26 BULLETIN 6 2, HAWAII EXPEKIMENT STATION 
Table 9 shows very large disparity in chemical composition of the 
various fractions in soil No. 547, and especially in soil No. 164. In 
these two soils the ratio of adsorptivity of the whole soil to that of 
the colloid extracted therefrom would not be expected to represent 
accurately the amounts of colloids present in the soil. 
DISPERSION OF HAWAII SOILS 
In connection with the mechanical analysis of soils, probably the 
greatest difficulty is had in obtaining a perfect dispersion. This dif- 
ficulty is especially noticeable with certain highly organic soils and 
with soils the colloidal complex of which contains in the adsorbed 
state certain bases, such as calcium, which have a flocculating effect 
upon the soil suspension. Several suggestions have been made in the 
past few years by various workers both as to the mechanical means 
to be employed and as to the chemical treatment the soil should 
receive to insure as nearly perfect dispersion as possible. In the 
light of recent studies on the exchangeable bases, much valuable infor- 
mation has been obtained regarding the role of various bases in the 
dispersion and flocculation of clays. For instance, it is known that 
sodium clay is much more readily dispersed than ammonium or 
potassium clays. It would appear logical, therefore, to employ an 
alkaline solution of a sodium salt to obtain maximum dispersion. 
Hawaii soils contain a larger amount of organic matter than do 
most soils of the mainland of the United States, and in certain areas, 
especially in the lowlands, highly calcareous soils exist. In addition, 
the soils are on the average heavier, containing upward of 50 per 
cent clay and fine silt. In view of these facts, an investigation was 
made to learn what manner of mechanical dispersion gives the most 
efficient results and what chemicals, if any are used, bring about 
complete deflocculation. 
In the studies undertaken in the station laboratory, the work on 
dispersion was therefore divided into two parts, one comparing dif- 
ferent mechanical means and their relative efficiency when the chemi- 
cal treatment is constant, and the other comparing the efficiency of 
various chemical treatments when the mechanical method is constant. 
In measuring the degree of dispersion, Bouyoucos' hydrometer was 
used: (1) Because under the same conditions (concentration and 
temperature) it gives comparable figures; (2) because mechanical 
analysis performed with the aid of it compared favorably with other 
standard methods; (3) because duplications carried out under iden- 
tical conditions gave good agreements; and (4) because of its 
rapidity. 
The five soils, varying widely in nature, selected for the work rep- 
resented certain Hawaii soil types and were in most part typical 
of the larger soil areas. A description of these soils is given (p. 5). 
MECHANICAL ASPECT 
Three methods widely used in mechanical soil analysis were tried : 
(1) Rubbing in a mortar with a rubber pestle under water contain- 
ing alkali (KOH) for varying lengths of time; (2) shaking in an 
end-over-end shaker in a machine revolving at the rate of about one 
revolution per second; and (3) using an electrical stirrer. The first 
method has been used by certain investigators, notably Engle and 
