16 BULLETIN 1193, U. S. DEPARTMENT OF AGRICULTURE. 
An indirect comparison was also obtained between the adsorptive 
capacities of the extract able and unextractable colloidal material 
in a manner which will be described in detail further on, but which 
was essentially as follows: All colloid possible was extracted from 
the soil and the residual soil material was quantitatively recovered. 
The quantity of unextracted colloid remaining in the soil residue was 
determined microscopically. The adsorptive capacity of the residue 
was also determined and the quantity of colloid present calculated 
by the ratio, 
adsorption per gram of residue 
adsorption per gram of extracted colloid 
A comparison between the quantity of colloid determined micro- 
scopically and the quantity of colloid indicated by the adsorption 
ratios was virtually a comparison between the adsorptive capacities 
of the extractable and unextractable colloids. 
Comparison of successive samples of extractable colloidal materials. — 
In order to determine how closely a small sample of colloidal material 
represented the rest of the extractable colloid, successive extractions 
of colloidal material were made from the following five soils : Sassafras 
and Huntington subsoils, and Marshall, Sharkey, and Vega Baja 
soils. The dye, water, and ammonia adsorption ratios all indicated 
widely different percentages of colloid material in the first two soils. 
The dye adsorption ratio was very different from the water and 
ammonia ratios in the case of the Marshall soil, while in the case of 
the Sharkey and Vega Baja soils, there was a fair agreement between 
all three ratios. If sampling of the colloidal material was the con- 
trolling factor affecting the agreement of the different ratios, it was 
to be expected that the lots of colloidal material successively isolated 
from the Sassafras and Huntington subsoils would vary from each 
other more than the successive samples isolated from the Sharkey 
and Vega Baja soils. 
The procedure adopted for extracting the colloidal material was 
somewhat different from that described on page 8, although the col- 
loidal material was graded from the larger particles in the same 
manner, by use of the supercentrifuge, and concentrated similarly by 
use of Chamberland-Pasteur niters. The method of dispersing 
the colloidal material was so much more effective than the method 
first used that large samples of soil were undesirable. The separa- 
tions described below were made with samples varying from 25 pounds 
of the Marshall soil to 1 pound of the Vega Baja soil. 9 
The colloidal material was dispersed by gently rubbing the soil 
when in a pasty condition with a rubber pestle, or, in the case of large 
samples, by kneading with the hands. From 5 to 20 times as much 
water as soil was added, and after thorough agitation the resulting 
suspension was allowed to settle overnight before decanting. One 
part of ammonium hydroxide in 3,000 parts of distilled water was used 
for promoting deflocculation. 
s As the work progressed the samples were reduced from 25 pounds to about 2 pounds, since it was found 
possible to eliminate losses of the finer soil material when the suspensions were passed through the super- 
centrifuge. In working with the Sassafras subsoil and the Marshall soil considerable suspended matter 
escaped through the bottom of the centrifuge bowl and became contaminated with oil on the bearing, so that 
it had to be discarded. These losses were avoided when working with the other three soils. By using water 
in place of oil for lubrication it was possible to recover and rework the material escaping from the centri- 
fuge bowl. In the case of the Huntington subsoil a slight amount of material was lost, but in the case of 
the Sharkey and Vega Baja soils extraction and recovery of the colloidal material was practically quantita- 
tive; hence the results show the total quantities of this class of material present in these soils. 
