During free tissue transfer, an artery and one or two veins must be sutured. Either the artery or the vein can be repaired first, but the optimal order of vascular repair has not been established. Maintaining a clamp on a vessel is associated with vessel injury, and releasing the clamp may cause bleeding and vascular stasis. The purpose of this study was to determine if the order of vessel repair and the length of vascular clamping affects surgical outcome. Free flaps were created on Sprague-Dawley rats (400-450 g) as follows: Bilateral abdominal skin flaps (3 x 3 cm) based on the superficial inferior epigastric vessels were elevated. The femoral artery and vein were tied and divided distal to the branching of the inferior epigastric vessels. Proximal to the branching, the artery and vein were divided. The animals were then randomized into four groups as follows: In group I (N=16), the artery was repaired and then the clamps were released to revascularize the flap. Venous stasis occurred as the vein was being repaired because of the venous clamp. In group II (N=15), the artery was repaired, but the clamp was maintained to prevent blood from coming in contact with the fresh arterial anastomosis and to prevent venous stasis. The vein was then repaired. In group III (N=15), the vein was repaired first but the venous clamp was not released until the artery was repaired. In group IV (N=15), the vein was repaired first and the clamps were released, allowing venous blood to contact the fresh anastomosis while the artery was being repaired. After final clamp removal, all anastomoses were assessed immediately for evidence of thrombosis. Five days postoperatively the skin flap was evaluated for evidence of necrosis and the anastomosis was inspected for evidence of late thrombosis. The anastomoses were resected for histopathological evaluation. Flap success was compared between groups using chi-squared analysis. Eleven of 16 flaps failed (69%) in group I, and 3 of 15 flaps failed (20%) in each of groups II, III, and IV. Statistical significance was reached in comparing group I with the other three groups (p < 0.01). All failures in group I were caused by immediate venous thrombosis. The other failures were secondary to arterial thrombosis. Histopathological analysis failed to demonstrate any differences between the groups. According to these results, arterial repair followed by clamp release prior to the completion of the venous repair results in a low success rate, probably secondary to venous stasis within the draining vein while the venous anastomosis is being completed. If the arterial repair is performed first, then it is recommended that the arterial clamp be maintained until the venous repair is completed. If the vein is repaired first, then it can be clamped or unclamped with similar results.