NATURAL SELECTION SINCE DARWIN 91 



it in the Contemporary Review. According to Spen- 

 cer, there is no such analogy except within very nar- 

 row limits, and natural selection is, in the majority of 

 cases, unable to do what artificial selection does. 

 For instance, a breeder can select one character to the 

 exclusion of all the others and thus modify the spe- 

 cies as far as this character alone is concerned. Na- 

 ture can not exert such a choice. If an individual 

 possesses a character useful from a certain point of 

 view, another individual may possess a character use- 

 ful from another point of view. Only a dominant 

 character can attain a high development in the natural 

 state; and this is not what we mean when we speak 

 of slight individual variations. The only thing nat-i 

 ural selection can do is to maintain all faculties at a 

 certain level of development, destroying all the indi- 

 viduals which fail to reach that level. 



In later years other naturalists, Morgan, Plate, and 

 De Vries, carefully studied the analogies and the dif- 

 ferences between these two modes of selection and 

 came to the conclusion that they had fewer features f 

 in common than Darwin had^wginally supposed. To 

 them the main difference lies in the fact that races 

 and varieties created by artificial selection are un- 

 stable and revert to the original type as soon as they I 

 are left to their own devices, while the new forms \ 

 produced by natural selection remain constant as long \ 

 as the conditions of their existence are unchanged. 



