1 66 AMPHIMIXIS OR ESSENTIAL MEANING OF [XII. 



belonged to the latter type. She produces ofTspring of the 

 variety B, and the type is retained for many generations. 

 In the illustration described above the type B would be retained 

 indefinitely ; for I assumed that only four idants were present, 

 and that all these became of the variety B. In reality, however, 

 this would occur but seldom, since the constitution of the germ- 

 plasm must be far more complex : not only are the idants more 

 numerous, but their composition out of ids does not remain 

 entirely the same throughout long periods of time, as I have 

 attempted to show in the first part of this essay. 



If the idants are not entirely unchangeable in this respect, if, 

 when they are freshly formed out of ids scattered through 

 the nuclear network, there is an occasional alteration in the 

 arrangement, we might then even assume that, by such dis- 

 placements, a germ-plasm a which contains no purely b idants, 

 but only a few ids belonging to the latter variety included 

 within the a idants, could, nevertheless, in course of generations, 

 undergo reversion to the variety B. But these are niceties, 

 which it is as yet too early to consider ; for w^e are only on the 

 threshold of knowledge concerning hereditary phenomena in 

 parthenogenesis. 



But something at any rate has been proved ; for we can 

 safely affirm that in parthenogenesis individual variation exists, 

 which, as in bisexual reproduction, has its foundation in the compo- 

 sition of the germ-plasm itself, and thus depends on heredity, and is 

 itself inheritable. I thus erred in former times, in believing that 

 purely parthenogenetic species entirely lacked the capability of 

 transformation by means of selection ; they do possess this 

 power to a certain extent. I was, however, right upon the 

 main point ; for their capability of transformation must be 

 much smaller than in bisexual species, as is evident from the 

 observations described above as well as from theoretical con- 

 siderations. The latter indicate that, in the course of genera- 

 tions, the constitution of the germ-plasm must ever become 

 simpler ; while the observations confirm this suggestion, inas- 

 much as they prove that a remarkable similarity exists between 

 the descendants throughout numerous generations. The ad- 

 vantages of that complex intermingling of many individual 

 predispositions which was brought about in the amphigonic 

 ancestors of parthenogenetic species become gradually lost, 



