194 THE FUNCTION OF SEX 



which, as it happens, are able to cross successfully. On the 

 Mendelian theory, however, the effect of bi-parental reproduction 

 should be calamitous ; and one is left to wonder how it happened 

 that this form of reproduction replaced parthenogenesis, in which 

 such a disastrous mingling of incongruous elements would, at 

 least, not occur. Moreover, self-fertilization is not uncommon, and 

 is a product of evolution. Here, since, almost invariably, only 

 like characters meet, there can be, speaking practically, no mixing 

 of unlike allelomorphs, and we must suppose that conjugation 

 that sex, which has burdened living beings with so many 

 characters has no function whatever. It is surely not surprising, 

 then, that Mendelians, conscious of the tangle, should feel moved 

 to declare that they " look on the manner and causation of adapted 

 differentiation as still wholly mysterious," and that "we can 

 profitably reject, as I believe, much of the theory of Natural 

 Selection, and more especially the idea that adaptations have 

 arisen because of their usefulness." 1 



320. On the other hand, if the mutation theory as conceived 

 by Professor de Vries is true, that is, if the variations on which 

 progressive evolution is founded are large discontinuous mutations, 

 and if mutants tend to be sterile when crossed with the parent 

 type, or if fertile, to have offspring which blend the parental char- 

 acters, then, again, one is left to wonder at the extraordinary 

 function assigned by implication to sex. When reproduction is 

 parthenogenetic, useful mutations, if they occur, have an excellent 

 chance of surviving. But when reproduction is bi-parental, they 

 can very rarely, if ever, escape being lost through sterility, or 

 swamped through blending. According to this hypothesis, there- 

 fore, it would appear that the function of sex is to render nature 

 powerless to produce adaptive changes. And yet we must suppose 

 sex to have been evolved by nature. In this case also how shall 

 we account for the occurrence of self-fertilization ? 



321. Examine now the theory that evolution proceeds through 

 the constant selection of ' continuous,' 'normal,' or 'fluctuating' 

 variations, the inheritance of which is blended. 2 As we have seen, 



1 See 306. 



* The term ' blended/ with the meaning given it in the present work, is not 

 altogether correct ; but it is convenient. It is used in contradistinction to 

 ' alternative ' and includes all shades of blending from an equal mixture of parental 

 characters to an exclusive reproduction of the character of one parent with total 

 loss of the character of the other. Exclusive inheritance in this sense differs 

 from dominance in that the former implies the complete loss of the character of 

 one parent, whereas the latter implies only its latency. To the Mendelian 



