The potential role of autoreceptors in regulating the activity of serotonin-containing nucleus raphe dorsalis (RD), raphe medianus (RM) and raphe pallidus (RPA) neurons was examined by recording the activity of these neurons under a variety of conditions both in vivo and in vitro. Raphe neurons recorded in vivo displayed the characteristic slow, rhythmic discharge pattern previously described for rat and cat raphe cells. The activity of these neurons was suppressed in a dose-dependent manner by tryptophan, LSD and chlorimipramine administered intravenously. There were no significant changes in the spontaneous discharge rate of raphe neurons over time when recorded in vitro, even though tissue serotonin and its metabolite, 5-hydroxyindoleacetic acid, decreased dramatically. RPA neurons fired significantly faster than either RD or RM neurons both in vivo and in vitro. Prior depletion of brain serotonin by p-chlorophenylalanine administration resulted in no significant change in raphe unit activity recorded in vitro. Elevation of brain serotonin by monoamine oxidase inhibition produced a total inhibition of raphe unit activity in vitro. Similarly, increasing the concentration of serotonin in the tissue slice by adding serotonin directly to the incubation medium resulted in a profound, though transitory, depression of unit activity. This depressant effect of serotonin was rapidly reversible upon drug wash-out. Serotonin receptor blockers, methiothepin, cypoheptadine, and methysergide, produced no significant change in unit activity. The serotonin reuptake blocker, fluoxetine, produced a total inhibition of raphe unit activity in all three nuclei in vitro. These data suggest that excess serotonin suppresses the activity of raphe neurons, apparently by an action on autoreceptors, but that a deficiency, or normal concentration, of serotonin does not influence the spontaneous activity of these cells. The data also show that RD and RM are much more sensitive to the depressant effects of serotonin than the caudal RPA neurons. More generally, these studies provide a data base for examining the electrophysiological and pharmacological characteristics of serotonergic neurons in the three major serotonin-containing nuclei in mouse brain. The mouse has proven to be a much easier species than the rat to use in these types of studies, based on the finding that mouse brain slices are more viable in vitro than are rat brain slices.