Nederlandse Archeologische Rapporten 081 Early medieval glass production in the Netherlands A chemical and isotopic investigation Hongjiao Ma, Julian Henderson, Yvette Sablerolles, Simon Chenery and Jane Evans Early medieval glass production in the Netherlands a chemical and isotopic investigation Hongjiao Ma, Julian Henderson, Yvette Sablerolles, Simon Chenery and Jane Evans Colophon Nederlandse Archeologische Rapporten 81 Early medieval glass production in the Netherlands: a chemical and isotopic investigation Authors: Hongjiao Ma, Julian Henderson, Yvette Sablerolles (University of Nottingham), Simon Chenery and Jane Evans (British Geological Survey) With a contribution from Menno Dijkstra (University of Amsterdam) Authorisation: Hans Huisman and Rik Feiken (Cultural Heritage Agency of the Netherlands) Illustrations (unless otherwise stated): Julian Henderson, Yvette Sablerolles and Marjolein Haars (BCL-Archaeological Support) Cover design: J. Ranzijn (Happyfolio) Cover: A selection of tesserae from the Wierum terp (province of Groningen) (Photograph: Jelle Schokker of the Noordelijk Archeologisch Depot). Design and layout: Xerox/Osage ISBN/EAN: 978-90-76046-85-3 © Cultural Heritage Agency of the Netherlands, Amersfoort, 2023 Cultural Heritage Agency of the Netherlands P.O. Box 1600 3800 BP Amersfoort the Netherlands www.cultureelerfgoed.nl 3 — Contents Summary 5 Samenvatting 7 1 1.1 9 1.2 1.3 1.4 1.5 2 2.1 2.2 2.3 2.4 2.4.1 2.4.2 3 3.1 3.2 3.2.1 3.2.2 3.2.3 3.3 3.4 3.4.1 3.4.2 3.5 3.5.1 3.5.2 3.6 3.6.1 3.6.2 3.7 3.8 3.9 3.10 3.11 3.12 Introduction The research project ‘Early medieval glass production’ Research questions Research approach Structure of the monograph Acknowledgements Technology and raw materials used for Dutch early medieval glass Introduction Furnaces, raw materials and glass sources Evidence for the glass industry in early medieval Europe outside the Netherlands Scientific analysis of early medieval glass in Europe The principal glass types Summary of existing scientific analyses of early medieval Dutch glass before the start of this project Evidence for early medieval glass-working in the Netherlands Introduction Maastricht, Limburg Province Maastricht-Jodenstraat (MAJO) Maastricht-Mabro Maastricht-Rijksarchief Susteren-Salvatorplein, Limburg Province Wijk bij Duurstede (Dorestad), Utrecht Province Wijk bij Duurstede – Parkeerplaats Albert Heijn (PPAH) Wijk bij Duurstede – Veilingterrein and Frankenweg/Zandweg Utrecht, Utrecht Province Utrecht-Domplein Utrecht–Oudwijkerdwarsstraat Leidsche Rijn, Utrecht Province Leidsche Rijn-LR 51/54 Leidsche Rijn-Leeuwesteyn Noord Oegstgeest–Nieuw Rhijngeest Zuid (Rijnfront), Zuid-Holland Province Rijnsburg-Abdijterrein, Zuid-Holland Province Valkenburg-De Woerd, Zuid-Holland Province Den Haag-Frankenslag, Zuid-Holland Province Bloemendaal-Groot-Olmen, Noord-Holland Province Wijnaldum-Tjitsma, Friesland Province 9 9 10 10 10 3.13 3.14 Wierum, Groningen Province Deventer-Stadhuiskwartier, Overijssel Province 4 The materials, analytical techniques and methodology Introduction An overview of the sites and glass samples Electron probe microanalysis (EPMA) for major and minor chemical composition Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) analysis for trace element compositions Thermal ionization mass spectrometry (TIMS) analysis to determine Nd and Sr isotopic compositions How analytical data is used in this study 4.1 4.2 4.3 4.4 4.5 13 13 13 16 17 17 19 29 29 30 31 35 36 37 39 40 41 43 43 46 47 47 48 48 49 51 52 53 53 4.6 5 5.1 5.2 The analytical results and discussion Introduction Glass samples from Maastricht (Jodenstraat and Mabro sites) 5.2.1 Naturally coloured and cobalt blue bead-making glass waste 5.2.2 Highly coloured opaque glass 5.2.3 Vitreous and semi-vitreous materials attached to the crucibles 5.2.4 Glass artefacts 5.2.5 Crucibles from the Mabro site, Maastricht 5.3 Glass samples from Gennep 5.4 Glass samples from Wijnaldum 5.4.1 Highly coloured opaque glass beads 5.4.2 Colourless glass beads 5.4.3 Vessel glass 5.4.4 Bead production materials 5.5 Glass samples from Utrecht 5.6 Glass samples from Wijk bij Duurstede (Dorestad) 5.6.1 Vessel glass 5.6.2 Other glass 5.7 Glass samples from Susteren 5.7.1 Trail decorated glass beads 5.7.2 Window glass 5.7.3 Glass attached to crucibles 5.7.4 Vessel glass 5.8 Glass samples from Deventer 5.8.1 Wood ash glass 5.8.2 Natron glass 5.8.3 Mixed alkali glass 5.8.4 Plant ash glass 5.9 Nd-Sr isotope analysis 5.10 Discussion 5.10.1 The base glass used for bead making at Jodenstraat 55 56 59 59 59 61 61 62 64 67 67 67 67 68 70 72 72 74 74 74 75 77 78 78 79 79 80 81 81 82 82 82 83 83 84 85 85 86 89 89 4 — 5.10.2 The use of crucibles in on-site lead tin yellow colourant production in early medieval northwestern Europe 5.10.3 The separate production of a tin-based white opacifier at Maastricht, Jodenstraat 5.10.4 The other chemical characteristics of the glass and its archaeological implications 5.10.5 A comparison of 7th–11th century vessel glass from Comacchio with early medieval Dutch glass and the suggested supply of raw glass in the two areas 5.10.6 Wood ash glass and mixed alkali glass 5.10.7 Plant ash glass 5.11 Summary 6 6.1 6.2 Synthesis and conclusions The Early Merovingian period (450-550 AD) The Middle Merovingian period (550– 650 AD) 89 6.3 6.4 6.5 The Late Merovingian period (650 – 750 AD) The Carolingian period (750 – c. 850 AD) The late phase, including the Ottonian period (c. 850 – c. 1000 AD) 101 102 104 Answering the research questions 107 91 7 92 95 96 97 98 99 99 100 Bibliography 115 Appendices Appendix I sample list Appendix II major and minor chemical compositions of samples analysed by electron probe microanalysis Appendix III trace element chemical compositions of samples analysed by LA-ICP-MS Appendix IV photos of the samples from Maastricht and Utrecht 125 126 154 170 183 5 — Summary This monograph brings together for the first time comprehensive combined archaeological, technological and scientific investigations, using chemical (major, minor and trace element concentrations) and isotopic (87Sr/86Sr and 143Nd/144Nd) analysis, of early medieval glass production in the Netherlands. We selected 276 samples of glass from Gennep, Maastricht (the Jodenstraat and Mabro sites), WijnaldumTjitsma, Utecht (the Domplein and Oudwijkerdwarsstraat sites), Susteren-Salvatorplein, Wijk bij Duurstede (the Hoogstraat and vicus sites) and Deventer-Stadhuiskwartier, dating to between the late 4th and 11th centuries covering the Merovingian and Carolingian periods for compositional analysis. In addition, 20 samples were subjected to isotope analysis. The results of our trace element and isotopic analyses have provided new and highly significant insights into early medieval glass production in the Netherlands. Several different compositional types of glass have been identified. Both high and low lead glasses have been found along with pristine and recycled (sub-types of) natron glass (HIMT, Foy 2, Egypt II, Levantine II), plant ash glass, wood ash glass and mixed-alkali glass. The best evidence in early medieval Europe for the onsite production of the lead-tin yellow and tin oxide colorant/opacifier in crucibles excavated from the 6th-7th century AD Jodenstraat site in Maastricht is discussed in detail. It is associated with comprehensive evidence for the manufacture of brightly coloured monochrome glass beads (also found at Wijnaldum), a craft specialisation. The base glass for the beads was imported ‘pristine’ Foy 2 glass. Our results show that a higher proportion of Merovingian glasses were imported ‘pristine’ (Egyptian) glasses than the glasses used in the Carolingian period, when the majority of glasses were recycled, potentially multiple times. While it is often suggested that elevated levels of antimony and lead started to occur in the Carolingian period we have found elevated levels already in Merovingian glasses, an indication that a small proportion of Roman tesserae and/or coloured Roman vessel glass was being added to the glass melt then. By combining trace element and isotope analysis we have been able to demonstrate that Carolingian recycled glass contained a small proportion of wood ash glass. Wood ash glass with elevated concentrations of Cs, Rb, Ba and Sr started to be manufactured in Europe from around 800 AD and added to natron glass to produce mixed alkali glass. Even though mostly recycled glass was in use by this time much of the Dutch natron glass can still be attributed an ultimate source in Egypt (recycled Foy 2 glass), with limited evidence for the use of ‘pristine’ glass. Our analyses provides evidence that recycled Foy 2 was still in use as late as the mid 10th century AD. When the results of our analyses for Carolingian natron glasses are compared with contemporary (7th-11th century AD) northern Italian glasses from the site of Comacchio and Spanish glasses from Tolmo de Minateda an interesting contrast is revealed. Both sets of glasses are largely recycled: whereas Dutch recycled natron glass has an ultimate source in Egypt, the recycled natron glasses from Comacchio and Tolmo de Minateda show far more evidence for the use of imported Levantine glasses instead. Only a single example of Levantine (II) glass has been found amongst Dutch early medieval glass with no evidence that such glass formed part of the recycling process. This is a clear reflection of differing trade contacts and glass supplies between northern and southern Europe. By the 9th-11th centuries AD the widest range of glass types was in circulation and, perhaps surprisingly, glass from Deventer includes pristine glass. The glass from Deventer consists of thirteen natron glasses (three Roman, three pristine Egyptian II, six recycled Foy 2, one pristine Foy 2), one plant ash glass, four mixed alkali glasses and nineteen wood ash glasses. Like plant ash glass production in western Asia, the primary production of wood ash glass would have formed a decentralised network because wood ash was also widely available. Although glass beads were certainly made in the Netherlands along the Meuse valley in the Merovingian period, probably with a ‘permanent’ workshop in Maastricht (even if the bead makers took part in other industries) we suggest that bead workers were mobile further north, visiting Rijnsburg, Wijnaldum and perhaps ValkenburgDe Woerd. A very likely source of vessel glass probably existed in Cologne. Scientific analysis of Merovingian bowls and beakers (450-550 AD) shows a correlation between vessel type, chemical composition and colour, suggesting that glass of specific colours were selected during vessel manufacture. This may simply 6 — have been part of batch production but we suggest that colour selection was related to a memory of the exotic origin of the glass and drinking rituals, including the colour of liquid the vessels contained. We have shown that the key types of raw glass were from Egypt- Foy 2 and HIMT in the Merovingian period and Egypt II in the Carolingian period; a single Levantine II (punty) glass is the only example of Levantine glass we have found; mixed-alkali glass probably derived from northern France; wood ash glass perhaps from Belgium, northern France or more likely from Germany (perhaps using the Viking trade network); plant ash glass was imported from western Asia as beads and raw glass (also perhaps by the Vikings); the raw glass was made into characteristic early medieval vessel types and incorporated into glass beads. 7 — Samenvatting Deze monografie combineert voor het eerst uitgebreid archeologisch, technisch en natuurwetenschappelijk onderzoek, met chemische (hoofd- en sporenelementen) en isotopenanalyses (87Sr/86Sr and 143Nd/144Nd), van vroegmiddeleeuwse glasproductie in Nederland. We selecteerden 290 glasmonsters uit Gennep, Maastricht (Jodenstraat en Mabro), WijnaldumTjitsma, Utrecht (Domplein en Oudwijkerdwarsstraat), Susteren-Salvatorplein, Wijk bij Duurstede (Hoogstraat en vicus) en DeventerStadhuiskwartier, die met dateringen tussen de late vierde en de elfde eeuw de Merovingische en Karolingische perioden beslaan. Daarnaast werden twintig van deze monsters geselecteerd voor isotopenanalyse. De resultaten van de sporenelement- en isotopenanalyses geven belangrijke nieuwe inzichten in de vroegmiddeleeuwse glasproductie in Nederland. Er zijn verschillende typen glas met verschillende samenstelling geïdentificeerd. Zowel glas met hoge als met lage gehaltes aan lood zijn aangetroffen, naast vers en gerecycled natron glas (sub-typen HIMT, Foy 2, Egypte II, Levantine II), glas op basis van de as van planten of hout en gemengd alkali glas. Het meest overtuigende bewijs voor de lokale productie van lood-tin geel en tin oxide als kleurstof en opacifier in vroegmiddeleeuws Europa - smeltkroesjes uit de zesde-zevende eeuw n.Chr. opgegraven in Maastricht Jodenstraat - wordt in detail besproken. Tijdens de opgraving is er uitputtend bewijs gevonden voor het maken van helder gekleurde monochrome kralen, een gespecialiseerd ambacht. Het basisglas voor de kralen was geïmporteerd “vers” glas van type Foy 2. Onze resultaten laten zien dat een hoger percentage van Merovingisch glas bestond uit geïmporteerd “vers” (Egyptisch) glas, vergeleken met glas uit de Karolingisch periode, toen het meeste glas werd gerecycled – één of meerdere keren. Hoewel vaak wordt gesuggereerd dat verhoogde gehaltes aan antimoon en lood pas beginnen in de Karolingische tijd, doordat kleine hoeveelheden Romeinse tesserae of glas van Romeins glazen vaatwerk werden toegevoegd aan gesmolten glas, vonden we al verhoogde gehaltes van antimoon en lood in Merovingisch glas. Met een combinatie van sporenelement- en isotopenanalyses hebben we kunnen aantonen dat Karolingische gerecycled glas een klein aandeel hout-as glas bevat. Hout-as glas met verhoogde concentraties van cesium (Cs), rubidium (Rb), barium (Ba) en strontium (Sr) werd voor het eerst geproduceerd in Europe vanaf ongeveer 800 n.Chr., en het werd toegevoegd aan natronglas om gemengd alkaliglas te maken. Hoewel in deze periode vooral gerecycled glas in gebruik was, kan veel van het Nederlandse natronglas nog steeds worden gelinkt aan een oorspronkelijk herkomst in Egypte (gerecycled Foy 2 glas), met beperkte aanwijzingen voor het gebruik van “vers” glas. Onze analyses laten zien dat gerecycled Foy 2 glas zelfs nog in gebruik was in het midden van de tiende eeuw n.Chr. Als de resultaten van onze analyses van Karolingisch natron glas worden vergeleken met contemporain (zevende-elfde eeuws) NoordItaliaans glas uit Comacchio en Spaans glas uit Tolmo de Minateda komt een interessant contrast aan het licht. De assemblages bestaan vooral uit gerecycled glas, maar terwijl het Nederlandse gerecycled glas in oorsprong uit Egypte komt, bevat het glas uit Comacchio en Tolmo de Minateda aanwijzingen voor een Levantijnse oorsprong. Slechts één voorbeeld van Levantijns (II) glas is aangetroffen onder Nederlands vroegmiddeleeuws glas, en er zijn geen aanwijzingen dat dit soort glas een rol speelde bij recycling. Dit is een duidelijke weerslag van verschillen in handelscontacten en glasleveringen tussen Noord- en Zuid-Europa. In de negende-elfde eeuw was de variatie in glastypes die werden gebruikt het grootst, inclusief – wellicht verrassend – “vers” glas uit Deventer. Het glas uit Deventer bestaat uit dertien stuks natron glas (drie keer Romeins, drie keer “vers” Egyptisch II, zes keer gerecycled Foy 2, een keer “vers” Foy 2), een keer plant-as, vier keer gemengd alkali en negentien keer hout-as glas. Net als plant-as glasproductie in West-Azië vormde de productie van hout-as glas een decentraal netwerk omdat hout as algemeen beschikbaar was. Hoewel Merovingische glazen kralen zeker werden gemaakt in Nederland in de Maasvallei, waarschijnlijk met een permanente werkplaats in Maastricht (zelfs als de kralenmakers ook andere ambachten uitoefenden), waren de kralenmakers verder naar het noorden waarschijnlijk mobiel. Daar bezochten ze Rijnsburg, Wijnaldum en mogelijk ook Valkenburg – de Woerd. Keulen was hoogstwaarschijnlijk een belangrijke bron voor glazen kommen en bekers. Natuurwetenschappelijke analyse van 8 — Merovingische kommen en bekers (450-550 n. Chr.) tonen een correlatie tussen typologie, chemische samenstelling en kleur, wat suggereert dat glas met specifieke kleuren werd geselecteerd voor deze toepassing. Dit zou simpelweg onderdeel kunnen zijn van grootschalige productie, maar we suggereren dat de selectie van de glaskleur te maken had met herinneringen aan de exotische herkomst van het glas en drankrituelen, inclusief de kleur van de vloeistof die in de glazen had gezeten. We hebben laten zien dat de belangrijkste types ruw glas tijdens de Merovingische periode Egypte – Foy en HIMT - waren, en tijdens de Karolingische periode Egypte II. Het enige voorbeeld van Levantijns glas is een Levantijns II puntige glastype. Gemengd alkaliglas kwam waarschijnlijk uit Noord-Frankrijk, hout-as glas mogelijk uit België, Noord-Frankrijk of Duitsland (wellicht via het Viking handelsnetwerk verkregen). Plant-as glas kwam uit West-Azië als kralen en ruw glas (ook wellicht via de Vikingen); het ruwe glas werd verwerkt tot typische vroegmiddeleeuwse kommen en bekers, en tot glazen kralen. 9 — 1 Introduction 1.1 The research project ‘Early medieval glass production’ The research project ‘Early medieval glass production’ was one in a series of studies referred to as Pre-Malta research (‘pre-Malta onderzoek’), and as such falls under the programme Knowledge for Archaeology. The programme aims to obtain datasets from (not yet fully elaborated) excavation data before the introduction of the Valetta convention in 2007. With the use of currently developed research methods and techniques, the Pre-Malta research programme enables the study of archaeological remains from old excavations, which means that substantial new knowledge can be obtained about the past. This research project involved a series of important goals related to the production technology, provenance, glass supply, recycling, trade and use of early medieval glass in the Netherlands, building on, and expanding significantly on, existing published research. The main goals were: • To carry out a full chemical and isotopic analysis of all available glass samples using cutting edge techniques. • To establish the raw materials used to make the transparent, translucent and opaque glasses samples. • To consider whether the glass has been recycled. • To attempt to suggest a source for the glass (i.e. provenance). • To investigate the change in glass raw materials over time. • To establish if there are any sub- and supraregional supply patterns for unrecycled early medieval glass found on Dutch early medieval sites from within Europe, the Levant, Iraq and Iran. • To investigate in more detail whether there are chronological changes in the use of pure imported as opposed to recycled glass moving from the Merovingian into the Carolingian period. • To compare the glass compositions and technologies used in the manufacture of glass beads and vessels and investigate if there is evidence for the use of raw material specialization. • To establish if there is any evidence for local specialisation of glass bead production as reflected in their chemical compositions. • To investigate the evidence for the production of lead-tin oxide opacified glass found in the Netherlands given the large number of crucibles containing a yellow substance that have been discovered in early medieval Dutch contexts especially in Maastricht and whether there is evidence for primary glass making. Scope of the research project The project focuses on the simple, monochrome beads, raw materials and production waste from Merovingian and Carolingian contexts along with contemporary vessel glasses.1 A small number of glasses dating to c. 900-1000 AD are included for comparison. The examination of polychrome beads and glass vessels have been included because they provide information to (better) answer the research questions. The starting point for this project’s research is Henderson and Sablerolles’ research plan from 2020. This plan is included as an appendix in the report ‘An Overview of Dutch Early Medieval glassworking, published chemical and isotopic analyses of glass beads and vessels, raw material provenance of beads and vessels, changes in raw material use over time and a plan for future scientific analysis’, which served as preparation for this project.2 The substantive information from that report has been largely incorporated into this report. 1.2 Research questions The central questions of the study are: 1. What raw materials were used in the local production of simple, monochrome Early Medieval beads? 2. Where were these raw materials obtained from? Sub-questions here are: i. What substances were used to make the different colours of glass in the artefacts tested? ii. What compositional groups can be distinguished in the glasses based on chemical analyses? 1 2 Between AD 480 and 987. Henderson & Sablerolles 2020. 10 — iii. What does this tell us about dating of primary glass production of these groups? iv. What do the isotope ratios (Sr, Nd) obtained from the glasses of selected compositional types tell us about the their origin and dating? v. What networks inside and outside the Netherlands were used in obtaining glass, including the colourants used? vi. To what extent were the raw materials or semi-finished products derived from primary production, or to what extent from systematic recycling of glass, including Roman? The research also provides building blocks for two NOaA questions3: • What are the nature, manifestations, extent and context of craft specialization? (NOaA 2.0 question 67) • Where do non-local raw materials of utilitarian objects come from? (NOaA 2.0 question 139) 1.3 3 https://noaa.cultureelerfgoed.nl. Research approach We have been able to carry out a comprehensive scientific analysis of a wide range of early medieval glass samples and production waste. The scientific techniques we used were Scanning Electron Microscopy, Electron Probe Microanalysis, Laser Ablation Inductively Coupled Plasma Mass Spectrometry and Thermal Ion Mass Spectrometry, resulting in the largest database of chemical and isotopic analyses for early medieval Dutch glass. The Covid-19 virus prevented us from travelling and taking new samples which made it difficult to plan and led to delays in scientifically analysing some samples. We had hoped to work on new glass samples excavated from secure archaeological contexts. Although possible for glass from Maastricht, Utrecht, Gennep, Wijnaldum, Susteren and Deventer, the context information for Dorestad glasses studied here were unavailable at the time of sampling in the early 1990s. It was nevertheless possible to provide dates for the Dorestad glass according to vessel form. This project has involved a collaborative team of archaeologists, archaeological scientists and a geologist: Hongjiao Ma, Julian Henderson, Yvette Sablerolles, Simon Chenery, Jane Evans and Menno Dijkstra linking archaeologists, archaeological scientists and geologists. 1.4 Structure of the monograph The remaining chapters of the monograph develop in a logical sequence. Chapter 2 provides essential information about early medieval European glass technologies, including raw materials, evidence of European early medieval glass production outside the Netherlands, the types of glass found in early medieval Europe followed by a review of published results for early medieval glass from the Netherlands. Chapter 3 discusses the existing archaeological evidence for early medieval glass production on the Netherlands. In Chapter 4 the sites from which glass samples used in this study are introduced briefly followed by a description of the three main analytical techniques used to investigate the samples chemically and isotopically: electron probe microanalysis (EPMA), laser ablationinductively coupled plasma- mass spectrometry (LA-ICP-MS) and thermal ion mass spectrometry (TIMS). Chapter 5 presents and discusses the results of the chemical and isotopic analyses of the glass and crucible samples. Chapter 6 is a synthesis of archaeological and scientific results according to chronological periods (450-550, 550-650, 650-750, 750-850 and 850-1000 AD) for the work together with conclusions. Chapter 7 provides succinct answers to the research questions and sub-questions listed in Section 1.2 above. 1.5 Acknowledgements We are grateful to W. Dijkman, Senior Conservator Archeologie en Erfgoed, Team Programma en Innovative, Centre Céramique– Kumulus – Natuurhistorisch Museum, Maastricht for allowing us to sample glass and glass working material from Maastricht, to H. Wynia, the municipal archaeologist of Utrecht for allowing us to sample glass from Utrecht, to J. Schokker of the Noordelijk Archeologisch Depot for allowing us to sample glass from Wijnaldum, A. Peddemors then curator at the Rijksmuseum van Oudheden, Leiden for permission to sample glass from Dorestad, W. van Es of the Rijksdienst voor het 11 — Oudheidkundig Bodemonderzoek also for permission to sample Dorestad glass and to E. Mittendorff, project leader archaeology, Deventer for sending us glass from Deventer so that we could sample it and both the former director of the Provincial Limburgs Museum in Venlo, G. Jansen and H. Stoepker, for permission to sample glass from Susteren. We are also grateful to Dr. A. Kronz of Göttingen University for sharing some unpublished data. We are very grateful for the support of H. Huisman and R. Feiken of the Cultural Heritage Agency of the Netherlands for being extremely supportive and for being very flexible throughout the project, at the very difficult time of Covid. Finally we are grateful to Dr Matthew Delvaux for permission to reproduce Figure 3.7 and Bill Bolton for producing publishable versions of all the Figures. 