Critical Values for Duncan's New Multiple Range Test

Author(s): H. Leon Harter

Source: Biometrics, Vol. 16, No. 4, (Dec., 1960), pp. 671-685
Published by: International Biometric Society
Stable URL: http://www.jstor.org/stable/2527770
Accessed: 20/05/2008 17:40
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=ibs.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

http://www.jstor.org

CRITICAL VALUES FOR DUNCAN'S NEW

MULTIPLE RANGE TEST
H. LEON HARTER

AeronauticalResearchLaboratories
Wright-Patterson
Air ForceBase, Ohio, U. S. A.
SUMMARY

David B. Duncan [2] has formulated a new multiple range test

making use of special protection levels based upon degrees of freedom.
Duncan [Tables II and III] has also tabulated the critical values (significant studentized ranges) for 5 percent and 1 percent level new
multiple range tests, based upon tables by Pearson and Hartley [8]
and by Beyer [1]. Unfortunately, there are sizable errors in some of the
published critical values. This fact was discovered and reported by
the author [4], who instigated the computation at Wright-Patterson Air
Force Base of more accurate tables of the probability integrals of
the range and of the studentized range than those published by Pearson
and Hartley [7, 8]. This extensive computing project, of which one of
the primary objectives was the determination of more accurate critical
values for Duncan's test, has now been completed. The purpose of
this paper is to report critical values (to four significant figures) which
have been found by inverse interpolation in the new table of the probability integral of the studentized range. Included are corrected tables
for significance levels a = 0.05, 0.01 and new tables for significance levels
a=
0.10, 0.005, 0.001-all with sample sizes n = 2(1)20(2)40(10)100
and degrees of freedom v = 1(1)20, 24, 30, 40, 60, 120, o.
INTRODUCTION

Multiple range tests are used for testing the significance of the range
of p successive values out of an ordered arrangement of m means of
samples of size N, where p = 2, ***, m. First one tests the significance
of the range of all m means by comparing it with the critical range for
the desired level of significance. If the range of all m means is found
to be significant, one next tests the significance of the range of (m - 1)
successive means, omitting first the largest and then the smallest (or
vice versa-order is unimportant) if either of these tests on (m - 1)
671

BIOMETRICS, DECEMBER 1960

672

means shows significance, one then proceeds with tests on (m - 2)

successive means, and so on until no further groups are found to have
significant ranges. Whenever the range of any group is found to be
non-significant, one concludes that the entire group has come from a
homogeneous source, and no test is made on the range of any subgroup
of that group. Multiple range tests differ from fixed range tests in
that the critical range of p means usually decreases as p decreases,
rather than remaining constant.
The new multiple range test proposed by Duncan [2] makes use of

special protection levels based upon degrees of freedom. Let -y2,e

1 - a be the protection level for testing the significance of a difference

between two means; that is, the probability that a significant difference
between sample means will not be found if the population means are
equal. Duncan reasons that one has (p - 1) degrees of freedom for
testing p means, and hence one may make (p - 1) independent tests,
each with protection level y2,af . Hence the joint protection level is
'Yv,a = (2,

(1

()

that is, the probability that one finds no significant differences in making
(p - 1) independent tests, each at protection level 72,a is 'y2P , under
the hypothesis that all p population means are equal.
CRITICAL VALUES FOR DUNCAN'S TEST

On the basis of protection levels 'yp,,e given by (1) for tests on p

means, Duncan [2, Tables II and III] has tabulated the factor Q(p, v, a)
by which the standard error of the mean must be multiplied in order
to obtain the critical range for Duncan's new multiple range test, for
a = 0.05, 0.01. In the sequel, this factor Q(p, v, a) will be called the
critical value or the significant studentized range for Duncan's test.
As mentioned earlier, Duncan's tables of significant studentized
ranges are based upon tables by Pearson and Hartley [8] and by Beyer
[1]. The tabular values for 2 < p < 20 and 10 < v < co were obtained
by inverse interpolation in the Pearson-Hartley tables of the probability
integral of the studentized range, while the remainder of the values
were computed by Beyer, using new methods. The Pearson-Hartley
tables of the probability integral pP,(Q) of the studentized range, with
v degrees of freedom for the independent estimate S2 of population
variance, are based upon their earlier tables of the probability integral
Pn(Q) of the range of n observations from a normal population. To
correct for finite degrees of freedom, they use the relation

,Pn(Q) Pn(Q)+ vFan(Q) + -2nbn(Q)

(2)

MULTIPLE RANGE TEST

673

The tables give values (to four, two and one decimal places, respectively)
of P.(Q), an(Q) and b.(Q) for Q = 0.00(0.25)6.50 and n = 3(1)20, with
the observation that the results are somewhat inaccurate for small
values of v(<10) and large values of Q(> 6). Actually, the tables are
inaccurate not only for v < 10, but also for values of v up to about 20,
and the inaccuracy for high values of Q is much greater than was
anticipated. The inaccuracies in the Pearson-Hartley tables, which
were due to the limitations of formula (2), in turn caused errors in the
published critical values for Duncan's test. Beyer was aware of the
difficulty for v < 10, and attempted to correct it by adding a term of
the form v-3cn(Q) to the right-hand side of (2). This alleviated the
difficulty to some extent, but did not remove it, and nothing at all was
done to correct the inaccuracies for v > 10. Having first become
aware of this situation during the course of an investigation of the
relation between error rates and sample sizes of multiple comparisons
tests based on the range (see reference [3]), the author [4] reported it
in a paper, presented to the American Statistical Association, which
included an outline of plans for the computation of more accurate
tables.
COMPUTATION OF THE TABLE

The computation of more accurate critical values for Duncan's

test required the computation of a more accurate table of the probability integral of the studentized range, and this in turn required the
computation of a more accurate table of the probability integral of
the range. Dr. Gertrude Blanch gave invaluable assistance in the
numerical analysis. Donald S. Clemm programmed the computation
of the probability integrals of the range and of the studentized range
for the Univac Scientific (ERA 1103) computer. Eugene H. Guthrie
programmed for the ERA 1103A the inverse interpolation necessary
to obtain the critical values for Duncan's test.
The methods of computation of the probability integrals of the range
and of the studentized range, together with voluminous tables, have
been reported by Harter and Clemm [5] and by Harter, Clemm and
Guthrie [6], and will not be repeated here. The method of inverse
interpolation employed, an iterative one suggested by Major John V.
Armitage, involves the following steps:
1. In the table of the probability integral of the studentized range for
n = p and the desired value of v, find the two successive probabilities,
yo and Yi , between which the requiredprotectionlevel P = ay, =
(1 - a)-' lies. Call the two corresponding arguments (studentized

BIOMETRICS, DECEMBER 1960

674

ranges) x0 and x, , respectively. The required studentized range

Q = J(p, v, y,a) will lie between x0 and xl .
2. Compute the tolerance T for P corresponding to a tolerance 5 X 1Ou-5
for Q by means of the equation T = (AP/AQ) X 5 X 10"-, where
AP = y- Yo, AQ = X1-xo
and u is the numberof digits before
the decimal point in numbers between xo and xl .
3. Perform linear inverse interpolation to find an approximation x to
the required R(p, v, 'Y,,) using the relation
X = [(Xi - x0)(P

Yo)/(Y-

Yo)] + Xo .

4. Perform direct interpolation, using Aitken's method with a tolerance

of 5 X 10-7 and with provision for up to 16-point interpolation if
the tolerance is not met for fewer points, to find the probability y
corresponding to the value x of the studentized range.
5. Compare the result y of step (4) with the required probability P,
using the tolerance T computed in step (2):
a. If Iy- P I < T, stop and set R(p, v, Yp.a)=X.
b. If (y - P) > T, replace Yi by y and xl by x, then repeat the
process, starting with step (3).
c. If (y - P) < - T, replace yo by y and xo by x, then repeat the
process, starting with step (3).
Once R(p, v, 'yr.a) has been found, the critical value Q(p, v, a) for
Duncan's test is determined as follows: Q(p, v, a) = R(p, v, z a,) for
p = 2 and Q(p, v, a) = max [R(p, v, 'y,,,a), Q(p - 1, v, a)] for p > 2.
The results are given in Table 1.
Values for v = co, obtained by inverse initerpolation in the table of
the probability integral of the range, are included for convenience in
interpolation (linear harmonic v-wise interpolation is recommended).
ACCURACY OF THE TABLE

The table of the probability integral of the studentized range, on

which the table of critical values for Duncan's test is based, is accurate
to within a unit in the sixth decimal place (except for values of the
probability greater than 0.999995, which are giyen as 1.00000), and
the interval is small enough to make interpolation possible. The

tolerancefor the direct interpolationwas set at 5 X

10-7

So

that the

interpolation error would not add appreciably to the error already

present, and hence the interpolated values are substantially as accurate
as the values in the input table. Inverse interpolation is, of course,
not as accurate as direct interpolation, the error being AQ/LAP times
as great for inverse interpolation as for direct interpolation. Thus the
tolerance for P was found by multiplying the tolerance for Q(5 X 10u5)

MULTIPLE RANGE TEST

oo

0)
cl

-4

1-4
.

00

00

,d4
0
m
-4

00

0)

cli

0
m

1-4
oo

0
t-

--I
oo

0)
N
o)

ce) m

00

-4

00 t.

