Fifty-five adult male patients with alcoholism and seizures were divided into three groups. In group I, all seizures were consistently produced by alcohol withdrawal (N = 16); in group II, some seizures were clearly precipitated by alcohol withdrawal and others were not (n = 18); and in group III none of the seizures had a consistent relationship with drinking (N = 21). The three groups were compared for age, frequency of seizures, compliance, Bear-Fedio and Boston-Naming scores, those for memory of words and faces, and finger-tapping scores, and neuropsychological impairment ratings as well as EEG background frequencies. The incidence of severe head injury preceding the onset of seizure disorder was significantly higher in groups II and III. This was further supported by a greater prevalence of asymmetric alpha rhythm, abnormal CT scans and multiple seizure types in those two groups. These data indicate that a prior head injury is a major variable accounting for the heterogeneity of the alcohol-epilepsy relationship.