Sampling Plans For Inspection

rrat fo.
- Technical Heport
#13
Sampling Plans for Inepectio; by Variables

by
ormld J. Liebertaan and George J. Besnikoff
aplac Table V - Matter Table for Bornal and Tightened Inspection

for SampllQg riant Bt<d on the veraê Hanê - with the enclosed, table
Corrected Tabl V.
July 15. 1955
!
i
o
o
VA
t^
.*
o
p.
o
o
VA
NO
en
^_^
o
r-t
NO
VA
J*
VA
VT\
vO
CM
o-*.
CA
r-t
cn
O
CM
j-
\o
c
*%.
CM
C-
CM
o
NO
CM
CM
CM
vO
0
VCN
>;
0
l-l
N
r-t
ON~
VA
CM
o-
CM
o-
*-
d-
^
t-4
ON
v-"1
CN-
CM
ON
CM
CT
CA
vO
5v
S.
a>
o
o
cn
cn
o
o
vO
NO
3-
vO
>
t-i
cn
CM
CM
-*
C>
EN-
CM
O1
sn
fvi
CM
CM
CM
CM
vr\
O
CM
VN,
CN,
CM
CM
CM
a.)
A
C-
c >
ON
11
CN(H
ON
ou
en
-1
CM
-J- 1
r-i
en
c
*n
*
vO
CM
o-
o-
Ni"
r-i
r-l
r-t
VO
r-i
-^
cn
VO
vn
t-i
CM
vn
e-t
CM
O
r-l
*A
H
^J-
.3-
J*
l-t
rH
t-i
cn
oo
en
o
c>-
O
vr\
VO
<n
CM
3ou
vn,
l-H
o-
VA
J-
CM
CM
CM
r-l
r-i
VA
ON
NO
[N-
UN.
NO
vr>
CM
-4-
rH
ir4
l-H
ON
rv
Cvj
UN
ON
crj
OO
ou
ou
C^
vn
NO
CN-
oo
o
e
vn
OU
v\
ou
H
VO
t>-
VO
VO
O^
CM
C
o
O
CM
r-i
r-1
VA
r-i
1&
ON
l-l
CM
OU
tv
^J-
en
r-i
VA
ON
r--
OU
CO
Jr
O-
J*
VO
i-t
CM
t>
vr>
CN-
ON
Jr
ou
CU
O
VO
CNVA
VO
NC
1-4
o-^
CM
J*
CN-
VO
ao
VA
r^
t-
J-
t-i
<
!>.
C-
VO
^J
&
-*
UN
u^
J-
-3-
-S
--t
J-
en
ON
CM
CM
CM
ou
o
ou
CVi
CM
CM
CM
CM
CO
VA
VO
NO"
t-i
r-i
t-i
CM
CM
VO
CNCN
r-i
o
*%,
f^J^
p*^
^
\>r
rH
C^
VA
VA
VA
CM
CN-
cn
J-
VT\
-*.
J
-3-
en
C^l
en
t>-
r^v
CA
NO
cn
CO
t)
[>
u*^
.*.
Hi
r-l
r^
l-H
,0
.--
r-i
r-t
CU
CM
CM
CU
NO
NO
C^V
C">
r>
CM
CM
CM
CM
CM
CM
C3s
ON
VO
ou
CU
ou
O
ve
3-
r-i
r-i
CM
CM
r-i
fs SO
-*
m
^
xrx
CM
iT>J
i-H
r-i
r-i
en
CM
CO
Cn
r-l
!_
c~.
.*
NO
_H;
/-**
J*
VA
VO^S,
f*
r-i
cn
cn
tv
ON
r-i
cn
tn
1 'CO
at
CM
ON
r-l
P^
C^CJ
r-i
r-i
r-i
r-t
CN-
VO
*.
VO
CN-
r-i
ON
1-4
r-i
t-i
rH
CD
\f\
OO
VO
to
fv.
VA
eg Tz.
IN soo
! -
'
s
/
\S
(0
(4
o
o
*
CM
.*
O.*
CO
Vi *
r-t
CM
CM
OvJ
CA
Jt
vn
c~
ft
- e
a;
Ou
o
*
VC
xr\
VO
CM
en
vO
r\
c
VT\
NO
VO
en
r-t
r-t
CM
CM
r-t
VO
vr,
CM
ON
cn
t-i
CN-
i.O
CM
&
en
\T<
CA
VO
vr\
vrv
r-l
ON
<r\
^
o
r-l
CO
en
rn
r-l
NO
CM
CM
VO
CM
kf\
VO
O
VO
VO
CM
vrv
(V
\_
<">
CM
to
VA
VA
VO
l-l
VO
VO
CN-
CM
cz
VA
cu
CN
N*
\J
V
VO
i
I
CO
>CM
_,
ON
oo
vO
CM
cn
en
cn
CO
VO
en
en
o
cn
-.
ou
v\
VO
cn
VA
VA
CM
cn
en
in
o
t-i
VA
t-i
CM
CM
CM
cn
1
V
t> M
3 r-l -H
WO,
l-l
jj-
cn
<n
en
cn
cn
cn
cn
cn
en
cn
cn
CM
CM
CM
CM
CM"
CM
CM
O
l;
VA
r-l
r-i
VA
i-i
CM
PH
CV
ON
vO
CM
3-
(^
ou
VT\
cn
C\
cn
CM
CM
CM
CM
CV|
O
r-t
vrv
VA
CM
o
en
m
n
J-
-a
u>
. -v
cn
cn
ON
CN-
-- ;
e
o
r-i
(X,
Q
f,0
ON
VA
o
I
VO
CO
o-
en
u
D ) V
N - *>
-rl +>
a U {
rH 1
o
cn
>
>
r-l
_.
. __ _.
S3
SAMPLING PLANS FOR INSPECTION BY VARIABLES
BY
GERALD J. LIEBERMAK AND GEORGE J. RESNIKCFF
TECHNICAL REPORT NO. 13

MARCH 12, 1954
PREPARED UNDER CONTRACT N6onr-25126

(NR-042-002)
FOR
OFFICE OF NAVAL RESEARCH
GERALD J. LIEBERMAN, DIRECTOR
APPLIED MATHEMATICS AND STATISTICS LABORATORY

STANFORD UNIVERSITY
STANFORD, CALIFORNIA
SampDing Plans for Inspection by Variables
By
Gerald J. Lieberman and George J. Eesnikoif
I.
Introduction,
In lot-by-lot acceptance sampling by attributes each item of a
sample drawn from a lot of manufactured items is classified simply as

defective or non-defective.
A random sample is drawn from the lot and
the lot is either accepted or rejected depending solely upon the number
of defectives in the sample.
When the classification into defective
arid non-defective is made on the basis of the measurement of a variable

quality characteristic, and the decision to accept or reject the let is
a function of these measurements (as opposed to the number of defectives)
such inspection procedures are known as sampling inspection by variables.
Since inspection by variables makes greater use of the information
concerning the lot than does inspection by attributes, whenever the
testing of the individual items involves measurement, variables plans
require smaller sample sizes to furnish the same degree of protection
than do attributes p.lens.
Variables sampling plans pertain to a single
quality characteristic, and it is usually assumed that measurements of

this quality characteristic are independent, identically distributed
normal random variables.
Such an assumption will be made throughout the
paper.
Sampling inspection by variables in divided into three categories,
known standard deviation plans, unknown standard deviation plans, and
- 2 -
average range plan3.
Known standard deviation plans are based upon the
sample mean and the known standard deviation.
Unkr?>wn standard deviation
plans are based upon the sample mean and the sample standard deviation.
Average range plans are based upon the sample mean and the ;. erage range
of k subgroups of r items (the sample size is then kr).
The purpose of this paper is to present a matching collection of
variables sampling plans based on known standard deviation, unknown
standard deviation, and average range.
Thcsa pi ins are classified
according to sample size and Acceptable Quality Level (AQL).
The prob-
ability of acceptance at the AQL was done to be consistent with

MIL STD 105A.
The plans presented below have the following properties which differ
from those plans already published:
1.
Each CC curve shown represents a variables sampling plan oased

on known standard deviation, unknown standard deviation, and
average range.
In other words, if the user chooses an OC curve,
he nas at his disposal the choice of fhe three types of plans,

guaranteeing, for all practical purposes, the same protection.
2.
The usual acceptance-rejection procedure in lot-by-lot sampling

inspection by attributes is to draw a random sample and accept
the lot if the number of defectives is sufficiently small.
This is equivalent to the procedure of estimating the percentage
defective in the lot by its best estimate, namely, the percentage
defective in the sample, anc' accepting the lot if this estimate
is siuall.
Similarly, sampling inspection by variables can be
viewed in the same manner.
In this paper, the acceptance-rejection
- 3 -
procedure is based upon estimating the percentage defective in

the lot-/ with acceptance occuring if and only if the estimate
is not too large.
3.
All existing sampling procedures for two-sided specification

limits have the disadvantage that the probability of accepting
a submitted lot with given percentage defective p does not
depend on p alone, but is also a function of the actual lot
mean n., or equivalently, on the actual division of the percentage defective into components lying above and bslow the
specifications U and In/
For this reason, it is not possible
to compute a single OC curve giving the operating characteristics

of a two-sided plan.
Such plans do not yield a constant prob-
ability of acceptance for a given percentage defective, but

rather a spectrum of probabilities.
Computing OC curves for
all possible divisions of p results in a band of curves.

twosided procedure presented is consistent with (2).
The ..
The
percentage below the lower specification is estimated and the

percentage above the upper specification is estimated.
The lot
is accepted if and only if the sum of these stimates is

sufficiently small.
This procedure ha3 the property that the band of OC curves
due to all possible divisions of the percentage defective is so
1/
The estimates used for the known and unknown standard deviation cases
are the uniformly minimum variance unbiased estimates of the percentage
defective in the lot.
2/
The two-sided regions for unknown standard deviation plans in Sampling

Inspection by Variables, by Bowker and Goode, McGraw-Hill, 1952, are
equivalent to those presented here. Both pieces of work stem from
research conducted at the Applied Mathematics and Statistics Laboratory,
Stanford University, during the summer of 1951.
- 4narrow as to be, for practical purposes, u single curve.

Furthermore, it then follows that the OC curve of the related-one-sided teat very closely approximates the OC band of
the two-sided test.
II.
General Inspection Criteria.

Associated with each inspection characteristic are the design
specifications.
If only an upper specification licit is given, the item
is considered defective if its measurement exceed? U,
If only a lower
specification is required, the item is considered defective if its

measurement is smaller than L.
If both upper and leper specification
limits are specified, the item is considered defec-ive if its measurement

either exceeds or is smaller than L.
If the percentage defective of a submitted lot is known, no sampling
inspection is necessary to determine whether or not the lot is to be
accepted.
If the percentage is sufficiently small, the lot is accepted,
otherwise rejected.
Since' such knowledge about the percent defective is
rare, a logical procedure is to estimate this percentage defective fr JO.

a sample, and accept or reject the lot on the basis of this estimate.
A sampling plan is then described as consisting of the sample size n;
a method of estimating the percentage defective, and a maximum allowable
estimated percentage defective allowing acceptance of the lot, p ,
If
an upper specification limit is given, the percentage above this limit,

$.., is estimated from the sample of sis,e n.
If p.. ip , l-hc lot is

A
accepted.
If a double specification limit is given, both p and p. are
computed.
If p p + p. s p , the lot is accepted.
In the ensuing aaerinns sampling plans based on known standard
- 5 -
deviation, unknown standard deviation, and average range will be presented.

These plans will be indexed according to code letter-^ and AQL.
The OC
curves for the corresponding plans (same code letter and AQL) based on
known standard deviation, unknown standard deviation, and average range
are essentially the same and are presented in Figures 1-16.
III.
Known Standard Deviation Pl*ns.

In this section, it will be assumed that the standard deviation
of the characteristic measurements is known.
<T
A complete set of sampling
plans indexed according to AQL and code letter is given in Table 1.

Operating characteristic curves for these plans are provided in the
appropriate figures.
Before presenting the general inspection procedure for known
standard deviation plans, some discussion about two-sided specifications
is in order.
For this case information is available about the percentage
defective in the lot before a sample is drawn.
There is a sufficiently
small value of -- which is a function only of the AQL that assures incoming quality worse than the AQL.
Consequently, before sampling, z-
should be calculated and the lot rejected (with no sample taken) if this
value is smaller than the minimum value of -^- on the grounds that the
incoming quality is poorer than the AQL.
found in Table 1.
Minimum values of - can be
Intuitively, this implies that the combination of
narrow specification limits and large standard deviation insure against

the submittal of good quality.
The code letters run from B to Q, wait ting A. Again this was done
so that there would be somewhat of a semblance between these tables
and the corresponding ones in MIL STD 105A.
- 6 -
The inspection procedure is as follows:

n
Draw a random sample of n items and compute x i-ln
.
a. For an upper specification limit compute C.t (-rr-)v
1.
2.
whore v
b.
Jn-l *
For a lower specification limit compute C' * (-~-)v

where v >lI en- .
c.
For a double specification limit, compute both

U " ^?)v and CL " ^^^y if ^ is rea'i'er than the Bdn
Otherwise, the lot is rejected before a sample is drawn.
C
3.
Enter Table II with C.1 and/or C' and read out pn and/or p.
i/
whichever is applicabler'
For an upper specification limit accept the lot if p.. p .
4.
For a lower specification limit accept the lot if p_ ip .

