We report the results of a double-blind cross-over study comparing clobazam and placebo in the treatment of refractory childhood epilepsy. Fifty-two percent of patients had greater than 50% reduction in their seizure frequency when taking the clobazam. During the placebo phase no patient recorded a greater than 50% reduction in seizure frequency. Sex, age, seizure type, intellect did not appear to differentiate clobazam responsive from nonresponsive patients. Only 2/21 patients had behavioral changes on the drug sufficiently severe to require the patient to drop out of the study prematurely. Drug interactions between clobazam and the other anticonvulsant medicines did not occur.