Ortho-methoxyphenylpiperazine (OMPP) and meta-substituted chlorophenylpiperazine (MCPP) blocked conditioned avoidance responding (CAR) in the rat (ED50 values = 5.6 (4.6, 7.3) and 2.4 (1.9, 2.9) mg/kg i.p. (95% confidence limits), respectively) without markedly altering escape responding. Since this test predicts antipsychotic efficacy, the piperazines were examined in radioligand binding assays and found to have no affinity for dopamine (DA) binding sites, but were active at serotonin binding sites. OMPP displaced ligands for the 5-HT1A binding site with high affinity (Ki = 9.5 (5.4, 17.9) nM) but was inactive at 5-HT2 sites (Ki greater than 1000 nM). MCPP, on the other hand, displaced ligands for 5-HT1, 5-HT1A and 5-HT2 binding sites with similar potencies (Ki values = 25 (3, 67), 23 (14, 40) and 40 (33, 48) nM, respectively). Pretreatment with metergoline (1.0 mg/kg i.p. -30 min) reduced MCPP- but not OMPP-induced block of CAR. OMPP, on the other hand, acted as a DA receptor antagonist in vivo blocking amphetamine-induced stereotyped behavior, whereas MCPP did not. Neither produced catalepsy even given in doses 8-10 times those required to block CAR. Insofar as these compounds lack antidopaminergic activity in vivo, yet are active in a test (CAR) predictive of antipsychotic activity in which DA receptor antagonists are active, they may be novel antipsychotic agents, or, perhaps, false positives in the CAR paradigm.