675

00

t-

00

t-

0)

00

00

Lo m N m
mmmmmmm

0)

LO "
.

cli cli

m
.

00

00
0

00

00

00

00

cli

cli C

cli

LO

"

0)

oo
LO
00

t"
00

0)
M
00

0
t0)

--l
--I
0)

00
t00

00
LO
00
.

00

m
00

cl

cl

cli

0)
cl
o)

0
m
m

oo
0

0
t0)

00
t00

t"
00

0)
M
00

00
Lo
00

tn
m

00
m

00

C)
m

-4

oo

00
CA

o
m
00

00

00
m
00

00

t- m
00 0
.

00

0)

0)
0
cli m
.

0)

0)

cli

cli cli cli cli

cli

tLO

0)
o

m
00

t0)

cli
-4

C)
"

0
0)
00

m
0
0)

00
-4
0)

000000000) C)
C

0
"
00

M
"
00

V
"
00

o
LO
00

to
00

cli cli

cli

cli

cli

cli cli

cli

cli cli

LO

"MMMMLO
tLO o
00
00
00
.
.
.

00
00

0)
00

00

00

00

00

00

00
t00

0
47

(o

0
.
00

Cl
.
00

Lo
.
00

-4
LO
00

0)
LO
00

o
o
00

m
00
00

cli

cli C

0)

LO

00

00

00

m
LO
00
.

00
LO
00
.

cli

cli

cli Cl cli (T

00
t-

00

t-

0)

00
t-

00
Lo

t"

0)
M

LO

00

00

00

00 00 00

00

00

cli

cli cli

C) C) -4

C)

00 00

0)

M
M

oo
C)

t0)

t00

10
00

00

M
00

cli cli cli cli

cli

t-

0)

oo

00

"

00

00

cli cli C

00

00

00 00 00

00

00

t- 40 t-

00

m
ko
cl
mmmmmmm
00
00
00
.
.
.

0
t-

C)

oo

00

lf)
.

cli cli

cli cli

00

C)
m

00

00

cli

00

00

oo

00

cli cli

0
m

-4d4

00

cli

cli

00

0)

t-

00

cli cli

"
00
C

cli C

t0)

m
.

oo 0
.

00

00
Lo
00

oo

00
t00

cli N

00

00

Cj cl

cli

00
t00

Cl cl

00

m
m
00

ol

0
t0)

m
-

"

00

00

cli cli

oo

m
ko
N
m
MMMMMMM""
00
00
00
00

00

00

--l
0)

00

0)
m
00

00

t-

-4

00

cli N

0)

cli cli

Ml

m
m

00

cli

00

00

00

00

00

00

00

00

oo

00

00

cli cli

00

00

LO

00

0)

0)

cli

t-

cl

cli

cli cli

cli

cli cli

LO
m

tm

0)
m

cli
"

00 00

00 00

00

cli C

N.

N.

N' N'

N' N.

CII

0)

00

00

m
00

C)
m
00

0)

m
00

00

00

00

00

00

cli
0)
00

00

00 00 00
N.

N'

00

N
00

00
m

"tv
1114

00 00 00

N' cli

cli cli

LO

LO

"

"

cli

00

00

00

00

cli

cli cli

t0

n4
C)

00

00

cli

cli cli
M
-4

Z
C)

00

Mm .
m

00
0)
cq

00

tn

C9

W
t-

00
Lo

t"

0)
M

oo
.
. (o
. oo
. (o

oo

0
t-

-4
-4

00
t-

00
to

LO
M

m 0 a, 0) oo oo (o (o

M.C cliC cliclicli

W
4
z

0)

00

M
Lo
00 00 0 (O
. . . .

oo

0.

00
t-

-4
oo

0
t-

-4

00
t-

tLo

oo oo

o
00

cli cli

m
m

00

t-

m
m

oo
C)

N
M

0)
N

tN

ko
N

0
m

00
cli

cl

0
C'l

t-4

LO
.-I

cli

cli

C'i cli

cli

cli cli

tcl

-4
cq

N
0

00
0)

n4

00

cl

cj

(o

t-

cli

"

C'i cli C'I cli

0)
-4

O
-4

0
-4

00
0

0)

0)

00

cli

cli

cli

cli

cli

m
oo

o
t-

0)

o
0)

M
0)

00

a,

0
to)

0)
cq

o)

00

cli

0)
C9
o)
.
00

-4
00
0

0
m
m

00

m
--I
.

cl
oo

oo

oo
0
a,

00
0
oo

oo

oo

00
(o

t-

t-

t-

LO

cl

cl

cli

cli C

00
t- t-

Lo

Lo

t-

-4

t-

t-

t-

cli

cli cli cli N

"

00 0)

cli

"
o
0)

00
m
00

1-4
tt-

0
M

0
0)

-4
00

-.14
-4
t-

'n cli cli

cli cli cli

M

00

0
m

"
t-

"
m

o
00

N
00

0)

o"o

cli cli cli

cli cq cli

cli

cli

cli N

"
0)
LO
.

m
o
LO
.

0
"
Lo
.

N
Lo
.

ko
0
Lo
.

cli N

t-

0)

00
N
M

ko
.

0
LO
00
.

00
"
t-

00

cli cli

0
00
o

0
m
o

"
t"- t-

Lo
C)

N
.

"

0)
"

t-

C4C C

0
t-

ko
o

0)
Lo

cli

:,i ol

00
00

o
0)
.

0)
0)
tO
-4d4 ,d4

C4N'

N'

t-

00
.

0
LO

cli
cli

00

N
00

0)

"

00

t-

cli

Cq cli C

N'

Cj
m
kO
.

n4
-4

t,
C)

C'l C'l N

C'l

"
"
M

to
cla
M

cli

cli
M

ko
-4

ol

oi

0
LO
ko
.

t-

cli

-4d4 00
to
LO kll
.
.

cli

cli cli

-4

cli

(7,i cli C,

00
"
t-

00

(7,; (7,;

0
C)
00

cli

0)

oo 00

C11 C

0
0)
00

t-. t-.

t-

n4
00

cli cli

cli

(t:

00

0)
00

ko

-4
0

00 00 00 00 00 00
cli cli cli cli cli cli

. . . . . . . . . cli cliclicliC C'IC

8

Cl

0
N

00. 00. 00. 00. 00. 00. t-. t-. t-.

00
0

(o

0
-4

-4
N

t-. t-. t-. t- t00000000000000000000

00
.

0
0-

N
N

00 00 00 00 00 00 00 00 00 00
. . . . . . . . . .

cli C)

00

cli

cli

0)
t-

'N

0
O

N
Lo

ko
,d4

0)
M
-4d4

0
0
cli
0
-4,J4 "

00
M

N' C

cli

cli cli

cli cli

t-

00

0
cq

C14 m

m
o
M

CII
co
.

-4

P4

00

-4

-4

-4

0)

"Coco

11,54 o

cl

676

BIOMETRICS, DECEMBER 1960

o
o
O000r co
CS

O o

?
Co

o
o0b

_ 000
o '

Cd

CII

0:

c7
00

_
o
co4 o e

o
0:
000Q
00? .X M0

~ ~ ~

,See; 11

?o

Z*

-4

CO

Cl

O_

aoo

00

":

00 0000

cn

.0 00 .

_0u04

-40
d1
CD
ceo
0 cl

000
?? mi4
X
0

Cl

ClClC

oo

_/

cC

c0

CS

-4

C0

D m

C) O

0o

lC C

sce

Cl

CD

00a~C

cl clCl000
~~~cl

00

00?
00 00
Cl ClC
0 e
d4
"

CClCC0c
Lo t
tt-C)
0
ClClClC

Cl
ko

00 00 00Q 0
cl

cl

Cl C0

?00 c) eo
tmr

clC Clcll

L
00 0e 0C0 m " co
0
00 00 00 0000
000
Ca Cl Cl

oO

t-

'

clclMl0
-

00 00 00 C) O 0

00
00

Cl Cl
Cl CS
Cl CCl C1
Cl ClQ CS
Cl0
Cl C1
Cq C
Cq Cl
Cli CS
.............. CS
CS Cl

..

00

SC

00

Q0O

t- a

00 00 00
CS CQ CS Cli CS M

CS CA CS CS
to
CD

t-

oo

00 0

"

0to

O0C0l000i
CS CS Cli Cli

Cl

0 0 c0u .0

Cli c

CID

w CCClC0
w

?O

CS Cli CS ..........
CS
o

00 00 00 0000

t- M00
000a,
Cl
c cccco
c

ClCCCCCCClll

C)

0000

CCl
oi

.
Lo

000000 00d
CiC

00 t

t- co00000tCq, CS CS C]

ol

O t-

00 00

0 Coo

0,

CC

cl0
ClCCl
C
kwl 0000'-4t.-0C
oO ..

o
- 0
t- Cl
m ClC
tm Cl
c
O4 Cl
Cl

0
m
co

"

00
1- Cl
C)
Cl

C0

00
Cl

m em cm

to t-

u"

0000000000)
ko mClN Cl mCl " Cl Cl Cl 0 Cl m
Cl
Cl
CS

~O

u e
00 00 c0

-o00

ClCl

00000
0

_C)

0
0000000000000000
cococoCID0.

oo

> 8t

00

Cl CCl

.Q .

Cl Cl00040000
Cl Cl Cl
Cl
Uww
0~0 ~000
o 00

0000

. 0000000t000
.
.
.

CID
C OD
lCl CO Cl .Cl. .

o0 00cl00
Clclclclclclclclcl

Lo

00

00000
00?

o
Cl

00 00

o0-

M 0

Cl oo000
Cl CS CS
Cl

ooo000-I
cO

cli Cs
00

cs

to 00 0
CO0 CO
l
CID CID c :CI co> ;>C H

Cl

r
O C) C) O
~~~~~~
??
Cl Cl ClC
C0
o,
000

