14 BULLETIN 1452, IT. S. DEPAKTMENT OF AGKICULTURE 
Zsigmondy (57), Katz (28), Rosenbohm (48), and many other in- 
vestigators have supplied such information for silica gel and gelatin. 
It was beyond the scope of this investigation to study all phases 
of the SAvelling of the different soil colloids. An attempt was made, 
however, to determine (1) the volume changes taking place when 
the colloids were in contact with liquid water, (2) the quantities 
of water the colloids would imbibe against gravity, and (3) the 
quantities of water the colloids would hold against 1,000 gravity 
after they had been in contact with liquid water (moisture equiva- 
lent determination) (10). Samples of silica gel, gelatin, and bento- 
nite were in luded in these determinations in order that the behavior 
of the soil colloid might be compared with that of well-known col- 
loids. 16 Possibly more light would be thrown on the natures of 
the different soil colloids by determining the vapor pressures of the 
colloids with different moisture contents, but it was thought advisable 
first to make determinations which have a more obvious application 
to the behavior of soils. 
Fischer (13, p. 44-151), Loeb (32), and others have determined 
the volume changes of fibrin and gelatin in different solutions by 
merely measuring the volume occupied by the granulated material 
in a graduated C3dinder containing the solution. This method is 
doubtless fairly accurate for estimating differences produced in a 
single material that swells several hundred per cent. Applied to 
more pulverulent or " sandy " materials, such as the soil colloids, 
the method is probably onry qualitative. In order to estimate 
variations in the swelling of different materials in the same liquid 
it is necessary to modify the Fischer method by including a de- 
termination of the volumes occupied by the materials in a liquid 
in which they do not swell. Therefore, in comparing the swelling 
of different soil colloids with the swelling of silica gel, gelatin 
and bentonite measurements were made of the volumes occupied 
by the materials in benzene and in water. It was assumed that 
the materials did not swell in benzene, and the volumes occupied 
in this liquid were used as the base on which to calculate the changes 
in volume, or swelling, in water. 
The procedure was as follows : The air-dried colloids were coarsely 
pulverized and passed through sieves. Only material passing a 
1-millimeter sieve and retained on a 0.5-millimeter sieve was used. 
Samples of 5 grams each were placed in graduated tubes with 15 
cubic centimeters of water or benzene. In order to remove air 
the tubes were evacuated to the boiling point at room temperature. 
After standing three hours the volume occupied by the colloid was 
observed. 
The observed volume of the colloidal material was, of course, 
the volume of the colloid lumps plus the volume of the liquid be- 
tween the lumps. In the process of swelling there was presumably 
an alteration not only in the size of the lumps, but also in the 
space between the lumps, although it is assumed in the method 
that the total pore space between aggregates does not change. Thus 
16 The silica gel used was a commercial product prepared by Patrick's method. The 
gelatin was a preparation for bacteriological work ; it had a pH of approximately 6.0 in 
water. The bentonite was procured from a company supplying mineralogical specimens ; 
it was a tvDicai sampV of this highly colloidal mineral, the characteristics of which are 
described by Ladoo (30). 
