77 primary depressive in-patients aged 18-70 years (mean = 46/47 years) were assigned randomly to experimental and control groups and treated in double-blind conditions in ten centers. Butriptyline and amitriptyline were both administered with an identical increasing schedule, up to 150 mg daily in the first week and a flexible schedule for the last 3 weeks of trial. Mean daily doses were 145 mg butriptyline and 142 mg amitriptyline after 2 weeks, and 77.5 mg amitriptyline and butriptyline after 4 weeks. Nitrazepam (5-10 mg) and haldol (5 mg) were also allowed, only if necessary. Symptomatology and antidepressant efficacy were assessed using the rating scales of Hamilton, Overall, BPRS, CGI and a side effect checklist. After initial comparison of the two treatment groups, the results showed that the antidepressant effects were significantly better with butriptyline on the number of dropouts, on the total score and on the following factors of the Overall Depression Scale: depression, guilt, anxiety, somatization and somatic complaints. Frequency of haldol prescription was significantly lower with butriptyline than with amitriptyline. The overall frequency of side effects and of autonomic symptoms did not differ in the two groups. The effects on other parameters (hematological and biochemical variables, ECG and EEG) were similar for both drugs. In conclusion, butriptyline has the same indications as amitriptyline but shows a better antidepressant efficacy at the same dosage.