A free radical mechanism of cataractogenesis involving enzymatic and nonenzymatic reactions, is proposed. Supporting experimental evidence is briefly reviewed. H2O2, which is one of the toxic metabolites of oxygen, was significantly increased 2-3 fold in ocular humors in several experimental cataracts and in human senile cataract. Various cataractogenic agents were also found to increase H2O2 in ocular humors in vivo prior to cataract formation. Enzymatic defenses against O2-. and H2O2 provided by superoxide dismutase, catalase and glutathione peroxidase were impaired in cataracts. In some cataracts, catalase and superoxide dismutase were affected earlier. Malondialdehyde (MDA), a major breakdown product of lipid peroxides was significantly increased by 2-4-fold in human senile cataract, in cataracts induced in rabbit and rat, and in hereditary cataracts in mice. All the reactive species of O2 (O2-., H2O2, OH. and 1 delta gO2) may participate in initiating lipid peroxidation of lens in vitro. Various scavengers of these species were capable of preventing lenticular lipid peroxidation, amongst which OH. scavengers were found to be the most effective. Biological antioxidant, vitamin E afforded 44% prevention of lipid peroxidation in lens. The important observation was that vitamin E was therapeutically effective in about 50% of animals in arresting cataract induced in rabbit by 3-aminotriazole. In these rabbits, H2O2 and ascorbic acid of ocular humors and MDA of lens were close to normal. It is our working hypothesis that the carbonyl groups of MDA and amino groups of amino acids, proteins, nucleic acids and their bases, and phospholipids could interact in a cross-linking reaction producing high molecular weight aggregates by Schiff-base conjugate formation in addition to disulfide cross-linking of proteins, and finally resulting in cataract.