NH3 as well as sulphide S is split off from insulin under the action of dilute NaOH, the amount being dependent upon the reaction conditions. A part of the NH3 obtained probably originates from one or more of the free amino groups of insulin, the presence of which seems necessary for the physiological action of the hormone. At the same time the cystine content is decreased; the arginine content remains unaltered. Treatment of insulin with hot, dilute HC1 results in the liberation of ammonia without a significant decrease in activity. This NH3 probably originates elsewhere in the molecule than that obtained by alkali treatment, possibly from amide-groups. During the action of acid alcohol on insulin, according to Carr''s method, formation of diketopiperazine rings may take place. These results are explained by means of a structural model of the hypothetical active group of insulin.