One thesis concerning the pathogenesis of "stress ulcer" states that the combination of (1) bile acid-induced H+ "back diffusion" and (2) gastric mucosal ischemia is acutely Hg) to induce ischemia, this thesis was further tested by selectively mitigating either the ischemic or the back diffusion components of the model. Vascularized wedges of proximal canine gastric wall mounted on Lucite chambers were studied. With the mucosa directly visualized, control group A (6 dogs) was subjected sequentially to (1) topical acid test solution alone (ATS), (2) ATS + S, and (3) ATS + S + topical 5 mM Na taurocholate (TC). Study group B (6 dogs): (1) ATS, (2) ATS + S + the beta adrenergic agonist, isoproterenol, 1.5 microng per kg-min infused into the splenic artery, and (3) ATS + S + TC + isoproterenol. Study group C (6 dogs): (1) ATS, (2) ATS + S + the bile acid binding resin, cholestyramine (C), 4 g per liter, and (3) ATS + S + TC + C. During each period the net flux of H+, the electrical potential difference, and the aminopyrine clearance (AC) were determined. Mucosal damage (intramucosal hemorrhage, erosions, and ulcers, graded 0 to 5) was assessed blindly by an independent observer using photographs. The results indicate (1) that, despite H+ back diffusion comparable to ATS + S + TC, intraarterial isoproterenol significantly protects against lesion formation by increasing AC, and (2) that, despite a reduction in AC comparable to ATS + S + TC, topical C significantly protects against lesion formation by preventing excessive back diffusion of H+.