I have studied hercynite - cordierite symplectites in LP-HT
metapelites in the Bushveld complex in South Africa. In that case
they appeared to replace chiastolitic andalusite at pressures
probably as low as 2 kbar and temperatures not exceeding 750°C. All
the phases were rather Fe-rich and the matrix was often Qtz-rich.
Just as Dave Waters pointed out, the replacement appeared
non-isochemical with equilibration probably only being achieved at
very small scale, producing namely strong a(SiO2) gradients between
the matrix and the Crd-Spl intergrowths.
This was described in our paper published this May in the Journal of
metamorphic geology (JMG 19, p. 311-325), where you may find a PT
grid calculated with THERMOCALC (containing some of the Spl-Crd
producing equilibria) together with some other potentially
interesting phase diagrams, as well as some discussion and references
to other similar occurrences.
Geosciences Rennes email: [log in to unmask]
Université de Rennes 1 tel: (++33) 126.96.36.199.79
Campus de Beaulieu - Bat. 15 fax: (++33) 188.8.131.52.80
F - 35 042 RENNES CEDEX