Oct. 22,11,17 Direct Determination 0} the Hygroscopic Coefficient 151 
that the conditions of their experiments were not such as to insure that 
the exposed soils became saturated with hygroscopic moisture. The 
work of Patten and Gallagher (25), confirming that of the earlier investi¬ 
gators mentioned, has unfortunately had the effect of leading to the gen¬ 
erally accepted view that, in his conclusions, Hilgard was in error (20, 
P- 36 7). 
WORK OF UPMAN AND SHARP 
As a reply to the criticisms of Patten and Gallagher, Upman and Sharp 
(18) report a study of the effect of the thickness of the layer of exposed 
soil and of the temperature of exposure, using boxes 12 by 18 by 19 
inches, similar to those employed by Hilgard in his later work (15, p. 198), 
but longer and wider than those mentioned in the above quotation (14, 
p. 76). They also used glazed paper placed on wooden tables with the 
tops only 1 inch above the surface of the water. They compared three 
depths (1.5,4, and 8.5 mm.) of an adobe soil exposed for various intervals, 
ranging from 1.5 to 455 hours, and at temperatures from 12 0 to 34.75 °C. 
With two other soils the periods of exposure were quite similar, but the 
range in temperature was less, while the thickness of the soil layer did not 
exceed 2 mm. 
Tipman and Sharp found in general that a greater amount of moisture 
was absorbed at the higher temperatures, and emphasize the necessity of 
using very thin layers of soil, having found about 1 mm. the best. They 
conclude that 8 to 10 hours’ exposure of air-dried soils is sufficient when 
the depth of the layer is only 1 mm. and the boxes are of the ditneticmn s 
they used. They call attention to the difficulty of securing the saturation 
of the air at any considerable distance from the water surface. 
These authors mention that duplicate samples agreed remarkably 
well. However, if they were run side by side, this agreement, as pointed 
out below, does not serve as evidence of the accuracy of the determination. 
EXPERIMENTAL WORK 
Two of us had previously (1, p. 590; 2; 5, p. 277-281) employed the 
method on a very limited scale, using an improvised place for the absorp¬ 
tion boxes. We found the loose sheets of glazed paper so inconvenient 
that we introduced a modification that greatly facilitated the work, 
substituting light pasteboard trays 7 inches long, 5 inches wide, and 0.75 
inch high, lined with glazed paper. On starting the determinations at 
the University of Nebraska we again used this modification, having had 
a large supply of the pasteboard trays lined with the glazed paper manu¬ 
factured by a local box company. As these one by one became somewhat 
limp from exposure to the damp atmosphere maintained in the absorp¬ 
tion boxes, they were discarded and new ones substituted. 
We used wooden boxes of the dimensions given by Hilgard, Titling both 
the boxes and the covers with blotting paper as specified. To prevent 
the warping of the sides and cover, it was found desirable first thoroughly 
