122 



THE AMEBIC AN NATURALIST [Vol. XLIV 



that these tendencies are separately and independently trans- 

 mitted. The fact is that there is nothing in the scheme of color 

 factors presented by Castle inconsistent with the facts of color 

 production presented by Kiddle, unless it be that they call for 

 two or three more or less distinct oxidizing processes instead of 

 only one. Even the view that there is only a single process 

 involved is not excluded if we can find a mechanism operative 

 in development that stops the oxidation partly at one stage and 

 partly at another. 



I agree with Riddle that the placing of the "uniformity- 

 spotted," etc., factors on the part of Castle in the germs of 

 rabbits as alternative factors is a virtual surrender of the whole 

 theory of discontinuous variation. According to my view, this 

 theory has never had any sound basis except as a result of 

 irregularities in the behavior of chromosomes, and Mendelian 

 facts are not dependent on any such theory or in any way re- 

 lated to it, as the writer has repeatedly pointed out. Professor 

 S. J. Holmes has also pointed out the fallacy in the assumption 

 that .Mendelian phenomena necessarily prove the theory of dis- 



That there may be more than one oxidation process concerned 

 in the development of color in organisms is shown by a number 

 of facts cited by Riddle. He gives a table of oxidation processes 

 in which the end reactions result in various colors. He further 

 cites the fact that lipochronies and guanin may possibly be the 

 source of color in certain amphibia. 



The author attempts to explain the variability in the second 

 generation of hybrids as a result of unstable equilibrium of 

 oxidation processes in the first generation. According to his 

 view each of the gametes represents a stable condition. The 

 first generation may be a compromise between the stable condi- 

 tions found in the two gametes. In the second generation there is 

 a tendency to revert to one or both of the stable points. This 

 view is hardly in keeping with observed phenomena. The 

 regularity with which certain types appear and their subsequent 



such defin 

 that the ( 

 processes < 

 Riddle i 



