624 



THE AMERICAN NATURALIST [Vol. XLVII 



brown, black and chestnut, five colors; showing that in 

 the germ cells of a gray horse there must be the factors 

 for five colors. In view of all the evidence which I have I 

 adhere to my first interpretation, adopt the long series 

 and place brown recessive to bay. Bay I place between 

 brown and the two colors which are dominant to it, gray 

 and roan. 



That gray and roan are dominant to bay there can .be 

 no doubt. Nine hundred and six foals from matings 

 gray X not gray produce 46 per cent, gray and 54 per 

 cent, not gray. It is known that homozygous gray when 

 mated with any of the four popular colors will always 

 produce a gray. It is only from a heterozygous gray that 

 other than a gray can be produced. Eoan behaves 

 exactly the same way. I have no records that would indi- 

 cate the comparative strength of roan and gray. For 

 the present I place them at the top of the series as of 

 coordinate strength. It is just possible that there is a 

 white that is dominant to both the gray and roan, but this 

 has not come under my observation. Nor do I have any 

 data to enable me to place dun in a series. 



The cause of the different shades of roan, bay and 

 chestnut must be left to another paper, as well as the in- 

 teresting behavior of the white markings to be found on 

 most horses, and also the dappled condition of certain 

 grays, bays and chestnuts. 



Sturtevant has suggested that C represent the factor 

 for chestnut; H for black; B for bay (or brown) ; G for 

 gray ; R for roan, and W for white. I now suggest this 

 change : Add the factors for brown, Br, and dun, D, and 

 change black to Bl. The series then becomes : C, hypo- 

 static to all others. Bl epistatic to C but hypostatic to 

 Br, which in turn is hypostatic to B. G and R are both 

 epistatic to B, and perhaps are hypostatic to W. This 

 leaves D (dun) unplaced except that it is known to be 

 near the top of the series with G and R. 



