The detailed arterial anatomy of the foot in severe limb ischemia is not well known. This study was undertaken to define foot arterial anatomy and correlate these findings with the early results (6 months) of femoral-distal bypass. After completion of the bypass, operative arteriography was performed by direct injection of contrast media into the graft. A lateral view of the distal limb and foot was obtained. Foot vessel anatomy was classified into primary and secondary pedal arches, analogous to the superficial and deep volar arches of the hand. For peroneal bypass, special attention was paid to perforating branches and their communications with these two pedal arches. A total of 56 distal bypass operations was analyzed. Femoral--anterior tibial bypass was performed in 26 cases. When either a primary or a secondary pedal arch was intact, early graft patency (6 months) was achieved in 20 of 21 patients. When neither pathway was present, graft failure occurred in four of five cases. Similarly, of 10 femoral--posterior tibial grafts, seven remained patent with at least one pedal arch intact, whereas six of seven grafts failed when neither arch was patent. Bypass to the peroneal artery was successful in eight of nine limbs when a patent primary or secondary arch was reconstituted via either the anterior or posterior perforating branches. None of four peroneal grafts remained patent when both arches were occluded. As a whole, in 40 limbs with either a patent primary or secondary pedal arch, early graft success was achieved in 35 limbs (87.5%). In contrast, in 16 limbs with no patent arch, only two (12.5%) were successfully reconstructed (P less than 0.001). Analysis of the results of femoral-distal bypass based on a single plantar arch as the sole determining anatomic factor in graft patency is not adequate. The secondary pedal arch and communicating branches of the peroneal artery are also of surgical significance. Operative arteriography can define runoff in the foot, and this information has prognostic significance. It may allow rational judgment regarding reintervention in patients with failed grafts.