2 Technology and raw materials used for Dutch early medieval glass 2.1 Introduction This chapter provides the basic background information about Dutch early medieval glass technology including a discussion of the raw materials used which can be suggested from glass chemical compositions and the furnaces used for making the likely sources of glass in Europe and western Asia. The main primary raw materials are an alkaline flux, a silica source and a calcium source. Fluxes are provided by natron, plant ash or wood ash; silica is normally provided by sand; calcium is provided by shell fragments in sand or calcium compounds in plant ash or wood ash . Mineral rich colorants were added separately to these glasses or opacifiers were developed from them by heat treating the glasses. The chapter covers glass production dating to between the late Hellenistic period and the early Islamic period as well as evidence for centralized and decentralized production organisations. It also discusses the compositional evidence for the recycling of glass in the second half of the 1st millennium AD. A separate sub-section (Section 2.3) is devoted to a discussion of the evidence for the early medieval glass industry outside the Netherlands, especially in Belgium, Denmark France, Germany, Hungary, Ireland, Italy and the United Kingdom. The next sub-section (Section 2.4.1) discusses the main compositional types of glass in some cases associated with primary glass making sites such as in SyroPalestine, Egypt, Syria, Iraq and northern Europe. The main compositional types are pristine Egyptian and Levantine natron glasses, Roman glass, HIMT and its variations, plant ash glass, mixed-alkali glass and wood ash glass. Section 2.4.2 is a summary of the existing published chemical analyses of early medieval Dutch glass from Maastricht, Susteren, Dorestad, Rijnsburg, Wijnaldum, Lent, Borgharen and Sittard with interim interpretations. 2.2 13 — Furnaces, raw materials and glass sources Most early medieval glass found in the Netherlands is what is known as soda-lime (natron) glass. From about the 2nd century BC, during the Late Hellenistic period, this kind of glass was fused from raw materials in massive rectangular tank furnaces. The earliest example of a tank furnace yet discovered is in Beirut, possibly dating to the 2nd century BC.4 Roman glass tank furnaces have also been found, in Egypt5 and Syro-Palestine, such as Jalame.6 The Levant continued to be the primary centre for the production of raw furnace natron glass on a massive scale into the Byzantine period, especially in the 6th–8th centuries AD7 with a probable dip in the scale of production in the early Byzantine period. The glass fused in these tank furnaces from raw materials attached itself to the floor of the furnace and, once it had cooled down, would have been removed, perhaps with a pickaxe, to produce chunks of raw furnace glass.8 These chunks would then have been reheated in a crucible within a second furnace type, perhaps of a beehive shape.9 This would then have enabled the glass-workers to work the glass into a range of glass artefacts using metal implements such as gathering rods to make beads, and hollow tubes, known as blowing irons, for blowing glass into vessels. This second stage of glass production is known as secondary production. It could have occurred on the same site as where the glass was fused (primary glass production) or on other sites at a distance from where the glass was fused.10 Raw furnace glass manufactured at primary glass making centres was sometimes traded by boat. Excavations of shipwrecks have revealed the extent of the trade, such as the 3rd century BC Sanguinaire found off the coast of Corsica which had at least 550 kg of glass including raw glass11, the 2nd-3rd century AD Mljet wreck off the Croatian coast produced about 100 kg of raw glass12 and the 2nd-3rd century AD Ouest Embiez 1 (Var) which produced between 350 and 700 kg of raw glass chunks, each weighing up to 25 kg.13 Furthermore raw glass has been excavated from the Golfe de Fos near the mouth of the Rhône14 and two metric tons of raw glass came from the early 11th century wreck at Serçe Limani, Turkey.15 4 5 6 7 8 9 10 11 12 13 14 15 Kouwatli et al. 2008. Nenna 2015, 19. Phelps et al. 2016. Gorin-Rosen 2000; Tal, Jackson-Tal & Freestone 2004; Nenna 2015; Freestone et al. 2000; Phelps et al. 2016. Gorin-Rosen 2000. Henderson 2000, 38–42. Henderson 1989; Freestone et al. 2000; Phelps et al. 2016, Henderson et al. 2021. Alfonsi & Gandolfi 1997. Rossi 2009. Fontaine & Foy 2007. Foy & Nenna 2001. Bass 1984; Bass et al. 2009. 14 — 16 17 18 19 20 21 22 23 24 25 26 Henderson 1989; Freestone et al. 2000. Shortland 2004; Henderson 2013, 51–53. Henderson 2013, 51–52. Henderson 2013, 52. Nenna 2015. Brems et al 2013a; 2013b. Boschetti et al. 2016; Henderson, Sode & Sablerolles 2019; Crocco et al. 2021. Lahlil et al. 2010; Boschetti et al. 2016; Boschetti et al. 2020. Henderson 1991a; Barber, Freestone & Moulding 2009. Schibille & Freestone 2013; Boschetti et al. 2016; Henderson, Sode & Sablerolles 2019. Henderson, Sode & Sablerolles 2019; Crocco et al. 2021. The presence of geographically separated primary and secondary glass-making sites has led to a suggested decentralized model for classical glass production in western Asia.16 The discovery of raw furnace glass on sites where there is no evidence for primary glass production either suggests that it was being traded through the site or that it was worked on the site. The existence of crucibles with a layer of glass on the inside supports the latter suggestion and there are examples of this from early medieval contexts in the Netherlands (see Chapter 3 for information about the industrial evidence for glass production in the Netherlands). There is no archaeological evidence for the primary manufacture of translucent or transparent natron glass in the early medieval Netherlands. Natron glass was manufactured from a combination of sand and a mineral flux called natron or natrun.17 The sand that occurs on the coastal beaches of the Levant is ideal for glass production and is referred to as such by both Strabo and Pliny.18 The second primary raw material was natron. The main source of this evaporite mineral flux was in the Egyptian western desert at Wadi el Natrun,19 close to some primary production sites for Roman glass.20 This mineral is an evaporite which is formed seasonally and would have been shipped or traded to glass makers on the Levantine coast. A third crucial component of natron glass, which gives it durability, is lime. Lime was provided by the marine shells in the sand. It appears that the proportion in the Levantine coastal sand was just right for the production of durable natron glass. It is possible that the shell fraction was separated by glass makers and mixed with sand in the correct proportion prior to glass production, though no archaeological evidence for this has been found. The availability of sand with these characteristics would have been one reason why primary glass-making furnaces were located on the Levantine coast. Both archaeological and scientific evidence confirms that this is the case. Strontium and neodymium isotope and mineralogical analysis of multiple beach deposits around the Mediterranean has suggested which sands would have been suitable for glass making.21 In spite of the existence of important Byzantine glass-making sites in the Levant in the 6th–8th centuries a range of political, social and economic factors would not necessarily provide a guarantee that fresh natron glass would have found its way to early medieval glass-working sites in the Netherlands. While it is widely accepted that some form of natron glass (whether pristine or recycled) was used for the manufacture of early medieval objects in the Netherlands (to be discussed in much more detail below and in Chapter 5) there is one source of fully fused coloured glass that was also used: glass tesserae.22 Evidence for the reuse of these cubes of generally opaque glass has been discussed in many archaeological and scientific studies and this study is no exception. Their discussion is relevant in this section because they were coloured and opacified: such colourants and opacifiers were also sometimes used in Dutch early medieval glass. Roman glass tesserae are invariably made from natron glass. Most are opacified with small crystals, especially of calcium antimonate (Ca2Sb2O7 or Ca2Sb2O6). Without additional colourants this produces an opaque white colour.23 Opaque yellow tesserae are coloured by lead antimonate crystals (Pb2Sb2O7) with a smaller number of opaque yellow tesserae coloured with lead stannate crystals (Pb2Sn2O7). Dull red tesserae are opacified and coloured with copper droplets also found in Roman enamels.24 Opaque turquoise blue tesserae are coloured with copper and calcium antimonate, opaque yellow-green tesserae with copper and lead antimonate; opaque blue tesserae are coloured by a combination of cobalt and calcium antimonate. Therefore if elevated levels of antimony, lead, copper and sometimes tin are found in translucent early medieval glass a likely source is recycled Roman glass tesserae.25 Alternatively, such elevated levels of colourants can be explained by the use of fragments of highly coloured vessel glass. The most significant collection of glass tesserae in the Netherlands has been discovered at Wierum.26 Dutch early medieval glass was also coloured deliberately with low levels of transition metal ions: cobalt to produce a deep translucent blue colour, copper for a turquoise colour and manganese for a purple colour. Various shades of green, amber and pale blue could be produced by modifying the furnace atmosphere in which the glass was melted if the glass contained iron and manganese. Amber and 15 — pale blue colours are produced in an oxygendeficient furnace atmosphere, green in a more oxidising atmosphere. There is evidence for the production of one particular colour of glass in northwestern Europe before c. 800 AD and as early as the 6th century: opaque yellow. Opaque yellow vitreous materials have been found in crucibles from several early medieval sites in Ireland, Denmark and the Netherlands. The evidence for its production and its scientific analysis will be discussed in more detail in Chapter 5. From around 800 AD, glass technology in western Asia underwent a technological transition, especially during the Abbasid caliphate. The Abbasid glassmakers had a marked effect on glass technology in the western Asia and the Mediterranean: instead of natron they made glass using ashes of salt-tolerant shrubby plants. These plants grew in semidesert, evaporitic and maritime environments in western Asia and parts of the Mediterranean basin. Because the plants used for the flux could grow in inland locations, one result was that primary glass production became more widespread across western Asia and into central Asia. This led to a fully decentralized production system with multiple primary production centres, many located in cosmopolitan hubs on the silk road.27 In inland locations suitable plants were far more accessible as a source of flux than the far more limited sources of the evaporitic mineral used by the Romans to make natron glass. Like the Romans, the Abbasids fused glass raw materials (plant ashes and sand) in large tank furnaces28 and added colourants to the glass to produce deeply coloured glasses as part of the secondary phase of production. It is worth noting that, unlike natron, these plants had a highly variable composition depending on a range of environmental factors. One of these is the geological nature of the soil in which the plants grew. Although this might be viewed as potentially confusing, using scientific analysis has enabled plant ash glasses to be provenanced in increasingly more geographically defined ways (see below). Although the Muslims were partly responsible for this transition in glass production the other possible influence on this technological change was the pre-Islamic manufacture of plant ash glass by the Sasanians between the 3rd and 7th centuries29 in modern Iran and Iraq for which there is no published direct evidence for primary glass production from raw materials. It is nevertheless likely that the glass was made at sites like Veh Ardašīr and Ctesiphon30 and Brill31 has suggested -based on the presence of tank furnace fragments found on rural sites, some of probable Sasanian date - that this is evidence for primary glass production. Most of the glass found in early medieval northwestern Europe, including Dutch contexts, dating to after c. 800 AD therefore shows a dependence on the import of ready-made glass combined with a transition that occurred in western Asian glass technology, with the production and export of plant ash glass, especially in the Carolingian period. The exception to this dependence on imported glass made in the Mediterranean basin and western Asia (with associated recycling during in the Carolingian period) was the use of some of the earliest glass fused from raw materials in northwestern Europe, from tree ashes. Some of the earliest examples date to the 8th century, for example from the Loire valley in France.32 However, its production became widespread in the 11th century and later, especially in response to the massive demand for cathedral and church windows, such as in the Weald of Kent in southern England, and France,33 but the period between c. 800 and 1000 AD was one of transition too.34 One of the hallmarks of the technological transition in early medieval northwestern Europe is the occurrence of mixed-alkali glass, a likely combination of different proportions of wood ash glass and natron glass. It is more likely that fully fused glasses were mixed than that wood ash was added to natron glass. This would have formed part of a period of experimentation with the new alkali raw material – wood ash. Like the plant ash used to make glass in western Asia discussed above, wood ash has a highly variable chemical composition depending on the geological characteristics of the soil in which the tree grew, the tree species, the season in which the ash is burnt and the part of the tree.35 As with plant ash glass it is becoming increasing clear that in some cases scientific analysis can help to provide a geographical provenance (often regional) for such glasses.36 From around 800 AD glass linen smoothers make an appearance in early medieval Europe. Scientific analysis has revealed that these were made in Europe using glassy slags derived from lead-silver cupellation.37 27 Henderson et al. 2021; Henderson 2022. 28 Aldsworth et al. 2002; Henderson et al. 29 30 31 32 33 34 35 36 37 2005a; Henderson et al. 2021. Mirti et al. 2008; Mirti et al. 2009. Simpson 2014, 204. Brill 2005, 66. Aunay et al. 2020. Wedepohl 2008; Meek, Henderson & Evans 2012; Henderson 2013, 104–108. Henderson 2013, 97–108; Aunay et al. 2020. Jackson, Booth & Smedley 2005. Meek, Henderson & Evans 2012; Adlington et al. 2019. Gratuze et al. 2003. 16 — Summarizing, the main types of glass in use in Early Medieval North-Western Europe, and therefore potentially available in the Netherlands, were: • imported pristine raw furnace natron glass; • recycled and mixed natron glass (including tesserae); • imported pristine plant ash glass (made from shrubs); • wood ash glass made in northwestern Europe; • mixed-alkali glass (mixed wood ash and natron glass) 2.3 38 Willmott & Welham 2013; 2015. 39 Wedepohl, Winkelmann & Hartmann 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 1997; Gai 2005. Willmott & Welham 2015. de Sigoyer et al. 2005. Andersen & Sode 2010. Henderson & Ivens 1992. Van Wersch et al. 2014. Dell’Acqua 1997. Sanke, Wedepohl & Kronz 2002. Szöke, Wedepohl & Kronz 2004. Dodt, Kronz & Simon 2021. Dodt, Kronz & Simon 2021, Abb. 10. Koch 1987. Stepphun 1998, 94–96, Abb. 24. Kronz et al. 2016. Haevernick 1979. Dodt, Kronz & Simon 2021, Abb. 11. Evidence for the glass industry in early medieval Europe outside the Netherlands Evidence for the glass industry in early medieval Europe is quite sparse. Glass furnaces have been found at Glastonbury (UK), associated with the evidence for glass working,38 and Paderborn (Germany).39 The evidence from Paderborn and Glastonbury is only for glass working, not glass making (the fusion of raw materials) and Paderborn is probably later. The evidence at Glastonbury consists of a furnace, a crucible fragment with green glass adhering, lumps of glass, spills of glass, a moil (the glass encircling the tip of a blowpipe constituting production waste which does not get recycled), pulls of glass, a cast slab and a bichrome cable as well as vessel and window glass fragments. A possible glass furnace associated with glass fragments and bichrome cables has also been found at Barking in London.40 Glass furnaces have also been found at Huy (Belgium).41 Evidence for glass working in the form of crucibles containing glass, dribbles and drops of glass, melted glass rods and cables, half made beads, vessels and window glass has been found (either separately or together) at a variety of places. Evidence for beadmaking has been found at Ribe (Denmark)42 and at Dunmisk (Ireland),43 where glass studs were also made. Some other places have yielded evidence for glass blowing, including the abbeys of Stavelot (Belgium)44 and San Vincenzo al Volturno (Italy).45 Other evidence has been found at the Carolingian monasteries of Lorsch and Corvey (Germany)46 and Zalavar (Hungary).47 There is also clear evidence that glass was worked at Cologne during both the Merovingian and Carolingian empires.48 Evidence of glass production consists of fragments of glass furnace floors, which were presumably beehiveshaped furnaces,though no plans of the excavations are published so it is difficult to judge. Multiple crucible fragments with glass adhering, dribbles and drops of glass, reticella rods, scraps of glass, tesserae and vitrified bricks have been found. A distribution of Merovingian loop decorated bowls down the Rhine has been recorded suggesting that they were made in Cologne mainly from HIMT-2 glass (see below). 49 Koch has shown that there is also a distribution of early Merovingian cone beakers down the Rhine, further supporting Rhenish production, probably in Cologne. 50 At Hedeby (Germany) a possible glassworking area was found, including a possible furnace.51 Two crucible fragments containing wood ash glass, a single one with high lead glass combined with wood ash and soda-lime glass and raw glass have been reported.52 Evidence for glass production has also been found at Cordel (Germany) although an early medieval date has been called into question.53 Scientific analysis of the glass from Cologne using electron probe microanalysis alone shows that secondary glass making involved HIMT-2, which was originally probably fused in the early to mid 4th century AD, and was mainly used to make funnel and bell beakers – as well as funnel beakers from Hedeby (Germany).54 A plot of weight % Fe2O3/TiO2 versus Fe2O3/Al2O3 provides evidence of a single funnel beaker from Cologne made with Egypt-2 glass (originally made between c. 720 and 780 AD), and funnel beakers from Hedeby made from Egypt-1 glass (originally made between c. 760/780 and 870 AD). However, unless failed examples of funnel beakers and moils, both of the appropriate chemical composition, are discovered it is difficult to be absolutely certain that the vessels were blown in Cologne from weak HIMT (HIMT-2), although it remains likely. Being the commonest glass compositional type at the time, weak HIMT is clearly not diagnostic to a specific production centre. Therefore, Cologne has provided evidence for glass working but not definite proof for the manufacture of funnel beakers there. 17 — 2.4 Scientific analysis of early medieval glass in Europe 2.4.1 The principal glass types Syro-Palestinian and Egyptian glass The primary characterization of pristine Levantine and Egyptian natron glass was carried out by Nenna et al.,55 Foy et al.,56 Freestone et al.,57 Phelps et al.,58 Freestone et al.59 and Schibille et al.60 These studies focused on glass which derived from primary glass-making sites in the SyroPalestinian area and Egypt and shows that different proportions of minerals such as zircons, chromite and feldspars can characterize the sands used to make glasses at different production sites and at different times. Two compositional groups of natron glass produced in Israel in the mid-late first millennium AD have been widely recognized so far. Levantine I is defined according to the chemical compositions of sixth to seventh century glass from Dor and Apollonia. The sand used for making Levantine I glass was probably derived from the Bay of Haifa, close to the mouth of the river Belus of antiquity. As Phelps et al.61 have noted the use of the term Levantine I has masked other compositionally related but distinct glass, such as that made at 4th century Jalame. Levantine II was defined using the chemical composition of furnace glass produced at Bet Eli’ezer and is dated to the 8th century Late Byzantine – Umayyad period.62 Two compositional groups were recognized by Gratuze and Barrandon63 in their study of early Islamic glass coin weights from Egypt. Since then the two groups have been referred to as Egyptian I and Egyptian II.64 Egyptian I glass typically has high alumina (3–4.5 wt%) and low lime (3–4 wt%).65 It has been suggested that this glass was produced in factories near the famous natron source at Wadi el Natrun. Egyptian II glass has relatively high lime (c. 9%) and low alumina (typically 1.5–2.5%). Recent detailed analysis of Egyptian natron glasses has revealed the existence of Egypt 1A dating to before 725 AD, Egypt 1B dating to between 720 and 780 AD and Egypt 2 dating to between 760/780 and 870 AD.66 Because these glasses from different primary production sites have clear compositional characteristics, they should be identifiable amongst early medieval Dutch glass. It has been noted that the levels of sodium oxide decreased67 and aluminium oxide increased over time68 in these pristine glasses, due to a shortage of natron, and the use of different sand deposits, respectively. Other natron glass Tesserae When glass tesserae made out of natron glass were mixed with other natron glass to extend its volume, certain compositional characteristics in the tesserae are passed on to the bulk glass. As discussed above the occurrence of elevated levels of antimony, copper, lead and sometimes tin in translucent vessel glass suggests that a stock of glass tesserae has been mixed into the bulk glass.69 Elevated levels of antimony indicate this especially because calcium antimonate was used almost universally as the opacifier in a high proportion of Roman glass tesserae. HIMT and its variations High iron, manganese and titanium (HIMT) oxide levels that are found in 4th–5th century HIMT natron glass indicate a probable Egyptian source.70 High levels of these oxides as well as zirconium show that sands with high proportions of minerals bearing these elements were used to make the glass. A higher iron variation has also been identified.71 Much work has been carried out to investigate variations of HIMT glasses which have been found in the Mediterranean and in northern Europe. The most important variation of HIMT sensu stricto is the Foy 2 compositional group. Foy 2 was originally identified in glass from Carthage72 and includes 6th century series 2.1, which contains elevated V, Ti and Zr, and series 3.2 as originally described by Foy and colleagues.73 These types of glass are regarded as ‘weaker’ types of HIMT with lower concentrations of iron, manganese and titanium and variously labelled HLIMT (high lime, iron, manganese and titanium),74 weak HIMT75 and HIMT 2. Foy 2.2 was originally defined by Foy and colleagues as a recycled version of Foy 2.1.76 This type of glass may have been recycled multiple times and is found across the Mediterranean and Europe in contexts dating to as late as the 9th century.77 The ultimate origin of the original base glass used in these glasses – which would then have been recycled 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 Nenna, Vichy & Picon 1997. Foy et al. 2003. Freestone, Gorin-Rosen & Hughes 2000. Phelps et al. 2016. Freestone et al. 2018. Schibille et al. 2019. Phelps et al. 2016. Freestone, Gorin-Rosen & Hughes 2000. Gratuze & Barrandon 1990. Freestone, Gorin-Rosen & Hughes 2000. Wt% = percentage of each oxide by weight. Schibille et al. 2019. Henderson 2002. Phelps et al. 2016. Henderson 1991a; Schibille & Freestone 2013; Boschetti et al. 2016; Henderson, Sode & Sablerolles 2019; Crocco et al. 2021. Foy et al. 2003. Ceglia et al. 2015. Schibille, Sterrett-Krause & Freestone 2016. Foy et al. 2003. Ceglia et al. 2019. Conte et al. 2014. Foy et al. 2003. Bertini, Henderson & Chenery 2020. 18 — and mixed with other glasses – was probably Egypt, with elevated proportions of heavy minerals such as zirconium characterizing Egyptian glass. Foy 2.1 high iron found in Byzantine glass weights78 has also been recognized in AngloSaxon Britain, Serbia, Merovingian France and Spain.79 Furthermore, another compositional variation, HIT (high iron and titanium), has been recognized from 5th–6th century Bulgaria80 and possibly from 5th–6th century Albania.81 HIT is unlikely to have been made in Egypt or the Levant; where it was made precisely is unknown. A plant ash variant of the Foy 2 family has been recognized by Schibille et al.82 in Byzantine glass weights. 