RESEARCH METHODS IX STUDY OF FOREST EXTIROXIUEXT. 87 
sary to subtract 21 from the percentage of capillary moisture to 
obtain a quantity having a fairly constant ratio to the wilting co- 
efficient. 
This is given by S^^S^ =2.90±0.06, or± 2.01 per cent. ' 
' T\ lltmg Coeff. 
The probable error of a single determination by this means was 
found to be 8.3 per cent of the wilting coefficient. 
In the treatment of forest soils Bates (105), at the Fremont Ex- 
periment Station, has found it necessary to use much larger caiis 
than those employed by Hilgard, and has also reversed the process, 
so that the result is rather a measure of the ability of the soil to 
hold the water of saturation than to lift water from below. A gal- 
vanized can 5^ inches deep and 1 inches in diameter, is filled to a 
depth of 5 inches with air-dried soil, which is jarred and tamped 
until no appreciable settling occurs. This can is perforated in the 
bottom and a filter paper is used to keep the soil from sifting out. 
The can is immersed to its full depth in water, but no water is al- 
lowed to flow on the top of the soil. As the water rises from the 
bottom by its own pressure, the air is pushed out, so that few, if 
any, air spaces are left. The samples are allowed to soak at least 
21 hours to insure complete absorption by the larger, permeable rock 
fragments. 
The weight attained at the end of this period, or a longer period 
if it appears necessary, is an index to the saturation capacity. 
The cans are now placed on a drain board, covered, and allowed to 
stand for 48 hours. In rehandling the cans care must be used to 
avoid jarring, as some of the water is held in a very delicate balance. 
The amount of water held at this time is a measure of the capillary 
moisture. In the vast majority of soils that have been treated, the 
capillary moisture is about 90 per cent of the saturation capacity. 
Clay does not affect this ratio appreciably, but humus increases it. 
The same cans are now used for the centrifugal test or moisture 
equivalent determination, which will shortly be described. After this 
they are oven-dried, to give the basis for dry- weight calculations. 
The apparent density is also computed from the weight and volume 
after this treatment. 
In Table 2, there is presented a comparison of the capillary mois- 
tures and wilting coefficients of 10 soils of one general type (granitic) 
from an Engelmann spruce forest, but varying widely in state of 
decomposition, clay content, and humus content. Each soil repre- 
sents a sample extending from the surface to a depth of 1 foot. The 
wilting coefficients for Douglas fir and Engelmann spruce were most 
carefully determined, the only objection that might be brought 
against the treatment being that the seedlings were given more direct 
