A method for mean flow velocity measurement using digital subtraction an giography (DSA) was evaluated, along with the results of phantom and animal experiments. The validity of its clinical use was studied by applying this tech nique in the arteries and bypass grafts in patients with arterial occlusive dis ease. Mean flow velocity was reliably determined by this method, but it could not be used as a determinant of arterial occlusive disease, because of a consider able overlapping between limbs with and without significant arterial lesions. A low value of mean flow velocity was not itself a reliable prognostic indicator of bypass failure. However, measurement of mean flow velocity was valuable in detecting functional abnormalities of the graft.