BREWSTER : VARIATION AND SEXUAL SELECTION IN MAN. 49 
computed from 8,551 measurements of 406 individuals belonging 
to 11 different races. In Table B are four sets of coefficients, cor¬ 
responding to four groups of individuals of the same race ; and it is 
noticeable how nearly alike are these coefficients for the two sexes, 
and for individuals from two different localities. This indicates 
that the coefficient of variability for a particular dimension in a 
particular race is a rather definite quantity. This datum is taken at 
random, the choice being determined, for the most part, by the 
difficulty in finding recorded measurements of both sexes. Every 
coefficient computed appears in the final result, including even 
those computed to find how small a number of individuals could be 
taken before the method would break down so that if it should 
turn out that variability is not correlated with conspicuousness, the 
mistaken conclusions of this paper would be due rather to falla¬ 
cious reasoning than to insufficient evidence or to data improperly 
selected. 
In discussions of variations it has commonly been taken for 
granted that differences in variability are caused by the greater or 
less severity of the struggle for existence. But if variability is in 
any way correlated with conspicuousness, this correlation can 
hardly be the result of natural selection, since it is difficult to ima¬ 
gine any connection between conspicuousness and utility. Sexual 
selection will, however, account admirably for this correlation; and 
the significance of this discussion is this suggestion of the imj)or- 
tance of sexual selection in man. 
It is not clear just how sexual selection operates to increase vari¬ 
ability instead of diminishing it as natural selection is commonly sup¬ 
posed to do. It is possible that sexual selection is pushing men and 
women in the direction of an aesthetic ideal, and in consequence 
conspicuous parts of the body cannot settle down to a constant con¬ 
dition. Perhaps variety itself is attractive, so that the individuality 
which comes from looking unlike one’s fellows is an advantage. 
To sum up then, I have tried to show that sexual selection has 
brought it about that parts of the body tend to be more variable in 
proportion as they are of greater aesthetic value. There is, there¬ 
fore, this much additional evidence in favor of Darwin’s well-known 
views of the importance of sexual selection in human evolution. 
The data here brought together have a certain bearing on some 
previous studies in variation. Prof. Karl Pearson ('97) has col- 