Cl

ol

Cq

,ot

~~ce)

Cl cli

0l

SC

t-C0

Lo

_o- o
Co0000000000

001? 00
Cl Cl
m4C
oi _crC) oo)C)0)

Cq

Cl 0 0
P:

t-

0,-00t-0Cq
CD

CS
0co

0
Cs C
cq, cs
W

00 00 00 00 00 00 00 00 00 00

t-

00
Cl

o_o tH

cs

C)

00 00 00

%
00

SCSC

00~~~0000ClClClClClClw
>

LO t-

_
00 c o
C0)o

O_ C
t- CS
CO 00
CS CS
CS

0000

00 00

M
000

co co
u
00
0~~~~~~~~~~~~~~~~~~~~~~~~~t-~~0
C
Z

t-

SC
0O
a

t
00

d4
" d4ab

;5

00000000000000000000

00 0000 0
0o oO00

000
00??

t
00

?? d10>XC

00

Cs cs

Cl

oO

C
000 0
tkllt-0000-

~~ ~ ~~~~c
11

cl
co

clCl

o-t

Q)9~40-m
Q O
.............00+00~....
cq o-0OO
o tcra _ 0) 0 o o

CO

00

U0

00

0000
0000000000
Cl Cl CS
ClC
Cl Cl

Cl ClC

t-

t-

co

ClClClClClClClClClCl
m
m
m

00 00-)

o mC lC

00

0 (Z

Qa)Q00
00 00 00
X C
? 0 .0C CS CS lcl c
Cq Cl CS S
Cq Cq cCCCl

m c0

CO co

0000000000
ClClClClClClClClClCl

0 oO00

Cl Cl

0
D0
C
dc
"
0 00 00 00
00

c.dl0

u CO
0 Cli 0C'I CS CS
CO
CA CS0 CA CS 0CS CS

o000

c0000

00

o:

0014C0C0ClC~~t-LOCClC
t- to "
M w

z Z 00
300
>
crW
cn
_^

d
"

o
CS CS cs cs cs cs
COtN0M
0

X~~~~~~ ~

X 0C
m
00

00 CS
00 C]
C
t- cS
CS

cr

c
d4c.1O00040
0

)
a
0?

Cl~~~~I~W

0
o

X ??

e
0

0-4
: :) 0000
00~~40000ClClClClClCl

C-

-to0

oo tO _O
CO0OCs N'

?0

o o
t-o

0o okoo
oo
oO ?I
C0t
C) 00 00 X

oo00

a)

.z

_
-

0 0

"

00

CJ

c 00 to
c
00 000 0C C0 0) oo
Cl Cl Cl Cl Cl Cl

00 00 00 00 00 0000 000000
ClCCl
Cl Cl Cl Cl Cl ClC

00
to000000000

cC

t-

CY)0o

o0

C)

0 000

000

?
00

w0
00

00
00000000

000
0 w

C?S

cra

00

_ X ocra
ooo
Cl

e0
0 00

C00

O00
C)
000 C)c000

~I00Cl%%

MULTIPLE RANGE TEST

to)

~OeO"
m

oo

tkr

t- bO
0) N

00

O)U

677

ca
t-

t-

cO

cra

.
H cO

C"

de

CS

~O
~O
o
.
.
.

ko
.

+X

co

~O~Oo

ko

CO Cd

t-

cn~~~~~~~~~~~0

co-4

t-

O.

eo

D~~~~~~~~~~~t

cO
o

O
ko

LO LO

tco

t_

t-

0)
cr

~~~~~~~~~~~~~o)

t-

tO

00eO
.
.

cra
0)

oo

crO

a,

cO

~Oeoo

ca
o

...

t-

oO

b
00

Co

t-

oo cra

eO X

CS Cd

LO
Ce cra

"

eo

Cd Cd Cd Cd C
eo

csH

uO

LO

cra

XXI LO
X 0

LO "

O
e

LOeOO

O
.

C t Cd u:

eo eOLO

t-

tLO

+O

oeOe
.

eo

eo

cra

00

00

" M CS
tt-

t-

Oe

dC

co

t-

Oo
00

O C
CS
N

OL

ko
.

0)Nr CS

O00

CO cra

e+t

cao) 00

Oe

00t_
.

u0 CO
de
O
O00
t_u

u:

0o
0) 00
de de

ks'

00

eO eO eO eO eOe

o)

-4 oO
00
t-

ca

_I

cra

oo

+N

c)

00

t-

cIO

? Hb oU:
;

O 0

cra 00

cl

Xt-

cr

Cl t-

Cl

O ca

oeo

eo

eo

t+"

ksl
N

00

HO CO tee

b~~~~~tb X O X
m O
+
m
t~~~~~~~~~~~~~~o
t <o

_s

) +O

-4

P t

t-

o
oo

00

-4

t-

CO LO LO d" de
eo

eo

eo

eo

C) 0)

oo
t-

t- t-b 0)
O ko

eo

eo

eo

eo

o 0)
m cl

r Q H b o X o mOO
eX
+
X
+
LO

u:>

EA

b u:
od

oo

t-

CO X

de

eo

) LO 0

OcO
LO
-4aCS t- d0
M

eo m 00
00
CO kO

~OLO LO LO

~~~~~~~~~~~~~~~o

O%
C!

C!

Af

00

to

O e

LO

t-

w
r

o
t-r

C) 00

t- O

"

"

to
m

t-

cq

00

cli

ko cli

t-

0)

kl
00 X

CO C t

co

t- Cd

?00 - O

Oc
dt-

Cf

"

(0 u:

0 t- a,

oO

cra

CD
C+ C+
M. co co coc

m
C
cra

co

co

co

tCO+C
Odq

t-

00

Cd
(M

CS

cra
H

O o
t- >X>

Cj

oo

b
M

i:
H

oo
ri
O

u:

CS

co

de
m u:

-!

0 t-)

t-

O
00
cri
O

cr

cl

CO CO CO

to

CI

O cr
-N c
-1
co

co

" ut
0C
ksl

co

M kl

oot-4
00 t- H

to

cr)

t,
oo
cr)

ts
ca

M
cO
cr)

co

tO ks

co

co

CO
to

co

t-Y0

CCt Ct CliCt
(D

U0 00 (

o o

CA

bt

Ct cl Ct CS Cli Cl

r
"c

r)

I Ct "DCr
! (: C! o

O ot
o

o0
(m

CO
"
Ct t4
OO

Cl

O
COCt CO
C t Ct

L
o

c CtC CO C' Ct CO Ct Ct CO Cli

Ct
-4
C
N

co

oo0

-4

o co c

to C

co
COLa

C Oko

~ CCtC

COOC tC

k t- -4

cl

) 00

LO

cO

m a,

0)

" O 00 eO
r00LO
cra
c cra oo
ci

e de

00

Ob
m cl

o Xo

t-

LO cra oo tcl
t-

eo~~~~C~
c c'! CS
!

C)

00

t-

ooe

) 00

M
X

eo eo eo eo eo eo

u:

oe

-4

de de de 4

H O+tCt
0

S C
0)

ksl t-

de

eo

ti

sH CS
N eoC

eoe.

??
"

N
o
*

cl

XX
tC) 00 t- LO m

O 00 O O e

?O

Cdo) 00

eo

M M m CX
M CO co CO
X M
m X

~~~~~~~~~tb
O m00de
O) O5 tO~~o
td Cb
oO
-< -O
~~~~~~~~~~~~cra
. oO CS
e

Al
CS

"

