In a cross-sectional design to address the effects of the course of depression on rapid-eye-movement (REM) latency, we have matched patients in their first-episode with (1) age-matched patients with recurrent depression, (2) onset-matched patients with recurrent depression, and (3) age-matched normal control subjects. Patients were also matched for sex and treatment site (inpatient or outpatient). No differences were found in REM latency for the three depressed groups, and all had lower REM latency than normals. This finding is taken as support for stable REM latency throughout the course of depression.