These experiments tested the efficacy of the 5-hydroxytryptamine3 antagonist ondansetron (OND) in reversing various aspects of benzodiazepine withdrawal in rats. Three tests were used in which the benzodiazepine antagonist flumazenil was administered to rats receiving chronic administration of chlordiazepoxide. In one test, the elevated plus-maze, flumazenil produced a reduction in time spent in the open arms of the maze; OND completely reversed this effect of flumazenil in a dose-related fashion. However, OND failed to block the effects of the anxiogenic drug pentylenetetrazole (PTZ) in the elevated plus-maze. In a second test, rats were trained to discriminate PTZ. After chlordiazepoxide, flumazenil substituted for PTZ; OND failed to block flumazenil. In a third test, rats maintained on a chronic base line of chlordiazepoxide were trained to discriminate flumazenil from vehicle. In this discrimination, PTZ substituted for flumazenil, and pentobarbital blocked the flumazenil stimulus; OND, however, failed to block the flumazenil stimulus. In a separate set of experiments, OND also failed to reverse the suppression of responding produced in a conditioned emotional response paradigm. Thus, some data from the elevated plus-maze are consistent with the hypothesis that benzodiazepine withdrawal shares common effects with other stimuli known to be anxiogenic, and that OND blocks this aspect of withdrawal. However, all other data are inconsistent with the hypotheses that OND is anxiolytic or has efficacy in reversing benzodiazepine withdrawal. We suggest that ondansetron is likely to have minimal efficacy in humans for the treatment of sedative-hypnotic withdrawal.