• journal_title：The Canadian Mineralogist
• Contributor：J. Victor Owen
• Publisher：Mineralogical Association of Canada
• Date：2012-10-01
• Format：text/html
• Language：en
• Identifier：10.3749/canmin.50.5.1255
• journal_abbrev：Can Mineral
• issn：0008-4476
• volume：50
• issue：5
• firstpage：1255
• section：Articles

Double corona structures separating relict glass frit from metakaolinite in a teapot lid excavated from the Bow porcelain factory site in Stratford, east London record a history of partial melting and subsequent subsolidus reaction between the quenched melt phase and metakaolinite. The sample has an aluminous-siliceous-calcic composition consistent with1<sup>stsup> patent (1744) Bow porcelain. This patent specified a ceramic paste containing an alkali-lime glass frit and kaolin. The glass behaved as an alkali delivery system to the clayey matrix of the ware. Diffusion of alkalis commenced at subsolidus temperatures (probably in the glass transition), rendering the relict frit relatively refractory. Partial melting subsequently occurred near the interface between the relict frit particles and the fluxed clayey matrix. This created inner, quenched melt (glass) coronas on the frit, now represented by symplectic intergrowths of a silica polymorph, diopside ± pseudowollastonite. The outer edge of the glass coronas served as a crystallization front from which calcic plagioclase (~An<sub>62sub>) microlites grew, extracting CaO from the coronal glass as they advanced into the metakaolinite matrix, gradually consuming it. These outer, labradorite coronas are porous, suggesting that this was a subsolidus process. Unlike soluble alkali carbonates used in some other porcelain pastes, the glass frit represented a point source of fluxing components. The proportion, composition, grain size, and distribution of this phase therefore controlled the melt fertility of the paste on a domainal scale in this sample.