118 Selection Does Not Lead to Origin of Species. 



WIN IS thereby sliown to be only apparent. In reality 

 it is only an isolation of previously constant types. From 

 horticultural practice it only differs by the fact that the 

 races to be isolated are old constituents of long cultivated 

 mixtures whereas in horticulture mutations are usually 

 isolated as soon as they appear. 



§ 13. THE LIMITS TO THE AMOUNT OF CHANGE THAT 

 CAN BE EFFECTED BY SELECTION. 



Selection does not lead to the origin of specific char- 

 acters.^ If I have succeeded in showing that this general- 

 ization is in accord with the facts of plant-breeding, the 

 chief support of the doctrine of selection has been under- 

 mined. I propose therefore to summarize the most im- 

 portant parts of my argument in a series of short para- 

 graphs. 



1. Linear Variation. Statistical methods of dealing 

 with variability are so generally employed at the present 

 time that an acquaintance with their principles may be 

 taken for granted. The chief thing however that we 

 learn from these curves is that the characters vary only 

 in two directions: plus and minus. The old-fashioned 

 vague talk about variation of single characters in all 

 directions has died a natural death. All variations in 

 mass, and weight, and number (meristic variations) con- 

 form to this law. ^ 



The character can be diminished or increased, but 

 nothing new can arise in this way. The differentiation 

 of organisms in the main lines of descent consists in the 

 development of new characters ; and the materials for 



^ In men abnormal characters fluctuate: but they soon disappear 

 and no new monstrous variety arises. See Kollmann in Correspon- 

 densblatt d. deutsch .Ges. f. Anthropologic, 1900, No. i, p. 3. 



