134 ACQUIRED CHARACTERS SEC. 



after the action of moderate warmth from the same second 

 generation of Prorsa. Dorfmeister kept caterpillars in process 

 of changing into pupse, or the pupae themselves, twenty-two 

 days at a temperature of 12.5 C., and then let them develop 

 in the room. A portion of these pupae, which belonged to the 

 second Prorsa generation, emerged the same autumn, another 

 portion, like Levana, in the following spring. The former 

 were intermediate between Prorsa and Levana (Porima) in two 

 stages, the latter were Levana partly approximating to Porima. 



The Prorsa obtained agree with forms which also occur in 

 natural conditions. I find them all in my collection. Whether 

 this is true for the various Levana or Porima forms I cannot 

 say. The possibility that artificial varieties are produced 

 which do not occur under natural conditions is meanwhile 

 not excluded. 



If I interpret Dorfmeister's experiments rightly, it is 

 possible to satisfy the greatest demand that can be made as 

 to the influence of external conditions on the modification of 

 organisms ; it is possible to call forth, thermometer in hand, 

 definite varieties which are possibly absolutely new, not occur- 

 ring under natural conditions. 



I cannot sufficiently emphasise the fact that the warmth and 

 cold forms of Prorsa and Levana, differing as they do, not by 

 one, but by several new characters, together afford the most 

 conspicuous example of the correlative, " kaleidoscopic " origin 

 of new forms. 



This fact leads at once to a question closely connected 

 with it, namely, whether such discontinuous modifications 

 depending on external influences are not only apparently dis- 

 continuous, whether they are not stages in the course of a 

 modification of the species which once proceeded gradually, 

 so that they are merely reversions to forms formerly dominant. 

 I am quite of opinion that this explanation applies in many 

 cases, and equally so that in others it does not that rather in 



