Forty-five patients who met Research Diagnostic Criteria (RDC) for secondary depression were assessed by St. Louis criteria, and by demographic, illness history, REM latency and dexamethasone suppression test measures. Fully one-third of the RDC secondary sample met St. Louis criteria for primary depression; only age at onset and length of illness discriminated St. Louis primary from secondary depression. RDC depressed patients secondary to alcoholism were compared to those secondary to nonsubstance abuse disorders (excluding schizophrenia). The subgroup with a history of alcoholism reported less severe depression and were preponderantly male. Neither dexamethasone response nor REM latency differentiated the RDC secondary subtypes. Little support was found to validate separation of the RDC secondary subtypes studied.