Smith (1904-1905a, b) was the first to show that human and bovine types of tubercle bacilli gave different reaction curves in glycerol broth. Although both types produced an initial alkalinity, the human type alone gave terminal acidity. Smith (1910) claimed that reaction curves were useful in distinguishing human from bovine types, but atypical curves led subsequent workers to abandon the test. The cause of terminal acidity is problematical. Smith (1910), Long et al. (1922), and Henley and Le Duc (1939) suggested that acids are formed from glycerol. Merrill (1930, 1931a,b), who cultivated pathogenic and saprophytic mycobacteria on broth and on synthetic media, believed that glycerol like available carbohydrates was oxidized completely without the accumulation of intermediate products. These divergent opinions may arise from the use of nitrogenous sources with different effects on glycerol metabolism, or from the use of media with different glycerol/nitrogen ratios (Goyl, 1936; Henley and Le Duc, 1939); and the uncertainty is aggravated by failure to identify the organic acids supposed to accumulate (Dingle and Weinzirl, 1932; Wedum, 1936). During the routine cultivation of acid fast bacteria on glycerol broth, it was noticed that most species gave a terminal alkalinity, whereas Mycobacterium butyricum regularly showed a terminal acidity. In view of the important bearing acidic products may have on the theory of terminal oxidation in aerobic bacteria, the phenomenon has been investigated in some detail. MATERIALS AND METHODS Organisms. The following species from the National Collection of Type Cultures, Lister Institute, London, were examined: Mycobacterium phlei (no. 525), Mycobacterium smegmatis (no. 523), Mycobacterium stercoris (no. 3820), and Mycobacterium butyricum (no. 337). Cultures. Stock cultures were maintained on 5 per cent glycerol broth for several generations before experimental use. In this paper the term "broth" is used to denote beef infusion to which