16 HEREDITY IN RABBITS, RATS, AND MICE. 



(text-figure 1, W); (6) 7 (text-figure 1, Fi); (c) 7 (text-figure 4, Fi); 

 and (d) 1 or 2 (text-figure 2, Fi). That is, this cross should, under the 

 hypothesis considered, produce a trimodal figure, with a mode for 

 25 per cent of the individuals at either end of the grading scale and 

 with a still larger mode (for 50 per cent of all individuals) at an inter- 

 mediate point. But what is actually observed is very different. The 

 figure (text-figure 5, Fi X W) is not trimodal but bimodal, and neither 

 of the modes is where the hypothesis demands that modes should be. 

 This is conclusive evidence against the correctness of the hypothesis 

 in question, but is entirely in harmony with the alternative one, that 

 dark and tan are allelomorphs but segregate in modified form, one 

 on the whole darker, the other on the whole lighter than before they 

 were crossed with each other. 



This matter of modification on crossing is one deserving further 

 consideration. It is in evidence in all the crosses made. It is clearest 

 where the races crossed differ most in grade, and it seems to consist 

 in a partial obliteration of those differences. Thus the uncrossed 

 white race has its mode at 17, the self race at 0. (See text-figure 2.) 

 In F2 the extracted whites contain no individual as high in grade as 

 17, and the highest mode lies at 12 to 14, facts which indicate that 

 white has been lowered in grade by its association with self in the Fi 

 zygotes. The back-cross of Fi with white also shows modification but 

 intermediate in amount, as might be expected from the fact that, in 

 the case of each zygote formed, only one of the two conjugating 

 gametes had been subjected previously to modifying influences, viz, 

 that one which was furnished by the Fi parent. The mode in this 

 case lies at 15, instead of at 17, as in the uncrossed race, or at 12 to 14, 

 as in the F2 extracted whites. 



That the extracted white has less influence in whitening an Fi zygote 

 than the uncrossed white is shown further by a comparison of Fi self X 

 white (text-figure 2) with the back-cross of Fi with self. Fi had its 

 mode at grade 1 and ranged upward to grade 3, but contained no self 

 individuals. If extracted white were identical with uncrossed white in 

 its whitening influence, then in the back-cross with self half the zygotes 

 should be of grade 1 or higher. But in the observed back-cross less 

 than one-third of the zygotes show any white, viz, 29 out of 94, and these 

 are lower in grade than the original Fi's, viz, 1.10 as compared with 1.51. 



The self character also emerges modified after the cross with white. 

 For the Fi zygotes, consisting of pure self united with pure white 

 (text-figure 2, Fi, S X W) were all close to self in grade, with a mode 

 on grade 1 and ranging upward only to grade 3. But the zygotes 

 formed by extracted self united with pure white produced in the back- 

 cross of Fi with white (text-figure 2, Fi X W) range in grade from 1 to 

 9 -with a broad low mode at 4 to 7. Evidently they have been much 

 increased in grade in the direction of white. 



