This study compared the effects of the beta-carboline anxiolytic, abecarnil, with other benzodiazepine receptor (BZR) ligands, including the full agonists diazepam and alprazolam, and the partial agonists ZK 95962 and bretazenil (Ro 16-6028), and alpidem, in the mouse four-plate test and plus-maze. The efficacy and potency of each compound was related to the fraction of BZR occupied by the drug. Abecarnil was efficacious in both tests and showed anxiolytic effects comparable with alprazolam and diazepam. In the four-plate test, abecarnil, bretazenil, and ZK 95962 had selective effects on releasing exploratory locomotor activity suppressed by footshock (punished crossings). None of these compounds significantly altered non-punished crossings. In contrast, diazepam and alprazolam increased both unpunished and punished crossings at low to medium doses (receptor occupancies of approximately 20-60%). The number of punished and unpunished crossings fell to control levels or below at higher, more sedative doses (approximately 80% receptor occupancy). Alpidem had very weak anxiolytic-like effects in this test and markedly reduced unpunished crossings at relatively low receptor occupancies (> 15%). In the plus-maze, abecarnil increased the time spent in the open arms and the percentage open arm entries to an extent equal to that observed following diazepam or alprazolam administration. Bretazenil and ZK 95962 had weak effects on the measures of anxiolytic activity in this test. Alpidem also had little anxiolytic-like activity in the plus-maze but markedly reduced the total number of arm entries.(ABSTRACT TRUNCATED AT 250 WORDS)