Seeking the Invisible

In this chapter, the results of a large-scale programme of glass chemical data collation and analysis are presented to argue that glass recycling in the Roman period was far more extensive than has been realized to date. Over 6,000 samples of glass from published analyses since 1999 are interrogated in order to produce a long-term and large-scale data set for recycling evidence. In addition, literary evidence, and data from shipwrecks and other archaeological sites are summarized and considered, and basic modelling is used to suggest the sorts of patterns we should be looking for in compositional data for glass. It is argued that the quantity of Roman glass that was recycled has been underestimated due to several factors, including a lack of consideration of ‘like with like’ mixing, a lack of consideration of the formation processes of the archaeological record, and analytical sampling bias towards colourless vessel glass.