The value of reactive hyperemia in improving visualization in aortofemoral arteriography has been previously stressed (1, 2, 3) but the method has not yet found general acceptance. We have performed angiography after tourniquet occlusion almost routinely for the last three years, and have been impressed by the simplicity and safety of the method and the improvement in quality of the arteriograms. From a group of 200 patients with arterial disease studied in this manner, we have selected 55 in whom the timing of arterial blood flow with and without reactive hyperemia could be evaluated in serial films. We have analyzed the effect of reactive hyperemia and attempted to determine the influence of various clinical factors upon the response to tourniquet occlusion. Material and Methods Arteriography was performed either by percutaneous femoral catheterization or by translumbar injection. For this study, however, only retrograde femoral arteriograms were selected. A PE 240 J-tipped catheter was inserted via a common femoral artery. With the catheter positioned at the T-12-L-1 level, a serial abdominal aortogram was obtained. The catheter was then withdrawn to above the aortic bifurcation, and additional serial runs were made, in order, of the pelvis, thighs, and legs, each time following an injection of 25–30 ml of 60 per cent methylglucamine diatrizoate (Renografin 60) or iothalamate (Conray). A Cordis injector and Schonander AOT film changer were used. Accurate timing of the films was obtained either with the help of an automatic film-counting device or by direct recording on a Sanborn multichannel recorder. Following the control study, one or more injections were made utilizing reactive hyperemia. Blood pressure cuffs were first applied above the knees. The thighs were then flexed at the hips to 60° for one minute to elevate the legs, after which the blood pressure cuffs were rapidly inflated to 100 mm Hg above systolic pressure. The legs were then lowered and the patient positioned for the injection. After seven minutes of occlusion, both tourniquets were released and the injection started as quickly as possible thereafter, usually within thirty seconds. results Results Application of tourniquets above the knees was well tolerated by nearly all patients. Only a few individuals noted significant pain, and on some occasions it was necessary to shorten the period of occlusion. When pain occurred it appeared most often to be due to venous congestion. Patients who complained of pain were often those with a history of phlebitis or venous insufficiency. At other times, pain resulted from too brief an elevation of the legs or slow inflation of the tourniquets. No patient had clinical exacerbation of his peripheral vascular disease as a result of this procedure. The effect of reactive hyperemia was almost immediately apparent.