DSM-III-R requires the exclusion of causal organic factors to permit a diagnosis of major depression. This exclusion criterion may be unreliable because (1) neither the biomedical nor the historical data base for determining organic causation is operationally specified, (2) rules for weighting various causes of depression are not specified, and (3) the distinction between causal and contributory organic factors often is arbitrary. The exclusion criterion may not be valid because (1) it is unstable with time and (2) clinical criteria for CNS involvement may correlate poorly with neurodiagnostic tests. Both reliability and validity of diagnosis would be improved by keeping Axis I strictly phenomenological and identifying relevant organic factors on Axis III.