The effect of lesion of 5-hydroxytryptamine (5-HT) neurons, produced by p-chloroamphetamine (p-CA; 2 x 10 mg/kg), and the influence of the benzodiazepine receptor antagonist flumazenil (10 mg/kg) on the anticonflict action of N-[3,5-dichloro-2-(methoxy)phenyl]-4-(methoxy)-3-(1-piperazinyl)benzenesulfonamide (SB-399885), a selective 5-HT6 receptor antagonist, were investigated in the Vogel conflict drinking test in rats. In addition, the interaction between SB-399885 (0.3 mg/kg) and diazepam (2.5 mg/kg) was evaluated in that test. All the compounds tested were administered intraperitoneally. The anticonflict activity produced by SB-399885 (3 mg/kg) was not modified in p-CA-pretreated rats, but it was totally blocked by flumazenil. Combined administration of non-active doses of SB-399885 (0.3 mg/kg) and diazepam (2.5 mg/kg) produced a pronounced anticonflict effect in rats. The present results suggest that the anticonflict activity of SB-399885 is not conditioned by the integrity of 5-HT neurons, and that benzodiazepine receptors are indirectly involved in its effect, possibly due to a functional interaction between 5-HT6 receptors and the gamma-aminobutyric acid (GABA)/benzodiazepine system.