Glass production at Jalame, Israel: Process, composition and relationship to Roman glass in Europe

Following an initial micro-XRF assay, 42 samples of glass from the twentieth-century excavations of the fourth-century CE production site at Jalame, Israel, were selected and analysed by EPMA, with subsets for trace elements by LA-ICP-MS (n=29 samples) and for Hf isotopes by MC-ICP-MS (n=6). Comparison of major elements for ten samples originally measured in Brill (1988) was good, although the earlier SiO2 values are slightly high. Analyses of furnace debris and contaminated glass imply that glass furnaces were made of, or lined with a lime-rich material, leading to contamination by CaO, rather than Al2O3. This confirms that higher alumina contents measured for Jalame glass relative to Roman Mn-decolourised glass of first-third century Europe are significant and it is suggested that the source of the earlier Roman glass lies further North on the eastern Mediterranean coast. A wide range of trace elements are shown to have entered the glass with the addition of manganese, so caution should be exercised when interpreting trace element and isotopic data on glass with significant amounts of Mn, although the effects on REE and some of the lithophiles most often used in provenancing appear minor or insignificant in Jalame glass with MnO below 1.3 wt%. For example, the hafnium isotope composition of the Jalame material does not appear to have been significantly affected and falls within the previously determined Levantine range, distinguishable from Egyptian glass. Associated trace elements indicate that two Mn sources were exploited and this raises the possibility of further investigation of Mn sources in early glass production. The wide range of added manganese suggests problems in obtaining a homogeneous batch and it is suggested that (a) a single initially inhomogeneous furnace load melted to a range of colours, which were subsequently colour-sorted as chunks, and/or (b) the recycling of poorly fused manganese-decolourised glass into later manganese-free batches resulted in the observed variation. The cobalt pigment used was the high-nickel late antique type and allows the introduction of this source to be pushed back from the sixth to the fourth century CE. The use of antimony in glassmaking ceased at around the same time, and it is speculated that these changes were related.