Conventional high temperature Fe(Ni)CrAl based alloys rely for their oxidation resistance on the formation of protective alumina scales. These form on the alloy surfaces during high temperature exposure due to selective oxidation of the alloying element aluminium. Several authors have shown, that, in spite of their high aluminium content, {gamma}-TiAl based alloys do not easily form protective alumina scales during oxidation at temperatures of about 700--900 C. Therefore it was frequently assumed, that in equilibrium with two-phase ({alpha}{sub 2}-Ti{sub 3}Al + {gamma}-TiAl) alloys, titania rather than alumina is the stable oxide phase. This assumption was verified by thermodynamical considerations by Rahmel and Spencer and Luthra. The expected formation of an {alpha}{sub 2}-Ti{sub 3}Al sub-surface depletion layer would even increase the larger stability of titania relative to alumina. However, recent careful analyses of oxide and depletion layer composition of {gamma}-TiAl alloys revealed, that in spite of the observed formation of rapidly growing mixed titania/alumina scales, alumina rather than titania is the stable oxide phase in equilibrium with {gamma}-TiAl based alloys. In the present paper it will be shown, that this apparent contradiction can at least partly be explained by the probable existence of a ternary Ti-Al-O compound, which has, as farmore » as known to the authors, not been accounted for in the explanation of {gamma}-TiAl oxidation so far.« less