The general medical community in the United States has been rather slow in adopting short-course bactericidal chemotherapy for tuberculosis despite the clear demonstration of the advantage by several carefully controlled clinical trials. Reported herein is experience between January 1976 and December 1982 in 1,028 patients with bacteriologically proved pulmonary tuberculosis treated for nine months with isoniazid (300 mg) and rifampin (600 mg) daily for one month followed by twice-weekly isoniazid (900 mg) and rifampin (600 mg) for the other eight months. They were treated by 45 local practitioners and supervised by public health nurses through 60 Arkansas Department of Health chest clinics in the state. Outpatient therapy was mostly self-administered in the routine treatment program. Overall success was achieved in 95 percent of the 751 patients who completed therapy; in 21 (2.8 percent), sputum cultures failed to convert to negative, and 15 (2.1 percent) have had relapse since therapy was stopped. Therapy could not be completed in 26.9 percent due to deaths, drug toxicities, relocation, refusal, etc. Of 21 bacteriologic failures, 18 patients developed isoniazid resistance and were treated with additional two bactericidal drugs. Most of the relapses (nine of 15) occurred within 12 months after chemotherapy was stopped. However, four relapses occurred quite late during follow-up. Only three of 15 patients with relapse showed isoniazid resistance. Side effects of the drugs were encountered in 10.3 percent, but major toxicities occurred in 3.2 percent (hepatitis in 2.6 percent, hematologic effects in 0.6 percent). Clinical surveillance for toxicity is preferred over routine and regular biochemical monitoring. Patient acceptance of the regimen was excellent, and compliance was good. Short-course chemotherapy is effective, with low drug toxicity, reduced cost of drugs, and ease of direct supervision when needed, and is acceptable to patients in routine treatment.