The Meyer-Overton hypothesis, predicting that the potency of an anesthetic correlates with its affinity for lipid, is a cornerstone of modern anesthetic theory. Several halogenated compounds were recently found to deviate from this prediction, whereas others did not. We tested the abilities of enflurane and five of these compounds to potentiate gamma-aminobutyric acid (GABA)A receptor responses in Xenopus oocytes expressing alpha 1 beta 2 or alpha 1 beta 2 gamma 2S GABAA receptors. Enflurane and the anesthetic 1-chloro-1,2,2-trifluorocyclobutane (F3) strongly potentiated chloride currents produced by 5 microM GABA with both alpha 1 beta 2 and alpha 1 beta 2 gamma 2S receptors. This potentiation decreased as the GABA concentration was raised. The transitional compound (less potent than predicted by its lipid solubility) 2-bromoheptafluoropropane produced modest enhancement, whereas three nonanesthetics (neither causing anesthesia in vivo nor decreasing the requirement for known anesthetics), 1,2-dichlorohexafluorocyclobutane, 2-chloroheptafluoropropane, and 2,3-chlorooctafluorobutane, did not affect GABAA receptor currents. Although all five compounds were predicted to be anesthetics by the Meyer Overton hypothesis, only F3 behaved as an anesthetic in vivo and only F3 markedly potentiated GABAA receptor responses in oocytes. These results strongly implicate the GABAA receptor in general anesthesia. Fluorescence polarization studies showed that anesthetics (enflurane and F3), but not nonasthetics (1,2 dichlorohexafluorocyclobutane and 2,3-chlorooctafluorobutane) disordered membrane lipids. Thus, for the compounds studied actions on both GABAA receptor function and lipid order distinguish between anesthetics and nonanesthetics.