36 BULLETIN 1452, U. S. DEPARTMENT OF AGRICULTURE 
characteristics of the colloid are usually associated in a particular 
way, it appears that there is not an absolute dependence of one 
fundamental characteristic on another, since there are so many 
irregularities in the parallelism between properties. Some of these 
exceptions may be due to incomplete or imperfect methods of deter ^ 
mining the properties or behavior of the different colloids, but it is 
believed that all irregularities could not be explained in this way. 
Several exceptions to the parallelism between properties have been 
noted, aside from the data given in Table 19. The Aragon colloid T 
for instance, gives a heat of wetting about the same as that of Hunt- 
ington colloid while it requires only one-eighth as much methylene 
blue to render it isoelectric (Table 18). 
Very little information is available in the literature on the rela- 
tion between different properties of the soil colloids, since it is only 
in the last few 3 r ears that much work has been done on the isolated 
colloidal material. In a previous paper the authors pointed out the 
correspondence between variations in heats of wetting and in am- 
monia adsorption for a series of 13 colloids (3). Joseph, working 
with certain samples supplied by this bureau, noted a correspondence 
between tensile strength and aclsorptive capacity (27). In the case 
of certain clay soils and fine soil fractions, which were doubtless 
largely colloidal, Hardy found a correspondence between the water 
held at maximum plasticity and the linear shrinkage of the 
material (00). 
Since there is a fair degree of parallelism between variations in 
most of the laboratory properties of different soil colloids, it seems 
that a fairly good idea of the character of the colloidal material 
present in a given soil might be obtained by determining one or two 
properties. Possibly further investigation is needed before suggest- 
ing just what determinations should be selected for indicating the 
general character and behavior of a soil colloid. It appears, however, 
from the data already obtained, that the general behavior of a soil. 
colloid might be predicted fairly well from a determination of the 
exchangeable bases and from a determination of either the heat of 
wetting or the adsorption of water vapor over moderately strong sul- 
phuric acid. Also, the character of the colloid and its behavior would 
SiO 
be fairly well indicated by its ,. ^ J; ^ ratio, as shown in the 
subsequent discussion. 
Although these determinations would apparently characterize 
fairly well the general behavior of a colloid, they might not indi- 
cate the state of aggregation of a colloid in the soil. Possibly other 
determinations would prove more satisfactory for measuring chang- 
ing conditions of a given colloid. In the case of several colloidal 
materials, viscosity determinations have been utilized for measuring 
changing states of aggregation in toils, but these may not be ap- 
plicable to colloidal material as it exists in the soil. 
RELATION BETWEEN PROPERTIES AND CHEMICAL COMPOSITION 
The fact that different soil colloids tend to show corresponding 
variations in many of their properties was explained on the ground 
that these properties are probably dependent upon more funda- 
mental properties of the colloids, such as the chemical nature, size, 
