July is, 1921 
Relation of Hardiness to Protein Content of Wheat 519 
with the percentage of protein. It seems that if any correlation exists, 
those varieties having the lowest percentage of kernels cut—the hardest 
varieties—should have the highest percentages of protein, and vice versa. 
The results as calculated by the writer are given in Table IX. The 
correlation table is given in Table X. 
Here also the correlation between the hardness of the wheat kernels 
and the protein content is not significant. 
Table X .—Correlation between hardness of grain ( subject) and protein content ( relative) 
in iq varieties of Kentucky-grown wheat « 
a First three columns at left denote, respectively, classes of percentages of flinty kernels cut under 
pressure of 4 pounds (1,818 gin.), mid points of classes, and deviations from the mean. The three lines in 
the boxes at the top denote, respectively, classes of protein percentages, mid points of classes, and deviations 
from the mean. 
Table XI. — Coefficient of correlation constants 
Number of 
strains. 
Mean. 
Standard deviation. 
Correlation 
coefficient. 
Specific 
gravity. 
Protein 
content. 
Volume. 
Specific 
gravity. 
Protein 
content. 
Volume. 
Between 
specific 
gravity 
and 
protein 
content. 
Between 
volume 
and 
protein 
content. 
67. 
I.32 
± O. 003 
Per cent. 
II. 67 
±0. 062 
II. 67 
±0. 059 
Mm. 
O. 04 
±0. 002 
Per cent. 
754 
±0. 062 
73 
db o- 042 
M tn. 
68. 
20. 60 
db O. 082 
I. 0089 
i°* 057 
0. 0=587 
dbo. 081 
0 
The correlation coefficient between hardness and protein content prov¬ 
ing to be practically zero, a study was then made to determine the corre¬ 
lation between specific gravity and protein content and between kernel 
volume and protein content of the Kansas pure strains. The correla- 
