THE CHEMICAL COMPOSITION OF SOIL COLLOIDS 33 
No definite conclusions concerning the manner in which silica and 
alumina are present in soil colloids can be drawn from the results 
of the hydrochloric acid and sodium hydroxide treatments. The 
results would seem to indicate that only a part of the silica and 
alumina, the part soluble in hydrochloric acid or sodium hydroxide, 
is free. The undissolved silica and alumina would seem to be com- 
bined or more insoluble than familiar forms of free Si0 2 and xVL0 3 . 
If all the silica and alumina were free, one would expect the pro- 
portion of A1 2 3 to Si0 2 dissolved by HC1 to differ greatly from 
the proportions dissolved by NaOH; but this is not the case. This 
consideration would point to the silica and alumina being in some 
combination which is less soluble than free silica and free 
alumina, possibly a kaolinitelike compound. On the other hand, the 
silica and alumina remaining undissolved do not correspond to any 
definite compound. The fact that the kaolinite suspension we used 
was less soluble in acid than the soil colloids does not prove that 
silica and alumina are not present as kaolinite, because the particles 
of the soil colloids were probably finer than the kaolinite particles 
and it is known that solubility depends considerably on the size 
of particles. 
The quantity of iron oxide dissolved by the hydrochloric acid is 
about the same as that dissolved by oxalic acid. The only red and 
brown colloids in the series are those of the Cecil and Chester soils. 
It will be noticed that the hydrochloric acid dissolved considerably 
more iron from these colloids than from the four light-colored col- 
loids. The digestion with hydrochloric acid removed the red and 
brown colors from the Cecil and Chester colloids just as the treat- 
ment with oxalic acid had done. The digested residues still con- 
tained large quantities of iron. This is considered further evidence 
that there are two forms in which iron is present in soil colloids. 
Lime is practically all dissolved by dilute lrydrochloric acid. Only 
half or considerably less of the magnesia is dissolved, however, and it 
would seem that the lime is held in a more easily decomposable con- 
dition than the magnesia. This same generalization is brought out 
by the compositions of the soil colloids as discussed on page 16 and 
also by the relative extractabilities in water. 
The phosphoric acid in soil colloids appears to be present in a 
form comparatively soluble in both hydrochloric acid and sodium 
hydroxide. 
Other concentrations of acid and alkali might be employed as 
partial solvents of soil colloids, though it is doubtful if any sharp, 
clear-cut conclusions could be drawn from the results. It has fre- 
quently been assumed that the ordinary acid-digestion method of 
analysis dissolves the colloidal matter in soils. The work ivith 
Sharkey and Cecil colloids showed that only about half was dis- 
solved. Analysis of the Sharkey residue showed that the iron had all 
been dissolved by the acid digestion, but the ratio of silica to alumina 
in the residue was the same as in the original colloid. The residue 
from the hydrochloric-acid digestion was treated with hot fuming 
sulphuric acid and then with sodium hydroxide, and there remained 
