No. 2.— Variation and Sexual Selection in Man. 
By Edwin Tenney Brewster, Andoyer, Mass. 
The proverbial ‘inch on a man’s nose’ suggests the rather 
obvious fact that certain parts of the human body determine the 
personal appearance far more than do other parts. These portions, 
which may be no larger or more useful than others, I shall (for 
want of a better term) call conspicuous. This paper attempts to 
show that there is a relation between the conspicuousness of any 
part of the body, and its variability as measured by the coefficient 
of variability of its dimensions. ( Cf. Pearson, ’96, p. 265-277; 
Brewster, ’97, p. 269-273. I here follow Professor Pearson in 
multiplying all coefficients by 100.) The data of the investigation 
are given in such of the appended tables as are designated by 
letters, while comparisons and synopses are given in the numbered 
tables. 
I shall first offer evidence to prove that conspicuous dimensions 
tend to be more variable than other dimensions. Table A, the first 
to be considered, gives the 4 coefficient of variability ’ of ten bone 
measurements for New England Indians of both sexes. From this 
it appears that the several dimensions of the head and face have 
coefficients of variability which may be as low as 2.5 or as high as 5. 
In other words, some dimensions, when measured by this method, 
are twice as variable as others. Now it is evident that a person’s 
appearance is determined by the dimensions of the face rather than 
by those of the head; that it is, as a whole, the face rather than the 
head, which is noticed and remembered; which is, in short, con¬ 
spicuous. This, of course, does not mean that every feature of 
the face is more conspicuous than any feature of the head. The 
width of the jaw, for example, since it is partly masked by the 
cheeks, affects the appearance decidedly less than the size and shape 
of the forehead. But the head, as a whole, is less conspicuous than 
the face as a whole. If, then, conspicuousness is correlated with 
variabilitv, the dimensions of the face should be more variable than 
those of the head, and the mean of the six coefficients of the face 
dimensions should be greater than the mean of the four coefficients 
of the head dimensions. A glance at Table A shows that this rela- 
