THe identification of Salmonella cultures from cases of intestinal infection forms an important part of the routine duties of many bacteriological laboratories, and in this connexion the most important problem is to differentiate between B. paratyphosus B and the closely related food-poisoning bacilli. To the public health officer this differentiation is often a matter of great importance, for it is essential for him to know whether the case is one of paratyphoid fever or of infection by one of the food-poisoning organisms. The routine carbohydrate fermentation tests do not help to distinguish between these organisms but, as a general rule, agglutination tests are of service. Consequently, preliminary agglutination tests with “O” sera are carried out, and serve to place the organism in one of several subgroups. In this paper we are mainly concerned with organisms falling into that “0” subgroup containing B. paratyphosus B, B. aertrycke, and the “Stanley”, “Heidelberg”, “Chester”, “Derby”, “Reading”, Abortus equi and certain other strains of Salmonella (see Kauffmann, 1937). Later, tests with specific “H” sera can be performed and the cultures often accurately identified, but always the point of practical importance in such investigations is to distinguish between B. paratyphosus B and the foodpoisoning group. While it is often not of any practical importance to know the precise name of a food-poisoning bacillus, it is important to exclude the possibility of its being a strain of B. paratyphosus B.