Relation of Xonauailable Water to Hygroscopic Coefficient 45 
the semiarid soil was lowered by reducing the foliage of the 
plants, but the depth ol penetration of the roots was not thereby 
lessened. 
The final water content of the humid subsoil in this experi- 
ment bore no relation to either the wilting coefficient or the hy- 
groscopic coefficient. In the two cylinders of semiarid soil in 
which there was barely enough water to mature seed there was a 
direct dependence of the final water content upon the hygroscopic 
coefficient, the former being only a little above the latter. In the 
semiarid subsoil in which roots were well developed, the final 
content of free water was independent of the distance of the sub- 
soil from the surface, except in one cylinder, and in this there 
remained an abundance of free water even after the death of the 
plants. In the latter the free water content increased slightly 
with the distance from the surface. 
In all 6 of the cylinders which bore plants a hard crust de- 
veloped under the surface mulch but appeared to have no in- 
jurious effect on the plants. 
GROWTH OF CROP PLANTS WITH DIFFERENT AMOUNTS OF 
WATER IN THE SUBSOIL. EXPERIMENT OF 1910. 
The 22 metal cylinders used in this experiment were of the 
same material, form, and dimensions as those employed in 1909. 
After being filled, 2 were left unplanted, 6 were planted to Red 
Fife wheat, 6 to Kubanka wheat, 6 to milo, and 2 to Mexican 
pink beans. The 20 were planted on February 5, the two un- 
planted cylinders like the others being provided with a one-inch 
mulch of dry soil. All the cylinders, as soon as the mulch of air- 
dry soil had been added, were removed to the greenhouse and 
sunken in a rectangular pit four feet deep. The portions above 
the surface were enclosed by planks and banked up with soil as 
shown in figure 6, thus protecting all but the surface 3 to 6 
inches from the heat of the direct rays of the sun and the rapidly 
varying air temperatures. (When the photograph was taken 
some of the soil was removed from the sides so as to show the 
numbers on the cylinders.) 
Beginning near the middle of March a record was kept of the 
daily maximum and minimum temperatures and of the evapora- 
tion from a free water surface. A little later a record of the 
humidity was begun. While the humidity did not differ much 
from that in the open air, the maximum daily temperature on 
more than half the days exceeded 100° F. and the experiment 
was finally cut short by a temperature of 124° F. on June 12. 
(Table 4.) 
This temperature in itself would not have been fatal to the 
plants had the soil moisture not already been reduced to such a 
