The number of variant forms which can develop in cultures of the anthrax bacillus seems to be very large. (Gratia 1924; Nungester 1929; Stamatin 1934, 1935a, 1935b; Januschke 1935.) Some changes in colony form are so abrupt and occur so regularly that they appear, at first sight, to be stages of a life cycle or a life sequence. For example, all virulent strains investigated here (Sterne 1937a and 1937b) have produced a smooth mucoid growth on serum agar in an atmosphere containing certain concentrations of carbon dioxide; and all have, under such conditions, rapidly and continuously produced avirulent daughter strains. Nevertheless, the majority of variant colony types occur in a more haphazard manner so that it is usually impossible to predict, with any assurance, the changes that will occur in a culture subjected to certain conditions. The results summarized below were obtained during the course of experiments done to see whether any “trends" could be recognized in the dissociation pattern of B. anthracis.