34 2 Journal of Agricultural Research voi. viii, No. 9 
In general, the nitrates in the first 3 feet vary inversely with the water 
applied. 
The moisture content of the fallow plots receiving manure at the rate 
of 15 tons to the acre averages highest, the unmanured slightly lower, 
mainly in the third foot, and those on which 5 tons were applied consider¬ 
ably below the two others, especially for the larger irrigations. 
With 15 tons of manure the average nitrate content of the soil was 
highest, and with no manure lowest, these differences occurring mostly 
with 5 and 10 inches of water. The nitrates of the manured plots show 
a rather marked rise when a 5-inch application of water had been used, 
while more than this amount causes a decrease. A tendency for a steady 
decrease of nitrates with increasing percentages of water was not noticed 
in the fallow plots as it was in the cropped. 
£2Percent Water 
■ P R M. NaNOj 
Manure 
Applied 
Water 
Applied 
1 
1 i ll II11II 11 
None|5Tons|l5T£)n5 
None |5Tons|i5Tons 
■ 5 Inches 
10 Inches 
^Inches 
100 
80 
60 
4oX 
u 
a 
d 
CL 
Fig. 7.—Graph showing the soil moisture and nitrates on 
to which different quantities of water 
July 21, 1913, to a depth of 10 feet in cropped soil 
and manure had been added. 
Comparing cropped and fallow plots, we find that comparatively more 
of the nitrates are used by the plants on the 15-ton-of-manure soil than 
on the 5-ton, and more on the 5-ton than on the unmanured. For the 3 
feet shown the difference in nitrates between fallow and cropped soil 
decreases with increasing depth. Except for the unirrigated soil the 
nitrates, as shown at this part of the season, are influenced much more 
by the crop than is the percentage water. The cropped soil contained 
on the average approximately 15.3 p. p. m. of nitrates, expressed as 
sodium nitrate (NaN 0 3 )., and the fallow a little more than double this 
quantity. 
SUMMER OF 1913 
On July 21, 1913, the plots were again sampled, this time to a depth of 
10 feet. The moisture and nitrates for all cropped plots are shown in 
figure 7, and for all fallow plots in figure 8. These results show that 5 
