474 Journal of Comparative Neurology and Psychology. 



mean age and stature, we shall get rid of only about 

 29 per cent, of the usual sex difference in brain-weight. 

 Similarly, if the male and female groups have the same 

 mean stature and body weight only about 28 per cent, 

 of the usual sex difference in brain-weight will be removed. 

 Finally, if the male and female groups have the same mean 

 skull length and skull breadth about 47 per cent, of the 

 sex difference in brain-weight will be removed. It will be un- 

 derstood that the effect of these selections is not cumulative in 

 the proportions given by the percentages. In other words, a 

 selection of males and females to the same type with respect to 

 characters other than brain-weight (excepting, of course, other 

 characters of the brain itself) will reduce the usual sex differ- 

 ence in this character by probably less than 50 per cent. So 

 then we must conclude that the adult female human brain 

 weighs less than that of the adult male, irrespective of differ- 

 ences in other bodily characters. 



The sexes are equally variable in brain-weight. This is 

 shown by Table V. which gives the differences between the 

 male and female coefficients of variation with the probable 

 errors of the differences. The value tabulated is the male min- 

 us the female constant in each case. 



TABLE V. 



Relative Variability of the Sexes. 



Male — Female. 



In three cases out of the eight the male has a larger coeffi- 

 cient of variation than the female, but in no instance can the 

 difference between male and female coefificients be considered 

 significant when compared with the probable error. Therefore 

 we are forced to the conclusion stated in the preceding para- 

 graph. 



