Contamination of ground water in domestic water wells has been documented in the vicinity of a uranium mill near Canon City, south-central Colorado, U.S.A. Acidic tailings fluid (raffinate) was passed through a core collected from the subjacent calcite-bearing sandstone to evaluate the effect of interactions between the raffinate and bedrock on the fluid pH, and on the mobility of Al, Ca, Cl, Fe, Mn, SO 4, and Zn. In the experiment, the pH initially increased from 2.3 to 8.0 as calcite in the core dissolved and neutralized the raffinate. Concurrently, amorphous ferric hydroxide precipitated in the micro-environment surrounding the reacting carbonate grains. This led to a gradual decrease in pH to 3.4 due to the armoring of the remnant calcite cement by amorphous ferric hydroxide. The results were modeled using the mass transfer computer program, MINTEQ. The pH was modeled by simulating the dissolution of calcite in the raffinate, while the Eh was set at the values measured in the experiment. The behavior of Mn was described by the dissolution of manganocalcite, but an adequate model for dissolved Ca required both calcite dissolution and ion exchange of Ca for Na. Aluminum behavior was simulated by assuming a hydroxide solubility constraint above pH 5.7 and by AlOHSO 4 in more acidic regimes. Iron was modeled by the precipitation of an amorphous ferric hydroxide. From chemical analyses, measurements of Eh and pH, and MINTEQ calculations, the log K sp of the amorphous ferric hydroxide in the experiment was determined to be from −33.5 to −37.6. Zinc was modeled by means of the triple-layer sorption algorithm in MINTEQ, assuming the amorphous ferric hydroxide phase to be the sorbent. Comparison between the experimental effluent and ground-water metal concentrations downgradient from the site are in general agreement. Specifically, SO 4 appears to be the best indicator of the encroaching front of acidic contaminants in the subsurface. The general concurrence between the observed experimental results, the computed model predictions, and the downgradient ground-water metal concentrations indicate that mass transfer models such as MINTEQ are useful in predicting the interaction between bedrock and acidic tailings fluid.