We examined the effects on sleep and memory of a nighttime dose of triazolam, 0.125 mg, in seven subjects with Alzheimer's disease (AD) who were reported by caregivers to be frequently up at night. Subjects were admitted to an intermediate care hospital ward for the 8-day ABA design protocol (placebo baseline-drug-placebo washout). Drug or placebo was given each evening at 2100 h. Sleep was assessed with a wrist-worn activity monitor. Memory was evaluated using a computerized delayed-matching-to-sample (DMTS) task administered at 0800 and 2130 h. Triazolam had no significant effects on total sleep time at night, latency to sleep onset, number of arousals, or time asleep during the day. DMTS performance was significantly worse at night compared to morning during baseline, but there were no significant drug effects. Our results suggest the standard geriatric dose of triazolam, 0.125 mg, may not be an effective hypnotic in AD patients with disrupted sleep, but neither does it substantially worsen the recent memory deficits of AD.