Alice Lee 
213 
and ultimately the two components 
Male 
Female 
Mean 
Population 
Standard Deviation 
Max. Ordinate . 
133-25 
49-83 
43-72 
145-75 
2-231 
2-662 
23-83 
21-84 
While the means agree roughly with those obtained by Parsons' sexing (49 and 
43), we see that this analysis much more nearly equalises the number of male and 
female bones, and indeed makes the female population rather larger than the male, 
while Parsons bas 79 more males. The "truncated tail" method would probably 
give results in better accordance with the present had we not truncated at the 
quite arbitrary D wight-Parsons' divisions. 
These examples may suffice to illustrate the application of the Tables to 
anthropometric measurements on man, where we can feel fairly confident that 
the material, if sufficient in quantity, would be adequately described by a Gaussian 
or normal distribution. Such cases may arise when material for the two sexes, 
or for two races, is commingled and we can be fairly certain that one or other or 
both '■ tails " of the material present homogeneous parts of the mixture. 
Another illustration drawn from Galton's data for American trotters will be 
found in the Tables for Statisticians, p. xxvi. The chief weakness of the method, 
besides the assumption of the Gaussian, often quite legitimate, is the absence as 
yet of the values of the probable errors, which values must be very considerable 
for slender material such as that used above. 
See following page for Table of Gaussian " Tail " Functions. 
Biometrika x 
28 
