This study was undertaken to compare the effects of 0.25 mg of brotizolam, 15 mg of flurazepam, and placebo on the sleep and performance of elderly subjects with chronic insomnia during a 2-week period of administration. Thirty-six male and female subjects who ranged in age from 60-72 years were divided into three treatment groups. All groups received placebo on the first three study nights, the active drug or placebo on the next 14 nights, and placebo again on the two following withdrawal nights. Sleep was assessed by means of questionnaires, and residual effects during the day were studied by means of the multiple sleep latency test and a variety of memory, performance, and vigilance tests. Sleep improved with all treatments. Rebound insomnia was noted on brotizolam withdrawal; flurazepam withdrawal had a milder impact. At the end of this 19-night study, only the placebo-treated group was sleeping significantly longer than at baseline. Both drug treatments increased daytime sleepiness and impaired performance on the first day after their administration. These effects waned after 2 weeks of treatment with brotizolam, but not flurazepam. The results of this study affirm the increased sensitivity of elderly subjects to benzodiazepine hypnotics and their indication for acute or intermittent insomnia, rather than for the more chronic forms of this disorder.