20 
ment of the texture in heavy soils. Indeed this is said to be the 
primary function of hme, which is one of the most universally applied 
soil amendments. 
The above studies indicate that the maintenance of a crumb 
structure is seriously menaced by the presence of even a trace of 
certain compounds. Hence flocculating or deflocculating agents 
alter the soil structure. The latter not only destroy the compound 
aggregates but also bring about a diffusion or swelling of the coUoidal 
clay. This results in a closing of the pore spaces, shutting out the 
air, development of acid conditions, and menacing the whole cycle of 
normal soil transformations. 
Since it is conceded that the best physical state, known as a crumb 
structure, is due to flocculation of the smaller grains into aggregates, 
the conclusion is obvious that the study of conditions conducive to 
the formation of a colloidal state and the relation of salts to this 
state may be of considerable local application. 
As a means of studying this property of Hawaiian clay a sample 
of highly puddled soil was chosen, one in which the clay would 
remain in suspension for weeks. A stock suspension of this soil 
sufficient for all experiments was prepared, so that a suspension of 
known concentration would be available. The degree of floccula- 
tion, to a certain extent, depends upon the relation of the amount of 
clay in suspension to the strength of the flocculating agent. 
Normal solutions were prepared of all the salts used in previous 
experiments, except the slightly soluble ones, in which case saturated 
solutions were made. By a series of preliminary experiments those 
salts having a negative or deflocculating effect were eliminated. 
These include potassium and sodium phosphates and carbonates. 
Secondly, salts causing a flocculation of the clay but not sufficiently 
soluble to form a normal solution were eliminated. These include 
the oxid, carbonate and sulphate of calcium and the oxid and car- 
bonate of magnesium. 
The comparisons were made in glass cylinders of 400 cubic centi- 
meters capacity in which were placed 2 cubic centimeters of normal 
salt solution, 10 cubic centimeters of clay suspension, and 188 cubic 
centimeters of water, making a total of 200 cubic centimeters. This 
mixture was shaken and allowed to settle. In the case of the stronger 
flocculants this proved too great a concentration, hence the experi- 
ments were repeated with a much weaker solution. The results are 
given in Table IX. 
