CLASSIFICATION AND ADAPTATION 17 



attributed variation partly to the union of gametes 

 containing various determinants, which he termed 

 amphimixis : this, however, would introduce nothing 

 new. Then he proposed his theory of germinal 

 selection, determinants growing and multiplying 

 in competition, some perhaps disappearing alto- 

 gether, though this does not satisfactorily account 

 for entirely new characters. With Weismann, 

 however, every species was a different adaptation, 

 and natural selection was the deus ex machina ; 

 to quote his own words, Alles ist angepasst, 



Romanes and other writers, on the other hand, had 

 always maintained that in many cases the constant 

 peculiarities of closely allied species had no known 

 utility whatever, so that the problem presented 

 by these characters was not explained by any 

 theory of the origin of adaptations. 



Mendelism, since 1900, has studied experimentally 

 the transmission of definite characters, and main- 

 tains that the characters of species are of the same 

 nature as the characters which segregate in Mendelian 

 experiments. Such characters are not in any way 

 related to external conditions, and cannot, therefore, 

 be adaptive except by accident. Professor Bateson 

 goes so far as to admit that such large variations 

 or mutations offer more definite material to selection 

 than minute variations too small to make any impor- 

 tant difference in survival, but as regards species the 

 important factor is the occurrence of mutations 

 which are inherited and at once form a distinct 

 definite difference between allied species which is 

 not due to selection and has nothing to do with 

 adaptation. 



In a book entitled Problems of Genetics ^ 1913, 