to

ksD e

de de
mm
C~~~~~~~~~O
mc

P;
ro

;~~~~~~~~~~~t
V

C~~~~~~~~~~~O
~~~~~~~~~~~~~~~~~~~

X~~~~~ ~
?

).

eo

o +

>~~~~~~~~~~~~~t
ksl

0
(
MOCDb

i oo
m
Cr)oo

0C)O
"
oo

tl
oo

BIOMETRICS, DECEMBER 1960

678

. o
to

0X

<;

CD~

;>

o) 00

0X

00

LO

CO

CO

CO

t-

00

O>

cl

ux

*^

V 0r
Effl~~~~~~0
z
_

;~~~~~0

;~~~~~t

o?)

Lo -4

00

~od

OO

O4

00

00

t-

t-

00

00

t-b

t-

.
.D oo

00

t-

to.q

to

cl

-4

to.Q

10

4
Cl

t0S

O
.

4 4

O O 4

OC

.D

"C

to

"

0) O

-ICH:d!

t0

00

00

Hcr

ot-

to

t-

d"

"i

to 0Q
l

t-

to

U)

U)

Cs

oo

U)

CO

0)

cl

m
c

oO O

CtO

t-

Co

t-

r-

tC

"

LO

4o

co co

t- tt t

???
.

to
00
t- wOO4
t-

O
cl

dq

00

O m to
0,

00 0 Cli to b
to to to t

t4

dq

O OC

mOC

0OO4
O

CY)

.
t-

.
.
tt>o
tt-

tt-

to
00

t-

00

t0t

t-

t-

t-

co

co

co

.
t-

t-

1-

to

CQ
cl

co
to

0q

OC

0)

O OC

coO

co CO CO C

to
w

00
b

to
"
d"
t- t-

1
w

w
0

qd
q
I"!
C C OC

tO

CO'

CO' CO' CO' CO

-CO

t<d>

to to r- 00 0)
<r

>

0q dqd

coOC co

t
..

cl
...

to

00

4..

t-

co

O
C

t-

-4
00
dq

C0

~o

COC

t-

co

Cr

o
to
00
dq

0 0 cot

t-

t-

CSl

ttC
L O
CO

OC

eo

00 0

t-

0 t- t- t- t- t- t

t- 00 0 0
cl co
cl
P~~~~~~~~~~~~~~~~~~~'
C:+no

t-

00

t- t
t t

00

C)

CO

CO
m
m
CO

t-

o
t-

t
t-

00

ICOICOICOCOC

dq

to

sm

oO O

t-

OC

t-

00t- t- t- t-

dq

CO

t-

4 t 0t o 0 t- t t- t- t-

4 4

44

t-bObo
to

"
CO
m

to
o
O

-4 V

" t-

to

t-

o1

.Dm
.

OC

0b

0) 00

toD

Wo
.

OC

00

c t;

10

X~~~~~t
r~~~~~~o

O
k
CO

0M

00

.q

LO

00

.q

co co co co co co

oo
CYO

o1

CO

O m O O 4

O 4

Co

00

00 00 t-

oo O O 4

oHc

cr oo H
bO 4

t-

LO

C O 0 CIt-

t-

oO t-

ko

CO

O~~~~~~~~~~~~~~c
U

HC)

co

gt"

sK

sO

bO

0)

co
co

co
-4
o0

ca

(r

co

co

0r

CO

to

m m t
co

co

,)

cra
0)

to

Cs

DbO

to

qo

t-

Sbd

X 3
o

t-

mm

crcs
0)

4O

0R

> e

4
LO

Oob
o

Cs

O<C

tosdqd

O~~~~t

ct
m

cD oo

|t-

oo

~ ~~~~~

cn

>

mm

C>0

o)C

t-

X m 3

CCY)~~~c
cso

0?

OC

COC
to to

bCb

cli

679

MULTIPLE RANGE TEST

m 'Ili 1-4

O
ko
t-

"
m

o)
1-4

m
0
t-

t1-4
m

"t

to
0
cli

-.14
"

08D co LOLOLOio LO LO -.14

co "IV

11t,

44

1.14 00
ko
(M
Oo)

00-144
ko
1-4
0)
0)

t00

0m
to
(m
0(M

14
4z)

-4

(O
1-4
0)

-4

(m

to

to

10

t-

-t

tce)

to0
0N
cl

-4

io

io

Lo,

Lo,

ko
c)

t-4

ul

-.t
t-o
o

000co

C'l
M
06

'O

'.14mcl
t-ott-

t-4
-14MN-4
.
.

'.14 co
t-0t-

cl

cl
10
Mt-o
oo

(m

to

to

di
C'lpt-ormt-ot.

C!
O1114
(M

Co
o

LO

LO

LO

cl

-O

(m
m0

<0

MCII
0t-

LO

to

00

1-4

N10
Mt-ot.
.
to
00
-4

M"14

Mcl
t-0t-0ww
.
to

co

to

d4

C14 1-4
(M tt.
t-

c
1-4
00

tto
00

cli
to

to
co
ko

`V
C)
to

-4
-.14
t-

00
tO

ul
-q
to

m
U-)
ll'

0eom
cl
(m

t(M
t.

C14 (n
.I'
ttt.
.

(M
C'l
t-

to
c0mtC.0
10
.
.

t-

0)
to
ll'm-4
tI-

00
t-

-4
to
to

to
00
to

1-4 O
II-)
cli
ll`"mm

(Mm00
to0mt-0
ll'
to

tomLO
1114 -4
00
00

00

0)

t01
-4
t-m0t-""0
a)w00wt-

00c
ko"WkoN0w

1-4 00
Nko
0(Mmwwt-

10

cl

-I

t-

t-

ec

. . . . .

to

(M

ul

.114

t-

t-

t-

-O0mt14
ko
O
ko

lq"41.114'4.4

(m

10

00

C)"0)MCO
00
0'.14
oo
0(m
.
.
.

00
-140C.0M-4
t0c) oo
.
.
.

C)

t-

t-

10"I-iq

to

0(O
(M
c

10
C'l
00
.

O
t-

LO

ko

cli

-4
U-i
t-

t-m1-1
000
t00

ll'

LO

u-i

(M
tm

CD
m

-II0ko
.14
-4
"dim
.

.14
.

cli
0)

tcli

cli

t-

00
t-

-4
t-

mco

00
"Ili

ko
0(

1-4
1-4 ko
"IVmcli

.4

4"V.44-44

to
oo
to

CIO 00
(M cli
to
ko

1-4"t-

eo
to

1-4
(M C14 CIO
mCeD
cli

to

Cl

CYD

to

00

ko

"14

CO0cl

CD "
ko
c
1.144 CID

ko
M
00

LO

1,14 COD -4
.
.
.
ko

00
00
t-N00
(M

mto
0(M

M00
m

10

00
-4
ko

44444

LO

-4
tO
"ko
mt-ot-"m-

00

O
(M
'.14 tO
ko

t-

ll`

(M
co

CD

ll`

M"14

(M

10

t-mcl

. . . . . . . .

E-4

io

ko

.c

00

1-4
co
t-

0
00
t-

-.14

"t

I*
0CD
00
a)

tto
00m(M
a)
(M
.
.

to

to

m'.14-O

ll'

m..i4-CO

-I

O
ko
t-

t-

04

tto
-I0"

Ilqi CID cli

t-0t0t-"MCl
.
.
.

co
LO

1-4
0cl
.Mt"14 06
-4

1.114 1-4
N
0
00
00

1.4
ko
00

eo
, I
0)
00

tko

IV

cl

t-

CO

00

t-

(m

1-4

ko

mmC14
(M

1-4

14M"14
eo

t-0t-

co
o001
(M
4
P

:)

-4

-O

1-4

co"-4
LO t-0ko

(R

C!

-4
cq

t-

c!

(R
-4

m
C!

10

CID

c
10

-4

to0-v

Cl

to

'O

to

000
-4
--4

"Nm
cl
to
(M 00

--4

'I

c!

ul)

III)

'O

-.,v

t00

t-

ko
'.

U
.0

tO0

-O

to
-4

00

t LO' 1

1-4
to
Cl

ci
1ko

cl
0
t-

(M

-4

00

to

0ld4

-4

"

"

ko

U-)

t-

tO
(=I

t-

ko

C)
(M
t-

00

t-

ko

cli

cli

mcli

(M

t-

(M
ko

to

t00
U,

CD
-4
ll'
.

ko

-.14

c
tt-0mtmmcli
.
.

ci
.

r-4

1.14

cl
ko
ko

00
(M0cli
000mt-0m
mcli

co

-4

tmco

to
.

(M0

t t- t t
14

ko
cli

t(M
to

t4

"

10

kO

t4

1.14

-4
tC.0

0)"
t-0
10

10

'14,

1114

11144.44

"
00
14

clia0C)
(M
co
co

0
4

-4

N
-4

(Mmm

M
-4

-4
ci

14
-4

co
cl

00
'o

10
.
w-4"t-0
mmcli

.t-

10

LO

-4
10"'114
14
t4

CID
-14
.

ul

CII
14
.

c
-Ji

Ntow
mcli
4
4

,,
00
co

kO
-4

co

D
-4

(M
(m

t-4

"I' t-

cr

C)

t(M
-4

1-4

1-4

C)

0)

4"'4

-4
(M 00
00
m-O
0m10
cli
1-400m(M
14.

cli

-14

00

cli0

-14 Co
4

-4
CII III)N
(M OM.
CY)MMCD
.
.
.
.
14"
-,14

0,

1-40
14

(M

-4
140t0)
-I'MMMN

-O
cli

D
0

't! c ci cI

-.14

(M"0t-

t-

CII000
CII tO0CO
tO
ll'
10
.
.
.
"4
-d4 ,II

0)

mm
1114 1114 "
to)
Cj

10

eo

III)

-O

0N(M
-t-m0wto

t-

t4

(M0t1-d4 -O
ko"00)
.
k
ko

OM-4
c
c
.
.

U'
-O

"14OOl
C)
toN00

000

mw

C)
--14
t0
.
.

,14

.I'-mmoo

1-4
Cj

eo

C)
eo

0 0 t r,

CO4

ec

t-4
eo

(m
eo

.1.14W"lt4

M0M0-4
(m90Om0ttt-