For a double specification limit accept the lot if Pp pT ^ p
Example.
The specified minimum yield point for certain steel castings Is
The standard deviation is known to be <P - 3000 psi.
55000 psi.
AQL plan is to be used with a sample size of 6.
A 1%
The yield points of the
sample specimens are

62000; 61000; 68500; 59500; 65500; 63900.
The following are computed from the data:
1. x - 63400.
3.
C - (^^00)(1.095) - 3.07.
From Table II p^ - .107$.
1/ It is shown in Section VI-3 tbat these estimate are the uniformly

minimum variance unbiased estimates of the true percentage defective.
- 7 -
4.
TV.
From Table I p
is 2.57$.
Hence the lot Is accepted.
unknown Standard Deviation Plans.

In this section it will be assumed that the standard deviation of
the characteristic measurement is unknown.
A complete aet of sampling
plans indexed according to AQL and code letter is given in Table III.
Operating Characteristics Curves for these plans are provided in the
The inspection procedure is as follows:
n
1.
x
i
il
Draw a random sample of n items and compute x
n
t (xx-x)2
and s
2.
i-1
n-1
U-x
a.
For an upper specification limit, compute C
b.
For a lower specification limit, compute (L -= .
c.
For a double specification limit., compute both
C - S* and CT 2=i .
Us
3.
Enter Table IV with C^ and/or C. and read out jL and/or

p. whichever is applicable^
4.
For an upper specification limit accept the lot if p\. i p .

For a lower specification limit accept the lot if p\ <L p
For a double specification limit accept the lot if jL + p. $ p

Example.
The specifications for electrical resistance of a certain electrical
component is 660 ohms and L - 620 ohms.
with a sample of size 10,
-'
A .4% AQL plan is to be used
The data are as follows:
It is shown in Section VI-4 that theses estimates are the uniformly

minimum variance unbiased estimates of the true percentage defective.
- 8 -
639, 640, 650, 647, 662. 637, 652, 643, 657, 649.
The follcvring are ssjapoted from the data:
1.
x 647.6
3.
From Table IV
s *i 65.38 - 8.04.
PL - 0
P - 5.31*
P = K + Pu " 5'31*'
4.
From Table III

p* - 1.3056.
Since p > p , the lot is rejected, furthermore, for possible review

board evidence, tha best estimate of the percentage defective in the lot
is 5.31^.
Furthermore, it is evident that all the trouble lies in the
upper specification limit.

7.
Average Range Plans.

In this section it will be assumed that the standard deviation of
the characteristic measurement is unknown.
A complete set of sampling
plans indexed according to AQL and code letter is given in Table V.

Operating characteristic curves for these plans are provided in the
The inspection procedure 1? as follows:
1.
Draw a random sample of u items and group the measurements

into k subgroups of 5 (where n 5k) with the exception of
sample sizes of 3, 4, and 7
For these sample sizes use a
single subgroup of 3, 4, and 7 respectively.
For each
- 9 -
2.
3.
subgroup compute the range (R.).

Determine the average
x
k
_
n
range R (/Z. & )A* Calculate x (T" x* )/n#
il x
i-1 x
U-c 1/
a. For an upper specification limit, compute JJl (-=-)h.=f
b.
For a lower specification limit, compute C (^r-)h.=/
c.
For a double specification limit, compute both
Enter Table VI with C" and/or C and read out p,T and/or
p. whichever is applicable.
JA
4.
For an upper specification limit accept the lot if p\. < p .

For a lower specification limit accept the lot if pL < p .
For a double specification limit accept the lot if /Sp.. Ap. < p4
Example.
The maximum temperature of operation for a certain device is specified
as 160 .
A U% AQL plan is to be used with a sample of size 15.
The
sample items have operative temperatures of

178, 175, 174, 158, 172
1^ - 20
177, 166, 172, 167, 163
\ - 14
174, 173, 162, 182, 170
R0 - 20 .
The following are computed from the data:

1. x - 170.87.
2b. CL - (1*0-170.87)
3.
PL
4.
From Table VI
2#394
1#21>
- 11.08*
p* - 9.61*.
j*
Since p. > p the lot is rejected.
h is a factor found in Table V which is a function of the sample

code letter.
tW JBUMTU Jatt fee hnd t TlgbtaMd XMBMUSB for ?U=s But on Kaon 81O4M4 ftBTisMi
f3M
1V
\V
.41
\nf
taammtmMm te.Uk* l*rl (na

d..l
6.
nt2
.0
".16
.15
.<*
6.M
.2*
40
144
tfl
s
I
-'
2 1.28 1.414
: .310 1414 2 .510 1414 3 1.94 1.225
3 .369 1.225 3 .968 1.225 3
.959 1.225 4 1.8 1.153
3 .079 1.225 3i.U4 1.225 4 .290 1.155
4 111 L.195 4 .161 1.155
5 .130 1.11 6 .230
6 US 1.09J 6 .234 1.095 1 .343 1.080 7 .507 i.oao .791 1.0% f 1.1 1.061 10 1.79 1.094
7 .141 1.080 7 .226 1.060 .330 1.069 <f .469 L.061 9 .760 1.061 10 1.14 1.054 11 1.73 1.049
.153 1.069 .243 1.069 9 .351 1.061 10 494|l.094 11 .76 UCH9 12 1.15 1.045 13 1.74
399 1.155 4 .681 1.155 5 1.05 1.113 5 1.76 1.11S
5 .296 1.11 5 445 1.118 6 .721 L09S 6 1.S4 1.095 7 1.75 1.080
uaes 6
.321 1.09J 6 .478 1.095 7 .756 1.080 8 1.U 1.069 1.80 1.069
X-OU
10 .141 1.094 11 .a7 L.049 11 .32* 1.049 12 46i|l.045 13 .721 L.041 U 1.08 1.03 16 1.62 L033
19 .134 1.087 20 .207 1.026 22 .296 1.024 23 423 jl.023 25 655 1.021 27 .980 1.019 30 147 1.017
27 JOB 1.019 30 .193 1.017 31 .2*3 1.017 34 .397L015 37 .615 LOU 40
.921 1.013 44 1:39 1.012
37 .130 1.014 40 .196 1.013 42 289 1.012 45 402 [l.Oll 49 .620 1.010 54
.920 1.009 99 1.39 1.009
14 .138 1.03 15 .211 1.035 16 .308 1.033 17 43tt.031 19 .673 1.027 21 1.00 1.029 23 1.51 1.023
.10
.069
40
.25
.15
1.00
.65
iooaptcbla ltaali.tr hmaim (UcktaMd la*paotloa)
ftbl IM
Sod*
tar
&Ua for anal tad Tlfktawd UspmHan far TUam IIMI ban SUsOmsi OmiMOm (,
-SS
1.00
1.90
^r^g^^tt
4.00
2.90
B p
lUMj
fit/)
10.00
331
19.00
ft I>
~rar
^50
vl
2.73 14X4
1414
6.11 1.4U
9.27 1414
17.74,1.22j
3 2X. 1.225
33.67 1.229
2.23 1414
!
3 .00 1. 4U
7.5* 1.225
10.79 1.225
15.60 1.223
22.97 1.195
31.01 1.155
2.7 1.225
'- 1.229
3.5
6.99 1.195
9.97 1.155
15.21 1.111
20.80 l.Ufl
.64 1.095
1946 i.(
16.64 1.069
2.58 1.159
3.87 1.155
6.09 1.11
8.92 1.U8
13.89 1.099
6 2.57 1.099
3.77 1.099
5.83 1.080
8.62 1.069
1.061
17.88 1.0*9
L.049
2.62 1.080
3.6 1.069
5.6 1.061 10 43 1.094
12.3$ 1.045
17.36 1.03
23.96 1.033
2.99 1.061
33 1.094
5.60 1.049113 8.13 1.041
12.04 l.OJJ
17.09 1.031
13.43 1.086
2.57 1.049
3.61 1.049
5.5 1.04115 .*; 1.035
11.
1.029
U.71 1.025
23.13 1.022
2.49 1.045
343 1.03
5.34 1.035
7.72 1.039
11.57 1,026
l*.i'3 1.022
22.63 1.019
14 2.51 1.03
3.94 1.035
5.29 1.029
7.80 1.026
11.56 1.023
16.37 1.019
22.57 1.017
2.35 1.1031 19 3
4.9 1.024129 7.34 1.021 29 10.93 1.011
2.19 L.023
4.6
0
T
2.00
15.61 1.016
a.77 1.013
016
4.99 1.01* 42 1040 1.012
14.87 1.010
20.90 1.009
2.12 U016 36
014** 4.59 1.012
.79 1.011155 10.17 1.009
14.5 l.C
204 1.007
2.00 010154
009 61 4.39 1.00
64 1.007 12 9.7 1.006
14.09
005 111 119.90 1.009
005127 24 .02
004 147 19 .34 1.003
1.50
.00$ n
4.34 1.006
2.90
4.00
646 1.009 169

6.90
ink
10.00
4Mptabl* 9jMlil7 I*l* (4*UM4 iaatto!*)
15.00
Bo
"1"
8
ft
en
8
8
J^fe^
CM
UN
en
en
en
ft
cS
8
STF"
rid -e>
o o o
>4li-B
-iJ u -o
o o
1
1
c
cv
3 a
8
3
8"
3
en
en
CM
t
S
tM
*
8
8
*
CM
1
8
u
o u
8
CO
t
c!
U\
a
8
Os
.
CM
VN
CI
-W) fc. -B
o o
a.
o
o o
ft
CM
Cl
CJ
oo
CJ
Is
I
s"**"
ci*&eP
8
CM
r1.
R
S
8
CM
3
CM
*
8
cv
CD
CJ
ft
1-1
<n
ft
s*>
<
3
-3 )6
4 A
ft
*
1
5
8
4
l-l
CJ
a?
en
en
8
CM
8
5
CJ
CJ
CJ
J 1
8
CM
CO
CO
r-l
l-l
CO
CM
C-
CM
CO
CO
CM
CO
CM
CO
CO
CM
CM
CO
in
l-l
P?
1A
CM
H
H
en
eft
e4
rvr.
CD
CO
CN
CO
CM
CM
3
8
CM
8
CM
CM
<n
o
CM
-it
8
3
CO
en
en
R
Tn
<n
1
8
8
&
CM
pp;
CM
i-J
en
H
3
3
CM
CM
CM
CM
CJ
reo
M3
n
O
CM
s
O
1*5
en
M?
I
l-l
CM
-I
8s
"A
o\
CM
OS
t
*
8
CM
!8
CJ
tn
8*
i 1
CM
C-
8 8
? V0
rn
--n
rn
CM
CM
*n
-a
* i
t %
CM
8
8
8 8
8 8
H
C"
CM
CM
en
*n
en
f*
8
8*
CM
8
OS
i
CM
ft
8
CM
CM
8
OS
8
Q
CM
o\
c5
CM
*n
1-
tv
CO
R
8
ft
8
en
tt
CM
p-
ft
8
3
8 8
* ,9,
CM
CJ
CM
CM
CM
el ft
H
CM
m
CO
CJ
U*t
*1A
CJ
CM
CM
CM
r-l
CM
CM
CJ
CM
CM
CM
s>V
CM
CM
CSI
vr
8
CO
ON
CM
CM
OS
v\
OS
en
o-
e?
l-l
CM
O
1 5 S
Os
SO
NO
*n
CM
CD
2 1
en
p-
CM
CM
CM
en
CJ
ft 3 3
3
CO
CM
e2
CSI
sQ
ja
-a
SO
en
CJ
ft 8
tn
.
8
Os
SO
CM
r-*
a
en
o
CM
CJ
o
*n
CM
CM
CM
CM
CM
CM
*IA
CM
CM
UN
3
8
R
8
n
<n
pO-
CO
ft 8
CM
en
>
l-l
en
1 *
CM
so
g
S
3 5
CM
ft
-=1
8
e4
ft
CC
CM
"T" "ST
8
p-
. * S *.
e^
pCM
OS
o*
i
CO
N0
8
CO
r4
r-l
H
8
4
|
3 1 3
a*
CK
SI
CM
* I i
t*-
CD
a a
a a a a a a \esa a a a a
CO
C*-
o.
CO
id
p-
ft
\A
en
a a a' a
p a
a a a a
R ft ft ft
R ft R
M3
en
o
o
CO
Os
a 1
*>
s
8
en
p\A
3 R CM. a & 8

