SUMMARY: Three phases of Proteus vulgaris, A, B and C, are distinguishable by cellular and colonial morphology, and to some extent serologically. Phase A is the modal form of freshly-isolated strains ; it has a uniform bacillary morphology, swarms intermittently on nutrient agar, and forms stable suspensions in 0.85 yo saline.Phase B strains, though possessing flagella, are usually non-motile, non-swarming, and highly pleomorphic; usually form unstable suspensions. Phase C strains are uniformly filamentous, motile, often swarm in a continuous film, and stability in saline varies with the strain.The somatic surface of phase A strains is characterized by a dominant typespecific antigen, and traces of a non-specific and a possibly strain-specific antigen. In phase B strains the type-specific antigen is largely lost, and the other two antigens dominate the surface. The antigenic surface of phase C strains does not differ markedly from that of phase A.The phase variations A-tB and A+C are reversible. Phase B resembles a partially R form, in which the mouse-virulence is significantly less than that of the parent phase A strain.The first serological study of colonial variation in Proteus vulgaris is that of Weilk Felix (1916,1917) in their classical paper describing the 0 and H variants. This was amplified later by Weil (1920, 1923) and Felix (1923) who described variations of colonial morphology in the 0 strains of P. vulgaris. Both authors described a modal colony, and one which was opaque grew more slowly than the modal on nutrient agar at room temperature, and gave a finely granular growth in broth. In addition, Felix described a non-swarming (0) colony which had an irregular or ill-defined edge, and was granular ('griessig'). Although the small-colony variant and the apparently 'normal' type of colony to which it occasionally reverted, differed serologically from the parent ' normal ' strain, no association was demonstrated between changes in antigenic constitution and colonial form. Both authors regarded this variation as a specific kind of group transformation, largely because Weil found that his small-colony variant of an X 19 strain had developed an antigenic receptor identical with that found in his X 2 strains. Felix confirmed this, and regarded many of his variant strains as intermediate stages (' ubergangsformen ') in this antigenic reconstruction.Few of the later workers on the morphology and serology of the Proteus group mention colonial or antigenic variation (Taylor, 1928 He described a granular colony with a rough surface and irregular edge, which formed an auto-agglutinable suspension in 0.85 yo saline and which he regarded