Human growth hormone and saline were administered for one night each to normal volunteers in a cross-over study. A dose of 2 units im given 15 min before bedtime had no effect on sleep EEG parameters. In contrast, 5 units resulted in a 19% decrease in slow-wave sleep (p < 0.01) and a 13% increase in REM sleep (p < 0.05). Neither doe, when given during daytime, affected tests of affect or serial learning.