Cloninger's clinical method of classifying alcoholics into two groups (Types I and II) was examined with data obtained from 360 VA hospitalized male alcoholic patients. For operational criteria, the Cloninger clinical method of subtyping alcoholics employs age-of-onset of problem drinking and symptom-clusters supposedly associated with each subtype. Marked overlap was found between the symptom-clusters used to define the two subtypes. Ninety-one percent of the entire sample satisfied criteria for both symptom-clusters. Dividing the sample by early-onset (Type II, less than or equal to 25 years) and late-onset (Type I, greater than 26 years) alcoholism did not substantially reduce the overlap between symptom-clusters; i.e., 96% of the early-onset and 83% of the late-onset subgroups were positive for both symptom-clusters. Only 21 men (6%) could be classified when both age-of-onset and the type-appropriate symptom-cluster were used to separate patients. In hospital settings, at least, these findings suggest that the two-group clinical alcoholism typology proposed by Cloninger basically reflects the age-of-onset of problem drinking.