Lorazepam, an anxiolytic drug, was evaluated in a 2-mg dose using a 16-night protocol including 7 nights of drug trial. Initially and with continued use the drug was moderately effective in inducing and maintaining sleep. Side effects included episodes of memory impairment and confusion in 2 subjects and group mean increases in daytime anxiety and tension with continued drug use. Following drug withdrawal, there was a marked and significant worsening of sleep above baseline levels (rebound insomnia) on the third night as well as significant increases in tension and anxiety the next day. The peak degree of withdrawal sleep disturbance was several times the peak degree of sleep improvement with drug administration.