No. 639] INHERITANCE OF BLUE IN POULTRY 315 



other. The F^'s from such a cross have been secured by 

 mating an extracted white of the composition pp(2?e) (Re) 

 with a black Andalusian, PP{rE){rE), which gave all 

 blues. From these an attempt will be made to secure F, 's 

 in considerable numbers. Back-crosses to the parental 

 types will also be made. The Fg's should approximate 

 the same ratios as appear in Table IV and also give some 

 evidence on the second possible explanation of the per- 

 sistent deviations about to be discussed. 



The calculation of theoretical expectancies presup- 

 poses the equal viability of all phenotypic and genotypic 

 classes. If for any reason the individuals of one or more 

 of the obtained classes tend to be less viable than certain 

 other classes, deviation from expectancy will occur if the 

 lack of viability expresses itself prior to making the 

 counts. 



As has already been pointed out, the lumping of the 

 data presented in Table IV brings forth a poorer fit than 

 is shown in any of the constituent groups. The deficient 

 classes are the blue-splashed and the white, while the 

 most preponderant class relatively is the black. 



It seems to be a rather tacit assumption among poultry- 

 men, particularly, it must in truth be said, among those 

 breeding pigmented varieties, that the recessive white 

 varieties are less vigorous (and so in all probability less 

 viable) than the pigmented varieties of the same breeds. 

 In how far this assumption is based on fact there is no 

 critical evidence to call upon. 



Regarding the relative viability of splashed and self- 

 colored races there is no suggestion from any source. 

 Splashed varieties are, so far as I am aware, nowhere 

 bred as such, and the experience of practical breeders 

 may accordingly not be appealed to. 



While in the case in hand the assumption of low via- 



