An audit of patients receiving antiepileptic drug therapy for epilepsy was carried out in 25 general practices in Glasgow. Patients were identified from computerized records of repeat prescriptions for antiepileptic drugs. Overall, 1052 (0.72%) of 145,609 screened patients had treated epilepsy. Only 5% were children, while 19% were over 65 years. Twenty-nine per cent were diagnosed by a neurologist, and in 24% no record was available of who had made the diagnosis. Fifty per cent had tonic-clonic seizures only. Partial seizures occurred in 39%, absences in 4%, and myoclonic jerks in 3%. In only 39% of case records was current seizure control documented. Seventy-four per cent and 41% of patients had surface electroencephalography and computerized tomographic brain scanning, respectively. In more than 80% of patients the presence or absence of birth injury, febrile convulsions in childhood, and a family history of epilepsy were not mentioned. Seventy-six per cent of patients were receiving anticonvulsant monotherapy. The most commonly prescribed drugs were carbamazepine (43%), phenytoin (34%), sodium valproate (22%) and phenobarbitone (15%). Eighty-four per cent had attended a hospital clinic with their epilepsy, and 19% had been admitted to hospital with seizures or complications. A standard record form for the assessment and follow-up of epileptic patients in general practice would help in providing optimal management and in facilitating the setting up of a shared-care programme.