NATURAL SELECTION 97 



where there is rigorous elimination to check the backward flow. 

 Even among the spermatozoa there is keen competition and 

 :onsequent weeding out of the unfit, and to this perhaps we 

 may attribute the fact that pammixis does not work with greater 

 rapidity. 



Though I attribute stability as a general rule to selection, I 

 im quite aware that a new character is sometimes constant from 

 the first. But such a thing is exceptional. If we refer to the 

 experience of breeders we find that constancy is due, as a rule, 

 to systematic selection. And wild nature must be subject to 

 the same law. When individuals belonging to a wild species 

 have by the aid of some marked variation survived a change of 

 environment and become the founders of a new species, they are 

 to some extent in the position of a domestic breed in which some 

 particular point has been recently accentuated by selection. 

 Under the circumstances, great stability is not to be expected 

 till some generations have passed and the tendency to retro- 

 gression has been checked. 



As to the tendency among domestic animals, the testimony of 

 breeders is emphatic. Mr Baker, quoted by Darwin, 1 thus sums 

 up, " Whenever there has been neglect, the breed (of horses) 

 has proportionately deteriorated." Darwin adds what may seem 

 like a qualification of this. " From what we see of the many 

 common races of dogs, cattle, fowls, pigeons, etc., which with- 

 out any particular care have long retained nearly the same 

 character, we have no grounds for believing that they would 

 altogether depart from their type." 2 He admits, however, that 

 they would be " very far from excellent of their kind," i.e. t as I 

 understand it, they would rapidly after a score or a hundred 

 generations is Darwin's estimate lose their most recently 

 acquired characters. 



I cannot help thinking that Darwin greatly underestimated 

 the tendency to retrogression. Surely there is a good deal of 

 unmethodical selection in the cases where the fixity, the very 

 slow deterioration of which he speaks, is observable. There is 

 plenty of evidence to show that stability, once it has been lost, 



1 Animals and Plants under Domestication, vol. ii. p. 139. 2 Loc. (it, 



G 



