A prospective study of male in-patients over 55 years old who met Feighner criteria for non-bipolar depression was performed to determine if a previous history of alcoholism significantly influenced treatment or response to treatment. Among 58 subjects with complete follow-up information, the 16 who had a history of alcoholism had a presentation at index which differed from that of the non-alcoholics, and on follow-up they clearly had more chronic illness. This elderly sample with alcoholism resembles 'neurotic-reactive' depressives described in younger samples, and supports a past history of alcoholism as being a risk factor for chronicity of depression on follow-up in the elderly population.