eS 3 3 -3 3 3 3 3 3 -3 3 -3
CM
en
P"Ft K
p8 l3
3
8
8 si?
H
-I
A H
A l-l i-i H
en
CO
SC
CM
E
r-n
CO
R OS
c- o "5s co M3
* c p- SO
. P-
Po
3 8 8 8
8 8 o o o O
Os
*
ft* 3 3
3
5
H
4 e-i
en
cv
H
O
a a
en
pH
iH
GO
8*
8
ft
s 8
ft R
4
*
i 1
9
CO
ft
s it s
SO
l-l
.
8
R
a
CM
it
8*
CO
iH
so
t-
%o
<n
CM
CM
CM
MO
*n
psO
r-4
VO
o\
CM
en
3
8 S
CM
CM
ft
8
O
O
8
8
CM
pCO
OS
CM
&
CM
o\
u\
en
CO
CM
O
CM
en
<n
en
CM
SR
"ft
o
CM
Os
CM
CM
8
8
en
CJ
rj 3
a
3
3
A 3
f 1 3
8
c*\
*n
8*
ON
8
8
-a
S
CO
CO
p-
8
3
\A
CO
p-
o o
8 8 8 8 8 8 8
ft 3 3 3 3 3 3 3
*
8
f^ f** 1
ao
C3
i S
s 3
1 1
1?
3 3
8
CO
rTJ
p-3
CM
ft
R
8
8
en
3?
CJ
CO
(S
CM
en
8
8
in
o
p-
1 en
R
in
en
8
PI
8
8
p
*
8
Pi
1 t
eS
en
VO
1 i
ft
O
CD
\A
CM
CM
r-t
VO
en
O
en
CM
R.
R
o
R
c
CM
CM
p-
3
g 8 -3
8
&
8
3
8
8
8
8
en
o\
85
CM
**>
O
CM
CM
en
5!
CO
CM
ft
5
8
5
.
8
pCO
CM
ft ft
rCM
8
CM
OS
CO
CM
CM
CM
a ft
-3 h -p
ej
o o
csi
\e\
f-fcs*
.ôEP
&
CM
r-l
O
C3
P-
en
irr-
d i
CM
CM
P-
R*
TO
8
(-1
V
CD
en
o
IN
CO
CJ
CM
CO
en
CM
ft. s
A
l"i
en
CM
Si
p-
8
a
1-1
8
R *
8
en
8
S
1 i i
P-
.3
a s
CM
ON
en
a*
CJ
CO
en
3* s
CM
\A
. 8
l-l
1A
p-
CM
2$
I
CK
CM
CM
en
o
1
8
00
en
m
en
en
o
p-
CM
CD
CM
CJ
CM
en
o
-4 S -B
pCSI
*4 h -B
en
en
ft
UN
CM
s s
a 3
s
CM
CM
Os
ES
CM
1 1
CM
tO
8
S
CM
'ex
*o
CM
us
CM
CM
UN
CM
en
p-
CJ
*
8
CJ
8
3
<n
CO
<n
(V
IT
3
8 8
ft R a
CM
1 1
s 8
C0.003
00.003
co.ooU
OO.OOli
CO
<?
it
^
i^
1A
n
'n
ft 3 ^ ^
-^
R
O
en
-3
OS
os
<5
l-l
m
e>
Os
HI
~PT
CO
CM
rn
en
CM
CM
CM
3
CJ
CJ
CM
CM
Pi
li
Ft
VS
SO
8
CSI
in
3 3
k -f
a s
o 8 CM
3 3 3 3
3 3 3 3
"en
c\:
5 a a 8 CM c:
pi
^s toS
8
8
3
eo
o
ft*
a.
**
I
C
o
ft
CM
"P.
Si
5
*p.
*o.
ft
jI
8
ft
*D.
>
ft
s
a
si
8 5ft
ft
>
c\
ft
a
*\
>
ft
ft
ft
IT
ft
ft
St
ft
>
ft
ft
a a a a
a
si
*
NO
ft
ft
*
ft
>
ft
ft
ft
ft
ft
ft
3
8 8 a
*>
8 a 8
15
CM
ft
o
CM
CM
CM
CM
8
HI
-4
CM;
a
ft
ft
MB
s 5 5 8
-*
ft
*
*
CM
CM
CM
5
CM
CM
ft
o
3
cv
Pi
a 3
8
PI a a
3 a Si ft
&
r-
St
3 si Si Si
a 5 5 a a
-4
ft
ft
fv
-=0
ft
ft
>
r>
ft
cv
fft
p.
&
%
a*
m
4*
St
C?
CM
CM
CM
CM
ft
"2
3 8
i
I
4
8
4
ft
ft
JS
V
*o.
3
o
-IX
-v>
o.
*o.
lal
f-c
B 8
M9
ft
sd
>
*
o
so
tft
ft
ft
ft
o
5o 5o 3o o3
*
o
*
o
so
ao
^>
ft
ft
o
1o
1o
ft
o
ft
5 g
o o
CM
>
Si
A
CM
1o
ad
a*
3 3
o
ft
o
ft
>*
t*
8 8
ft
ft
IP
S 1
h.
C5
a*
(H
31
^>
*.
r-l
a 3o 3o o 1
fs
iX
ft
rt
8 R s a
1
1
r-4
4*
I.
m
m
4
.
!*
3-
11
assRaaaaaa
,RRR;1IJ.R8R*,
aRBRiisaaaa
s ^ t |: t t r K
asBSjiaaaas
K ^ A (I ft ^ s M
aRRScoasa
< f ^ S E M a B 8
aBRRRuaaaa
lei ii a a' A
nseea miij t
iiisssaaa
a'aRftsasaaa
sspââa?
a R R R a a a a a
^ 3 e, : i| > s it t r
asRRa'aaas
s s t
' M
'
3
3
e.
1 Grts
o
K
aaaaaaaaaa
iJRSRRSrt}*.
i i i ii it as
8.R^R^R^#^
i it a t t sa a
R<^RR8,SI}|tp
aaaaaaaaaa
aaassaaaa
qa,S;aRa*aR3
aaaaBsaaa
aaaaiaiiia
aaaaaaaaaa
aaasssRaa
H
M M ( ^ * ^
aaaaaassaa
R J ( S SI Jt
Sfjaaa'saaa
^SfiRR,t^R3q
a a a a a a a a a a
fcfc*JR3*a^R
4a;iRR8S
asssaaasss
***sssaaafl
( t l| < t 9 ( t 1
ssaasKcsss
p^j^pjP1*^:??;
sassiihia
Si******
aaaaizxa
aasssassaa"
R^RajiasiRP
aaaa MSS is
n a a a' a a a a a'
sssaaa'RRR
aasaasaaa*
ast3?,a*3*a
aassssssiaR
aqaasaaaaa
^Sj^^59^s|Ra
aassssaaaa
âs[^Ra85(t;a
aaassaaaaa
aaRRRaaa'a
SS888ftSt99
aassssaaaa
i| * M f ? s M
aaaasaaa'sa
aaaaaaaaaa
aaaasaa*
3 : };i| t
t -' ,J5 * . ? n * p
aaaaaaaia
aaaaaaaaaa
tsR^caâ,**
asasiâaaa
t^R5S^âj5
aaaaaaaaaa
iaR.aRârr
aaaaiaaa-ia
ij a l I
aaasiirtaaa
aaaazsuxaa
n is a ^ tu
aaaaaîaa
:pâ,p-4asp
asaaaiaaaa
ac*Rra*.8P
sassarsisa
aaaaaa
8Sa.R3S8*
6 ^ s * i; *
St, CIS 1
RRBaaaa"88
8 S R 8. $ R 8 R 3
ssaaasaaaa
SSSSS'SSsSS
aaaaaaaa'a
bttâsRS*
s s s ii i 3 j 3
*R8SR8SR8^
aaaaia
sasa*R6p
a a a a ' ' i
aaaaaaaaaa
aaaaaaaaaa*
asaaaasaaa
,*,
sr-=*~
o o o <s & B S 8 S
'H
^'
^*
,;
g
V
888S*38RR**a*.
R*!}RRR;*RRRR*'
*
R*3Rft6R*RRR*
o sssa*3apR6aR*
iiUUM t til
* 883>?6âp*aaaB.
U ^ t S f M
MAi;f|iiantBiii|t
R43RRRRRRRRR*
,
4 8*3*ft9lM8 S|t
R4!>RRRK*A*RR*
I
3
ii |
I.
uaaa
a' a t
tRB.câRâp
t a t t t it ^
Ra^âa!}3R
B; a i i i t
i, tK a, 3 ^ 4
i a t ' i i
RaR^6f!RR58
a j
*aq:?spRs*R
''
RRRRRftaaat:
I;ISII ^
ii '
*RXRSRRRRRR
M 1 5 . M
<^<?fti^*gj
Mid Mn M i
Ra^5a#q*^R
I a a '
RRRRftftRKR
M J U S U U
it RB a
RRRRRRaft
R * 3 t RRRRRRRK
RRRRRRRRRft
RRRasaaaR
R^SRRRRRRRRRR
***RRRRRRR
55S!*2M3t3
a
RRRRRtsaaa
5 3*. 5**$**.
till
SRRRRE,**
SantssRRRn.
RRRRRRRR
a*R3*R*Ra
ii*;;; a a ii
8aRffSRS*
s d li '
R3:q<R#^
uuituut it m
8 n % t ?. 3, it q * R * q
RHa. RRRRRRRRR
S22532S535
RRRRRRRRRR
( a i t
ja'SSfS.a.'ia.}*
tJ:RRaa
âRâaâ.*s
!}asâ^rta.es
RRRRRJ?*^
a t a , a, t
fi6R*'RRR*a
aa*RJjSjKR,RR
stiiii t H
sq?RRR*,3ta
t l t
6RRRRRSSS
SRR^RRSRSR
<
t t CM mi
t tmcit
RR^, C^R8.3R
KfilltltUS
tuuttut cu
RRRRRRXRRR
R$*!3!3SRRRRR
Pld* o-nÎJRRRf;, XR
RRRRRRRRHR
RRRRRRRRRR
J25S251S35
aaafiRRRRs
M
8 M M
RftBRR'aRRfcR
33*^;?^*^
RRRRRRRRRR
**M*!*i**
RttRRRKRKRR
i J. * 9 3 S * % 3 t
RRRRRRRRRR
R8.$RSSB8$R
RRRRRRRRRR
*
o
<
- s.
1
I
1
s
I
1
I
ft
*s
ft
ft
d
*
m
a
*,*
s a
3.
9R.a98RaRS
-*
in
w%
ft
ft
ft
irt
*A
ft
o\
t\
<r*"A-4-4-#4--4
9 9 fi it M H * o
**nn444*<44
ssP33a
tf\
tfS
TV
p*
v>
tft
VA
ft
ft
in
44''*4'444
a.
a 8. 8.
PR^sasstpa
*s
8.
-4
d t) 9 s 6 1 s
J;
d<4*4<4>-4d-4-4#s
^ 3 3 8
$^ p 3
^*4"4Apsft
in4t4<i'<t<f4n
8 8
8 8.
888888888
OOOOOOOOO
ooooooooo
R SRRURSRRR
* P 3
? ? 8.
^\
tf
88K3RSI>SJ{P
3.
RRPRqqaaRP
S
o
R
o*
3 8 8 8. 8 8 8. 8 8 8
ooo'ooooo'oo"
5o
8888888888
oddoododoo
So o8,
AAA*\A-AAtii*
<4
8 3 3 8 3) 8 S 4 8. *
4
*4,-(HiH>-4H<-t<-<,-<
a s
i -i
4
-*
S
a
A
if
* * 1^
1 3
1 *
K
ft
3 *
ft
A
or
A
8. 8
S
o
(5
v4
T*
3 5
O
ft
8|
3
R
3 2
R R 3
ci
ft
ft
a
A
ft
8
O*
a
d
8
d
8
d
8
d
8
d
8.
e
8 8
PS
a s
PS
PS
ft
a
PS
PS
PS
Pi
&m Kw
S P
(4
* 8
PS
PI
R
(4
8. * R
S P
<4
PI
3 f
apj
P*
P:
e4
*
1
W P*
8 3
P*
01
8 8.
PS
P
.4
to
if a
ft
at *
(4
8.
(4
(0
PS
8 8
*.
(4
3 X 3
a 8
R 5! s
ft ft ft
a ? 8.
8
A
PS
a a
p*
(4
<4
fi
(4
3
pi
r4
P*
et
TC
8.
(4
*4
. a
R S 3
4
A
SI
a.
"1 aA aft
R R
ft
s it
8
4
F\
ft
A
IT*
ft
R R
1^
8.
ft ft
R 3
*4 1**
ft*Sftft*flNWN
8
d
4 *>>
^ a P
3
*
-4 ^
a 3 5 a, *
Mnn(n4<t4444
R I 11
9 if R H
R S S 8
4 -*
os
8BR8R*3*
4 - 4
3 a a
8S*3
SS8RPSR3R
R. R. S 3
4 -* t -i
1 M S fi 1 i! ^ ;;
v>
s 5
4
*6
<1
p*
5,*3SRSPaa
3 a s
**<o)ft>AM\sft
tnn4M
- 1
I
s
8 qR5at8
*
M
an ae t i; t x
*J
8
d
P4
8. 8
8
o
8. 8
O
8
d
8. 8.
O
8
d
8
d
8
o
8 8 8. 8.
o d d o
8
o
8
o
8 8 8
8
o
p R R
8 8 3 S *
8
d
8 8
o
8. 8
r4
8. 8
1
I
or
fc
*
3
3
1
i1
R
A
aaagaaa'aa
919
M> ! 1*
aaaaaaaaa
na MH^t
a aaaaaaaaa
R
Jt
3
Oh
^M( it (mi; t a
9 9 j 9 S a a 8 '
t; i s >
a a
u 9
i i
a a a a a a a a *
9 9 n
a a a a' a" a a * *
aaaaaaaa
^ssassRax*
asaaa*
0"
8. ?
qa.sai^tjag.e.
aaaaaaaa**
aaanaaasa
aaaaaaaa*
A
3
* 5 3 <! 8. *. *
999999999
i$aiti*^qs**
aaassaaa**
a
a
assays?
eaaaaaaaa
aaatia**
.88fc*933
fj 9999'a'as as
aaaaaaa-**
* >
8 P S 8 ft * 8
999399933
as8S}a,Rpa*a
aaaaa*5-'**
8R~ac*;8
aaaaaau'ia
2-154
3 333 9
9 9; >
CSR8.tR8iR
939fa:Sa93S
8.SP8.SR8SR8
^ 5
03 8*88.88
8. 8888
MOOOJOO
OOOOOOOOO
daaaaaaaa
3 3 3 3 33353-3
OK
**
*4
*.
ri
,4
P*
r4
*. "?
1 >?
a 55 8
a st
8. 8888
5 a
*
*
8.
w\
8. 8.
o e
8. 8.
O
St It 3
R
m
3*
R a
OB
5
m
p-
If
If
*3
8
m
8
a
? R
p a
P-
1r*
ft
^ *
3.
8.
* n n
* n 4
r^

