The reaction between S(IV) and O-2 in aqueous solution is of importance because of its involvement in flue gas desulfurization processes and in acid rain formation. Despite considerable research, a complete mechanism for the reaction has not been established. In this work a detailed study of the uncatalyzed reaction in the region of pH 4.5 has been carried out, The overall rate law was found to be -d[O-2]/dt = k(obs) [HSO3-](2)[H+](-2)[O-2](0) with k(obs) of 3.6 x 10(6) M s(-1) at 25 degrees C and ionic strength 0.05 M. The yield of the intermediate S2O72- formed during the reaction has been determined over the pH range 4.1-6.3. By comparison with the yield of S2O72- produced in the reaction of HSO3- with HSO5-, it is concluded that S2O72- formed in the HSO3--O-2 reaction comes from the reaction between HSO5- and HSO3-, and therefore HSO5- must be an intermediate in the HSO3--O-2 reaction in this pH region. From the above yields it is concluded that 60% of the reaction proceeds through the intermediate HSO5-. A detailed mechanism consistent with the overall rate law of the reaction has been proposed. A key feature is the proposal that the chain initiation occurs through the reaction of HSO5- With HSO3- to form SO3.- SO4.-. In order to study the individual rate laws for the initiation, propagation and termination reactions, a relaxation technique was applied. In these experiments, the reaction at a steady state rate was perturbed by suddenly introducing a change in the concentration of HSO3- or H+, and the relaxation of the reaction to the new steady state was recorded. Treatment of these experimental data not only confirmed the HSO3- dependence of the rates of initiation, propagation, and termination in the proposed mechanism but also gave the pH dependence of these reactions. A study of the remarkable inhibitory effect of methanol produced strong support for the proposed initiation reaction, The effect of the introduction of chain initiator S2O82- was also found to be consistent with the mechanism. Relative values of rate constants of reactions in the proposed mechanism have been evaluated from the present experimental data. Evaluation of absolute values of these rate constants requires one additional piece of experimental information, which, depending on the choice, yields a range of values, An arbitrary choice is used to give;rough values of the rate constants for initiation, propagation and termination and typical values for the steady state concentrations of chain carriers. The chain length appears to be about one thousand at pH 4.5, independent of the bisulfite concentration.