Alice Lee 
209 
form a truncated portion of a normal curve, d, n and 2 can be found (see Tables, 
pp. xxvii and 25). 
We have = ^V^^^ (i), 
■\^i = ajd (ii), 
ylr3 = N/n (iii)- 
These are tabled for each value of h', at first proceeding by 01 and then by 10 as 
unit. Now ^frl being known we find h' from the table, and hence deduce -v/r, and 
"^3- '<|^2 gives us the value of tr from known d. Hence h = h' xa can be found, 
lastly (iii) gives us the total population from which 7i is drawn. Thus the constants 
iV, <7 and h which fix the total Gaussian are determined. 
It will be sufficient to illustrate the method of using the tables on certain data 
as to the English thigh-bone, recently published by Parsons*. 
Dwight"f- has adopted a method of sexing human femora on the basis of a 
markedly bimodal distribution obtained by him for American bones. He terms 
female any femur with diameter of head less than 4-5 mm., and male any femur 
with diameter of head over 47 mm. Parsons follows this rule and sexes by other 
points femora with heads from 45 to 47. As unsettled remainder he has 20 femora 
of 45 mm, and he gives 12 to $ and 8 to {/" ; of 46 mm. and 47 mm. he has 41 
femora and he gives 4 to $ and 37 to ^. As a result of this process he obtains 
a female frequency curve which rises very abruptly at high values of the diameter, 
and a male frequency curve which rises very abruptly for low values of the femora. 
But, if there really be any marked skewness in frequency of the parts of the 
human skeleton, which is very unusual, we should anticipate that it would be of 
the same sense. Parsons' distributions are as follows {loc. cit. p. 256) : 
36 
37 
38 
39 
40 
41 
42 
43 
44 
¥ 
46 
47 
48 
49 
50 
51 
53 
53 
54 
55 
? 
1 
1 
3 
8 
14 
12 
18 
12 
12 
3 
1 
8 
8 
29 
17 
3l" 
19 
13 
10 
6 
8 
2 
The $ 48 mm. femur according to the rule should have been treated as a 
male but presumably it had marked female characters. Were there no marked 
male characters in an}' bone below 45 mm. ? It will be seen that there is a 
remarkable dip in the total material at 46 mm. which corresponds to Dwight's 
division. In material measured six years ago in the Biumetric Laboratoiy, where 
every bone in a relatively large series was measured, no such dip occurs and there 
is in those data no justification for Dwight's method of sexing|. The group of 
29 cT bones at 47 mm. and the sudden cut off at 45 mm. seems to condemn this 
method of sexing, at t'jy rate from the statistical standpoint. 
* Journal of Anatomy and Physiology, Vol. xlviii. pp. 238 — 267. 
t American Journal of Anatomy, Vol. iv. p. 19. 
% This material lias been statistically reduced and will shortly be published. 
