308 The Genetics of the Dutch RahUt 



on a strictly factorial basis: whether we can assign to each- member of 

 the series a definite genetic constitution, the validity of which we can 

 test by definitely devised experiment. To some extent Castle's in- 

 terpretation is a factorial one in that he offers an explanation in terms 

 pf the four factors forming his system of multiple allelomorphs. He 

 admits however that some of the data can hardly be brought into line 

 with his interpretation, a point to which I shall recur later. On the 

 other hand, by postulating " mutual modification " between members of 

 his allelomorphic series he appears to me to be introducing a complica- 

 tion without the justification of fact. I confess however that I am not 

 clear as to the nature of the process involved in this term. It must be 

 something distinct from heterozygous forms, because it involves the 

 conception of the several members of the allelomorphic series tending, 

 as the result of crosses, to become more like one another. Yet we 

 are told that it "need not be regarded as change in the nature of a 

 single gene, but merely as equalization of the residual heredity addi- 

 tional to the single genes which produce hybrid ratios ^" So far as I can 

 understand this it appears to be an unnecessarily complicated and round- 

 about way of saying that there are other factors concerned of which the 

 nature has not yet been demonstrated experimentally. I propose there- 

 fore to dismiss the terms "mutual modification" and "residual heredity" 

 from the discussion for the present, and to enquire whether we cannot 

 frame a simpler hypothesis to cover the facts than that adduced by 

 Castle. 



Before doing so, however, I wish to draw attention to several points 

 in connection with which Castle's hypothesis appears to me to be un- 

 satisfactory. In the first place before we can offer an interpretation on 

 factorial lines we must be in a position to present evidence for the 

 existence of the various homozygous forms which that interpretation 

 involves. We must determine by experiment which of the terms of 

 our apparently continuous series can be got to breed true, and which 

 cannot. By breeding true is not of course meant that any individual 

 in a homozygous strain should be an exact replica of any other. For 

 reasons that are at present beyond us, the individuals of a homozygous 

 strain may be expected to present some amount of fluctuation in the 

 extent of the white markings. This however need not be very great. 



1 It would appear from a sentence on p. 16 that Castle does not bind himself by this 

 admission, for he states his opinion here "that dark and tan are allelomorphs but separate 

 in a modified form, one on the whole darker the other on the whole lighter than before 

 they were crossed with each other." 



