A N S (Tissier and Martelly, 1902) and CZ. sordellii (Sordelli, 1922; Hall and Scott, 1927) are in many respects similar organisms. Clark and Hall (1937) found that they are morphologically and culturally indistinguishable and serologically related-only the non-pathogenicity of the first and high pathogenicity of the second distinguish them. Stewart (1938), after investigation of some few strains, suggested that CZ. sordeZZii is a pathogenic variant of CZ. bifermentans and that they should be regarded as a single species: CZ. bijiermentans. This suggestion has been accepted by some workers (Breed, Murray and Smith, 1957). However, Huet and Aladame (1952) and Tataki and Huet (1953) were able to distinguish these species on the grounds of urease production and agglutination. Strains of CZ. sordeZZii are ureolytic ; sera of rabbits immunised with ureolytic strains agglutinate ureolytic strains only.Cf. bifermentans is non-ureolytic and never pathogenic, but sera against CZ. bi$ermentans agglutinate both ureolytic (CZ. sordellii) and non-ureolytic (CZ. biferrnentans) strains. Brooks and Epps (1959) investigated many pathogenic and non-pathogenic, ureolytic and non-ureolytic strains and found that in addition to ureolysis other features distinguish C Z . sordellii and CZ. bi&errnentans-colonial morphology, fermentation of carbohydrates, the results of the methyl red test, the production of indole and acetoin, gelatinolysis and the nature of the amino acids present in cultures.They concluded that separation of CZ. sordellii from CZ. bifermentans is justifiable.The organisms are antigenically related (Clark and Hall; Stewart), but could be serologically distinguished (Tardieux and Nisman, 1952;Tataki and Huet). Also the spore antigens are distinctive (Meisel and Rymkiewicz, 1959), although antisera to spores of Cl. bifermentans reacted to low titres with spores of CZ. sordeZZii (Walker, 1963). By precipitation tests Ellner and Green (1963) were able to detect in CZ. sordellii and CI. bqermentans species-specific antigens in addition to a common antigen shared by both species as well as some other clostridia studied. Recently Meisel and Switalska (1967) studying 30 strains of Some ureolytic strains are non-pathogenic. Cl. sordellii and 35 of Cl. bqermentans found that neither " species " forms an antigenically homogeneous group ; several distinct serological types could be detected, although more than 80 per cent. of strains could be classified by means of agglutination. These studies also seem to support the view that the species are distinct. However, this taxonomic relationship is again a subject of controversy. It is relatively simple to obtain fully atoxic substrains from toxic strains of Cl. sordellii, but Huang (1959, cited by Huang, Tamai and Nishida, 1965) isolated a toxic strain from the heart of a rabbit infected with an atoxic strain; this toxic strain became atoxic on repeated subculture. This observation inspired the study of conditions that induce the changes of some properties of the strains. Nishi...