtp"y\
;
mately a relation of the form,
xs-\-yt
w-p
p'
p'^W
or for our having ulti-
CHAP.
III.]
VECTOR EQUATION OF OSCULATING CIRCLE.

xiL
C**+yOp'
1 +
1 + isp'p'-^
cj-p
= x + ^(xs + ?/t)p"p'-^ + &c.
7/
539
itp"p'-i
:
both members are generally quaternions.

of the scalar parts gives here no useful information, on account of the arbitrary/ character of the coefficients x and y ; but these disappear, with
the two other scalars, s and t, in the comparison of the vector parts, whence follows
in
which
last equation,
The comparison
(5.)
the determinate
and
limiting equation,
XIIL
2Yp'(<o-py^ = -Yp"p-\
which evidently agrees with V.

(6.) It is then found, by this little quaternion calculation, as was of course to be
expected,* that the circle (3.), through any three near points of a curve in space,
coincides ultimately with the osculating circle, if the latter be still defined (389) with
reference to a given tangent, and a near point, which tends to coincide with the given
point of contact.
393. An osculating circle to a curve oi double curvature does

not generally meet that curve again; but it intersects generally a
plane curve, of the degree n, to which it osculates, in 2w 3 points,
distinct
real,
from the point p of osculation, whereof one at least must be

it may happen to coincide with that point P
and
although
such a circle intersects also generally a spherical curve of double

curvature, and of the degree n, in w 3 other points, namely in
those where the osculating pâne to the curve meets it again.
An
example of each of these two last cases, as treated by quaternions,
may be
useful.
(1.) In general, if
we
becomes,
I.
clear the recent equation, 892,
2p'2Vp' (w-p)
V. or XIII., of
fractions, it
+ {(o-pJYp'p-,
which p = op = the vector of the given point of osculation, and p', p" are its first
and second derivatives, taken with respect to any scalar variable t, and for the par-
in
ticular value
point p
(whether zero or not) of that variable, which answers to ihQ particular

w denotes generally the vector of any point upon the circle, which
while
osculates to the given curve at that point p.

(2.)
Writmg then (comp.

II.
and
389, (10.)),
= â +
III.
to express that
we
ri/3,
.
ft)
p=a- r2/3,
= OQ = aa +
p"
are seeking for the remaining intersection
* This conclusion
is
and
2r3/3,
of a plane hyperbola
indeed so well known, and follows so obviously from the docdeduced here as a verification of previous for-
trine of infinitesimals, that it is only
mulae,
a;"'j3,
for the sake o^ practice in the present Calculus.
540
III.
[bOOK
with its osculating circle at p, the equation I. becomes, after a few easy reductions,
including a division by Va/3, the following biquadratic in x,
IV.
in
which the cubic factor
= (ar - ty
(tâ^x
- (3^)
to be set aside, as answering only to the point
is
(3.) Substituting then, in III., the remaining value IV. of
a;,
we
itself.
find the ex-
presssion,
OQ
'
^^
^.,^
comparing which with 371,
we
(3.),
2ta
2^2ri/3
J'
see that if the tangent to the hyperbola at the
given point p intersects the asymptotes in the points A, B, then the tangent at the
sought point q meets the same lines oa, ob in points a', b', such that
VI.
whence Q
A still
(4.)
known
results,
VII.
whence
two
OA
oa'
OB ob' = 0A2
= Ob2,
at once found, as the bisecting point of the line a'b'.
is
more simple
may
.
construction, and one more obviously agreeing with

be derived from the following expression for the chord pq
:
PQ = w
-p=
- r2a-2) (ta^(3 - t'â(3^)

(<2/3-2
= (<3/3-2 - t-â-^)ap'ft
II
ap-'^lS
follows (comp. 226) that if this chord pq, both ways prolonged, meets the
asymptotes ob and OA in the points r and s, we have then the inverse similitude
it
of triangles (118),
VIII.
A EOS
Ce'
AOB.
As
regards the cqijality of the intercepts, rp and QS, it can be verified

without specifying the second point Q on the hyperbola, or the second scalar, x, by
observing tiiat the formula III., combined with the first equation II., conducts to
(5.)
the expressions,
IX.
which
. .
OR = ^ _
= (a;-i + n) |8,
give, generally,
X.
os
RP = QS = fa -
= (a; + tâ
a;-i/3.
And
as regards the general reduction, of the determination of the osculating

circle to a spherical curve of double curvature, to the determination of the oscu-
(6.)
lating plane,
it is suflScient
when we
to observe that
take the centre of the sphere for
the origin, and therefore write (comp. 381, XIV.),
XI.
then
if
p'3,
we
p2
Spp'
const.,
operate on the vector equation
= 0,
I.
Spp" = -
p'2,
with the symbol V.
p,
and divide by
there results the scalar equation,
XII.
which expresses that the

the curve
= 2Sp(w-p) + (w-p)2 =
circle is entirely
while the other scalar equation,
XIII
obtained
a>2_p2
contained on the same spheric* surface as
by operating on
I.
Sp"p'(w
- p),
with S p", expresses (comp. 376, V.) that the same

.
* This conclusion is
geometrically evident, but
sake of practice in the quaternions.
is
here drawn as above, for the
541
CASE OF A SPHERICAL CURVE.
CHAP. Ill]
ofthe perpendiplane :* so that its centre K is the foot

on that plane from the origin, and we may therefore write (comp.
circle is in the osculating
cular
fall
let
385, VI.),
XIV.
and with the
oK=K = 5^^,
verification that the expression
389, IV., because
when
with the relations,
XIV.
XV.
S-=S^=1;
agrees with the general formula,
XVI. ..pVp"p'+|o'3 = S|oVP.
the conditions XI. are satisfied.
(7.)
And
even
the given curve be not a spherical one, yet
if
general expression for
XVII.
and operate on
I.
/c
with S.p" and S.p"p',
culating plane, combined with a
new
=p+
we
we
XVIII.
-^
Vp p
retain the
389, IV.,
find again the equation
scalar equation,
tions be written thus,

.. (a)-fc)2
which
may
XIII. of the os-
after
a few reduc-
= (p-fc)2;
and which represents a new sphere, whereon the osculating

great
if
k,
circle to the
curve
is
circle.
394.
To
now an
give
example of a spherical curve of double cur-
vature, with its osculating circle and plane for any proposed point p,
and with a determination of the point Q in which these meet the
curve again (393),

conic, of
we may
consider that spherical conic, or sphero-
which the equations are (comp. 357,

I.
= 0,
./9'-l-r2
11.),
II. ..^/>HSX/>/i/>
= 0;
namely the intersection of the sphere, which has its centre at the
of the second order, which has
origin, and its radius =r, with a cone
the same origin for vertex, and has the given lines X and /t for its
two (real) cyclic normals. And thus we shall be led to some suffias their plane
ciently simple spherical constructions, which include,
limits,
the
the analogous constructions recently assigned for the case of
common
hyperbola.
= 2S\pS/xp p^SXjit (comp. 857, 11'.), the equations I. and II.

(1.) Since SXp//p
allow us to write, as their first derivatives, or at least as equations consistent therewith,
III.
Spp'
0,
sxp'
+ sxp =
0,
Sup'
here arbitrary, so that
S/ip
Compare
= o,
we may
because the independent variable

derived vector p' to be multiplied by any convenient scalar
is
conceive the
first
in fact, it is only the
542
which
direction of this tangential vector p'
and
tinue the derivations consistently,
equations,
IV.
Spp" +
p'2
is
here important, although we must conwrite, as consequences of III., the
must
so
S\p" + S\p' = 0,
0,
III.
[bOOK
S/ip"
- S/Jip = 0.
(2.) Introducing then the auxiliary vectors,
V. ..j;=VX/t,
ff
= \Sfip+fiSXp,
= p + p',
= p p',
whence
= 87/0- =SXr = S/it;,
VI. ..
and by new
VII.
= 2SXpSfxp,
Sp(r
= 2S/ip,
SXi;
= 2S\p,
derivations,
cr'
= V?7p,
t'
= p+p",
S\r' = Sjuu' =
v'^p'-p",
S\i;'
we
S/ir
0,
S/ir'
= S/iT,
= -S\u,
r and v are the vectors ox and on of the points in which the
see first that
recti-
p meets the two cyclic planes, perpendicular respecand because the radius op is seen to be the perpendicular bisector
tively to \ and /x
of the linear intercept TU between those two planes, so that
linear tangent to the curve at
;
VIII.
p'
= PT = UP
if
-i-
X.
or
we have
OP,
n AP
IX.
uop = pot,
= n PB,
the tangent arc on the sphere, to the same conic at the same point p, meet the two
CA and CB in the points A and b : the intercepted arc ab being thus bi-
cyclic arcs
sected at
its
point of contact p, which
is
a well-known property of such a curve.
Another known property of a sphero-conic is, that for any one such curve
the sum of the two spherical angles cab and ABC, and therefore also the area of the
(3.)
We
is constant.
can only here remark, in passing, that
quaternions recognise this property, under the form (comp. II.),
spherical triangle ABC,
XI.
(4.)
The
XII.
or
by
cos (A
+ b) = - SUXpup = - 9 TX/x = const.

'
scalar equations III.
p.
^
and IV. give immediately the vector expressions,

XIII.
Îp(^jif5^,
S\/zp
'
^
^
p=p-
(lfl)IV,
SX/xp
(2.),
XIV. ..p' =
^^
and
XV. ..p" = p-?,
if
XVI.
?= ^^
= r r' = V +
the
new
plane to the conic at p meets the line

we have the geometrical expressions,
XVII.
p" = xp,
and the lines* r and v are the
v',
auxiliary vector ^ being thus that of the point x, in which the osculating
We may
r'
ij
of intersection of the cyclic planes
= XT,
t;'
= xu,
if
and
v',
so that
= ox,
traces of the osculating plane on those
also consider the derived vectors r'
or the lines
two
cyclic
xx and xu,
as corresponding tangents, at the points T and u (2.), to the two sections, made by
the cyclic planes, of that developable surface which is the locus of the tangents tpu
to the spherical conic in question.
CHAP.
III.]
CONSTRUCTION FOR A SPHERICAL CONIC.

on the former
planes, or of the latter
p and
tively to
while each 4-
p,
?;,
while a and
543
as being perpendicular respec-
c',
are the traces on the plane
Xyt*
of the
two cyclic
normals, of the normal plane to the conic at the point p, and of the tangent
or at least these lines have the directions of those
plane to the sphere at that point
:
traces.
(5.) Already,
from the expression XVI. for the portion ox of the radius oc (2.),
which is cut off by the osculating plane at P, we can
or of that radius prolonged,
derive a simple construction for the position of the spherical centre, or pole, say E,
of the fimaZZ ciVc/e which osculates at that point p, to the proposed sphero-conic.
For if we take the radius r for unity, we have the trigonometric expressions,
XVIII.
sec
CE
cos
EP = (T|
= Tr^
SU7?-ip =) sec^ pb sec cp
Fig. 80.
or letting fall (comp. Fig. 80) the perpendicular
XIX.
DE = cos DP
cos
XX.
or finally,
cos
PB
dbe
COS
(or
PB
cd on
COS
dae) =
PE
the normal arc pe,
= COS DB COS BE
(6.) But although it is a perfectly legitimate process to mix thus spherical trigonometry with quaternions (since in fact the latter include the former), yet it may
be satisfactor}' to deduce this last result by a morepwre/yquaternionic method, which
can easily be done as follows.
XXI.
in
The values
(4.) of p and p" give,
= pSo'p"-(rSpp"=pSp(7+p'2<T
Vp'p"S;7p
= (r - p') S(rr + crSpV = rS(Tr + Yrp'a
r,
Yrpa^
denotes a vector J- p (because SpV = 0), and

?, p' (because
this line p'c has therefore the direction of the projection of the line
which
= o)
pa
and we are thus led
S);p'p'(r
??
on a
through the line oc of intersection of the cyclic planes, a, plane cod perpendicular to the normal plane to the conic
at p, or to let fall (as in Fig. 80) a perpendicular arc cd on the normal arc pd; after
plane perpendicular to
which the normal

circle sought, as
to draw,
or the axis oe of the osculating

Yp'p", will be contained in the plane through the trace r, or
perpendicular to the plane of r and p'<r, or to the plane dob ;
to the sought osculating plane,
being
OT, or ob, which
p',
is
||
and therefore the spherical angle dbe (or dae)

(7.)
dered in
We may
its
by XXI.,
also observe that
own plane,
if
will be a right angle, as before.
be the centre of the osculating
circle, consi-
or the /oo< of the perpendicular on that plane from O, then
544
and
[bOOK
therefore
XXIII.
OK
- PC
i--
which gives again the angular
XXIV.
r2
XX.
relation
tan ep
= siu^ pb
XX V.
= PT of = sin POT = sin PB,
<r -J-
but
pV,
The rectangularity
(8.)
equation,
XXVI.
which we
in proving
two
or
pV,
p,
but
cr,
p,
is also
being ob-
cot pd,
<t.
expressed by the
employ the values,
= 1,
= (- T-y^Qrjp =)
Sp'cK-i
also interpret these equations
may
(V: S)p(r=Up'.
-1-
p'cr
of the planes of r, k and r, p'a
Src-i
XXIV.
= S (Ykt VpVr) = S/crSpVr - r^Sp'ffK
may
XXVII.
(9.) "We
and
Tp'r-i
because
cot pd,
the quotient XXIII. being thus a vec-
ought by 393, XV. to be and the trigonometric formula

tained from its expression, by observing that
tor, as it
the
III.
XXVIL,
SpVr-i.
as expressing the system of
relations,
XXVIII.
from which
follows that
it
drawn through
And
(10.)
r,
- r-i
c-
J-
r,
- r-i -L pV
as to be perpendicular to the plane through r
the two relations
XXVIIL
pression,
XXIX.
We may
K-
also that k, is a line in the plane so
and therefore
k"',
/c-i
and
p'cr,
as before.
are both included in the following ex-
r-
= Vr-ipV
Sp<T.
from the foregoing spherical construction, the

= EP of the osculating
following trigonometric expressions, for the arcual radius r
small circle (5.), and for the angle a = pae = ebp which it subtends at A or at B
(11.)
also easily deduce,
XXX.
A and B
tan r
= sin -
tan a
XXXI.
tan a
= ^ (cot A +
here denoting, as in XI., the base angles of the triangle
and
cot b)
abc with c
for ver-
denoting as usual the base ab, namely the portion of the arcual tangent
(2.) to the conic, which is intercepted between the cyclic arcs.
(12.) The osculating /)/ane and ci7'cle at p being thus fully and in various ways
tex,
determined,
again.
we may next
inquire (393) in
In symbols, denoting by
lar equations,
XXXII.
which are
also
all
Sfcw
= Skp,
vector value,
We satisfy
which
XXXIII.
which x
is
S\wS/iw
any
it is
= SXpS/xp,
w =p,
scalar
in fact
w=
we have
w2 = p2,
but can in general be
satisfied
the object of the problem to assign.
the two first of these three equations
expression,
in
what point q do they meet the conic

we have the three sca-
the vector of this point,
evidently satisfied by the value
by one other
(13.)
XXXIL, by assuming the
+ |(a:-Jr'-a;u'),
the relations,
Sk$ = Sep, SXv' = - 2S\p, S/tr' = 2S/ip,

= SX^ = S/i? = SXr' = S/xu' = Sk/ = Skv',
XXXIV.
whence XXXIII.
gives,
XXXV.
SXw = a;SXp,
S/tw
x-^Sfip, &c.
CHAP.
And
INTERSECTION WITH OSCULATING CIRCLE.
III.]
545
because
XXXVI. ..p = ^ + Kr'-^'),

we shall satisfy also the third equation XXXII., if we adopt for x any root of that
new scalar equation, which is obtained by equating the square of the expression
XXXIII. for to, to what that square becomes when x is changed to 1.
(14.) To facilitate the formation of this new equation, we may observe that the
relations,
K = p-p",
which have
all
r'
p'
+ p",
v'
XXXVII.
the resulting equation
is
Spp'=0,
Spp"=-p'2,
ASKV =
- 4S$r' = 3t'2 + v'^,
therefore, after a
dratic in X,
XXXVIII.
of which the cubic factor
the point p itself.
(15.)
= p'-p",
occurred before, give
We have
XXXIX.
comparing which
= (:r -
3v'^
1)3
{v^x
r'2)
to be rejected (comp. 393, (2.)), as answering only to
then the values,
is
r'2
few slight reductions, the following biqua-
ic
XL.
and
r'2i;'-2,
last expression
OQ=a; = $-f |
^^
- '^
with the formulae XVII., we see that the required
point of intersection q, of the sphero-conic with its osculating circle, can be constructed
by the following rule. On the traces (4.), of the osculating plane on the two cyclic
planes, determine
XLI.
two points Ti and
XT.XTi=XU2,
Ui,
by the
conditions,
XU.XUi = Xt2;
then
XLII.
TiQ
= QUi,
or in words, the right line TiUi is bisected hy the sought point Q.

(16.) But a still more simple or more graphic construction may be obtained, by
The
of the chord PQ.

and XL.,
investigating (comp. 393, (4.)) the direction

this rectilinear
XLIIL
chord
PQ
XXXVI.
by
is,
= io-p = l(y"^ - r'2)
(w'-i -f t'-i)
)'2
f p'2
= hCr-^-
vector value of
v''^) t'^t'
v') v'
o'2^
p'2\
pq has therefore the direction (or its opposite) of the fourth proportional
v', or pt, xt, and xu; if then we conceive
(226) to the three vectors, p\ t', and
this chord or its prolongations to meet the traces xt, xu in two new points T2, U2,
we shall have (comp. 393, VIII.) the two inversely similar triangles (118),
the chord
XLIV.
T2XU2
cc'
uxr.
we may conceive /owr

(17.)
planes, through the axis oke, perpendicular respectively to the /our following right
To deduce hence a
lines in the osculating
XLV.
plane
.
r',
spherical construction for Q,
v',
(o-p,
p',
which planes will cut the sphere iufour great
XLVI.
are parts,
if F,
G,
or
xt, xu, pt,
circles,
pq
whereof the four arcs,
EF, EG, EP, EH,
(see again Fig. 80) be the feet of the three areual perpendiculars
from the pole e of the osculating

areual chord pq.
circle
on the two cyclic arcs cb, ca, and on the
4 A
546
[boOK
111.
XLVI. are therefore connected by the same angular relaXLV. and we have thus the very simple formula,
XLVII.
GEH = PEF,
These four arcs
(18.)
tion as the /owr lines
to
expressing an equahty between two spherical angles at the pole e, which serves
determine the direction of the arc eh, and therefore also the positions of the points
and
Q,
by means
of the relations,
XL VIII.
PHE =
= O HQ.
n PH
i
or any chord of the conic,
(19.) If the arcual chord pq, both ways prolonged,
cut the cyclic arcs cb and CA in the points k and s (Fig. 80), it is well known that
there exists the equality of intercepts (corap. 270, (2.)),
XLIX.
and conversely
this equation,
RP = n Qs
-^
(11.), or with the trigono-
combined with the formulae
metric expression,
L.
tan PE = tan r
| sin
(cot A + cot b),
for the tangent of the arcual radius of the osculating circle, enables us to determine
may
what
be called perhaps the arcual chord of osculation pq, by determining the
or simply p, from principles of spherical trigonometry alone,

spherical angle rpb,
in a way which may serve as a verification of the results above deduced from quaternions.
rh = hs, and by
(20.) Denoting by t the semitransversal
the oblique-angled triangles rpb, spa give the equations,
{
LI.
cot {t
cot {t
s)
sm - =
cos
s)
sm - =
cos p cos -
the semichord
p cos - + sm p cot
ph = hq,
b,
<i
while the right angled triangle
phe
LII.
- sm p
cot
gives,
tan s
= sin p tan r.
Equating then the values of cot 2s, deduced from LI. and LII., we eliminate s and f,
and obtain a quadratic in tan p, of which one root is zero, when tan r has the value
L.
such then might in this
new way be
inferred to be the tangent of the arcual ra-
dius of curvature of the conic, and the remaining root of the equation
cos -
LIII.
tan p
is
then,
- cot a)
(cot B
cot
a formula
PQ
to the
cot
B+
cos2
which ought to determine the inclination

tangent pa, but which does not appear at
tan2 r
p, or
first
rpb, or qpa, of the chord
sight to admit of any simple
interpretation*
*
limit;
We
might however at once see from this formula, that p = A - B at the plane
which agrees with the known construction 393, (4.), for the correspond? ii"'
chord PQ in the case of the plane hyperbola.
CHAIMII.] HYPERBOLIC CYLIDER, ASYMPTOTIC PLANES. 547

(21.) On the other hand, the construction (17.) (18.), to which the quaternion
analysis led us, gives
LIV.
and
HEP = GEP - GEH = GEP - PEF = FEB -f GEA,
therefore, by the four right-angled triangles, phe, bfe, age, and
bpe
or
epa,
conducts to this other formula,
LV.
cofi (cos r cot p)
= cot-M
cos r cos
- tan
(b + a)
cot-J
in
which a
is
the same auxiliary angle as in
XXXI.
the proposed verification (19.), that this last equation

relation between A, b,
and
c,
cos r cos - tan
(a + a
j,
we ought therefore to find, as

LV. expresses virtually the same
;
the formula LIII., although there seems at
first to
and such agreement can accordingly be proved

by a chain of ordinary trigonometric transformations, which it may be left
be no connexion between
exist,
p, as
them
to
to
the reader to investigate.
(22.) k geometrical proof oi ihQ validity of the construction (17.) (18.) may
The product of the sines of the arcual perpendibe derived in the following way.
culars, from a point of a given sphero-conic on its two cyclic arcs, is well known to
be constant ; hence also the rectangle under the distances of the same variable point
from the two cyclic planes is constant, and the curve is therefore the intersection of
the sphere with an hyperbolic cylinder, to which those planes are asymptotic.
It
may
then be considered to be thus geometrically evident, that the circle which oscu-
lates to the spherical curve, at
which
is
any given point
p, osculates also to the hyperholuy
made by
the osculating plane at this point ;

the point in which this hyperbola is
the section of that cylinder,
and that the point q, of recent investigations,

met again, by its own osculating circle at p.
is
But the determination 393, (4.) of

such a point of intersection, although above deduced (for practice) by quaternions,
is a plane problem of which the solution was known ; we may then be considered to
have reduced, to this known and plane problem, the corresponding spherical proband thus the inverse similarity of the two plane triangles XLIV.,
although found by the quaternion analysis, may be said to be geometrically exlem (12.);
- v', of the osculating

plained, or accounted for the traces xx and xu, or t and
plane to the conic on the two cyclic planes (4.), being evidently the asymptotes of
:
the hyperbola in question.

(23.) In quaternions, the constant product of sines, &c.,
II. of the cone,
is
expressed by this
LVI.
SUXp SU/ip = (g- SX/i)

.
2TX/A
= const.
and the scalar equation of the hyperbolic cylinder, obtamed by eliminating p2 between I, and II., after the first substitution (1.), is
LVII.
SXpS/ip
XXXIII.
= ir^ (g - SXfx) = const.
be considered as the vector equation of

the hyperbola, of which the intersection Q with the circle, or with the sphere, is de tcrmined by combining that equation with the condition w^ = p^ (=: r-).
while the expression
for
a>
may
548
III.
[bOOK
(24.) In the foregoing investigation, we have treated a sphero-conic in connexion

its ci/cUc arcs (2.)
but it would have been about equally easy to have treated
with
the same curve, with reference to

of which
its focal
points
or to the focal lines of the cone,
the intersection with a concentric sphere.

called the bifocal transformation, in 360, (2.)).
it is
(Compare what has been
We
can however only state generally here the result of such an application
(25.)
of quaternions, as regards the construction of the osculating small circle to a spherical conic, considered relatively to
its^ci
which construction* can indeed be also
Two
geometrically deduced, as a certain polar reciprocal of the one given above.
focal points (not mutually opposite) being called f and G, let pn be the normal arc
at p,
which
arcs, FP,
thus equally inclined, by a well-known principle, to the two vector

so that if the focus G be suitably distinguished from its own opposite,
is
GP
the spherical angle fpg is bisected by the arc pn, which is here supposed to terminate on the given arc fg. At N erect an arc qkr, perpendicular to pn, and termiPerpendicidars, qe, re, to these last
nating in q and E, on the two vector arcs.
arcs, will meet on the normal arc pn, in the sought pole (or spherical centre^ e, of
the sought small circle, which osculates to the conic at the given point p.
(26.) The two focal and arcual chords of curvature from p, which pass through
F and
G,
and terminate on the osculating
circle,
in virtue of the foregoing construction, which
The great
which
circle qb,
is the
are evidently bisected at q and k,
may
common
therefore be thus enunciated
of the two focal and arcual

the normal arc pn on the fixed
bisector
chords of curvature from a given point p, intersects

arc FG, connecting the two foci ; that is, on the arcual major axis of the conic.
(27.) The construction (5.) fails to determine the position of the auxiliary point
in Fig. 80, for the case
and
when
the given point p
in fact the expressions (4.) for p'
is
on the minor axis of the conic
and p" become
infinite,
when the denominator
But it is easy to see that the auxiliary vector <r, which represents
SX/XjO is zero.
generally the trace of the normal plane to the curve on the plane of the two cyclic
normals, becomes at the limit here considered the required axis of the osculating
circle;
and accordingly,
LVIII.
we have
LIX.
(28.) In general,
if
we assume simply
(corap. (1.)
and
(2.)),
= Vptr, and therefore p" = VpV + Vpo-',

= 0, and Vp'p" a, when SX/tp = 0.
p'
o-'
|1
if
we determine
three points l, m, s in the plane of X/u,
by the
formulae (comp. again (2.)),
r ^
= ^P^
LX...OL
A^P'
,
0==g-^,
then L and
^P^
OS=
=l(OI. + OM),
-I
will be the intersections of the cyclic normals X,
p.
with the tangent
* The reader can easily draw the Figure for himself. As regards the known
and 394, (22.)), for determining the chord of
intersection of a plane conic with its osculating circle, it will be found (for instance)
rule, lately alluded to (in 393, (4.),
in page
194 of Hamilton's Conic Sections
(in Latin,
London, 1758).
The two
sphe-
rical constructions, for the small circle osculating to a spherical conic, were early
deduced and published by the present writer, as consequences of quaternion calculations.
Compare the
first
Note to page 535.
CHAP.
III.]
OSCULATING CONE, OSCULATING SPHERE.
549
plane to the sphere at p, and the normal plane to the curve at the same point will
bisect the right line lm in the point s ; we shall also have this proportion of sines,
LXI.
sin
= cos LOP
if
LOS
:
cos
sin som = SUXp SUfip

POM = sin ppi sin PP2,
:
pPi, PP3 be the arcual perpendiculars
comp. (23.),
from the point p of the conic on the two cyclic
and this general rule for determining the position of the line OS, or tr, applies
even to the limiting case (27.), when that variable line becomes the axis of the osculating circle, at a minor summit of the curve.
arcs
II.
(29.) As an example, let us suppose that the constants^, X,

are connected by the relation,
LXII.
.g
= - SX/i,
whence
LXIII.
fi
in the equation
S(VXp .V/zp) =
the cyclic normals are therefore in this case sides of the cone, and the two planes
which connect them with any third side are mutually rectangular ; so tha^ the conic
is now the locus of the vertex of a right-angled spherical triangle, of which the
is
hypotenuse
And by
given.
applying either the formula LXI., or the construction
we find that th^ trigonometric tangent oii\i% arcual radius

(28.) which it represents,
of the osculating small circle to such a conic, at either end of the given hypotenuse,
equal to half* the tangent of that hypotenuse itself.
(30.) It is obvious that every determination, of an osculating circle to a spherical
curve, is at the same time the determination of what may be (and is) called an osis
which rests upon that curve,

Applying this remark to the last example (29.), we arrive at the following theorem, which can however be otherwise
culating right cone (or cone of revolution), to the cone
and has
its
deduced
vertex at the centre of the sphere.
If a cone he cut
in
circle
right cone which osculates to
it
hy a plane perpendicular to a side, the axis of the

along that side passes through the centre of the sec-
tion.
395.
curve,
When
we may
a given curve of double curvature is not a spherical

propose to investigate the spheric surface which ap-
An osculating
proaches to it most closely, at any assigned point.
circle has been defined (389) to be the Zmz^ of a circle, which, touches
a given curve, or its tangent pt, at a given point p, and cuts the same
curve at a near point q; while the tangent pt itself had been regarded
(100) as the limit of a rectilinear secant, or as the ultimate position

of the small chord pq.
It is natural then to define the
osculating
being the limit of a spheric surface, which passes through

the osculating circle, at a given point P of a curve, and also cuts that
curve in a point Q, which is supposed to approach
indefinitely to P,
and ultimately to coincide with it. Accordingly we shall find that
sphere, as
this definition
* This
may
conducts by quaternions to formulce sufficiently simalso be inferred
by
limits
from the formulae (11.)
in
a were used, provisionally, to denote a certain spherical arc and angle.
which r and
550
pie;
and that their geometrical
known
[bOOK
interpretations are consistent
III.
with
results: for example, the centre of spherical curvature, or the
centre of the osculating sphere, will thus
be shown to be, as usual, the
point in which th& polar axis (391, (5.)) touches the cusp-edge of the
polar developable (391, (6.))- It will also be seen,that whereas in

if R be a point in the normal plane (370, (8.)) to a given
geneiâl,
curve at p, we can only say that the difference of distances, rq-rp,
is
small of an order higher than the
first,
if
the chord pr be small
of the^rs^ order ; and whereas, even if r be on the polar axis (391,

that this difference of distances is
(4.)), we can only say generally
small, of an order higher than the second ; yet, if R be placed at the
centre s of spherical curvature, the difference sQ-spis small, of an
order higher than the third: so that the distance of a near point ot.from
the osculating sphere at the given point p, is generally small
of the fourth
order, the chord being still small of ihe first.
where X
(1.) Operating with S.X,
392, V. of the osculating
circle,
is
an arbitrary
we obtain
line,
on the vector equation
the scalar equation of a sphere through
that circle under the form,

I..
which may however, by 393,
.0 =
(7.),
28-^+8^;
w-p
p
be brought to this other form, better suited to
our present purpose,

II.
= (p - k)3 + 2cSp"p' (w - p)
k)2
(a>
scalar constant, while k
being any
and the vector
or
which evidently expresses that
To express now
substitute for
(T
the vector of the centre
still
of the centre s of the sphere

III.
(2.)
is
is
of the circle
= K + cYp"p',
this last centre
is
on the polar axis.
that this sphere cuts the curve in a near point Q,
we
are to
the expression,
IV.
.<jj
= pt=p + tp' + ^ty + ^tûtp'",
with
tto
but K has been seen (in 391) to satisfy the three equations,
V. ..
reducing then, dividing by ^t^,

sphere the condition,
VI.
so that finally the vector
VII.
by which
<t
Sp"(K-p)-jo'2,
and passing
0=Sp"p'(K-p);
to the limit,
we
find for the osculating/
Sp"\p-K) + 3Sp'p" = cSp"'p"p';
the three scalar equations,
satisfies
.0=Sp'{<T-p\
it is
= Sp'(K-p),
Sp"(tr
- p) - p'S
= Sp"'(<r - p) ~ SSp'p",
completely determined, and of which the two last are seen to be the
first, while that first is the equation of the normal plane
successive derivatives of the
CHAP.
CONSTRUCTION FOR CENTRE OF SPHERE.
III.]
whence the centre
s of this sphere is
also be thus written (comp. 391, (4.)),
may
VIII.
be substituted for derivatives in the equations VII.,
may
(3.) Differentials
(by the sub-arts, to 386, comp. 391, (6.)) the
ks touches the cusp-edge of the polar developable.
point where the polar axis
which
551
= dT(p-fr),
= d3T(p-(r),
= d2T(p-<r),
if
d(T
= 0;
^
the distance of a near point
of the given curve from the osculating s/jAere is there-
fore small (as above said), of an order higher than the third,
if
the chord pq be small
of the Jirst order.
The two
(4.)
first
equations VII., combined with V., give also
IX...O = Sp'((X-K),
= Sp"((T-K),
= S(K:-p)((r-K);
which express that the line ks is perpendicular to the osculating plane and absolute
normal at p, as it ought to be, because it is part of the polar axis.
(5.) Conceiving the three points P, K,
results the following
X.
of
or their vectors p, k, c, to vari/ together,
s,
and VII., combined with
the equations V.
their
own
derivatives, give
among
other
= Sk'p = S<7'p' = Sff'jo" =
S<t'(k
- (o) = S(t"p'
which the geometrical interpretations are easily perceived.

(6.) Another easy combination is the following,
XL
..0
= S'((T + p-2:),
by derivating the last equation IX., with attention to other relations;

but 2k: - p is the vector of the extremity, say M, of the diameter of the osculating
circle, drawn from the given point p ; we have therefore this construction
as appears
On
the tangent
kk'
of the centre of the osculating
to the locus
perpendicular from the extremity
circle, let fall
of the diameter drawn from the given point P
a
;
perpendicular prolonged will intersect the polar axis, in the centre s of the osculating sphere to the given curve at p.
this
(7.) In general, the three scalar equations VII. conduct to the vector expression,
XII.
(r
+ p'^Yp"'p'
= p+ 3Yp'p"Sp'p"

c
,
Spp
or with differentials.
XIII.
..^ = p+
the scalar variable being

(8.)
And
if,
XIY.
still left
as an example,
.
3Vdpd2pSd,od2p + dpSVd^pdp
'
Sdpd2pd3p
arbitrary.
we introduce
= cta + a%
p'
= ca +
the values for the
â+i|3,
Iielix,
p"=-(^)'/3,
(^P
occurred before, we find after some slight reductions the expression, in which a denotes again the constant inclination of the curve to the axis of
whereof the three
first
the cylinder,
XV.
but this
is
= p- a*(3 cosec2 a = eta a*j3 cot^ a
what we found for k, in 389, VIII. for the

K and s, of absolute and spherical curvature, coincide.
precisely
two centres,
helix,
then, the
552
III.
[boOK
(9.) This known result is a consequence, and may serve as an illustration, of the
general construction (6.) ; because it is easy to infer, from what was shown in 389,
(3.), respecting the locus of the centre k of the osculating circle to the helix, as being
another helix on a co-axal cylinder^ that the tangent kk' to this locus is perpendicular to the radius of curvature kp, while the same tangent
(kk' or k) is always
perpendicular (X.) to the tangent (pp' or p') to the curve ; kk.' is therefore here at
right angles to the osculating plane of the given helix, or coincides with its polar
axis : so that the perpendicular on it from the extremity m of the diameter of cur-
vature
falls at
the point
case coincides, as found
(10.) In general,
XVI.
p'
if
itself,
with which consequently the point s in the present
by
calculation in (8.).
we
introduce the expressions 376, VI., or the following,
= s'r>sp,
p" = s"^Ds^p
+ s"t>sP,
p'"
= s'^Ds^p + Bs's'Ds^p + s"t>sP,
in which s denotes the arc of the curve, but the accents

respect to an arbitrary scalar
lations,
XVII.
Dp2
and
= - 1,
if
we
still
indicate derivations with
observe (comp. 380, (12.)) that the re-
S DspD,2p
.
0,
T>spT)s^p
Wp^ = 0,
which Dp' and Dj^p^ denote the squares of D^p and Ds'^p, and S DspD^^p denotes
S(Dgp.Ds'p), &c., exist independently of the form of the curve ; we find that s" and
s'" disappear from the numerator and denominator of the expression XII. for a- p,
in
and that they have
for
s'6
common
factor: setting aside which,
we have
thus the
simpler formulae,
S DspDs^ piVp
IDspDs'^pDs^p
And
accordingly the three scalar equations VII., which determine the centre of the
osculating sphere, may now be written thus,
XIX.
S((r-p)D,p
= 0,
S(fr-p)D,2p+
0,
S (tr - p) D,'p =
0.
(11.) Conversely, when we have any formula involving thus the successi^se derivatives of the vector p taken with respect to the arc, s, we can always and easily
generalize the expression, and introduce an arbitrary variable

or by writing (comp. 390, VIII.),
equations XVI.
t,
by inverting the
XX.
T>sp =s'"ip',
= s'-2p"_
Ds^p ='''(s'-'p')'
s''^s"p\ &c.
may happen (comp. 379, (2.)) that the independent variable t is only
proportional to s, witliout being equal thereto ; but as we have the general relation,
(12.) It
XXI.
Di^p
= s'Ds"p,
if
'
= Dts = Tp' = const.,
nearly or quite as easy to efiect the transformations (10.) and (11.) in the case
here supposed, or to pass from t to s and reciprocally, as if we had ' = 1.
(13.) If the vector a be treated as constant in the derivations, or if we consider
it is
for a
moment
riations
the centre s of the sphere as
of distance
= (p ay) we
a.
fixed point, and attend only to the va o-)^

it, then (remembering that T(p
of a point on the curve from
not only easily put (comp. VIII.) the three equations
the forms,
XXII.
= D,T (p - a) = D,2T (p - cr) = D,3T(p - a),
but also obtain by XVII. thh fourth equation,
XXIII.
T(p -
(t)
D,^T(p
<7)
((7
- p) D/p +
D,2p2.
XIX.
under
CHAP.
COEFFICIENT OF DEVIATION FROM SPHERE.
III.]
we
(14.) If then
XXIV.
553
XXV.
= T(k p) = TOs-p-^ = radius of osculating circle ;

R = T(a - p) = radius of oscidating sphere ;
.
and
S = ^C^JL^lPl^
- T>s^p^
XXVI
we
S.D,p3p,3pD,V
S DspD,-p3D^3p'
.
must be constantly equal to unity, for every spherical

a curve which is 7wn-spherical, the distance sq of a near point
see that this scalar, S,
curve
Q,
,3
but that for
from the centre s of the osculating sphere at
p, is generally
given by an expres-
sion of the form,
XXVII.
SQ =
/2
with o =
^^^^',
80 that, at least for near points Q, on eacfi side of the given point p, the curve lies
without or within the sphere which osculates at that given point, according as the *calar, s, determined as above, is greater or less than unity.
(15) In the
XXVI. may
case (12.), the formula
XXVIII...
be thus written,
5=|:eV>Z.
whence, by carrying the derivations one step farther than in (8.),

helix,
XXIX.
and accordingly
5'=cosec2a>l,
it is
or
XXIX'.
easy to prove that this
curve
lies
we
find for the
5- l = cot2 a>0
wholly vnthout
its
osculating
sphere, except at the point of osculation.
(16.) In general, the scalar
and which enters as a
SQ
sp
S1, which
vanishes (14.) iorall spherical curves^

XXVII. for the deviation
coefficient into the expression
any other curve from its own osculating sphere, may be

of Nan- Sphericity ; and if qt be i\iQ perpendicular from that
tangent pt to the curve at the given point P, we have then this
of a near point of
called the Coefficient
near point Q on the

limiting equation,
by which the value of that
XXX... ^-l = lim.
coefficient
may
be expressed,
/sq2-S
V
QT^
(17.) Besides the forms XVIII., other transformations of the expressions XII.
XIII. for the vector

it
seems
<t
of the centre of an osculating sphere
sufficient here to suggest that
that those expressions for
some
useful practice
a reduce themselves generally to
XXXI.
..
might be assigned
may
zero,
but
be had, in proving
when
the condition,
Tp = const.
is satisfied.
r-i is often called (comp. 389,

(4.)) the
(18.) It may just be remarked, that as
absolute curvature, or simply the curvature, of the curve in space which is consi-
dered, so iJ-i
R are
is
sometimes called the spherical curvature of that curve
called the radii* of those
two curvatures
while r and
respectively.
* "We shall soon have occasion to consider another scalar

radius, which we proby the small roman letter r, of what is not uncommonly called the
pose to denote
torsion, or the
second curvature, of the same curve
4b
in space,
ELKxMKNTS OF QUATERNIONS.
554
When
396.
the arc (s) of the curve
we have
variable, the calculations (as
while no essential
plified,
[bOOK
III.
made the independent

become considerably sim-
is
seen)
generality/ is lost,
because the transforma-
tions requisite for the introduction of an arbitrary scalar variable
follow a simple and uniform law (395, (U.), &c.).

the expression (comp. 395, IV.),
I.
in
p^
= p-^s7 +
4-
^sûy\
t'
= D3V,
Js^t'
(t)
Adopting then
with
Mo
which
II.
BsP,
-t"
D//>,
and therefore
III.
we
shall proceed to
t'-
= 0,
Stt'
deduce some other
0,
Stt" +
t^2
affections of the curve^ besides
curvature (395, (18.)), which do not involve the consipower of the arc (or chord). In particular, we
shall determine expressions for that known Second Curvature (or
its spherical
deration of
t\\Q fourth
which depends on the change of the osculating plane, and is

measured by the ultimate ratio of that change, expressed as an angle,
to the arc of the curve itself; and shall assign the quaternion equa-
torsion),
tions of the known Rectifying Plane, and Rectifying Line, which are
respectively the tangent plane, and the generating line, of that known
Rectifying Developable, whereon the proposed curve is Sk geodetic (382)
:
so that
would become
it
a right
linCy by the unfolding of this last surmay be well to express, in this new
2k
But first it
plane.
notation, the principal affections or properties of the curve, which
depend only on the three first terms of the expansion I., or on the
face into
three initial vectors
which
fying
line
(1.)
/>,
t,
t',
or rather on the two last of these ; and
we
shall see, the rectifying plane, but not the rectinor what has been called above the second* curvature.
include, as
Using then
instead of
first,
I., this less
expanded but
still
rigorous expres-
sion (comp. 376, I.),
IV.
* In a
Note
to
ps
=p+
sr
^s'hisT,
with
uq
1,
a very able and interesting Memoir, ** Sur les lignes courbes non
by Dr. Salmon in the Note to page 277 of his already cited
planes''' (referred to
and published in Cahier XXX. of the Journal de VEcole Polytechnique),

Saint- Venant brings forward several objections to the use of this appellation,
Treatise,
M. de
and
which he proposes to intro" camhrure ;" but the

" second curvature'^
expression
may
for the present, as being at least not unusual, and appearing to be suffi-
also to the phrases torsion, flexion, &c., instead of
duce the new name,

serve
xis
ciently suggestive
CHAP.
EMANANT LINES AND PLANES.
III.]
and with the relations
II.
and
we have
III.,
at once the following system of three
rectangular lines, which are conceived to be
curve
V.
r=
VII.
r being a line
unit tangent
.
v=
drawn
which the arc
IX.
r~^
while
drawn from the given point p
of the
v (which
has the same length (Tv

from r to t is positive.
at P,
as
the vector
is
now,
= p + r^T = p + rUr',
r = Tt'-^ = radius of curvature ;
IX'.
or
and
r',
(4.);
;
r''i
normal at P
= r-i)
of curvature (389,
a conceived motion along the curve, in virtue

towards the centre of curvature,
which centre
is
vector
binormal (comp. 379, (4.))
r' is directed
OK = K = p -
= TT' = curvature
line
r'=
in the direction of
(s) increases
VIII.
and the third
VI.
= -r'T = t't~^ =
tt'
or of the osculating circle, of
if
all
and
of
555
to the surface
of tangents to the curve)
directed so that the rotation round
is
it
At the same time, we have evidently a system of three rectangular vector

from the same point p, which may be called respectively the tangent unit, the
normal unit, and the binormal unit, namely the three lines,
(2.)
units
X.
the
The
(3.)
Ur' = r/,
Ur = r,
normal unit being thus directed
Uv = rTT
(like r') towards the centre of curvature.
vector equation (comp. 392, (2.)) of the circle of curvature takes
now
the form,
bJ
with the verification that
in
which
fx
by the
it is satisfied
XII.
(comp. 395. (6.))
a*
the vector
is
value,
= /w = 2/c-p = p-2f'->,
om
of the extremity of the diameter of
curvature pm.
(4.) The normal plane, the rectifying plane, and the osculating plane, to the
curve at the given point, form gLjrectangular system of planes (comp. 379, (5.)),
so that their scalar equations are, in
perpendicular respectively to the three lines (1.)
;
the present notation,
..Sr(w-p) = 0;
XIII.
XIV.
Sr'
(<>
p)
XV.
Sr (ai-p) =
by pairing which we can represent the tangent, normal, and binormal

regarded as indefinite right lines
XVI.
Vr(w-p) = 0;
C5.) In general,
XIX.
if
XVII.
or
.
by the three vector
Vr'(w-p) = 0;
to the curve,
equations,
XVIII.
.. Vj/ (o>
p)
0.
the two vector equations,
Vj7 (w
represent two right lines,
= 0,
p)
ph and
p^Hs,
and
XIX'.
V;;, (w,
which are conceived
to
- p,) = 0,
emanate according to
any given law from any given curve in space, the identical formula,*
* It
is
obvious that
we have
draw a common perpendicular
to
thus an easy quaternion solution of the

problem to
any two right lines in space.
ELEMKNTS OF QUATERNIONS.
556
[bOOK
III.
shows that the common perpendicular to these two emanants, which as a vector is remember of this formula XX., intersects the two lines in the two
presented by either
points of which the vectors are,
XXI.
w = p + jb
=r
XAl
u)s
Yv^s
= ps +
J740
Yriris
(6.) In general also, the passage of a right line from any oe given position in
space to any other may be conceived to be accomplished by a sort oi screw motion, with
the common perpendicular ior the axis of the screw, and with two proportional velocities, of
translation along, and of rotation
round that axis
the locus of the two given
(when Mms interpolated^ being a

Screw Surface, such as that of which the vector equation was assigned in 314, (1 1.),
and was used in 372, (4.).
and of
a^^ the intermediate positions of the line
(7.) kga\n,iov
any quaternion,
q,
we have (by 316, XX. and XXIII.*)
the two
equations,
XXII.
= ^ g.UV^,
IU5
XXir.
VUg = sin
L 9.UV9
comparing which we see that
XXIII.
Y\Jq
= m\ Lq: Lq=
lU^
(very nearly)
1,
the angle of the quaternion be small ; so that the logarithm and the vector of the
versor of a small-angled quaternion are very nearly equal to each other, and we may
write the following general approximate formula for such a versor :
if
XXIV.
Ug = (iU2 =) jTUg^
nearly, if ^ g be small;
the error of this last formula being in fact small of the third order,
small of the^rs^.
(8.)
if
And
thus or otherwise (comp. 334, XIII. and XV.),
XXV.
if
we
perceive that
= J7,j;-i,
with
XXVI.
.. 17,=
= V -,
and
XXVIII.
ri
we
??
+?'
shall then
have nearly,
XXIX.
neglecting
- = *,
.,
A = S -,
V
and
XXX.
Tg =
T- = 1 + sA
?/s
= (1 + sh) sÔr] = t'Orj + shrj,
last binomial, the _^rs^ (or exponential)
of the near
emanant
s^,
XXXI.
which
be small, the expressions,
if
.Ug = U
f]
or,
write for abridgment,
XXVII.
it
we may
the angle be
the quaternion q have the form (comp. (5.)),
and
in
if
term alone influences the direction
line (5.).
*
Although the expression XXII'. for VUg is here deduced from 316, XXIIL, yet
might have been introduced at a much earlier stage of these Elements ; for instance,
in connexion with the formula 204,
XIX., namely TVU5' = sin Lq.
CHAP.
(9.)
VECTOR OF ROTATION, AXIS OF DISPLACEMENT. 557
III.]
At the same
by supposing
time,
s to tend to 0, the
formula
XXI.
gives, as
limit,
TTl
XXXII.
OH = wo = P + JyS-h = |0-J?S
for the vector of the point,
say h, on the given emanant ph, in which that given
by the common perpendicular
line is ultimately intersected
(5.),
or
by the axis of
the screw rotation (6.) ; but the direction of that axis is represented by the versor
U9, and the angular velocity of that rotation is represented by the tensor TO, if the
velocity of motion (1.) along the given curve be taken as unify
say that the vector 9
itself,
or the factor which multiplies the arc,
we may
therefore
in the exponential
XXXI., if set oS from the point u determined by XXXII., is the Vector of

Rotation of the Emanant, whatever the law (5.) of the emanation may be.
(10.) And as regards the screw translation (6.), its linear velocity is in like
term
manner
by
in direction,
by the following expression (obtained
from XX.),
XXXIII.
= BS
= projection
And
and
represented, in length
limits
(set off
from h)
Fector of Translation of Emanant,
of unit-tangent on screw-axis (or of r on 9).
the indefinite right line through the point H, of which this line
a part,
is
may
be called the Axis of Displacement of the Emanant.

(11.) It
is
easy in this manner to assign what
may
be called the Osculating
Screw Surface to the (generally gauche^ Surface of Emanants, or indeed to any

proposed skew surface ; namely, the screw surface which has the given emanant
(or other) line for one of its generatrices, and touches the skew surface in the whole
extent of that right line.
(12.) It
surface of
is
however more important here
emanants
conversely this vector
XXXIV. ..t =
is,
(13.) In the
or
0,
= 0,
Sr6>
same
case, of
7/
or
||
r,
may
is
if
or
was
when
the
and that
that surface be undevelopable.
XXXIV.
is
therefore expressed
Sj7J7V
to be expected)
= 0;
when
j/
= r,
that
of tangents.
the vector 9 of rotation becomes equal (by
the expression XXXII., for the vector wq
XXVII. and VII.) to the binomial v and

of the foot H of the axis reduces itself to p
indeed
that in the case
of translation vanishes
of the Surface of Emanants
this condition is satisfied (as
for the surface
cannot be constantly zero,
The Condition of Developability

by the equation,
and accordingly
to observe,
developable, the vector
is
otherwise evident),
ihdit t\\Q
and thus we might be led to see (what

passage from a given tangent to a wear one
;
be approximately made, by a rotation round the binormal, through the small

sTv =r"i = arc divided by radius of curvature.
angle,
(14.) Instead of emanating lines, we may consider a system of emanating planes,

which are respectively perpendicular to those lines, and pass through the same points
of the given curve.
to another of
XXXV.
may
lint,
It
may
be sufficient here to remark, that the passage from one

planes, represented by the equations,
two such near emanant

.
be conceived to be
Sj7
(w
- p) =
made by a
0,
XXXV.
rotation through
S;?,
- p) = 0,
= aT9, round the
(w
an angle
right
558
- p) = 0, SrjXui - p) - S;r= 0,
Sj? (w
XXXVr. Ye(^(o-p) + r)-iSr}T=0,
XXXV. touches its developable envelope, and which
XXXVI.
or
in
[bOOK
which the plane
III.
to the recent vector 6, or to the vector
of rotation (9.) of the emanant
is
line
parallel
so that
an equal vector be set off on this new line XXXVI., it may be said to be the Vector Axis of Rotation of the Emanant Plane.
= r, so that the equation XXXV. repre(15.) For example, if we again make /
if
sents
now
we
the normal plane to the curve,
and
that plane with its derived equation,

equations,
XXXVII.
Sr(w -
are led to combine the equation XIII. of

so to form the system of the two scalar
= 0,
p)
- p) +
Sr'(w
= 0,
whereof the second represents a plane parallel to the rectifying plane XIV., and
drawn through the centre of curvature VIII. and which jointly represent the polar
;
axis (391, (5.)), considered as an indefinite right

by the one vector equation,
XXXVIII.
(16.)
And
if,
J/
(w
- *c) = 0,
we
is
represented otherwise
Vv (a* -
p)
=-
r.
a portion equal to thebinormal v,

line,
conveniently be measured from the centre k) may be said,
may
such portion (which
which
XXXVIII'.
or
on this indefinite
line,
set off
by (14.), to be the Vector Axis of Rotation of the Normal Plane ; or briefly, the
Polar Axis, considered as representing not only the direction but also the velocity of
that rotation, which velocity =Tj/ = r"i = the curvature (IX.) of the given curve
;
Uv =
the binormal unit (2,), set off on the same axis from
the same centre of curvature, may be called the Polar Unit.
while another portion
(17.) This suggests a new way of representing the osculating circle by a vector
equation (comp. (3.), and 316), as follows:
XXXIX.
= p^-ST^which agrees, as we
which,
when
Is^T
-I-
(t'"
the expansion
I.
^s^v^T
or IV., if s^ be neglected;
continued, the next term
is
XL.
1-sv)t'-'^
-1-SV- ^S2j^2) /-I
with the expression
see,
= p + (S''-l)r'-
Ws=/e + ''(p-K)
= p + ST + (t^''-
and of
is,
= ^s^vt' =
(18.) The complete expansion of the exponential form XXXIX., for the variable
vector of the osculating circle, may be briefly summed up in the following trigono-
metric (but vector) expression
XLI.
Ws
=K+
cos -
Ui^ sin
.
in which,
so that
XLII.
we may
=-
r^T-*,
also write, neglecting
XLIII.
a>s
and
(p
k),
Uv. (p-K) = ri'r'-i =rr;
no power ofs,
=p +
rr sin - + r2/ vers -
and if this be subtracted from the full expression for the vector p, the remainder may
be called the deviation of the given curve in space, from its own circle of curvature :
which deviation, as we already see, is small of the third order, and will soon be de-
CHAP.
VECTOR AND RADIUS OF SECOND CURVATURE. 559
III.]
composed into its two principal parts, or terms, of that

normal and the hinormal respectively.
(19.)
order, in the directions of the
Meantime we may remark, that if we only neglect terms of the fourth

T. gives, by III. and IX., for the length of a small chord pp,,
order, the expansion
the formula
XLIV.
= T(p.-p)=:T(sr + |2/ +
= V{-(r+|s2r' + ^^'r")2}
PP,
=
>/t*'-li73)
*-2i^
^s3r")
=
2^^^2-r'
this length then is the same (to this degree of approximation), as that of thecAord of
an equally long arc of the osculating circle : and although the chord of even a small
arc of a curve is always shorter than that arc itself, yet we see that the difference is
generally a small quantity of the third* order,
if
the arc be small of the frst.
397. Resuming now the expression 396, 1., but suppressing here
the coefficient w^ of which the limit is unity, and therefore writing
simply,
with the relations,

II.
if 5
arc,
=-
T^
and
r~^
ture (> 0), while

III.
with
1/
tt'
Stt'
r-i
=S
r'T^rS
z=
r-\
St't^'
= r-V,
= hinormal,
= _ t'2
= Tt'= curvature^'\ as before, or r= radius of curvar' = D^r

and introducing the new scalar,
= T-^V - = Secondl
or the
t"
IV.
Stt'^
0,
new
Curvature,
vector,
= V- = Vector of Second Curvature,

v'
supposed to be set off tangentially from the given point p of the
new
curve, or finally this other
V.
r=
scalar (> or
<
0),
= Radius of Second
Curvature,
* This
ought to have been expressly stated in the reasoning of 383, (5), for
it was not sufficient to observe that the arc and chord tend to bear to each other
which
a ratio of equality, without showing (or at least mentioning) that their

difference
tends to vanish, even as compared with a line which is ultimately of the same order
as the square of either.
t Whenever
it is
this
word curvature
is
thus used, without any qualifying

adjective,
always to be understood as denoting the absolute (or first) curvature of the curve
in space.
Compare the Note
to
page 554.
560
[bOOK
III.
which gives the expression,

t" = - r-^r - r-W + r'TT'
= - r-Ûr + (r-')'UT' + (rrY'TJu
VI.
we proceed
to deduce some of the chief affections ofa curve in space,

which depend on the third power of the arc or chord. In doing this,
although everything new can be ultimately/ reduced to a dependence
on the two new scalars, r' and r, or on the one new vector t"^ or even on
- Vtt'^,
yet some auxiliary symbols will be found useful, and almost

necessary. Retaining then the symbols i^, /c, o-, R, as well as t, t', r,
and therefore writing as before (comp. 396, VIIL),
v'
VIII.
VII.
-
we may now
k)-^
(/5
OK = = /)- T'-' = /jf rUT'=/3 + rV,

= r'^\J{K -p) = t' = D,V= Vector of Curvature,
/c
write also, by 395,
IX.
OS
<T
XVIIL,
=o -
I/'
K + rr'iv
fc -i-
r'vUv.
and
X.
. .
(/)
o-)"'
= i?-*U(o- - rt =
^projection of vector
XI.
Fecôr o/" Spherical Curvature^

ofcurvcUure on radius {R) ofosmlating sphere ;
we have now, by
because
VI.,
= (tt')' = Vtt'' = - r-'r' - r-W,

= {rvy = - rrW^ = - rÛr^
{Uv)'
= - r-'r'^ = r* r-^
StV' = - Stt
../'
or
XF.
and
XII.
If then
(j')
i^'"'StV
we denote by p and P
the linear and angular elevations, of the

above the osculating plane, we shall
centre s of the osculating sphere
have these two new auxiliary
scalars, which are positive or negative

together, according as the linear height KS has the direction of + v
or of - V :
XIII.
while
=
XIV. P= kps = tan-^^ = sin-i^=cos-' ^;
r
Uv r'r;
M
R
=
XV.
+
i?=T(<r-/>)=y(r2+i>0 \/0^ ^''r');
P being treated as generally acute. Another important

and an accompanying angle of elevation, are given by the for-
the angle
line,
mulae,
XVI.
= Yv'v'^ +
-=
\ = V^' =
?-2VTV^
= Rectifying Vector
= r-V + TT'=r-'UT + riUi^

(set off
from given point
Vector of Second Curvature plus Binormal;
p),
CHAP.
XVII.
CYLINDER, HELIX, AND PARABOLA. 561
OSC. CONES,
III.]
\
r
11= Z - = tan-^ - = Elevation of Rectifying Line (> 0, <
T
= the
tt),
angle (acute or obtuse, but here regarded as positive),
which that known and important

to the curve; so that (by XIII.,
H and P,
(396) makes with the tangent

XIV.) these two auxiliary angles^*
line
from which (instead of deducing them from r' and r)

the affections of the curve depending on s^ can be deduced, are
connected with each other and with r' by the relation,
all
XVIII. ..tan
P = r' tan jy.
other combinations of the symbols offer themselves easily,
Many
by
the rules of the present calculus; for instance, the vector a may be
determined by the three scalar equations (comp. 395, XIX.),
XIX.
= 0,
ST(<r-/>)
St'(^-/))
= -1,
St'/(<t-/))=0,
whence, by XVI,,
XX.
rV = r2 V(Vt't'^ {a - p)) = Y\{<t - p),
a result which also follows from the expressions,
XXI.
t'^=
V -+ SJT'
a - p=r^r' \-rpv = r1J'T' -^p\Jv,

XXII.
XXIII.
rpY\v = - rpv'W = - rr^T'
and
because
we may
= (\-r-V)T',
therefore replace the formula I. for the vector of the curve

is true to the same order of
approximation,!
by the following, which
XXIV. ..ft=^t^ +
+
^('^-/') |;VX(<T-rt:
and may thus exhibit, even to the eye, the dependence of all affections connected with 6^ on the two new lines^ X and <r- />, which were
not required when s^ was neglected, but can now be determined by
and
as before).
The
the two scalars r and p (or r and r', or
geometrical signification of the scalar p is evident from what precedes,
namely, the height (ks) of the centre of the osculating sphere above
that of the osculating circle, divided by the hinormal unit (Ui^) ; and
* The
appears to have been first considered by Lancret, in connexion
angle
with his theory of rectifying lines, planes, and surfaces but the angle hei'e called
was virtually included in the earlier results of Monge.
t As regards the homogeneity of such expressions, if we treat the four vectors

and <T, and the five scalars s, r, i?, p, and r, as being each of t\\Q first dimension, we are then to regard the dimensions of r, r', k', H, and P as being each zero ;
pty p, K,
those of
r', v,
and \ as each equal
to
and that of
4c
either r" or
v as being =
2.
562
[bOOK
III.
what has been called the radius r of second curvature (V.),

shall see that this is in fact the geometrical radius of a second cir-
as regards
we
cle, which osculates, at the extremity of the tangential vector rr, to the
principal normal section of the developable Surface of Tangents ; and
thereby determines an osculating oblique cone to that important sur-
and also an osculating
face,
H, and the
right cone* thereto, of
which
latter
cone the
the axis of revolution :

also
a
an
of
side
osculating right cylinder, on which is traced
being
what is called the osculating helix.
shall assign the quaternion
is
semiangle
rectifying line
is
We
equations of these two cones, and of this cylinder, and helix ; and shall
show that although the helix has not generally complete contact of the
third order with the given curve, yet it approaches more marly to that
curve (supposed to be of double curvature), than does the osculating
But an
osculating parabola will also be assigned, namely, the

which
osculates
to the projection of the curve, on its own osparabola
and
will
it
be shown that this parabola represents
:
culating plane
circle.
or constructs one of the two principal and rectangular components (396,

(18.)), of the deviation of the curve from its osculating circle, in a
direction which is (ultimately) tangential to the osculating sphere, while
the helix constructs the other component.
An osculatijig" right cone to the
drawn /row a given point of the curve, will also be assigned by quaternions and will be shown to have in general a smaller
tt tt acute
C
C), than the acute
cone of chords,
semiangle
H),
(or
of the osculating right cone (above mentioned) to the surface of tangents, or (as will be seen) to the cone of parallels to tangents (369,
semiangle
(6.), &c.)
(or
the relation between these two semiangles, of two osculating

by the formula,
right cones, being rigorously expressed
XXV... tanC = ftani?:
weif? oJ/ig-Me
cowe of the second order will be assigned, which has con-
tact of the same order with the co7ie of chords, as the second right cone
(C), while the latter osculates to both of them; and also an oscuculating parabolic cylinder, which rests upon the osculating parabola^
and
is cut
plane
perpendicularly in that auxiliary curve by the osculating

given curve. And the intersection of these two last sur-
to the
faces of the second order (oblique cone and parabolic cylinder) will
These two osculating cones, oblique and right, to the surface of tangents
appear to have been first assigned, in the Memoir already cited, by M. de Saint Venant the osculating (circular) helix, and the osculating (circular) cylinder, having
:
been previously considered by
M.
Olivier.
CHAP.
CONTACT OF THIRD ORDER, TWISTED CUBIC.
III.]
563
be found to consist partly of the hinormal at the given point, and

a certain twisted cubic* (or gauche curve of the third degree)^
partly of
which latter curve has complete contact of the third order with the
given curve in space. Constructions (comp. 395, (6.)) will be assigned,
which will connect, more closely than before, the tangent to the
locus of centres of curvature, with o^êr properties or affections of that
And finally we shall prove, by a very simple quaternion

consequence of the formula XF., the known theorem,f
that when the ratio of the two curvatures is constant, the curve is a
given curve.
analysis, as a
geodetic
scalar expression III., for the second curvature of a curve in space, as
The
(1.)
defined
on a cylinder.
m 396,
may
be deduced from the formulse (396, (5.), &c.) of the recent
theory of emanqnts, which give,
XXVI.
= Vi;V' = r-ir,
wo = |0,
if
'-,
>7
= v,
its envelope, coinwhile the line of pontact (396, (14.)), of the emanant jo?arae
cides in position with the tangent to the curve; in passing, then, from the given
with
the binormal (j/) and the osculating plane (4- v) have

round that tangent (r) as a common axis, through a
small angle =x~^s, and therefore with a velocity =r~i, if this symbol have the value
assigned by III., or by the following extended expression, in which the scalar vapoint
to the
near point
Ps,
(nearly) revolved together,
riable (t) is arbitrary
XXVII.
(comp. 395, (11.), &c.),
r-i
=S
-^, = S rrT = Second
Vp p
while the binormal has at the
same time been translated
perpendicular to the tangent r, through the small interval

sent order of approximation) represents the small
(2.)
As an example,
XXVIII.
.. pt
if
we
Curvature
Vdpd2p
(nearly), in a direction
is
= st,
which
(in the pre-
chord pp^.
take this new form of the equation of the helix,
= b(at cot a -f </3),
with
Ta = T;3 = 1,
and
Sa]3
= 0,
which gives the derived vectors,
XXIX.
and
XXX.
we
p/
= ba (cot a + e"'^),
this expression for the arc s
easily find (after
s= s't,
pi'
= - b^%
(supposed to begin with
where
s'
pi"
= api',
<),
Tp' = 6 cosec a = const.,
a few reductions) the following values for the two curvature^
* This convenient
appellation (of twisted cubic) has been proposed by Dr. Salsee pages 241, &c,, of his
for a curve of the kind here considered
already cited
Treatise.
The osculating twisted cubic will be considered somewhat later.
mon,
t This theorem was estabhshed, on sufficient grounds, in the cited Memoir of M.

Venant (page 26) but it has also been otherwise deduced by M. Serret,
de Saint
m the
Additions to
M.
Liouville's Edition of
Monge (Pans,
1850, page 561, &c.).
564
XXXI.
r-1
= 6-1
sia2 a,
= fti
r->
sin
[bOOK
111.
a cos a;
while the common centre (395), of the osculating circle and sphere, has
now
for its
vector (comp. 389, (3.)),
XXXIP.
= = pt- be^^^lS cosec2 a = 6 cot a {at - oj8 cot a)
ff
6 being here the radius of the cylinder, but a denoting
still
the constant inclination
of the tangent (p') to the axis (a).
ThQ
(3.)
line
rectifying line (396), considered merely as to it^ position, being the
of contact of the
sented
by the
XXXIII.
XIV.) with
rectifying plane (396,
its
own
envelope,
is
repre-
equations,
= Sr'(a> - p) = Sr"(w - p)y
with the signification XVI. of \

as emanants, or
change
to r',
t]
XXXIII'.
or
= VX(a> - p),
and accordingly, if we treat the rectifying planes

we find the value d=Vr"r'-^ =X, which shows also
;
that in the passage from p to Pj the rectifying plane turns (nearly) round the rectify= sT\, or with a velocity of rotation represented by
ing line, through a small angle
the tensor,
XXXIV.
T\ = V (r-2 + r-2) = r'l cosec S"= r"!
sec
H;
what we have
called the rectifying vector, X, coincides in fact (by the general

theory of emanants) with the vector axis (396, (14.)) of this rotation ôf the rectify-
so that
ing plane : as the vector of second curvature (r-'r) has been seen to be, in the same
full sense (comp. (1.)), the vector axis of rotation of the osculating plane, when velocity, direction,
(4.)
and position are
all
the derivative
of the arc
When
s'
taken into account.
unity (comp. 395, (12.)), the expression
general form,
XXXV.
X = V4-7, = V
and accordingly
only constant, without being equal to

be put under this slightly more
= Rectifying
Vector;
asfip
we have thus
for the helix (2.)
XXXVI.
is
XVI. may
the values,
= as'-^ = a6'i sin a = ar-^ cosec a,
UX = a
the rectifying line is therefore, for this curve, parallel to the axis, and coincides with
the generating line of the cylinder, as is otherwise evident from geometry.
The
value,
TX = 6"*
sin a, of the velocity
of rotation of the rectifying plane, which
here the tangent plane to the cylinder,
motion along the curve,

give, for the
is
when compared with a
also easily interpreted;
same helix (by XXXI.), the
XXXVII.
r'
is
conceived velocity of
and the formulae XVII., XVIII.
values,
= 0,
P = 0.
H=a,
The normal
(5.)
(or the radius of curvature), as being perpendicular to the
rectifying plane, revolves with the same velocity, and round a parallel line ; to de-
termine the position of which
we have only
to
change
r}
new
XXXVIIL
OH =
=P+
=
or the point h in which it cuts the normal,

formula 396, XXXII., which then becomes,
line,
to r' in the
o;o
r-2 Qic
pcos-
=p -
r'S -^,
p)
r-p
H+
AT
+ T^K
Ksin^H;
X'^t'
CHAP.
AXIS OF DISPLACEMENT OF NORMAL.
III.]
565
and
the vector of rotation (396, (9.)) of the normal is therefore a line
divides (internally) the radius (r) of curvature into the two segments,*
||
XXXIX.
fl = r sin2 H,
which are proportional
namely, into segments
HK = r cos2
=X, which
H-,
to the squares (r-2
and
r"2) of
the j^rs<
and second curvatures.

(6.)
of an
At the same
emanant
line
time,
what we have
called generally the vector
of translation
becomes, for the normal (by 396, (10.), changing 9 to X), the
line
XL.
= XS - = UX cos H=: r"'X"',
set off
from the same point
and the indefinite right line, or axis, through that point h,
XLI.
= VX (w - Wo),
XLI'.
or
= VX (w - p
along which axis the normal moves, through the small
cos2
^ - sin2 ^),
line si, while it turns
round
the same axis (as before) through the small angle sTX, may be called (comp. again
39,6, (10.)) the Axis of Displacement of the Normal (or of the radius of cm*vature).
(7.)
As a
verification, for the helix (2.)
XLII. ..PH =
(1)0- pt
6,
we have
thus the values,
hi<**^ = hat cot a,
= acosa;
so that the axis of displacement (6.) coincides with the axis (a) of the cylinder, as
was
of course to be expected.
(8.)
When
and XL., of
is not a helix, the values VI., XVI., XXXVIII.,

enable us to put the expression I. for p* under the
the given curve
r", X, wo,
and
t,
form,
XLIII.
ps= W0 + SI +
1^'^
(p
- Wo)
or
by this last small vector of the third order,

namely by that part of the term ^s^t" which has the direction of the normal t', or of
r', and which depends on r',from the osculating helix,
the curve therefore generally deviates,
XLIV,
and from the osculating right
XLV.
whereon that helix
while
its
is
traced,
= wo + SI +
^(p
Wo),
cylinder,
.
TVX(w - wo) = sin H,
and of which the rectifying
axis of revolution (comp. (7.))
is
line
(XXXIII.)
is
side,
the axis of displacement (XLI.) of the
normal.
(9.)
curve,
is
Another general transformation, of the expression

had by the substitution,
<2r'
XLVI.... = +
in
which
< is
a new scalar variable
I. for
the vector of the
<3
-+-,
for this gives the
new
form.
* This law of division of a radius of curvature into segments, by the common

its consecutive, has been otherwise deduced by
perpendicular to that radius and to
M.
de Saint Venant, in the
Memoir already
referred to.
566
[bOOK
III.
XLVII. ..pt = p + tT+it^lT' + '^\^ ith-^v,

and therefore shows that the curve
deviates,
by
order,
XLVIII.
^th-^v
this other small vector of the third
= }s^r-^rT',
the part of the term }s^t" which has the direction of the hinormal v, and
is, by
which depends on r, from what we propose to call the Osculating Parabola, namely
that
new
that
auxiliary curve of which the equation
XLIX.
or from the parabola
given curve on
its
a)t=^p
which osculates
own
is,
+ tT + it^
{-i}
at the given point p, to the projection of the
osculating plane.
(10.) And because the small deviation XLVIII. of the curve from the parabola
is also the deviation of the same curve from this last plane, if we conceive that a
near point Q of the curve is projected into three new points Qi, Q2, Q3, on the tangent, normal,
and binormal
L.
respectively,
shall
= r-i =
lim.
we
have the Umiting equation,
Second Ourvature
PQ1.PQ2
the sign of this scalar quotient being determined
(11.)
But we may aIso(comp. 396,
formation of
I.,
analogous to the forms XLIII. and

LI.
with the value
by the
(17.), (18.))
XI. of
v'
sum
is
of the two first terms gives the vector of
distant from the given point pp
of that circle equal to the arc s of the given curve

.
XL VII.,
.p.^K^l'îp-K^^-v'T,
in which the
the point of the osculating circle, which
LIL
rules of quaternions.
employ this Âird^'encraZ ^rans-
i3^V = As3(r" + r-2r)
by an arc
and the third term,
= --is3r-irV' + is3j.-i^^
which represents the deviation from the same circle, measured in a direction (comp.
IX. or X.) tangential to the osculating sphere, is (as we see) the vector sum of two
rectangular components, which represent respectively the deviations of the curve,
(8.), and from the osculating parabola (9.).
from the osculating helix
(12.) It follows, then, that although neither helix nor parabola has in general
complete contact of the third order with a given curve in space, since the deviation
from each is generally a small vector of that (third) order, yet each of these two
auxiliary curves, one on a right cylinder XLV., and the other on the osculatmg
plane, approaches in general more closely to the given curve, than does the osculating
circle : while circle, helix, and parabola have, all three, complete contact of the second* order with the curve, and with each other.
* It
appears then that we may say that the helix and parabola have each a con
which is intermediate between the second and third or-
tact with the curve in space,
ders
or that the exponent of the order of each contact
But
it
must be
left to
is
the fractional index, 2|.
mathematicians to judge, whether this phraseology can pro-
perly be adopted.
CHAP.
(13.)
SECTION OF SURFACE OF TANGENTS.
III.]
As regards
t,
in
T{(l
l+0 + il...,
;^).+ <rj
therefore (to the present order of approximation),
LIV.
Arc of Osculating Parabola (from wo
Ti^'tdt
= ^ +_-+-_ = s
or
Jo
Arc
if
scalar,
of the parabola, that equation gives,
LIII.,.T.-, =
and
new variable
the geometrical signijication of the
XLIX.
the equation
567
=s
then an arc
br^
(by
of Curve in Space (from po
be thus set
to
XL VI.)
to
p^
same
the parabola, with the
dSupon
o)t)
initial
point p,
parabolic arc, or its chord tJt (Oq, be obHquely projected on the initial tangent r, by drawing a diameter of the parabola
through its final point, the oblique tangential projection so obtained will be =*r by
XLIX. ; and its length, or the ordinate to that diameter, will be the scalar t.
and the same
(14.)
and
initial direction,
And
if this
as regards the direction of the diameter of the osculating parabola,
drawn as we may suppose from p, if we denote for a moment by

to the normal + t, regarded as positive when towards the tangent +
XLIX. and XVIII.) the formula,
LV.
its
r,
inclination
we have (by
tanZ)=- = ÂtanPcoti7:
an instance of the reducibility, above mentioned, oiall affections of the curve

and P.
depending on s^, to a dependence on the two angles,
which
is
(15.) Some
of these affections, besides the direction of the rectifying line X, can
be deduced from the angle
H alone.
tor equation of the surface
of tangents
LVI.
in
which
and
As an example, we may
Ws,
are two independent
= ps -f tp's =
and
terms depending on
LVIII.
which
is,
s* in pj.
.
Sr(a>
If then
we
and
sr'+^r",
cut this developable
- p) = c= any
relatively to the surface, a
ps-\- tTs,
scalar variables,
LVIL..rs=r +
+
observe that the vec-
of the form,
is
LVI. by the plane,
given scalar constant,
normal plane at the extremity of the tangen-
is also a generating line, we get thus a.

principal* normal section, of which the variable vector has for its approximate expres-
tial
vector CT from p, while this tangent
sion,
LIX.
.. b>s
= {p^-cT)^(cs +. .) t' + (^cs2r-i +
.) j/
the terms suppressed being of higher orders than the terms retained, and having no
find then thus, that the vector
influence on the curvature of the section.
of the
We
centre of the osculating circle to this
given curve
is,
normal section of the surface of tangents
to the
rigorously,
Some general acquaintance with the known theory oi sections of surfaces

here supposed, although that subject will soon be briefly treated by quaternions.
is
568
LX.
= p + c(r+rv) = p + crX
p-\-cr+
we make
= r,
III.
XXXI IT'. And
so that the locus of all such centres is the rectifying line

cular,
[bOOK
if,
in parti-
or cut the developable at the extremity of the tangential vec-
tor rr, the expression
LX. becomes then p-rrT + rUr
which expresses that the
radius of the circle of curvature of this normal section of the surface is precisely
what has been called the Radius (r) of Second Curvature, of the given curve in
But this radius (r = r tan H) depends only on the angle 21, when the radius
space.
(r) of (absolute) curvature
is
given, or has been previously determined.
The cone of the second order, represented by the quaternion

= 2rSr (w - p) Sv (w - p) + (Vr (a> - p))2,
LXI.
(16.)
equation,
its vertex at the given point p, and rests upon the circle last determined ; it is
then the locus oi all the circles lately mentioned (15.), and is therefore (in a known
has
sense) an osculating oblique cone to the developable surface of tangents : its cyclic
normals (comp. 357, &c.) being r and r + 2rv, or t and rr + 2v\Jv. But, by 394,
cone to this cone LXI., and therefore also (in a sense
(30.), the osculating right
likewise
known)
of centres
to the surface of tangents itself
(15.),
namely the rectifying
tangent (r) to the curve
is
is
one which has the recent locus
line (\), for its axis
one of its sides
its
semiangle
of the quaternion equation of this osculating right cone
LXII.
is
is
of revolution, while the

= S, and a form
therefore
the following (comp.
XLV.),
TVU\(a> - p) = sin H.
(17.) The right cone LXII., which thus osculates to the developable sur/ace of
tangents LVI., along the given tangent r, osculates also along that tangential line
to the cone of parallels to tangents, which has its vertex at the given point P ; as is
at once seen (comp. 394, (30.)), by changing p' and p" to r' and r", in the general
expression Yp'p" (393, (6.), or 394, (6.)), for a line in the direction of the axis of
the osculating circle to a curve upon a sphere. And the axis of the right cone thus
determined, namely (again) the rectifying line (X), intersects the plane of the great
which is parallel to the osculating plane, in a point
circle of the osculating sphere,
L of which the vector
is,
LXIII.
OL = p + rpX = p + rr'r + rpv.
We
have thus, in general, a gauche quadrilateral, pksl, right-angled ex(18.)

cept at L, with the help of which one figure all affections of the curve, not depending
on s*, can be geometrically represented or constructed: although it must be observed
that
when r'=
0,
then itself a. great

points
and
s,
which happens
for the helix
(XXXVII.), the
circle of the osculating sphere,
osculating circle
and the points p and
is
L, like the
coincide.
it may assist the conceptions to suppose lines set

(19.) In the general case,
from the given point p, on the tangent and binormal, as follows
off,
LXIV.
PT = BL =
rr'r
pb = tl = ks = rpv
have a right triangular prism, with the two right-angled triangles,

TPK and LBS, in the osculating plane and in the parallel plane (17.), for two of its
the two
faces, while the three others are the rectangles, pksb, pblt, kslt, whereof
for thus
first
we
shall
are situated respectively in the normal and rectifying planes.
CHAP.
III.]
INCLINATIONS DETERMINED BY QUATERNIONS. 569
AH
scalar properties of this auxiliary

(20.)
neral methods, from the three scalars, r, r, r', or
may
prism
H,
r,
be deduced, by our ge-
and
all vector properties
same prism can in like manner be deduced from the three vectors r, r', r", or
from r, r, v', which (as we have seen) are not entirely arbitrary, but are subject to
of the
certain conditions.
(21.) As an example of such deduction (compare the annexed Figure 81), the
equation of the diagonal plane spl, which contains the radius
(R) of spherical curvature and the rectifying line (\), and V
of the trace, say pu, of that plane on the oscuwhich trace is evidently parallel (by the con-
the equation
lating plane,
tk
struction) to the edges ls,
of the prism, are in the recent
notations (comp. XX.),
LXV.
ith
= Sr"(w - |o)
LXVI.
= V(r-lr)'
(w
- p)
rSrV" = r'Srr"= r^r', by II.

by 204, (22.), if a and (3 be any two
"^
the verification that
(22.) In general,
actors,
we have the
LXVII.
Fig. 81.
= tan^^ =
^
a
.tan
expressions,
tan
Z.
/3a
= - tan
Z.
a/3
(3
(TV: S)
a/3,
the angles of quaternions here considered being supposed as usual (comp. 130) to be
> 0,
generally
LXVIII.
but
< tt
example, we have thus.
for
-= (TV: S) \r-i = (TV:
tanir=tan L
as in
XVII.
and
in like
manner we have
S) (r-i
by
generally,
r')
= rTr' = rr-,
principles already ex-
plained (comp. 196, XVI.),
LXIX.
cos
Z.
- = cos
a
Z.
= - cos Z
/3a
= - cos
Z.
ai3
(3
= S^:T^=SU^=-SUai3.
a
(23.) Applying these principles to investigate the inclinations of the vector r",
is perpendicular to the diagonal plane LXV. of the prism, to the three
which
rectangular lines
r, r', v, or
the inclinations of that diagonal plane itself to the nor-
mal, rectifying, and osculating planes, with the help of the expressions deduced from
VI. for the three products,* rr", t't", vt", we arrive easily at the following results :
*
to
A student,
who
should be inclined to pursue this subject, might find

all the binary products of the nine vectors,
it
useful
form for himself a table of

T,
T, t", V, V, X, o-^')
a+6r +
<T
/u,
and
k',
many quaternions, and reduced to the common quadrinomial form,

cr'+ej/, in which a, 6, c, e are scalars, whereof some may vanish, but
considered as so
which are generally functions of
r, r,
and
r'.
4 D
570
r-2
r"
with the
sum
verification, that the
cause
LXXI.
LXXI'.
or
Or we may
(24.)
LXXII...tan^- =
r
r~h-^
r"
r-^r
of the squares of these three cosines
is
unity, be-
+ r-2i?2) = V(l + r'2 + r-^-r'^)

= V(r-2r'8 + T\), Tr" = V(r-4 + Tv').
r2TT"= V(l
rXr"
t"
III.
[bOOK
same general
write, on the
^;
Tr
plan,
tan^^=^^^;
r
r
- = - V(l + r'2);
tan Z
or
LXXIII.
tan ^ tt" =
UTr^
tan L
rV" = rr'-iT\
vr"=- rr-i V(l + r'2)
tan Z
and P,
and may modify the expressions, by introducing the auxiliary angles
with which may be combined, if we think fit, the following angle of the prism,
LXXIV,
PKT = BSL = tan-ir'.
(25.) Instead of thus comparing the plane spl with the three rectangular planes
(379, (5.)) of the construction, we may inquire what is the value of the angle spl,
which the radius (R)

we find, on the same
of spherical curvature
makes with the
rectifying line (\)
and
by quaternions, the following very simple expression for

which may however be deduced by spherical trigonometry
plan,
the cosine of this angle,

also.
LXXV.
cos SPL
pr =
= - SUX((T - p) = sin P sin if
ill
LXXV.
or
cos spl = cos spb
A.
cos bpl.
(26.) In general, it is easy to form, by methods already explained, the quaternion equation of a cone which has a given vertex, and rests on a given curve in space ;
and
also to determine the right cone
any given side of it.

