SummarySeveral methods were used for extracting humus from a number of widely different soil types. The cementing power of the various humates was tested by mixing them with soil and determining the weight of water‐stable aggregates surviving wet sieving. In general, α‐humic acid>Tiulin's group‐2b humates >group‐2a humates >group‐1 humates >fulvic acid. Greater stability was obtained when the humates were fixed as films over the soil particles by H or Ca ions.An attempt was made to fractionate humic acid into its component colloidal cements without success. Various chemical treatments, such as hydrolysis, deaminatipn, esterification, and acetylation modified the glueing qualities of humic acid, suggesting that colloidal substances with NH2, COOH, and OH polar groups were partly responsible for the cementing action of humates.A study was made of the influence of various organic substances on soil aggregation. Resins, fats, and some proteins cemented soil into aggregates so impervious to water that slaking was prevented. Some microbial gums and other polysaccharides acted as good soil glues, while others were useless. Lignates were only slightly inferior to numates as soil cements, but tannin and cationic detergents were of little value. The glueing properties of proteins and polyuronides were considerably altered by various chemical treatments.The structure of puddled soils was sometimes restored by repeatedly wetting, incubating, and drying them. Acid, lateritic soils high in sesquioxide cements were rapidly restored; chernozems containing much colloidal humus and clay took longer, but other types showed little improvement even after prolonged treatment. Sterilization of soil rendered the humus more available for micro‐organisms, which were then able to improve soil structure temporarily. Micro‐organisms reduced the stability of aggregates cemented by films of calcium humate.This paper examines the possibility of colloidal humus constituents acting as soil cements. Geltsef (1936, 1937, 1943) claims that only readily decomposable organic materials improve the physical condition of soil. She considers that resistant humus is practically inert and plays little part in cementing soil particles into water‐stable aggregates. Many soils, however, such as lateritic red loam, chernozem, rendzina, and terra rossa possess good crumb structure even after the nutritive humus is exhausted through long periods of continuous cultivation. It is doubtful whether they owe much of their permanent structure to fungal hyphae and capsulated bacteria. McHenry (1945) also considers that too much attention has been given to the frail crumbs formed by microorganisms and not enough to the strong aggregates present in many soils.Demolon and Hénin (1932) studied the relative aggregating powers of humus and clay. Sideri (1930, 1938a, 1938b) postulated that soluble humates were orientated by clay particles, thus forming films. On drying, these films became irreversible so that they cemented aggregates together even when re‐wetted. Myers (1937) showed that humates acted as better cements in the presence of Ca and H ions than Na ions. He considered that the humates were bound to negatively charged clay micelles through polar groups, such as COOH.Tiulin (1938a, 1938b, 1946) classified soil organic matter into loosely bound humates of Ca and Mg and those firmly bound by Fe, Al, or clay. These could be separated by fractional peptization using different concentrations of sodium hydroxide. He found that chernozems with good structure contained relatively high amounts of the former, while podzols with poor structure contained large quantities of the latter combinations. Jung (1943) and Meyer (1941) considered that part of the humus was adsorbed to clay by polar bonds and part by molecular and van‐der‐Waals forces. Chapek and Sakun (1944) thought that most of the humate was held mechanically and only a small fraction was adsorbed onto clay. McHenry (1945) and Peterson (1946) studied the influence of clay minerals on the stability of organo‐mmeral aggregates and concluded that clays with a high cation‐exchange capacity gave better aggregation than those with a low exchange value. Electron‐microscope studies of natural soil aggregates by Kroth and Page (1946) showed that clay particles were covered with humus, and they concluded that humus was adsorbed to the clay and was not merely entangled.Practically nothing is known concerning the cementing power of the various constituents in humus. Fuller (1946, I947a, 1947b) has obtained evidence that uronides occur in humus, while Forsyth (1947) has isolated a small amount of polyuronide from fulvic acid. Martin (1945, 1946), McCalla (1946), and Geoghegan and Brian (1946) have found that bacterial and other polysaccharides aggregate soils. Kojima (1947a, 1947b) detected appreciable quantities of amino acids in peat, thus suggesting the presence of proteins, and McCalla (1946), Kroth and Page (1946) have observed the glueing power of undecomposed proteins. Waksman (1938) and Gillam (1940) gained evidence that lignin‐like substances were present in humus, and McCalla (1945), and Alderfer, Gribbins, and Haley (1944) have shown that such substances are capable of aggregating soils.