ANIMAL EVOLUTION 29 



earlier in the segmentation phases than is the case in 

 species in which the larva is fully developed and feeds 

 itself. The more the larval stage is obscured, the earlier 

 are the adult constituents segregated during the segmen- 

 tation of the egg, and the more completely does the 

 development manifest itself as a ' mosaic- work '. In fresh- 

 water worms and earthworms there is no larval stage, 

 and here we find that the constituents for the formation 

 of the adult body are segregated at an early period, 

 while those for the formation of larval organs are perhaps 

 indicated, for these worms are descended from marine 

 worms, but as there are no larval organs to be formed 

 the constituents necessary for their formation are never 

 segregated. They have dropped out of the life-history, 

 and cannot any longer be present in the germ, for if they 

 were there they would make their presence manifest in 

 the course of the development. 



There is evidence that this process of abbreviation of 

 the development, involving the suppression of the larval 

 stage and the precocious segregation of the factors 

 appropriate to the formation of adult organs, has not 

 occurred once only, but has happened again and again 

 within the limits of the class. Every time that it has 

 happened there must have been some change in the 

 constitution of the germ-plasm ; some loss of tendencies 

 to produce larval organs, some hurrying up and strengthen- 

 ing of tendencies to produce adult organs. 



Ultimately this hurrying up process, which we call 

 precocious segregation, must affect the germ-cell itself. 

 For in cases where the first two, four, eight, or even 

 sixteen blastomeres are what we call totipotent, in other 

 words, when each contains all the factors necessary to 

 produce the specific structure of the entire organism, 

 there cannot be any pre-localization of the factors in 

 the germ-cell : they must be equally distributed through 

 it. On the other hand, the cases of Dentalium and the 

 Limpet — several others could be quoted — show that this 