(27.) But if we merely wish
which osculates (394, (30.))
to this general
cone, along
chords from
may
to assign the osculating right cone to the cone
p, or to the Iocjjs of the line pps,
observe that
if
this
we may
imitate a recent process
new cone be cut by the normal plane LVIIL, the
of
and
vector of
the section has the following approximate expression, analogous to LIX., and like
suflScient for our purpose,
LXXVI.
from which
new
it
may
.ujs
= p + CT + IcsT + \cs'^v-^v
be inferred (comp. (15.), (16.)), that the
it
aocis
of revolution of thQ
right cone has for equation,
LXXVII.
This axis
is
= V(r-ir + fv)
(w
- p).
therefore situated in the rectifying plane^ between the rectifying line
the
(\ or r-ir + v), and the tangential vector (IV.) of second curvature (r'^r) while
from -f r towards + v) has the
semiangle C of the same new cone (measured like
value already assigned by anticipation in the formula XXV., and is therefore less
:
if both be acute, but greater than

if both be obtuse; so that,
than the semiangle
in each case, the new right cone (C) is sharper than the old right cone (^H).
uni
(28.) The same result may be otherwise obtained, by observing that an
CHAP.
III.]
INTERSECTION OF OSCULATING SURFACES.

XLIV., and 397,
vector in the direction of the chord pp has (by 396,
571
I.) the
ap-
proximate expression,
LXXVIII...x.= U(p,-p) =
(l
j^)(r "|+?|:)
whence the axis of the osculating right cone to the cone of chords (27.) has rigorously
the direction of the line V;^'x" (for s
LXXIX.
= 0),
^ = Vr'(r2r" + ^r)
or of the vector,
=X-iv = r-ir + |j/,
as before.
(29.) This axis ^ makes (if we neglect s^) the same angle C, with the chord
whereas Hlq former axis X makes unequal angles with
PP,, as with the tangent r
;
same order
those two lines, within the
for our
methods
easily be
deduced
(or degree) of approximation
conduct to the expression,
24rr
XXV., between the two right cones,
from which the relation
may
anew.
(30.) Neglecting only s*, and employing the substitution
for the vector of the given curve becomes,
XL VI., the expression
XLVII.
LXXXL..pt = p+tr + it^v-]-lth-^v,

where the variable scalar
bola,
t denotes, by
and the constant vector v has the
if
LXXXIL
(13.), the ordinate

direction,
by
= r'+
3r
of the osculating para-
(14.), of the diameter of that
parabola.
(31.) In the present order of approximation, then, the proposed cwrt-e J7i space
the common intersection of the three following surfaces of the
second order, all passing through the given point P :
may be considered to be
LXXXIIL
2(Sr'(w-/)))2=3rSi'Ca>-p)Syv(a>-p);
=
LXXXIV.
2Sr'(fe'-p) -r2(SuK<^-p))3;
LXXXV.
3rSv (w - p) = - r'-Sr\(o - p) Svv^co - p)
.
whereof the frst represents a new osculating oblique cone, which has a contact of the
same {second) order with the cone of chords, as the osculating right cone (27.) the
;
second represents an osculating parabolic cylinder, which

the osculating parabola (9.),
by the osculating plane
is
to the
cut perpendicularly in
curve
and the third
represents a certain osculating hyperbolic (or ruled) paraboloid, whereof the tan-
gent (r) is one of the generating

rabola is another.
lines,
while the diameter (y) of the osculating pa-
(32.) Each of these three surfaces (31.) has in fact generally a contact of the
third order with the given curve; or has its equation satisfied, not only (as is ob-
by the point P itself, but also when we derivate successively

with respect to the scalar variable t, and then substitute the values (comp. LXXXL),
vious on inspection)
LXXXVI.
r, r, p, r',
v,
w = po = P>
w'
and v being treated
each of these derivations,
= po' = r,
a>"
= po" = v,
w"'
= po'" = t^v;
as constants of the equation, or of the

surface, in
572
The cone LXXXTII., and the cylinder LXXXIV., have a common geneand in like manner, another generating line of the
(33.)
ratrix,
same
line
III.
[boOK
namely the hinormal* (v)
cone,
on the
namely the tangent (r) to the curve, has just been seen (31.) to be a
and although the cylinder and paraboloid have
paraboloid LXXXV.
:
no finitely distant right

finity, in the
line
common, yet each may be said
which plane the quaternion equation
LXXXVII.
to contain the line at in-
diametral plane of the cylinder, namely in the plane of v and
= Sry (o) - p),
is
or
v,
of
(comp. (14.)),
LXXXVII'.
= S(rrV' - 3r)
(w
- p)
or the line in which this diametral meets the parallel axial plane.
(34.) On the whole, then, it is clear, from the known theorj" of intersections of
common generating line, that the given curve of

double curvature (whatever it may be) has contact of the third order with the twisted
cubic,f or gauche curve of the third degree, which is represented without ambiguity
by the system of the two scalar equations,
surfaces of the second order having a
LXXXVIII. ..y = a;2,

if
we
= x^,
LXXXIX.
(35.)
theory,
it
fx=(t =)-r2Suv(w-p),
= (t^=)- 2r2Sr'(w - p),
(z=(<3=)_6r2rSv(w-p).
)y
As another
may
geometrical connexion between the elements of the present

be observed that while the osculating plane to the curve, of which
plane the equation
is,
XC.
touches the oblique cone
Sr((u-p) = 0,
as in 396,
XV.,
LXXXIII., along the tangent t to the same curve,
the cfmrnf-
LXXXVII.
touches the same cone along the binormalv, which was lately
seen (33.) to be, as well as r, a side of that oblique cone ; but these two sides of
contact, T and v, are both in the rectifying plane (396, XIV.), and the two tangent
fraZ/)?ane
planes corresponding intersect in the diameter v of the parabola (9.)

therefore this theorem :
The diameter of the osculating parabola
to
we have
a curve of double curvature
is the
polar of the rectifying plane, with respect to the osculating oblique cone LXXXIII.
that is, with respect to a certain cone of the second order, which has been above de;
duced from the expression LXXXI. for the vector p< of the curve, as one naturally
suggested thereby, and as having a contact of the third order with the curve at p,
* The
geometrical reason, for the osculating cone LXXXIII. to the cone of chords
containing the binormal (v), is that if the expression LXXXI. for pt were rigorous,
and if the variable t were supposed to increase indefinitely, the ultimate direction of
the chord pp* would be perpendicular to the osculating plane.
And the same binormal is a generating line of the parabolic cylinder also, because that cylinder passes
through
p,
and
It is sufficient
all its
generating lines are perpendicular to the last mentioned plane.

to observe, on the side of calculation, that the
equations
however
LXXXIII. and LXXXIV.
are satisfied,
when we suppose
oj
p\\v.
t Compare again page 241, already cited, of Dr. Salmon's Treatise

285, in page 225 of the same work.
also
Ark]
CHAP.
TANGENT TO LOCUS OF CENTRES.
III.]
second order with the cone of chords from that
also a contact of the
and therefore
573
point.
(36.) Conversely, this particular cone LXXXIII. \& geometrically distinguished

all other* cones of the same (second) order, which have their vertices at the
from
the given
given poirit p, and have each a contact of the same second order, with
cone of chords from that point, or of the third order with the given curve, by the
condition that it is touched (as above), along the hinormal (v), by the diametral
plane (vv) of the osculating parabolic cylinder LXXXIV.
the simultaneous variations of
(37.) We have already considered, in 395, (5.),
the points p and K, or of the vectors p and k.

p, we have the following
expression fj, = 2K
Jirst derivative of the latter vector,
and
"With recent notations, including the

other transformations, for the
among
therefore for the tangent
centres of curvature, of a given curve in space
XCI.
kk'
to the locus
of
- r'-i)' = r + T'-'rV -i
(p
= (p 4- r^r')' = T + r^T"+ Irr'r
= rr'r + rV^ v = rr (r -\-p~^rv) = rr-' (pr' + rv)
kk'= BsK = k' =
rr
rr'
p-K
(T-K
rr'(a
u)
c = r'^(<y-/^)7\
((T-k)(k-P)
,
= cot H(Ut' tan P f Uv) = r-i7?(Ur' sin P + Vv cos P)

= r^j/v'r' = r*T'v'v v'^v't'^ = t''Wv~^
= x-^v{p-G) (K-|t?) = r-i(K-p) (p-<t)v
= r-'i2U(x/(p -
(t) (/c
p))
= &c.
the diameter of curvature pm, and let fall

a perpendicular kn from the centre k of the osculating circle on the new radius sm of the osculating sphere (as in the
if
then
we draw
annexed Figure 82), this perpendicular will touchf the locus of the centre k, a result which agrees with the construcand we see, at the same time, that the
tion in 395, (6.)
;
length of the line kk', or the tensor Tk',
may
be expressed
(comp. LXXIII.) as follows,
XCII.
kk' = Tk' =
= r^Tv' = tan Z
i?Tr-i
rr".
\,.
.--''
project the tangent kk', into its two rectFig. 82.
angular components, kk, and kk on the diameter of curvature and the polar axis, we shall have by XCI. the expressions
(38.) If
we
* The cone of
this system (36.),
normal plane, and which therefore
which
is
touched along the binormal by the
intersects the parabolic cylinder
LXXXIV.
in
twisted cubic (comp. (34.)), having also contact of the third order with the
the following
curve, is easily found to have, for its quaternion equation,
2r2 (Sr'(w - p))2 = 3rSr (w - p) Sr (w - p) ;
new
and with respect to this cone (comp. (35.)), thQ polar of the rectifying plane is the
(absolute) normal (r') to the curve.
t Geometrically, and by infinitesimals, if we conceive k' to be an infinitely near
point of the locus of k,
PKs)
will be right,
this semicircle
drawn
the tangent kk'
is
and therefore
and the point k'

to
in the
will be
(comp. Fig. 82)
normal plane at
p,
the angle pk's (like

but the radius of
;
on the semicircle pks

is
parallel to the line sm, to which line
therefore perpendicular, as above.
574
XCTII.
XCIV.
KK,
rr'r'
r'Ur'
and
3,
or
r',
H and P,
but not on
(39.)
Sec.
p-K
^^ = &c.
KK = r2rV = rr-iUv =
these two projections then, or the vector-tangent
-^ =
[bOOK
and thereby
a-K
kk'
itself,
a/Z the affections of the
III.
would
suffice to
determine
curve which depend on
s*.
We have also the similar

XCV.
triangles (see again Fig. 82),
A K k'k oc k kk' a kms
and the vector equations,
XCVI.
kk'
= r-V =
whence
SM = KK, sk = kk' Kai = kk
Vector of second curvature (IV.)
pk
also result the scalar expressions,
XCVII.
tan KSK, = tan
kpk = r-' = Second*
Curvature (III.)
this last scalar being positive or negative, according as the rotation
ksk, (or kpk')

appears to be positive or negative, when seen from that side of the normal plane,
towards which the conceived motion (396, (1.)) along the given curve, or the unit
tangent
+ r,
is directed. f
(40.) Besides the seven expressions, III., XXVII., L., and XCVII., this important scalar r"> admits of many others, of which the following, numbered for reference
as 8, 9, &c.,
and deduced from formulae and
amples; and
may
principles already laid down, are exserve as exercises in transformation, according to the rules of the
present Calculus, while
some of them may
also be found useful, in future geometrical
applications.
(41.)
XCVIII.
We
.
have then (among others) the transformations
Second Curvature
= p-'r' = r-i cot H= T\ cos
r"i
(=
seven preceding expressions)
H = r'lr
cot
CHAP.
EXPRESSIONS FOR THE SECOND CURVATURE.
III.]
TE-i
rr'v
rr'r'
a-K
(a-K)T
^_
tan L xtt"
= iJi
tan L
T(a-p)
rr'+pif
((T-k)(p-k)
r(K p)
rr'r
c- k
(<'-)(?-)
r'
575
(40, 41, 42, 43)
(44, 45, 46, 47)
= S y^
r
7
((T-ic) (p-jc)
=S
(48,> 49,I
\
KS.KP
60)y
-dcosZ a
(Sari^)
(51, 62, 53)
rd cos
Z. -
PKSL, in the forms 50 and 51, being points of the same gauche quadrilateral as in
(18.)
and
a, in
form
varieties of
52 and 53,* denoting any constant vector : -while several other

be deduced from the foregoing by very simple processes, such
may
as the substitution of
Uv for rv,
which gives
&c.,
for instance
(comp. XI'.), from the
form 38, these others,
XCYIir.
"We
r.
= llM' = -J^^ = -_^

rdr
TT
Ur
also write, with the significations (10.) of Qi
may
sion analogous to L.,
XCVIIl".
r-i
= GKP.lim.
and
(54, 65, 66).
Q3, the following
expres-
(67),
which contains the law of the inflexion of the plane curve, into which the proposed
curve of double curvature is projected, on its own rectifying plane : the sign of the
scalar, to
which
this last expression ultimately reduces itself, being
determined by
the rules of quaternions.

(42.)
And
besides the various expressions for the positive scalar r"^,
which are
immediately obtained by squaring the foregoing forms, the following are a few
others
Square of Second Curvature = r-2 = Tr-2

= T\2 - r-2 = r2Sr"r'\ - r-2 = r2Ti/'2 - r-2r'2
= r^Srv'r" - r-^r'i = r^Tr"^ - r-2 - r-2?'2 = i2-2 (r<Tr"2 - 1)
= R-^riTv'^ = 22-2T/c'2 = ij 2 tan2 i tt"
XCIX.
while the important vector r", besides
lowing
C.
among
.
its
other expressions (comp.
two
(1, 2,
(7, 8, 9)
XX. XXI.)
r"=D^p (= the two expressions VI.)
=
(43.)
it
original forms VI., admits of the fol-
= r-2VX((T - p) = \t - rVr = v't - r-^r

= rVv'X = r-2r-ir ((t - p - r) = r-^p + r-2X (a - p)
curves,
3)
(4, 5, 6)
As
((p-K)-i)' = r'(K'-r)r' =
(6, 7, 8)
-r-2r_^^--^
(9, 10, 11).
regards the general theory (396, (5.), &c.) of emanant lines
might have been observed that
* This
(3, 4, 5)
last
to in the second
form 53 corresponds
Note to page 563.
to
if
we
{r])
from
write,
and contams a theorem
of
M.
SeiTet, alluded
576
CI.
^=V
CII.
Avith
^,
=V
[bOOK
-, as in 396,
III.
XXVIL,
t]
XXXII. takes the simplified form,

PH = wo- p = JjSrj'^K = projection of vector on emanant rj
= r"Jr, and ^ = 0, ph = 0, orwo = p, asin(l.);
for example, when t] = v, then
so that the projection ph again vanishes,
and when n = r, then 9 = v, = r^r' -^
the equation 396,
cm.
Z,
t^
ij,
as in 396, (13.).
(44.) In an extensive class of applications, the emanant lines are perpendicular
and since we have, by (43.),
to the given curve (t? ->- r)
;
we may
write, for
tliis
case of normal emanation, the formula,
of vector of curvature
_ y _ projection
p_^
(r')
on emanant
line (tj)
square of velocity (T^) of rotation of that emanant

for example,
= X,
then
when
the emanant
as in (3.),
CVI.
(jj)
coincides with the absolute
normal
(r'),
we have
and the recent formula CV. becomes,
PH = oio - p =
^ = r'T\-2
= rV sm2 H={k-q) sia2 H,
which agrees with the expression XXXVIII.
And
emanant planes, hy making

passing to a second derived
equation, we find for the intercept between the point p of the curve, and the point,
say R, in which the litie of contact of the plane with its own envelope touches the
(45.)
Sr;
in the corresponding case o^ tangential
XXXVI., and
in the second equation 396,
cusp-edge of that developable surface, the expression,
evil.
PK =
-r-;r- =
SrjT]
which accordingly vanishes, as
p) =
S??(aj
plane
(46.)
of a curve
tions
Some
it
T]
:p
n
:r^,
projection of Tj on d
ought to do, when
= v,
r]
coincides with the osculating plane
additional light
tn space depending
may be thrown on
this
that
is,
when
the emanant
XC.
whole theory, of the affections

arc, and even on those affec-
on the third power of the
which depend on higher powers of s, by that conception of an auxiliary sphewhich was employed in 379, (6.) and (7.), to supply constructions (or
rical curve,
geometrical representations) for the directions, not only of the tangent (p') to the
given curve, to which indeed the unit-vector (r) of the new curve is parallel, but
also of the absolute normal, the binormal, and the osculating plane ; while the same
auxiliary curve served also, in 389, (2.), to furnish a measure of the curvature of
the original curve, which is in fact the velocity* of motion in the new or spherical
curve, if that in the old or given one be supposed to be constant,
and be taken
for
unity.
Accordingly the vector of velocity t, of this conceived motion in the auxiliary

what we have called (389, (4.), comp. 396, VI.) the vector of cur-
curve, is precisely
vature of the proposed curve in space

of the radius (r) of that curvature.
and
its
tensor (Tr')
is
equal to the reciprocal
CHAP.
III.]
CASE OF CONSTANT RATIO OF CURVATURES.
577
We
might for instance have observed, that while the normal plane to tlie
(47.)
curve in space is represented (in direction) by the tangent plane to the sphere, the
plane (as being perpendicular to the absolute normal) is represented simito the spherical curve : and it is not difficult to prove
that the rectifying line has the direction of that new radius of the sphere, which is
rectifying
larly
by the normal plane
drawn
to the point (say l)
where the normal arc
to the auxiliary
curve touches
its
own envelope.
(48.) The /Joi/i< L thus determined
is the common spherical centre

(comp. 394,
of curvature, of the auxiliary curve itself, and of that reciprocal* curve on the
same sphere, of which the radii have the directions (comp. 379, (7.)) of the hinor-
(5.))
mals to the original curve; the trigonometric tangent of the arcital radius of curvais therefore ultimately equal to a small arc of that
curve,
ture of the auxiliary curve
divided by the corresponding arc of the reciprocal curve (or rather by the latter arc
with its direction reversed, if the point L fall between the two curves upon the
sphere)
and therefore to the first curvature
second curvature (r"i)
(r"') of the given curve, divided
by the
and thus we have not only a simple geometrical interpreta-
tion of the quaternion equation XI'., but also a geometrical
proof (which may be
important but known relation XVII., which

connects the ratio (r r) of the two curvatures, with the angle (^R) between the tansaid to require
no
calculation'), of the
:
gent (r) and the rectifying line (\), for

(49.) In whatever
manner
this
any curve
known
in space.
H = r:r) has once been esof the two curvatures be con-
relation (tan
tablished, it is geometrically evident, that if the ratio
because the cui-ve crosses the generating lines of its own rectifying developable (396) under a constant angle (H), that developable surface must be cylin-
stant, then,
drical
or in other words, the proposed curve
of double curvature must,
supposed, be a geodeticf on a cylinder (comp. 380, (4.)).
in the case
Accordingly the point
L,
becomes then a fixed point upon the sphere, and is the

common pole of two complementary small circles, to which the auxiliary spherical
curve (46.), and the reciprocal curve (48.), in the case here considered, reduce tliemin the
selves
two
last sub-articles,
so that the tangent
and the binormal
to the curve in
space make (in the
* The
reciprocity here spoken of, between these two spherical curves, is of that
known kind, in which each point of one is a pole of the great-circle tangent, at the
corresponaing point of the other : and accordingly, with our recent symbols, we have
not only v
= Ytt',
but
also,
Yvv' r'^Vv'v'^ =
r-^r-^r
r.
|1
t The writer has not happened to meet with the ^reome^rtca/proo/of this known
theorem, which is attributed to M. Bertrand by M. Liouville, in page 658 of the
already cited Additions to
Monge; but
the deduction of
it
as above, from the fun-
damental property (396) of the rectifying line, is sufficiently obvious, and appears to
have suggested the method employ ed by M. de Saint- Venant, in the part (p. 26) of his
Memoir sur
les lignes
courbes non planes, &c., before referred
to, in
which the result
is
Another, and perhaps even a simpler method, suggested by quaternions,

o{ geometrically establishing the same theorem, will be sketched in the present suba proof by the quaternion anaarticle
(49.); and in the following sub-article (50.),
enunciated.
lysis will
be given, which seems to leave nothing to be desired on the side of simpli-
city of calculation.
4 E
578
radius
case) constant angles, with the fixed
same
drawn
[bOOK
to that point
III.
and the curve
therefore (as before) a geodetic line, on some cylindrical surface.

itself
when the two curvatures have thus a constant ratio, the
(50.) By quaternions,
is
equations XI'. afld
XVI.
give,
(rXy = (Uv + rr-ir)' = (rr-i)V

r\ = a constant vector ;
CIX.
CVIII.
or
0,
the tangent (r) makes therefore, in this case, a constant angle ( JZ") with a constant
and the curve is thus seen again, by this very simple analysis, to be a
line (rX)
:
geodetic on
And
cylinder.
because
easy to prove (comp.
it is
XXXI.),
that v/e
have in the same case the expression,
ex.
or
<rf
the
secft'on
r sin^ If
of the cylinder
radius of curvature of base,
made by a
plane perpendicular to the generating
known theorem results, with which we shall conclude the present seWhen both the curvatures are constant, the curve is a geodetic
of sub-articles
lines, this
ries
other
on a right circular cylinder (or cylinder of revolution')

*
above, for simplicity and by eminence, a helix,
or
it is
what has been
called
398. When the fourth power (5*) of the arc is taken into account,
the expansion of the vector ps involves another term^ and takes the
form (comp. 397,

I.
in
I.)
../>,
= /> f 5T + I5V' f
I^^t''
+ i^sh'",
which
II.
7"'
so that the
= D,V,
new
on two new
III.
Stt'^'
affections of the curve,
scalars,
We must
and
such as
r'
be content to
and
=-
= - 3r^r'
3St't'^
thus introduced, depend only
and
r'\ or r'
R\
or
H' and
P',
here a very few remarks on the

of
such
and
on
the
manner in which it may be exaffections,
theory
tended by the introduction of derivatives of higher orders.
&c.
offer
* In
general, the expression XLIV. for the vector u)s of the osculating helix, in
r~iX"* = r X~ir', and p - wo = ^~^r', gives Tui's = 1 so that the deviaI =
which
tion (8.) maybe considered (comp. (13.)) to be measured from the extremity of an
arc of the helix, which is equal in length to the arc s of the curve, and is set oflf from
the same initial point p, with the same initial direction
note the value of
the formula
(Ds
answering to
XXXVIII.
It
may
while wo does not here de-
bat has a special signification assigned by

also be noted that the conception, referred to in
0,
(46.), of an auxiliary spherical curve, corresponds to the ideal substitution of the
motion of a point with a varying velocity upon a sphere, for a motion with an univelocity in space, in the investigation of the general properties oS curves of dou-
form
ble curvature:
and that thus
it is
intimately connected (comp. 379, (9)) with the
general theory of hodographs.
CHAP.
FOURTH POWER OF ARC, ROT. OF RADIUS
III.]
The new vector
(1.)
on which everything here depends,
r'",
is
the following forms,* analogous to the expressions 397, VI. for r"
easily reduced to
:
t'
= 3r-Vr + (r(r-i)"+\2)r*+
The
579
R.
(r-2r-i)V2i/.
derivatives of the four vectors, v, k\ X, a, taken in like

>vith respect to the arc s of the curve, are the following :
(2.)
V.
manner
first
(Vrr")' = Vrr"' + r-2\

= r-2r-ir + (r-r-i)'r'-i 4 (r(r- 1)" - r-2) v
k"= - r-irV + (rr" - r2r-2)r' + (r2r-)V
= (rr-i)V
\' = Cr-')V + (r-^yrv,
or VII'.
(r\)'
a ={k +prvy = (p' + rr-i)rj/ = RRp-^rv
v"=
Vr.
VII.
VIII.
in
which
last the scalar derivatives p'
by the equations,
IX.
and
X.
We have
i2'
R~^(pp'
and E! are determined, in terms of
p' =. Cr'r)' = r"r +
+ rr) =p' sin P+ r
cos
and
r',
r'r',
P= (/>' + cot R) sin P.
rV-r-ir'
^'^
=
r- ^^'"
r2
XII
'
rr\2
r2 -h r2
and the
r"
also the derivatives,
rr'-r'r
~ ^'^^ ^
^^^"
"*"
^^'^
-
'
+ p3
i22
relations,
XIII.
XIV.
XV.
.
which
(comp. 397,CVIII.)
may
SrrV"=Svr"' = -(r-2r-i)';
S/'r"' = - r-3r-2(p' - rrX2)
Srr"r"'=
.
SrV"r"'=r-2SXr"'
be proved in various ways, and by the two
which, the derivatives
r'
and />', and therefore
calculated, as scalar functions of the
of them, including the
We may
(3.)
new vector
also deduce,
= -r-5(rr-i)';
also
four vectors
first
H' and
(or the
two
last) of
P', can be separately
r, r', r", r'",
or of
some three
t'".
from either V. or VIII., the following vector expresis evident from the recent theory
(396,
which the geometrical signification

397) of emanant lines and planes :
sions, of
XVI.
. .
Vector of Rotation of Radius (^) of Spherical Curvature
Vector of Rotation of Tangent Plane to Osculating Sphere
= (say)0 = V^ = V^-^ = i2-2r(i/-i(T' + (r-p)

V
whence follows
<T
(1, 2, 3)
common angular velocity of these two conwith the velocity of motion along the curve,
this tensor value for the
nected rotations, compared
still
* In these new
expressions, on the plan of the second Note to page 561, the
r', p\ R', and the vector a', are to be regarded as of the dimension zero ; r",
H\ P', and k" of the dimension 1 X' of the dimension - 2 and v" and r'", as
scalars
being each of the dimension
- 3.
580
XVII.
Velocity of Rotation
= T0 = TV ^ =
i2-
of Radius (R), or of Tangent Plane
V(l +
i?'2
cot2
to Sphere^
P) = 7?-i V { 1 + (p' + cot Ey
which p = 0, />'= 0,
XVI. and XVII. become,
and XVir.
T0 = T\ = r"! cosec H,
with the verifications, for the case of the helir, for
R = r,
that these expressions
XVI'.
= \,
which agree with those found
cos"
P}
P= 0,
and
and velocity
before, for the vector
radius (r) of absolute curvature.

(4.) As another verification,
III.
[bOOK
we have R' =
for
of rotation of the
every spherical curve, and the
general expressions take then the forms,
XVI".
^,
XVII".
and
<T-p
of which the interpretation
(5.)
is
easy.
In general, the formula XVII.
XVIII.
i22^2 +
XIX.
or thus,
may
= - ij'2 C0t2 P= i2'2 -p-2R2R'2 =
i2'2 + ^'2 = (r'2 cos2 P

= V(l + T<r'2 cos2 P) = V(l + Ta'^ - R'i)
= V(i22 _ r2(7'2) = ^(^rz + r^Tcr"^)
E'-T:<p
i?T0
XX.
or finally,
T0 = R-\
so that the stnall angle, sT0, between the two near radii of spherical curvature,
ultimately equal to the square root of the sum of the squares of the two
small angles, in two rectangular planes, sR~^ and rsR'^Ta', or psp^ and sps^, which
and
jRs, is
are subtended, respectively, at the centre s of the osculating sphere by the small arc
s of the given curve,
and at the given point p by the small corresponding arc sTa'
of the locus of centres s of spherical curvature, or o{ the cusp -edge (3d o, (2.)) of the
polar developable
exactly* as the S7nall angle sTX, between two near radii (397,
r and rg, is ultimately the square root of the

(5.)) of absolute curvature,
squares of the ^2^0 oÂer maW an^ês, sr"i and sr-', or pkp and kpKs,
and are subtended
likewise situated in two rectangular planes,
sum
of the
which are
at the centre
k of the
osculating circle by the small arc s of the curve, and at the given point p by the
corresponding arc sTk' of the locus of the centre K (comp. 397, XXXIV., XCIV.).
in which the radius R of the osculating sphere at p ap(6.) The point, say v,
proaches most nearly to the near radius Eg from p^, is ultimately determined (corap.
397, CV. and X.) by the formula,
Vector of Spherical Curvature
_y_
~
~
Square of Angular Velocity of Radius (/i)
(p
- .)-1Tr^ =
the vector of this point
__^
XXII.
(in its
ultimate position)
is
therefore
r2i2'2p+7,2^
r2i2'2p
ov= p + ?^ =
,
-^^^ = ^^
:
/^
r^R'i-^p^
with the
verification, that
^
r^R'
rV2(T
\-x^r^
(by X., comp. XVII.) the scalar />iriJ' or R' cot
rc-
* It will soon be seen that these two

results, and others connected with them,
depend geometrically on one common principle, winch extends to all systems of
normal emanants (397, (44.)).
CHAP.
duces
CIRCUMSCRIBED DEVELOPABLE, AUX. ANGLE
Tir.]
rr^ for the case p-0, p' = 0, P=0 (comp. (3.)) and
XXXVIII., for the vector oh of the point of nearest
or to
itself to cot //,
5&1
J.
that thus the expression 397,
approach, of a radius (r) of absolute curvature to a consecutive* radius of the
same kind,
is
reproduced.
In general,
(7.)
if
we
new auxiliary
introduce a
angle, J, determined
by the
formula,
XXIII.
the expression
cot/=p-iriJ'
XXII. takes
= iE'cotP=(p' + coti7) cosP=ff(r- + P'),
the simplified form (comp. again 397,
XXIV.
ov =
^ = p cos2
4-
XXXVIII.),
/+ a sin2 /;
and the segments, into which the point v divides (internally) the radius
sphere, have the values (comp. 397, XXXIX.),
XXV.
pv = i2 sin3 J,
A geometrical signification may be
(8.)
known
analogous to the
assigned for this
signification of the angle //^(397,
S((T-p) (w-p)
of the
\s= R cos2 J.
tangent plane to the osculating sphere at P touches its

along a new right line, of which the scalar equations are,
XXVI.
= 0,
new angle
J,
which
XVII.).
In
fact,
own
^(rr'-r) (w
is
the
developable envelope
p)
such lines can be shown to be circumscribed,

along the given curve, to the locus of the osculating circle, which is at the same time
and because the developable locus of
all
the envelope of the osculating sphere,
XXVI.
we
shall briefly call this locus
of the
line
And
the Circumscribed Developable.
the inclination of the generatrix of

this new developable surface, to the tangent to the given curve at p, if suitably measured in the tangent plane to the sphere^ is precisely the angle which has been
above denoted by
J.
To render
(9.)
this conception
more completely
clear,
pj
is set off
as tq represent,
by
its
plane to the osculating sphere
and
suppose that a
so to be, in the phraseology (396, (14.)) of the
general theory of emanants, the vector-axis of that rotation.

the values,
XXVII.
let us
from the given point p, on the indefinite line XXVI., so

length and direction, the velocity oit\\Q rotation of the tangent
finite right line
We
shall then
pj= 0(= the six expressions XVI.)

= i2-ir(cot /+ U ((T - p)) = R-^ cosec J(t cos /+ TU((r - p) sin /)
(10.)
being determined by the formula XXIII., and a

cosec/, being thus obtained for the velocity XVII.
the angle
T0 = i?-i
have
Hence the new angle
J, if
new
(7, 8)
expression,
conceived to be included (like i7) between the
be considered to be measured /rom r to 0, ox from the unit-tangent to the curve at P, to the generating line pj of the circumscribed developable
which last tangent to the osculating sphere
p)
(8.), in the direction from r to r (tr
limits
and
tt,
maj'-
* This usual
expression, consecutive,
is obviously borrowed here from the lansupposed to be interpreted, like those ustd in other
parts of the present series of Articles, by a reference to the conception of limits.
guage of infinitesimals, but
is
582
III.
[boOK
generally, like the tangent <p or pj itself, an acute angle with the positive
binonnal v, as appears from the common sign of the scalar coefficients of that vec-
makes
tor, in their
developed expressions.
may also be remarked, as an additional point of analogy, and as servbecomes right, when the
ing to verify some formulte, that while the older angle
(11.) It
given curve
(12.)
is
new angle J=-xy
plane, so the
As another geometrical
some other
for every spherical curve.
illustration of the properties of the angle /,
which
results of recent sub-articles,
may
and
serve to connect them,
of
still
more
let
closely, with the general theory of normal emanants from curves (397, (44.)),
us conceive that ab, bc, cd are three successive right lines, perpendicular each
each
to
us denote by a and b the angles bca and cbd, and by c the inclination
to bc
and let us suppose that these two lines are intersected by their
let
ad
of the line
perpendicular in the points G and h respectively.

(13.) Then, by completing the rectangle bcde, and letting
common
fall
the perpendicular
BF on the hypotenuse of the right-angled triangle abe, we obtain the projections,

AE and fb, of the two lines ad and gh, on the plane through b perpendicular to bc
and hence, by elementary reasonings, we can infer the relations
XXVIII.
= tan ade = tan2
a + tan2 6
AG AF Ab2~
=
=
= sin2 aeb,
=
XXIX.
ad ae ab-*
bc
XXIX'.
BH = bc sin^j, if tan^' = tan a cot 6
.
and
or
tan2 c
nothing here being supposed to be small.
BCA and cbd,
rectilinear angles,
ACD
of the plane
or a
and
It
b,
may
also be observed, that the two
represent respectively the inclinations
the plane BCD, and of the plane
to
BH
^^,,,
abd
to the plane
abc.
(14.) Conceive next that pq and PsQs are two near normal emanants, touching
the polar developable in the points Q and Qs, whereof q is thus on the given polar
axis Ks, and q^
is
cut, in the points p'
The
eraanant pq.
and the
and let the second emanant be

by planes through p and q, perpendicular to the first
on the near polar axis k^Qs
and
q',
line pp' will then be
very nearly tangential to the given curve at

very nearly situated in the corresponding normal plane
so that these two new lines will be very nearly perpendiciilar to each
line qq' will be
to that curve
and the gauche quadrilateral p'pqq'

recently considered quadrilateral abcd.
other,
will ultimately
have the properties of the
(15.) This being perceived, if we denote by e the length of the emanant line pq,
the small angle a is very nearly = e'h ; and if the small angle b be put under the
new coefficient b' is ultimately equal (by XXIX'.) to e~^ cotj

an auxiliary angle, not generally small, and is such that we have ultimately PH = PQ.sin2j, if H be the point in which the given normal emanant pq
approaches most closely to the consecutive emanant PsQ.
form
b's,
where y
then the
is
(16.)
XXX.
We
.
have then the ultimate equation,

e6' = PQ X lim. (si qpQs)
coty =
length of emanant line (pq)
X angular
velocity
of the tangential plane (p'pq) containing
it ;
being here conceived as turning, for a moment, round the tangent to

the given curve at p, and the velocity of motion along that curve being still taken
this latter plane
for unity.
CHAP.
GENERAL THEORY OF SUCH AUXILIARY ANGLES. 583
III.]
when we change e to r, b' to r'S and^ to H, we recover in

(17.) Accordingly,
rr"J (397, XVII.), for the cotangent of the
way the fundamental value cot
older angle H; and when, on the other hand, we treat the radius of spherical curva-
H=
this
ture as the
normal emanant, supposing q
to coincide with s,
and therefore changing
R, and b' to r-' + P', we recover the last of the expressions XXIII. for the co(r'^ -\- P'), together
tangent of the new but analogous angle J, namely cot J =
^vith an interpretation, which may not have at first seemed obvious : although that
e to
expression itself was deducible, in the following
other ways, from equations
among
previously established,
pR
(18.)
As regards
the angular velocity, say
p \R
w,
sin
of the
emanant
P
line pq, or the ul-
timate quotient of the angle between two such near lines, divided by the small arc s
of the given curve, we see byXXVIII. (comp. (5.)) that this small angle vs is ulti-
mately equal to the square root of the sum of the squares of the two other small anabove denoted by a and b, and found to be equal, nearly, to e-'s and e'^s cotj
gles,
respectively
we may
XXXII.
then establish the general formula,
Angular Velocity of Normal Emanant
= = e-i cosecj
which reproduces the values, /-"icosec H, and R"^ cosec /, already found
gular velocities of the two radii, r and R.
(19.)
the
And
if
emanant PQ,
we observe
is
easily
that
t\iQ
for the
an-
projection of the vector of curvature, Kp-', on
proved to be =qp"1
= e'^.pQ, we
see
by XXXII. that
if
be divided by the square of the angular velocity (w) of the line

PQ, the quotient is the line pq. sin^J, or ph (15.): which reproduces the general
result, 397, CV., for a\i systems of normal emanants, together with a geometrical
this projection
interpretation.
(20.) As still another geometrical illustration of the properties of the new angle
we may observe that in the construction (12.) and (13.) the corresponding auxiliaiy angle y was equal to aeb, or to abf, and that the line bf (= hg) was perpenJ,
dicular to both
BC and ad, although not intersecting the latter. Substituting then,

abcd, and passing to the limit, we may say
as in (14.), the quadrilateral p'pqq' for
that
if
new
line
pj be a common perpendicular, at the given point p, to two conse-
PQ and p'q', the general auxiliary angle j is simply the

common perpendicular pj, to the tangent pp' to the curve.
instead oi normally emanating lines PQ, we consider a system of
cutive* normal emanants,

inclination p'pj, of that
(21.) And if,

tangential emanant planes (as in 397, (45.)), to which those lines axQ perpendicular,
we may then (comp. 396, (14.)) consider the recent line pj as being a. generating
line of the developable surface,
the auxiliary angle, f j,
is
which
is
the envelope of all the planes of the system ;

by (20.) the inclination of this gene^
therefore generally
*
t In these geometrical illustrations, the angle _/ has been treated, for simplicity,
as being both positive and acute ; although the general formulce, which involve the
and /, permit and require that we should occasionally attricorresponding angles

bute to them obtuse (but still positive) values : while those angles may also become
right, in
some particular cases (comp. (11)).
584
[bOOK HI.
a result which agrees with, and uicludevS, the known and fundamental property (397, XVII.) of the angle H, in connexion with the Rectifying
ratrix to the tangent
Developable (396) and also the analogous property of the newer angle /, connected
(8.) with what it has been above proposed to call the Circumscribed Developable.
;
We shall soon return briefly on the theory of that new developable surface
and of the new locus (of the osculating circle^ or envelope of the osculating
uphere) to which it ha been said to be circumscribed: but may here observe,
(22.)
(8.),
that
if
we
abridgment (comp. VIII. and XXIII.),
\vrite for
XXXIII.
..
= - =
rv
'
cot
^= cot /sec P,
then what has been called the coefficient of non-spheridty (comp. 395, (14.) and
is easily seen to have by XIV. the
values,
(16.))
XXXIV.
I^V -
S~l =
..
Xp- -rv\^-)-l
= - HrSvV*' - 1
L[p'
=coti7cot/secP=
<r
(I, 2)
(3, 4, 5)
^L^.
tRR'
TV
(6,7,8):
pr
whence also the deviation of a near point Vg of the curve, from the osculating sphere
at p, is ultimately (by 395,
XXXV.
XXVII.).
=
=
'~^- --^ ^
sp.- sr=
and accordingly, the square of the vector
p,
-a
is
given
now (comp.
I.)
by the ex-
pression,
(p.-0^ = (p-^?-l^{r2S((T-p)r"'-l},
r'^S (or - p) t" = /S = 1 + wrr'i = &c., as above.
in which
and
n enters into the following expressions for the arc^

(23.) The same auxiliary scalar
for the scalar radii of VaQ first and second curvatures^ of the locus of the centre
s of the osculating sphere., or of the cusp-edge of the polar developable (comp. 391,
(6.),
and 395, (2.)):
XXXVI.
XXXVI'.
XXXVl".
+Jds = Arc of that Cusp-Edge
ri
(or of locus ofs)
7? 7?'
ri = nr = r + p'i =
-7- =
Radius of Curvature of same edge
Radius of Second Curvature of same curve
(^Scalar)
= wr = <tV = {Scalar^
being here called scalar radii, because the^zrs^ as well as the second
(comp. 397, V.) is conceived to have an algebraic sign. In fact, if we denote by ki
these
two
latter
the centre of the osculating circle to the cusp-edge in question,
its
vector
is
(by the
general formula 389, IV.),
XXXVII.
OKi
= Ki=
<r
V(7
with the signification
XXXVI'.
XXXVIII.
a'
of
n;
= tirv,
(7
- nrrr' =
p
because by
- prrr' + prv = a - rirr',
XXXIII. (comp. 397,

= n'rv nrr-^r',
XI'.),
a" =^n'rv + n(rvy
and therefore
XXXIX.
(T
= - n2
Xa'a" = nh-W.
CURVATURES OF CUSP-EDGE OF POLAR SURFACE. 585
CHAP.
III.]
We may
also observe that
the relation
or'
||
XL.
V r = V = x'^T =
v gives (by 397, IV.),
v'
o''
Vector of Second Curvature of given curve
and that we have the equation,

-,rx
=
PK
îS
XLI.
K\
=
-p
according as the cusp-edge turns
r\
with
r
its
> 0,
concavity or
its
but
ri
>
< 0,
or
convexity towards the given
curve at P.
(24.)
sitive
The radius
quantity,
is
XLII.
and
of {first) curvature of that cusp-edge,
therefore represented
by the
when regarded
^n^ = r, = T:r, = ET ^=^~

r
ar
as a
po-
tensor,
(> 0)
-,
XXXVI". of second curvature of the same cuspby XXXVIII. from the general formula 397, XXVII.,
as regards the scalar radius
edge, its expression follows
which gives here,
XLIII.
rr'
= S =^ =
the two scalar derivatives,
nr
Y(j(T
-^,
Yvv
= /i-ir-i,
because
XLI II'.
-^ =
Yvv'
n and n", which would have introduced the derived vec-
Dg^p and Bg^p, of the fifth and sixth orders, thus disappearing
from the expressions of the two curvatures of the locus of the centre s of the osculating sphere, as was to be expected from geometrical* considerations.
tors
r'''
and
(25.)
r^, or
For the
formula
helix, the
thus, as a verification, the
known
XXXVII.
result, that
gives ki
= p,
or Ki
the given point
= p we have then
;
oithis curve
is
the centre of curvature Ki of that other helix

(comp. 389, (3.), and 395, (8.)),
is in this case the common locus of the two coincident
centres, K and s. It is
itself
which
scarcely
we have
necessary to observe that for the helix
also
/= H.
(26.) In general, the rectifying plane of the locus ofs is parallel to the rectifying plane of the given curve, because the radii of their osculating circles are parallel ;
the rectifying lines for these two curves are therefore not only parallel but
equal ;
and accordingly we have here the formula,
XLIV.
Xi
=
V
\ (by 397, XVI.),
T
n V-,=
,
which will be found to agree with this other expression (comp. 397, XVII.),
XLV.
tan
^1 =
=
-^
Tri
- Uri
r
= -foot
H,
-
the upper or lower sign being taken, according as the

Figs. 81, 82), on is
and the new angle
* In
fact,
new curve
is
concave (as in
convex at s (comp. (23.)), towards the old (or given) curve at
Hi
being measured in the new rectifying plane,
from the new
n represents here the velocity of motion of the point s along
its
own
represent respectively the velocities of rotation of the tangent

and binormal to that curve
so that nr and nr must be, as above, the radii of its
locus,
while
r"^
and
r-i
two curvatures.
4 F
586
[bOO'K
III.
tangent a or nrv^ to the new rectifying line Xi, and in the direction from that new
tangent to the new binormal vi, or (corap. XL.) to a line from s which is equal to
the vector of second curvature T'^t of the given curve^ multiplied
namely by Tn-i, or by the coefficient n-i taken positively.
by a
positive scalar,
(27.) The former rectifying line \ touches the cusp-edge of the rectifying developable (396) of the given curve, in a new point r (comp. Fig. 81), of which by
397, (45.), and by XV., the vector /rom the given point is, generally,
Yt'H"
^^^^
UXsin/T
r\
r-2\
with the verification that this expression becomes infinite (comp. 397,
when the curve is a geodetic on a cylinder.
(4.9.), (50.)),
(28.) In general, the vector or of the />otn< of contact b, which vector

here denote by v, may be thus expressed,
xLVii.
..
= oR=p + mx,
XLViii.
if
we
shall
z=^^=;i^,;
E'
(rr-')
and because
= (nr')'r, by
(rX)'
XLIX.
y'
= rX f
VII'., its
^-^
first
derivative
is,
= UX cosec H (I sin H)' = UX (V +
cos
H)
which however the new derived scalar I' involves H", and so depends on t"'
while the scalar coefficient I itself represents the porfiou (+pR)of</te rectifying line,
intercepted between i\\Q given curve, and the cusp-edge (27.) of the rectifying devein
and considered as positive when the direction of
lopable,
with that of the
line
+ X,
but as negative in the contrary
this intercept
pr
coincides
case.
(29.) For abridgment of discourse, the cusp-edge last considered, namely that of
the rectifying developable, as being the locus of a point which we have denoted by
''
the curve (r) ;" while the /ormer cusp-edge
the letter r, may be called simply
(23.), or that of the po/ar developable,
(s) ;" the locus of the centre
(k)
and the given curve
:"
itself
may
may
be called, on the
plan,
L.
this
be called in like manner "</ curve

be called " the curve
(comp. again Figs. 81, 82)
" the curve

(p)."
The
arc rr of the curve (r),
(30.)
same
may
of (absolute) curvature
is
XXXVL),
(by XLIX., comp.
['l!v'dis=ls-l+ [*cosird;
arc being treated as positive, when the direction of motion along it coincides with
that of
(31.)
XVI.
LI.
X.
The expression VII. for X', combined with the former

by the general formula 389, IV.,
expression S
for X, gives easily

.
Vector of Centre of Curvature of the Curve (r)
whence
LII.
Radius of Curvature of Curve (r)
v'
v'
T = T
,
-,
the scalar variable being here arbitrary.

see, at the same time, that the angular velocity of the rectifying lu.'
(32.)
X, or of the tangent to this crtrve (r), is represented by H' ; or that the tmi^
We
CHAP.
CURVATURES OF EDGE OF RECTIFYING SURFACE. 587
III.]
angle* between two such near lines, X and Xj, is nearly equal to sM', or to Ha Hi
while the vector axis (V/VX~i) of rotation of the rectifying line, set off from the
- /JVr', for its
point R, has -H'IJt', or
expression.
As
(33.)
regards the second curvature of the same curve (r), Ave
and
that the expression (comp. VII.
LIII.
X"= (r-i)"r + (r-i)"rj/ +
r'l
(rr-)'r'
combined with the parallelism (XLIX.) of v to

397,
observe
may
LI.),
(r-i)"r
X,
gives,
(r-^)"rv
+ VXX',
by the general formula
XXVII.,
LIV.
Radius of Second Curvature of Curve (r)
cos
Yv'v"
XV
VXX'
~x
TX
'
with the verification, that while l' + cos

represents, by (30.), the velocity of motion along this curve (r), TX represents, by 397,
(3.), the velocity of rotation of
its osculating plane,
namely the rectifying plane of the given curve (p) and it
:
worth observing, that although each of these two radii of curvature, LII. and
LIV, depends on t'^ through V (28.), yet neither of them depends on r^ (comp.
As another verification, it can be shown that tha plane of the two lines X
(24.)).
and t' from p, namely the plane,
is
LIV.
which
Sr'X(a>-p)
= 0,
the normal plane to the rectifying developable along the rectifying line, and
contains the absolute normal to the given curve
(p), touches its own developable enis
or
XXXVIII.,
given curve to
fying
if H be the point determined

by the formula 397,
point of nearest approach of a radius of curvature (r) of that
consecutive (comp. (6.)
this line'RH must therefore be the recti-
along the line rh,
velope
t\\Q
its
of the curve (r)
line
tangent of
its
and accordingly (comp. 397, XVII.), the trigonometric
inclination to the tangent
rp
to this last curve has for expression
(abstracting from sign),
LIV".
tan
PRH = PH p5 = + I'^r sin2

:
H= rH' sin H = T\-^H'
Radius (LIV.) of Second Curvature of Curve (r)

Radius (LII.) of First Curvature of same Curve
(34.) Without even introducing r''', we can assign as follows a twisted cubic
(comp. 397, (34.)), which shall have contact of the fourth orrfer with the given
~"
curve at p
or rather an indefinite variety of such cubics, or gauche curves of the
third degree.
LV.
Writing, for abridgment,

.
Sr(aj -p),
LVL
so that
y = - Srr'(w - p), z
p + ccr + yrr + zrv,
Srv((>
p),
w=
the scalar equation,
Lvii...(?^f=6(lY-t('^yy*
result substantially equivalent to this
analysis, in the
deduced, by an entirely different
above cited Memoir of M. de Saint- Venant, and
geometrical considerations
(or, as
is
+ ',
which
also lead to expressions for the
is
illustrated
by
two curvatures
he calls them, the courbure and cambrure), of the cusp-edge of the rectifying
and to a determination of the rectifying line of that cusp-edge.
developable
588
[bOOK
III.
which e is an arbitrary but scalar constant, represents evidently, by Mi

form, a
cone of the second order, with its vertex at the given point p and this cone can be
proved to have contact of the fourth order with the curve* at that point or of the
in
third order with the cone of chords from it

In fact the
(comp. 397, (31.), (32.)).
coefficients will be found to have been so determined, that the difference of the two
members of this equation LVII. contains s^ as a factor, when we change w to

ps, as
given by the formula I., or when we substitute for xyz their approximate values for
the curve, as functions of the arc s ;
namely, by the expressions IV. for r'", and
397, VI. for r",
- rV ^2
*2
LVIII.
'
^*=
"i
27
r*
s3
where the terms
set
down
s*
+
24 ^^''"''y
''"'+ ""''-'^^
^,
are more than sufficient for the purpose of the proof.
-*
in y which
It
be added that the
may
here,
may
coefficient of
be transformed as follows
LVIII'.
Srr'r"'
is
the only one at all complex
= -(r-)"-r-iX2 = r-3^ + p(r-2r-i)'5
S being that scalar for which (or more immediately for its excess over unity) several
expressionsf have lately been assigned (22,), and which had occurred in an earlier
investigation (395, (14.), &c.).
(85.)
With the same
significations
LV.
of the three scalars xyz, this other equa-
tion,
LIX.
LIX'.
or
ISry- {Sx - r'y^ =

.
(9
2ry-(a;-irV)2 = (l
when we
+ r'2 -
Zrr"
- 3r2r-2) y^,
-|r^(r^)"-lr2r-2)y2,
x and y the values LVIII. of x,

and neglect or suppress s^ it therefore represents an elliptic (or hyperbolic)
cylinder, which is cut perpendicularly, by the osculating plane to the given curve at
will be found to be satisfied
and
ys,
p, in
substitute for
an
having contact of the fourth order with the prq/cc-
ellipse (or hyperbola'),
(comp. 397, (9.)), of that given curve upon that osculating plane : and the cylinder itself has contact of the same (fourth) order with the curve in space, at the
tion
* In the language of
infinitesimals, the cone
LVII. contains five
consecutive
but it contains ouly four conpoints of the curve, or ha.3 five-point contact therewith
secutive sides of the cone of chords from the given point, or has only four-side con:
tact with that cone, except for one particular value of the constant,
e,
which we
shall
be observed that xyz form here a (scalar) system oi three

rectangular co-ordinates, of the usual kind, with their origin at the point p of the
curve, and with their positive semiaxes in the directions of the tangent r, the vectw
presently assign.
of curvature
It
we have
r',
It
may
and the binormal
v.
might have been observed,
in addition to the eight
also,
XXXI v.
5-
= Hr-^
cot
forms
/= n cot H
XXXIV.,
thai
(9, 10)
CHAP.
OSCULATING ELLIPTIC CYLINDER.
III.]
same given point p, so that we
may call
it
589
(comp. 397, (31.)) the Osculating Elliptic
to the osculating plane.

(or Hyperlolic) Cylinder, perpendicular
if we suppress the second member of either LIX. or LIX'.,
(36.) As a verification,
we obtain, under a new form, the equation of what has been already called the Oscu-
ParahoUc Cylinder (397,
lating
LXXXIV.)
verification, the co-
and as another
of y^ in that second member vanishes, as it ought to do, when the given

curve is supposed to be a parabola : that plane curve, in fact, satisfying the differential equation of the second order,
efficient
LX.
r be
still
3rr"
LX".
or
if
r'2
r-
= 9,
LX'.
or
^ j>
r^ (r^)" =
2,
= const. =p%
1
J
the radius of curvature, considered as a function of the arc,
s,
while
is
here the semiparameter.
The hinormal v
(37.)
and although
LIX.
make
it
its
such,
by the
is,
by suppressing the
will intersect the cylinder
LVII., yet
we can
particular value zero to the arbitrary constant,
by assigning the
equation, or
construction, a generating line of the cylinder
this line is not generally a side of the cone
term, ez^.
And when this
is
e,
in
done, the cone LYII.
LIX., not only in this common side v (comp. 397, (33.)),

of th.Q fourth order with
cribic, which will have contact
but also in a certain twisted
the given curve at p, as stated at the commencement of (34.).

such cubics can be de(38.) But, as was also stated there, indefinitely many
scribed,
at the
which shall have contact of the same
same point.
For we
LVL),
LXI.
in
which
a, b, c are
any
OE =
which
^ is
jO
+ ar +
three scalar constants
LXIL
in
a)
order, with the
{foui-tli)
may assume any point e
of space, or
same
curve,
any vector (comp-
brr' + crv,
and then the vector equation,
= p, + /(-|o),
a new scalar variable, will represent a cylindric surface, not generally

and having the line pe for
of the second order, but passing through the given curve,
We
can then cut (generally) this new cylinder by the osculating

a generatrix.
to the curve at p, and so obtain (generally) a new and oblique projection of
the curve upon that plane ; the x and y of which new projected curve will depend on
plane
the arc s of the original curve
LXIII.
by the
.
relations,
with the approximate expressions LVIII. for
new scalar
constants,
expressions LXIII. for
and
y = ys
x = Xs-ac'^Zs,
C,
x and y
And
XsysZg.
if
we then determine two
by the condition that the substitution of these

shall satisfy this
LXIV.
bc-^zs;
2ry
= rcs +
new
<2Bxy
last
equation,
+ Cy\
be neglected (comp. (35.)), or by equating the coefficients of s^ and s*,

in the result of such substitution, then, on restoring the significations LV. of xyz,
if
only
s^
and writing
for abridgment,
LXY.
X=x-ac-%
the equation of the second degree.
Y^y-bc-^z,
590
LXVI.
2rF=X2 + 25^F+
III.
[bOOK
CV^,
represent generally an oblique osculating elliptic (or hyperbolic) ei/linder, which

has contact of the/o7/r/A oj-der with the given curve at p, and contains the assumed
will
we determine
If then
line PE.
substitution of abc for xyz, or
finally the constant e in
by the
LVII., by the result of the
condition,
-(Ti-(Fr-(r.y- +
the cone LVII,, and the cylinder
cc,
LXVL, will have that line pe for a common side ;
and
will intersect each other, not only in that line, but also (as before) in a twisted cubic,
although
now
new one, which
will
have the required (fourth) order of contact, with
the given curve at the given point.

(39.) If, after the substitution (38.) in
three powers,
tion
and then eliminate
s^, *, s^,
of condition, which
which the
ratios of
abc
(40.) Again, if
scalars Xs,
(comp.
t/s,
are assignable functions of

this condition
r, r, r', r', r", r",
LXVIII.
is satisfied,
has contact of the fifth order with the given curve at P.
we improve
and
I.) of ps,
and when
the approximate expressions LVIII. for the three
by taking account
^s,
coefllcients of the
are conducted to an eqria-
serve to determine the direction of the generating line
LXVI.
we
a&H b5V 4- chc^ + ec^ = ac(bg + ch)
still
PE, while the coeflScients a, b, c, e, g,

and r'", depending on the vector t^^
the cylinder
C,
found to be of the ^brm,
is
LXVIII.
in
LXIV., we equate the
B and
if
we
in the equation of the cone

cient of s^ in the difference
of
s^,
or
by introducing the new term
substitute the expressions so improved, instead of x, y,
z,
LVII. and then equate to zero (comp. (34.)) the coeffiof the two members of that equation, we obtain a definite
expression for the constant,
e,
which had been arbitrary
before,
but becomes
now
given function of rrr'r'r" and r" (not involving r"), namely the following
'
'
*~'5 Vh~ rî'^'^~"7^'^'^~
and when the constant
"^
4r2r2
r2r
J"'
e receives this value,* the cone has contact of the
j^M
order
with the curve at the given point.

(41.) Finally, if we multiply the equation LXVII. by 6g+ch, we can at once
eliminate a by LXVIII., and so obtain a cubic equation in b: c, which has at least
one real root, answering to a real system of ratios

direction of the line
pe
in (38.).
a, b, c,
and therefore
to
a real
It is therefore possible to assign at least one real
cylinder of the second order (39.), which shall have contact of the fifth order with
the curve at p, and shall at the same time have one side pe common with the cone
of the second order (40.),
which has contact of the same (fifth) order with the curve
and consequently it is possible in

(or of the fourth order with the cone of chords)
this way to assign, as the intersection of this cylinder with this cone, at least one real
;
Compare
the
first
Note
to
page 688.
OSCULATING TWISTED CUBIC.
CHAP. Ill]
591
hoisted cubic, which has contact of the fifth* order with the given curve of double
And such a cubic curve may be called, by

curvature, at the given point thereof.
eminence, an Osculatingf Twisted Cubic.
on the subject of such cubic
(42.) Not intending to'returu, in these Elements,
we may take
curves,
this occasion to
LXX.
represents a curve of this kind,
remark, that the very simple vector equation,^

.
yap = pYtip,
a and
if
(3
be any two constant and non-parallel
by the symbol S.X, in which \

an arbitrary but constant vector, the scalar equation so obtained, namely,
In
vectors.
fact, if
we
operate on this equation
LXXI.
representa
a,
SXap =
S\(oS/3p
- pS/3X,
surface of the second order, on which the curve
making then
successively
LXXII.
\ = a and \ =
S (Yap .Yf3p) =
0,
j8,
we
and
is
is
wholly contained;
two equations,
get, in particular, the
LXXIII.
(V/3p)2
-f
Sa(3p
= 0,
representing respectively a cone and cylinder of that order, with the vector [3 from
the origin as a common side : and the remaining part of the intersection of these
two
surfaces, is precisely the curve
known
sense already referred
LXX., which
therefore is a twisted cubic, in the
to.
same curve, would be

(43.) Other surfaces of the same order, containing the
obtained by assigning other values to \ for example (comp, 397, (31.)), we should
get generaWy an hyperbolic paraboloid from the form LXXI,, by taking X-i-j3.
;
But it may be more important here to observe, that w//Aou^ supposing any acquaintance with the theory of curved surfaces, the vector equation LXX. can be shown, by
known
(see page 242 of Dr. Salmon's Treatise, already

can generally be described through any six given points ;
and also (page 248), that three quadric cylinders (or cylinders of the second order
or degree) can be described, containing a given cubic curve, their edges being pait
Accordingly,
is
cited), that a twisted cubic
rallel to the three (real or
imaginary) asymptotes.
f Compare the first Note to page 563.

X This example was given in pages 679,
&c., of the Lectures, with some connected transformations, the equation having been found as a certain condition for
the inscription of a gauche quadrilateral, or other even-sided polygon, in a given
spheric surface (comp. the sub-articles to 296)
the 2n successive sides of the figm-e
It
being obliged to pass through the same even number of given points of space.
was shown that the curve might be said to intersect the unit-sphere (p^ = 1) in two
imaginary points at infinity, and also in two real an^ two imaginary points, situated
on two real right lines, which were reciprocal polars relatively to the sphere, and
might be called chords of solution, with respect to the proposed problem of inscripand that analogous residts existed for even-sided polygons in
tion of the polygon
and other surfaces of the second order : whereas the corresponding problem, of the inscription of an odd-sided polygon in such a surface, conducted only to
the assignment of a single chord of solution, as happens in the known and analogous
ellipsoids,
theory of polygons in conies, whether the number of sides be (in that theory) even
or odd.
But we cannot here pursue the subject, which has been treated at some
length in the Lectures, and in the Appendices to them.
592
[bOOK
quaternions, to represent a curve of the third degree, in the sense that
an arbitrary plane, in
t?iree
points (real or imaginary).
In
fact,
it is
we may
III.
by
cut,
write the
equation as follows,
LXXIV.
Vgp = -
LXXV.
if
a,
q being here a quaternion, of which the vector part

is
arbitrary
we may
still
and then, by resolving (comp. 347)

further transform
LXXVL
which conducts
to a cubic
/3 is
.q=ff + ^,
given, but the scalar part
this linear equation
LXXIV.,
as follows,
it
.^(^-/32),o=^S/3a+^V/3a-i72,
equation in g, when combined with the equation,
.
LXXVII.
Sfjo
e,
of any proposed secant plane.

(44.) The vector equation
LXX., however, is not sufficiently general, to represent an arbitrary twisted cubic, through an assumed point taken as origin ; for
which purpose, ten scalar constants ought to be disposable, in order to allow of the
curve being made to pass through ê* oÂer arbitrary points : whereas the equation referred to involves only Jive such constants, namely the four included in Ua
Ta (comp. 358).
(45.) It is easy, however, to accomplish the generalization thus required, with
the help of that theory of linear and vector functions (0p) of vectors, which was asand U/3, and the one quotient of tensors T/3
We
have
signed in the Sixth Section of the preceding Chapter (Arts. 347, &c.).
only to write, instead of the equation LXX,, this other but analogous form which
includes
it,
LXXVIII.
and which
gives,
by
Yap + Ypfp = 0,
principles
or
LXXVIII'.
0p + cp
= a,
and methods already explained (comp. 354,
(1.)),
the transformation,
^^ + c)-.a^
^
LXXIX...p^ = (^
m ''"'!''^"!f\
+ m'c + m c^ + c^
a, //a, and â being here fxed vectors, and in, m', m" being fixed scalars, but c
being an arbitrary and variable scalar, which may receive any value, without the
expression
LXXIX.
ceasing to satisfy the equation
LXXVIII.
first Note to page 591.

In general, when a curve in space is
be represented (comp. 371, (5.)) by two scalar equations, each new arbitrary point, through which it is required to pas*, introduces a necessity for <7o new
Compare the
supposed
to
and accordingly each new order, say the

contact with such a curve, has been seen to introduce a new vector,
or
Ds'|0,
disposable constants, of the scalar kind

n", of
7-(-i),
=-
subject to a condition resulting from the general equation

1 (comp. 380,
TDsp =
l,
or
XXVI., and
396, III.), but involving virtuaWy two new scalar

constants. Thus, besides the /our such constants, which enter through r and r' into
the determination of the directions of the rectangular system of lines, tangent, norr2
mal, and binormal (comp. 379, (5.), or 396, (2.)), and of the length of the radius
of {frst) curvature, r, the three successive derivatives, r', r", r", of that radius, and
the radius r of second curvature, with its two first derivatives, r' and r", have been
seen to enter, through the three other
vectors, t", t"', t^"^, into the determination
(41.) of the osculating twisted cubic.
CHAP.
III.]
VECTOR EQUATION OF A TWISTED CUBIC.
The curve LXXVIII.
(46.)
LXXVII.
therefore
is
in three points (real or imaginary),
three roots of the cubic in
c,
which
is
593
cut (comp. (43.)) by the plane

answering to and determined by the
formed by substituting the expression LXXIX.

and consequently it is a curve of the third
for p in the equation of that secant plane;
which have directions correspondobtained by equating to zero the denominator of that

so that
expression LXXIX., or by making M=0, in a notation formerly employed
thej' have the directions of the three lines /3, which satisfy this other vector equadegree, the three (real or imaginary) asymptotes to
ing to the three values of
c,
tion (comp. 354, I.),
LXXX.
V/3^/3
0.
(47.) Accordingly, if /3 be such a line, and if y be any vector in the plane of a

and /3, the curve LXXVIII. is a part of the intersection of the two surfaces of the
second order,
LXXXI.
whereof the
first is
side (comp. (42.))
Sap0p = 0,
LXXXIL
and
a cone, and which have the line

:
the curve
ft
anew
therefore found
is
Syap + Syp^p =
from the origin
0,
for a
common
to be a twisted cubic.
(48.) And as regards the number of the scalar constants, which are to be conceived as entering into its vector equation LXXVIII., when we take for (pp the form
V50P + VXpjU assigned in 357, I., in which go is an arbitrary but constant quaternion, such as g + y, and X, /x are constant vectors, the term gp of 0p disappears
under the symbol of operation V.p, and the equation (45.) of the curve becomes,
LXXXIII.
in
which the four
versors,
Yap + pVyp + YpYXpfi -
while the two tensor quotients,
Ty Ta
:
introduce each two scalar constants,

and TXfx Ta, count as two others : so that
Ua, Uy, UX,
JJfi,
is exactly made up, the curve

being
supposed to pass through an assumed origin, and therefore to have one point
It is scarcely worth observing, that we can at once remove this last restricgiven.
the required
number
of ten such constants (44.)
still
by merely adding a new constant vector'to p, in the last equation, LXXXIII.

(49.) Although, for the determination of the osculating twisted cubic (41.), to
a given curve of double curvature, it was necessary (comp, (40.)) to employ the
vector r''' or Ds^p, or to take account of s^ in the vector ps, or in the connected sca-
tion,
x^sZs of (34.), and therefore to improve the expressions LVIIL, by carrying in

each of them (or at least in the two latter) the approximation one step farther, yet
lars
many other problems relating to curves in space, besides some that have been
already considered, for which those scalar expressions LVIIL are sufficiently approximate : or for which the vector expression T. suffices.
there are
(50.) Resuming, for instance, the questions considered in (22.) and (23.), we
additional light on the law of the deviation of a near point Pg of the
curve, from the oscidating sphere at p, as follows.
Eliminating n by XXXVI'.
may throw some

from
XXXV., we
find this
new expression,
LXXXIV.
.iP5-SP =
'
24rr2i2
the direction of this deviation from the sphere (i?) depends therefore on the sign
of the scalar radius
(23.) of curvature of the cusp-edge (s) of the jpo/ar deve-
outward or inward
lopable: and
it is
edge turns
concavity (comp. XLI.) or
its
(cormp. 395, (14.)),

its
4 G
according as that cusp-
convexity, at the centre s of the oscu-
594
lating sphere, towards the point
p of the given
curve, that
[bOOK
is,
III.
towards the point of
osculation.
(51.) Again, if we only take account of s^^ the dcAdation of Vsfrom the osculating circle at p has been seen to be a vector tangential to the osculating sphere, which
may
be thus expressed (comp. 397, IX., LII.),
LXXXV...c,P. = 'ivV='^ll^,
&rn
be the point on the circle, which is distant from the given point P by an arc of
that circle =s, with the same initial direction of motion, or of departure from p, re-
if c
presented by the
common
unit tangent t
the quantity of this deviation
s^R
expressed by the scalar
from
(P)
the osculating
that
by the deviation
is,
by the secant
at p, multiplied
plane*
is therefore
s^
(comp. 397,
(9.), (10.))
(r'^R') of the inclination
of the radius (72) of sp/ierica/ curvature, to the radius (r) of absolute curvature, and positive when this last deviation has the direction of the hinormal v.
(52.) On the other hand (comp. (5.)) the small angle, which the small arc ss,
of the cusp-edge (s) of the polar developable subtends at the point p, is ultimately
expressed by the scalar,
LXXXVI.
sps,
= (PS5 -
ps)
this angle being treated as positive,
i2-i
when
cot
J*
P=
7?
TlVS
(by XXXIII.),
the corresponding rotation f round
Besides the nine expressions in 397, (42.) for the s^ware
r~3 of
t from
the secowd
curvature, the following may be remarked, as containing the law of the regression of
XhQ projection of a curve of double curvature on its own normal plane :
r-2
PQ32
~,
11m.
2kP
being still the centre of the osculating circle, and Qi,

(10.)) the projections of a near point q (or p^), on the
(or inward radius of curvature pk), and the binoi-mal
terms of the three vector projections corresponding, of
are (comp. LVIII.)
PQi
= ST
397, XCIX., (10)

>"
V
'
PQ2^
Q2, Q3 being still (as in 397,

tangent, the absolute normal
at p.
In
the principal
fact,
the small chord
pq
(or pp^),
PQ3
= (fsV =) -
Ur'
PQ3
zr
= (is^ri^
=,)
Uj^
orr
whence, ultimately,
9
- PQ32
2 PQo^
.
t Considered
IT
= r-2^r\jT
=
'
as a rotation, this small angle
vector, rp'^R'R'^ST',
and
if
the vector deviation
be multiplied by this, the quarter of the product

tion from the osculating sphere, under the form,
S*(p-tT)
*
24/2
^
rrp
r-2
kp.
may
be represented by the small
LXXXV.
is
(comp.
from the osculating
XXXV.) the
circ/e
vector devia-
CHAP.
DEVIATIONS FROM CIRCLE AND SPHERE.
III.]
595
rs to ps^ is positive and if we multiply this scalar, by that which has just been
assigned (51.), as an expression for the deviation CsP* from the osculating circle, we
:
by
get,
XXXV.,
the product,
s^R
(53.)
Combining then the recent
lowing Theorem
tR's
R's^
results (50.), (51.), (52.),
we
arrive at the fol-
The deviation of a near point Vs of a curve in space, from the osculating sphere
at the given point P, is u/timaieli/ equal to the quarter of the deviation
of the same
near point from the osculating circle at p, multiplied hy the sine
the small an'
of
of spherical curvature (s), or of the
and this deviation
cusp-edge of ih.Q polar developable, subtends at the same point P
SP) from the sphere has an outward or an inward direction, according as the
(sFs
gle which the arc ss^,
of the locus of
centres
SSs is concave or convex towards the same given point.

(54.) The vector of the centre Sj, of the near osculating sphere at
same order of approximation, comp. I.),
same arc
LXXXVIII.
and although a
oSs
= o-s = + sa + Is'â" + \s^(t"' +

(7
^V^cr'^
p,, is
(in the
already a function (by 397, IX., &c.) of r, r', r", so that a'
is
(as in (2.) or (22.)) a function of r', t", t", and o", a", o^^ introduce respectively
the new derived vectors r"^, r^, r^*, or D^^p, D/p, Dj'^'p, which we are not at pre\&
sent employing (49.), yet we have seen, in (23.) and (24.), that some useful combinations of a" and a'" can be expressed without r'^, r^
and the following is another
remarkable example of the same species of reduction, involving not only a" and c"'
:
but also
(t'^,
but
admitting, like the former, of a simple geometrical interpreta-
still
tion.
(55.)
Remembering (comp.
LXXXIX.
(<r
(22.),
- p)3 + 722 = o,
and 397, XV.) that
and
XC.
Sr"' (ff
- p) = r-^S = r-2 + nr-^r^
and reducing the successive derivatives of LXXXIX. with the help of the equations
397, XIX., and of their derivatives, we are conducted easily to the following system
of equations, into which the derived vectors r, t', &c. do not expressly enter, but
which involve
(t\ <t", a'", a^^,
XCI.
= 0;
= 0;
XCII.
Sc7'((T-p) +
S(T'(T"(<r-p)
- p) + ff'2 + {RR)' =
XCIII.
Scr"((r
XCI V.
S(t"' ((T-p)-\- 3S(T
+ (RR')" =
.
R', R", R'", R^^
i2/?'
XCV.
and
S<T'' (o"
- p) + 4S<rV" + 3(r"2 +
(RR')'"
-^ = --',
rr
rrp
auxiliary equations being,
XCVI.
XCVII.
and
S(T'r
.
S(t"V
0,
SffV =
=-
S<t"t'
0,
Sff'V
= So-'r" =
= 0,
comp. 395, X.
- S (<r - p) r'"
= _r-2("5_l) = _r-ir-i.
Srr"
(56.) But, if Rs denote the radius of the near sphere,
we have,
and
if
we
still
neglect
XCVIII.
pX^ = - (t. - p*)2 =
Rs^
= iJ2 + 2sRR' 4- s-i^RRy +
'-
(RR')"
+ ^ (RRJ"
s^,
596
whence
by LXXXVIII., and by the recent equations,
follows,
from which (comp. (24.)) everything depending on
pression,
[bOOK
this
r'^,
111.
very simple exr^'' has disap-
t%
peared^
XCIX.
- p)2 +
((T.
R,2
ZJ^
IZrrp
and which gives (within the same order of approximation, attending
the geometrical relation,

C.
PS,
PsSs
or
^'*
N
^
= ^,
T((Ta- p)- Iis=
C'.
ns*
to
XXXV.)
SP
SPs
,7-;
sp - sp = s,p - SP = i2g - jB.
(57.) This result might have been foreseen, from the following very simWhen the coefficient
ple consideration.
of non-sphericity (395, (16.)), or
S-1
of the deviation of a curve from a sphere,
curve
is
(what we may
by an amount which
curve at p,
s,
is
p must
of the curve, then the given point
son, exterior to the near sphere,

tions, PS,
positive, so that
call) the
exterior to
of the arc,
a near point p, of the

given sphere, which osculates to that
ultimately proportional to the fourth power
is
PsSs,
and sp,
which osculates
be, for the
at the point p
wh\ch have been found by calculation
sp,
same
rea-
and the two deviato be equal
be neglected, must in fact bear to each other an ultimate ratio of equality,

because the two arcs, + s and s, from p to p,, and from p, hack to p, are equally
= (*)* And precisely the
long, sdihovLgh oppositely directed ; or because (+0^
(C),
if s5
same reasoning
applies,
when
the coefficient 5'- 1
negative, so that the deviations,
is
equated in the formula C, are both inwards.

