Sleep disturbances and need for hypnotics are disproportionately greater in the elderly compared to younger adults. The present study provides a placebo-controlled subjective hypnotic efficacy of zolpidem, triazolam, and temazepam in elderly insomniacs. After a single-blind placebo screening week (DSM-III-R criteria), 335 elderly insomniacs (ages 60 to 85) were randomized to 28 days of double-blind treatment with zolpidem 5 mg, triazolam 0.125 mg, temazepam 15 mg, or placebo, followed by a 4-day single-blind, placebo withdrawal period. The primary efficacy parameters were self-reported sleep latency (SSL) and self-reported total sleep duration (SSD); they were measured by responses on daily Morning Questionnaires. SSL values were compared by Cox proportional hazards regression technique. SSD values were compared by ANOVA. Compared to placebo, zolpidem and temazepam produced significantly shorter SSL over the 4 treatment weeks, but triazolam did not. In the zolpidem group, SSL was significantly shorter than in the placebo group at all four treatment weeks; in the temazepam group, SSL was significantly shorter than in the placebo group at weeks 1, 3, and 4. SSD was increased above baseline levels in all groups. No tolerance to the subjective effects or rebound above baseline levels occurred in any of the treatment groups. Overall, the drugs were well tolerated. No difference was found among the placebo and treatment groups in overall adverse event incidence rates. However, compared with zolpidem and placebo, temazepam produced significantly higher incidences of drowsiness and fatigue, and triazolam produced a significantly higher incidence of nervousness than zolpidem.