28 THE INHERITANCE OF COLOR IN MICE. 



If these results are added to those obtained by Miss Durham in similar 

 matings we have: 



The result is not strikingly close to the expected numbers. There is in my 

 experiments a deficiency of pink-eyed animals too extensive to be overcome by 

 the excess of pink-eyed obtained by Miss Durham. The difference in result in 

 the two cases may be considered as probably too great to be due entirely to 

 chance.* Furthermore, in the following cross the result is still more aberrant 

 and again there is a lack of pink-eyed animals. The cross is that of heterozy- 

 gous dark-eyed animals (Pp) with pink-eyed animals (pp). Equality of dark- 

 eyed and pink-eyed young is expected. 



Here, if the higher numbers (gained by adding Miss Durham's results) are 

 considered, the numbers are closer to equality, though still widely unequal: 



It is possible that in the "pink-eyed" animals of my strain some constitu- 

 tional weakness exists, either directly connected with the "pink-eyed" modi- 

 fication or to a marked degree correlated with it, which tends to cause a higher 

 degree of mortality among the pink-eyed than among the dark-eyed young. I 

 have tried, so far as possible, to record the eye-color of young at birth, but have 

 failed in many cases. Even if it had been possible, it would not be safe to state 

 that animals with apparently unpigmented eyes were all pink-eyed colored 

 forms, for albinos at birth are indistinguishable from the pink-eyed colored 

 series. 



When "pink-eyed colored animals are crossed inter se, only "pink-eyed" 

 animals have been produced; 496 young have been thus obtained. 



♦Using the method described bjrEast & Hayes (1911) p. 37, we find that the deviation 

 from the expected ratio is 0.0810, while the probable error is =^0.0752. 



