i6o 
Journal of Agricultural Research 
Vol. XI. No. 4 
absorbed moisture to as high a point as was found in the smaller boxes 
at the end of 24 hours, while in the second even this was not a sufficiently 
long exposure. 
In the small absorption boxes we also tried using two tables, one placed 
1 to 2 inches above the surface of the water and the other 2 inches still 
higher. In this manner four trays, two on each table, could be placed 
in a single absorption box. In all cases the results were unsatisfactory, 
the soil in the trays on the lower table absorbing less moisture than those 
on the upper. 
CONCORDANCE OF DETERMINATIONS IN PRACTICE 
An illustration of the variation in the values found for a particular 
soil from day to day is afforded by a series of 30 samples of loess soils 
used in cylinder experiments (3, p. 78-115). On four consecutive days 
a single determination was made on each (Table X), using as controls 
two soils, H and K, with coefficients of 6.1 and 14.2, respectively. Three 
trays with the former and two with the latter were exposed each day. 
It will be observed that the values obtained for the different samples 
rose and fell together day by day and that the control soils indicate 
both the direction and in general the magnitude of this fluctuation. 
Both controls showed the highest values on the second day and the next 
highest on the first, while on both the third and the fourth they were 
slightly lower than on either of the previous days. Of the 30 other soils 
28 showed the highest values on the second day, 25 the next highest on 
the first, while 22 were third highest on the third day, and 14 on the 
fourth day. As the values obtained for the controls on the last two days 
were nearest those commonly found, we used the averages of the values 
found on these two days for the 30 other soils in estimating the amount 
of free water (3, p. 97). 