a
6
!5
<^
v9
8
P-
<
8
P 3,
^ ^
a R
p^
1 *
3 j
H 8
*1 5
5 3
B 8
ft ft M
8.
e
8.
o
8
d
8.
o
8. 8 8.
o e e
? ? 8.
s6
R
4
<
8 3.
*.ft 1w.
<0
at
R R R
p^
5p- 3r-
R 8.
a *> *
9 t
R
m
8
o
fi
R 3 fl
I P
d
P*
s a
8. *
a ofe
a s
l^
8 P
p^
at
5 a
a;
pi
It
& 8.
. *
1
8.
Of
a a
IS
* m1
p-
"1 5
r
R 8
a J
a
m
R
W
K R
* 1
9339399
. R a
o
3
3
J-fccf
a 9
a a a a a a a a * *
>
5Rscp**a8B
9 3 a ai a a a*
ssax^sarats
asaaaaaa'
a 3
aaaaaaaai
âspRRsg
^ni t U9tit
99 9 9 9
H3 > Mt S tt
a 9 a* g a 9 a a a
8.
8.
8.ft 1
s
53 3 3 .
P4
8. P
ft A
8.
O
8
4
8.
8. 8.
e o
^ 1 * 9 * * 1 !? ^ ^
HHHMM
MMMMM
H-HOOOOOOOO
*.
*40
odoooodo
S*??e3333;
o
o'
o*
0*00
o"
0*0*0*00
at
HOOTQOOOO
OOOOOOOOOO
MMMgM?
oooooooooo
H U H 3 a I
5 1 3 3 E S $ it &
$3$3SS3*8
oooooooooo
1 MM.tMM
oôooo-oooo
oooooooooo
f"
0*000*0*0000*0
t-t
HMHSSH
1 U H U M
SHM?3H
oooooooooo
OOOOOOOOOO
oooooooooo
oooooooooo
oooooooooo
oooooooooo
oooooddo^-
MMMMH
MMMnH
-IMHIHM
MM M* M
oooooooooo
MMMMM
MMMMM
000000-000
oooooooooo
oooooooooo
oooooooddd
oooooooooo
oooooooooo
oooooooooo
oooeoddodr*.
oooooooooo
1 H ! s : a M *
u s s s s n } *
doooo'ooooo
oopob*oo*r*oo'
oooooooooo
oooooooooo
MM34MM
MMMMM
MMMMM
oooooooooo
H H H S H s
Q*
MMUHH
oos
ooododb*
o'
MMMMM
3 5 8 8 e 2 1 5 8
8*SS5S333
o*
o*
-i
.*">
">
M M. * M M. 8
oocooooddd
oooo
j'ododd
oooooooooo
Ik
S32S5S35S3
oooooooooo
oooooooooo
i H H. s S g e ?
ooooooddo'o*
8 8 8
8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8 8 8
8 1 g
S 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8
8 8 |
8 8 8 3 8 8 8 8 8 8
1 8 | H 8
8 8 8 8 8 8 8 8 8 8
o"
oo
o"
-j
ddc*daoe>
oooooooooo
oooooooooo
8 8 8 8 8 8 8 8 8 8
0.000
OOOOOOOC3Q
8 8 8
o.ooo
oooooooooo
0.000
s !>*rf*
888888888g
o.ao
o.ao
o.ooo
ooooododsd
i o.oa*
o.ao
o.ao
j
1
^nHuM
ooo
MM|MM
oooooooooo
o.a
.*S*Spp,i,S3
oooo
H H K a *, M *
RsaR*R*S;Kfc
oooooooooo.
-S ? 3 S 2 $ i S S S
**
rt
#t
*.-%****
.-t
r*
t*
r*<*C4c*e*ciftrieici
9
a
<V
* t U i J } ? < (|
tt
ti
ti
ti
vi
ri
rt
et
rl
t*
'
'
'
ni
<
'
fi
i
hi
s
1
3
*-*
"*
R 5 3 3 ft
fi ' *4
**
rl
r+
r+
r-{
sasasarsapp
P 3 3 S *
& ^ q R
i-4
"
-J
r<
*4
r*
*4
<M
r*
SS^SSf:p
3 ft S ft
j.3 3 *
MMMNH
|4
|4
.-.
r.
gstj^Rstjaa
3 S 8 P 8 S 8 * P
3. ft 3t R ^ !|URIi
MOIftMMCriraHMM
r-t
'
CI
fJ
* 8 S 8 S 8 * P * 3
eieitttttteir-ttitiei
3 * ^ 5 R a 1} S 3 !
fl
cl
fj
fJ
rt
r4
r4
-H
* S ? P 3 w w ^
^-4HHHH-**-4MH
8. 8 8 8
S 8
oooooo
8 8. 8. 8.
oooo
a]
r-i
r-i
r-i
~*
**
r*
1-4
r*
f*
-<
-1
-<
3*R>? ^ * ^ ^
-
--I
^4#^#4*4*4
R P 8: c c
oooodcoddo
Ul^
oddoo
s a a s e 8 s a s u.
8. 8 8. 8 8. 8, 8. 8. 8 8.
8 8. 8. 8 8 8. 8 8 8. 8.
ooododdodd
oooddddddd
88888S888S
88888.8,8888
00*00000000
oooooooooo
oooooooddd
8 8 8, 8 8. 8. 8 8. 8
8 8. 8 8 3. 8 8. 8 8, 8
8. 8. 8.8888?88
8 8, 8 8 8 8. 8. 8. 8. 8
OOOOOOOOOO
oooooodddo
gt?Pjy*ti*
8e,8SaJJP*B88
^âasa^sq^
>s
i-i
*1
t-t
*4
e 3 3 S 3 S g J s
-Hr-Mr*->4H*H-H4it4
1MMMMM
MMMMM
^Hv4r4v4tH(HiHf4r-4
MMMMM
HHH*4HHOOod
i"**--lr-l
P400
0000
3 S 2 $ 5 2 ; 3 S if
oooooooooo
OOOOOOOOOO
MMMMM
ooooc
rt
<*H!3 S Si S 8 8.
M
i P * 3
ooooo
r*
-4-HOOOOOOOO
o*ddod
o
~i
Hr4r1cH4*(
-H
HMHJO
s fc a * R * s * a !j
00
--<
EqS8<ait
#|
<-<
S3-? * ?***!?
r-
0000000
r4r44(^rf4(Hr4(-4M
SRSfCaCP3#2
f**MM||NMMN
4JJS388R
r*
rrf
*-*
*4
SJ a 8 af 8
ft*
-j-S
P S S 3 fc
'MSSMiiM
NMN'MMMftftf*
a R S 3 ft
-4
a a H * * t a *
i-N-M-NMlitNN-NM
P 8 3 3 *
MMMMM
oooooooooo
e 8 | I | 8
oooooooooo*
1 1. . 1 . 8. 1 8
oooooooddd
1 8 8 8 8 8.
ooeodddddd
8 8
oooooooddd
cr
1 8 1 E E 8 5 * 8
8?*28813?9
HHUSHS
H1S3i8B88
ooooooo
At
SEggS8?2&
&M*333*3*
SUHUH!
MMMMM
oooooooooo
cooooooodd
ouoooooddd
IHHHHS
M33S8&538
S H n S 3 i S 1
MMMMM
i
i
ooc
ooooooo
ooo
! 1
i
3
1 a.o o 3o 1o 3o 3o 3o
ooo
H it H H s 3
8 8 8 3 8
333gg33gg
3 3 8 8 8 8
3S583?S&8
I 5 S 3 g g S
g 3 g g S 3 3 1 I
8 8 8 8 8 8 g 8 8
SeslSsISS
3 g 1 H H 3
9 g 1 1 8 8 S 8 8
3333333318
ooo
1 3 8* g
88888gg
gggggggg8
8 8 8 8 8 8 8 8 8 8
888888S
8 8 8 8 8 8 8 8 8 8
8 g 2 g 5 g
o.noo
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
g 8 8
o
*- 8 8 8
N
*"V
#-\
#\
:?i
1^
MMMMM
a n K n. n x
odo
oooooo
ddddoddooo
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8 8
8 M. 8 M M. M
M. 11 11 11 8 8
i 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8 8 8
MMMMM
oooooooooo
**
0666666606
dad
ddoooo
oooooooooo
dc'
".
oooo
dodo
8888888888
!{
15
i^rs*^f^-j*Ni<N^#^-mN
ft
*N
*N
-P*.
#\l-N*n
913 ^ ?^ 5 ?S
oddoooodoo
MMMMM
oooooooooo
do
5oooooooooo
2 S 3 3 H 3 H;
aoddducd
#N
*.
3 3 if S H S S 3 2
F"l
CN
f\
f\
1-1
f\
f\
I-",
*r\
aUMUMI
MMMMM
M M M 1 111
MMMMM
0
oodd
00000
oooooooooo
ddooooodo'o
oooodocodd
oooooooooo
oddoododdo
ddddodoooo
oooooooooo
oooooo
odocooddoo
oooooooooo
oooooooooo
oooooooooo
MMMMM
MtMMM?
oooooooooo
dododo
oooooooooo
oooooooooo
oooooooooo
dddoedoood
->
oooo
QOOowOGO
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8
8 8 8 M 8 8
8 8 8 8 8 8 8 8 S 8.
8 8 8 3 I 8 5. S .
8 8 8 8 8 8 3 3 8 8
o666or>6666
6666066666
i 8 8 8 8 8 8 8
0006006090
8oooooooooo
8 8 8 8 8. 8 8 8 8
8 1 8 8 8 8 8 I
dododooood
MM 1 MM 1
8 8 M 1 1 I 8. I
P(S-sp-<piff:tp
s. 0 a * *. c a
t*rtt4t4f*rit*r****t
ooooooo
RRj
8 8 8 8 M M. M
odedo~doo
000
ddododo
dod
oddoodoodo
oooooooooo
odd
0666660
oooooooooo
oooooooooo
1 g 8 8 8 8 8 8 |
dododooood
0666606666
8 8 8 8 8 8 8
oooooo
6666666666
8 8 8 8 8 8 8
1 e ? s M 11
dodo
odoodo
8 8 8 8 8 8 8 8 8 8
g $ I 8 8. 8 9 8 g e
ssssnun
000
ddoooo
8 8 8 8 8 8 8 8 8 8
M1M111M
OOOOOOOOOO
ddo
000
oooooooooo
8 8 8 8 8 8 8 8 8 8
oooooooooo
ddoooo
oooooo
8 8 8 8 8 8
3 3 3 3 5 8 S S
a 5 M a ? M a
oooo
000
oooooo
8 8 8 8 8 8 8 8 8 8
MMMMM
H H H M H HSSMMM
snaJHHS 3832S8S?1
MMMMM MMMMM
333339 3 332 MMM. MM
MMMMM litfiiiiii
S H H U 8 H MMMMM
MMMMM
ddooooodoo
000
oooooooooo
TBrf*
?S3aiS88gg
8 8 8 8 8 8 8 8 8 8
666606
M M M 1 11 8
8 8 8 8 8 8 8 8 8 8
888 USH88
0066666666
8 8 8 8 8 8 8
ooooooo
*-
MMMMM UMIM'M
oooooooooo
oooooo
ooooooo
ododdoodoo
S
8
ooooooo
MMMMM
odd
ooc
odd
i
tell
oooooooooo
9 a 8 S S g 3
oooooooooo
n ? n y s
IUHIHH
do
MHiigsn
M M 1 11 8
"
oooooooooo
oooooooooo
OOOOOOOOOO
OOOOOOOOOO
oooooooooo
oooooooooo
HHH8H?
oooooooooo
oooooooooo
oooooooooo
MMMMM.
MM. 11 MM
8M1111H1
M 18 M. 8 8 M
MM I 8 8 8 8 8
8 M 8 8 M 11 8.
8 8 8 8 8 8 MM
8. M 8 M 11 8 8
oooooooooo
dddooooodo
oooooooooo
oooooooooo
dddooddodo
6666
ft
HM-MMMMM
60660
6606000060
8
*?<
af
3 5 S 1 S i 8 s
38388339333
oooooooooo
oooododoooo
8 8 8 8 8 8 8
38833333338
oooooooooo
oooooooodoo
oooooooo'o
3 8 8. 8 8 M o8
odddoddooc
oooooooooo
ooooooooooo
8 3 8 8 3 3 8 8
3 3 3 8 g 8 g g g g g
er
i
i
-= *
1 j
)
**
oooooooooo
ooooooooooo
8S38883
ggggggggggg
g g 8 8 g g
g g g g g g 8 8 8 8
oooooooooo'
oodooodoco
*%
gggggggg
8 8 3 8 8 8 8 3 8 8 8
3 8 8 8
8 8 8 8 8 8 8 8 3 8 8
51
s
!
<?
ooooooooo
oooooooooo
8^
R
8
Ni
i^
*J
i **
s
.
il
(a
j * -**
8 8 8 8 8 8 8 8 8 8 8
8 8 3 8 8 8 3 8 8 3
8 8 8 18 8 8 8 8 8 8
8338888888
OOO
ooooooooooo
ooooooooooo
8 8 8 8 3 3 8 8 8 8 8
OOOOOOOOOO
oeoooe
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 3 3 8 8
oooooooooo
dwOoodooooo
3 8 8 3 8 8 8 8 8
<
8 8 3 8 8 3 8 8 3 8 8
oooooooooo
ooooooooooo
8 3 8 8 8 8 8 8 8 8
8 8 8 8 8 8 3 3 3 3 8
OOO
OOOOOOOOOOO
8 8 8 8 8 8 8 3 8 8
8 8 8 8 8 8 8 3 8 8 3
OOOOOOOO
ododooodddo
MM^ffR
s Mstm u
gaggSSelis
S 3 3 3 3 1 8 8
OOOOOOOOOO
doddododod
OOOOOOOOOO
oooooooooo
