In a recent study, we reported the presence of neurones expressing 5-HT2B receptor protein in the medial amygdaloid nucleus of the adult rat brain. In the present study, bilateral micro-injection of the 5-HT2B receptor agonist 1-[5-(2-thienylmethoxy)-1H-3-indolyl]propan-2-amine hydrochloride (BW 723C86, 0.09 and 0.93 nmol, 5 min pretest) into the medial amygdaloid nuclei increased the total interaction time of a pair of male rats in the social interaction test, to a comparable extent to chlordiazepoxide (5 mg/kg p.o., 1 hr pretest) without altering locomotor activity; indicative of anxiolytic activity. The increase in social interaction was prevented by pretreatment with the 5-HT2C/2B receptor antagonist N-(1-methyl-5-indoyl)-N′-(3-pyridyl) urea hydrochloride (SB 200646A, at 2 but not 1 mg/kg p.o., 1 hr pretest), which did not alter behaviour when given alone.Intra-amygdala BW 723C86 (0.09, 0.31 and 0.93 nmol, 5 min pretest) did not significantly alter the number of punished responses made when the same rats were examined seven days later in a Vogel punished drinking test, although chlordiazepoxide (5 mg/kg p.o., 1 hr pretest) produced the expected anxiolytic profile.The results are consistent with the proposal that activation of 5-HT2B receptors in the medial amygdala induces anxiolysis in the social interaction model but has little effect on behaviour in a punished conflict model of anxiety. These data suggest that serotonergic neurotransmission in this nucleus may selectively affect specific kinds of anxiety generated by different animal models. © 1997 Elsevier Science Ltd.