14 



TEE AMERICAN NATURALIST [Vol.LVI 



cliaiige, it is clear that the character altered was not a 

 permanent one. And if after long re-subjection to the 

 original environment the induced change disappears, it 

 is equally clear that the new character was no more per- 

 manent than the original one. If we now assume that 

 there are other characters that are permanent, not alter- 

 able by environmental action, of course we can obtain no 

 light on these by changing those characters that can be 

 changed. To me it appears that we have no right to as- 

 sume, at the present stage in the game, that any such 

 absolutely permanent characters exist. If this be true, 

 then the production of changes persisting through many 

 generations of uniparental, and even of biparental, re- 

 production, with the further fact that the greater the 

 number of generations the altering agent has acted, the 

 greater the number of generations the change persists, 

 seems of the greatest interest. It perhaps would, if 

 action of the environmental agent continued sufficiently 

 long, lead to production of inherited characteristics that 

 are as permanent as any such characters are. It is cer- 

 tainly, as Jollos agrees, capable of producing such di- 

 versity of biotypes as we find within a species; and it 

 might perhaps, if the results of diverse agents are cumu- 

 lative, produce any of the inherited diversities found in 

 organisms. This is the most promising lead that we 

 have found in the study of uniparental production.^ 



In sum, the study of variation in uniparental repro- 

 duction yields the following: The germinal or genotypic 

 constitution in most organisms is extremely stable; in 

 many stocks it changes not at all, so far as observation 

 goes. To alter it by physical or chemical agents is usu- 

 ally to kill it. In some of the lowest organisms -rhizo- 

 pods, bacteria, some infusoria— it changes with some- 

 what greater frefpiency, though still rarely. The nature 

 of til' cli.iiiu.-. .111.1 wlicthei- they may be permanent, or 

 I'l"-' - !>' 1 .itiuii^ revert to the original condi- 



