In an open study, 18 patients suffering from an acute episode of schizophrenia and 18 patients with severe mania were given haloperidol at different dosage levels. Haloperidol plasma concentrations were measured and the status of the patients was evaluated at intervals using the Brief Psychiatric Rating Scale (BPRS). No correlation was found between the oral dose and the plasma concentrations in the schizophrenic group, but there was a low correlation for manic patients. In both groups, however, there was a correlation between these two parameters when the drug was given intramuscularly. There was no correlation between haloperidol plasma concentrations and the BPRS scores. Of the 36 patients 28 responded well to the treatment and were discharged from hospital. A good relationship between the clinical status as observed by the clinicians and the BPRS score was found. Plasma haloperidol concentration measurement was found to be a useful tool for the detection of non-compliance or excessive plasma levels. Accordingly, the present study indicates the value of drug level monitoring as a means to improve therapy.