(58.) As regards the deviation (51.) of the near point p, of the curve from the
osculating circle at P, we may generalize and render more exact the expression
LXXXV., by considering a point Ct of that circle, which is distant by a circular arc
=t
from the given point p

CI.
and
oct
of which the vector
(>)t
is,
= p + rr sm - + r^r
.
or
if
we only
neglect
way we
shall
cm.
with the scalar
p+r(<-^]
neglecting s^ and
t'^,
c<Ps= p
(18.),
r.-(|-5|j].
-m= XT+YrT'-\- Zrv,
JV=a:,-r sin-,
F=y,-rvers-,
r
and attending
to the expressions
= ^-'
67^
2r
6rr
which
vers r
have (comp. (34.)) the vector deviation,
in
by 396,
coefficients,
CIV.
or,
t^,
CII...oc, = <.,=
(59.) In this
rigorously,
r, r', r,
p,
24
and n have the same
^=2^;
LVIII. and LVIII'.,
873'
24r3
^
24r2r'
'
significations as before.
CHAP.
RELATIONS BETWEEN THE DEVIATIONS.
III.]
(60.) Assuming then for the circular arc
597
the value,
r's*
cvi... =, +
which
differs (as
curve,
we
we
see)
_,
by only a quantity of the ybwrM order from the arc

same order of approximation, the expressions,
s of the
shall have, to the
F=^Z--^,
r
24r2r
..^=0,
CVII.
Z = Z5 = &c.,
as before,
the deviation at p^ from the circle being here measured in a direction parallel to the
normal plane at p ; and if * be neglected (although the expressions enable us to
take account of
it),
to the osculating
this deviation is also parallel (as before) to the tangent t{<t
sphere in that plane
which agrees with the
CVIII.
such then
represented in quantity
without neglecting
also,
= - - =
by
p)
iJr'^'zs,
sp
s^,
sps
24rri2
the component of the deviation from the osculating circle, which is panormal PS to the sphere at p and we see that it only differs in sign
is positive when its direction is that of the inward normal, or inward ra-
is
rallel to the
(because
it is
result in (51.).
The expressions CVII. give
(61.)
while
it
sp
(comp. C), for the outward deviation sps
same osculating sphere at the given point p.
(62.) This latter component (61.) is small, even as compared with the /ormer
and the small quotient, of the latter divided by the former,
small component (60.)
XXXV.
dius Ps), from the expression

from the
of the near point Ps,
is
ultimately (by
LXXXVI.),
-nrs
rV-VpZ
^^^
where the small angle spSs is positive or negative, according to the rule stated
(52.), and may be replaced by its sine, or by its tangent.
is
in
in (60.), by a plane which

(63.) Instead of cutting the given osculating circle, as
parallel to the given normal plane at P, we may propose to cut that circle hy the
near normal plane at
ex.
which
is
small,
and
Ps,
or to satisfy this
= Srs (ps -
<o),
easily found to give

s*
CXI. ..* =
being
still
and
so that in passing to this
new
.
condition,
= X^tTs + FSrr'n +
by CV. the values
neglected)
3,
CX'.
or
(s
and
being
still
ZSrvTs
supposed to be
CXII.
F=&c., Z=
^=^,
new near point
CVII.
we only change
from
make no change in the values of Y
Ct of the circle,
zero to a small quantity of the foxirth order, and
&c., as in
and Z.
(64.) The new deviation CtFs from the given circle may be decomposed into two
partial deviations, in the near normal plane, of which one has the direction of the
p^ to the near sphere at Ps, and the other has that of the
unit-normal Rs'^(jTs ps) to the same sphere at the same point (or the opposites of
these two directions)
and the scalar coefficients of these two vector units, if we atunit-tangent Its~Ws(jTs
tend only to principal terms, are easily found to be,
598
CXUl./l^^p;,
R
and
CXiy../Il^^^^^.
R
8rrR
6r2r'
(65.)
We may
CXV.
then write
which
it
Deviation nf near point Ps from given oscuhiling circle,

measured in the near normal plane to the curve at Ps,
= new
in
III.
[boOK
= ;rv
6r-r
CtFs
- ^ps) +
Urs(frs
^
^
5 U((7
SrrR
- ps)
be observed, that the second scalar coefficient is equal to three times

sp (XXXV. or C), of the near point Vg of the curve, from
may
the scalar deviation sp
the given osculating sphere (at p).
(66.)
But we may
ing a new deviation

osculating sphere at
that near sphere
also interpret the
new
mentioned, as represent-
coefficient last
namely, that of the point Ct of the given
P^,
when
considered as positive
circle,
that new point Ct
or as denoting the difference nf distances, SgCt
therefore (comp. (56.)) this
CXVI.
new geometrical
SsC^- SsPs= 3(sPs
CXVr.
or
relation, of
(67.) Supposing, then, at
SsCt
sp)
3ssP
from
is
SsP.
the near
exterior to
We
have
an extremely simple kind
3 (ssV
- s^p*)
2sPs.
S-Ws
that the coefficient ofnon- sphericity

posiconceive a point to move backicards, upon the curve,
first,
(comp. 395, (16.)), if we

from Ps to P, and then forwards, upon the circle which osculates at P, to the new
point Ct (63.), we see that it will^^rs^ attain (at p) a position exterior to the sphere
tive
which osculates at
ation,
with respect
(at the
new
Pj,
have an amount, determined in (56.), of outtvard devinear osculating sphere ; and that it will afterwards attain
a deviation of the same character (namely outwards, if 5'> 1),
or will
to that
point Ct)
from the same near sphere, but one of which the amount
this last relation holding also
(68.) It
Ps,
is
on the near
circle,
which
from the arc
when S <1,
or
will be threefold the former

both deviations are inwards.
when
easy also to infer from (65.), (comp. (57.)), that if we go hack from
circle which osculates at that near point, through an arc (t) of that
will only differ
of the
()
by a small quantity of the fourth order (comp. (60.))

moment we
curve, so as to arrive at a point, ^Yhich for the
shall simply denote by c, and in which (as well as in another point of section, not
necessary here to be considered) the near osculating circle is cut by the given normal plane at P, the vector deviation of this new point c of the new circle, from the
given point p of the curve, must be, nearly
CXVII.
PC =
6r2r
Ur((T
^
- ^^
p)
SrrR
the coefficients being formed from those of the formula
- s, and
then changing the signs of the results

CXVI'. takes now the form,
CXVIII.
sc
SP
3 (sps - sp),
or
U((T
^
- ^^
p)
CXV., by
;
'
first
changing
while the relation
CXVIII'.
sc
BTps
s to
CXVI.
2sp.
going from p to Ps along the curve, we go forward

(69.) Accordingly if,
backward, through any positive or negative arc, t, of the circle which osculates
after
that point
Ps,
we
shall arrive at a point
which we
may here
denote by
Cs,
vector (comp. again 396, (18.)) of this near point (more general than
hitherto considered)
Avill be,
rigorously,
or
t ;
or
at
and the
any of
those
CXIX.
And
if
\so find
599
LOCUS OF OSCULATING CIKCLE.
CHAI'. III.]
w,
<
= 0C5, = ps +
<
TsTs sin
-t +
rs^T
vers -.
we develope this new expression to the accuracy of the /uur<A order inclusive,
that we satisfy the new condition Tcomp. (63.))
CXX.
and that then the expression
CXIX.
CXXL
when
Sr(w-p)=0,
<
= -s -
|^;
agrees with CXVII., within the order of ap'
proximation here considered.

(70.) A geometrical connexion can be shown to exist, between the two equivalents which have been found above, one for the quadruple (LXXXVII,, comp. (53.)),
and the other
for the triple
(CXVIII.), of the deviation sps "sp of a near point Pj

which osculates at the given point p in such a manner
of the curve, from the sphere
that
if
ferred from it.

(71.) In fact
if
two expressions be regarded
either of those
we draw,
in the
as
known, the other can be
normal plane, perpendiculars pd and pe
m-
to the
and determine points d and e upon them by drawing a parallel to

PS through the point c of (68.), letting fall also a perpendicular cf on ps^, the two
small lines PD and DC will ultimately represent the two terms or components CXVII.
lines
of
PS and
PC
PSs,
and the small angle dpc will ultimately be equal to three quarters of the
sps*-, and will correspond to the same direction of rotation round r, be-
small angle
cause
CXXII...^ = |.^=fV-^,
or
CXXIII.
we
so that
DPC=
|spSs
= |dpe
have the ultimate ratios (comp. the
shall
annexed Fig. 83*):
CXXI V.
But the
line
cf
is
DC de ce
:
(or FP)
=3 4
:
1.
ultimately the trace, on the given
normal plane, of the tangent plane at c to the near

osculating sphere; the small line fp (or ce) represents therefore the deviation
s^P-SsPs of the given point p from that near sphere, or the equal deviation (57.)^
its ultimate quadruple, de, represents the product mentioned in (52.) ;
SPs SP
;
and the ultimate
triple,
dc, of the same small line ce, is a geometrical representation

sp, which has been more recently considered.
of that other deviation sc
(72.)
When
the two scalars, s and
t,
are supposed capable of receiving any vaLocus (8.) of the Osculating
the point Cs,t in (69.) may be any point of the

and if
Circle to the given curve of double curvature
lues,
normal
to this superficial locus,
we
seek the direction of the
at this point, on the plan of Art. 372, writing first
the equation of the surface under the slightly simplified, but equally rigorous form,
* In
s is to be conceived as terminating there,
Figs. 81, 82, the little arc near
preceding arc of the curve which is the locus of s, if r', r, n, and there-
or as being a
fore also
p and
ri,
he positive (comp. the second Note to page 574). In the new Fipde is to be conceived as being in fact much smaller than
gure 83, the triangle
PKs, though magnified to exhibit angular and other relations.
600
ELEMENTS OF QUATKHNIONS.
with
CXXV.
tt=pt + TsTs sin + Ts^t's vers m,
CXXVI. .u = rs-H = pKsCs,
.
so that u
here a
new
III.
(jjs,
is
[bOOK
t,
scalar variable, representing the angle subtended at the cen-
tre Ks, of the osculating circle at p^,
by the arc,
we
of that circle,
t,
are led, after a
few reductions, to the expression,
CXXVII.
V(DWs, M
I>s(jJs^
which proves, 6y quaternions, what was
= TsTg-^ (w, - (Ts)
u)
vers
be expected from geometrical* conside-
to
rations, that the locus
of the osculating circle

the Envelope of the Osculating Sphere.
is
also (as stated in (8.)
and (22.))
The normal to this lociis, at any proposed poiCtt Cs, t of any one osculating
thus the radius of the sphere to which that circle belongs, or which has the
same point of osculation Pg with the given curve, whether the arc (s) of that curve,
(73.)
circle, is
and the arc
(t) of the circle,
be small or large.
We
must
therefore consider the tan-
gent plane to the locus, at the given point p of the curve, as coinciding yvith the tangent plane to the osculating sphere at that point and in fact, while this latter plane
;
(J- Ps) contains the tangent r to the curve, which is at the same time a tangent to
the locus, it contains also the tangent r(<T-(o) to the sphere, which is by CXVII.
another tangent to the locus, as being the tangent at p to the section of that surface,
which is made by the normal plane to the curve.
(74.) But when we come to examine, with the help of the same equation CXVIL,
is the law of the deviation DC (comp. Fig. 83) of that normal section of the
what
locus, considered as
is
a new curve (c),/rom
its
own tangent pd, we
find that this law
ultimately expressed (comp. (71.)) by the formula,
CXXVIII
:=î^'
PD*
32
^;s
R^
const.
;
'
hence i>c varies ultimately as the power of pd, which has the /rac^zon |- for its expothe limit of pd^ DC is therefore null, and the curvature of the section is infinite
;
nent
atT.
p is a singular point of the curve (c), in which

(75.) It follows that this point
theZocM* (8.) is cut (73.), by the normal plane to the given curve at that point but
it is not a cusp on that section, because the tangential component pd of the vector
;
chord PC is ultimately proportional to an odd power (namely to the cube, by CXVII.,

comp. (71.)) of the scalar variable, s, and therefore has its direction reversed, when
that variable changes sign : whereas the normal component DC of the same chord re
is
proportional to an even power (namely the fourth, by the same equation CXVII.)
same arc, s, of the given curve, and therefore retains its direction unchanged,
of the
when we
P.s
pass from a near point
on the other side of
Ps,
on one side of the given point
p, to a near point
it.
(76.) To illustrate this by a contrasted case, let G be the point in which thetansecgent to the given curs'ê at p* is cut by the normal plane at p ; or a point of the
shall then have
tion, by that plane, of the developable surface of tangents.
We
* In the
language of infinitesimals, two consecutive osculating spheres, to any
curve in space, intersect each other in an osculating circle to that curve.
CHAP. HI.] ENVELOPE OF SPHERE WITH VARYING RADIUS.
601
the sufficiently approximate expressions,
CXXIX. ..pq = p,-p-Is + \t,=
= -PQ2-2rQ3,
-^
with the significations 397, (10.) of Qo and Q3; hence the point p of the curve is
(as is well known) a cusp of the section (g) of the developable surface of tangents
(comp. 397, (15.)), because the tangential component ( PQ2) of the vector chord
(pg) has here Si fixed direction^ namely that of the outward radius (kp prolonged)
while it is now the normal component ( 2PQ3)
of the circle of curvature at p
:
which changes direction^ Avhen the arc

time
we
pressed
s of the curve
changes sign.
see* that the equation of this last section (g)
At
the
same
ultimately be thus ex-
may
_,__^_.
CXXX.
(- 2pq3)2
^
-^
8pk
9r2
(- PQ2)3
const.
:
'
comparing which with the equation CXXVIII,, we see that although, in each case,
the curvature of the section is infinite, at the point P of the curve, yet the normal
component (or co-ordinate) varies (ultimately) as the power | of the tangential component, for the section (g) of the Surface of Tangents : whereas the former component varies
of the
the power f of the

bj-- (74.) as
Locus of the Osculating Circle.
latter, for the
corresponding section (c)
(77.) It follows also that the curve (p) itself although

the last-mentioned locus (8.), while
it
is
it is not a
cusp-edge of
such on the surface of tangents, is yet a
the nature and origin of which line will
Singular Line upon that locus likewise :

perhaps be seen more clearly, by reverting to the view (8.), (22.), (72.), according to which that Locus of a Circle is at the same time the Envelope of a Sphere.
(78.) In general,
if
we suppose
that
vector and scalar kinds, of any one real

a",
<r
and
are
and scalar
ang two real functions, of the

and that a', R', and
t,
variable,
R'\ &c. denote their successive derivatives, taken with respect to
it,
then a
be conceived to be the variable vector of a point s of a curve in space, and

to
be the variable radius of a sphere, which has-its centre at that point s, but alters generally its magnitude, at the same time that it alters its position, by the motion of
may
its
centre along the curve (s).
(79.) Passing from one such sphere, with centre s and radius R, considered as
given,
and represented by the scalar equation, f

(^
- p)2 + i?2 = 0,
LXXXIX.,
which p is now conceived to be the vector of a variable point p upon its surface,
to a near sphere of the same system, for which a, s, and R are replaced by ot, St, and
Rt, where t is supposed to be small, we easily infer (comp. 386, (4.)) that the equa-
in
tion,
S(r'(<r-p)+i?72'
which
is
formed from
LXXXIX.
= 0,
by once derivating
XCI.,
tr
and
with respect to
t,
but

t This equation, and a few others which we shall require, occurred before in this
series, but in a connexion so different, that it appears convenient to repeat them
here.
602
[booK
III.
treating p as constant, represents the real plane (comp. 282, (12.)) of the (real or
imaffinart/) circle, which is the ultimate intersection of the near sphere with the
given one
we
the radius of this ciVc/e, which
shall call r, being found
by the
follow-
ing formula,
CXXXI.
r2(T'2
= R2 (i2'2 +
CXXXl'.
or
(t'),
r^Ttr'^
= H^ (T<r'2 -
'*),
and being therefore real when
CXXXII. ..R'^+(7'^< 0,
while the centre, say k, of the
circle is
CXXXI".
and the plane XCI. of the same
CXXXir.
or
always
real,
OK = = (T +
and
iJiJ'ff'-'
AC
circle is parallel to the
i2'8
its
< T<r'2
vector
is,
normal plane of the curve
(s).
With
(80.)
the condition
is easily
386, (6.))
represented
CXXXII.,
XCI., represent than jointly a real
and the
circle ;
LXXXIX.
and
locus of all such circles (comp.
proved to be also the envelope of all the spheres, of which one is

LXXXIX. alone ; each such sphere touching this locus,
by the equation
in the whole extent of the corresponding circle of the system.
(81.)
The platie XCI., considered as varying with
for its envelope ;
and the real right
line,
has a developable surface

t,
or generatrix, along which one touches the
other, is represented (comp. again 386, (6.))

tions,
by the system of the two
scalar equa-
XCI. and
Sff"((r-p)
<T'2
(ii;i2')'
= 0,
XCIIL;
whei'e p is now the variable vector of the line of contact, although it has been f reaped
as constant (comp. 386, (4.)), in the process by which we are here conceived to pass,
by a second derivation, from LXXXIX. through XCL to XCIII.
(82.) This real right line (81.) meets generally thepêre, and also the ciVc7c (as
its plane), in two (real or imaginary) points, say Pi, P2 ; and the curvilinear
being in
locus of all such points forms generally a species of singular line,*
upon the
superfi-
forms in general two

which generally two-branched line (or
cial locus (or envelope) recently considered (80.)
or rather
it
branches (real or imaginary) of such a line :

curve) is the (real or imaginary) envelope (comp. 386, (8.)), of
all
the circles of the
system.
(83.)
The
equation,
S(T'(T"(<r-p)
XCII.,
0,
which now represents (comp. 376, V.) the osculating plane
to the curve (s), shows
* Called
by Monge an arete de rebroussement, except in the case to which we
next proceed, when its two branches coincide. The envelope (80.) of a varying
shall
sphere has been considered in two distinct Sections, XXII. and XXVI., of the
Application de V Analyse a la Geometric; but the author of that great work does
not appear to have perceived the interpretation which will soon be pointed out, of the
condition of such coincidence.
Meantime
it
may
be mentioned, in passing, that quawhen the two branches (pi)
ternions are found to confirm the geometrical result, that
(Pa) are distinct, then each is a cusp -edge of the surface

coincident, the singular line (p) in
but that when they are
which they merge has then a
different character.
CHAP.
CONDITION OF COINCIDENCE OF CUSP-EDGES. 603
III.]
that this plane through the centre s of the sphere
is
perpendicular
to the right line
contains the perpendicular let fall from that centre on that

(81.), aiid consequently
line: thQ foot p of this last perpendicular is therefore found by combining the three
linear
and
scalar equations,
XCI., XCII., XCIIL, and
CXXXm.
op = p = <T-f
vector
Jtf^^"^
..g=-(T'i-JR'i- RR" =
CXXXI V.
if
its
is,
^'^
T(t'2
>
- {RE)\
or
(84.) T\\Q condition of contact of ihQ right line (Ji\.) with the sphere (78.),
with the circle (79.), or the condition of contact between two consecutive* circles Of
the system (80.), or finally the condition of coincidence of the two branches (82.)
upon the surface which is touched by all those circles^ is at the

same time the condition of coexistence of the/owr scalar equations, LXXXIX., XCI.,
XCII., XCIII. it is therefore expressed by the equation (comp. CXXXIIL),
of that singular line
CXXXV.
which
may
or thus,
i?2 (V(rV")2
= {ga + RR'a'y
also be thus written,f
CXXXVI.
CXXXVII.
the scalar variable
maining
still
(78.),
- i2V)2 = {E^ + (r'3) (^ma"^ + ^),

= {ga'^ + RR'ScrVy
+
<t'2) (Ya'ay
R:^{R'-^
(i2S(T'(T"
with respect to which the derivations are performed, rewhich is determined by the formula
entirely arbitrary, but the point p,
CXXXIII., being now situated on both the sphere and the circle : and its curvilinear
locus, which we may call the curve (p), being now the singular line itself, in its re-

t In page 372 of Liouville's Edition already
cited, or in
page 325 of the Fourth
Edition (Paris, 1809), of the Application de V Analyse, Sec, it will be found that
this condition is assigned by Monge, as that of the evanescence of a certain radical,
under the form (an accidentally omitted exponent of
member being
here restored)
[a(^>" + xp'V + ttV") -h-^y + h^ [a2(^"2

in
which he writes,
tt" in the
second part of the
first
+ 4,"^ +
7r"2)
- A*] =
for abridgment,
h^=:l-<p'2-ily'^-7r'^,
and
co-ordinates of the centre of

0, ^, TT are the three rectangular
considered as functions of
its
radius a.
Accordingly,
if
we change
a moving sphere,
to a,
and a to
= =
= =
^j^ 4- JcTT, supposing also that ig* a' 1, and R" a" 0, whereby g is changed
to - A2, and R"^+ o'"- to ^2, in the condition CXXXVI., that condition takes, by the
i^
rules of quaternions, the exact
sake of reference,
we
cited in this
shall call, for the present, the
Note
which, for the
Equation of Monge, although
does not appear to have been either interpreted or integrated by that illustrious
Indeed, if Monge had not hastened over this case of coincident branches,
on which he seems to have designed to return in a subsequent Memoir (unhappily
it
author.
not written, or not published), he would scarcely have chosen such a symbol as A?
- A3), to denote a quantity which is essentially negative, whenever (as
(instead of
here) the envelope of the sphere
is real.
604
duced and one-branched
And
state.
be expected from geometry, that

earlier condition
of reality
CXXXVII.
the last form
when
CXXXII.
also
III.
shows, what was to
of coincidence
this condition
is satisfied
[bOOK
is satisfied,
the
together with this other in-
equality,
CXXXVIII.
which then
results
from the form
(85.) The equations

give the expressions,
.. J22(T"2+i72<0,
CXXXVI.
CXXXI., CXXXIV., and
the general formula 389, IV.,
r is still the radius of the circle of contact of the sphere with its envelope^ and
the radius of curvature of the locus of the centre s of the same variable sphere;
whence it is easy to infer, that the condition CXXXV. may be reduced to the fol-
where
Ti is
XXXVI'. and XLII.)
lowing very simple form (comp.
CXLI.
(rVi)2
= {RRy
CXLl'.
or
ndr = + EAR ;
the independent variable being still arbitrary.

(86.) If the arc of the curve (s) be taken as that variable
of the
same condition
CXLII.
is
i22
t,
the form
CXXXVI,
easily reduced to the following,
= (^RR'f + ^zr^s,
derivating then, and dividing
by
Ig,
CXLIII.
with
we have
this
new
.g=l- {RRJ
which
is
of linear form with respect to RR', whereas the condition itself m^y be considered as
a differential equation of the second degree, as well as of the second order,*
CXLIV. ..RR^ = ri {gn)'

CXLVI. .u = RR' = RDtB,
;
if
or
by CXXXI.
or
CXLIV.
or
+ r,ri'(u' - 1) + = 0,
CXLVII. .u^=R^-r^,
riû"
therefore
CXXXI'., because we have now,
CXLVIII.
80 that the
CXLV.
and
new scalar
CXLV.
is
(t'2
= -1,
or
Ta'=l,
or
d<
= Td(r:
variable, RR*, or u, with respect to which the linear equation
only of the second order, represents the perpendicular heightf
of the centre s of the sphere, above the plane of the circle, considered as
a,
function
curve (s), and as positive when the radius
of the sphere increases, for positive motion along that curve, or for an increasing value of its arc.
(87.) If the curve (s) be given, or even if we only know the law according to
of the arc
which
its
equation
(t) of the
radius of curvature (r{) depends on its arc (f), the coefficients oi the. linear
are known ; and if we succeed in integrating that equation, so as to
CXLV.
We shall
ternions (84.),
soon assign the complete integral of the differential equation in quaalso that of the correspondmg Equation of Monge, cited in the
and
preceding Note.
t It will be found that this
precisely to the
new
scalar u,
if
we
abstract from sign, corresponds
of earlier sub-articles, although presenting itself in a different con-
: for the
sphere (78.), and the circle (79.), under the condition (84.), will
soon be shown to be the onculating sphere and circle to the recent curve (p), or to
the singular line (84.) upon the surface at present considered, that is, ou the locut
nexion
or envelope (80.)-
CHAP. Ill] DETERMINATION OF
THE SINGULAR
605
LINE.
find an expression for the perpendicular u as a function of that arc t, we shall then
and r of the sphere
be able to express also, as functions of the same arc, the radii
and
circle,
by the formulae,
CXLIX. ..+r = <7ri=ri(l-tt'),
and
CL.
i22
= 2 jMd<= tt2 + ri2(l
-m')2
which the integral 2jMd< would otherwise introduce into

the expression for R^, being in this manner determined, by means of the other two,
which arise from the integration of the equation above mentioned.
the third scalar constant,
(88.) For example, it may happen that the locus of the centre s of the sphere
has a constant curvature, or that ri = const. and then the complete integral of the
linear equation CXLV. is at once seen to be of the form,
;
CLI.
= a sin fri-i^ +
a and b being two arbitrary (but scalar) constants
6),
after
which we
may
write,
by
(87.),
CLII.
= ri - a cos (n'l^ + 6)
CLIII.
and R, of
so that, in this case, both the radii, r
R'^
= rx^ -
2ari cos
(rrH +
ft)
a2 ;
and sphere, are periodical
ci7'cle
functions of the arc of the curve (s).

(89.) In general, if that curve (s) be completely given, so that the vector cr is a
known function of a scalar variable, and if an expression have been found (or given)
for the scalar
which
satisfies
same variable; and so
of the forms of the condition (84.),
any one
we cau
by the formula CXXXIIL, as a function of the

can assign the point p of the singular line (84.), which cor-
then determine also the vector
p,
For this purpose we

responds to any given position of the centre s of the sphere.
have, when the arc of the curve (s) is taken, as in (86.), for the independent variable t, the formula,
CLIV.
= MO-' - (1 m') <t"-i = Ki u(j rû'a",

<r
be the vector of the centre, say Ki, of the osculating circle at s to that given
curve, so that (comp. 389, XI.) it has the value,
if K\
CLV.
OKI
= Ki = ff - <r"-^ = + riêr",
CLV.
with
<T
a"^
r^r^
= 0.
we
denote by v the distance of the point p from this centre Ki, and attend to
the linear equation CXLV., we see that
If then
CLVI.
and
or
CLVI'.
more generally,
if
t;
= K^ = Tfp - Ki) = V(2 + ri2'2),

w' = riri'ui, with T(t' = 1
.
CLVII.
cw'si'
CLYW. .u = RR'si'-\
and
CLVI".
while
u2=
u^
so that i denotes the arc of the curve (s),
This distance,
when the radius of curvature
left arbitrary.
v, is
= rir'iw',
CLVII". ..si = JTdcr,

ri2'25i'-2
therefore constant
ri of
that curve
(= a)
is itself
variable
t is
in the case (88.),
again
namely
a constant quantity.
When si=T(t'= 1, as in CXLVIII., the part - ua' of the first expresCLIV. for p becomes = k, by CXXXI". and CXLVI. attending then to CLV.,
(t
(90.)
sion
when the independent
we have
tiie
scalar quotient,
CLVin.
'i^ = l-u'i
(T Kl
whence generally.
606
being again arbitrary.
Accordingly,
general expression CXXXIII. for p, with the expression

the following for k\ (comp. 389, IV.),
CLIX.
we
easily deduce this
Ki
III.
we combine
if
CXXXI".
for
/c,
the
and with
(t'3
(T
new form
CLIX'.
[bOOK
an arbitrary scalar variable,
of the scalar quotient,
^^ =
a
for
,,
Ki
+ ({RRJ - MR'Sa'-W) a'-^
- a'2 = si
which agrees with CLVIII'., because
2,
and S
= -7-.
8\
(7
(91.) It has then been fully shown, how to determine the vector p as a. function
are two known functions of that variable, which satisfy
of the scalar t, when er and
of the forms of the condition (84.).
any one
also the derived vectors, p, p",
S;c.,
must then be
It
as functions of the
possible to determine
same variable; and accord-
by derivating any three of the/owr scalar equations, LXXXIX.

XCI. XCII. XCIII., of which that condition (84.) expresses the coexistence. Now
if we derivate a first time the two first of these, and then reduce by the second and
ingly this can be done,
fourth,
we
CLX.
get the equations,

.
and although
Sp' (o-
- p) = 0,
Sp
V=
0,
whence
CLX'.
p'
||
V(t' (jr-py,
formula only determines the direction of the tangent to the singular line at p, namely that of the common tangent at that point to two consecutive
circles (84.), yet it enables us to infer, by the remaining equation XCII., that
this last
CLXI.
p'-^
<r'\
II
and
VffV",
rednciog by which the derivative of XCIII.,

Str'"
the scalar variable being
- p) +
(a
still
we
Sp'(T"=
+ {RRy = 0,
XCIV.,
A.n6. conversely,
the system* oiilxQ four
SSctV"
arbitrary.
CLXI'.
find,
equations LXXXIX. XCI. XCIII. XCIV. gives the three equations CLX. CLXI'.,
and so conducts to the equation XCII., and thence to the condition (84.) unless we
;
suppose that p
is
a constant vector
a,
or that the variable sphere passes through a
fixed point A, a case which we do not here consider, because in

(p) would reduce itself to that one point.
(92.) Derivating the two equations CLX.,
CLXI'., we find these new equations,
CLXIL
Sp"(<r
CLXIII.
whence
the singular line
and reducing with the help
- p) - p'2 = 0,
Sp"'(<r
it
=
Sp"<r'
- p) - 3Sp'p" = 0.
of
* In the
language of infinitesimals, this system of equations expresses that four
consecutive spheres intersect, in one
common point
to be a fixed one, the condition (84.) requires that

S(T'<7"((r
a)
p.
When
that point happens
we should have
the relation
or geometrically, that the curve (s) should be in a plane through
the fixed point, which
is
then a singular point of the envelope.
CHAP. HI.]
ENVELOPE OF OSCULATING
We
by elimination of the
are led then,
SPHEliE.
607
derivatives of(T, to the system of the three
equations 395, VII. ; and we conclude, that the point s is the centre^ and the radius
is the radius^ of the osculating sphere* to the
whence it is easy
singular line (p)
to infer also, that i\iQ plane of contact (79.) of the sphere with its envelope is the
and that the
osculating plane y
(72.)), to the
same curve
circle
of contact (80.) is the osculating circle (corap.

where two consecutive circles touch one
(p), at the point
another (84.).
(93.) In general, and even without the condition (84.), the tangent to a branch
(82.) of the curvilinear envelope of the circles of the system, at any point Pi of that
branch, has the direction represented by the vector Y(t'((t- pi), of the tangent to the
circle at that point
but when that condition
of the singular line coincide, the point

to the
so that the two branches
is satisfied,
p of that
line is in the osculating
plane (88.)
curve (s)
and then the equation XCII. shows that the tangent p', or
p), to the line, is perpendicular to <t", or parallel to Vcr'o-"
(comp. CLXI.),
:
V(j-'((r
and therefore that the singular line crosses that plane at right angles.
(94.) It follows that, with the condition (84.), the singular line (p)
an ortho-
is
gonal trajectory to the system of oscidating planes to the curve (s) ; and whereas,
when this last curve is given, there ought to be one such trajectory for every point
0^ a. given osculating plane, this circumstance is analytically represented, in our recent calculations, by the biordinal form of the differential equation
CXhY.,
of which
the complete integral must be conceived (87.) to involve generally, as in the case
(88.), two arbitrary constants.
same condition of coincidence of branches,

must have the curve (s) for the cusp-edge of its polar developthat the sphere, with s for centre, and with H for radius, must be the osca-
(95.) It follows also that, with the

the singular line (p)
able
or
lating sphere to the curve (p), as otherwise found

circle (80.)
must
by
calculation in (92.)
while the
be, as before, the osculating circle to that curve.
(96.) Accordingly, all equations, and inequalities, which have been stated in the
recent sub-articles (79.), &c., respecting the envelope of a
moving sphere with va-
and without any
special selection of the
riable radius, under that condition
(84.),
independent variable, admit of being verified, by means of the earlier formula9 for
the osculating circle and sphere to a curve (p) treated as a given one, when the arc
() of that
curve
is
taken as such a variable.
(97.) For example,
and
+^3 <
we had
lately the two inequalities,
CXXXVIII.
R"^+
a'^
<
0,
CXXXII.,
And
accordingly the earlier sub-articles (22,),

(23.) give, for those two combinations, the essentially negative values,
iE2(T"2
CLXIV.
o,
ij;'2
ff'2
= - p-2r22j'2
CLXV.
Râ'^+g^ = - (Snr)y
* In the
language of infinitesimals (comp. the preceding Note),
if
every four
consecutive spheres of a system intersect in one point of a curve, then each sphere
Simple as this geometrical
passes through four consecutive points of that curve.
reasoning
is,
the writer
is
not aware that
it
has been anticipated
and indeed he
is
at present led to suppose that this whole theory, of the Locus of the Osculating
Monge had however
Circle, as the Envelope of the Oscidating Sphere, is new.
considered, but rejected (page
circles
374
of Liouville's Edition), the case of a system of
having each a simple contact with a curve in space.
608
which
in obtaining
CLXVI.
last,
.
(t"2
the following transformations have been employed
= - w'2 _ n2r-2
CLXVII.
.g = -n'p^ nrt-\
As
(98.)
example
III.
[bOOK
regards the verification of the equations, it may be sufficient to give one

and we shall take for it the last general form CLVII. of the differential
For this purpose we
equation of condition (84.).
may now
write,
by
(22.) and
(23.),
CLXVIII.
and have only
si'
= + ,
CLXIX.
XCIV., then
'
= +p',
riiV'=/>Vin"'
because
^(/;2-|-p'2i.2)'=:p'r(r +j9'r)',
we denote by
(99.) If
M = ji>,
ci, C2, cz
besides the equation
the
first
members
LXXXIX.,
which
but
p = r'r.
of the equations XCI.,
may
= for the whole of

i2, we have ci
we have not also c^ = 0, except for a
finition of the radius
lope (80.)
=j3'r;
to observe that
XCIIL,
be regarded as a mere de-
the superficial locus or enve-
point on one or other of the
And
two (generally distinct) branches of the singular line (82.) upon that locus.
cut the surface by a plane perpendicular to
if, at any other and ordinary point, we
the circle at that point,
we
find,
by a process
of the
same kind as some which have
been already employed, expressions for the tangential and normal components of the
vector chord, whereof the principal terms involve the scalar c-z as a, factor, while the
latter varies (ultimately) as the ^g-ware of the former, so that the curvature
section
is
finite
and known, but tends
to
become
infinite
when
cg
of the
tends to zero.
(100.) If the condition of coincidence (84.) be not satisfied, so that the two
branches of the singular line (82.) remain distinct, and that thus C2=0, but not
C3
(comp. (91.)), for any ordinary point on one of those two branches, then
if Ave
by a plane perpendicular to the branch, or to the circle

we find an ultimate expression for the vector chord which
cut the surface at that point
which touches
it
there,
involves the scalar c^ as a factor, and of which the normal component varies as the
sesquiplicate
power of the tangential one
so that
we have
cubical cusp, and each branch of the singular line
here the case of a semi-
a cusp-edge* of the surfiice,

exactly in the same known sense (comp. (76.)) as that in which a curve of double
curvature is generally such, on the developable locus of its tangents.
is
(101.) But wlien the condition (84.) is satisfied, so that the two branches coinand that thus (comp. again (91.)) we have at once the three equations,
cide,
CLXX.
..ci
C2=0,
0,
03=0,
then the terms, which were lately the principal ones (100.), disappear
and a
neio
expression arises, for the vector chord of a section of the surface, made by a plane
perpendicular to the singular line, which (when we take t = s, as in (96.)) is found
CXVII., and of course conducts to

same system of consequences the tewêniiaZ component now varying
the cube, and the normal component ag the fourth power of a small
to admit of being identified with the formula
precisely the
ultimately as
variable, so that the cuspidal property of the point
p of the
section
no longer
exists*
although the curvature at that point is still infinite, as in (74.) and the Singular
Line, reduced now to a single branch, to which all the circles of the system osculate,
:
Compare
the Note to page
CHAP.
INTEGRAL OF THE EQUATION OF CONDITION.
III.]
609
a cusp -edge of the Surface, as had been otherwise found before

line of a different character* which may thus be regarded, with refea
but
(77.),
rence to a more general Envelope (80.), as the result of 3l Fusion (84.) of Two Cuspnot
(92.)> (95.), is
Edges.
The condition
(102.)
expressible
of sxxch fusion (or coincidence) has been seen (84.) to be
and second
differential equation of the second order,
by the
- R'gY =
iRS(T'<T"
(ir + (^2)
(i22(T"2
g = -(T'i-{RRy,
with
And we
and with the independent variable arbitrary.
are
now
the complete general integral^ of this differential equation
the two following equations (comp. 395, (7.) and (14.)), of
degree,
CXXXVI.
CXXXIV.
+ g^),
prepared to assign
namely the system of

the vector and scalar
kinds,
CLXXI.
in
which p
press,
is
= ^^ ^^''^'l'y;,\yyi,
CLXXII.
and
i2
= T(cr - p)
an arbitrary vector function of any scalar variable, t, and which exinterpreted, that a is the variable vector of the centre s,
when geometrically
R is the
and that
of
variable radius, of the osculating sphere to
which the variable vector of a point p is p.

(103.) In fact, if we met the cited equation of condition
a.n
arbitrary curve (p),
CXXXVI., g
represent-
without any previous knowledge of its meaning or origin, we might first, by the rules of quaternions, and as a mere affair of
calculation, transform it to the equation CXXXV. ; which would evidently allow
ing therein the expression
CXXXIV.,
the assumption of the formula CXXXIII., p being treated as an auxiliary vector,

which satisfies (in virtue of the supposed condition) the system of the foîr scalar
equations,
in (91.)
whence derivating and combining, as

new systemX of ^Mr scalar eqtiations, whereof one
LXXXIX., XCI., XCII., XCIII.
and
(92.),
we
are led to a
Compare the Note to page 602. Monge (in page 372 of Liouville's Edition) has
(when a certain radical vanishes) "les deux branches de la courbe
the remark, that
touchee par toutes les caracteristiques se confondent en une seule et cette courbe,
sans cesser d'etre une ligne singuliere de la surface, n'est plus une arete de rebrous:
The propriety of this last name, "line of

sement, elle est une ligne de striction."
although he has confirmed, as
striction," appears to the present writer questionable
referred to, the
above, by calculations with quaternions, the result that, in the case
:
singular line
the
is
not a cusp-edge.
lated,
by
all
line in question is
not merely touched, but oscu-
the circles of the system.
t Compare the
less
does not seem to have perceived that, in
Monge
same case offusion, the curved

first
Note
to
We
page 604.
say here, general integral, because a
general one, although involving one arbitrary function (of the scalar kind),
soon be pointed out.

Note to page 603) may be consiX The Equation of Monge (comp. the second
dered as the condition of coexistence of the /or following equations, in which 0, i//,
will
77
are supposed to be functions of ,
and
to be differentiated or derivated as such
610
[bOOK
III.
again the equation LXXXIX., and may be written under the form CLXXII.
while the three others are those formerly numbered as 395, VII., and conduct (except in a particular case which we shall presently consider) to the vector expression
is
CLXXI., which
conversely
is sufficient
to represent them, all derivatives of
<t
and
of
R being thus eliminated.

(104.) The case just now alluded to, in which the general integral (102.) is reby a less general form, is the case (91.) when the variable sphere passes
through &. fixed point A, to which jaomf, in that case, the singular line reduces itself.
placed
And
the integral equations,* which then replace
thus written
CLXXIIL
(1).
(2).
(3).
(4).
(T
=a+
</3
+ y,
with u
CLXXI. and CLXXII., may
= F(t), and CLXXI V.
72
be
= T(^/3 + uy)
..(ar-^)2 + (y-4/)2 + (z-7r)2=:a2;
+ (y-4^)^p'+ (z-7r)7r' + a = 0;
2
..(x- (p)<p"+{y ^) V'+Q^ 7r)7r"+ 1 - f _;p'2_ ,r'2=
.(x-ip) (xjj'ir" tt'-^") + (y 1//) (ttV" 0V") + (z 7r) (f'^p"- ^V) =
.
{x-<l>)(p'
have been employed by Monge himself, but the fourth does

have been perceived by him, the condition of evanescence of a radical
whereof the
first tliree
not seem to
its stead.
And by a translation of quaternion results, above
deduced, into the usual language of analysis, it is found that the complete and general integral, of the non-linear differential equation of the second order, which is ob-
having been used in
tained by the elimination of x, y, z between these four, is expressed by a new system

and the three others, in which
of four equations, the equation (1) being one of them
;
a;,
y,
z are
now
the following
treated as arbitrary functions of a,
- ^//)y' + (z - 7r)2' =
- <p)x" + (y - >p)y"+ (z - 7r)z" + x^ + y'^ + z'a = o
..(x- <p)x"' + (y - ^)y"' + (z - 7r)z'"+ Z\(x'x" + yy" + z'z") = 0.
(5).
{x-if)x'-\ (y
(6).
(a;
(7).
and are derivated as such, being
By treating a as a function of some other independent variable, t, the terms + a and

+ 1, in (2) and (3), come to be replaced by + aa and + aa" + a'^ and the slightly
;
more general form, which Monge's Equation thus assumes, has still its complete
general integral assigned by the system (1) (5) (6) (7), if x, y, z (as well as a) be
now regarded as arbitrary functions of the new variable t, in the place of which it is
=
only two arbitrary
pemiitted (for instance) to take x, and so to write x' \, x"
functions thus entering, in the last analysis, into the general solution, as was to be
expected from the form of the equation.
* T\\Q
particular integral corresponding, of the Equation of Monge, is expressed
:
by the following system

(p
= a + et + lu,
(et
luy-
= b+ft + mu, 7r = c^gt + nu,

+ {ft + muy + {gt + n) = a^
\l/
abcdefglmn being nine arbitrary constants, while t and u are two functions of
whereof one is arbitrary, but the other is algebraically deduced from it, by means
the fourth equation.
assigned before.
The writer
is
of
not aware that either of these integrals has been
CHAP.
VECTOR EQUATION OF LOCUS OR ENVELOPE.
III.]
611
an arbitrary function of the first

and , /3, y being three arbit,
the second scalar coefficient, , being here
scalar coefficient, or of the independent variable
trary but constant vectors
so that the curve (s)
is
now
obliged to
lie
in
some one
plane* through the fixed point A, but remains in other respects arbitrary. Accordingly it will be found that this last integral system, although less general than the
former system (102.), and not properly included in
tion
CXXXVI.
this
common
differential equait, satisfies the

whereof the two members acquire, by the substitutions indicated,
value,
CLXXV.
{Râ'a" - i22 = &c.
= i2-2f2 (<' _ )2 "2 (V/3y)*.
(105.) Other problems might be proposed and resolved, with the help of formulsef
already given, respecting the properties or affections of curves in space which depend
on
i)\Q
tor
fourth power (s*) of the arc, or on {h.Q fourth derivative D/p or r"' of the vecbut it is time to conclude this series of sub -articles, which has extended to a
ps',
much
greater length than was designed,
form 396, XI.

lating Circle
by observing
for the equation of a circle
may
that, in virtue of the vector
of curvature, the Locus (8.) of the Oscu-
be concisely but sufficiently represented
CLXXVI.
by the Vector Equation,
V -^^ +vs=0.

might for example employ the formula VI. for k", in conjunction with
one of the expressions 397, XCI. for k', to determine, by the general formula 389,
IV., the vector (say ^) of the centre of curvature of the curve (k), and therefore also
t
We
the radius
of curvature of that curve, which
is
the locus of the centres of curvature of
After a few
the given curve (p), supposed to be in general one of double curvature.

reductions, with the help of XII., we should thus find the equations,
CLXXVII.
CLXXVIIL
.. ^
=
V-,
K
k:
^^+
rK
-4 =
fc
(r-'
+ -2^-^-2^,
ds
k'
in
which
whence
last the
denominator
is
'
P')t,
rdfc
a quaternion, and the scalar variable
is
arbitrary
also,
CLXXIX. . Radius of curvature of curve (k),

or of locus of centres of osculating circles to a given curve (p) zn space,
.
Mr r/l
-ij^lVr
_
dPY^f pVVi
ds
'^{'Rrj
''
with the verification, that for the case of a plane curve (p), for which therefore
1'
and - =
CLXXX.
r
being
still
we have
ds
.
thus the elementary expression,
Radius of Curvature of Plane Evolute
the radius of curvature, and
rdr
the arc, of the given curve.
612
[bOOK
III.
which apparently involves only one scalar variable^ s, namely, the arc of the curve
t, which corresponds (69.) to the arc of the cir-
(p), the other scalar variable, such as
disappearing under the sign

the Circumscribed Developable,
cle,
is
now
and that the surface, which was called
cus, or
Envelope, in a certain singular (or eminent) sense, as touching

Singular Line.
When we
399.
take account of the Jifth power
the expression for

398,
receives a
p^
in
(8.)
seen to be in fact circumscribed to that Loit
along
its
of the arc,
(s^)
new term^ and becomes (comp.
I.),
I. ../>.
= /> + ST + I5V' +
^^T'f
.,1^T'''
+ ylo^V- ;
and although some of the consequences of such an expression have

been already considered, especially as regards the general determination of what has been above called the
Osculating Twisted Cubic
to a curve of double curvature, or the gauche curve
of the third degree which has contact of the fifth order with a given curve in space,
yet, without repeating any calculations already made,
tional light may be thrown on the subject as follows.
(1.)
it
may
addi-
As regards the successive deduction of the derived vectors in the formula

be remarked that if we write (comp. 398, LVL, LXL),
11.
we
some
DVV =
'^'^
=^ + ^nrr
-f
I.,
CnrVf
shall have, generally,

III.
with the
IV.
whence
a+i
= a'n - r-^bn,
6n+i
= h'n + r-ân - r-'c,
Cn+i -c'n
+ r->6,
initial values,
..ao=l,
60
= 0,
co=0,
/"^ -'"'''
\a3 = 3r-3f',
^2
or
= (r-0',
63
IV'.
..ai
= 0,
= (r-i)" - r-3 - r-ir-2,
as in the expressions 397, VI. for r", and 398, IV. for
efficients of r'^ being in like manner found to be,
= r-i,
ci
0;
= r (r-2r- )',
C3
t"
the corresponding co-
= - 2 (r-2)" + ((r-i)')2 + r-2(r-2 + r-2)

= (r-i)"' - 2 (r-3)' - 3 (r-ir-i)'r-i
64
= r-i (r-i)" + 3 ((r-O'r-i)' - '"'r"! (r-2 +
( Ci
/a4
VI.
6i
c^=r-^x-\
r-2)
and being sufficient for the investigation of all aflfections or properties of a curve
space, which depend only on the fifth power of the arc s.
(2.) For the helix the two curvatures are constant, so that all the derivatives
is
of
the expressions become therefore greatly simplified,

easily perceived, allowing us to sum the infinite series for ps, and so
the two radii r and r vanish
and a law
in
to obtain the following rigorous expressions for the eo-ordinates*
Xg,
Zs of this
* We have here, and in this whole investigation, an instance of the facihty with
which quaternions can be combined with co-ordinates, whenever the geometrical no-
CHAP.
OSCULATING TWISTED CUBIC TO HELIX.
III.]
particular curve, instead of those
but only as far as s* inclusive
VII.
where
and
ar,
613
which were developed generally in 398, LVIII.,
= P {x-H + r'Z sin
7/s= Pr' vers t;
Ze=
t);
Pr-î-^ (t
- sin )
are an auxiliary constant and variable, namely,
VIII.
(r-3
+ r-2)-i = r sin jy,
= ?*,
/ being thus what was

by T\-i, and t being the angle between two axial planes; while the origin is still placed at the point p of the curve,
and the tangent, normal, and binormal are still made the axes of xyz.
denoted in earlier formulae
The cone
(3.)
of the second order, 398, (40.), which has generally a contact of

a proposed curve in space, at a point P taken for vertex, has in
the fifth order with
this case of the AeZia; the equation
(comp. 398, LVII.* and LXIX.),
/3r
2r\ ^VlOr
3rr
^
Accordingly
it
X.
f(f)
3 (^
- sin
(3<
7 sin
if
we
write, for a
mo-
- 20 vers2 1,
the eight evanescent values,
XL
whence
can be shown, by elementary methods, that
ment,
we have
1 r
lOr
/O =/'0 =/"0 =/"'0 =/'^0 =/^0 =/^'0 =/^"0 =
IX. has
(in the present example, although

not generally) a contact as high as the sixth orderf with the curve, of which the
it is
easy to infer that this cone
co-ordinates have here the expressions VII.

tion
and consequently that the cone
must wholly contain the osculating twisted cubic
ture of a question
may render
it
in ques-
to that curve.
convenient so to combine them, by offering to our
any obvious planes of reference. If it be thought useful to pass to a system

connected more immediately with the right cylinder than with the helix, we may
notice
write,
l(r-^Xs
r-'z)
= l-r-^ sin tj
=l^r-i cost,
ys=Pr-i-ys
Zs = Z(r-ia;s + r-i2s) = /2r-i<,
fx,
where
l^r'^
= r sin^ H is the
radius of the cylinder, with converse formulae easily as-
signed.
* In the
corresponding equation 398,
have been printed as
LXVIL,
like the coefficient of
the coefficient of 6ac ought to
6xz in the equation LVII.
t Or
in
in this case,
modern language, seven-point
contact, in the sense that the cone passes,

It may be remarked
through seven consecutive points of the curve.
that the gauche curve of the fourth degree, or the quartic curve, in which this cone
cuts the cylinder of revolution whereon the helix is traced (cutting also in it a certain other cylinder of the second order),
and which has the point p for a double point,
by one of its two branches at that point, while it has seven-point

contact with the same helix by its other branch : and that thus the fact of calculacrosses the helix
tion,
expressed by the formula
XL,
is
geometrically accounted for.
614
[bOOK
III.
(4.) In general, to find a second locus for such a cubic curve, the method of
recent sub-articles (398, (38.) &c.) leads us to form the equation (398, LXVI.)
of a cylinder of the second order, or briefly of a quadric* cylinder, which like
the quadric cone (3.) shall have contact of the fifth order with the proposed
curve in space, at the given point p; the ratios of uhc, which determine the
direction of a generating line pe, being obliged for this purpose to satisfy a certain
equation of condition (398, LXVIIL), of which the /bnn indicates that the locus of
this line PE is generally a certain cubic cone, having the tangent (say rx) to the
curve for a nodal side : along which side it is touched, not only (like the quadric
cone) by the osculating plane (z

plane, whereof the equation (gy
= 0) to that
+ hz = 0, or
given curve, but also by a second

after reductions
y |r'z 0) shows
that the second branch of the cubic cone crosses the first.branch, or the quadric cone,
or the osculating plane to the curve, at an angle of which the trigonometric cotanis equal to halfj.he differential of the radius (r) of second curvature, divided
by the differential of the arc (s) ; so that this second tangent plane to the cone coincides with the rectifying plane to the curve, when the second curvature happens to be
The tangent pt therefore counts as three of the six common sides of the
constant.
gent
two cones with p for vertex
and the three other common
sides, for the assigning of
has been shown (in 398, (41.)) how to form a cubic equation in b: e, are
the parallels from that point p to the three real or imaginary asymptotes^ of the
twisted cubic, and are generating lines pe of three quadric cylinders, whereof one at
which
it
least is necessarily real,
and contains, as a second
locus, that
sought osculating gauche
curve of the third degree.

(5.) In applying this general method to the case of the helix, it is found that the
cubic cone breaks up, in this example, into a sj^stem of a new quadric cone, which
touches ihQ former quadric cone IX. along the tangent pt to the curve (the two other
sides of these two cones being imaginary^, and of a plane (y = 0), namely
the rectifying plane (comp. (4.)) of the helix, or the tangent plane to the cylinder of
revolution on which that given curve is traced and that this last plane cuts the
common
first quadric cone in two real right lines, the tangent being again one of them,
and
the other having the sought direction of a real asymptote to the sought osculating
Without entering here into details of calculation, the resulting equatwisted cubic.
on which that sought gauche curve

be at once stated to be (with the present system of co-ordinates),
tion of the realX quadric cylinder,
may
* So called by Dr. Salmon, in his Treatise

already
to page 591 of these Elements.
cited.
is situated,
Compare
the
first
Note
f Compare again the Note last referred to.

As regards the two imaginary quadric cylinders,
by the same general method, employing
mon
sides
which
is
(5.), of the
cone IX., and of that other quadric cone above referred
seems
to,
here a separable part of the general cubic locus, and has for equation,
IX'.
It
their equations can be formed
as generating lines the two imaginary com-
sufficient here to
tions of those
y y3 = 6 ^xz-l-f 3^-^- 2^2.
remark, that by taking the sum and difference of the equacylinder?, two new real quadric surfaces are obtained,
two imaginary
CHAP.
DEVIATION OF CUBIC FROM HELIX.
III.]
in such a
if we set aside the right line,
manner that
xm..., = o,
which
is
common
Gl5
side of the cone
.+
(l-l:).
= o,
IX. and of the cylinder XII., the curve, which is

is the twisted cubic sought.
As an
the remaining part of their complete intersection,
elementary verification of the fact, that this gauche curve of intersection IX. XI I,
has contact of the fifth order with the helix at the point P, it may be observed that
if
we change
the co-ordinates xgz in XII. to the expressions VII., and write for
abridgment,
XIV.
F(t')
= (3< +
we have then (comp. X. XI.) the
XV.
JTO
7t sin
ty - 200 vers t + 60
vers^
t,
six evanescent values,
= F'O = F"0 = i^"0 = FiÔ = FÔ =
0.
(6.) As another verification, which is at the same time a sufficient /)roo/ of the
a posteriori kind, that the gauche curve IX. XII. has in fact contact of i\\Q fifth order with the helix, it can be shown that while the co-ordinates y and of the latter
may (by
VII., writing simply x for

f
__,
and neglecting
Xs,
^ /3
_a:2
1 \
ri6
x"^)
be thus developed,
M5
24
^*"2;+24r V'''"r2J^7207V^
1\
r^x'^^"'7^r
"^
\_^'~ 6/r
120rr
the corresponding co-ordinates
y and z
Vr^*
oitliQ former, that
is,
of the curvilinear part
of the intersection of the cone IX. with the cylinder XII., have (in the
of approximation) developments
which
may
(r-2
same order
be thus abridged,
+ r-2)2 ^-6
(7.) The deviation of the helix from the gauche curve IX. XII. is therefore of
the sixth order (with respect to x, or s), and it has an inward direction, or in other
words, the osculating twisted cubic deviates outwardly from the helix, with respect
to the right cylinder
according to which
it
the ultimate (or initial)
amount
of this deviation, or the
law
tends to vary, being represented by the formula,
xvir...y.-y=^^::^^^ =
which also contain the osculating twisted
cubic,
and
^;
intersect each other in that
gauche curve namely two hyperbolic paraboloids, which have a common side at infinity, and of which the equations can be otherwise deduced (by way of verification), without imaginaries, through easy algebraical combinations of the two real
:
equations IX. and XII.
616
where
denotes as in (2.) the angle, which

of the right cylinder,*
di
[bOOK
III.
plane drawn through a near point
p,
and through the axis
XVIII.
whereon the helix
is
traced,
..2ry=fa:-^zy+
f l
!^'\y2,
makes with the plane drawn through the same
axis of
revolution, or through the right line,
XIX.
= -z,
.a;
y=
r-i(r-2
+ r-2)-i=ZV,
and through the given point p while ys is still the (iuAvard) distance of the same
p^, from the tangent plane to the same cylinder at the same given -point p.
:
near point
(8.) If
we
cut the cone IX., and the cylinder XII., by any plane,
XX...
2^=.
{..(11 -l:).}.
drawn through their common side XIII., we obtain two other sides, one for each of
and these two new right lines, in this plane XX., interthese two quadric surfaces
;
sect each other in
tions of the
new
xr^x
XXI.
a a new point, f of which the co-ordinates xyz are

given, as func-
variable w,
x=
"''^VrS
>
by the
r
r2
three fractional expressions,^
60
2ry
m;2
"^20/2
Qrrz
^3
^"^20"^
^"^20l2"
while the twisted cubic, which oscidates (as above) to the helix at p, is the locus of
all i\i& points of intersection thus determined.
Accordingly, if we develop xyz by
XXI.,
in ascending
powers of w, neglecting w^ (or a;"), we are conducted, by elimiy and z in terms of x, which agree with those found
nation of w, to expressions for
in (6.), and thereby establish in a

t\xQ fifth
order, between the
new way
the existence of the required contact of
two curves of double curvature.
* With the co-ordinates VII'. of a recent Note

(to page 612), the equation of
this cylinder
would
be,
XVIir.
t The plane XX.,
therefore
it
.x2 + y2 = ZV-2.
as containing the line XIII.,
and therefore meets the cubic at
and
is
parallel to
an asymptote,
through the -given point p

can only cut the twisted cubic in one other point, of which the posiinfinity
it
also passes
by the equations XXI.

Quaternions suggest such fractional expressions, through the formula 398,
LXXIX. for the vector (^ +c)"i a ; but it is proper to state that expressions of
fractional form, for the co-ordinates of a curve in space of the third order (or degree)
tion is expressed
were given by Mobius, who appears to have been the first to discover the existence
of such gauche curves, and who published several of their principal properties in his
Bary centric Calculus (der barycentrische

Notes to pages 23 and 35, and Note
Calcul,
at the
Leipzig,
1827).
end of these Elements.
Compare
CHAP.
The real asymptote
(9.)
variable
REAL AND IMAGINARY ASYMPTOTES.
III.]
to the cubic
curve
to tend to infinity in the expressions
is
found by supposing the auxiliary
XXI.
namely the second side

through the side XIII.
XIX., we
in
;
10/2
--,
which the
3 r
it is
(comp. XX.),
^.^^^
XXII..
., =
617
therefore the right line
r\
-+(iJ--i^p)"0,
XII.
elliptic cylinder
and by comparing the value of
is
its
cut by a normal plane
y with the equation
see that the least distance between the real
asymptote to the osculating

twisted cubic, and the axis of revolution of the cylinder on which the helix is traced,
equal to seven-thirds of the radius of that right cylinder,
ima(10.) As regards the two imaginary asymptotes, they correspond to the two
is
XXI. to
ginary values of w, which cause the common denominator of the expressions
vanish ; but it may be sufficient here to observe, that because those expressions give,
generally,
XXUh
+ l-]z = w,
x+llbr
\5r
j
the two imaginary lines in question are to be considered as being contained in two
imaginary planes, which are both parallel to the real plane* through
XXIV.
namely
to a certain
or to the elliptic
.x +
lr\
/6r
Urnr-)^
common normal plane
P,
to the two real cylinders XII.
and
XVII I.,
and right cylinders already mentioned.
(11.) In general, instea.d of seeking to determine, as above, a cylinder of the

second order, which shall have contact of the ffth order with any given curve of
double curvature, at a given point p,
we may
propose to find a second cone of the
same (second) order, which shall have such contact with that curve at that point,
vertex being
its
Sit
some other point of space
(âbc).
Writing (comp. 398, LXVI.)
the equation of such a cone under the form,
XXV.
2r(cy
substituting for
- bz) (c-z) = (ex - az^ + 25(caJ - az)
xyz the co-ordinates
{cy
- bz) + C{cy - bz^
under the forms (comp.
XgysZg of the curve,
398, LVIIL),
^*~
s^
+
6^2
=
XXVI...^v.
^*
1
azs*
which the
coefiicients 036303
ing according to powers of
we
s,
120'
--
+^
+
2r
120'
^24 ^,
6r3
ôTt'^'m'^
in
ttis^
"^
"24
120'
and Oib^c^ have the values assigned in (1.)

^, and comparing coefiicients of
neglecting
develop-
s^,
s\ s^
find first the expressions,
*
The right line at infinity, in this plane XXIV., is the common side of the
two hyperbolic paraboloids mantioned in the third Note to page 614, as each containing the whole twisted cubic.
4k
618
^^^TT
XXVII...
[bOOK
III.
b r\i
ar\ r^f
b \
-^1
^f\ r.
4:[
4[
+
i>=-^r.+--). c=-5^.'+--) +5^1 --)+-(^.3--C3),
,
-r.
which are the same
for cone as for cylinder
and then are led
to the
new equation of
condition,
xxvni...^(.-^^.)=.-^3.:^.(.-,*.-l-^^)
which
differs
from the corresponding equation for the determination of a cylinder

(fifth) order of contact with the curve, but only by the one term
having the same
in the second
err
member, which term vanishes when the co-ordinate
c of the vertex
is infinite.
(12.) Eliminating
and C, and substituting
for
V. and VI., we find that the condition XXVIII.

398,
LXVIII.)
acf 6
for
_ 4r
*~9;'
XXX.
their values
- rc2 = a63 +
hbê
cbc^
ec^
-f
abridgment,
^ ^
**
^"3~;2'
+
^ (6r"r- - 9r-irV"r2 + Ar-i/h^ +
=
c=
I
The
and 046404
be thus expressed (corap.
XXIX.
in
ctsftacs
may
3rr"
2r-^r'h
6r'r'
Grr'ir'z
86r-2r'r3
(9r"'r2
locus of the vertex of the sought quadric cone
XXV.
is
ISrir
12rr-');
+ 18/ - 27rr-ir').
therefore that cubic sur-
face, or surface of the third order, which is represented by the equation XXIX. in
abc; this surface, then, is a second locus (comp. (4.)) for the osculating twisted cubic,
whatever the given curve in space
being
still
may
be
a first locus for that cubic curve
the quadric cone (comp. (3.)), of which the equation in abc
is
(by 398,
LXVII.* and LXIX.),
XXXI...4(:)^ 6(^],
r*/ 9
(^y*.
21
r'2
3r"
3r'r'
"^6Vr4~^"^7*~73'"^737~
and which has contact of theîtA order with the
9r"\
27r'2
4^
"^
727
curve, while its vertex
is
j*^''
at the
given point p of osculation.
* After
making the correction indicated in a former Note (to page 613), so as to
bring the cited equation into agi-eement with the earlier formula 398, LVII. The
quadric cone XXXI. may be said to have five-side contact with the cone of chords
of the given curve
(compare the
first
Note
to
page 588).
CHAP.
RULED CUBIC LOCUS, CUBIC CYLINDER.
III.]
619
of the two radii of

(13.) Instead of thus introducing, as data, the derivatives
curvature, r and r, taken with respect to the arc, s, it may be more convenient in
many applications to treat the two co-ordinates y and z of the curve as functions of
the third co-ordinate x, assumed as the independent variable
and
so
to write
(comp. (6.)) these new developments,
XXXII...
V+V
y.=- +
and then the equation of the quadric cone
XXXIII.
XXXI.
y2
^=6-;;+ir
+120'
<2gyz
m'
become
will be found to
a:2
(in xyz),
+ hz^,
with the coefficients,
XXXIV.
^ = rr
y"'
rz-
fe
rr^
j,
- r2r2
XXIX.
while the cubic surface

the
same form as
before,
XXXV.
in
which the
will also
namely
.
jcz
(y
come
y"'2
+ - rz'^y"'
rV^2
;
j
by an equation
to be represented
of
xyz) by the following,
(in
+ hz) -
rz2
= ay^
-i-
hy'^z
+ cy22 ^ e_j3^
coefficients are,
a = - - (as before)
XXXVI. .A
r3ry"'2
=-
"
r2y
- |r3r2y"'z" - |r2ry^^ +
h = - rry'" + irr2'^
z'^;
^r2r2z^
_4
_ rh-2y"'3 +
ir3r2y"y^ -^r^xy-.
(14.)
rz
^^
(^y-
set of expressions for the coefficients we may adopt, some gemaybe drawn from the mere /or/ns of the equations, XXXI.
XXXIII. and XXXV., of the quadric cone and cubic surface, con-
Whichever
neral consequences
and XXIX., or
sidered as two loci (12.) of the osculating twisted cubic to a given curve of double
curv'ature.
Thus,
if
we
eliminate ac (comp. 398, (41.)) from
XXIX. by XXXI.,
by XXXIII. from XXXV., we get an equation between 6,

which rises no higher than the third degree, and is of the for?n,
or xz
XXXVII.
2r22
with the same value of a as before
= ay3 +
b^y^z
such then
c,
or between y, z,
+ c^z2 + e,z3,
the equation of the projection of

the twisted cubic, on the normal plane to the curve ; and we see that, as was to be
expected, the plane cubic thus obtained has a cusp at the given point p, which
(when we neglect
tion of the
(15.)
cylinder,
^ or a;') coincides
is
with the corresponding cusp* of the projecitself, on the same normal plane.
given curve of double curvature
The equation XXXVII. may

which
is
also be considered as representing a cubic
a third locus of the twisted cubic
Compare
the
first
formula of the
first
and on which the
Note
to
page 504.
tayigent
pt
620
to the curve is
line,
a cusp-edge, in such a manner that an arbitrary plane through
suppose the plane
XXXVIII.
where v
time in
[bOOK
.. 3rz
III.
this
= tjy,
any assumed constant, cuts the cylinder in that line twice, and a third
a real and parallel right line, which intersects the quadric cone in a point at
is
pt
infinity (because the tangent
is
a side of that cone), and in another real point,
on the twisted cubic, and may be made to be any point of that sought curve,
by a suitable value of v : in fact, the plane XXXVIII. touches both curves at p, and
therefore intersects the cubic curve in one other real point. And thus may fractional
which
is
expressions (comp. (8.)) for the co-ordinates of the osculating cubic be found genewhich we shall not here delay to write down.
rally,
(16.) Without introducing the cubic cylinder XXXVII., it is easy to see that
any plane, such as XXXVIII., which is tangential to the given curve at P,
cuts the cubic surface XXXV. in a section which may be said to consist of the
tangent twice taken, and of a certain other right line, which varies with the
direction of this secant plane, so that the locus XXXV. or XXIX. is a Ruled
Cubic Surface, with the given tangent pt for a singular* line, which is inby all the other right lines on that surface, determined as above and if we
tersected
set aside this line, the
remaining part of the complete intersection of that cubic surface with the quadric cone XXXIII. or XXXI. is the twisted cubic sought. We
may then consider ourselves to have completely and generally determined the Oscuculating Twisted Cubic to a curve of double curvature, without requiring (as in 398,
(41.)), the solution of any cubic or other equation, f
(17.) As illustrations and verifications, it may be added that the general ruled
cubic surface, and cubic cylinder, lately considered, take for the case of the helix
(2.), the particular forms,J
* If the cubic
surface be cut by a plane perpendicular
to the
tangent pt, at any
point T distinct from the point p itself, the section is a plane cubic, which has T for
a double point ; and this point counts for three of tlie six common points, or points of
intersection, of the plane cubic just
quadric cone
is
cut
mentioned with the plane conic in which the

plane, because one branch, or one tangent,
by the same secant
of the plane cubic at
t touches the plane conic at that
point, in the osculating plane
to the given curve at p, while the other branch, or the other tangent, cuts that plane
conic there.
f
zero,
It
may be remarked
and changing
c,
by equating the second member of XXXVII.

we obtain generally the cubic equation, referred
to
to
and that by suppressing the term rc2 in XXIX., or the term - rz'
XXXV., we pass, in like manner generally, from the cubic surface of recent sub-
in 398, (41.)
in
that,
y, z to b,
articles, to the earlier
By
cubic cone (4.).
suppressing the term
rz^,
case of the helix from the equation

in the last Note to page 614
dividing
XXXIX.
namely
by
ry
or
and transposing, we pass for
t^
of the cubic locus, to the equation IX'.

which forms
to the equation of that quadric cone
a separable part of the general cubic cone, the other part being here
(in this example)
the tangent plane (y = 0) to the right cylinder.
CHAP.
III.]
INVOLUTES AND EVOLUTES IN SPACE.

XXXIX....
,.-r.=^;^,3
621
^^r_8r^^^,_
and
XL..
.rz^
?.-y3
^ +
9r
ll-+-]yz2;
r
'
and that accordingly these two last equations are
XXI.
the fractional expressions
400.
satisfied,
independently of w,
when
are substituted for xyz.
The general theory* of

by quaternions, as
briefly treated
evolutes of curves in space may be

follows: a second curve (in space,
or in one plane) being defined to bear to

evolute to involute,
j^
lOVr
when the fij'st
2i
first
curve the relation of
cuts the tangents to the second at right
angles.
(1.)
and
Let p and a be corresponding vectors, op and os, of involute and evolute,

o-', p", c" denote their first and second derivatives, taken with respect to
let p',
Then the two funa scalar variable t, on which they are both conceived to depend.
damental equations, which express the relation between the two curves, as above
defined, are the following
L.
S(<T-p)p' = 0;
n.
V(<r-p)(T'
= 0;
which express, respectively, that the point s is in the normal plane to the involute
at p, and that the latter point is on the tangent to the evolute at s
so that the locus
:
ofv
(the involute)
is
a rectangular trajectory to
all
such tangents to the locus of 8
(the evolute).
(2.)
EHminating c p between the two preceding equations, and taking
derivatives,
we
IIL..SpV = 0,
whence
their
find,
IV.
S((T-p)p"-p'2
VI.
also,
= 0,
V. .. V((T-
p)<T"-VpV =
0;
Sp'a'a'O.
we see first, by IV. combined with I. (comp. 391,

point s of the evolute is on the polar axis of the involute at p, and
therefore that the evolute itself is some curve on the polar developable of the invo(3.) Interpreting these results,
(5.)), that the
lute; and second, by VI. (comp. 880, I.), that this curve is a geodetic line on that
polar surface, because the osculating plane to the evolute at s contains the tangent
normal to the locus of evolutes.

of curvature (395, (6.)) of a curve in space is not generally an evolute of that curve, because the tangents f kk' to that locus do not genebut a given plane involute has always the locus just
rally intersect the curve at all
to the involute at p,
(4.)
The
and
therefore also the (parallel)
locus of centres
* InventecT
by Monge.
t It might have been remarked, in connexion with a recent
(397), that this tangent kk' or
gle of which the cosine is,
ticles
k' is
series of sub-arti-
inclined to the rectifying line X, at an an-
- SLV\ = i2-iT\-i = + sin HcosP;

upper or lower signs being taken, according as the second curvature v^
or negative, because Bk'\ = -r"'.
is
positive
622
mentioned
are
as
all
its
one of
for
evolutes
its
and
[bOOK
has, besides, indefinitely
III.
many others* which
geodetics on the cylinder which rests perpendicularly on that one plane evolute
base.
An
(5.)
easy combination of the foregoing equations gives,
VII.
(TC<r
- p))' = - S (U((T VIII.
or with differentials,
- p')) = +
((t'
- p)= Tder
p)
dT((T
S(T'Urr'
= + T(r',
whence by an immediate integration (comp. 380, XXII. and 397, LIV.),

IX.
AT (o- - p) = J Td(T = + arc of the evolute :
.
two points such as
this arc then, between
and
Si of the latter curve, is
difference between the lengths of the two lines, PS
and
equal to the
between the
PiSi, intercepted
two curves themselves.

(6.) Another quaternion combination of the same equations gives, after a few
steps of reduction, the differential formula (comp. 335, VI.),
X.
dcosoPs=-dSU
^^ = ^/'^P
,.S^;
T((r-p)
p
if
then the involute be a curve on a given sphere, with its centre at the origin o, so
is a geodetic on a concentric cone, this differential X.
vanishes, and
that the evolute
we have
the integrated equation,
XI.
cos
OPS = const.,
or simply,
XI'.
ops
= const.
the tangents ps to the evolute being thus inclined (in the case here considered) at a
constant angle,f to the radii OP of the sphere.
(7.) In general, if
we denote by Ji the interval ps between two

we shall have the equation,
corresponding
points of involute and evolute,
XII.
or
.((r-p)2+i22=o,
XII'.
and the formula VIL may be replaced by the

XIII.
in
i2'2 4-
<t'2
T((T-p) = i2;
following,
Xlir.
or
0,
Di2 = +TD<(T,
which the independent variable t is still left arbitrary.

(8.) But if we take for that variable the arc soS< of the
some fixed point of that curve, we may then
XIV.
Compare the
appears, that
if
XV.
<=jTd<r,
first
Note
to
di2,
evolute,
measured from
write,
= +d?,
XVI.
Di?,
1;
page 534; from the formulaa" of which page it now

= ob and y = oc for its major and
/3
the involute be an ellipse, with
minor semiaxes, and therefore with the scalar equations,

(S/3-V)2
+ (Sr'p)2=l,
S/3yp
= 0,
the evolutes are geodetics on the cylinder of which the corresponding equation
(S/3a)i+(Sy(r)i
f This property
is,
= (i;32-y2)i.
of the evolutes of a spherical curve
was deduced by
Professor
Paper On the Connexion of Involute and Evolute in Space (Cambridge and Dublin Mathematical Journal for November, 1851); in which also a
definition of involute and evolute was proposed, substantially the same as that above
De Morgan,
adopted.
in a
CHAP.
INVOLUTES AS LIMITS OF ENVELOPES.
III.]
623
whence
XVIL
+0 = 0,
D^ (Ht
the integral IX. being thus
In
(9.)
mode
this last
XIX.
+ < = const. = i?o,
i2
JRt
Iio=t- arc SoSi
of evolute,
made of infinitesimals* or even of small differentials. We only infer, as

XVIII. (comp. 380, (9.)), that the quantity Rt + t is co7istant,f because its deriva-
no use
in
of obtaining the result,
Afs =
XVIIL
and
under a new form reproduced.
is
tive is null:
it
having been previously proved (380,
a consequence of our
(8.)), as
(320, 324) that if s be the arc and p the vector of ant/

= Td|0 (380, XXII.) is rigorously satisfied; y/hsitever
curve, then the equation d
the independent variable t may be, and whether the two connected and OTM/<aeo
definition
of differentials
differentials
be small or large.
But when we employ the notation of integrals, and introduce, as above,

symbol jTds, we are then led to interpret that symbol as denoting the limit of a
(10.)
the
sum (comp. 345, (12.)); or
XX.
to write, generally,
J Td,o
= lim. STAp,
Ap = 0,
lim.
if
with analogous formulae for other cases of integration in quaternions.
Geometri-
cally, the equation,
XXI.
and
Tdp =
or
A*,
XXI'.
J Td<r
= A #,
denote arcs of curves of which p and a are vectors, comes thus to be interpreted as an expression of the well-known principle, that the perimeter of any
if s
curve (or of any part thereof)
is
the limit of the perimeter of an inscribed polygon
when the number
(or of the corresponding portion of that polygon),

indefinitely increased,
(11.)
and when
The equations
I.
and XII. give,
XXII.
XXII.
of the sides is
their lengths are diminished indefinitely.
Scr'((7-p)
+ i?ir =
being again arbitrary
coincide with the formulae 398,
0,
but these equations XII. and
LXXXIX.
and
XCL we may
;
then,
by
898, (79.) and (80.), consider the locus of the point p as the envelope of a variable
for radius, and is represphere, namely of the sphere which has s for centre and
= OP be the vector of a variable point
sented by the recent equation XII.,
ifp
thereon.
(12.)
is
But whereas such an envelope has been seen to he generally a
real or imaginary (398.