C)

oo

-O

-t44-I,

Lo
t-

'-q!
ciR

Mt-oOm

000

10

10

ul)M(mM
U-)ww"

ko
O

U-i

to
oo

ul)

t-

tomO
t-m(m
t.
.

-4

ko

t-

ol
co
LO OC) cj
(mWwt.
.
.

tc')

oo

oo

(m

ci

M14

t-0ClO

CO"0O
'O
1-

-4
Nto

ko

111), 1

ll''

1-4 t-m(M
tC)m"moo
cl
M10W

co

0)

to

c t (::

,14
-4

eD

to,

wec

10
t(M

U-)

<0
'.14"C)
tko
(MMU-D
t-010"cli

06

cli

ll'

co

-4
cl

co

ko

00

(Om

10

10

-40

M
R

Mn4

14

00

>

00

10

mt-ot-

11
04

U-i

-q000
0
eo
-40

LO

CPO O

c
10
(M

"'

10

t4-

CO'

cli
cli
(M

00
10
00

CO
(M
t-

c c c CO'

-4
t0C)

-O
-.14

10
io
C)

(M III) cli
oo"Oo'.14
t00
00

clim

1.14

-iq

ce).

CVD' CVD' c

COD' M'

1-4

-4

cli

10

t-

cli
1-4

to

BIOMETRICS, DECEMBER 1960

680

1-4

00

-o

co

C',

-o
U-)

-4

"

'Ili

m
m

cli
0
00

ttt-

'Ili
-o
t-

co

Cl

t-

,I'

cli
C)

tt-

"Ili

to

co
m
00

oo

ttt-

t-

t10
1.114 cli
tt-

t0
t-

ul)
.Iq
oo

m
m
oo

cli
0
00

ttt-

m
t-

0
t-

t10

-,I'

-,I'

.Iq

.114 .114 'It

'It

'It

'It

it
00

00
ul)
00

ul)
114
00

m
m
00

cl
C
00

cj
tt-

0
t1.14 0
tt-

M
t'O

10
m
'O

N
0oo

.t
too

oo
U-5
oo

w)
14
oo

m
m
oo

.t
o)
t-

t-

w)
-I
t-

CO
t-o

0
M
10

14
00
to

cli
0oo

too

00
U-5
oo

.t
oo

cli
m
oo

t-

t-

00
0
t-

10
10
10

0
10

CII
tto

cli
0-

t-

tw)
oo

co
.t
oo

(2)
m

00
00

lll

0)
0
-o

0)
lll
to

0-

t-

w)

CII

Cj
t-

"

'Ili
t-

w
ul)

cli

"Ili

00

00

00

w
10
00

00

M
00

cli
000

t00

00
to
00

14
00

00
U-)
oo

.I'

cl
000

ul)
'.14

to
-o
t-

0
tt-

-o
tt-

-o
ul)

to

"Ili

.I'
t-

m
0

Cl
t-

t-

t,7 C t17 ll

104

.I'
"

-o
o

cli
t-

C Ci

C! :

C C

ul)
1-4
cli

CO

Cj

14
t-

cli

ci
t-

-1
C'l

cli
p
t,

CO
0
cl

t-

to

cli
t-

cli

q
cli

to

U-)

tC

-o
0
Cl

t-

"D

-4

0
.Iq
0-

cli
0oo

. . . . . . . . . .

C Cl!

00

CO

t-4
m

cl

-I'

14

.I'

14

Cl
t-

0
-I'

.I'

0
.t
0

14

. . . . . .

00

-t
0

cli

"It
C'l

C Ci-!

10
o

t0

00

oo
m

CO
o

14

o
0

CO

cj

t-

to

ci

"It
C'l

cli

00

CID "'ll
0

-4
cli
co

to

00

to

14

to

CO

Cl

t-4
co

-o
0
cli

"It
C'l

1-0
0

-4
0

10

10

1-4
10

1-4
-

cli
t0

0
0

cli

0
-t

't

14

ul)

to

0-4

E-4

m
cli
m

"'ll
C!

-4
CII

to

-I'

t-

14
cli

to
0

-14 C

Cl -!

-4
10

1-4
-

CII
t-

0
14

-4

-4

14

to

CO

z
o

"t

C!

1-4
CII

10

cli
tt-4
14 co

CO

co

cli

cli

t'

-4

00

14

CO
.

t-4

Cj
t-

14 M
.

14

o
.

to

"It
o

0
to

m
0

t-I

to

-4

00

-o
0

'14
0

-4
N

0.
oo

0
14

Cl
-4

0.

0.

t00

14

14

ko
oo

14

w)
-O

00

co

14

-o
14

00 00
.

14

M
oo
14

lll

cli

cli
00

co
co

m
00

0
m

t'
14

00
-4
oo

14

t-

14

-I
t-

0
Cl

-4

0ul)

t0

14

14

oo
N

m
-4

00

00

14

-4
Cl

Cj

t-

14

00
-o
14

14

14
-4
lll
4

10 10

It)

t0.14

to

C'l
ll,

1-t1-t1-t114
114
N
t-

00
m

(Z)
0

0-. 00. 00. 00. 00. w.

m
00

't
t- -o -o

cli
ul)

0
(Z)

ll,
m
it)

00
t14

I'll

I'll

I'll

.114 .114 -14

1-t

1-t

I-t'

114

tt-

m
ll,

cli
M

cli
-4

-4
14

to
00

-o
-4

tul)

cli
0

111)
0

1-4
0it)

ul)

I'll

I'll

-4
1-4

0
t-

U-,)
m

8 00000000t-

t- -o -o o -,o -44

0
-4

0)
N

t0

tcli

eo
0

-14
N

-4

t1-4

o
0

-14
N

1-4

t-

-I!C ci -!

ul)

ul)

ul)

ul)

to

t-

-14
N
1-4
-

0ll,
0

to

U-)

U-)

1-4
-4

Cl
-4

C!

0.

10
oo

14

00
m

00
co

to
t1-4

C; C

14

-14

-4

,14

cl
t-

CO
o
cli

10
10

14

14

OD'

'1114 -4
0
CII

14

-4

Cj

00

14

co
00

00

Cl
cli

M
o

"II
0

--4
cli

0
to

-,II

C;'IliC t-

14
cli

-o
M

14
1-4

00

-o

cli

co

.114 It)

-o

to

ul)

to

tl.

00

0-

-o
0
0

0
14

C0

0 0 t,

co

v
-14

00
.

.
14

m
-4

.
14

-4

00

.
14

-4

00

.
14

to
-4

--4
0-

-4

0-o

00

.
14

t-4

114

00
-4

0)
-4

0
N

114
114

140 0 0 0
cli

co

.114 -o

cli

-,II

00
0

MULTIPLE RANGE TEST

--4

C)

C)
-4

C4

ko

4-4

00

t-

CD

to

<0

to

to

00

t-0to

-I

00

t-

(O

mmOC)

LO

co

-.114 (M

CO

CO

U-J

'IO

C,Oto

-4

tO

(O

io co
-I, -4
CD ce)-

mN
Ncq
-4

14 --4 "0
--( mw
.
.
.

t-

06tl: 1 ll

C)

co

00

mm00
mo

P.
fA

w
Q

-4
mmoo
C
oo (7)0
-4 -4 -4

com00
(::

(m 01 t- ol
0'o
com
CYDC> Mt-

00 t-

CO tO LO U'

oo

co

00 (M C)
C4
o
-om

mm00
(7) o

o,,

4-4"I0-4

E--4

00

Pk

co

c;

00

C
-4

C)
-4 00

1
CO O

(: C9mo0om
NNco

to

to

CO 10
o0
ko
10

-4

co

U-) ko

ko

C)
mmC)-40CO

CO
0CO
'
,CII -4
I
oo (M C)
-4 -4 00 t- CO

cli

oo

10

CD co

to

00
CD kil) C)
oo co

t-m
k.0 tC> t- CD 10

LO 10
D0
10
1

06CD,
u-i (M0
C"

t-

CD

to

cl0

(O t-

0(M

(O t-

-4"M

C'I

,II

to

to

m
cra

ko to

to

tCD

tkO

CD

to

t-m
m

Ntm

10ko

-114

--q (O CD
o
t- 00 oo (m
00 t- t0(M

10

14

14

m00
ko to

to

14

t(O t-

t-

00
LO

t-

cli

cli

tmto
-4 -4

00 o to
-4 Nm
(O t- to

t- m
(M0-4

mm

00 t"- to

U-i

t- -4
C0O
C C

to

cl
CO

0(M

ko tko 0-

U- -4

ko

LO

ul

C'l t-

io

co
-4

. . coo. t- . . U'.

LO

cli (M 00 (M
mteo o t- t- 00 0)
(7) 00 t- to to

co to
W Ntto

C)
(M

10 00ko

to

10

-4

(7) "

oo C'l 00

m00

ot-

M
cli clim

( 0 t,:

'"!

t-

-4

t-

10 10

'o

'o

'o

Co to

10M
t0

N0
tN wko"
C) (M (M (M

.
t-

t- 00 00 C (=) -q
t00

00,

C) C) -4
0CO

CII CO CII -4
"
cqNm
00 t- CO to

CII kO
oco
m

ko

m
w
CD kom
W0
0
ko0
(8 C> (7) 0.) 00 00

CY) 00 CO eD 00 ce)
ko ko CO t- 000
mm
ko
t-

CII)
N t- WM
to
-4 CO "(M
(M cl o
C> (D. (M. (M. oo 00 t- . t.
.
.
to

w0
NN
CD t- 00 (M C) cli
m mcli
o . to"
.
.
.
1.14.

N0o
t- t-W
C")N

000
wtcli
U

C) o)

wui
1-4 c
NC)

U-. . .

0(M0
""to) 00

00

t-

00 00

t-

CD t-

"'O

t-

0) C'D00

t-

U- -4t-

0tt-

10

