336 
Journal of Agricultural Research 
Vol. VIII, No. 9 
reduce evaporation, although air could circulate freely. The percentages 
of moisture were maintained by replacing the water lost at the end of 
each month. The soil was a sod and was pulverized and sieved before 
being placed in the jars. An analysis of the soil was made after it had 
been held under the treatment from January, 1912, to August, 1914. 
The results are given in figure 1, which shows the largest amount of 
total soluble salts in the soil kept at nearly 29 per cent moisture. As the 
soil moisture increased 
beyond this amount 
there was a sharp de¬ 
crease of soluble salts 
present. 
The nitrates varied 
much the same as did 
the total salts, the 
main difference being 
that the maximum 
quantity of nitrates 
was found with an 
initial soil-moisture 
content of about 24 
per cent. It must be 
remembered that the 
soil dried out some¬ 
what between water¬ 
ings; hence, the aver¬ 
age moisture would .be 
a little below this 
amount. The most 
rapid rise occurred 
practically at the same 
places for nitrates and 
total salts, and the 
two curves followed 
each other rather 
closely. This seems 
to indicate that the 
Should the nitrates in 
using the lowest per- 
Fig. i.—G raph showing the total soluble salts and nitrates in soil held 
in the laboratory, with different quantities of moisture for 2% years. 
rise in total salts is due largely to the nitrates, 
the soil be considered as NaNO s , the increase, 
centage as the basis of comparison, will account for nearly half of the 
increase in total salts. The ratio of total salts to nitrates decreases 
from 4.34 to 1 at 2.95 per cent moisture to 2.09 to 1 at 23.55 P er cent, 
and then rises to 4.31 to 1 at 48.3 per cent. 
These results indicate that in a soil of the nature used in this experi¬ 
ment a moisture content between 23 and 28 per cent is the most advan¬ 
tageous for liberating plant food, especially for the formation of nitrates. 
