16 BULLETIN 1452 ; L T . S. DEPARTMENT OF AGRICULTURE 
The soil colloids, as a class behaved more like bentonite than like 
silica gel or gelatin, although they all differed somewhat in behavior 
and none of them resembled bentonite closely. They all broke down 
suddenly in water into irregular masses of the same general appear- 
ance as the original particle. The lumps of Fallon colloid showed 
least disintegration and the Norfolk showed most. Part of the 
swelling of the soil colloids recorded in Table 6 may have been due 
to fractured particles packing less closely than the original lumps, 
but all the volume increases of certain colloids could not be accounted 
for in this way. In the case of the Fallon colloid, which swelled 
most and fractured least, the swelling must have been due chiefly 
to an increase in volume of the material without any change in 
homogeneity that was apparent under the low-power microscope. 
Certain of the soil colloids thus appear to possess some degree of 
elasticity, though it may be slight. 17 
Hofmeister (25) and many subsequent investigators, in determin- 
ing the quantity of water imbibed by gelatin and similar colloids, 
merely immersed disks of the material in water, wiped the surface 
with filter paper, and weighed. Katz (28) modified the procedure 
for pulverulent materials by drying between porous plates. These 
procedures were not well adapted to soil colloids ; hence the fol- 
lowing method was devised : 
The air-dried colloids were thoroughly mixed with about 25 per 
cent of water and were then compressed under 3,000 pounds pres- 
sure per square inch into cylindrical briquettes 16 millimeters in 
diameter by approximately 15 millimeters high. The briquettes 
were placed on a blotter, one end of which dipped into a dish of 
water, and were left over night covered with a bell jar. In the 
morning the briquettes were weighed, and after drying at 110° C. 
for 18 hours they were weighed again, the loss in weight being im- 
bibed water. Some of the briquettes cracked as water was imbibed, 
but moistening the material prior to briquetting greatly diminished 
cracking. 
The above procedure gives the quantity of water the briquetted 
colloids will take up and hold against gravity. The quantities of 
water the powdered soil colloids will hold against a centrifugal force 
equivalent to 1,000 gravity were determined by the moisture-equiva- 
lent method. This method was applied to the soil colloids essen- 
tially as Briggs and McLane (10) developed it for soils. 
Duplicate determinations by either the modified Hofmeister or 
moisture-equivalent methods usually varied less than 5 per cent 
when the same lot of colloid was used; but frequently different lots 
of the same kind of colloid prepared under slightly varying condi- 
tions did not agree so closely. Probably the results by both meth- 
ods are affected somewhat by a state of aggregation of the colloid 
that is not controlled in the experimental procedures. 
17 Additional data regarding swelling of the colloids were obtained by W. H. Fry 
through microscopic observation of single lumps. The volume of a single lump before 
immersion in water was estimated by the aid of an eyepiece micrometer, and the com- 
bined volumes of the aggregates into which the single lump broke up in water were also 
estimated. Volume increases estimated in this way approximated roughly those given in 
Table 6. The microscopic estimates of swelling were as follows : The Sassafras colloid, 
37 per cent ; the Sharkey, 87 per cent ; and four values were obtained for the Fallon, 
which ranged from 44 to 193 per cent and averaged 110 per cent. Thus it seems that 
the volume increases given in Table 6 were largely due to volume increases in aggregates 
which appeared homogeneous under the low-power microscope and not to such aggregates 
packing diifferently in water and in benzene. 
