322 Geofifrey Smitli 



testes. This case is of particulav interest to us because it proves 

 that the form aud development of the chela is correlated with the 

 primary sexual development, and also that the phenomeuon of high 

 and low dimorphism is due to the assumption of particular states of 

 sexual development at jìarticular stages of growth. 



I know of no observations upon the primary sexual development 

 of high and low insects, but it can hardly be doubted that this class 

 offers no exception to the universal rule that the secondary and 

 primary sexual characters are here too closely correlated. But in 

 the case of insects a special word of warning is needed. It has been 

 proved in several cases (Oudemans 4 and Kellogg 5) for butter- 

 flies and moths that the destruction of the rudiment of the gonad in 

 the larva has absolutely no effect on the development of the secon- 

 dary sexual characters. This apparently contradictory result is only 

 contradictory in appearance; because it is possible to prove, as I 

 intend to show in another work, that the development of the secon- 

 dary sexual characters in general is not due to the presence of a 

 corresponding differentiated sexual gland, but that the diflferentiation 

 of both the secondary aud primary sexual characters is due to a 

 third factor, which I will call the primary sexual development, aud 

 which may be uninfluenced by the destruction of the germinai cells 

 at an early period. 



It is conhdently assumed bere that the phenomenon of high and 

 low dimorphism, both facultative and definitive, is due to a parti- 

 cular kind of evolutiou of the primary sexual development in the 

 growth of the individuai, such that the males tend to become split 

 up into high and low forms difiteriug both structurally and physiolo- 

 gically in their sexual nature? 



5. Evidences of high and low dimorphism as a meaus of 

 progressive differentiatioii. 



The principle of high and low dimorphism has been shown to 

 affect the males of various widely diflerent species, and to constitute 

 a differeutiating factor within the limits of these species. We may 

 now enquire whether this principle of differentiation may not extend 

 outside the limits of a species aud exert a Controlling influence in 

 the general evolution of organic groups. It appears that if this is 

 the case we ought to be able to observe a certain relation between 

 the male secondary sexual characters of the species which constitute 



