76 PROCEEDINGS OF THE AMERICAN ACADEMY. 



expected. The deviations from the INIendelian ratios ma}', therefore, be 

 due simply to the rehitively small numbers of young obtained. 



Cueuot ( : 02) in similar experiments procured a much larger number 

 of mice, and his figures are much more nearly in accord with the theo- 

 retical expectation. By breeding together the F^ heterozygotes, he 

 obtained 198 pigmented mice and 72 albinos, or 2G.6 per cent albinos. 

 By back-crossing the heterozygotes with albinos, Cuunot obtained au 

 a[tproximate equality of pigmented and of albino mice, but the j)recise 

 numbers are not given. 



The pigmented animals obtained by interbreeding the Fi heterozygotes 

 should be of two sorts according to INIendelian principles, the one of pure 

 dominants, the other of heterozygotes ; and these two classes should be 

 to each other numerically as 1 : 2. No actual test of this numerical ratio 

 has as yet been made iu the case of mice. To make such a test, it is 

 necessary to breed each of the F2 pigmented individuals to an albino or 

 to a heterozygote (containing the albino character recessive), and a suf- 

 ficient number of young must be so obtained to determine whether or not 

 the particular individual is producing germ cells having the albino char- 

 acter. It is obvious that the cross with an albino is preferable, in this 

 case, to the cross with an individual known to be heretozygous. For the 

 albino, if bred to an animal containing the albino character recessive, will 

 (by formula 6) produce 50 per cent of albinos, whereas from two hetero- 

 zygous animals, only half as many are to be expected. The writer has 

 tested 10 of the pigmented mice of generation Fo by breeding them to 

 albinos. Five of these individuals produced only pigmented young, 8, 1 1, 

 9, 5, and 11 respectively. We may therefore conclude that they were 

 in all probability of the class D, or "extracted" (Bateson) dominants. 

 Each of the five other animals of generation Fo produced both pigmented 

 and albino offspring when bred to albinos. They were, then, of the class 

 D(R). The totals of their litters are 23 pigmented, 26 albino, offspring. 

 It is therefore plain that, qualitatively, the Mendelian expectation is 

 realized, in the case of generation F.,, where the three classes, D, D(R), 

 and R should be present in the proportion 1:2:1. The fact that the 

 ten mice of classes D and D(R), although selected at random, were 5 of 

 them of class D and 5 of class D(R), instead of being one-third D and 

 two-thirds D(R), may be not without significance in this particular case, 

 wherein there is some slight evidence of irregularity. Yet the result 

 may be purely a chance occurrence, for the numbers are small. 



It is of interest to examine the separate litters making the above total 

 of 23 pigmented and 26 albino young (see Table B). 



