308 F. B. SUMNER AND H. H. COLLINS 



36 normals (wild type) 

 12 yellows 

 12 pallids 



4 double recessives. 

 Since the peculiarities of both the yellows and the pallids de- 

 pend upon the absence of pigment, and since the former do not 

 lack any pigment which is not likewise lacking in the latter, it 

 would seem probable that the double recessives would be 

 indistinguishable from ordinary pallids. The proportions above 

 stated would accordingly become 36:12:16. 

 The actual numbers^ ^ proved to be: 

 36 normals 



6 yellows 



7 doubtful (normals or yellows?) 

 15 pallids. 



The exact agreement of the first of these figures and the almost 

 exact agreement of the last are sufficiently striking. If it were 

 permissible to regard the 'doubtful' specimens as being geneti- 

 cally 'yellows,' a single further transposition would bring these 

 figures into exact conformity with 'expectation.' Unfortunately, 

 the case is not as simple as this. 



The thirty-six 'normal' mice, while differing among themselves, 

 presented no greater range of variability than would a similar 

 random collection of wild gambeli, which race, indeed, they re- 

 sembled pretty closely in appearance. In this respect, they 

 agreed with their Fi parents. Despite evident differences of 

 color, likewise, the pallid specimens probably showed no greater 

 range of variation than was to be found in the original pallid 

 stock. The 'yellow' class in the list includes only those concern- 

 ing which no doubt was felt. While none of the pelages were as 

 rich in color as the brightest of the '6' strain of yellows, they were 

 probably an average lot. 



The difficulty here relates to the status of the 'doubtful' 

 specimens, intermediate in appearance between yellows and 

 normals. One of these died early, so that its status could not 



18 The classification of individuals was made without any reference to these 

 totals, which were not computed until later. 



