ESTIMATION OF COLLOIDAL MATERIAL IN SOILS. 33 
colloid and non colloid at a size of 1 micron should be checked by 
microscopic observation even if the same kind of centrifuge as that 
described on page 20 is used for the purpose. The dispersed col- 
loidal material may be conveniently collected and concentrated as 
described on page 8. The adsorptive capacities of the colloid and 
soil for water vapor may then be determined as described on page 10. 
According to the data given in Table 10, the ratio of the two adsorp- 
tive capacities should, without correction, give very nearly the true 
amount of colloidal material (particles less than 1 micron in diameter) 
in the soil. A further check on the results may be obtained by deter- 
mining the adsorptive capacities of the soil and colloid for dye and 
ammonia. The dye and particularly the ammonia ratio should, 
however, be corrected for alteration in the adsorptive capacity of 
the colloid produced by extraction. To get this corrective factor as 
calculated on page 28 is a laborious task; it is necessary to determine 
the adsorptive capacities of the residues left after extraction of the 
colloidal material, and the quantities of soil residues and of colloid 
separated must be determined accurately. If a marked disagree- 
ment should occur in the corrected ratios, it would presumably be 
due to the unextractable colloid being particularly low in adsorptive 
capacity as compared with the colloid extracted, and in this case the 
ammonia ratio would probably be the highest and nearest the true 
value. 
The method of estimating the colloidal content by actual extrac- 
tion of a large part of the colloidal material and determination of the 
remainder unextracted by observation with the petrographic micro- 
scope can also be used, but considerable familiarity with microscopi- 
cal work of this character is necessary to secure accurate results and 
extraction of much of the colloidal material is a long operation. 
The adsorption method of determining soil colloids as proposed is 
obviously not accurate in the sense that most chemical analyses are, 
but it is plainly much more accurate than any method of mechanical 
analysis that has yet been proposed, in view of the fact that the 
quantities of colloidal matter which we were able to isolate by an 
especially long continued treatment of the soil (Tables 3 and 4) were 
only 40 to 60 per cent of the total quantities present (Table 10). 
This paper is concerned chiefly with an examination of the adsorp- 
tion method of estimating colloids, but the quantities of colloid 
shown by this method to be present in the more important soil types 
are of interest as enlarging our general knowledge of soils. This is 
discussed in the following section. 
QUANTITY OF COLLOIDAL MATERIAL IN SOILS. 
It seems to have been a quite general opinion that soils contain 
only a small quantity of colloidal material. Geclroiz (14) holds that 
soils contain less than 0.02 per cent of material which will go into 
colloidal suspension on treatment with water (1 part soil to 5 parts 
water). 15 Schloesing (34) and recently Ehrenberg (8, p. 99) state 
that soils almost never contain more than 2 per cent of colloids. 
Lvon (19), Russell (29), and Wiegner (38) do not state the amounts 
of colloidal material that may be expected in soils of different char- 
15 Gedroiz makes a distinction between the gel material in the soil and the colloidal material obtained in 
the water extract. 
