Permanence and Evolution. 21 



appears that with careful selection fixed races 

 of uniform colour may be formed. What is this 

 but to say that colour in the horse is not fleeting 

 but permanent, and that if it appears otherwise, 

 the cause lies in the absence of selection, i.e. y in 

 intermixture. We must take care not to be 

 deceived by words ; there are in nature per- 

 manent races so much alike that they intermix 

 naturally when found on the same area ; for the 

 breeding together of these we have no name, 

 breeding them apart we call applying selection. 

 There are other races somewhat more distinct 

 which only breed together when secluded from 

 their own immediate kind ; for the separate 

 breeding of these we have no name, but their 

 commixed breeding we call crossing. Another 

 fallacy of language consists in giving an inde- 

 pendent existence to abstractions, as in saying, 

 "the horse is very variable, the ass less so," 

 which means the type we call horse exists con- 

 joined with much more diversity than the type 



