Feldspar powders, An 0An 76, were dissolved in flow-through reactors at 25°C, pH 3, to investigate the effect of feldspar composition, electrolyte concentration, and cation identity upon dissolution rates. BET surface area increased 1.5–7 times over the approximately 2000 hour reaction times; we, therefore, calculated dissolution rates with the final, rather than the initial surface area. This correction resulted in calculated rates which were, correspondingly, 1.5–7 times lower than several previously published rate estimates. Dissolution rates increase linearly with increasing anorthite content over the composition range studied. Rates decreased with increasing NaCl, and to a lesser extent, increasing (CH 3) 4NCl concentrations. We interpret our rate data with a surface-controlled rate model: rate = k · [SOH ex] n , where [z.tbnd;SOH ex] is the concentration of H + which reacts with the feldspar surface through proton-cation exchange reactions. Previous workers have used [z.tbnd;SOH ex] to represent protons adsorbed to surface hydroxyl sites. We express [z.tbnd;SOH ex] with a Langmuir competitive adsorption isotherm, and fit our rates to the model: rate = kN s K H{H +} 1 + K H{H +} + K Na{Na +} 0.5 , where k = the rate constan, N s = the surface site density, K H = the H + constant for adsorption at the exchange site, K Na = the Na + constant for adsorption at the exchange site, and { i} denotes the activity of species i. Aluminum and the network-modifiers, Na, K, and Ca, were preferentially released compared to Si during the initial phase of dissolution. After 500–1000 hours in H 2O-HCl, dissolution became stoichiometric for the microcline, albite, and bytownite compositions. Oligoclase and labradorite continued to exhibit preferential Ca and Al release even after 3000 hours of dissolution. Exsolution texture, observed in labradorite, may provide a structural control for preferential Ca and Al release. Apparent nonstoichiometric dissolution in oligoclase is due to the presence of Ca- and Al-rich accessory phases, present in the original feldspar samples. This work suggests that in the absence of accessory phases and mineral defects, steady-state feldspar dissolution is stoichiometric for all compositions.