138 THE SHORTER SCIENTIFIC PAPERS 



The investigations of Kammerer, Woltereck, Ferro- 

 niere, etc., are of decided interest, but to those critically 

 inclined they offer no evidence giving pronounced sup- 

 port to the proposition that environmental stimuli form 

 new genetic factors. 



Thus, in turn, have the theories as to the method by 

 which evolutionary progress occurs been undermined by 

 doubt. Feeling the insufficiency of small chance varia- 

 tions, of environmental variations, and of larger germinal 

 variations, as a summation process, it is not to be wondered 

 that the truth-seeking pilgrim has become wearied in his 

 journey and longs for a more secure resting place. 



III. 



Let us return to the problem as suggested in the open- 

 ing paragraph, namely the actual origin of heritable 

 characters, and consider somewhat more carefully as to 

 whether theories exist justified by facts, which will fur- 

 nish acceptable evidence. There are two well-developed 

 hypotheses, the general one of DeVries and the more 

 specific one of Morgan and his associates, founded on dis- 

 continuous variations, and that of Castle based on con- 

 tinuous variations. 



Considering the views of DeVries and his followers in 

 the light of experimental investigations made during the 

 last ten years, it has become more and more evident that 

 by far the greater number, if not all, of the so-called 

 mutations thus obtained, were explainable on the basis of 

 the combinations of preexisting units of the germ cells. 

 This rests upon the proposition that there are present in 

 the gametes certain hypothetical entities to which the 

 name gene or factor has been applied and which give rise 

 to the heritable characters of an organism. Thus it is at 

 once recognized that the problem relates to the origin of 

 the gene, rather than to the origin of the apparent char- 

 acters with which it is correlated, and that by far the 

 greater number of so-called new characters are not new, 

 but were performed at remote periods of time. So far 

 as the present arguments are concerned, it matters not 

 whether the results are assumed to be brought about by 



