THE EEQAPITULATION THEOEY. 31 



inevitably lead to starvation. The difficulty is evaded by 

 retaining the external form and habits of one particular stage for 

 an unduly long period, so that the relation of the animal to its 

 surrounding environment remains unaltered, while, internally, 

 preparations for the later changes are in progress. 



5. The tendency t o the acquisition of new character s. This 

 has been dealt with already ; it arises from the fact that the 

 larval forms of animals, like the adults, are exposed to the action 

 of natural selection, and so are liable to acquire characters that 

 do not belong to the ancestral history. 



Before leaving the subject it is worth while inquiring 

 whether any explanation can be found of recapitulation. A 

 complete answer can certainly not be given at present, but a 

 partial one may, perhaps, be found. 



Darwin himself suggested that the clue might be found in 

 the consideration that at whatever age a variation first appears 

 in the parent, it tends to reappear at a corresponding age in the 

 offspring ; but this must be regarded rather as a statement of 

 the fundamental fact of embryology than as an explanation of it. 



It is probably safe to assume that animals would not 

 recapitulate unless they were compelled to do so : that there 

 must be some constraining influence at work, forcing them to 

 repeat more or less closely the ancestral stages. It is impossible, 

 for instance, to conceive what advantage it can be to a chick or 

 a rabbit embryo to develop gill clefts which are never used, and 

 which disappear at a slightly later stage ; or how it can benefit 

 a whale, that in its embryonic condition it should possess teeth 

 which never cut the gum, and which are lost before birth. 



Moreover, the whole history of development in different 

 animals or groups of animals offers to us, as we have seen, a 

 series of ingenious, determined, varied, but more or less 

 unsuccessful efforts to escape from the necessity of recapi- 

 tulating, and to substitute for the ancestral process a more 

 direct method. 



A further consideration of importance is that recapitulation 

 is not seen in all forms of development, but only in development 

 from the egg. In the several forms of asexual development, of 

 which budding is the most frequent and most familiar, there is 

 no repetition of ancestral phases ; neither is there in cases of 



