The standard-state thermodynamic properties of 14 Å Fe-amesite [Fe4Al4Si2O10(OH)8] were calculated from the reequilibration experiments of Parra et al. (2005) and natural data bearing on the Fe-Mg partitioning between chlorite and chloritoid over a wide range of pressure (P), temperature (T), and rock composition. The combination of these data with the three-site mixing model and thermodynamic properties of daphnite proposed by Vidal et al. (2001) led us to reappraise the evolution of the Margules parameter WGAlFe with pressure and temperature. The new data [H0fFe-Am = −7 607 460 J/mol, S0Fe-Am = 514.8 J/(mol·K), V0Fe-Am = 20.90 J/bar, WGAlFe = 1200 – 31T + 0.7(P – 1)] are compatible with all the natural data and reequilibration experiments, and in fair agreement with the results of synthesis. They are used to calculate the compositional evolution of chlorite with T, P, or aO2 at different bulk-system compositions. The calculated phase relations and chlorite compositions are compared with previous experiments conducted in the FeO-Al2O3-SiO2-H2O system. The addition of the Fe-amesite end-member increases the number of equilibria that can be calculated for any natural assemblage involving chlorite. Two independent equilibria are obtained for natural (Fe-Mg)-chlorite-quartz or diaspore assemblages. It is therefore possible to estimate metamorphic conditions for high-variance assemblages from the composition of chlorite. We discuss the reliability of such P-T estimates using natural chlorites occurring in low-pressure and low-temperature or medium- to high-pressure and temperature samples. Assuming that the proposed model and thermodynamic parameters are correct, these various natural data suggest that XFe3+ [= Fe3+/(Fe2+ + Fe3+)] in metamorphic chlorites is low (<0.1), whereas it can be as high as 0.3 in Si- and Mg-rich chlorites formed at low-P-T conditions.