Mecapit Illation and Conclusion. 323 



thus to give rise to variation, than gemmules which are 

 formed in the female body, we should expect to find, in 

 a variation which first appears in a male, much more 

 tendency to become hereditary than in a variation which 

 first appears in a female. For the same reason we should 

 expect to find an organ which is confined to males much 

 more likely than one confined to females to give rise to 

 hereditary modifications. 



For a similar reason we should expect the males of 

 unisexual animals to vary more than the females. 



We can form some conception of the amount of modi- 

 fication which an animal has recently undergone by com- 

 paring the adults with the young, and hf comparing 

 allied species with each other. 



AVhen we make comparisons of this kind w^e find that 

 throughout the animal kingdom, with very few excep- 

 tions, wherever the sexes are separate and differ from 

 each other, the males of allied species differ from each 

 other more than the females do, and the adult male differs 

 more than the adult female from the young. This law 

 is so j)ronounced and conspicuous that its existence has 

 long been recognized by all naturalists. 



We also find that organs which are confined to males, 

 or which are of more importance or are more perfectly 

 developed in the males than they are in the females, are 

 very much more likely to give rise to hereditary modifi- 

 cations than parts which are confined to or are most 

 developed in females; that a part which is thus confined 

 to males is much more likely to vary than a similar female 

 part; that males are, as a rule, more variable than females; 

 and that the male leads and the female follows in the 

 evolution of new races. 



The scientific accuracy of most of these generalizations 

 regarding secondary sexual characters has long been 



