Seventy volunteers had 3 nights of sleep recordings during a period of marital separation, and 61 returned for repeat studies 1 year later. At that time, the divorce was final for 42. Forty of the volunteers were depressed when first screened, and 30 were not. Initially all those undergoing marital separation had less delta sleep than an age-matched married comparison sample. Delta increased at followup for those whose divorce was completed. Rapid eye movement (REM) latency was reduced and REM percent was elevated only in the depressed. Among the not depressed, those whose divorces remained incomplete at the followup had lower delta, higher REM percent, and shorter REM latency than did those whose divorces were finalized. This suggests that prolonged emotional stress may put these subjects at some risk for a mood disorder.