Polygraphic sleep recordings were compared between patients with primary major depression (MDD), patients with primary alcoholism and secondary MDD, and normals. Patients differed significantly from normals on the following measures: both patient groups showed short REM latency, and REM latency corrected, along with prolonged sleep latency. Secondary depressives differed from controls on several other measures: sleep onset time, total sleep time, delta sleep, REM percent, stage one sleep, stage three sleep, non-REM sleep, stage three and delta sleep in the first non-REM period. Prior research has shown a decrease in non-REM sleep and total sleep time in alcoholic patients who are not currently depressed, and short REM latency in patients with MDD. Thus, our findings suggest an additive effect of two disorders known to affect sleep: alcoholism and depression.