This was a double-blind study designed to compare 'Compazine', Librium and placebo in a psychoneurotic outpatient population with dominant symptoms of anxiety. The patient participants were accepted for the study when their pre-treatment symptoms of anxiety met a minimum or greater score on the Hamilton Anxiety Rating Scale. The study interval was 4 weeks. Patient psychopathology was rated by the treating physician on the Hamilton Anxiety Rating Scale, the New Physician Rating List and Global Evaluation. Patients completed the Self-Rating Symptom Scale and the Profile of Mood States. All efficacy criteria ranked the study medications in the same order. Prochloroperazine was most effective, chlordiazepoxide was also active and placebo was least effective. There are statistically significant active medication versus placebo comparisons supporting desirable drug action. Chlordiazepoxide versus placebo comparisons approached statistical significance for all efficacy criteria.