Aug. 26,1918 
Variation and Correlation in Wheat 
371 
For height of tallest culm, average height of culms, and average length 
of spikes the means in 1914 are lower, but the standard deviations are as 
high or higher than the standard deviations for these characters in any 
of the three other years. The comparatively high variability in 1914 of 
average height of culms is indicated on the frequency distribution graph 
(fig. 1). The differences in centimeters between the height of tallest 
culm and average length of culms are 18.188 ±0.602 in 1914, i5.2i4±o.322 
in 1915, 6.35o±o.2i8 in 1916, and 4.28o±o.284 in 1917. The greater 
difference in height between the tallest culm and the average of the culms 
in 1914 indicates that the drouth in early July and the stemrust in late 
July and early August of that year prevented the secondary culms from 
approaching in height the main culm as closely as they did the three other 
years. This would tend to increase the variability of these two charac¬ 
ters as well as that of average length of spikes. After making due allow¬ 
ance for the abnormal conditions in 1914, it is of interest to note the com¬ 
paratively low variability of height of tallest culm as indicated by the 
standard deviations. 
For number and average weight of kernels and total yield of plants the 
means were lowest in 1914, but the standard deviations for these charac¬ 
ters in the same year were either next to the highest or equal to the 
highest. 
The mean for yield of kernels in 1914 was reduced materially, owing 
to the drought and black stemrust, and the standard deviation is also low. 
The mean for average weight of kernels per plant was highest in 1917, 
but it is accompanied by the lowest standard deviation in the 4-year 
peribd. This is indicated on the frequency distribution graph (fig. 4). 
For number of culms, total length of culms, total length of spikes, and 
weight of straw per plant the means for 1914 were either almost equaled 
or exceeded by those for 1915 and 1916, more favorable years; but the 
standard deviations are equaled or exceeded only by those for the 1915 
crop. 
In general, the standard deviations tend to follow the same order as 
the means, the variability being greatest where the means are the greatest, 
due in both instances to favorable conditions for development. Excep¬ 
tions to this tendency may be due in part to the frequent smaller differ¬ 
ences between standard deviations as judged by their probable errors 
compared with the differences between means as judged by their probable 
errors. Average weight of kernel had the highest mean in 1917 accom¬ 
panied by the lowest standard deviation, which is an exception. A 
number of exceptions occurred in 1914, owing to the very favorable 
condition for development during the first part and the opposite con¬ 
ditions during the latter part of the growing season. 
