A PRIMER FOR THE FIBONACCI NUMBERS; PART IX

MARJORIE BICKWELL
A. C. Wilcox High School, Santa Clara, California
and
VERWERE. HOGGATT,JR.
San Jose State College, San Jose, California

TO PROVE; F

DIVIDES F .
nk

For many years 9 it has been known that the n

divides F

Fibonacci number F

if and only if n divides m, n > 2. Many different proofs have

been given; it will be instructive and entertaining to examine some of them.

Some special cases are very easy* It is obvious that F,

divides F 2 , 9

for F 2 k = F k L k . If we wish only to prove that F^ divides E^ k when k is

a power of 2, the identity
F j. = F n L n L 2n
0 L.
4n Ln j .- l0
2Jn
2J n
suffices.
1. PROOFS USING THE BMET FORM
Perhaps the simplest proof to understand is one which depends upon
simple algebra and the Binet form (see [1]),

(1)

{l)

r,n
- P
a - p

where
a

= (1 + N / 5 ) / 2 ?

p = (1 -

N/5)/2

are the roots of x2 - x - 1 = 0.- Then

F =
n

nk

0 nk

-g
a - p

/ k

0k\

= (?-^4\ a - p /
529

(M) = P. M ,
k

530

A PRIMER FOR THE FIBONACCI NUMBERS

[Dec.

where

M = (n"1)k

<n-2V

^ - 3 ) ^ 2k

...

+ a k / J (n-2)k +

^(n-l)k ^

If M is an integer, F. divides F , , k f 0.
Since a/3 = - 1 , if (n - l)k is odd, pairing the first and last terms,
second and next to last t e r m s , and so on,
M = (a< n - 1 ) k + ^ n " 1 ) k ) + ( - l ) V n - 3 ) k + ) 3 ( n - 3 ) k )

+
L

where the n

(n-l)k

(-D 2 V n - 5 ) k ^ ( n - 5 ) k )

(-\-m

+ ( 1)2kL

(n-5)k

---

Lucas number is given by

L n = an + /3 n .

(2)

Thus, M is the sum of integers, and hence an integer.

If (n - l)k is even,

the symmetric pairs can again be formed except for the middle term which is
M)(n-l)k/2

^(n-Dk/2

again making M an integer. Thus, F,

divides F , , or, F
nKi

K.

divides F
n

if n divides m . See also H-172, this i s s u e .

2.

PROOFS BY MATHEMATICAL INDUCTION

O t h e r proofs can be d e r i v e d , s t a r t i n g with a known identity and using

m a t h e m a t i c a l induction.
(3

>

F o r e x a m p l e , u s e the known identity (see [2])

m+

n =

mFn+l

m-lFn

Let m = nk:
(4)

nk+n "

n(k+1) "

nkFn*l

nk-lFn

1971]

A PRIMER FOR THE FIBONACCI NUMBERS

Obviously 5

531

divides F
and F
divides F 0 , for F 0
= F L , so
n
n
n
AD.
zn
n n
that F M divides F.
for k = 1, 2* A s s u m e that F divides F . for i =
n
Kn
n
in
1, 2 , , k. T h e n , since F M divides F
and F divides F. , identity
J
n
n
n
kn
(4) f o r c e s F
also to divide F n 1 W so that F divides F,
for a l l p o s n
n\K"r/
n
Kn
itive i n t e g e r s kB
Another identity ? e a s i l y proved using (2) and (2), which l e a d s to an e a s y

proof by m a t h e m a t i c a l induction i s
(5)

L F
+ F L
= 2F
n m-n
n m-n
m

L e t m = nk,

yielding

(6)
If F

L F ,. -v + F L ,. -* = 2F . .
n n(k-l)
n n(k-l)
nk
n

divides F

and

divides

-v,
n n( k - l )

then

m u s t divide

F , ,
nk

A l e s s obvious identity given by S i l e r [3] also y i e l d s a proof by m a t h e m a t i c a l inductions

(("I)" + 1 - V ( E Finj = ( - D \ - Fn(k+1) + F n .

If F

divides F . for i = 1, 2 S 3 S * - , k 9 then F

i s a factor of t h e l e f t n
in
n
hand m e m b e r of (7). Since F divides F
and F divides F, , F
must
n
n
n
Kn
n
also divide F l l W so that F divides F.
for all positive i n t e g e r s k*
n(k+l)
n
kn
^

3B PROOFS FROM GENERATING FUNCTIONS AND POLYNOMIALS

Now l e t u s look for elegance*

Suppose that we have proved the g e n e r -

ating function identity given in [ 4 ] ,

F , xR
1 - L x + (-1) x'
n

.
k=0A

532

A PRIMER FOR THE FIBONACCI NUMBERS

[Dec.

