:\22 



THE AMERICAN NATURALIST [Vol. XLVTI 



dence of the red. He states that the development of 

 black does not hinder the development of some red pig- 

 ment also in the hybrid, but the red so developed is con- 

 cealed by the black. Black he regards as epistatic to red. 

 Castle states in his recent book (1912) that in the F 2 gen- 

 eration three blacks to one red are produced. 



Spotted animals contain white in patches. These 

 patches may be very small in extent, or, at the other ex- 

 treme, extend over the whole coat so that the eyes alone 

 have dark pigment. These black-eyed whites, however, 

 do not breed true, but produce spotted offspring, the 

 spotting being variable. Black-eyed white mice give this 

 result, and are to be sharply separated from albinos that 

 have pink eyes and white hair. Albino guinea-pigs often 

 have small patches of black, especially on the feet and 

 ears, but this is not true for albino mice or rats. 



In guinea-pigs the spotted animals may be black and 

 white ; or red and white. These races are said to breed 

 true, or at least certain bicolor races of these kinds breed 

 true. In addition there are races having red, black and 

 white in their coats. These are the tricolors and it is 

 with this race that we are here chiefly concerned. It has 

 just been said that the tricolor is a distinct race, but this 

 must not be understood to mean that they do not pro- 

 duce bicolor animals. In fact, amongst the offspring, bi- 

 color animals continually crop out. It is this fact that 

 has led Castle in his recent article to state that tri- 

 colors do not breed true. The bicolors produced in this 

 way seem to differ from the pure races of bicolor in that 

 they may produce tricolors again. For the present the 

 question may be left open whether pure races of bicolors 

 could be produced by selection of bicolors thrown by tri- 

 color parents. Of course, if bicolor races had originally 

 been incrossed, such a separation would be expected. In 

 our experiments, at least, some bicolor individuals have 

 appeared that seem to breed true, although the experi- 

 ments are not extensive enough to settle the question. 



In the following account, therefore, it should be under- 



