The effects of a short-acting benzodiazepine hypnotic, triazolam (0.5 mg), on sleep, performance, and arousal threshold were assessed in 20 male poor sleepers (age 21 +/- 2.37 years). Following in a laboratory screening night, all subjects received placebo for 3 nights (single-blind), ten received triazolam and ten placebo for 6 nights (double-blind), and all received placebo on 2 withdrawal nights (single-blind). All effects described below were statistically significant. Triazolam reduced sleep latency and increased total sleep time and sleep efficiency. Percent Stage 2 was increased and percent Stage 4 was reduced during treatment. Morning performance, measured 8.25 h post-drug, showed no decrements. Acute effects were assessed on treatment night 6 during arousals from sleep at 1.5, 3, and 5 h post-administration: performance was impaired in triazolam subjects on the Wilkinson 4-Choice Reaction Time Test, Digit Symbol Substitution Test, Williams Word Memory Test, and Card Sorting Task. In the morning following treatment night 6, long-term memory was tested using a recognition task requiring subjects to identify words presented during night-time test batteries: triazolam subjects correctly identified fewer target words. Triazolam administration produced anterograde amnesic effects. However, in a Paired Associates Test learned prior to drug ingestion on the previous evening, triazolam did not impair morning recall of word pairs. Threshold for arousal from slow wave sleep was elevated during treatment, and triazolam subjects did not show increased sensitivity to the arousing tone over nights as did placebo subjects.