ALLEN. — THE HEREDITY OF COAT COLOR IN MICE. 81 



Male lieterozygote, No. 51, gave six litters by albinos, and in four of 

 these the exact Meudelian proportion, 1 : 1, is maintained with a total of 

 ten young. In the remaining two litters three pigmented young are 

 obtained to one albino. The total excess of pigmented over albino mice 

 in this case is thus but two in a total of 18 young. A third heterozygote, 

 9 gray 208, produced three litters, in two of which there was an excess 

 of pigmented young, while in the third there was an excess of albinos, a 

 total of eight pigmented to four albino young. Two other animals pro- 

 duced each two lots of young ; in the case of one there is Mendelian 

 equality in each litter, while in the case of the other, a shortage of albinos 

 in one litter is exactly offset by an excess in the second litter. The re- 

 maining cases show a similar state of affairs, for a single large or small 

 litter may contain equal uuml)ers of pigmented and of albino young, 

 while subsequent lots show an excess of the one or the other sort in vary- 

 ing degrees, amounting in some cases to a total exclusion of the pig- 

 mented or of the albino class, and this too in litters of as many as three 

 or four young. The most extreme case obtained was one in which there 

 was but one albino in a litter of seven mice, the expectation being 

 equality of albino and pigmented individuals. 



It is evident from this table that there is a considerable variation in 

 the relative numbers of gametes of each class brought by the heterozygote 

 to each union. Frequently there is an exact numerical equality of the 

 two classes in separate litters, as is theoretically expected on Mendel's 

 assumption that a heterozygote produces gametes of the two sorts in 

 equal numbers. In other cases the gametes of such an individual pro- 

 duced at one time, or at least all which unite with those of another ani-. 

 mal to produce young, may be entirely of the one class or the other, while 

 between these two extremes are found various degrees of variation in 

 the relative proportions of the two sorts. This variation appears to be 

 quite independent of sex. The numerous deviations from the 1 : 1 ratio 

 tend to counterbalance one another, often to a remarkable degree, so 

 that as -the numbers obtained become larger, the approximation to the 

 theoretical ratio becomes constantly closer, and in the long run we may 

 say that the heterozygote does produce equal numbers of gametes of the 

 two classes considered. 



In order to show that Mendelian ratios are due simply to the chance 

 combinations of gametes, the following experiment, suggested by Profes- 

 sor Castle, was tried. Equal measured quantities of small white beans 

 were placed in two bags of the same shape and size. One-half the beans 

 in each bag were marked with a spot of red ink, and the other half were 



VOL. XL. — 6 



