88 DARWINISM TO-DAY. 



two categories, the category, namely, of cultivated plant 

 races which are reproduced practically exclusively by 

 division (i. e., by cuttings, roots, scions, buds, etc.). The 

 reason for relying upon this kind of reproduction is, in 

 nearly every case, that these races do not breed true to seed, 

 i. e., the races are not fixed, are unstable. And even among 

 those races which we are accustomed to allow to reproduce 

 by seed how necessary it is to maintain the unusual environ- 

 ment, the exaggerated excellence of conditions of food 

 supply, humidity, protection from natural enemies, etc., if we 

 are to be successful in maintaining the parental characters of 

 the plant. Let a few individuals escape from the hothouse 

 or fertilised and sprinkled garden and see how soon, if they 

 can persist at all, they lose their characters of amelioration, 



* 



and become most pitifully unadorned. 



In Pfeffer's " eyes the fundamental difference between 

 -the two selection processes rests on the fact that the breeder 

 or plant ameliorator selects his individuals (the "to be 

 saved") on the basis of the character or condition of single 

 characteristics, while in nature survival is not determined 

 by such conditions, but on a basis of total or all-around fitness 

 or advantage. 'The moment," says Pfeffer, "that one be- 

 lieves one's self to be able to place in parallel, simply and 

 directly and in general, the activity of the breeder and the 

 activity of the struggle for existence, and from this false 

 generalisation deductively to compare the selective work of 

 the breeder based on definitive special characters with the 

 automatic selective work of nature based on similar specific 

 characteristics, that moment one enters the camp of the 

 teleologists, whether he is doing it knowingly and will- 

 ingly, or not. In short it is a logical fallacy when one as- 

 sumes to substitute for the selective action of the breeder a 

 mechanically-working natural selection. Only in a single 

 kind of case has this position any justification, and this not 

 on account of logical correctness but on account of the pecu- 



