36 CLASSIFICATION AND ADAPTATION 



this end. But how do we know that feathers in 

 their origin were connected with flight ? It seems 

 equally probable that feathers arose as a mutation 

 in place of scales in a reptile, and the feathers were 

 then adapted for flight. Nothing shows the dis- 

 tinction better than convergent adaptation. Owls 

 resemble birds of prey in bill and claw and mode 

 of life, yet they are related to insect-eating swifts 

 and goat-suckers and not to eagles and hawks. 

 Swifts and swallows are similar in adaptive characters, 

 but not in those which show relationship. It may 

 be said that the characters believed to show true 

 afiinities were originally adaptive, but we do not 

 know this. Similarly, in reptiles the Chelonia are 

 distinguished by the most extraordinary union of 

 skin-bones and internal skeleton enclosing the body in 

 rigid armour : it may be said that the function of this 

 is protection, that it is adaptation, and can be ex- 

 plained by natural selection, but the adaptation in 

 this case is so indefinite that it is diflicult to be 

 convinced of it. 



Systematists have always distinguished between 

 adaptive characters and those of taxonomic value — 

 those which show the true affinities — and they are 

 perfectly right : also they have always distrusted 

 and held aloof from theories of evolution which 

 profess to explain all characters by one universal 

 formula. In my opinion, those who, like Weismann, 

 consider all taxonomic characters adaptive, are 

 equally mistaken with Bateson and his followers, 

 who regard all characters as mutational. No system 

 of evolution can be satisfactory unless it recognises 

 that these two kinds of characters are distinct 

 and quite different in their nature. But it may be 



