80 CONTROLLED NATURAL SELECTION 
one sex; this conception is easier. Such a 
variation would be correlated with the sexual 
determinant ; variations not so correlated or 
bound to a sexual determinant cannot, one 
would conceive, be made to do so by Natural 
Selection: for instance, favour their fixation 
in the male, and their eradication in the 
female. 
Sexual Selection, if it explains the origin 
of extra sexual characters, does not explain 
their use, yet, as has just been shown, they 
must be useful, or otherwise Natural Selec- 
tion would remove them. 
Beside this argument, there are many facts 
that sexual selection fails to correlate and 
which it ought to correlate. They mostly 
come under the following headings : 
1. Sex selection accounts for the brilliant 
colours of males, but fails to account for 
equally brilliant colours which are frequently 
found in both sexes. 
2. Sexual selection, if it has been observed, 
has not been observed to be especially asso- 
ciated with animals which show the greatest 
sexual differences in characters. Birds, for 
instance, in which the sexes are similar, do 
not show less selection by the female of 
the male, than birds whose males are of 
