Journal of A gricultural Research 
Vol. XXI, No. 8 
5 x 6 
Table VI .—Constants involved in the calculation of correlation coefficient of mean 
breaking point and protein content of California wheats 
Mean. 
Standard deviation. 
Correlation 
coefficient (r). 
Crushing point. 
Protein con¬ 
tent. 
Crushing point. 
Protein con¬ 
tent. 
Gm. 
631 
±5.9176 
Per cent. 
9.89 
±0. 1194 
Gm. 
89. 90 
±4- 1853 
Per cent. 
I. 8058 
±0. 0841 
O. OI 11 
±0. 061 
The results corroborate the Kansas data, with a correlation coefficient 
of practically zero, and a probable error larger than the correlation 
coefficient. 
In the Kentucky experiments, conducted by Harper and Peters (2 ), the 
hardness of the kernels was determined by means of a piston with a 
Fig. 2.—Machine used at the Kentucky Agricultural Experiment Station for testing the hardness of 
wheat, designed by J. N. Harper (2). 
cutting edge, which is pressed down upon the kernel by means of weights 
added directly from above (fig. 2). The kernel is cut in two and is not 
crushed. 
The opinion of the Kentucky investigators regarding the apparatus 
used is as follows: 
This device for testing the hardness of wheats appears to us to be superior to those 
that measure the pressure required to crush a kernel. 
The data obtained are presented in Table VIII. 
