Metal prills in primary glass: A puzzling aspect of the production process

Primary glassmaking in the first millennium CE during Roman and post Roman periods was conducted in large heating installations, the so-called tank furnaces, in two main regions of the East Mediterranean, namely the Levant and Egypt. Especially in Israel, a large number of tank furnaces in various sites such as Apollonia, Bet Eli'ezer (Hadera) and Jalame have been discovered, giving new insights regarding the technology of glassmaking in the first millennium CE. In this paper, we investigate the existence of metal prills embedded in raw glass at the contact layer with a tank furnace floor. The glass chunks were found in a refuse pit at Umm ez-Zinat, southern Carmel, Israel, and are dated to the Late Roman-Early Byzantine periods (fourth-early fifth centuries CE) based on glass vessels found with the debris of primary and secondary production. Another three prills from Bet Eli'ezer, the massive glass industry dated to the seventh-early-eighth centuries CE, were also analysed to compare any similarities/differences between the prills from the two sites. Results show that the prills vary greatly in shape and composition. Their morphology ranges from rounded spheres to asymmetric ellipses, all showing inclusions indicating the source material was chemically heterogeneous. The prills from Umm ez-Zinat are mainly lead, with minor contribution of copper, tin, antimony and silver as inclusions. Two of the three prills from Bet Eli'ezer are copper alloyed with lead, and one has a lead matrix with various copper inclusions. We investigate possible scenarios to understand the nature of the metal in the glass, how and when it was added to the glass melt, and possible technological and ritual interpretations for its presence.