xm VARIATION AND EVOLUTION 147 



nomenon, but natural selection plays the part of a con- 

 servative, not of a formative agent. 



It is interesting to recall that in earlier years Darwin 

 was inclined to ascribe more importance to "sports" as 

 opposed to continuous minute variation, and to consider 

 that they might play a not inconsiderable part in the 

 formation of new varieties in nature. This view, how- 

 ever, he gave up later, because he thought that the rela- 

 tively rare sport or mutation would rapidly disappear 

 through the swamping effects of crossing with the more 

 abundant normal form, and so, even though favoured by 

 natural selection, would never succeed in establishing 

 itself. Mendel's discovery has eliminated this diffi- 

 culty. For suppose that the sport differed from the 

 normal in the loss of a factor and were recessive. When 

 mated with the normal this character would seem to dis- 

 appear, though, of course; half of the gametes of its prog- 

 eny would bear it. By continual crossing with normals 

 a small proportion of heterozygotes would eventually be 

 scattered among the population, and as soon as any two 

 of these mated together the recessive sport would ap- 

 pear in one quarter of their offspring. 



A suggestive contribution to this subject was recently 

 made by G. H. Hardy. Considering the distribution of 

 a single factor in a mixed population consisting of the het- 

 erozygous and the two homozygous forms he showed that 

 such a population breeding at random rapidly fell into a 



