In order to investigate the role of the 5-HT1A receptors of the amygdala in modulating anxiety, rats were implanted with bilateral cannulae aimed at the basolateral nucleus of the amygdala complex and infused with either artificial cerebrospinal fluid (aCSF) or the selective 5-HT1A receptor agonist 8-OH-DPAT (50-200 ng) and tested in two animal models of anxiety. In the elevated plus-maze test, no significant effects were detected in this dose range. In contrast, 8-OH-DPAT caused an overall reduction in levels of social investigation, thus indicating anxiogenic actions in the social interaction test. At 50 ng, 8-OH-DPAT had a selective action on anxiety, while at 200 ng there was a concomitant reduction in locomotor activity and, in some animals, signs of the 5-HT1A syndrome. Evidence that the anxiogenic effect of 8-OH-DPAT (50 ng) was due to activation of 5-HT1A receptors came from the finding that (-)-tertatolol, a 5-HT1A receptor antagonist, reversed this effect at a dose (1.5 micrograms) which was silent when given alone. The benzodiazepine receptor agonist, midazolam (1 and 2 micrograms) was bilaterally administered into the basolateral nucleus of the amygdala and evoked clear-cut anxiolytic effects in the social interaction test. These data indicate that the agonist activation of post-synaptic 5-HT1A receptors in the basolateral nucleus of the amygdala may produce anxiogenic effects, while agonist activation of BDZ receptors in the same areas evokes anxiolytic effects. Our results from the social interaction test are similar to those previously reported from tests of anxiety using punished paradigms, but contrast with those found in the elevated plus-maze. Thus, it is concluded that either the two tests have different sensitivities to midazolam and 8-OH-DPAT or more intriguingly, the tests are evoking fundamentally different states of anxiety, with that evoked by the plus-maze being mediated via brain areas or receptors different from those studied here.