NATURAL SELECTION 



l 9* 



and that there are more opportunities for the rapid achieve- 

 ment of results (e.g. by pedigree breeding). If this is true, the 

 processes of Artificial and Natural Selection differ rather in 

 the relative abundance of their material and the means for 

 rapidly producing and stabilising new combinations than in 

 any more fundamental difference. Though we may admit 

 that much polymorphism occurs in nature, there is nothing 

 equivalent to the judicious utilisation of suitable crosses 

 coupled with the isolation 

 of desirable combina- 

 tions, when once estab- 

 lished. It seems then 

 that the analogy does on 

 examination become di- 

 vested of much of its 

 original force. If it is 

 argued that selection is 

 nevertheless the trans- 

 forming agency, it is only 

 reasonable to admit this, 

 but it is a selection ap- 

 plied in circumstances 

 that can scarcely be ever 

 realised in nature. 



(b) Experimental selec- 

 tion. — Since Johannsen's 

 classical ' pure-line ' ex- 

 periments several at- 

 tempts have been made 

 to modify inbred stock 

 by selection. Results 



similar to those obtained by Johannsen have been obtained 

 by Ewing (191 6), Jennings (1910), Ackert (1916), Lashley 

 (1916), and Zeleny and Mattoon (191 5). In these experi- 

 ments selection shifted the mean of a given character 

 to some extent and was subsequently ineffective. More 

 definite progressive modification was obtained by Banta 

 (1921), Jennings (191 6), and Castle (1919). It is as well, 

 however, to remember that the ' residual heredity ' (*.*. the 

 amount of variation that a strain heterozygous for several 

 characters is capable of manifesting) of one stock may be more 



Fig. 2i. — Individuals of two different 

 Clones of Hydra, kept under similar 

 Conditions. 



(From Lashley, 19 16.) 



