Oct. 22, I917 Direct Determination of the Hygroscopic Coefficient 163 
In the case of 30 sets, each consisting of 6 samples, hygroscopic- 
coefficient determinations had been made in duplicate on each. The sets 
had been taken from five fields at each of six different places, and to 
secure a single set of composite samples representing each area we mixed 
equal weights of soil from the corresponding depths of the five fields, 
making allowance for the unequal amounts of hygroscopic moisture at 
the time. Accordingly the average of the values for the various depths 
in each area represents the mean of 10 determinations. These data are 
given in the third column of Table XI. Later, samples of each of the 30 
composites were exposed singly on three consecutive days. Only the 
one control soil, A, was used; on the first day the average value for this 
was found slightly higher and on the third slightly lower than on the 
second day. Of the 30 other soils 2,7 showed the highest value on the 
first day and 22 the lowest on the third day. 
From the above the advantage of using samples of control soils will 
be evident. The mere concordance of determinations on different days 
and even by different operators is in itself no evidence that such data 
express the correct hygroscopic coefficients. Especially striking illus¬ 
trations of this were afforded by the data obtained using pasteboard 
trays described above—data that we later wholly discarded. The de¬ 
velopment of an indirect method for the determination of the hygroscopic 
coefficient from the content of hygroscopic moisture was suggested by 
the parallel rise and fall of the two control soils from day to day (4, p. 
348 - 359 )- 
reliability of the method described by hilgard 
To determine this, we followed his method exactly, except that we 
used a table of paraffined wood, which he does not specify. To give 
the method the most severe test we each day employed fresh and, accord¬ 
ingly, air-dry sheets of glazed paper. Four soils, two of them of high 
hygroscopicity, were employed, they being exposed in duplicate for 12 
hours and also for 24 hours (Table XII). During the whole of the time 
the temperature of the room was between 14 0 and 15 0 C. 
Table XII. —Results showing the reliability of Hilgard's method when carried out as he 
has described it 
Hygroscopic coefficient. 
Soil. 
On glazed paper. 
In 
aluminum 
12 hours' 
exposure. 
24 hours* 
exposure. 
trays 
(24 hours’ 
exposure). 
Control soil L. 
15-4 
19-5 
5 : 2 6 
15-7 
20. 0 
5 :l 
16.0 
Control soil I... 
2I 'I 
*6 
Control soil A. 
Control soil C.... 
