88 Selection Does Xot Lead to Origin of Species. 



of the second part of our question : what may be expected 

 in the future as the result of continued selection. 



It is generally assumed that individual variability is 

 unlimited, moreover that as a result of continued selec- 

 tion in any direction variation continues to extend in the 

 same direction. But this assumption is based on no em- 

 pirical foundation whatsoever. The experiments of 

 breeders, particularly those on acclimatization tell in 

 favor of strictly limited (although agriculturally highly 

 important) changes only.-^ The one thing about which 

 there is absolutely no doubt whatsoever in this question, 

 is the fact of regression : the definite backsliding of the 

 mean of the race as compared with the extreme indi- 

 viduals chosen as seed-bearers. 



By improved methods the selection process can be 

 considerably accelerated, and the goal reached some years 

 earlier. In practice important results have been obtained 

 in this way, but tliey are not of such a kind as to be any 

 help in this discussion. 



I have already had occasion to remark once or twice 

 that our knowledge of regression is very scanty, con- 

 sidering the great importance of the phenomenon : and 

 that much remains to be found out about it. If it should 

 turn out that as a result of continued selection the yearly 

 reversion became gradually smaller we should have a re- 

 sult of extraordinary importance for the theory of selec- 

 tion. If the contrary happened, this theory would have 

 to be definitely abandoned. And so long as a decision 

 is wanting, the theory evidently lacks the necessary foun- 

 dation. 



The data that I have been able to collect tell in favor 

 of the view that the maximum change which can be 



* Cf. the following sections. 



