ESTIMATION OF COLLOIDAL MATERIAL IX SOILS. 7 
•differing only in size. The uncertainty of such separations and the 
unsatisfactonness of the results obtained are due to the presence of 
colloidal material which it is difficult to disperse into discrete particles 
of any definite size. A determination of the colloidal material in 
soils is, therefore, needed to supplement the results afforded by any 
system of mechanical analysis based on size alone, which does not 
disperse colloidal aggregates (Davis, 7). 4 
ADSORPTION METHOD OF ESTIMATING COLLOIDAL MATERIAL IN 
SOILS. 
DESCRIPTION OF METHOD. 
The earlier investigators in estimating the colloidal material in 
soils actually isolated at least a part of the colloids, but they did not 
utilize the marked adsorptive properties of the material as a means 
of characterization nor as a test for complete separation. Later 
investigators recognized high adsorptive capacity as a distinguishing 
characteristic of the colloidal material, but failed to isolate the 
material. Knowing the adsorptive capacity of the kind of colloidal 
material present in a given soil and the adsorptive capacity of the 
whole soil, one should be able to estimate the quantity of colloidal 
material present. If the adsorptive power of the soil is due wholly 
to the colloidal material, the ratio, 
adsorption per gram of soil 1 nn 
adsorption per gram of colloid 
should give the percentage of colloidal material in the soil. 
Moore, Fry, and Middleton, of this Bureau {22), suggested this 
method of estimating the colloidal material in soils. They deter- 
mined the adsorption ratio in the case of one soil, measuring the 
quantities of malachite green and ammonia gas adsorbed by the soil 
and by a sample of colloidal material extracted from the soil. The 
adsorption ratio obtained by malachite green agreed exactly with the 
ratio obtained by ammonia. 
l>i It might seem that adsorption by noncolloidal soil particles would 
constitute a fundamental error in this method, making the ratio of 
the adsorptive capacities of soil and colloid indicate a higher per- 
centage of colloid in the soil than was actually present. However, a 
subsequent investigation (Anderson, 1) showed that in the case of 
ordinary soils more than 95 per cent of the total adsorptive capacity 
of the soil for malachite green, water, and ammonia is located in the 
colloidal constituents. Only in the case of highly micaceous soils 
is the adsorption by noncolloidal particles significant. 
Since the adsorption method of determining the colloidal content of 
a soil would not, as a rule, be seriously affected by noncolloidal 
constituents, it seemed advisable to try the method on a number of 
different soils, 32 in all. It also seemed advisable to determine the 
adsorptive capacities of the soils and colloids for water vapor as well 
as for malachite green and ammonia. Theoretically, any substance 
which is not appreciably adsorbed or acted upon by the noncolloidal 
soil particles, but which is adsorbed by the colloidal material, might 
be used for determining the colloidal content of the soil by this method. 
4 It is doubtful if the method of mechanical analysis proposed by Sven Oden (23, 24) attains as complete 
a deflocculation of the colloidal material as the older method of Williams, since the residues apparently 
are not "worked" to the same extent. 
