The present studies show that ischemia and reperfusion cause severe injury without evidence of generation of large amounts of reactive oxygen. Calculations show that it is necessary that amounts of reactive oxygen several orders of magnitude greater than that measured during ischemia-reflow injury be generated intracellularly by redox-cycling compounds such as diquat or by direct oxidants such as t-butyl hydroperoxide before endogenous cellular antioxidant systems in ischemia-reperfused rat liver or canine heart are overwhelmed. Thus, our data are incompatible with the hypothesis that oxygen radicals cause ischemia-reflow injury by lipid peroxidation or tissue thiol oxidation as postulated in Figure 1. How might one reconcile these results with the important experiments documenting protection against ischemia-reflow injury by treatment with antioxidants, antioxidant enzymes, allopurinol, cyclooxygenase/lipoxygenase inhibitors, cobra venom factor, and antisera against leukocytes (Table 1)? It seems most likely that the small amounts of reactive oxygen that can be generated in a reduced, acidic environment, such as occurs in tissue ischemia, could only be playing a catalytic function and not producing a massive oxidative attack on tissue membranes. Thus, future studies might best be directed at the catalytic function of superoxide/hydrogen peroxide in modulating the availability of substances such as lipoxygenase and cyclooxygenase products and of endothelium-derived relaxing factor during ischemia-reflow injury.