Alprazolam was evaluated in chronic insomniacs in a 1-mg bedtime dose. The 16-night sleep laboratory protocol included four placebo-baseline nights followed by seven nights of drug administration and five placebo-withdrawal nights. On the first three drug nights (nights 5 to 7), the drug was highly effective in inducing and maintaining sleep with this short-term use. By the end of the one week of administration (nights 9 to 11), however, the drug had lost about 40% of its efficacy. During drug use, one subject reported some difficulty in controlling expression of inappropriate emotions when interacting with others, which suggested the presence of disinhibition. On the third night following drug termination, there was a significant increase in sleep difficulty above baseline levels (rebound insomnia). This worsening was of comparable magnitude to the peak improvement of sleep with drug administration. Thus, the clinical utility of alprazolam when administered to insomniac patients appears to be limited because of a relatively rapid development of tolerance and possible disinhibitory reactions during drug use and the occurrence of rebound insomnia following withdrawal.