(79.)) according as R'-
XIII. the intermediate or limiting case (comp. 398,
* In
general,
it
may have
been observed that
+ o-'2 <
or
>
CXXXI.),
we have
0,
for
surface,
we have
which
here
which the
by
circles
hitherto abstained, at
whole Chapter of Applications, from making any use of

although they have been often referred to in these Notes, and employed
least in the text of this
infinitesimals,
therein to assist the geometrical investigation or enunciation of results.
regards the
mechanism
But as
of calculation, it is at least as easy to use infinitesimals in
quaternions as in any other system
as will perhaps be
farther on.
t Compare the Note
to page 516.
shown by a few examples,
624
of the system become />otn/5, and the surface
here the involute (p)above considered.
The
itself
[bOOK
III.
degenerates into a curve, which
is
involutes of a given curve

(s) are there-
fore inc/wrferf, as a. limit, in that general system of envelopes which was considered in
the lately cited subarticles, and in others immediately following.
(13.) The equation of condition, 398, CXXXVI., is in this case satisfied by

XIII., both members vanishing; but we cannot now put it under the form 398,
CXLL, because in the passage to that form, in 398, (85.), there was tacitly effected
a division hy r^, which is not now allowed, the radius r of the circle on the envelope
For a similar reason, we cannot now divide

being in the present case equal to zero.
g, as was done in 398, (86.); and because, in virtue of II., the two equations
398, CLX. reduce themselves to one, they no longer conduct to the formulae 398,
hy
CLX'. CLXI. CLXr. CLXIII. XCIV. ; nor to the second equation 398, CLXII.
(14.) The general geometrical relations of the curves (p) and (s), which were
investigated in the sub-articles to 398 for the case
when
the condition* above re-
are therefore only very partially applicable to a system of involute and evolute in space
at least if we still consider the former curve
(the involute)
as being a rectangular trajectory to the tangents to the latter
(the evolute), instead
ferred to
is satisfied,
of being, like the curve (p) previously considered, a rectangular

trajectory (398, (94.))
to the osculating planesf of the curve (s).
If,
without thinking of evolutes,
CXXXVI.
it
as lately in
is satisfied,
we merely suppose
(13.),
XX.
will be found (comp. the symbolical expression 274,
nary solution
in 353, (18.) of the
+ i?2 =
that the condition 398,
by our having the
system Syp
relation R'^
for Oi,
= 0, p2= o),
a"^=
0,
and the imagi-
that the envelope of the
which such spheres are
(conceived to be) cut by the (tangent) planes, Sc' ((t p) + Ii^= 0, may be said to
be generally the system of all those imaginary points, of which the vectors (or the
bivectors, comp. 214, (6.)) are assigned by the formula,
sphere
((T
where
is
p,
p)^
o,
making
/it
we reduce
the K of 398,
centres
an arbitrary
or the locus of the (null) circles in
cr
- Hit'-
vector,
i<t'
and
+ (U(t' +
y ~1)
Ytr'fx
-\/'- 1 is the old
this expression for
imaginary of algebra.
By
p to the real vector form,
CXXXI."; and
of the null
thus the curve (p), which is here the locus of the

of contact, and coincides with the involute in the present
may still be called a Singular Line upon the Envelope of the
circles
series of sub-articles,
Sphere (with One Variable Parameter^, as being in the present case the only real part
of that elsewhere imaginary surface.
f The curve to the osculating planes of which another curve is thus an orthogonal trajectory, and which is therefore (398, (95.)) the cusp-edge of the polar
developable of the latter curve, was called by Lancret its evolute by the plane (de-
veloppee par
le
plan)
whereas the curve (s) of the present
series (400) of sub-ar-
whose tangents the corresponding curve (p) is an orthogonal trajectoiy, has

been called by way of distinction the evolute by the thread (developpee par le 61)
ticles, to
this last curve.
geometers
It
to delay here on subjects so well known

be invited to read again, in connexion with then
would be improper
but the student
may
the sub-articles (88.) and (89.) to Art. 398.
CHAP.
CURVATURE OF HODOGRAPH OF EVOLUTE.
III.]
625
(15.) If the arc of the evolute be again taken for the independent variable
may
t,
of motion along that arc be always towards the involute^
if the positive direction
write,
XXIII.
= + Ra',
iJ'=-l,
(T
and
we
(t'2=-1, &c.;
whence
XXIV.
if
then k
= ok
=
p'
- <r",
p" = 2i(T"'
i? j",
be the vector of the centre
Yp"p'
XVI. XXXIV., and XCVIII.
XXV.
/c
=p+
-^, =
Ypp
= <T 4-
<T
+i2(
-Ct,1
V(7
(T
j
(t'+
ES(T'(T"'a"
=0'
RSa'o'-W
of curvature
is
i?rriXr
V(T"'<7"-1
= er + UXi. R cos
Hi,
397 become, when we pass from the curve

with the present relations between those two curves this centherefore the foot of the perpendicular let fall /rom the point p
5i, and \i be what
(p) to the curve (s),

tre
(15)),
V(t"'(7"
<r -
=
if ri,
= i?2V(r"'<T"
of the circle which osculates to the invo-
formula 389, IV. gives, after a few reductions,* the expression
lute at p, the general
(comp. 397,
H, and X
r,
in
of the involute, on the rectifying
line Xi
geometrical considerations, because
by
of the evolute
as indeed
is
evident from
(3.) this rectifying line of the curve (s)
is
the
the curve (p).

(16.) If we conceive (comp. 389, (2.)) an auxiliary spherical curve to be described, of which the variable unit- vector shall be,
offis of
polar
XXVI.
and suppose that v
the point
XXV.
is
OT=r = <T'=U(p-(7) = -R-i(p-<^),
the vector
t which corresponds
ou
of the centre of curvature of this
to the point s of the evolute,
we
new
curve, at
shall then
have by
the expression,
XXVII. ..Tn = v-r =
-^,=
-^,=
^=PK:ps;
y<j
R
Vr r
<T
we have
therefore this theorem, ihsit ih.&
inward radius of curvature of the hodograph
of the evolute (conceived to be an orhit described, as in 379, (9.), with a constant

velocity taken for unity') is equal to the inward radius of curvature of the involute,
divided by the interval
between the two curves (p) and (s) and that these two
of curvature, Tu and pk, have one common direction, at least if the direction
o{ motion on the evolute be supposed, as in
(15.), to be towards the involute.
radii
(17.)
ted
The following is perhaps a simpler enunciation of the theoremf just stahe corresponding points of involute and evoand s, Si, Sg,
-(/*P, Pi, P2,
Especially
=-
by observing that Yc'Yo'c" =
o"^,
because
St'(t"
0,
and So-V"
be thrown on this theorem, by comparing it with

and by observing that the equations 397, XVI.
XXXIV. give generally, in the notations of the Article referred to, for the vector of
the centre of curvature of the hodograph of any curve, the transformations,
t Some additional light
may
the construction in 397, (48.)
T+ =
Vr'r'"'
+ ^=-r->\-i = U\.cos/r.
A
4 L
626
[bOOK
III,
= sp, the
if we draw lines STi SiPi, ST2 S2P2, tîth a common length
the curve
the
order
with
contact
second
. will then
have
curve
PT1T2
of
spherical
PP1P2 ., that is with the involute at p.
and
lute,
|1
1|
401. The fundamental formula 389, IV., for the vector of the
centre of the osculating circle to a curve in space, namely the for-
mula,
which has been
employed throughout the present

and used in connexion with
so extensively
Section, has hitherto been established
and
derivatives
rather than with
differentials of vectors,
We may
great or small.
however
establish, in another
differences^
way, an
es-
sentially equivalent formula, into which differences enter by their

limits (or rather by their limiting relations'), namely, the following,
III.
Ao'
/>
+ lim. .=7
if
and
lira.Ao=0,
\ a.^pAp
lim.
AV
-=0,
A/>
being understood to signify the same thing

as V(AV.A/9); and then may, if we think fit, interpret the differential expressionll. as if d/3 and di^p in it denoted infinitesimals* of the
the denominator
first
VAÂ^
and second orders
with similar interpretations
in other but
analogous investigations.
(1.) If iu the second expression 316, L., for the perpendicular from
AB, we change a and
/3
to their reciprocals
we
reciprocal of the result,
obtain this
oD-g-
TV
new
^"'~^"^
denominator of which, OB
diameter
od
of the circle
y = oc = |oD = ^5
gab
&c.
V/3a
VI.
may
be replaced
oa.ab.ob
V(ob.oa)'
by ab, or by AC + ab,
so that if c be the centre of this circle,
for the
its vector
p, q, r are any three points of any

as usual an arbitrary origin, and writing
Supposing then that
given curve in space, while
V.
line
expression,
a(/3-or)/3
V/3-ia-
in the
o on the
(comp. Figs. 58, 64) and then take the
is
OP = p,
OQ = p +
A,o,
OR = (0 + 2Ap +
A2|C),
and therefore
VII.
QR = Ap + A^p,
PQ = Ap,
the centre c of the circle
pqr has
ipR = Ap + i A2p,
the following rigorous expression for
+
+
DC = y = p + Ap (Ap A2p) (Ap iA2p)
V(A2p.Ap)
Compare 346,
(17.),
and the
first
Note
to
page 623.
its
vectoiN
CHAP.
OSCULATING CIRCLES BY INFINITESIMALS.
III.]
whence passing
we
to the limit,
obtain successively the expressions III. and II. for
the vector k of the centre of curvature to the curve
Q and
R, being both
627
pqr
at
the two other points,

p, accord-
supposed to approach indefinitely to the given point
ing to any law (comp. 392, (6.)), which allows the two successive vector chords,
and QK, to bear to each other an ultimate ratio of equality.
oi thus, first
(2.) Instead
the differences
Ap and A2p
which become
of terms,
pq
forming a rigorous expression, such as VII., involving

then simplifying the formula so found, by the rejection
with respect
indefinitely small,
to the
terms retained
and
changing differences to differentials (comp. 344, (2.)), namely Aptod/o, and

A*p to d2|0, in the homogeneous expression which results, and of which the limit is to
be taken we may abridge the calculation, by at once writing the differential sym-
finally
bols, in place of differences,
we have
and at once suppressing any terms, of which we foresee
must disappear from the final
that they
by quaternions, that
perceived,
Thus, in the recent example, when

be the centre of the circle pqr, the
result.
if
equation
..PK =
VIII.
PQ QR I (PQ + QR)
.
V{(qr-pq)pq}
is
we may at once change each of the three factors of the numerator to

qr pq in the denominator is to be changed to d^p and thus
rigorous,
while the factor
inward vector-radius of curvature k
differential expression II., for the
p,
djO,
the
is
at
once obtained.
(3.) It is scarcely necessary to observe, that this expression for that radius, as
with and includes the known expressions for the same radius of curvature of a curve in space, considered as a (positive) scalar, which has been denoted in
vector, agrees
the present Section
by the
(because the more usual symbol p would

Thus, while the formula II. gives immediately (be-
italic letter r
have here caused confusion).

cause
Tdp =ds)
the equation,
IX.
it
gives also (because dp^
X.
and
it
= Tr',
=- ds2,
.
.r-ids3
= TVdpd2p,
= - dsd^s)
and Sdpd^p
r-id2
conducts (by 389, VI.) to this
= V(Td2p2 - d'42)
still
the transformed equation,

;
simpler formula (comp. the equation r"!
396, IX.),
XI.
(4.) Accordingly, if we
.r-ids
= TdUdp.
employ the standard trinomial form (295,

XII.
which gives, by the laws of the symbols
a vector^
ijk
(182, 183),
= idx +jdy + kdz,

6s = Tdp = V(dz^ + dy + dz^),
= V{d^x-2 + d^y^ + d22),
dV = id'^x +yd2y + M^z,
Td'^p
Vdpd2p = i(dyd2z dzd^y) + j^dzd^x dxd-'-z) + k {dxd^y dyd^x),
f dp
XIII.
1.) for
p=zix +jy + kz,
.dx
^^,
the recent equations
.dy
dz
,,^,
.,dar
IX. X. XI. take these known forms
IX'.
X'.
r-id3 = V((dyd22 - dzd^y^ + .)

r-ld2 = V(d2;i: + d2yM d2z2-d22);
.
628
The formula
(5.)
which
IV.,
[bOOK
lately served us to determine a diameter of a circle
through three given points, may be more symmetrically written as follows,

be a diameter of the circle abc, then
XIV.
AD V(aB BC) = AB BC CA
III.
//"ad
an eqxiation* in which V(ab.bc) may be changed to V(ab.ac), &c., and in which

it may be remarked that each member is an expression (comp. 296, V.) for a vector
at, which touches at A the segment abc while its length is at once a representa:
tion of the prorfwcf
o/ fê
Zen^rÂs
o/ <Ae sides of the triangle abc, and also of the
double area of that triangle (comp> 281, XIII.), multiplied by the diameter of the
circumscribed circle.
In general,
(6.)
if
pqrs be any four concircidar
points, they satisfy (by 260,
IX., comp. 296, (3.)) the condition of concirculariiy,
xv...v(!?.?5
RP
\SQ
which
be thus transformed :t
may
XVI.
V ( ?2 + Z5^=v( I.PQ. 5I^\

VR
PR
J
\PS
\TS
Writing then (comp. VI., and the remarks in (2.)),

PS = a> - p, PQ = dp, PR = 2dp+d2(0,
XVII.
.
the second
order,
member
which
first order ;
is
is
seen to be, on the present plan, an infinitesimal of the second

member is only of the
therefore to be suppressed, because the first
and thus we obtain at once the following vector equation of the
ing circle to the curve
QP + QR = d2p,
pqr
osculat-
at p,
student might find
useful practice to verify, that
it
if
we
write in like man-
ner,
XIV'.
be.V(bc.ca)=bc.ca.ab,
a second diameter, then AB = ed, or Abde is a parallelogram. He may

employ the principles, that a^y = y/3a, if Sa/3y = 0, and that /3y - y/3 = 2V/3y ; in
so that
BE
is
virtue of which, after subtracting XIV'. from
by
its
equal V(ab.bc), the equation
lation mentioned.
It is
therefore that
abde
is
or
obtained, and proves the re-
easy also to prove that
XIV".
and
XIV., and dividing by V(bc.ca),
ad - be = 2ab
is
BD V(bC CA) = AB S (bC

.
CA),
a rectangle.
t Without having recourse to this transformation XVI., we might

XV. by infinitesimals, as follows
condition
treat the
xvir.
..
i^^
2qR _
PR
j_
j
oi-p-dp
QP+ QR ~
d2p
_
PR
2dp -t- d2p
uj-p
_
~
d2p
2dp
and
equating then to zero the vector part of the product of these two expressions,
suppressing the infinitesimal of the second order, the equation XVIII. of the osculating circle
is
obtained anew.
CHAP.
OSCULATING SPHERES BY INFINITESIMALS.
III.]
XVIII...
vfJ^
+ 'Ûo;
2dp
\(jj
which agrees with the equation 392, VI., although

ner,
and conducts anew to the expression
XIX.
if
Again,
od =
II. for
-^ + V ^,
K~p
(7.)
629
deduced
a quite different man-
in
k p, under the form,
as in 392,
VIIL
dp
^ be the diameter from the origin, of any sphere through that
point o, which passes also through any three other given points A, B, c, with
&c., we have by 296, XXVI. the formula,
XX.
^Sa/3y
= Va(/3-a)
OA =
a,
(r-i8)y;
XVI I. ),
XXI. ..a = d,o, (3-a = dp + d^p, y - j3=dp + 2dV + d3p,
XXII.
^ = 2ps = 2 (<r - p),
writing then (comp.
and
where a
is
(as in 395, &c.) the vector os (from an arbitrary origin o) of the centre
s of the osculating
sphere to a curve of double curvature at P,
we have by infinitesi-
mals, suppressing terras which are of the seventh and higher orders, because the
member
XXIII.
is
first
only of the sixth order, and reducing* by the rules of quaternions,

((T
- p)Sdpd2pd3p = |Vdp (dp + d2p)
(dp
+ 2d2p + d^)
= 3Vdpd2pSdpd3p + dp2Vd3pdp
(3dp
+ 3d2p + d^p)
which agrees precisely with the formula 395, XIII., although obtained by a process
so diflFerent.
(8.) Finally as regards the osculating plane,

in space, infinitesimals give at once for that
XXIV.
and
if
(oi
and the second curvature, of a curve
plane the equation,
- p) dpd2p = 0,
agreeing with 376, V.
three consecutive elements of the curve be represented (comp.
diflferential
XXI.) by the
expressions,
XXV.
the second curvature
QR = dp + d2p,
PQ = dp,
r"',
defined as in 396,
is
RS
= dp +
2d2p + d^p,
easily seen to be connected as follows
with the angle of a certain auxiliary quaternion
from unity
q,
which
differs infinitely little
^r^r^T^
XXVI.
,^
r-id
V(qr Rs)
Vdpd^p
vvArtT
XXVII.
..2=-^
^=1+:ft^?-?
VdpdV
V(pq.qr)
.
Z ^5,
<
if
* Of the
eighteen terms Avhich would follow the sign of operation ^V,
if
the se-
cond member of XXIII. were fully developed, one is of the fi)urth order, but is a
scalar ; three are of the//M order, but have a scalar sum ; nine are of orders higher
than the sixth
and two terms of the sixth order are scalars, so that there remain
;
only three terms of that order to be considered.

second member in question reduces itself to the
In this manner
sum
fV. (dpd2p)2 = 3Vdpd2p Sdpd2p,

=
|dp2V(dpd3p + 3d3pdp) dp2Vd3pdp
.
and
and thus the third member of XXIII.
is
obtained.
it is
found that the
of the two vector parts,
630
we have
[bOOK
HI.
then the expression,
XXVIII.
Second Curvature
= r"! = -~ =
,^,
Ydpd-p
dp
which agrees with the formula 397, XXVIL, and has been illustrated,
397 and 398, by numerous geometrical applications.
in the sub-
articles to
On
it appears that although the logic of derived vectors,

of vectors considered as finite lines, proportional to such derivatives, is perhaps a little clearer than that of infinitesimals, because it shows more
evidently (especially when combined with Taylor's Series adapted to Quaternions,
(9.)
and of
the whole, then,
differentials
342, 376) that nothing
is
nions, in practice, with
neglected, yet
as infinitely small differences
abridgments of calculation
Section
On
7.
it is
perfectly possible to
combine* quater-
methods founded on the more usual notion of

:
and that when
arise,
this
combination
is
Differentials,
judiciously made,
without any ultimate error.
Surfaces of the Second Order
and on Cur-
vatures of Surfaces,
As
402.
early as in the First
Book
of these Elements^
some
spe-
cimens were given of the treatment or expression of Surfaces of the

Second Order by Vectors ; or by Anharmonic Equations which were
derived from the theory of vectors^ without any introduction, at that
Thus it was shown, in the
stage, of Quaternions properly so called.
sub-articles to 98, that a very simple anharmonic equation {xz = yw)
might represent either a ruled paraboloid, or a ruled hyperholoid^ ac-
cording as a certain condition {ac = bd) was or was not satisfied, by

the constants of the surface. Again, in the sub-articles to 99, two
examples were given, of vector expressions for cones of the second order (and one such expression for a cone of the third order, with a
conjugate ray (99, (5.)); while an expression of the same sort,
namely,
I.
was assigned
a,
/3,
/)
CCa
4-
3/^3
+ 27,
with
(99, (2.)) as representing generally an ellipsoid^j with
7, or OA, OB, DC, for
three conjugate semidiameters.
in

any future applications of quaternions, to
or as
finite differential,
f In
like
x'^-\-f->rZ'^=\^
however be of course
necessary,
specify in which of these two senses, as a
an infinitesimal, such a symbol as dp
manner the
11. ..
It will
And finally,
is
employed.
expression,
p=xa+y (3+ zy,
with
x^ + y^
- z^ =
1,
or
=-1,
represents a general hyperboloid, of one sheet, or of two, with a/3y for conjugate semidiameters while, with the scalar equation x'^ + y^ z^= 0, the same vector expres'
:
sion represents their
common asymptotic cone
(not generally of revolution).
631
QUATERNION EQUATIONS OF SPHERE.
CHAP. Ill]
in the sub-articles (11.) and (12.) to Art. 100, an instance was furnished of the determination of a tangential plane to a cone, by means
o{ partial derived vectors.
403. In the Second Book, a much greater range of expression

was attained, in consequence of the introduction of the peculiar symhols,
or characteristics of operation, which belong to the present Calstill with that limitation which was caused, by the con-
culus; but
and notation of a Quaternion being confined, in that Book, to

admitQuotients of Vectors (112, 116, comp. 307, (5.)), without yet
ception
Powers of Directed Lines in Space : although verand even norms* of such vectors were already intro-
ting Products or
sors, tensors,
duced (156, 185, 273).

(1.)
The Sphere^f
for instance,
vector OA, or a,
with a length
sented, not only
(comp. 402,
I.
which has
Ta = a,
I.)
by the
for
centre at the origin,
its
one of
its radii,
and has the
admitted of being repre-
vector expression,
=
p a;a + yj3 + zy,
+ ^24. 22 =1^
0:2
with
r.
= Ty = a,
..Ta=T/3
i^
'
'
and
I".
S^ = S ^= S | = 0,
a
a
(3
but also by any one of the following equations, in which

to
-o
II...- =
K^,
pa
III...^K^=1;
a
a
XIII...Nfs^
= l;
145, (8.), (12.)
186, C2.),
i87,\l.)
200, (11.),
215,(10.),
X...Np = Na;
V^'\
IV...N^=1;
a
VII...T^=1;
IX...N^ = N-;
..S^::i^=0;
p+a
XI...(s^Y-(v^Y=l; XII...NS^
or
permitted to change a
VI...Tp = Ta;
V...Tp = a;
VIII.
it is
XIV...Tfs^
273,(1.)
204, (6.),
NV=1;
+
XXV., XXVI.
204,(9.)
V^^=l;
by the system of equations,
XV...
representing a sr/stem
8^
= 0.,
fv^>
= ^2-l(^0),
of circles, with the spheric surface
204,(4.)
for their locus.
The notation Na, for (Ta)2, although not formally introduced before Art. 273,
had been used by anticipation in 200, (3.), page 188.
t That is to say, the spheric surface through a, with o for centre. Compare
the
Note to page 197.
632
(2.)
OÂer/orms of equation,
principles be assigned
XVI.. .t = K^;
a
p
for
for the
example we
same spheric
may
-vvrite,
surface,
[bOOK
may
IIT.
on the same
CHAP, in.]
marks apply
EQUATION OF ELLIPSOID RESUMED.
to the representation,
633
of other surfaces of the second
by quaternions,
order.
A brief review,
404.
or recapitulation, of some of the chief ex-
with the Ellipsoid, for example, which have been

pressions connected
already established in these Elements, with references to a few others,
may
not be useless here.

+zy, with the scalar relation
and with arbitrary vector values of the constants a, /3, y, which
cited (402) from the First Book, or the equations 403, I., without
the vector expression pz= xa-\-y(5

(1.) Besides
a^ + y2+2?=l,
was
lately
the conditions 403, I'.^
Book (204,
the Second
H*
which are peculiar to the sphere, there were given in

which differed from those lately num-
(13.), (14.)) equations
bered as 403, XI. XII. XIII.

for instance, there
XIV. XV., only by
the substitution of
for
V^
was the equation,
as the
analogous to 403, XI., and representing generally* an ellipsoid, regarded
locus of a certain system of ellipses, which were thus substituted for the circlesf
XV.) of the sphere, by a species of geometrical deformation, which led to the

establishment of certain homologies (developed in the sub-articles to 274).
(403,
although
it
has since been found possible and useful, in this Third Book, to identify
own indices or axes (295), and so to treat them as a
those right versors with their
system of three rectangular lines, as above.

* In the case
the equation I.
of parallelism of the two vector constants (j3 1| a),
its axis in the direction of a;
with
re
a
vo?M<ion,
-S'/jêroirf
o/
represents generally
while in the contrary case of perpendicularity (j3 -^ a), the same equation I. represents
an
elliptic Cylinder,
204, (10.), (11. )
t The equation
cond Book,
whence
it
with
its
and the Note

I.
might
of
generating lines in the direction
(3.
Compare
to page 224.
also
have been thus
Avritten,
on the principles of the Se-
would have followed at once (comp. 216, (7.)), that the ellipsoid
with a common radius = T(3, by the two diametral planes,
I. is
cut in two circles,
/3
form, while I. itself is

what was there called a focal form, of the equation of the surface ; the lines a'^ ^'i
two (real) cyclic normals, while /3 is one of the two
being, by the Third Book, the
cone.
Compare the Note to page
of
the
(real) /ocai lines
(imaginary) asymptotic
In
fact, this equation
I', is
what was
called in
474.
859 a
cyclic
634
[BOOK
III.
of Aer pain
(2.) Employing still on?y quotients of vectors, but introducing <2f?o
of vector-constants, y, 5 and t, k, instead of the pair a, (3 in the equation I., which
were however connected with that pair and with each other by certain assigned re-
lations, that equation
was transformed successively

II.
and
to a
form which
may
III.
and
was
this last ^brm
(6.), (7.)),
.Tf^ +
to
be written thus (comp. 217, (5.)),

.
Tf + K
I
= Tt2-
p'JTp
interpreted, so as to lead to a
which was
216,
K|'j=l,
illustrated
by a Diagram
217,
Tk2;
XVI
Hule of Construction* (217

and from which manj
(Fig. 53),
geometrical properties of that surface were deduced (218, '219) in a very simpl(
manner, and were confirmed by calculation with quaternions the equation and construction being also modified afterwards, by the introduction (220) of a new pair o
:
vector-constants,
and
K, in
(3.)
and
i'
k',
which were shown
to
admit of being substituted
for
the recent form III.
And
although the Equation of Conjugation,

316, LXIII
IV...S-S^-s(v^.V^,Vl,
a
a
'
which connects the vectors

plane of the
other,
X,
fi
(i
of any two points l, m, whereof one
with respect to the ellipsoid
of the First Chapter of the present Book, yet

processes of the Second
li j
Book alone
it
is
on the polai
was not assigned till near the enc

was there deduced by principles anc
I.,
which thus were adequate, although not
ir
the most practically convenient way, to the treatment of questions respecting tangen,
playies and normals to an ellipsoid, and similarly for other surfaces^ of the sam(
second order.
* This Construction
of the Ellipsoid, by means of a Generating Triangle and f
Diacentric Sphere (page 227), is believed to have been new, when it was deducec
by the writer in 1846, and was in that year stated to the Royal Irish Academj
(see its Proceedings, vol. iii. pp. 288, 289), as a result of the Method of Quater
nions,
which had been previously communicated by him to that Academy
(in tht
year 1843).
f The following
are a few other references, on this subject, to the Second Book
Expressions for a Eight Cone (or for a single sheet of such a cone) have been giver
in pages 119, 179, 220, 221.
In page 179 the equation S - S
- = 1,
has been
as-
signed, with a transformation in page 180, to represent generally a Cyclic Cone, oi

a cone of the second order, with its vertex at the origin ; and to exhibit its cyclic
planes, and subcontrary sections (pp. 181, 182).

Hight Cy/mrfers have occurred in
case of an Elliptic Cylinder has been
pages 193, 196, 197, 198, 199, 218.
-J- a in
and a transformation of the equaalready mentioned (the case when /3
I.)
tion III. of the Ellipsoid, by means of reciprocals and norms of vectors, was assigned
;
in page 298.
And
several expressions (comp. 403), for a Sphere of which the
ori-
CHAP.
C4.)
Iir.]
But
TRANSFORMATIONS OF THE EQUATION.
in this Third
Book we have been
the simpler form,*
V.
635
able to write the equation III. under
T (tp + p/c) = k2 _ t2,
282,
XXIX.
which has again admitted of numerous transformations for instance, of all those
which are obtained by equating (k'^ i^)^ to any one of the expressions 336, (5.),
;
for the
square of this last tensor in V., or for the norm of the quaternion ip + pK ;
of equation thus arising, which are easily converted into focal forms
while a rectangular transformation (373, XXX.) has subsequently been
cyclic formsf
(359);
assigned,
whereby the lengths
(âbc),
and also the directions, of the three semiaxes of
the surface, are expressed in terms of the two vector-constants,
t,
the results thus
with those previously deduced, from the

in the Second Book.
geometrical construction (2.)
The equation V. has also been differentiated (336), and a normal vector
obtained
by calculation being found
to agree
(5.)
= ^p
has thus been deduced, such that, for the ellipsoid in question,
VI.
Srdp =
0,
and
VII.
Svp
a process which has since been extended (361), and appears to furnish one of the
best general methods of treating surfacesX of the second order by quaternions : espevector functions (^(pp) of veccially when combined with that theory of linear and
tors,
which was developed in the Sixth Section of the Second Chapter of the pre-
sent
Book.
gin was not the centre, occurred in pages 164, 179, 189,
and perhaps elsewhere,
without any employment o( products of vectors.

* Mentioned
by anticipation in the Note to page 233.
t Compare the second Note to page 633. The vectors i and k are here the
and t k is one of the /oca/ lines ; the other being the line i k of
cyclic normals,
page 232.
I
The following
are a few additional references to preceding parts of this Third
much greater length than was designed (page 302).

In the First Chapter, the reader may consult pages 305, 306, 307, for some other
In the Second Chapter, pages
forms of equation of the ellipsoid and the sphere.
Book, which has extended to a
416, 417 contain some useful practice, above alluded
to,
in the differentiation
and
= T(ip + px). As
regards the Sixth Section of

that Chapter, which we are about to use (405), as one supposed to be familiar to the
reader, it may be sufficient here to mention Arts. 357-362, and the Notes (or some
transformation of the equation
r'*
of them) to pages 464, 466, 468, 474, 481, 484.
In this Third Chapter, the sub-
373 (pages 504, &c.) might be re-perused; and perhaps the

investigations respecting cones and sphero- conies, in 394 and its sub-articles (pages
541, &c.), including remarks on an hyperbolic cylinder, and its asymptotic planes
articles (7.)-(21.) to
page 547). Finally, in a few longer and later series of sub-articles, to Arts.
of sur397, &c., a certain degree ot familiarity with some of the chief properties
faces of the second order has been assumed ; as in pages 571, 688, 691, and generally
(in
in the recent investigations respecting the osculating twisted cubic
&c), to a helix, or other curve in space.

It appears that this Section may be conveniently referred
similarly in other cases.
to,
(pages 691, 620,

as III.
ii.
and
636
[bOOK
III.
405. Dismissing then, at least for the present, the special consideration of the ellipsoid^ but still confining ourselves, for the moment, to Central Surfaces of the Second Order, and using freely the
principles of this Third Book, but especially those of the Section
ii.
(III.
6) last referred to, we may denote any such centra/ and nowthe scalar equation (comp. 361),
b}'-
conical surface
I.
..//>
= S/>0/>=l;
the asymptotic cone (real or imaginary) being represented by the

connected equation,
II.. .fp
= Sp^p = 0;
and the equation of conjugation, between the vectors />, />' of any two
points p, p^ which are conjugate relatively to this surface I. (comp.
362, and 404, (3.), see also 373, (20.)), being,
III.
/(/>, />0
=f(p', p) = Spc^p^ =
while the differential equation of the surface
IV.
= dfp = 2Svdp,
is
V.
with
S/>'0/>
1 ;
of the form (361),

.
i/
0/>
which represents the normal v to the surface,

and self-conjugate (361, (3.)) ; and the surface
this vector-function 0/>,
being at once linear

itself
hQing the locus of
selves, so
all
that its equation
IhQ points p which are conjugate

may be thus written,
because
V...f{p,p)=l,
Such being the form
(1.)
and rectangular
satisfy
of ^p,
it
unit-lines^ ai, aa, as,
them-
by 362, IV.
/(p,p)=fp,
has been seen that there a.realvta.ys three real
and three real
scalars, Ci,
C2, C3,
such as to
(comp. 357, III.) the three vector equations,
VI.
whence
<pai
= -ciai,
<f)a2
= C2a2,
<pa3--cza3i
also these three scalar equations are satisfied,
VII. /on =
and
to
I.
ci,
/as =
Co,
fas
= C3
therefore (comp. 362, VII.),
VIII.
./(criai)=/(c2-Ja2)=/(c3â3)
= l.
(2.) It follows then that the three (real or imaginary) rectangular
IX.
/3i
= ci-iai,
/32
= C3-ia2,
^3
lines,
= cfâs,
are the three (real or imaginary) vector semiaxes of the surface I. ; and that the three
(positive or negative) scalars, ci, co, C3, namely the three roots of the scalar and cubic
equation* Af =
(real or
(comp. 357, (!.))> ^^^ the (always real) inverse squares of the three
imaginary) scalar semiaxes, of the same central surface of the second order.
* It
tion
<p
is
Mo=
0, as in page 462, &c., because the funcunnecessary here to write

here supposed to be self- conjugate ; its constants being also real
is
CHAP.
GENERAL CENTRAL SURFACE.
III.]
For the reality of that surface
(3.)
I., it is
637
necessary and sufficient that one at
least of the three scalars cj, co, C3
should be positive ; if all be such, the surface is an

two, but not the third, it is a single-sheeted hyperholoid ; and if only
ellipsoid ; if
a double-sheeted hyperholoid
it is
one,
and
QSich finite,
(4.)
different
We have already
mation,
which
to be
tors /3i/32i83
(5.)
how
to obtain the rectangular transfor-
IX., be thus written,
may now, by
it is
seen (357, (2,))
X,..fp = ci cSaip)2 + C2(Sa2io)2 + C3(Sazpy,
XL
but
those scalars being here supposed to be
from zero.
We
..fp
remembered
may become
had
= (S/3i-ip)2 + (S/32-ip)2 + (S/33-ip)2

by
that,
(2.)
and
(3.),
one or even two of these three uec-
imaginary, without the surface ceasing to be real.
also the cyclic transformation (357, II. II.'),
XII. ..fp
= gp^ + SXpfxp=p^(g-8Xfi) + 2SXpSfxp,
which the scalar g and the vector \, p, are real, and the latter have the directions
of the two (real) cyclic normals * in fact it is obvious on inspection, that the surface
in
is
cut in circles,
by planes perpendicular
to these two last
lirtes.
proved that the four real scalars, cic-zCsg, and the^we real vecaia23^/i> are connected by the relationsf (357, XX. and XXI.),
(6.) It has been

tors,
XIII.
XIV.
oi
= -5--TX/z,
C2
= -g+S\fi,
at least if the three roots C1C2C3 of the cubic
A/=
C3
= -5'i-T\/*;
03 =
a2 = UVX;u,
.ai=U(XT^-/iT\),
U (XT/^ + /iTX)
be arranged in algebraically as-
cending order (357, IX.), so that ci<C2<C3.

(7.) It may happen (comp. (3.)), that one of these three roots vanishes
that case (comp. (2.)), one of the three semiaxes becomes infinite,
and in
and the surface
I.
becomes a cylinder.
TXfi, so that the two other roots are

(8.) Thus, in particular, if ci= 0, or g =
both positive, the equation takes (by XII., comp. 357, XXII.) a form which may
be thus written,
XV.
and
it
represents an
(SX/xp)2
+ (SXpT/i + S/ipTX)2 = TX/i -
0,
or
g = SX/i,
XVI.
and represents an hyperbolic cylinder

the remaining root C3
rical,
its
is
the equation becomes,

.
2SXpS//p
1,
the root ci being in this case negative, while
see also the proof by quaternions, in 373,

(16.),
theorem, that any two subcontrary circular sections are homosphe-
with the equation (373,
XLIV.)
of their
common sphere, which
centre in the diametral plane of the two cyclic normals X,
t These
positive.
Compare the Note to page 468
known
>
elliptic cylinder.
(9.) Again, if cg
&c., of the
SX/^
relations
and a few others mentioned are
occurred in an earlier part of the work,
it
is
found to have
p..
so useful that, although

^gi.
seems convenient to restate them h
638
[bOOK
III.
= 0, or ^ = TX/x, so that ci and cz are both nega(10.) But if we suppose that C3

the equation may (by 357, XXIII.) be reduced to the form,
tive,
XVII.
it
(SXfipy
(SApT/i
S/ipTX)2
= - T\/i -
represents therefore, in tîs case, nothing real^ although
same
an imaginary*
case, the equation of
it
SXft
<
may be said to be, in the
elliptic cylinder.
(11.) It is scarcely worth while to remark, that we have here supposed each of
the two vectors X and /i to be not only real but actual(Ji.rt. 1) ; for \i either of them
were to vanish, the equation of the surface would take by XII. the form,
XVIII. ..p2=^-:,
XVIir. ..Tp = (-^)-i,
or
and would represent a real or imaginary

was negative or positive : X and fx have
sphere, according as the scalar constant
also distinct directions, except in the case
of surfaces of revolution.
(12.) In general,
cyclic normals, X,
fi,
from the relations
(6.), that
the plane of the two (real)
it
results
is
perpendicular to the (real) direction of that (real or imagi-
nary) semiaxis, of which, when considered as a scalar (2.), the inverse square c% ia
algebraically intermediate between the inverse squares ci, C3 of the other two ; or that
and cyclic planes (SXp = 0, S/zp = 0) intersect in that real line

(YXfi) which has the direction of the real unit-vector a^, (1.), corresponding to the
all which agrees with known results, remean root ci of the cubic equation 3f =
the two diametral
specting the circular sections of the (real) ellipsoid, and of the two hyperboloids.
Some
406.
additional light
405, II.
may be thrown on
the theory of the
the consideration of
its asymptotic cone

by
which cone, by 405, XII., the equation may be thus
central surface 405,
of
I.,
written,
I.
and which
is
//>
=9P^ f S V/"/> = pH9- S V) +
2S\/)S/t/>
real or imaginary^ according as
we have
the inequa-
lity,
II.
that
is, by 405,
roots of the cubic
as the surface fp
/ < XV,
or
m...g''> \^/ji?
(6.), according as the product C1C3 of the extreme

is negative or positive ; or
finally, according
M=
=
\ is
a (real) hyperholoid^ or an ellipsoid (real or
imaginary!").
* In the Section
(III.
ii.
6) above referred to,
many symbolical results have been
imaginary cyclic normals, or focal lines, &c., on which it ia

But it may be remarked that as, when the scalar function
return.
established, respecting
unnecessary to
fp admits
of changing sign, for a change
of direction of the real vector
positive for some such directions, and negative for others,
the two equations, /p
=+
loids, of different species

-^^se
1,
:
/p
so,
=-
1,
when
represent then two real and conjugate hyperbo-
the function /p
is
either essentially positive. Of
essentially negative, for real values of p, the equations /p
be said to represent two conjugate ellipsoids, one

tion
?/:;ompare the
p, so as to be
although/( p)=/(+(o),
Note immediately preceding
also
= 1 and/p = - 1 may
and the other imaginary.

real,
CHAP.
CONES OF THE SECOND ORDER.
III.]
As
(1.)
tisfied, it
639
regards the asserted reality of the cone I., when the condition II.
suffice to observe that if we cut the cone by the
plane^
is
sa-
may
IV.
the section
..SX(p-/i)=-^,
a circle of the real and diacentric sphere,
is
Y...p^ = 2Sfip,
and a real
because
circle,
VI.
v. ..(p-))2 = ^2;
or
on the real cylinder of revolution,
it is
TV(p - ;i)UX = (T/i2 _^2TX-2)i,
so that its radius is equal to this last real radical.
For example, the cone
(2.)
VII.
S^ S
^ = 1,
or
Vir.
2(SapS/3|0
- a^p^) =
0,
which under the form Vll. occurred as early as 196, (8.), and for which \ = a,
= (3, g = Sa/3 - 2a^, and therefore TX/x +g>0, the condition II. reduces itself to
fi
iXfi-g >
or after division
VIII.
S=
P
that the sphere
&c., to the
..KT4S)^>1,
a
when
and accordingly,
by 2Ta2,
either of these
1 is cm<
Vlir.
or
two
form (comp. 199, XII.),
..S^/^>1;
ya
will be
found
1 in a real circle, the base of
a real
last inequalities exists,
by the plane S - =
it
cone VII.
As an example of the variety of processes by which problems in this Calcube treated, we might propose to determine, by the general formula 389, IV.,
the vector k of the centre of the
osculating circle to the curve IV. V., considered
The first derivatives of the equations
merely as an intersection of two surfaces.
(3.)
lus
may
would allow us
to
TY
IXthe section
radius
p'
=.
VX(p- jw), and
is
- = P+
1
P'^
^,
therefore p"
= Xp'
= P + P' = SpX+V/iX = /*-^X

^
,
therefore a circle, because its centre
of curvature
is
whence, by the
constant
and
its
is,
X.
assume
we have
formula,
= T(p-jc; = T(p-/i + ^X-') =
CT/i2-^2TX-2>,
the radius of the cylinder VI.
AVhen the opposite inequality III. exists, the radius X., the cylinder VI.,
and the cone I., become all four imaginary ; the plane IV. being
then wholly external to the sphere V., as happens, for instance, with the plane and
(4.)
the circle IV. V.,
sphere in (2.),
when
the condition VIII. or VIII'.
(5.) In the intermediate case,
XI.
i^2
is
reversed.
when
= X2/i2,
or
Xr. ..^ = + TX),
X. vanishes ; the right cylinder VI. reduces itself to its axis ; and
In
IV. V. becomes a point, in which the sphere is touched by the plane.
then, the cone I. is reduced to a single (real*) right line, which has
the radius r in
the circle
this case,
It
may however
be said, that in this case the cone consists of a pair ofimagi-
nary planes, which intersect in a real right
line.
640
[bOOK
III.
(compare the equations of the elliptic cylinders^ 405, XV. XVII.) the direction of
- jiiTX, if ^ = - TX/i, but the perpendicular direction of XT/x +
XT/i
^TX, SS g =
+ TXfi.
(6.) In general (comp. 405, X.), the equation of the cone
admits of the rect-
I.
angular transformation,
XII. ..fp
and the two sub-cases
= ci(Saip)2 + C2(Sa2p)2 + 03(803,0)' =
last considered (5.) correspond respectively
the evanescence of the roots
3f=
(by 405, (6.))
to
with the resulting directions ai,

03 of the only real side of the cone. An analogous but intermediate case (comp. 405,
= 0,org = S\fi in which case, the cone I. reduces itself to the
(9.)) is that when C2
Ci, C3
of the cubic
0,
pair of (real) planes^
XIII.
SXp
S/xp
0,
namely to the asymptotic planes of the hyperbolic cylinder 405, XVI., or

which are usually the two cyclic* planes of the cone.
(7.) The case (comp. 394, (29.)),
XIY..
for
which the equation
XV.
may
g = -S\p,
or
=fp = 2(S\pSpp -
deserve a moment's attention.
same time
ci-
C2
+ C3 = 0,
cyclic
normals X
And
= 2S(VXp .Ypp),
case, the two planes, of Xp and pp,

and p with an arbitrary side p of the cone,
and these two normals
which thus
sides of the cone,
dicular to those two sides.
p^SX)^)
In this
are alwaj's rectangular to each other

are at the
of the cone becomes,
I.
which connect the two
XIV'.
to those
is
to the cj'clic planes
cut in circles,
by planes perpen-
because the equation of the cone
may
(in the
same
case) be thus written,
XVI.
TV(X +p)p = T V(X p) p,
while the lengths of X and
p may vary, if their product TXp be left unchanged, so

represent a7iy two lines from the vertex, in the plane of
the two cyclic normals, and harmonically conjugate with respect to them, it follows
that, /or this cone XV., the sines of the inclinations of an arbitrary side p, to these
that
X + /1 and X p may
two new
lines,
have a constant ratio
to each other.
I. of/p shows (comp. 394, (23.)), that the constant product of the sines of the inclinations, of a side p of the cone to the two cyclic
(8.)
Tn general, the second form
planes, has for expression,
XVII.
while the
first
by showing
form
that
I.
of the
cosz^.cosz
X
A*
V TXfi
a real cone) the cosine of a real angle, namely, that
of the quaternion product Xppp, since
XVIII.
fi
same function^ reproduces the condition of reality IL,
is (for
g TX/x
J 9 + cos^.
-=M
=i(^
;^ _ X
P
.
it
gives the relation,
= SUXp/xp = cos L Xppp = cos L '-^^.

-^
IXp
A.
* The cones and surfaces which have a common

the vectors
known
X and
centre,
p, but different values of the scalar g,
phraseology, to be biconcyclic.
and common values
may
thus be said,
of
in a
CHAP.
(9.)
ARCUAL AXES AND FOCI 6F A SPHERO-CONIC.
III.]
We may
if
sub-case (6.),
also observe that in the case
of reality
IT.,
641
with exclusion of the
03 have the direction of the internal axis of the cone, so that
XIX. ..ci<0,
C2<0,
XX.
XIX'.
or
C3>0,
= ca-âa + (- ci)-irti,
pi
X^ have the
the two sides (of one sheet) in the plane of
pz
sf>S\fx,
ff<T\fi,
directions,
= c^-ia^ (- ci)-iai
then their mutual inclination, or the angle of the cone in the plane of the cyclic
normals, be denoted by 2b, we have the values,
if
XXI.
tan2 b
- ci
XXI'.
cos 2b
= -^^^^ =
-
ci
JTXfj,
+ C3
the angle of the quaternion Xpfip is therefore (by XVIII.), equal to this angle 2b,
namely to the arcual minor axis of the sphero-conic, in which the cone is cut by the
concentric unit-sphere.
(10.) The same condition of reality II. may be obtained in a quite different way,
as that of the reality of the reciprocal cone, which is the locus of the normal vector,
XXII.
Inverting this linear function
(p,
XXIII.
v = ^p=gp + YXpfi.
by the
method
and 361,
the expression (comp. 354, (12.),
in
of the Section III.
ii.
6,
we
find first
(6.)*),
mp = rPv = fi^XSXv + \^nSnv-g(\Sfiv + n8Xv) + (^g^ - X^)^,

m = (^ - SX^) (^3 - X^^) = - ci^scs
XXIV.
which
and next the reciprocal equation (comp. 361, XXVII.),
XXV.
which
may
= Sr^v = /ti2(SXv)2 + X2(S/xv)2-2^SXvS/xv + (^2_x2^2)^2^
be put under the form,
xxvi...cs(.^
the quotient
g TXp, thus presenting

;
itself
.^)
anew
^:^,
as a cosine,
namely
supplement of the sum of the inclinations of the normal v (to the cone
cyclic
normals
X,
(of that cone)
or as the cosinef of
tt
- A B,
if
as that of the
I.), to the
A and b
two
denote
(comp. Fig. 80) the two spherical angles, which the tangent arc to the sphero-conic
(9.) makes with the two cyclic arcs : so that by comparison of XXI'. and XXVI,
we have
the relation,
XXVII.
(11.)
A+B = Z7 +
X
Comparing the expression XXI'.
* In the
expression 361,
XXVI.
/l-
= 7r-2b.
for cos 2b,
with the
last expression
term ought to have been

have been changed.
t This relation was mentioned by anticipation in 394, (3.) and the relation in
XXVII. may easily be verified, by conceiving the point of contact p in Fig. 80
printed as
-YXfiSXvp
for \pv, the second
or else the sign should
(page 543) to tend towards a minor summit of the conic, or the tangent arc
tend to pass through the two points c, c', in which the cyclic arcs intersect.
4 N
apb
to
642
XVIII.
may
for
TX/it,
Ave derive the following construction for a sphero-conic,
easily be verified
round one of them (say
which
:*
by geometry
Having assumed two points

circle
III.
[bOOK
m) on a
(l,
sphere,
l), bisect the arcs
and having described a
(mq) which are drawn
small
to its circum-
ference from the other point ; the locus of the bisecting points (p) will be a spheroconic, with the two fixed points for its two cyclic poles (or for the poles of its cyclic
arcs), and with an arcual minor axis (2b) equal to the arcual radius of the small
circle.
(12.) As regards the arcual major axis (say 2a) of the same sphero-conic,
(with the conditions XIX.) the angle between the two sides (comp. XX.),
XXVIII.
whence (comp.
XXIX.
and
therefore,
p3
XXL),
tan2 a =
C3~â3f
H
-
( C2)"a2,
XXIX'.
or
= czâz (- C2)-âz
p4
cos 2a
= I^Hl^ = (sav)
-
C2
it is
e,
C3
a few easy reductions being made,
sinb
XXX.
^{i(i
from which we can at once
su^)}=cosuf;
focus of the conic be determined, by drawing from a minor summit to the major axis an arc equal to the major semiaxis a,
the minor axis subtends at this focus (or at the other) a spherical angle equal to the
infer,
that
if
a,
angle between the two cyclic arcs.

(13.) For the two real unifocal transformations of the equation of the cone,
the /orms,
XXXI.
aiyagy +
with one common
be
and
0,
XXXI'.
a(Va2 + 6(S/3'jo)2 = 0,
a and
set of real values of the scalar coefficients,
real focal unit lines a,
may
6(S/3p)=
a',
and two real
normals ^,
directive
sufficient here to refer to the sub- articles to
and
noticed, that if the cone be real,

axis, so that the inequalities
XIX.
XXXII.
if
C3-1
>
instead of the inequalities 358, III., or 359,
the earlier formulae referred
to,
358
j3'
6,
but with two
corresponding,
except that
the line n^ have the direction of
are satisfied,
or
and therefore
also
it
it
should be
its
(by 405,
internal
(6.)),
> ci-i > C2-1,

XXXVII., we
are
now
to change,
in
the symbols ciczczaiaâs to czcic-zasaiaz, so that we
have now the values,
XXXIII.
(14.)
And
last values, it
..
a = -ci,
XXXIV. ..sml^.seclt^
the sines
* In
of the
fact,
c3-ci-\-C2,
if
T/3
= T/3'=1.
as regards the interpretation of the

unifocal form
is evidently contained in this other
equation,
inclinations of
11^ =
XXXI., with these
g3-Cl+C
".^.liJJ:^2\| =
\*
const.
an arbitrary side (p) of the
cone, to
a.
focal line
(a),
the bisecting radii op are parallel to the supplementary chords u'q, if

of the sphere
and the locus of all such chords is a cyclic cone,
mm' be a diameter
resting on the small circle as its base.
CHAP.
and
SYSTEM OF CONFOCAL SURFACES.
III.]
643
corresponding rfjVec/or />/ane (-1-/3), thus bearing to each other (as
to the
is
known) a constant ratio, which remains unchanged when we pass to the other
line (a'), and at the same time to the other
(real) director plane (J- /3')
(real) focal
well
and the focal plane of
two lines (a, a') being perpendicular to that one of the
tliese
which corresponds
three axes,
which the reciprocal
to the root
(here ci,
by XXXIT.) of the
cubic, of
algebraically intermediate between the reciprocals of the other
is
two.
It
(15.)
is,
however, more symmetric to employ the bifocal transformatioti
(comp. 360, VI.*),
XXXV.
in
(Sajo)3
which the scalar constant
2^SapSa'p
a,
The equation XXXV.,

XXVI. XXXVI.),
XXXVII.
the
- e2)p2
= cos2a;
are the twof real and /ocaZ unit lines, recently considered (13.).
(16.)
(comp.
(1
has the value (comp. XXIX'.),
XXXVI.
and
+ (Sa',o)2 +
for the case of a real cone,
..
may be
thus written
= cos-ie=2a;
Z^+Z^
a
a
sumX of the inclinations of the side p to the two focal lines a, a' being thus conand equal (as is well known) to the major axis of the spherical conic : and
stant,
when e> 1, the cone becomes imaginary, yet

we shall shortly see.
although,
it is
then asymptotic to a
real ellipsoid, as
may
I.
in
The
bifocal form (406, XXXV.) of the equation of a cone

the
suggest
corresponding /orm,
407.
a=(7//> = (S/9)2-2eSapSaV + (SaV)2 + (l -e')p\
which a and a' are given and generally non-parallel unit-lines,

e and C are scalar constants, as capable of representing gene-
while
rally
(comp. 360,
(2.), (3.))
a central hut non- conical surface (fp = 1)

shall find that if, in passing from one
And we
of the second order.
such surface to another^ we suppose a and a' to remain unchanged,,

but e and C to vary together, so as to be always connected by the
relation,
II.
in
which
* It
is
I is
to be
some
C = {e^-l)
real, positive,
remembered
(e
+ Saa')l\
and given
scalar,
then
that, in the formula here cited, the
all
the sur-
symbols
a, a' did
not denote unit- vectors.
t
arises
When
these two vectors a, a' remain constant, but the scalar e changes, there
with a concentric
:
or, by their intersections
a system of biconfocal cones

system of biconfocal sphero-conics.
between two opposite directions, for
the
choice
to
the
Or
according
X
difference,
The angular transformation XXXVII. may be accomone of the two focal lines.
plished, by resolving the equation XXXV. as a quadratic in e, and then interpreting
sphere, a
the result.
644
faces
the
60 deduced, or in other
I.
common
Tir
'"
with
words the surfaces represented by
+ (SaV)^ +
(a^-l)(6 + 8aaO
(Sarf'- 2eSapSaV
nf
='^''=
e for the only varialle parameter^
(1.)
The scalar form
Cv=
so that the function
And
we
write, for
shall
C(pp
= (a - ea) Sap +
is self-conjugate, as it
(e2
(a'
- ea) Sa'p + (1 - e^) p,
ought to
be.
thus,
1) I2<pa
= a' - ea,
(e^
1)
I'-^â'
= a - ea',
^2
= (g +
1) /2,
62
(e
+ Saa') ^2,
c^
(e
1)1^,
have the values,

VIII. ..
Lvaa'
=-6-2Vaa',
^(a - a') = - c-'^(a - a')
(
comparing which with 405,
(1.), (2.),
we
IX.
any one
aU(a +
of which the direction
the surface fp
as asserted.
=1
a'),
see that the three

(real or imaginary)
lines,
of
abridgment,
VII.
we
form
(e2-l)(e + Saa')
we have
because
VI.
if
'
also be thus written, with the value II. of C,
V.
(2.)
(1 -e')^"
compose a Confocal System.
III. o^fp gives the connected vector
^^
which may
III.
equation,
^
[bOOK
cU(a -
bJJYaa,
a'),
be reversed, are the three vector semiaxes of

and therefore, by VIL, that the si/stem III. is one of confocals,
may
(3.) The rectangular transformations, scalar and vector, are

X., and 357, V. VIII.)
now (comp.
405,
-''
XI.
+1
PV = IHp =
+ Saa'
U(a-a').SpUCa-a')
which can both be
established,
ways, and from the
first
by the
of which
rules of the present Calculus, in several other
follows that (as
is well
known) through any proposed point p of space there can in general be drawn three confocal surfaces, of a
one an ellipsoid, for which e> 1, and therefore a^>h->c^>{i;
given system III.
it
another a single-sheeted hyperholoid, for which
and the third a double- sheeted hyperholoid,

>62>c2.
e<\,
for
> - Saa',
a^
which c<-Saa',
> 62 > q > c2
c>
1,
o2>0
CHAP.
III.]
(4.)
From
the oÂer rectangular transformation XI.
= pip what
the normal vector v
= <pp
it
follows, that if
we
645
denote
when
p remains the same, but

changed to a secoyid root e\ of the equation III. or X. of the surface, considered
\)y VI
e is
RECTANGULAR SYSTEM OF THREE NORMALS.
as a cubic in
becomes,
then
e,
XII.
^^^^= P<pvi=
e
ci
XIII.
but
= P(pi(pp = l^(p<PiP
Spvi = Spr =/ip =/p =
l-(piv
the substitution of
/ip being formed from/p, by
XIV.
and the known theorem
(5.) It follows,
as follows
ei for e
therefore,
= Sp(pvi = Svi(pp = Sviv,
results, that confocal surfaces cut
each other orthogonally.*

(p'^p can be expressed
from V. and VI., that the inverse function
XV.
or that p
1,
may
= l^(aSa'p + a
0-ip
Sap)
b^p
be deduced from v by the formula,
XVI.
= ^-ij/ = Z2 (aSaV + a'Sai/) - b^p,
which can easily be otherwise established.

of the surface reciprocal to the surface
I.
Hence (comp. 361, (4.)), the equation

new surface which has v
or III., or of that
is
(instead of p) for its variable vector,
XVII.
= Fv = Sv<p-^v = 2Z2SavSaV - b'^v^
the fixed focal lines a, a' of the confocal system III., or of the corresponding system
becoming thus (in agreement with known results) the^xec?

normals (or cyclic lines, comp. 361, (6.)) of the reciprocal system XVII.
from III., no use has been made of
(6.) In thus deducing the equation XVII.
of the asymptotic cones,

cr/elic
and 0p.
Without the
could therefore have inferred, by a slight modifi-
the rectangular transformations X. XI., of the functions /p
transformations last referred
to,
we
XVII., that the reciprocal surface (Fv= 1) with v for its variwhich has the same rectangular system of directions for its three semi-
cation of the form

able vector,
axes as the original surface (fp
= 1),
but with inverse squares (the roots of
its
cubic) equal to the direct squares of the origitial semiaxes, has for equation (comp.
40*, XII.),
XVIII.
= Fv = P (Sava'v - ev^) = SXvfiv + gv^,

n = la', g = - el^ = - eTXfi
.l
XIX. ..X = Za,
if
being thus deduced anew, but by a process quite

rent from that employed in (2.), under the forms (comp. 405, XIII.),
the values VII. of a',
XX.
a-2
= C3 = -g + TXfi;
b^
= C2 = -g+8Xii
while the directions IX. of the corresponding semiaxes

as, a2, ai,
diffe-
b^, C'
c^
may
= ci=-g-TXfi',
be deduced as those of
from the formulae 405, XIV.
(7.) If the
symbol w (v), or simply
<ov,
be used to denote a new linear and
self-
conjugate vector function of v, defined by the equation,
XXI.
We shall
wv = pSpv P (aSa'v +
a'Sav),
soon see that the same formula XII., by expressing that v,
vi,
and
fvi or 0iv are complanar, contains this other part of the known theorem referred
that the intersection is a line of curvature, on each of the two confocals.
to,
646
with p here treated as a vector constant, then (because

be thus written (comp. 354, &c.),
III.
[bOOK
XVI.
1) the equation
Spv=
may
XXII.
(a;
6^-)'
0;
the three rectangular directions, of the three normals v, Vi, v^ to the three confocols through p, are therefore those which satisfy (comp. again 354) the vector quadratic equation,
XXIII.
..
and they are the directions of the axes of

357, &c.),
XXIV.
in which p
= 0;
new surface of the second order (comp.
Svwv = (Spv)3 - 2Z2SavSa'v =
treated as a constant vector, but
is still
The
(8.)
Vrwj/
this
new
inverse squares of the scalar semiaxes of this
are the direct squares

three confocals
b"^,
XXV-
<p
= w + m'6 + m" {h-^y +
1),
(&2)3,
m", deduced here from the new function w, as they

in the Section III. ii. 6, have the values,
in which the coefficients m,
were deduced from
surface (Sj/wv
what may be called the mean semiaxes of the

squares must therefore be the roots of this new cubic,
&i2, hz^ of
these latter
1,
v as a variable one.
to',
Im =Z4(Saa'p)2;
XXVI.
m'
= Zi (Vaa')2 + 2Z2S {Yap .Va'p)
(m"=p2we
if
Accordingly,
observe that (because
2^2 Saa'.
Ta = Ta' =
1)
we have among
others the
transformation,
XXVII.
(Saa'p)2=p2(Vaa')2-(Sap)2-2Saa'SapSa'p-(Sa'p)2,
we can
express this last cubic equation

efficients, under the form,
XXVIII.
which,
when we change
XXIX.
these values
XXVI.
of
its
= (63+p2) {(62_Z2Saa')2_;4}
+ 2/2 (62 _ Z2Saa') SapSa'p - U {(Sapy + (Sa'p)2)
comes the cubic
XXV., with
in
e,
b^
by VII.
to its value l^{e
+ Saa'),
and divide by
I*,
co-
be-
or the equation III. under the form,
= (62 -
1)
{p{e + Saa') + p2}
+ 2eSapSa'p - (Sap)2- (Sa'p)2.
(9.) As an additional test of the consistency of this whole theory and method,
the directions of the three axes of the new surface XXIV., or those of the three
normals (7.) to the confocals, or the three vector roots (354) of the equation
XXIII., ought to admit of being assigned by three expressions of the forms,

=il/(T
nivi =
+b^x^
\pffi 4-
b i^x^^
+b*<T,
+ *i^î>
n2V2 = ipaz + bo'ix<T2 + bzâz

inv
in
which
a,
<Ti,
(T2
b^,
bi% bo} are the three scalar roots of the cubic
are three arbitrary vectors ; n, i,
712
XXV.
or
XXVIIL,
while
are three scalar coefficients, which
can be determined by the conditions Spi/= Spi/i = Spj/3= 1 (comp. XIII.); and \f',
X are two new auxiliary linear and vector functions^ to be deduced here from the
function
co,
referred to.
in the
same manner as they were deduced from
<p
in the Section lately
CONE RESTING ON FOCAL HYPERBOLA.
CHAP. Ill]
647
for convenience the

(10.) Accordingly, by the method of that Section, taking
the subject of the opegiven* vector p (instead of the arbitrary vectors cr, (Ti, cz) as
rations \p and X, we find the expressions,
XXXI.
.ipp
XP = i^(aâ'p + a Sap 2pSaa')
= l*Yaa'Saap,
whence, after a few reductions, with elimination of n by the relation Spv

the cubic in b-, the first equation XXX. becomes
1,
and by
XXXII.
= (62 V + p)
{(62
_ Z2Saa')2 - 1*}
+ li(b9_ pSaa)
which
is in fact a
form
as appears (for instance)
(aSa'p + a'Sap)
between v and
of the relation
by again changing
P (c
62 to
- 1* (aSap + a Sap)
any one of the confocals,

+ Saa'), and comparing with
p, for
the equation IV.
(11.) Another and a

still
or
more
interesting auxiliary surface, of
the directions of the normals v, is found
inverting the
by
which the axes have
new
linear function
o*,
by forming from XXII. the inverse equation,
XXXIII...
(w-
+ 6-2)j/=0;
in which,
XXXIV.
and from which
a)-iv.(Saa'py
it
directions of the axes of this
XXXV.
=YaaSaa'v + l-^(YapSapu + Ya'pSapv);

v
follows that the normals
Sj/w- v
new
XXXVI.
or
0,
to the confocals
through p have the
cone,
.
= ^^ (Sau'v^ + 2SapvSapv,
with p treated as a constant, as before.

(12.)
The
vertex of this auxiliary cone being placed at the given point p, of in-
we may inquire in what curve is the cone cut, by

the plane of the given focal lines, a, a', drawn through the common centre o of all
the surfaces III.
Denoting by a = ta + t'a' the vector of a point s of this sought
tersection of the three confocals,
section,
and writing
XXXVII.
the equation
XXXVI.
= (r-p=<a-f
t'a
- p,
gives the relation,
XXXVIII. ..' = - =
2
const.
is therefore an hyperbola, \i]\\c\\ \& independent of the point -p, and has
the focal lines of the system for its asymptotes. And because its vector equation may
the section
be thus written (comp. 371, II.),
XXXIX.
or
what may be
XL.
it
satisfies
(j
= ta + im-W,
called its quaternion equation as follows (comp. 371, I.),

.
2Ya<T.Y<Ta'
= P(Vaay,
XLI.
with the significations
which is known by the
XXVI.
name
of
m = 0,
and m'
;|/(t
and
P-
to
= 0,
therefore that important curve,
of the Focal Hyperbola :f
* The
general expressions for
t Compare the Notes
m'
it is
pages 231, 505.
namely the
x'^ include terms,
limit to
which
which vanish when
648
[bOOK
III.
the section of the confocal surface by the plane of its extreme* axes tends, when the
axis (26) tends to vanish. "We are then led thus to the known theorem, that
mean
ifJ with any assumed point F for vertex, and with the focal hyperhola\ for base, a
cone he constructed, the axes of this focal cone have the directions of the normals
to the confocals through P.
As
(13.)
regards the Focal Ellipse,
its
two scalar equations
may
be deduced
from the rectangular form X., by equating to zero both the numerator and the denominator of its last term they are therefore,
;
XLII.
S(a-a')p = 0,
2/2
= (SpU(a +
a'))^
+f
^^^^J
the curve being thus given as a perpendicular section of an elliptic cylinder, with
?V2 and ^V (1
or of the ellipse
The same curve may
(14.)
or
- c2)i and
Saa'), or (a^
_ c2)^,
(62
for the
semiaxes of
its
base,
itself.
XLIII.
XLIir.
be represented by the equations,
also
Sap =
TVap = (62 _ c2)i,

TVa'p = (62 - c2)i
Sa'p,
Sa> = Sa|0,
which express that it is the common intersection of its own plane (-i- a a) with two
right cylinder s,X which have the two focal lines a, a' of the system for their axes of
revolution,
and have equal
radii,
denoted each by the radical last written.
(15.) In general, the unifocal
namely,
XLIV.
in which
a and
(1
((Vap)'
e2)
+ 62) +
(S(a'
- ea) p)\
may be interchanged, shows that the two equal right cylinders,

XLV.
(Vap)2 + 62 = 0, XLV.
(Va'p)2 + 62 = 0,
XLVI.
TVap = 6, XLVI'.
TVa'p = 6,
real if their common radius 6 be such, that is, if the confocal (e) be either
.
which are
an
III.,
a'
or
(13.)) of the equation
form (comp. 406,
and which
ellipsoid (supposed to be rea7), or else a single-sheeted hyperboloid,
have the ^ca^
lines a, a' of the
confocal surface
system for their axes of revolution, envelope^ that
the planes of the two ellipses of contact (which again are real
curves, if 6 be real) being given
XLVII.
by the equations,
= 0,
S(a' -ea)p
XLVII'.
S (a - ea) p
of the surface (or of the system), and are the

(real) director planes (comp. 406, (14.)) of the asymptotic cone (reaX or imaginaxy),
so that they pass through the centre
to the particular confocal (e).
Namely, those two of which the squares algebraically include between them
this latter being, for the same reason, considered here as the mean.
that of the third
We shall soon see that
theorem, in which this
J The reader
may
is
consult page 513 of the Lectures, for the case of this theorem
which answers to a given

rally
quaternions give, with equal ease, a more general known
included as a limit.
ellipsoid.
p
or
The focal
ellipse
may
also be represented gene-
by the expression (comp. page 382 of these Elements),
= (cfi -
c2)i
V. a<U (a
+ a') (
by the same expression, with a and a interchanged.

Compare pages 199, 228, 233, 299.
CHAP.
CENTRO-FOCAL ELLIPSES.
III.]
649
(16) Whetker the mean semiaxis (b) be real or iraaginary, the surface III.
be itself real) is always, by the form XLIV. of its equation, the locus
(supposed to
of a system o^ real ellipses (comp. 404, (!.)) in planes par aZ/e/ to the director plane
XLVII., which have their centres on the focal line
on a plane perpendicular to that line.
a,
and are oxihogonaWy-projected
into circles
(17.)
The same surface
is
also the locus of a second system of such ellipses, re-
and
lated similarly to the second focal line a',
to the second director plane
XLVII'.
appears that these two systems of elliptic sections of a surface of the second
order, which from some points of view are nearly as interesting as the circular sections^ may conveniently be called its Ceutro'Focal Ellipses.
and
it
when
(18.) For example,

the equation of the ellipsoid
404, L) of
thej^rs^ quaternion form (204, (14.),
employed, one systerai of such ellipses coincides with
the system (204, (13.)) of which, inihQ first generation^ of the surface, the ellipsoid
is
Besides that first generation (I) of the Ellipsoid, which was a double one, in
the sense that a second system (17.) oi generating ellipses might be employed, and
which served to connect the surface with a concentric sphere, by certain relations of
homology (274)
and the second double generation or construction
(II),
by means
oitwo diacentric spheres (217, (4.), (6.), (7.), and 220, (3.)), which was.
several other generations of the same important
illustrated by Fig. 53 (page 226)
of either
were deduced from quaternions in the Lectures, to which

refer. A reader, then, who happens to hav a copy of that
surface
it is
only posaLble
work, may
system of two reciprocal ellipsoids, with
a common mean axis (2b), by means of a moving sphere, of which the radius (= 6)
is given, but of which the centre has the original ellipsoid for its locus ; while the
here to
consult page
499
earlier
for degeneration (III) of a
corresponding point on the reciprocal surface, and also the normals at the two points,
are easily deduced from the construction. In page 502, he will find another and per-
haps a simpler generation (IV), of the same pair of reciprocal ellipsoids, by means of
mean sphere, comp. 216,
quadrilaterals inscribed in a fixed sphere (the common
the directions of the /owr sides of such a quadrilateral being given^ and one
(10.))
pair of opposite aides intersecting in a point of one surface, while the other pair have
;
point of the other (or reciprocal) ellipsoid.

In the page last cited, and in the following page, there is given a new double generaits circular sections (of either system) being contion (V) of any one ellipsoid
for their intersection the corresponding
of two equal spheres (or spheric surfaces), of which the line

of centres retains a fixed direction, while the spheres slide within two equal and
centre of the generated surright cylinders, whose axes intersect each other (in the
structed as intersections
face),
and of which the common radius
of the
same volume, there
soid (abc), analogous to
new
is
the mean semiaxis (b).
directions of the sides of the quadrilaterals,
tion) inscribed in the circles of a certain
revolution,
Finally, in page
699
new generation (VI) of the original ellipthe generation (IV) by the fixed (mean) sphere, but with
will be found a
mean
which are
also (in this last genera-
ellipsoid (or prolate spheroid) of
which has the mean axis (2b) for its major axis, and has two medial
whose common distance from the centre is represented by the ex-
foci on that axis,

pression,
V(a2_62)V(fe2->^
V(a-6* +
4o
'
c2)
650
[bOOK
III.
treated as the locus ; and an analogous generation of the two hyperboloids, by

geometrical deformation of two corresponding surfaces of revolution, with certain
was
resulting homologies (comp. sub-arts, to 274), through substitution of (centro-focaT)

ellipses for circles, conducts to equations of those hyperboloids of the same unifocal
form
a and
if
namely,
)3
have
tion of the ellipsoid (so that
/3
significations analogous to those in the cited equa-
and not a
here
is
a.
focal
XLYIIL..(s-^y+(v|y
line),
1;
the upper or the lower sign being taken, according as the surface consists of one
sheet ^r of two.
(19.) It may also be remarked that as, by changing /3 to a in the corresponding
equation of the ellipsoid, we could return (comp. 404, (1.)) to a form (403, XI.) of
the equation of the sphere, so the same change in XLVIII. conducts to equations
of the equilateral hyperboloids of revolution, of one sheet and of two, under the very
simple forms* (comp. 210. XI.),
XLIX.
..
^Y = -l,
sf
and
=
..s[^y +l;
L.
which it seems unnecessary to insert points after the signs S, and of which the
geometrical interpretations become obvious when then they are written thus (comp.
in
199, v.),
LI.
where
T - = op
T^ = Vsec2f ?- ^^\
(i-'^)
oaj while Z
axis of revolution oa, and
is
Z.
LIL
the inclination
is
aop
T^=\/sec2z-:
of the semidiameler
op
to the
the inclination of the same semidiameter to a
plane perpendicular to that axis.

(20.) The real cyclic forms of the equation of the surface III. might be deduced
from the unifocal form XLIV., by the general method of the subarticles to 359 but
since we have ready the rectangular form X., it is simpler to obtain them from that
;
form, with the help of the identity,
LIII.
by
- p = (SpU (a + a)y + (SpU Vaa')2 + (SpU (a - a')),
eliminating ihQ first of these three terms for the case of a single-sheeted hyperho-
the common tangent planes, to this mean (or medial) ellipsoid, and to the given (or
generated) ellipsoid (âbc^, which are parallel to their common axis (26), being parallel also to the two umbilicar diameters of the latter surface.
* The same
forms, but with <t for p, and (3 for o, may be deduced from XLVIII.
the
on
plan of 274, (2.), (4.), by assuming an auxiliary vector c such that
pa
S = + S -,
and
V=V
P
the homologies, above alluded
to,
between the general
hyperboloid of either species, and the equilateral hyperboloid of revolution of the

same species, admitting also thus of being easily exhibited.
CHAP.
III.]
loid (for
which
EXPONENTIAL FORMS, EQU. OF CONFQCALS.

> a-2 > > c-2)
b'^
the second for an ellipsoid (c
>
651
> a-2 >
fc*
0)
and the third for a double- sheeted hyperboloid (a-2 > > c-2 > b'^).
of the surface III. maybe, we can always derive from
(21.) Whatever the species
XLIV.
the unifocal form
of its equation
what may be
an Exponential Trans-
called
namely the vector expression,
formation
LIV.
= ara +
LV.
with
y Va'/3,
+ y2/U Vaa' =
x^fa
remaining arbitrary, but the two scalar coefficients, x and y,

being connected by this last equation of the second degree
provided that the new
constant vector /3 be derived from a, a', and e, by the formula,
the scalar exponent,
t,
^
which gives
Ta =
+ Saa'
a few reductions (comp. the expression 315, III. for
after
a',
when
1),
LVII.
LVIII.
..
Va/3=UVaa', S(a'-ea)/3=
Va'i3 = /3S.a<+ UVaa'.S.a<-i
LIX.
Saa'/3=0;
0,
Y.aYa'(3 =
a*VYaa=T-n
LX.
S(a'-ea)p = a;(e + Saa'), Vap=yaUVaa'
= *-.
LXI.
while
/a = a-262c-, and LXII.
//3 =/UVaa'
If
we
treat
the
in
the
as
the
variable
expression LIV.
exponent, t,
only
(22.)
for p, then (comp. 314, (2.)) that exponential expression represents what we have
.
called (17.) a centra-focal ellipse ; the distance of its centre (or of its plane)
centre of the surface, measured along the focal line a, being represented
from the
by the co-
z ; and the radius of the right cylinder, of which the ellipse is a section, or
the radius of the circle (16.) into which that ellipse is projected, on a plane -^ a,
being represented by the other coefficient, y while ^tTr is the excentric anomaly.
efficient
(23.)
X and y
If,
on the contrary, we treat the exponent
as given, but the coefficients
as varying together, so as to satisfy the equation
LV.
of the second degree,
the expression LIV. then represents a different section of the surface III.,
made by
a plane through the line a, which makes with the focal plane (of
an angle
this latter section (like the former) being
but being an ellipse for an

because
be such
loid,
LXIIL
(24.)
And it
is
ellipsoid,
always
real,
if
and an hyperbola
a, a')
the surface itself
for either
hyperbo-
./a./UVaa' = a-2c-2 by LXI. and LXIL
scarcely necessary to remark, that
by interchanging a and
a'
we
obtain a Second Exponential Transformation, connected with the second system

(17.)
o{ centro-focal ellipses, as the^r*^ exponential transformation LIV. is connected with
the first system (16.).
= has likewise its two systems of centro-focal

(25.) The asymptotic conefp
and its equation admits in like manner of two exponential transformations^
ellipses,
of the
form LIV.
the only difference being, that the equation
the following,
LXI V.
z^fa
LV.
is
replaced
by
+ y2/UVaa' = 0,
and y* have opposite signs by (28.).

(26.) Finally, as regards the confocal relation of the surfaces III,, which may
in which, for a real cone, the coeflGlcients of x^
represent
any confocal system of surfaces of the second
order, it
may
be perceived
652
from (4.) that an essential character of such a relation

tion,
LXV.
Yv^<t>v,
= Yv(l>^v
is
[bOOK
expressed
III.
by the equa-
perhaps be called, on that account, the Equation of Confocals.

(27.) It is understood that the two confocal surfaces here considered, are repre-
which
may
presented by the two scalar equations,
Sp^p =
LXVl
Sp0,p
1,
= 1,
LXVI'.
or
/p
= 1,
/,p
=1
and that the two linear and vector functions^ v and v of an arbitrary vector p,
which represent normals to the two concentric and similar and similarly posited sur,
faces,
LXVII.
passing through
'
= const, fp = const.,
fp
p, are
any proposed point
LXVIII.
= ^p,
v^
= <p^p.
(28.) It is understood also, that the two surfaces LXVI. or LXVI'. are not only
concentric, as their equations show, but also coaxal, so far as the directions of their
tixes are
concerned
or that the two vector quadratics (comp. 354),
LXIX.
are satisfied
V|0^|0
= 0,
by one common system

it will
understandings,
and
LXX.
Yp^,p =
0,
And
with these
LXV., which has been
be found that the equation
above the Equation of Confocals,
of three rectangular unit lines.
called
not only necessary but sufficient, for the estab-
is
lishment of the relation required.

(29.) It is worth while however to observe, before closing the present series of
subarticles, that the equations XII,, and those formed from them by introducing
C2
and
give the following
V2,
LXXL
LXXII.
and
and
among
other relations
fVvi = (i^ - 6i)- = -fiVv

.
fiVv^ = (h^ - h'^Y^ = "/aUvi
.f{vi, vz) =/i(v2, v)
=/> {v, vi) =
&c.
therefore,
LXXIIL
./i{(628_6i2)JUj/2CM-62)mj/} =0;
easy to see that the two vectors under the functional sign/i in this last
expression have the directions of the generating lines of the single- sheeted hyperboloid (ej) through p, if we suppose that b-y^ > 6i2 >
> 6^, so that the confocal (eo) is
whence
it is
here an ellipsoid, and (e) a do2ible-sheeted hyperboloid.

