The anxiolytic effects of buspirone and its major metabolite, 1-(2-pyrimidinyl)-piperazine (1-PP) have been investigated with a conflict (shock-induced suppression of drinking) paradigm in rats. Buspirone (10 mg/kg, p.o.) showed an anticonflict activity with a bell-shaped dose-response relationship without any effect on spontaneous water consumption. Higher doses of buspirone reduced the punished response. Diazepam (20 and 40 mg/kg, p.o.) also showed an anticonflict activity in a dose-dependent manner, but animals with diazepam showed an increase in spontaneous water consumption at these doses. On the other hand, 1-PP (6.25-200 mg/kg, p.o.) showed a weak anticonflict activity with a significant effect at 25 mg/kg without any effect on spontaneous water consumption. In the 7-day treatment test, buspirone (5 and 10 mg/kg, p.o.), 1-PP (5 and 25 mg/kg, p.o.) and diazepam (10 and 40 mg/kg, p.o.) did not develop the tolerance to the anticonflict activity. Conversely, the anticonflict activity of diazepam was increased by the repeated treatment. Diazepam (10 mg/kg, p.o.) showed an anticonflict activity without any effect on spontaneous water consumption in this test. These results demonstrated that buspirone clearly exhibited an anticonflict effect similar to that of diazepam in a Vogel-type conflict test, and its real anxiolytic effect may not be always based on 1-PP, the main metabolite of buspirone.