g 3 g 8 8
SSSsSSobo
83888888
OOOOOOOOOO
OOOOOOOOOO
S 3 3 3 s 1 3
8888883838
oooooodooo
OOOOOOOOOO
8888888888
8888333338
oooooooooo
OOOOOOOOOO
8333888883
3 3 8 8 8 8 8 8 8 8
oooooooooo
OOOOOOOOOO
8833333888
8 8 8 8 8 8 8 8 8 8
oooooooooo
oooooooooo
8 8 8 8 8 8 8 S 8
g g g 3 g g g g
oooodtioo
883ggggggg
oooooo
oodo
8 g g g g g 8 8 8 8
oooooooooo
dododd
8 3 8 8 3 8 8 8 8 8
oodo
8 8 8 8 8 8 8 8 8 8
oooooooooo
dodododddd
8 8 3 8 8 8 8 8 8
8 3 8 8 3 8 3 8 8 8
oooooooooo
OOOOOOOOOO
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 3 8 8 8
oooooooooo
oooo-oooooo
8 8 8 8 8 8 8 3 8 8
8 8 8 8 8 8 8 8 8 8
OOOOOOOOO
oooooooooo
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8 8 8
ooooooo
oooooooooo
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 3 8 8 8
OO'
o
8 8 8 8 8 8 8 8
ododooooeo
oooooooooo
oooooooo
>
oooooooooo
OOOOOOOOOO
oooooooooo
8 | 8 8 8 8 8 8 8 8 8
8 3 3 3 3. s 8 8 8 8
ooooooossc
OOOOOOOOOO
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8 8 8
oooooooo'
oo
ob
ftAMSAKfc&R
oooooooo
k d fTTTTm
8
3
?
a
1I
H
8
!
3
S
p.
*Q.
*p.
p.
>5
>2
r-l
>o
O*
(-1
C*
CM
CM
to
=t>
H>
CM
CM
St
to
i-i
CM
2 5; 8
a
3
w\
ft a 8
r-S
r-l
;*
s
IT
C>
c>
Tk
UN
en
c>
r-J
r-i
5 3 8
fc
-^
ON
3
<0
>o
^
-
^
-
>o
CM
u\
CM
3 3 5J
cn
CM
8 8
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
Si
&
&
3 5 8
CM
TV
CM
el
i> So
I^v
u>
hs
[y
f^v
u>
5
CM
a 8
CM
H
so
o
5o
*>
O
c s
CM
e
*\
<o
WS
*
o
TV
A
IS
n
CM
CM
CM
i-4
a 3 2
SI
w\
w\
*
o
en
en
ft
o
S 2
o
CM
CM
3 3 1
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
en
-4-
W\
r>
8 a
ft
&
CO
3 S 8
*-*
I*
CM
<V
w\
l-i
3 a
CM
r-l
09
IA
CM
CM
**
f-
11
O
CM
3 3 3
o
H
O
w\
UN
a
gIf
g t
O
t-i
i
i
9 si
1o 1o 1o
p.
fill
CM
-
Itl
>o
$
**
3
S
st
p,
a a a
-Tv
H
CM
3 3 8
8 5 5; $5
s 8 3
_js^
"p.
5
3 3
CM
8 3
p.
CM
CM
ex
ft
TV
a st
p.
I 1
$ 3
3 3
si
'p.
p.
'P.
J 8
4
O
5 g
at
tr
si ttt I t u n
; ; j J a si j j
a 8 u * J ! F *
rf*RRaaaaaa
asSRSSKtss.
SRRRa'aaaa
RRsasusasa
a R R a s, a a a" a" a"
RRsaaaaaa
KSRijiSsa;*
^RRSRaaaaa
1! *
saasaaada
ill
sjJasaa
8-
jssss'ass
f-
R6*S34SR*.3
asa'a'aaa
u t ; s s J K
i^RRRaa'ad
Raasa
a a * a t s 3 s: 3 s
RSnSRRRaa'a
a3SR33RJ5a8
RSS444RRRa
i R 8 U U s S. i
a 8 s 8 a a a* a a a
aaaaaaaaa'
aasaasaaaa
tl R9ni|8R
3 siiidij
^HRRtU
aR*r3*
a s 8 a 8 a a s a a
a 5 JM t US t
aaaaaaas
aaas'sa
asaaiSSa
t<|URS ?l| tt
asssfisaaaa
q5S^fe383*3
tJURjnrt
a a' a s a a a a a a
â**R3fe3,fca
aasaa'aaaa*
fc?RS8S^R*
aaassass
aaaisssaa
3&sRa8.3.fca
a a a" a 8 a s a a'
&>8SS}aa.*B,8P
aaa'aa'aaaa
RP3n*Ki^fr
a a" 8 a a a a a a a
RSPâfc asp
a'aaas'aa
)t M t s u c
a a a a 8 a a a*
âsKRPâi
aasssaaaaa
^S*SJ38SIR.
asaasasa'
a s a a a a a a a*
aaatBEa"aa
a.a.sJ?â*aaR
aassaaaaa
Ra.R^qaasR;*
aaa&saaa
sasRaaitsâ
a'assaa'aaa
a a" a a* a a a
a'sa'ssaaa
I ) ( M 3 < M
aaa'saaa
aaaasssaa
S5C*aR3?, 8*
R6**a**j?**.
aa'aaaaaaaa
aaaa3s3aaa
aaaaaa
aaaîiaas
*ftasS4ap
aaaiiiiaa
8J3aSS^8.S:a
99 s: jiu9!
XS3a*s3t:P
9 9 93^^99
SSSiJRP^SSt
9 933J J9i99
auSRpR!;P
aaaaaiâa
*SHS35!3RS#
aaaîafa
8SR8SR8SR8
a" s a a a a a a a s"
aaasaa
RasaRRatRs
Raa'sRRR^jj*
i ; t ; n i| i ? t
s a J ^ ;; * R
a s a a a a s s s
a'aaaiaajJaa
gepp^psfcsfc
8-a = 4 = *C8*
S J } J S S J 5
S8!?3S?&8S
8
4
8 8 s 3 s * a * R S '
4
S S4*33P*-346
4
B 3UUU1 1
33*ft8:Sft*S5
**RRsa
sisisiRRsasR"
tsaca.agâit
t c t u u e j j
j 8 S R a
SapR8SRR34
R K t u' n s
t i; u t t
âasseRats
n i t i ; i n
Hi?
3a3*3*:$*3
R8SR8SR8SR
2' i
*
SB3*s**a.rs,
aaaaaaaaaa
SR8SR8SR8S
aaaaaaaii
a 6
R
8
Si
8
o
Si
8
ttl
9. 4 ;> )t 1; t t * *
8 *
- Si 4
*4 8 9
Si 4
i1
8 S
Ul ^t RSI
a a R a a a as a
sag^SRas*.
aaRRRaaa
Raa^SFRsjR
aRRRsia
RC8*-84a
5 } < * 8 tt S 1) * *
a RRRaaaa
SRRRR88S5
R8aa^8^x
6 Eat#Raa^tBR*
4 iifest;****'
iaRRRRHaaa
H ? a ? 5 ? ^
RRRRSR888R
6 U^ Rt f Rt
RRRfiRRiafi
RRaSHS44
âtjassaac
fiR88 8*
MKt t UM
RRR8#^
r- si S
o 8
m
m
a
St *' 3*Ri;ftt;**RRR
8 8 3gR#StR3r4
si 4 3 * K t t U * li S 5
8 *i SRRRCaarqfcfl
a 4 iiir*fcfeft**<iR
8 4
8US UMM
si * RftR ******
aaRRRnaaa
m s n n R t ^
aa*RRRaa
sun^R *.s,
Rsr. ti t r, t j
R R
8*RRSi4$R81
RRai^RRasia
i * a * R a a $ a, ?
RRRaaaRRR
o 8. *i R 8 4 S 8.
R 8 38
ft 9 3 H * X t t R'
8. 4
as
aa*R8,tRR,SK
RR8S^4
RRRRRR8 8 8R
8 8
**
aaaaa**qss
u'iissnii
i ; H s M * M
8 8
RRSR88SR
R8sg*atRa
R R K R R ; R a fc
âaaaass*
aC^9*S}8$ft
H RR RR C '
Si 4
I " si
iI - SSi
s t
3RSS3HS84R8
5 li t ft t t j j X
a*8Ra*3$s8.*
SRRRftSRRRKR*
aassasasia*
R
astgsisia
ftsassadâa
R j (' t j ' 6
ssasaa^âa
R %
e *
, , 4 S "U H I U
* t R M R A *
a^3**^#SRS
RRR t t RRRRR
HIHHM4
D t N K ' t 1 9 ;
aaaRRRRRBR
âi|8it^PR
aaaaaRRRR
*8$R8$R8$R
RRHr, RRRRR
pâart88,i?R
RRRaaaRSRR
RaRBafSftRRR
88RRR8*H'S
SBRSPR8P9B
'
t li II u' Hs
^ S R ? * ? 8 * , 8
RRft*8**R;i,
MURtlUt
<U^M^
*
O
<
I1
I
f5
ft
<*
a 8.
d *
*
N
j
n
3
*
ft
R
a. t * ft * ft
ft
a
%
in
ft * & ^
=?
ft * 8 *
s ft
a 8 6 ft 5 3
4 4 <
A
4
8 *
=1 1
ft
3
1?
6.
<
a e * r * s
8 ft a 8 ft
*
4 4
ft
ft ft
a ft
ft
A
R a ft t $
R a 8 ft 8
4 4 4 A
R 3
5 R
ft rt ft
8
4 S; R
r.
r-
5 ft
i 4
ft s 3 ft
4 4 * 4
8, 3
ft
8 *
s
ft
ft ft ft
4 4 4
ft
s *. 8
4
353
555 553
555 353
555 4R 45 4R
553 333
555 355
535 333
3 ft 3555
3 5 a4 4a e4^vt
*
4 5 3535
ft
ft 8 ft I
4 3 4 4 2
3 a4 % S *. .
3
P.
A
r; t
A A
fA
?^
-
ft a * a ft
^
P
-4
r4
* ê 1*
is
!|
g
S 3 t
M
Pi
J
O
O
P4
ft
A
in
tn
*\
*\
!?^
(4
ft s
A A
P
ft
8 * ^ $
ft
f
ft
p
.ft 3rft r*
<5 et* cc*.
*ft r* n1
? ^ ^
R S s 4
MM
fft
a 8
- 5 3
115
M
ft
*
. ; ^
et
* a $ $ *
* A A N N
r<
P * a
2
a ftp
o
5? * 8 8
o d d d d
8 8.
d O
8. 8. 6 8 8 8 8 8
8 8. 8. 8 8 8 8. 8
odd d d o o d
odd
8 8
d d
8 8 8 ft 8 8 8 8 8 8
d O o d d d d d o
8 8 8 8 8 8 8 8 8 8
d d d d odd o d o*
8. 8 8 8 8 8 8 8 8 8
d d O O *> d d d d
8 8 8 8 8. 8. 8 8 8 8
d d odd d d odd
ft
.4 3
A
A
ft
*
4
3 3
*
*
=<
*4
*A
<H
* . ?
-<
*-
^RaftPftfta.8^
aaaa'aaa*
sR.aRPX*a.8e.
a a* a a' a a a *
p* p ft ft1 ft R ft. ft
r*
r4
r4
r4
J 4
4
ftftftaftsftfta^
ftifftftâftftftR
aaaaaaaaaa
* *a^*rft*:
aaaaaaaaaa
w4
ft
rt
* ft ftrS ^ ft a. ft ft ft
i*r^
*rt*t**-*<
5l!!RftR58t.
aaaaaaaaaa
aaaaaaaaaa
M
83âs^3(
aaaaaaa'a'p?
^S?^^?8.ft
aasaaaaaaa
a
aaaaaaaaaa
ft 3&8ftRft8*93
aaaaaaaaaa
&4*^J?8. ft*
aaaaaaaaaa
^8RftRafi*i^g
aaaaaaaaa
aaaaaaaaaa
5 5 * 3 5 1 1 1 2 8-
aaaaa'aa*
S 1| a t 1 ^ 9 9 t t
a a a a' a' a a a' * *
3*$$!iasR.
aaaaaaaa
n3 SJ 3 3 5 5 s
iiiniiiH
5555555551
drirfaaaaa*
5 3 h a 5 5 5 * * 3
R4*?3ftpR:;fta
JKR^SSSS
{it
ft%^8*.pî;ft
tMH<!MM
5^^?5 5~^g
RKft^Raasp^
5535555255
i^MJ^H
< f * q H 9 * P
#>oftiiii
*.R*8.*.35
R 8 JJ 3 ;, 8 I * ft
aaaaaaa''
fAKtet#d*-wO*>r
^s*^!^?*J
q^pftRftft8..9
* *
li 8. M H M M
5315555355
55îa?'?*:?!t
-<n!?5n115?5
aaa****-*
1) ( p S if 0 f: ^ A S
8Rft8ft3R?ft
55^5555555
^?5^58.^5^
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8. 8 8
8 8 8 8 8 8 8 8 8 8
339555***5
r*rtMrtrt***4)*r+rt
aaaaa***
9>
>
OOa
OOdOOQOOOO
ft
tf
^FJ^4H
5i!'?3<5j,!5
ajta>t>.**v>*<ifit<f
aaaaasas''
* r
ri
8SjPRqi?asCi
aaaaaaaaad
S38^!|S8ft
aaaaaaaa'aa.
a e a * <
4
^-rfi>dc'*rBfft
s * t ?
A
M
AI
rt
8. 8 8 8 8 8 8 8 8. 8.
d d dodo d d d d
** **flftP^R81
asaaaaaa
*. *3t3ft?^R8t
9 33999999!)
1
i
i
**
8 8 8 8 8 8 8 8 8 8
d d d O odd odd
hi
* *. ? a
ft 5 s? 1
#A *\ m\ *s
a
A
A
A
ft
d
5 5 3
3 8 ft 5
<**
3 * >t
d o o o
ft
ft R ft R
ft ^
3 3 *.
ft 8 * 8. P ^
A
ft
tj
t*
r*
fi
Pt
< ft ? j * ft ft 8 ft
f ^ ^ P 3

