Purpose: The aim of this study was to determine the accuracy of Doppler waveform characteristics in grading femoropopliteal stenoses and to determine the interobserver and intraobserver reproducibility of measuring the same waveform characteristics. Methods: Thirty patients with isolated areas of stenosis found by arteriography were evaluated by color duplex sonography. Each patient underwent scanning by two observers on two separate occasions. Each observer was blind to the other's results. Doppler spectra were recorded in areas where color change suggested the highest velocity and also at the nearest normal proximal area. Peak systolic velocity, spectral broadening, and waveform configuration were measured at each site. Results: An increase in peak systolic velocity of more than 200% accurately predicted a 50% or greater reduction in luminal diameter on angiography (70% sensitivity, 96% specificity). The presence of spectral broadening and an abnormal waveform shape were found to correlate poorly with the degree of stenosis. Analysis of variance showed no significant difference between observers in velocity measurements (p = 0.78). Conclusions: We conclude that although stenoses of greater than 50% can be distinguished from minor stenoses, more precise definition of the degree of narrowing is unlikely. The good repeatability of the velocity ratio makes it an excellent tool for monitoring major changes in the progression of disease.