THE CHEMICAL COMPOSITION OF SOIL COLLOIDS 13 
ISOLATION AND DESCRIPTION OF THE COLLOIDAL MATERIAL 
The method of isolating the colloidal material 6 (from the soils 
above described) was as follows: A suspension containing the col- 
loidal matter was obtained by shaking up the soil with water, gen- 
erally in the proportion of 1 to 5. In some cases it was necessary 
to use one part of ammonia to 3,000 parts of water to aid dispersion. 
After settling over night, the suspension was p'assed through a high- 
power centrifuge. The colloidal material passing through the cen- 
trifuge was concentrated to the consistency of a jelly by sucking off 
the water through Pasteur-Chamberland filters. The excess of 
water was then evaporated. Prepared in this manner, the colloid 
was not contaminated with an appreciable concentration of soluble 
salts from the evaporation of a large volume of the soil solution. 
The filtered solution from the colloidal suspensions contained soluble 
matter varying from 8 to 200 parts per million. 
The particles in the suspension passing through the centrifuge 
were practically invisible in the ordinary microscope. As examined 
with the ultra-microscope, the apparent size of the particles seldom 
exceeded 0.3 micron in diameter. The unaggregated or primary 
particles had a most active Brownian movement. When the col- 
loidal material was separated from the bulk of the solution it was 
in the form of a stiff, very sticky gel. This dried down, with great 
shrinkage, to a hard, horny mass which adhered strongly to the 
tongue and broke with a conchoidal fracture. The color of the 
different colloids, in powder form, varied greatlv, as shown in 
Table 6. 
In obtaining the colloidal material for the analyses given in Table 
3 only from 5 to 10 per cent of the total colloidal matter calculated 
to be present was extracted. It is shown in Table 5 that the samples 
analyzed are fairly representative of the composition of all the col- 
loidal matter it is possible to extract, with procedure followed. 
In a previous bulletin (17) the difficulty of extracting all the col- 
loidal material from the soil has been explained. With the methods 
which the writers employed they have been able to extract from cer- 
tain soils only about half of the total colloidal matter estimated to be 
present, even when a complete extraction was the object. 
It should be pointed out here that while 1 micron has been taken in 
previous publications as the size distinguishing a colloidal particle 
from a noncolloidal particle (7), the analyses given in Table 3, as 
stated in the description of the colloidal material, were made on 
samples consisting of particles of 0.3 micron or below in diameter. 
The amount of material between 0.3 micron and 1 micron is probably 
small ; hence there would probably be little difference in the compo- 
sition of fractions having 0.3 micron and 1 micron as the upper 
limit of particle size. 
6 Most of the colloidal materials analyzed in this work were prepared by M. S. Anderson, 
of this bureau. 
