324 
GENETICS: S. WRIGHT 
Proc. N. a. S 
It is clear that measures of variability have little value unless the grades 
are properly corrected. Two methods have been tried for making this 
correction, one empirical, the other theoretical. The distribution of grades 
of white was tabulated, males and females separately, in the 17 families 
which were still on hand during 1916 and 1917. With a properly corrected 
system of grades the variability should be approximately equal in all cases. 
The points were found which divided each distribution into four equal 
parts, i.e., the median and two quartiles. The distance between the 
median and each quartile was then correlated with the grade midway be- 
tween the two points. The average of the values which fall within one 
grade is shown in the solid line in figure 2. It will be seen that a given 
variation near the middle of the range corresponds to one about half as 
great near the ends. 
An effect of this kind is to be expected on theoretical grounds. Although 
the skin of a piebald guinea-pig is divided sharply into areas in which pig- 
A B 
93 % White 50 % White 
FIG. 3 
The'~two curves are intended to represent the distribution of color potentialities among 
areas in the skins of two guinea-pigs. X is the critical potentiaUty, above which 
color is developed. The same change in average potentiality increases the amount 
of color in the guinea-pig with a large amount of white (A) only half as much as in 
the guinea-pig which is half white and half colored (B). 
ment is either produced to the full amount characteristic of the animal, 
or is wholly absent, it is not to be supposed that the influences which at 
some critical period in ontogeny determine whether a region is to be colored 
or white, are so sharply alternative in themselves. It seems more reason- 
able to suppose that the sums of the favorable and unfavorable influences 
in different parts of the skin could be arranged in a graded series. Doubt- 
less in certain white regions a slight difference in the conditions would 
have enabled color to develop, while in others, a great change would have 
been necessary. Similarly with colored areas. Suppose, then, that the 
skin is divided into a large number of equal areas and that it were possible 
