The effects of phenobarbital, clonazepam, valproic acid, ethosuximide and phenytoin were examined in pigeons performing under a repeated acquisition procedure. Clonazepam (0.06-0.75 mg/kg), valproic acid (40-120 mg/kg), ethosuximide (40-160 mg/kg) and phenytoin (2.5-15 mg/kg) produced generally dose-dependent decreases in rate of responding, whereas phenobarbital (5-50 mg/kg) had little consistent effect on response rate across the dose range studied. Phenobarbital and clonazepam produced dose-dependent increases in error rates. Although valproic acid and phenytoin generally increased errors relative to control values, this effect was not directly dose-dependent or consistent across subjects. A within-session analysis of the distribution of drug-induced increases in errors revealed that the main effect of phenobarbital, clonazepam, valproic acid and phenytoin was to increase errors during early acquisition (i.e., before the procurement of 15 or fewer reinforcers). Later in the session, a similar number of errors per reinforcer was made during drug and control sessions. In contrast to the other anticonvulsants examined, ethosuximide had little effect on error rates. These results suggest that there are qualitative as well as quantitative differences in the effects of anticonvulsant drugs under the repeated acquisition procedure.