The binding of [3H]bicuculline methochloride (BMC) to mammalian brain membranes was characterized and compared with that of [3H] gamma-aminobutyric acid ([3H]GABA). The radiolabeled GABA receptor antagonist showed significant displaceable binding in Tris-citrate buffer that was improved by high concentrations of chloride, iodide, or thiocyanate, reaching greater than 50% displacement in the presence of 0.1 M SCN-. An apparent single class of binding sites for [3H]BMC (KD = 30 nM) was observed in 0.1 M SCN- for fresh or frozen rat cortex or several regions of frozen and thawed bovine brain. The Bmax was about 2 pmol bound/mg of crude mitochondrial plus microsomal membranes from unfrozen washed and osmotically shocked rat cortex, similar to that for [3H]GABA. Frozen membranes, however, showed decreased levels of [3H]BMC binding with no decrease or an actual increase in [3H]GABA binding sites. [3H]BMC binding was inhibited by GABA receptor specific ligands, but showed a higher affinity for antagonists and lower affinity for agonists than did [3H]GABA binding. Kinetics experiments with [3H]GABA binding revealed that low- and high-affinity sites showed a similar pharmacological specificity for a series of GABA receptor ligands, but that whereas all agonists had a higher affinity for slowly dissociating high-affinity [3H]GABA sites, bicuculline had a higher affinity for rapidly dissociating low-affinity [3H]GABA sites. This reverse potency between agonists and antagonists during assay of radioactive antagonists or agonists supports the existence of agonist- and antagonist-preferring conformational states or subpopulations of GABA receptors. The differential affinities, as well as opposite effects on agonist and antagonist binding by anions, membrane freezing, and other treatments, suggest that [3H]BMC may relatively selectively label low-affinity GABA receptor agonist sites. This study, using a new commercially available preparation of [3H]bicuculline methochloride, confirms the report of bicuculline methiodide binding by Möhler and Okada (1978), and suggests that this radioactive GABA antagonist will be a valuable probe in analyzing various aspects of GABA receptors.