1. The effects of a series of nine GABAB receptor antagonists of widely varying potencies on electrically stimulated release from cortical slices of [3H]-GABA in the absence or presence of 10 microM of the GABAB agonist, (-)-baclofen and of endogenous glutamate in the presence of (-)-baclofen were compared. 2. The concentrations of the compounds half maximally increasing [3H]-GABA release (EC50's) at a stimulation frequency of 2 Hz correlated well with the IC50 values obtained from the inhibition of the binding of the agonist, [3H]-CGP 27492, to GABAB receptors in rat brain membranes (rank order of potency: CGP 56999 A > or = CGP 55845 A > CGP 52432 > or = CGP 56433 A > CGP 57034 A > CGP 57070 A > or = CGP 57976 > CGP 51176 > CGP 35348). 3. Likewise, the concentrations causing half-maximal increases of [3H]-GABA in the absence or presence of (-)-baclofen, and of endogenous glutamate in the presence of (-)-baclofen, correlated well with each other. Reports in the literature suggesting the CGP 35348 exhibits a 70 fold preference for inhibition of (-)-baclofen's effects on glutamate over [3H]-GABA release, and that CGP 52432 shows a 100 fold preference in the opposite sense, could not be confirmed in our model. 4. Therefore, our results suggest that, if there are pharmacological differences between GABAB autoreceptors and GABAB heteroreceptors on glutamatergic nerve endings in the rat cortex, they are not revealed by this series of compounds of widely different potencies. 5. In particular, our results with CGP 35348 and CGP 52432 do not support the hypothesis that GABAB autoreceptors and GABAB heteroreceptors on glutamatergic nerve endings represent subtypes with different pharmacology.