BREEDERS' EVIDENCE 217 



founded upon the contrary belief, are delusive and expensive 

 mistakes." 



We have given this argument at some length, since it deals 

 with a subject of great practical importance, and since it is pre- 

 sented to us with the double authority of Cope and Brewer. It 

 is, however, on every count most disappointingly inconclusive. 

 If the size is a function of four variables, — (a) the inheritable 

 constitution of the stock (statistically determinable in certain 

 of its expressions at the beginning of a period of observation) ; 

 (b) the individual modifications produced by altered nutrition 

 (approximately determinable by control experiments and ob- 

 servations) ; (c) the possible occurrence of modification-inherit- 

 ance ; and (d) the amount of discriminate selection within a 

 given period (also admitting of mere or less precise statement), — 

 then the only feasible way of reaching a conclusion as to the 

 importance of any one factor — say the third in this case — is to 

 eliminate the others one by one. 



As to the Alderney cows, it is admitted by all that the skilful 

 breeder can breed small or breed large, either by relying wholly 

 on the selection of a sufficiently variable stock, or by assisting 

 selection by modification kept up for each generation ; but this 

 does not touch the question at issue. 



And if it be a fact that large-sized races always come from 

 regions of abundant nutrition, and vice versa, it is plainly 

 consistent both with natural and artificial selection. 



As to the argument that unless modification-inheritance be 

 a fact the practice of breeders is an expensive mistake, one is 

 tempted to retort that the latter is at least as likely as the former ; 

 but the sufficient answer is that breeders, even though they may 

 think they do, never put their stake on the doubtful card. 



Finally, it may be noted, though this is a point rather for the 

 biologist than for the breeder, that experiments on increased 

 size of parts are more decisive than those which refer only to 

 the size of the whole. 



