380 
Journal of Agricultural Research 
Vol. XIV, No. 9 
that between those for height of tallest culm. Therefore it is evident 
that on soil of medium productivity, the more favorable weather condi¬ 
tions in 1917 as compared with that in 1916 did not influence correlation 
to any marked extent. 
In 1914 and 1915 the coefficients for all plant characters at maturity 
are comparatively low; and, with the exception of those for number of 
culms, yield of kernels, 
- /A/ DSC/G&AMS an d average weight of 
kernels, they are sig¬ 
nificantly lower in 1915 
than in 1914. This 
difference is due to 
the highly favorable 
environmental condi¬ 
tions in 1915. 
Considered as a 
whole, there is a dis¬ 
tinct tendency toward 
correlation between 
weight of seed sown 
and the characters of 
the resultant plants. 
However, the correla¬ 
tion, even under av¬ 
erage conditions, is 
stance, and is subject 
to the influence of en¬ 
vironmental condi¬ 
tions to so marked an 
extent that with some 
characters the relation 
may be obliterated en¬ 
tirely; and with other 
characters, including 
yield, may be made 
so slight that under ordinary conditions of experiment no relation could 
be detected. 
INTERRELATION OF PLANT CHARACTERS 
not high in 
any m- 
Fig. 7. —Graph showing regression for weight of seed and yield of 
kernels per wheat plant in 1916. 
The coefficients given in Table XIII show that only in the group where 
yield of kernels is correlated with other characters is there a general 
tendency toward less correlation in 1914 and 1915 than in 1916 and 1917. 
This condition although much less marked, is similar to that found when 
weight of seed was correlated with plant characters at maturity and is 
