Rats were treated for 10 days with haloperidol (0.2 or 3 mg/kg/day intraperitoneally [IP]), with either S(+) or R(-) N-n-propylnorapomorphine (NPA; 3 mg/kg IP three times daily) or saline as a placebo control. Brain was microdissected into 11 regions for radioimmunoassay of neurotensin (NT)-like activity under coded ("blind") conditions. Concentrations of NT in rat brain regions ranked central amygdaloid nucleus greater than ventral bed nucleus greater than ventral tegmentum greater than dorsal bed nucleus greater than arcuate nucleus greater than substantia nigra greater than nucleus accumbens septi (accumbens) greater than piriform cortex greater than nucleus caudatus (caudate) greater than mesoprefrontal cortex greater than cingulate cortex, similar to previous observations. Since untreated and placebo-injected control results were indistinguishable, a stress artifact is unlikely to account for the findings. Haloperidol at the lower dose produced no significant changes but, at the higher dose, yielded relatively large (42%-143%) increases of NT concentrations in accumbens, caudate, and substantia nigra. S(+)NPA, which has some properties as a limbic-selective dopamine antagonist, yielded smaller (55%-66%) but significant average increases of NT in accumbens and piriform cortex, and lesser trends toward increases (40%-51%) in caudate, nigra, and mesoprefrontal cortex--all persisting for 5 days after treatment, whereas the R(-) enantiomer, a potent dopaminergic agonist, increased NT only in nigra (by 112%). These observations confirm previous results with haloperidol and add to the impression that S(+)-NPA shares some properties of atypical antipsychotic agents.