The predictive validity of catalepsy as a rodent model for detecting the extrapyramidal side effects (EPS) of antipsychotic drugs was recently questioned when the novel antipsychotic savoxepine produced little catalepsy in rodents while producing significant EPS in schizophrenic patients. Because catalepsy is viewed as an important model for predicting EPS, we decided to re-evaluate the effects of savoxepine. Savoxepine, clozapine, haloperidol, olanzapine, ORG 5222, raclopride, and risperidone were examined in two tests for catalepsy (grid and bar tests) in male Sprague-Dawley rats. The ability to antagonize amphetamine-induced hypermotility was also examined, since this measure is believed to predict clinical efficacy. With the exception of clozapine, all drugs produced dose-dependent catalepsy in both tests. For each drug, the minimum effective dose for producing catalepsy was greater than or equal to the ED50 for antagonizing amphetamine-induced hyperactivity (defined as the dose producing a 50% reduction in hyperactivity). Clozapine resulted in the widest separation of effective doses in the catalepsy and activity models. Raclopride produced the next largest separation while the remaining drugs resulted in only a one- or two-fold dose separation between the two behavioral tests. The results with haloperidol and clozapine are consistent with the clinical effects of these drugs (severe versus mild EPS). The ratios of effective doses in catalepsy and activity for the remaining novel drugs are also consistent with preliminary clinical findings indicating some EPS with each of these compounds. Thus, catalepsy remains a suitable rodent model for detecting compounds with EPS liability in humans.