503 



resulted in a gradual shifting of the period of growth to the 

 beginning of the free living period; and have at the same 

 time forced the larvae to emerge from the egg in increasingly- 

 earlier periods of individual development. The processes of 

 accelerated growth and of premature emergence from the egg 

 have reached their maximum not among Muscids, as is usually- 

 supposed , but among the chalcid wasps. 



By this gradual concentration of the growth period to the 

 beginning of the post-embryonic life a new organism which 

 may have to fit into a wholly new environment has been 

 produced — the larva. Coenogenetic modifications, almost as 

 interesting as the adaptations of the adult insect, have 

 arisen ; but it is in the protected parasitic larvae that we must 

 look for cases in which the processes of premature hatching 

 and rapid feeding, unfettered by a complex environment, 

 have proceeded the farthest ; and it is among the chalcid wasps 

 that the most profound recovery from larval specialization — 

 metamorphosis — is to be found. 



It is possible now to obtain a clearer conception of the 

 significance of the pupa. When specialization first developed, 

 return to the imaginal condition was doubtless by means of 

 cellular rejuvenation. Increasing specialization of the imago 

 produced, automatically, as I have shown above, increased 

 specialization towards rapid food absorption. When recovery 

 by rejuvenation from this specialization became impossible, 

 continued specialization of some larval cells must have been 

 accompanied by decreased specialization of others; ultimately 

 the latter would have remained as embryonic cells, and 

 imaginal "discs" were the result. 



Now in the early stages of evolution of the larval form, 

 the resulting metamorphosis must have been of a very simple 

 type, differing little, indeed, from direct development; a 

 simple rejuvenation of tissues, such as occurs still to a certain 

 extent amongst Coleoptera, must have occurred. Nothing is 

 known of the metamorphosis of the dragon-flies, but a cell 

 rejuvenation metamorphosis among these forms may be pre- 

 dicted with considerable confidence. There is no extensive 

 tissue death, apparently, and it is not possible here to speak 

 of a pupa in the usual sense of the word. But as the specializa- 

 tion became more marked, extensive tissue death, and 

 corresponding tissue regeneration occurring late in larval life, 

 brought about an absolute break in the orderly developmental 

 process, and the actual development of the less generalized 

 characters of the insect did not begin till after the larval 

 tissues had died. This metamorphosis begins considerably 

 before the last ''larval" moult; eventually when the larva 

 does moult there comes to view a wholly different organism — 



