446 
Journal of Agricultural Research 
Vol. VII, No. io 
under favorable conditions before these exact points of change can be 
determined. 
In the summer of 1912 nine large galvanized-iron pans, or tanks, 
5 H feet long by 1 foot wide by 3 inches deep were employed to give 
larger surfaces and deeper soil. The equivalent of 10 kgm. of dry Green¬ 
ville loam was put into each tank and puddled with excess water to 
firm the soil and smooth its surface. Enough water in addition to the' 
water already present—5.4 per cent—was added to make a series with 
5 per cent intervals from 5.4 to 35.4 per cent. These tanks were set on 
the floor of the laboratory and were weighed on Monday, Wednesday, 
and Friday of each week. The losses were made up at the time of 
weighing. This study ran from June 12 to July 15, a period of 34 days. 
The next year seven tanks were set up and run during the 81-day 
period from June 18 to September 6. This time 7.2 kgm. of Greenville 
loam, containing 2.6 per cent moisture, were used. Intervals of 5 per 
cent were again made, bringing the moisture content up to 7.6, 12.6, 
etc., as high as 37.6 per cent. Weighings were taken three times a 
week as in the first trial and water added to make up the evaporation 
loss. Table IV contains the results of the two trials. 
Table IV. Effect of initial percentage of soil moisture on evaporation from Greenville 
loam in galvanized-iron tanks feet long and I foot wide 
1912 
1913 
Average of 2 tests. 
Moisture. 
Loss in 34 
days. 
Moisture. 
Loss in 34 
days. 
Average 
moisture. 
Loss in iis 
days. 
Per cent. 
Gm. 
Per cent. 
Gm. 
Per cent. 
Gm. 
5 * 4 
3 , 744 
7. 6 
10, 710 
6 * 5 
14 , 454 
10. 4 
7,575 
9.518 
12. 6 
17 , 530 
n* 5 
25,105 
IS* 4 
17. 6 
19,275 
16. 5 
28,793 
20. 4 
9, 200 
22. 6 
20, 395 
21. 5 
29 ,595 
25*4 
10,128 
27. 6 
20,300 
19,870 
26. 5 
30, 428 
30 * 4 
10,340 
32. 6 
3 1 * 5 
30,210 
35*4 
11,045 
37 * 5 
19,990 
36 . 5 
31,035 
Table IV shows that the loss was rapid to 16.5 per cent and then slow. 
Of course, the wide intervals prevented locating exact points, but in 
general this test corroborates rather closely those already reported for 
loam. 
SAND, CLAY, AND MUCK 
During the period from December 26,1913, to January 3,1914, similar 
tests were made with sand, clay, and muck. One hundred gm., dry weight, 
of sand and clay and 50 gm. of dry muck were used in the same kind of 
tin plates as those used in the other trials. All percentages were run in 
triplicate, the sand for each percentage from 1 to 33, the clay for each 
percentage from 1 to 55, and the muck for each 20 per cent interval from 
