The effect of triazolam (0.25 mg) and placebo was investigated in healthy, male subjects who slept in a sitting position. After the intake of placebo, sleep efficiency, rapid eye movement (REM) sleep and subjective sleep quality were lower than in the preceding sleep episode in bed, while stage 1 and REM sleep latency were higher. Triazolam did not prevent this impairment of sleep. However, in comparison with the placebo condition, the percentage of slow wave sleep was higher in the first third of the night, and in the morning sleep was rated as more quite. EEG power density in nonREM sleep was reduced in the frequency range of 1.25-10.0 Hz and enhanced in the range of sleep spindles (12.25-13.0 Hz). These changes were still present in the last third of the night. In REM sleep, triazolam reduced spectral activity in some frequency bins between 4.25 and 10.0 Hz. The sitting position itself affected the nonREM sleep spectra, since the placebo level in the 2.25-21.0-Hz range exceeded the baseline level. We conclude that a 0.25 mg dose of triazolam does not effectively counteract a posture-induced sleep disturbance, but induces changes in the EEG spectra which are typical for benzodiazepine receptor agonists.