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 Schibille et al. 2016. Ares et al. 2019. Smith, Henderson & Faber 2016. Conte et al. 2014. Schibille et al. 2016. Henderson et al. 2021. Henderson et al. 2016; Siu et al. 2020; Schibille et al. 2020. Henderson 2002, Table 1, analyses 13, 22–25, 31, 41, 48 and 49. Henderson 1999, 288, Table 2, analysis 32. Sablerolles & Henderson 2012, analysis 129. Henderson 2023. Dekówna 1980, Table 44. Kronz et al. 2016. Langbroek 2021b. Jankowiak 2021. Aunay et al. 2020; Pactat et al. 2017. Kronz et al. 2016. Dekówna 1978, Tables 2 and 4; Dekówna 1980, Table 48, nos. 3, 4. Kronz et al. 2016. Wedepohl, Winkelmann & Hartmann 1997. Kind, Kronz & Wedepohl 2002, Table 2. Henderson 1991b, 129, Catalogue numbers 240 and 250, Appendix; Henderson 1993. Henderson 2012. Henderson 2012. Henderson 2023. Pactat et al. 2017. Kronz et al. 2016. Van Wersch et al. 2014. Plant ash glass From c. 9th century AD plant ash glasses manufactured in the Islamic domain in western Asia and the southern Mediterranean were made in cosmopolitan centres along the silk roads stretching from Spain to central Asia.83 Because local sources of plant ashes were used to make the glass their chemical and isotopic characteristics relate to the geological or geographical location in which they were made. A range of minor and trace elements, including their ratios, can be used to characterize the raw materials used to make the glasses, such as Ca, Mg, Ba, Zr, Ti, Cs/K, Li/K, Li/Na, B/Na, 1000Zr/Ti, La/Ti, Cr/La, Ce/Zr, Y/Zr, Mg/Ca.84 Examples of plant ash glasses that have been found amongst early medieval European glass are millefiori, blob decorated, melon and chopped beads from 10th century Viking age burials at Peel on the Isle of Man,85 a colourless silver foil glass bead from Wijnaldum,86 the blue cable applied to the rim of an 8th–9th century pale green funnel beaker from Dorestad,87 in the matrices of two 9th century trail-decorated funnels and in the body of a funnel beaker as well as beads found at Susteren.88 For a closer consideration of Susteren examples see below, and the use of trace element analysis in Chapter 5. Dekówna89 reports the analysis of eight artefacts, a glass drop, a rod, five segmented beads and an annular bead from Hedeby (Germany) that are plant ash glasses opacified with lead stannate. Kronz et al.90 note that some plant ash glasses from Hedeby were mixed with lead. Moreover, typical Islamic millefiori glass beads have been found at Dorestad.91 Like the distribution of Islamic coins (dirhams),92 it appears that Islamic plant ash glasses are mainly restricted to the northern Netherlands, Scandinavia and the Baltic, coinciding with the presence of Vikings, and far fewer examples have been found in France,93 with negligible numbers in Germany until the 9th century.94 Mixed-alkali glass Most glasses with mixed-alkali compositions that have been published mainly date to the 9th–10th centuries, reflecting the transition from natron to wood glass technology in northern Europe. There is no doubt that wood ash glass was made in northern Europe. Mixed-alkali glasses have been found at Hedeby (Germany)95 including a single funnel beaker,96 and at Paderborn (Germany)97 with two examples from the abbey at Fulda (Germany) with relative potassium oxide/soda levels of 7.7/6.6 and 6.0/9.1.98 Two pre-851 and an 11th–12th century AD (possibly redeposited) mixed-alkali green window glass have been found at Lurk Lane, Beverley (UK).99 Two examples of mixed-alkali glass have also been found at Dorestad: a late 8th century yellow-green palm funnel100 and a 9th century AD yellow-green funnel beaker.101 The analysis of a yellow-green tubular base of a classic funnel beaker from Susteren was also of a mixed-alkali composition.102 Comprehensive evidence for mixed-alkali glass working has been found at the c. 800– 866 AD Carolingian site of Méru, Oise (France). The archaeological evidence is in the form of furnace walls and crucibles containing glass. Both natron glass and glasses with potassium oxide levels of between c. 2% and 12% were found. This potentially suggests that Méru was one possible location where the mixture of wood ash and natron glass or, less likely, the addition of wood ash to natron glass, actually occurred103 although the case is not completely proven. It is also suggested that mixing of natron and wood ash glass occurred at Hedeby in Germany.104 Wood ash glass High potassium levels were introduced using wood ash as a flux. High potassium glasses have been published from the abbey of Stavelot (Belgium).105 These glasses are quite early, from 19 — contexts dating to between the second half of the 7th century to the early 9th century AD. The twelve examples are characterized by soda levels of below 3%, potassium oxide levels of between 10.14% and 20.21% and calcium oxide levels of between 13.84% and 20.23%. All contain high magnesium oxide and phosphorus pentoxide levels; the rest of the glass from the site is natron glass – there is no plant ash glass. In addition, full wood ash window glasses characterized by very high potassium and calcium oxide levels like those from Stavelot have been found in probable late 8th to early 9th century AD contexts at Baume-les-Messieurs, Jura (France).106 Wood ash glasses have also been found at the church of St Hermès-et-Alexandre (Belgium).107 In Germany wood ash glass has been found at ecclesiastical sites, at the abbeys of Lorsch, Corvey, Brunshausen–Gendersheim and Fulda.108 Excavations of the 10th century site of La Milesse, Sarthe (France) has produced large-scale evidence for working high potash glasses. The glasses that were scientifically analysed had a relatively restricted compositional range which led Pactat et al.109 to suggest that the glass was fused there. Kronz et al.110 have reported the presence of nine funnel beakers made with wood ash (out of a total of 61 wood ash glasses) and two crucibles containing wood ash glass from Hedeby (Germany). It was noted that wood ash glass contains higher calcium oxide levels in glass that dates to post 1200; the range detected was between c. 10% and 22% with the majority being between c. 10% and 16%.111 Some of the earliest wood ash glasses contain up to about 11% calcium oxide, such as from Hedeby. Six funnel beakers from Borg in Norway112 dating to between 800 and 1000 AD contain between 11.2 and 12.7% calcium oxide. High lead glass has been reported from a variety of northern European sites. The first high lead glass appears c. 10th century AD. Four examples have been reported from Hedeby with 22 other glasses being a mixture of plant ash glass and lead.113 Glass making in early medieval Europe therefore only involved wood ash to produce high potassium glasses. Before around 800 AD glass-workers relied on the import of scrap and raw glass which was either formed directly into objects or mixed with other glass. 2.4.2 Summary of existing scientific analyses of early medieval Dutch glass before the start of this project Scientific analysis of early medieval glasses from Maastricht, Susteren, Wijk bij Duurstede (Dorestad), Rijnsburg, Wijnaldum, Lent, Borgharen and Sittard had already been carried out prior to this study. What follows is a brief summary of the results from these investigations. Maastricht The early medieval glass and glass-bearing objects from Maastricht114 include a good number of glass beads (including failed beads), glass rods and crucibles with vitreous materials attached. Seventeen glass crucibles with colourless, pale green and opaque yellow residues attached were discovered at Jodenstraat and are discussed in much more detail in Chapter 5. Apparently sintered material was attached to one of the crucibles found at the Mabro site and it was suggested that this might possibly be frit (though see Chapter 5). The results of the analyses are given as means and standard deviations for each glass colour, for window glass samples and for two separate samples of crucible glass.115 Using these results, a translucent blue splinter of glass resulting from glass working could possibly be of a pristine Levantine glass (but see Chapter 5). The glass is coloured with 0.1% cobalt oxide in the presence of 0.2% cupric oxide. The three window fragments contain elevated MgO, K2O, TiO2 and MnO, with low antimony oxide concentrations, so are likely to be recycled/mixed glass varieties of weak HIMT. Yellow-green glass from inside a crucible is contaminated with 8.7% Al2O3; it contains low CaO (3%) as well as elevated TiO2 and Fe2O3. It is therefore difficult to estimate what its original composition was, especially if the same elements (e.g. Al and Fe) were present both in the original glass and in the crucible wall from which they migrated into the glass. Colourless glass attached to another crucible wall has very similar elemental contamination. The opaque yellow residues (n=4) contain the highest PbO levels (30.6±10.02) associated with tin oxide because they relate to the production of lead-tin oxide (Pb2Sn2O7/Pb2SnO4). 106 Van Wersch et al. 2015. 107 Van Wersch, Mathis & Hoffsummer 2009. 108 Sanke, Wedepohl & Kronz 2002; Kind, 109 110 111 112 113 114 115 Kronz & Wedepohl 2002; Wedepohl 2003. Pactat et al. 2017. Kronz et al. 2016. Kronz et al. 2016, Abb. 9. Henderson & Holand 1992. Kronz et al. 2016. Sablerolles, Henderson & Dijkman 1997. Sablerolles, Henderson & Dijkman 1997, Table 2. 20 — They contain slightly elevated MgO and relatively high manganese and iron oxides of above 1%. The opaque white residues (n=2) contain high tin oxide and lower lead oxide because they are opacified with tin oxide crystals. These opaque white residues also contain higher magnesia levels (1.3±0.2%) than the opaque yellow glasses, something that has been found in Roman opaque white enamels and tesserae.116 The four opaque red glasses analysed contain only slightly elevated levels of magnesia and potassium oxide (means of 1.15% and 0.83% respectively). High levels of ferrous oxide (3.8±0.63%) may indicate that iron-rich crystals contribute to the red colour (see Chapter 5 for a more detailed interpretation); 0.37% ZnO possibly indicates that brass filings were used as a source of copper colourant. Two opaque green glasses which contain elevated MgO (with a mean of 1.2%) are coloured with cupric oxide (mean of 4%). They are opacified with lead stannate. Zinc has also been detected, suggesting that brass was added as the copper source. A single opaque turquoise glass was probably opacified with lead-tin oxide. It is coloured with 2.3% CuO and contains elevated MgO at 1.2%. 116 Henderson 1991a. 117 Henderson 2023. 118 Henderson 2012. Susteren Six beads, eleven vessels, two crucibles and ten windows were analysed using electron probe microanalysis and Scanning Electron Microscopy.117 The glass beads and vessels are primarily of natron glass having similar compositional characteristics to those from Wijnaldum and Maastricht. All contain between 0.74% and 0.83% MgO. In beads 4 and 6 the K2O levels are above 1%. In most cases the TiO2 levels are between 0.1% and 0.2%; bead 5 contains 0.22% TiO2. These natron glasses contain between 2.49% and 2.76% Al2O3 with the exception of bead 6 which contains 3.05% Al2O3. Being beads the glasses are coloured in various ways: the ‘black’ (deep translucent brown) body of bead 1 is coloured with 5.09% Fe2O3; a combination of MnO and Fe2O3 has produced the green colour in the bodies of beads 2, 5 and 6. Most of the translucent bead bodies (and opaque decoration) contain low levels of Sb, Pb and Cu oxides – an indication of a level of recycling involving the addition of coloured Roman glass, including tesserae (which are invariably coloured and opacified with Ca2Sb2O7). The bodies of beads 3 and 4 contained high levels of magnesia, potassium oxide and phosphorus pentoxide and are therefore plant ash glasses. Bead 4 contains the lowest concentration of Al2O3 at 2.06%. The eleven glass vessels from Susteren that were analysed consist of nine soda-lime natron glasses, one mixed-alkali glass (no. 19) and one plant ash glass (no. 28). The natron glasses contain between 0.76% and 1.2% MgO, up to 1.24% K2O and some have elevated P2O5 (e.g. no. 23 with 0.37%). TiO2 concentrations range from 0.09% (no. 25) to 0.4 (no. 26) so are variations of HIMT. The Susteren vessel glasses are colourless, cobalt blue, blue-green, pale green and pale yellow. The pale yellow glass is probably coloured with ferrous oxide. All glasses contain elevated concentrations of CuO, Sb2O3 and PbO. No copper was detected in the colourless sample (no. 26). These oxides are indications of mixing and recycling. The mixed-alkali glass (no. 19), a pale green funnel beaker with a tubular base, contains 8.76% Na2O and 8.5% K2O. It also contains 3.69% MgO, 1.5% P2O5 and 9.73% CaO all of which are quite distinctive characteristics associated with the inclusion of an organic flux. The blue-green plant ash glass trechterbeker fragment (no. 28) contains the lowest Al2O3 (2.03%) of the vessel glasses showing that a purer silica source was used. It is also characterized by 4.2% MgO and 2.48% K2O. The opaque red, yellow and white decorative elements used on beads 2, 3, 4, 5 and 6 mainly follow the same pattern of colourant use as discussed for the Wijnaldum and Maastricht opaque glasses: elevated Fe (4.59%) and CuO2 (1.74%) in red glass, high PbO and SnO2 in opaque yellow (probably in the form of Pb2Sn2O7 crystals) as well as a combination of Pb and Sb which are probably responsible for opaque yellow (in the form of Pb2Sb2O7 crystals) in the decoration of bead 5. Wijk bij Duurstede (Dorestad) Forty two vessels, one glass chip, five tesserae, one rod and two linen smoothers were analysed using electron probe microanalysis.118 Of these, 39 of the vessels are of a soda-lime natron glass composition. The remaining three consist of one plant ash blue trail from the rim of a trechterbeker 21 — (no. 129b), and two potassium (wood ash) palm funnel (late 8th century) and funnel beaker (9th century) glasses (nos 103 and 136). These contain relatively low potassium oxide levels at 8.6% and 7.9% respectively, associated with 1.6% and 1.2% soda levels and 14.1% and 13.8% calcium oxide levels. These represent early examples of wood ash glasses. All the natron glass has very similar characteristics to those already discussed for Wijnaldum, Maastricht and Susteren. The glasses contain between 2.1% and 3.34% Al2O3 but mainly fall between 2.5% and 3.0%, CaO concentrations of between 5.85% and 8.59% and P2O5 at 0.37% and 1.8%. Some contain slightly elevated levels of MgO and K2O and some elevated levels of P2O5 and MnO. Unsurprisingly, such characteristics suggest that the glasses have been recycled or mixed and are weak HIMT. The 32 Dorestad blue-green and pale green vessel glasses are coloured by a combination of MnO and Fe2O3, some with Fe2O3 levels up to 2.7%. The greenish translucent glasses are also coloured with CuO levels up to 2%. There is a statistically coherent number of pale green/ nearly colourless as opposed to blue-green vessel fragments from Dorestad. There is some evidence that the pale green and especially the nearly colourless glasses contain lower Al2O3 than blue-green glasses, though there are exceptions. MnO levels in all shades of green are at similar levels, mainly at between 0.5% and 0.7%. The combination of MnO and Fe2O3 will impart colour to the pale green glass given a specific oxidizing/reducing atmosphere in the glass furnace. The elevated CuO levels in the blue-green glass therefore appear to provide the deeper green colour. The lower Al2O3 in the pale green and colourless glass suggests that the glass was made with a slightly different sand source from the blue-green glasses. The blue plant ash glass trail used to decorate a trechterbeker (no. 129a) is coloured with cobalt oxide (0.06%). The body of the nearly colourless beaker decorated with gold foil might be expected to be a purer glass but it has a very similar ‘intermediate’/Foy 2 composition to other green Dorestad glasses with no obvious use of a decolourizer or evidence that a ‘special’ glass was used. However, the furnace atmosphere must have been controlled carefully to produce the colourless glass. The red beaker base contains 2.73% Fe2O3 and 1.51% Cu2O associated with 0.66% ZnO and 1.4% PbO. The copper-rich colourant (probably in the form of cuprite droplets) used may therefore well have included scrap brass. Rijnsburg Thirteen Merovingian glass samples from the glass-working site of Rijnsburg were analysed using electron probe microanalysis. Six were beads or unfinished beads, six were rods and one was a sample of crucible glass.119 The only translucent glass is a turquoise rod of a natron composition. All the yellow glasses were opacified with lead-tin oxide and both white glasses opacified with SnO2 but unusually they contained low levels of MgO, whereas seven other opaque white early medieval glasses from Maastricht and Wijnaldum contain elevated MgO levels.120 It can be suggested that the Rijnsburg white glasses were made at a separate source from other white glasses. Although only two red glasses were analysed the same thing is true for them: neither of them contain elevated MgO and K2O, something that is found almost universally in other early medieval opaque red glasses. The chemical compositions of the two red and two orange glasses from Rijnsburg are quite different from each other. The red glasses contain much higher Fe2O3 than detected in the orange glasses (5.7% and 3.2% versus 0.8% and 0.5% respectively). On the other hand, the orange glasses contain 20% and 18% Cu2O as opposed to 5.7% and 3.2% Cu2O in the red glasses. The orange glasses are therefore likely to contain denser and larger Cu2O crystals than the opaque red glasses. The latter are liable to be in the form of micron sized copper droplets or cuprite. The opaque yellow material on the inside of the crucible fragment consists almost entirely of Pb(O) and Sn(O2): 79.8% and 11% respectively, with 4% SiO2 and may be evidence for production of lead-tin pigment on site. Wijnaldum The scientific analysis of the glass and glassbearing artefacts from Wijnaldum reported on 38 electron microprobe analyses of twelve vessels dating to between 450 and 900 AD, 24 beads dating between 550 and 900 AD, a vitreous blob attached to a crucible dated to 250–350 AD and glass attached to a crucible 119 Dijkstra, Sablerolles & Henderson 2011. 120 Dijkstra, Sablerolles & Henderson 2011, fig. 12. 22 — 121 122 123 124 125 126 Henderson 1999. Henderson 1985; Matin 2019. Steppuhn 1998, 100–101. Henderson 1991a; Silvestri et al. 2012. Corbella 2017. Van Os et al. 2014. dating to between 575 and 620 AD.121 The translucent glasses are mainly glass vessels; three are glass beads. The beads are mainly deliberately coloured with the oxides of transition metals, manganese, iron and copper. Opacification is due to the formation of tin oxide, SnO2 (white), lead-tin oxide Pb2Sn2O7 (yellow) and cuprous oxide Cu2O (red).122 High iron associated in the opaque red glasses would have acted as an internal reducing agent. Apart from one, the translucent glass samples are natron glass containing elevated MnO and Fe2O3. The exception is a single drawn segmented silver foil bead which has a full plant ash composition with much higher levels of MgO (5.5%) and K2O (2.2%) which are comparable with Islamic glasses found at Hedeby.123 Elevated MgO levels of c. 1% are present in two vessel glasses; the earliest glass vessel dating to c. 450 AD contains 0.3% TiO2 which is also unusually high – its significance will be discussed below. Three colourless glasses were analysed, one of which is the silver foil bead. One glass is probably decolourized with 0.4% antimony trioxide. Some translucent glasses also have trace levels of TiO2, CuO, Sb2O3, SnO2 and PbO, all indicators of glass recycling or mixing, including the use of Roman glass tesserae. The dull opaque red glass bead contains high potassium and magnesium oxides, a characteristic of Roman enamels and tesserae124 so this is evidence that this kind of glass continued to be used in the early medieval period. An opaque red globular bead contains unusually high CaO (11.7%) and relatively low Al2O3 (1.9%). The presence of tin suggests that scrap bronze was used as a copper-rich colourant in the bead; tin is absent from the other opaque red beads. All red glass beads contain lead ranging from 1.2% to 14.2%. An orange bead is opacified with copper in the presence of iron, the orange rather than red colour was possibly attributable to differences in the sizes of copper crystals in the glasses. Lead stannate in crystalline form is the opacifier in opaque yellow-green glass beads (numbers 18 and 23) which are otherwise coloured with copper and iron respectively. The opaque yellow glasses contain between 21.5% and 54.5% PbO and are opacified with Pb2Sn2O7/Pb2Sn2O6 crystals. The opaque yellow material on the inside of the flat tray or less likely a furnace fragment (number 38) contains 63.5% PbO. It may have been a flat plate on which the opaque yellow residue was heated probably at c. 650°C. It is clear from the composition that the glass-like material is contaminated by interaction with the ceramic substrate. A single vitreous blob from a crucible that was analysed (number 37) contains high Al2O3 (9.7%), 63.5% PbO and 2% total alkali – it is probably a fuel ash slag. Both an opaque yellow tessera and the yellow spiral trail decorating the rim of a beaker from Wijnaldum are opacified with lead antimonate. Though opaque glasses contain elements associated with their colour and opacity they also can contain elevated MgO, K2O, CoO, CuO. The presence of these oxides could suggest that the base soda glass used to make the opaque glasses was recycled, though more detailed analysis of opaque yellow glasses in Chapter 5 provides interesting new information. Lent Corbella125 used p-XRF to analyse 30 glass beads that were found in four graves in the Merovingian cemetery of Lent. Most glasses are probably of a natron composition with additives to modify the glass colours. It was not possible to analyse sodium. In this case the data should be treated as indicative because it’s possible that different colour compositions were combined in one analysis. The red glasses analysed contained elevated potassium, iron and copper as found in other such glasses. The detection of MgO was unreliable. Two of the glasses analysed apparently contained alumina levels above 4% with relatively low CaO so potentially might have originated in south Asia. The white glass analysed contained low tin and elevated antimony so may be opacified with calcium antimonate. Two green glasses may be coloured with a combination of iron and manganese oxides. Borgharen The results of semi-quantitative analyses of the glasses from Borgharen produced using p-XRF are listed according to their colour.126 The opaque white glasses mainly contain tin with low levels of antimony and are therefore likely to be opacified with SnO2 crystals. However, some very high levels of P2O5 have been listed. If these are so high, which does not seem likely, then bone ash needs to be considered as a possible white opacifier. Opaque blue glasses are also apparently opacified with tin oxide. The glasses 23 — contain relatively low levels of lead oxide. According to these results, the same opacifier was used in opaque brown glasses. The colourless glass may have been clarified with manganese oxide. The translucent green glass seems to contain tin and lead so may be semiopaque. The opaque orange glass contains c. 6% PbO and unusually high CaO. A compositionally quite consistent series of eight opaque red glasses indicate that they contain high tin and iron. It is not completely clear from these analyses that copper is responsible for the colour and/or the opacification. However, a comparison with analyses using ICP techniques and SEM EDX show that these analyses are fully quantitative. 