214 Hum bert. 
Width 
250—13 U, V, W, X, Y, and Z (averaged) 5.588 20.08 
250—I3a 32501 10.79 
Number of Branches 
250—13 U, V, W, X, Y, and Z (averaged) 1.932 10.96 
250—13 a 1.414 7.00 
Number of Seed Pods 
250—13 U, V, W X, Y, and Z (averaged) 28.092 33.02 
250—13 a 50.520 27.50 
We see that for each character for each line studied the coefficient 
of variability is greater in those lines from the Mitchell Farm (poor 
soil) than in those lines from the Plant Breeding Garden (good soil). 
Using the standard deviation as the index of variability the results 
are almost as striking though not quite so consistent. The standard 
deviation is greater for three of the characters out of four, in the 
lines grown on the Mitchell Farm plots. This is true for each pair 
used in comparison. The standard deviation for the character “number 
of seed pods”, however, is greater in each case for the lines grown 
upon the Garden plot. 
It should be pointed out that the lines grown upon the Mitchell 
Farm plot contained a much smaller number of variates than the 
lines grown upon the Garden plot. This is shown in tables No. 4, 
No. 5, No. 6, and No.7. From this cause alone the Mitchell Farm 
lines would have somewhat smaller standard deviations. Without 
putting too much weight upon this fact, we may at least say that 
it reduces the difference which is in favor of greater variability in 
case of good soil as shown by the character “number of seed pods”. 
In nearly all of the comparisons made, then, the constants for 
the poor soil lines are very much greater than the corresponding 
constants for the good soil lines. This is contrary to the generally 
accepted idea that increase in nutrition is the principle cause for 
increase in variability. This idea was first expressed as a law by 
Knight!). It was adhered to by early writers and by Chas. Darwin?) 
who states that “of all the causes which induce variability, excess 
of food, whether or not changed in nature, is probably the most 
powerful.” Within the last few years definitely planned experiments 
have been reported which give evidence, one way or the other upon 
1) Knight, T. A. Treaties on the Culture of the Apple. 
2) Darwin Chas. Animals and Plants under Domestication. Vol, II page 236. 
