RESEARCH METHODS IN STUDY OF FOREST ENVIRONMENT. 115 
It should be borne in mind that the ideal temperature conditions 
were not attained in this test, and, as has been stated, that at no time 
has complete saturation of the vapor existed, except possibly for 
short periods during cooling. This ideal has, however, probably been 
approached more closely than in any previous test, and the long 
period employed gives us a new insight into the phenomena of 
absorption. 
The following comments on Table 5 may assist in an understand- 
ing of these results. The comparative behavior of various soil com- 
binations will not be discussed, as these merely substantiate the ob- 
servations of others. 
1. The amount absorbed by the unwashed fine sand in 382 clays 
is only one-third more than the absorption in 5 days. It is, however, 
evident that even at the end of the longer period the unwashed sand 
was not in an atmosphere of saturated vapor, but rather in one whose 
pressure was quite as much controlled by the presence of soils still 
absorbing vapor, and particularly by the sample containing the 
largest amount of potassium nitrate. Assuming that all the moisture 
absorbed by the last entered into the salt solution, the latter would 
be a 22 per cent solution and would stand for an osmotic pressure of 
more than 45 atmospheres. It is therefore not surprising to find that 
a soil which contains not over 20 parts per million of soluble mat- 
ter should make little gain in this atmosphere. 
2. On the other hand, the amount absorbed by the fine sand in 382 
days is just about equal to the wilting coefficient for this sand, as 
nearly as can be estimated from a test on the original soil, which 
contained about equal proportions of material finer and coarser than 
the fine sand. 
3. The continued and relatively large absorption, especially by 
the soils containing active salts, might be ascribed to the low tem- 
peratures under which the test was conducted. It is believed, how- 
ever, that the evidence of a condition slowly approaching saturation 
vapor pressure, and never quite up to it, is convincing, and that this 
explains not only the present results but nearly all the phenomena 
that have been reported in a similar connection. 
A number of other tests somewhat similar to the above were made 
during 1918, but for short periods only. Several attempts were 
made to compare the osmotic pressures of soil samples in their 
natural moisture conditions, by placing the fresh samples under a 
single bell jar without a control solution, to note whether the samples 
gained or lost moisture in the common atmosphere. While these 
gains or losses indicated the relative dryness of the several samples, 
the tests were not continued long enough to produce any results of 
value. It was found that a period of two or three weeks was inada- 
quate to bring about equilibrium between the many samples, whose 
