In seven subjects, a large amount of alcohol, taken at bedtime, reduced the total duration of stage 1 electroencephalographic-rapid-eye-movement periods in the first 5 hours of sleep; a moderate amount of caffeine produced no significant change.