202 THE FUNCTION OF SEX 



since persistent variations are so many in parthenogenesis, we must 

 suppose that mutations are a thousandfold more numerous under 

 that condition than when reproduction is bi-parental, which is, to 

 say the least, unlikely. 



332. Speaking generally, among animals that reproduce bi- 

 parentally, varieties are few when the powers of locomotion are 

 great that is when mating between individuals derived from 

 widely separated districts is common. Thus, avian species have 

 few varieties except when spread over a very wide area or an area 

 that is interrupted by natural barriers, and often, unlike partheno- 

 genetic types, only a single variety inhabits a whole region. 

 When there are effective barriers, varietal differentiation is usually 

 noticeable. When the powers of locomotion are less, the number 

 of varieties is greater. Thus, as Gulick noted, every valley in 

 Samoa has its local species of land-shells, the differentiation 

 between the varieties, species, and genera being proportionate to 

 the distances which separate them. 1 All this indicates strongly 

 that the effect of sex is to blend characters. If the theory of unit 

 segregation and gametic purity were true, and, as a consequence, 

 that of blended inheritance untrue, species which reproduce sexually 

 should possess at least as many varieties as those the reproduction 

 of which is parthenogenetic, and those which range widely as many 

 as those with poorer powers of locomotion. In that case there 

 would be, not only no approach to uniformity through conjugation, 

 but such a shuffling of characters that almost every individual 

 would form a variety by itself. 



333. The fact that, apart from the sexual traits, all the individuals 

 of the same variety differ, with rare exceptions, only to the extent 

 of fluctuations, is presumptive evidence that the function of sex is 

 to blend parental characters. In that case the average experience 

 of the whole race, rather than the, perhaps, exceptional experience of 

 single individuals, becomes the determining factor in evolution. The 

 advantage of this is obvious); for most individuals are more likely 

 to have an average than an exceptional experience. Moreover, as 

 I say, blending when combined with Natural Selection is a selecting 

 agency. Selection eliminates harmful progressive and retrogressive 

 variations, and, on rare occasions, old-established characters that 

 have become harmful. Blending, with its swamping effects, its 

 general tendency to produce retrogression, eliminates useless char- 

 acters and variations characters and variations which are merely 

 useless to the individual, but which, if accumulated, would become 



1 See Darwinism, by Alfred Russel Wallace, p. 43. 



