Triazolam reportedly causes phase advances in hamster wheel-running rhythm after injection during subjective daytime. However, it is unclear whether benzodiazepine affects the PER: gene expression accompanying a behavioural phase shift. Brotizolam (0.5 - 10 mg kg(-1)) induced large phase advances in hamster rhythm when injected during mid-subjective daytime (circadian time 6 or 9), but not at circadian time 0, 3 or 15. Brotizolam (5 mg kg(-1)) significantly reduced the expression of PER:1 and PER:2 in the suprachiasmatic nucleus 1 and 2 h after injection at circadian time 6, and slightly reduced them at circadian time 20. Injection of 8-OH-DPAT (5 mg kg(-1)) at subjective daytime induced similar phase advances with a reduction of PER:1 and PER:2 expression. Co-administration of brotizolam with 8-OH DPAT failed to potentiate the 8-OH DPAT-induced phase advances and reduced PER: expression. Both phase advance and rapid induction of PER:1 and PER:2 in the suprachiasmatic nucleus after light exposure (5 lux, 15 min) at circadian time 20 was strongly attenuated by co-treatment with brotizolam 5 mg kg(-1). The present results strongly suggest that reduction of PER:1 and/or PER:2 expression during subjective daytime by brotizolam may be an important step in causing a behavioural phase advance. The co-administration experiment suggests that common mechanism(s) are involved in brotizolam- or 8-OH DPAT-induced phase advances and the reduction of PER: gene expression. These results suggest that brotizolam is not only a good drug for insomnia but also a drug capable of facilitating re-entrainment like melatonin.