127 Sittard A plot of relative levels of Pb and Sn in all analysed glass from Borgharen and Sittard using p-XRF shows very similar patterns for both sets of beads.128 These data for Sittard, in particular, have a strong positive correlation. The beads from Sittard contain Sn levels of between 3% and 12.5%. Such high levels must be due to a high density of Pb2Sn2O7 crystals in the glasses – analysis of the matrix glass between the crystals using ED-XRF would likely produce a lower level of Sn in the matrix. The Sittard beads contain PbO levels of between 12% and 35%. It is notable that the Borgharen data are more scattered in terms of relative Pb and Sn levels, especially at levels below 12% PbO and c. 3% SnO2. The overall positive correlation between the Pb and Sn suggests that these colouring/ opacifying elements were added together to the glass melt. There is a contrast in the levels of CaO between the yellow glasses from Sittard and Borgharen,129 with Sittard containing significantly lower levels. A majority of glasses from Sittard contain c. 2% CaO whereas Borgharen glasses have a peak of CaO levels at 4% so this suggests that different recipes may have been used to make the glasses at the two sites. Discussion ‘Roman’ natron glass was still in circulation c. 800 AD, and later130 so by examining scientifically European early medieval glass assemblages of vessels and beads it becomes possible to assess the extent to which glasses have been imported from the Middle East (whether natron or plant ash glass). Thus the scientific investigation of this technological transition also provides a way of demonstrating how local glass production developed in Europe, especially with the emergence of glass made using wood ash. Characteristic Merovingian and Carolingian vessel types were blown and bead types made in Europe from imported raw furnace glass or glass cullet in a secondary production process. For the high lead glasses, rather than removing the PbO and recalculating the totals to 100% the use of ratios in Figs. 2.1 and 2.2 is considered to be an alternative and equally acceptable means of presenting the data so as to investigate its provenance. All natron glasses (the majority of the analyses considered) including the base glass for opaque glasses were possibly fused in Egypt or the Levant. The natron source would have been Wadi el Natrun in Egypt. Therefore, what follows is a consideration of the provenance of the glasses and by inference also the provenance of the raw materials since local or easily accessible raw materials would have been used. The Wijnaldum (Wij) glasses have a wide compositional range according to major (Na2O, CaO and SiO2) and minor components (Al2O3). There are three outliers (Wij 16, 17 and 32) which obscure some of the finer detail of the remaining data in Fig. 2.1. They plot as outliers because Wij 16 contains a very high CaO level (11.7%), Wij 17 is probably not a glass and Wij 32 is a plant ash glass (5.5% MgO and 2.2% K2O) rather than a natron glass. Wij 37 has a very unusual composition. It is possibly a plant ash glass, but contains 9.7% Al2O3 and 3.7% Na2O. Therefore in Fig. 2.1 the data for Wij 16, 17, 32 and 37 have not been included. This allows for a more sensible comparison between the results for glass from other early medieval Dutch sites as well as other selected contemporary sites. This plot is an approximate way of defining compositional types and potential for raw material provenance used by other researchers.131 The plot suggests that Wij 1, 2, 3, 9, 13, 18, 20, 23, 25, 30, and 31 are potentially Levantine glasses (though see Chapter 5) and that Wij 24 is of the HIMT type and therefore potentially made in Egypt. The main bulk of early medieval Dutch natron glass fall into the Foy 2 compositional types (and its variants).132 These glasses have undergone 127 128 129 130 131 132 Personal comment Hans Huisman. Huisman et al. 2019. Van Os et al. 2014. Phelps et al. 2016. Bertini, Henderson & Chenery 2020. Foy et al. 2003; Ceglia et al. 2015; 2019; Schibille, Sterrett-Krause & Freestone 2016; Bertini, Henderson & Chenery 2020. 24 — Fig 2.1 A plot of Na2O/SiO2 versus CaO/Al2O3 in early medieval glass beads and vessels compared to glasses from Jarrow and Monkwearmouth, UK (Brill 2006; Freestone & Hughes 2006). 133 Freestone, Gorin-Rosen & Hughes 2000. 134 Phelps et al. 2016. 135 Gallo et al. 2015. multiple recycling/mixing episodes. Even if they were originally pristine Egyptian glasses before they were mixed their provenance cannot now be identified exactly. The glasses typically contain elevated Fe2O3, TiO2, MnO and also some elevated MgO, P2O5 and K2O. The plotted points for the mean results of both translucent and opaque glasses from Maastricht all fall within the area that coincides with the intermediate/‘Foy 2’ glasses. This is to be expected given some of their other compositional characteristics discussed above, including quite high MgO, MnO, and Fe2O3 and therefore come into consideration for HIMT or related glasses. None of the glasses contain high TiO2 levels. A single translucent blue chip of glass appears to have a pristine Levantine composition: with a Na2O/SiO2 ratio of 0.14 it falls well below other glasses considered here – due to its 74% SiO2 – and can therefore be tentatively suggested as a product of Bet Eli’ezer characterized by such high SiO2 levels133 although its Al2O3 level of 2.4% is c. 0.5% lower than would be expected. Fig. 2.2 should be a more precise way of defining different compositional types of natron glasses. There are five outliers: Wij 11, 13, 14, 15 and 33. The reason why these are outliers is that they contain high levels of iron: blue-green no. 11 (2.2%), opaque red no. 13 (4.6%), opaque red no. 14 (4.1%), opaque red no. 15 (4.4%) and dark olive green no. 33 (3.2%). Without plotting the outliers in Fig. 2.2 a better classification of Wijnaldum glasses than seen in Fig. 2.1 becomes evident. By far the largest proportion of samples fall into the same ‘intermediate’/‘Foy 2’ composition as seen in Fig. 2.1. None of the Wijnaldum glasses fall below the Fe/Ti value of 0.5, a value that may be technologically significant for the few Susteren and Dorestad intermediate/‘Foy 2’ glasses that do. The low values of Fe/Al of <0.2 for four Wijnaldum glasses (numbers 1, 9, 30 and 31) could suggest that they are unrecycled Levantine glasses, but see Chapter 5. Wijnaldum 1 could potentially be of a Levantine–Apollonia type134 and Wijnaldum 30 perhaps of the Levantine–Jalame type.135 25 — Fig 2.2 A plot of Fe2O3/Al2O3 versus Fe2O3/TiO2 in Dutch early medieval glass beads and vessels compared with results from Jarrow and Monkwearmouth UK (Brill 2006; Freestone & Hughes 2006). These glasses are characterized by low MgO and K2O at c. 0.5%–0.7% weight, Al2O3 c. 3%, and low iron and titanium. Overall, they lack the elevated MgO, K2O, TiO2, Fe2O3 and MnO (and sometimes P2O5) that characterize ‘Foy’ glasses, found in the bulk of the remaining data. However a more detailed interpretation is given in Chapter 5. Three Wijnaldum glasses, 5, 7 and 12, appear to plot with HIMT (high iron, manganese and titanium) glasses in Fig. 2.2. Careful evaluation of the results indicates that glasses 5 and 7 are unlikely to be HIMT, glass 12 is more typical of the composition (i.e. it contains 0.3% TiO2). The difficulties in allocating a glass type to these natron glasses are overcome significantly by analysing them with the more sensitive technique, laser ablation inductively coupled plasma-mass spectrometry (LAICP-MS): see Chapter 4. Maastricht mean results are plotted in Fig. 2.2 (triangles). All of the glasses, irrespective of whether they are translucent or opaque, fall into the area of the plot occupied by ‘transitional’/‘Foy 2’ glasses. This suggests provisionally that all the glasses analysed from the site are recycled/mixed and therefore conform to the bulk of other analyses of early medieval glasses from the Netherlands using the relative values of Na2O/SiO2 and CaO/Al2O3. None fall into the area of the plot occupied by pristine Levantine natron glasses or HIMT glasses. See Chapter 5 for a more detailed consideration. Although it was suggested above that the translucent blue chip of glass from Maastricht might be a pristine Levantine glass, when plotted in Fig. 2.2 this appears not to be the case. As noted above most have elevated MgO, MnO and Fe2O3 concentrations, above the values found in pristine Levantine glasses (and pristine Egyptian HIMT glasses); trace element analysis (Chapter 5) reveals a more definitive means of classifying the glass. Early medieval opaque and translucent glasses have not been considered together before in this way. This initial evidence therefore suggests that recycled/mixed early medieval glass was used as a base glass for 26 — 136 137 138 139 140 Bertini, Henderson & Chenery 2020. Freestone, Gorin-Rosen & Hughes 2000. Kato, Nakai & Shindo 2009. Ceglia et al. 2015. Schibille, Sterrett-Krause & Freestone 2016. 141 Bertini, Henderson & Chenery 2020. 142 Henderson et al. 2016. 143 Henderson, Ma & Evans 2020. adding opacifying compounds – and this is also apparently true for the single translucent blue glass sample from Maastricht. Due to the high level of contamination of both translucent glasses in the crucible fragments from Maastricht the results have not been included in Fig. 2.2. The Susteren glasses plot with 8th and 9th century Dorestad glass having Na/Si oxide ratios of above 0.24. Nevertheless, Levantine glasses can also plot within the area occupied by glasses with ‘Foy 2’ and ‘intermediate’ compositions136 so as noted above a better discrimination is needed; an attempt at this is given in Fig. 2.2. It appears that none of the Susteren glasses are pristine (unrecycled) Levantine I, Levantine II137 or N1 Levantine glasses as defined by Kato et al.138 which are manufactured at primary production sites. The compositions for glass beads 1 and 6 as well as vessels 21, 23, 24 and 27 from Susteren are possible contenders for Levantine glasses but all contain higher levels of MgO, P2O5, K2O, TiO2, MnO and Fe2O3 than pristine natron glasses. This is also true of all other natron glasses from the site. These elevated levels of impurities show that almost all of the Susteren natron glasses have been recycled, perhaps several times. A nearly colourless funnel beaker from Susteren with a dark blue incalmo rim (no. 26) is the only example of an HIMT composition. It has the highest Fe2O3 (1.34%), TiO2 (0.4%) and MnO (1.79%) levels of all the Susteren samples. It also contains a low calcium oxide level of 6.2%, another characteristic of HIMT glass. All other Susteren natron glass is of the ‘intermediate’/‘Foy 2’ type with higher calcium oxide and elevated levels of MgO, P2O5, K2O, TiO2, MnO and Fe2O3 compared to pristine Levantine I and other natron glasses. The Susteren glass plots close to Jalame (Levantine I) on a major/minor component plot of Na/Si oxides versus Ca/Al oxides with relatively high Na/Si oxide values of above 0.24. However it is clear that when sand impurities are plotted in Fig. 2.2 apart from the HIMT sample they fall into the same plotted area as ‘Foy 2’ compositions as defined by Ceglia et al.139 and Schibille et al.140 and ‘intermediate’ glasses from Comacchio as defined by Bertini and colleagues.141 In Fig. 2.1 Dorestad ‘8th’ century (pre-750) glasses are plotted separately from Dorestad ‘9th’ century (750–850 AD) glasses. Almost all ‘8th’ century glasses have CaO/Al2O3 values of between 2.5 and 3.09 or close to this range, whereas ‘9th’ century (750–850 AD) glasses have a far wider range of such values, of between 1.86 and 3.5. This suggests that raw materials with a wider compositional range were used and/or that mixing/recycling introduced a wider compositional range in the glasses that date to before 750. On this basis all Dorestad glasses can be classified as ‘intermediate’/‘Foy 2’. Figure 2.2 (from which Dorestad outliers have been removed) on the other hand provides further interesting insights into the Dorestad glasses. Almost all natron glasses are of the ‘intermediate’/‘Foy 2’ composition according to this plot. However, four Dorestad glasses have low Fe/Ti oxide values of between 0.48 and 4.63 and this distinguishes them from the vast majority of early medieval European ‘intermediate’/‘Foy 2’ glass and extends the values of such glass. All ‘8th’ century Dorestad glasses have Fe/Ti ratios of above 5. This increasingly wide range of Fe/Ti oxide values in early medieval European glass deserves to be revisited in more detail using trace element analysis. The single Dorestad glass that plots in the ‘HIMT’ area is in fact a plant ash glass with low Al2O3 (1.9%). This however raises the issue of whether the silica sources used to make plant ash glasses would benefit from such an approach. Other trace element ratios have been used to investigate Islamic plant ash glasses,142 as have radiogenic isotopes.143 Rijnsburg glass tends to have lower Na/Si oxide values than most translucent early medieval glass (Fig. 2.1) which is intriguing because it suggests that the base glass could be pristine (unrecycled) Levantine glass, though this would still need to be confirmed with trace element analysis. If this is the case it would show that relatively pure glass was imported and used as the base glass for making the opaque Rijnsburg material. Three glasses with Na/Si values above 2.2 are (based on Fig. 2.1) probably of an ‘intermediate’/Foy composition. As can be seen in Fig. 2.2 six Rijnsburg glasses plot in the area of Fe/Al versus Fe/Ti oxide values that is consistent with being pristine Levantine glasses too, so this substantiates the above suggestion. As noted above the use of ratios allows a direct comparison with translucent glasses even though the opaque glasses contain high levels of PbO 27 — and Sn. This is an unexpected and intriguing result which indicates how important further research using trace element and isotopic analysis would be. The two Rijnsburg glasses that have been badly contaminated when being worked, as well as the material attached to the Rijnsburg crucible, and glasses with very high Fe levels have not been included because, for different reasons, they cannot be compared with the other plotted glasses with any validity. It is impossible to plot Lent data on Fig. 2.1 because no soda was quoted. An attempt to plot the data for Lent in Fig. 2.2 was made and, where data was available, all plotted well away from the other data. This suggests that the quality of the data is too low to be considered further here. A small number of early Islamic plant ash glasses were used to make early medieval beads and vessels. The glasses have elevated MgO and K2O concentrations and are quite distinct from natron glasses as discussed above. A plot of MgO versus CaO in relation to Levantine, northern Syrian and Iraqi/Iranian early Islamic glasses (not given here)144 suggests that glasses from Wijnaldum and Susteren ultimately derive from Iraq or Iran, that a second glass from Susteren probably derives from Syria and a single glass from Dorestad probably derives from the Levant. Their provenances using trace element analysis are discussed in more detail in Chapter 5. As already discussed, the bulk of Dutch early medieval glasses are of the ‘intermediate’/‘Foy 2’ composition. Given the range of impurities in these glasses they have undergone potentially multiple episodes of recycling and mixing. Most glasses of these highly variable compositions date to, at the earliest, the 8th century and, using major and minor chemical analysis, there is no clear development or change in their chemical compositions over time. However, a number of observations can be made when the compositions of glass from different early medieval sites are compared. Both Rijnsburg and Wijnaldum glasses appear to include more examples of imported pristine glass than the later sites of Susteren and Dorestad. At Rijnsburg such glass would have been used as the base glass for making beads. However, currently there is no actual evidence for mixing colourants with pristine glass there. Based on these results no pristine glass was detected at Susteren or Dorestad and only a single probable example has been detected amongst the Maastricht glass. This suggests that there may have been a more direct supply of pristine glass c. 600 AD to some sites where the glasses were coloured/opacified (and possibly made into beads or vessels on the same sites) than for the Foy 2 glasses where a number of recycling events would probably indicate that more intermediate sites were involved. Moreover, the variation in composition of natron glasses as defined by major oxides (Na2O, SiO2, CaO) as well as Al2O3 appears to decrease with time when glasses dating to before c. 750 AD (‘8th century’ in Fig. 2.1) are compared with those that date to after c. 750 AD (‘9th century’ in Fig. 2.1). All of these observations would benefit from further scientific analyses using more sensitive techniques of chemical analysis and also isotopic analysis (see Chapter 5). Matthes145 has suggested that there was an increase in the levels of As and Sb in opaque yellow glass over time from c. 600 AD. The implication could be that the lead sources changed over time. However significant levels of As have not been detected in Dutch early medieval opaque yellow glass using electron microprobe analysis and the increase in Sb levels might relate to increasing levels over time in the base glass used (see Chapter 5). Early Islamic plant ash glass was introduced in the early 9th century so we can be sure that these few early medieval examples date to this time, or later. Further analyses of early medieval Dutch glasses will certainly reveal more examples and it may eventually be possible to suggest the key centres that such glasses and glass artefacts were imported from. Based on this provisional review we can suggest that glass used to make early medieval vessel, bead and window glass found in northwestern Europe was mainly imported and reworked there. Primary production for natron glass occurred in the Levant and Egypt where coastal sand and natron from Wadi el Natrun were fused in large tank furnaces. Glass furnaces have been found on a number of early medieval northwestern European sites but all of these show evidence of secondary production (glass working). It is conceivable that glass was reheated and blown into characteristic early medieval vessel types such as bowls and beakers or made 144 Henderson et al. 2016. 145 Matthes 1998, fig. 10. 28 — 146 Foy 2003; Ceglia et al. 2015; 2019; Schibille, Sterrett-Krause & Freestone 2016; Bertini, Henderson & Chenery 2020. 147 Henderson 2012; Henderson 2023. into beads and windows; the weight of evidence suggests vessels were made in Germany. The exceptions are the production of an opaque yellow material (originally suggested to be glass), especially at Maastricht and, later, when the first wood ash glass was made from about 800 AD. It is possible that mixed-alkali glass was made at Méru by adding wood ash to natron glass but a more likely technique was to mix the two types of glasses. Although most early medieval northwestern European glass is natron glass only small numbers of pristine (unrecycled) Levantine and Egyptian glass occur. Most glass that has been found dating to between the 4th and 10th centuries is HIMT (characterized by high levels of iron, manganese, titanium and zirconium) and its recycled variants. Based solely on the major and minor elemental analysis of early medieval Dutch natron glass, most appear to fall into the recycled Foy 2 compositional types.146 These Dutch early medieval glasses typically contain elevated Fe2O3, TiO2, MnO and also some elevated MgO, P2O5 and K2O.147 A much more detailed consideration of the probable origins of early medieval Dutch glass using trace element analysis is given in Chapter 5 of this book. Concluding remark Glass from the earlier periods, between 450–550 and 550–650 AD contain somewhat more examples of unrecycled (‘pristine’) glasses and a smaller number of recycled glasses. By the time of the Carolingian empire from about 750 AD almost all the glass is recycled as exemplified by glass found at Dorestad. Exceptions are Egyptian II glass from Hedeby and Cologne. By 800 AD primary glass production of wood ash glass occurs in northwestern Europe and, as noted above, at a time of technological transition, the production of mixed-alkali glasses occurred. Plant ash glass was imported from western Asia in the form of beads and raw glass; it was used in the manufacture of early medieval European beads and occasionally glass vessels. 29 — 3 Evidence for early medieval glassworking in the Netherlands Introduction ISIA N FR TER S EA R T OO W E N O E ST R ers STER GO Lauw S L FRISIA TRA N CE H Wierum GO Wijnaldum BloemendaalGroot Olmen Almere Deventer ValkenburgDe Woerd Leidsche Rijn Utrecht Wijk bij Duurstede (Dorestad) AA SL AN D Den Haag - Frankenslag Rijnsburg Oegstgeest IJss el RIJ NL AN D WE STE RN FR ISIA K E NNEMERLAND TE XE ie L Vl M Rh ine Ma as ine Rh Susteren t eld Sch Maas Glass production can be divided into two different stages. The first is primary production involving the fusion of raw materials (see Chapter 2). The second is remelting the glass made in the first stage to make glass objects. There is no recorded evidence for primary glass production in the early medieval Netherlands. There is, however, evidence for secondary production, not just the remelting of glass but also the manufacture of pigments for glass coloration. In the Merovingian period, comprehensive evidence for glass bead making has been found at Maastricht-Jodenstraat, Rijnsburg-Abdijterrein and Wijnaldum (fig. 3.1). The evidence falls between the last quarter of the 6th and first half of the 7th century. As Pion has shown, this period coincides with a steep reduction of imported oriental beads in the west, giving birth to a new production of highly coloured Merovingian bead types.148 Maastricht, developed from an old Roman town was one of several production centres in the middle Meuse valley149, well within the borders of the Merovingian empire. On the other hand, the production of popular Merovingian bead types in the central places150 of Rijnsburg in the Rhine delta and the northern terp site of Wijnaldum shows that the demand for fashionable beads reached well beyond the borders of the Merovingian empire (fig. 3.1). There is scantier evidence for glass production in the Carolingian period in the form of waste, perhaps in part because the custom of wearing colourful beads became less popular in the Frankish heartlands151, though they were still being worn in the northern periphery of the Carolingian empire. Imports of Islamic period beads from the end of the 8th century onwards152 in settlements along the Rhine and in cemeteries north of the Rhine, traded through Viking trade networks to the west, may also have contributed to a reduction in Frankish glass bead production. Another notable difference with the preceding period is that that the evidence for glass working derives from a more varied range of site types and now includes an early emporium (Dorestad) and ecclesiastical centres (Utrecht-Domplein, Susteren) (fig. 3.1). From the Carolingian period, not all phases of the production processes are E A 3.1 Maastricht Frisian inhabited area Holocene coastal area 0 50km Pleistocene hinterland Fig. 3.1 Sites in the Netherlands with evidence for glass working in the early medieval period (adapted from Dijkstra 2011, 12). represented by the glass working waste, so it is more difficult to attribute what evidence there is to specific products. Glass beads were probably made at Dorestad, while possible evidence for the production of the earliest highly coloured stained window glass was found at the early medieval monastery of Susteren. In the following the sites are considered starting with southern sites and moving northwards. 148 149 150 151 152 Pion 2014. Van Wersch 2012. Nicolay 2017. Delvaux 2017. Sablerolles & Van der Linde-Louvenberg 2019 (Leiderdorp-Plantage); Langbroek 2021b (Dorestad); Van Es & Schoen 2007/2008 (Zweeloo cemetery). 30 — 3.2 Maastricht, Limburg Province Excavations carried out in Maastricht in the 1980s and early 1990s yielded evidence of an impressive range of craft activities carried out during the 6th and 7th centuries. The location of Maastricht is shown in fig. 3.1. In total, nine different sites were identified in quite different parts of the settlement.153 Apart from one, these were all located on the west bank of the Meuse. Here a Roman castellum had been constructed in the 4th century which developed into an important political, religious and economic centre in the Middle Meuse region during the Merovingian period. Maastricht had its own bishop, see and mint. Gold coins minted in Maastricht have been found all over northwestern Europe, with a significant concentration in the northern Netherlands, especially in Friesland Province.154 Evidence of glass-working was found at six locations, the Mabro site (Onze Lieve Vrouweplein), the Derlon site, the Jodenstraat site, the Rijksarchief site, the Boschstraat area and the Lage Kanaaldijk site.155 On all of these sites glassworking was combined with one or two other high-temperature crafts (table 3.1), namely pottery production (Lage Kanaaldijk), ironworking (Boschstraat, Rijksarchief), copper-alloy working (Derlon, Boschstraat, Jodenstraat) and possible gold-working (Mabro). Antler-working took place at the Mabro site, Derlon site, Boschstraat area and Rijksarchief site, while waste from amber-working was mixed in with glass bead production waste from the Jodenstraat site. It is likely that glassworkers worked closely with other craftsmen and shared knowledge about furnace technology, fuel and other materials. For instance, the glass crucibles used by the glassworkers on the Jodenstraat, Mabro and Rijksarchief sites were all repurposed ovoid coarse-ware cooking pots (Wölbwandtöpfe) which were most probably made in Maastricht. At the Céramique site on the east bank of the river Meuse, four cross-draught kilns were found filled with wasters, dominated by coarse-ware Wölbwandtöpfe (fig. 3.2).156 Remarkably, these pots are also used as glass crucibles in Merovingian Rijnsburg (see Section 3.8) and Valkenburg-De Woerd (see Section 3.9) and there is continuity in the reuse of Wölbwandtöpfe as glass crucibles into the Carolingian period (globular pots type Dorestad W III) (see discussion by Menno Dijkstra, Section 3.5.1) 1:4 0 20cm Fig. 3.2 Maastricht-Céramique site: A restored cooking pot (Wölbwandtopf ) from the fill of a furnace (Photograph: Wim Dijkman). Table 3.1 Maastricht: Early medieval sites with evidence for glass-working in combination with other crafts. Maastricht sites High-temperature crafts glass Derlon X pottery Other crafts iron copper-alloy gold? X Mabro X Jodenstraat X 153 Dijkman 2013. 154 Pol 1999, fig. 7. 155 Sablerolles, Henderson & Dijkman 1997, Boschstraat area X X Rijksarchief X X 295, fig. 1; Dijkman 2013, fig. 1. 