Three measurements of diastolic filling were compared in 29 patients with essential hypertension and 27 age matched normotensive controls. Systolic function was normal in all but one of the patients. The mean (1SD) first one third filling fraction (a measurement of early diastolic filling) was significantly lower in the hypertensive groups (0.27 (0.24] than in the control group (0.45 (0.16)). The hypertensive group was subdivided into those with electrocardiographic abnormalities and those without. In the subgroup with a normal electrocardiogram the mean (1SD) first one third filling fraction measurement (0.28 (0.16)) was significantly lower than in the control group. In the subgroup with an abnormal electrocardiogram, the first one third filling fraction was even lower (0.24 (0.9)). In addition, the time to peak filling rate (213 (56) ms) was significantly longer in the subgroup with the abnormal electrocardiogram than in the control group (164 (45) ms). However, the interobserver reproducibility of the time to peak filling measurement was poor. The peak filling rate was low in the subgroup with an abnormal electrocardiogram, but not significantly different from the normal controls. The discriminatory value of the three diastolic measurements did not improve with exercise. These results showed an early diastolic filling abnormality in essential hypertension that did not appear to be caused by disease of the large coronary vessels as it was present in patients with normal wall motion and a normal exercise electrocardiogram. The occurrence of diastolic abnormalities when systolic function is still normal may mark an early stage in the development of hypertensive heart failure, at a time when the process is still potentially reversible.