1. Benzodiazepines (BZ) and barbiturates both potentiate chloride currents through GABA(A) receptors to enhance inhibition. However, unlike barbiturates BZ do not impair autonomic control of heart rate. We hypothesised that BZ might not significantly potentiate GABAergic transmission in the caudal nucleus of the solitary tract (cNTS), which is critically important for mediating the baroreceptor reflex. 2. In rat brain slices the BZ agonists chlordiazepoxide and midazolam (2 and 50 microM) did not significantly enhance currents evoked by GABA in voltage-clamped cNTS neurones. Chlordiazepoxide (50 microM) reversibly increased electrically evoked IPSPs in 5/10 rostral NTS (rNTS) neurones but only in 2/10 cNTS neurones. Pentobarbitone (50-100 microM) was effective in enhancing GABA(A)-mediated responses in all NTS neurones. An inverse BZ agonist, methyl 6,7-dimethoxy-4-ethyl-beta-carboline-3-carboxylate (DMCM; 1 or 10 microM), failed to depress GABA-induced currents in the cNTS. 3. Microinjections of midazolam (10 and 100 microM solutions) into the cNTS did not affect the baroreceptor reflex (P > 0.2) while pentobarbitone (100 microM) significantly and reversibly depressed it (gain decrease to 53 +/- 11 % of control, P < 0.01). 4. Reverse transcriptase polymerase chain reaction revealed the presence of alpha(1), alpha(2), beta(2), beta(3) and gamma(2) GABA(A) receptor subunit mRNA in the cNTS. No alternatively spliced variants of the alpha(1)- and gamma(2)-subunits were revealed. Moreover, GABA(A) epsilon-unit mRNA was found in both the cNTS and rNTS as two alternatively spliced transcripts. 5. Immunocytochemical analysis revealed numerous GABA(A) epsilon-subunit-positive neurones within the cNTS with significantly fewer epsilon-subunit-positive cells in the rNTS. 6. As incorporation of the epsilon-subunit in recombinant GABA(A) receptors may confer BZ insensitivity we propose that the paucity of BZ actions in the cNTS is due to a high level of epsilon-subunit expression. This is the first demonstration of a possible physiological impact of the epsilon-subunit on native GABA(A) receptors.