Chap. XVIII.] ORNAMENTAL COLORS. 281 



result, without the aid of selection, of one or more varia- 

 tions having occurred, which from the first were sexually 

 limited in their transmission. ^Nevertheless, it can hardly 

 be admitted that the diversified, vivid, and contrasted col- 

 ors of certain quadrupeds, for instance, of the above-men- 

 tioned monkeys and antelopes, can thus be accounted for. 

 We should bear in mind that these colors do not appear 

 in the male at birth, as in the case of most ordinary va- 

 riations, but only at or near maturity ; and that, unlike 

 ordinary variations, if the male be emasculated, they 

 never appear or subsequently disappear. It is on the 

 whole a much more probable conclusion that the strong- 

 ly-marked colors and other ornamental characters of 

 male quadrupeds are beneficial to them in their rivalry 

 with other males, and have consequently been acquired 

 through sexual selection. The probability of this view is 

 strengthened by the differences in color between the sexes 

 occurring almost exclusively, as may be observed by 

 going through the previous details, in those groups and 

 sub-groups of mammals which present other and distinct 

 secondary sexual characters ; these being likewise due to 

 the action of sexual selection. 



Quadrupeds manifestly take notice of color. Sir S. 

 Baker repeatedly observed that the African elephant and 

 rhinoceros attacked with special fury white or gray horses. 

 I have elsewhere shown 3a that half-wild horses apparently 

 prefer pairing with those of the same color, and that herds 

 of fallow-deer of a different color, though living together, 

 have long kept distinct. It is a more significant fact that 

 a female zebra would not admit the addresses of a male 

 ass until he was painted so as to resemble a zebra, and 

 then, as John Hunter remarks, " she received him very 

 readily. In this curious fact, we have instinct excited by 



82 'The Variation of Animals and Plants under Domestication,' 1868, 

 toI. Si. pp. 102, 103. 