156 Dijkman 2013, fig. 4. Lage Kanaaldijk X antler X X X X X amber X X X X 31 — 3.2.1 Maastricht-Jodenstraat (MAJO) Site: 1 Site type: specialist craft zone Province: Limburg Municipality: Maastricht Place: Maastricht Toponym: Jodenstraat 30 Start date 580/590 (late 6th century) End date: 610/620 (early 7th century) Description: In 1988 excavations by the Gemeentelijk Oudheidkundig Bodemonderzoek Maastricht (GOBM), the Municipal Archaeological Service of Maastricht, took place at Jodenstraat 30.157 The site was situated north of the late Roman castellum, near the Via Belgica which connected Tongeren and Maastricht. A rubbish pit filled with the debris from glass bead making was found.158 The pit also contained waste from copper-alloy-working and amberworking.159 Based on the pottery finds, the pit was filled sometime in the late 6th to early 7th century. To date, the glass assemblage represents the most comprehensive evidence for 6th–7th century glass bead making in Europe. Glass production waste: The debris consisted of 750 glass objects which represent the full range of waste from glass bead production. The production waste was divided into eight main groups: glass rods (n=369), ‘punty’ glass from a beadmaker’s tool (n=36), glass threads with and without tweezer marks (n=17), glass drops (n=39), finished and failed beads (n=123), crucibles (n=38, EMN=17), cullet or scrap glass (n=20), glassy slags/fuel ash slags (n=53) and non-diagnostic fragments (n=55) which include (small lumps of) melted glass and fragments that are too small to classify. All waste categories are dominated by opaque yellow glass (apart from scrap glass and glassy slags). Opaque yellow, green, red, white, red and blue make up 57%, 17%, 16%, 7% and 0.9% of colours respectively. The balance are translucent light green (2%) and blue-green (0.6%), and transparent colourless (0.2%) of glass. Almost all beads are wound and have tapering perforations showing they were made by winding melted glass around a mandrel, a bead-making tool with a conical point. Such a tool may have been found at the Rijksarchief site (see Section 3.2.3). A few beads were made by perforating a section of a ‘composite’ glass rod which was made by fusing strands or slender 157 The site remains unpublished, apart Fig. 3.3 Maastricht-Jodenstraat site: From left to right: opaque yellow melted rods ends and threads with tweezer marks, finished pentagonal beads, drawn and composite rods, failed, cracked beads (Photograph: Gemeentelijk Oudheidkundig Bodemonderzoek Maastricht (GOBM)). from the rubbish pit with production waste. 158 Sablerolles, Henderson & Dijkman 1997. 159 For waste from amber-working, see Dijkman 2013. 32 — rods together in order to make a glass rod with a large enough diameter for making a bead. The vast majority of the beads (79%) are cylindrical in shape, predominantly with a pentagonal section (76%), while some have a square (2%) or round (1%) section. The polygonal beads were believed to have been shaped with tongs but are much more likely to have been shaped by hand with a small wooden tool known as a paddle. The remainder of the beads are made up of small globular beads of opaque yellow glass and medium-sized bi-conical beads. The majority of the beads are split in half along the length of the perforation (72%) (fig. 3.3). This is a commonly observed phenomenon on bead-making sites (see Section 3.8 and Section 3.12) and occurs when the glass is overheated during the making of the beads, or when the beads have not been annealed or cooled down properly, causing the beads to crack.160 Crucibles are represented by 38 fragments from at least 17 coarse-ware cooking pots (Wölbwandtöpfe). In 15 cases, only the lower halves of the pots were used to melt what was originally believed to be highly coloured opaque yellow glass (see Section 5.2.3). More recently, a crucible base with an opaque white deposit was identified by one of the authors of the 1997 160 Gam 1990; Heaser 2018. 161 Dijkman 2013. 162 Cf. Sode 2004; Risom 2013, 56–57, 60. publication (fig. 3.4).161 Table 4.2 lists further selected crucibles and associated glassworking debris from Jodenstraat (MAJO 1-27); images are in Appendix IV (figures appendix IV.11-38). No crucibles with opaque red, green and blue glass have been found. Once made, opaque glass would have been worked into beads, for instance by gathering some melted glass onto a solid metal rod or punty and winding it around a mandrel, a beadmaking tool with a conical point. However, the presence of opaque yellow and white glass rods suggests that these were an essential phase in the bead-making process. The rods were made by attaching two metal rods (punties) to a gather of melted glass, then pulling the glass in opposite directions creating a drawn glass rod of several metres’ length with a more or less round cross-section The long glass rods could have been fragmented into shorter sections, as suggested by the presence of a particular type of glass rod breaking splinter. Short rod sections could have been pre-heated, picked up with a punty, as indicated by the presence of punty glass among the waste products, and wound around a mandrel.162 Alternatively, a production technique in use by modern beadmakers may have been used. Fig. 3.4 Maastricht-Jodenstraat site: Crucible base with opaque white glass, drawn and twisted opaque white rods and cracked, failed beads (Photograph: Gemeentelijk Oudheidkundig Bodemonderzoek Maastricht (GOBM)). 33 — Table 3.2 Maastricht, Jodenstraat (MAJO): selected artefacts and photo numbers. Find number Fragment Glass colour(s) Sample 1-1-7-4 crucible base opaque yellow MAJO 1 1-1-5-17 crucible body deep translucent and yellow MAJO 2 21 figure appendix IV.12 1-1-7-11 crucible base opaque yellow MAJO 3 22 figure appendix IV.13 1-1-7-3 crucible base opaque yellow MAJO 4 23 figure appendix IV.14 1-1-7-2 crucible base opaque yellow MAJO 5 (inside) & MAJO 5 (outside) 24 figure appendix IV.15 (inside) & figure appendix IV.16 (outside) 1-1-7-8a crucible base opaque yellow MAJO 6 25 figure appendix IV.17 1-1-7-8b crucible base opaque yellow MAJO 7 26 figure appendix IV.18 1-1-7-8c crucible base dark translucent MAJO 8 29 figure appendix IV.19 1-2-3 possible brick fragment white and opaque yellow MAJO 9 30 figure appendix IV.20 1-1-7-451 a and b fragments blue MAJO 10 39-40 figure appendix IV.21 1-1-7-21-22a scrap red MAJO 11 41 figure appendix IV.22 1-1-7-21-22b scrap red MAJO 12 42 figure appendix IV.23 1-17-28 scrap green MAJO 13 43 figure appendix IV.24 1-2-3-VG3-2 window yellow-green MAJO 14 44 figure appendix IV.25 1-2-3-VG3-1 window yellow-green MAJO 15 45 figure appendix IV.26 1-2-3-VG3-3 window pale yellow-green MAJO 16 46 figure appendix IV.27 1-1-7-500 thin rod green MAJO 17 47 figure appendix IV.28 1-1-7-431 drop yellow weathered MAJO 18 50 figure appendix IV.29 1-1-5-388 drop red MAJO 19 51 figure appendix IV.30 1-1-7-463 drop dark green MAJO 20 52 figure appendix IV.31 1-1-7-462 pulled rod milky blue MAJO 21 53 figure appendix IV.32 1-1-7-583-594 thin rod red MAJO 22 54 figure appendix IV.33 296 twisted rod opaque white MAJO 23 58 figure appendix IV.34 1-2-5a beaker base green MAJO 24 60 figure appendix IV.35 1-2-5-b punty glass blue MAJO 25 61 figure appendix IV.36 1-1-7-503 ribbed flat blue-green MAJO 26 68 figure appendix IV.37 309-349 rod fragments green MAJO 27 73-74 figure appendix IV.38 This involves heating up the end of a longer section of a glass rod, attaching the hot glass to a mandrel and winding it directly around a mandrel. According to experimental glass beadmaker Sue Heaser, finds of glass rods of about 5–10 mm thick on many early medieval bead-making sites suggest that that beadmakers in the first millennium AD knew the technique.163 She states that this method gives more control than using a gather of molten glass attached to a punty. She goes on to explain that only the end of the glass rod is heated to liquid point so that the rest of the rod remains cool and rigid. The cool end is used as a handle by the beadmaker to control the hot glass at the other end as it is applied to Sample number 20 the mandrel. This could be another explanation for the presence of ‘composite’ glass rods at the Jodenstraat site made up of lots of strands or slender rods which are fused or twisted together (fig. 3.3 and fig. 3.4). Modern beadmakers use these leftover bits to fuse onto the end of glass rods when they have become too short to handle.164 Several examples of rods with melted ends (fig. 3.3) among the waste material from the pit could have been destined for this kind of recycling by melting the ends and pressing two of the same colours together.165 Chemical analyses published in 1997 suggested that the lead-tin-yellow crucible residues has the same composition as the opaque yellow glass rods and other waste Photo number figure appendix IV.11 163 Heaser 2018. 164 Personal comment Ingrid Pears, Hot Glass Studio, Thoresby, Notts., UK. 165 Cf. Heaser 2018. 34 — products, including (failed) beads. Moreover, the chemical compositions of opaque white, red and blue-green glass rods was also linked to waste products of corresponding colours. However, only trace analysis can give incontrovertible evidence that this is indeed the case (see Section 4.4). No crucibles fragments bearing opaque red, blue-green or blue glass were found and it was believed that these must have been imported in the form of glass rods that were directly worked into beads. In that case, the base glass used for the presumedly imported glass colours is likely to be different from the base glass used for the locally made opaque yellow and white glasses. A more detailed scientific investigation looking into this matter is given in Chapter 5. Fragments of the rim, (upper) bodies and base of two crucibles show that complete pots were used to melt colourless glass and translucent blue-green glass. Drops of translucent greenish glass amongst the waste products suggest this glass was worked on or near the site (fig. 3.5). Fragments of translucent greenish scrap glass – old vessel glass and the earliest early medieval grozed window glass from the Netherlands – were interpreted as cullet destined to be remelted to create this type of base glass. Further detailed investigations of the base and 166 Callmer & Henderson 1991, fig. 2. 167 Andersen & Sode 2010, 32, fig. 11, 34, table 11. 168 Preiß 2010, nr 38, fig. 10; Sablerolles & Henderson 2012, afb. 6.19, NO 5041. scrap glass using trace element analysis are given in Chapter 5. Not included in the 1997 publication are a couple of fragments of Roman faience melon beads and a handful of very small fragments of crushed translucent dark blue glass (fig. 3.5) as well as an almond-shaped bead of translucent dark blue glass. Although it is impossible to establish from what kind of object(s) the translucent blue crushed fragments come from, it is tempting to suggest they could derive from early planoconvex glass ‘cakes’ of translucent dark blue glass which are found on later 8th–9th century bead-making sites in southern Scandinavia, such as Åhus (Sweden)166 and Ribe (Denmark)167 and at Dorestad168 (see Section 3.4.2). With this in mind, it will be interesting to see whether the translucent dark blue crushed fragments can be chemically linked to a translucent dark blue bead found among the waste. Moreover, could this translucent dark blue glass have been used as a base glass to make opaque blue glass on site given there are three beads and a drop of opaque blue glass as well as an opaque greyish blue glass rod. This will be further discussed in Chapter 5. Fig. 3.5 Maastricht-Jodenstraat site: From left to right: fragments of Roman faience melon beads, crushed translucent dark blue glass and yellowish green window and vessel glass of translucent greenish glass with glass drops in matching colours (Photograph: Gemeentelijk Oudheidkundig Bodemonderzoek Maastricht (GOBM)). 35 — 3.2.2 Maastricht-Mabro Site: 2 Site type: dump zone Province: Limburg Municipality: Maastricht Place: Maastricht Toponym: Mabro Start date: 500 (6th century) End date: 700 (7th century) Description: Excavations carried out in 1981 by the GOBM or Municipal Archaeological Service of Maastricht at the site of the Maastrichtse Broodfabriek (Mabro) or the Maastricht Bread Factory, located at the Onze Lieve Vrouweplein (MAVP) 16–18 remain unpublished. Glass production waste: The site produced twelve fragments of glass crucibles, a substantial number of beads, glassy slags, one of which might be a fragment of melted furnace wall, and at least one red-burnt fragment of clay covered with translucent greenish glass.169 Eleven crucible fragments can be dated to the 6th or 7th centuries (Wölbwandtöpfe), of which ten were made available for sampling (table 3.3). Additionally, one rim fragment (find no. 1-5-OA) is of a late 4th–early 5th century bowl (type Alzey), perhaps recovered from a late Roman grave and reused in the Merovingian period. Chemical analysis of the latter was included in the 1997 Maastricht-Jodenstraat publication.170 The fragment is covered with translucent greenish glass below the rim and a vitrified, off-white granular material on the inside and outside of the rim. At the time, this was tentatively interpreted as overheated frit. Frit is a half-product of glass-making, so this could constitute the earliest evidence of glass-making in the early medieval west. This material has now been re-analysed using scanning electron microscopy and the results are presented in Section 5.2. Two rim fragments of Merovingian crucibles show similar deposits (table 3.3). As is the case on the Jodenstraat site, the crucibles either contain colourless or translucent natural green glass or opaque yellow glass; no crucible fragments with other colours were found. Table 3.3 lists ten crucible fragments together with selected images (MABRO 1-10) from Maasricht-Mabro and the sample numbers used in scientific analysis. The images are at Appendix IV (figures appendix IV.1-10). Judging from photographs of the beads, they are all made by winding and are either monochrome or decorated with trails in contrasting colours.171 Using the bead typology developed by Pion in 2014 for beads from six Merovingian cemeteries in Belgium, later adapted by Vrielynck, Mathis and Pion,172 the Mabro beads cover a long period between 480–530 (P1) and 620–670 (P5), but they mostly date in period 560–610 (P3) and period 610–640 (P4) (table 3.4). This site has not been published, so it is impossible to state which beads are likely to be local products, but given the dates for the crucibles, those beads dating to roughly the Table 3.3 Maastricht-Mabro: Selected crucibles and their glassy contents, together with photo numbers and sample numbers. Find number Fragment Glass colour(s) Sample Photo number 03-04-2000 rim white: frit-like MABRO 1 figure appendix IV.1 3-OA-55 body colourless/white: frit-like MABRO 2 figure appendix IV.2 01-03-1951 body? translucent green MABRO 3 figure appendix IV.3 3-OA-1 rim colourless MABRO 4 figure appendix IV.4 1-5-OA rim colourless/white: frit? MABRO 5 figure appendix IV.5 3-OA-40 (= 3-AA-40) base opaque yellow MABRO 6 figure appendix IV.6 02-02-2018 base (burnt clay?) colourless/pale green MABRO 7 figure appendix IV.7 03-05-2024 base opaque yellow/green MABRO 8 figure appendix IV.8 03-05-2024 body? deep translucent MABRO 9 figure appendix IV.9 03-04-2012 base green MABRO 10 figure appendix IV.10 169 Information and photographs of the finds were kindly provided by Wim Dijkman, Senior Conservator Archeologie en Erfgoed, Team Programma en Innovatie, Centre Céramique – Kumulus – Natuurhistorisch Museum. 170 Sablerolles, Henderson & Dijkman 1997, 307–308, pl. 25, 1. 171 With many thanks to Wim Dijkman for providing the photographs. 172 Pion 2014; Vrielynck, Mathis & Pion 2018. 36 — Table 3.4 Maastricht-Mabro: Typology and bead periods (Pion 2014, Vrielynck, Mathis & Pion 2018). Find number Form Colour Decoration Typology Period 01-02-1963 globular opaque white translucent light blue crossing trails B3.3-3a P4 02-01-2000 cylindrical, round section? opaque yellow - B1.4-1a? P3? 03-04-2025 short-cylindrical opaque red - B1.4-2a P3 03-05-1932 bi-globular opaque yellow - B1.2-1b P4 03-05-1933 globular, medium opaque white - B1.1-4a P1-5 03-06-2004 globular, small opaque yellow - B1.1-2a P2-5 3-OA-0 disc opaque white translucent dark blue crossing trails B3.3-2a P3 3-OA-40.1 cube opaque red opaque yellow borders & crosses B11.1-5 P2 3-OA-52 cylinder, square section opaque red opaque yellow dots B6-2-1b P5 Period: P1 = 480-530 AD; P2 = 530-560 AD; P3=560-610 AD; P4=610-640 AD and P5=620-670 AD. 2nd half of the 6th and first half of the 7th century are the most likely candidates: a small globular bead, a bi-globular bead and a possible cylindrical bead of opaque yellow glass, a medium-sized globular bead of opaque white glass and a short cylindrical bead of opaque red glass. There are two polychrome, trailed beads: an opaque white globular bead with translucent light blue narrow crossing trails and an opaque white disc-shaped bead with translucent dark blue crossing trails. 3.2.3 Maastricht-Rijksarchief Site: 3 Site type: mixed crafts zone Province: Limburg Municipality: Maastricht Place: Maastricht Toponym: Rijksarchief Start date: 480–490 (late 5th century) End date: 600? Description: Excavations carried out in 1990–1991 by the Municipal Archaeological Service of Maastricht (GOBM) are briefly discussed by Hulst.173 The site was situated in the middle of the delta of the river Jeker, south of the 4th century Roman castellum and not far from the old Roman road which connected Tongeren, the Roman capital of the civitas Tungrorum, to the new centre in the region, Maastricht. Dozens of rubbish pits were found in this area containing (late) 5th and 6th century pottery, Roman tiles, chunks of local sandstone (‘kolenzandsteen’) and flint, waste from glass-working, antlerworking and iron-working. No traces of buildings were found, apart from three scattered postholes and remains of a small ditch. Glass production waste: Hulst174 lists 55 fragments of early Merovingian vessel glass, two glass rods, (fragments of) 25 beads, drops of glass, melted glass and small, dark vitreous spheres (‘glasbolletjes’) which may also be linked to ironworking on the site. Furthermore, one fragment of a glass crucible, glassy slags and vitrified fragments of (a) furnace floor with glassy slags and iron slags adhering to them were found. One pit contained traces of firing and may have been the firing pit of a dismantled furnace. A truly remarkable find is that of a forged iron rod which is round in section at one end and square in section at the other (fig. 3.6). Hulst remarks that the perforations of the beads match the circumferences of the iron tool. The rod is interpreted by Hulst as a bead-making tool or ‘mandrel’, a forged iron rod with a conical point on which beads are formed by winding a glass thread around it. 1:1 0 173 Hulst 1992. 174 Hulst 1992. 5cm Fig. 3.6 Maastricht-Rijksarchief site: Forged iron mandrel, square-sectioned, with a round-sectioned conical point (Photograph: Wim Dijkman). 37 — The opposite, square-sectioned end of the rod was probably originally inserted into a wooden handle.175 A similar tool from a bead-making site at Paviken on the Baltic island of Gotland was deemed too slender for an awl and has been interpreted as a mandrel (fig. 3.7).176 The squaresectioned end is hollow and could have been wrapped around a wooden handle. A mandrel with the remains of a handle was found at RibeDommerhaven, where beads were made on a very large scale in the 8th–9th centuries.177 According to Hulst it is obvious that beads were made on or near this site sometime in the 6th century. He mentions that the beads are mostly made of monochrome glass, including blue, brownish-red and yellow. Some beads are decorated with glass trails of blue, white or yellow glass. The beads include finished specimens and failed beads. Among the failed beads are examples where the winding of the spiral glass trail around the mandrel had gone wrong. Cracked specimens have been found. This can occur after the glass has been overheated or when the finished beads have not been annealed properly. Both rods are monochrome opaque brownish-red. Recent photographs of some of the finds178 show that the scrap glass includes a thick-walled fragment of early/mid-Roman blue-green glass (probably window glass or a square bottle), a yellow-green knocked-off rim of a late Roman cup of Isings type 96a,179 a rim fragment of a Merovingian bell beaker of yellowish-green glass with dark inclusions and lots of bubbles, decorated with vertical optic blown ribs and an opaque white trail below the straight, firerounded rim. The latter is probably contemporary with the bead-making. Among the beads, there is a sub-biconical opaque red bead with an opaque yellow spiral. This is the only specimen which can be securely dated to Pion’s Period 3 (560–610 AD), while a seemingly dark/dirty bi-globular bead with whitish deposits is probably of Pion’s Period 4 (610–640 AD).180 A spiral bead of translucent light greenish glass is very similar to one found at Leidsche-Rijn Leeuwesteyn Noord (Rijnfront) (see Section 3.6.2). 1:1 0 5cm Fig. 3.7 Paviken, Gotland, Sweden: possible mandrel with a bead added for museum display (Photograph: Matthew Delvaux). 3.3 Susteren-Salvatorplein, Limburg Province Site: 4 Site type: Monastery Province: Limburg Municipality: Echt-Susteren Place: Susteren Toponym: Salvatorplein Start date: 714 End date: 1802 Description: In the 1990s, excavations by the former Archaeological State Service (ROB) took place immediately north of the basilica of St Amelberga, the Romanesque church which is still standing today. These revealed remains of a monastery which was inhabited from the 8th century until it was dissolved in 1802 during the French occupation. The monastery was almost completely demolished in the early 19th century. The results have recently been published by Henk Stoepker.181 The foundation of the monastery and a small church is recorded in a charter from 714 when Pepin II and his wife Plectrude donated a small estate on the river Suestra to the AngloSaxon missionary Willibrord. The stream was part of a larger drainage system in the Limburg Meuse Valley. The excavations yielded only scant evidence for late Merovingian habitation, consisting of a timbered building with several associated waste pits and an oven, a cistern and a few graves. Very few portable finds were recovered from this period. The period of the late 8th century and 9th century sees an increase in habitation and there were now two stone buildings, one of which was circular and may have been a funeral chapel, a timbered building, as well as the above-mentioned cistern; to the east of the buildings was a craft zone with a bell-casting pit 175 Heaser 2018. 176 Lundström 1981, 99–100, fig. 10:4. With 177 178 179 180 181 many thanks to Matthew Delvaux, Princeton University, for providing the reference and translating the Swedish text into English. Sode 2004, 86, fig. 3. Information and photographs are kindly provided by Wim Dijkman, Senior Conservator Archeologie en Erfgoed, Team Programma en Innovatie, Centre Céramique – Kumulus – Natuurhistorisch Museum. Isings 1957, 113–114. Bead type B5.2-2a, Period 3; Bead type B1.2, Period 4 (Vrielynck, Mathis & Pion 2018). Stoepker 2021. 38 — 182 Sablerolles 2023 (basispublicatie chapter 29). 183 Pottery identification by Jan de Koning. 