A single blind, within-patient clinical study was carried out on sixteen profoundly mentally retarded in-patients. Five of them had primary generalized, four had partial, six secondary generalized and one had mixed epileptic seizures. With the exception of two patients, all had for months or years been on CBZ combined with other anticonvulsive drugs. Eleven cases had received diphenylhydantoin, eight valproate, six had been given phenobarbital and five nitrazepam. Without changing other anticonvulsive therapy, CBZ was replaced by ox-CBZ up to a dosage of 30 mg/kg b.w. in two or three daily doses. The anticonvulsive efficacy of ox-CBZ was considered better than that of CBZ in eight patients. In three cases, CBZ was preferred and in the remaining five no preference could be stated. A desirable psychotropic effect was found in three patients on ox-CBZ. Unwanted side effects occurred in seven patients. Two of them had their first episode of status epilepticus during the trial. Two out of sixteen patients had to discontinue the therapy. One of them experienced several episodes of status epilepticus, and the other patient lost appetite and had to be fed by tube. No signs of hepatitis occurred in any patient. Three patients are still on ox-CBZ, three and a half years from the start of the trial.