The synthetic steroid anesthetic alphaxalone (3 alpha-hydroxy-5 alpha-pregnane-11,20-dione) was studied in two behavioral paradigms known to be sensitive to anxiolytic drugs. In an elevated plus maze, alphaxalone produced an anxiolytic profile, significantly increasing the percentage of entries made into the open arms as well as the percentage of time spent on the open arms. In the conflict test, alphaxalone (6 and 8 mg/kg) produced a significant dose-dependent increase in punished responding and a decrease (8 mg/kg) in unpunished responding. The pattern of responding was similar to that observed with the benzodiazepine agonist chlordiazepoxide (2-8 mg/kg). The increase in punished responding was not altered by the benzodiazepine antagonist Ro 15-1788 and only partially blocked by the picrotoxinin receptor ligand isopropylbicyclophospate (10 and 15 micrograms/kg). The gamma-aminobutyric acid agonists picrotoxin (1 mg/kg) and bicuculline (1 mg/kg) also failed to suppress the rate-increasing effects of alphaxalone in the conflict test. Chronic administration of alphaxalone for 1 week produced no tolerance to the anxiolytic behavioral effects. In addition, no changes in pain threshold were noted with alphaxalone (8 mg/kg) in the tail-flick analgesia test. These results suggest that the pharmacologic substrates for the anxiolytic actions of alphaxalone may be independent of either the benzodiazepine or picrotoxinin binding sites of the gamma-aminobutyric acid/benzodiazepine receptor complex.