In order to determine whether sleep onset per se might precipitate the sleep growth hormone (GH) response, sleep EEG-EOG and plasma GH characteristics were studied in 15 healthy young males (AM) sleeping from 8 AM to 10 AM, when REM sleep is known to predominate, and in 14 subjects (PM) sleeping from 4 PM to 6 PM, when slow-wave sleep (SWS) predominates. PM subjects obtained significantly more SWS and less REM sleep than AM subjects. There was wide individual variability in the level and timing of the GH rise during the naps, but GH release was significantly greater during PM naps than during AM naps for the groups as a whole. The difference between conditions reflected primarily the fact that more subjects exhibited frank GH peaks during PM naps than during AM naps. These results indicate that the occurrence of the sleep GH response is not dependent upon sleep onset, and confirm previous reports of an association between the sleep GH response and SWS.