Twenty-two patients with mild or moderate mitral or aortic incompetence were randomly assigned to treatment with either hydralazine (mean 127 mg/day, range 37-225) or placebo. Eight patients in the hydralazine group and ten patients in the placebo group completed the study. Two of the patients in the hydralazine group and one patient in the placebo group were withdrawn because of suspected side effects. One patient dropped out because of influenza. Over a period of seven weeks the patients were monitored clinically as well as non-invasively with echocardiography and exercise testing. The systolic blood pressure fell from 152 +/- 10 to 135 +/- 9 mm Hg (mean +/- s.e.m.) (17%, P less than 0.01). The diastolic blood pressure fell from 63 +/- 8 to 58 +/- 8 mm Hg (5%, P = 0.09). The heart rate was unchanged. Left ventricular internal diameter in systole decreased from 49 +/- 3 to 45 +/- 3 mm (9%, P = 0.05) and in diastole from 73 +/- 4 to 70 +/- 3 mm (4%, P = 0.03). Left ventricular systolic wall tension fell from 200 +/- 16 to 152 +/- 18 mm Hg (24%, P less than 0.01). Left ventricular shortening fraction increased from 32 +/- 3 to 36 +/- 3% (12%, P less than 0.01). Maximal exercise capacity improved from 3200 +/- 800 to 3800 +/- 700 kpm (19%, P = 0.02). No significant responses were observed in the placebo group. Oral hydralazine reduces left ventricular internal diameters, improves left ventricular performance, presumably at a lower level of oxygen consumption, and improves exercise capacity in patients with modest mitral or aortic incompetence.