oct. sa. 1917 Indirect Effects of Selections in Breeding Indian Corn 121 
figures 1 to 4 and as the standard deviations have the smaller probable 
errors, we very naturally inquire for a plausible explanation of these 
results. Shall we conclude that the mean value is more stable and char¬ 
acterizes the strain better than the standard deviation does ? To be sure, 
we often give a mean value to characterize a type; but in accord with 
what is stated above about these four strains of com, the ratio of their 
differences of means to the probable errors of such differences are in 
general much greater than the ratio of the differences of their standard 
deviations to corresponding probable errors. In more definite form it is 
found that the ratio of the average difference of means in a crop to the 
average of the.corresponding probable errors varies from 12 in the case 
of length of ears to 23 in the case of circumference of ears. On the other 
hand, the ratio of average differences in standard deviations in a crop to 
the average of the corresponding probable errors varies from 3.5 in the 
Fig. 8.—Graphs showing the standard deviation of the rows of kernels on the ears in the four Illinois strains 
of Indian com for n crops, i9°5"i9!5- 
case of rows of kernels to 7 in the case of weight of ears. Incidentally, 
in the latter case, we find that the standard deviation is strikingly higher 
for the low-protein and the low-oil than for the high-protein and the 
high-oil strains, and that this difference persists without exception for the 
11 crops. We should recognize, therefore, that the standard deviations 
of the four strains for a given year differ much less when compared with 
their probable errors than do the corresponding means. Hence, we 
should expect much less tendency to maintain an order of values of 
standard deviations than of means. If, on the other hand, the mean 
values were close together, and the standard deviations were significantly 
different, we could characterize strains better by standard deviations 
than by means. 
