鈥� Studies of severely depressed hospitalized patients suggest a shortened rapid eye movement (REM) latency as a specific biological marker for primary affective disease. To assess the validity of these findings, 40 outpatients referred to our Electroencephalographic Sleep Center for evaluation of depressive symptoms were studied. Concurrent with the all night EEG sleep studies, all patients received a brief clinical interview and a battery of self-rating scales. The entire sample was then subdivided into primary and secondary depressives on the basis of follow-up diagnoses. While there were no significant differences between groups on self-ratings of depressive symptoms, the group of primary depressives had significantly shorter REM latencies and higher measures of phasic REM than the secondary depressives. Furthermore, in this patient group, the delineation of primary vs secondary depression was greater than 80% on the basis of only two nights of EEG sleep. Such objective biological measures, if replicated, could provide a method for increasing the accuracy of differential diagnosis among depressed populations in clinical research.