Transmitral flow velocities were measured by Doppler echocardiography in nine healthy men who ingested 1 g/kg of ethanol within one hour. The measurements were made before the first drink and every hour thereafter for three hours. The peak mean (SE) blood ethanol concentration was 21.4 (1.0) mmol/l. Each man was also studied after drinking fruit juice. Ethanol increased the heart rate but did not change the peak transmitral velocities, the normalised peak filling rate, the deceleration of early flow, or the duration of relaxation as measured from the second heart sound to the peak early diastolic velocity. The ratio of the peak atrial to the peak early diastolic velocity rose from 0.41 (0.03) to 0.44 (0.03) after ethanol but remained unchanged after juice. The difference between juice and ethanol was independent of changes in heart rate. The fluid balance was more negative in the ethanol experiment (-727 (114) ml v -107 (70) ml), suggesting a reduction in preload, and the ethanol-induced net loss of fluid correlated with the concomitant change in the velocity ratio. A moderate dose of ethanol causes a small acute increase of the ratio of the peak atrial to the peak early diastolic velocity of mitral flow in healthy subjects. Although this change indicates altered diastolic function of the left ventricle, most of it may result from the diuretic effect of ethanol. Any major impairment of ventricular relaxation seems unlikely.