TACTICS IN PIGMENT-CELL RESEARCH ,35.9 



melanin content. One final point should be made: on Dr. Russell's pigment- 

 volume scale, albino skin would have a zero value. In our case, however, suspensions 

 of albino skin give a residual turbidity of about 80 Klett units. (Since making these 

 remarks originally, we have learned that there are important physiologic differences 

 between the intense black homozygote (aaBB) and the slightly less pigmented, but 

 enzymically heterotic, heterozygote (aaBb). 398 These recently discovered differences 

 do not, however, contradict our less detailed data summarized here.) 



When the data are subjected to an analysis of variance, it turns out that the 

 greatest source of significant biological variation is due to differences in color genotype. 

 The next major source of biological variation is attributable to differences between 

 litters within genotypes. (See sample analysis in table 59 for tyrosinase activity.) 



Table 59 



Sample analysis of variance for tyrosinase activity 

 (Maximum rate of net oxygen consumption in microliters per hour) 



F, P stand for variance ratios and corresponding probabilities when "within litters" mean 

 square is used as error mean square. F', P' stand for variance ratio and corresponding probability 

 when "within genotypes" mean square is used as error mean square. The data analyzed here are 

 more extensive than those previously summarized. 396 



Significant variance ratios of between : within genotypes permit ranking of genotypic 

 means for all measured melanogenic attributes according to the methods developed by 

 Duncan, 293 - 294 who also described an admirably compact way of summarizing the 

 results of all possible tests of significant differences between means. (See examples by 

 Foster. 396 ) 



While many statements are implied in such compact summaries of numerical 

 ranking, I wish to make only a few additional points. First, black skin, which is 

 obviously much more heavily pigmented than albino skin, is hardly distinguishable 

 from albino skin with respect to rate of respiration upon incubation with tyrosine. 

 Second, the skins of pink-eyed, dilute-brown mice react very strongly with added 

 substrates, despite their very low level of natural melanin content. (See figures 47 and 

 48.) Thus there is no simple correspondence between assayed enzymatic activity and 

 natural content of melanin. Third, detailed comparisons regarding allelic substitu- 

 tions on different genetic backgrounds indicate some regular pattern of effects. For 

 example, in black versus brown contrasts, brown skin is more active than black toward 



