Data on 24-hour urinary levels of catecholamines and metabolites were determined in 114 depressed patients. For each patient, a D-type score was calculated, using a discriminant function equation that was previously derived using data from an independent group of depressed patients. Of all measures, D-type scores provided the highest sensitivity and specificity for separating bipolar/schizoaffective-depressed patients from all remaining patients or from those patients with unipolar nonendogenous depressions. Using Research Diagnostic Criteria (RDC), bipolar I patients demonstrated significantly lower D-type scores than did all other RDC depressive subtypes, including bipolar II disorders. Similar findings were observed using the Clinical Inventory for the Diagnosis and Classification of Affective Disorders (CIDCAD) system: bipolar/schizoaffective patients demonstrated significantly lower D-type scores than all remaining subtypes, including diagnostically unclassifiable, probable bipolar patients (a category somewhat akin to RDC bipolar II disorder). Data pointed to the heterogeneity of bipolar disorders. Catecholamine and metabolite data in this study were compared with recent studies of others.