Oxidation of pyrite in aqueous solutions in contact with air (oxygen 20%) was studied at 25°C using short-term batch experiments. Fe2+ and SO42− were the only dissolved Fe and S species detected in these solutions. After a short period, R = [S]tot/[Fe]tot stabilized from 1.25 at pH = 1.5 to 1.6 at pH = 3. These R values were found to be consistent with previously published measurements (as calculated from the raw published data). This corresponds to a nonstoichiometric dissolution (R < 2) resulting from a deficit in aqueous sulfur. Thermodynamics indicate that S(−I) oxidation can only produce S(s)0 and SO42− under these equilibrium conditions. However, Pourbaix diagrams assuming the absence of SO42− indicate that S2O32− and S4O62− can appear in these conditions. Using these species the simplest expected oxidation mechanism is FeS2(s)+1.5 O2→Fe2++S2O32- followed by S2O32-+1.2 H+→0.4 S(s)0+0.4 S4O62-+0.6 H2O, and finally S4O62-+3.5 O2+3 H2O→4 SO42-+6 H+ possibly in several steps The overall reaction is FeS2+2.9 O2+0.6 H2O→Fe2++0.4 S(s)0+1.6 SO4 2−+1.2 H+, consistent with R = 1.6. In the most acidic (pH = 1.5) conditions, SO2 formation is expected as an intermediate step in the oxidation of S4O62− to SO42−. Degassing of SO2(g) would result in R < 1.6, again consistent with experimental observations. The above multistep mechanism, based on known aqueous redox chemistry of sulfur species, accounts for the deficit in aqueous sulfur noticed in all published experimental observations. The intermediate species cannot be detected, and it is consistent with calculated concentrations being below the detection limits. Under nonacidic conditions, S2O32− can be detected, but evaluation of the dissolution mechanism is hindered by precipitation of Fe(III) as iron oxyhydroxides.