AN AVERAGE WATERVILLE FORMATION PSEUDOSECTION: PREDICTED PHASE COMPOSITIONS AND COMPARISON TO GARNET-BIOTITE THERMOMETRY

An average Bt-zone pelite composition for the Waterville Formation, Maine, was used to construct a P-T pseudosection in the MnNaCaKFMASH system using thermodynamic data of Holland & Powell (1998). Predicted Grt, Ms, Bt, and St compositions are discussed, followed by a comparison of Grt-Bt thermometry for aluminumsilicate-bearing samples to the pseudosection Grt-in line and the And-Sil boundary.

Pelitic assemblages of the Waterville Formation define Bt, Grt, St-And, and Sil isograds with increasing T developed at 3-4 kilobars (Ferry, 1980). Pseudosection topologies at these P-T conditions consist of 3 dominant assemblages: Bt-Chl-Ms-Pl-Qtz (AS1), Grt-Bt-Chl-Ms-Pl-Qtz (AS2), and Grt-Bt-Ms-Pl-Sill-Qtz (AS3). As generally observed in nature, Bt octahedral Al and Fe/(Fe+Mg) increases with temperature in all 3 assemblages at constant P, while Grt Fe/(Fe+Mg) decreases with increasing T in AS2 and 3. With increasing T, the phengite component and K/(K+Na) of Ms decreases in AS1 and 2, while in AS3 the Al content remains relatively constant, Fe content decreases, and Mg and K/(K+Na) increases for Ms.

As previously reported, Fe-Mg partitioning between St and Grt predicted by the Holland and Powell thermodynamic data is opposite to that typically observed in nature. The pseudosection predicts St to have a higher Fe/(Fe+Mg) ratio than Grt, contrary to what is observed in the analyses of Ferry (1980), suggesting a problem in the relative thermodynamic properties of Grt and St, or the spurious effect of non-system components (e.g., Zn in St) on partitioning.

Utilizing activity models used in pseudosection construction, Grt-Bt thermometry was applied to And and Sil bearing samples using compositions reported in Ferry (1980). Temperatures calculated for And- and Sil-bearing samples plot in the And and Sil stability fields, respectively, of Holland & Powell (1998). However, many temperatures calculated for And bearing samples actually lie at T’s below the predicted Grt-in line. This discrepancy is likely due to either variation in bulk-rock compositions (especially Mn), or problems with thermodynamic data and activity models.