358 
Journal of Agricultural Research 
Vol. VII, No. 8 
with samples of the same soils dried at ioo° C. and still others that 
were in an air-dried condition. At the end of from 3 to 7 days those 
of the second and third sets were alike in moisture content, while those 
of the first were distinctly moister. A difference was frequently shown 
even after an exposure of 21 days. 
Using 5-, 10-, 15-, and 20-gm. samples of the same three control soils 
we compared paper with aluminum trays. Both gave the same results 
with H and La, but in some cases S, which had a coefficient of 22.0, gave 
a somewhat lower result on the paper trays, 7.7 as compared with 8.0, as 
though, following an increase in the relative humidity of the air, the 
material of the trays, being in itself hygroscopic, had competed with 
the contained soil for moisture from the atmosphere. 
From the above statements it would appear that once having accu¬ 
rately determined the hygroscopic coefficient of suitable control soils 
it would only be necessary to expose for a few weeks on trays in a closed 
room thin layers of the air-dried or oven-dried soils under investigation, 
together with a sufficient number of trays containing the control soils. 
After sufficient exposure the mere determination of the hygroscopic 
moisture in all would permit of the calculation of the hygroscopic coeffi¬ 
cients. Thus, the hundreds or thousands of samples which might be 
collected during the summer in connection with a soil survey might be 
placed in trays on the shelves as they reached the laboratory, the drying 
being left until winter. There could then be obtained that single¬ 
valued factor expressing texture which appears to us the most desirable 
of all those so far proposed (1, p. 214). 
In connection with the method there remains to be determined, among 
other things, the minimum time of exposure necessary, the most suitable 
material for trays, the desirability of providing for the agitation of the 
air in the exposure room and maintaining the humidity of the air in the 
room between definite limits. 
This indirect method of estimating the hygroscopic coefficient appears 
to give more reliable results than those to be obtained by the use of a single 
formula applied to either the mechanical composition (2, p. 411) or the 
moisture equivalent (3, p. 842), while at the same time requiring only the 
simplest equipment, as well as a minimum of skill on the part of the 
operator, and being economical of time. 
SUMMARY 
The estimation of the hygroscopic coefficient from the maximum water 
capacity, while with many soils giving values in full accord with those 
directly determined, with so many soils gives such erroneous results that 
it is to be regarded as too unreliable for use in connection with either 
studies of available soil moisture or for soil-survey purposes. 
From the studies reported it appears that the hygroscopic coefficient 
may be calculated from the hygroscopic moisture found in a soil which 