(30.) But if (T be taken to denote the variable vector of the auxiliary surface
the equation of that surface may by (7.) and (8.) be brought to the followform, with the meaning XXI. of lo,
rectangular
ing
XXIV.,
LXXIV.
SfTwtr
= CSp(T)2 _
2Z2SaaSa'(T= ft2(S(TUv)2
-f
Sc7W(r
&i2(S(tUj/i)'+ &2'(SffUv2)*
cone
cyclic normals, or those of its asymptotic
= 0, that is of the cone
0, or the ybca^ lines of the reciprocal cone S(Tw''(T
or finally the /oca/ lines of the focal* cone (12.), which rests on the focal
hence, with the inequalities (29.),
its
XXXVL,
hyperbola, have the directions of the lines
nions.
LXXIIL
more general known theorem, including
those focal lines are therefore
this, will
soon be proved by quater-
CHAP.
CONDITION OF CONTACT WITH RIGHT LINE.
III.]
(by what has just been seen) the generating
of the hyperboloid
lines
653
(ci),
which
the given point p.

passes through
And
(31.)
LXXV.
for
The
408.
an arbitrary a we have the transformation,
Z-2(Spff)a- Sa(ra'(7= e(S(TUi/)2
+ ^..(SffUvO^.
ei(S(rUvi)'
general equation* of conjugation,

I...
(//>,/>')
405,111.
!,
connecting the vectors p^ p' of any two points p, p' which are con1, may
jugate with respect to the central but non-conical surface fp
be called for that reason the Equation of Conjugate Points ; while

the analogous equation,
11.
../(^,/>0
o,
which replaces the former for the case of the asymptotic conefp -
may be
0,
by contrast the Equation of Conjugate Directions : in

fact, it is satisfied by any two conjugate seniidiameters, as may be at
once inferred from the differential equation f{p, dp) = of the surface
fp = const, (comp. 362). Each of these two formulae admits of numerous applications, among which we shall here consider the
deduction, and some of the transformations, of the Equation of a
called
Circumscribed Cone,
la.
which may
line pp'
also
p')-\f^{fp-l)(fp'-\);
{f{p,
be considered as the Condition of Contact, of the right
with the surface fp=
1.
(1.) In this last view, the equation III.
(30.)),
I V.
V.
or
=x2(/p -
be at once deduced, as the condi-
may
and quadratic equation (comp. 216,
tion 0^ equal roots in the scalar
=f{xp + afp') -
1)
4-
(2.),
+ xy,
- 1) + x'H/p' - 1)
and 316,
(a:
2a:x'(/(p, p')
which gives in general the two vectors of intersection^ as the two values of the ex-
xp +
x'p'
pression
-.
x-\-x'
(2.) If
in
any
we
treat t\iQ point p* as given^
giveri direction
quadratic,
r by
ti'^T
and
and denote the two secants drawn from
tz-^r,
then
ti
and
/(p'+ '"'0 = 1, or
VI.. 0=/(<p' + r)-< = i2(/p'-l) +
.
denoting then by
and writing p
ô'^^-
the harmonic
= p' + fo'^, we
VII.
mean
of these
t^
2</(p',
two
r)+fn
secants, so that 2to
have
ô(l
-/pO =/(p',
* For the notation used, Art. 362
r),
may
it
are the roots of this other
f(p, p')
be again referred
to.
= ti-\- fg,
654
we
are then led in this
way
to the formula
I.,
[bOOK
as the Equation
of the point p', if that plane be here supposed to be defined
by
III.
of the Polar Plane

well-known har-
its
monic property (comp. 215, (16.), and 316, (31.), (32.)).

(3.) At the same time we obtain this other form of the condition of contact
III.,
as that of equal roots in VI.,
the
first
VIII.
is
VIII.../(p',r)2=/r.(/p'-l),
member being an abridgment of (/(p', r))' and becanse
:
homogeneous with respect to
r, it represents
= 1, from the given point

gents (r) to the given surface /p
to prove that the equation III. may be thus written,
IX.
- p'Y =f(p - p')
/(p',
this last equation
a cone, namely the Cone of Tan-
(/p-
p'.
Accordingly
it is
easy
- 1),
last form it is seen to be homogeneous with respect to p p'.

Without
expressly introducing r, the transformation IX. shows that the
(4.)
equation III. represents some cone, with the given point p' for its vertex and be-
under which
cause the intersection of this cone with the given surface is expressed by the square
of the equation I. of the polar plane of that point, the cone must be (as above stated)
circumscribed to the surface /p

in
which that surface
is
cut
touching
1,
by that plane
it
along the curve (real or imaginary)
I.
(5.) Another important transformation, or set of transformations, of the equation
In general, for any two vectors p and p', if the

m, the vector function ;//, and the scalar function F, be derived from
the linear and vector function 0, which is here self-conjugate (405), by the method
III.
may
be obtained as follows.
scalar constant
of the Section III.
ii.
6,
we have
successively,
X. ../(p, p')2-/p./p' = Sp0p'.Sp'^p-Sp^p.Sp>p' = S(Vpp'.V0p^p')

= S pp'\pYpp' = mS. pp'ip-^Vpp' = mFYpp'
.
and thus the equation

XI.
if r = p p'
mFYpp'+f(ip
be a tangent from
p'.
p')
XIII.
= Sn//-i r +
XIV.
0,
XII.
or
mFVrp' -\-fr = 0,
m = C1C2C3 = - a'^b
Or because <pip = m, and
by 406, XXIV., we may write (with T

or
III. of the circumscribed cone becomes,
= p p')
Sv<p-iv,
'c'^,
either
if
= Vrp' = Ypp',
FYpp' = a^b^c^f(p -
p'),
as the condition of contact of the line pp' with the surface fp = 1.

geometrical interpretation, of this Isistform XIV. of that condition, can
(6.)
as follows.
be
Supposing at first for simplicity that the surface is an
assigned
easily
p be the point of contact, so that /p = 1, /(p, r) = ; and let the tangent

pp' be taken equal to the parallel semidiameter ox, so that/r =/(p p') = 1. Then,
with the signification XIII. of v, the equation XIV. becomes,
ellipsoid, let
XV.
in
\/Fv
= Tv.VFVv = abc;
which the factor Tv represents the area of the parallelogram under the conjugate
while the other factor VFUw resemidiameters op, ot of the given surface /p = 1
of the semidiameter of the reciprocal surface Fv 1, which is
the
reciprocal
presents
;
perpendicular to their plane
pot
parallel plane
equal,
by elementary principles,
formula
surfaces.
XV.
And
or the perpendicular distance between that plane,
which touches the given
and a
the result
ellipsoid
so that their product
^1
Fv
is
to the product of the three semiaxes, as stated in the
may
easily be
extended by squaring, to other central
CHAP.
AXES OF CIRCUMSCRIBED CONE.
III.]
655
(7.) It may be remarked in passing, that if p, c, r be any three conjugate semidiameters of any central surface /p = 1, so that
XVI.
and
if
= 1,
/p =/<r =/r
xp + i/ff + zrhe any
XVII.
and
/(p,
scalar equation,
XVIII..
=/(<r, r) =/(r, p)
(t)
other semidiameter of the
same
surface,
= 0,
we have then the
+ y<T + zr) = .r2 + y2 + 23=i;
./(a:p
a relation between the coeflScients, x, y, z, which has been already noticed for the
ellipsoid in 99, (2.), and in 402, I., and is indeed deducible for that surface, from
and real vectors alone
principles of real scalars
perholoids, one at least of those three coefficients
but in extending which to the hybecomes imaginary, as well as one
at least of the three vectors p, c, r.
Under the same
(8.)
XX.
ver}'^
the
(- m)i0-iVp(T = + (- m)-iV0p0(T
Sptrr
same
conjugate directions, that
is,
if
must
Strr
= Srwr,
if
(comp. II.) the equation,
satisfy
xx' + yy'
The equation VIII., with p

.
=
Sp(rr.S0p^<T0r -l.
xp + y<T + zr and x'p 4 y'ff + z't have only
they have the directions of any two conjugate semi~
XXIII.
XXIV.
m)-J
conditions, if
diameters, the six scalar coefficients
(10.)
= + aftc = + (-
simple relation,
XXII.
Under
(9.)
also,
Vp(T= + a6c0r= + (-m)-l0r;
XXI.
together with this
XVI. XVII., we have
conditions
XIX.
for p',
XXV.
+ zz =
may
,
(t
0.
be written under the form,
= wr = ^pSp^r + ^r (1 -/p),
a new linear and vector function, which represents a normal to the cone of tangents from p, to the surface yp = 1.
Inverting this last function, we find
XXVI...r = a>->^ = ^^-:i^^

the equation in a of the reciprocal cone, or of the cone
scribed cone from p,
XXVII.
is
S(ra>-i(T
of normals
to the circum-
therefore,
0,
or
XXVIII.
F<t
XXVIII'. .,F{a:
= (Sp(T)2,
Spcr)
or finally
a remarkably simple form, which admits also of a simple interpretation.

In fact,
the line <t Spff is the reciprocal of the perpendicular, from the centre o, on a tan:
is also a tangent plane to the surface ; it is therefore one

of the values of the vector v (comp. (6.), and 373, (21.)), and consequently it is a
semidiameter of the reciprocal surface Fv= 1.
(11.) As an application of the equation XXVIII., let the surface be the confo-
gent plane to the cone, which
by the equation 407, III. or X,, of which the reciprocal is reby 407, XVII. or XVIII. Substituting for F<t its value thus deduced,
cal (e), represented
presented
the equation of the reciprocal cone (10.), with
XXIX.
if
2^2Sa(rSaV
- (Spfr)2 =
62^2,
or
<t
for a side,
XXIX'.
then the vertex p he fixed, but the confocal vary,
becomes,*
Saaa'a
- /-2 (Sp(T)2 = ea"^
by a change of
e,
or of 6' which
* It
may be observed that, when 6 = 0, this equation XXIX. represents the
asymptotic cone to the auxiliary surface 407, XXIV. ; and at the same time the reciprocal of that /ocaZ cone, 407, XXXVI., which rests on the /ogaZ hyperbola.
656
varies with
tem
the cone
it,
whence
XXIX.
III.
[bOOK
will also vary, but will belong to a hiconcydic sys-
follows that the (direct or) circumscribed cones from a given point are
it
:
and also, by 407, (30.), that their common focal lines are the generating lines of the confocal hyperholoid* of one sheet, which passes through their
all hiconfocal
common
vertex.
(12.)
Changing
e to
e^
XXIX'., and using the transformation 407, LXXV.,
in
with the identity (comp. 407, LIII.),
we
find that if
= (S(tUv)2 +
cr
(StrUvi)
+ (S(tUj/2)2,
be a normal to the cone of tangents from p to (e^\
<7
the
it satisfies
equation,
XXX.
= (e - cj (S(tUj;)2 + ê^ _ ^ )
and therefore that
= ( - O'^
t be a tangent
(SrUvi)2
from the same point
- cj
(eg
p, to the
(SaUi/z)^
same oonfocal
(e,),
other condition,
it satisfies this
XXXI.
if
(SrUy)2
+ (ei - c,)-i
(SrUvi)
+ (e?- O'*
(SrUrz)^,
which thus
a form of the equation of the circumscribed cone to (e^), with its ertex at a given point p
the confocal character (11.) of all such cones being hereby
exhibited anew.
is
(13.) It follows also from
XXXI.,
have the directions of the normals
that the axes of every cone thus circumscribed

vo, to the three confocals through p ; and
r, vi,
known theoremf may be otherwise deduced, from the Equation of Confocals

That equation gives
(407, LXV.), by our general method, as follows.
this
f^ V
XXXII.
(v,
Ij
- v)
^v
Svi/^
= 0,r), and therefore,

(because ^v,
= ip,v{fg - 1), Yvv,S,vv^-\- Vi/^,v(l -/p) =
changing then "V to S, and v to
r,
we
XXXII., have
this last vector quadratic
the directions of the axes of the cone, with
r for side,
XXXIII.
that
is,
by
(14.)
./Xp, r)2 +/r.(l
-/p) =
P to (ej.
application of the formula XIV., with the abridged symbols r and v
VIII., the directions of the axes
As an
see that v, vi, V2, as being the roots (354) of
of the cone of tangents, from
of (5.) for p - p' and Vpp', the condition of contact of the line pp' with the confocal (e) becomes, by the expressions 407, III., XVIII., and VII. for the functions
f, F, and the squares a^, 62^ c2, the following quadratic in e
:
XXXIV.
- e^) t^ = f^ (Sawa'v - ev^)

(Sar) 2eSarSa'r + (SaV) + (1
there are therefore in general (as
system, which touch a given right line

roots of the last equation
confocals, say (e)
and
their parameters,"!^ e
and
known) two
is
and
for instance, their sum, is
XXXV.
(e
(e^),
of
a.
given
e^ are the two
+ c,)r2 = /-2i,2 _ 2SarSaV.
* This theorem
(which includes that of 407, (30.)) is cited from Jacobi, and
proved, in page 143 of Dr. Salmon's Treatise, referred to in several former Notes.
f Compare the second Note

X This
e
a^
-J
+ c^
-,
name
of parameter
to
is
is
page 648.
here given, as in 407, to the arbitrary constant
of which the value distinguishes one confocal (e) of a system from another.
CHAP.
CIRCUMSCRIBED RIGHT CONES.
III.]
667
and
(15.) Conceive then that p is a ffiven semidiameter of a given confocal (e),
is a tangent, given in direction, at its extremity ; the equation XXXIV. will
(1/0
then of course be satisfied,* if we change r to dp, and v to Vpdp, retaining the given
that
value of e
j)\
but
it
same p and dp (or
will also be satisfied, for the
uhen we change
XXXVI.
e,
for the
same r and
new parameter,
e to this
= - e + 2SaUdp
Sa'Udp
- 1'^ (VpUdp)^
say, the neiv confocal (e ), with a parameter determined by this last formula, will touch the given tangent to the given confocal (e).
Z2 = Q in the equation 407, III. of a Confocal System
(16.) If we at once make
tliat is to
parameter e finite, we fall back on the system 406,

but if we conceive that l^ only tends to zero, and
at the same time tends to positive infinity, in such a manner that their pro-
of Central Surfaces, leaving the
XXXV.
that e
of Biconfocal Cones
duct tends to
a.
finite limit, r^, or that
XXXVII.
.. lim.;
0,
lim.c = oo,
lim.cZ2
= r2,
then the equation of the surface (e) tends to this limiting form,
XXXVIII.
p2
a system of biconfocal cones
is
+ r2=
0,
XXXV II
or
I'.
Tp =
therefore to be combined with a system of concentric
make up a complete confocal system.

(17.) Accordingly, any given right line ff' is in general touched hy on]y one
cone of the system just referred to, namely by that particular cone (e), for which
spheres, in order to
XXXIV.) we have the value,

e= Sava'v-\ or XXXIX'.
XXXIX.
with V = ypp\ as before, so that v is perpendicular
(comp.
contains the vertex
406,
XXXV., when
meter,
and the
+ Saa' = 2Sai;Sa'u-,
to the given
plane opp', which
line ; in fact, the reciprocals of the biconfocal cones
but e as a variable para-
a, a' are treated as given unit lines,
compose the biconcyclicf system (comp. 407, XVIII.),
XL.
Sava'v
= ev^.
But, besides the tangent cone thus found, there

centre
of which,
by passing
from the same formula
to the limits
XXXIV.
is
a tangent sphere with the same
XXXVII.,
the radius r
may
be found
to be,
XLl...r = T'i=T^;
T
and such
is
in fact
from the origin on the given line pp'.

(18.) In general, the equation XXXIV.
with p for
its
is
for the length of the perpendicular
a form of the eqtiation of the cone,

a
p', and is circumscribed to
variable vector, which has a given vertex
given confocal (e).
* In fact
p-p
an expression (comp. 316, L.)
it
Accordingly,
by making
= Saa'
in that formula,
we
are
follows easily from the transformations (5.), that
/p./dp-a-26-2c-2/?'Vpdp=/(p, dp)2.
t The bifocal form of the equation of this reciprocal system of cones XL. was
given in 406, XXV., but with other constants (X, /x, g), connected with the cyclic
form (406,
I.) of the
equation of the given system.
4 P
658
led (after a few reductions, comp. 407,
XXVII.)
to
III.
[bOOK
an equation which
may
be thus
written,
XLII.
with the variable side t
and T
of p'
and
for p
0=l\Saa'ry + 2Sap'rSa'p'T,
= p p\
v,
as before
and which
from the equation 407,
differs
XXXVI.
only by the substitution
for thsLt focal cone,
which
on the focal hyperbola. The other (real) focal cone which has the same arbitrary vertex p', but rests on the focal ellipse, has for equation,
rests
XLIII.
as
found by changing e to
is
/2(S(a-a2 = Saua'v-i>2,
in the same formula XXXIV.
however simpler, or at least it gives more symmetric results, to change

Saa' for the focal hyperbola, and to + 1 for the focal ellipse, in
e, in XXXI. to
order to obtain the two real focal cones with p for vertex, which rest on those two
(19.) It
is
curves; while that third and wholly imaginary focal cone, which has the same vertex, but rests on the known imaginary focal curve, in the plane of b and c, is found
by changing
e^
to
1.
This imaginary focal cone, and the two real ones which rest
may thus be represented by the
as above on the hyperbola and ellipse respectively,

three equations,
XLIV.
XLV.
XLVI.
. .
= fl-2(SrUv)2 + i-2(SrUvi)2 + af^(SrVvî

= 6-2(SrUj/)2 + 6i-2(SrU/i)2 + bi'^lSrVv^y
= c-2 (SrUv) + ci-^ (SrUvi)2 + 03-2 (SrUj/2)2
T being in each case a side of the cone, and
v, vi, v^
having the same
significations
as before.
(20.) On the other hand, if we place the vertex of a circumscribed cone at a point
of a. focal curve, real or imaginary, the enveloped surface being the confocal (cj,
we find first,
same order
by XXX.,
for the reciprocal cones,
or cones
of normals
<t,
with the
of succession as in (19.), the three equations,
XLVII...a2(sUvff)2
= a,2;
XLVIII...62(SUv(r)2=6,2;
XLIX. ..c2(SUj/(t)2=c,2;
and next,
for the circumscribed cones themselves, or cones of tangents r, the con-
nected equations
L. ..
LI...
LH.
all
cal
a2(VUrr)2+a^2 = 0;
= 0;
+ c,2 = 0;
62 (VUvr)2+ 6,2
c2(VUi/r)2
which have the forms of equations of cones of revolution, but on the geometrimeanings of the three last of which it may be worth while to say a few words.
(21.) The cone L. has an imaginary vertex, and is always ?Ve{/" imaginary but
the two other cones, LI. and Lll,, have each a real vertex p, with ^2 >o for the
;
first,
and
c2
<
for the second
6 being the
mean semiaxis
of the ellipsoid, which
passes through a given point of the focal hyperbola, and c2 being the negative and
algebraically least square of a scalar semiaxis of the double-sheeted hyperboloid,
which passes through a given point of the focal ellipse: while, in each case, v
has the direction of the normal to the surface, which is also the tangent to the curve
at that point, and is at the same time the axis of revolution of the cone.
(22.)
The semiangles
tive sines the
of the
two quotients.
two
last cones, LI.
and LIL, have
for their resp.
CHAP.
TWELVE UMBILICAR VECTORS.
III.]
LIII.
two cones
&,
aud
6,
LIV.
(-
c^2)i
(-
659
c*)^
circumscribed to a single-sheeted hyperboloid, because, for such an euveloped surface (cj, b^ is real, and less than the b of
any eonfocal ellipsoid, while c, is imaginary, and its square is algebraically greater
eacfi of those
is
therefore real,
if
(or nearer to zero) than the square of the imaginary semiaxis c of every doublesheeted hyperboloid, of the same eonfocal system
but the cone LI. is imaginary, if
;
the enveloped surface (ej be either an hyperboloid of two sheets (b, imaginary), or
an exterior ellipsoid (b^> i) and the other cone LII. is imaginary, if the surface
;
(e) be either any ellipsoid

loid
(^2<
bola,
or else an exterior
-c^> c^).
c^<c^,
Accordingly it
the locus of the vertex of the cone LI.,
a^,
which
(c, real),
is
sheeted hyperboloid of the system
while the focal
the loctcs of the vertex of the cone LII.,
is
is
and dowiZe-sheeted hyperbo-
known
lies
that the focal hyper-
entirely inside every double-
ellipse,
which
is
interior to every ellipsoid
in like
manner
and real tan-
gents to a single-sheeted hyperboloid can be drawn, from every real point of space.
(23.) The twelve points (whereof only four at most can be real), in which a
surface (e) or (âbc) is cut by the three focal curves, are called the Umbilics of that
surface ; the vectors, say w, w,, w^^, of three such umbilics, in the respective planes
of ca, ab, be, are :
LV...a>
=^(a
a')
1-Saa'
c{a-a')
LVII.
"
+ Saa'
|(a-a');
1-Saa'
V^^bYaa'
1 + Saa'
'
and the others can be formed from
these, by changing the signs of the terms, or of

four real umbilics of an ellipsoid are given by the formula LV.,
The
some of them.
and those of a double-sheeted hyperboloid by LVI., with the changes of sign just
mentioned.
(24.) In transforming expressions of this sort,
it is
useful to observe that the ex-
pressions for the squares of the semiaxes,
= Z3(e + l), 6^- = /2( + Saa'), c2 = ^2(e-l),

=
Ta Ta'= 1, give not only a^ c^ = 2P, but also,
fl2
combined with
LVIII.
T_^ = ^ -^^ = eoslL~ = [-,^]
a-a'
and
407,
II
_
=
LIX...T-^
J-^-=s.niz-=( ^-^
2
"A/"
LX.
TYaa = V(l - (Saa')2) = sin ^ - = (az _ 62)J (52 _ c3),
.
Z-2
with the verification, that because
LXI.
therefore
VIL
LXI',
- a) (a + a') = 2Yaa',
(a
T (a - a) T (a + a') = 2T Vaa'.
.
"We have also the relations,
LXII.
LXIII.
- a)-3 =
a')-2 + T(a
(TVaa');'
+ aO"' - T(a - a')'^ = Saa'. (TVaa')-s
T(a +
T(rt
with others easilv deduced.
660
(25.)
which
all
Fig. 84.
The expression LV. conducts
to the following
admit of elementary verifications

Let u, u' be the two real points
in which an ellipsoid (ahc)
cut
is
* and
may
[BOOK
III.
other consequences,
among
be illustrated by the annexed
by one
branch of the focal hyperbola, with h for

summit, and with F for its interior focus
;
summit
the adjacent major
of the surface
being e, and r, r' being (as in the Figure)

the adjacent points of intersection of the
same
surface with the focal lines a,
a',
that
with the asymptotes to the hyperbola.

Let also v, t be the points in which the
is,
same asymptotes a, a' meet the tangent to

the hyperbola at u, or the normal to the
ellipsoid at that real umbilic, of which we may suppose that the vector ou is the ot
of the formula LV.
and let s be the foot of the perpendicular on this normal to the
;
surface, or tangent
tv
to the curve, let fall
from the centre
Then, besides the
o.
obvious values,
LXIV.
and the obvious
is
summit
at once a
the surface
OF =
(a^
of the focal ellipse,
- c2)i,
tv
and 407, VIII. LXI.)
LXVI.
ov = (a +
su-i = ^w = -
LXVII.
(a +
^
OR = - =
V/a
bisected at u,
and that the point

ellipse in which
a')
ab'^ca,
we
shall
have these vector expres-
ot = (a-c)a\
c) a,
3h = (2 - &^)J,
is
and a focus of that other
cut by the plane (ac) of the figure,
is
sions (comp. 371, (3.),
LXV.
OE =
relations, that the intercept
TV = a(a-a') + c(a +
(a
or'
- a'),
su
-;
=-
ah~^ca'
tu
a');
;
V/a'
whence follow by (24.) these other values,
LXVIII. ..ov = a+<?,
OT-a-c, tv = 26;
TU = uv = &, su = or = or' = 6"'c;
LXX. .6u = Ta> = (2_62^c2)i.
OS = (a2 - 62 + C2 _ a26-2c3)i = 6-1 (2 _ 52)i (62 _ c^)\,
LXIX.
LXXI.
(26.) It follows that the lengths of the sides OV, OT, xv of the umhilicar triangle
TOV are equal to the sum and difference {ac) of the extreme semiaxes, and to the
mean axis (26) of the ellipsoid ; while the area of that triangle = os TU = (j^^ ^*)^
(62
c2)i =
as real.
the rectangle under the two semiaxes of the hyperbola, if both be treated
The length (T^w)-', or sij, o the perpendicular from the centre o, on the
tangent plane at an umbilic u, is ab'^c and the sphere concentric with the ellipsoid,
which touches the four umbilicar tangent planes, passes through the points K, r' of
;
intersection of that ellipsoid with the focal lines , a', that
is,
as before, with the
* Some such verifications were

given in the Lectures, pages 691, 692, in connexion with Fig. 102 of that former volume, which answered in several respects
to the present Fig. 84.
CHAP.
EIGHT UMBILICAR GENERATRICES.
III.]
asymptotes to the hyperbola

meter ou,
by
or,
And
scribed right cylinders.*
given by the formuhi,
is
An
known
-which agrees (25.) with
(27.)
(21.)(22.), with the axes of the tioo circumsay u, of the umbilicar semidia-
finally the length,
LXXII.
all
661
= 2 _
52
c2
results.
umbilic of a surface of the second order
be otherwise defined
may
(comp. (23.)), as a real or imaginary point at which the tangent plane

a cyclic plane
bilicar
normal
&c.)
and
shown that
ters
is
(28.)
p'2
A less elementary
the
then
^2
+ ^2,
and
LXXIV.
may
if
LXXII.
verification of the value
For
assigned in 407, (8.).
1,
= p2 +
r'2
(t'2 4.
useful for other purposes,
ei, ^2
is
r and
<r,
common radius of the diamemean semiaxis b (comp. 216, (7.)
observe that the
by a slight extension of the analysis in (7.), (8.), (9.), it can be

p', cr', r' be any two systems of three conjugate semidiame-
that,
if p,
to
only necessary
sections of the ellipsoid
of any central surface, fy
LXXIII.
it is
in verification of the recent expressions (25.), (26.), for the lengths
and circular
(9.),
LXVl.
in
<p(xj
known property
of ou and su, it
tral
is parallel to
easy to prove (comp. 407, (20.)) that the umTo employ this
has the direction of a cyclic normal.
and accordingly
(Sp'ff'r')2
= (Sp(rr)2.
of 2j but one
be obtained from either the cubic in
which
or that in
b^,
ô^ b\^, b-^ be the roots of the former cubic, and
the roots of the latter, inspection of those equations shows at once that
is
e,
eo,
we
have generally,
LXXV.
-p^=5o2 + 5i2-}62--2^2S' = /2(^^+g^ + e2 + Saa');

OP 2 = Tp2 = ao^ + bi^ + ca^ = &o2 + ci'^ + 2^ = &c.,
LXXVI.
or
where the semiaxes

point p.
Making
o, 61, C3
LXXVII.
we
belong to the three confocals through any proposed
then,
.
o2
= a%
bi^
0,
= c^- 6^
c-z^
recover the expression assigned above, for the square of the length u of an
bilicar semidiameter of
an
um-
ellipsoid.
(29.) For any central surface, the principle (27.) shows that if X, p. be, as in
405, (5.), &c., the two real cyclic normals, and if 5- be the rea/ scaZar associated with
them as before, then the vectors of th.Q four real umbilics (if such exist) must admit of
being thus expressed
LXXVIII.
LXXIX.
and thus we
see anew, that
Compare 218,
84.
+ ^-IX VFX = ahc (ÛX + /iTX)

:
0- V
MFfi
=abc (gVp + XT/*)
an hyperboloid with one sheet has (as
is
well
known) no
and 220, (4.); in which the points
B, b' (comp. also

be conceived to coincide with the points R, k' of the
obvious that the theory of circumscribed cylinders is included
(5.),
may now
Fig. 53, page 226)
new Figure
It is
in that of circumscribed cones; so that the cylinder circumscribed to the confocal

(c),
with
its
generating lines parallel to a given (real or imaginary) semidiameter y of

= 1), may be represented (comp. III. XIV.) by the equation,
that surface (/y
III'.
/(p, y)2 =:/p
-1
or
XIV'.
FYyp = a'^b^c^
with interpretations easily deduced, from principles already established.
662
[boOK
real nmbilic, because for that surface the product abc of the seniiaxes
or because
it
has no real tangent plane parallel to either of
its
two
is
III.
imaginary
real planes of
circular section.
(30.) Of whatever species the surface may be, the three umhilicar vectors (23.),
of which only one at most can be real, with the particular signs there given, but
which have the /orms of lines in the three principal planes, must be conceived, in
virtue of their expressions
line,
of
LV. LVI. LVII.,
which the vector equation
Lxxx.
on an imaginary
to terminate
right
is,
+ 'Mz^.,
p=--îf:+l)_viKfli^^
Yaa'
aa
a-\-a
e'
Saa, + 1, and 1, when

And such an imaginary right
being a scalar variable, which receives the three values,
p comes to coincide with
io,
w^,
and
u),^
respectively.
line, which is easily proved to satisfy, for all values of the variable e\ both the rect'
angular and the bifocal forms of the equation of the surface (e), or to be (in an
imaginary sense) wholly contained upon that surface, may be called an Umhilicar
Generatrix.
(31.) There are in general eight such generatrices of any central surface of the
second order, whereof each connects three umbilics, in the three principal planes,
two passing through each of the twelve umhilicar points (23.) and because e'^ disappears from the square of the expression LXXX. for p, which square reduces itself
;
to the following,
LXXXI.
p2
= - Z2 (2e' + e + Saa') = -
62
_ 2Pe',
they may be said to be the eight generating lines through the four imaginary points,
in which the surface meets the circle at infinity.
(32.) In general, from the cubics in e and in b^, or from either of them, it may
be without difficulty inferred (comp. (28.)), that the eight intersections (renl or ima,-
ginary) of any three confocals (eo) (ei) (es) have their vectors p represented by the
formula
:
-^__^_^
LXXXII.
+aoaiff2
^^^-^
V-IM162
^,^^^,
C0C1C3
LXXX., we
comparing which with the vector expression
^^(^37)'
see that the three confo-
through the point determined by that former expression, for any given value of
are (c), (e'), and (c') again; and therefore that two of the three confocal surfaces
cals,
c',
through any point of an umhilicar generatrix (30.) coincide

in a
new way (comp.LXXV.) the expression LXXXI.

(33.) The locus of all such generatrices, for all the
a result which gives
for p^.
confocals (c) of the system,
a certain ruled smface, of which the doubly variable vector may be thus expressed,
as a function of the two scalar variables, e and e'
is
TVYVTTT
LXXXIII.
'
pe,e
^
= ^r^ l)Ke'+l) ~
+ '^~U( e+Saa')l{e'+Saa)
Yaa
a a
,
and because we have
thus, for
any one
LXXXIV.
l(e-l)He'-l)
'
set of signs, the differential relation,
D,p, = fD,,,p.,/,
CHAP.
DEVELOPABLE LOCUS AND ENVELOPE.
III.]
follows that this ruled locus
it
is
that wholly imaginary curve, of
(32.) the locus of
all
a Developable Surface
which the vector
is
p<.,<?,
663
its
edge of regression being

and which is therefore by
the imaginary points, through each of which pass three coinci-
dent confucals.
real part of this imaginary developable consists of the two real
(34.) The only
double tines upon it, as are also the imaginary focal, and the
focal curves, which are
circle at infinity (31.) ; and the scalar equation of the same imaginary surface, obtained by elimination of the
r
T
YXXV
V
'Zni'-aP
Hm^-aP
+
-
scalars e
and e\
is
found to be of the
22:m (m - n^x^y"^
+ 2 (p2 _ 6mn)a:V*
i+22(3m2 np)xY-^'^ + 22m^(n -p)x^ + 2'2m(mp - 3n^)z*yi

^
,
+2(m-n) ( - P) (P - ") ^V-- + 2m2 (m2 - 6np)x<
two arbitrary
namely the following
eighth degree,
+ 2 Smn (mn - Bp"^) x^y^ + 2 ^m^np (ji-n)x^ + m'hi^p^
in
which
Ave
have
LXXXVI.
.x =
- S|oU(a + a'), y = - SpU Vaa', z = - SpU(a - a'),

m = 62 - c^ n = 2_a2^ pâ-^-b^,
m + n+p = 0;
LXXXVIII.
LXXXVIL
and
50 that
-while each sign
written, for abridgment,
indicates a
sum
of three or of six terms, obtained
by
cyclical or
binary* interchanges.
(35.) From the manner in which the equation of this imaginary surface (33.) or
we easily see by (32.) that it has the double property
(34.) has been deduced,
:
Lst of being (comp. (20.)) the locus of the vertices of all the (real or imaginary)
and II. nd of
right cones, which can be circumscribed to the confocals of the system
;
being at the same time the common envelope of all those confocals which envelope
accordingly is known to be a developablef surface.
lines (31.) will come to be mentioned again, in con(36.) The eight imaginary
nexion with the Zi we* of curvature of a surface of the second order ; and before closing
:
the present series of subarticles,
it
may be remarked that the equation
in (15.), for the
determination of the second confocal (e^ which touches a given tangent, dp or pp', to
a given surface (e) of the same system, will soon appear under a new form, in con-
nexion with that theory of geodetic

we next proceed.
lines,
on surfaces of the second order, to which
When
xyz and abc are cyclically changed to yzx and bca, then mnp are
changed to npm but when, for instance, retaining x and a unchanged, we
make only binary interchanges of y, z, and of 6, c, we then change m., n, and p, to
- and - n
similarly
p,
j,
respectively.
t This theorem
cited Treatise
face
is
given, for instance, in page 157 of the several times already

also mentions the doiihle lines &c. upon the sur-
by Dr. Salmon, who
but the present writer does not yet
know whether
the theory above given, of
the locus (33.) of which imathe eight umbilicar generatrices, has been anticipated
ginary right lines TSO.) is here represented by the vector equation LXXXIII., from
:
which the scalar equation LXXXV. has been above deduced (34.), and ought to be
found to agree (notation excepted) with the kno\\Ti co-ordinate equation of the
developable envelope (35.) of a confocal system.
664
409.
general theory of geodetic
as treated
lines,
III.
[bOOK
by quater-
was given in the Fifth Section (III. iii. 5) of the present

Chapter ; and was illustrated by applications to several diflferent
We can only here spare room for applying the
families of surfaces.
nions,
same theory to the deduction, in a new way, of a few known but

principal properties of geodetics on ceiitral surfaces of the second order ; the differential equation employed being one of those formerly
used, namely (comp. 380, IV.),
I.
that
is, if
Yvdi^p
if
0,
II.
Td/j
const.
made the independent variable.
the arc of the geodetic be
(1.) In general, for any surface, of which r is a normal vector, so that the first
equation of the surface is Svdp = 0, the second differential equation
diflTerential
by
gives,
dSj'djO=
III.
(2.) Again, the sur face /jb
363, X'., 373,
IV.
and
surface, the expression,
= -v-'Sdj/dp.
d2jD
= const,
being
still
shall
have
quite ^feneraZ, if
we
write (comp.
III., &c.),
= 2Svdp = 2S<ppdp,
d/p
therefore,
on that
for a geodetic
I.,
by
we
V.
d/dp
= 2S(^dp
d2p)
III., for a geodetic,
+ 2S^=0.
VI...J^
SdjOd^p
0|O
(3.)
may
For a central surface of the second
order, <pp
is
a linear function, and we
write (comp. 361, IV.),
VII.
the general
diflferential
0dp
Sdpd^p = Sdp0dp =/dp
= d^p = dv,
equation VI. becomes therefore here,
+ 2S^=0,
VIII...^
V
/dp
and
gives,
by a
IX.
first
.
X.
or
where
and
integration, with the condition II.,
v^fdp
= Mp^,
P-sZ)-2
= h,
or
IX'.
X'.
or
= A = const.
= A"! = const.
Ti/2/Udp
P. i)
P = Tv~^ = perpendicular from centre on tangent plane,

D = (/Udp)* = semidiameter parallel to tangent
;
these two last quantities being treated as scalars, whereof the latter
imaginary,* together with the
* For the case of the

foregoing deduction,
P.D = const,
last scalar
ellipsoid, for
of Joachimstal's
(in substance) in
may
be
real or
h%
which the product P.
by quaternions,
was given
constant
D is necessarily real,
celebrated
page 580 of the Lectures.
the
first integral,
CHAP.
GEODETICS ON CENTRAL SURFACES.
III.]
(4.) Tiie following
a quite different
is
way
665
of accomplishing a first integration,
which conducts to another known result of not
less interest, although rather of a

Operating on the equation 407, XVI. by S.djO, and
graphic than of a metric kind.
remembering that
Spv=
XI.
I.
and
in which the
first
that
by
is,
and Svdp =
1,
.
0,
we
obtain the differential equation,
=;2(Sa'x^Sadp
S(0/Spd|0
SavSa'd/j);
II.,
XII.
Spdjo .
member,
XIII.
SpdV - p'SdpdV = Zd(Sadp

an exact
like the second, is

.
Sa'dp),
differential,
S(Vpdp.Vpd2p) = id(Vpdp)2
because
hence, for the geodetic,
XIV.
XV.
or
A'
being a
Z-2(Vpdp)2-2SadpSa'dp
2SaUdp Sa'Udp -
l-^
= /i'dp2,
(VpUdp)2 =
//,
scalar constant.
Comparing
(5.)
the
new
new constant
this last equation
h' is
the sum, e
with the formula 408, XXXVL, we find that

of what have been above called the parame-
e^,
* of the
given surface (e) on which the geodetic is traced, and of the confocal(e^
which touches a given tangent to that curve whence follows the knownf theorem,
that the tangents to a geodetic, on any central surface of the second order, all touch
ters
one
common confocal.^
The new constant eX^h'-e) may, by 407,
(6.)
(with for
Co),
be thu transformed
XVI.
e,
LXXV.
and 408,
LXXV.
= ei(TVUj/idp)2 + c2(TVUi'2dp)a
= ei(SUf 2dp)2 + 2(SUiîdp)2 = const.
the parameters of the two confocals through the point P of the geodetic on (e), and vi, v^, are as before the normals at that point, to those two surfaces
where
ci, ez are
(*l), (^2).
(7.) In fact, the
two equations
XVII.
last cited give the general transformation,

.
= e (V<tUv)2 +
<T
Z-2(Vp<T)
being an arbitrary vector, which
then this last expression to (e

obtain the first
and
~ 2Sa<7Sa'<r
ci(V(tUvi)2
+ eziYaVv^y
for instance
may
+ ej <T, or
be replaced by dp.
to e(V(TUi')*
e,T(T2,
therefore also the second transformation
Equating
since Sj/<t
= 0, we
XVI., because the three
normals vv\vz compose a rectangular system (comp. 407, (4.), &c.).

XVI. from the equa(8.) It is, however, simpler to deduce the second expression
and
tion 408, XXXI. of the cone of tangents from p to (ej, by changing r to Udp
;
then
if
we
write
XVIII.
.. i
dp
= Z,
V\
Compare the last Note to page 656.

t Discovered by M. Chasles.
a sphere, when the given confocal
X This touched confocal becomes
Compare 380, (6.), and 408, (16.), (17.) ; also the Note to page 617.
4 Q
is
666
'
[bOOK
so that vi denotes the angle at which the geodetic crosses the normal vi to
considered as a tangent to the given surface (e), the
first
integral
XVI.
III.
(i),
takes
the
form,*
XIX.
XX.
or
which the constant
a}
e^-ei sin2 vi + eo cos2 rj,

a-^ sin* i + a^^ cos^
&c.
n,
the primary semiaxis of the touched confocal (5.).

a^
(9.) "Without supposing that Tdp is constant, we may investigate as follows thi
differential of the real scalar h in IX. or X., or of the product P"2. D'^, for any curvt
in
is
on a central surface of the second order.
(L) and (2.), and resolving d^p

we get the general expi'ession,
XXI.
Leaving at first the surface arbitrary, as ii

and vdp
j/, dp,
in the three rectangular directions of
= - v-'Sdvdp + dp-iSdpd^p + (vdp)-iSvdpd2p
d2p
of which, under the conditions
Without assuming those
I.
and
conditions, if
the two last terms vanish, as in III.
II.,
we now
belong to a central surface of the second order,
introduce the relations VII. which
we have by V. and IX.
the expres-
sion, f
XXII.
= v2Sdvd2p + Si/dj/Sdvdp - ASdpd2p = Si/dvdp-i.Si/dpd2p,

dA = d.j/2Sd,/dp-i =d.P-2Z)-2=2Svdj/dp-iSi/dp-id2p;
^dA. dp2
XXIIL
or
XXIV.
or finally,
d^ dp* = 2Svdi;dp Si/dpdZp,

.
the scalar variable with respect to which the diflferentiations are performed being
hert
entirely arbitrary.
(10.) For a geodetic line on any surface, referred thus to any scalar variable
380, II. the differential equation,
we have by
XXV..
and
therefore
by XXIV.,
have again, as in
with
(1 1.)
= 0;
on a central surface of the second order,
as in
dA
= 0,
XXVI'.
or
= const,
X.
But we now
see,
by XXIV., that
for
such a surface the condition
not only by this differential equation of the second order

by this other diflferential equation,
is satisfied,
XXVII.
the product P-^D-^ (or
PD itself)
is
Si^dvdp
H'i
cos2
XXV.
XXVI
but
als(
therefore constant, not only as in (3.) for ever^
* Under this form

XX., the integral
Liouville,
cited in
w<
(3.),
XXVI.
h = p-2i)-2
Svdpd2p
for such a line
is
easily seen to coincide with that of
+ ^2 sin2 i = fi'i = const.,
page 290 of Dr. Salmon's Treatise.
t In deducing
this expression, it is to be
remembered that
dSdvdp = d/dp = 2Sdi;d2p

in fact, the linear
and self-conjugate form of v = ^p
gives,
Sdpd'v =/(dp, d2p) = Sdrd2p.
CHAP.
LINES OF CURVATURE.
III.]
667
geodetic on the surface, but also for every curve of another set* represented by this
last equation XXVII., which is only of the
Jirst order, and the geometrical meaning
of which we next propose to consider.
410. In general, if v and v + Av have the directions of the nora?i7/ surface^ at the extremities of the vectors p and p + A/?,
mals to
the cojidition of intersection (or parallelism) of these

I.
on an arbitrary surface,
is
The
called the Lines of Curvature,
St'di^d/>
to surfaces of the second order.
differential equation of the surface being, as usual,
III.
the normal vector
because the surface
is
is
Svdp =
0,
generally some function of
as yet arbitrary
p,
and dp, which is linear relatively to the

the dependence of dj^ on dp, we write
IV.
results
tion
4>
is
dv is therefore generally some
And if, attending
latter.
only
dv=0dp,
from what has been already proved (363), that this linear and vector funcat the
same time self-conjugate.

by r a tangentX pt
(2.) Denoting then
we
given extremity p of p,
lar equations, in
although not generally linear,
its differential
function of p
it
XXVII.),
= 0;
deduce a few general consequences, toge-
to
some that are peculiar
ther with
(1.)
now
= 0;
therefore (comp. 409,
II.
from which we are
SiÂj/A/>
what are
the differential equation] of
to
two normals
rigorously,
is,
which v
Namely, the
lines
by quaternions.
t In this equation
is
drawn
at the
t must satisfy the two following sca-
supposed to be given,
of curvature, as
II.,
to a line of curvature,
see that the vector
is
known, and as
dp and dv are two simultaneous
will presently be
differentials,
proved
which
(according to the theory of the present Chapter, and of the one preceding
it)
may
be at
pleasure regarded, either as two finite right lines, whereof dp is (rigorously) tangential to the
and to the line of curvature; or else as two infinitely small vecsurface,
tors,
dp being, on
this latter plan,
an infinitesimal chord Ap.
(Compare pages 99,
Notes to pages 623, 630.) The treatment of the equations is the same, in these two views, whereof one may appear clearer to some readers,
and the other view to others.
392, 497, 626,
and the
This symbol r
may not be
ples of this
is
first
used here partly for abridgment, and partly that the reader
the princis^
finite tangent, although
obliged to interpret dp as denoting
work allow him
so to interpret
it.
668
and
V. ..Svr=0,
VI.
Si/r^r
[bOOK
III.
tangent r admits therefore (355) of two real and rectangular directions, but not
in general of more : opposite directions being not here counted as distinct.
Hence, as
this
is
indeed well known, tlirough each point of any surface there pass generally two
:
and these two curves intersect each other at right angles.
lines
of curvature
(3.)
as follows.
direction:,
construction for the two rectangular directions of t can easily be assigned

Assuming, as we may, that the length of the tangent r varies with its
according to the law,
VII.
Sr0r=l,
which gives
VIII.
S (^r. dr) =
0,
by the properties above mentioned of ^

stant in v., so that we may write,
IX.
Si/dr
= 0,
or briefly
VIII'.
S^rdr =
and remembering that v
X.
and therefore (by VI.),
is
0,
treated as a con-
Srdr =
we see that, under the conditions of the question, the above mentioned length Tr, of
this tangential vector r, is a maximum or minimum : and therefore that the two
directions sought are those of the two axes of the plane conic V. VII., which has
p of the
surface,
(4.) This plane conic V. VII.

face at the given point p ; in fact
may be
centre at the given point
it is
and
is
in the tangent
Index
called the
easily
mere dimensions, with the known indicatrix
plane at that
Curve., for thi3
its
point.
given sur-
we
abstract from
(la courbe indicatrice) of
Dupin,* who
proved to coincide,
if
pointed out the coincidence (3.) of the directions of its axes, with those of the
and also established a more general relation of conjugation belines of curvature
tween two tangents to a surface at one point, which exists when they have the direcfirst
tions of
any two conjugate semidiameters
of that curve
so that the lines of curvature
are distinguished by this characteristic property, that the tangent to each
is
pendicular to its conjugate.

(5.) In our notations, this relation of conjugation between two tangents
which satisfy as such the equations,
V.
is
Sj^r=
0,
and
V.
or
XI'.
Svt'
per-
r, r',
= 0,
expressed by the formula,
XI.
we have
Sr^r' =
0,
Sr'^r
r'
therefore the parallelisms,!
XII.
so that the equation VI.
may
11
XII'.
Yv(}>t',
II
Vv0r
be written imder the very simple form,
XIII..
which gives at once the rectangularity
Srr'=0,
lately mentioned.
Developpements de Geometrie (Paris, 1813), pages 48, 145, &c.

t The conjugate character of these two parallelisms, or the relation,
V. v(pYv(pr
may
easily be deduced
formula 348, VII.,
in
II
T,
if
Si/r
0,
from the self-conjugate property

page 440.
of
(p,
with the help
of the
CHAP.
INDEX CURVE AND SURFACE, CONJ. TANGENTS. 669
III.]
The
(6.)
parallelism XII'.
may
be otherwise expressed by saying (comp. (4.))
that
XIV.
XV.
and
dp
have the directions of conjugate tangents
Yvdv
or that the
two
vectors,
Ap and YvAv,
when TAp diminishes
.
But whatever
have ultimately such directions,
indefinitely.
may be this length of the chord Ap, the vector YvAv has the direction of the line
of intersection of the two tangetit planes to the surface, drawn at its two extremities
another theorem of Dupin*
able be circumscribed to
is
therefore reproduced, namely, that if a develop-
any proposed curve thereon, the generating lines of this developable are everywhere conjugate, as tangents to the surface, to
the corresponding tangents to the curve, with the recent detinition (4.) of such con-
any
surface, along
jugation.
(7.) The following is a very simple mode of proving by quaternions, that t/a
tangent r satisfies the equation VI., then the rectangular tangent,
XVI.
satisfies
the
same equation.
conjugate property of
gives,
XVII.
(8.)
t'=vt,
we have only to
by VI. and XVL,
For
this
purpose
observe, that the self-
= Sr>r = Sr^r'=v-2Si^r>r'.
Another way of exhibiting, by quaternions, the mutual rectangularity of
the lines of curvature,
is
by employing (comp. 357,
I.) the self-conjugate /ort,
XVIII. ..>r=^r+VXr/i;
which the vectors
p, and the
scalar ^r, depend only on the surface and the point,

X,
and are independent of the direction of the tangent. The equation VI. then becomes by v.,
in
XIX.
= Svr\r/i = SvrXS)ur + Svr/iS\r;
..
assuming then the expression,
XX.
we
T=^xYv\ + yYvp,
easily find that
XXI.
y\Yvpy- =
x^
(YvXy,
or
XXI'.
the two directions of r are therefore those of the
XXII.
..
two
yTVv/x = ^TVvX
lines,
UVj.X+UVr/1,
which are evidently perpendicularf to each other.
Dupin proved^rsi (Der. de Geometric, pp. 43, 44, S:c.), that two such tangents
have a relation of reciprocity to each other, on which
as are described in the text
account he called them " tangentes conjuguees :" and afterwards he gave a sort of
image, or construction, of this relation and of others connected with it, by means of
the curve which he named " Vindicatrice" (in his already cited page 48, &c.).
t This mode, however, of determining generally the directions of the lines of

an illusor}' result, when the normal v has the direction of
curvature, gives only

either
X or p, which happens at an timbilic of the surface. Compare 408, (27.), (29.),

and the
670
(9.)
An
interpretation^ of
some
interest,
may
[bOOK
III.
be given to this last expression
XXII., by the introduction of a certain auxiliary surface of the second order which
may be called the Index Surface^ because the index curve (4.) is the diametral sec^
With the retion of this new surface, made by the tangent plane to the given one.
cent signification of 0, this index surface is represented by the equation VII., if r
be now supposed (comp. (2.)) to represent a line pt drawn in any direction from
the given point
Or
p,
and therefore not now obliged
to satisfy the condition
V. of tan-
we
denote by p + p' the vector from the origin

to a point of the index surface, the equation to be satisfied is, by the form XVIII.
of
^ (comp. 357,
gency.
for greater clearness,
if,
II.),
XXIII.
l=Sp'<pp'=gp'^ + SXp'np'',
the centre of this auxiliary surface being thus at p, and its two (real) cyclic normals
being the lines X and fjt so that Vj^X and Yvfx have the directions of the traces of
:
two cyclic planes, on that diametral plane (Sv|o'= 0) which touches the given
We have therefore, by XXII., this general theorem, that the bisectors of
surface.
its
the angle formed by these two traces are the tangents to the two lines of curvature,
whatever the form of the given surface may be.
(10.) Supposing now that the given surface is itself one of the second order, and
that its centre is at the origin o, so that it may be represented (comp. 405, XII,)
by the equation,
XXIV.
XVIII.
of ^,
= Sp(pp=gp^^-S\pfip,
which
will reproduce with those values the form

X, p.,
g,
that the index surface (9.) becomes in this case simply that
centre transported from o to p
and therefore with a tangent
we see
given one, with
plane at the
and
with constant values of
its
origin,
which
is
parallel to the given tangent plane.
And
thus the
traces (9.), of the cyclic planes on the diametral plane of the index surface, become
recover then,
here the tangents to the circular sections of the given surface.
as a case of the general theorem in (9.), this known but less general theorem : that
We
any point of a surface of the seof curvature, which pass through the same
the angles formed by the two circular sections, at
cond order, are bisected by the
lines
point.
(11.)
And
because the tangents to these latter lines coincide generally, by (3.)
(4.) (9.), with the axes of the diametral section of the index surface,
made by
the
tangent plane to the gioen surface, they axe parallel, in the case (10.), as indeed is
well known, to the axes of the parallel section of a given surface of the second
order.
(12.)
And
if
we now look back
to the earlier formulae of 407, (4.),
to the
we
Equation of Confocals
in 407, (26.),
and
shall see that because the vector v\, in the
last cited sub-article, represents a tangent to the given surface
with the normal v and the derived vector
cpvi,
so that
= 1, complanar*
(comp. 407, XII.
Sp^p
it satisfies
XIV., and the recent formulae V. VI.) the two scalar equations,
XXV.
which are likewise
^vvi
satisfied
= 0,
and
(comp. (7.))
XXVI.
Svvi^vi
when we change
Compare the Note
to
= 0,
v\ to the rectangular tan-
page 645.
CHAP.
III.]
follows that these two vectors, vi
it
gent V2
LINES OF CURVATURE ON CENTRAL SURFACES. 671

and
vj,
which are the normals
to the
two confocals to (e) through p, are also the tangents to the two lines of curvature on
th&t given surface of the second order at that point
whence follows this other theo:
rem*
of Dupin, that the curve
surfaces, is
line
of orthogonal intersection
of curvature on each.
(-407, (4.)),
of two confocal
(13.) And by combining this known theorem, with what was lately shown respecting the umhilicar generatrices (in 408, (30.), (32.), comp. also (35.), (36.)),
we may see that while, on the one hand, the lines
curvature on
central surface
of the second order
have
sense, they
for their
common envelopef
which connect the twelve
lines (4:08, (31.)),

face, three
of
a,
have no real envelope, yet on the other hand, in an imaginary
the system of the eight imaginary right

imaginary) umbilics of the sur-
(real or
by three, and are at once generating lines of the surface itself and also of
known
developable envelope of the confocal system.

may be added, as another curious property of these eight imaginary
right lines, that each is, in an imaginary sense, itself a line of curvature upon the
the
(14.) It
surface: or rather, each represents two coincident lines of that kind.

In fact, if
denote the variable vector 408, LXXX. of such a generatrix by the expression,
XXVII. ..p =
in
which
vectors,
e' is
a variable scalar, but c,
e'<T+(r',
two given or constant but imaginary
are
<t'
we
such that
XXVIII.
XXIX.
and
./(7
= 0,
v,
= - Z2, r'2 = - 53,

= S(7>(r = 0, /(t' = 1,
(t')
S(T(t'
= S(r0 (7=0,
we have the imaginary normal
XXX.
(t2
/(tr,
with (for the case of a real umbilic) a real tensor^
= e'dxr + 0(t'
o.
(T,
XXXI.
Tv = + (g-O^^
abc
De'v. de Ge'ome'trle, page 271, &c.

t The writer is not aware that this theorem,
to which he was conducted by quahas evidently an intimate connexion

with a result of Professor Michael Roberts, cited in page 290 of Dr. Salmon's Treatise, respecting the imaginary geodetic tangents to a line of curvature, drawn from an
ternions, has been enunciated before
but
it
umbilicar point, which are analogous to the imaginary tangents to a plane conic,
drawn from
a.
focus of that curve.
An
illustration,
sentation, of the theorem (13.) is supplied
by Plate
which
is
almost a visible repre-
II. to Liouville's
Monge (and
by-
the corresponding plate in an earlier edition), in which the prolonged and dotted
parts of certain ellipses, answering to the real projections of imaginary portions of
the li7ies of curvature of the ellipsoid, are seen to touch a system of four real right
lines, namely the projections (on the same plane of the greatest and least axes), of
the four real umbilicar tangent planes, and therefore also of what have been above
umbilicar generatrices of the surface.
called (408, (30.), (31.)) the eight
(imaginary)
observes (page 150 of Liouville's edition), that "toutes les

dans ce parallelogramme
ellipses, projections des lignes de courbure, seront inscrites
dont chacune d'elles touchera les quatre cotes :" with a similar remark in his expla-
Accordingly
Monge
nation of the corresponding Figure (page 160).
672
and we
find, after reductions, the
XXXII.
.. v(t
= + V-1*tTv,
[l500K
III.
imaginary expression,
XXXIII.
whence
..
Si/OP
= 0,
Sv(T(p(r
= 0.
The
differential equations V. VI. of a line of curvature are therefore symholicnllt/

satisfied^ when we substitute, for the tangential vector r, either the imaginary line
(T
itself,
tor
va
or the apparently perpendicular but in an imaginary sense coincirffni* vec-
and the recent
assertions are justified.
As regards the real lines of curvature, on a central surface of the second

order, we see by comparing the general differential equation II. -with the expression 409, XXIII. for the differential of h, or of PW-^, that this latter product, or
(15.)
P.D itself, is constantf for a line of curvature, as well as for a geoon such a surface, as indeed it is well known to be although this last
constant (^P.D) may become imaginary, for the case of a single- sheetedX hyperbothe product
detic line,
loid,
and must be such
a line of curvature on an hyperboloid of two sheets.
for
(16.) And as regards the general theory of the index surface (9.), it is to be obin the
served that this auxiliary surface depends primanVy on the scalar function
= 1, or generally /p = const., of the given surface ; and that it is not enequation /p

For if we write, for instance,
tirely determined by means of that surface alone.
XXXIV.
we
shall have, as the
new
of III.,
XXXV.
and
if
we then
XXXVII.
df/p
with
d/p
= 2S'dp
as before,
equation of the same given surface, instead
= 2Sni/dp,
by analogy
write,
.
= fl,
first differential
f/p
XXXVI.
with
= i'fp;
to IV.,
nv = (idp = n^djo + nVSrdp,
with
XXXVIII.
the new index surface, constructed on the plan (9.), will have for
analogous to XXIII. the following

,
XXXIX.
n'
its
= 2f "/p,
equation,
^p6p = nSp'ipp' +
n'
(Svp'y
= const.
* As
regards the paradox, of the imaginary vector a being thus apparently perpendicular to itself, a similar one had occurred before, in the investigation 353, (17.),
(18.), (19.)
and
it is
explained, on the principles of
ing that this imaginary vector is directed

to page 459.
(31.), and the Note
modern geometry, by observ-
to the circle at infinity.
Compare 408,

J Although the writer has been content to employ, in the present work, some of
these usual but rather long appellations, he feels the elegance of Dupin's phraseology,
t Compare the
adopted also by Mbbius, and by some other authors, according to which the two central hyperboloids are distinguished, as elliptic (for the case of two sheets), and hy"
The phrase quadric," for the general surface of the

second order (or second degree^, employed by Dr. Salmon and Mr. Cayley, is also
very convenient. It may be here remarked, that Dupin was perfectly aware of, or
perbolic (for the case of owe).
rather appears to have
first
discovered, the existence of what have since his time come
; which important curves were considered by him, as

de
being at once limits of confocal surfaces, and also loci ofumbilics. Comp. Dev.
to be called the /ocaZ conies
Geometric, pages 270, 277, 278, 279

by M. Chasles (Brussels, 1837).
&c.,
see also
page 390 of the Apergu Historique,
CHAP.
But
(17.)
if
will
XXXIX.,
673
FORMS OF DIFFERENTIAL EQUATION.
III.]
we take
= n,
this last constant
have a common diametral
the two index surfaces^ XXIII. and

made by the given tangent plane,
section,
namely the index curve (4.) and they will touch each other, in the whole extent of
And it will be found that the construction (9.), for the directions of the
;
that curve.
of curvature, applies equally well
lines
liary surfaces:
Svdvdp = 0,
if
in fact,
is
it
to the
one as to the other, of these two auxi-
evident that the differential equation II.,
receives no real alteration,
when v
is
namely
multiplied by any scalar, n, even
that scalar should be variable.
(18.) And instead of supposing that the variable vector p

373, to satisfy a given scalar eqriation, of the form*
is
thus obliged, as in
fp = const.,
= ix +jy + kz, and v =fp = F (x, y, z), and if we write,

dw = pdx + qdy + r6z,
dp =p'dx + r"dy + q"dz,
dq = q'dy + p"dz + r"dx, dr = r'dz + q"dx \-p"dy,
we may then write also, on the present plan, which gives dfp = 2SvdjO,
v = -l(ip +jq + kr),
dp = idx +jdy + kdz,
dv = -^ (idp +jdq + kdr), Sdpdj/-= i (dxdp + dydq + dzdr)
and theindex surface, constructed as in (9.), and with p' changed to Ap = iAx +jAy
* If
|0
+ hAz,
will thus
(a).
have the equation,
ip'Aa;2
^ ig'Ay2 + ^r'Az'^ +p"AyAz +
more generally = const.
or
so that
it
may
be
made in
AzAx 4- r"Aa;Ay =
this
way
1,
depend upon, and

of the second order^
to
be entirely determined by, the six partial differential coefficients

,p" ., of the function v ovfp, taken with respect to the three rectangular coAnd by comparing this equation (a) with the following equation
ordinates, xyz.
p'
same auxiliary surface, which results more directly from the principles employed in the text (comp. XVIII. XXIII.),
of the
(b).
we can
SApÂp =gAp^ + SXAp/iAp =
for those six partial

easily deduce expressions
1,
coefficients, in
terms of ^, X,
fi.
Thus, for example,
= SXfi -g + 2SaSj>
SXijxi
+ SjXSjn + SkXSkp, = - SXp. therefore,
- 3g = ci + C2 + C3 = - m",
=
^ (Px^v + T>y^v + D^sr) SXfi
|Da;2
but
= ip' = -g+
SiXSi/i
(c).
if ci, 02, C3
be the roots and
m"
a coefficient of a certain cubic (854, III.), deduced

= ^dp, on a plan already explained. If
from the linear and vector function dv
then the function v satisfy, as in several physical questions, the partial differential
equation,
(d).
the
tic
is
sum
of these three roots,
ci, C2, cz,
+ BzH = 0,
will vanish
and consequently, the asympto-
in the second
cone to the index-surface, found by changing 1 to
the property that
real, and has (comp. 406, XXL, XXIX.)
(e).
if
Dx2 + DyS
a,
b denote
its
cot2a
two extreme semiangles.
+ cot2b = l,
An entirely
4 R
different
member
method
of (a),
of trans-
674
we may
bles,
[bOOK
III.
suppose, as in 372, that p is a. given vector function of two scalar variay, between which there will then arise, by the same fundamental formula
X and
a differential equation of the first order and second degree, to be integrated

For example, if we write,
possible) by known methods.
II.,
(when
XL.
we
= ix +ji/ + kz,
XLI.
= ip +jq - k,
and thus the general equation
dz
=pdx + qdy,
by assuming (with a constant
shall satisfy the equation III.
XLII.
whence
II., for
factor understood),
dv = idp +jdq
the lines of curvature on an arbitrary surface,
receives (by the laws of ijk) the form,
XLIII.
dp(dy + qdz) = dq(dx+pdz')',
form has accordingly been assigned, and in several important questions

but which is now seen to be included in the still more conemployed, by Monge*
cise (and more easily deduced and interpreted) quaternion equation^
which
last
Svdvdp
= 0.
411. For a central surface of the second order, we have as usual

Ai/ = 0A/!>, and therefore
(by 347, 348, and by the self-conjugate form of 0),
V
(j)p,
I.
Yv/^v = Y(/>p^^p = yJrVpAp = m0"^V/>A/9 ;
the general condition of intersection 410, I. of two normals, at the

extremities of & finite chord A/j, and the general differential equation
410, II. of the lines of curvature, may therefore for such a surface
receive these new and special forms :
forming, by quaternions, the well
writer,
may
and
will
be remarked, that because
m' =
have then, by the general theory

cially
subarticles to 350,
by the
known
equation (d), occurred early to the present

In the mean time it
be briefly mentioned somewhat farther on.
mulae,
(f).
dv
III.
by
ii.
when the equation (d)
(c),
6 of linear
remembering that
+ xdp = 0,
and
(g).
and vector
<p
is
^<r
is satisfied,
functions,
and
here self-conjugate, the
^ff
we
espefor-
= mV,
being auxiliary functions, and m' another coefficient of the cubic, while <r is an
For the same reason, and under the same condition
arbitrary vector.
(d), the
X,
function ^ itself has the properties expressed

(h).
in
which the
0VtK
= K0i -
ttvo vectors
i,
i0K,
and
by
(i).
the equations,
.
(jt'^YiK
k are arbitrary, and m'
is
= Yî<pK - m'Vt/c
the same scalar coefficient as
before.
* See the enunciation of the formula here numbered as

XLIII., in page 133 of
Monge compare also the applications of it, in pages 274, 303, 305, 357.
Liouville's
(The corresponding pages of the Fourth Edition are, 115, 240, 265, 267, 312.)
The quaternion equation, ^vdvdp = 0, was published by the present writer, in a
communication to the Philosophical Magazine, for the month of October, 1847
See also the Supplement to the same Volume xxxi. (Third Series);
(page 289).
and the Proceedings of the Royal
Irish
Academy
for July,
1846.
CHAP.
CONDITION OF INTERSECTION OF NORMALS.
III.]
II.
= 0,
Sdpcp-'Ypdip = 0,
SA/>0-'V/>A/)
III.
or
II'.
or
III'.
S/)A/90-iA^
.
675
= O;
S/>d/>0-'d/>
which admit of geometrical interpretations, and conduct to some

new theorems, especially when they are transformed as follows:
X and
fji
IV.
V.
S/) A/30->
S\d/j.S/)d/>0->
still
+ S/nAp
S/) A^0-i\,
0,
S/td/).S/)d/)0-'X,=O,
more simple forms,

VI.
VIL
a,
SX A/)
being (as in 405, (5.), &c.) the two real cyclic normals of
while the same equations may also be written under
the surface
the
..
SaAp.Sa'pAp + Sa'Ap SapAp = 0,

Sad/>.SaVd/> + Sa'd/).Sa/r>d/) = 0,
.
a being, as in several recent investigations, the two real focal

which are common to a whole confocal system.
unit lines,
(1.)
The vector (p'^YpAp
in II. has
by
I.
the direction of VvAj/; whence,
by
410, (6.), the interpretation of the recent equation II., or (for the present purpose)
of the more general equation 410, L, is that the chord pp' is perpendicular to its
oum polar, if the normals at
its
extremities intersect.
Accordingly,
intersection be called n, the polar of pp' is perpendicular at once to
therefore to pp'
if their
pn and
point of
p'n,
and
itself.
The equation
II'. may be interpreted as

expressing, that when the normals
thus intersect in a point n, there exists a point f'' in the diametral plane
a result which may bo
OPP', at which the normal p''n" is parallel to the chord pp'
(2.)
at
P and
p'
otherwise deduced, from elementary principles of the geometry of surfaces of the
second order.
on the converse propositions, that when either of

intersection (or parallelism) of the two normals
(3.) It is unnecessary to dwell
these conditions
at
p and
III.
p'
is satisfied,
there
is
or on the corresponding but limiting results, expressed
by the equations
and II r.
(4.) In order, however, to
we must
select
some
make any use in calculation
of these
new forms II.,
suitable expression for the self-conjugate function
a corresponding expression for the inverse function
VIII.
which has already several times
<pp=pp +
occiu-red,
<p'^.
<p,
III.,
and deduce
The form*
Y\pfi,
has also been more than once inverted
but the following new inverse f form ,
* The vector
form VIII. occurred, for instance, in pages 463, 469, 474, 484,
and the connected scalar form,
641, 669
fp
= ffp^
-i-
357,11.
SXpfip,
has likewise been frequently employed.
t Inverse forms,
for ^"ip or m"^pp,
have occurred
in pages 463, 484,
641 (the
676
IX..
[bOOK
(^-SXju).^-V = |0-XSp^-V-)uSp<^-iX,
In
has an advantage, for our present purpose, over those assigned before.
form IX. gives at once the equation,
X.
(^-SX/i).^-^VpAjO=VpAp-\SpA|O0-i/i
and so conducts immediately from
II, to IV., or
The equation IV. expresses ^'eneraZ/y,
(5.)
III.
a certain cone of the second
order,
which has
from
-/iS/[)A|O0-\;
V. as a
III. to
limit.
that the chord Ap, or pp',
its
fact, this
is
a side of
vertex at the point p of the given
and passes through all the points p' for which the normals to that surface innormal at p and the equation V. expresses generally, that the two
surface,
tersect the given
aides of this last cone, in
which
it is
cut
by the given tangent plane at the same
point
of curvature.
(6.) But if the surface be an ellipsoid, or a <fo5/c-sheeted hyperboloid, then
(comp. 408, (29.)) the always real vectors* (p'^X and ^~^p, have the directions of
P, are the tangents to the lines
semidiameters drawn to two of the four real umhilics ; supposing then that p is such
a semidiameter, and that it has the direction of <p'^\ the second term of the first
of the equation IV. vanishes, and the cone IV. breaks
member
up
into a pair of
planes, of which the equations in p' are,
XI.
SX (p' - p) =
0,
and
XII.
Sp'<}>-^\6-^p
whereof the /ormer represents the tangent plane at the umbilic

presents the plane
of the four real
p,
and the
is
not intersected by
latter re-
umhilics.
normal at
(7.) It follows, then, that the
any real normal
the real umbilic
to the surface, except those which are
drawn
at points p' of that

principal section, on which all the real umhilics are situated: but that the same real
umbilicar normal
pn
is,
in
an imaginary
sense, intersected
by
all the
imaginary nor-
mals, which are draicnfrom the imaginary points p' of either of the two imaginary
generatrices through p.
p', under the condition of intersection of its
with a given normal pn, is generally a quartic curve, namely the intersection of the given surface with the cone IV.
but when this cone breaks up, as in
(8.) In fact, the locus of the point
normal
p'n'
(6.), into
the general quartic
is
namely the principal

ttco
normal, and the other tangential to the surface,

likewise decomposed, and becomes a system of a real conic,
two planes, whereof one
is
section (7.),
and a pair of imaginary right
lines,
namely
the
umbilicar generatrices at P.
We
(9.)
generatrix
normals
see, at the
is (in
same
time, in a
an imaginary sense) a
new way (comp.

line
of curvature
410, (14.)), that each such

: because the
(imaginary)
to the surface, at all the points of that generatrix,
are situated
by
(7.) in
one common (.imaginary) normal plane.

(10.)
Hence through a
correction in a
Note
with the form IX.,
to
it
real umbilic, on a surface of the second order, there pass
which
last
page should be attended
will easily be seen
ffX-Xy
*
to)
(comp. page 661) that
Compare the Note immediately preceding.
gjx-j^
In comparing these
CHAP.
THREE LINES THROUGH AN UMBILIC.
III.]
three lines
of curvature
imaginary right
whereof one
a real conic
is
(8.),
677
and the
ttvo others
are
namely, the umhilicar generatrices as before.
lines,
(11.) If we prefer differentials to differences, and therefore use the equation V.

we find that this equation takes the form 0=0, if the
of the lines of curvature,
point
and that if the normal at that point be parallel to

of the equation V. breaks up into two factors, namely,
P be an umbilic
differential
XIII.
SXd2p
XIV.
and
0,
Sdp^-'X^-i/i
X, the
whereof the former gives two imaginary directions, and the latter gives one real direction, coinciding precisely with the three directions (10.).
(12.)
And
if p,
instead of being the vector of an umbilic, be only the vector of a
point on a generatrix corresponding,
we
shall
still
satisfy the differential equation
belongs to the same imaginary right line

shall then have, as at the umbilic itself,
by supposing that
v.,
djo
XV.
SXilp
Spdp0->X
0,
because
we
= 0.
An
umhilicar generatrix is therefore proved anew

(comp. (9.)) to be, in
extent, a line of curvature.
its ivhole
(13.) The recent reasonings and calculations apply (6.), not only to an ellipsoid,
but also to a double-sheeted hyperboloid, four umbilics for each of these two surfaces being real.
But if for a moment we now consider specially the case of an ellipsoid,
and
if
we denote
substitute in IV.
abridgment the real quotient
for
and V.
for X,
p.,
<p-%
,
ha ^ a
a\ha
:=^
and then, after division by 7*2-
1,
by
h,
we may then
2b\jp
:
a+c
-ha-a ^
ac{a\c)
there remain only the
the equation IV. reducing itself to VI.,
\-
<p-^p the expressions,
26UX
XVI. ..a fta=
a\ c
XVII.
and V.
j^
ac{a-\-c)
two vector constants a
a',
to VII.
The simplified equations thus obtained are not however peculiar to ellipbut extend to a whole confocal system. To prove this, we have only to combine the equations II. and III. with the inverse form,
(14.)
soids,
XVIII. . m-'p = aSa'p + a'Sap p (e + Saa'),

which follows from 407, XV., and gives at once the equations VI. and VII., what.
ever the species of the surface
The
(15.)
may
be.
VII. must then be
satisfied
by the three rectan-
gular directions of dp, or of a tangent to a line of curvature, which answer to the

orthogonal intersections (410, (12.)) of the three confocals through a given point P ;
ought therefore, as a verification, to be satisfied also, when we substitute v for dp,

V being a normal to a
that is, we ought to have the
confocal through that point
it
equation,
XIX.
SavSapv+Sa'vSapv =
0.
And
accordingly this is at once obtained from 407, XVI., by operating with S.pv ;
so that the three normals v are all sides
of this cone XIX., or of the cone VII. with
dp for a
side,
with which the cone V.
And
is
found to coincide (13.).
because this last equation XIX., like VI. and VII., involves on/y the
"
two foeal lines
Ifindea, a' as its constants, we may infer from it this theorem :
(16.)
678
[bOOK
III.
many surfaces of the second order have only their asymptotic cones hiconfoand pass through a given pointy their normals at that point have a cone of the
finitely
caly*
second order for their locus ;" which latter cone is also the locus of the tangents, at
the same point, to all the lines of curvature which pass through it, when different
values are successively assigned to the scalar constant fl2_
gs (or 2/2)
that
is,
when
the asymptotes a, a' to the focal hyperbola remain unchanged in position, but the
semiaxes (a2 _ ^2)1^ (^52 _ f2)J of that curve (here treated as both real) vary together.
The equation VI. of the cone of chords (5.) introduces the fixed focal
by their directions only. But if we suppose that the lengths of those
(17.)
a'
lines ,
two
is
lines are equal,
we
unity,
tem of unit
without being here obliged to assume that each of those lengths

have (comp. 407, (2.), (3.)), the following rectangular sys-
shall then
lines, in
the directions of the axes of the system,
XX.
which obey in
U (a + a),
all respects the
U (a - a'),
UVaa',
laws ofijk, and
often be conveniently c?no/ed by
may
those symbols, in investigations such as the present.
semidiameter
p,
and the chord Ap,
And
by decomposing
XX., we easily find
then,
in these three directions
the
the
following rectangular transformation^ of the foregoing equation VI.,
XXI
in which
^(" + ")"'P
*
'
S(a + a')A|0
permitted to change
it is
differential equation of the lines

find, in
a new way, a condition
S(a
Ap
a')A/)
to dp, in order to obtain
of curvature ; or
satisfied
S.Uaa'AjO
a new form of the
else at pleasure to v,
by the three normals,
and
so to
to the three confocaU
p.
through
The
(18.)
gular lines
cutting
S(-0'^P ^ S.(Vaa ')-ip
it
cone, VI. or XXI., is generally the locus of a system of three rectaneach plane through the vertex, which is perpendicular to any real side,
in a real pair of mutually rectangular sides
the section of the same cone,
same
while, for the
its
reason,
vertex
p,
generally an equilateral hyperbola.

however, the point p be situated in any one of the three principal
but cuts any side perpendicularly,

(19.)
by any plane which does not pass through
If,
is
planes, perpendicular to the three lines XX., then the cone XXI. (as its equation
shows) breaks up (comp. (6.)) into a pair of planes, of which one is that principal
* That
is, if the surfaces (supposed to have a common centre) be cut by the
plane at infinity in biconfocal conies, real or imaginary.
t The corresponding /orm, in rectangular co-ordinates, of the condition of intersection, of
normals at two points (xyz) and (x'y V), to the surface,
is
the equation (probably a
meet with
it),
(fc2
_ c'-^X
x'
x
\\\
which
it is
known
22
y2
Ofl
one, although the writer has not
(c2
ff2)
y'
y - y
evident that xyz and x'y'z'
may
(ff.2
_ 52) z'
_
z
be interchanged.
happened
CHAP.
plane
CENTRES OF CURVATURE.
III.]
itself,
while the other
is
mer
which
is
plane usually cuts the surface in
section at right angles,
But
And while i\xQ former plane

always a line of curvature through
another conic, which crosses the for-
perpendicular thereto.
cuts the surface in a principal section,

p, the latter
679
and gives the direction of the second
line of curvature.
we
further suppose, as in (6.), that the point p is an

wnbilic, then
the second plane is a tangent plane; and the second conic
(as has been seen)
(19.)
is itself decomposed, into a pair of imaginary right lines : namely, as before, the two
(20.)
if
umbilicar generatrices through the point, which have been shown to be, in an imaginary sense, both lines of curvature themselves, and also a portion of the envelope
of all the
others.
We
shall only here add, as another transformation of the general equation

VI. of the cone of chords, which does not even assume Ta = Ta, the following
(2 1.)
XXII.
S(a + a')Ap.S(a+a')|oA|0=S(a-a')A/o.S(a-a')pAp;
where the directions of the two new
a + a and a a,
lines,
are only obliged to be
harmonically conjugate with respect to the directions of thejixed focal
lines of the
any two conjugate semidiameters
of the focal
system
or in other words, are those of
hyperbola.
412.
The
subject of Lines of Curvature receives of course an

when it is combined with the known concep-
additional illustration,
Without yet en-
of Curvature.
tion of the corresponding Centres
tering on the general theory of the curvatures of sections of an arbitrary surface, we may at least consider here the curvatures of those
normal
which
sections^
touch at
any given point the
Denoting then by o- the vector of the centre

such a section, and by R the radius ps, considered
ture.
is
positive
when
has the direction of +
it
we have the two fundamental

a = pÊ\Jv',
I.
.
II.
of curva-
as a scalar
which
easy to see that
+ dUi' = 0;
new form of the general
this
lines
of curvature of
II. .. i2-'d/>
whence follows
410,
equations
it is
v,
of the lines of curvature,

III.
Vd/)dtr,.
= 0;
with several other combinations or transformations, among which

may be noticed here
the following
(1.)
The equation
by merely
I.
requires no proof;
differentiating* as
follows at once,
and IV.
is
if
ff
and
and from
it
R were constant
the equation II.

after
is
obtained
which the formula
III.
easily deduced.
To students who are accustomed
to infinitesimals, the easiest
way
is
here to
680
To obtain from
(2.)
assume
may
this last equation
V.
XXII.) one
a more developed expression for R^ we
form (comp. 410, XVIII.),
(1.)), the general
dv = gdp + Y\6pfx,
and then, while the tangent dp has (by
are independent of dp
ft
III.
considered as a linear and self-conjugate function of dp (410,
for dv,
in which ^Tj X,
[bOOK
410-,
or other of the two directions,
VL. .dp\\JJYv\VYvfiy
i
^
"'
the curvature R'^ receives one or other of the two values corresponding,
VII.
7^l
One mode
(3.)
= - Tv-' (g + SXUj/. SfiVv TVXUv TVÛv).

