TESTING WHEAT FOR PROTEIN 21 
Experience in the chemical research laboratory of the grain divi- 
sion has been that the smaller the sample ground, the larger the mois- 
ture loss. In the experiment listed in Table 14, sample 11998 con- 
tained 10.34 per cent moisture when 8 grams were ground, and 11.41, 
11.26, 11.58, and 12.20 per cent, respectively, when 15, 30, 60, and 125 
grams were ground. Sample 12002 contained 9.25, 10.29, 10.63, 
11.82 and 12.34 per cent moisture when the same quantities of seed 
were ground; and sample 12005 had 9.52, 10.73, 10.35, 10.85, and 
11.42 per cent moisture under the same conditions. 
Moist or wet wheat is not only difficult to grind but will lose 
a greater percentage of moisture during grinding than dry wheat. 
Questions have been frequently asked as to how great this loss is 
and how the protein results would differ if they were reported on 
a moisture test, such as the Brown-Duvel 12 test; also how other 
results would vary from the Brown-Duvel test if different methods 
of making moisture tests were used. 
To obtain data on this point, 20 samples of wet wheat were 
obtained and moisture determinations made by the following meth- 
ods : (1) By the Brown-Duvel method; (2) by heating the whole 
grain for 120 hours under atmospheric pressure at a temperature 
between 99 and 100° C. ; (3) after grinding, by the vacuum-oven 
method for a period of five hours at a temperature of 99.5° C. and at 
a pressure of 4 inches of mercury; (4) after grinding, by the air- 
oven method under atmospheric pressure and at a temperature of 
108° C. for a period of five hours; (5) after grinding, by the water- 
oven method, under atmospheric conditions at a temperature of 
99.5° C. for a period of five hours. 
By the Brown-Duvel method the average moisture content of the 
samples was 14.9 per cent. Drying the whole seed to constant weight 
in the water oven gave an average of 15.01 per cent moisture. On 
the ground seed by the vacuum-oven method the moisture content 
was 14.17 per cent; by the electric-oven method, 13.90 per cent; and 
by the water-oven method, 13.68 per cent. The maximum difference 
by all methods was 1.39 per cent. Some of this can be attributed to 
the loss of moisture due to grinding. Necessarily, if a uniform 
method is not used for making moisture tests, uniform protein 
results can not be obtained. 
For each per cent change in the moisture content of the sample 
there will be a corresponding change, equivalent to 1 per cent of the 
protein content originally present. This change in protein content 
will be in the inverse order to the change in moisture content. That 
is to say, if a sample of wheat contains 12 per cent of protein at a 
moisture content of 13 per cent and dries to 12 per cent, the protein 
content on the new moisture basis will be 12.12 per cent. Tables for 
converting protein values at any given moisture content to a unit 
basis of 13.5 per cent moisture will be found in Miscellaneous Circu- 
lar No. 28, United States Department of Agriculture. 
Lack of a uniform method for making moisture tests is further 
illustrated by the moisture results obtained by 9 of the protein test- 
ing laboratories on the 12 samples of wheat submitted for protein 
tests. The .data which are given in Table 15 show a range of 1.46 
12 See footnote 4- 
