Two techniques are compared for determining the ferric iron content in FexO using the electron microprobe: (i) a direct method termed the flank method (Höfer et al., 1994), which is based on the concomitant changes of both the intensity ratio and the wave length of the Fe Lα and Fe Lβ X-ray emission lines with Fe2+/Fe3+, and (ii) an indirect method by calculating Fe2+/Fe3+ from elemental microprobe analyses which may include oxygen. The precision and accuracy of Fe3+/∑Fe determined with the flank method is ± 0.02 (1s). Similar values are achieved with the indirect method by determining the Fe and O contents, while measurement of iron alone and calculation of the oxygen content from the difference to 100% is less accurate. The results are valid only for this relatively simple two component system and cannot be transferred to multiple component systems without further calibration. At present, the flank method can be applied if the bulk chemical composition and crystal structures of the samples are known and calibration curves with similar materials have been established. The indirect method is restricted to samples with Fe as the only multivalent cation. The precision of the flank method is largely affected by the accuracy of the spectrometer position. Thus, a method is developed to optimize the spectrometer calibration using the Fe Kα1 9th order emission line.