10 
SOIL CHEMISTRY: E. W. HILGARD 
more slowly than with the crude substance, and although aided by 
heating to 100° for three days, did not recover the full volume of 11 cc. 
A repetition of the experiment with 1.70 grams of dry clay which had 
been precipitated from diffusion by ammonium carbonate, gave an in- 
crease of volume to 32 times that of the clay. 
A parallel experiment with some ordinary plastic clays showed an in- 
crease of only two to two and a half times the original bulk. 
Chemical Composition of the Clay. — On leaching the original substance 
with water, it was found to contain 2.60 % of soluble salts, of which 1.74 
was sodium sulphate, 0.74 sodium carbonate, and 0.12 sodium chloride. 
Black rings from the evaporation of the alkali solution were reported to 
exist in the depressed spots, as is usual in the case of 'black alkali.' 
In attempting to obtain for analysis a sample free from carbonates, it 
was found that silica went into solution even when only cold tenth- 
normal acid was used. It was therefore necessary to analyze the entire 
mass. On ignition the loss was 19.6 %; the ignited material decrepi- 
tated when wetted, but did not swell. 
Two analyses of the original material were made; one by acid diges- 
tion for five days; another, after ignition, by fusion with alkali carbonate. 
Unfortunately the material at command was too scanty to permit the 
use of more than one gram and 1.7 grams for these analyses, respectively, 
the results being as follows: 
Partial Analyses of Clay from Mexico. 
Fusion with 
carbonate of soda. 
Acid digestion. (Ignited substance) 
Insoluble 1.83% 0.00 
Soluble silica 43.00 51.43 
Lime 9.06 8.97 
Magnesia 17.11 27.07 
Ferric oxid 
1.76\ 
3.48/ 
. ^ 10.63 
Alumina 3 
Ignition loss (water, carbonic acid and organic matter) 19.60 0.00 
Soluble salts. 2.60 
Total 98.44 98.10 
The above data, deficient as they are, show clearly a totally different 
composition from any 'clay^ on record. The alumina present is far be- 
low any reasonably assumable compound with the soluble silica; the 
predominant base being evidently magnesia, and that greatly in excess 
of the lime present. There is thus an apparent relationship to the 
saponite or sepiolite group of minerals, but the extremely ready decom- 
position by acids (even by acetic acid in dilute solution) differentiates 
