Mechanical analysis comprises two problems: the preparation of the suspension and the analysis thereof. Preliminary treatment aims at maximum dispersion. The state of coagulation of a soil suspension depends upon the nature of the adsorbed ions and on the interaction of these ions with those of the dispersing agent. The stability of a suspension seems to depend upon the height of the potential (-potential) between the inner and outer layer of ions, but it is possible that high potential is merely a consequence of high hydration. The most stable suspensions are formed by clays saturated with highly hydrated ions--Li or Na. The nature of the adsorbed cation also influences dispersion, Li and Na ions effectin the best dispersion. For colloidal particles the notion of a definite size has to be given up. The smallest grains will always be aggregates, the size of which depends upon the nature of the medium and of the ions present. The object of preliminary treatment is to free the particles of a definite size from the coating of colloidal material, which varies in size. This can best be achieved by saturation with highly hydrated ions and subsequent removal of excess electrolyte.

Dispersion methods are of five different kinds: (1) destruction of organic matter, (2) decomposition of carbonates if present, (3) removal of excess electrolytes by washing, (4) addition of peptizers, and (5) mechanical treatment, such as shaking or stirring. In later years the importance of washing has been stressed by many investigators. Peptizers are mostly alkaline; their most important function is probably to neutralize exchange hydrogen ions on under-saturated soils or sediments. Peptizers are used at random if the degree of undersaturation is not known; therefore Puri advocates determination of the exchange capacity. Stirring is probably the most effective mechanical treatment.

All methods of analysis are founded on Stokes' law which is valid for particle sizes of less than 50 microns; above that dimension Oseen's equation has to be used. The conditions for the validity of Stokes' law are discussed. It is pointed out that varying hydration of the smallest particles may cause variations in specific gravity. Methods of analysis are of two kinds: those in which the material is separated into size fractions and those in which the size fractions are computed from analytical data without separation. The first group involves the use of sieves, elutriation, and decantation methods, the second group requires sedimentation methods of different kinds. In elutriation methods the soil suspension is allowed to settle in a rising current of constant velocity; particles who e settling velocities are lower than that of the current will be carried away. In decantation methods the suspension is allowed to settle for definite periods, after which the liquid above the sediment is removed. The procedure is repeated until the liquid is clear. In the pipette method the concentration of the settling suspension is determined at definite depths after definite time intervals, by sampling and weighing. The size fractions are ascertained by the difference. In the hydrometer method the concentration is found from the specific gravity, which is determined at definite time intervals. If the specific gravity of the dry sediment is considered constant, the hydrometer may be calibrated in grams per liter, thus giving the concentration directly. The size fractions are again fou d by difference. The sedimentation tube of Wiegner and the automatic balance of Oden are examples of methods based on the sedimentation curve. The interpretation of this curve in terms of grain sizes is given.