220 DEVELOPMENT AND HEREDITY. 



The principle which I have illustrated by the case 

 of the butterfly seems to me to be applicable to all 

 cases of animal metamorphosis. Wherever animals 

 with an established method of development are in 

 active intercourse with their environment, and strug- 

 gling for their existence through a long period of 

 immaturity, it must follow that when the conditions 

 of their larval life change, the course of their larval 

 development must be affected. In any change of 

 environment it is hardly possible that all the stages 

 of larval life could be equally affected. The change 

 would probably be more favourable to some stage or 

 stages of life, and less favourable to others. The 

 result would be a prolongation of the favourable 

 stages of life, and an abbreviation and compression 

 at other stages. Thus may be explained the origin 

 of periods of sudden changes, or metamorphosis. 



In the same category with metamorphosis should 

 be placed another class of phenomena, which is 

 usually described as " alternation of generations." 

 But we must except here dimorphic species, already 

 discussed, and that kind of alternation of genera- 

 tions which consists in the alternation of sexual and 

 parthenogenetic generations, as in Aphis and other 

 arthropoda. This latter is a phenomenon that truly 

 deserves its name. The cases, however, of many of 

 the Hydroids and Tunicates are of another char- 



