The effects of a phenothiazine, on sleep in ten healthy volunteers were investigated in two double blind polygraphic studies. The first part consisted of a single dose study with promethazine 50, 100, and 200 mg, using pentobarbital 100 mg as a reference substance. In the second part, four subjects spent 20 consecutive nights with nine drug nights (promethazine 100 mg), followed by a placebo withdrawal period of six nights, in the sleep laboratory. Promethazine showed a dose related REM-depressing effect with a greater decrease, the higher the dose. The placebo value was 20.7% REM of total sleep time and gave 16.3, 13.5 and 11.4 respectively, for promethazine 50, 100 and 200 mg. Pentobarbital gave 16.2% REM. There was also an increase of stage II and with the highest dose an increase of stage III + IV. An increase of REM-latency together with a decrease of REM-periods was also seen, and while pentobarbital gave a decrease in REM-density, promethazine did not cause any changes in the phasic REM-component. A REM-rebound was seen in the first night of withdrawal with an increase of per cent REM from 19.9%-25.1%. The mean for the whole withdrawal period was 23.1%. Promethazine in the highest dose, 200 mg, gave drowsiness and hangover effects in 14 nights out of 20. The REM-depressing effect of promethazine together with its relatively weak REM-rebound effect may explain its value in the treatment of withdrawal symptoms following abuse of alcohol and barbiturates.