340 



THE AMERICAN NATURALIST [Vol. XL VII 



assume that there is one factor whose "absence" makes 

 possible the development of the spotted coat. It seems 

 to us that the experimental evidence, more especially 

 the selection experiments of Cuenot and of Castle, sug- 

 gest the possibility that the "spotted coat" is a very 

 complex affair, depending presumably on a number of 

 factors. 



Although this possibility has been repudiated by 

 Castle and not considered by Cuenot, it may be at least 

 worth serious examination; for, if it should prove true, 

 an entirely different appearance will be given to the se- 

 lection experiments referred to above. Now, the fact that 

 the modal class changes when much spotted (with white) 

 and little spotted (with white) animals are selected, and 

 the fact stated by Cuenot that much spotted behaves like 

 a dominant to little spotted, suggests that we may be deal- 

 ing here with a mixed population that may be treated in 

 conformity with a Mendelian interpretation of the prob- 

 lem. 



If much spotting has arisen through a series of prog res- 

 sive mutations, the following hypothesis may serve at 

 least to put the facts in a new light. 



It may be expresed in a general way as follows : If one 

 special condition must be realized before any spotting 

 can occur (the first realized stage may be simply due to a 

 recessive spotting factor ss. Such an animal, mated to 

 pure uniforms, will give : 



8 



S Uniform 





s Spotted 



S 



s F t ? 



s 



s F lC ? 



SS Ss 



Ss — ss 



1SS— 28s— Iss F 2 

 which is the simple Mendelian ratio of 3:1. In other 

 words, the first realized stage of the spotted is a modifi- 

 cation of the original factor and therefore its allelo- 

 morph. This means that in all ss animals the spotted 