.
of arriving at this last transformation, or of
showing that
if
(comp. again 410, XXII.) we assume,
r= (or UVXv UV/iv,

= SXUv S/*Uv TWVv.TYfiVv,
.^SXr.S/ir-i =S(VXUv.V/iUv) + TVXUv.TV/iUv,
2SUXr. SU/ir'^ =S(VUXv.VU/iv) TVUXv.TVU;tv,
VIII.
IX.
then
or
X.
or finally,
XI.
II)
SXr/xr-i
introduce the auxiliary quaternion,
is to
..q = YVXv.YXJiMV
XII.
and
to prove that, with the value (or

direction) VIII. of r,
tion (in which Vg^^ as usual, represents the square of
Yq),
XIII.
(4.)
And
this
may
= Sa^2SUXr SUur-
+ To^ =
^
^^f
we have thus
the equa-
Sg + Ty
be done, by simply observing that
we have
thus (with the
value VIII.) the expressions,
XTV
XV
XV.
- ^^^^^
SrTTX
SrUX
SrUu SrUX.SrU^
XVI.
because
XVIL
and
r2
q.TT
_ ^9'
_I(^f'^y =
^^^^^^^^^ _^,
.Yq = -Vv. SUXfiv
-^W^
= - 2 + 2SUg = l^ÎM.
Tg
(5.)
that
it
Admitting then the expression VII.,
may
be thus transformed
XVIII.
and that the
R-^
for the
curvature
22"*,
we
easily see
=-Tv-i
L+ TX/x.cos li^ + i-\\;
difference of the two (principal) curvatures, of normal sections of an
arhitrary surface, answering generally to the two (rectangular) directions of the
conceive the differentials to be such. But it has already been abundantly shown, that
this view of the latter is by no means necessary, in the treatment of them by quaternions.
(Compare the second Note to page 667.)
DIFFERENCE OF CURVATURES.
CHAP. IM.]
'
681
curvature through the particular point considered, vanishes when the normal
V has the direction of either of the two ct/clic normals, X, /u, of the index sttrface
(410, (9-)) ^^^^ ^^' when the index curve (410, (4.)), considered as a section of
lines of
that index surface,

sense,
when the
a circle : or finally,
is
point in question
is,
in
a received
an umbilic* of the given surface.
(6.)
That
although considered to be a given one, has hitherto (in these

been treated as quite general. But if we now suppose it to be a
surface,
last sub-articles)
central surface of the
second order, and to be represented by the equation,
XIX.
.fp
= gp^~-\- SXpfip = 1,
which has already several times occurred,
we
see at once,
from the formula VII. or
XVIII. (comp. 410, (10.)), that the difference of curvatures, of the two principal
normal sections of any such surface, varies proportionally to the perpenrf/cM/ar (Tv"'
or P) from the centre on the tangent plane, multiplied by the product of the sines of
the inclinations of that plane, to the
two
cyclic planes of the surface.
(7.) In general (comp. 409, (3.)), it is easy to see that
XX.
if
of T
..
= Sr-i.4r
= -Z>-2
S^
^
dp
denote the (scalar) semidiameter of the index surface, in the direction of dp or

but for the two directions of the lines of curvature, these semidiameters become
of the index curve.

Denoting then by ai and aj these
(410, (3.), (4.)) the semiaxes
semiaxes, the two principal radii of curvature oi any surface come by IV. to
last
be thus expressed
XXI.
And
if
Ri
= ai^Tr
E^ = aa^Tv.
the surfaae be a central one, of the second order, then ai, a2 are the semiaxes
Tv is (comp. again 409,

on that plane from the centre.
of the diametral section, parallel to the tangent plane ; while

let fall
(3.)) the reciprocal P-i of the perpendicular,
Accordingly (comp. {6.), and 219, (4.)), it is known that the difference of the interse squares of those semiaxes varies proportionally to the product of the sines of
the inclinations, of the plane of the section to the
(8.) And. as regards the squares themselves,
they
be thus expressed, in terms of
may
faces,
and
in
the.
agreement with known results
XXII.
two
it
cyclic planes.
follows from 407,
LXXI., that
principal semiaxes of the confocal sur:
_ i2
ai2 = 2
aa^
= a^- fl2*
being thus both positive for the case of an ellipsoid ; both negative, for that of a
double- sheeted hyperboloid ; and one positive, but the other negative, for the case of
an hyperboloid of one sheet (comp. 410, (15.)).
(9.) In all these cases, the
normal
is
drawn towards
the
same
side of the
is situated (because
tangent plane, as that on which the centre O of the surface
I. and XXI.) both the radii of curvature Hi, R^, are drawn
Svp = 1) hence
;
(by
towards
in this direction, or
sin^Zs-sheeted hyperboloid,
this side, for the ellipsoid
and both
but one such radius for the
radii for the hyperboloid of
two sheets, are di-
rected towards the opposite side, as indeed is evident from the forms of these surfaces.
Compare
4s
682
[bOOK
IH.
(10.) The following is another method of deducing generally the two principal
curvatures of a surface, from the self-conjugate function,
XXIII.
which
some good
affords
dv = 0d|O,
410, IV.
practice in the processes of the present Calculus.
Writing,
for abridgment,
XXIV.
-^ = i2-iTv = - S ^ = - Sr-i0r,
cr-p
where r
the form,
XXV.
where
rr v~^Yv(pT = (pT
denotes a new linear
4>
and
we have
self-conjugate, because
not generally
the form,
XXVI.
ii.
6,
v'^Srtpv
II. is easily
brought
to
Or,
which however
vector function,
tion on the plan of the Section III.
and with the
dp
a tangent to a line of curvature, the equation
is still
not in general
is
Treating then this neiv funcwe derive from it a new cubic equation, oi
=M+
^v
\\
v.
MV + M"r2 + r^,
coefficients,
XXVII.
..
3/=
M' = Sv->i//v,
0,
M" = m"-S/-i^j/;
a certain auxiliary function (= m^'i), and m" being the coefficient* analo\p being
= is foreign
gous to M", in the cubic derived from the function
itself The root r
to the present inquiry
but the two curvatures, Ri'^, R^'^, are the two roots of the
;
following quadratic in
that foreign root
It'^,
obtained from the equation
XXVI. by
the rejection
of
XXVIII.
As a
(11.)
first
as in v., the form
application of this general equation
gr
4-
y\T[i
we
=a
shall then
again,
with those assigned by VII.

(12.) As a second application,
and the two
let the surface
semiaxes (real or imaginary)
its scalar
^r have
lt
have the values,
great variety of transformed expressions
abc for
iVf .
XXVIII.,
XXIX.
M"=2(g+ SXJJv.SfxVv),
J/' = (^ + S\Uv SfiVvY - (VXUv)2 (YfiVvY,
.
XXX.
and
= (R-^TvY + WR-YIv +
resulting curvatures agree
be central of the second
order, with
then the symbolical cubic (350)
in
becomes,
XXXI.
and the
= 03 - m"02 + m'0 - m = (0 +
coefficients of the quadratic
XXVIII.
a-^) (0
b'^) (0
denotes the semidiameter of the surface in the direction of the normal
XXXII.
Ri-^
+ i?3-i = - J/"Tv-i = - (m" +/Uj/) P=
Compare the Note
to page 673, continued in
evanescence of the coefficient

that
we have
here $0^1^
tor (comp. 351).
= 0,
M,
(-2
page 674.
+ 63 + c-2 - N^) P;
The reason
so that the
symbol $-'0
may represent
same kind,
of the
from a surface of the second order
concentric, similar,
is
an actual vec-
Geometrically, this corresponds to the circumstance that
the normal does not change.
or of the occurrence of a null root of the cubic,
pass, along a semidiameter prolonged,
surface of the
c-2)
in R'^ take the values, in which A'
and similarly placed, the
when we
to another
direction of
CHAP.
PRODUCT OF CUIIVATURES.
III.]
XXXIII.
i2i-Ji?2-i
683
= 3nv-2=-mv-< = a-6-2c-2p4;
both of which agree -with known results, and admit of elementary verifications. *
= x (360, XVI. ), we shall see that
(13.) In peneral, if we observe that m"-<p
the quadratic XXVIIl. in r (or in H^Tv) may be thus written:
XXXIV.
or thus
more
briefly
= Sv-^(r^v-\-rxv + -4^v);
(comp. 398, LXXIX.),
XXXV.
= Si/-i(^ + r)-iv.
XXV. gives the expression,

XXXVI.
v2r = (0 + r)-iv. Sr^r
condition Svr = 0, the equation XXXV. follows
(14.) Accordingly, the formula

.
from which, under the

(15.)
tures
We
have therefore
of sections of
^'cweraZ/y, for the procfac^ of the
ani/ surface at
XXXVII.
any
at once.
two principal curva-
point, the expression:
Ei-^B2-^=rir2Tv-^ = -v-'Svrpv
which contains an important theorem of Gauss, whereto
=- S
;!/
we
i
V
shall presently proceed.
(16.) Meanwhile we may remark that the recent analysis shows, that the squares
al^ ao^ (7.) of the semiaxes of the index-curve are generally the roots of the follow-
ing equation,
XXXVIII.
when developed
= Sv(0 + a-2)-'j/,
as a quadratic in s?.
(17.) And that the same quadratic assigns the squares of the semiaxes of a diametral section, made by a plane -L. v, of the central surface of the second order which
has
Sp^p =
1 for its equation.
(18.) Accordingly, Vpîp has the direction of a tangent to this surface,

perpendicular to p at its extremity ; and therefore the vector,
XXXIX.
= p-^yp^p ^fpp- p-^ =
(7
which
is
{1>- p'^) P,
perpendicular to the plane of the diametral section, which has the semidiameter p
The equation,
so that it is perpendicular also to p itself.
for a semiaxis
is
XL.
S(T(0-p-2)-i(r
assigns therefore the values of the squares
= O,
( p^) of the scalar semiaxes of the cen-
a which agrees with the formula XXXVIII.

of the second or(19.) If then a surface be derived from a given central surface
to the diameder^ as the locus of the extremities of normals (erected at the centre)
tral section -^
tral sections of the given surface, each such
normal (when
real)
having the length of
one of the semiaxes of that section, the equation of this new surfacef (or locus) will
admit of being written thus
:
XLI.
..
Sp(^-p-2)-ip=0.
* As an
it may be
easy verification by quaternions of the expression XXXII.,
remarked (comp. 408, (27.)), that if a, (3, y be any three rectangular unit lines,
then
fa +//3 +fy = const.

t
When
the given surface
the celebrated
farther on.
Wave Surface
is
an
= Ci + C2 + Ci= a"^ + h"^ + c-2.

ellipsoid, this derived surface
of Fresnel,
which
will
XLI.
is
therefore
be briefly mentioned somewhat
684
The
(20.)
of the values
first
= ^jO,
$ion (if
as
by
and
(r
= p+r-ir = (l+r-'^)p = r-i(^ +
and operation with
inversion,
XLIII.
for the auxiliary scalar r, gives the
II'
expres^
for a central surface of the second order),
it is
XLIL
whence,
XXIV.,
[bOOK
r)p;
0,
= r(0+r)-i(r;
XLIV.
v= r(0 + r)-i0(T;
therefore, because
XLV.
Spv= 1,
r-2 = S((0 + r)-(T.(^ +
The following
(21.)
useful for other purposes.
of Centres^ that
on the locus of
is,
all the centres of
XLVI.
S<Ti;
Sucr
is
may
369, IV.)
r)-2^(T.
=1
XLVII.
Si;d(T=
connected with a (comp. 373,
XLVIIL
Sddv = 0;
as before, a tangent to the line of curvature
is,
riable vector of the
by the vector expression,
all satisfied
XLIX.
where r
S.ff(0
curvature of principal normal sec-
the vector (say v) of the Reciprocal Surface

*
(21.)) by the equations of reciprocity
tions,
which are
r)-i0(T)
a quite different way of arriving at this result, which is also

Considering a as the vector os of a point s on the Surface
is
normal plane
so that, if
a>
denote the va
to this last curve, the equation of that plane (comp.
be thus written,
L.
This normal plane, to the
line
Sy((u-p) = 0.
of curvature at p,
is
therefore at the
known
same time
the
from simple
geometrical considerations, independently of the form of the given surf^^ce, which remains here entirely arbitrary.
tangent plane to the surface of centres at
(22.)
The expression XLIX.
for
LI.
giving
It
of centres,
as indeed
it is
v gives generally the

.
to be,
relation,
Spi;=l;
by 410, V. and VI., these two other equations.
also,
XLVIIL,
s,
understood that
is
and dy,
dcr
in
the differential
equations
XLVIL,
are in general only obliged to have directions tangential to the surface
and
to
its
reciprocal, at corresponding points
so that the equations
might be in some respects more clearly written thus, Svda = 0, SctSv = 0, the mark
d being reserved to indicate changes which arise from motion along a given line of
curvature, while S should have a more general signification. Accordingly if, in parwe write dp = vdp, for a variation answering to motion along the other line,
ticular,
and denote the two

i22,
we
shall
radii of curvature for the
have by
II.,
d(7=d/?i.Uv,
so that
we have
iBr'dp +
d(T
dLV =
two
directions
0, R-r^dp
+ SUt^ =
dp and dp by i?i and

and therefore by I.,
0,
= Sp + d(RiVv) = (l-RiRrir')v6p+SRi.Vv;
both Sdpd(r= 0, and 8dpd(T= 0, and therefore the tangent dp or r
normal v to the corresponding

otherwise visible from geometry.
And when
we have thus found an equation of the form <y=r, operation with S.<t giv-= hv
XLVI. the value t = Spr, as in XLIX., because <t - p v -^ t.
to the given line of curvature has the direction of the
sheet of the surface of centres, as
is
\\
CHAP.
RECIPROCAL OF SURFACE OF CENTRES.
III.]
LII.
Avhich are
Svu =
and
0,
LIII.
Svv(pv
685
0,
independent of the /orm of the given surface.

(23.) But if that surface be a central quadric,* then the equation LI.
still
be
may
thus written,
LIV.
combining -which with
LIL and
= 8v(p-^v = Sv(p-^v
we
LIII.,
L\...v=
LVI.
-;
derive the expressions:
^:
= (fi-'j^=
/y = Sy0u, and Fv Sv^-^v, as usual.

(24.) Operating with S. v on this last expression for p, and attending to LII.
and LIV., we find the following quaternion forms of the Equation of the Reciprocal
A\
herein
of the Surface of Centres
LVn.
(Svp =)
,2f^F
LVIII.
""^
'
i;4
= (iî; - )/y
]
or
LIX.
= (Fu-l)/-;
LX.
or
Fu
&c.,
/-
when
whereof the second,
with a
translated into co-ordinates,
of the
knownf equation
same
(25.) Differentiating the form LX,,
LXL
we
find,
iVi
/
.
v^
/-
and observing that
d.v4 = 4Sw3di;,
found to agree perfectly
is
reciprocal surface.
d> = 2S^fdy, dFu = 2S^-ivdu,
by comparison with XLVI. and XLVIIL, the expression:
LXIL..,T = 0-'t;-?^+,^;
or finally
LXIIL
or
= r'-+
by XLIX., with the recent signification
XXIV.
= J-2 (0 + r)2 ^-u,
LXV.
LX IV.
(7
because
LX.
and, for the same reason, the equation
of
-^^
r,
=/Ur =flJv
of the reciprocal surface
may
be thus
briefly written,
LXVL
Fv +
r-iy2
= i,
(26.) Inverting the last form for

find for
LXVII.
/u +
and using again the
Compare the
= r2(0 + r)-2 ^cr
to the equation
LXVIII.
(7,
LXVI'.
n;2
relation
= o.
XLVI., we
first
V the expression,
and then are conducted anew
'*
^vhile
last note to
XLV.,
= S. (7(1 +
page 672
or to the following,
r-V)'20(7.
see also the use
made
of this
known name
quadric," for a surface of the second order (or degree), in the sub-articles to 399
(pages 614, &c.),

Tlie equation alluded to, whicli is one of the fottrth degree, appears to have been
assigned by Dr. Booth, in a Tract on Tangential Co-ordinates (1840), cited in
page 163 of Dr. Salmon's Treatise. See also the Abstract of a Paper by Dr. Booth,
first
in the
Proceedings of the Royal Society for April, 1858.
686
(27.) This last equation
LXIX.
III
may also be thus written,

l = S.a(l + r-i0)-3 (^)(T;
but by combining XLIII. LI.
LXVII. we
have,
= (Spu=)S.<r(l+r-i^)-3^<T;
= S. (t(1 + r-i0)-3^2o.^
LXXL
LXX.
hence
a result which
(28.) If
LXXII.
r'
we
XLIV. LXVII.
for
v and
v.
write,
r = Udp,
be otherwise and more directly deduced, under the form Svu =
may
(LIL), from the expressions
r and
[bOOK
t'
= V(vdp),
LXXIII.
and therefore
.tt'
= Uv,
being thus unit-tangents to the lines of curvature, the equation III. gives,
generally,
LXXIV.
= Vrd(rr') = - dr' + rSr'dr,
whence
LXXIV.
dr'
|1
of which general parallelism of dr' to r, the geometrical reason is (comp. again llj.)
that a line of curvature on an arbitrary surface is, at the same time, a line of curvature on the developable normal surface which rests upon that line, and to which
the vectors r' or vdp are normals.
The same
(29.)
LXXIII.
substitution
for Ui/ gives
by
II., if
we
denote by the
arc of a line of curvature, measured from any fixed point thereof, so that (by 380.
(7.), &c.),
LXXV.
Tdp = d,
dp
= rds,
Bsp
r,
the following general expression for the curvature of the given surface, in the direction r of the given line, which by LXXIV. is also that of dr'
:
LXXVL
= S.rD,(rr')=-S.rr'D,r = S(Uj/-.D,V);
72->
(by 389, (4.)) what we have called the vector of curvature of the line of
curvature, considered as a curve in space, and R'^TJi^ is the corresponding vector of
curvature of the normal section of the given surface, which has the same tangent r at
but Ds^p
is
the given point

-the former,
hence the latter vector of curvature
on the normal v
is
(generally) the projection of
to the given surface.
(30.) In like manner, if we denote for a moment by i?/i the curvature of the developable normal surface (28.), for the same direction r, the general formula II.
gives,
by LXXIV.,
LXXVII.
the vector
H^'W
of this
R-^ =
tT>st'
new curvature
is
=-
two
last results are
= S. r'-^D^V
therefore the projection on the
t\ of the vector of curvature Dg^p o? the given

see that these
Sr'D.r
line of curvature.
new normal
But we
shall soon
included in one more general,* respecting all plane
sections of an arbitrary surface.
(31.)
The general parallelism LXXIV. conducts easily, for the case of a central
known and important theorem, which may be thus investigated. Writ-
quadric, to a
ing, for such a surface,
LXXVIIL..r=/r,
r=fr',
*
Namely in Meusnier's Theorem, which can be proved generally by quaternion
with about the same ease as the two foregoing cases of it.
CHAP.
SURFACE OF CENTRES AS AN ENVELOPE.
III.]
here
o that r retains
LXV., and r
we have by LXXIV. the
recent signification
its
the other direction of curvature,
LXXIX.
because
Sr0r'=
We
(32.)
is
is
the analogous scalar for
differential,
= 0,
dr'=2S?ir'dr'=2Sr^r'Sr'dr
by 410, XI.
0,
have then the
relation,
LXXX.
that
is
687
/U (vdp) =/r' = / = const.
to say, the
square (r'-i) of the scalar semidiameter (/)') of the surface, which

the second tangent (r'), is constant for any one line of ctirvature
;
parallel to
and accordingly (comp. XXII., and the expression 407,

of this square
LXXI. for/Uvi),
(r)
the value
is,
LXXXI.
(/Ui/dp)-i=r'-i=a2-a'2 = 62_6'2 = c2_c'2,
be the scalar semiaxes of the confocal, which cuts the given quadric
if a\ h\
{abc)
along the line of curvature, whereof the variable tangent is r,
for instance, the
(33.) This constancy of /Uvdp may be proved in other ways
c'
Svdvdp
general equation
LXXXII.
dv = vSv-'dv + dpS
LXXXIV.
and
gives, for
dp,
||
LXXXVI.
we
=
.
we
0,
^p = v)
1,
curvature on an arbitrary surface,
LXXXIII.
LXXXV.
which gives Vi/d^p
line of
S dp0 {vAp) =
while for a central quadric (/p
hence, for such a surface, if
it is
Vdvdp = vApSvÂv
dv
because
= ^dp
easy to show that
we have
also,
0(/dp)=Vpdp/(x;Udp);
suppose for simplicity that d or
Tdp
is
constant,
have,
d/(vdp)
= 2S(0(vdp).d(vdp)) =
a differential equation of the second order, of which
LXXXVII.
/(vdp) = Cj/2dp2,
a.
LXXXVII'.
or
2Sv-idi/./(j/dp),
first integral
.
is
evidently,
/U(a/dp) = C = const.
(34.) But we see that the lines of curvature on a central quadric are thus included in a more general system of curves on the same surface, represented by the
differential
equation
LXXXVI.,
of which the complete integral would involve two
and which expresses that the semidiameters parallel to those tangents to

the surface, which cross any one such curve at right angles, have a common square,
constants
and therefore
(if real)
common
length, so that (in this case) they terminate on
sphero-conic.^
(35.) Admitting however, as a case of this property, the constancy LXXX. of
the scalar lately called r', namely the second root of the quadratic XXXIV. or
XXXV.,
to
of which the coefficients
another of what we
to conceive r
and v
ceive this theorem^
may call
and the
for the
first
moment
root r vary, in passing from one point

a line offirst curvature, we have only
to be accented in the equations
which perhaps
Compare
is
new
LXVI. LXVl'.,
in order to
per-
the sub-articles (6.) (7.) (8.) to 219, in page 23l.
688
The Curve* on
central quadric,
[bOOK
III.
of the Surface of Centres of curvature of a

second curvature of that given surface for all
the Reciprocal (24.)
which answers
to the
the points of a given line oi first curvature, or which
is itself \i\
a known sense the
reciprocal (with respect to the given centre) of the developable normal surface (28.)
which rests upon that line, is the intersection of two quadrics ; whereof one (LXV I'.)
while the other (LXVI.) is a
is a cone, concyclic with the given surface (fp = 1)
;
surface concyclic with the reciprocal of that given quadric
{Fv =
1).
(36.) Again, the scalar Equation of the Surface of Centres (21.) may be said
to be the result of the elimination of r-i between the equations LXVIII. and LXXI.,
whereof the
latter is the derivative^ of the
have therefore
this
theorem
former with re?pect to that scalar
we
An Auxiliary Quadric (LXVIII. or XLV.) touches the Second Sheet of the

Surface of Centres of a given quadric, along a Quartic Curve, which is the locus of
the centres of Second Curvature for all the points of a. Line of First Curvature (35.)
;
and (for the same reason) the same auxiliary quadric is circumscribed, along the
same quartic, by the Developable Normal Surface (28.), which rests on thsit first
line: with permission, of course, to interchange the ysoxAs first and second, in this
enunciation.
(37.) "When the arbitrary constant r is thus allowed to take successively all vacorresponding to both systems of lines of curvature, the Surface of Centres is
lues,
therefore at once the Envelope^ of the Auxiliary
of the Quartic Curve
(36.), in
auxiliary quadric in one of
its
Quadric LXVIII., and the Locvs

which one or other of its two sheets is touched, by that
successive states, and also
by one of the developable
surfaces of normals to the given surface.
(38.)
To
obtain the vector equation of that envelope or locus,
* The variable vector of
this curve
is
r'
easily seen (comp.
we may
XLIX.)
proceed
to be,
VT
and the reciprocal surface (21.) or (24.) is by (25.) the locus of this quartic (35.).
f The analogous relation, between the co-ordinate forms of the equations, was
perhaps thought too obvious to be mentioned, in page 161 of Dr. Salmon's Treatise;
or possibly it may have escaped notice, since the quartic curve (36.) is only mentioned
there as an intersection of two quadrics, which is on the surface of centres, and
answers to points of a line of curvature upon the given surface. But as regards
the possible novelty, even in part, of any such geometrical deductions as those given
in the text
from the quaternion analysis employed, the writer wishes to be under-
stood as expressing himself with the utmost diffidence, and as most willing to be
The power of derivating (or differentiating) any symbolical
corrected, if necessary.
scaexpression of the form LXVIII., or of any analogous /brm, with respect to any
which it involves explicitly, as if the expression were algebraical, is an important

but an easy consequence from the principles of the Section III. ii. 6, which has been
lar
!?o
often referred to.
^ Compare the Note immediately preceding.
CHAP.
VECTOR EQUATION OF SURFACE OF CENTRES. 689
III.]
using a new expression for <r, in terms of v or of p, which may then be

transformed into a function of two independent and scalar variables.
Denoting
as follows,
(comp. (32.)) by
ai, 61, ci
the semiaxes of the confocal which cuts the given surand by 03, 62, C2 those of the other confocal, so
face in the given line of curvature,
two confocals have the directions of the tangents r',

not only the expressions LXXXI. for r'-', with a'b'c'
but also the analogous expressions (comp. 407, LXXL),
that the normals vi, V2 to these
T lately considered,
changed to
we have
ai, 61, ci,
LXXXVIII.
We have
therefore
r-J
= a2 - as^ = 62 _ 63^ = c2 - c-z^.
by XLIL, combined with 407,
XVL,
this
very simple expression
for <r:
LXXXIX.
(T
= ((6- +
r-i)v
and as a
containing, in the present notation,
= 03-iv = 02"'0|O;
result of the present analysis, a
known
and interesting theorem,* on which however we cannot here delay.

(39.) It follows from this last value of o-, combined with the expression 408,
LXXXII.
for p, that
we may
write,
Yaa'
\aA-a'
as the sought Vector Equation of the Surface
quadric (abc)
of Centres of curvature of a given

ambiguous signs being virtually included in these three terms, be-
cause in the subsequent eliminationsf the semiaxes enter only by their squares
while I, a, a' are constants, as in 407, &c., for the whole confocal system, and ahc
:
are also constant here, but a^
and
may
_ ^2
and a^
a-^,
or r'"i
and
r-i
(38.), are variable,
be considered to be ihQ two independent scalar s of which
cr
is
a vector func-
tion.
413.
Some
brief remarks
may
here be made, on the connexion
of the general formula,

I.
in
Si^-'
(0 + r)-'u =
0,
412,
XXXV.
which r-R'^Tv (412, XXIV.), and which when developed by the

ii. 6 takes
(comp. 398, LXXIX.) the form
rules of the Section III.
of the quadratic,
*
Namely Dr. Salmon's theorem (page 161
of his Treatise), that the centres of

a given point are the poles of the tangent plane,
The connected theorem (page 136), respecting
with respect to the two confocals.
the rectilinear locus of the poles of a given plane, with respect to the surfaces of a
curvature oi
a,
giv&a. quadric at
confocal system, is at once deducible from the quaternion expression 407, XVI. for
although the theorem did not happen to be known to the present writer, or at
0'iv,
least
remembered by him, when he investigated ih^t formula of inversion

which some have been already given.
for other
applications, of
t The corresponding elimination in co-ordinates was first effected by Dr. Salmon,

who thus determined the equation of the surface of centres of curvature of a quadric
162 of his already cited Treato be one of the
twelfth degree. (Compare pages 161,
tise.)
4t
690
II.
j-2
[bOOK
+ rSv-\u + Sp-'irf = 0,
412,
III.
XXXIV.
with Gauss's* theory of the Measure of Curvature of a Surface; and

especially with his fundamental result, that this measure is equal to
the product of the two principal curvatures of sections of that surface:
a relation which, in our notations, may be thus expressed,
III.
V.dUv BJJv = Ei-'Ei'Ydp^p.
(1.) As regards the deduction, by quaternions, of the equation III,, in which d

aud S may be regarded as twof distinct symbols of differentiation, performed with respect to two independent scalar variables, we may observe that, by principles and
rules already established,
IV.
and that
dU;
= V-.Uv,
therefore the first
V.
dVv = V
member
V.dUvÛv =
(2.) Again, since we have dv

= 0, Svdp
^^p, the relations Si/dp
of III.
may
be thus transformed
.V
v(
.Vv = -Vv.Y V
= 0d|O (410,
= 0, and the
= -i/->S'-id/^v.
IV., &c.), and in like

self- conjugate
manner ^v =
property of ^, allow
us to write,
VI.
whence follows
sion,
Vdr^v =
at once
and
^pYdp^p,
VII.
by V. the formula
deduced from the quadratic

VIII.
III., if
Vdo^p = v'^SvApSp
we remember
the general expres-
II.,
= - r-Vira = - S -
i2i-ii?2-
t//
-.
412,
XXXVII.
(3.) If then we suppose that p, Pi, P2 are any three near points on an arbitrary surface, and that K, Ki, R2 are three near and corresponding points on the
unit sphere, determined by the condition of parallelism of the radii or, ori, 0R2 to
the normals pn, PiNi, P2N2, the two small triangles thus formed will bear to each
other the ultimate ratio,
APP1P2
a result
which
justifies
* The reader
is
Calthough by an entirely new analysis) the adoption by Gauss
referred to the Additions to Liouville's
in which the beautiful
Memoir by Gauss,
Monge
(pages 505, &c.),
Disquisitiones generates circa

superficies curvas, is with great good taste reprinted in the Latin, from the Commentationes recentiores of the Royal Society of Gottingen.
He is also supposed to look
back,
if
necessary, to the Section III.
especially to the deduction in page
ii.
437
function (and therefore also the former)
t Compare page 487, and the Note
entitled
6 of these Elements (pages 435, &c.), and

of 7p from 0,
is
to
remembering that the
here self-conjugate.
page 684.
latter
CHAP.
MEASURE OF CURVATURE OF A SURFACE.
III.]
691
product* of curvatures of sections, as the measure of the curvature of the

with his signification of the phrase.
of this
surface,
As another form of this important product or measure, if we conceive that

(4.)
the vector p of the surface is expressed as a function (372) of two independent scalars, t
and
X.
which
and
if
we
..Dtp = p',
will allow us
Dp = p
Bt^p
= p",
DtDup =
= p^^,
D2p
p/,
(comp. 372, V.) to assume for the normal vector v the expres-
sion,
XL..v = Vp'p.
easy to provef that
it is
we have
XII.
generally,
=S
i?rii23-i
^ S^ V
S^'Y;
V
I
which takes as a verification the well-known form,
xm...ie,.ij.H =
when we write (comp. 410,
XIV.
XV.
p=ia?+jy + Ar, p =T)p = i + kp, p^ = Dyp=j + kq]

V = Vp'p^ = k ip -jq, p" = ^r, p/ = ksy p,^ = kt.
(5.) In general, the equation
XVI.
also
XVIII.
and
if
..
XII.
may
be thus transformed,
v4i?rii?3- = S (Yvp".YvpJ - (Jvpiy + v2 (Sp"p,^ - p/s)

XVII.
Tdp2 =r ed<2 + 2/dMM + gû\
e =
v"^ =P - eg
9 = -p,^ whence XIX.
-p'2, f=-Sp'p,,
if
_^_^,
(18.)),
we still denote, as in X., derivations relatively to t and u by upper and

we may substitute in the qtiadruple of the equation XVI. the values,
lower
accents,
XX.
XXI.
and
2 (Sp"p,,
hence the measure of curvature
XXlL.,e, fg;
and
therefore, as
* If
it
2Vvp/ = -^'p' + e,p^,
2Vj/p"= (e,-2/)p'+e'p^,
is
an
p/')
= e - 2// + g"
explicit function
e\f\g';
e^^f^g^;
was otherwise proved by Gauss,
of the ten scalars,
and
this
2Yvp,=-g/
e-2f'+g":
measure depends onlyX on the
be supposed to be in any manner known that a limit such as IX. exists,

two vector areas in III. is a scalar independent of the di-
or that the quotient of the
rections of ppi, PP2, or of dp, ^p,

tions of the lines
-Ri~î?2"'
we have
only to assume that these are the direcby 412, II., the product
in order to obtain at once,
of curvature,
as the value of this quotient or limit.
The quadratic in R'^ may be formed by operating on 412, II. with S.p' and
S.p and then eliminating d< d.
of this exclusive dependence, is perX The proof by quaternions, above given,
shorter than the correspondhaps as simple as the subject will allow, and is somewhat
t'
ing proof in the Lectures
in
page 606 of which
is
given however the equation,
692
[bOOK U
expression (XVII.) of the square of a linear element, in terms of two independfl
and of their differentials (dt, d).

Hence follow also these two other theorems* of Gauss
a surfacQ be considered as an infinitely thin solid, and supposed
scalars
(^, ?/),
(6.)
If
h^ flexible
to
but inextensible, then every deformation of it, as such, will leave unaltered, 1st, the
Measure of Curvature at any Point, and Ilnd, the Total Curvature of any Area;
that
by
the area of the corresponding portion of the unit sphere, determined as in (3.)
is,
radii parallel to normals.
t and are geodetic co-ordinates, whereof the former re(7.) Supposing now that
presents the length of a geodetic ap from a. fixed point a of the surface, and the latter
represents the angle bap which this variable geodetic makes at A with a fixed geo-
detic AB, it is easy to see that the general expression
XXIII.
so that
Td/)2
we have now
= df2 + 2dtt2,
in
which
XVII. takes the
XXIV.
shorter form.
= Tp, = Tv
the values,
XXV. ..e=l, /=0,

and the derivatives of
= n^,
the measure of curvature reduces itself
XXVI.
g'=2nn',
g"
= 2nn" + 2n'^
And
and /all vanish.
thus the general expression XII.

by (5.) to the very simple form,
ier>i?2-i
= - n-in" = -
n-^Dt^n
for
generally a function of both t and u, although here twice derivated

with respect to the former only.
(8.) The point p being denoted by the symbol (t, ), and any other point p' of
the surface by (t -\- At, u + Au), we may consider the two connected points Pi, Pa, of
in
which n
is
which the corresponding symbols are (t + At, u) and (t, u + Am) and then the
quadrilateral PPiP'F2, bounded by two portions pPi, P2p' of geodetic lines from A,
and (as we may suppose) by two arcs PP2, Pip' of geodetic circles round the same
;
fixed point, will have
its
area ultimately
=nAtAu
(by XXIII.), and therefore (by
XXVI., comp. (3.), (6.)) its total curvature ultimeLtely = n" At Au, or = Ann'. Aw,
when At and Am diminish together, by an approach of p' to p.
= ^, n' = 1 chang(9.) Again, in the immediate neighbourhood of A, we have n
and
with
t from * = 0, we obtain
to
to
then
Atn'
respect
d^n',
integrating
ing
;
- n'
as the coeflScient of
Aw
in the result,
and are thus conducted
to the expres-
sion:
XXVII.
Total Curvature of Triangle
APp'=
(1
n')
A, ultimately,
AP, Ap' be any two geodetic lines, making with each other a small angle
and if pp' be any small arc (geodetic or not) on the same surface.
if
4 ieg-pyHi-^Rî =
= Au,
(/^- 2gf + ^,^J
- e,g' - 2e,f - 2g'f + 4/'/,)

+ 9 {.e? - 2e'f + e'g') - 2 (êg -/2) (e,^ - 2// + <?"),
"rfê'g,
Avhich
XXL,
may now
and
be deduced at sight from XVI., by the substitutions

only in notation from the equation of Gauss
differs
page 623, or Salmon, page 309).

* See
page 524 of Liouville's Monge.
XIX. XX.
(Liouville's
Monge,
CHAP.
TOTAL CURVATURE, SPHEROIDAL EXCESS.
III.]
then that
(10.) Conceive
which therefore
is
finite arc oi
a,
XXVIII.
may
n',
Total Curvature of Area
the area here considered being
make with each other the
finite
any curve upon the
surface, for
be conceived to be a function of u
APQ=J(1 -n')dM = Aw - Jn'dM
bounded by the two geodetic lines ap, aq, which

angle A, and by the arc pq of the arbitrary curve.
(11.) If this curve be itself a. geodetic, and if we treat its co-ordinates t, m,

its vector p, as functions of its arc, s, then the second differential of p, namely,
XXIX.
must be normal
we
same kind,
this other expression of the
have
shall
pq
and consequently
t,
693
and
= p'^^t + P,d2M + p"dt^ +
2p;dtdu + p^du^,
and consequently perpendicular to p' and p^.
with S.p', and attending to the relations XVIII. and XXV.,
d2p
to the surface at p,
Operating* therefore
which give
XXX.
we obtain the
=-
p'2
we observe that
XXXIII.
I'
dt
= Sp'p', =
Sp'p"
Sp'p =
0,
Sp^p;
= nn,
XXXI.
if
Sp'p,
1,
d2<
we may
= nn'du^,
XXXII.
or
du
= - n'du,
write,
ndu = sin
= cos vds,
vds,
XXXIV.
because
being here the variable angle, which the geodetic
pq makes
d<
at
nK\u2 = d
P with ap pro-
longed.
for -n'dw, in XXVIII.,
(12.) Substituting then
the integration becomes possible, and the result is
angle at A, and
tt
+ Aw = (tt
v)
+ Ar)
(w
its
value dw given by
Au + Av; where Am
the
and
q, in
arrive at another
most
the sura of the angles at p
is
XXXII.,
is still
the geodetic triangle APQ.
(13.) Writing then B
and c instead of P and
remarkable Theorem f of Gauss, which
XXXV.
may
Q,
we thus
be expressed by the formula
Total Curvature of a Geodetic Triangle
= what may be called the
Spheroidal Excess
abc = a-1-b+c
A, b, c, in the second
used to denote the three angles of the triangle
and the
;r,
member, being
total surface of the unit
sphere (= Att) being represented by 720, when the part corresponding to the geodetic
o 1 80.
triangle is thus represented by the angular excess, A + b -j
(14.)
And
it is
heing extended, as
easy to perceive, on the one hand,

it
was by Gauss,
how
it
give
what would on the same plan be
may
require to be modified, as
of curvature
cipal sections
is
negative, as
have
it is
how
to all geodetic polygons

it
this
:
theorem admits of
and on the other hand,
was by the same eminent geometer, so as to

when the measure
called a spheroidal defect,
for surfaces (or parts of surfaces) of whichj^the prin-
their curvatures oppositely directed.
To operate with S.p, would give a result not quite so simple,^but reducible to
XXXI., with the help of d^s = 0.
t The enunciation of this theorem, respecting^ which its illustrious discoverer
" Hoc
justly says,
theorema, quod, ni fallimur, ad elegantissima in theoria superficierum curvarum referendum esse videtur," ... is given in page 533 of the Additions to
the form
694
[bOOK
III.
414. The only sections of a surface, of which the curvatures

have been above determined, are the two principal normal sections at
any proposed point; but the general expressions of III. iii. 6 may
be applied to find the curvature o^ any plane section, normal or oblique, and therefore also of any curve on a given surface^ when only
its osculating plane is known. Denoting (as in 389) &c.) by p and k
the vectors of the given point p, and of the centre k of the osculating
circle at that point, and by s the arc of the curve, we have generally
(by 389, XII. and VI.),
I.
Vector of Curvature of Curve - kp~'
= {p-
1^)'^
= DV =
j~
^ i"
the independent variable in the last expression being arbitrary. And

we denote by a and | the vectors of the points s and x, in which
if
the axis of the osculating circle meets respectively the normal and
the tangent plane to the given surface, we shall have also, by the
=
right-angled triangles, the general decomposition, kp"'
(as vectors), or
SP"^
xp''
II...D,V = (/-'c)-^ = (/>-^)-^ + (p-fr;

where the two components admit of being transformed as follows:
Normal Component of Vector of Curvature of Curve (or
III.
.
= {p-
Section)
ffi,
0-2
<t)"'
i^~'S
dv
(/>
- o-i)-'
cos^ v
{p
o-g)-'
sin^ v
Vector of Normal Curvature of Surface for the direction of

the given tangent;
being the vectors of the centres
s,,
Sa
(comp. 412) of the two
principal curvatures, and v being the angle at which the curve (or
its tangent d/>) crosses the first line of curvature (or its tangent tj),
while
<T
is
the vector of the centre s of the sphere which
osculate to the surface, in the given direction (of d/?)
IV.
is
said to
and
Tangential Component of Vector of Curvature
= (p=^
|:)-i
v-'dp-'Svdp-'d^p
Vector of Geodetic Curvature of Curve (or Section);
this latter vector being here so called, because in fact its tensor re-
Monge. A proof by quaternions was published in the Lectures (pages

606-609, see also the few preceding pages), but the writer conceives that the one
given above will be found to be not only shorter, but more clear.
Liouville's
CHAP.
CURVATURES OF SECTIONS,
III.]
OSC. SPHERES.
695
presents what is known by the name of the geodetic* curvature of a

curve upon a surface : the independent variable being still arbitrary.
As
(1.)
regards the decomposition
OA, OB, and if
= oc =
XLL),
408,
To prove the
dSpdp = 0,
'
'
'
p
(3.) Hence,
B^Vv, we have
VII.
for the case of
by 412,
we
transformation III.,
first
(2.)
that
II., if a, ^ be any two rectangular vectors

the perpendicular from o on ab, then (comp. 316, L., and
djO
dp2
djO
we denote
(7.), if
have, by
(r= p + i2Ur,
a central quadric
surface (410, (9.),
S;c.),
= J>2.Tj/,
i?
;
and XL, observing
dp*
(ip2
the vector III. of normal curvature by
the general expressions (comp. 412,

.
I.
I.
with
XXI.),
VIII.
D being generally the
Tj/
= i>-i,
semidiameter of the index
or for a quadric the semidiameter of that surface
which has the direction of the tangent (or of dp) and P being, for the
face, the perpendicular from the centre on the tangent plane, as in some
itself^
latter sur-
earlier for-
mulae.
To deduce the second transformation
(4.)
n, T2 denote unit tangents to the
Euler, let r,
which contains a theorem of
III.,
section
and the two
lines of curvature,
so that
IX.
we may then
= Ti
cosr
+ r3sinr,
X.
R-^lv^
shall
t^
= t\^^t-^=. \-^
dv
V
(j~ p
and
and
write generally (comp. 412, IV.),
= -S = -Sr->0r = Sr0r,
dp
have the values (comp. 410, XI.),
XI,
Sn0Ti = i?r'Tv,
whence
XII.
Sr2^r2 = -R2"^Tj/,
i?"'
Srj0T2 = Sr2^ri
= 0;
= i?r> cos2 r + jR2"' sin^ r,
and the required transformation is accomplished.

(5.) The theorem of Meusnier may be considered to be a result of the elimination
(2.) of d^p
from the expressions for the normal component III. of what we may call
; and it may be expressed by the equation,
the Vector Dg'^p of Oblique Curvature
XIII.
S^^=l,
-K
or
XI II'.
if it
be
now understood that
=
S^^
K
0,
which gives XIII".
the point
s,
of which a
is
the vector,
is
pks = ^,
2
the centre of the
* The
name, ''courhure geodesique,^^ was introduced by M. Liouville, and has
Compare pages 568, 575, &c.
been adopted by several other mathematical writers.

of his
Additions to Monge.
which osculates
circle
the
[bOOK U
696
same
normal
to the
section
direction to the surface, as will be
(6.) In general, \f p
+ Ap
or of the sphere
more
clearly seen
which osculates
by what
be the vector of anj/ second point
in
follows.
the given sur-
p' of
face, the equation
XIV.
u)
,
,
with
for a variable vector,
10
Ap
which touches the surface at the given point
represents rigorously the sphere
P,
and
passes through the second point p' ; conceiving then that the latter point approaches
to the former, and observing that the development* by Taylor's Series of the equation yp
XV.
even
= const,
.
if
= Ap-Âfp = 2S
= 2SvdjO,
gives (if dfp
and dv = 0dp),
+ S -^ +
V
(bAp
they be not always null,
we
terras
which vanish generally with Ap,
are conducted in a
new way, by
the
known
con-
same
cen-
ception of the Osculating Sphere for a given direction to a surface, to the

the equation of this sphere being,
tre s, and radius R, as before
:
XVI.
-^
p
(7.) Conversely, if
mediate between
Mm.
s'^ = -Ito.S*-^ =
Ap
Ap
we assume a radius R,
Ri^ and
or
10
is
algebraically inter-
= /e.
xvir...s^
lO
will cut the surface in two directions of osculation^ assigned

if
dp
such that R-^
R'i'\ the tangent sphere,
xvii...s-i^=:^,
p
Ji
but
V S*"
Rr^ be outside those
limits, there will
by the formula XII.
be only contact, and not any (real) in-
tersection, at least in the vicinity of p.
any second point of the surface, and if we denote

moment by (IT) and (2) the normal /)Zane pnp' and the normal section corresponding, we may suppose that n is the point in which the normals to the plane
and if we then erect a perpendicular at n to the
curve (2) at p and p' intersect
in (6.),
(8.) If p' be again, as
for a
plane (n),
section,
and
which we
this
it
will be crossed
p' to the tangent p't' to the
therefore in
p' to the
surface, in
may
call n'
by every perpendicular at
particular by the normal at
so that the line p'n is the projection^
a point
on the plane pp'n,
of
second normal p'n' to the surface.
but the point p'

of the
Conceiving then the plane (11) to hQ fixed,

to approach indefinitely to p, we see that the centre s of curvature
normal section (2), which is also by (6.) the centre of the osculating sphere
same direction, is the limiting position of the point N, in which
to the surface for the
Compare Art. 374, and the Second Note
to
page 508.
The
occasional use,
there mentioned, of the differential symbol dp as signifying a finite and chordal vector, in the development of /(p + dp), has appeared obscure, in the Lectures, to some
friends of the writer
and he has therefore aimed^
for the
sake of clearness, in
at least
the text of these Elements, and especially in the geometrical applications, to confine
that symbol to its first signification (100, 369, 373, &c.), as denoting a tangential
and to a curve or surface) p itself being generally
regarded as a vector fjtnction, and not as an independent variable (comp. 36'?, (3.j).
vector (finite or infinitely small,
CHAP.
tlie
VECTOR OF GEODETIC CURVATURE, DIDONIAS. 697
III.]
given normal at r
the given
(9.)
intersected
is
hrj
the projection* of the near
normal
on
p'n',
normal plane.
The two components
XVIII.
and IV. are included in the binomial expression,
III.
Vector of Oblique Curvature (or of Curvature of Oblique

Section)
= (p - k)-i = v-iSdi^dp-i + v-idp-iSj^dp-id2p,
obtained by substituting in I, the general equivalent 409, XXI. for

d'p,
and in which (as before) the independent variable is arbitrary ; and the tangential
which
is
component IV.
may
be otherwise found by observing that, by
XIX.
(10.) Another
way
p-K
of deducing the
the vector ^
must
satisfy
f^,
dp
= 0.
Sj^dp
because
same component IV.,
ing system of three scalar equations, which
II.,
=8^:^ = S^^ = -S.dp-.dV,

r
p-K
(vdp)-^ = v~^dp-\
and that
and
I.
by
is to
resolve the follow-
the geometrical definition of the point
XX. ..S(^-p)v =
S(?-p)dp = 0;
0;
S(^-p)dV = dp2;
and which give,
XXI
0^
^
""^P'
^^^P
Svdpd2p
or (p
^)-i = &c.,
We have also the
as before.
XXII.
= (vdp)-i
transformations,
Vector of Geodetic Curvature
St'dp-idV'
S(j/Udp.dUdp) =
i;dp
(p-4)-
1^1^= &c.
x shows also easily, that if a developable sura given surface (s), along a given curve (c), and if in
the unfolding of the former surface, the point x be carried with the tangent plane,
originally drawn to the latter stoface at P, it will become the centre of curvature, at
(11.)
The
definition of the point
face (d) be circumscribed
the
new point (p),
that the radius
to the
to
new or plane curve
(px) of geodetic curvature
is
(c') obtained
by
this
development
equal, as indeed
it is
knownf
the radius of plane curvature of the developed curve.

(12.) Thisj9/awe curve (c') is therefore a circle^ (or part of one)
tion,
XXIIl.
* The reader
PX = T (^ -
p)
if
so
to be, to
the condi-
= const.,
calculations and constructions, in pages 600, 601

In the language of infinitesimals, an infinitely near normal v'^'
intersects the axis of the osculating circle, to the given normal section,
may compare the
of the Lectures.
t Compare page 576 of the Additions to Liouville's Monge,

become circles by development,
% The curves on any given surface, which thus
have also the isoperimetrical property expressed in quaternions (comp. the first Note
to
page 530) by the formula,
XXVI.
=
5jS(Ut/.dp^p) + c^jTdp
0,
which conducts to the differential equation,
XXVII.
c-idp
= V.UvdUdp
4 u
(comp. 380, IV.),
698
but
be
satisfied
be
infinite, that
it
degenerates into a right
[bOOK
radius of geodetic curvat
line, if this
is, if
XXIV.
T (p - 0'^ =
0,
or
XXV.
Svdpd^p
= 0,
or finally (by 380, II., comp. 409, XXV.), if the original curve (c) be a geodetic line
on the given surface (s), and therefore also on the developable (d) which agrees
with the fundamental property (382, 383) of geodetics on a developable surface.
:
(13.) Accordingly it may be here observed that the general formula IV., combined with the notations and calculations of 382, conducts to the expression
ZQOC -4- QU
for the geodetic curvature of any curve on a developablij
(z + o')T|o'-\ or
,
surface, whereof the element ds crosses a generating line at the variable angle v, while
is the angle between two such consecutive lines : a result easily confirmed by geometrical considerations, and agreemg with the differential equation z + v' =
(382,
zdx
IX.) of geodetics on a developable.
415.
We
shall conclude the present Section
with a few supple-
mentary remarks, including a new and simplified proof of an important theorem (354), which we have had frequent occasion to
employ for purposes of geometry^ and which presents itself often
m physical applications of quaternions also

and
namely, that if the linear
be self-conjugate, then the Vector Quadratic,
vector function
I.
Yp(pp =
354,
0,
I.
represents generally a System of Three Real and Rectangular Directions ; and that these (comp. 405, (1.), (2.), &c.) are the directions
of the Axes of the Central Surfaces of the Second Order, which are
represented by the scalar equation,
II.
more generally,
III.
S/30P = C/J^-l- C, where
S/)0/>
= const.
or
. .
C and
C are any two scalar constants.
(1.) It is an easy consequence of the theory (350) of the symbolic and cubic
equation in ^, that if c be a root of the derived algebraical cubic iVf=0 (354), and
if we write * =
+ c (as in that Article), the new linear and vector function $p must
be reducible to the binomial form (351),
and
in
which the scalar constant
XXVIII.
(^
p) U.j/dp
radius of developed circle
can be shown to have the value,
= + T (? - p) = Radius
and each such
fllJ
of Geodetic Curvature,
curvê includes,
by XXVI., on
the given
maximum area
with a given perimeter : on which account, and in allusion

to a well-known classical story, the writer ventured to propose, in page 582 of the
surface, a
Lectures, the
name
" Didonia" for a

curve of this kind, while acknowledging that
the curves themselves had been discovered and discussed by
M. Delaunay.
CHAP.
NEW PROOF
III.]
IV.
OF RECTANGULAR SYSTEM.
.^p = <l>p+cp = fiSap + jS'Sa'p,
as the condition (353,
XXXVI.)
V.
with
V/3a
With
o{ self- conjugation .
699
+ V/3'a' = 0,
we may
this condition
then write,
YI.
p=zAa + Ba',
= A'a' + Ba;
(5'
easy to see that no essential generality is lost, by supposing that a and a

are two rectangular vector units, which may be turned about in their own plane, if
be suitably modified so that we may assume,
/3 and /3'
and
it is
VII.
a2
IX.
and
= a'2 = V/3'a'
(2.)
The equation
X.
and
it
1,
whence
= Baa = - Y(3a,
I.,
Yp^p = 0,
is,
XL
by
XIV.
and
*a = - /3, $a' = = - A'aa'.
/3',
V/3'a
XII.
or
Yaa'p =
suppose,
V (a?/3 + x'/3')
= 0,
$jO
we
(xa + x'a) =
by IX.,
XV..
B(ix^-x^) + (A-A')xx'=0:
while conversely the expression XIII. will satisfy

this quadratic in x'
x, of
which the
has necessarily two real roots, with a product
ists,
as asserted, a system of three real
lowing,
XYI.
xa +
x'a',
satisfy the equation I.
and
under
I.,
and rectangular
x'a
xa',
this
system
this condition
XV. But
B and A A'
coefiicients
nish,
which
under the form,
is satisfied
p=xa + x'a',
VIII.
= Aaa,
VjSa'
cannot be satisfied otherwise, unless
XIII.
that
Saa =
and
is
do not generally vahence there always ex-
directions, such as the fol-
aa' (or
Va a'),
generally definite
which proves
thej^rs^ part of the Theorem.
The
(3.)
lines a, a'
coincide with the
two
be made by (1.) to turn in their
may
first
directions
..5 = 0,
XVII.
and
XVI.
which
^=Aa,
own
till
plane,
they
will give,
^'
= A'a',
therefore,
XVIII.
0p
= - cp + AaSap + A'a'Sa'p
= (c + A) aSap + (c + A') a'Sa'p +
and thus the scalar equation
XIX.
Sp<pp
II. will
= Cc + A)
(Sap)2
caa'Saa'p
take the form,
(c
+ A')
(Sa'p)^
+ c(Saa'p)a=
const.,
which represents generally a central surface of the second order, with its three
axes in the three directions a, a', aa' of p ; and does not cease to represent such a
and with such axes, when
surface,
pression
XX.
for
Sp0p we
substitute, as in III., this
new ex-
Sp^p
- Cp2 = Sp0p + C((Sap)2 +
(Sa'p)2 + (Saa'p)^)
the second surface being in fact coney die (or having the
same
= C'=
const.
cyclic planes) with the
- Cp, in (pp, disappearing under the sign V.p so that the

first, and the new term,
second part of the Theorem is proved anew.
here on the cases, in which the surfaces XIX.,
(4.) It would be useless to dwell
:
XX. come
to be of revolution, or even to
directions of their axes, or of
But
as
in
be spheres, and when consequently the

partially or even yihoWy indeterminate.
become
p
an example of the reduction of an equation in quaternions to the ybrm
I.,
I.,
700
without
we may take
at first presenting itself under that form,
its
[bOOK
III.
the very simple
equation,
XXI.
which
ptpK=
XXII.
and which
with K not
ijOKrp,
be reduced (comp. 354, (12.)) to
may
V.pVf,oc
XXIX.) by
accordingly satisfied (comp. 373,
is
rections,
XXIII.
Ui - Uk,
Ij
= 0;
Vt
YiK,
the three rectangular
di-
Uk,
of the axes (âbc) of the ellipsoid,
XXIV.
which
T (tp + pc) = /e2 - i\
one of the surfaces of the concyclic system (comp.
is
XXV.
Stpicp
Cp2
282, XIX.
III.),
+ C,
as appears from the transformations 336, XI., &c.

(5.) In applying the theorem thus recently proved anew,
occasions used the expression,
XXVI.
in
which
is
(6.) It
dr
on
havê
= ^dp,
a vector normal to a surface whereof p
is
i/
function
we
several
410, IV.
is
the variable vector, and the
treated as self-conjugate (363).
is,
however, important to remark that, in order to justify the assertion
of this last property, the following expression of integral form,
XXVII.. .JSvdp,
must admit
of being equated to some scalar function of p, such as
being assumed that p itself \s
^/p +
const.,
function, of any determinate form, of a

scalar variable, t.
The self-conjugation of the linear and vector function <p in
XXVI., is the condition of the existence of the integral XXVIL, considered as re-
without
its
a.
presenting such a scalar function (comp. again 363).

(7.) There are indeed several investigations, in which
it is
sufficient to regard
V as denoting some normal vector, of which only the direction is important, and
which may therefore be multiplied by an arbitrary scalar coefficient, constant or
variable, without
tic lines, in
(8.)
any change
And
in the results (comp, the calculations respecting ^'codc-
5, and many others which have already occurred).

there have been other general investigations, such as those regarding
the Section III.
iii.
the lines of curvature on an arbitrary surface, in which dv was treated as a selfconjugate function of dp, while yet (comp. 410, (17.)) the fundamental differential
equation Sj/dj/dp = was not affected by any such multiplication of v by n.

(9.) But there are questions in which a factor of this sort may be introduced,
with advantage for some purposes, while yet it is inconsistent with the self-conjugu'
Hon above mentioned, unless the multiplier n be such as to render the new expression Sravdp (comp.
XXVIL)
an exact
differential of
some scalar function
of p.
(10.) For example, in the theory oi Reciprocal Surfaces (comp. 412, (21.)),
is convenient to employ the system of the three connected equations,
XXVIII.
when
the length of v
Si^p
1,
Sj^dp
= 0,
Spdv =
it
373, L. LI.
determined so as to satisfy the j^rs^ of these equations,

v~i being then the vector perpendicular from the origin on the tangent plane to the
bxit
is
CHAP.
CONDITION OF INTEGRABILITY, FACTOR.
III.]
701
piven but arbitrary surface of which p is the vector, while p"'
is the corresponding
with v for vector, the differential dv Zosc
self- conjugate character, as a linear and vector function of dp
although
for the reciprocal surface
perpendicular
its
generally
retains that character if the scalar function
it
fp be homogeneous,
in the equation
= const, of the original surface, as it is for the case of a central quadric* for
fp
which v = <pp, Av = (pdp, &c., as in former Articles.
In
(1 1.)
the introduction of the
fact,
multiplication of
v by the
XXIX.
we
shall
have
XXX.
and these
But
this
ni/
new
factor
d/p
first
= (Srp)-i
= 2Svdp,
dv
and
= 6dp = n0dp + vSodp,
if
= Strdp,
0dp,
and vector
functions,
XXXI.
6'6.p
pair of conjugate linear
will not be equal generally, because
this last parallelism exists in the case
XXVIII. is equivalent to the

we write (comp. 410, (16.)),
equation
we
= n^dp +
shall not in general
(rSvdp
have
becauise
of homogeneity (10.),
<t
||
v.
we have
then the relations,
XXXII.
if
r be the
2Svp = rfp,
= dSvp = rSvdp,
d .n-i
number which represents the dimension

On the other hand it may happen, that
of /p (supposed to be whole).
the differential equation Svdp =

represents a surface, or rather a set of surfaces, without the expression Svdp being
an exact differential, as in (6.) ; and then there necessarily exists a scalar /acior,
(12.)
or multiplier, n,
(13.) For
XXXIII.
which renders
example the
it
such a
Sypdp = Svdp =
itself is
Sypdp
XXXIV. ..v = Vyp,
with
0,
dv = Vydp = ^dp,
of planes), drawn through a given line y

not an exact differential, and the inê^ra/ XXVII.
represents an arbitrary plane (or a
but the expression
differential.
set
of p, with the present form of v, of which the differenaccordingly a linear function ^dp, which is not conjugate to itself but to
= ^dp.
opposite (comp. 349, (4.)), so that we have here <p'dp
if we
But
the
v
by
(14.)
multiply
factor,
represents no scalar function

tial
its
dv
is
XXXV.
i/-2
=(Vyp)-2, which gives
XXXVI.
dn
= S(Tdp,
XXXVII.
!/
=d
v-i
=-
v-i Vydp
.v-^=6dp =
= 22yVyp,
(T
and therefore Sy<T = 0, Spo- = 2, then the new normal vector nv, or
to have the self-conjugate
differential,
6'dp
j/-', is
found
and accordingly the new expression,
XXXVIII.
is easily
i\iQ
Snvdp
dp
= Sj^->dp = S r-^,
Yyp
seen to be an exact differential, namely (if
plane of y and p makes with
di
with y constant,
Ty =
fixed plane through y
1),
:
that of the angle which
so that,
when v
is
thus
* It was for this reason that the

symbol Tj/ was not interpreted generally as
the origin on the
denoting the reciprocal, P-\ of the length of the perpendicular from
although, in several of those fortangent plane, in the formulse of 410, 412, 414
for the case of
mulae, as in an equation of 409, (3.), that symbol was so interpreted,
:
a central surface of the second order.
702
changed
XXVII.
to nv, the integral in
[bOOK
III.
a,
geometrical signification, which is
then represents the change of this angle,
acquires
often useful in physical applications, since
it
in passing from one position of p to another ; or the angle through which the variable
plane of yp has revolved.
(15.) In fact, the general formula 335, XV. for the differential of the angle of
a quaternion gives, if we write
XXXIX.
= Vyp
the two connected expressions
XL...d^g =
:^~-i
const.,
po
Ty =
const.,
1,
+
XLI...JS;^= A^(Vyp:Vypo);
S^;
which contain the above-stated
result,
and can
be otherwise established.
easily
= 0dp be not self-conju(16.) In general, if the linear and vector function dv

= <idp be formed from it as in (11.), it results from
gate, and if the function d.nv
that sub-article, and from 849, (4.), that we may write,
XLII.
with the
.{<p- f) dp
= 2Vydp,
relation,
XLIII.
2y,
(6
= 2ny +
- 6') dp = 2Vy,dp,
Yv(T;
where y, y, are independent of dp, although they mat/ depend on p itself. If then
the new linear function <idp is to be self-conjugate, so that y< = 0, we must have
XLIV.
which
latter
2ny + Yva =
0,
XLV.
and therefore
very simple equation, not involving either n or
quaternions, of the Condition of Integrahility*

if the vector y be deduced from v as above.
of the
Sy v
c,
is
thus a form, in
Svdp = 0,
(17.) The Bifocal Transformation of Sp(pp, in 360, (2.), has been sufficiently
considered in the present Section (III. iii. 7) ; but it may be useful to remark here,
that the Three
Mixed Transformations
of the
same
scalar function fp, in the
same
whole known theory of the Modular and

Umhilicar Generations of Surfaces of the Second Order.
= 1, t is the vector of an
(18.) Thus, in the formulas of 360, (4.), if we make e
Umhilicar Focus of the surface /p = 1, and ^ is the vector of a point on the Umhili'
series of sub-articles, include virtually the
car Directrix corresponding; vih^ncQ iho, umhilicar focal conic and dirigent cylinder (real or imaginary) can be deduced, as the loci of this point and line.
= 1, in the formulae of 360, (6.), we ob(19.) Again, by making ei and e^ each
tain
Two Modular Transformations
of the equation of the
same surface;
* If the
proposed equation be
Svdp=/)da; + gdy + rdz
we
= 0,
so that
= zP + j Q + kR, where
P= D,3 - Dj,r, Q = Da;r - J)~p,
easily find that
v^-Qp^-jq^-kr"),
2y
R = T),jp- D^j
XLV. becomes therefore here,

ri? = 0, which agrees with known
the condition of integrability
pP +qQi
results.
ci, 3
being
CHAP.
MODULAR AND UMBILICAR GENERATIONS.
III.]
703
Modular Foci, in two distinct planes, aud î, ^3 being vectors of points
upon the Modular Directrices corresponding whence thQ modular focal conies, and
or imaginary), are found by easy eliminations.
cUrigent cylinders (real
vectors of
by assuming that
(20.) Thus,
either
S\(p-^3) = 0,
- ^3) = 0,
S/i (p
XLVI. ..SX(p-^0 = 0,
- ^1) = 0,
XLVII.
S/i (p
or
the equations 360,
XLVIII.
(p
XVI., XVII. may be brought to the forms,
- 1)2 = /nx2 (p - ^0^
XLIX.
(p
3)2
= m^? (p - ^3)2,
with the values,

L.
mi-
=1
and
^,
LI.
^132
=1
Cl
in
which
co,
ci,
C3 are
?
;
C3
the three roots of a certain cubic (iV/= 0), or the inverse
semiaxes (real or imaginary) of the surface, arranged in

squares of the three scalar
order
ascending
(357, IX., XX. ; 405, (6.), &c.): and mj, mz are the
algebraically
two (real or imaginary) Moduli, or represent the modular ratios, in the two modes
of Modular Generation* corresponding.
(21.) It is obvious that an equation of the form,
LIL
represents a central quadric, if
Mac
C = const,
T<pp=
^p be any linearf and vector function of
p, of the
Cullagh's rule of modular generation, which includes both those modes,

page 437 of the Lectures by an equation of the form,
wa.s expressed in
T(p-a)=TV.yV|3p;
which the origin is on a directrix, /3 is the vector of another point of that right
line, a is the vector of the corresponding focus, y is perpendicular to a directive
in
(that
is,
generally, to a cyclic) plane, p
+ S/3y
and
is
the vector of any point p of the surface,
the constant modular ratio, of the distance
is
ap
of p from the focus, to
same point p from the directrix ob, measured parallel to the diThe new forms (360), above referred to, are however much better
the distance of the

rective plane.
adapted to the working out of the various consequences of the construction ; but it
cannot be necessary, at this stage, to enter into any details of the quaternion transformations
still less need we here pause to give references on a subject so interest:
but by this time so well
ing,
known
bilicar generations of surfaces of the
to geometers, as that of the modular and umsecond order. But it may just be noted, in order
to facilitate the applications of the formulae L.

for all the central quadrics,
then the roots
ci, co, C3
a^
>b^ >
and
LI., that if
whether 62 and
c^
we
and
write, as usual,
be positive or negative,
coincide, for the ellipsoid, with a"2, 6-2, c-2
sheeted hyperholoid, with c^, a"^, J

h'^, c"2, a-2^
c^,
for the single-
for the dowJZe- sheeted hyperboloid
with
(comp. page 651).
t In page 664 the notation,

dp
= 2Sa/dp =
409, IV.
2S^p6p,
was employed for an arbitrary surface; but with the understanding that this function
<pp
(comp. 363) was generally non-linear.
It
may
be better, however, as a
704
kind considered in the Section III.
more attention
quires a little
LIII.
ii.
6,
[bOOK
whether self-conjugate or not; but

an equation of this other form ^
III.
it
re-
to perceive, that
T(p-V./3Vya) = T(a-V.yV)3p),
The
represents such a surface, whatever the three vector constants a, /3, y may be.
discussion of this last /orm would present some circumstances of interest, and might
be considered to supply a new mode of generation^ on which however we cannot
enter here.
(22.)
The
surfaces of the second order, considered hitherto in the present

Section,
But if, retaining the significations of ^, /, and F^
had the origin for centre.

we compare the two equations,
have
all
LIV.
we
shall see
/(p -
/c)
and
C,
LV.
/p
- 2Sp = C,
(by 362, &c.) that the constants are connected by the two
LVI. ..c
= ^/c,
relations,
C'=C-/k= C-Sk:=C-F;
80 that the equation,
LVII.
is
an
/p
2S6P
=/(p - <p-h) - Ft,
identity.
(23.) If then
we have
still
dimension,
we meet an
/p= Sp0p =
we
shall
know
equation of the form LV., in which (as has been usual)

a scalar and homogeneous function of p, of the second
that it represents generally a surface of that order, with
the expression (comp. 847, IX., &c.),
LVIII.
(24.) It
may
= 0-' =
m-ii//
Vector of Centre.
LIX. ..m=0,
Ti//c>0,
and then the centre may be said to be at an infinite distance, but in a

direction : and the surface becomes a Paraboloid, elliptic or hyperbolic, accord-
exist together
definite
happen, however, that the two relations,
ing to conditions which are easy consequences from what has been already shown.
(25.) On the other hand it may happen that the two equations,
LX. ..m=0,
i//f
= 0,
and then the vector k of the centre
acquires, by LVIII., an
indeterminate value, and the surface becomes a Cylinder, as has been already suffi-
are satisfied together;
ciently exemplified.
(26.) It would be tedious to dwell here on such details
but
may
it
be worth
= (pp, except for central quadrics ; and to confine

general rule, to avoid writing v
ourselves to the notation dv = ^dp, as in some recent and several earlier sub-articles,
when we
wish, for the sake of association with other investigations
treat the function
as linear (or distributive)
and
results, to
because we shall thus be at Hberty
to treat the surface as general, notwithstanding this property of
<p.
As
regards
the methods of generating a quadric, it may be worth while to look back at the Note
to page 649, respecting the Six Generations of the Ellipsoid, which were given by
the writer in the Lectures, with suggestions of a few others, as interpretations of
quaternion equations.
CHAP.
III.]
CUBIC CONE, SCREW SURFACE, SKEW CENTRE.
705
while to observe, that the general equation of a Surface of the Third Degree maybe thus written
LXI.
C and
y being any
scalar
+ Sptpp + Syp + C=0;
Sqpq'pq'p
and vector constants
any
<pp
y,
and self-con-
linear, vector,
jugate function ; and q, q, q" any three constant quaternions

the variable vector of the surface.
while p
as usual,
is,
(27.) In fact, besides the one scalar constant, C, three are included in the vector
in the function
(comp. 358) and of the ten which remain to be
and six others
introduced, for the expression of a scalar

degree, the three versors Ug-, Uj',
T.qq'q"
U9"
And
for the
LXII.
(29.) If then
we
we
Sqpq'pq"p
0,
q", represents the general Cone of the Third

vertex at the origin of vectors.
combine this last equation with that of a secant plane, such as
5,
Degree, or Cubic Cone, which has
the third
same reason the monomial equation,
with the same significations of
p, of
supply nine (comp. 312), and the tensor
the tenth.
is
(28.)
and homogeneous function of
q',
its
shall get a quaternion expression for a
Plane Cubic, or plane curve

of the third degree : and if we combine it with the equation p^ -f 1 = of the unitsphere, we shall obtain a corresponding expression for a Spherical Cubic,* or for a
curvê upon a spheric surface, which is cut by an arbitrary great circle in three
pairs
Sf p
0,
of opposite points, real or imaginary.
(30.) Finally, as an example of sections of surfaces, represented by transcendental equations, let us consider the Screw Surface, or Helicoid,f of which the vec-
may be thus written (comp. the

LXIII.
with
p = c (a; + ) a + ya^y,
tor equation
sub-arts, to
Ta =
1,
314)
a being the unit axis, while (3, y are two other constant
constants, and x, y two variable scalar;?.
(31.) Cutting this surface
LXIV.
we
easily find that the scalar
Section
may
LXVI.
by the plane
be thus written
.
and
y>0;
two
vectors, a, c
scalar
of j8y, or supposing that
= Sy/3p = /32Sap - Sa/3S/3p,
y = Ya{3,
LXV.
and writing
and vector equations of what we may
= 6Sa/3,
call the Screxo
LXVII.
6(a;+)=yS.a-';
(32.) Derivating these with respect to x,
p=y(yS.a*-
and eliminating
/3
/3S.a*-i).
and
g',
we
arrive
at the equation,
LXVIII.
Compare
(a;
+ a) p' +
the Kote to page 43
2y,
if
LXIX.
26z
= Try'
see also the theorem in that page,
new mode
which con-
o( generation of cubic curves in a given plane

an easy modification, of the corresponding curves upon a sphere.
The condition
t Already mentioned in pages 383, 502, 514, 557.
tains perhaps a
or,
y>0
by
an-
swers to the supposition that, in the generation of the surface, the perpendiculars
from a given helix on the axis of the cylinder are not prolonged beyond that axis.
4 X
706
but zy in LXVIII.
tion at the point
is
the vector of the point, say G, in which the tangent to the secnamely the line in the plane
we see then, by LXIX., that

is perpendicular to the axis a
of intersection depends only on the constant, b, and on the variable, y,
:
being independent of the constant,
To
(33.)
III.
y), or p, intersects the given line y,
(ar,
of that section which

this point
[bOOK
a,
and of the
variable, x.
interpret this result of calculation,
which might have been otherwise
found with the help of the expression 372, XII. (with /3 changed to y) for the normal V to a screw- surface, we may observe, first, that the equation LXVII., which
may
be written as follows,
LXX.
would represent an
ellipse, if
section of the right cylinder
LXXI.
and gives
p=yV.a+'^,
the coefficient
LXXI. by
TVap = yTy,
y were treated as constant
the plane
LXIV.
; namely, the
the vector semiaxes (ma-
jor and minor) of this ellipse being y/3 and yy (comp. 314, (2.)).
(34.) By assigning a new value to the constant <?, we pass to a
new screw
sur-
face (30.), which differs only in position from the former, and may be conceived to
be formed from it by sliding along the axis a while the value of x, corresponding
;
by LXVI., and thus we shall have a new screw section (31.),

but the tangent to the section at
ellipse (33.) in a new point Q
to a given y, will vary
which
will cross the
this point will intersect
by (32.) the minor axis of the
ellipse in the
same point o
as
before.
if
(35.) We shall thus have a Figure* such as the following (Fig. 85);
F be a focus of the eUipse bc, and g (as
in which
above) the point of convergence of the tangents to the screw sections at the points p, q,
&c., of that ellipse, it is easy to prove, by
pursuing the same analysis a little farther,
1st,
that the angle {g), subtended at this

by the minor semiaxis oc, which is
focus F
also a radius (r) of the cylinder
LXXI.,
is
equal to the inclination of the axis (a) of

that cylinder to the plane of the ellipse, as may indeed be inferred from elementarj'
principles
same
and Ilnd, what
focus (f)
by the
is less
obvious, that the other angle (h), subtended at the
interval og, or
present constniction, in which
by what may be
<
called (with reference to the
or that the angles made by

with a are either both acute, or both
obtuse) the Depression (s) of the Skew
Centre (g), is equal to the inclination of the same axis
(a) to the helix on the same
Bxp and
it is
supposed that 6
0,
/3
which is obtained (comp. 314, (10.)) by treating y as constant,

equation LXIII. of the Screw Surface.
cylinder,
* Those who are

acquainted, even
slightly,
in the
with the theory of Oblique Arches
(or
skew bridges), will at once see that this Figure 85 may be taken as representing rudely
such an arch and it will be found that the construction above deduced agrees with
:
the celebrated Rule of the Focal Excentricity, discovered

practically by the late Mr.
Buck. This application of Quaternions was alluded to, in page 620 of the Lectures.
CHAP.
STATICS OF A RIGID BODY.
III.]
Section
8.
70T
On a few Specimens of Physical Application of
Quaternions, with some Concluding Remarks.

416. It remains to give, according to promise (368), before concluding this work, some examples* oi physical applications of the
we
present Calculus: and as a first specimen,
shall take the Statics
of a Rigid Body.
rtn be n Vectors of Application, and let ]8i, . . /3,j be n correspond(1.) Let a\,
ing Vectors of Force, in the sense that n forces are applied at the points Ai, . A of
a.
free but rigid system, and are represented as usual by so many right lines from
.
which
ob are equal, though drawn from a

y (= oc) be the vector of an arbitrary point c of space.
Then the Equation^ of Equilibrium of the system or body, imder the action of these
n applied forces, may be thus written
those points, to
common
origin
and
lines the vectors OBi,
let
I.
(2.)
or r., into the
two vector equations

II.
which
or thus,
2V(a-y)/3=0;
The supposed arbitrariness
S/3
I'.
VySj3= SVoiS.
(1.) of y enables us to break
up the formula
I,
III.
0;
SVa/3 = 0;
easy to assign, as follows, the physical signification.

II. expresses that if the forces, which are applied at the points
of the body, were all transported to the origin o, their statical resultant, or
Ai
vector sum, would be zero.
of each of
(3.)
it is
The equation
(4.)
in the
The equation
usual
III. expresses that the resultant of all the couples, produced
way by such a
transference of the applied forces to the assumed origin,
is null.
(5.)
And
the equation
I.,
which as above includes both
that if all the given forces be transported to

arising will balance each other
which
is
II.
any common point
and
III.,
expresses
c, the couples
hence
a sufficient condition of equilibrium of the
system.
(6.)
When we
have only the

IV.
relation,
.
S(S/3.2Va/3) =
0,
without S/3 vanishing, the applied forces have then an Unique Resultant = S/3,
variable vecacting along the line of which I. or I', is the equation, with y for its
tor.
The reader may compare the remarks on hydrostatic
pressure, in pages 434,
435.
We
^^
I. or I'.,
t
say here,
equation :*' because the si7igle quaternion formula,
contains virtually the six usual scalar equations, or conditions, of the equilibrium at
present considered.
708
(7.)
And
[bOOK
III.
the physical interpretation of this condition IV. is, that when the
and (4.) the resultant force is in the plane of
forces are transported to o, as in (3.)
the resultant couple.