4 n tn *>\
a.
ft 3. ft ? * i?
A A **. P\ tA m
s ft * H "
8 ft a P a *
4 A A m % *
s
4 4
8. *
S ft q H 9
4 -* 4 4 4 4
* * a
* * R ^
8 R ft 8 S 3
4
^
A
ft*.
f*.
3 s
1^
4 8 8 a R ft ft ft ft R
4 4 4 4 4 4 2 J
ft
A.
ft P 51 ft
A A
A
? 3
A % ?
R ft R ft S a 8 ft ft ft
? 1
ft S 5 t a ft ft
4 4 #^ ;. <n #\
8
4 4
<?^
ft if a
* * * ft
Sn
0%
a 8 ft 8 ^ P
4 4 m m 0% m
ft
A
C
A
f>
WS
t | i;
8. 8. 8 8. 8 8 o 8 8 8.
d O d O O O d d d d
A.
3
*
ft
*
*
> <>
*
*a
ft.
rtrtrtrtrtiirtrtrtrt
HMMMM
oooooooooo
oooooooooo
oooooooooo
oooooooooo
MMMMM
1 l * M M * ? IS MUMU
iUflUiH MMMMM
oooooooooo
0000000*000
oooooooooo
oooooooooo
Ioooooooooo
? M M M M
oooooooooo
oooooooooo
H
UUf M ^
oooooooooo
oooooooooo
U
H HH S S
oooooooooo
oooooooooo
OOOOOOOOOO
oooooooooo
oooooooooo
oooooooooo
oooooooooo
SHHHH3
oooooooooo
ftS&&M35
oooooooooo
s. HUHHM
E 3 % 3 9 S 3 5 S
*
o
KKS&IMMI
oooooooooo
HMMMM
MMMMH
oooooooooo
XrfOOOOOOOO
1 M IIIII! 4
oooooooooo
nooeooo~
oooooooooo
i
]
ooooooooo
o'
oooooooooo
dooooooo
MMMMM
oooooooooo
oooooooooo
000000000*0
oooooooooo
oooooooooo
oooooooooo
oooooooooo
88888881
g 8
n H H! SS S
oo
S 3 ! 1 8 1 1 3 *
8 8 g g
oooooooooo
a
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8 8 8
8 8 8 S 8 8 8 8
oooooooooo
M M M
8 8 1 8
8 8 8 8 8 8 8 8
8888888888
S 8 8 8 8
8 8 18 8 8 8 8 8 8
j>&
8 8 8 8 8 8 8
s
M
oooooooooo
oooooooooo
oooooooooo
oooooooooo
oooooooooo
oooooooooo
oooooooooo
i
1
oooooooooo
a 1 s152 M 1 g
oooooooooo
oooooooooo
M
!!!!!!!-! if
oooooooooo
oooooooooo
oooooooooo
oooooooooo
oooooooooo
RaaaaSARRS
8 3 3 3*38538
V. ft R 3$ S * P. * R
vtr*ctctt*tM**ctctr4
fM
oooooooooo
W
(t
oooooooooo
oooooooooo
888888188 8
oooooooooo
C4t4tieititiei<it*iei
?! R 3 3 S * S ^ % R
cr
3SSS8S*8f
*
MrartMNOiMMNct
I
Si
t it i x a ^ s^ i;
asasat
ft
3 3R33KasS:J
as 3 * 3 3 3 3
HHSHH?
**PPP
SSRC;*:!*
SMUUtU
n 3 u U U
s ais. a * e s *
N
<i
fi
! s i ; t i i! : j
*4v4H*-tr4*-<*4000
s M t i i
-4
**,:|%*R$!$a
r*
,H
NW
sscappsssa
Jt t P f P u s s s
i s I; P s M ^ 3
s 3 a n 9 i % % g
4p4v4r4^4#40OOO
3 1 * s %% 1 I I '4
r<-*i-iHdodddd
i
I
^ * n 3 s? a *
aR48as8.')ft
H*4*4ooooooo
sstgsasss
-I
C\
<fl
cl
flt-iOOO
* R 1 B 8 P P P 3 3
ptOOOOOOOOO
oooooooooo
oooooooooo
to s a H )j 3 S 3 S J
S 4 P e 3. 8 * 5
oooooooooo
$oooooooooo
* 1 3 * * 5 * * .
!} 5 8 a 3 S 8 3 a I
4decdeodddo
3**^5|fiRR*.
oooooooooo
ssqsassss'ss
8888888888
8 8 8 8 8 8 i 8
8 8 8 3 8 8 8 8. 8. 8
8 8 8 8 8 8 8 8 8 8
odcddddodd
8 8 8 8 8 8 8 8 8. 8
8 8 8 8 8 8 8 8 8 8
d
o
*l
'j
ddoddd^doo
8888888888.
oooooooooo
s5"
oooooooooo
8 8 8 8 8 8 8 8. 8 8
OOOOOOOOOO
8 8 8 8 8 8 8 8 8 8
dddoodddod
8 8 8 8 8 8 8 8 8 8
ddddooodod
8 8 8 8 8 8. 8 8 8 8
oddododddo
dddooddood
8 S 8 8 <* 8 * 6 8 8
a a a. a 4 s, 'a s 9
O*
04
*V
-f
ft
Oi
oooooooooo
oooooooooo
3. 8 8 8 8 8 8 8 8 8
oooooooooo
MM 1 8 M M
oooooooooo
M 8 M 1II I 1
ddodooddoo
i
9
|1 || MMM
HHHHU
Ms *.|| is ||
oooooooooo
ft.
OOWwOOOOOO
OOOOOOOOOO
cOOOO
ooooooooo
3 3 3 1 & f i 1! 8
OOOOOOOOOO
i
3
| 1 J || Ml ||
!!!!!!!
j | 111 H M
OOOOOOOOOO
OOOOOOOOOO
H U i 8 M M
S || | 3 3 | || a
OOOOOOOOOO
8. J | | J || || |
OOOOOOOOOO
oooooooooo
OOOOOOOOOO
S I U 8 & S S
3 3 3 3 3,
3 3 8
8 8S.8g8.9Mff
ddoddododd
oooooooooo
1 3 e I I 3 i 3
3 3 8 5 8 8 8 8 8 1
OOOOOOOOOO
OOOOOOOOOO
MMMMM
oooooocooo
3oo858
MMMMM
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8 8 8
oooooooo-
OOOOOCOOOO
ododddoood
U H H 3 | u
OOOOOOOOOO
oooooooooo
OOOOOOOOOO
i f 11 HM
MMMMM
>-
ÔOOJOOOO
OOOOOOOOOO
S 8 8 8 H 8 8 8 i
oooooodddd
1 i 8 8 S g 1 8 8
oooooodddd
8 8 6 8 8 8 8 8 8 8
dooooooddd
8 8 8 8 8 8 8 8 S 8
8 8 8 8 8 1 8 8 | 8
8 8 8 8 8 i 8 8 8
8 e s
8 8 8 8 8 8 8 8 1
ooobbdddb
oooodo*>ooo
dododddooo
OOOOOOOOOO
OOOOOOOOOO
8 \ 1 8 1
o
8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8 8 8.
OOOOOOOOOO
18 8 8 8 8 8 8 8 8
OOOOOOOOOO
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 I
8 8 8 8 f. 1 8 8 8
r>
o*
o"
o"
o'
o'
888. 8 8 8 8
oodooddddo
OOOOOOOOOO
oododdddd
8 8 8 8 8 8 8 8 8 8
8 8 8 8 8 8 8 8 8 8
bodddodddd
OOOOOOOOOO
î si u s n u
OOOOOOOOOO
8 8 8 8 8 8 8 8 8 8
O
o"
O*
O*
8 i 8 i 8 8 8. 8
oooooodddd
ft3t*^t:
U ? ? * ? * !} U
MMMMM
dddooooood
MMMMM
^MH3U3
HSHHU
OOOOOOOOOv
OOOOOOOOOO
MMMMM
OOOOOOOOOO
M. MMMM
MMMMM
|| M 1 1 M M
MMMMM
MMMMM
MMMMM
MMMMM
MMMMM
M. ME
- 35883
bddboo oooo
SSMMMM
5 3 I 1 1 8 I f 8
OOOOOOOOOO
OOOOOOOOOO
fEHIHIU
dooooooooo
M | M M M8
M*MMM3
ft 8 8 8 8 18 8 8 8 8
8 8 8 8 8 8 8 8 8 8.
dddddddodo
MMMMM
tlttlMIJ]
MMMMM
MMMMM
MMMMM
MMMMM
} | ] ]| J | M
MM MM ||
M M M It j!
MMMMM.
MMMMM
MMMMM
MMMMM
MMMMM
MMMMM
MMMMM
o*tf*
or
rr*?*74r*r*wA'+mr+
MMMMM
OOOOOOOOOO
* a 11111111 s
OOOOOOOOOO
o
OOOOOOOOOO
MMM9SI3
OOOOOOOOOO
M
OOOOOOCCC9
OOOOOOOOOO
8 S 8 8 8
OOOOOOOOOO
OOOOOOOOOO
bdobbdodob
OOOOOOOOOO
8 8 8 8 8 8 8 8 8
Wi[
OOOOOOOOOO
8 8 8 8 8 8 8 8 8 8
OOOOOOOOOO
OOOOOOOOOO
o*
OOOOOOOOOO
ooovoooooc
oooooooooo
OOOOOOOOOO
MMMMM
MMMUM
OOOOOOOOOO
ooooocoooo
M M i * M II
OOOOOOOOOO
oooooooooo'
OOOOOOOOOO
8 \ 3 g 3 S 3 3 8
UHHHM
MMMIIM
|| |j] 11 Ml
OOOOOOOOOO
3 3 M S ? M 3
OOOOb
MMMMM
MMMMM
ddddooooo
OOOOOOOOOO
OOOOOOOOOO
oooo
oooo
OOOOOOOOOO
ooooododoo
OOOOOOOOOO
OOOOOOOOOO
oddddoddoo
OOOOOOOOOO
OOOOOOOOOO
dooooooooo
oooodooooo
000000*0000
OQOOOOO****
OOOOOOOOOO
OOOOOOOOOO
OOOOOOOOOO
doooooooo
OOOOOOOOOO
OOOOOOOOOO
MMMMM
oddcdddodd
ooooooooo
MMMMM
IMMMMI
OOOOOOOOOO
1188888888
o o a o c s 6 o a
OOOOOOOOOO
odoooddooo
OOOOOOOOOO
oodooooooo
dooooooooo
M II II II M
8 8 8 8 8 8 8
8 8 8 8 18 8 8 8 8
8 8 1 M 8 8 8 8 8
8 8 8 8 8 M 8 M
M. 1 It M M 1
MMMMM
OOOOOOOOOO
M
I | | 8 || |8
obobobo<*ob
8 8 8 8 8 1 1 I
8. 8 M M. M M
M M. II II II
8 MM M 8 8 8
8o 8 8o- 5M II
o o 13 e"M2
o'
booboooboo
O
dddddoodod
s
o*lo*
88S8888888
8
.
. 1 S s. -. n n "
OOOOOOOOOO
tititifititlliiiliti
OOOOOOOOOO
OOOOOOOOOO
1 8 8 8 8 8 8 8 8 8
oodododddd
n M n t
<VM**MMMNIMM
O'
o'
I M. 1 1 1 1 II
|| | M 8 M. M.
oddooddob
ooo
oodoodo
t-\
-*
#,
w\
#*
m*
MMMMMS
MMMMMI
M M M. M M. S
MMMMM
M M M. M M
8 8. MMMM
HHHUU
8 8 8 8 18 8 8 8 8
oooooooooo
8ddoooddooo
8 8 M S 8 8 8 g gd
8 8 8 8 1 8 8 8
8 8 8 8 8 8 8 8 8 8 8
odd. oododdc
oooooooooo
8 g 8 g g g g 8 8
oooooooooo
odooddddoo
eddeeodoooe
oooooooooo
<
88888888888
ooodeiddoodd
ddododeddod
8 8 8 3 8 8 8 8 8 8 8
oooooooooo
'1
8 8
18 8 8 8 8 8 8 8 8 8
28
8 8 8 8
8 8 8 8 8 8 8 8 8 8 8
Ii
oooooooooo
8 I g 8 8
888888 3 888
oooooooooo
fc 8 8 8 8 8 8 8 8 8 8
oooooooooo
M
8 8 8 8 8 8 8 8 8 8
OOOOOOOOOO
8 8 8 8 8 8 8 8 8 8
oooooooooo
3
3
8 8 8 5 8 8 8 8 8 8 8
o
OOOOOânee
oodoooooood
8 8 8 8 8 8 8 8 8 8
oooooooooo
8 8 8 8 8 8 8 8 8 8
oooooooooo
8 8 8 8 8 8 8 8 8 8
oooodoooood
8888888888
oooooooooo
8 8 8 8 8 8 8 8 8 8 8
d*y&o6<
S 8 8 8 8 8 8 8 8
oooooooooo
8 8 8 8 8 8 8 8
oooo^do
8 8 8 8 8 8 8 8 8 8 8
B
3 rr fteeettpfttic*
8S8?lfS*6SCg
3 3 3 8 8 8 8
o*
3 3 3 S 3 3 8 3
oooooooooo
3311888888
3 3 5 2 S S 1
8888888888
oo
oooooooooo
oooooooo
oo
oo
oooooooo
oo
8 8. 8 M M. M ?
oo
M M M M M
MMMMM.
ddoddddo
do
8888888888
5 f 8 8 8 8
oooooooooo
oooooooooo
dddoe-coooo
MMMMM
oddodooodd
8oooooooooo
8 8 8 8 MM
8ddodddoood
8 1 t 8 8 M. M
8 8 8 8 g 8 8 8 8
oooooooooo
goooooooooo'
8 M. 8 M. 8 8. 8
8. | 1 8 8 8 |
8. 8 8 1 M. M 1 8.
oooooooo
3ooocoooooo
3 II M M. M.
oooooooo
oo
*. g 3 3 3 3 9 i 3 3
oooooooo
oooooooo
addd'o'ddddo
M M M M. M
oddodoodoo
ddddodddod
8 8
8.dddddooddo
8 1 8 8 M 1 1 1
8 8 8 8 8 8 8 8 8 8
S 1 8 8 8 8 8 8 8
888B8888M
M. M. 1 MM 1
M M. M M M.
8. 8. | 8 8 1 I ! ! 8
odddddoooo
8. 8 8 8 M II M
8 | 8 8 8 8 8 8 8 8
8888888885
RWR*R*R**
HÛ^M
IN
oooooooooo
ot-
oooooooo
oooooooooo
JIV
oocooocooo
oooooooooo
o'
o"
o'
o'
e>
o'
o"
o'
o"
o'
o"
o"
o'
o"
'
PVltw J o p.oixi ! JO wsnj
p3
:q--..--...."
ll
I
II
i i
ill
iJ
Is
15
ft
EUiillliHiiii'liiFFrJrffi
1
3
t
S H
19
j
J fi
IS
*
\
TKt-nr tTTTe www rrrnm;
Ssf
tu iii-î-TiTrriW^tttenigtttf
i
ft aw r la BM*ty
Mnl4lTali
ltt.'
* nvw ar Jam mi fealitr
IMOI
far nonal
IMVMU*
*>./
Uli % run f-r a^l
HM
C** Mn oc.iu
IM (bl WTOMt dtftotiw)
tli* r
V nwtr* UM 'V *M 4>
"lIMNl W Ulli ||| |
MMM* IM
(la i-*wrt MMUn)
frf?
itaMBp.
IM11II >1JH
ftr iJ v ;. t*nr. t-Ontujr*
MMmH mi m MB mm taam m SBH K-UMUB BIIUH

