94 BULLETIN 1059, U. S. DEPARTMENT OF AGRICULTURE. 
Or. in other words, there is a probable error of less than 1 per cent 
in this average relationship. Single determinations, however, show 
a probable variation of 2.9 per cent of the wilting coefficient, as 
measured by this direct means. 
This work, though extremely thorough, was confined wholly to 
soils encounted in agricultural regions, and while these varied be- 
tween 1.6 to 57 per. cent moisture equivalent, they were undoubtedly 
more homogenous than forest soils in general, and lacked the compli- 
cating features of both rocks and large quantities of organic matter. 
It is not desired to suggest that, if this method were readily applicable 
to forest soils, and if experimental error both in wilting coefficient 
and moisture equivalent determination could be largely eliminated, 
the general relationship would be found different in the case of 
forest soils. Unfortunately, no one has made sufficient use of the 
moisture equivalent, in connection with wilting tests on forest species 
and forest soils, to determine whether the formula of Briggs and 
Shantz holds good. It is hardly to be doubted, however, that a for- 
mula must be worked out for each species, or the species of each 
general climatic region. Also, there is little doubt that occasional 
soils will be found in which, owing to exceptional alkalinity or 
acidity, the wilting coefficient is extremely high, and hence the 
formula breaks down. 
In connection with the capillary moisture determinations by 
Bates (105), data on corresponding moisture equivalents have also 
been given in Tables 2 and 3. These, as pointed out, were deter- 
mined on samples which had just passed through the capillarity 
tests. The 4 by 54 inch soil cans were placed in a machine of such 
speed and radius as to develop a centrifugal force of 100-gravity, 
the radius being computed to the center of the 5 -inch column of 
soil. Ordinarily, 30 minutes of revolution suffices to extract the free 
water susceptible to this force, but with a heavy clay an hour may 
be required. 11 
11 In order to show the importance of the time element, where such large masses of 
soil are being treated, and also to illustrate the very great difference between the water- 
holding powers of sand and clay, two samples were weighed repeatedly after short 
periods on the centrifugal machine The one sample consisted of very fine, thoroughly 
washed sand from granitic soils, the other entirely of silt and clay from innumerable 
sources, the clay probably not constituting over one-fourth of the whole mass. Both 
samples had previously been compacted by oentrifuging, so that the rapid loss of mois- 
ture in the first period can not be ascribed to loose structure. The test wasi somewhat 
complicated by a freezing atmosphere which, in fact, necessitated cessation before an 
end point for either soil was plainly reached. From a mass of soil of about 1,070 
grams in either case, the sand gave up in 80 minutes 276.3 grams of water, of which 
230.7 grams (84 per cent) was released in the first 2\ minutes of oentrifuging. The 
corresponding figures for the silt and clay were 76.2 grams, and 12.7 grams or 17 per 
cent. In the last 20 minutes of the 80-minute period the loss for the sand was 3.1 
grams and for the clay 12.8 grams. 
