Abstract In the Mt. Franks area of the Willyama Complex, microfabric evidence suggests that the alteration of andalusite to sillimanite has taken place by a process similar to that suggested by Carmichael (1969). Andalusite is pre- to syn- S 2 in age. Alteration to “sericite” has resulted in the formation of “sericite” laths, some of which are crenulated about S 2, and some which are syn- and post- S 2. “Fibrolite” occurs in these andalusite—“sericite” aggregates within the sillimanite zone and is wholly embedded in “sericite”. “Fibrolite” is pre- to syn- S 2 in age. This evidence is interpreted as suggesting that the formation of sillimanite from andalusite took place via a “sericite” phase. Further microfabric observations are interpreted to imply constant volume for the reaction aluminosilicate → “sericite”. This suggests a situation in which Al 3+ is relatively mobile but Al 4+ is relatively immobile. This suggestion differs from Carmichael's (1969) idea of Al 3+ immobility.