In order to assess the stability of the primary triphylite + sarcopside assemblage, we performed hydrothermal experiments between 400 and 700 °C (Ni/NiO oxygen fugacity buffer, P = 1 kbar), starting from the Li(Fe2+2.5–xMn2+x)(PO4)2 (x = 0.0, 0.5, 1.0) compositions, which represent the ideal compositions of triphylite + sarcopside assemblages in which both minerals occur in a 1:1 molar ratio. The triphylite + sarcopside assemblage is observed in all experiments, associated with other phosphates like (Fe2+,Mn2+)2P2O7, (Fe2+,Mn2+)Fe3+2(PO4)2(OH)2·nH2O, or Fe3+4(Fe2+,Mn2+)3(PO4)6. Electron-microprobe and SIMS analyses show a progressive decrease of the Li contents in the triphylites, balanced by an increase of their Fe2+-contents, when the temperature increases. These compositional changes are due to the increase of the triphylite–sarcopside miscibility along the Li2(Fe2+,Mn2+)2(PO4)2–Fe2+(Fe2+,Mn2+)2(PO4)2 solid solution; the experimental phase diagrams can consequently be used as a geothermometer to calculate the exsolution temperatures of the assemblages. A linear fit of the experimental data leads to the general equation: T (°C) = (–142 * XFe) – (773 * Li pfu) + 1131, where XFe = Fe/(Fe + Mn). The uncertainty is around ±15 °C, and the influence of pressure is assumed to be negligible. By using this equation, exsolution temperatures were calculated for nine triphylite–sarcopside assemblages from pegmatites; these temperatures do not represent the crystallization temperatures of the phosphate nodules, but correspond to the closing temperature of the triphylite–sarcopside element exchange. Nevertheless, these temperatures, between 276 and 397 °C, are in fairly good agreement with those generally accepted for the crystallization of primary phosphate assemblages in granitic pegmatites.