*l w* * tmmm i^f nt hM in lirl MMM M IMM>M1| 4tl*)
*mU mm o*i U*%g;
ft Aotalttad Lot (lit i
lm*mm * * n mi rf u* gm,.
-UWltntelki d iMkM
k'li7
*f *!* Lrt ii ^roj* fefWttl)
9f9-f
11
HlMlUllllllKI*
l>awalaUalr<
i iiiiih unl
I
I Lit (! prwt Iu]
0*11%- ml Mttii Jm (U fMMM tab)
ggggj
l.;i.M|.l,i;Jl.i.|
TO:i!.-:*hi%-t{
3$
m *>*' <t|i1< MiaM

x UM I mua mat +m*
"-- -
ii i ii ) r 11
a;.
(ma <w mm urn* MM i
I IM IIIMlinij *i*lM%)
tm*A*
MM
c*t* UMrt I
r{rtftfrt
Ml at li
MIWlMMi
I U (la a MMM)
alur x in 1I
- 10 -
71-1.
General Theory.
D. Blackwell [1] has shown that if x has density p~(x), g is any-
unbiased estimate of , and T is a sufficient statistic for 0, then

E(g|T) is also unbiased, and furthermore, has a variance no greater than
tfcat of g.
Lehmann and Scheffe [8] have extended this result and have
shown that if T is completey then every estimable function^ h(0)

possesses an unbiased estimate with uniformly smallest variance and
this estimate i3 the unique unbiased estimate of g(O) which is a function
of T.
It has already been shown [8] that the sufficient statistics for
the normal distribution are complete.
Furthermore, it is evident that
there exists an unbiased estimate of the fraction of a normal population

lying outside a fixed interval, e^g,. the fraction of independent observations in a sample from this normal population which lie outside this
interval.
Consequently, the fraction of a normal population lying out-
side a fixed interval (p) is an estimable function and has an unbiased

estimate with uniformly smallest variance, and this estimate is the
unique unbiased estimate of p which is a function of the sufficient
statistics.
1/ The idea of basing the two-sided acceptance procedure on the uniformly

minimum variance unbiased estimate of the fraction defective stems
from research carried out at the Applied Math, and Stat. Lab. by a group
including
Professor A. K. Bowker, K. Rubin, and L. E. Moses,
2/ A statistic T (more properly a family of distributions of T) is said

to be complete is E^f(T) - 0 on all 0r implies f(T) 0 except
possibly on a 3et N for which P(N) 0 for all dr.
3/
Given a random variable x with density p(x), a function g(&) is said

to be estimable if there exists a function h(x) such that EQh(x) g().
-li-
ne are concerned with estimating the parameter
P - 1 - J ~- e
L -Tir^
-
2
*>
(T
dz .
Two cases will be considered; namely, (1) /a unknown, <?-> known
(2) both
/
LA.
and c7> unknown.

Let x, ,Xp,... ,x be independent random variables from the above
normal distribution.
Define "p' as the usual attribute estimate of the
fraction defective i.e., the ratio of the number of defective items to

the sample size.
It is evident that 'p'1 is unbiased.
Hence it follows
fron; Blackwell [1] and Lehmann and Scheffe's [8] results that p - ECp\T)
is the unique uniformly minimum variance (UMV)unbiased estimate of p,
where T are the sufficient statistics for the normal distribution.
Since /p' is the sum of independent identically distributed random variables taking on the values 0 and 1, it is evident that p E(p'\T) is
equivalent to E(^|T) where 'p'is defined as follows:
Let y be any one
of the observations (x, ,x_,...,x ) 3ay x..
(1)
VI-2.
Pfr^X^,.,.^) - 0
ifL<y<U
"pCyn^jX,,...,^) 1
otherwise.
The Estimate When the Population Variance is Known. If the observations are drawn from a normal population with unknown
mean /,<, and known variance <f , then x - X_ x. is a sufficient

n
i-1
statistic. In this case p E(p|x) is the UMV unbiased estimate of p.
- 12 -
If p is the estimate (1) then
p(x) - S(^|x) - Pr{p - l\x] - i - Pr{L < 7 < u)x} - 1 - f *&&- d7

h
where g(y,x) is the joint probability density of y and x, and h(x) is

tne probability density of x.
n x.
Consider the joint probability density of y and x' >_ ^>
im2 a~
- -** -Uy-M)2 (n-lJCx-^)2)

21TC72
The transformation
j , fc-l)x"'
leads to
n
g(y,x) -7=
5e
2<r
and division by
h(i)-~g-ee
J2~TC^
results in
ri
sZ _
1 / -
>2 "i
-7-2i(y-r) 53(m*yO j
2^
'
n^x;
- 13 -
n
,)n 1
h(x)
(7-D2
^F<r
Therefore
<
!~n L-x
J
2(n-l)^
(7-i)2
.2
===: e
.2
^ dt +
-== e
^ dt
Jn-1 &
713.
The KuCmncî Acceptance Criterd.cnThe acceptance procedure is formulated as follows:
Accept a lot
A
*
of Items if in the sample pip
where p* is so chosen that if the
population percentage defective is p, i.e., if the portion of the

population lying outside (U,L) is p, then the probability of acceptance
will be L .
It is shown that in the one-sided case this is equivalent
to the well-known and widely used procedure:

Let K
be defined by
co
s,
JTTT
Accept if 5c < U-k</\
- 14 -
In the one-sided case, when L " -co, then
oc
r u^
n-l <r
we must have
I n~ Dx ^
_, T,
I n1 .
fry
If we take k "J K ^ the OC curves of the two acceptance procedures

will be identical.
For the case where there is a double specification limit and the
standard deviation is known, it is evident that if -~ < 2 K,r J, the
'
<7^
AQL
2
incoming value of p, the percentage defective, is greater than the AQL.
Consequently, the lot should be rejected without a sample being drawn.
The OC curve of this two sided test.takes the form of a band, the lower
bound of which is the one sided OC curve.
For incoming quality which is
no worse than the AQL the upper bound corresponds to equal division of
the defective below the lower specification limit and above the upper
specification limit.
The band here is relatively wide, but this is
desirable since this gives "good" quality a better chance of acceptance.

