Given that secondary magnetite is common in serpentinites, it is clear that serpentinites are oxidized rocks. Questions remain, however, concerning the distribution of ferric iron among magnetite and serpentine minerals and the role of ferric iron-rich serpentine in the formation of secondary magnetite. Direct determination of ferric iron in serpentine is not possible using an electron microprobe. We show, however, that the stoichiometic effects of ferric iron substitutions are detectable, although not quantifiable, by microprobe. First, we demonstrate that for studies that provide both microprobe analyses of major elements of serpentine and Mössbauer analysis of ferric iron, substitution effects are obvious. Next, it is equally clear that the early veins forming at the onset of olivine hydration (type 1 veins) show no indication of the presence of ferric serpentine, although a small amount of ferric 'brucite' may occur. Finally, we show that secondary (type 2) veins, which form as the system becomes open to fluids in equilibrium with plagioclase or pyroxene, contain, in addition to significant alumina, stoichiometric indications of ferric iron substitution. The serpentine in these veins is magnesian, usually with Mg#s around 96-98. Thus, even if a significant proportion of this iron is ferric, it comprises only a small fraction of the total ferric iron budget of the rock. Given that reduced iron is known to be abundant in early-formed brucite and early-formed serpentine and given that brucite, in particular, is absent from evolved serpentine veins, we propose that most magnetite in serpentinites forms as a tertiary product via oxidation of brucite.