38 BULLETIN 1193, U. S. DEPARTMENT OF AGRICULTURE. 
SUMMARY. 
The noncolloidal soil constituents are defined in this paper as the 
distinctively mineral particles above 1 micron in diameter. The fol- 
lowing are classified as colloidal constituents : All the organic matter, 
inorganic material dispersing into particles below 1 micron in diam- 
eter, and aggregates of particles less than 1 micron in diameter. 
The earliest method of determining the colloidal material in soils 
consisted in actually separating the material from the rest of the soil. 
The probable error in this method is the difficulty of obtaining a 
complete deflocculation or peptization of the colloidal material. This 
error is enormous in some rapid methods recently suggested. Ad- 
sorption methods previously proposed probably do not indicate in 
most cases even the relative colloidal contents of soils, since they do 
not include a determination of the adsorptive capacity of the kind of 
colloidal material present. 
The adsorption method of determining the colloidal content of soils 
investigated in this paper involves determinations of the adsorptive 
capacities both of the soil and of colloidal material extracted from 
the soil. The ratio of the two adsorptive capacities multiplied by 
100 is taken as giving the percentage of colloidal material in the soil. 
Tests of this method on 32 soils are reported, the percentages of collo- 
idal material being calculated from the adsorptions of malachite 
green, water, and ammonia. 
The colloidal materials extracted from different soils vary greatly, 
in their adsorptive capacities for malachite green and ammonia, but 
are much more constant in their adsorptive capacity for water. 
In the case of 13 soils the percentages of colloid calculated from 
the adsorptions of malachite green, water, and ammonia agree; in the 
case of 14 soils the percentages indicated by the adsorptions of two of 
the substances agree but disagree with the percentage indicated by 
the adsorption of the third substance; and in the case of 5 soils 
different percentages of colloid are indicated by the adsorption of all 
three substances, malachite green, water, and ammonia. 
The two most probable sources of error in the adsorption method of 
determining the colloidal content of a soil appear to be the difficulty 
of extracting a sample of colloidal material which is exactly repre- 
sentative in adsorptive capacity of all the colloidal material in the 
soil, and the possibility of a change taking place in the adsorptive 
capacity of the colloid on extraction. These two possible sources 
of error are investigated in the case of several soils. 
The adsorptive capacity of a small sample of extracted colloidal 
material differs by about 10 per cent from the adsorptive capacity of 
all the colloidal material which is extract able by the methods em- 
ployed. However, the colloidal material which can not be extracted 
has, in many soils, a much lower adsorptive capacity than that which 
is extractable. 
Evidence is presented which indicates that the colloidal material 
has a somewhat different adsorptive capacity after extraction than 
it has in the untreated soil. The data indicate that the colloidal 
material of most soils in its extracted condition is slightly more 
adsorptive of malachite green, slightly less adsorptive of water vapor 
and considerably less adsorptive of ammonia than in its condition 
in the untreated soil. 
