To determine the correlation between electrocardiographic (ECG) findings and anatomy utilizing echocardiography in children with pressure- or volume-loaded left ventricles, we analyzed the preoperative ECG tracings of 19 patients who underwent surgery for significant aortic stenosis and 12 patients who underwent cardiac catheterization or surgery for clinically significant ventricular septal defects. We then compared them with a group of 21 normal controls. The left ventricular muscle mass in these patients was calculated from echocardiograms using the simplified cubed formula. Posterior and septal wall thickness and cavity size were significantly greater in the aortic stenosis group than in the normal group. Only cavity size was significantly greater in the ventricular septal defect group than in the normal group. Eighteen aortic stenosis patients (95%) and ten ventricular septal defect patients (83%) had a left ventricular muscle mass greater than 2 standard deviations above the mean for the normal group. Significant differences were found in the voltages of SV1 + RV6 and in the voltage of RV6 alone between normals, aortic stenosis patients, and ventricular septal defect patients regardless of age. Using conventional ECG criteria for left ventricular hypertrophy, the highest sensitivity in aortic stenosis patients (67%) and ventricular septal defect patients (60%) was modest. The likelihood ratio for a positive test in either group was the best for SV1 + RV6 > 98th centile for age; RV6 > 98th centile for age was the best single measurement. No correlation was found between voltage and any measurable hemodynamic or anatomic data. Conventional pediatric ECG criteria for left ventricular hypertrophy have only modest sensitivity regardless of whether the heart is under pressure or volume load. Because left ventricular muscle mass can be precisely determined by echocardiography, these ECG criteria should be applied cautiously.