Aqueous N2O/O2-saturated solutions of poly(U) were irradiated at 0 degrees C and the release of unaltered uracil determined. Immediately after irradiation G(uracil release) was 1.5 which increased to a value of 5.3 +/- 0.3 upon heating to 95 degrees C. Thereby all of the organic hydroperoxides (G = 6.8 +/- 0.7) and some of the hydrogen peroxide (G = 1.7 +/- 0.2) was destroyed leaving G(peroxidic material; mainly hydrogen peroxide) = 1.0 +/- 0.7. G(chromophore loss) = 8-11 was measured immediately after irradiation, but no increase was observed upon heating. Addition of iodide destroyed the hydroperoxides and caused immediate base release to rise to G = 4 and further heating brought the value to that observed in the absence of iodide. In contrast, on reducing the hydroperoxides with NaBH4, immediate uracil release rose to only G = 2.8 and no further increase was observed on heating. A major product (G = 2.7) is carbon dioxide. There are also osazone-forming compounds produced (G = 2.7), all of which are originally bound to poly(U). Heating in acid solutions, as is required for this test, releases glycoladehyde-derived osazone (G = 0.8) and further unidentified low molecular weight material (G = 0.9). It is concluded that the primary radicals which cause these lesions are the base OH adduct radicals. In the presence of oxygen these are converted into the corresponding peroxyl radicals which abstract an H atom from the sugar moiety. In the course of this reaction base-hydroperoxides are formed. However, such base hydroperoxides cannot be the only organic hydroperoxides, but some (G congruent to 2.5) sugar-hydroperoxides must be formed as indicated by the increase in base release by the addition of iodide. It is speculated that a sugar-hydroperoxide located at C(3') is reduced by iodide to a carbonyl function at C(3'), a lesion that releases the base, while reduction with NaBH4 reduces it to an alcohol function at C(3') thus preventing base release.