The anticonvulsant potency of nitrazepan, diazepan and trimetadiones in five patients with myoclonic absences was compared. A placebo was also used. Therapeutic tests were performed by 4-days successive treatments designed to avoid the residual effects of one drug upon the other administered afterwards. In two cases statistical analysis was not necessary for demonstration of the superiority of nitrazepan and diazepan in regard to trimetadiones; in two others the application of Tukey's test allowed this same conclusion. In the last case the numeric variations of the seizures during the therapeutic essay suggest identical conclusion allthough the statistical analysis was not significant. There are indications that nitrazepan is more effective than diazepan against this type of epileptic seizures.