NATURAL SELECTION 41 



a certain favorable variation arise in but a few individuals of 

 a species, it seems improbable that these divergent individ- 

 uals will breed with one another rather than with the much 

 more numerous non-divergent members of the species. If, 

 however, a divergent individual crosses with a non-diver- 

 gent individual, the useful character which has appeared in 

 the divergent individual will be less marked in the offspring. 

 In the following generations it would be still more dimin- 

 ished by the same process, until finally it will be entirely lost. 

 This swamping of variations by interbreeding has seemed to 

 some to make the development of new characters by natural 

 selection improbable. 



The force of this objection is great. Doubtless many 

 divergent characters are swamped by their possessors inter- 

 breeding with those individuals of the species in which these 

 characters do not appear. If it were not for this fact evo- 

 lution might be much more rapid. Evolution is slow, and 

 the swamping effect of interbreeding may largely account 

 for the slowness of the process. But while evolution may 

 be retarded by intercrossing, we have no indication that it 

 is prevented. 



Two individuals of different species ordinarily will not 

 breed together in a state of nature, though occasionally they 

 will do so ; and in those rare cases in which species do 

 cross, the offspring only very rarely are fertile. Nature, by 

 this infertility, has provided against promiscuous interbreed- 

 ing between species, and has thus prevented the species 

 already developed from being lost by confusion with one 

 another. Does she in some similar way prevent promis- 

 cuous intercrossing between the individuals of a single 

 species, and thus secure the perpetuation of favorable varia- 



