Left ventricular hypertrophy (LVH) detected by electrocardiography (ECG) and, more recently, by echocardiography has been shown to be an extremely strong predictor of morbidity and mortality in patients with essential hypertension and in members of the general population. Increased left ventricular mass (LVM) is strongly related to both increased blood pressure and overweight. Indexation of LVM by body surface or height has advantages for the detection of LVH related to hypertension or obesity. Indexation of LVM for height to the power 2.7 revealed by analysis of growth (allometric) relations may accomplish both these goals. In validation studies, the sensitivity of echocardiography to detect LVH has been reasonably high (85-100%), whereas that of ECG has ranged from as high as 50% in severely diseased necropsy populations to as low as 6-17% in recent studies in Cornell and Framingham. ECG sensitivity can be improved by using Cornell multivariate regression equations or by consideration of the Cornell voltage-QRS duration product. Obesity dramatically decreases the sensitivity of the ECG for detection of LVH, and recent research suggests a lower specificity and a higher rate of false-positive ECG diagnoses of LVH in black than in white subjects. Standard criteria for ECG LVH are less useful than echocardiographic findings for stratifying populations into high- and low-risk subgroups because of lower sensitivity, but improved ECG criteria need further evaluation in this respect.