An experimental model of surgically-induced arterial thrombosis was devised using the femoral arteries of dogs. Within 7 days, 67% of the arteries became completely thrombosed and only 12% remained compeletly patent. In the group of dogs that received low-dose heparin, 69% of the vessels were completely thrombosed and 6% remained completely patent. In the group of dogs treated with low-dose Ancrod to induce partial defibrination, 75% remained completely patent while only 19% of their femoral arteries were completely thrombosed. Although the ancrod was effective in preventing arterial thrombosis, 88% of the wounds showed moderate to severe separations. Most likely the absence of a fibrin lattice, necessary for the securement and growth of fibroblasts as the wound heals, explains this latter effect. Thus while Ancrod may become useful as an anticoagulant in certain clinical situations, it should not be used in proximity to surgery. Finally, in these studies of acute arterial thromboses, low-dose heparin therapy offered no protective effect.