The Earth’s past climate, sea level and ice volume have been reconstructed from the oxygen isotope ratio (18O/16O) of inorganic and biogenic carbonates. These reconstructions assume that temperature is the dominant controlling factor for the carbonate-water oxygen isotope fractionation (αc/w) but there is evidence for temperature-independent variations in αc/w. Uncertainties surrounding the reasons for αc/w variations complicate paleoenvironmental reconstructions based on the carbonate 18O/16O. A major advance in understanding oxygen isotopes in carbonates has been to view carbonate-water fractionation as the result of kinetic and/or equilibrium fractionations occurring between water and dissolved inorganic carbon (DIC) species and between the DIC species and carbonate. However, quantifying the intermediate fractionation steps is in its infancy. This thesis presents a new general model of oxygen isotope fractionations in the CaCO3-DIC-H2O system that quantifies DIC-H2O and CaCO3-DIC fractionation as a function of temperature, pH, salinity, calcite or aragonite saturation state (Ω), DIC residence time in solution and the activity of the enzyme carbonic anhydrase. The model is used with new and previously published oxygen isotope data to re-examine the cause of αc/w variations for inorganic calcite, ostracod calcite and coral aragonite. Implications for paleoclimate reconstructions based on the carbonate-18O/16O proxy are discussed.