Vol. 6, 1920 
GENETICS: S. WRIGHT 
325 
to determine the true potentiality of each area at thejcritical period in 
development. Let us assume that these potentialities are distributed, 
at least roughly, according to the normal probability curve. 
In figure 3 variation in the piebald potentiality is measured along the 
horizontal axis. Curve A is intended to represent the number of areas 
with each potentiality in one guinea-pig, while curve B represents the same 
for another guinea-pig in which the tendency toward color production is 
higher as a whole. The critical potentiality is at X. All areas which ex- 
ceed this produce color, while those which fall below, however slightly, 
remain white. Guinea-pig A is thus about 93% white, while B is about 
50% white. It is easy to see that any cause which increases the potentialit}' 
FIG. 
A normal probability curve, the area of which is divided into tenths with the exception 
of the end areas which are twentieths. The deviations from the middle of the curve 
to the middle of each area (mean of the end areas), measured in terms of the standard 
deviation, are used as a corrected series of grades in place of the actual percentages 
of white in piebald guinea-pigs. 
of A and 5 by a given amount makes a greater change in the area"of color 
in B than in A, From a table of probability integrals, one can find at 
once the change in the area of the curve, brought about by a unit change 
in the value of X, in the neighborhood of any given area. These results 
should be comparable with those obtained empirically for physiologically 
equivalent variations at different areas of white. 
The theoretical curve, as shown by the dotted line in figure 2, does not 
fit the empirical curve perfectly, but appears to give a sufficiently accurate 
correction of the grades for practical purposes. Aside from its greater 
smoothness, the theoretical curve has the advantage that by its use the 
variability in the amount of color in a stock of guinea-pigs can be compared 
