THE AMERICAN NATURALIST [Vol. XLIV 



that is to say confined to the generation immediately 

 affected, or do they reappear, if only to a slight degree, 

 in the offspring? I have long felt that satisfactory evi- 

 dence for or against the transmission of such modifica- 

 tions would be lacking, so long as zoologists confined 

 themselves to a search for directly visible qualitative dif- 

 ferences. With few exceptions, however, such has been 

 the method by which the problem has been attacked. 



The following conditions must, I believe, be realized 

 before we may hope to attain to any satisfactory test 

 of this time-honored question: (1) We must select for 

 experiment such an organism and such a physical agency 

 that the latter may modify the former without directly 

 influencing the germ cells. The action of temperature 

 upon a warm-blooded animal seems to realize this condi- 

 tion most fully. (2) We must discover readily meas- 

 urable, quantitative changes in the parent generation, 

 before we can hope to test the reappearance of such 

 changes in the offspring. 



Having discovered such modifications in the parent 

 generation, there are theoretically two methods by which 



raise the offspring of the experimental and control 3 lots 

 under identical conditions, or (2) we may raise the off- 

 spring of the modified parents under the same conditions 

 as were employed to effect the original modification. In 

 the first case, we should compare the two sets of animals 

 having different parentage. Assuming a given modifi- 

 cation of the value a-, and supposing that 1/n represented 

 the proportional part of this to be transmitted, the off- 

 spring of the two lots would be found to differ by the 



should compare the second generation with the parent 

 generation, the two being measured at the same age. 



