Two groups of six depressive psychogeriatric patients (age 69-93) have been treated for three weeks with either 25 mg or 50 mg of amitriptyline (AT) per day. Before the beginning of the treatment and after 8, 15 and 22 days of medication, the patients (7 outpatients and 5 hospitalized patients) underwent various psychopathological tests (Hamilton, Crichton) and biochemical investigations, i.e. clinical chemistry and hematology, analyses of AT and nortriptyline (NT), in the plasma. In some subjects the hydroxylated metabolites and the binding of these substances to plasma proteins were also measured. Except one patient, all responded favourably to the treatment. No difference between the two doses has been observed as regards their clinical efficacy and incidence of side effects. The plasma levels of AT and NT were below those generally recommended in the literature. The monitoring served to identify some cases of noncompliance and, in one patient, a deficiency of hydroxylation of AT, confirmed by the debrisoquine test. These results suggest that the "therapeutic window" needs redefinition and that low-dose medication with antidepressants is indicated in a psychogeriatric population.