Patients suffering from mixed anxiety/depressive reactions, referred to the out-patient department of a large psychiatric hospital, were treated with either amitriptyline tablets or fluphenazine with nortriptyline (f/n) tablets for a period of four weeks. The study utilized a double-blind, completely randomised design, and patients' progress was assessed by means of the Wing Present State Examination, the Wakefield Self-Assessment Depression Inventory, and a side effects inventory. Both the symptom rating scales showed that purely depressive symptomatology improved significantly in each treatment group, but the patients' self-ratings showed that only f/n produced a significant alleviation of anxiety symptoms and panic attacks. The patients receiving f/n rated themselves as significantly (p less than 0.05) less irritable, as well as less anxious, after 4 weeks treatment, than those receiving amitriptyline. The PSE schedule did not differentiate between the 2 treatment groups, but self-rating, which is a more sensitive method of eliciting drug effectiveness in patients suffering from mild psychiatric disorders, did demonstrate patient preference for f/n. This should have important implications for compliance with treatment.