I materiali vitrei di Herdonia (Foggia, Italia): studi di caratterizzazione e ipotesi di provenienza

The sample set from the cistern of the domus B consists of 21 glass sherds (cups, lamps, glasses/lamps, goblets, jugs/bottles, unidentified typologies) and 4 production wastes dated to the first half of the 5th century A.D. They are characterised by various shades of colour (yellow, olive-green, green-yellow, yellow-red, blue-green, blue, green, light and dark green), but colourless glass is also represented. Samples were analysed by optical microscopy (MO), scanning electron microscopy (SEM-EDS), electron Probe Micro Analysis (EPMA), X-ray fluorescence (XRF), inductively coupled plasma optical emission spectrometry (ICP-OES), inductively coupled plasma mass spectrometry (ICP–MS) and instrumental neutron activation analysis (INAA) in order to characterise both the texture and composition of glasses, and to identify vitrifying, stabilizing, flushing and colouring agents. Our data were compared with those available in the literature in order to investigate the provenance of materials. All finished products were produced from a mixture of siliceous sand and natron; however, they can be divided into GROUP A, comprising yellow, olive-green and green-yellow glasses, and GROUP B, represented by yellow-red, colourless, blue-green, light green and blue glasses. GROUP A is characterised by lower SiO2, K2O and CaO contents and higher Na2O and MgO contents than GROUP B, suggesting a different source of sand and a different recipe. Furthermore, Sr and Zr contents seem to indicate for GROUP A the use of shell and a different sand from that of the Levantine coast, while for GROUP B the employment of a Levantine sand rich in shell. The colour of the artefacts is probably due to both the Fe3+/Fe2+ ratio and Fe2O3 and MnO contents, but further studies on the oxidation state of Fe will be necessary to clarify the question. Unfinished products were produced from natron and a siliceous sand and have an intermediate composition with respect to GROUPS A and B, even if more similar to GROUP B. Flushing contents indicate a Na2O-richer recipe or shorter production cycle with respect to finished products. As for colorants, the considerations relative to the finished products are also valid for unfinished products. As for the provenance of materials, GROUP A shows analogies with glasses included in the HIMT group (Carthage, Cyprus) of unknown origin. The composition of GROUP B samples is similar to that of Levantine glasses, suggesting an import of raw glass or artefacts from the Syro-Palestinian coast. They were apparently produced from Belus sand mixed with natron. Lastly, although the provenance of unfinished products is unclear, they are compositionally similar to the glasses produced from natron and Belus sand.