The most numerous type of micro -organisms in untreated soils was found to be pleomorphic Gram-positive, non-spore-forming bacteria, characterised by the production of very short coccobacillary or coccal forms. Their cultural and physiological characters do not offer a basis for classification but the organisms have been arranged into groups, according to the presence or absence of. motility. Among the non-motile organisms (Group II) two morphological types have been distinguished. Some of the organisms exhibit the characters of corynebacteria and are regarded as belonging to the Corynebacterium liquefacIens group. Others under certain conditions produce a finely branched mycelium, and belong to the Proactinomyces. Experiments on microbial dissociation showed that the Corynebacterium and Proactinomyces types were very closely related. The motile organisms (Group I) resemble the non-motile in many respects, and are unusual in producing branched forms. Their classification has not been settled, but a genus analogous to Gray and Thornton' s (1928) Mycoplana might be formed to include motile organisms resembling corynebscteria. Small Gram-negative rods occurred less numerously than the Gram-positive types in untreated soils. Motile and non-motile, gelatin-liquefying and non-liquefying organisms, producing white or yellow colonies, were all encountered. Spore-forming rods and typical Actinomyces were numerous in only a small proportion of the soils examined, and were not investigated. A study of the numbers and types of organisms in several soils, led to a comparison of different plating media used in soil work. A medium suited to the growth of poor-growing strains of Groups I and II, and composed of 0.25 per cent yeastrel and 0.25 per cent peptone, gave higher colony counts than any of the other media tested. The counts were not, however, comparable with those obtained by other workers using the direct microscopic method. An attempt was made to justify the numbers obtained from direct counts, by inoculating soil dilutions into sterilised soil and sand, a medium which must approach very nearly to that which is natural to soil bacteria. The numbers thus obtained did not exceed those resulting from the inoculation of yeastrel bouillon, and it was found that all the organisms which were capable of growing in sterilised soil and sand, also grew on yeastrel agar. Thus colony counts counts from yeastrel agar may be regarded as giving an approximate estimate of the total number of live micro-organisms in soil. Experiments on the decomposition of certain organic materials in soil, showed that organisms belonging to Groups I and II did not multiply noticeably. This fact, considered along with their morphology and general occurrence in untreated soil establishes their identity with Winogradsky's autochthonous microflora, the function of which is as yet unknown.