Substitution of peripheral arterial pressure for ascending aortic pressure is a common but poorly validated practice in the assessment of aortic valve gradients by catheterization. The accuracy of this practice was assessed by comparing the left ventricular-ascending aortic mean gradient in 26 cases of aortic stenosis with the left ventricular-femoral artery gradient, both with and without compensation for temporal delay in femoral artery pressure. Aligned left ventricular-femoral artery gradients (matching upstrokes to compensate for peripheral time delay) underestimated the left ventricular-ascending aortic gradient by 10 mm Hg (range 0 to -17). Unaltered simultaneous left ventricular-femoral artery gradients overestimated the left ventricular-ascending aortic gradient by an average of 9 mm Hg (range +1 to +18). For both peripheral techniques, the error was relatively constant throughout the range of aortic valve gradients. The most accurate estimate of both aortic valve gradient and area was obtained by averaging the gradients and areas derived from aligned and unaltered left ventricular-peripheral arterial simultaneous tracings. Although only occasionally critical for clinical decision-making, these errors may be overwhelming in certain types of research applications, such as comparisons of valve prosthesis gradients and serial evaluations of aortic stenosis. An additional source of error is a coexistent peripheral arterial gradient that was present in 21% of otherwise technically suitable patients in the screened study group.