Superoxide (O2−) is a key intermediate in the cycling of organic matter and trace metals in natural waters but production rates are difficult to determine due to low steady-state concentrations, rapid decay rates, and unstable standards. On the other hand, superoxide’s dismutation product, hydrogen peroxide (H2O2), is relatively stable in filtered water. Thus, if the stoichiometry between O2− and H2O2 is known, one can derive superoxide data from H2O2 measurements. The relationship between O2− and H2O2 remains uncertain in seawater but work by Petasne and Zika (1987) presented a method for examining the relationship between O2− and H2O2 during irradiations of coastal seawater using superoxide dismutase (SOD), which forces a 2:1 stoichiometry between O2− and H2O2. Here we report the first O2− apparent quantum yield (AQY) spectra following their approach; performing irradiations of various fresh and seawater samples and measuring H2O2 accumulation with and without added SOD. For all but a single riverine sample, H2O2 AQY spectra fell in a narrow range, but O2− AQY spectra varied such that O2−:H2O2 ratios were always greater than 2 and were highest for the clear waters of the Gulf Stream (~3.4 O2− per H2O2 generated). Because this approach eliminates the need to measure O2− production rates directly, it represents a simple way to refine the stoichiometric relationships that would potentially allow global estimates of O2− photoproduction rates, O2− steady-state concentrations (O2−ss), and related surface ocean redox reactions based on more manageable H2O2 photochemical studies.