Previously, we reported that the rate (R) of hydroxyapatite dissolution in acetic, lactic, and phosphoric acid solutions is a function of the degree of saturation with respect to the dissolving mineral, DS (defined as the ratio of the mean ionic activity product for hydroxyapatite [Ca5OH(PO4)3] in solution to its solubility product constant), and the sum of the acid activities (sumBiH) in solution: R = K(1-DS)m(sumBiH)n. The present study was undertaken to explore the general validity of this model in describing the kinetics of enamel demineralization. Thin sections of human enamel were exposed to partially saturated 0.1 mol/L lactic acid solutions, at two different DS levels, and at pH values of 4.3 to 6.0. Thin sections of human enamel were also exposed to solutions with four different concentrations of acetic and lactic acids (pH 4.3) with three different DS values and, at one DS value, to solutions of propionic acid. Mineral loss was monitored by quantitative microradiography. In solutions with pH values of 4.3 and 5.0, "lesions" were formed with well-defined surface layers, whereas, in solutions with pH 6.0, "lesions" were produced with no apparent surface layers. The formation of relatively intact surface layers was consistent with predicted phase transformations. Rates of mineral loss were found to be inversely proportional to both the degree of saturation with respect to enamel mineral, DS(En), and the pH of the solution and increased with increased activities of each organic acid, consistent with the proposed model. However, at the same DS(En) and acid activity, rates of demineralization were the same in the acetic and propionic acid solutions, whereas rates of demineralization in lactic acid were greater. It is suggested that specific interactions of acid species with enamel mineral may modify the rate of enamel demineralization. These in vitro findings suggest that relatively small differences in DS(En) values found in plaque fluid may result in very significant differences in the rate of enamel demineralization in vivo.