184 Stoepker 2021, 229, afb. 11.11, V12-053185 186 187 188 GL-10. Pottery identification by Jan de Koning. Henderson 2023 (basispublicatie chapter 31). Stoepker 2021, 231, table 11.1, V09-129GL-01, V04-194-GL-01. Stoepker 2021., 230, afb. 11.12. Freestone 2015,. and five ovens. This development is mirrored by an increase in the number of graves; these are presumably associated with the early medieval abbey church which is likely to be found beneath the present-day Romanesque church. Finds of highly coloured quarries are typical for ecclesiastical contexts and are a testament that stained glass windows were in use, most likely in the abbey church. A watercourse (complex 4300/4400) north of the habitation zone was used as a refuse dump from which many Carolingian period finds were retrieved. As much as 85% of Carolingian pottery was imported from the German Rhineland, showing the monastery was firmly embedded in the Rhenish trade system, despite the location of the monastery on the Meuse. Wine glasses such as (palm-)funnels probably came from the same Rhenish production centres as the ceramics. No indications were found for animal husbandry, although it can be assumed that nearby farms on (a) monastic estate(s) would have provided the monastery with animal products. The over-representation of certain skeletal elements of pigs shows these were specifically imported for the consumption of meat, one of the few signs of luxury enjoyed in the monastery. During the first half of the 10th century, the habitation zone was cleared and a large amount of settlement refuse, including remains of buildings, was dumped in watercourse 4300, perhaps as a result of a Viking raid in 881–882, although no evidence was found to support this. For the period of c. 900 to c. 1050 only one wooden building, a well and graves are discernible, and possibly some ovens. There are significantly fewer portable finds than in the previous century. The building of a new, Romanesque church in the second half of the 11th century ushers in a new phase of the monastery, characterized by the construction of stone-built cloisters. The digging of watercourse 4200 in the 11th century, intersecting the early medieval watercourse 4300/4400 caused a lot of early medieval material to be redeposited in later contexts. Glass production waste: Five fragments of glass production waste make it likely that glass was worked in the monastery, probably during the early medieval period.182 The finds consist of two fragments of glass crucibles, a partially melted Roman tessera, a glass fragment from glassblowing tool and a possible fragment of opaque yellow raw glass. The crucible fragments derive from the same context in watercourse 4310 (800–1300 AD) which mostly contains redeposited Carolingian material (60%) and some 10th century (17%), 11th–12th century (22%) and Iron Age/Roman period (1%) material. One crucible fragment is probably made of Carolingian Badorf ware183 and presumably derives from the same type of cooking pot that was used at Utrecht-Domplein (Dorestad type W III) (see Section 3.5.1) and probably also at Leidsche Rijn (see Section 3.6.2). Unfortunately, the translucent light (bluish-)green natron glass is too contaminated to be linked to either the window or the vessel glass found at Susteren. The other crucible fragment of possible grey Meuse Valley ware is covered on the inside with a thin layer of translucent dark blue natron glass with a small area of colourless glass (fig. 3.8a).184 The chemical composition of the dark blue glass in the crucible can be chemically linked to the dark blue glass of two translucent dark blue window quarries:185 an irregularly shaped quarry from a grave dating to the late 10th/11th century and a small triangular quarry from a context in watercourse 4400 with predominantly high medieval material with some early medieval finds (8%).186 High concentrations of antimony in all three glasses show that antimony-rich Roman tesserae were mixed in to colour the glass. The find of a partially melted, opaque dark blue Roman tessera (fig. 3.8b) could suggest that dark blue glass for the production of window glass was made in the monastery by adding blue tesserae to a colourless base glass.187 The tessera was found together with the above-mentioned triangular dark blue quarry and can be either high or early medieval. The number of artefacts involved is small but the presence of the crucible clearly shows that dark blue glass was being worked. The practice of recycling Roman tesserae, especially blue tesserae for colouring window glass, was carried on into the 12th century.188 A thick-walled fragment of translucent dark bluish-green glass is covered on the concave inside with dark grey iron scale from a glassblowing tool (fig. 3.8c). The fragment comes from the intersection of high medieval watercourse 4200 (11th–13th century) and early 39 — such as a spiral trail. In the Carolingian period, self-coloured spiral trails were especially popular on the necks of globular jars. One bluish-green vessel fragment from Susteren may derive from such jars. A fragment of opaque yellow glass from a 17th century cesspit, which also contains some early medieval material, may be a raw glass fragment struck from an early medieval glass ingot. In northern France yellow glass was worked in several monasteries during the 8th and early 9th centuries to produce reticella wares, especially bowls decorated with yellow spirals and reticella rods.191 A body fragment of such a bowl was also found at Susteren.192 a b c 3.4 0 1:1 Wijk bij Duurstede (Dorestad), Utrecht Province 5cm Fig. 3.8 Susteren-Abdijterrein: Glass-working production waste: a) fragment of a crucible with translucent dark blue glass on the inside (V12-053-GL-10); b) a partially melted, opaque dark blue tessera (V04-194-GL-02); c) a glass fragment from a gathering or bit iron (V12-078GL-01) (Photographs: Limburgs Museum, Venlo. Drawing: SAGA Archeologie). medieval watercourse 4300; the find context contains 25% early medieval pottery. It is possible that both crucible fragments and the glass fragment from the gathering iron were deposited closely together in watercourse 4310 and that the latter was redeposited when watercourse 4200 was dug.189 The glass is well preserved and is likely to be natron glass rather than wood ash glass which makes an early medieval date more likely. Whether early or high medieval in date, this fragment represents the only direct evidence for glassblowing in the Netherlands since the Roman period. The fragment is interpreted by the archaeologist and experimental glassblower Mark Taylor190 as glass that was broken off a gathering iron or bit iron. A bit iron is a long, thin iron rod with a flat end which is used by glassblowers to add handles, feet or decorations to a glass vessel. Since no early or high medieval vessels with handles or added feet are known from the Netherlands, the fragment therefore most likely results from decorating a translucent bluish-green vessel with a self-coloured decoration An overview of glass-working waste from Dorestad was published by Preiß in 2010.193 His inventory counts 84 finds which are made up of a few old finds without contexts as well as in situ finds from excavations in the late 1960s and ’70s in the harbour area (Hoogstraat excavations) and the settlement on the river bank (vicus). However, the majority of the finds (60%) were retrieved during more recent excavations in the agrarian settlement, the Parkeerplaats Albert Heijn (PPAH) excavations in 1992–1993 and the Veilingterrein excavations in 2007–2008, probably due to wet-sieving. The largest category of glass waste is formed by tesserae (43%) which were recycled on a large scale in the early medieval period. They were melted down to make beads, especially in Scandinavia, or to increase the volume of glass batches intended for the production of window and vessel glass.194 Preiß’s group of deformed glass (29%) may also include accidentally melted vessel fragments. The remainder is made up of glass drops and threads (15%), glass lumps (7%) and miscellaneous (6%). A find worth mentioning in the last category is that of a crucible which was found before 1978; it comes from the vicus on the river bank and is now lost.195 Isings described the fragment as follows: ‘Fragment of a crucible, pinkish grey ceramic. Covered by a layer of greyish green glass on the outside and a thick layer of green to bluish green glass on the inner surface.’196 189 Stoepker 2021, 229. 190 Many thanks to Mark Taylor of ‘Heart of 191 192 193 194 195 196 England Glass’ (https:// heartofenglandglass.co.uk) and ‘The Glassmakers’ (http://www. theglassmakers.co.uk). Louis 2015, fig. 4b, c, d; Cabart, Pactat & Gratuze 2017; Henton 2020. Stoepker 2021, 201, table 10.03, V08-190GL-07. Preiß 2010. Henderson, Sode & Sablerolles 2019; Schibille & Freestone 2013, 2–3, fig. 1C, D. Isings 1978; Preiß 2010. Isings 2015, 444, No. 16035. 40 — Perhaps the glass on the outside seems greyish green because the outside of the crucible had discoloured to a dark grey colour as the result of reheating, as can be seen on the crucible fragments from Utrecht-Domplein (see Section 3.5.1). The glass production waste from the PPAH excavations and the Veilingterrein excavations is discussed in more detail below. 3.4.1 Wijk bij Duurstede – Parkeerplaats Albert Heijn (PPAH) 197 Van Dockum 1997. 198 Nyst 2003, 13. 199 Nyst 2003, 32, catalogue 4.2; Preiß 2010 passim. 200 Van Es & Verwers 1980. 201 De Koning 2012, 186. Site: 5 Site type: Settlement Province: Utrecht Municipality: Wijk bij Duurstede/Dorestad Place: Wijk bij Duurstede Toponym: Parkeerplaats Albert Heijn (PPAH) (Steenstraat/Zandweg) Start date: 750–775 End date: c. 1250 Description: In 1992 excavations by the former Dutch National Service for Archaeological Heritage (ROB) were carried out at the intersection of Steenstraat and Zandweg before the construction of a car park for a planned new supermarket (Albert Heijn), hence the toponym Parkeerplaats Albert Heijn (PPAH).197 Fig. 3.9 Dorestad-Parkeerplaats Albert Heijn (PPAH) site: Glass working crucible with melted opaque white glass, probably from a Roman tessera. Defects in the transparent base may indicate where crushed tesserae were attached but have fallen out. Rim diameter 17.5 cm (Preiß 2010, 125, fig. 107). Not to scale. Both Carolingian and later settlement traces were found. In the Carolingian period this area belonged to the settlement west of the vicus on the river bank. This settlement has a more agrarian nature as evidenced by farm buildings discovered later during the Veilingterrein excavations (2007–2008) further north along Zandweg (see Section 3.4.2). Carolingian features include ditches which seem to enclose (farm) yards on which posthole clusters, pits and wells were found. Most of the contents of the pits were sieved which yielded an enormous amount of glass fragments, including a rare gold-foil beaker, glass beads, bone artefacts and birds and fish bones. The bone artefacts included production waste. The adjoining Albert Heijn supermarket site also produced Carolingian period loom weights and traces of metalworking.198 Pottery finds were made up of the usual range of Carolingian wares, for instance Rhenish and Eifel ceramics, as well as younger wares from Pingsdorf, Andenne and Paffrath. The finds prove this part of the settlement remained inhabited after the Viking attacks and only shifted in a south-easterly direction in the mid-13th century, where the town of ‘Wiic bi Duerstede’ would develop. Glass production waste: Some thirty-six fragments of glass-working waste, including eight tesserae, were found distributed throughout this part of the settlement (trenches 810–815).199 Among these was a large fragment of a glass crucible (fig. 3.9). It was found in a pit together with other glass-working waste made up of two blue tesserae, a regular and an irregular drop of translucent pale green glass, a small, dark sphere, six melted lumps of translucent pale green vessel glass and a melted fragment of ‘black’, deep olive-green glass. The crucible belongs to a pluriform group of bowls (type WX in the Dorestad typology)200 which are late Merovingian in origin and are made in different production centres in the Rhineland. On the Dorestad-Veilingterrein site they are mostly 8th century in date and are also found in Carolingian yards.201 The inside of the crucible is covered with a thin layer of almost colourless glass of c. 1 mm thickness with a thicker patch of opaque white glass which probably represents (a fragment of) a melted tessera. Preiß points out that defects in 41 — the translucent glass probably indicate locations where other (crushed?) tesserae had been attached.202 The crucible can be linked to beadmaking, but also to vessel or window glass production. 3.4.2 Wijk bij Duurstede – Veilingterrein and Frankenweg/Zandweg Site: 6 Site type: Settlement Province: Utrecht Municipality: Wijk bij Duurstede Place: Wijk bij Duurstede Toponym: Veilingterrein (Zandweg) & Frankenweg/Zandweg Start date: c. 600–650 End date: 900 Description: Due to their proximity these two sites have been treated as one. The archaeological remains belong to the same agrarian settlement west of the vicus and harbour works as the PPAH-excavations (see Section 3.4.1). The excavation Frankenweg/Zandweg was carried out by the Archeologisch Diensten Centrum (ADC) in 2001 before the planned construction of an apartment block on this site.203 During the Merovingian phase (c. 600– 725) no buildings were found, but four wells indicate there must have been some. It was only possible to reconstruct one Carolingianperiod building with a boat-form, possibly a farmhouse. Other Carolingian features are ditches and eight wells. Metal slags were found as well as waste products from antler-working and possibly glass-working. Habitation decreased dramatically during the third phase (late 9th–10th century) and was discontinued during the high middle ages when the area was in use for arable farming. Immediately south of the 2001 excavation, the ADC carried out another, much larger excavation (1.7 ha) on the site of a former fruit auction (Veilingterrein) along Zandweg.204 The archaeological remains belong to the same agrarian settlement west of the vicus and harbour works as the PPAH excavations (see Section 3.4.1). Habitation started in the Merovingian period around 650. Three large farm yards were identified in this area. On these, the remains of two farmhouses were identified as well as twelve wells, many waste pits and inhumations. In the third quarter of the 8th century a new partition of the land took place into rather narrow strips of land, clearly delineated by ditches. These boasted eight buildings, 112 wells, many pits, latrines and oven-pits. Seven boat-shaped farmhouses were identified and one building with straight sides, its function is uncertain. The farm houses do not show obvious differences in size or layout. Several yards yielded evidence for ironsmithing and weaving wool, one (yard K4) provided clear evidence for specialized crafts, namely amber-working and the production of brass (terminus post quem 800), while finds of two touchstones with traces of gold and the largest concentration of coins point to trade activities in this yard. A large amount of pottery from the German Rhineland and the Eifel, mill stones from the Eifel, wine glasses most probably from the Rhineland, wine (in barrels) from the middle Rhine region, combs from Scandinavia, and Roman tesserae, possibly from the Mediterranean, underline the international character of the settlement and the importance of trade. The period between 875 and c. 1050 saw a steep decline in habitation and only one farmhouse can be identified, while three were found dating between c. 1050 and 1300. From c. 1300 onwards, the area was used for arable farming. Glass production waste: The glass production waste from the Veilingterrein was published in 2012 in a monograph on the excavations, and recently in an overview article on the beads from Dorestad.205 Tesserae make up the largest category (n=13), almost all in the blue/green colour spectrum. Fragments of translucent bluish-green and dark blue glass point to imports of (chunks or ingots of) raw glass. One bluish-green flake was struck off from a larger lump of raw glass and has an imprint of a glassworker’s tool (fig. 3.10). The small diameter of the tool makes it more likely it was a beadworking tool (punty). Two convex fragments of dark blue glass (fig. 3.10) clearly belong to plano-convex ‘cakes’ also found on Scandinavian bead-making sites such as Åhus, Sweden.206 There are three glass drops in corresponding colours (fig. 3.10). Two glass rods of opaque yellow glass (square- 202 203 204 205 Preiß 2010. Sier, Van Doesburg & Verwers, 2004. Dijkstra, 2012. Sablerolles & Henderson 2012, 326–333; Langbroek 2021b. 206 Callmer & Henderson 1991, fig. 2. 42 — 6071 5041 0 1:1 4834 5cm 6500 Fig. 3.10 Dorestad-Veilingterrein site: A translucent bluish green flake of raw glass with imprint from a probable punty (6071), a convex fragment from a translucent blue glass ‘cake’ (5041), a plano-convex drop of bluish green glass (4834), an opaque yellow glass lump with a section of an opaque yellow composite rod and a fragment of punty glass (6500) (Sablerolles & Henderson 2012, fig. 6.19). 207 Sablerolles & Henderson 2012, 329–330, 208 209 210 211 212 213 afb. 6.21, findnr. 5791; Langbroek 2021b, table 7, findnr. Veilingterrein 5791. Langbroek 2021b, fig. 12.5, 64, table 7, findnr. Veilingterrein 5791. Lassaunière et al. 2016; Henon 2020. Langbroek 2021b, table 7, findnr. Veilingterrein 1195. Sablerolles & Henderson 2012, 326–333, afb. 6.21, 5791; Langbroek 2021b. Van Doesburg 2004. Langbroek 2021b, table 7, findnrs Zandweg WD 754.2.63b, WD 754.2.63b. sectioned) and white glass (round-sectioned) are likely to be associated with bead-making. A lump of opaque yellow glass has part of a composite opaque yellow glass rod and yellow punty glass from a glassworker’s tool melted onto it (fig. 3.10). Composite glass rods were also found on the Merovingian bead-making site of Maastricht-Jodenstraat and can be linked to bead-making (see Section 3.2.1). An opaque green object, first believed to perhaps represent the pinched end of a glass rod,207 is actually a failed, wound bead and represents waste from glass bead production.208 A fragment of a twisted bichrome rod has a translucent bluish-green core with an opaque white trail twisted round it. In the Carolingian period reticella rods with translucent greenish cores and opaque twisted trails were used to decorate vessel glass. Twisted bichrome cables have been found at the French monasteries of Hamage and St Amand-les-Eaux on the river Scarpe, where they were used to decorate globular jars and bowls.209 A thick, flat piece of opaque yellow glass shows black traces of iron oxide scale in the fractures on the sides as well as what seem to be small amounts of ceramic (from a crucible?) or red-baked clay.210 This yellow glass could have been intended for bead-making or for decorating 8th century vessel glass, funnels, jars and bowls, with yellow trails. Finally, a quantity of melted bluish-green glass adheres to a fragment of red-baked clay.211 In several contexts on the abovementioned yard K4, a small concentration of glass bead production waste, including four tesserae, was found together with waste from amber-working, raising the possibility that either a beadmaker was working here side by side with an amber-worker or that both crafts were carried out by the same craftsman. A combination of waste from amber and a small concentration of bead-working waste, including five tesserae, is also found on the northernmost yard (yard K1), which saw most activity around 800. Finally, from the 2004 excavation (Frankenweg/Zandweg), a tessera of ‘bright blue glass’ and a lump of clay covered with a thick layer of ‘blobby greenish glass’ were found immediately north of yard K1 on the Veilingterrein site.212 These are not included in Preiß’s 2010 overview of glass production waste from Dorestad. Nor are two small spheres of whitish glass, possibly weathered translucent glass.213 These finds point to the local production of beads, but the production of vessel glass cannot be excluded. There is a relatively small amount of glass production debris, but it is worth considering that, in contrast to the situation in the famous Danish bead-making site of Ribe, no original floor surfaces were preserved, so only finds from pits, wells and ditches were recovered, and that wet sieving was only carried out in specific 43 — instances.214 It is therefore hard to gauge whether this constitutes small-scale production at household level, production for local demand or (supra-)regional production. 3.5 Utrecht, Utrecht Province Utrecht may be one of those rare places in the Netherlands where there are indications for Merovingian glass-working – though we need to await the final results of the post-excavation work to establish if there is more comprehensive evidence – as well as evidence for glass-working in the Carolingian period. 3.5.1 Utrecht-Domplein Site: 9 Site type: Proto-urban settlement Province: Utrecht Municipality: Utrecht Place: Utrecht Toponym: Domplein Start date: c. 40 End date: present day Description: The Domplein (Dom Square) in Utrecht city centre is named after the Domkerk (Dom church or St Martin’s Cathedral) which dominates the central square. Excavations carried out in the 20th and 21st centuries have made clear that the square was continually inhabited from the Roman period to the present-day. Small-scale excavations took place between 1927 and 1949.215 These revealed parts of an auxiliary fort – Traiectum – which was first constructed in the 40s of the 1st century. After four wooden phases, it was rebuilt in stone around 200 and abandoned during the course of the 3rd century. Furthermore, evidence for Carolingian and high medieval habitation was found, as well as remains of the 11th century imperial palace – Lofen. Excavations in 1993 of part of the Heilig Kruiskapel (church of the Holy Cross), which was demolished in 1829, made clear that the church was probably founded around 700, rather than in the 10th century as was suggested in 1929. This has led to speculation that this simple hall church may be identified as one of two churches reputedly built here by the Anglo-Saxon missionary Willibrord around 695: either the now-lost church of St. Salvator (SintSalvatorkerk) or the church of St. Martin (SintMaartenskerk), the predecessor of the presentday Dom church. Glass production waste: In total, 17 fragments of glass crucibles were found during excavations in 1933 at the site of the Roman castellum on or near the Domplein (table 3.5).216 Of these, five were published in 1934 and illustrated by the famous Dutch artist and illustrator Anton Pieck (fig. 3.11).217 Three of the published fragments (body fragments 203, 204 and B8) were not among the crucibles made available for sampling carried out at the Cultural Heritage Agency of the Netherlands (RCE).218 Two of the latter were found just west of the Domplein in ‘Flora’s hof’ (courtyard). Judging from the illustration (fig. 3.11), the glass colours in these 214 Dijkstra 2012, 25–27, 591. 215 Wynia 2013. 216 Vollgraff & Van Hoorn 1934, 63–64. The objects are part of the collection of the Provinciaal Utrechts Genootschap van Kunsten en Wetenschappen (PUG) or the Provincial Society for the Arts and Sciences (PUG findnumber 234). 217 Vollgraff & Van Hoorn 1934, pl. XXII. 218 The crucibles were sampled by Hans Huisman, Dutch National Heritage Agency (RCE), Amersfoort. Fig. 3.11 Utrecht-Domplein site: Illustration of glass crucible fragments by Anton Pieck (Vollgraff & Van Hoorn 1934, 63-64, pl. XII); top row: findnrs 234 (rim), 203 (body); middle row: findnrs 204 (body), B8 (base); bottom row: findnr B8 (body) (Vollgraff & Van Hoorn 1934, pl. XXII). 44 — crucible fragments are a deep blue-green rather than ‘dark green’ and pale bluish green rather than ‘pale green’. Six photos of the glass production waste are given at Appendix IV (figures appendix IV.39-45). 219 Isings, Rauws, Lägers & De Kam, 2009, 48–49. 220 Isings, Rauws, Lägers & De Kam, 2009, 221 222 223 224 225 48–49. Barfod, Feveile & Sindbæk 2022. Sablerolles and Henderson 2012; Langbroek 2022. See for instance Baumgartner & Krüger 1988, 71, No. 14, 72, No. 16. Stoepker 2021, 230–232, Table 11.1; Sablerolles 2023. Van Es & Verwers 1980, 81–87. At the time of the excavation, the fragments were erroneously believed to be Roman in date, but in a publication of 2009 a rim fragment (DPL-234) and a base fragment (DPL 1933-zn3) were published as fragments of (a) Carolingian globular pot(s).219 Judging from photographs of the objects, all the other fragments probably also belong to this type of pot (see Dijkstra’s contribution below). Based on the find locations, the glasses inside the pots and their fabrics, at least eight pots are represented, but this number may have to be adjusted in future when the fragments themselves are studied. Two rim fragments have off-white deposits on the inside below the rim, on top of the rim and on the outside, just below the rim (DPL 1933 234, DPL 1933 zn2. The composition of these white deposits will be discussed in more detail in Chapter 5. A rim fragment (DPL 1933 234) has two very thick deposits of translucent glass which start just below the rim and are sticking out above it. This is likely to be the result of an attempt at removing the last remains of viscous glass from the crucible which solidified just before it could be poured out. It was suggested by Isings et al. that glass production on the Domplein is linked to making red glass inlays for jewellery or beads.220 A recent discovery of a glass bead workshop at Ribe dated to between 760 and 790 yielded fragments of a crucible with the remains of recycled Roman green glass with streaks of red and brownish coloration caused by iron and copper. It is suggested that this glass was used for the production of opaque red cylindrical beads and black wasp beads221, bead types that were also found at Dorestad.222 The glass inside the crucibles at Utrecht Domplein may well represent an attempt at making opaque red glass. However, red marbled translucent blue-green/bluish-green glass was popular in the late Merovingian and Carolingian periods for the production of vessel glass, especially bowls and bulbous jars,223 and for flat glass intended for stained glass windows. A fragment of a Carolingian red marbled bluegreen glass quarry was found at the early medieval monastery at Susteren (Limburg Province) (fig. 3.12), where there are indications for the production of window glass, most likely in the Carolingian period (see Section 3.3).224 In the context of the Domplein, production of window glass is certainly a possibility since there was at least one church here in the 8th century (see above). Interestingly, the monastery at Susteren had special links with the church of Utrecht as it had been founded by the AngloSaxon missionary Willibrord (see Section 3.3). 1:1 0 5cm Fig. 3.12 Susteren-Abdijterrein: A fragment of quarry of translucent blue-green glass with red marbling (Stoepker 2012, table 11.1; Sablerolles 2023). Glass crucibles: Utrecht-Domplein Contribution by Menno Dijkstra (University of Amsterdam) Based on the photographs, the crucible fragments probably all derive from Carolingian globular pots (Dorestad type W III)225 dating between 750 and 875/900 (fig. 2.13). This is indicated by the three rim fragments as well as the relatively thin-walled body fragments and the lenticular base. 1:4 0 20cm Fig. 3.13 Dorestad-Veilingterrein: Example of a complete Carolingian globular pot of Badorf ware (Dorestad type W IIIA) (De Koning 2012, afb. 4.28, findnr 503) (Photograph: Archeologisch Diensten Centrum (ADC). Height c. 17 cm. 45 — Table 3.5 Utrecht-Domplein: fragments of glass crucibles together with photo numbers and those scientifically analysed (Description of missing fragments from Vollgraff & Van Hoorn 1934, 63–64). Find number (modern) Find number (old) Crucible Glass Sample DPL 1933 zn2 ? 1 rim translucent bluish green & off-white deposit DOM 5 35 figure appendix IV.43 DPL – zn3 (Flora’s hof) B8 1 base translucent dark green glass (1 mm thick) marbled with opaque (purplish) red glass (0.5 mm thick) DOM 3 33 figure appendix IV.41 Missing (Flora’s hof) B8 1 body translucent pale green to dark green, red marbled - - - Missing 203 1 body translucent dark green layer (1-6 mm thick) - - - Missing 204 1 body translucent pale green glass, cracked (0.5–3 mm thick) - - - DPL 1933 234 234 1 rim translucent bluish green marbled with opaque red DOM 4 34 figure appendix IV.42 DPL 1933 73-84 ? 