THE exchange of metal ions between an oxide mineral surface and water occurs in a wide range of processes, including corrosion(1), the breakdown, of inhaled dusts(2,3), soil formation(4) and the cycling of toxic substances in the environment(5). In studies of the mechanisms of dissolution, the measured rate-law order with respect to protons(6-15) cannot be reconciled with the number of protons needed to form any reasonable assumed activated complex. Here we suggest that this discrepancy can be avoided if one takes into account the number of protonation and deprotonation steps leading to detachment of the hydrated metal ion. We show that the experimental proton rate order reflects a net balance of protons removed and attached in these steps. Our mechanism explains why the rate order generally coincides with the metal valence(8,9,11,12,16-18), and why there is a similarity between rates of water ligand liability in dissolved complexes and rates of mineral dissolution(19-22) and metal desorption(23). It eliminates the need to invoke catalysis by protons, and establishes a close consistency between reactions at surfaces and (better understood) ligand-exchange reactions in solution.