Gas chromatography and radioreceptor assay technologies made possible the measurement of nanogram quantities of neuroleptic in the blood. They revealed large differences in blood concentrations among patients receiving the same oral dose. This led to the hope that measuring neuroleptic concentrations would allow more effective oral dosing for individual psychotic patients. Early attempts to study neuroleptic concentration/clinical response patterns, using chlorpromazine, produced confusing results. Many of the studies had serious design flaws, and were complicated by the many active metabolites of chlorpromazine. Adequate blood level/response studies should use fixed drug dosage schedules, drug-responsive patients, and neuroleptics without active metabolites. Brain concentration of drug is probably better reflected by erythrocyte than by plasma concentration. Therefore, recent investigators have usually used erythrocyte haloperidol or butaperazine concentrations to study level/response relationships. These recent studies strongly suggest an "optimal concentration" level/response pattern for haloperidol and butaperazine. Patients tend to respond best to moderate concentrations of neuroleptic, and nonresponse can be due to either too low or too high concentrations. In the latter case, decreasing the oral dose should be seriously considered. Blood level findings remain preliminary. The clinician should continue to fine-tune his clinical skills, using neuroleptic blood concentrations to complement his clinical judgment, in the attempt to determine the lowest effective dose for the individual patient.