Abstract Events which initiate postoperative deep venous thrombosis occur in the intraoperative or early postoperative period in patients undergoing major surgery. Also endothelial alterations and accumulation of blood cells and noncellular material were observed in canine veins 4 hr. after splenectomy, hysterectomy and intestinal anastomosis. In the present investigation we examined canine jugular veins and carotid arteries 1 hr. after hysterectomy. Hysterectomy increased the amount of plasma (as measured by 125I-labeled protein) in interstitial space of veins and arteries 3.6-fold over control values. In control vessels mg plasma/gm vessel were: jugulars, 23 and carotides 14, and 1 hr. after hysterectomy: jugulars 78 and carotids 52. Differences were significant at the 0.02 level. Scanning electron microscopy showed that endothelium of control veins was smooth and free of adhering blood elements while endothelium of posthysterectomy dogs had extensive pseudopods, blebs and pits. Adhering and invading leukoctyes were found on all postsurgical veins singly or in small patches. Despite increased permeability postsurgical arteries were free of blood cells and little affected at the electron microscopic level. This model resembles clinical postoperative deep vein thrombosis in that vessels to be studied were distant from the surgical site and were not damaged directly. We suggest that the changes caused by blood born products of tissue injury may represent the venous lesion that contributes to the initiation and propagation of deep venous thrombosis, other conditions being favorable. Failure of blood elements to adhere to arteries may help explain their resistance to postoperative thrombosis.