268 
Journal of Agricultural Research 
Vol. XXI, No. 4 
of the bases on the surface of the soil particles, itself becoming combined 
and remaining on the soil particles, partly at least, in the form of a 
silicate. This silicate, being in the form of a colloid, a hydrogel, will in¬ 
crease the effective size of each soil particle and thus greatly retard the 
movement of water through the soil and may, in fact, stop percolation 
entirely. Furthermore, on drying, this colloidal gel will tend to bind 
the soil particles in a solid mass. With solutions of calcium or alumi¬ 
num, on the other hand, when the displacement of bases on the surfaces 
of the soil particles takes place colloid gels are not formed, and, in fact, 
such sodium silicate hydrogels as may exist there are almost completely 
eliminated or masked by the formation of practically insoluble precipi¬ 
tates of calcium or aluminum. 
From these preliminary experiments it is also evident that the salts of 
aluminum, like the salts of calcium, when applied to the soil release 
into the soil solution other bases that may be present but not in solu¬ 
tion. It is assumed that such bases as ammonia, sodium, potassium, 
calcium, and magnesium are displaced by the aluminum from their com¬ 
bination in the silicates of the soil and unite with the acid radical of 
the aluminum salt. If this assumption is correct, it would naturally 
follow that soluble salts of aluminum may be useful both in improving 
the physical condition of a soil and also in rendering other bases more 
soluble in the soil solution and presumably more available for the nutri¬ 
tion of plants. 
EFFECT OF SODIUM SALTS 
The injurious effects resulting from leaching a neutral sodium salt 
from the soil may be demonstrated by a very simple experiment. If a 
sample of a permeable soil be placed in a funnel over a filter paper or 
in an ordinary laboratory percolation tube, its percolation rate with 
water may be determined. If the soil is then leached with a solution 
of sodium chlorid or of sodium sulphate or a mixture of the two salts 
in constant or in increasing concentrations the rate of percolation will 
remain approximately constant, or it may be increased slightly over that 
observed with pure water and the percolate will continue clear and nearly 
or quite neutral in reaction. On the other hand, if the concentration 
of the leaching solution is reduced or if the soil is leached with pure 
water after treatment with the salt solutions three phenomena develop, 
namely, the rate of percolation is reduced, the percolate becomes turbid, 
and it also becomes strongly alkaline. 
If the soil so treated is rich in clay, these phenomena are developed to a 
marked degree. With some clay soils the effect is so pronounced that 
it is not possible to collect a sufficient quantity of the percolate for 
analysis. The course of events just described is illustrated in Table I. 
In this case the soil used was a sample of mesa sand from the Yuma 
Mesa in southwestern Arizona. It was taken from a point on the mesa 
which had not been irrigated. The proportion of clay was very small, 
