68 
Journal of Agricultural Research 
Vol. IX, No. a 
completely the precipitation. The deep samplings for moisture deter¬ 
minations should be sufficiently numerous, frequent, and deep to afford 
a reliable history of the moisture content in the moist lower subsoil, in the 
central dry zone, and in the intermittently moistened surface layer. To 
illustrate the difficulty of securing a satisfactory site for such a field study, 
it might be mentioned that in no part of Minnesota would it appear 
feasible, either the climate being too humid, the water table too near the 
surface, or, lastly, the subsoil being too shallow where free of rock frag¬ 
ments and, where deep enough, carrying too many such fragments to 
permit satisfactory sampling. 
SUMMARY 
Uniform columns ot soil of known hygroscopic coefficient and moisture 
equivalent were employed in various laboratory experiments, the 13 
soils used ranging in texture from a coarse sand to a silt loam with hygro¬ 
scopic coefficients of 0.6 and 13:3, respectively. 
Five of the loams, placed in capillary connection with the natural 
subsoil mass, saturated with water and allowed to stand protected from 
surface evaporation for several months, lost water until the amount 
retained bore a close relation to the hygroscopic coefficient, being from 
2.1 to 3.1 times this value, according to the particular soil. When a 
layer of coarse sand or gravel separated the column of loam from the 
natural subsoil mass or interrupted it, the downward movement of the 
water in the soil above this layer was much delayed. Where the col¬ 
umn consisted of successive 2-inch layers of loams differing widely in 
texture, the order of their arrangement exerted no influence upon their 
final water content. 
Soil columns 30 to 36 inches long, while protected from all loss of 
of moisture at the sides and bottom, were freely exposed to evaporation 
at the surface for periods varying from a few weeks to half a year. The 
moisture content, originally uniform and lying between 2.0 and 3.0 
times the hygroscopic coefficient, fell until it reached, at depths below 
the first foot, an almost constant minimum with the ratio 1.9 to 2.2. 
Employing 2-foot columns of 12 different loams, each with an initial 
moisture content approximately equal to its hygroscopic coefficient, 
enough water was added to raise the average moisture content of the 
column to 1.5 times the hygroscopic coefficient, the water being applied 
in one experiment to the top and in another to the base of the column. 
After the cylinders had stood for three or four months fully protected 
from evaporation the distribution of moisture, with regard to the sur¬ 
face to which it had been applied, was found to be the same in both 
experiments. The maximum distance through which an effect was 
shown was about 2 feet, but in most cases much less. The maximum 
final ratio of moisture content to hygroscopic coefficient was found in 
