The short-acting benzodiazepine triazolam has been shown to cause phase-dependent phase shifts of the circadian activity rhythms of squirrel monkeys maintained in constant light. The present study sought to determine whether properly timed triazolam administration can also accelerate the reentrainment of circadian rhythms following phase shifts of the daily light-dark (LD) cycle. Circadian rhythms of telemetered body temperature and locomotor activity were recorded from squirrel monkeys exposed to an 8-h phase advance of the LD cycle, followed 16 days later by an 8-h phase delay. On the day of the phase advance, each animal received a single injection of triazolam (0.3 mg) or of vehicle alone in midsubjective day (circadian time 6 [CT6], where CT0 represents the time of light onset and the beginning of subjective day, and CT12 the time of dark onset and the beginning of subjective night]), 2 h after the new time of dark onset, while on the day of the phase delay the animals received triazolam or vehicle in late subjective night (CT20), just before dark onset. This procedure was then repeated, giving triazolam to animals that had previously received vehicle alone, and vehicle to animals that had received triazolam. The daily acrophases of the temperature and activity rhythms were calculated by cosinor analysis, and exponential functions were fitted to the acrophases that followed each of the phase shifts. The rates of reentrainment of the temperature and activity rhythms, defined as the time required for the exponential functions to reach 90% of the new steady-state phase of entrainment, were slower after the phase advance than after the phase delay, but did not differ significantly between drug and vehicle conditions. Since triazolam was administered either in darkness or just before dark onset, these results suggest that the phase shifting effects of triazolam in squirrel monkeys may be entirely light-dependent.