Anxiolytic drugs, such as the benzodiazepines and the azapirones (ipsapirone, gepirone, buspirone), are well known to affect states of vigilance and to decrease the firing rate of serotoninergic neurones within the dorsal raphe nucleus in rats. In order to examine whether the newly developed 5-HT3 antagonists with potential anxiolytic properties act through similar mechanisms, the effects of several of such antagonists: MDL 72222, ICS 205-930, ondansetron and/or zacopride on both sleep-wakefulness and the discharge of serotoninergic neurones within the dorsal raphe nucleus were investigated in rats. When tested in a wide range of doses (0.05-10 mg/kg, i.p.), none of these drugs significantly affected the states of vigilance, except ondansetron, at 0.1 mg/kg, which increased paradoxical sleep for the first 2 hr after administration and MDL 72222, at 10 mg/kg, which reduced both paradoxical and slow wave sleep and increased wakefulness for the same initial period after treatment. In vivo, in chloral hydrate anaesthetized rats, as well as in vitro, in slices of brain stem, none of the 5-HT3 antagonists tested affected the firing rate of serotoninergic neurones. Similarly, no change in the electrical activity of serotoninergic neurones could be evoked in vitro by superfusion of the tissue with the 5-HT3 agonists, phenylbiguanide (10 microM) and 2-methyl-5-HT (1 microM). At a larger concentration (10 microM), the latter compound reduced the neuronal discharge probably through the stimulation of somatodendritic 5-HT1A autoreceptors since this effect, as that of ipsapirone, could be prevented by 10 microM l-propranolol. Comparison of these data with those obtained with benzodiazepines and 5-HT1A agonists of the azapirone series, supports the concept that different mechanisms are responsible for the anxiolytic-like properties of 5-HT3 agonists, compared to those of other anxiolytic drugs.