(8.) When the equation
II.,
but not
III., is satisfied, the applied forces
= ^YaPt
themselves into One Couple^ of which the Axis
whatever
compound
maybe
the posi-
tion of the origin.

so to place the
(9.) When neither II. nor III. is satisfied, we may still propose
auxiliary point c, that when the given forces are transferred to it, as in (5.), the
resultant force Sj3 may have the direction of the axis 2V(a-'y)/3 of the resultant
couple, or else the opposite of that direction
so that, in each case, the condition,*
v...v--i|zzM.,
shall be satisfied
by a suitable limitation of the auxiliary vector
y.
(10.) This last equation V. represents therefore the Central Axis of the given
system of applied forces, with y for the variable vector of that right line or the axit
:
of the screw-motion
which those
forces tend to produce,
when they are not in balance,
as in (1.), and neither tend to produce translation alone, as in (6.), nor rotation
alone, as in (8.).
(11.) In general,
5 be an auxiliary quaternion, such that
if
VI.
its
vector part, Yq,
.92/3
= SVa/3,
equal by (V.) to the Vector-Perpendicular, let fall from the

; while its scalar part, S5, is easily proved to be the quobe called the Central Moment, divided by the Total Force : so
is
origin on the central axis

tient, of
that Vgexists
what may
when
the central axis passes through the origin, and S5
= when there
an unique resultant.
(12.) "WTien the total force
does not vanish,
2)/3
let
be a new auxiliary qua-
ternion, such that
VIII.
with
= SQ = S^,
and
IX.
= oc = VQ,
and vector parts
then cS/8 represents, both in quantity and in direction,

;
the Axis of the Central Couple (9.), and y is the vector of a point c which is on the
for its scalar
central axis (10.), considered as a right line having situation in space: while the
position of this point on this line depends only on the given system of applied
and does not vary with the assumed origin o.
Under the same
(13.)
X.
XII.
2Va/3 = c2/3 + VySi3;
* The
equation V.
when y
is
conditions,
2a|3 = (c + y) S/3
may
also
v.
we have
XI.
XIII.
.
T2a/3 = (c2 - y2)JT2/3

(2VaiS)2
= 0^(2)3)2+ (Vy 2/3)2;
be obtained from the condition,
T2V(a -
y)|3
treated as the only variable vector
of the Central Moment.
forces,
=
;
a minimum,
which answers
to a
known
propt
CHAP.
GENERAL EQUATION OF DYNAMICS.
III.]
known
whereof XII. contains the

obtained
by transferring
law^ according to
which the axis of the couple
an assumed point
all the forces to
709
o,
(4.),
varies generally in
in direction with the position of that point

while XIII. expresses the
quantity and
known corollary from that laAv, in virtue of which the quantity alone, or the energy
(TSVa^) of the couple here considered, is the same for all the points o of any one
:
which has the central axis of the system
right cylinder,
for its axis
of revolution.
(14.) If we agree to call the quaternion product pa.aa' the quaternion momenty
or simply the Moment, of the applied force aa' at A, with respect to the Point P, the
quaternion sum âfS in X. may then be said to be the Total Moment of the given
system of forces, with respect to the assumed origin o and the formula XI. expresses that the tensor of this sum, or what may be called the quantity of this total
;
constant for all points o which are situated on any one spheric surface,
its centre : being also a minimum when o is
is
moment,
with the point c determined in (12.) for
placed at that point c itself, and being then equal to what has been already called
moment, or the energy of the central couple.
the central
(15.) For these and other reasons,
appears not improper to call generally the
it
point c, above determined, the Central Point, or simply the Centre, of the given
system of applied forces, when the total force does not vanish and accordingly in
the particular but important case, when all those forces are parallel, without their
sum being zero, so that we may write,
;
XIV.
= 6i/3,
/3i
pn =
bnfi,
T2/3 >
0,
and the vector y becomes (comp. Art. 97 on hary-
the scalar c in (12.) vanishes,

centres),
^_
XV.
oc = y
+
= feiai
7
so that the point c, thus determined, is
what
coincides with
I.,
may
XVI.
(17.)
Total
hnttn
;
:2ba
independent of the common direction
/3,
and
which have been already expressed by
(1.),
Moment =
what may perhaps be
Any
also be included in this other quaternion equation,
of which the value is independent

represents
usually called the Centre of Parallel Forces.
is
The conditions of equilibrium
(16.)
the formula
infinitely
its
a scalar constant,
and which, with
its
sign changed
called the Total Tension of the system.
small change, in
the alteration pf each of
'S.a^
of the origin
i\\Q
position of a rigid body,
is
equivalent to
vectors a to another of the form,
XVII.
+ daâ+ei- Yia,
being two arbitraiy but infinitesimal vectors, which do not vary in the pasand thus the conditions ofequilibrium
sage from one point a of the body to another
(1.) may be expressed by this other formula,
6
and
XVIII.
which contains,
2S/3â=0,
for the case here considered, the Principle
of Virtual
Velocities,
and
admits of being extended easily to other cases of Statics.
417.
The general Equation of Dynamics may be thus

I..
2mS(D;-|)â =
0,
written,
710
[bOOK
III.
with significations of the symbols which will soon be stated; but as

only propose (416) to give here some specimens of physical application, we shall aim chiefly, in the following sub-articles, at the de-
we
duction of a few formulae and theorems, respecting Axes and Moments of Inertia^ and subjects therewith connected.
(1.)
In the formula
of the system
Sa
I.,
is
the vector of position, at the time
any variation of that
is
of
t,
the mutual connexions between the parts of that system; the vector
a moving
| an accelerating
force, or
and S are marks, as
denoted by
which acts on the element
to all
m^
represents
of
mass
and taking the scalar and the summation

the elements, and is generally equivalent to a triple in-
usual, of derivating
extends
an addition of
tegration, or to
force,
an element
vector, geometrically compatible with
triple integrals in space.
And
the formula
is
ob-
tained (comp. 416, (17.)), by a combination of D'Alembert's principle with the prin-
which
ciple of virtual velocities,
is
analogous to that employed in the Mecanique
Analytique by Lagrange.
(2.) For the case of a/rec but rigid body, we may substitute for â the expression i-\-Yia, assigned by 416, XVII.; and then, on account of the arbitrariness
of the
two
ing.
and
infinitesimal vectors t
t,
the formula
IL.. 2m(D2a-O = 0;
HI.
I.
breaks up into the two follow-
^mYa(T>tâ-K)^0',
which correspond to the two statical equations 416, IL and III., and contain respectively the law of motion of the centre of gravity, and the law of description of
areas.
(3.) If the body have a./?a;e(f jaoinf, which we may take for the origin o, we
eliminate the reaction at that point, by attending only to the equation III.
and
;
may
theu express the connexions between the elements
IV.
D(a
= Yia,
whence
V.
by
the formula,
D2a = lYia - VaD^t
being the Vector- Axis of instantaneous Rotation of the body, in the sense that its
versor Ut represents the direction of the axis^ and that its tensor Tt represents the
t
angular
(4.)
velocity, of
By
such rotation at the time
t.
v., the equation III. becomes,
VL
2maVaDi=2m(VtaSm-VaO;
and other easy combinations give the laws of areas and living
(5.)
VII.
VIII.
When
'2maDt<t
- 2?V a^d< = = a
J
y
force,
under the form
constant vector
i2m(D<a)2 - 2iSjta^d< = c=
a constant scalar.
the applied forces vanish, or balance each other, or more general!
when they compound themselves
into
a single force acting at the fixed
point, so that
in each case the condition
IX..
is satisfied,
2TOVa^=0
the equations (4.) are simplified; and
and self-conjugate function

X.
if
we introduce a
0, such that
01
'S.maXai
= i2ma2 - 2maSat,
linear. 'vect
CHAP.
MOMENTS AND AXES
III.]
and write A^
XL
- 2c,
for
OF INERTIA.
711
they take the forms,
.>D<i4-Viî = 0;
XII.
^t
+ y = 0;
XIII.
St0t= ^2;
real constants, of the vector and scalar kinds, connected with each
y and h being two
other and with i by the relation,
XIV.
It
may be added that y
is
Sty
-h
now the
XV.
also
<pT)ti
= Vty.
sum of the doubled ureal velocities of all the

by the mass m of that element, and each re-
vector
elements of the body, multiplied each
by a right line oDta perpendicular to the plane of the area described

round the fixed point o in the time dt and that h^ is the living force, or vis viva of
the body, namely the positive sum of all the products obtained by multiplying each
element m by the square of its linear velocity, regarded as a scalar (TD<a).
presented
(6.)
"When
is
regarded as a variable vector, the equation XIII. represents an

in the body, but moveable with it ; and the equation XIV.
which isjixed
ellipsoid,
represents a ân^'cni />/ane to this ellipsoid,
which plane isjixed
in space, but changes

thus the motion of that body may
shown by Poinsot, to be performed by the rolling
generally be conceived, as was
{without gliding) of an ellipsoid upon a plane ; the former carrying the body with it,
And
in general its position relatively to the body.
while
its
centre
rQxnBxns fixed
and the semidiameter
(t)
of contact being the vec-
tor-axis (3.) o{ instantaneous rotation.

(7.) The elHpsoid XIII. may be called, perhaps, the Ellipsoid of Living Force^
on account of the signification (5.) of the constant h^ in its equation ; and the fixed
plane
XIV., on which
it
rolls,
is
parallel to
what may be
called the
Plane of
Areas (Sty = 0) no use whatever having hitherto been made, in this investigation,
But if we here admit the usual definition of such
of any axes or moments of inertia.
:
we may say that the Moment of Inertia of the body, with respect to any
through the fixed point, is equal to the living force h'^ divided by the square*
of the semidiameter Ti of the ellipsoid XIII. because this moment is,
a moment,
axis
XVI.
(8.)
2m(TVaUt)2=t-22m(Via)2=-Si-i^t = /t2Tt-2.
The equations XII. and XIII.
XVII.
= y2Sl0l - A2(0i)2 =
give,
St^^
if
XVIII.
V = y0t -
A202,
and this equation XVII. represents a cone of the second degree, fixed in the body
(comp. (6.)), but moveable with it, of which the axis t is always a side, and to which
the normal, at any point of that side, has the direction of the line v.
But it follows
* Hence
it
may easily be inferred,
with the help of the general construction ofan
by Figure 53
in page 226, that for any solid

body ,
can always be found (indeed in more ways than
one) two other points, B and c, which are W^QVfi&e fixed in the body, and are such that
the square-root of the moment of inertia, round any axis AD, is geometrically constructed by the line bd, if the point D be determined on the axis, by the condition that
ellipsoid (217, (6.)), illustrated
and any given point A
A and D
thereof, there
shall be equally distant
produced,
was given
from
c.
This theorem, with some others here re-
in the Abstract of a Paper read before the
Eoyal
Irish
Academy
on the lOth of January, 1848, and was published in the Proceedings of that date.
712
[bOOK
from XI., or from XII. XV., and from the properties of the function 0, that D<t ia
perpendicular to both ^t and ^î, and therefore also by XVIII. to v the cone XVII.
;
is
therefore touched, along the side
by that other
We
of the instantaneous axis of rotation.
which
cone,
are then led,
by
is
the locus in space
this simple quaternion
of the body, which

of one cone on another.
analysis, to a second representation of the motion
by Poinsot
(9.) To treat
posed
namely, as the rolling
briefly
by quaternions some of Mac CuUagh's
also
was pi4B^
on
results
this sub-
be noted that the line y, though îcecf in space, describes in the body
cone of the second degree, of which the equation is, by what precedes,
ject, it
may
XIX.
AV =
.i7^Sy^-y
XX.
if
0,
..i7
= Ty,
XXI.
or
y2
+^2 =
write y = oc, the point c is indeed fixed in space, but describes a

sphero-conic in the body, which is part of the common intersection of the cone
we
while, if
XIX., the sphere XXI., and the reciprocal
XXII.
(10.) Also, the normal to the
direction of ^^^-ly ^ }^2y^ or of t
tion be
drawn through the
Sy0-'y =
new cone
ellipsoid
y, has the
this direc-
be the side of contact of the plane
it will
(7.),
any point of the side

and if a line in
XIV.)
A^y-i (comp.
fixed point o,
A3..
(9.), at
with the cone of normals at o

procal) cone rolls on that plane of areas.
of areas
(comp. XIII.),
XIX.
to the cone
which
last (or reci-
(11.) As regards the Axes of Inertia, it may be sufficient here to observe that
the body revolve round a permanent axis, and with a constant velocity, the vector axis I is constant
and must therefore satisfy the equation,
if
XXIII.
Vi0i
0,
because
XXIV.
Dî
has therefore in general (comp. 415) one or other of Three Real and Rectanptdar
Directions, determined by the condition XXIII.: namely, those of the Axes of
Figure of either of the two Reciprocal Ellipsoids, XIII. XXII.
it
(12.)
And
the Three Piincipal Moments, say A, B, C, corresponding to those

by XVI. the three scalar values of i~^<pi so that the
three principal axes, are
symbolical cubic (350) in
may
XXV.
Forming then
(13.)
III.
by
ii.
6,
we
be thus written,
+ J) (0 + ^) (0 + C)
(0
by the general method of the Section

moments A, B, C, are the three roots (always real,
the algebraic and cubic equation,
XXVI.
this symbolical cubic
^3 _ 22^2 + (4 4-
'2)
in which, n^, n'2, n''2 are three positive scalars,

.
0.
find that the three
this analysis) of
XXVII.
2 =
- 2ma2;
and the combination
n'^n'^ -
n'2
A - (n2n'2 _ n"2) =
= - ^mm' (Yaa'Y
n'"^ is
n"2= 2:mm'm"(Saa'a")2;
another positive scalar, of which the value may
be thus expressed,
XXVIII.
+
if a, a',
a",
<S:c.
namely,
ABC=nhi^ - "2 = 2i2m'a2 (Vaa')2
22w27n'm" (Taa'Ta'a"Ta"a
+ Saa'Sa'a"Sa"a),
be the vectors of the mass-elements m, m', m", &c.
CHAP.
SYSTEiM OF ATTRACTING BODIES.
III.]
And
(14.)
XXIX.
u
-ABCr^
XXX.
round one of
B+C)(p+BC+CA + ^B,
^-iQ
a body, with a fixed point, &c., bepin to revolve

three principal axes of
inertia, it will continue to revolve round that
by XV., &c., that
therefore,
gives this other symbolical result,
=f'+(iA +
follows that
and
XXV.
713
its
if
with an unchanged velocity of rotation.

(15.) It has hitherto been supposed, that
axis,
the
all
moments
of inertia are referred
axes passing through one point o of the body but it is easy to remove this reFor example, if we denote the moment XVI. by /o, and if / be the corstriction.
to
responding
moment
which the vector
an axis parallel
for
to
t,
but dfewn through a new point Q, of
w, then
is
XXXI.
4 = C^^m(Yi(a - w))2
= /o 4- 22m. S (wi-iVtK) pZSm,

KSm = Sma, and XXXIII. /> = TVwUt,
-j-
XXXII.
if
so that K is the vector of the centre
the distance between the

(16.) If then
two
we suppose
that
is, if
the axis
very simple relation,
which agrees with
418. As
we propose
w,
m\
m",
vectors are
and p
is
parallel axes.
that the condition
XXXIV.
is satisfied,
of inertia (or of gravity) of the body,
..Vifc
pass through the centre of inertia,
we
shall
have the
XXXV. ../, = /o+/)22m;
known
results.
a third specimen of physical applications of quaternions,

to consider briefly the motions of a System of Bodies ^
.
regarded as free material points, of which the variable

and which are supposed to attract each other
a^\
a, a',
according to the
law of the inverse square: the fundamental
for-
mula employed being the following,

I.
P thus
2mSD/aa + aP = 0,
if
II..
P=S
'/''"'
{a-
a')
denoting the Potential (or force-function) of the system, and

â',
being infinitesimal, but otherwise arbi-
the variations ca,
trary.
(1.)
To deduce the formula
equation 417,
1,
of dynamics,
I.,
with the signification
we have
first,
II. of
P, from the general
two
bodies, the following
for the case of
expressions for the accelerating forces,
in...^=^_,
(a -a')r
r=
r-r^'
- a)r
(a
if
= T(a-a');
714
III.
[boOK
whence follows the transformation,*

IV.
- SCmKda+m'KSa) =
-^^
a- a
-= d
a result easily extended, as above. If the law of attraction were supposed different,
there would be no difficulty in modifying the expression for the potential accordingly.
(2.) In general, when a scalar, f (as here P), is a function of one or more vec_
tors, a, a',. ,. its variation (or differential) can be expressed as a linear and scalar
function of their variations (or differentials), of the form Sy3â + S/3'â' +
(or
.
SS^da)
which
in
. .
j3'
j3,
are certain
shall find
and are them-
finite vectors,
derived from the given scalar function/. And

.,
convenient to extend the Notation-]- of Derivatives, so as to denote
selves generally /unctions of a, a',
we
new and
it
these derived vectors
&c.,
/3, /3',
by the symbols, D^/,
'Da'f,
In this manner we
&c.
shall be able to write,
v..
and the
I.
differential equations of
the forms
more
^P=SS(DP.â);
mWa + DP=
m'DM' + Da'P =
0,
Wa =
VIII.
SmDfa = f3;
X.
and
which
(3,
force of the
by
0,
&c.
y are constant vectors, jH
system (comp. 417,
first integrals,
in I. to one
common but
to Yia, Sa' to Yia', &c.
force,
result
imme-
is
a constant scalar, and 2
P is
the living
(5.)).
(2.)) of deducing the three equations, of which these
the following.
is
and of living
IX.
SmVaDâ = y;
r=-|Sm(D<a)3 = P+ir;
One mode (comp. 417,
(4.)
da
will take
^^^_^^ + ___^__) + ..,&,
(3.) The laws of the centre of gravity, of areas,

diately from these equations, under the forms,
are the
fully,
vn.
in
VI.
or
motion of the bodies m, m', m",
To
obtain VIII., change every variation
arbitrary infinitesimal vector,
For IX., change da
c.
being another arbitrary and infinitesimal vector. Finally,
to arrive at X., change variations to differentials (â to da, &c.), and integrate
once, as for the
(5.)
two former equations, with respect
The formula
I.
tion on the variations da,
XI.
by means
S (Df2a â) = D^S (D<a

.
XII.
may
t.
..P=
da')
- ^d. (Ptay,
definite integral
(comp, X.),
i'(P+ T)dt.
j;(p-fP),
not be useless here to compare the expression in page 417, for the dif-
ferential of a proximity.
t In
this
restric-
of the general transformation,
combined with the introduction of the following
* It
to the time
admits of being integrated by parts, without any
extended notation, such a formula as d/p
v=J^Dpfp.
2SvdjO would give,
CHAP.
INTERMEDIATE AND FINAL INTEGRALS.
III.]
In
(6.)
we denote by ao, a'o,

when ^ = 0, and by Doa,
fact, if
or their values
Dto',
we shall thus
have, as a
XIIL
in
which no variation
the initial values of the vectors a,
a',
the corresponding values of D<a,

integral of the equation I., the formula,
first
SmS (Pta
7lo
Doa',
da - Doa Sa^) + SF=0;

and which conducts to important conse<,
St is assigned to
quences.
To draw from
(7.)
some
it
of these,
we may
observe that
if
the masses m, m',
VL
be treated as constant and known, the complete integrals of the equations

or
VIL must be conceived to give what may be called lheâZ vectors oi position a,
of velocity Dâ, D<a', . . in terms of the initial vectors ao, a'o, . . Doa,
. . and
',
and of the
time, t
whence, conversely, we
may
'velocity to be expressible as functions of the initial
and
conceive the initial vectors
and
final vectors of position,
In this way, then, we are led to consider P, T, and

as being scalar
ictions (whether we are or are not prepared to express them as such), of a, a',
and thus, by (2.), the recent formula XIII. breaks up into the two
o'o, . . and t
'the time.
lowing systems of equations
XIV.
X V.
ad
mBta + T)aF= 0, m'Dâ' + 'DjF= 0, &c.

- m'Doa' +
- wDoa +
=
=
Ba^F 0,
Da'^F 0, &c.
.
"whereof the ybrwier may be said to be intermediate integrals, and the latter to be
final integrals, of the differential equations of motion of the system, which are included in the formula
I.
(8.) In fact, the equations
Dtâ,
XIV. do not
involve the final vectors of acceleration

VI. or VII. had done; and the equations XV.
the dependence of the â? vectors of position a,
on
as the difierential equations
express, at least theoretically,
the time,
t,
and on the
initial vectors of position ao,
(7.) the complete integrals
ought
and
of velocity
Doa,
. .
as
by
And on
to do.
tant properties, the function here denoted
by
account of these and other impormay be called the Principal* Ftmc-
Hon of Motion of the System.

. . and
be given, that is, if we consider the
Doa,
(9.) If the initial vectors ao,
actual progress in space of the mutually attracting system of masses m, . from one
set of positions to another, then the function
depends upon the time alone ; and
.
by
its definition
spect to
is
XII.,
its
rate or velocity of increase, or its total derivative with re-
thus expressed,
XVI.
(10.)
But we may
same definite integralF, when

which involves
of position a,
ao,
. .
D<F=P+
T.
the partial derivative, say (D^F), of the

initial vectors
regarded (7.) as a function of the final and
inquire what
is
also the time explicitly,
and
is
now
to be deri-
were constant :
vated with respect only to that variable t, as if the final vectors a,
whereas in fact those vectors alter with the time, in the course of any actual mo.
tions of the system.
"
On
* This function was in fact so called, in two Essays by the present writer,
a General Method in Dynamics," published in the Philosophical Transactions (Lon1835 ; although of course coordinates, and not ^modon), for the years 1834 and
the latter not having been discovered until 1843
were then
ternions,
and the notation
employed,
for scalar,
S, since adopted
was then used instead of F.
716
(11.) For this purpose,

vative BtF,
which
it is
and X.),
XVII.
observe that the part of the total derimentioned changes of a, .. is (by XIV.
2S(DaF.D<a) = 2T;
last
and therefore (by XVI. and X.), that the remaining part must
XVIII.
III
sufficient to
from the
arises
[bOOK
(DiF) =
be,
P-T=-H.
(12.) The complete variation of the function

as well as a, . and ao, is treated as varying,
Fis
XIIL), when
therefore (comp.
XIX.
dF= - Hdt -
-EmSDtada + ^mSDoadao.
X. XIV. XV., it is easy to infer

(13.) And hence, with the help of the equations
must satisfy the two following Partial Differential
that the principal function
Equations in Quaternions
XX.
XXI.
in
which Pq denotes the

(14.) If
initial
(Dii?')-^Sm-i(DaF)2
(D,F)-^Sx-i(D^F)2
= P;
= Po;
value of the potential P.
we write
XXII..
so that Frepresents
what
r=('22cl^
called the Action, or the accumulated living force, of
is
the system during the time t, then by X. and XII. the two definite integrals
V are connected by the very simple relation,
XXIII.
Fand
V=F + tH;
whence by XIX. the complete variation of F, considered as a function of the final

and initial vectors of position, and of the constant
of living force, which does not
explicitly involve the time,
XXIV.
may
be thus expressed,
dV = tdH- ^mSDtada+m8Doadao.
(15.) The partial derivatives of this

useful than F, and may be called,
more
teristic*
new
function V, which
by way
some purposes
the Charac-
it,
Function of the motion of the system, are therefore,
XXV.. .DF=-mD,a,
XXVI.
The intermediate
&c.;
D(,F= + TnDoa, &c.
and XXVII.
(16.)
is for
of distinction from
.T>^V=t.
integrals (7.) of the differential equations of motion,
which were before expressed by the formulae XIV., may now, somewhat less simply
be regarded as the result of the elimination of ff between the formulae XXV. XXVII.
and the ^naZ integrals
VI. or VII., which were expressed by XV.,

are now to be obtained by eliminating the same constant ^between the recent equations
of those equations
XXVI. XXVII.
* The
Action, V, was in fact so called, in the two Essays mentioned in (he pr*'The properties of this Characteristic Function had been perceived by
ceding Note.
the writer, before those of that which he
Function, as above.
came afterwards
to
call the
Principal
CHAP.
PRINCIPAL AND CHARACTERISTIC FUNCTIONS. 717
III.]
The Characteristic Function^
(17.)
F, is obliged
(comp. (13.)) to satisfy the two
following partial differential equations,
XXVIII.
XXIX.
I2m-i (Da Vy + P+H=0iSm-i (D F)2 +Po+H=0;

.
each of these
vanishes, like F, when t = 0, at which epoch a = Oo, a' = a'o, &c.
two functions, F and F, depends symmetrically on the initial and final vectors of poand each does so, only by depending on the mutual configuration of all those
sition
;
and
initial
final positions.
(18.) It follows (comp. (4.), see also 416, (17.), and 417, (2.)), that the funcmust satisfy the two conditions,
tion
XXX.
S(DaF+DF) = 0;
XXXI.
2V(aDaF + aoDF) =
which accordingly are forms, by XIV. XV., of the equations VIII. and IX., and
therefore are expressions for the law of motion of the centre of gravity, and the law of
description of areas.
And, in
two analogous conditions,
manner, the function
V is
obliged to satisfy these
XXXIII.
SV(aDaF+ aoDoF) = 0;
2(DaF + DoF) = 0;
accordingly, by XXV. XXVI., are new forms of the same equations VIII. IX.,
XXXII.
which
like
and consequently are new expressions of the same two laws.

(19.) All the foregoing conditions are satisfied when t is s7naU, that is, when the
time of motion of the system is short, by the following approximate expressions for the
functions
for
Fand
H and
F, with the respectively derived
and mutually connected expressions
XXXIV.
..F=^(P+Po) ;
F= s (P + Fo + 2^)>
XXXV.
.
XXXVI.
iZ
XXXVII.
n which
s denotes
XXXVIII.
= -(D,F)=-i(P+Po)
a real and positive

s2
= _ 2m (a -
i-^2'
t=DMV=s(^F+Po + 2Hy^;
scalar,
ao)2,
or
such that
XXXIX.
= V 2mT (a -
ao)\
419. As Q. fourth specimen, we shall take the case of a free point

or particle, attracted to a fixed centre* o, from which its variable
vector is a, with an accelerating force = Mr^^ if r = Ta = the distance
*
When
two free masses,
and
m', with variable vectors
a and a,
attract each
other according to the law of the inverse square, the differential equation of the relative
motion of
r.
lid this
+ m\
m
.
about m'
Dâ -
is,
a')
by 418, VII.,
= {in + in')
equation F. reduces
itself to
(a
I.,
- a'yh-\
Avhen
we
if
= T(a - o')
write a for
a~-n\ and
for
718
of the point from the centre, while

differential equation of the
I.
J/ is the
[bOOK
III.
attracting mass: the
motion being,
I>â = Ma-^r-\
.
if D (abridged from D^) be the sign of derivation, with

respect to
the time i.
on
(1.) Operating
I.
we
with V.a, and integrating,
obtain immediately the
equation (comp. 338, (5.)),

II.
VaDa =i3= const.
it is
proportional to the time
plane, and also that the area described in

U/3 being the fixed unit-normal to the plane, round
which expresses at once that the orbit

;
is
angular motion, revolves positively ; and T/3 representing in

quantity the double areal velocity, which is often denoted by c.
(2.) And it is important to remark, that these conclusions (1.) would have been
which the
point, in its
obtained by the same analysis,
had been replaced by any other scalar
if r-i in I.
function, f(r), of the distance ; that is, for any other law of central force, instead of
the law of the inverse square.
(3.) In general^ we have the transformation,
= dUa:Vada,
III. .. a-iTa-i
because,
by 334, XV.,
IV.
the equation
may
I.
therefore
V.
and thus
it
&c,,
have,
dUa = V(da.a-i).Ua = a-2Ua.Vada = a-Ta-i.Vada;
we
gives,
VII.
by
II.
by an immediate
.
be transformed as follows,
D2a = -yDUa,
if
VI.
=- Jlf/Si
integration,
.Da = 7(Ua-0,
or
VII'.
Da = ( - Ua)
being a new constant vector, but one situated in the plane of the
plane /3 and y are perpendicular.
if,
orbit,
to
which
a. Da, Dâ are here (comp. 100, (5.) (6.) (7.)) the vectors of /)05t<ion,
and acceleration of the moving point; and it has been defined (100, (5.))
But
(4.)
velocity,
that
y,
for
any motion of a point, the vectors of velocity be set oflF from any common
on which they terminate is the Sodograph* of that motion.
origin, the curve
(5.)
Hence a and Da,
if
the latter like the former be
drawn from the fixed point

; and because the
and hodograph
o, are the vectors of corresponding points of orbit
formula VII. gives,
VIII.
SyDa = 0,
it
follows that the hodograph
of
which
Compare
it
IX.
(Da + yt)' =
y2,
in the present question, a Circle, in the plane of the
see also pages 100, 616, 578, from the two latter
be perceived, that the conception of the hodograph admits of some
Fig. 32, p. 98
may
is,
and
purely geometrical applications.
CHAP.
with
orbit,
LAW OF THE CIRCULAR HODOGRAPH.
III.]
Ty =
cy) for the vector of its centre, and with

shall also denote by h.
ye (or
719
MT/3-' for
its
radius, which radius we
The
(6.)
Law
of the Circular* Hodograph
is
therefore a mathematical conse-
the Inverse Square ; and conversely it will soon be proved, that

quence of the Law of
no other law of central force would allow generally the hodograph to be a circle.
For the law of nature, the Eadius (h) of the Hodograph is equal, by (1.) and
mass (Af), divided by the double ureal velocity
(7.)
(5.),
to the quotient of the attracting
and
(T/3 or c) in the orbit ;
if
we
write
X. ..e=Tf,
this positive scalar e
be called the Excentricity of the hodograph, regarded as a

with respect to i\iQ fixed centre offorce, o.
may
circle excentrically situated,

if e
(8.) Thus,
< 1,
the fixed point o
is
interior to the hodographic circle
if e
1,
the point o is on the circumference ; and if c> 1, the centre O of force is then exteTxor to the hodograph, being however, in all these cases, situated in its plane.
The equation VII.
(9.)
XI.
gives,
.
6-Ua=-y-Da = Da.y-';
operating then on this with S. a, and writing for abridgment,
XII.
so that
/) is
.p= /3y-i = M- iT/32 = e^M-\
and
a constant and positive scalar, whiles
XIII.
is
SUac = cos v,
the inclination of a to
e,
we
find,
XIV.
the orbit
is
(10.)
+ Sa=p;
or
XV.
1
therefore biplane conic, with the centre of force
and p
for its excentricity,
And we
mass
attracting
see,
that this constant c
may
for
u
a,
focus, having e
for its semiparameter.
by XII., that
Af, the
e cos
if this
semiparameter
be multiplied by the
product is the square of the double areal velocity c ; so

be denoted by (Mp)^, which agrees with known results.
on the other hand, we divide the mass (iW) by the semiparameter (/>),
the quotient is by XII. the square of the radius (i/Tj3'^ or A) of the hodograph.
(11.)
If,
(12.) And if we multiply the same semiparameter p by this radius MTjS'^ of

the hodograph, the product is then, by the same formula XII., the constant Tj8 or
double areal velocity in the orbit, so that h = Mc~^ = cp~K

(13.) If we had operated with V. a on VII'., we should have foimd,
c of
XVI.
/3
= V. a (c - Ua) y = (Sa + r) y
which would have conducted to the same equations XIV.
XV.
as before.
* This law
of the circular hodograph was deduced geometrically, in a paper read
Koyal Irish Academy, by the present author, on the 14 th of December
before the
1846
was
virtually contained in a quaternion formula, equivalent to the rewhich had formed part of an earlier communication, in July, 1845.
(See the Proceedings for those dates ; and especially pages 345, 347, and xxxix.,
;
but
it
cent equation VII.,
xlix.,
ofVol. III.)
720
we
(14.) If
operate on VII. with S.a,
XVII.
but
XVIII.
and
XIX.
we
if
- y2 = A2 =
we
(by VI. and XII., comp. (1 1.)),
- (Vaey =e^r^-(p - ry=p (2r -p
-r'â-i),
XX.
write
III.
find this other equation,
- rDr = SaDa = y Vac
M
_
[bOOK
<
hence squaring XVII., and dividing by
r"^,
e^
we
obtain the equation,
= A/
XXI...,|J
obvious that this last equation, XXI., connects the distance,"/; with
the time, t, as the formula XV. connects the same distance r with the true anomaly,
V that is, with the angular elongation in the orbit, from the position of least dis(15.) It
is
But it would be improper here to delay on any of the elementary consequences of these two known equations although it seemed useful to show, as above,
how the equations themselves might easily be deduced by quaternions, and be contance.
nected with the theory of the hodograph.

(16.) The equation II. may be interpreted as expressing, that t\xQ parallelogram
Fig. 32) under the vectors a and Da of position and velocity, or under any
(comp.
two corresponding vectors (5.) of the orbit and hodograph, has a constant plane and
area, represented by the constant vector /3, which \s perpendicular (1.) to that plane.
But it is to be observed that, by (2.), these constancies, and this representation, are
not peculiar to the law of the inverse square, but exist for all other laws of central
force.
(instead of Mr'^^ represent the acce(17.) In general, if any scalar function

lerating force of attraction, at the distance r from the fixed centre o, the differential
equation of motion will be (instead of I.),
XXII.
and
if
we still
XXIII.
in
write
VaDa =
D3a = -
which
Di2
/3
/3,
D2a = i?m-i = - R\Ja
as in II., the formula IV. will give,
Ua -
= cU/3,
i?/-2/3Ua,
if
= Tf3,
and
XXIV.
V ^ = r'^B
D3/Y
as before.
(18.) Applying then the general formula 414, L,
we
have, for any law* offorce,
the expressions,
XXV.
Vector of Curvature of Hodograph
XXVI.
- =
D3a
Dâ
Dâ
Rr^
Rr^c'^
= - V
Radius {h) of Curvature of Hodograph
VaB
Force x Square of Distance

Double A real Velocity in Orbit'
* The
general value XXVI., of the radius of curvature of the hodograph, was
geometrically deduced in the Paper of 1846, referred to in a recent Note.
CHAP.
of
PRODUCT OF OPP. VELOCITIES, POTENTIAL.
III.]
only conducts, in a new way, for the law of nature^ to the conA/c~', but also proves, as stated in (6.), that for any other luw
last not
which the
stant value (7.). ^
of central force the
hodograph cannot be a
indeed
circle, unless
be such, and to have moreover the centre of force at
tlie
orbit
happens
to
its centre.
(19.) Confinhig ourselves however at present to the

,;id
721
law of the inverse square,
writing for abridgment (comp. (5.)),
XXVI L
k:
= OH =
ey
h of Hodograph^
Vector of Centre
which gives, by (5.) and (7.),
XXVIII.
o of vectors being
the origin
circle, that
that this constant product*
by
T/c
eh,
we
the centre offorc^
any two opposite
the product of
XXIX.
still
by the properties of the

; and
sec
velocities in the orbit is constant
be expressed as follows,
may
(e-l)AUK.(e+l)/tUc = A2(l-e2)= Ma-\
Xnil. and XX.
The expression XXIX. may be otherwise written as k^ - y2; and if v be

any point u external to the circle, but in its plane, and u the length
of a tangent UT from that point, we have the analogous formula,
(20.)
the vector of
XXX.
M2=y2_(_K)3=T(v-K:)2-/t3.
(21.) Let r and r' be the vectors ox, ot' of the two points of contact of tangents thus drawn to the hodograph, from an external point u in its plane ; then
tach must satisfy the system of the three following scalar equations,
.Syr = 0;
XXXL.
XXXIL
(r
k)^
y2
XXXIIL
S (r-
k)(i;
- ) =
y2
whereof the first alone represents the planej the two ^rsi jointly represent (comp.
and the third expresses the condition of conjugation of the points

T and u, and may be regarded as the scalar equation of the polar of the latter point.
(5.)) the circle ;
It
is
understood that Sy v
0,
as well as
SyK =
&c., because
0,
is
perpendicular
(3.) to the plane.
(22.) Solving this system of equations (21.),
XXXIV.
in
= K:+y(y + tt)(i;-fc)-i;
which the scalar has the same value as in
pressions give,
* In
by what
strictness, it is
we
XXXIV.
find the
.
two expressions,
r'= k + y (y -) (u - k)-i
As a verification, these ex;
(20.).
precedes.
only for a closed
orbit,
that
is,
for the case (8.) of the centre
hodograph (e < 1), that two velocities can be opposite ;

their vectors having then, by the fundamental rules of quaternions, a scalar siaA
positive product, which is here found to be = Ma~^, by XXIX., in
consistency with the
of force being interior to the
known theory
of elliptic motion.
Other cases also.
It is
The
result
however admits of an interpretation,
obvious that when the centre o of force
is
in
exterior to the hodo-
graph, the polar of that point divides the circle into two parts, whereof one is concave, and the other convex, towards o ; and there is no difficulty in seeing, that the
former part corresponds to the branch of an hyperbolic orbit, which can be described
under the influence of an attracting force : while the latter part answers to that
other branch of the
same complete hyperbola, whereof the
the force to be repulsive.
4 z
description
would require
722
XXXV..
In
XXXVI.
fact
it is
= 0; XXXV.
S(r'-/c) (''-)= 0;
- v)2 = (r' - 1;) = - u^.
(r
S(r-ic)(r-i;)
and
III.
[bOOK
found that
XXXVII. ..T-v = u{u + y)(v- k)-'
XXXIX.
and
XXXVIII.
- k) = y
y) (r
.
(r
= T (v -
T(u-\-y)
k)
w + y being here a quaternion.

(23.) If u' be the vector ou' of ani/ point
respect to the circle, then changing r to
to z, in
witli
XXXIV,, we find this vector
of the equation of that polar,
form (comp. (21.))
XL.
.5;'
=K+y
(y
+ 2)
(w
fc)-l,
by an easy transformation,
or,
XLI.
in
on the polar of the point u
u',
and u
v',
which k
(h^
an arbitrary
is
we suppose
(24.) If then
+ m2)
v'
h^v
u^K
+ 2y
(k
v),
scalar.
that u'
the intersection of the chord tt' with the
is
right line ou, the condition
XLII.
\v'v =
XLIV. ..Ykv.^k-
but
XLI II.
will give
v) = kS
zy
- y2) +
{kv
^^S
(ku
^2)
the coefficient then of k, in the expanded expression for v\ disappears as

do : and we find, after a few reductions,
XLV.
j/= V
Skv
y2
V- v-^Skv
it
ought
to
a result which might have been otherwise obtained, by eliminating a new scalar y
between the two equations,
XLVI.
v'
= t/v,
{i/v
k) (w
(25.) Introducuig then two auxiliary vectors, X,
XLVII.
XLVII'.
and therefore
XLVIIL
and
we have
X = v-'^Skv,
X - k = v iV/cv,
+f
= y3.
Skv = vX = Xv,
SkX = X2, (X - jc)2 = x:2 _
or
f 1
k)
such that
^-^^
V\
whence
fi
\\
X,
(/i
\2^
- k)2 = y2,
from the
(u
- X)
(y'
- X) =
(/I
or
X)2,
L.
lu. lu'
= lm2^
= om. Accordingly, the point l is the foot of the perpendicular let

centre h on the right line ou, while m is one of the two points m, m' of
X = OL, and
fall
^ =X
/x,
the very simple relation,
XLIX.
if
/i
intersection of that line Avith the circle
so that the equation L. expresses, that the
points u, u' are harmonically conjugate, with respect to the chord

the middle point, as is otherwise evident from geometry.
Aiai',
of which L
is
(26.) The vector a of the orbit (or oi position), which corresponds to the vector
Da) of the hodograph (or of velocity'), and of Ivhich the length is Ta = r = the
r (=
distance,
be deduced from r by the equations,
may
LL
= r(K-T)y-i,
and
LIL
Vra = -/3= My->;
whence follow the expressions,

LI II.
Potential
= M-i =
(say)
P= Sr
(c
- 7)
= Sy (k -
r)
CHAP.
CONSTRUCTIONS FOR THE POTENTIAL.
111.]
the second expression for
being deduced from the
723
by means
first,
of the relation
XXXV.
shows that the potential is equal, 1st, to

(27.) The first expression LIII. for
the rectangle under the radius of the hodograph, and the perpendicular from the
centre
o of
on the tangent
force,
at
to that circle
tangent from the same point T of the hodograph, to

Exccntricity, namely to that new circle which has
il/,
to the square of the
oh
for
called the Circle
a diameter.
And
the
of
first
otherwise deduced from the equality (7.) of the

h of the hodograph, multiplied by the constant
may be
of these values of the potential
mass
and Ilnd,
what may be
to the product he of the radius
c of double areal velocity, or by the constant parallelogram (16.) under any two corresponding vectors.
(28.) The second expression LIII. for the potential P, corresponding to the
point T of the hodograph, may (by XXXIV., &c.) be thus transformed, with the
help of a few reductions of the
same kind
as those recently
uy...P=^ = *^-p,
if
employed
LV..., = v(.-.),
t'
q being thus an auxiliary quaternion ; and in like manner, for the other point
we have the analogous value,
lately considered,
LVI...P'=-='^'^^""^'^Â2
r'
rlience
LVII.
and
+ M2y2-)
A2+2
h^ (Sg2
P.F'
therefore,
LVIII...- = P-I?^^~-^.
and
finally,
2M
. = r^--^=Sq+-^=v(\-v') = OV.V'L.
F+P
Sq
^
2PP'
LX.
'M2y2
r+r
same second expression LIII. shows, that if

(29.) In fact, the
from t and t* on hl, then the potentials are,
v and
v' be the
feet of pei-pendiculars
LXI.
P= ou
TV,
P' = ou
and
tV
and it is easy to prove, geometrically, that the segment u'l is the harmonic mean between what may be called the ordinates, xv, t'v', to the hodographic axis hl.
the point u to take any new but near position u, in the plane,
(30.) If we suppose
the polar chord tt', and (in general) the length u of the tangent ut, will change ; and
we
shall
have the
differential relations
LXII.
LXir.
LXII I.
.
and
that
(31.) Conceiving next
dr = (r - v)-S (r - k) dv
dr' = (r' - v)-S (r' - k) dv
u moves along the
write,
hXlY.
we
shall have,
du = m'IS (k -v)dv.
.v = (x-e){fi-X),
line
LU
if
ou, or lu, so that
LM
we may
ELKMENTS OF QUATERNIONS.
724
LXV,
Av = {ix-X)dx
= v{x-
e')-ida;,
u be on LM prolonged, and if o be on the concave

by LIII., the differential expressions (30.) become,
LXVI.
111.
> 1 > c',

tmt'
and
thus,
dr = (u-r)-jP(^-0-^d:c; dr' = {v - r'y^P' (x-e'y^dx-,

dw = u-^Sq.(x - e'y^dx, with 8q = v{\-v)
P'dx
Pdx
, .
LXVII.
so that
a;
side of the arc
if
and
with
[bOOK
LXVIII.
Tdr
t(a;
e')'
Tdr
if
da;
>
0.
u(^x-e'y
Such then are the lengths of the two elementary arcs tt, and t*!/ of the hodograph,
two near secants ntt' and nt,t/ drawn from the pole n of the
intercepted between
chord
Mjtf',
and having u and
u, for their
own
poles
and we
see that these arcs are
and P', or by LXI. to the ordinates, tv, t'v', or

proportional to the potentials,
and accordingly we have the inverse similarity (comp.
finally to the lines NT, nt'
118), of the two small triangles with n for vertex,
:
LXIX.
as appears on inspection of the
A NTT, cx'ntV,
annexed Figure 86.
CHAP.
THEOREM OF HODOGRAPHIC ISOCHRONISM.
III.]
because, for the motion here considered, the measure or quantity of the force
725
is,
by
I.
and LIII.,
LXXI.
TD2a = Mr-^ = M'^F^.
(33.) The times of hodographically describing the two small
and t't/, are therefore inversely proportional
and
tional to the distances in the orbit ;
circular arcs, t^t
to the potentials, or directly propor-
sum
their
is,
LxxII...d.+dr=l^+^^""^^-('+^'>^^
P P jx-e (a;-e')'
that
is,
by LX. and LXIV.,
LXXIII. ..d<+d<'
We have
(34.)
LXXV.
LXXIV.
if
-,
.p=T(ii-\)=i:M.
also the relations,

.
= (a;2 -
7^
and
l)\ g,
LXXVI.
M
=
(1
e'2)
g^
two small times may be thus expressed,
so that the sura of the
..xxviT.
d,.d<-
..
mL:r^Xlzlf^
_
Mi
X
l)J
(a:2
or finally,
dw
i\l-e'-iy\^
e'
LXXIX.
if
in
which Figure u'w
x = s&cw,
or
an ordinate of a
is
m;
= z.mlw
cos u>)*
in Fig. 86,
semicircle, with the chord
mm'
of the hodo-
for its diameter.
graph
(36.)
The two near
secants (31.), from the pole
of that chord,
have been here
supposed to cut the half chord lm itself, as in the cited Figure 86 but if they were
to cut the other half chord lm', it is easy to prove that the formulas LXXVIII.
LXXIX. would still hold good, the only difference being that the angle w, or ml"vv,
;
would be
now
obtuse,
and
secant
its
x<l.
(36.) A circle, with u for centre, and u for radius, cuts the hodograph orthogonally in the points t and t' ; and in like manner a near circle, with ti^ for centre,
and M
+ dw
points
for radius, is
T and t/
(31.).
another orthogonal, cutting the same hodograph in the near

conceiving a series of such orthogonals, and observ-
And by
LXXVIII. depends only on the four scalars,

known when the mass Mand the Jive points o,
they do not change when we retain that mass and those
ing that the differential expression
M~â?, e\ w, and
(\w,
which are
all
M, u, u, are given, so that

but alter the radius h of the hodograph, or the perpendicular hl, let fall from
its centre ii on the fixed chord mm', we see that the sum of the times (comp. (33.),
L,
points,
of hodographically describing
any two circular
arcs, such as
t t and
t't/,
even
if
they be not small, but intercepted between any tivo secants from the pole n of the
fixed chord, is independent of the radius (Ji), or of the height hl of the centre h of
the hodograph.
such as the two in Fig. 86, having a com(37.) If then two circular hodographs,
mon chord mm', which passes through, or tends towards, a common centre offorce o,
with a common mass Jf there situated, be cut by any two common orthogonals, the
sum of the two times
of hodographically describing (33.) the two intercepted arcs
(small or large) will be the same for those two hodographs.
726
(38.)
And
'^
III.
we have the following Theorem*
as a case of this general result,
Hodographic Isochronism (or Synchronism)
[boOK
of
common chord, which passes throvgh, or

centre of force, be cut perpendicularly hy a third
circle,
the times of hodographically describing the intercepted arcs will be equal."
If two
circular hodographs, having a
common
tends towards, a
For example,
we have
in Fig. 86,
LXXX.
the equation,
(39.) The time of thus describing the arc tmt' (Fig. 86),
out concavef towards o (so that a;>l>e', as in LXV.),
LXXVIII.) by the definite integral,
LXXXI.
Time o/tmt'
Time ofTMT'=time o/avmw'.
=2
^\,
^
if this
is
expressed (comp.
r.
Jo
arc be through'
i^-e
cos m;)2
and the time of describing the remainder of the hodographic circle, if this remaining
arc t'm't be throughout concave towards the centre o of force, is expressed by this
other integral,
LXXXII.
= 2
T^me of.'u'.
^
^(1^/^^
^ r-^~^M
J,^(l-e'cosM;)2
j
(40.) Hence, for the case of a closed orbit (e'^
mean angular
LXXXm.
we have
velocity,
Perioiic 77.
<
1, e
<
1,
> 0),
if
n denote the
the formula,
= ?^=
LXXXI V.
^V
\M
(1
^
M=a3n2,
^^
-e'^Y
[%-,
'Jo
(1-e
cosm;)2
-"[^'
as usual.
The same result, for the same case of elliptic motion, may be more rapidly obtained,
by conceiving the chord mm' through o to be perpendicular to oh for, in this position of that chord, its middle point l coincides with o, and e'by LXIV.
(41.) In general, by LXXVI., we are at liberty to make the substitution,
;
LXXXV.
/a'(l-c'2)3 \i
.
j^
supposing then that
e'
=-
-3
1,
or placing
with g
half chord of the hodograph
at the extremity m' of the chord,
by LXXXI.,
LXXXVI.
Parabolic time of tmt'
= 2Mr^
g^ ]q (I
for,
we have
wy'
dw
cos
when the centre offeree is thus situated on the circumference of the hodographic
we have by (8.) the excentricity e = 1, and the orbit becomes by XV. a para-
circle,
* This
Theorem, in which
is
occupied by a
common mass
it is
common centre
was communicated to the Royal
understood that the
(Af),
of force (o)
Irish
Aca-
the 16th of March, 1847. (See the Proceedings of that date. Vol. III., page
It has since been treated as a subject of investigation by several able writers,
demy on
417.)
to
whom
f-pace
the author cannot hope to do justice on

which now remains at his disposal.
t Compare the Note
to page 721.
this subject, within the
very short
CHAP.
bola.
CHORD OF ORBIT, SUM OF DISTANCES.
III.]
For hyperbolic motion (e'2>
without the substitution
lm' prolonged; and the formula
tuated between
(42.) For
any
and
a<0), the formula LXXXI. (with or

if c* < - 1, that is, if o be on
I,
to be
is
employed
LXXXII.,
e'>l, e'<secw, that
if
t, t' of the
tangents to the orbit at p,
hodograph, and
if
o be
si-
which corre-
q be the point of meeting of the
if
as in the annexed Figure

87, while the tangents to the
p',
meet as before
t, t'
i.-^,
u.
laio of central force, if r, p' be the

points of the orbit
spond to the points
hodograph at
e>
1,
LXXXV.)
727
we
in u,
shall
have the parallelisms,
I
Fig. 87.
LXXXYII.
OP
II
op'
UT,
II
pq
t'u,
||
qp'
ot,
ot'
||
writing then,
LXXXVIII.
.OP=a,
= a', OT = Da = r,
OP'
most of which notations have occurred
LXXXIX.
thus
and
XC.
XCI.
= Da' = r', ou =
we have
i;,
OQ =
w,
the equations,
= Va(r-v;=Va'(w-r')=Vr(a;-a)=Vr'(a'-w);
Vav = Var = /3 = VaV' = Va'u, a'-a||i;, pp' oxr,
Vrw = Vra = - /3 = Vr'a' = Vr'w', r - r' w, t't oq.
.
before,
OT'
||
||
||
Geometrically, the constant parallelogram (16.) under op, ot, or under op', ot', is
equal, by LXXXVII., to each of the four following parallelograms : I. under op, ou ;
II.
under
t't
II
op',
ou;
III.
under oq, ot
and IV. under oq,ot'; whence pp'||ou, and
OQ, as before.
(43.) The parallelism XC. may be otherwise deduced for the law of the inverse
square, with recent notations, from the quaternion formulae,
=
xciL..î^=-^
r
r
\- V
+
^
u
in which,
and which may be obtained in various ways whence

if s denote the length T {a' - a) of the chord pp' of the
;
XCIII.
u
,=;-7T
T
(\
xcir...'=^!îi:.',
r
+ r
it
= UT UL =
v)
&c.
being the same auxiliary angle as in (34.),
may
orbit,
&c.
also be inferred, that
then (comp. Fig. 86),
= sm
m;
728
(44.) It
[boOK
easy to prove that
is
xc.y...x-. =
(,+^)f, x-/=(i-"-)f
whence
XCV.
= :^'=^,
..t'^
r
T\
JP
XCVI.
and
= K.P-i(r-X)
P'-i(r'-X) w
the lines lt, lt' are therefore in length proportional to the potentials, P,
their directions are equally inclined to that of ou, but at opposite sides of
the line
lu
drawn
it,
y;
and
so that
Accordingly (see Fig. 86), the three points t, l, t'

which has hu for diameter so that the
bisects the angle tlt'.
are on the circle (not
P'
in the Figure)
tlu
are equal to each other, as being respectively equal to the angles

utt', tt'u, which the chord tt' of the hodograph makes with the tangents at its ex-
angles ult',
tremities
the triangles tlv, t'lv* are therefore similar, and

to P', or as / to r.
by LXI., as
lt
to lt' as
is
tv
to
t'v', that is,
(45.) Again, calculation with quaternions gives,
xcvii.
whence
= (izllC^ = (-.)(- X) (. - )-.,

(îzl^n)
V r
V T
v'T r^v' ~
=:TXCVIII...T^
T
,
= r^T
T-v =uf:uL = sin>;
is the common ratio, of the segments tu', u't' of the base tt' of the triangle tlt', to the adjacent sides lt, lt', which are to each other as r' to r (44.)
and because this ratio is also that of to r \- /, by (43.), we have the proportion,
such then
XCIX.
OP
op'
PP' =':
= lt' LT
:
tt',
and the formula of inverse similarity (118),

Alt't a' opp'.
C.
.
Accordingly (com p. the
two
triangle are respectively equal,
and
by the
therefore,
last Figures),
by the
circle (44.), to the
but the two rotations, round
o from p
the base angles opp', op'p of the second
parallelisms (42.), to the angles tul, t'ul,
base angles tt'l, t'tl, of the

to p',
and round l from
first
triangle
t' to t, are oppo-
sitely directed.
(46.) The investigations of the three last subarticles have not assumed any knowledge of the form of the orbit (as elliptic, &;c.), but only the law of attraction according to the inverse square, or by (6.) the Law of the Circular Hodograph. And
the
principles give not only the expression
same general
Ma-\ but
also
CI...
which
last
(by LX.
2M
may
LXIV. LXXIV. LXXIX.)
=(1 e'cosu?)*/^;
whence
LXXVl.
for the constant
this other expression,
r+r
CII.
e'*
:
be considered as a quadratic in e, assigning two values (real or

scalar, when the first member of CII. and the angle w are given
imaginary) for that
the sine of this latter angle being already expressed
by XCIII.
(47.) Abstracting, then, from any ambiguity* of solution,
we
see,
by the
definite
* That there
ought to be some such ambiguity is evident from the consideratior,
when a, focus o, and two points v, p' of an elliptic orbit are given, it is stiil
that
CHAP.
LAMBEirr'dTHEOREMjPARTlALDIFF. EQUATIONS. 729
III.]
integrals in (39.), that the time
the inverse square,
is
of describing an arc pp' of an orbit, with the law

a function (conip. (36.)) of the three
ratios,
r^^r'
'
^"'.>/
,+.
ich is a form
of Lambert's Theorem, but presents itself here as deduced from the

Theorem of Hodographic Isocbronism
without the employment
recently stated
(38.),
any property of conic
sections.
(48.) Tlie differential equation

tten
(comp. 418,
CIV.
in 418, X., if
S.D2â + ^P =
we now
whence
0,
Series
CV.
418)
T=
may
be thus
P+IT,
write,
r=-iDa2 = -|r2,
CVI. ..
fact
of the present relative molion
I.
and generally the preceding
I.,
and
CYll.
H=~;
..
2a
(by LIIL, comp. (20.) (21.)),
CVIII.
F may
that
2fl-=2(P-r)=2P+
CIV. by
(49.) Integrating
cix.
.-
r?
= K^
and writing (as
parts, &c.,
..F=P(r+p)d^,
cx.
and
y2=
a
in 418,
r=
XII. XXII. ),
p2rd<,
and
again be called the Principal Function

we have the variations,
V the
Characteristic
iction of the motion,

:
CXI.
il
dF= SrSa - Sr'Sa - Hdt
which
vectors (op, op') of the orbit ;
CXIII.
CXIV.
F being
if
The two
^r
dV= Srâ -
and
a',
t,
while
= - r-'Saâ,
r, r', s
^r'
confining ourselves then, for the
T is
r';
a scalar function of
a, a',
H,
two scalar function^, only
(comp. 418, (17.))
moment,
derivatives,
D'F= D^'F=CXV.
D^V=ti
.
but
h = - -iS (a
= - r'-iSa?a',
Sr'da' + tSH;
be called the initial and
vectors a, a' can enter into these
through their dependent scalars
CXVI.
CXIII'.
(PtF)=-H',
here a scalar function of a,
now what may
whence follow the partial
DaF=DaF=T;
M be treated as given.
(50.)
CXII.
a) denote
a, a' (instead of ao,
to the function V,
- a)
(â'
da)
and observing that we
have by CXII. the formula,
CXVII.
in
S (râ - r'U) = D^ F. Sr +
which the variations ca, da are arbitrary, we
CXVIII.
CXVIir.
D^^ V. 6r
+ D. V.
Ss,
find the expressions,
r--ar-iD,F + (a'-o)s-'D,F;
= + aV-'D/ F+ (a'- a)s-iDsF;
r'
permitted to conceive the motion to be performed along either of the two elliptic arcs,
which together make up the whole periphery. But into details of this kind
pp', p'p,
we cannot
enter here.
5 A
730
which give these
[bOOK
others,
CXIX. ..DrV= rY(â' -a)T: Yaa'
D/ F= rY(a - a') r' Yaa'

CXX.
Ds F = s/3 Yaa,
Var = Va'r' = j8.
CXIX'.
and
because
(51.) But,
III.
by XCir.,
CXXI.
rr +
rV = (r + r') v'
\\
v\\a'- a,
the chord TT of the hodograph, in Figures 86, 87, being divided at u' into segment
Tu', u't', which are inversely as the distances r, r', or as the lines op, op' in the
orbit
we have
CXXII.
therefore the partial differential equation,

.
D,F= D/ F,
two
so that each of the
and
functions,
CXXIII.
similarly,
F and
D^F= Dr'F;
F, depends on the distances
depending on their sum, r + r.
r,
r\ only by
as variable, the Principal

(52.) Hence, if for greater generality we now treat
Function F, and therefore by CXIV. its partial derivative
(JitF), are func-
H=-
tions of the four scalars,
CXXIV.
.r^r',
s,
M.
and
t,
And in like manner,
the Characteristic Function (or Action-Function') V,

(53.)
and its partial derivative (by CXV.) the Time, t^DnV, may be considered as
functions of this other system of four scalars (comp. (47.)),
CXXV. ..r + r',
and
H,
s,
no knowledge whatever being here assumed, of the form or properties of the

but only of the law of attraction.
orbit,
new
(54.) But this dependence of the time, t, on the four scalars CXXV., is a
form of Lambert's Theorem (47.) which celebrated theorem is thus obtained in a
;
new way, by
(55.)
the foregoing quaternion analysis.
Squaring the equations CXVIII. CXVIII'., attending to the
CXXII., and changing
CXXVI.
signs,
2P+
we
get these
new
relation
partial differential equations,
2-H = (DrF)2 + (D,F)2
+ !LliL-f D^r. DsV;

'^^
cxxvr. ..2P'+2H=(Prry + (PsVy+ ''^~[

d,.f.d,f;
CXXVII. ..a^=-r^, a 2 = - r'2, (a' -a)^ = - s^.
because
Hence, by merely algebraical combinations (because
P = Mr-\
= Jl/r'-), we
and P'
find:
CXXVIII.
i ((Dr F)2
CXXIX.
cxxx.
(D, F)2)
D,.F.D,r=
= H+
r'
^^
r+r
M
r +
4:M
r^r+s
r
+ r s
(4
(D,.F+DJ02 = 2^+ r+r'rr

+ r'-\-s
+ s =^[\r-T-T^
^
+ r -s
\r+r -s
1\
;
aj
a
CHAP.
DEFINITE INTEGRALS FOR ACTION AKD TIME.
III.]
by CXII. CXVII. CXXII., we have the
(56.) But,
CXXXl
V vanishes
and the function
ment of the motion
cxxxii...
cxxxiii.
As
variation,
SV-tSH= i(DrV+ DsV) d {r + r'+ s) + ^(PrV-DsV)
..
with
whence
t,
and
therefore with
s,
a verification,t
-^
and
- s)
commence-
easy to infer the expressions,*
it is
when
5(r + r'
at least at the
r=r(_^-+fVd.=rf-^
=i f
731
f yi d, = J
s are small,
-â*d.;
^'4- - ^y*<u.
and therefore
nearly
= r, we have
thus the approximate values,
CXXXIV.
..
CXXXV.
in which s
may
V={2P+2H)ls = {2T)U = 2Tt;

= (2P+ 2Hyis = (2 Tyh
<
be considered to be a small arc of the
with which that arc
is
orbit,
and (2 T)i the
velocity
described.
(57.) Some not inelegant constructions, deduced from the theory of the hodograph, might be assigned for the case of a closed orbit, to represent the excentric and
mean anomalies ; but whether the orbit be closed or not, the arc tmt' of the hodo-
graphic
circle, in Fig. 86,
represents the arc of true anomaly described
tends at the hodographic centre
ii
an angle tht', which
is
for it sub-
equal to the angular mo-
tion pop' in the orbit.
We may add that,
(58.)
CXVIII. CXVIII'.
tions
whatever the special form of the orbit
give,
CXXXVI.
from which
it
r'
= (Ua
-|-
be, the
equa-
Ua) D^ F;
follows that the chord tt' of the hodograph
of the angle pop' in the orbit
may
by CXXII.,
and
therefore,
is
by XCI., that
parallel to the bisector

this angle is bisected
by
in Fig. 87, so that the segments pb, rp', in that Figure, of the chord pp' of the
of the chord tt' of the hoorbit, are inversely proportional to the segments tu',
OQ
uV
dograph.
(59.)
We arrive then thus,
known theorem
in a
new way, and
as a
new
verification, at this
tha,tiftwo tangents (qp, qp') to a conic section be drawn from
Expressions by definite integrals equivalent to these, for the action and time
motion of a binary system, were deduced by the present writer, but by
in the relative
an
in the First Essay, &c., already cited, and will be

entirely different analysis,
It is supposed that
II., pages 285, 286.
found in the Phil. Trans, for 1834, Part

the radical in
tion
if it
CXXXIII.
does not become infinite within the extent of the integra-
did so become, transformations would be required, on which
we cannot
enter here.
t An analogous
which however
riable
constant.
it is
verification
to
may
be applied to the definite integral
be observed that both r +
whereas, in the recent integrals
r'
and
s vary,
CXXXII. CXXXIII,,
LXXXI.
iu
along with the var
r' is
treated as
732
any common point
(q), they subtend equal angles at
form of the conic
cial
(60.)
structions
And
may
a focus
[bOOK
III.
whatever the spe-
(o),
be.
although, in several of the preceding sub-articles, geometrical conillustrate (and so to confirm^ if confirmation were
have been used only to
needed) results derived through calculation with quaternions
suggestive nature of the present Calculus enables us, in this as in
own
tions, to dispense with its
processes,
when once they have
yet the eminently

many other quesindicated a definite
train o^ geometrical investigation, to serve as their substitute.
(61.) Thus, after having in any manner been led to perceive that, for the motion
above considered, the hodograph is a circle* (5.), of which the radius ht is equal
(7.) to the attracting
the vectors op,
ot
mass M, divided by the consta.nt parallelogram (16.) under

and velocity, in the recent Figures 86 and 87, which
of position
parallelogram is equal to the rectangle under the distance op in the orbit, and the
perpendicular oz let fall from the centre o of force on the tangent ut to the hodo-
graph,
we
see geometrically that the potential P, or the
tance, for the point
equal (as in (27.)) to the rectangle under

triangles htv, uoz, to the rectangle under
the
ht and oz, and therefore, by

ou and tv (as in (29.)).
is
the similar
perpendiculars
HL
dis-
(62.) In like manner, the three potentials corresponding to the second point t' of
first hodograph, and to the points
and w' of the second hodograph, in Fig. 86,
are respectively equal to the rectangles under the
mass divided by the
of the orbit corresponding to the point t of the hodograph,
t'v',
wx,
so that, for these
common chord mm',
w'x', on
what we have
same
line
ou, and the three other
called (29.) the hodographic axis,
two pairs ofpoints, in which the two circular hodograpfis, with

are cut by a common orthogonal with u for centre, the /owr
potentials are directly proportional to the /owr hodographic ordinates (29.).

(63.) And because the force (Jf-"2) is equal to the square of the potential
(Mr-^), divided by the mass (^M), the four forces are directly as the squares of the
each force, when divided by the square of the corre-
four ordinates corresponding
sponding hodographic ordinate, giving the constant or common quotient,
CXXXVII.
ou2
M,
(64.) It has been already seen (31,) to be a geometrical consequence of the two
pairs of similar triangles, ntt , nt't', and ntv, nt V, that the two small arcs of the
Jirst
hodograph, near T and
ofthefxed chord uyi,
t',
intercepted between two near secants from the pole
or between
two near orthogonal
circles,
with u aud u, for
centres, are proportional to the (w.o ordinates, tv, t'v'.
(65.) Accordingly,
if
we draw,
as in Fig. 86, the near radius (represented by a
* This
follows, among other ways, from the general value
XXVI.
for the radius
of curvature of the hodograph, with any law of central force ; which value was geometrically deduced, as stated in the Note to page 720, compare the Note to page
719, by the present writer, in a Paper read before the Royal Irish Academy in 18-16,
and published in their Proceedings. In fact, that general expression for the radius
of hodographic curvature may be obtained with great facility, by dividing the element /d< of the hodograph (in which /denotes the force), by the corresponding
element cr ^df of angular motion
in the orbit.
CHAP.
DISTANCE OF COMET OR PLANET FROM EARTH.
III.]
733
dotted line from ii) of the first hodograph, and also the sniaU perpendicular uv,
u of the first orthogonal to the tangent UT, and terminated in
erected at the centre
Y by
the tangent from the near centre u the two
ITTY,
and THV, uu^Y, give the proportion,
CXXXVIII.
XT,
new
pairs of similar triangles,
TV = uu, UT
'
tht
not merely confirms what has just been stated (64.), for the case of the Jirsi
dograph, but proves that the four small arcs, of the two circular Jiodographs in
5.
86, intercepted between the
IT
ordinates already mentioned.
two near orthogonals, are
directly proportional to the
(^66.) But the time of describing any small hodographic arc is the quotient (32.)
that arc divided by the force; and therefore, by (63.), (65.), the four small times
And ^the harmonic

inversely/ proportional to the four ordinates.
een the two ordinates tv, t'v' of the^rs^ hodograph, does not vary
i
mean u'l bewhen we pass
or to ani/ other hodograph, with the same fixed chord mim', and the
orthogonal circles ; it follows then, geometrically, that the sum (33.) of the
two small times is the same, in any owe hodograph as in any other, under the condi-
the second,
Iiich
same
tions
above supposed
and that
this
sum
is
equal to the expression,
2M.uu'
OU'.UT-U'L
CXXXIX.
2M.ul7'.ui)
OU-L,M'.UT
which agrees with the formula LXXIIT.

(67.) On the whole, then, it is found that the Theorem of Hodographic Isochronism (38.) admits of being geometrically* proved, although by processes suggested
and sufficient hints have been already given, in connexion
(60.) by quaternions
:
with Figure 87, as regards the geometrical passage from that theorem to the wellknown Theorem of Lambert, without necessarily employing any property oi conic
sections.
420.
tion,
As
which
2i
of a cornet^ or
(1.)
Let
fifth specimen,
ilf
new
planet,
Write
from the
and
deduce by quaternions an equadetermination of the distance
earth.
also,
L..Ta = r, Tw =
so that r
shall
be the mass of the sun, or (somewhat more exactly) the sum of the
and let a and w be the heliocentric vectors of earth and
masses of sun and earth

comet.
we
assist in the
adapted to
is
tt',
T(w-o) = e,
are the distances of earth
distance from each other, and p
Then (comp. 419,
is
U(w-a)-^,
and comet from the
sun, while z
is
their
the unit-vector, directed from earth to comet.
I.),
* It
to have been nearly thus that the
appears from an unprinted memorandum,.
author orally deduced the theorem, in his communication of March, 1847, to the
in cases of invention, his own
Royal Irish Academy ; although, as usually happens
had involved principles and methods, of a much
previous processes of investigation
less
simple character.
734
II.
and
III,
with
Mr-Sa,
- a) = M(r-^-
I V.
M>2
D2(a>
I!?"
Bho = - Mwôj,
- Mzw^p,
w-^) a
D2a=-
D2. zp
[bOOK
= _ (a + zpy =
;.2
22
4.
_ 2zSap.
The vector
a, with its tensor r, and the mass M, are given by the theory of
(2.)
the earth (or sun) ; and p, Dp, D^p are deduced from three (or more) near observations of the comet; operating then on III. with S.pDp, we arrive at the formula,
Sp DpD^p
'
'
'
~
"
_rlM_M\_
which becomes by IV., when
z V73
spDpUa
cleared of fractions
and
'
radicals,
algebraical equation of the seventh degree, whereof one root
is
and divided by
2,
the sought distance*
an
z,
of the comet (or planet) from the earth.
As
421.
we
a sixth specimen^
sliall
indicate a method, suggested
by quaternions, of developing and geometrically decomposing the

disturbing force of the sun on the moon, or of a relatively superior
on a relatively inferior planet.
(I.) Let a, <T be the geocentric vectors of moon and sun; r, s their geocentric
AI the sum of the masses of earth and moon and S the
distances (r = Ta, s = Ta)
mass of the sun ; then the differential equation of motion of the moon about the
;
earth
may
be thus written (comp. 418, 419),

I.
if
be
still
the
mark
D2a =
M.0a + 5.(0o--^(<T-a)),
of derivation relatively to the time,

II.
and
â=0(a) = a-iTa-i;
so that <pa is here a vector-function of a, but not a linear one.
(2.) If
fall
we
confine ourselves to the term
back on the equation 419,
1.,
and
member of I., we
anew to the laws of undisturbed
in the second
Mâ,
so are conducted
relative elliptic motion.
(3.) If we denote the remainder of that second member by 7;, then tj may be
called the Vector of Disturbing Force ; and we propose now to develope this vector,
according to descending powers of

the quotient r:
(4.)
s,
The expression
III.
of the
for that vector
Ss-â-^
1843).
two bodies
is
1
{
5s-iff-i { 1
Compare the equation
tion, Paris,
(r
a), or according to ascending powers of

the earth.
moon and sun from
may
be thus transformed
- (1 -
(1
= Dâ Mâ
a(r->)-iT(l
acr"!)-!}
Vector of Disturbing Force
of the distances of
aa-"^)-^ (1
7;
- T-ia)-4
(Tom. I., p. 241, new ediby inspection of a globe, which
in the Me'canique Celeste
Laplace's rule for determining,
the nearer to the sun, results at once from the formula V.
CHAP.
that
lY. ..
is,
V.
if
?7i
=-
r]
&-i(r-i (^(T-â
Tji
+112
'\rm+ &c.
+ f a<T'0 = ^ri (" +

2a-^
35r2
VI.
735
DEVELOPMENT OF DISTURBING FORCE.
III.]
(a(Ta"i+ 2(r+ 6(ra(Ta"i(T-*)
73
3<ra(T-')
J72,i
+ 72,2+
?i,2
>72,3;
&c.
7i, i
the general term* of this development being easily assigned.

have thus a. first group of two component and disturbing forces, which
(5.)
We
Sr
are of the
same order as
-=
s^
as
a second group of three such
a third group of four
The first component
(6.)
forces,
same order
forces, of the
and so on,
of the^rsf group has the following tensor and versor,
VII...T,. = ^,
Uj/i,i=Ua;
it
is
therefore a purely ablati-
mn, acting along the

moon's geocentric vector em protious fierce
longed, as in the annexed Fi-
gure 88.
(7.)
mn', of
Fig. 88.
The second component

the same first group, has an exactly
triple intensity^
mn'= 3mn
and
its
di-
such that the angle nmn', between these two forces of the first group, is
bisected by a line ms' from the moon, which is parallel to the sun's geocentric vector
rection
is
ES.
(8.) If then
we conceive a
line
em' from the
earth,
having the same direction as
new line will meet the heavens in what may be called for the
moon ])i, such that the arc ])])i of a great circle, connecting it
the last force mn', this
moment & fictitious

with the true moon D
in the heavens, shall be bisected by the sun
0,
as represented
in Fig. 88.
(9.) Proceeding to the second group (5.),
we have by VI.
of this group,
VIII.
Tno,
~,
U,?2,
8s*
a line from the earth, parallel to this
new
may be called
1,
necting
it
a.
first fictitious sun,
with the true sun,
is
= Uatra-i =
_
force,
component
meets therefore the heavens in what
such that the arc of a great
bisected by the
moon
J),
as in the
circle,
01,
same Fig.
con-
88.
* Such a
general term was in fact assigned and interpreted in a communication
Irish Academy (^Proceedings, Vol. III., p. 514)
The development may also be obtained, although
June 14, 1847, to the Koyal

and in the Lectures, page 616.
of
for the first
by Taylor's Series adapted to quaternions. Compare pp. 427, 428, 430,

431 of the present work and see page 332, &c,, for the interpretation of such sym-
less easily,
bols as (ratj-K aca^^.
73G
III.
[l300K
(10.) The second component force, of the same second group, has an intensity ex= 2T)y2j i) ; and in direction it is parallel to the
actly double that of i\\Q first (Tj;2,2
sun's geocentric vector es, so that a line
drawn
meet the heavens in the place of the sun
The
(11.)
Âjrcf
in its direction
from the earth would
component of the present group has an intensity which is ^vqthe^rs^ component (Tt]2iZ = 5T?;o, i); and a line drawn in its
ci&ely five-fold that of
from the earth meets the heavens
direction
connecting these two
Qi 0^,
the arc
(12.) There
no
is
in a second fictitious sun
fictitious suns,
is
extending this analysis, and
difficulty in
Qo,
such that
bisected by the true sun
0.
this interpretation, to
subsequent groups of component disturbing forces, which forces increase in number^

and diminish in intensity^ in passing from any one group to the next ; their intensities, for
each separate group, bearing numerical ratios to each other, and their direc-
tions being connected
by simple angular
relations.
(13.) For example, the third group consists (5.) oi four small forces,
j?3, i,
by the
four
j/s,
Sr"^
of which the intensities are represented
bv -r
multiplied respectively
whole numbers, 5, 9, 15, and 35 and which have deVeciiows respectively parallel to
lines drawn from the earth, towards a second fictitious moon ])3, the true moon, the
first fictitious moon ])i (8.), and a third fictitious moon J)3
these three fictitious
;
suns lately considered, being all situated in the momenand the three celestial arcs, ])o]), DI)i, DiDa,
tary plane of the three bodies E, M, s
moons, like the two
fictitious
of apparent e/onâ<ion of sun from moon

above cited Fig. 88.
being each equal to double the arc

in the heavens, as indicated in the
J0
(14.) An exactly similar method may be employed to develope or decompose the

disturbing force of one planet on another, which is nearer than it to the sun ; and it
is important to observe that no supposition is here made, respecting any smallness
of excentricities or inclinations.
As a seventh specimen of the physical application of quatershall investigate briefly the construction and some of the
we
nions,
properties of FresneVs Wave Surface, as deductions from his princi422.
ples or hypotheses* respecting light.

(1.) Let |0 be a Vector of Ray- Velocity, and fi the corresponding Vector of
Wave- Slowness (or Index -Vector^, for propagation of light from an origin o, within
a biaxal crystal
I.
his
S/XjO
so that
=-
II.
S/t^p
and therefore
* The
present writer desires
to
own, respecting these or any
rival hypotheses.
III.
SjO^/i
0,
be understood as not expressing any opinion of

In the next Series (423), as an
eighth specimen of application, he proposes to deduce, from a quite different set of
physical principles respecting
Mac
light,
expressed however
still
Cullagh's Theorem of the Polar Plane
in the
language of the
intending then, as a
ninth and final specimen, to give briefly a quaternion transformation of a celebrated
equation in partial differential coefficients, of the first order and second degree, which
present Calculus,
occurs in the theory of heat, and in that of the attraction of spheroids.
BAP.
jjp
CONSTRUCTION OF FRESNEL*S WAVE SURFACE. 737
III.]
and
dfi
be any infinitesimal variations of the vectors p and

consistent with
/i,
unknown), of the Wave-Surface and its
e scalar equations (supposed to be as yet
Reciprocal (with respect to the unit-sphere round o), namely the Surface of WaveSlowness, or (as it has been otherwise called) the Index* -Surface : the velocity of
a vacuum being here represented by unity.

The variation dp being next conceived to represent a small displacement,
tangential to the wave, of a particle of eth^ in the crystal, it was supposed by Freslight in
(2.)
nel that such a displacement dp gave rise to an elastic force, say dt, not gener ally in
a direction exactly opposite to that displacement, but still 9. function thereof, which
function is of the kind called by us (in the Sections III. ii. 6, and III. iii. 7) linear^
and self-conjugate ; and which there will be a convenience (on account of its
connexion with certain optical constants, a, h, c) in denoting here by 0"i^p (instead
of (pdp)
so that we shall have the two converse formulae,
vector,
lY.
(3.)
The
= <}>d;
dp
y.
= (p-'^dp.
ether being treated as incompressible, in the theory here considered, so
that the normal component
ij.'^Sp.Se
of the elastic force
may
be neglected, or rather
suppresicd, there remains only the tangential component,
VL
n-^Yfi5e
= de-fi-iSfid{,
as regulating the motion, tangential to the wave, of a disturbed and vibrating particle.
(4.) If then
tangential to a
it
be admitted that, for the propagation of a rcciiVmear
wave
which the velocity
rjftraîon,
the tangential force (3.) must be

exactly opposite in direction to the displacement dp, and equal in quantity to that
of
is
Tî,
displacement multiplied by the square (Tp.-^) of the wave-velocity,

and VI., the equation,
VII.
= n'^dp,
i}>-'^dp-ii-^SiJids
or
VIII.
combining which with II., we obtain at once

equation of the Index Surface,
IX.
or
this
^p
we
have, by V.
(0-1-^-2)-! /t-lS/i^^
Symbolical Form
of the scalar
= S/it-(^-'-/x-!)->-i;
by an easy transformation,
X.
or finally,
XI.
= S/U0-1 (0-1 - jw-2)->/i-i

1 = S/i (p"^ - <py^fi
.
* This brief and

expressive name was proposed by the late Prof. Mac Cullagh
(Trans. R. I. A., Vol. XVIII., Parti., page 38), for that rfctprocaZ of the wave-surface
which the present writer had previously called the Surface of Components of
for various purposes: for instance, to pass from
Wave- Slowness, and had employed
the conical cusps to the circular ridges of the Wave, and so to establish a geometrical connexion between the theories of the two conical refractions, internal and exter~
his own methods had conducted him (Trans. R. I. A., Vol. XVII ,
pages 125-144). He afterwards found that the same Surface had been
otherwise employed by M. Cauchy {Exercises de Mathe'matiques, 1830 p. 36), who
did not seem however to have perceived its reciprocal relation to the Wave.
nal, to
Part
which
I.,
5 B
738
[bOOK
III.
while the direction of the vibration dp, for any given tangent plane to the wave,
i/^L
determined generally by the formula VIII.
'^^P~
That formula
(5.)
combined with the expression V.
for the displacement,
for the
elastic force resulting, gives
XII.
XIV.
if
and
dp=- (pvSfiSe,
{(p-ix^)v
= fi,
XIII.
XV.
.or
= - vSfiSs,
Si
- fi^y^fi,
(0
V being thus an auxiliary vector; and because the equation XI. of the index
gives,
XVI.
it
S/xu
follows that the vector w,
plane
=-
if
XVII.
while
1,
drawn
and
like p
/x
Yvdt =
by XIII.,
0,
from o, terminates on the tangent
and is parallel to the direction of the

The equations XIV. XVI. give,
to the wave,
(6.)
surface
elastic force.
- Sv<pv = 1, whence XIX.