4 rim remains of translucent ‘garnet’ red glass in translucent bluish green glass DOM 6 36 figure appendix IV.44 DPL 1933 77-53 B7 1 body colourless glass (crucible: thin-walled beige fabric) DOM 2 32 figure appendix IV.40 DPL 1933 77-56 B26 1 body remains of opaque red glass - DPL 1933 77-36 B42 2 body colourless glass DOM 1 DPL 1933 77-57 B42 1 body colourless glass - - - DPL 1933 77-31 B45 1 body remains of opaque bright red glass - - - DPL 1933 77-58 B45 1 body opaque bright red glass - - - Total N Sample number Photo number 31 figure appendix IV.39 17 Perhaps this type of globular pot was preferred because of its closed form. Crucible fragments of similar pots were probably found at the monastery of Susteren (see Section 3.3) and in Leidsche Rijn – Leeuwesteyn Noord (See Section 3.6.2). The photographs show that the fragments are of a reddish colour on the inside and grey on the outside. One thin-walled fragment is beige on the inside (DPL 1933 77-53). The photographs show that the outside surfaces of the sherds are marked by small blisters which are located in places where the temper pierces the surface. These blisters are probably caused by reheating of the pots when reused as glass crucibles and may represent the early phases of vitrification. The presence of rims and the fact that the base fragment and the body fragments are not covered by spilt glass on the fractures, indicate only complete pots were used. This is in contrast to the late 6th-century Maastricht-Jodenstraat (see Section 3.2.1) and early 7th-century Rijnsburg-Abdijterrein sites (see Section 3.8) where complete pots contained colourless or translucent blue-green glass, while only the lower parts of the pots were used for working relatively small amounts of highly coloured opaque yellow glass: by removing the upper parts of the pots, the beadmakers created wide-open ‘bowls’ that allowed for easy access to the glass.226 It is difficult to make a statement about the types of fabrics represented on the basis just of photographs (table 3.5). Moreover, it can be hard to distinguish the different types of tempers in rather thin-walled Rhenish fabrics from Badorf and Walberberg in the Vorgebirge and Mayen in the Eifel. Parts of the photographs of the crosssections show the presence of quartz sand, but this can be present in both Mayen and Vorgebirge wares. Seven sherds show black specks, which are mostly tiny cavities, but some of them could be fragments of augite or hornblende, which are typical for Mayen fabrics (see table 3.6). Other typical Mayen volcanic temper-like pumice grains and off-white specks (clay pellets) are almost absent, but this is not uncommon in thin-walled fragments. This thinness could also explain the absence of small red particles, probably ferronic nodules, which are sometimes noted in Vorgebirge fabrics.227 Only the fabric of the sectioned rim fragment DPL 77-234 can with 226 Sablerolles, Henderson & Dijkman 1997, 304; Dijkstra, Sablerolles & Henderson 2011, 185–186. 227 Fabric details are based on Redknap 1988, 5 and 11; Bardet 1995, 221–230; Keller 2012, 213. 46 — Table 3.6 Utrecht-Domplein: Rim type and fabric of the Carolingian globular pots used as glass crucibles. Find number Crucible Rim type Fabric (based on photographs) DPL 1933 zn2 rim W IIIA coarse quartz sand, Walberberg fabric (Dorestad w-4?) DPL 1933 234 rim W IIIA black specks etc., Mayen fabric (Dorestad w-12) DPL 1933 73-84 rim W IIIA coarse quartz sand, Walberberg fabric (Dorestad w-8) DPL 1933 zn3 base - black specks, Mayen ware? DPL 1933 77-31 body - black specks, Mayen ware? DPL 1933 77-36 body - black specks, probably Mayen ware DPL 1933 77-53 body - not determinable DPL 1933 77-56 body - Mayen or Walberberg fabric? DPL 1933 77-57 body - black specks, probably Mayen ware DPL 1933 77-58 body - black specks, Mayen ware? W IIIA = globular pot of Badorf ware, dating between 750 and 875/900 (Dorestad type W IIIA). confidence be identified as typical for the production centre at Mayen (Dorestad fabric w-12). Rim fragment DPL 73-84 can be identified as a Walberberg fabric (Dorestad fabric w-8) due to the temper with coarse quartz sand. 3.5.2 Utrecht–Oudwijkerdwarsstraat 228 Post-excavation is ongoing at the time of writing. Many thanks to Marieke Arkema, external archaeologist, Municipality of Utrecht, Ontwikkelingorganisatie Ruimte, Duurzame Stad, Erfgoed, for all initial information on the site. Site: 10 Site type: Proto-urban settlement Province: Utrecht Municipality: Utrecht Place: Utrecht Toponym: Oudwijkerdwarsstraat Start date: c. 600 (?) End date: c. 750 Description: During building activities in Oudwijkerdwarsstraat just east of Utrecht old city centre, archaeologists of the municipality of Utrecht recovered remains of an early medieval settlement.228 It comprises part of a large settlement that stretched along the former bank of the Kromme Rijn for at least 150 metres. An initial assessment of the pottery points to habitation in the 7th and first half of the 8th centuries. This find is very important for two reasons: not much is known about early medieval life in this part of Utrecht, and the finds are extraordinarily well preserved. No house plans were found, since only the rear parts of farm yards were found. However, there are remains (postholes, clay) of probable subsidiary buildings, while ditches may represent yard boundaries. Waste pits yielded pottery, animal bones, millstones and loom weights, offering an insight into the economy of the settlement. About 1400 amber fragments, ranging in size from splinters to fairly large lumps, and including half-products, show amber was worked in the settlement. Never before has so much amber been found in Utrecht. The amber finds were concentrated in two shallow pits which yielded many other finds including 13 glass fragments. Furthermore, silver and copper coins and glass beads were also found. Glass production waste: Glass production waste is scanty. An irregular drop/melted lump of translucent bluish-green glass (with a yellowish tinge) (WP 5-1-135) could be the result of glassworking, but could also be accidentally melted vessel glass. However, when vessel glass is accidentally melted, for instance in a hearth, it is usually possible to tell it is a deformed vessel fragment. The glass drop comes from a pit (S135) which yielded animal bones, pottery and burnt clay. This pit was only discovered when sectioning an adjacent pit (S62) which yielded pottery, metal, burnt clay, animal bones and three amber fragments. The pits are at a distance of 30 metres from the two shallow pits mentioned above with concentrations of amber. A small amount of glass (WP 6-1-170) seems to be crushed, perhaps in preparation for being melted down. Two photos of the glass production waste -OUDWIJ 1 (sample 78) and OUDWIJ 2 (sample 79)- are given at Appendix IV (figures appendix IV.45 and 46). 47 — 3.6 Leidsche Rijn, Utrecht Province The large early medieval settlement excavated at Leidsche Rijn has yielded sparse evidence for glass-working in the late Merovingian and Carolingian periods. 3.6.1 Leidsche Rijn-LR 51/54 Site: 11 Site type: Riverine settlement Province: Utrecht Municipality: Utrecht Place: Utrecht Toponym: Leidsche Rijn 51/54 Start date: 575 End date: 775 Description: In the period 2000–2001 a field survey, coring and a trial excavation were conducted as a result of plans to expand the A2 highway in Leidsche Rijn located in the western part of the city of Utrecht.229 The archaeological research demonstrated that an early medieval settlement was situated here. In 2005 a part of this settlement was excavated by the municipality of Utrecht (project LR51 and LR54). The excavations revealed that the early medieval settlement was located on relatively high land along the northern bank of the Oude Rijn. The second half of the 7th and the early 8th century formed the heyday of the settlement. In the course of the second half of the 8th century habitation declined. However, a few stray finds from the 9th century suggest that the area was also inhabited later, in the Carolingian period. In total 88 early medieval buildings were found, which comprised 14 farmhouses and 74 outbuildings. Noteworthy are 34 large outbuildings with very long, pointed wooden posts, maybe granaries. As only pollen (no grains) of oats, rye and wheat were found, the arable fields were probably not very close to the settlement. Farm animals were reared, especially cows and sheep. The lack of bones of 2-to-4-year-old cattle in the oldest and youngest habitation phases may point to the export of cattle during these periods. The diet was supplemented with riverine and a b 0 1:1 5cm Fig. 3.14 Leidsche Rijn - Leeuwesteyn Noord: A fragment of a glass crucible, probably from a Carolingian globular pot (Dorestad type W III), a) covered on the inside with red marbled translucent greenish glass (left) and b) on the outside with a blob of translucent greenish glass (right) (Sablerolles 2019, fig. 7.15). imported marine fish as well as game. The inhabitants were engaged in (occasional) shipping activities and artisan production: iron was produced from imported iron ores and was also worked in the settlement; bronze and lead were worked. Combs were produced from antler and seven fragments of raw amber point to amberworking. As much as 80% of the pottery was wheel-thrown, imported from the Rhineland together with the remains of at least seven late Merovingian drinking glasses (palm-funnels). Trade activities may have taken place, judging by the discovery of 37 early medieval coins. It is suggested that imported wares such as pottery, glass, millstones, iron ore, coal and amber could have been obtained with agricultural surplus stored in the granaries, as well as animal products. Glass production waste: An opaque blue tessera was found in a waste pit on yard 3 in the eastern part of the settlement.230 This yard also yielded most evidence for craft production: metalworking (iron and bronze), bone- and antlerworking and amber-working. Moreover, almost all of the twelve glass beads were found on this yard, one of them in the same pit as the blue tessera, as well as most fragments of old, Roman and Migration period glass, which might represent cullet intended for recycling.231 The pit is dated to the late Merovingian period which would make the tessera the earliest specimen in the Netherlands together with a tessera from Oegstgeest – Nieuw Rhijngeest Zuid (Rijnfront) (see Section 3.7). 229 Nokkert, Aarts & Wynia, 2009. 230 Isings 2009, 247, afb. 11.2, table 11.1. 231 Isings 2009, 249, afb. 11, table 11.2, 250– 251. 48 — Although there is no glass bead production waste to accompany the blue tessera, the fact that it was found together with raw amber may point to a glassworker and amber-worker being active in this part of the settlement. 3.6.2 Leidsche Rijn-Leeuwesteyn Noord 232 Norde 2019. 233 Sablerolles 2019, 136–138. 234 Norde 2019, 274–275, fig. 14.1, 7. Site: 12 Site type: Riverine settlement Province: Utrecht Municipality: Utrecht Place: Utrecht Toponym: Leeuwesteyn Noord Start date: 575 End date: 850 Description: In 2015 and 2016 excavations, commissioned by the municipality of Utrecht and carried out by RAAP, took place immediately east of the early medieval settlement Utrecht-Leidsche Rijn LR 51/54 which was excavated a decade earlier (see Section 3.6.1).232 The settlement, on the northern bank of the Old Rhine, was probably located in an outer, not an inner bend of the river as had previously been assumed. The combined excavations have unearthed 275 metres of a large settlement which continues in easterly and westerly directions. The excavations confirm the idea proposed by the 2005 excavators that the settlement continued into the 9th century, probably till around 860. Just as in Dorestad-Veilingterrein, the area was reorganized in the Carolingian period and (parts of) yards are delineated by north/south oriented ditches and picket fences. Remains of two farms and several subsidiary buildings were found. This brings the total of houses in the Leidsche Rijn settlement to 14, while a total of 57 large subsidiary buildings were found. Of the latter, 37 are large, two-aisled buildings constructed with very long posts, deeply driven into the ground. They are interpreted as warehouses. It has been pointed out that similar structures are also found in some other 7th–8th century settlements along the Old Rhine (see Section 3.7) and it has been suggested that these could point to the existence of specialized settlements. It was argued that the paucity of cereal grains in botanical samples collected during the 2005 excavations may indicate that (this part of) the settlement was geared towards craft and trade rather than agriculture. There is evidence for antler-working; metal-working evidence consists of iron-smithing and bronze-working. There is possible evidence for lime-burning. A glass crucible indicates glass was worked in the settlement. Glass production waste: A body fragment of a crucible was recovered from one of the uppermost layers of the fill of the Old Rhine which contained both Merovingian and Carolingian pottery (fig. 3.14).233 The fragment is reddish on the fracture, and grey on the outside, and is probably of a Carolingian globular pot (Dorestad type W III, fabric 12), similar to the crucibles from Utrecht-Domplein (see Section 3.5.1). The crucible is covered on the inside with a thin layer (1–1.5 mm) of translucent (bluish-) green glass marbled with opaque red glass, comparable to some of the Utrecht-Domplein specimens. A drop of translucent greenish glass was spilt on the outside. In this context, it is less likely that this glass was being worked for the production of window glass as was suggested for Utrecht-Domplein. Perhaps, therefore, this glass was used for the production of beads. The only bead recovered from the site is a spiral bead from a Carolingian pit, made of faintly translucent greenish glass full of small bubbles, dark inclusions (iron fragments from a beadmaker’s tool) and what looks like black/dark red streaks.234 This bead may be one of the local products. It is similar to a bead found on the Maastricht-Rijksarchief site where beads were made (see Section 3.2.3). 3.7 Oegstgeest–Nieuw Rhijngeest Zuid (Rijnfront), Zuid-Holland Province Site: 13 Type: Riverine settlement Province: Zuid-Holland Municipality: Oegstgeest Place: Oegstgeest Toponym: Nieuw Rhijngeest Zuid (Rijnfront) Start date: c. 550 End date: 725 Description: From 2009 until 2014 the University of Leiden excavated a settlement dating to Merovingian times. Smaller parts of the settlement had already been excavated by 49 — ARCHOL and ADC.235 The Oegstgeest settlement has been excavated almost completely. It was of a modest size, with c. five or six contemporary farmsteads during its existence. Its population will not have been larger than c. 60 persons. It was located along the northern bank of the Old Rhine close to its estuary, some five km from the coast. The settlement was set in a landscape that was subjected to both riverine and maritime influences and was intersected by river arms, gullies and creeks, creating four quarters or ‘islands’ interconnected by one bridge and several small dams. Structures connected to shipping are jetties, quay works, and land abutments. As is the case at Rijnsburg-Abdijterrein (see Section 3.8) and Valkenburg-De Woerd (see Section 3.9), the farmhouses were laid out on a grid at right angles to a main gulley of the Oude Rijn. Each yard, enclosed by fences made of wattle, consisted of a farmhouse with several associated outhouses, multiple pits and wells. The postholes of the outhouses were remarkably deep and it is suggested that the buildings may have had raised floors enabling safer storage of agricultural produce. Animal husbandry in the floodplains (cattle), some agriculture on the river levees, and fishing were the backbone of the economy. They also practised crafts such as smithing, and above all casting of copper alloy objects. Small and unexpectedly large crucibles show that this must have taken place to satisfy the needs of themselves and those in other settlements. Antler combs were made and amber-working was a widespread activity in the settlement. There is some tentative evidence for glass-working. Imported pottery from the German Rhineland, grain from löss areas, probably the Main area, wine (barrels) from the middle Rhine area and some exotic imports testify that this settlement made good use of its advantageous location along one of the most important early medieval supraregional waterways, with the possibility to engage in exchange with the wider, early medieval world. Glass production waste: Five fragments of possible glass production waste were found in five different locations: two fragments of tesserae, a possible fragment of a glass rod, a black drop and a trapezium-shaped fragment of translucent blue-glass with rounded edges.236 Two weakly transparent green-blue fragments from a pit and a well from probable Roman tesserae could be the earliest evidence from the Netherlands for the reuse of tesserae by early medieval beadmakers, together with a specimen from the Merovingian settlement of Leidsche Rijn L51/54 (see Section 3.6.1). A colourless fragment is interpreted as part of glass rod used for winding beads. Since wound colourless glass beads were not common in the 7th century, it could perhaps also be a fragment of a reused Roman glass stirring rod. A drop of brown, almost black glass was split when it was heated. Small, spherical black drops have been found in association with glass bead production waste in Åhus, Dorestad, Maastricht-Rijksarchief and Wierum.237 The wound, monochrome and trailed beads found in the settlement (n=28) mostly date to the 7th century and show many similarities with half-products and wasters of beads found in the contemporary nearby bead production site at Rijnsburg-Abdijterrein (see Section 3.8). 238 The shapes, colours and decorations used are strikingly similar, including small flattened globular beads of opaque yellow, red and white glass, white beads with translucent bright blue crossing trails and a red bead with white crossing trails, and it is suggested that these beads could have been made in RijnsburgAbdijterrein.239 Alternatively, they could have been made by travelling beadmakers who visited riverine settlements along the Old Rhine, including Rijnsburg-Abdijterrein and possibly Valkenburg-De Woerd. 3.8 Rijnsburg-Abdijterrein, Zuid-Holland Province Site: 14 Site type: Riverine settlement Province: Zuid Holland Municipality: Katwijk Place: Rijnsburg Toponym: Abdij Start date: 600 End date: 12th century Description: Between 1944 and 1966 a series of excavations by the archaeological institutes of the universities of Groningen and Amsterdam and the former National Service for Archaeological Heritage (ROB, now RCE) took place on the site of a former Benedictine abbey 235 Hemminga & Hamburg 2006; 236 237 238 239 Hemminga et al. 2008; Dijkstra 2011, 134; De Bruin 2018, 20–25; De Bruin, Bakels & Theuws 2021. Langbroek 2021a, table 12.2, fig. 12.6. Callmer & Henderson 1991, Table 1C, 1; Preiß 2010, number 47; Henderson, Sode & Sablerolles 2020, 78. Langbroek 2021a, fig. 12.2. Langbroek 2021a. 50 — 240 Dijkstra, Sablerolles & Henderson 2011; Dijkstra 2011, 114-133. 241 Nicolay 2017. 242 Koch 1977, 207, Farbtaf. 3, Gruppe 34. 243 Pion 2014; Vrielynck, Mathis & Pion 2018. (c. 1130–1574), immediately east and north of the present-day church.240 The excavations of this convent led to the discovery of various older settlements lying underneath. The oldest phase of the settlement dates to around 600 AD, when the southern bank of the Vliet was divided with wattle fences into small north-south oriented plots where several rectangular longhouses and secondary buildings were built, that can be associated with three or four generations of occupation lasting until around 720. It was situated in a tidal saltmarsh at the southern bank of a creek (the Vliet) of the Oude Rijn, close to the former mouth of the river. The economy was primarily based on farming, although traces of oven-/furnace-like structures, possible outdoor hearths, smithing slags and a few tuyères point to iron-working, while two crucibles, bronze fragments and a probable casting mould point to bronze casting. One of the smaller buildings close to the creek is believed to be a possible home of a smith and his family. A few lumps of amber may point to amber-working. The evidence for the production of Merovingian type beads also comes from this phase of the settlement. During the late 6th century and the early decades of the 7th, the settlement can be considered part of a probable central place complex located in the mouth of the Rhine.241 Therefore, it could have been at the invitation of a local or regional ruler that a Merovingian beadmaker travelled to the settlement to produce fashionable glass beads. In the second phase (720-890) a new type of boat-shaped house appears; apart from its shape, there was a difference in orientation pointing to a different organisation of the plots. It is not clear whether there was a short hiatus in habitation. In the Carolingian period a chapel and a cemetery were added. Settlement traces which can be identified with the fortress of Rinasburg date between 890-1050 (phase 3), followed by the building of a new church accompanied by a farm or possibly a rectory (1050-1130, phase 4) and a Benedictine nunnery in the 12th century (phase 5). Glass production waste: Virtually all waste from glass bead production debris was found in a feature which mainly consisted of fired clay, the remains of a hearth or possibly a rudimentary glass furnace. There may be a connection with one of the small buildings in the immediate surroundings, a possible home of a smith and his family. Glass bead production took place during the second generation (phase 1b, c. 610–640) or third generation of building (phase 1c, c. 640–680). The glass bead production waste (objects) consists of finished, unfinished and failed beads (n=68), glass rods (n=45), punty glass from a beadmaker’s tool (n=3), crucibles (n=8) and one undiagnosed fragment. Two lumps of fired clay covered with translucent greenish glass may be from the furnace floor. Several categories represented among the bead production waste from MaastrichtJodenstraat (see Section 3.2.1), such as glass drops and pulled threads, are missing at Rijnsburg. Scrap glass is also lacking. The latter is perhaps a coincidence, but it was noted that not a single fragment of Merovingian glass vessel was found in the entire settlement or in the nearby cemetery. Excluding the crucibles, the waste categories are dominated by opaque yellow glass (an average of 49.6%), followed by opaque white (24.8%), opaque red (18.8%), opaque turquoise (5.1%) and opaque orange (1.7%). Beads that were produced in the settlement include monochrome globular, bi- and tri-globular beads of opaque yellow glass, and bi-globular beads of red glass. Trailed beads include biglobular beads of red glass with both white crossing trails and a white spiral, tri-globular beads of opaque red glass with opaque white crossing trails and white beads with translucent blue crossing trails. In the context of the beads from the cemetery of Schretzheim, Koch was quoted as stating that beads with narrow crossing trails represent billige Massenware and are ubiquitous in necklaces of the later 6th and 7th centuries.242 This date can now be refined by a more recent bead typology developed for beads from cemeteries in Belgium by Pion and Vrielynck, Mathis and Pion. 243 The above-mentioned bead types are all typical for Pion’s Bead period 4 (600–640) (Table 3.7). Given the types of beads that were produced here, it is therefore most likely that the rudimentary furnace was in use during settlement phase 1b (c. 610-640). Most of the beads were split during manufacture, either due to overheating of the glass or not annealing the beads properly after manufacture, a common occurrence on beadmaking sites (see Section 3.2.1 and Section 3.12). 51 — Table 3.7 Rijnsburg-Abdijterrein: A selection of locally produced beads, their typology and bead periods (Pion 2014; Vrielynck, Mathis & Pion 2018). Form Colour Decoration Type Period Bi-globular opaque yellow - B1.2-1b P4 Tri-globular opaque yellow - B1.2-1c P4 Bi-globular opaque red - B1.2-2b P4 Bi-globular opaque red white crossing trails & 1 white spiral B5.2-1h P4 Tri-globular opaque red white crossing trails B3.2-1c P4 Globular opaque white translucent blue crossing trails B3.3-3a P4 Period: P4=610-640 AD. Chemical analysis of the glass from Rijnsburg-Abdijterrein indicates that the rods are of a very similar composition to the beads and therefore that the beads are very likely to have been made from the rods on site. Furthermore, the opaque yellow glass from the crucible fragments – although not associated with the production waste from the furnace – is proven to be of the same general chemical lead-stannate composition as the opaque yellow glass production waste. Five fragments of glass-bearing crucibles were retrieved from different contexts within the Merovingian settlement, although one may derive from a section through the furnace. During the campaign of 1963 three fragments of crucibles were found, of which two smaller fragments derive from contexts which had intrusions from later periods. There are six base fragments and two body fragments. These belong to coarse-ware cooking pots (Wölbwandtöpfe) which were deliberately broken to obtain their bases for use as shallow, dish-like crucibles, comparable to those from the Maastricht-Jodenstraat site (see Section 3.2.1). When initially investigated, the crucibles were thought to have glass attached to them. However, no detailed scientific analysis was carried out as part of this project to ascertain whether this material is glass or not. A base covered with what appears to be yellow glass on the inside and on the fracture may come from a section through the ‘furnace’. A body fragment has what appears to be opaque yellow glass over a white layer, together with a greenish spot with streaked colourless and yellowish glass-like material lying over it; the streaked material covers the fracture while opaque yellow and yellowish/white spots can be seen on the outside. A small base fragment has opaque yellow material sticking to the outside of the base and colourless glass with small opaque yellow spots on the inside; a second small base has the same characteristics. A body fragment has a thin layer of colourless glass on the inside and an irregular, bubbly glass layer on the outside, indicating this was probably overheated and bubbled over. None of the fragments show any traces of vitrification. The remaining three bases show traces of vitrification on the outside and may have been used for glass- or metal-working. It was suggested that a colourless base glass was modified on site using lead-tin-yellow pigment. Further scientific research needs to be carried out in order to investigate/ confirm whether fully formed yellow glass is present. 