294 F. B. SUMNER AND H. H. COLLINS 



white, and 'color' (in this case a yellow-orange) are likewise 

 shown in table 1.* 



These figures may be profitably compared with similar ones 

 giving results for a series of the normal wild gambeli race. Table 

 2 is based upon ten skins of adult individuals, five of which were 

 trapped in December and January, five in June. Care was 

 taken to select a representative series including lighter and 

 darker specimens.^ The conspicuously darker shade of even 

 the paler mice of the wild type is indicated by the higher per- 

 centages of black, while the comparative lack of color is also 

 obvious. The ratio of red to green on the other hand (see below), 

 is very close to that in the 'a' yellows, being 3.09, as compared 

 with 3.17. Thus the difference between these 'yellow' mice 

 and the normals appears to depend chiefly upon the relative 

 proportions of black and of yellow pigment, not upon the char- 

 acter of the latter. 



The 'b' strain of yellows (figs. 3, 4, and 8) 



A single pair of normally colored wild mice (P 9 15 and 

 P cf 59) in the cultures of the junior author became the parents 



* These last figures represent the proportional magnitudes of three sectors on a 

 color wheel which would combine to produce the shade in question, assuming that 

 the black disk was of zero luminosity, and the white equal to that of the standard, 

 and that the colored sector was of maximum saturation and intensity. As a 

 matter of fact, the commercial cardboard disks are very far from fulfilling these 

 conditions, so that large corrections (over 25 per cent in the case of 'white') must 

 be made, in order that color wheel and photometer determinations shall agree. 



For reasons which cannot here be discussed, the proportion of black is regarded 

 as equal to the difference between the highest color-screen reading and 100 per 

 cent, the proportion of white being equal to the lowest color-screen reading, and 

 the 'color' constituting the balance. In an earlier paper (Sumner, '21) the pro- 

 portion of black was computed differently, following the instructions contained 

 in a pamphlet issued by the manufacturers. The values thus obtained were 

 somewhat too high. The procedure here adopted has the authority of Mr. F. E. 

 Ives, the inventor of the instrument, and, furthermore, the figures thus arrived 

 at correspond fairly well with those obtained by the color-wheel method (due 

 corrections being made) . 



^ It is true that this series includes none as pale and as highly colored as 

 certain extreme variants of the 'normal' stock. At least one of these last (result- 

 ing from selective mating for two generations) resembles an average 'yellow' in 

 appearance, though undoubtedly quite different genetically. 



