KYANITE COLOR AS A CLUE TO CONTRASTING PROTOLITH COMPOSITIONS FOR KYANITE QUARTZITES IN THE PIEDMONT PROVINCE OF VIRGINIA

Kyanite quartzite at Willis Mountain (Buckingham County) in the central Piedmont Province of Virginia contains pale blue-gray to white kyanite, whereas kyanite at nearby Baker Mountain (Prince Edward County) is a deep blue to blue-green. Otherwise, rocks from both areas are mineralogically similar, and contain ubiquitous rutile, sporadic white mica, and locally abundant pyrite. We evaluated possible compositional causes for the kyanite color difference by analyzing grains from both localities by electron microprobe. Willis Mountain kyanite is almost pure Al₂SiO₅, but Baker Mountain grains consistently contain measurable amounts of Cr₂O₃ and/or FeO (up to ~0.4 wt% each). Thus, a combination of crystal field transition (Fe,Cr) and charge transfer (Fe²⁺®Fe³⁺) mechanisms is probably responsible for kyanite color at Baker Mountain. Relatively high Cr in Baker Mountain kyanite seemed surprising for such quartz-rich rocks, and prompted us to obtain whole rock major and trace element data for selected samples (3 from Willis Mt., 2 from Baker Mt.). The compositional data provide some key constraints on the nature of the quartzite protoliths, which have been interpreted either as sedimentary rocks (e.g., aluminous sandstone) or hydrothermally altered volcanic rocks. Major element compositions from both localities are dominated by Si and Al, consistent with nearly pure quartz-kaolinite protoliths. Levels of Cr are indeed high at Baker Mountain (~600 ppm), but below detection limits at Willis Mountain. Concentrations of many other trace elements (Ni, Co, Zn, Ga, Nb, Y, Rb) are extremely low, as are Ga/Al values. In contrast, Sn concentrations are uniformly high (10-26 ppm). Willis Mountain rocks have higher concentrations of Zr and higher Zr/TiO₂ values relative to rocks at Baker Mountain. Our results indicate that these deposits originated by metamorphism of volcanic rocks that had been previously altered by extreme leaching in acidic, hydrothermal fluids. The contrasting trace element signatures suggest that Baker Mountain rocks were derived from a mafic (basaltic) protolith, whereas the protolith at Willis Mountain was more evolved (possibly andesitic).