Auditory arousal thresholds of good (N = 12) and poor (N = 12) sleepers (sleep onset insomniacs) were obtained during stage 2, stage 4, and REM sleep at various times of the night. Despite claims of being "light" sleepers who are easily awakened by noise, poor sleeper auditory arousal thresholds were the same as those of good sleepers. Flurazepam (30 mg) increased the auditory arousal thresholds of poor sleepers (N = 6), but the increase was statistically significant only during the period of peak effect which occurred 1--2 hr after ingestion. Consistent with poor sleeper complaints of trouble falling asleep, the return to sleep (i.e., sleep latency) was significantly longer for poor than for good sleepers following stimulus arousals during the first stage 2 and first stage 4 periods of the night. Sleep latencies for good and poor sleepers did not differ significantly following subsequent arousals. The sleep latency following the first stage 2 stimulus arousal was significantly reduced in poor sleepers during flurazepam-induced sleep.