Then, since the leading coefficient of the divisor is one and the resulting operations of division are multiplying, adding, and subtracting integers, the quotient coefficients F

/F

of powers of x are integers, and F

divides

F , for all integers k ^ 0.

Let us develop a generating function for a related proof that L
L,

divides

whenever k is odd. Applying (2) and the formulafor summing an infin-

ite geometric progression,

oo

OO

2-# L(2i+l)nX

oo

^an(2i+l).xi+j-i8n(2i+l)xi

i=0

i=0

i=0

l - a

+
2

P
1 - p

/^

(an + pn)(l - (-l) n x)

/ 2n ^ 0 2nv , / flv2n 2
n
1 - (a
+ p )x + (ap) x

L n ( l - (-l) n x)
1 - L 0 x + x2
2n
Then

(2i+Dn

i _

1 - (-Dnx

LJ ~~
": :
~~1
. A
n
1 - L 0 x + x*
i=0
2n
so that by the same reasoning given for the Fibonacci generating function
above, L / r t . l l X / L
(2i+l)n

is an integer.
&

Next, we prove that L/o^+i^ / ^

is true that
L

so that

(2k+l)n

nL2kn "

^s a*1 integer another way.

(_1)

(2k-l)n

Now it

1971]

A PRIMER FOR THE FIBONACCI NUMBERS

L

(2k+l)n _ _
L
2n
n

, ..n+1

533

(2k-l)n
L
n

Thus, we are set up to use mathematical induction since when k = 1, it is

clear that L

divides L . Thus, if L,Q1

(2k+l)n / n
indue tion0

so a n m t e

. /L

1
\KVx)TL

is a

ers

T n e

i s an integer,

then

proof i s complete by mathematical

We can carry this one step further, and prove that L

is not divisible

by L n if m f (2k+l)n, n > 2e
L

(2k+l)n+)

= L

nL2kn+j

+( 1)IlL

"

(2k-l)n+j^

1. 2, 3, , 2n - 1 .

Thus, given that some j = 1, 2, 3 , , o r 2n - 1 exists so that L, , ..v ..

is divisible by L , then by the method of infinite descent, L/ ? , -v . i s
divisible by L for this sa
same j = 1, 2, 3, , or 2 n - l . This will ultiJ
n
mately yield the inequality
I m

-n+j

which is clearly a contradiction since the L

n
in that range are all smaller

than L , n ^ 2* The same technique can be used on F n ^ and F

that F

divides F

to prove

only if n divides m , n ^ 29 (Since F 2 = 1 divides

all F , we must make the qualification n > 2.)

If the theory of Fibonacci polynomials i s at our disposal, the theorem
that F
case.

divides F

if and only if n divides m , n > 2, becomes a special

(See [5],)
If the following identity is accepted (proved in [5]),

n l 2 (-1>UlLm-(2i+l)n ) + ( - ^ m ^ p n '

? *

when |n| < |m|, n f 0, the identity can be interpreted in terms of quotients
and remainders; the quotientbeing a sum of Lucas numbers and the remainder

534

A PRIMER FOR THE FIBONACCI NUMBERS

[Dec,

of l e a s t absolute value being a Fibonacci n u m b e r o r i t s negative. The r e m a i n d e r i s z e r o if and only if e i t h e r F

= 0 o r F
= + F , in which
0
0
J
m-2pn
m-2pn
n5
c a s e the quotient i s changed by 1 . In the f i r s t c a s e ,
m i s an even multiple of n; and in the second,
odd multiple of n. So, F
That F

divides F

divides F

m - 2pn = 0,

m - 2pn = i n ,

so that

with

an

if and only if n divides m , n > 2.

only if n divides m can also be proved through

u s e of the Euclidean Algorithm [2] o r as the solution to a Diophantine e q u a tion [ 6 ] to e s t a b l i s h that

(F
v

, F ) = F,
v
n
(m,n)

(m > n > 2) ,
' 9

o r , that the g r e a t e s t c o m m o n d i v i s o r of two F i b o n a c c i n u m b e r s i s a F i b o n a c c i

n u m b e r whose s u b s c r i p t i s the g r e a t e s t common d i v i s o r of the s u b s c r i p t s of
the o t h e r two Fibonacci n u m b e r s .
4.

THE GENERAL CASE

A second proof that L

2,

divides L

if and only if m = (2k + l ) n , n ^

p r o v i d e s a s p r i n g b o a r d for studying the g e n e r a l c a s e .

