Toxin of sufficient strength to kill a 400-gramme guinea-pig in three days and a half in a dose of 0·cubic centimetre developed in suitable bouillon, contained in ordinary Erlenmeyer flasks, within a period of twenty-four hours. In such boullon the toxin reached its greatest strength in from four to seven days (0·005 cubic centimetre killing a 500-gramme guinea-pig in three days). This period of time covered that of the greatest growth of the bacilli, as shown both by the appearance of the culture and by the number of colonies developing an agar plates. The bodies of the diphtheria bacili did not at any time contain toxin in cosiderable amounts. The type of growth of the bacili and the rapidity and extent of the production of toxin depended more on the reaction of the bouillon than upon any other single factor. The best results were obtained in bouillon which, after being neutralized to litmus, had about seven cubic centimetres of normal soda solution added to each litre. An excessive amount of either acid or alkali prevented the development of toxin. Strong toxin was produced in bouillon containing peptone ranging from one to ten per cent. The strength of toxin averaged greater in the two and four-per-cent peptone solutions than in the one-percent. When the stage of acid reaction was brief and the degree of acidity probably slight, strong toxin developed while the culture bouillon was still acid; but when the stage of acid reaction was prolonged, little if any toxin was produced until just before the fluid became alkaline. Glucose is deleterious to the growth of the diphtheria bacillus and to the production of toxin when it is present in sufficient amounts to cause by its disintegration too great a degree of acidity in the fluid culture. When the acid resulting from decomposition of glucose is neutralized by the addition of alkali the diphtheria bacilus again grows abundantly. Glucose is not present, at least as a rule, in sufficient amounts in the meat as obtained from the New York butchers to prevent the rapid production of strong toxin if the bouillon is made sufficiently alkaline. In our experiments, when other conditions were similar, the strength of the toxin was in proportion to the virulence and vigour of growth of the bacillus employed.