VI EVOLUTION OF BEES 269 



ceding considerations, appear in all cases as separate organs 

 of an original form in which both were united, a view which 

 is supported by other well-known facts. Accordingly, there 

 is reason to maintain that the separation of the sexes depends 

 partly on conditions of nourishment, and their peculiar 

 characters (especially in mental attributes), partly on correla- 

 tion. To this question I shall return, and then at the same 

 time attempt to explain parthenogenesis by considerations 

 closely connected with those I have just suggested. 



The case under examination thus appears to me to have 

 a peculiar importance in the treatment of wider questions, 

 and therefore I will shortly summarise the results of my 

 reflections upon it. 



1. The facts that queen bees can be reared from worker 

 larvse by better nourishment, and that the former have bodily 

 and mental characters quite different from those of the latter, 

 show clearly, 



'a. the influence of nutrition on the modification of 



forms, 

 h. the importance of correlation in the same process — 

 the importance of kaleidoscopic transformation. 



2. This remarkable process, in particular the fact that the 

 mental characters of the two kinds of bees, workers and 

 queens, are so completely different, and in each case so 

 distinctly form a complete whole, can only be explained on 

 the assumption that both sets of characters were originally 

 united in one form, and that their separation has resulted in 

 consequence of the necessity of division of labour, througli 

 kaleidoscopic transformation, so that the two complementary 

 forms appear as organs of that original form. 



3. But the orimnal form included the drones also. It is 



o 



in consequence of richer nourishment of the ^gg that even 

 w^orker larvae are produced : on nourishment by the sper- 

 matozoon, that is, on fertilisation. If this nourishment is not 