"to

CO

t- C t- -4w0
ko cli 00
"(M
CD UD 0 0

4"'

t-

-4

-4

(M"

1-

00 0)

C)

-4

co 00 ocq
m

U-NN

cli

cli

00

t- ori C) C-1"tmcli

cli

O(M

ul),U

-4

- clim

00

ko.

(M 00

CO (:T) O

-4

-,di

CO

('
-4

,4

ko co
(O t-

N
ko o
ko

ko oo 0)
"O
m

C'I

mcli

00 0M

-4

m"I"
o(M

-4

t-

Ma) -4

"o

(:

m(Z)

C Cl!Ci-! -! R
to

CII -4

to

oo CY)C'lm m
C4
40

mui
'o -4
oko-000N
(M t-

J4

to,
-o
0

C'l (m t-

CD Ml 'OO"C'l

.114 lq4

Cq

14 10M
00 (7)M-4
-4 ec cq (M eo
t- t- (M cq c
kO
Co M CII CII -4 -400

CII
CO

(0, t- kO ui U:)

to

-14

ro

140

-4

to

1-4 CY)m00
t- t(
(: C. ce) 10
-I -!
oo 0, (=>
00 t- to

W10N

ce) ce) C'l

oom

w
m

mko
ko"

00

P-4

N
M

-4

ko

U
-4

t-o

mmoo"C)

C)

t.

" 10 -4
ko -4w
oi oi

U"

t- t- C'l

t-

kOg

"14 ko 10 00 00 (M
t- co
-OC .
.
.
.
.
.
10 Lo --d4 "4 ""

00
m

CII
cli

C> oo "o C>

mm00NLo

Z
o

co

-14

10

ul)

C'l LO 00
t- 0
-14

Ci1 CiC O 1
t-

ko

00

io

-4
t-

10'U
06t 1 1 Co'
00

o
C'l
t-

Co ko

(O cli
co"t-o
CII CII 10 CO

oo 00-"
'! CY)moo
kO 00
(=, (: C.MCII

wcloCi

C.0

LO

ko ko to

CO ko

to

. . 00. t- .

co

U-i

ul) ul) 'O Lo ko to

C'l

0
""com
.
.
.
.
.
.
ul)
oo t- -o o

t-

(m

to, to,
CO0cl
-o t-M

00 ol
co"oo
C'l Cq co

cli
00 (MC)cli
06 .C
-4 -4 1-4
C>

1 1

ko

kt-J -4

kO
ko

00 ko W N
"
t-w0
CD00

(7)0
-4

oo

ko
-4
00

O
(Z) C)
oo (m 00 -4
t00

cli
o t-4

M0-I

oo

(M"

0tNt-

10

ko
0
C

U-j
cli

W 1-14 (=> tm m Mc'j

C) CO 00 CII
ko

(M t-

cli co
w

cl

CD U-) ko

(4

ko

m-4

ko
ko
00m
C
C)

ul
t-

t- _4 to
-4M
M
CD U-i -.J4

to,

mm00

ko

mcq

M Cj toM14
(M ko _4 00
m m mcli
ko

ui

N--14

00
m
(M 00
ko

C.0m

ko

0-o

.
10

E-4

PA

kf

Na,

LO

cli

-4

-4

U-.)

to

oo
<MCII
""CDm
.
.

-4mmoo
.(:

-4
-4

LO

10

-14 -4
moo
.
.
.

'o

-4

00

ko

m-o

ooM0.
-4

U-.)

--4

t-

'o

krj0

10

-4

-4

(:

E-4

LO

-4

'-4

kxll -114 n4m

.

14
CD C) O
_4 Cj io (M
t- to 10

(: C.MNW14
cli. cli I* co.
O
.
.
oo (m o

-4

t-

00 CII 10 0) N-4
-4MM00
00 14 -4 10N
.(:
CMCII
MW
Cl COM-4
C)
.
.
.
.
.
coomO.

mmwwNkow

-4

LO

U-)
.

Nmto

moo

-114 m

"di
.

N10

-.0N

NNoM

oo

-o

mU-.)

-'14

mNoo

C;

00
4

00

00

681

cli m
--I 1-4 -4

U' o
-4 -4

t-4

00 0)
-4 -4

0
N

0
cli m

o
"

C) c)
co cq

BIOMETRICS, DECEMBER 1960

682

-4

,=;

(:

M
t

to
'.14

0
-4

I-O

-o
m

11-14 0
'.14 1-

-4
-4

m
'o

m
cq

-4

0)

00

I-

-o

-o

to

to

-O
'.14
m

0
-4
14

I10
0

-o
M
00

'-I' 0
1.14 1I-o

-4
-o

m
-o
to

m
cq
to

0
00

t-O
0

10
m
00

".14 0
'-I' t-o
t-

-o

M
-o
to

M
N
to

0
00
-o
m
11-14 1-14 .14

t-O
o

'o
m
00

'-I'
g

0
t-o

-4
-4
-o

m
'o
to

m
N
to

0
00

-o
to

cq
to

00

-o

-o

00
N

cq

cq -o

-4

-4

00

1-

m
0

m
to

00
M
cli

00
N
cli

cl
00
-o

-4

C.

C.

00
m
cq

00
N
cq

N
00
-o

10
.I'
m

0
-4

N
00

-O
.I,

0
-4

-4
'o

N
00

00
m

0
00

-o
m

m
'.14

00
0

-4
'o

m
cq

cl

cli
00

I10
N

m
-4
N

-o

0
'.14
N

0)
0

co
'.14

00

to

-o
m

10'

10
-4
104

I0)
COD

M
00
'.14 0)
COD N

10
-4

t0)
m

'-q -o

10 t-

00

'o

-4

0
m

'-I'
M

0
N

m
N

t0
m

-4
00
N

-4
N
N

-4
-o

-o
m

00
I',

o
oo
-4

-4

00

t-

oo
m

00
N

10'

06t-, c c c 10'
00
m
cq

M
c;

-4
z
c;
oo

10
'.14
m

0
-4
-4

tto
0

N
00
-o

to
0
1.14 -4
m
-4

(:

00
m
cq

(=>
-4

06t c c c

COD 00

00
C'l
cq

CII
00
N

c'l

cli
o

to
'.14
M

10'

to

0
-4
-4

00
00

10'

0
'o

-O

10
m
00

-o

O' 101010to to

-O to,

tto
0

-o
m
oo

1.14 0
'.14 tt-o

-4
-4
-o

m
-o
ko

m
cq
LO

0
-o
00
-o
M
".14 ".14 '.14

1.14
0
M

0
N
m

,I'
-o
cli

t0

'.14

t10
0

-o
M
00

t-o

-4

m
-o

m
N

00
00

-o
10

00
N

m
0

0
00

00
0

to
M

0
-o

t10
0

-o
M
00

t-o

-4
-o

m
-o
to

cq
cq
10

to
00
'.jq

M
10
-.14

10
cq
.04

0
-o
0
tCOD COD

tcq
N
0
COD N

-4
10
-4

0
cq
10

m
oo

0
10
'.14

0
cq
'.14

to
0
M

0
tM)

10
0
N

00
-4
N

0
t'.14

10

10
-4

00
00
m

".14
-o
m

t00
N

cq
00
m

tto
m

00

00
N

co
to

00

t-

-o

-4
c'j
".14
-4

C'l
oo
0)

0
0
00

0
0
N

0
m
-4

0
to

00
-o
0)

to
oo
00

00
tN

00
0
-4

oo
-4
-

-o
m

m
to
0

oo

t-

-o

'.14

t-

-o
N

00 0
-

C'l

06t 1 1 1

"I-4 -4

00
N
cq

00
M
cq

m
c

oo

IZ2

cq
oo
co

ol
-4

00

00
cq
cq

06t-' -O'

10'

t0
N

0-04 1-114 (-=4>

cq
00
-o

to
'.14
m

0
-4
-4

t-

10'

06r c c c

-Q
z

En

-4

m
(:

m
C.

oo 4

"I
-4

(=>
-4

00
M
N

-4

M
-o

co

-z

co

-4

-4

-4

cli

00 "I (=>
4 14
-4
-4

z
o

00

(:

"I

00
N
N

cq
oo
-o

10
-.14
m

0
-4
-4

t10
0

-o
COD
oo

06t C.61 1
00
M
C'l

00
cq
C'l

cq
oo
co

10
'.14
m

0
-4
-4

-4

t10
0

-o
M
00

'.14
'.14
t-

0
t-o

-4
-4
-o

-o
M
00

'.14
t-

tco

-4
-o

00
M
cli

00
cq
cli

N
oo
-o

10
'.14
m

0
-4
-4

t10
0

00

00

cq

to

cl

00

.I' -

cq

cq

-o

0
-4

co

00

-O m

00

t-

C.6-O'-O'

00
N
cq

N
oo
-o

00
m
cq

(=>
-4

00

t C.6

-O'

-O'

00
m
cq

00
cl
cq

cl
oo
-o

-O
'-I'
m

0
-4
-4

-4

-4

00

t-

C.6Co'C.

'

oo

00 N

(:

C.

m
N

cl
N

"I
-4

(=>
-4

oo
co

00N
m m 00
m N 00
cli

-4

to
0

'-q

cli

;k

1.14 -4
COD -4

00
-o

oo

:$

0
10
to

to

-4
to

.14

t-

N
cl

"I
-4

00
-4
to

.I' tt-

-o

00

0
-o

to

-O 0
to

-o

cq

tm

to

".14
m

-4

cq m
0)

10
'-I'
M

0
-4
-4

-O 0
.I,
m

-4
-4

t-O
o)

m
00

'.1400 10 0 M N 10 M
0

1.14 -o
t-o

0
-o

to
to

-o

cq

00

10 N 0 10

m
m

10
cli

-o
m
00

M
'-I'
t-

10
'o
co

0
.I'
t-

0
'o
-o

m
0
to

t-4

00

c'l
0

m
00

10'10'10''.14,
'.14,

10'
10
0

M
-o
-O -4
1.14

M
00
m

M
-O
m

-o
cl
m

N
t'.14 10
CIA -4

N
t0

t00
0

-4
O
0

m
-4
00

10
00

cq

0
tm

0
'.14
m

m
