A simple progressive exercise test was performed before and after operation on five subjects undergoing mitral valve replacement and on five subjects undergoing aortic valve replacement. The responses of heart rate and ventilation were related to work rate )kilopond metres/min). The patients were also assessed clinically by the New York Heart Association grading and radiologically before each exercise test. The clinical grading was shown to be a poor guide to observed exercise tolerance, as the improvement noted in symptoms was not matched by the objective measurement of working capacity. Only two patients had normal exercise tolerance after surgery, although six of the ten patients claimed that they had no exertional dyspnoea after operation. The changes in simple ventilatory function tests before and after operation were generally small. We suggest that measurements of exercise tolerance before and after operation should be an essential part of heart valve replacement surgery.