Clinical trials designed to evaluate the efficacy of drugs in human aggression have been scarce until recent years. The potential antiaggressive action of minor tranquilizers in humans has received little attention in spite of the claimed "taming effect" in some animal studies. A recent report examining the literature regarding the effects of benzodiazepines on animal models of aggressive behavior has pointed out the lack of consistency in such findings. Similar observations have been noted in humans where reduction in aggressive manifestations is contrasted with an increase in hostility in a few studies, as well as with the appearance of "paradoxical" rage reactions. Some variables that could account for such discrepancies have been advanced. They include, among others, dosage, acute vs. chronic drug administration, and possible qualitive differences among this group of agents. Individual variations as to presenting clinical picture, initial levels of anxiety and hostility, and personality types have also been mentioned. Implications of some of these findings for future clinical research are discussed. At present, a study designed to test the efficacy of two benzodiazepines, at dosages higher than those usually recommended, is being carried out in a population of anxious, aggressive-prone individuals with poor impulse control. Thus far, and in agreement with our previous clinical experience, we have not seen "paradoxical" rage and such high dosages have been well tolerated.