A simple, inexpensive disk technique for testing bacterial sensitivity to antibiotics, was devised and evaluated to determine the value of such a test for an office practice. One hundred and nine patients were studied. Each patient was selected because of the clinical impression of an acute bacterial infection. Every patient had, singly or in combination, acute rhinitis, pharyngitis, tonsillitis, tracheitis, laryngitis, and bronchitis. Forty-two also had acute otitis media, and four had pneumonia. Twenty-seven patients were infected by bacteria resistant to the drug prescribed empirically but responded to a drug indicated as effective by the test. An error of 25% in empiric selection of an effective antibiotic was present. Application of this test will shorten the course of disease and reduce drug expenses for at least 25% of such patients.