6 BULLETIN 1193, U. S. DEPARTMENT OF AGRICULTURE. 
tion of the water a second treatment yields a heavy colloidal suspen- 
sion. In such cases the first treatment with water doubtless removes 
a certain amount of a soluble salt which has a flocculating effect. 
Konig and Hasenbaumer (17) showed that after addition of aluminum 
chloride to a soil, agitation with ten fresh portions of water failed to 
bring any colloid into suspension; but the eleventh treatment 
with water containing a small amount of ammonia produced a heavy 
suspension. 
Other methods which are based on adsorption — a positive property 
of the colloids — have been proposed for indicating the relative 
colloidal contents of soils. Ashley's method (#), involving a deter- 
mination of the quantity of malachite green adsorbed, has been 
most widely used on ceramic clays, which may be regarded as a 
special type of soil. Our results show, however, that while Ashley's 
method might indicate differences in the plasticities of ceramic 
clays, it does not indicate the relative colloidal contents of agricul- 
tural soils, since the colloids extracted from different soils vary 
greatly in their adsorptive capacities for this dye. 
Mitscherlich (21, p. 476) has suggested thai, the amount of water 
vapor adsorbed by soils under standard conditions is probably to be 
attributed almost entirely to the colloidal matter present, but that 
the water adsorption values of different soils probably do not indicate 
the relative amounts of colloids present, since different soils contain 
different colloids which vary in their adsorptive capacities. Our 
results would indicate that while colloids extracted from different 
soils may vary considerably in their adsorptive capacities for water 
vapor when in equilibrium with the partial pressure of water vapor 
afforded by 10 per cent sulphuric acid they have a relatively constant 
adsorptive capacity when in equilibrium with 3.3 per cent sulphuric 
acid. 
Stremme and Aarnio (36) applied the Van Bemmelen method of 
determining colloids by digestion with sulphuric and hydrochloric 
acids to various rocks, and they concluded that the method was 
subject to error in that certain noncolloidal minerals were dissolved. 
Van der Leeden and Schneider (18) conclude that the dye adsorp- 
tions of soils do not correspond to the colloidal contents, that 
unweathered silicates as well as colloidal material are dissolved by 
digestion of the soil with hydrochloric acid, and that there is not an 
exact parallelism between the quantities of water and dye adsorbed 
by soils. 
Tempany (37) has recently proposed that the colloidal content of a 
soil might be estimated from the shrinkage of the soil on drying. It 
is yet to be proved, however, that the colloidal materials in ail soils 
have the same shrinkage coefficient. 
In the earlier investigations, and in much of the modern work, the 
attempt has been made to determine the colloidal matter in soils 
through the methods of mechanical analysis ; that is, by actual separa- 
tion of the colloidal and noncolloidal soil constituents. Proceeding 
in this manner, it was, of course, necessary to get the colloidal matter 
in suspension before it could be recognized as colloid and size of the 
particles became the sole criterion of judging what was colloid. 
Such a separation would be definite and entirely satisfactory for the 
interpretation of soil phenomena if all the soil particles were of one 
kind; that is, if the particles were all fragments of crystalline minerals 
