Table 1.
Mean chemical compositions of mosaic tesserae and window glass from Khirbat al-Minya as measured by LA-ICP-MS divided according to primary glass type and colour.
Fig 1.
Base glass characteristics of the architectural glasses from Khirbat al-Minya.
(A) TiO2/Al2O3 and Al2O3/SiO2 ratios of the tesserae in comparison with selected primary glass reference groups; (B) separating the tesserae according to their affiliations with primary production groups as reflected in their TiO2/Al2O3 and Al2O3/SiO2 ratios; (C) K2O and MgO concentrations of the window glasses compared to glass reference groups distinguishing natron from plant ash glasses; (D) K2O/P2O5 and Cr/La ratios differentiate between plant ash glasses from different geographical regions. Data sources: [1, 10, 36–47].
Fig 2.
Base glass characteristics of the Levantine type mosaic tesserae and window glass fragments from Khirbat al-Minya.
(A) CaO/Al2O3 versus Na2O/SiO2 ratios compared to published data from Apollonia, Jalame and Bet Eli‘ezer; (B) MnO and CaO contents for the tesserae distinguish different base glasses and different degrees of recycling. The line at 0.03% MnO indicates the proposed threshold below which manganese is considered a natural impurity of the silica source; (C) potash and phosphorus contents are positively correlated in the bulk of the samples with P2O5 levels of up to about 0.25%, indicative of re-melting and/or recycling. Higher P2O5 contents are probably due to the use of calcium phosphate as an opacifier. Data sources: [1, 6, 10, 51]. Asterisks indicate reduced and normalized compositions.
Fig 3.
Selection of glass tesserae found at Khirbat al-Minya.
The majority of the tesserae were made with Levantine or Egypt 1a primary glass. Some colours were exclusive to one type of primary glass: all the purple tesserae were made with Egypt 1a glass, while the aqua, black, blue, red and gold leaf tesserae were made with Levantine glass. Green, yellow, turquoise and olive tesserae were identified in both primary glass groups. Scale bar is 1 cm.
Fig 4.
Backscattered electron images of selected samples in cross-section.
(A) Yellow Levantine sample NS-T12 shows characteristic incomplete mixing of base glass and colouring agents with phases and particles of different compositions, given as means of three point analyses for selected oxides. Phases analysed are bright inclusions, the bright lead-rich phase, a mid-lead phase, and a dark lead-poor phase; (B) manganese oxide inclusions identified in purple sample NS-T40 of an Egypt 1a base glass; (C) turquoise Egypt 1a tessera NS-T30 exhibits a large inclusion of calcium phosphate, with a reaction zone and bubbles surrounding it. Also visible are small crystals of opacifier with sodium from the surrounding glass substituting for calcium in a cation exchange [58]. Tables report point analyses of each inclusion; (D) copper sulphide prills and some calcium phosphate in olive Egypt 1a tesserae (NS-T36).
Fig 5.
Characteristics of the colouring and opacifying agents used for the tesserae from Khirbat al-Minya.
(A) Egypt 1a tesserae exhibit positive correlations between PbO and Bi that suggest a common lead raw material; (B) CuO and ZnO are positively correlated in the Egypt 1a glasses, while the Levantine glasses have more variable contents; (C) positive correlations of CaO and P2O5 in the Egypt 1a tesserae with a ratio of approximately 4:3 CaO:P2O5 is consistent with the use of hydroxyapatite Ca5(PO4)3(OH) as opacifying agent. A few Levantine samples have also been opacified by calcium phosphate.
Fig 6.
Workflow diagram for the Egypt 1a tesserae found at Khirbat al-Minya.
Correlations between lead and bismuth, and zinc (and tin) and copper, suggest that the same raw materials were used to create different colours within the Egypt 1a group.