Eighteen normal-sleeping young (mean age 25.6 years) volunteers received either ethanol (0.75 g/kg producing blood ethanol concentrations of 71.1 +/- 24.3 mg/100 ml on average) or caffeine (4.0 mg/kg dissolved in 300 ml of 97% caffeine-free instant coffee) at 0920-0950 h after spending 5, 8, or 11 h time in bed (TIB) the previous night. Latency to sleep onset was tested at 1000, 1200, 1400, and 1600 h. Mean sleep latency differed significantly between drugs on each day of testing, with subjects being sleepier after ethanol than caffeine. On day 2 the TIB manipulation produced significant differences in latency, with the 11-h condition differing from both the 8- and 5-h conditions. The significant interaction revealed that in fully rested subjects (11-h TIB), ethanol did not produce sleepiness to the degree it did after 5 or 8 h in bed. In this condition latencies were similar to those of the caffeine and 5- or 8-h TIBs.