Geochemical and mineralogical distinctions between Bonnin and Morris (Philadelphia, 1770–1772) porcelain and some contemporary British phosphatic wares

Abstract

AbstractThe major element compositions of 15 ceramic sherds from the Bonnin and Morris factory site were determined by electron microprobe. Thirteen samples are phosphatic; the others consist of (a) “soapstone” (magnesian/plombian) and (b) true porcelain, and are interpreted as exotic artifacts, as is one compositionally distinct (relatively SiO2‐poor, P2O5+CaO‐rich) phosphatic sample. Although long considered to be virtually indistinguishable from Bow porcelain (London: ca. 1747–1776), the phosphatic Philadelphia wares have a relatively low mean CaO/P2O5 ratio (3.3 versus 3.8; molecular proportions) and high alumina content (6.6 versus 5.4 wt % Al2O3). Furthermore, unlike Bow, the Bonnin and Morris samples contain calcic plagioclase (bytownite), and in some instances, an orthoclase‐rich ternary feldspar. The preservation of calcic plagioclase indicates that Philadelphia porcelain was fired at (rather than above) the thermal minimum in the An‐SiO2‐C3P system, although the presence of Na (and other fluxes) in these wares precludes the exact determination of the maximum firing temperature from this phase diagram. These wares are also distinctive insofar as the phosphate and melt phases can contain small amounts of lead; they have bulk lead contents of approximately 0.1–1.2 wt % PbO. This component has not been detected in the body of Bow or other contemporary British phosphatic porcelains. Their principal similarity lies in the fact that both wares contain sulfate. In addition, the glazes on Bonnin and Morris porcelain (e.g., PbO ∼ 35–50 wt %; SnO2 ∼ 1–2%) compositionally resemble those used at Bow. If feldspar is formed at all, then Al‐poor phosphatic porcelain (or those with low CaO/P2O5 ratios) will have comparatively low modal calcic plagioclase contents, thereby allowing the rapid depletion of this mineral via resorption by the melt phase during vitrification. Such appears to have been the case for analyzed Bow porcelain, which is therefore interpreted to have been overfired (sensu lato) relative to its Philadelphia counterpart. Conceivably, calcic plagioclase could be preserved in low‐Al wares that were fired only briefly at vitrification temperatures. Given the role of firing history in governing the mineralogy of porcelain, compositional criteria are more reliable for distinguishing these wares. © 2001 John Wiley & Sons, Inc.