For quality worse than the AQL, the Above restriction on ~^ eliminates
certain divisions, including equal division, so that the band quickly
tends to the one sided curve.
A diagram of the OC band is as follows:
-15 -
Prob, of
Acceptance
AQL
P..
VI-4.
A.
*%_.
L.
The Estimate cf p When the Population Variance is Unknown.
If the observations are drawn from a normal population with unknown

2
_
mean M. and unknown variance cr , then the pair of sample values x, and
X. (x^x)
are sufficient statistics.
In this case $(x,S ) - E(^]x,S )
where
-\2
S2 - Z K-x)
i-1
/O
If p is the estimate (1) then
P(>:,S2)
Pr{p-l|x,S?} -1-/ &'<?Ji v-2S

x
h(x,S )
^ dy
= 16 -
where f (y,x,S ) is the joint probability density of y, one of the

observations in the sample, the sample mean x, and the sample sum of
squares S*, and h(x,S ) is the joint probability density of x and S .
It is well known that h(x,S~) is given by
r=-/~2s 2
2^"
h(x,S ; -
2<T*
~~Z-j
i2Tf^r(~)(2c^2)
_
To find f(ypc,S ) consider the Joint density of the sample.
This
may be expressed as the Joint density of the mutually independent sample

_
2
statistics y, x1, and S' where
' -i-5rr
fe^
;
s<2 -fi
(^-x-)
i2
H(n^)(2c^2)^
Now let
- 1? -
nx.
y
n-1 " n-1
X'
S'2-S2-^- (x-y)2
y - 7
Undor this transformation the expression (2) becomes
n-1 ( W
K|S2 - JJj; (x-y)2}
n-1 /""
where
n
n-2
2
>|^1 2lV<r(2c? )
p(2f)
and the variable are subject to the restrictions
-co < y < oo

OSS 2 < 00
*-7
-1 n
= o2>
Dividing the expression (3) by h(x,S
) results in the conditional
density of j, given x and S .
n /- ..>2l
18 -
P<)
^v^ ^
If now the transformation z
**
= * = ^g^J-^r is made the density of
the random variable z is obtained:
nrr2)
(|-D-i x a (i-D-i
g(z) - --hrzir?-z
< ->
(5)
o<z<i
This will be recognized as a symmetrical Beta distribution with

parameters -s 1.
Hereafter, g(z)dz will be denoted by d/?(-S= _ l).
Since flU^S2) - FT-\J > U(x,S2) + Pry < l|x,S2) then in terms
of z
W) - *{ * i - i J5] {. *i - i J)
VI-5.
The unknown-
Acceptance Criterion.
Just as in the case whsre the population variance is known the

acceptance procedure is:
Accept if p(x,S ) < p
where p
is so chosen
that if the population fraction defective is p, the probability of

acceptance will be L .
It is shown that in the one-sided case,; i.e.,
hers there is only one specification limit, say U, this procedure is

equivalent to the well-known and widely used test x ks < U, there
- 19 -
s - S/Jn-l.
Let A
be defined by j
2 Njn-1
d/5(| - 1), then pip if and only if
rV îr
5 /,*. a110 since s S//n-l this is equivalent to
S -/^p*
x Jg (1-2^)3 s U
If k is taken equal to j^ (1-2/^*-) the procedures p < p
x -t ks < U will have identical OC curves.
A
criteria p based on the statistic
U-x
and
In this caper the acceptance
n
-x
are tabled rather than B to
conform with the common use of the square root of the unbiased esstimator
of O^ .
For the case where there is a double specification limit one of the
authors made nu&ievicai investigations [11] of the OC curves for various
divisions of the psrcentage defective.
It was evident that the band
was so narrow that for all practial purposes it can be assumed to be a

single OC curve, i.e., the OC curve of the one sided plan.
VI-6.
The Acceptance Procedure and Estimate Based on the Range.

Since the pair of sample statistics (x,H), where 2 is the sample
range or average range-/ is not sufficient for the normal distribution

when the mean and variance are both unknown, no uniformly minimum
variance estimate of p which ir .> function x and R" can be derived.
In
this paper the following estimate of p(x,ff) is used:
In this paper, whenever pno<i^lfl the subôiip size is taken as 5>

so that R is the average range of m subgroups of S (the sample
consisting of Sm observations). For sample siaes of 3, 4, and 7,
R" is taken as the range of the sample.
/.u -
P(x,5) - J
6fi [-
^ [î-lj J
here <X and v> are constants for fixed n, which will be explained in
a subsequent paragraph of this section.
No difficulty is encountered in connection with OC curves (which
were obtained by numerical integration) for the acceptance procedure
p(5c,R) < p
"ince Pr^p < *"j - Pr--J_x kS < j
whenever k - Jr (1-2^
The procedure to accept if x + kS <; is of course the one commonly

used in place of the sample standard deviation in variables sampling.
A heuristic justification for the statistic p(x,ff) as an estimate
of p is the following.
It has been verified by numerical investigation
U-x
that the statistic C-Tr" is approximately distributed as non-central t
"v* whenever /x, and

with degrees of freedom v1 and eccentricity ii
V +1 "^sr,
V-/U
/
& are such that -~- is sufficiently large, i.e., when the population
fraction defective is small.
The numbers ot and i-' are constants for
each fixed sample size n and are determined as follows Under the
U-x is distributed as non-central t , i , TK for
assumption thatcx.-^ff
v>,fv>+l p
small p then
where xlt1 and s, are the sample mean and sample standard deviation
based on \} observations. Equating the first two moments of the
k
statistics x &- fi and x ,
- s^ and solving for OL. and V
gives ot as a function of V and V as a function of n and k.
In
),
- 21 ~
d2[y"
fact oc
where d~ is defined by ER" = d-cr> and e is defined by
r*s~
E>J-=- c~7>.
In order to make the constants independent of k the
values ior the limiting case, i,es; when p goes to zero, were chosen^
This approximation was verified by the authors by numerical
U-x
integration for the statistic -= and reference to tables of the nonIt
central t. It was found to be very good for both large and small n for
most of the QL values used in this report.
small n regardless of p.
It is especially good for
For the larger values of p where n is large
the approximation did not hold with great- accuracy.
However, it
could ha. been improved by making >;, and hence oL, functions of p as
well as of n.
This was not done because this would have entailed
increasing the number of tables of estimates, thereby complicating use

of the tables by sheer bulk.
Furthermore, the effect of this en the
estimate pKx,R) would have been slight since on the average the sample
T1
values of lir- fall in the region of the tables where the estimate p
R
changes very little as n changes.
Essentially the above argument indicates that p(x,R") is approximately

distributed as pXx^.jS^).
Hence $(x,R") has sampling properties similar
to that of p(x^,Si) where the effective number of observations is V+1.
This value of \> obtained in this way is exactly that obtained by

Fatnaik who suggested that [n c =s= may be approximated by noncentral t with degrees of freedom ;/ ^'it with eccentricity Tn ~^r.
This approximation is excellent and will hold for all p but
would not fit into the framework of the proposed estimate ^(x,R).
Sss P. B. Patnaik, The use of mean range a? an estimator of
variance in statistical test, Biometrika, vol. 37, June 1950.
- 22 -
Bibliography
[I]
Blackwell, D.,"Conditional Expectation and unbiased Sequential

Estimation," Annals of Math. Stat., ?ol 1? (1947), p- 105*
[2j
Bowker, A. H., and Goode, H., "Sampling Inspection by Variables,"

McGraw-Hill, 1952.
[3]
Craig, Cecil C4, "Note on the Distribution of Non-Central t With

an Application," Annals of Math. Stat.t Voi 12 (1941), pp = 224-228.
[4]
Ireson, W. G., Sampling Tables for Inspection by Variables, Applied

Mathematics and Statistics Laboratory, Stanford University, Contract
N6onr 25126, Technical Report No. 7, 1952.
[5J
Ireson, W. G.. and Resnikcff, G., Sampling Tables for Inspection

by Variables Based on the Average Range, Applied Mathematics and
Statistics Laboratory, Stanford University, Contract N6onr 25126,
Technical Report No. 11, 1952.
[6]
Jennett, W J., and Welch, B. L., "The Control of Proportion

Defective as Judged by a Single Quality Characteristic Varying
on a Continuous Scale," Supplement to the Journal of the Royal
Statistical Society, Vol. 6 (1939), pp. 80-88.
[7]
Johnson, N. L., and Welch. B. Lv, "Application of the Non-Central

t-Distribution," Biometrika. Vol. 31 (1940), pp. 362-389.
[8]
Lehmann, E. L., and Scheffe, H., "Completeness, Similar Regions,

and Unbiased Estimation," Sankhya, Vol. 10, Part 4, November 1950.
L9J
MIL - STD - 105A, "Sampling Procedures and Tables for Inspection

by Attributes," Sup't of Documfjnts, Government Printing Office,
Washington D.C., 1950.
[10]
Patnaik, P. B., The Use of Mean Range as an Estimator of Variance

in Statistical Tests, Biomotrika. Vol 37, June 1950.
[II]
Resnikoff, G., A New Two Sided Acceptance Region for Sampling by

Variab3.es, Applied Mathematics and Statistics Laboratory, Stanford
University, Contract N6onr 25126. Technical Report No. 8, 1952,
- 23 -
[12]
Statistical Research Group, Columbia University, "Techniques of

Statistical Analysis," McGraw-Hill, 1947.
[13]
Statistical Research Group, Columbia University, "Sampling Inspection," McGraw-Hill, 1948.
[14]
Wolfowitz, J., "Confidence Limits for the Fraction of a Normal

Population Which Lies Between Two Given Limits," Annals of Math.
Stat., Vol. 17> December 1946.

Sampling Plans For Inspection

Încărcat de

Informații document

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

Sampling Plans For Inspection

Încărcat de

Drepturi de autor:

Formate disponibile

rrat fo.

Sampling Plans for Inepectio; by Variables

aplac Table V - Matter Table for Bornal and Tightened Inspection

July 15. 1955

SAMPLING PLANS FOR INSPECTION BY VARIABLES

GERALD J. LIEBERMAK AND GEORGE J. RESNIKCFF

TECHNICAL REPORT NO. 13

PREPARED UNDER CONTRACT N6onr-25126

GERALD J. LIEBERMAN, DIRECTOR

APPLIED MATHEMATICS AND STATISTICS LABORATORY

SampDing Plans for Inspection by Variables

Gerald J. Lieberman and George J. Eesnikoif

sample drawn from a lot of manufactured items is classified simply as

A random sample is drawn from the lot and

When the classification into defective

arid non-defective is made on the basis of the measurement of a variable

Variables sampling plans pertain to a single

quality characteristic, and it is usually assumed that measurements of

Such an assumption will be made throughout the

average range plan3.

Known standard deviation plans are based upon the

sample mean and the known standard deviation.

Unkr?>wn standard deviation

Thcsa pi ins are classified

according to sample size and Acceptable Quality Level (AQL).

ability of acceptance at the AQL was done to be consistent with

Each CC curve shown represents a variables sampling plan oased

In other words, if the user chooses an OC curve,

he nas at his disposal the choice of fhe three types of plans,

The usual acceptance-rejection procedure in lot-by-lot sampling

Similarly, sampling inspection by variables can be

viewed in the same manner.

In this paper, the acceptance-rejection

procedure is based upon estimating the percentage defective in

All existing sampling procedures for two-sided specification

For this reason, it is not possible

to compute a single OC curve giving the operating characteristics

Such plans do not yield a constant prob-

ability of acceptance for a given percentage defective, but

Computing OC curves for

all possible divisions of p results in a band of curves.

percentage below the lower specification is estimated and the

is accepted if and only if the sum of these stimates is

The two-sided regions for unknown standard deviation plans in Sampling

- 4narrow as to be, for practical purposes, u single curve.

General Inspection Criteria.

If only an upper specification licit is given, the item

is considered defective if its measurement exceed? U,

specification is required, the item is considered defective if its

If both upper and leper specification

limits are specified, the item is considered defec-ive if its measurement

If the percentage is sufficiently small, the lot is accepted,

Since' such knowledge about the percent defective is

rare, a logical procedure is to estimate this percentage defective fr JO.

an upper specification limit is given, the percentage above this limit,

If p.. ip , l-hc lot is

If a double specification limit is given, both p and p. are

If p p + p. s p , the lot is accepted.

In the ensuing aaerinns sampling plans based on known standard

deviation, unknown standard deviation, and average range will be presented.

Known Standard Deviation Pl*ns.

of the characteristic measurements is known.

A complete set of sampling