30 patients with delirium tremens were given in a double-blind trial – beside the basic treatment with chlormethiazol (Distraneurin®) – aprotinin (Trasylol®) or placebo. Duration of the delirium and the amount of chlormethiazol used were the criteria for successful treatment. It was shown that the additional application of aprotinin did neither shorten significantly the duration of the delirium nor save the amount of chlormethiazol used. Methodologically, special attention was given to the question of duration of the delirium and of registering symptoms. A delirium-rating scale was devised and its analysis showed a good randomization of the items. One main question was as to what extent the individual items were good indicators of a delirium. An item intercorrelation showed that there were two clusters of symptoms: psychological and sympathetic nervous system symptoms. It could be shown that the items ‘consciousness, orientation, hallucinations and short-term memory’ were good indicators of the delirium, while items of the autonomous nervous system, as tremor of hands and body, facial muscular twitching and exteroceptive reflexes, were less indicative of delirium. The duration of the delirium seems to be the best criterion for the question as to whether a drug is effective or not in delirium tremens. There is a highly significant correlation between the degree of the severity of the delirium and its duration. Other significant predictors for the severity of the delirium were the maximal pulse rate and change in blood pressure. Age, duration of alcoholism and psychological or physical depravation showed no influence on the duration of the