Abstract. The ferric iron content in hydrothermally altered ultrabasic rocks and their major minerals, serpentines and Mg-chlorites, is important for establishing the oxidation state budget from oceanic ridges to subduction zones, in carbonaceous chondrites, and for modeling phase equilibria. A compilation of literature Mössbauer spectroscopic data on serpentines and magnesian chlorites from high-pressure ophiolites yields much lower ferric-to-total-iron ratios (Fe3+ / Fetotal) than those obtained on similar samples by X-ray absorption near-edge spectroscopy (XANES), leading to contradictory estimates of the ferric iron budget of subduction zones. New Mössbauer analysis of antigorite and Mg-chlorite samples from suites of high-pressure ophiolitic terrains of various Phanerozoic ages confirms the low and homogeneous values previously obtained by this technique, while lizardite inherited from oceanic hydrothermal alteration is ferric iron rich. We argue that XANES values may be biased by photo-oxidation when samples have a high Mg content, which is the case for serpentines and chlorites from subduction zones. Photo-oxidation is less important in Fe-poor phyllosilicates of the mica and talc families and does not affect the Fe-rich serpentines (greenalite, cronstedtite) of meteorites or Fe-rich terrestrial phyllosilicates. Mössbauer Fe3+ / Fetotal ratios of serpentine confirm the occurrence of a major redox change at the lizardite–antigorite transition near 300–400 ∘C rather than at the dehydration of antigorite at 500–650 ∘C in serpentinites from high-pressure ophiolites.