-4
m

-o
cq
cli

0
'.14
-4

to

t-o
0

0
oo
00

c'j
ai
t-

'o

t-

.I'
0

-O
00

'o
t-

10'

t-O
o)

'o
m
00

to
m
to

10'
-O m

t-

'-I'

-O
(M

m
oo

-O
-o

-O 0
.I, -4

'o 0
-O N

tN

cq

-o

-4

00

t-

'.14

10

-o

t-

00

0
-4

-4

-O '-I' 0

'o

cl

t-

-O
m

cl
m

0
cq

0
cli

cl
-4

m
0

m
t-

00 -o
m 0

tt-

t00

00 00
0 0

0
-4

0
m

0
'.jq

cli

-4

00

t-

t-

00

cr) 0

g'

-4

-4

"

cli

-o

04

00
to

cl
to

t-

-4

to

to

cl

0
m
-O -

-O -0

"I
-4

to
oo

-4

oo 4

-4
-o
to

06t C6C6C6

oo 4

cli
cli

m
-o
10

10' 10'

06r -O'16C.6

m
cq

cli

-4
-4
10

(=>

oo
cli
-4

t-o

10'10'

-4 m M 00 00 N 10 0 t- -o
o

.04
t-

'.14 10 -o

r4

-4

-4

-4

-4

683

MULTIPLE RANGE TEST

0010

-om-o

r-I

-4

r-I

-o

-o

to

co

0)

-o

'.14mcqo'.14
-4m-oO1114
14
co
10

00

00

t-

t-

t-

t-

-o

-o

-o

to

-o

0mN

t-

N-4

m10 -4 too 0

to

t-

.Iq

cq0

-oN0

to

cli

-o

mto
t-

000

cli
-4

-4

00

00

t-

to

to

to

00

cq to

t-

to

1.14

01.14

-4

00

00.Iq

cq

oo

t-

14

to

0
Nt-

t-

-4

(=>
t-

tto

-o

-o

-o

to

1.14 COD01
'.14

m-

0
t-

to

-o

00

to

11-14t-

00

-o

00

"'I'll 11-04 N-4

-o

cq
to

00t0

cq
-o

00
t-

-o
cq

00 -o

to

m1.14

t-

00

10

to

t-

10

to

-o

--4 10
mcli

t-

to
.Iq

-o
-4

to

to

-om

to
-o

011-14
00
0

-4
00

0)

00mcli

cli

to

10
0M

10

-o

10

1.14
m

-o
cli
m

0mt-o
t-

10

00to
-4
t-

10

'.14
t-

to
-4
00N

-4
10

cli
oo-

10

ec4
00m00
C'I

00

-4

0)

00

t-

to

t-

N10

00
00mt-

'.14

00

-o

000

mmN

to
0

00

000

.Iqm
-o

t-

to

cli

tomcli

0)

-o

00

00

mNN

to

t-

-o

t -O'C6C6C6C6C C6C C6

0 06t t

(-=>4 (m,

-o

to

C'!
-o

-o

14 '.14mNN

"14014

oo
-o

t-

cq
t-

-4

t-

tomto

-o
aimt-

't!

-o

t-

to

to0
cli
.,vmm

0
'.14

to

t t

14
t-

COD

.Iq CODcq 0

co
-o

00N

t-

000LO
o00

0, oo,oo,t t

.Iq

10

t-

to

cq

-4
-0
10
-4
10M

-o

cli

-o

C'!(:R0 t'

(-=4>

-o

to

t t

t-

1-1

to

t- co to to

mto0tNo0

.04

0606t t,

'.14M

-o

00

-! C 0
00'.14
t-8mt-4

oo

0001.14
t-

cli
5>

-o

-o

cq

"14

10

-o

-o

10

-o

-o

(M

cq0cq
10
0Mt-

t-o
00m

to

00
-4

t-

-4

oo
oo

-o
-o

oo C'I 0

E-4

cq
10

t-

10

-o

1-1

too

-4

t-

000
00

t-

-o

I-

C'I000

"! '-l!

ci -! (: 0 t

co
-o

cli

P.

0)M
cq
cqmto

0cq
000

-o

to

d4m

t-

'114 t-

M0

t-0

d4

0)
cq

:114' 0-04 N-4

P.

1114 tcqmt-0

I-

to

00

M0010

ti
.4
FA

-o

-4
cq

,06

000

I-

I-

00

00

0)
cq

cli
-4

cl 0

t,

N-4

-O
t-

t-

10

-4

-4

to

o)

I-

-o

t-

-! t-:

06C (=;

00

00
00

00

0
0
1.14 cq

to
to
1-14 t-

m0
oo
r-i

0)

-4

-o

N14

14
'ION(=>

m0)

to

E-4

to

0N000Nt-

t-

1.14 t-o-o

o-114
0

m0

o
C

00

-40.I'

cli

C,

ci

to to
'.14 t-

-o

00

06C
10

t-

t-

00N

-4

CII

0N

00
-4

ooN
-4

8444
0

.
0

cli
clmNoo
0

-o

t-

t-

co

co

_I

-o

-o

-o

".14

o
o

-4

00

140

In

cli

00

cli

t-

cli

to

to

to

to0

-o

co

co

to

00

t-

cli

14 CII
00N-o
-o
t-

000

'.14MCl

-4
-4

t- cli
0mt-4
00t- to clicli
0

t-

C'!C'!-!Rc!
-o

-4
-o

10
00
".14 t-

-4

-o

cli
000Mt-

t00
-4M-o010

-o

-o
C'I
t-M0
.

00

t-

t-

1.14 toN
00
-4
to

-4

t-

to

t-Nto
tomcqmto
'.14mC'I
.
.
co

co

.
co

co

co
t10

0)0
co
1.14 co
0
'.14N-4m-O
1400
1.14mcq

ci

00

n4

-O
t-M
m

C)

N
00

t-

-O

co0-O

-4

'.14 0

t-

o0

t-

-o

-o

t-

00

t-m
00
00Nt0
.

00

0
-o
00
0
01.14
10MC'I
t-

0
0
1.140t-4

O'

00Nt00
00

t-

cli
-4
CO

'.14

00m0

t-

-o

00

d4

00

-O0'oN5
'.14M

.I,

t-

00m00

00

-O

-O

-ONp

00

-4

'.14

00

00

t-

00

-oN

to

to

to

to

00

(7)

""OOOO 8

-o

-O, t6

d4-

'o0oo

-O'o t-0

m'o 00 t-0t-! t- 't!

-o

-400
.
co

1 1.6C6C6C6C6

'o

000Nt-

't!

co

'.140

0
t-

t-

06t t t-6

.I' t-

oo cq
-4

-o

-o

t-

't!

-o

co

t00

00 03

oo

t_,I

co

co

0tom

00

R'.14

cl -1

cli

10

-4

-4

t-

cli

00

C 06t t C

-4

(7) oo 00

cli

-o

t-

-ON-Ommt-

C.9 .I!

t-

0to

10N00
10o-4

cli0
-4
tC

t-

00

t-

C'l0C'lm000

-o"t-

mmcli

m.14-o 00C'l0tci ci -! -! .

oC'l t00

t00
000Mt-

00

t-

oo,t-, t t

M10
t-M10

o
o0CII
o

-O0

0)

-4

00

cli

00

'ION(=>
M0

to

00 0

10

-o0000-4
LO t-

00

! '-I''o C'lmcl

1.14

co

cli

00

00

ci -!R0

ci -! (: 0 t-7
-o

-o

to

-4

1-1 -4

to

14

-I -4

to

to

to

10

-o

t-

-4

-4

-4

-4

-4

0 -ON0 t'.14

-q N

to

to

.14

.14

.14

684

BIOMETRICS, DECEMBER 1960

(7) to
co

ci

(7)

to

to

'-4

'-4

to

4m t-

C'l

14

to

to
oo

-4

t-

00

00

t-

14
co

cli
oo
C'D 00

o)
t

E-q
O

to
'I!

to
t'

oo

C'l

to

to

00
-4

to

-4

0
-o

0
-o
O

O
(7)

C
o
O
0

o
El

C)
co

cli

P'

co

co

10

00

CII

00

oo

00

00

t-

t-

t-

-4

".14 -4

co

LO m m LO 00 -4 -o -4 -o

t.
C'l
-o
.

-4
'o
1-4
.

N
m
oo

00

00

t-

t-

00

00
00
to

t-

C! O ll' t': t- 10

t-

t-

-o

-o

-o

-o

-o

0
-4
0
LO
".14 N
.
.

cli
m
-4

cli
(=>
t-

to
LO
-o

t-

11-14 0
m
'-I'
0
0
.
.
tco

m
LO
t-

t-

co

co

cl

cl

LO

00

00

LO

t-

t-

LO
t00
.

00

00

"Ili

10'

t-

cli

'o

-4

-o

Co'

Co'

tM

O cli m
00 N -o
0
00 'o

m
1
cli

cli
00
O

cli
N
0

m
co
t-

O
-o

0
o
0)

m
cli

0
-o
O

-4
00
m

cli

00

t-

to

00
10
m

m
O
C'l

O
00
1.14 t00

00
-4
t-

O
-o
to

co

cli

co

co

00 N co C) '-I'
-4 C) 00 t- 10

o Cr) t- 00
C'l -q C'l
-! (:R

tt-

LO 0
o

ci
t-

t-

t-

-o

to
.14

to
t-

0)
C'I

-o

N
M

oo
oo

t0

0
m

00
O

-o
0

0
O

-o
-

m
0
11-14 t-

-4
cli

0
10

-4

C'!-! C C 0 t-7

'o

to
m
cq

-o

00

00

t-

t-

t- t- t- 'o 'o 'o 'o 'o

