The 5-HT2C antagonist SB-242084 was examined in various anxiety tests at doses based on reversal of mCPP-induced hypoactivity (0.1--3 mg/kg ip). In the elevated plus-maze task, SB-242084 exhibited signs of anxiolysis (time spent, distance travelled, and entries into open arms), but this was potentially confounded by its general increase of locomotion; alprazolam selectively affected open-arm parameters. In a Geller--Seifter conflict test, SB-242084 produced a modest, nonsignificant increase in punished responding compared to the significant effect produced by diazepam. None of the treatments significantly affected unpunished responding. In the conditioned emotional response (CER) test, SB-242084 produced an increase in the suppression ratio (SR, smaller than diazepam). Since this 5-HT2C antagonist also increased lever pressing, an additional test was conducted with amphetamine that stimulated lever pressing but, nonetheless, failed to produce any change in SR. In the fear-potentiated startle task, SB-242084 was inactive in comparison to a significant effect of diazepam. The previously described reduction of schedule-induced polydipsia by fluoxetine and 5-HT2C receptor agonist Ro60-0175 was attenuated by SB-242084 pretreatment, however, the latter compound exhibited a potent increase in polydipsia when given alone. The present results demonstrate an anxiolytic potential of SB-242084, as well as an intrinsic response-enhancing property, however, both of these effects are task dependent.