Six normal men ingested thioridazine, high and low doses, and placebo on three occasions. Their plasma and urinary thioridazine and mesoridazine plus sulforidazine were measured over 5 hr, together with their plasma prolactin, and a battery of psychophysiologic variables. Drowsiness and EEG changes correlated highly with rise in prolactin, but not with drug plasma concentrations. Finger tremor increased, and some psychologic tests were impaired by thioridazine; other psychologic tests, the auditory-evoked response, and palmar skin conductance were unaffected and showed no relationship to drug or prolactin levels. This suggests that plasma prolactin may be a useful indicator of some aspects of the individual's response to a psychotropic drug, and possibly a better guide to the selection of a suitable drug and its appropriate dose in clinical practice than the measurement of plasma concentrations of the drug itself.