Although there is some evidence that compounds acting at 5-HT2 receptors show anxiolytic activity, little is known about the specific involvement of the different 5-HT2 receptor subtypes in the modulation of anxiety-related responses. In the present study, the behavioural effects of mianserin, a non-selective 5-HT2 receptor antagonist, MDL 100,907, a selective 5-HT2A receptor antagonist, and SB 206553, a selective 5-HT2B/2C receptor antagonist, were investigated in two rat (the Vogel drinking conflict and the elevated plus-maze tests) and two mouse (i.e. the mouse defense test battery (MDTB) and the light/dark choice test) models of anxiety. Diazepam was used as a positive control. In the Vogel drinking test, mianserin (10 mg/kg) and SB 206553 (3-30 mg/kg), but not MDL 100,907, increased punished responding. Similarly, mianserin (1 mg/kg) and SB 206553 (3-10 mg/kg), but not MDL 100,907, increased entries into the open arms of the elevated plus-maze. These effects are consistent with anxiolytic-like actions of mianserin and SB 206553, although the magnitude of the effects of these two compounds was less than those of diazepam. In addition, in the MDTB, the 5-HT2 antagonists did not clearly affect the defensive reactions of mice exposed to a rat stimulus and they failed to reverse the avoidance of the illuminated box in the light/dark choice test. These results indicate a lack of anxiolytic-like action of the compounds in mice. These behavioural profiles suggest that blockade of the 5-HT2A receptor may not reduce anxiety and demonstrate that 5-HT2B and/or 5-HT2C receptor subtypes may be primarily involved in the anxiolytic-like effects of mianserin and SB 206553 in rats.