BovieE: NON-AVAILABLE WATER IN SOILS 281 
culture case, where a relative humidity of approximately 10 per 
cent. was maintained. The culture case, as described in another 
paper,” contained a thermograph, a hygrograph, and an electric 
fan. The humidity was controlled by regulating the volume of a 
stream of air, dried by passing through sulphuric acid, and forced 
into the case by means of an air pump. 
Wilting the plants in a relatively dry atmosphere, practically 
solved the problem of knowing when to consider them wilted. 
These plants did not show temporary partial recoveries from the 
drouth effects, as others did which were grown under fluctuating 
conditions of humidity. Moreover, the leaves dried out imme- 
diately after death, so that it was possible to see the progress of 
death over the various members of the plant and to have each cul- 
ture stopped at the same stage of wilting—a very important detail 
in non-available waten determinations. A study to determine the 
effects of relative humidity on non-available water is now in prog- 
ress, but the results of this study are not necessary for the present 
paper, since in the experiments here reported, all of the plants 
were under like conditions of humidity. 
As soon as the plants had wilted, the tumblers were removed 
from the case, and the amount of non-available water determined 
in the following manner: The quartz was passed through a screen, 
having 20 meshes to the inch, which removed all the roots. A 
sample of about 80 grams of the quartz was put into an aluminum 
dish and weighed on a chemical balance.* The sample was then 
dried at 110° C. for one week. It was found that this amount of 
desiccation was sufficient to insure constant weight. 
From the loss of weight in drying, the percentage of non-avail- 
able water was determined, then the total water remaining in the 
soil was calculated, also the concentration which this water would 
have had, had it existed as free water,—assuming that the plants 
had used up none of the salts and neglecting any water of crystal- 
lization. The results of the determinations are given in the fol- 
lowing table: 
*These operations were carried on in the laboratory and probably permitted some 
experimental error to enter. However, as little time as possible was consumed before 
the weighings were made. The sifting was done in a small closed vessel. 
he numbers refer to the amount of salts added; thus 25a means that a 0.25 per 
cent. solution of sodium chloride and full nutrient salts were added to the quartz. 
Sets from a to e contain sodium chloride, sets from / to p contain no sodium chloride. 
