390 
Journal of Agricultural Research 
Vol. XIV, No. 9, 
When weight of seed sown is correlated with plant characters at 
maturity, it is noticeable that in 1915 there are four coefficients with 
the minus sign and that there is a tendency for the coefficients to be 
lower than in 1914. 
In contrast with the low and varying relation in 1914 and 1915 is the 
generally moderate and consistent correlation between weight of seed 
sown and plant characters in 1916 and 1917 when the plants were grown 
on the poorer soil and from somewhat lower mean weight of seed. 
From this study conclusive evidence is given that for the conditions 
under which the work was done, environment reduced radically or 
obliterated entirely the correlation between weight of seed sown and 
plant characters among which is yield. 
This information answers, in part at least, the questions raised in the 
introduction to this article regarding the r 61 e of weather and soil in 
comparisons of heavy and light seed for planting. 
If these results were applicable to the wheat crop in general during 
the 4-year period, it is clear that on soils of moderately high productivity 
with favorable weather conditions heavy kernels as compared with 
light kernels used for planting may be expected to give very moderate 
or no increase in yield. 
In the study of the interrelation of plant characters a substantial and 
fairly consistent correlation was found between yield of kernels and 
average weight of kernels, average height of culms, and a somewhat 
higher correlation with number of culms. Between average height of 
culms and average weight of kernels there is a moderately high corre¬ 
lation each year, except in 1915, when the coefficient is very low. 
If these relations held for the wheat crop during the 4-year period, 
separating from the crop each year seed of higher average weight would 
be selecting seed from plants which had a decided tendency toward 
higher yield, and, with the exception of the year 1915, from plants 
which were taller and at the same time higher yielding. In 1915 there 
was practically no relation between average weight of kernels and 
average height of culms, and separating the larger seeds from this crop 
would be selecting seed from both high and low yielding plants. 
The tendency of the tallest plants and the plants having the greatest 
number of culms to be the highest yielders is a valuable index in making 
individual plant selections from mixed populations. 
SUMMARY OF CONCLUSIONS 
Subject to the environmental conditions under which the work was 
done, the following conclusions may be drawn: 
(1) The magnitude of the means for any of the characters studied 
varied in response to environmental conditions. Tower yields of straw 
resulted from a reduction in number, total length, or average length 
of culms per plant; and lower yields of grain from a reduction in the 
