258 
AGRICULTURE: ALWAY AND McDOLE 
considered it to be largely responsible for the crops of extensive areas 
which would be unproductive if entirely dependent upon the rain and 
snow which fall upon their surface, and Cameron^ sees in it the means 
of indefinitely maintaining the mineral nutrients of the surface soil, 
Rotmistrov^ regards it of no importance to those plants whose roots 
do not reach down to it. HalP in a recent analysis of the situation 
has stated: "The evidence on either side is far from being conclusive 
and more experiments are very desirable." 
The present differences in views appears to be due to the failure in 
laboratory experiments and field studies to take into consideration 
some physical constant that is directly related to both the lower limit 
of available moisture and the water-retaining-capacity of the soil, if 
we define the latter as the maximum amount which a soil will carry 
after it has been saturated and then, protected from both direct evapo- 
ration and the indirect effects of this as well as the action of plant roots, 
allowed to come into approximate moisture equihbrium by the down- 
ward movement of the excess of water into the subsoil mass. The lower 
limit of available moisture as determined by plant-house experiments, 
in which crop plants were grown in 6-foot cylinders and left unwatered 
until they matured or died of lack of water, appears to be practically 
coincident with the hygroscopic coefficient. Up to the present a method 
of estimating the water retentiveness in the field from one of the physi- 
cal constants of the soil has not been developed. The laboratory ex- 
periments and field studies of the authors make it appear that in the 
case of soils with hygroscopic coefficients between 14 and 3 this bears 
a rather simple relation to the hygroscopic coefficient and that in coarser 
soils, while it bears a much less simple relation, this is still one that may 
be experimentally determined. As the great majority of the tillable 
soils of dry-land regions fall within the limits of hygroscopicity men- 
tioned it would appear that through the determination of both the 
moisture content and the hygroscopic coefficient in the case of samples 
of the deeper subsoil we could learn both the percentage of the physio- 
logically important water and the departure of this from the maximum 
which the particular subsoil could retain. 
The investigation was carried out in 1912 and 1913, while we were at 
the Nebraska Agricultural Experiment Station, the field studies being 
conducted partly in the humid eastern portion of Nebraska and partly 
in the semi-arid western end of that state. 
The soils involved in the experiments included both surface soils and 
subsoils and ranged in texture from a Dune sand, with a hygroscopic 
coefficient of 0.6 and a moisture equivalent of 1.5, to a silt loam with 
