SPECIES AND VARIETIES. 199 



Varieties can be given-off by a relatively pure species. In 

 habitually self-fertilized plants every geno-typically different 

 individual, is effectively isolated and protected from the swamp- 

 ing effect of random crossing. After every cross, such as 

 occasionally even take place in barley or rice, a multitude of 

 plants of divers genotype are produced, each one a potential 

 species, each one rapidly becoming pure for its own genotype. 



Natural selection must, if anywhere, affect the final quali- 

 ties of a population of autogamous plants by a selection be- 

 tween the divers pure lines. It must be remembered, that a 

 mixture of very many different equally well-fitted pure lines 

 will gradually become poorer by the dropping out of those 

 lines, which in a given generation do not happen to be included 

 in the number of germinating seeds. 



What should be the stand-point of the systematists ? We have 

 seen that they believe in the stability of species and that their 

 whole procedure depends upon this relative stability. The alien- 

 ation between systematists and geneticians has been mainly 

 brought about by the acceptance by the latter of Darwin's 

 hypothesis of the changes which species were thought to un- 

 dergo by selection. 



It is the very great merit of Wagner to have pointed out, 

 that species are not changed, that they cannot be changed. I 

 am firmly convinced that on this point Wagner has proved to 

 be in the right and Darwin in the wrong. Darwin's observa- 

 tions on this point were made on cultivated plants and animals. 

 These conclusions cannot be generalized, for in these organ- 

 isms we find that one of the main factors in their change is an 

 isolation, stricter than any found in nature. If species of culti- 

 vated plants and animals under domestication change by se- 

 lection, as we have seen the Airedale terrier and the Collie 

 changed, this change can easily be seen to be effected by a 

 rigorous isolation, the breeders using only very few animals in 

 every generation to continue the breed. It may be imagined 

 that in exceptional circumstances an analogous process caused 

 by an analogous chance combination of circumstances can oc- 



