There has been a long-standing interest in plasma prolactin as a potential in vivo indicator of blockade of tuberoinfundibular D2 dopamine receptors. Potential relationships between prolactin response and neuroleptic treatment have been obscured by the use of high doses which have caused prolactin to plateau. With lower doses of neuroleptic now commonly in use, prolactin may be more valuable as a correlate of clinical response. In this study, 23 acutely exacerbated schizophrenic and schizoaffective patients were washed out for at least 6 days and were then treated with haloperidol to achieve fixed low to moderate plasma levels under double-blind conditions. Clinical response, plasma prolactin, and haloperidol plasma levels were measured weekly for 3 weeks. Clinical symptoms at endpoint were related to both prolactin change and final prolactin level during haloperidol treatment. Specifically, fewer symptoms at endpoint were associated with a greater increase in prolactin over time and a higher prolactin level at endpoint. Thus, prolactin increase caused by low to moderate doses of haloperidol may be a correlate of endpoint symptomatology. As lower doses of typical neuroleptics are now in use, prolactin response as a predictor of clinical response may have more clinical utility. Further study of prolactin and clinical response to typical neuroleptics should focus on low neuroleptic doses.