To determine whether urinary sodium excretion (a rather rough estimate of sodium intake) can influence left ventricular mass independently of arterial pressure, 91 untreated subjects with essential hypertension and 50 normotensive subjects of similar age were studied. Left ventricular mass index (M-mode echocardiography) was positively correlated with urinary sodium excretion in hypertensive (r = 0.22, P less than 0.01) as well as normotensive subjects (r = 0.22, P less than 0.05), and systolic arterial pressure was correlated only in hypertensive subjects (r = 0.23, P less than 0.01). When hypertensive subjects were divided into groups with appropriate or inappropriate left ventricular mass by reference to a theoretical optimal left ventricular mass for each subject's level of systolic arterial pressure, left ventricular mass was appropriate in 68% and inappropriate in 32% of subjects. Urinary sodium excretion was higher in subjects with inappropriate left ventricular mass compared with those with appropriate left ventricular mass. In conclusion, sodium excretion may be an important modulator of the influence of arterial pressure on the left ventricle in normotensive subjects and subjects with essential hypertension.