1921.] 



Research in Animal Breeding. 



253 



pens of such Fl birds, an F2 generation of 239 birds was raised, 

 viz., 113 cockerels and 126 pullets. In contrast to the uni- 

 formity of the Fl generation these F2 birds exhibited a wide 

 range of variation. As shown graphically in Fig. 6, the weights 

 of the cockerels varied from about 550 to 1,600 grammes, while 

 those of the pullets were from 500 to 1,200. The majority of 

 the birds in this generation were between the weights of the 

 original parental breeds, but a few w^ere larger than the Ham- 

 burgh, and a few were smaller than the Sebright (Plate 1, Fig. 2 

 and 3). Here we have an apparent case of blended inheritance, with 

 fair uniformity in Fl, and a wide range of variation in F2. Can 

 such a case be interpreted in terms of the factorial theory? An 

 interpretation is possible if we suppose that the Hamburgh and 

 the Sebright differ in several factors, each of which affects the 

 weight of the bird. The explanation of such cases was first given 

 by Nilsson-Ehle, the well-known Swedish plant breeder, to 

 account for the results of certain of his experiments w'ith wheat 

 and oats at Svalof. The closeness with which the theory fitted 

 his results left little doubt of its being a true interpretation. The 

 essential part of his idea is that a similar effect may be brought 

 about by more than one factor, though such factors are inde- 

 pendently transmitted in the usual way. 



Let us suppose that there are several similar factors A, B, C, 

 D, &c., which influence the weight of poultry. When a bird 

 possesses none of these factors it will be the smallest type of 

 bantam ; when it contains A it will be rather larger ; when it 

 contains both A and B it wdll be larger again, and so on until 

 the largest breed is reached, which must be supposed to contain 

 a full collection of these factors. Again, let us suppose that 

 when a bird is pure for one of these factors, i.e., when it has 

 received it from both parents, the effect on its weight is greater 

 than when it has received it from one parent only. In other 

 words we suppose that dominance is not complete, and that the 

 Aa bird, for example, is not so heavy as the A A bird of otherwise 

 similar constitution. And so also for the other weic^ht factors, 

 B, C, D, &c. 



Now if we suppose that the Hamburgh contained three such ' 

 factors. A, B and C, w^hile the Sebright contained a different one, 

 viz., D, we obtain a theoretical explanation which covers the 

 observed facts : — 



(1) The uniformity of the parental breeds for a markedly 

 different average weight. 



(2) The uniformity of the Fl birds in weight. 



