3 i4 THE VARIATION OF ANIMALS IN NATURE 



briefly considered the joint evolution of conspicuous colours 

 and unpalatability, and conclude that the difficulties of such 

 a process have not been sufficiently considered. 



In less well-studied cases, which we consider next (pp. 265- 

 271), the same sort of difficulties arise, but there is much less 

 positive information. The features which are presumed to be 

 adaptive are found only in some members of the community 

 living in a given habitat ; the ' adapted ' species are often not 

 proved to be confined to that habitat, and there is little evi- 

 dence that Natural Selection is the only possible agency which 

 could have produced the results. 



With the available evidence, however, it is scarcely possible 

 to estimate the importance of selection. The negative evi- 

 dence in the second half of this section must also be given due 

 weight. 



The body of facts set out in our section dealing with the 

 mimicry theory forms the best documented argument bearing 

 on the selectionist view of the evolution of animal colour. 

 When all the evidence is considered, it is difficult to resist the 

 conclusion that selection has played some part in the evolution 

 of mimetic resemblances. As we have pointed out, the possibility 

 of the parallel evolution of similar colour-patterns in different 

 species has been little investigated. The first step in a mimetic 

 resemblance is always the most difficult one to account for, 

 and possibly parallel variation in different genera may help to 

 bridge this gap, for there is some evidence suggesting that if 

 birds do discriminate between colour-patterns it is chiefly 

 between those that are rather sharply distinct from one another. 

 We do not believe that there is as yet sufficient evidence to 

 affirm that selection by predators, especially birds, is very 

 highly discriminative. 



In the second half of this section we consider indirect 

 evidence against the Natural Selection theory. A survey of 

 the characters which differentiate species (and to a less extent 

 genera) reveals that in the vast majority of cases the specific 

 characters have no known adaptive significance. A few 

 special cases where such a significance has been suggested 

 are considered in detail (pp. 283-290). Most of these examples 

 still require confirmation. As we have frequently insisted, 

 without some sort of direct evidence for selection such examples 

 prove very little. It may be conceded that in a number of 



