METAMORPHOSIS. 211 



acquired characters, but must have developed solely 

 for the period of embryonic life for which they are 

 useful, and they must, further, be the result of 

 the peculiar conditions which came gradually to 

 surround that period of the life of the animal. The 

 placenta, etc., have probably been caused by the 

 new and peculiar stimuli which began to act on 

 the organism when the egg remained so long in the 

 body of the parent that part of its development was 

 performed under these novel conditions. This be- 

 lief is strengthened by viewing the similarity of 

 embryonic conditions, and the consequent similarity 

 of modifications which have arisen in such widely 

 different groups as the mammals, the selachian 

 fishes, the lophobranchiate fishes, and Doliolum 

 among the tunicates. We see therefore that this 

 associated chain of nervous reactions which form 

 the hereditary impulse resembles in its nature 

 the ordinary chain of associated reactions which 

 are established in an individual by practising any 

 regular series of movements. For in such a series 

 of reactions the middle part of the chain can be 

 gradually modified and changed without destroying 

 or weakening or changing the other associations in 

 the first and last parts. In such a series of well- 

 practised movements as the playing of a passage 

 of music upon an instrument, the first effort to 



