We have found that the partial inverse benzodiazepine agonists Ro 15-4513 and FG 7142 antagonize the depressant electrophysiological effects of locally applied ethanol in the cerebellum. Although absolute tissue concentrations are not known, dose-response curves constructed using pressure-ejection doses as previously described we found that FG 7142 was more efficacious, but less potent than Ro 15-4513. Our observation that ethanol and inverse benzodiazepine agonists have interactions which are not competitive might suggest that these two drugs act through separate, but interactive mechanisms in order to produce the observed ethanol antagonism. If such independent interactions were mediated at different sites on a given macromolecular complex, such as the GABAa/Cl- channel, then one might expect to find allosteric interactions between those sites as well as with the functional response of the complex to GABA activation. Indeed, this hypothesis is consistent with the recent finding of Harris and collaborators that ethanol potentiates the inverse agonist actions of Ro 15-4513 and FG 7142. On the other hand, we were unable to find large ethanol-induced potentiations of GABA effects on all neurons which showed depressant responses to ethanol administration in rat cerebellum. However we did find that the GABAa antagonist, bicuculline, blocks the depressant effects of ethanol on the same neurons. We conclude that the interaction between ethanol and GABA probably does not occur directly at the GABAa receptor site, but that the GABAa mechanism does play a permissive role in the ethanol-induced depressions of cerebellar Purkinje neurons.(ABSTRACT TRUNCATED AT 250 WORDS)