Ninety one patients with new episodes of minor affective disorder were selected by their general practitioners as suitable for anxiolytic medication. Half the patients were allocated randomly to a drug-group (anxiolytic medication), and half to a non-drug group (brief counselling without anxiolytics). Psychiatric and social assessments were made (i) at initial consultation when treatment was started; (ii) one month later; (iii) seven months later. Before treatment the two groups were similar on all main variables. On the General Health Questionnaire, 85 per cent of patients were psychiatric cases before treatment, 40 per cent at one month and 30 per cent at seven months. Similar improvements were found with other measures of psychiatric state (Profile of Mood States; Present State Examination) and social functioning (SAS-M). Improvements were similar and parallel in the two groups. Neither group of patients increased their consumption of alcohol, tobacco or non-prescribed drugs. The non-drug group did not make increased demands on the doctors' time.