i84 ANIMAL ECOLOGY 



tapir, the peccary, the jaguar, and the puma, than it plays in 

 Africa in the lives of such animals as the zebra, the sable 

 antelope, the wildebeeste, the lion, and the hunting dog." 

 Chapman ^oe gives evidence in favour of the same views. 



It can always be argued about any of these animals that 

 even if the colours are not directly adaptive they may be 

 correlated in development with some character (perhaps 

 physiological) which is adaptive. But such arguments cannot 

 apply to species which are dimorphic, like the arctic fox or 

 the white-eared cob. Similar colour dimorphism is found 

 also in the Tibetan wolf,^^ the African lion,i08 and the 

 American grey squirrel,^'^ but is comparatively uncommon in 

 mammals. In birds, however, it is often found. In the 

 Galapagos Islands there is a hawk {Buteo galapagoensis) 

 which has two phases (independent of age or sex), one of 

 which is dark, while the other is pale buff,^^^ and a species 

 of gannet {Sula piscatrix wehsteri) on the same islands which 

 has two phases, brown and white.^^^ Many more examples 

 could be given ; a good deal of the very considerable evidence 

 about birds has been summed up by Stresemann.^i^ Exactly 

 comparable colour dimorphism occurs in certain American 

 dragonflies on the genus Mshna.^^ 



7. There is another important Hne of evidence on the subject 

 of adaptation which has recently been investigated very care- 

 fully by Richards and Robson and reviewed in a paper .29 The 

 gist of their conclusions is that very closely allied species 

 practically never differ in characters which can by any stretch 

 of the imagination be called adaptive. If natural selection 

 exercises any important influence upon the divergence of 

 species, w^e should expect to find that the characters separating 

 species would in many cases be of obvious survival value. But 

 the odd thing is that although the characters which distinguish 

 genera or distantly allied species from one another are often 

 obviously adaptive, those separating closely allied species are 

 nearly always quite trivial and apparently meaningless. These 

 two authors say, after reviewing the whole subject : *' It thus 

 seems that the direct utility of specific characters has rarely 

 been proved and is at any rate unlikely to be common. Further- 



