RESEARCH METHODS IX STUDY OF FOREST ENVIRONMENT. 113 



osmotic pressure varies inversely as the moisture content. The 

 latter is probably true only within rather narrow limits. 



Further, since obtaining an equilibrium by vapor transfer is a 

 slow process, it is desirable to be able to treat a large number of soil 

 samples simultaneously. For this purpose a large-size bell jar, 

 resting upon plate glass, may be employed. A pressure cooker, the 

 cover joint being properly sealed, has also been found very useful. 

 There may be one or more vessels of the solution, and as many ves- 

 sels of soil as are desired, within the chamber. Using 2£-inch soil 

 cans for soil containers and similar beakers for solutions, about 

 60 soil samples may be treated at once under a 14 by 12 inch bell 

 jar. However, a great deal of evidence shows the desirability of 

 a relatively large vessel for the control solutions, and a decrease 

 in the number of soil samples. 



It should be remembered that this treatment will only give the 

 osmotic pressure for each soil at one particular moisture content, 

 which will depend upon the total amount of water in all of the 

 samples, as well as upon the concentration of the control solution 

 at the initiation of the test. To obtain a range of values for any 

 soil, separate tests must be made under varying conditions. 



The fundamental provision in such a test is that the vapor cham- 

 ber should have a constant temperature and be evenly heated on 

 all sides. To accomplish this most simply a deep excavation in the 

 ground is desirable. The very gradual seasonal change of tempera- 

 ture in such a situation will not work any harm, since the vapor 

 pressure within the chamber will adjust itself to such a change with- 

 out necessitating any condensation. In the lack of this, a dark 

 cellar may be chosen. 



In proof of the theory that a soil solution would continue to ab- 

 sorb water indefinitely in the presence of saturated water vapor, 

 and that, therefore, the hygroscopic coefficient as now known is a 

 purely empiric quantity, a test has been conducted for slightly more 

 than a year with clean sand and various modifications thereof which 

 represent the different elements encountered in various types of 

 soil. The sands and modified sands were all palced in the vapor 

 chamber in an oven-dry state, together with a bottle of distilled 

 water with linen wicks having about 10 square inches evaporating 

 surface. The successive weighings of samples and bottle indicate 

 that some vapor is constantly escaping from the chamber, so that 

 the vapor therein is never absolutely saturated. Early in the test 

 it was found impossible to heat the chamber as evenly as desired 

 in the cellar in which it was placed, so that it was removed for a 

 time to an electrically heated incubator. Here the fluctuations, 

 though small, were rapid and a slight overheating at one time caused 

 a very severe loss from all samples, the chamber being unable to 

 i < 1 1 < ;: 5 — 22 — i ; 1 1 1 i i or>n a 



