566 
AGRONOMY: L. T. SHARP 
be expected. That such an increase had occurred seemed evident from 
the data secured by the use of Briggs'^^ centrifugal method for moisture- 
equivalents. However, a study of the hygroscopic coefhcients of the 
variously treated soils failed to reveal any increase in surface. In all 
probabiUty Mitscherlich's^^ method would be the most appropriate one 
for obtaining information regarding this point, but it has not yet been 
tried by us. 
The interchange of ions when salt solutions are placed in contact with 
soils forms the working basis for the hypothesis that new colloids have 
been formed. The filtrates from soils which have received salt appli- 
cations followed by leaching with water contain calcium and magnesium 
in amounts chemically equivalents^ to the sodium taken up in exchange 
therefor by the silicates. It is to this new sodium silicate compound 
as the first possibility that we attribute the increased colloidal prop- 
erties noted above. Potassium and ammonium salts seem to produce 
similar compounds, while calcium and barium salts under the same 
conditions do not produce the colloidal hydra ted silicates. These new 
substances, together with the normal soil colloids, remain in a floccu- 
lated condition as long as free acid^^ or an excess of the soluble salts 
added or those formed by the interchange is present. 
As flocculated colloids their effect on the physical condition of the 
soil is either beneficial or passes unobserved. Hence as long as salts 
are present in sufficient quantity no objectionable feature appears, but 
when these flocculating agents are removed by leaching with water, 
the colloids become diffused and offer great resistance to the further 
passage of water. 
In explanation thereof, we might assume in accordance with Tolman's 
ideas that the normal lyophobic soil colloids when suspended in water 
have a positive surface-tension, hence greater than zero, but that the 
new hydrated sodium silicate when suspended in the same medium 
automatically provides by hydrolysis a sufficient quantity of OH ions 
to lower this surface-tension and to bring it nearer the zero value, so that 
the colloids, in this case, would have a far greater tendency to become 
and to remain dispersed. This assumption requires the consideration 
that both the normal colloids and those resulting from the interchange 
of ions carry negative charges, which brings us immediately to our 
next hypothesis which deals in the main, with the nature of the sus- 
pending medium, rather than with the changes occurring in the solid 
substances. 
This hypothesis is largely based on the contention that the behavior 
of soils under the influence of salts agrees in some measure with the 
laws which are thought to govern the behavior of dispersed systems to 
