The fate of arsenic in groundwater depends largely on its interaction with mineral surfaces. We investigated the kinetics of As(III) oxidation by aquifer materials collected from the USGS research site at Cape Cod, MA, USA, by conducting laboratory experiments. Five different solid samples with similar specific surface areas (0.6–0.9 m² g⁻¹) and reductively extractable iron contents (18–26 μmol m⁻²), but with varying total manganese contents (0.5–3.5 μmol m⁻²) were used. Both dissolved and adsorbed As(III) and As(V) concentrations were measured with time up to 250 h. The As(III) removal rate from solution increased with increasing solid manganese content, suggesting that manganese oxide is responsible for the oxidation of As(III). Under all conditions, dissolved As(V) concentrations were very low. A quantitative model was developed to simulate the extent and kinetics of arsenic transformation by aquifer materials. The model included: (1) reversible rate-limited adsorption of As(III) onto both oxidative and non-oxidative (adsorptive) sites, (2) irreversible rate-limited oxidation of As(III), and (3) equilibrium adsorption of As(V) onto adsorptive sites. Rate constants for these processes, as well as the total oxidative site densities were used as the fitting parameters. The total adsorptive site densities were estimated based on the measured specific surface area of each material. The best fit was provided by considering one fast and one slow site for each adsorptive and oxidative site. The fitting parameters were obtained using the kinetic data for the most reactive aquifer material at different initial As(III) concentrations. Using the same parameters to simulate As(III) and As(V) surface reactions, the model predictions were compared to observations for aquifer materials with different manganese contents. The model simulated the experimental data very well for all materials at all initial As(III) concentrations. The As(V) production rate was related to the concentrations of the free oxidative surface sites and dissolved As(III), as rAs(V)=kox′[Mn(IV)OH][H3AsO3] with apparent second-order rate constants of koxf′=6.28×10-1 and koxs′=1.25×10-2M-1s-1 for the fast and the slow oxidative sites, respectively. The As(III) removal rate decreased approximately by half for a pH increase from 4 to 7. The pH dependence was explained using the acid–base behavior of the surface oxidative sites by considering a surface pK_a = 6.2 (I = 0). In the presence of excess surface adsorptive and oxidative sites, phosphate diminished the rate of As(III) removal and As(V) production only slightly due to its interaction with the oxidative sites. The observed As(III) oxidation rate here is consistent with previous observations of As(III) oxidation over short transport distances during field-scale transport experiments. The model developed here may be incorporated into groundwater transport models to predict arsenic speciation and transport in chemically heterogeneous systems.