THE BEGINNINGS OF LIFE. 563 



as it lasts) after the homogenetic method, the process 

 itself must, in each of such cases, have been one of 

 Heterogenesis. These transforming molecular changes 

 must have been total and complete, — since all organic 

 memory of the previous phases of such forms seems 

 to have been effectually blotted out. This, however, 

 appears to be by no means the case with all equally 

 simple organisms. In some, the ascending grades of 

 evolution by which the more perfect sexual type has 

 been achieved, have probably been much more gradual, 

 and where this is the case some of these grades may 

 be preserved in the developmental changes which the 

 fecundated germ subsequently goes through. The germ 

 does not tend to develop immediately into a form 

 similar to that from which it had been derived. It 

 rather tends to grow at once into some antecedent 

 form, and to attain to the likeness of its parent, 

 by passing through developmental changes or meta- 

 morphoses, more or less similar to certain current 

 heterogenetic changes which had been previously apt 

 to occur. 



Whenever, therefore, a given series of develop- 

 mental changes or imperfect transformations has ter- 

 minated in the production of a sexual form, and when 

 these changes have not involve-d a period of total 

 molecular rearrangement, the germ, by a power of 

 ' reversion ' (which is not unfrequent even among the 

 higher forms of life) may tend to reproduce such changes 

 rather than undergo an immediate development into 



o o :j 



