The effects of ibogaine were studied in 12 rats trained to perform in an 8-arm radial maze. In Phase I, the mean number of sessions to criterion and cumulative errors to criterion, as well as mean response rate, were determined for two groups of six animals in a task where only four arms were baited. Group 1 received a potentially neurotoxic dose of ibogaine (50 mg/kg IP administered twice, with approximately 8 h between injections), and group 2 received vehicle. Both groups had similar levels of performance, but ibogaine-treated subjects had a significantly lower rate of responding in the maze. During Phase II, subjects were given a range of doses of ibogaine 20 min prior to working in the maze. Ibogaine produced a dose-dependent decrease in response rate, but efficiency (% arms correct) was not affected. In Phase III, subjects were divided into the same groups as they had been in Phase I. Ibogaine (30 mg/kg, IP) or vehicle was administered immediately following daily sessions in the maze. Ibogaine-treated rats committed significantly fewer errors than those in the vehicle treated group. Thus, in the present study, ibogaine failed to produce any deleterious effects on either acquisition of a novel task or efficiency in a previously learned task.