326 
GENETICS: S. WRIGHT 
Proc. N. a. S 
with the theoretical variabiUty in the potentiaHties within a single average 
guinea-pig. 
The relative importance of unit variations in area in different parts 
of the range is as the reciprocal of the variations which are physiologically 
equivalent. In figure 4 the area of a probability curve is divided into equal 
tenths except for the end areas, which are twentieths. The difference in 
potentiality corresponding to any difference in area can easily be estimated. 
The corrected series of grades as given in figure 4 has been made by finding 
the average of the deviations from the median of the curve to each of the 
class limits, with the exception that the deviation of the mean has been 
used for the end classes. A little consideration will show that the proper 
centering of the grades varies slightly with the mean of the population 
with which one is dealing. The grades as corrected here, are, however, 
sufficiently accurate for the present purpose. 
TABIvE 1 
Numbers, Means and Standard Deviations in Random-Bred Stock and Inbred 
Family No. 35 during 1916, 1917 and 1918 
EXP. B 
(random-bred) 
FAMII^Y 35 (inbred) 
No. 
Mean 
No. 
Mean 
cr 
Sire 
105 
-1-0.337 
± 0.050 
0.756 
=b 
0.035 
73 
—0.071 ± 0.044 
0.559 
± 
0.031 
Dam 
107 
-h0.390 
=t 0.054 
0.824 
=fc 
0.038 
73 
+0.220 ± 0.045 
0.570 
0.032 
Son 
498 
-H0.273 
± 0.024 
0.808 
=t 
0.017 
235 
—0.017 ± 0.026 
0.593 
=fc 
0.019 
Daughter 
488 
+0.543 
± 0.024 
0.796 
d= 
0.017 
236 
+0.140 ^ 0.027 
0.613 
=i= 
0.019 
Total Off- 
spring 
986 
+0.407 
± 0.017 
0.802 
=t 
0.012 
471 
+0.062 ± 0.019 
0.603 
=i= 
0.013 
Compare with figure 4 for conversion of grades into percentages of white. The 
standard deviation of 0.603 in the inbred family means about 22% of the total dorsal 
area of the coat. 
Two stocks were chosen for study of the relative importance of heredity 
and environment. One of these was the control stock {B) in which matings 
were made at random, except that even the mating of second cousins was 
avoided. The other was family 35, which since 1915 has been descended 
entirely from a single mating in the seventh generation of inbreeding and 
largely from a single mating in the twelfth generation. The records for 
1916, 1917 and 1918 were used in both cases. Table 1 gives the number 
in each experiment, the means and standard deviations. Table 2 gives 
the correlations between parent and offspring, between litter mates and 
between the parents, weighted by the number of young. Inspection of 
the tables shows that sex makes no significant difference in the correlations. 
This is in agreement with the results of direct experiments. Reciprocal 
crosses between inbred families at the opposite extremes, as families 39 
and 13 in figure 1, produced practically the same result, the progeny in 
