THE GROWTH OF GROUPS 89 



characters appear in two individuals is not due to accident 

 but to community of descent. This can hardly be disputed. 



(3) The four characters are similar to the marks of 

 species. Two of them have been used as marks of identi- 

 fication, in certain cases. 



From the view point of the Mutation Theory we regard 

 an animal as a combination of characters, and the species 

 question may be stated thus how is it that a number of 

 animals have the same combination of characters ? This 

 aspect of the problem is well illustrated by the following 

 words 



" Every taxonomic description testifies to the fact 

 that a certain set of characteristics is usually found 

 associated in each species or variety. The prevailing 

 theory has been that this association is a necessary one, 

 maintained because all the characters are necessary to 

 the success of the species in its relation to the external 

 environment, or else that they were physiologically 

 dependent. Modern work in hybridizing is establishing 

 the fact that few of the specific characters are inter- 

 dependent. Their association is, so far as interaction 

 goes, mostly accidental. Thus, in my experiments with 

 poultry, I have reached the same conclusions as have 

 been reached by Johannsen, De Vries, and indeed by all 

 recent workers. I find, namely, that of the scores of 

 evident external characteristics of poultry that are in- 

 herited in alternative fashion scarcely two can be found 

 that are always associated. The most striking exception 

 is the association of high nostril and absence of single 

 comb. (C.B. Davenport, "The Inheritance of Poultry.") 



Many other examples of correlation are known. To 

 mention one recently described, Mr. H. M. Leake working 



