In most research dealing with biological abnormalities in depression, the clinical diagnosis of depression is made and the occurrence of a biological abnormality, for example, reduced REM latency, is documented. In this study, that design was reversed; REM latency was used as a grouping variable to assess empirically the "biological" priority of Research Diagnostic Criteria endogenous symptoms. We found that terminal insomnia, pervasive anhedonia, unreactive mood, and appetite loss were most likely to discriminate among "reduced" and "nonreduced" REM latency depressions at various threshold values. Contrary to expectation, diurnal mood variation was found equivalently in all categories of REM latency studied. Implications for clinical decision making based on endogenous symptoms are discussed.