558 



THE AMERICAN NATURALIST [Vol. LIII 



the geolofrical history of brachiopods, echinoderms and mollusks. 

 If a .suitable environment continues, the specialized organism may 

 coiitiiuie indefinitely. The idea that genetic variation occurs only 

 in I 1 tion and is irreversible is widespread, but needs sub- 

 stantiation before we accept it into a category of fixed ideas. The 

 world indeed may wait long to see again a four-toed horse, but 

 the reason probably is that we already have a better type in the 

 one-toed horse, which replaced the former because it was better, 

 not because it was degenerate. If selection, natural or artificial, 

 saw at the present time a distinct advantage in a polydactylous 

 hoi-se, it is quite possible that the type might once again be pro- 

 duced. The animal breeder would ask only such a start as was 

 seen in Caesar 's three-toed steed, to produce a race of polydactyl 

 horses. W. E. Castle. 



BrssKY Institution, 



Forest Hills— Boston, Mass., October 1, 1919 



ANOMALOUS KATIOS IN A FAMILY OF YELLOW MICE 

 SUGGESTING LINKAGE BETWEEN THE GENES 

 FOR YELLOW AND FOR BLACK 



During the course of an experiment involving the breeding of 

 yellow and non-yellow varieties of mice certain anomalous ratios 

 were i)ro(luced by a family of yellow mice. Since an explanation 

 of these facts brings out coiLsiderations regarding yellows which 

 have not been treated in the literature of the subject, it seems 

 well to put the case on record. 



The peculiar family originated in a cross of black-and-tan (a 

 very dark form of yellow) with brown. F^ consisted of blacks 

 and yellows. The blacks when tested proved to be heterozygous 

 for brown and showed in their subsequent generations no pecul- 

 iarities of inheritance. The F^ y<5llows should theoretically have 

 been heterozygous for both black and brown for, 



Let YyBB = black-and-tan parent (yellow carrying black) 

 and yybb= brown parent; 



Then F, should consist of yellows, YybB. and bla<-ks, yyBb. 



These F, yellows were back-crossed to |)uiv lifowns. 



Tin' progeny distribution to be exi)('( ti 1 1 would !)'■ as follows: 



The F, vellow parent. YybB, would fofni -auirtes. YB, Yb. 

 yB and yb. 



The luown parent, yylili. would form only one type of gamete, 

 viz., yh. Th.' .'xiH-rt.'d /y-oii.- combinations would be 



The yellow youn<r obtained from this back-cross should be of 

 two genotypes, YyBb (carrying both black and brown) and 



