934 
Journal of Agricultural Research 
Vol. XXVIII, No. 
in accord with evidence of alteration which was obtained in the case of adsorp¬ 
tion studies where water vapor and ammonia were the substances adsorbed (4). 
It is not clear just how the process of extraction influences the heat of wetting 
of the colloid. It may be influenced to some extent at least by the structure 
of the aggregates of the colloidal material. The state of aggregation or structure 
of the particles in the colloidal material probably has considerable effect upon 
its heat of wetting and other physical properties. The same colloidal material 
will doubtless vary in its active surface according to whether the primary par¬ 
ticles are present in loose or in compact aggregations. The arrangement of the 
colloidal particles is no doubt changed through the process of dispersion, subse¬ 
quent concentration and drying. It is reasonable to assume that such changes 
as these are largely responsible for the alteration in heat of wetting values and 
may be responsible for the alteration in adsorptive capacity for ammonia or 
water. 
In estimating the colloid content of a soil from the ratio, 
heat of wetting of soil 
heat of wetting of colloid’ 
the value should be too high, provided alteration were the only source of in¬ 
accuracy. It has been shown, however, that the extracted colloids usually give 
higher heat of wetting values than the so-called “unextractable” portions. 
Furthermore, a small portion of the colloid extracted from a soil tends to give 
a higher heat of wetting than a larger sample (Table III). The error due to 
the quantity of colloid sample extracted and the error due to alteration, are, 
therefore, in opposite directions. If both errors were of the same magnitude 
the heat of wetting ratio should give a true expression of the quantity of colloid 
in a soil. Inasmuch as the sampling error tends to be less the larger the sample 
and the alteration error is probably greater the larger the sample, it is presumable 
that some size of sample from each particular soil would give a correct value for 
colloid content by the heat of wetting ratio. It is apparent that in case of very 
large samples (Table V) the alteration error was more than sufficient to balance 
the sampling error. 
Ammonia adsorption studies indicate an alteration in adsorbing power due to 
extraction and also a sampling error similar to that found for heat of wetting U). 
The colloid content of soils shown by heat of wetting ratios and by ammonia 
ratios (Table II) are as a rule in good agreement. Although complete con¬ 
firmation is lacking, it seems probable that the small samples of colloid used in 
these ratios give errors so nearly compensating that the results usually approxi¬ 
mate the true quantities of colloid present in the soils. 
The heat of wetting method and the ammonia adsorption method for deter¬ 
mining the colloidal content of soils appear to be of about the same value as 
regards practicability. Both require the extraction of a sample of the colloidal 
material, which involves considerable time as well as expense for apparatus. 
They do not appear to be as practicable as, and there is no evidence of their 
being more accurate than, the water adsorption method when the adsorption 
is conducted over 3 per cent sulphuric acid and a factor of 0.3 gm. of water per 
gram of colloid is taken as the constant value for soil colloids. 
SUMMARY 
The heats of wetting of soils and of colloidal material isolated from soils are 
given and the possibility of determining the colloidal content of soils from such 
data is discussed. 
Heats of wetting of the colloidal materials from different soils vary widely. 
In some cases they are comparable with the heat values for starch and for synl 
thetic inorganic gels. 
Practically all of the heat of wetting of a soil appears to be due to colloidal 
material. 
