In a group of 70 patients with endogenous depression entering a controlled trial of real versus sham ECT, urinary 3-methoxy-4-hydroxyphenylglycol (MHPG) excretion was significantly reduced by comparison with previously studied groups of control subjects, of acute and chronic schizophrenic patients and of anxious patients. However, urinary MHPG was unrelated to severity of depression, or to the presence of delusions, retardation or agitation. MHPG excretion did not predict clinical outcome, or the response to ECT. Urinary MHPG content at trial entry was unrelated to past tricyclic antidepressant or benzodiazepine medication, although an influence of the latter on the findings cannot be excluded, since all patients received benzodiazepine (nitrazepam) night sedation during the trial. During the 4-week trial MHPG excretion remained low and did not increase in relation to change in clinical state, although there was a small but significant increase in patients who received real ECT. The findings confirm that urinary MHPG excretion is reduced in depression, but establish that such reductions are not state dependent. Since the increase in MHPG excretion with ECT is not related to changes in clinical state, the former presumably does not reflect the mechanism of action of ECT.