(8)

i n d i c a t e s that L

The identity

^ = F ^-L + F L .
m+n
m+1 n
m n-1

divides L

, if L divides F .
m+n
n
m

Since

FQ
= L F
,
2p
p p
L

divides F 0 .
2p
F

whenever L

But since

2(k+l)p

2kp+2p

divides F,

divides F2>

T h e n , r e t u r n i n g to (8), if m = 2kn,

2kn+n

, it m u s t divide F / ,

by m a t h e m a t i c a l induction that L

2kpF2p+l

(2k+l)n

and we have p r o v e d

for all positive i n t e g e r s k.

divides L

2kn+lLn

2kp-lF2p

or,

2knLn-l

1971]

A PRIMER FOR THE FIBONACCI NUMBERS

divides L / 0 1 ,-v .
n
(2k+l)n
To p r o v e that L n divides

535

so that L

p r o v e that L n divides F
F

2n-j

nFn-j

only if

only if

m = (2k+l)n,

m = 2kn,

n > 2.
j

^ ^ - j '

n ^ 2,

we

We u s e the identity

= 1>2,

. . . , n - l

to show t h a t L

cannot divide F 0 ., If L divides F 0 ., then L m u s t

n
2n-3
n
2n-j
n
but L n ^ F n > |F__.|, c l e a r l y a contradiction. T h u s , L
di-

divide F .,
vides L m

if and only if m = (2k + l)n.

A proof of this s a m e t h e o r e m using

a l g e b r a i c n u m b e r s i s given by C a r l i t z in [ 7 ] ,
Now we c o n s i d e r the g e n e r a l c a s e . Given a Fibonacci sequence defined
by
Hi = p ,
1
^

H2 = q,
4
^

u n d e r what c i r c u m s t a n c e s does H

H ^
= H ^ + H
n+2
n+1
n
divide H

?
m

Studying a sequence such as

1, 4 ,

5 5 9,

14,

2 3 , 37,

60,

97,

157,

254,

411,

665,

1076,

eoe

quickly convinces one that each m e m b e r divides o t h e r m e m b e r s of the s e quence in a r e g u l a r fashion.

For example,

5 divides itself and e v e r y fifth

m e m b e r t h e r e a f t e r , while 4 divides itself and e v e r y sixth m e m b e r t h e r e after.

The m y s t e r y i s r e s o l v e d by the identity
H
m+n

= F
H + F H
* m+1 n
* m n-1

If H

divides F , then H divides e v e r yJ m

t e r m of the sequence t h e r e n
m
n
after. F u r t h e r , divisibility of t e r m s of {H } by an a r b i t r a r y i n t e g e r p can

be p r e d i c t e d using t a b l e s of Fibonacci e n t r y points. If H^ i s divisible by p ,

then H. ,

i s the next m e m b e r of the sequence divisible by p , w h e r e

is

the e n t r y point of p for the Fibonacci s e q u e n c e . F o r e x a m p l e , if 41 divides

ER9

then 41 divides H n + 2 Q

and 41 divides H

+2Qk

since 20 i s the s u b -

536

A PRIMER FOR THE FIBONACCI NUMBERS

Dec. 1971

script of the first Fibonacci number divisible by 41, but 41 will divide no
member of the sequence between H

and H

2.

REFERENCES
1. Verner E. Hoggatt, J r . , Fibonacci and Lucas Numbers (Boston; Houghton
Mifflin Co. , 1969), pp. 37-39.
2. N. N. Vorobyov, Fibonacci Numbers (Boston; D. C. Heath and C o . ,
1963), pp. 22-24.
3. Ken Siler, "Fibonacci Summations," Fibonacci Quarterly, Vol. 1, No. 3,
October, 1963, pp 67-69.
4. V. E. Hoggatt, J r . , and D. A. Lind, "A P r i m e r for the Fibonacci Numbers: Part VI," Fibonacci Quarterly, Vol. 5, No. 5, D e c , 1967, pp. 445460.
5. Marjorie Bicknell,

!T

A Primer for the Fibonacci Numbers: Part VII, M

Fibonacci Quarterly, Vol. 8, No. 4, October, 1970, pp. 407-420.

6. Glen Michael, "A New Proof for an Old Property," Fibonacci Quarterly,
Vol. 2, No. 1, February, 1964, pp. 57-58.
7. Leonard Carlitz, "A Note on Fibonacci Numbers," Fibonacci Quarterly,
Vol. 2, No. 1, February, 1964, pp. 15-28.

[Continued from page 528. ]

Perhaps the most carefully wrought example is the introduction to the
first movement of the Sonata for Two Pianos and Percussion. Here the divisions, while not conforming to numbers of the Fibonacci series (0,1), are all
determined by the golden mean. Measures 2-17 (the first measure is simply a roll on the timpani) contain 46 ternary (3/8)units, the most appropriate
for study in a passage which contains both 6/8 and 9/8 measures. The golden mean of 46 is 28, which is the dividing line between the second and the
third statements of the theme, and the place where the theme becomes inverted. The golden mean of 28 is 17.3, the juncture of the first and second statements of the theme. The two cymbal notes further subdivide the first and