A great need exists for antipsychotic drugs which will not induce extrapyramidal symptoms (EPS) and tardive dyskinesias (TDs). These side effects are deemed to be a consequence of nonselective blockade of nigrostriatal and mesolimbic dopamine D2 receptors. Nondyskinetic clozapine (1) is a low-potency D2 dopamine receptor antagonist which appears to act selectively in the mesolimbic area. In this work dopamine antagonism was assessed in two mouse behavioral assays: antagonism of apomorphine-induced climbing and antagonism of apomorphine-induced disruption of swimming. The potential for the liability of dyskinesias was determined in haloperidol-sensitized Cebus monkeys. Initial examination of a few close cogeners of 1 enhanced confidence in the Cebus model as a predictor of dyskinetic potential. Considering dibenzazepines, 2 was not dyskinetic whereas 2a was dyskinetic. Among dibenzodiazepines, 1 did not induce dyskinesias whereas its N-2-(2-hydroxyethoxy)ethyl analogue 3 was dyskinetic. The emergence of such distinctions presented an opportunity. Thus, aromatic and N-substituted analogues of 6-(piperazin-1-yl)-11H-dibenz[b,e]azepines and 11-(piperazin-1-yl)dibenzo[b,f][1,4]thiazepines and -oxazepines were prepared and evaluated. 11-(4-[2-(2-Hydroxyethoxy)ethyl]piperazin-1-yl)dibenzo[b,f][1,4]thiazepine (23) was found to be an apomorphine antagonist comparable to clozapine. It was essentially nondyskinetic in the Cebus model. With 23 as a platform, a number of N-substituted analogues were found to be good apomorphine antagonists but all were dyskinetic.