The effects of an ultra-rapidly eliminated hypnotic (midazolam 7.5-15.0 mg; mean elimination half-life ~2 h) and a rapidly eliminated hypnotic (brotizolam 0.125-0.25 mg; mean elimination half-life ~5 h) were studied on transient insomnia induced by sleeping in a reclining seat. Rest in the seat did not lead to delay in sleep onset, but there was increased wakefulness and drowsy sleep and less REM sleep. There were no differences in wakefulness or drowsiness between sleep with drugs in the seat and with placebo in bed, but with midazolam, though not with brotizolam, there was a reduction in REM sleep 300 min after sleep onset. Ultra-rapidly and rapidly eliminated compounds used in the management of transient insomnia should be given in doses that are as free as possible from central nervous system depression as indicated by suppression of REM sleep during the early part of the night. Low doses of ultra-rapidly eliminated drugs are indicated for sleep-onset insomnia and for short periods of sleep, while rapidly eliminated hypnotics with elimination half-lives of ~5 h have the potential to sustain sleep free of residual effects.