444 
Journal of Agricultural Research 
Vol. VII, No. 10 
with water. The soil was then set aside to dry. When the condition 
of dryness was approached, the moisture was made up to the desired 
content. Twelve percentages—5, 10, 12^, 15, 20, 25, 30, 35, 40, 
45, and 50—were run in triplicate. The pans were set on a table in a 
large laboratory room. Weighings were taken daily to the nearest 
one-tenth of 1 gm. and water was added to make up the loss. This 
experiment continued for 42 days, at the end of which time the losses 
were computed for each week and for the entire period. Table III gives 
the losses for each period. * 
Table III .—Effect of initial quantity of water on evaporation 
Soil mois¬ 
ture. 
Water evap- j 
orated. ] 
Soil mois¬ 
ture. 
Water evap¬ 
orated. 
Per cent. 
5 , 
7 X 
10 
1 
is 
20 
Gm. 
96.4 
229. 2 
3 ^ 3 * ° 
484. 8 
594 ° 
822. 2 
Per cent. 
25 
30 
35 
40 
45 
5 ° 
Gm. 
956-1 
h 013-3 
X, 048. I 
1,074. 5 
1,148. 1 
1,185. 8 
-An examination of Table III shows that evaporation increased rapidly 
to about 20 per cent, less rapidly from 20 to 30 per cent, and slowly 
from 30 to 50 per cent. 
Later Experiments 
Various later studies were made with loam, sand, clay, and muck.. 
The methods were similar to those used in the preliminary work, but 
the experiments were much more detailed and thorough. Most of the 
work was done in small tin plates and copper vessels 8 inches in diam¬ 
eter and 4 inches deep, though some tests were made in Petri dishes, 
some in long galvanized-iron tanks and some in deep cylindrical gal¬ 
vanized cans. 
LOAM 
The test with loam was conducted in the manner already described 
with 100 gm. of dry soil in tin plates. The percentages of moisture ran 
from hygroscopic water, which was about 1.8 per cent, to 50 per cent. 
The experiment was conducted for 13 days, from August 29 to Septem¬ 
ber 12, the pans being made up to the original weight each day and 
shifted on the table to eliminate, in part, the influence of air currents. 
A similar test was made during the winter in a steam-heated laboratory. 
. All percentages from 1 to 40 were run in triplicate with 1 per cent inter¬ 
vals. Weighings were made each day and losses were considerably 
greater than those for the same number of days in the other trial. A 
possible explanation for this is that the laboratory under artificial heat 
