ESTIMATION OF COLLOIDAL, MATERIAL IN SOILS. 17 
The residue left after decantation of the suspended material was 
repeatedly treated in the same manner as the original soil until it 
was apparent that no more material was to be dispersed by these 
means. This residue contained all the larger soil particles and con- 
stituted the " coarse fraction" discussed later. The decanted sus- 
pensions were run through the supercentrifuge and the colloidal 
material passing through was concentrated as previously described 
on pages 8 and 9. The material remaining in the bowl of the super- 
centrifuge was recovered and kept separate from the fraction of coarser 
soil particles. It was repeatedly rubbed and treated with water 
in the same way as the coarser material until it was apparent that no 
more material capable of passing the supercentrifuge was to be ob- 
tained. The residue left, called "fine" fraction or residue, con- 
tained chiefly particles of 1 to 50 microns diameter, although the 
material was not rigidly graded. By keeping the coarser and finer 
particles separate in this way, isolation of the colloidal material was 
facilitated considerably. It should be borne in mind that the fine 
and coarse fractions do not represent definite classes of soil material 
as the colloid fraction does. After the finer residues had become 
pretty well exhausted of extract able material, boiling the residues 
lor 6 to 12 hours was resorted to in some cases. This treatment, 
however, did not seem to increase the quantity of colloid obtainable, 
although it probably did reduce slightly the number of treatments 
required. The fine and coarse residues were rubbed and treated 
with water 24 to 81 times to disperse all the colloidal material, and 
the total amount of water used for the completed separation varied 
between 200 and 2,000 times the weight of the soil sample. 
The chief object in making this separation was to see whether the 
colloidal material brought into suspension by successive treatments 
of the same soil differed essentially in its adsorptive capacity. Hence 
in most cases the first fraction of colloidal matter collected was com- 
posed of the first four treatments of washings of the soil. This was 
comparable with the material isolated for determining the adsorp- 
tion ratios given in Table 2. The second, or second and third, frac- 
tions of colloidal material were obtained from a large number of suc- 
ceeding washings of the aoil and usually represented the greater part 
of the colloid material extractable by the means employed. The 
last fraction of colloidal material, except in the case of the Hunting- 
ton subsoil, was composed of material which yielded to dispersion 
less readily than the preceding fractions. The suspensions from 
which the last fraction was obtained were noticeably less concen- 
trated than those obtained from the preceding washings of the soil. 
The last or fourth fraction of the Huntington subsoil was different 
from any of the other fractions extracted for this particular study. 
It was composed of material which had remained for several days in 
suspension, but which had nevertheless been thrown out by the 
supercentrifuge. It contained practically no particles larger than 1 
micron in diameter, coarser particles having been removed by 
subsidence. 
The quantities of colloidal material constituting each successive 
fraction isolated are given in Table 3. In column 2 the quantities 
are expressed as percentage of the whole soil and in column 6 as 
percentage of all the colloidal material isolated. The adsorptive 
57580—24 3 
