160 HEREDITY AND SEX 



of factors that determines sex itself. If selection acts 

 here, it must act directly on germinal variations, that 

 are independent in origin of the sex-determining factor, 

 but dependent on it for their expansion or suppression. 



These considerations make many of the earlier state- 

 ments appear crude and unconvincing ; for, they show 

 that the origin of the secondary sexual characters is 

 a much more complex affair than was formerly im- 

 agined. 



These same considerations do not show, however, 

 that if a new germinal character appeared that gave 

 its possessor some advantage either by accelerating 

 the opposite sex to quicker mating or by being corre- 

 lated with greater vigor and thereby making more 

 certain the discovery of a mate, such a character would 

 not have a better chance of perpetuation. But in 

 such a case, the emphasis no longer lies on the idea 

 of selection with its emotional impHcations, but rather 

 on the appearance of a more effective machine that 

 has arisen, not because of selection, but, having arisen 

 quite apart from any selective process, has found itself 

 more efficient. Selection has always implied the idea 

 that it creates something. Now that the evidence 

 indicates that selection is not a guaranteed method 

 of creating anything, its efficiency as a means of easy 

 explanation is seriously impaired. 



