Fifty-two adults referred for evaluation of aortic stenosis (AS) were studied using continuous-wave and pulsed Doppler echocardiography. Three windows were used to determine which approach (apical, right parasternal or suprasternal) yielded optimal results. Doppler-derived peak aortic valve gradients were compared with the peak gradients measured at cardiac catheterization in 23 patients. High-velocity jets were best recorded from the cardiac apex and less frequently from the right parasternal and suprasternal areas. However, gradients from the right parasternal area correlated best with cardiac catheterization findings, although recordings could be made from this window in only 49% of the patients. Velocities from the suprasternal window were significantly (p less than 0.01) lower than those from the apex or right parasternal areas. Gradient underestimation from the suprasternal window tended to increase with age of the patient (p less than 0.01). When the maximal Doppler derived gradient from any window was compared with catheterization measurements, the correlation coefficient was 0.86. Gradients derived from Doppler velocities accurately predicted severe (more than 50 mm Hg) gradients at catheterization. Thus, Doppler echocardiography is useful in evaluation of AS when several windows are used for optimal assessment of aortic valve gradient.