0
N
M

N
-o

co
-4

cq
m
00

oo
00
to

0
m

t-

t-

00
Lo
.
t-

0
m
t-

oo
to

-o
00
m

ooC'lo . . . . .
-4

-4

-4

00

00

t-

to
14

to
t-

C
'.14

C
N
m

N
co

-o
-

N
m
oo

oo

C'l

00

00

t-

N
O
-

0)
00
0

N
0
00

00
O
00

. . . . . .

m
cq

0
O

oo
0

m
oo
00

0
0
t-

11-14 0
m
-o
t-o

0
o
to

00
Lo
to

cl
t-

00
to
to

t-

(m0000t- t-

t- r- r- co co co C.6C Co'Co'

10

0)

0)

C)

C'l 'O m oo
'"! t-7 -'4! C'l
m -o -4 00

t-

-o
-4

-4

m
00

C)

(m (m t-

-4

co

co

00

00
00
0

0)

m
t(M
.

t00

C'l

co
r-

t-

oo C)
Cl t-

00
00
t-

t-

-o

-o

m
t-

tt1.14 00
-o
to

-4
.

m
to

"44

co
co

r- N
10 10

C)

10

t-

t-

co

co

10

-o

0)

-o t10 co

co

0)
-4

0)
10

10
C)

10
m

a)
-o

LO
C)

co

co

co

-o

-o

10

-o

-o

-o
co

00 C) N
t-

r00

00
-4

-4

co
00

00

co

.,i,

C -! (:R(::

t- t- t- co co co co -o -o -o
a) C) m 1144m 00
m -O 00 N -o
00 t- co 10 10

N -o
00 10

-O 00 'g N
m -4 C) 0)

0
-4

a)

00

00

t-

t-

t- t- t- -o co co co co 16Co'

10

a)

a)

t-

C)

t-

cli

C)

00

10

1144 0

co

t-

C'l co
-'t! C'l -o
-4
m

"
m

co
-4

N
0

0
a)

co
-o

-o 0
10 10

oo

(m -4 -4 -4 0; 00'00'
m a) If) 10 a) a)N t-N -o-O 00o
-o

-4

oo

-O

cli

m
00

10
N

10

co
N

0
00 t-

C6

co
0

00 C)
(M m

cli

cli

10

0
g
00 -o

00

-4

r-

.I'

(M 00 co -O

-o
a)

10'
10

co

00

a)

tcli

C) I'll t-4 a) t-

-4

-4
-o

-o

t-

cli

-o -o 10
00
m
10

N
m

0
10 m 00 00 a) a) t- N
-'o 0
00 t- 00 0 m t- N
--I 0
00 t- -o -o 10 I'll 1.14
.
.
.
.
.
.
.
.
.

C)

-4

LO

cli

-'o 00 -- -'o (M m
C'l C) (M t- 10 .II

t- t- t- -o -o -o -o -o -o -o

10'

C)

ir) m
N -'o 00 N
C'l C) 00 t-

a)

,II oo

a) 0000t- t-

1.14 00 N 00 m -4 N N 0
1.14
0
N 00 -O -O -o 00 -4 -O (M
m -4 a) 00 t- -o 10 10 "Ili m
.
.
.
.
.
.
.
.
.
.

t- t- co co co co co co co -o

1-4

cl

-o t- 000) 0

-4 clico

t-:
M -o -4 t- 10
.
.
.
.
C'l C)
.

10

t-

10

0) I'll-4 -4 -4 a) 0000t- tm a) If) 10 o) a)N t-N a).I' (M

1.14 N
N

t-

00

t-: "R

cli
-4

-o

co co 'o C C6Co'

00

-o
.

t- t- t-

t-

-!

-o
-4

Co
.

t-

(m0000t- tN
-o

C c'! (:RC

t-

10 10 0) 0) t- -4
n4 t- -II Cl CII 10

m
N
N

00 t- 'O 'O tm cli C) a) 00 t.

t-

-1 0
00

161 16

10

N
co

00

00

-o

t-

C -'t! tm
C) "'il oo C'l C)

00

t-

LO

(m

o)

cli

'Ili o Cr) m
t00 ta,

N 00
CODoo

t0

t-

t-

10 10 a)
1 r
'"! N
m

m
0

00

oo
-4

00

t-

a)
C

00

m
-1

C'l

C)
a)

to

t-

0 C

'-4

N 00 M 00 M 00
to m cli 0 0 t-

to
N

to

a)

to

00

a)

LO

t-

.II N
C'iC) C

to

to

00

C tcli cli

(m

-4

-o

t-

-4

C'l

Lo
to

m
O

oo

m
1-4

t-

-o

tto

00

to

-o

-4

to

0
0

to

to

00
m

oo

oo

CO 0)
10

co

t-

"14 "14 00 tt- M 00 M

ci -! C 0

00

C)

0
t-

t-

(M -I -4 -4 (m0000t- t-

co

t-

to

-4

'o

co

-4

-4
C'l

t-

t-

C'l
m

-o

-4
00

t-:

-o

-4

a)

oo

-o

to

M
z

t-

(711 11-14 -4
m
m
LO

'.14

(M

t-

1-4
LO

Cl!

t-

co

t-

o
0

t-

t"

t-

-o
O

t-

-4
00

m
cli

"! ci -! (:R (: 0

00

(=! (7

C'!

06 C O
00

11-14-4 -4 -4 0
m LO 00 cli tO O t"

-4 N
to m
Cl -1

cli cli -o m 00

O
ti

-! 0
00

'6 '4C
(=;

t-

'4

t-

00
00

tcl

oo C'l
'4

N
m

10co t- 000) 0N

10 (M
'O 0
10 .II

C6-O-10'lo'10'

'-I'c'
0 c' c'
CYD114 o
cli 8

685

MULTIPLE RANGE TEST

by 1/(,AQ/AP) = AP/AQ. Since u is defined as the number of digits

before the decimal point in the studentized range interval under consideration, this would guarantee that the error in Q would not exceed
5 units in the fifth significant digit if the ratio of the change in P to
the change in Q were constant throughout the interval under consideration. This condition (P piecewise linear in Q) is obviously not satisfied
in practice, but as long as the weaker condition
max [AP0/AQ0Q
, AP,/AQ,] < 2 AP/?\Q,

i y - yi and AQi = Ix where AP

I (i = 0, 1) is satisfied, the
error in Q will not exceed a unit in the fourth significant digit. This
weaker condition is in fact satisfied, and hence it can be stated that
the error in the critical values for Duncan's test, which are given in
Table 1, does not exceed a unit in the fourth and last significant digit.
REFERENCES
[1] Beyer, William H. [19531. Certain PercentagePoints of the Distribution of the
StudentizedRange of Large Samples. Virginia Polytechnic Institute Technical
Report No. 4.
[2] Duncan, David B. [1955]. Multiple range and multiple F tests. Biometrics
li, 1-42.
[3] Harter, H. Leon. [1957]. Error rates and sample sizes for range tests in multiple
comparisons. Biometrics13, 511-36.
[4] Harter, H. Leon. [1957]. Critical values for Duncan's new multiple range test
(abstract), Jour. Amer. Stat. Assoc. 52, 372.
[5] Harter, H. Leon and Clemm, Donald S. [1959]. The ProbabilityIntegrals of the
Range and of the StudentizedRange-Probability Integral, PercentagePoints, and
Moments of the Range. Wright Air Development Center Technical Report
58-484, Vol. I.
[6] Harter, H. Leon, Clemm, Donald S., and Guthrie, Eugene H. [1959). The
ProbabilityIntegralsof the Range and of the StudentizedRange-Probability Integral and PercentagePoints of the StudentizedRange; Critical Values for Duncan's
New Multiple Range Test. Wright Air Development Center, Technical Report
58-484, Vol. II
[7] Pearson, E. S. and Hartley, H. 0. [1942]. The probability integral of the range
in samples of n observations from a normal population. Biometrika32, 301-10.
[8] Pearson, E. S. and Hartley, H. 0. [1943]. Tables of the probability integral of
the 'studentised' range. Biometrika 33,89-99.