A. O. POWYS 
39 
a confirmation of the result several times noted that the frequency of statures 
in a homogeneous group gives very nearly a normal distribution. 
With the females the case is somewhat different ; the observed distributions 
are much more irregular. This is a very common experience in statistics for this 
sex. It is probably to some extent due to greater difficulty in ascertaining the 
real age of women, and again to a very likely differentiation in stature between 
women who have borne and those who have not borne children. But admitting 
the irregularity, we see again that the mode and the mean are in sensible agree- 
ment. The distributions are closely normal. In all cases but one (ages 50 and 
under 60) the curves diverge from normality in the direction of the curve of 
Type IV. But in this one case of divergence, the curve, which is of Type II., is 
within the limits of the probable error of random selection a normal curve. It 
would seem reasonable to conclude that the stature distribution of these homo- 
geneous groups is nearly normal, but that what divergence there is lies in the 
direction of curves of Type IV. 
ISO 
I70 
160 
I 50 
I40 
ISO 
130 
no 
100 
so 
1 
1000 
Vi \', 
1 
/ \ 
1 . 
1 
1 
1 
i— 
1 
perl 
— r 
( 
a 
f ' 
-if 
II 
■ — 
1 
Centroid] 
-ir 
e 
CerUroid 
r 
r 
mi 
■ If 
j If 
Frequency 1 
— t 
1 
-t~- 
* \ i 
1 
■ / 
■ / 
- . 
iff 
ii 
\| 
\ 
-r 
1 
1 
1 
Hr- 
I 
-^ 
1 
t 
n !^ 
.< 1 
k. 
s 
53 54 55 56 57 58 59 80 61 62 63 0* 65 66 67 68 69 70 71 72 73 74 75 76 77 
Stat u r e i n i nc he s. 
Fig. 5. Statures per 1000, Male and Female, Ages 20 — 25. Male Curve, firm lines: 
1G6-7100 
y= 
1 + 
(99: 
9138 
Mean at 66" -95- 
Mode at 66"-99. 
Unit of 1 = 1". 
Female Curve, broken lines: 
y = 
96-9970 
L / X V-'l 43.S885 
\ ^ V21-5568 ) ] 
Origin at 67"-33. 
Modal Frequency =168-6. 
Mean at 62" -60. 
Mode at 62" -66. 
Unit of x = l". 
Origin at 65"-ll. 
Modal Frequency =170-07. 
