154 
Journal of Agricultural Research 
Vol. XI, No. 4 
In the case of each of three soils, we exposed two trays of the soil in 
each of three absorption boxes. From each box we transferred one 
sample to a weighing bottle as quickly as possible, this requiring from 23 
to 30 seconds. The other was transferred from a tray to a weighing 
bottle and back again three times, so as to consume 150 seconds in the 
operation. With each soil an appreciable loss of moisture is observable 
(Table II), but this is so small and the time purposely consumed in the 
one case so much longer than would be needed by any operator that it 
appears that no serious error from this source need be feared if the operator 
simply makes the transfer as speedily as he can. 
Table II. —Influence of speed of transfer of soil from tray to weighing bottle upon amount 
of hygroscopic moisture found 
Soil. 
Hygroscopic moisture. 
Transferred 
once, 
requiring 20 
to 30 seconds. 
Transferred 
three times, 
consuming 
150 seconds. 
C. 
0 - 5 
4.8 
18.7 
0-3 
4. 2 
16. 6 
d .:. 
E 
SUITABILITY of trays of various materials 
Using four control soils, A, C, F, and G, we compared trays of aluminum, 
copper, zinc, tin plate, glass, vulcanized rubber, granite ware, and those 
of pasteboard described above. Part of the last had been saturated with 
paraffin before being lined with glazed paper. In some, paraffined paste¬ 
board formed the sides, but the bottom was cut out, thus making a por¬ 
tion of the glazed paper lining the bottom of the tray. The metal trays 
were 7 inches long, 5 inches wide, and 0.75 inch deep. The trays of glass, 
granite ware, and vulcanized rubber were photographers' trays, having 
almost the same dimensions as the preceding. To distribute the thin 
layer of dry soil in the trays, an aluminum salt shaker was found very 
convenient. 
In order to determine the approximate maximum amount of moisture 
that was likely to condense on the layers of soil independent of its internal 
surface, we exposed coarse quartz sand of 0.5 to 1 mm. diameter in 
duplicate in all the different kinds of trays, both for 16- and for 38-hour 
periods. The amount of absorbed moisture was independent of both the 
kind of tray and the time of exposure, varying from 0.02 to 0.14 per 
cent, with an average of 0.06, an amount quite negligible in connection 
with soil-moisture studies. 