= Sfidv = - Svdî,
XVII I.
fi^v^
v-^SfiSp
because dSfivÔ, by XVI., and ^Su^w = 2S(0u.û), by the self- conjugate pro;
pertj' of
comparing then XIX. with III., we see that + p (as being -i- every dfi)
.
has the direction of /Lt
XX.
which
.p->
last
v~^,
and
= -/i-v-i;
therefore,
XXI.
by
p-2
I.
and XVI., that we may
equation shows, by (5.), that the ray
is
perpendicular (on Fresners prin-
ciples) to the elastic force Se, produced
The equations XX.
(7.)
XXIII.
we have
by the displacement
and XXI. show by XIV. that
(p-2-0)u =
..
XXV.
XXVI..
see that
to
of the
X. and XI.,
= Sp0(0-p-)-ip-;
we can
(4.))
.0 = Sp-i(^-p-2)-ip-i;
XXVII..
l=Sp(p2-0-J)-'p;
return from each equation of the wave, to the corresponding
equation of the index surface, by merely changing p to

result will soon be seen to be included in one
Jiule
(p-2- ^)->p-i;
by XXII., the following Symbolical Form (comp.

Wave Surface,
by transformations analogous
and we
v=
dp.
therefore,
Equation of the
or,
XXIV.
whence
p-i,
write,
XXII. ..Spv = 0;
= ;i2_^-2.
more
fi,
general,
and ^
which
to 0"^
may
but this
be called the
of the Interchanges.
(8.)
The equation
XXV. may
XXVIII.
but under this last form

(19.), the
it

.
Sp (0
- p-2)-ip =
coincides with the equation 412,
Wave Surface may be derived from the auxiliary
XXIX.
Sp^p
XLI.
hence,
by 412,
or Generating Ellipsoid,
= 1,
by the following Construction, which was in fact assigned by Fresnel* himself, but
as the result of far more complex calculations:
Cut the ellipsoid (abc") by an arbitrary plane through its centre, and at that centre erect perpendiculars to that plane,
which shall have the lengths of the semiaxes of the section

ties
of the perpendiculars so erected
* See Sir John F.
W.
will be the sought
wave
the locus ofthe extremisurface.
Herschel's Treatise on Light, in the Encyelopoedia J/e-
tropolitana, page 545, Art. 1017.
CHAP.
(9.)
CONNEXION OF RAY WITH INDEX-VECTOR.
And we see, by IX.,
that the Index

Surface
may
similar construction, from that

Reciprocal Ellipsoid, of
the same plan,
739
be derived, by an
exactly
is, on
which the equation
XXX.. .Sp^-V = l.
(10.) If the scalar equations,
oted by the abridged forms,
ir
XXVII. and
XXXI. ..fp = l,
then the relations
II. III.
I.
XI., of the
XXXII.
and
wave and index
surface, be
= 1,
.. F/i
enable us to infer the expressions

(comp. the notation in
418, (2.)),
XXXIII...
XXXV.
XXXVI.
^Sid
and the Note
f^;
to that sub-article),
V being the same auxiliary vector XV. as before, and

tor, such that (by XXIV. XXVII. and IX. XV.),
XXXVII.
iDpfp = (p2 _ 0-l)-lp _ pSp (p2 - 0-)-V = -lO- (o2p,

iD^F/t = (/i2 yin pSfx (p,^ (p)-^p = -v-v^p;
XXXIV...
^=zZf^;
rhich (comp. 412, (36.),
^B
I
III.]
w=
(0-1
being a
= i>v
XXXVIII.
p2)- V
=
XXXIX.
Spu)
;
new
Spa>
auxiliary vec-
=-
also w dp by XII., so that (comp. (5.)) if w be drawn (like p, p, and v)

from the point o, this new vector terminates on the tangent plane to the index surface, and is parallel to the displacement on the wave ; also dp d = u} v.
whence
||
(11.) Hence, by
XXXIII. XXXV. XXXVIIL,
"'
^^''=r^'
and
^.=-(''-'+^)-'by XXXIV. XXXVI. and XVI.,
similarly,
so that, with the help of the expressions
xLi...-.-.=p+.-.,
XV. and XXXVII.
for
v and w, the ray-vec-
and the index-vector p are expressed as functions of each other: which functions ax& generally definite, although we shall soon see cases, in which one or. other
tor p
becomes partially indeterminate.

(12.) It
as follows:
now to enunciate the rule of the interchanges, alluded to in (7.),

any formula involving the vectors p, p, v, lo, and the functional
some of them, it is permitted to exchange p with p, v with a, and
is
easy
In
symbol
with 0-1
0, or
;
provided that we at the same time interchange dp with de (but not* gene-
when
rally with dp),
may
either dp or de occurs.
It is true that, in
be said
to
passing from
ally, in the present investigation,

is
conceived
II. to
III. (instead of passing to
XLIIL), we
But usu-
have exchanged not only p with p, but also dp with dp.
to
have a
dp represents a small displacement
definite direction,
tangential to the
wave
(2.),
which
whereas dp
740
[bOOK
For example, we pass thus from XX. to XLI., and conversely from the
from II. we pass by the same rule, to the formula,
;
III.
latter]
to the former
XLIII.
Spde
0,
which agrees by XVII. with XXII.
and, as other verifications, the following equations
XLIV.
.. Sp
= fiYfide
XLV.
,. de
(13.) The relations between the vectors
gure 89
may
= pYpdp
be noticed,
XLVI.
Sfidt
= Spdp.
be illustrated by the annexed Fi-
may
in which,
^'
XLVII. .op = p, OQ = /i,

ou = V, ow = w,
and XLVIir. .op' = -p-i,
oq' = -)u-i, ou'= -v'^, ow'=-oi"i
.
^V-
in fact
it
evident on inspection,
is
that
XLIX.
= ou
OP
ou'
op'
= ow
and the common value
= OQ
.
oq'
;
of these four scalar products is here taken as negative unity.
As examples
(14.)
ow'
of such illustration, the equation
XX. becomes
= qu';
p'o
XLI. becomes oq' = w'p; XXIII. may be written as w + p"' = p~^v, or as

p'w ou = p'o OP &c. And because the lines pq'u and qp'w are sections of the
tangent planes, to the wave at the extremity p of the ray, and to the index surface
at the extremity q of the index vector, made by the plane of those two vectors p and
:
ft,
we
while dp and St (as being parallel to w and v) have the directions of pq' and qp'
see that the displacement (or vibration) has generally^ in Fresnel's theory, the
;
direction of
ih.Q
projection of the ray on the tangent plane to the wave ; and that the
has the direction oiihQ projection of the index vector on the
elastic force resulting
tangent plane to the index surface : results which might however have been otherwise deduced, from the formulae alone.
(15.) It
may
and p from each
be added, as regards the reciprocal deduction of the two vectors
expressions,
L.
(by XLI.
other, that
/i-i
XXXVIIL, and XX. XVI.) we

and
=w-iVwjO,
LI.
p-^ = v~^Yvii
{i
have the
which answer in Fig. 89 to the relations, that oq' is the part (or component)
-L. ou.
perpendicular to ow and that op' is, in like manner, the part of oq
of op,
We
(16.)
have
LIL
also the expressions,

.
- /i-i =
which may be similarly
interpreted
LIV.
and
w-iVwi;,
LIII.
- pi = y-i Vi/w,
and which conduct
to the relations,
-(Vuw)2 = v2p-2=(uV-2 = Suw.
Hence, the Locus of each of the two Auxiliary Points u and w, in Fig. 89,
face of the Fourth Degree ; the scalar equations of these two loci being,
LV.
=
(yvcpvf + ^v<pv
0,
continues, as in (1.) to represent
while St
still
and
any
LVI.
(Vw0-w)2 + Sw0->w
a Sur-
is
infinitesimal tangent to the index surface,
denotes the elastic force (2.), resulting from the displacement Sp.
CHAP.
from which
of the
EQUATIONS OF WAVE, LINES OF VIBRATION.
III.]
it
would be easy
two reciprocal
The equations XII. XXII., combined with
(17.)
LVII.
= S(0->p
LIX.
the self-conjugate property of
LVIII.
or
dp),
hence (between suitable limits of the constant),
Sp0-'p = h*
= ^Sp^V
evert/ ellipsoid of the
form,
= const.,
and coaxal with the reciprocal ellipsoid XXX., being also

fimilar to it, and similarly placed, contains upon its surface what may be called a
Line of Vibration* on the Wave ; the intersection of this new ellipsoid LIX. with
which
is
thus concentric
wave surface being generally such, that the tangent at each point of that line (or
curve) has the direction of Fresnel's vibration.
the
(18.)
The fundamental connexion
stants, a, b,
c,
of the crystal,
is
(2.) of the function cp with the optical conexpressed by the symbolical cubic (comp. 350, L,
and 417, XXV.),
LX.
from which
it is
with the help
for those surfaces,
XXIX. and XXX.
ellipsoids,
give
1^,
deduce constructions
to
741
and
if
we
LXL
this /ormwZa
LXII.
Changing
LXIII.
the IFave
+ <?-2)
C0
+ &-2)
(0
c-2)
by methods already explained, that if e be any scalar,
write,
we have then
(19.)
(0
easy to infer,
is
E=(e-<?-2) (e-6-2) (e-c-),
of inversion,
(0 + e)-i = e2 _ e (^ +
+ 6-2 + c-2) - -2j-2c-2^-i.

the equation XXVI IL of the wave becomes,
then e to
= p-2 + a-2 + 6-9 + c-2 + Sp-i0p - a-H-^c'^Sp(p-^p
(j-2
p~%
therefore (as
otherwise
is
known) a Surface of the Fourth Degree
and
known), the Index Surface is of the same degree, its equation

(found by changing p, 0, a, b, c to fi, 0-^, w^, 6"', c-i) being, on the same plan,
(as is likewise well
LXIV.
= /i-2+
a2
62
c8
S/i-i
0-
V - a-^b^c^SfKpfx.
(20.) These equations may be variously transformed, with the help of the cubic
LX. in (p, which gives the analogous cubic in <p-^,
LXV.
for instance, another
LXVI.
in
which
it
may
(0-1+ a^) (0-1
52-) (^^-1 ^. c2)
form of the equation of the \^V3
= Sp0-2p +
(p2
be remarked that Sp0-2p
is,
a2 + 62 + c) Sp0-V - a^^c^
= (0-ip)2 < 0, whereas Sp0-ip > 0.
;
in LXVI., its value A* from (17.), we find

(21.) Substituting then, for Sp0-ip
that this second variable ellipsoid, with h for an arbitrary constant or parameter,
LXVII.
= (0-ip) + h^ (p2 + a2 + 62 4-
c2)
- a262c2,
contains upon its surface the saine line of vibration as the first variable ellipsoid
which involves the same arbitrary constant h ; and therefore that the line in
LIX.,
* Such
lines
of vibration were discussed by the present
writer,
but by means of
a quite difi"erent analysis, in his Memoir of 1832 (Third Supplement on Systems of

which was published in the following year, in the Transactions of the Royal
Rays),
Irish Acadeniv.
See reference in the Note to page 737.
742
III.
a quartic curve, or Curve of the Fourth Degree, as being the intersection
is
question
[bOOK
of these two variable but connected ellipsoids
and that the wave
itself is
the locus
of all such quartic curves.
The Generating Ellipsoid (Sp<pp =1) has a, b, c for its semiaxes {a>b>c
for any vector p, in the plane of be, we have the symbolical quadratic
(22.)
>
0)
and
(comp. 353, (9.)),
LXVIII.
(0
5-) (0
+ c-2) =
0,
LXIX.
or
- b-^c-^(l>-^ = ^ +
6-2
c-2
making then this last substitution for ^ + 6"2 + c^ in LXIIL, we find, for the section of the wave by this principal plane of the ellipsoid XXIX., an equation which
breaks up into the two factors,
LXX.
|0-2
= 0,
a-2
and
LXXI.
b-^-c-^Sp^-ip
whereof the ^rs^ represents (the plane being understood) a circle, with radius = a,
which we may call briefly the circle (a); while the second represents (with the same
understanding) an ellipse, which may by analogy be called here the ellipse (a) its
two semiaxes having the lengths of c and b, but in the directions of b and c, for
and ^ + c-2 = 0, respectively, so that this ellipse (a) is
which directions <p +b-'^=
merely the elliptic section {be) of the ellipsoid {(^bc), turned through a right angle
:
in its
own
plane, as
by the construction (8.) it evidently ought to be. And an exwhat indeed is otherwise known, that the plane of ca
actly similar analysis shows,
wave
cuts the
of ab cuts the
(23.)
in the
system of a
same wave
The
circle (b),
surface, in
and an
circle (c),
ellipse (6)
and an
and that the plane
ellipse (c).
circle (a) is entirely ea;/enor to the ellipse {a)
and the
circle (c) is
wholly intetior to the ellipse (c) but the circle (6) cuts the ellipse (6), in four
real points, which are therefore (in a sense to be soon more fully examined) cusps
;
(or nodal points') on the
say
po
+ pi,
and
wave
Wave- Cusps
surface, or briefly
which are drawn from the centre o
called Lines of Single
Velocity, or briefly
Rag-
and the vectors
to these four cusps,
may
p,
be
Cusp-Rays.
that these lines or rays must have

(24.) It is clear, from the construction (8.),
the directions of the cyclic normals of the ellipsoid (abc) ; which suggests our using
here the cyclic forms,
LXXII.
for the function
(pp
<p,
fusion, instead of the

(25.)
XXVII.
p.
Fv
or Svcp-^v
Sp^p =gp^
+ SXpX'p =
1,
X' being written, to avoid con-
to X', v to p, and g to g-p-^, in the expression 361,

equating the result to zero, and resolving the equation so
obtained, as a quadratic in
Wave,
LXXIII.
ellipsoid (8.)
of 357, &c., to represent the second cyclic normal.
Changing then p
for
and
=gp + VXpV,
and the generating
we
tind this
new form
of the
Equation XXVIII. of the
LXXIV. ..gp^=l+ SXpSX'p + TVXpTVX'p
the upper sign belonging to one sheet, and the lower sign to the other sheet, of that
wave surface. The new equation may also be thus written, as an expression for the
inverse square of the ray-velocity Tp, or of the radius-vector, say
LXXV...r-2 =
because,
by 405,
LXXVI.
Tp-2 =
c-2
a- 2
c-2
_^+-
rt-2 4-
r,
of the wave,
p \
cos^^^+Z^,],
(2.), (6.), &c.,
..a-2
= -5'-TXX',
6-2
= -<7 +
SXX',
c-2
= -^+TXX';
CHAP.
CONICAL CUSPS ON WAVE AND INDEX SURFACE.
III.]
and we have the
verification, for
LXXVII.
(26.) If
we
write (comp.
LXXVIII.
the equation LXIII. of the
a cusp-ray (23.), that
= 6-2^
r-2
or
if
or \'.
p\\\
p-2(l + Sp0p) +-26-V2Sp0-p,
wave takes
.
the form,
= a-2 + 6-2
fp
and we have the partial derivative (comp.
= const.
c-2
XXXV.),
+ Sp0p) - p-20p 4- -26-2c-20-lp

= p-3 (1 - Vp0p) + -26-2c-20-lp
JDpfp =p-3(l
= Tp = 6,
XXXI.),
=fp
LXXIX.
LXXX.
743
which gives by XXXIII. the expression,
LXXXl. ../i_ p-3(yp0p-l)-g-26-2c-0-i

p-2
'
4.^-2j-2e-2Sp^-ip
and therefore a generally

the ray p
is
(comp. (11.)) for the index vector
definite value
(27.) If the ray be in the plane of ac, then (comp.

.
and therefore by
/X
_
_
P-' ( S/O0-'P
j_2"^Sp^_,p
an expression which gives,
LXXXV.
(comp. (22.)),
not also on the ellipse
LXXXVIII.
that
is, if
/i
if
- a'c^) - (p'^ + b-^) ^-^p

_ ^2^2) + (p-3 + 6-2) ^2^2
'
definitely,
n = ~ p-i,
LXXXVII.
but not
is
LXXXL,
TYYTTV
^ V
that
when
LXIX.),
LXXXII.
^p + (a-2 + c-2) p + a-'^c-^-^p = 0,
- p^P'^p)
LXXXIII.
Yptpp = a-2c-Wp(p- ip = a'^c^ (Sp^-'p
whence
fi,
given.
LXXXVI.
= a2c2,
if
.
p-2
6-2
o,
Sp0-ip
the ray terminate]_on the circle (6), at any point which
(6);
and with equal
= - a-zc-^<p-'p,
if
LXXXVII.
the ray terminate on the ellipse
is
defijiiteness,
(6), at
but not
LXXXVI.
any point which
is
hold good,
not also on the
circle.
(28.)
The normal then
incides with the
normal
to the wave, in each of the
to the section,
made by
two cases
the plane of ac
for a
moment from
may
on the same plan be called the ellipsoid (6).

An exactly similar analysis shows that the wave
last mentioned, co-
and
if
we
abstract
the cusps (23.), we see that the wave is touched, along the circle
= 6, which we may call the
(6), by the concentric sphere LXXXVI. with radius
sphere (6) ; and along the ellipse (6) by the concentric ellipsoid LXXXVII. which
(29.)
cles ()
and
(c),
briefly called the
(c).
touched along the
cir-
concentric and similar ellipsoids, which
and
is
by two other concentric spheres, with radii a and c, which may be

and along the ellipses (a) and (c) by two other
spheres {a) and (c)
And by comparing
may by analogy be called the ellipsoids

LXXXVII. of the ellipsoid (6) with
the equation
form LIX., we see that the three
elliptic sections (a) (b) (c) of the
(a)
the
wave, made by
the three principal planes of the generating ellipsoid (a6c), are lines of vibration
the constant /t* receiving the three values, b-c^, câ^, a'^b*, for these three
(17.)
ellipses respectively.
744
But at a cusp the two equations
(30.)
LXXXIV.
the expression
no reason
this case
plane of
for
fi
LXXXVI.
LXXXVII.
and
takes the indeterminate form -
coexist,
III.
and
in fact, there is in
for preferring either to the other of the two values, within the
ac,
LXXXIX.
= - po-\
ix
XC.
^ = Ho,
XCI.
if
which po is the cusp-ray (23.), and the first value of

but the second to the ellipse (6).
in
(31.)
this.
[bOOK
fi
/uo
= - a^c-s^-ipo
corresponds to the circle,
The indetermination of [x, at a wave-cusp, is however even greater than

we observe that the equations LXXIX. and LXXX. give, for this case,
For, if
LXXXVI. LXXXVII.,
= a-2 + 6-2 + c-2, and
fpo
by LXXXIII.
XCII.
XCIII.
= 0,
Dpfp
for
= po,
shall see that if p be changed to po+ p' in the expression LXXVIII. for fp, and
only terms which are of the second dimension in p' retained, the result equated to zero
will represent a co7ie of tangents p', or a Tangent Cone to the Wave at the Cusp :
we
which cone
dition I.,
is
of the second degree, and evert/ normal p. to which, if limited by the conhere to be considered as one value of the vector p, corresponding to the
is
value Po of p.
(32.)
dex
And
it is
evident,
surface, that if
values of
p which
vo,
by the law
+ vi
(12.) of transition from the
are analogous* to the four cusp-rags
end of each such new
po,
must be a Conical Cusp on

the Conical Cusp (31.) on the Wave, which is in
line,
wave
to the in-
be the Lines of Single Normal Slowness, or the four
there
logous to
such cusps.
ipi
(23.), then, at the
the Index Surface, ana-
manner one of four
like
(33.) In forming and applying the equation above indicated (31.), of the tangent cone to the wave at a cusp, the following transformations are useful
:
xciv.
- (p -h p')-2 = - p-2 (1 + p-'p'y (1 + p'p-')-^

= - p-2 2p-2Sp'p-l + p-V'2 - 4p-6(Spp')2 + &C.,
-1-
p', and p, p being

without neglecting any
the terms not written being of the third and higher dimensions in
any two vectors such that Tp'< Tp (comp. 421,

terms, the self-conjugate property of
XCV.
S(p +
p')
(4.))
also,
gives (comp. 362),
^(p + pO = Sp^p + 2Sp>p +
Sp'0p',
with an analogous transformation for the corresponding expression in 0-'

cubic LX. in ^, or LXV. in 0-i, gives for an arbitrary p,
XCVL
XCVIL
and
therefore,
XCVIII.
among
while the
0(0 + a-2) (0 + c-^-)p^~ 6-2 (0 + a-2) (^ + ,,-2)^,

= - 62(^0 + a-2) (0 + c-2)p;
0-1 (0 + a-2) (0 + c-2)p
other transformations of the same kind,
(0 +a-2)2 (0
c-2)2p
= (a-2 - &-2)
(c-2
5-2) (^
+ a-2)
(J,
5-2)p.
* This word " analogous" is here more proper than " corresponding" in fact,
the cusps on each of the two surfaces will soon be seen to correspond to circles on
;
the other, in virtue of the law of reciprocity.
CHAP.
CIRCULAR RIDGES ON WAVE
111.]
We have
also for a
XCIX.
msp, the
^po =
C.
(34.) In this
-1-
c-2) hi
the equation of the tangent cone
way
745
values,
- (-2 c-2) po
XCIX'.
1 4 Spofpo = (fl-2 +
= fl-<C-*Spo0-2|Oo = -2i2c-* - (a-2 + c-2),
//o^
/uo
.
INDEX SURF.
A2sD
easily found to take the
is
form,
CI.
and
0=b*Sp'((p + a-2)
(^+
by operating with Dp' (comp.
to give,
CII.
= 64(^+^-2)
a;/i
c-^)p'- 4Sp>oS|0>o
(10.) (26.) (31.)),
+ c-2)p'- 2poSp>o -
(0
2/ioSp'poi
the scalar coefficient x being determined, for each direction of the tangent p' to the
wave at the cusp, by the condition I., which here becomes (31.).
CIII.
also,
by CII.,
&c.,
we have
Sjupo= Sjuopo
some
after
CIV.
CV.
=-
slight reductions,
xS/ipo= 2(62Sp>o + Sp'po)
a;S/Lt/wo
= 2(Sp')ito-iWo^Sp'po);
= 4(A2/io2+ l)Sp'poSp>o+4(poSp>o + AtoSp'po)'

a;2^2
= - 462 (Sp>o)2 + 4 (62;io2 - 1) Sp'poSp>o + 4/io2 (Sp'po)2
CVI.
but this last expression is equal, by CIV. CV., to - x^SfipoSfifio the equation of
the cone of perpendiculars^ let fall from the wave-centre o on the tangent planes at
the cusp, takes then this very simple form,
i
CVII.
so that
and
tfiis
cyclic
ju2
+ S/ipoSju/io =
cone of the second degree has the two vectors po and
normals (comp. 406, (7.))
of which the diameter
CVIII.
and
/uo
by the plane
it is cut,
at once for sides
CIII., in a circle^
is,
T (juo +
po-0
= (TAto^ - 6^^ = & (P'"^ ~ "^)^
and therefore subtends, at the centre
CIX.
- =
o,
tan-1
and
.
(^"^
- ^^J"
in the plane of ac, the angle,
62 (5-2
_ ^-2)^
(c-a
6-2)^.
po
is
(35.) And by combining the equations CIII. CVII.,

a small circle of the sphere,
ex.
/i2
= Sfjifio,
CX'.
or
S/i-
we
see that this circle (34.)
Vo =
which passes through the wave-centre, and has the vector
for a diameter, passing

/uo
- po-K
through the extremity of the vector
curve
a
This
circle
III.,
of contact of the plane CIII. with the suris, by
(36.)
the curve corresponds, by
face of which fi is the vector, because every vector p, of
to the one vector po of the wave ; it is therefore one of Four Circular Ridges on
also
(31.),
the Index Surface, the three others having ejwaZ diameters, and corresponding to
the three remaining cusp-rags, ~ po, pi, - Pi (23.); and there are, in like manner.
Four Circular Ridges on
CXI.
vo)
I'l
the Wave, along which
Spvo = -l,
= +l,
Spi/o
it is
êing the four lines introduced in (32.)
diameters, of these four circles on the wave,
CXII.
where CXIII.
T((To
+ J/o-i) =
CQ^-a'^c^tpVo,
is
touched by the four planes,
Spvi = -1,
;
also the
Spi/i
=+
common
l,
length of the
(comp. CVIII.),
(T<ro2-62)i==i-i(2_62)i (Ja_c2)i,
Svq<Xo
TvQ=b-\ and CXV.
CXIV.
5c
= -1
746
finnlly,
and
Vq'^
new
four
circles
diameter of the
are the
(Tq
and the
two values* of
the plane of ac, for the
p, in
first
HI.
of the
between these two vectors, or the angle which the
anffle
iu the
circle,
[bOOK
same
plane,
subtends at the wave-centre,
is
(comp,
CIX.),
CXVI.
tan-i
6-2 (^2
52)1 (^2
_ c2)i.
(37.) In the recent calculations (33.) (34.), the circle of contact (36.) on the
index surface was deduced jôm the tangent cone at a wave-cusp, as a section of a
certain cone of normals CVII. to that tangent cone CI., made by the plane CIII. ;
is a simpler, and perhaps more elegant method, of deducing and
same circlehy means oiits vector equation (comp. 392, IX. &c.), and
but the following

representing the
without assuming any previous knowledge of the character, or even the existence, of
that conical wave-cusp.
(38.) In general,
we
by eliminating the auxiliar}' vector v between XX. and XXIII.,
arrive at the following equation,
CXVII...(^-p-2)(;, + p-l)-l=p-l;
which holds good for every pair of corresponding vectors p and
And
index surface.
rector
ft,
of the
wave and
in general, this relation is sufficient, to determine the index-
when the ray-vector p is given because (p + e)"0 is generally = 0.

But when e is a root of the equation E= 0, with the signification LXI.
:
/t,
(39.)
E, then, by the formula

nate form
and
of inversion LXII., the
symbol
of
4 c)"'0 takes the indetermi-
((p
therefore, for every point of each of the three circles (a) (5) (c) of
CXVII. /at7
the wave, the ^rmwâ
(23.), that the value of
/x
to determine fi
although it is only at a cusp
becomes in fact indeterminate (comp. (27.) (28.) (29.)
(p
= (0o),
(30.) (31.)).
(40.)
At such a cusp
CXVIII.
(At
the equation
CXVII.
takes the symbolical form,
+ po->)"'=(^ + &"')'Vo"^=C/ô + />o-')-' + (î-&-'^)-^0;
recent signification XCI., and the symbol (0 +6"2)-i0 denotmg any

/to retaining
vector of the form yj3i if /3 be the mean vector semiaxis of the generating ellipsoid
its
XXIX.,
so that
CXIX.
(41.) "Writing then
for
Si3^/3
= 1,
(0 + 6-2)
/3
= 0,
T/3
6.
abridgment (comp. XX.),
CXX.
vo
= -{fjio + Po-'y\
the Vector Equation of the Index Ridge (36.)
is
obtained under the sufficiently
simple form,
CXXI.
and
V/3
(/i
P0-O-'
this equation does in fact represent a Circle
* It
circle
is
and
common
not
diflScult to
ellipse (fc),
tangent.
show that these
wherein the wave
is
+ ViSvo =
(comp. 296, (7.)), which
is
easily
are the vectors of two points, in which the

cut
the plane of ac, are touched by a
by
CHAP.
DIAM. OF CIRCLES, DIRECTIONS OF VIBRATION. 747
III.]
proved to be the same as the circular section (34.), of the cone CVII. by the plane
its diameter CVIII.
being thus found anew under the form,
cm.
CXXII.
v.ith
Tyo-i=*TV\\' = 6(6-2_a
the significations (24.) (25.) of X, \'; in fact
cxxiii.
po=
(c-2_6-2)i,
2)5
we have now
the expressions,
VQ=po-'(yx\'y\
b\]\,
vith the verification, that
CXXIV.
And by
(42.)
tion (comp.
(^
+ 6-2)vo = \SX't/o + VS\vo = 6-'U\ = -po~'we have
a precisely similar analysis,
CXVII.),
first
new general
the
and
for anj/ ttvo corresponding vectors,

p
rela-
ju,
CXXV...(^-l_^-2)(p+^-l)-l=^-l;
and then
in particular
CXXVI.
60 that finally,
if
(comp. CXVIII.), for
(p
we
= ir^ + h^y^ vo-i =
vo-')-^
CXXVIII.
still
(Sp^p =
first
1)
and
is
found (comp.
V/3(p + vo-O-'
is
62)10
to be,
CXXI.)
+ V/3wo =
0,
Twq-K
Thus
diagram (13.).
.
-f
thus found anew (comp. CXXII.), without any reference to cusps
(43.) Several of the foregoing results
CXXIX.
^0-0-^+ C^->
mean
the
CXIX.)
(37.), as the value of
last
vector semiaxis of the generating ellipsoid

the diameter CXII., of this circle of contact of the wave with the
(as in
plane CXI.,
((TO
XLI. CXX.),
.<oo=-((To+x'o-0-S
the Vector Equation of a Wave-Ridge
being
= Vq,
write for abridgment (comp.
CXXVII.
/3
p,
op = jOq,
if
we
may
new
we have
be illustrated, by a
suppose, in that Fig. 89, that
ov = Vq,
OQ = jUc,
whence
CXXX.
use of the
the values,
op'= pQ-',
&c.,
then the index-ridge (36.), corresponding to the loave-cusp p (23.), will be the circle which has p'q for
diameter, in a plane perpendicular to the plane of the Figure,
which
wave
is
here the plane of ac
at p, has the wave-centre
having also
and has
its
the cone of normals
fi (34.),
rests
for its vertex,
for a subcontrary section that
and
to the tangent cone to the
on the last-mentioned
circle,
second circle which has pq' for diameter,
plane in like manner at right angles to the plane of poq also if r and s
first circles, such that oes is a right line, namely
;
be any two points on the second and

a side
of the cone here considered, then the chord
pendicular to this last
line,
and has the
pr
of the second circle is per-
which answers
direction of the vibration Sp,
here to the two vectors p (= po) and p

because (comp. (14.)) this chord is perpendicular to /i, but complanar with p and p.
(44.) Again, to illustrate the theory of the wave-ridge (36.), which corresponds to
:
a cusp (32.) on
tiie
in Fig. 89, writing
CXXXI.
index-surface,
now
.
(instead of
OQ= Vfl,
we may suppose that

CXXIX. CXXX.),
OP=<7o,
OW=Wo
this
oq'
cusp
= Vo"',
is
at the point
&G.
then the ridge (or circle of contact) on the wave will coincide with the second circle
(43.), and the cone of rays p from o, which rests upon this circle, will have the/r*t
for
circle (43.) for a
sub-contrary section
also the vibration, at
any point R
of the
wave-
748
have the direction of the chord rq',
rulge, will
[bOOK
same kind
for reasons of the
III.
as be-
fore.
K and
(45.) Let
same
Fig. 89
k' denote the bisecting points of the lines pq'
then k'
and
the centre of the index ridge, in the case (43.)

the centre of the wave-ridge.
is
qp', in the
;
while, in
the case (44.), k is

(46.) In thQ first of these two cases, the point k is not the centre of any ridge,
on either wave or index-surface ; but it is the centre of a certain suhcontrary and
o for vertex which rests upon an index-ridge ;

the direction (43.) of a vibration ^po, at the wave-cusp
so that this cusp-vibration revolves, in the plane of the circle last
circular section (43.), of the cone with
and each of
its
chords
P corresponding
PR has
mentioned, with exactly half the angular velocity of the revolving radius KR.
(47.) And every one of those cusp -vibrations dpo, which (as we have seen) are
one plane, namely in the tangent plane at the cusp P to the ellipsoid
all situated in
(b) of (28-), has (as
by (14.) it ought to have) the direction of the projection of the

cusp-ray po, on some tangent plane to the tangent cone to the wave, at that point P
to the determination of which last cone, by some new methods, we purpose shortly
:
to return.
(48.) In the second of the two cases (45.), namely in the case (44.), pq' is a
diameter of a wave-ridge, with k for the centre of that circle, and with a plane (perpendicular to that of the Figure) which touches the wave at every point of the same
and the vibration, at any such point k, has been seen to have the
circular ridge
;
direction of the chord rq',
the tangent plane at
which
is
in fact ih.Q projection (14.) of the
ray or upon
to the wave.
(49.) And we see that, in passing from one point to another of this wave-ridge,
the vibration rq' revolves (comp. (46.)) round the fixed point q' of that circle,
namely round ihQfoot of the perpendicular from o on the ridge-plane, with (again)
half the angular velocity of the revolving radius kr.

(50.) These laws of the two sets of vibrations, at a
ctisp
and at a ridge upon the
wave, are intimately connected with the tioo conical polarizations, which accompany
the two conical refractions,* external and internal, in a biaxal crystal ; because, on
the one hand, the theoretical deduction of those two refractions
is
and was
and ridges
in fact accomplished by, the consideration of those cusps
on the other hand,
* The
in the theory of Fresnel, the vibration is
wn-iter's anticipation,
from
theorj',
of the
associated with,
:
while,
always perpendicular
two Conical Refractions, was
a general meeting of the Royal Irish Academy, on the 22nd of October, 1832, in the course of a final reading of that Third Supplement on Systems of
Bays, which has been cited in a former Note (p. 737). The very elegant experi-
announced
at
ments, by which his friend, the Rev.
Humphrey
Lloyd, succeeded shortly afterwards
On the Phenomena presented by Light, in its passage along the Axes of Biaxal Crystals, which was read
before the same Academy on the 28th of January, 1833, and is published in the
in exhibiting the expected results, are detailed in a Paper
First Part of Volume XVII. of their Transactions.

Dr. Lloyd has also given
an account of the same phenomena, in a separate work since published, under the
title of an Elementary Treatise on the Wave Theory of Light (London, Longman
and Co., 1857, Chapter XL).
same
CHAP.
FORMS OF EQUATION OF CUSP-CONE.
III.]
But
the plane of polarization.
to
749
into the details of such investigations,
we cannot
enter here.
(51.) It
and
p-i,
not
is
difficult to
formation^ for a7iy vector
cxxxii.
two other vectors, pi'

we have iha general trans-
into
show, by decomposing p
perpendicular and parallel to the plane of ac, that

p',
.
64Sp'(^
-2) (0
c-2)p'=(S/ioPop';';
the equation CI. of the tangent cone at a wave-cusp
therefore be thus
may
more
briefly written,
CXXXIII.
and under
mals from the cusp,

centre
(S/iopop')-'-4Spop'Sjuop';
this form, the cone in question is easily
proved to be the locus of the norCVII., which has /x for a side, and the wavewhile the same cone CVII. is now seen, more easily than in
to that other cone
for its vertex
from o on the tangent planes
(34.), to be reciprocally the locus of the perpendiculars

to the
wave
at the cusp, in virtue of the
new equation CXXXIII.,
of the tangent
cone at that point.

(52.) Another form of the equation of the cusp-cone may be obtained as folThe equation LXXIV. of the wave may be thus modified (comp. LXXVI.),
lows.
by
pi
the introduction of the two non- opposite cusp-rays, po
CXXXIV.
where
it
22&2c2
+ (a2 +
(53.) Changing then p to po
the
= po.
we
(comp. (31.)),
CXXXV.
+ (^ _ c?) Spop Spip

= +(a2-c2)TVpop.TVpip;
member vanishes, as well as the second, at
C2) 62p2
will be found that the first
cusp for which p
p'
= 6U\ (CXXIII), and
= 6U\':
find
S(3op'
+ p',
and retaining only terms ofjirst dimension in
an equation o{ unifocal form (comp. 359, &c.),
= + TVaop',
or
CXXXV.
(Yuopy +
(S/Sop')^
with the two constant vectors,
CXXXVI.
ao= (6-2-a-2)i (c-2_
6 2)^^^^
CXXXVI'.
/3o
= /io-|0o*';
CXXXV. or CXXXV. represents the tangent cone, with p' for

being positive for one sheet, but negative for the other.
(54.) As regards the calculations which conduct to the recent expressions for
o Poi it may be sufficient here to observe that those expressions are found to give
the equations,
and
this equation
side, S/3op'
CXXXVII.
2a262c2ao= (2- c2) poTVpoPi
= 2 (a2 + c*) b^po + (a - c) (poSpopi - b^pi)
CXXXVir.
2a^b^c^(3o
.
and
that, in
deducing these,
CXXXVIII.
together with the formula
CXXXIX.
we employ
.
Spopi
the values,
(55.) It is not difficult to
J^TVXr
62SXX'
7FTTV,
TXX"
XCIX., and the
^ (po pi) = -2
TVpopi
^"'^
TXX'
'
following,
- Pi)
(po
(po
+ pi) = - c-2
show that the equation CXXXV.
into the equation

tangent cone at a cusp, can be transformed
(po
+ pi).
or
CXXXV'., of the
CXXXIII. but it
;
750
may
III.
[bOOK
be more interesting to assign here a geometrical interpretation^ or construction,
of the unifocal form last found (53.).
(56.) Retaining then, for a moment, the use made in (43.) of Fig. 89, as serving to illustrate the case of a wave-cusp at p, with the signification (45.) of the newpoint k' as bisecting the line p'q, or as being the centre of the index-ridge and
conceiving a parallel cone, with o instead of p for vertex, and with a variable side
;
OT = p'
then the cusp-ray op (=Po
line ok'
(= i
(/Uq
- po"0 iPo)
plane corresponding
II
CXXXV.
and the formula
ing*
CXL.
which
cos k'ot
& focal line of the
is
o)
new
cone,
and the
the directive normal, or the normal to the director
^^
is
found
to
conduct to the follow-
= sin pok' sin pot,
Equation of the Cusp- Cone : or (more im-'

mediately) of that Parallel Cone, which has (as above) its vertex removed to the
called a Geometrical
maybe
wave-centre o.
CXL. may
(57.) Verifications of
one of the two right

(or of ac)
CXLI.
that
ot = pi=pq
by the plane of the
cut
is
figure
either
by assuming
is,
be obtained, by supposing the side ot to be
pz, in which the cone
lines, pi',
+ Pq'^\\ou,
CXLI'.
or
OT =
p2'
= /Oo + /*o"M|ow
easy to show, not only that these two sides, ou, ow, make (as in Fig. 89)
an obtuse angle with each other, but also that they belong to one common sheet, of
and
it
is
the cone here considered, because each
mal
makes an acute angle with the
directive nor-
ok'.
(68.) Another way of arriving at this result,

takes easily the rectangular form,
is
to observe that the equation
CXXXIII.
CXLIL
(Sp'(UAio+Upo))2
= (Sp'(Uô-Upo))M T;ioPo(SpX>oPo?;
the internal axis of the cusp-cone has therefore the direction of
U/tto
+ Up^,
that
of the internal bisector of the angle poq, while the external bisector of the
angle
is
one of the two external axes, and the third axis
f>fPo5/*oj ^^^ Sp'(U)Uo+

itvo sides, pi
U(0o)<0, whether
and pz, belong
p'
is
is,
same
perpendicular to the ju /a we
= pi', or=p2': and
(as above) to 07ie sheet, because each
therefore these
is
inclined at an
acute angle to the internal axis VfXo + Upo.

line of the parallel cone (56.) is fiQ,
(59.) It is easy to see that the second focal
or OQ ; and that the second directive normal corresponding is the line ok (45.), in
whence
the same Fig. 89
equation of the cone at o,
;
CXLIII.
cos
may
be derived (comp.
KOT = sin koq
sin
qot
CXL.)
with
this second geometrical
koq = pok'.
(60.) And finally, as a bifocal but still geometrical form of the equation of the
cusp-cone, with its vertex thus transferred to o, we may write,
CXLIV.
Any legitimate form

when treated by rules
(61.)
Sp0"ip,
stated
POT + L QOT = const.
of
anyone
=:
of the present Calculus
and exemplified, not only conducts
L wou.
oi the four functions,
to the
tpp,
<P''^p,
Sp<pp,
which have been already
connected forms of the three other
functions of the group, but also gives the corresponding forms of equation, of the
IFave and the IndeX' Surface.
CHAP.
RIDGES RESUMED, ORTHOG. TO LINES OF
III-I
VIB.
751
(62.) For instance, with the significations (32.) of vo and vi, the scalar funcwhich is = 1 in the equation XXX. of the Reciprocal Ellipsoid (9.),
tion Sp0~^f),
may
be expressed by the following cyclic form, with
that ellipsoid,
CXLV.
Sp0-ip
= - 6'^^ +
x/q,
vi for the cyclic normals of
- c2)62SvopSvip
(2
reciprocating which (comp. 361), we are led to a bifocal form of the function
Sp^p, which function was made = 1 in the equation XXIX. of the Generating Ellipsoid (8.), and is now expressed by this other equation (comp. 360, 407),
^^^^'^-
(;^^3^2
CSP^P +
6-'P^)
= (Svop)^- + (3v,py -
^-^ SrôpSnp
vq, vi being here the two (real) /ocaZ lines of the same ellipsoid (8.), or of its (imaginary) asymptotic cone.
(63.) Substituting then these forms (62.), of Sp^p and Sptp'^p, in the equation
LXIII., we find (after a few reductions) this new form of the Equation of the
Wave
CXLVII.
(2p2
(2
c3)
SvQpSvip +
a2
= (a2 _ c2)2 { 1 - (Si/op)^}
c2)2
{l-(Svip)2};
whence
it
CXI. touches
follows at once, that each of the four planes
the circle in which
by equating
it
cuts the quadric,
with
to zero the expression squared in the first
member
example, the Jirst plane CXI. touches the wave along that
of which on this plan the equations are,
CXLVIII.
Svop
+1 = 0,
the wave, along
2p3 + (
- c)
Svip
of
circle,
(2
c2)
which
is
found
CXLVII.
For
Vq, vi for cyclic normals,
or wave-ridge,
Svop
and because
CXLIX.
and
therefore,
^ (vo
+ vi) = - -2 (vo +
with the value CXIII. of
CL.
the second equation
(To
= - 2c20vo =
CXLVIII.
CLI.
p2=
vi),
(p
-vi)=- c-2 (vo -
vi),
o-o,
i ((^2
-1-
C^-) J/o
or
CX.)
represents (comp.
S(Top,
{vo
CLI'.
(3
c2) J/i),
the diacentric sphere,
So-Qp-'
1,
which passes through the wave-centre o, and of which the ridge here considered is a
The diameter of that ridge may thus be shown again to have the value
section.
CXII.
and
it
may
CLII.
(64.) It
is
= - p2,
Sj/opS<Top
was shown
surface, or at the
and
be observed that the
circle is
or
a section also of the cone,
CLir.
SvopStrop-^
=- 1.
in (17.) that the vibration dp, at any point of the wavep, is perpendicular to ^-ip, as well as to // by II. ;
end of any ray
therefore iangentialto the variable ellipsoid LIX., as well as to the
wave
itself.
easy to infer, that this vibration must have generally the direction of the
= 0, by XXXIX., but also Sa0-'p
auxiliary vector w, because not only S/xw
Hence
it is
Sp^-Ja>
= Spv = 0,
results at once
by
by XXII. and
XXXVIL
XXXVII.
Indeed, this parallelism of dp to
from XII.
(65.) If then we denote by 5'p an infinitesimal vector, such as fiSp, which is tangential to the wave, but perpendicular to the vibration dp, the parallelism dp w
\\
will give,
752
CLIII.
CLI V,
whence
new
for this
= fidp
6'p
SpS'p
-L
II
= 0,
d'Tp
because
p,
fioi
CLV.
or
0,
CLIII'.
[BOOK
.
Sp/iw
III.
= 0;
.Tp^r= const.,
upon the wave.
direction 5'p of motion
And
thus (or otherwise) it may be shown, that the Orthogonal TrajectoLines of Vibration (17.) are the curves in which the Wave is cut by
that is, by the spheres p^ + r^ = 0, in which the
Concentric Spheres, such as CLV.
(66.)
ries to the
constant for any one, but varies in passing from one to another.
(67.) The spherical curves (r), which are thus orthogonal to what we have
called the lines (A) of vibration, are sphero-conics on the wave ; either^ because each
radius r
is
such curve (r)
CLVI.
or because,
it is
by XXVII.,
on a concentric and quadric cone, namely,
situated
by XXVIII.,
is,
0=SpC^+r-2)-ip;
on this other concentric quadric,
CLVII.
Sp(0-i +r2)-ip.
(comp. LXXV.)) that, for any real point of the wave,

nor greater than fl2 j and that the squares of the scalar
(68.) It is easy to prove

r2
cannot be
less
than
c2,
semiaxes of the new quadric CLVII.
with two, according as r

(69.)
in algebraically ascending order, r^
are,
a^,
so that this surface is generally an Ay/3er&o?otc?, with one sheet ov
r^_c2;
62^
And we
>
see, at
or
< 6.
the same time, that the conjugate hyperboloid,
CLVIII.
which has two sheets or
one, in the
= Sp (^-1 + r2)-ip,
same two
cases,
r>h, r<b, and has (in descend-
ing order) the values,
CLIX.
for the squares of its scalar
XXIX.
so that the quadric
To form a
(70).
upon the wave,
CLX.
it
_ r,
_ r^^
c^
- r\
confocal with the generating ellipsoid

the conjugate of such a confocal.
is
semiaxes,
CLVII.
62
itself is
distinct conception (comp. (67.)) of the course of a curve (>)
may
(a)..r = a;
be convenient to distinguish
(/3)
.r
ih.Q five
following cases
<a, >6; {y)..r = b; (J)..r<6,>c; (e)..r
= c.
(71.) In each of the three cases (a) (y) (t), the conic (r) becomes a circle, in
one or other of the three principal planes namely the circle (a), for the case (a)
:
(6) for(y); and
(c) for (s).
(72.) In the case
(j3),
the curve (r)
is
one of dotible curvature, and consists of
two closed ovals, opposite to each other on the wave, and separated by t\).6 plane [a),
which plane is not (really) met, in any point, by the complete sphero-conic (r) and
;
perpendicularly, in two (real) points of the

while the same oval crosses also the
ellipse (6), which are external to the circle (b)
plane (c) at right angles, in some two real points of the ellipse (c).
each separate oval crosses the plane
(b')
(73.) Finally, in the remaining case (S), the ovals are separated by the plane
and each crosses the plane (b) at right angles, in two points of the ellipse (6),
which are interior to the circle (6) ; crossing also perpendicularly the plane (a), in
(c),
two points of the elHpse
(a).
(74.) Analogous remarks apply to the lines of vibration (A); which are either
the ellipses (a) (b) (c), or else orthogonals to the circles (a) (6) (c), and generally
to the sphero-conics (r), as appears easily
from foregoing
results.
CHAP.
fresnel's theory.
III.]
753
(75.) It may be here observed, that when we only know the direction (Uyn),
but not the length (T^), of an index-vector fi, so that we have two parallel tangent
planes to the wave, at one common side of the centre, the directions of the vibrations
Sp
two planes, according to a law which
differ generally for these
it is
easy to as-
sign as follows.
(70.) The second values of p. and dp being denoted by p^ and Sp^,

the equation IX. of the index-surface, these two other equations
we have, by
cLxi.
- /i-2)-i p
S/i (0-'
cLxr.
= S/i
of which the difference gives, suppressing the factor p^'^
CLXII.
CLXII'.
or
- p-^y^n
-p-i)-^p
S/i (0-' -)u,-2)-i (0-1
= S (^-1 - /z-2)- p
(^-1
p'^,
(0-1
- p-^y^p,
because (,<P~^p~^y^, as a functional operator, is self conjugate, so that p may be

transferred from one side of it to the other ; just as, U v = (pp be such a self-conju-
= Sr^p = Spcpv = Sp<p'^p,

we have the parallelisms,
gate function of p, then v^

(77.) But, by Vill.,
CLXIII.
^p
II
(0-1
- p-^y^p
CLXIV.
that
CLXIir.
by CLXII'., we have the very simple
hence,
is,
&c.
dp,
\\
(^-i
ix;^y^p
relation,
.8dpdp=0;
the two vibrations, in the two parallel planes, are mutually rectangular.
The following quite different method has however the advantage of not
only proving anew this known relation of rectangidarity, but also of assigning quaternion expressions for the two directions separately : and, at the same time, that
of leading easily to what appears to be a new and elegant Geometrical Construction,
(78.)
simpler in some respects than the
By
known
one, which can indeed be deduced
from
it.
principles of Fresnel's theory (comp. (3.)), the vibration {dp),

one tangent plane to the ivave, is situated in the normal plane (through /z),
(79.)
the
first
on any
which contains the direction (de) of the
ehistic force ; that is to say,
we have
the
Equation of Complanurity,
CLXV.
(80i)
We
have then, by
II.
CLXVI.
.Spdpde=^
0.
and V., the system of the two equations,

.
SpSp =
0,
SpSp<l>-^Sp
comparing which with the equations of the same form,

Si^r
= 0,
Srr0r =
orem
410, V. VI.
O,
we derive at once the following Construction, which
may
also be expressed as a
The-
".4< cither of the two points Q of the Reciprocal Ellipsoid XXX., the tangent
plane at which is parallel to that at the given point v of the Wave, the tangents to
the Lines of Curvature on the Ellipsoid are parallel to the tangents to the Lines of
Vibration on the Wave ;" namely, to one at that given point p itself, and to another
same side of the centre,

two others above mentioned.
at the other point p', on the

parallel to each of the
5 D
at which the tangent plane is
754
Thus
(81.)
one of the
for
each of the two points
lines of curvature at
and
[bOOK
III.
p, p' the line of vibration is parallel to
it is
what
evident, from
precedes, that the
other of these last lines has the direction of the corresponding Orthogonal
(66.) at
p or p' nor is there any danger of confusion.
:
(82.) As regards quaternion expressions, for the two vibrations on a given wavefront, the sub-article, 410, (8.), with notations suitably modified, shows by its formulae XIX. XXII. that we have here the equations,
CLXVII.
O^SfidpvoSpvi
= Sfi dp
and
CXVIII.
vq Svi dp
dp vi Si'o dp,
Sfi
.dpW VVfivQ VYfivi,
if Vq, vi be, as in earlier formulae of the present Series

422, the ct/clic normals of the
reciprocal ellipsoid, which are often called the Optic Axes of the Crystal.
" for
(83.) And hence may be deduced the known construction, namely, that
of wave-front, the two planes of polarization, perpendicular

two vibrations in Fresnel's theory, bisectihe two supplementary
direction
any given
respectively to the
and diedral angles, which the two optic axes subtend at the normal
and
or that these planes of polarization bisect, internally
to the front :"
externally, the angle be-
tween the two planes, jxVq and pvi.

(84.) It may not be irrelevant here to remark, that if p, and p, be any two index-vectors, which have (as in (76.)) the same direction, but not the same length, the
equation
LXIV.
CLXIX.
enables us to establish the two converse relations:

.
abcTp^
= <3p<pp)--
CLXIX'.
ahcTp,
= (S/i,(/./i,)-i.
(85.) Either by changing a, 6, c, 0, p to a"^, 6-2^ c-2, 0-i, p, or by treating

form LXIII., in (19.), of the Equation of the Wave, as we have just treated
form LXIV., of the equation of Index Surface, in the same sub-article (19.), we
that
if
p and
p^
be any two condirectional rays (Up^
CLXX.
CLXX'.
and
= (Sp^- V)
*.
or,
= S(p,0-ipJ-^,
(6c)-iTp
or,
(a5c)-iTp,
= Up),
the
the
see
then,
= (Sp^'V)^
=
a6cTp-i
(Sp^^-'pJ*.
abcTp-^
somewhat interesting geometrical consequence may be deduced from

(86.)
these last formulas, when combined with the equation LIX. of that variable ellipsoid^
Sp^'ip = h\ which cuts the wave in a line of vibration (Ji). For if we introduce this
/t* for
Sp^-'p, and write r, instead of Tp^ to denote the length of the second
symbol
ray p, the
first
equation
CLXX.
will take this simple form,
CLXXL
which shows at once that
r,
r^
= abch--,
and h are together constant, or together variable
and
Vibration on one Sheet of the Wave is projected into an

Orthogonal Trajectory to all such Lines on the other Sheet, and conversely the latter
into the former, by the Vectors p of the Wave :" so that one of these two curves would
therefore, that
"a Line of
appear to be superposed upon the other, to an eye placed at the Wave- Centre o.
(87.)
Avith
The
is represented by the equation CLVI.,

and as being a surface of the 5eco<f degree, it ought
one of the fourth, in some cu7've of the eighth degree ; or in
visual cone, here conceived,
some constant value of r
to cut the wave,
some system
which
of curves,
is
which have the product of
their
dimensions equal to
eight.
CHAP.
fresnkl's theory.
III.]
755
#
CLVI. with the
Accordingly we now see that the complete intersection, of the cone

wave, consists of two curves, each of the fovrth degree ; one of these being, as in
a complete line of vibration (A):
(67.), a complete sphero-conic (r), and the other
a
new
geometrical connexion being thus established between these two quartic
curves.
As
(88.)
additional verifications,
we may regard
the three principal planes, as
of the cutting cones ; for then, in the plane (a) for instance, the circle () and
the ellipse (a), which are (in a sense) projections of each other, and of which the
limits
latter
has been seen to be a
line
of vibration, are represented respectively by the two
equations,
CLXXII.
in
CLXXIl'.
and
.r=a,
bc
= h%
agreement with CLXXI. and similarly for the two other planes.
(89.) It was an early result of the quaternions, that an ellipsoid with
;
at the origin
its
centre
might be adequately represented by the equation (comp. 281, XXIX.,
or 282, XIX.),
CLXXIII.
or,
(tp
= K- t3,
pk)
> Tk
Ti
if
without ay restriction on the two vector constants,
this other equa-
by
k,
e,
tion,*
CLXXIir.
(90.)
Ellipsoid,
Comparing
this
we
we
lation (comp.
see that
336)
with
Sp0p=
(tp
1,
t2)2.
as the equation
XXIX.
(/c^
t)2
Sp^p =
{ip
+ Kp
+ 2Stp(cp
pic) (pi
(i2 -f k2)
done by assuming (comp. again 336) this cyclic form
is
CLXXV.
of the Generating
are to satisfy, independently of p, or as an identity, the re-
CLXXIV.
which
+ pfc)2 = (k;2 -
(k2
= l2)2 (Pp = (i2 + k2) p + 2VKpt

= (t - K)2p + 2tSKp +
for ^,
2k;Sip
or as in (24.) comp. 359, III. IV.,
^p=gp + VXpX;
Sp<pp
* This
equation, CLXXIII'. or
=gp^ + SXpX'p =
CLXXIL, which had
LXXII. LXXIII.
been assigned by the
author as a form of the equation of an ellipsoid, has been selected by his friend
Professor Peter Guthrie Tait, now of Edinburgh, as the basis of an admirable
Paper, entitled: "Quaternion Investigations connected with Fresnel's Wave-Snr-
which appeared in the May number for 1865, of the Quarterly Journal of
Pure and Applied Mathematics; and which the present writer can strongly re-
face,"
commend to the careful perusal of all quaternion students. Indeed, Professor Tait,
who has already published tracts on other applications of Quaternions, mathematical
and physical, including some on Electro- Dynamics, appears to the writer eminently
carry on, happily and usefully, this new branch of mathematical science:
fitted to
and likely
to
become
sors to its inventor.
in
it,
if
the expression
may
be allowed, one of the chief succes-
756
with expressions for the constants
which
X, \',
5-,
give,
[bOOK
III.
hy LXXVI., the following
values for the scalar semiaxes,*
CLXXVI.
.a = Tt +
6=-^^^;
T(t-fc)
T/c;
= Ti-T,f;
whence conversely,
CLXXVII...Tt =
(91.)
Knowing
Tk=^;
CLXXV.
thus the form
(t
of the function
&c.
which answers
<p,
in the
present case to the given equation CLXXIII. of the generating ellipsoid, there
would be no difficulty in carrying on the calculations, so as to reproduce, in connexion
with the two constants

Series, respecting the
t,
k, all
the preceding theorems and formulas of the present

the Index- Surface.
But it may be more useful to
Wave and
show
briefly, before we conclude the Series, how we can pass from Quaternions to
Cartesian Co-ordinates, in any question or formula, of the kind lately considered.
(92.) The three italic letters, ijk, conceived to be connected by the four funda-
mental relations,
i-^=r'
= k^=ijk = -l,
(A), 183,
were originally the only peculiar symbols of the present Calculus

they are not now
.50
tain general signs
much
and although
used, as in the early practice of quaternions, because cer-
of operation, such as
yet they (the symbols ijk)
may be
S,
V, T, U, K, have since been introduced,
supposed to be
still familiar
to a student, as links
between quaternions and co-ordinates.
We
(93.)
shall therefore
which the meaning and
merely write down here some leading expressions, of

seem likely to be at once perceived, especially after
utility
the Calculations above performed in this Series.
(94.)
.,*
The vector semiaxes
we may
(40.) (42.)),
of the generating ellipsoid being called a,
/3,
y (comp.
write,
CLXXVIII.
a = m, /3 =jb, y = *c
CLXXIX.
0p = a-iSa->/o + /3-iSi3-V + y-'Sy-ip = 2a-iSa-V = Sia^T;
=
=
CLXXX.
2
'S.a-^x^
CLXXXI.
Sp^-ip = 2^2x2
Sp0p
(Sa-V)2
CLXXXII.
(0 + e) p = 2a {a^ + e) Sa'ip
.
tion
* The
reader, at this stage, might perhaps usefully turn back to that Construcof the Ellipsoid, illustrated by Fig. 53 (p. 22 6), with the Remarks thereon,
which were given in the few
last Series of the Section II.
i.
13,
pages 223-233.
It
k,
of which the lengths are expressed by CLXXVII., are the three sides, CB, CA, Ab, of what may be called the
Generating Triangle ABC in the Figure; and that the deduction CLXXVI., of the
will be seen there that the three vectors,
three semiaxes,
ahc,
i,
k,
from the two vector constants,
i,
k,
with
many
connected
can be very simply exhibited by Geometry. The whole subject, of the equa= ic^ - 12 of the ellipsoid, was very fully treated in the Lectures;
(tp + pic)
results,
tion
and the
calculations
may
be
made more
the long but important Section III.
unnecessary to dwell more on
it
ii.
in
general, by the transformations assigned

6 of the present Elements, so that it seems
in this place.
CHAP.
III.]
GEN. LAWS OF REFLEXION AND REFRACTION.

CLXXXIII.
CLXXXIV.
CLXXXV.
or
CLXXXVI.
(^
= Tp2 =
if r
4-
"2
ey^p = 2a (a
then v
2ar2,
+ e)-> Sa-'p
= r-2((f> + r-2)-ip
= Spv = 2 -^?^ = -f?^ + -^^- +

^ -!!fl
r2_a2
r2_^2^ r2-62
r2-c2'
= - Spu) = - Spc^v = - Su^p
a;2
and the Index-Snrface may be treated
r2-a2
^2
r2-62
or obtained from the
similar!}',
f-z
Wave by
to their reciprocals.
As an
423.
a;2
r3_nr2
vestigate,
.for TFave,
=2
changing ahc
757
eighth specimen of physical application
we
shall in-
*
by quaternions, MacCullagh's Theorem of the Polar Plane
and some things therewith connected, for an important case of incidence of polarized light on a biaxal crystal: namely, for what was
called
by him the
case of uniradial vibrations.
(1.) Let homogeneous light in air (or in a vacuum), with a velocity f taken for
unity, fall on a plane face of a doubly refracting crystal, with such a polarization
that only one refracted ray shall result

city of the incident, refracted,
and
let p, p',
p" denote the vectors of ray-velo-
reflected lights respectively,
p having the direc-
tion of the incident ray^ prolonged within the crystal, but p" that of the reflected
; andi let ^' \iQ i\iQ rector of wave-slowness, ox i\iQ index-vector (comp.
these /owr vectors being all drawn from a given
422, (1.)), for the refracted light
ray outside
point of incidence o, and

(2.)
we have
being within the crystal.

Then, by allX wave theories of light, translated into the present notation,
like p',
p.',
the equations,
I.
IL
where v
is
..p2
p" =
a normal to the face
vpv-\
;
=
p" + p
V.
and
2t,
p"
with
whence
III...p" =
IV.
= S/p'=p"2=-l;
II'.
= ix' -
p,
also,
pS^-2/S-^;
h-p
f^'-p
if
IV'.
= i;-' V/t'p = V- >Vvp
- p = - 2vSpv-^ = -
2i/-iSpi/
* See
pp. 39, 40 of the Paper by that great mathematical and physical philo"
On the Laws of Crystalline Reflexion and liefraction," already referred

Note to page 737 (Trans. R. L A., Vol. XVIIL, Part I.).
t Of course, by a suitable choice of the units of time and space, the velocities and
as may be thought
slownessses, here spoken of, may be represented by lines as short
sopher,
to in the
convenient.
from the principles of

X These equations may be deduced, for example,
ghens, as stated in his Tractatus de Lumine (Opera reliqua, Amst., 1728).
Huy-
758
so that the three vectors, p,
p", terminate
fi',
cular to the face of the crystal
and the
on one right
[bOOK
which is perpendibetween the^r*^ and
line,
bisector of the angle
third of them, or between the incident and reflected rays,
III.
is
the intersection
of the
plane of incidence with the same plane face.

(3.) Let r, t', r" be the vectors of vibration for the three rays p, p', p", conceived to be drawn from their respective extremities
then, by all* theories of tan;
gential vibration,
we have
the equations,
VI. ..Spr = 0;
to
VIT. ..
which Mac Cullagh adds the supposition
perpendicular to the refracted ray
IX.
Sp'r'
VIIL
S/t'=0;
or,
Sp"r"=
(a), that the vibration in the crystal
is
with the present symbols, that
X.
whence
r'
||
Yfi'p',
when
the direction of the refracted vibration r being thus in general determined,

those of the vectors p' and p! are given.
(4.) To deduce from r' the two other vibrations, r and t",
Mac Cullagh assumes, (6), the Principle of Equivalent Vibrations, expressed here by the formula,
XL
r-r' + r" =
0,
which the fAree vibrations are parallel to one common plane, and the refracted vibration is the vector sum (or resultant) of the other two
(c), the Principle
in virtue of
of the Vis Viva, by which the reflected and refracted lights are together equal to the
incident light, which is conceived to have caused them
and (d), the Principle of
;
Constant Density of the Ether, whereby the masses of ether, disturbed by the three
lights, are simply proportional to their volumes : the two last hypothesesf being
here jointly expressed by the equation,
XII.
Sv (p72 -
p'r'2
+ pV'2) =
0.
(5.) Eliminating p" and r'' from XII. by V. and XL, r^ goes
with the help of 1. and 11'., the following linear equation in r,
XIIL..2S- = 1 +
T
a second such equation
VIIL, and attending
XIV.
and a third
is
obtained
to I.
.
if
Xlir.
.. v'
and we
find,
= u'-p';
by eliminating p" and r" by
III.
and
XL
from
VI. VIL, namely,
2SpvSAtV =
linear equation in r
* The
equations VI.
1^=1^,
Spp
Spv
off;
is
(p2 -p'-i-)
Spr'=
- Sp'v'Spr
given immediately by VI.
VIL VIIL
hold good, for instance, on Fresnel's principles
but Fresnel's tangential vibration in the crystal has a direction perpendicular to that
adopted by Mac Cullagh.
t In the concluding Note
(p.
74) to this Paper, Professor
Mac Cullagh
refers to
an elaborate Memoir by Professor Neumann, published in 1837 (in the Berlin Transactions for 1835), as containing precisely the same system of hypothetical principles
respecting Light.
But
there
was evidently a complete mutual independence,
researches of those two eminent men.

in the Proceedings of the R.
1.
A., Vol.
Some remarks on
I.,
pp. 232, 374,
this subject will be
and Vol.
II., p.
96.
in the
found
CHAP.
MAC CULLAGh's POLAR PLANE.
III.]
(6.) Solving then for r,
three linear
and
by the
rules of the present Calculus, this system of the
scalar equations VI.
XII L XIV., we
find for the incident vibration
the following vectoj' expression,*
VovV'
^^-
'^
= ITZ.
01^
759
^V'.
2Spv
2rSpv = r'Spv' - v'Spr'

"^
and accordingly
it may be verified by mere

inspection, with the help of VII. and IX.,
that this vector value of t satisfies the three scalar equations (5.). And when the
incident vibration has been thus deduced from the refracted vibration r', the
reflected
vibration r"
is
at once given by the formula XI., or by the expression,
XVL..r"=r'-r,The
(7.)
relation
XVII.
if
r,
XV'.
^v'tt
giv-es at
0,
once the equation of complanarity,
or the formula
XVIII.
li'
-p
\\\t,
-y
then a plane be anywhere so drawn, as to be parallel (4.) to the three vibrations

also to the line fi' - p', which connects two correspondr', r", it will be parallel
but this is one form of

ing points, on the wave and index surface in the crystal
enunciation of Pi-ofessor Mac Cullagh's Theorem of the Polar Plane, which theorem
:
is
thus deduced with great simplicity by quaternions, from the principles above sup-
posed.
For example, if we suppose that op and OQ, in Fig. 89, represent the reand the index vector p,' corresponding, and if we draw through the
p',
(8.")
fracted ray
PQ
line
a plane perpendicular to the plane of the Figure, then the plane so
drawn
(on the principles here considered) the refracted vibration r', and will
be parallel to both the incident vibration r and the reflected vibration t" whence
the directions of the two latter vibrations may be in general determined, as being
and then
also perpendicular TespQCtiyely to the incident and reflected rays, p and p''
will contain
the relative intensities (Tr^, Tt'2, Tr"^) of the three lights may be deduced from the
relative amplittides (Tr, Tr', Tr") of the three vibrations, which may them elves be
found from the three complanar directions, by a simple resolution of one line t into
two others, of which it is the vector sum, as if the vibrations were /or<?e.
(9.)
^'
The equations IT.
(= p +
i(=p
v),
of the three vectors p, v,

.
../-
* The expressions
Ur"
V. and XIII'. enable us to express the four vectors,
p" (= p 2j/-'Srp), and p' (= p -f r v'), in terms

v, which are connected with each other by the relation,
t(=p-v-iSvp), p" (= p 2v-iSi/p), and p' (= p + v v'),

+ 2S^p = Sf' (p + v), because XIX'.
Svp'= S {v' - v)p,
XIX.
XIX.
IV'.
v'^Svp),
of the incident
XV. XVI.
and
enable us to determine, not only the directions Ur,
also their amplitudes Tr,

reflected vibrations, but
Tr", or
the intensities Tr^, Tr"^ of the incident and reflected lights, for any given or assumed
or intensity Tr'2 of the refracted light,
amplitude Tr' of the refracted vibration,
determined the direction Ur' of the refracted vibration by means of the
after
having
formula X.
760
as in
XIIL, or because
ju'2
- p2 = S/aV) by
[bOOK
III.
and XIII'.; and with which t
I.
is
connected (VII. and IX.), by the two equations,
XX.
S(p+v)r' = 0,
XXI.
and
Sj^V
while r and r" are connected with the same three vectors, and with
Vin. XI.
tions VI.
system (comp. (5.)) of three linear and scalar equations in
XXII.
and
= 0;
.. S/)7-
2S^|oSrr
by the
rela-
= Sv' (p +
v) Svr'
r,
= Sv'p
2Sj^pSr'->r
therefore to the vector expression,
2rSrp =
equation, in
which
XXIII.
and conversely,
relations (9.)
XI.
XXV.
<u
XV.
no
difficulty in
deducing this
be any vector,
may
Vj/V{(|O-w)r-(p'-w)r' + (p"-aj)n"}r' = 0;
when w
is
thus treated as arbitrary, the formula XXIII., with the

p, p', p", v, v', /i', but without any restriction (ex-
between the vectors
cept itself) on r,
which
Ypv'r', as in
these or other transfomations, there is
By
(10.)
new
in
r',
XIII., which conduct, by elimination of r", to the following
r", is sufficient to give the
r',
r- r'+
r"
= 0,
and
XXIV.
two vector equations,

.
pr
- pV' + p"r" = ccv- +

y,
'
= ^v
{pT
p't' -^ p"t") = Si/vt\
and XXI...y
=S
{pr p'r'v + p"r")
and which conduct to the two scalar equations (among others),
XXVII.
Sfc
(pr
XXVIII.
and
p'r'
+ p"r")
S/p
(Spr
= 0, if XXVII'.
- Sp'V") = Srp'S/r'
S/cv
we now suppose the equations VI. VIII. IX. to be given,

by XXVIII. while, as a case of XXVIL, and with
IV. or IV'. of t, we have the equation,
so that if
VII.
tion
vi\\\
follow,
XXIX.
St (or
p't'
f p"t")
0,
the equation
the significa-
= 0.
And
thus (or otherwise) it may be shown, that the three scalar equations
(11.)
VI. VIII. IX., combined with the one vector formula XXIII., which (on account of
the arbitrary w) is equivalent to five scalar equations, are sufficient to give the same
direction of
t',
and the same dependencies of r and t" thereon, as those expressed by

XV. XVI. and therefore (among other consequences), to the for-
the equations X,
muliB XII. and XVII.

VI. VIII. IX. contain what may be called the Princi(12.) But the equations
of vibrations rectangular to rays); and the forple of Hectangular Vibrations (or
mula XXIII. is easily interpreted (416.), as expressing what may be termed the
Principle of the Pesultant Couple : namely the theorem, that if the three vibrations
be regarded as three forces, rt, r't', r"t", acting at the
(or displacements), r, r', r",
ends of the three rays, p, p', p'', or or, or', or" (drawn in the directions (1.) from
r't', r"t" (conceived as applied to a solid body), is equivalent to a single couple, of which the plane
is parallel (or the axis perpendicular) to the face of the crystal.
the point of incidence o), then this other system of three forces, rt,
CHAP.
PRINCIPLE OF EQUIVALENT MOMENTS.
III.]
(13.) It follows then,
and
we can
(II.),
(III.)
waves)
i\iQ
by (10.) and
infer all the following
(11.), that
76 1
from these two principles,* (I.)
Principle of Tangential Vibrations (or of vibrations tangential to the
(IV.) the Principle of Equivalent Vibrations (4.) ;

(V.) the Principle of the Vis Viva, as expressed (in conjunction with that of the
Constant Density of the Ether) by the equation XII. ;
(VI.) the Principle (or Theorem) of the Polar Plane;

And (VII.) what may be called the Principle of Equivalent Moments,^ namely
* The word "

Principle" is here employed with the usual latitude, as representing
an hypothesis assumed, or a theorem deduced, but made a ground oi subsequent
either
The principle (I.) of rectangular vibrations coincides, for the case of an

ordinary medium, with the principle (III.) of tangential vibrations but, for an ex-
deduction.
traordinary medium, except for the case (not here considered) oi ordinary rays in an
uniaxal crystal, these two principles are distinct, although both were assumed by
Mac CuUagh and Neumann. The present writer has already disclaimed (in the Note
to page 736) any responsibility for the physical hypotheses ; so that the results
given
above are offered merely as instances of mathematical deduction and generalization
attained through the Calculus of Quaternions.
t In a very clear and able Memoir, by Arthur Cayley, Esq. (now Professor
Cayley), "On Professor Mac Cullagh's Theorem of the Polar Plane," which was
read before the Royal Irish Academy on the 23rd of February, 1857, and has been
Academy (pages 481-491), this name

given to a statement (p. 489), that "the
equal to the sum of the moments of Rt and
printed in Vol. VI. of the Proceedings of that
"principle of equivalent moments,"
moment
of R't' round the axis
P"t" round the same axis"
AH,
is
is
AH
the line
being (p. 487) the intersection of the

plane of incidence with the plane of separation of the two media, that is, with the
face of the crystal: while Rt, R't', R"t'' are lines representing (p. 488) the three
;
vibrations (incident, refracted,
and
reflected), at the
AR", which are drawn from the point of incidence A,

within the crystal.
the sign of the
the line
ing
it
And
moment
of
ends ot the three rays

so as to
lie,
in fact, if this statement be modified, either
R"t"
(p.
A R, AR'
all three (p.
487),
by changing
491), or by drawing the reflected ray
AR",
like
or"
of the present investigation in the air (or in vacuo), instead of prolongbackwards loithin the biaxal crystal, it agrees with the case XXIX. of the
more general formula XXVII., which
is itself
the Principle of the Resultant Couple.

ject obliged
him
to do,
what seemed
to
included in what has been called above
In venturing thus
him
to point out, as the sub-
to be a slight inadvertence in a
Paper of
such interest and value, the present writer hopes that he will not be supposed to be
deficient in the admiration (long since publicly expi-essed by him), which is due to the
vast attainments of a mathematician so eminent as Professor Cayley.
Notes (so far), was copied and sent

to a later Paper by Mac Cullagh (read December 9th, 1839, and published in Vol. XXI., Part I., of the Transactions of the Royal Irish Academy, pp. 17-50), entitled "^ Essay towards a
Since the preceding Series 423, including
its
to the printers, the writer's attention has been
drawn
Dynamical Theory of crystalline Reflexion and Refraction
;" in
which there
is
given
at p. 43) a theorem essentially equivalent to the above-stated "Principle of the
5 E
762
theorem that the Moment of the Refracted Vibration (rV)
[bOOK
is
Moments of the Incident and Reflected Vibrations (rt and

any line, which is on, or parallel to, the Face of the Crystal.
the
equal to the
III.
Sum of
r"t"), with respect to
that the Author had

[It appears by the Table of luitial Pages (see p. lix.),
tended to complete the work by the addition of Seven Articles.]
in-
Resultant Couple," but expressed so as to include the case where the vibrations are
not uniradial, so that the double refraction of the crystal is allowed to manifest itself.
Mac CuUagh speaks, in his enunciation of the theorem, of measuring each ray, in the
direction of propagation
tion
of the
which agrees with, but of course
anticipates, the direc-
reflected ray, adopted in the preceding investigation.
that subsequent experiments,
by Jamin and
The
writer believes
others, are considered to diminish
the physical value of the theory above discussed.
much
MAR 19
BINDINQ SECT.
1981
^2>
PLEASE
CARDS OR
DO NOT REMOVE
SLIPS
UNUVERSITY OF
THIS POCKET
TORONTO
LIBRARY
Hamilton (Sir) Williajn Rowan

Elements of quaternions
QA
257
H35
physical
FROM
&

William Rowan Hamilton William Edwin Hamilton Elements of Quaternions PDF

Încărcat de

Informații document

Descriere originală:

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

William Rowan Hamilton William Edwin Hamilton Elements of Quaternions PDF

Încărcat de

Drepturi de autor:

Formate disponibile

H2>

WILLIAM ROWAN HAMILTON,

FELLOW OK THE AMERICAN SOCIEIT OF ARTS AlTD SCir.NCKS;

AND ROYAL ASTRONOMER OF IRELAND.

EDITED BY HIS SOX,

WILLIAM EDWIN HAMILTON, A.B.T.C.D., C.E.

RIGHT HONOEABLE WILLIAM EAEL OF ROSSE,

Will no instructions were

to the publication of his Writings,

for his late

Elements of Quaternions," which, but

in all their completeness, in the

hands of the Public.

brother, the Rev. A. H. Hamilton,

named Executor, being

undertake the publication, deputed this

my triple duty in this matter First,

secondly, to the present public

thirdly, to succeeding generations.

had several con-

him on the subject of the

to be reserved for the

intimate, that after dealing with

he intended to advert to the great labour which

the writing of the

both mental and mechanical;

subsidiary and unprinted calculations, he wrote out

thus incurring the main expense, of the publication of

gratifying as it was, surprised me the less, when

former book, " Lectures on Quaternions,"

Preface to the " Lectures," which doubtless he would

have acknowledged, had he lived to complete the Pre"

He intended also, I know, to express his sense of

nuscript by my father, and which he might possibly

thus best fulfilled

part as trustee of the unpubElements," I now place them in the hands of

the scientific public.

William Edwin Hamilton.

[1.] The volume now submitted to the public is founded on

entirely new, and the present work can in no sense

Table of Contents^ by collecting into one view the headings of

remarks, especially as regards the

here a few intro-

has been thought convenient on this occasion

Three Books, each

develope one guiding conception or view, and to

treatise is divided into

relates to the Concept

of a Vector^ considered as a directed right

The Second Book introduces

ception of a Quaternion^ considered as the Quotient of two such

of Vectors^ regarded as constituting a Second Principal

ON VECTORS, CONSIDERED WITHOUT REFERENCE TO

FUNDAMENTAL PEINCIPLES EESPECTING VECTORS,

On Sums of Three more Vectors, ....

This short First Chapter should be read with care by a beginner

of the result, essential but not pecuUarX to quaternions, that (what

Section of the second Chapter of the Third Book, as III.

AND LINES IN A GIVEN PLANE,

On Linear Equations connecting two

ternions as Quotients of Vectors.

Curves in a given Plane,