The advantages and limits of empirical, semi-empirical and thermodynamic methods devoted to the estimation of chlorite-formation temperature are discussed briefly. The results of semiempirical and thermodynamic approaches with different assumptions regarding the redox state of iron in chlorite are compared for a large set of natural data covering a range of pressure conditions from a few hundred bars to 18 kbar and temperature from 100 to 500°C. The T-XFe3+ evolution estimated using the thermodynamic approach of Vidal et al. (2005) shows a systematic increase in XFe3+ with decreasing temperature, which is compatible with the decrease in aO2 buffered by magnetite- or hematite-chlorite equilibrium. This trend as well as the observed increase in vacancies in chlorite with decreasing temperature is interpreted as the incorporation of Fe3+-sudoite. The standard-state properties of this endmember have been derived to reproduce the observed T-aO2-XFe3+ evolutions. It can be used to estimate T-aO2-XFe3 values with a Chl-Qtz-H2O multi-equilibrium approach. When combining our results with those of other studies published recently, it appears that thermodynamic approaches and mapping techniques developed for metamorphic rocks can be used to discuss the conditions of formation of very low-grade rocks where kinetics is much more sluggish than in metamorphic rocks. This requires use of appropriate analytical tools and techniques with a spatial resolution of a few hundred nanometres.