6 BULLETIN 1311, U. S. DEPARTMENT OF AGRICULTURE 
lated by Schloesing was not exactly of the composition of kaolinite, 
though it was considered by him to be a hydrous aluminum silicate. 
Van Bemmeleivs (3) studies are a distinct contribution to our 
ideas of the composition of soil colloids. He treated some very 
heavy clay soils from river bottoms in Holland and from Java 
with hot moderately concentrated hydrochloric acid, then the residue 
with hot concentrated sulphuric acid, each acid treatment being 
followed by an extraction with sodium hydroxide. He considered 
that this treatment dissolved all the colloidal matter in the soil and 
did not sensibly dissolve the noncolloidal material present. 4 In 
the hydrochloric acid-sodium hydroxide extract, which dissolved 
varying amounts of silica and alumina and practically all the iron, 
lime, and magnesia, the molecular ratio, .Tvrrf '* varied from 1 to 5 
in different soils. The sulphuric acid-sodium hydroxide treatment 
dissolved considerable silica, alumina, and iron, but practically 
nothing else, and in several series of soils the molecular ratio of 
the silica to alumina was very nearly 1 to 2, which is the ratio found 
in kaolinite. Later work, however, showed that this strict relation 
between the silica and alumina did not hold with all soils, though 
the variation was not great. Failing to find a definite ratio between 
the silica and alumina in these extracts. Van Bemmelen concluded 
that these constituents were not present as definite compounds but 
rather as absorption compounds of indefinite chemical composition. 
Van Bemmelen was able to remove the red color from certain 
red soils by dilute hydrochloric acid, and from this fact he reasoned 
that some form of free ferric oxide was present in these soils. He 
also found that dilute sodium hydroxide extracted a large amount 
of alumina and very little silica from some very heavy tropical 
clays, and from this concluded that free aluminum hydroxide was 
a constituent of the colloidal matter in these soils. 
From the experimental work just given and from a general re- 
semblance of the adsorptive properties of soils to the adsorptive 
properties of artificial gels of silicic acid and the hydroxides of iron 
and aluminum which he prepared and studied. Van Bemmelen (If) 
concluded that the inorganic colloidal constituents of the soil were 
the following gelatinous materials : Free iron oxide, free silicic acid, 
and indefinite adsorption compounds or complexes of silica with 
alumina, iron oxide, and water, and small amounts of lime, magnesia, 
soda, and potash. 
Hall (18) is very guarded in his statement concerning the com- 
binations of the elements in the fine fractions. His recent statement 
on the subject is as follows : 
Taking a mean of a number of analyses of the elay fraction of Rothamsted 
soils, then if all the alumina may be supposed to be combined as kaolinite 
A1 2 3 , 2Si0 2 , 2H 2 0, the clay would contain 72 to 75 per cent kaolinite. 11 to 
12 per cent ferric oxide, with 9 to 10 per cent' silica and 4 to 6 per cent of 
alkalies and alkaline earths. This means that the clay contains practically 
no quartz particles, all the silica being in combination with the bases. 
In another publication Hall and Russell (20) state that " the clay 
is clearly a complex silicate or a mixture of several, and in our soils 
two varieties were found " — one from highly fertile soils, " yielding 
* This conclusion has been questioned by Stremme and Aaraio (U) and others. 