3.9 Valkenburg-De Woerd, Zuid-Holland Province Site: 15 Site type: Riverine settlement Province: Zuid-Holland Municipality: Katwijk Place: Valkenburg Toponym: De Woerd Start date: 525 End date: 950 Description: Excavations between 1986 and 1988 by the former National Service for Archaeological Heritage (ROB, now RCE) revealed the remains of an early medieval settlement at Valkenburg-De Woerd. The provisional findings were published in 1987 and 1990.244 The settlement was laid out 244 Bult & Hallewas 1987; Bult, Van Doesburg & Hallewas 1990. 52 — 245 246 247 248 Jezeer & Jongma 2002 (in Dijkstra 2011). Dijkstra 2011, 172. Magendans & Waasdorp 1989. Magendans & Waasdorp 1989, 33; see also Dijkstra, Sablerolles & Henderson 2011, 192. on a natural levee along the inner curve of a meander of the Oude Rijn, between gulleys on either side. The settlement was just a few kilometers away from the early medieval settlement at Oegstgeest-Nieuw Rhijngeest Zuid (Rijnfront) on the opposite side of the river. The history of Valkenburg-De Woerd begins at the establishment of the Roman limes. In the mid-first century CE a military entrepôt harbour was laid out here, which must have been part of the vicus of castellum Valkenburg. Roman occupation ceased around 230 AD. Ceramic finds from a transect cut across a gulley of the Oude Rijn date between 525 and 950, with most finds dating to the 8th–9th centuries.245 The shore was divided into plots laid out on a grid at right angles to the river, similar to the situation at Oegstgeest-Nieuw Rhijngeest Zuid (Rijnfront) (see Section 3.7) and RijnsburgAbdijterrein (see Section 3.8). There were probably six to eight yards (width c. 50 m) divided by ditches, simultaneously at any given time during the Merovingian and Carolingian periods. The plans of the buildings are very unusual, mostly two-aisled, and are difficult to interpret. Farms like those found at RijnsburgAbdijterrein and Oegstgeest Nieuw Rhijngeest Zuid (Rijnfront) are lacking. Dijkstra points out that in the Merovingian period, two-aisled buildings functioned as barns and he hypothesises that the buildings on De Woerd may have combined two functions: traders may have lived and worked in them, while the buildings were used to store products or agricultural produce during the trading high-season, drawing a comparison with two- and three-aisled buildings on the dams in the Dorestad harbour.246 In the south-eastern part of the settlement the remains of revetments and a jetty were found, underlining the importance of the river as a mode of transport. There is evidence of boneand antler-working and of livestock rearing. Glass production waste: A glass crucible fragment (Find No. 510-4-307) was found in one of the trenches (trench 510) cut across the river. The crucible is covered on the inside with a thin, even layer of translucent pale greenish glass. A recent examination of the crucible by Epko Bult, University of Leiden, revealed it is a lower body fragment of a Merovingian Wölbwandtopf dating to the 7th rather than the 6th century. The crucible may, therefore, be contemporary with the crucibles from the nearby settlement at Rijnsburg-Abdijterrein and could be linked to bead-making (see Section 3.8). 3.10 Den Haag-Frankenslag, ZuidHolland Province Site: 16 Site type: Coastal settlement Province: Zuid-Holland Municipality: Den Haag Place: Den Haag Toponym: Frankenslag (Johan van Oldenbarneveltlaan 91–95) Start date: 500–550 End date: around 700 Description: Small-scale excavations (385 m2) carried out by the municipality of Den Haag in 1984 yielded part of a Merovingian settlement located on the eastern side of a coastal barrier.247 The settlement started in the first half of the 6th century and ended in the late 7th or early 8th century. Shortly afterwards, there is evidence for arable farming until the settlement was covered by drift sands (the Younger Dune formation phase-0). The remains of pits, hearths, three houses, and two successive sunken huts were found which were probably used for weaving. The inhabitants grew rye and barley on the nutrient-poor sandy soils, and reared cattle and sheep. They supplemented their diet with locally caught marine and riverine fish, game and wild fruit. Locally sourced bog iron was processed for the production of iron. Finds of Rhenish pottery and millstones, bronze and lead are believed to have been obtained by generating agricultural surplus. The Meuse and Rhine river systems could have been accessed over land (by way of the coastal barriers or the beach) or by sea. Glass production waste(?): A few sherds of brittle hand-made pottery were found in a sunken hut. They are covered on the inside and outside with dark, deep blue-green ‘glass’, perhaps due to vitrification of the fabric of the crucible.248 It is not clear if these fragments represent waste from glass- or metal-working. 53 — 3.11 Bloemendaal-Groot-Olmen, NoordHolland Province Site: 17 Site type: Coastal settlement Province: Noord-Holland Municipality: Bloemendaal Place: Bloemendaal Toponym: Groot Olmen Start date: 675 End date: 850 Description: In the dunes of the National Park Zuid-Kennemerland near Bloemendaal early medieval remains were found at 14 different locations.249 The remains, which had been buried under the Younger Dunes (formed between 1200 and 1600), appeared when the area was restored to its former ‘driftsand’ state by de-turfing. In 2006 and 2007 Hollandia excavated a settlement (location 1-3) dating between the 5th and 7th centuries. A survey combined with some small trial trenches carried out by the ROB prior to the Hollandia excavations showed that locations 4, 5, 8 and 14 were in use during the 8th and early 9th century. Hollandia excavated locations 8 and 14 which were part of the same settlement. In total, seven buildings were discovered, one barrel-lined well and remnants of fences. Site 8, where habitation layers were partially preserved, represented a single building dated to the 9th century. Site 14 yielded the remains of six buildings, including three house plans with a distinct boat-shaped form comparable to ‘urban farms’ found in Dorestad-De Heul. This imported building tradition most probably originated in the central riverine area and the Veluwe. Evidence points to the agrarian nature of the settlements, while marine fish and molluscs played a more important role in the diet than in the older settlement (location 1-3). Pottery, glass, millstones and whetstones were imported from the Rhineland and the Eifel. In the 8th century the North-Holland coastal region was incorporated into the Carolingian empire and it has been suggested by de Koning that the settlement at Bloemendaal-Groot Olmen may have been connected to a royal domain which, according to historical sources, was located in the area around nearby Velsen.250 Glass production waste: A surface find of an opaque dark blue tessera was found near location 14 (8th–9th century).251 This location also yielded a few fragments of thick-walled, blue-green Roman glass which could be cullet intended for recycling, perhaps to make the kind of globular ‘bottle’ blue-green bead that was also recovered from this location.252 3.12 Wijnaldum-Tjitsma, Friesland Province Site: 18 Site type: Terp settlement Province: Friesland Municipality: Harlingen Place: Wijnaldum Toponym: Tjitsma Start date: c. 50 AD End date: 950–12th century? Description: In 1990 fragments of a 7th century gold cloisonné royal brooch were found in a field on the Tjitsma terp near present-day Wijnaldum by metal detecting. Its footplate had already been found by chance in the 1950s. These finds were the catalyst for the excavations that were carried out by the Universities of Groningen and Amsterdam on the eastern crest of the terp settlement between 1991 and 1993.253 Although they yielded a wealth of information, no tangible remains of the king or a royal residence were found. A second volume on the ceramic assemblage was published in 2014.254 The early medieval artificial mound or terp settlement at Wijnaldum was located on a salt marsh ridge which was oriented east-west. There is evidence it was settled since the 1st century AD. It was one of a number of closely spaced terps by the salt marsh which was open to the sea. It is assumed that it was quite densely populated since the Roman period, including during the early middle ages. A recent field survey has shown that the beginning of habitation probably started as early as the 1st century.255 The end of habitation on the terp may have come in the 12th century when the last farmstead may have moved to a separate house terp, just like other farms in the terp region. The heyday of the terp settlement was the period between 550 and 650 when the area 249 250 251 252 De Koning 2015. De Koning 2015. De Koning 2015, 317–318, afb. 11.6. Sablerolles & De Koning 2015, 311–316, afb. 11.1, 3, 4, 10. 253 Besteman et al.1999. 254 Nieuwhof 2020. 255 Kaspers 2020. 54 — surrounding Wijnaldum, northern Westergo, had developed into the centre of a kingdom that covered the entire terp region of the northern Netherlands. According to Nicolay, the distribution of gold jewellery in a distinctive style suggests that the king residing at or near Wijnaldum had retainers across this entire area (see Section 3.13).256 The evidence for glass working on the terp dates to this period and it seems likely that the elite status of the settlement played a role in attracting a travelling Merovingian beadmaker to visit the settlement. Traces of habitation in this period are modest though and include the remains of six buildings divided over four households: four (possible) sod houses, two granaries and a sunken hut. Each house was built on a house platform built from sods. The houses were N-S orientated towards two large boundary ditches running east-west immediately south of the platforms. One (possible) house yielded evidence for two hearths and evidence for metal-working. Wheel-thrown pottery imported from the Rhineland makes up 63.7% of the total ceramic assemblage on the terp during this period and it is thought that Wijnaldum or northern Westergo was a distribution centre for Merovingian pottery; traders of Frankish goods such as pottery (or its contents) may have depended on the Wijnaldum elite for access to markets in the northern coastal area. The northern Netherlands became incorporated into the Frankish empire during the 8th century, an area equivalent to present-day Friesland in 734, and Groningen in 784 AD; northern Westergo was no longer the political centre controlling the area. An increasingly more 256 Nicolay 2014, 20–23. 257 Sablerolles 1999, 263–266. even distribution of imported Carolingian pottery across the northern coastal area probably shows that traders were able to access the area and were no longer obstructed or controlled by the political centre. A reflection of this was the percentage of imported pottery at Wijnaldum during the Carolingian period, which increases to c. 13.3%. Habitation was concentrated in the south-eastern part of the excavated area of the site during the Carolingian period, on the southern flank of the terp. The highest parts of the terp were used as arable fields, also found on other terps during the 1st millennium. Glass production waste: Glass-working evidence is sparse.257 The most important object is a very thick fragment of baked clay, possibly part of a tray or a glass furnace, covered with a thick layer (1.0–1.3 cm) of weathered opaque yellow glass which has permeated through the porous, pinkish-orange fabric (fig. 3.15). It was found amongst waste from metal-working by a blacksmith/bronze-caster. The dump is very closely dated to the last quarter of the 6th and the first quarter of the 7th century and is contemporary with the glass-working evidence from the Jodenstraat site in Maastricht. Two small (flattened) globular beads of opaque yellow and white glass accompanied this find and are among the likely local products. This type of bead was also found among the bead production waste from the Jodenstraat site in Maastricht, where only yellow examples are represented. Many glass beads found on the terp were in one of two large boundary ditches (550–600) and many of these simple, wound beads – including small flattened globular beads of opaque yellow, Fig. 3.15 Wijnaldum-Tjitsma: Detail of possible furnace floor or tray with opaque yellow glass permeating through the fabric (Photograph: Henk Faber Bulthuis, Noordelijk Archeologisch Depot, Nuis). 55 — red and white glass – could have been local products.258 Several halves of short cylindrical beads were retrieved from the above-mentioned large boundary ditch. They are split lengthwise, along the perforation, and are probably failed beads due either to overheating of the glass or as a result of not annealing the beads properly, causing them to crack (see Section 3.2.1). The same ditch also yielded three fragments of unworked amber, suggesting glass and amber bead-making could have been carried out at the same time.259 An inhumation burial on the terp from 550–600 AD contained a necklace with at least 22 small, rather roughly shaped amber beads which were perhaps made locally.260 A transverse breaking splinter of an opaque greenish-white rod (see Section 3.2.1) comes from a 5th century context. It is, however, not securely dated, so perhaps this fragment is contemporary with the above-mentioned furnace or tray fragment. Moreover, almost all context-dated opaque white beads from the terp date to the second half of the 6th century or between 575 and 625. In view of the paucity of the material, this production waste was interpreted as relating to just one production event. Because the glass waste production was found among that of a bronze-caster, it was suggested that beadmaking could have been a secondary activity carried out by, for instance, a bronze-caster or a gold- or silversmith who visited the terp occasionally. The possibility of a travelling beadmaker, however, cannot be excluded as it would be logical for such a craftsman to seek out (more) permanent high-temperature craftsmen on the terp. There is also some very limited evidence for bead-making on the terp in the Carolingian period. It consists of a fragment of translucent deeply coloured blue-green (turquoise) ‘punty’ glass from around a beadmaker’s tool. It is from a context with a reliable date between 750–800. An opaque yellow tessera from a ditch is probably dated between 750–770.261 The tessera clearly shows thin swirling layers of colourless glass within the yellow matrix, indicating that the yellow opacifier is not fully homogenized with the translucent base glass (fig. 3.16). 2:1 0 2,5cm Fig. 3.16 Wijnaldum: Opaque yellow tessera of opaque yellow glass streaked with colourless glass. The dimensions of the yellow tessera are: length 1.25 cm; height 0.81 cm and width 0.97 cm. (Photograph: Henk Faber Bulthuis, Noordelijk Archeologisch Depot, Nuis). These two fragments do, of course, not constitute solid evidence for bead-making on the terp in the second half of the 8th century, but they at least raise the possibility, especially in view of the recently published tesserae finds from the terp of Wierum (see Section 3.13).262 3.13 Wierum, Groningen Province Site: 19 Site type: Terp settlement Province: Groningen Municipality: Winsum Place: Wierum Toponym: Wierum Start date: c. 400 BC End date: late middle ages. Description: The largest find of Roman tesserae in the Netherlands originates from the terp of Wierum near Wierumerschouw (Groningen Province) in the northern coastal region, which was a frequently flooded salt-marsh area.263 The terp was located on the wide river Hunze, later renamed Reitdiep, which connected the Wadden Sea and North Sea with inland locations . The find is regrettably without a context and is likely to have been discovered between 1912 and 1916 when an estimated 3.5 ha of the original 5 ha of the site was dug commercially for its fertile soil. Only c. 1.5 ha of the original terp remained. In addition to the results of a coring programme, that provided information on the original circumference and the subsoil of the terp, in 1983 an overview of the finds was published.264 258 These finds come from boundary ditch 1233 (Sablerolles 1999, 270–273 passim). Sablerolles 1999, 277, cat. nr 226–228. Sablerolles 1999, 276, cat nr 191–213. Sablerolles 1999, cat nr 216. Crocco et al. 2021. Nieuwhof 2006; Crocco et al. 2021 and references therein . 264 Miedema 1983. 259 260 261 262 263 56 — 265 Nicolay 2014. 266 Preiß 2010, 124, 130 number 47; Callmer & Henderson 1991, Table 1C, 1. In 2004 Groningen Province decided to restore the terp to its original size and shape using soil dredged from the river Reitdiep. Archaeological excavations carried out by the Groninger Instituut voor Archeologie (GIA) of the University of Groningen (Rijksuniversiteit Groningen) revealed that the site was inhabited from the 4th century BC or slightly earlier, until at least the late middle ages, probably with an interruption in the 4th century AD. Unfortunately, no farmhouses and outbuildings or artisanal areas were excavated. It is unlikely that there was much labour specialisation because Wierum was mainly a self-sufficient agricultural settlement, like all terps. The find of a crescent-shaped gold pendant suggests that one of the retainers of the king who resided in or near Wijnaldum lived at Wierum in the Merovingian period (see Section 3.12).265 During the 8th or possible the 9th century the region became incorporated into the Frankish empire. In the course of the Merovingian period a local leader may have made the settlement of Wierum his home. Because of its advantageous position on the river Hunze, it is surmised that by the Carolingian period it may still have had regional political significance. The combination of its convenient location and political status may have attracted itinerant craftsmen, including beadmakers. Glass production waste: The assemblage has been interpreted as a supply of ‘raw’ glass of an early medieval glass beadmaker, most likely active on the terp in the 8th/9th century. This may have been a travelling beadmaker visiting terp sites such as Wijnaldum in the northern coastal region, which was most easily accessible by boat from the central riverine area with Dorestad at its centre. The glass finds are dominated by (fragments of) 201 tesserae. Most tesserae are affected by heat: something which can be the result of having been in a high-temperature workshop environment. Other glass finds are made up of five fragments of highly coloured early Roman vessel glass, one fragment of possible naturally tinted Roman or early medieval vessel glass, three plano-convex drops of opaque green glass, almost colourless glass and translucent dark blue glass, and 13 irregular drops/melted lumps of (almost) colourless, pale green and pale blue-green glass. The latter may be recycled Roman vessel glass or Roman gold- foil tesserae stripped of their gold-foil. A small, matt grey sphere may be a globular glass drop of a type also found at bead-making sites of Maastricht-Rijksarchief (see Section 3.2.3), Wijk bij Duurstede (Dorestad) (see Section 3.4), and Åhus in Sweden.266 Chemical analyses confirm the Roman date of the analysed glass finds. Apart from the glass finds, the assemblage includes four stone tesserae: two of green porphyry, one of purple porphyry and a white tessera, probably white marble, which is still embedded in mortar showing it was robbed from an ancient mosaic. Three more tesserae show the remains of mortar adhering to one side. Furthermore, there is a fragment of Egyptian blue and two fragments of amber. The latter may indicate that the production of glass and amber beads was closely linked. Two fragments of basalt may derive from millstones imported from the Eifel. It is argued that the stone tesserae and the Egyptian blue pebble could have been collected accidentally with glass tesserae during the frequent spoliation of lavishly decorated Roman buildings. The Egyptian blue and highly coloured vessel fragments may even indicate that the collection originates from one or more buildings that contained a combination of first century AD shell mosaics, early glass mosaics and glass tesserae mosaics, or transitional forms thereof. 3.14 Deventer-Stadhuiskwartier, Overijssel Province Site: 20 Site type: (Proto) urban settlement Province: Overijssel Municipality: Deventer Place: Deventer Toponym: Stadhuiskwartier Start date: c. 850 End date: c. 1200 (thereafter medieval city) Description: Deventer is situated in the east of the Netherlands, on the river IJssel, a tributary of the Rhine, which flowed into Lake Almere, now the IJsselmeer, which gave access to the Wadden Sea and the North Sea. The excavations in the inner city of Deventer, project 312 (2007–2009) and project 434 (2012– 2013) revealed multi-period occupation, 57 — including a late-mesolithic camp, lateprehistoric settlement traces and especially many remains of the medieval city and its early medieval predecessor.267 The earliest phase of the medieval settlement consists of several scattered buildings and a layer of arable land dating to the 8th and early 9th centuries. In the third quarter of the 9th century the land was reorganized on a large scale. The area was levelled and divided into new, regular plots. This development can be seen as the start of the process of urbanisation in Deventer. A new type of urban house was introduced which is clearly different from farmhouses in the surrounding agrarian settlements, e.g. they lack a stable. Many floor remains of this type of house were found, as well as a large number of cesspits, waste pits and wells. The finds indicate a large increase in craft activities in the late 9th century. At the end of the 9th century a defensive rampart was constructed around the settlement. From the 10th century onwards several timber houses with cellars were present as well as secondary buildings with cellars which probably had an artisanal function. From the late 9th and 10th centuries there is evidence for boneworking, iron-working (smithing slags) and textile production from different locations. Production waste from different crafts is found together in the same waste pits on the same plots. During the 10th and especially the 11th century large tuff (stone) houses appear. Initially, the tuff is sourced from old Roman building material, transported along the Rhine from the Roman fort at Xanten, Germany. In the late medieval period the area developed as the centre of the medieval city with a town hall and houses belonging to members of the urban elite. The glass finds include vessel glass, window glass and some glass beads. Among the vessels are fragments of very thinly blown funnel beakers which are mostly made of a wellpreserved, clear bluish-green glass. There are also fragments of thick-walled, curved vessels of heavily weathered glass. The window glass is mostly made of heavily weathered light glass that is greenish where it is possible to see the colour. Several quarries have preserved sides that were nibbled with a grozing iron in order to give them a distinct shape. The quarries would have been mounted in lead strips. Glass production waste: Glass production waste is scanty and dates between c. 850 and c. 1050 (unpublished data). All glass production waste products were found in waste pits or cesspits. Two pits dating between 900 and 925 (project 312, K60 and K74) and two pits dating to the first half of the 9th century (project 343, K116 and K174) also yielded production waste of smithing, bone-working and textile production. A hollow, glassy slag dates to 850–900 (434/16203) and is very similar to glassy slags found in a 10th century glass workshop in La Milesse (Sarthe, France) where wood ash glass was made from raw materials and blown into glass vessels.268 Three fragments date between 900 and 925. A heavily weathered chip of glass with a conchoidal fracture and with characteristic concentric ribs (312/29057) was struck off a larger chunk of raw glass. A heavily weathered fragment has one convex surface and is more or less triangular in section (312/29028). It may be a transverse breaking splinter struck off a glass ingot with at least one curved side. A small, heat-affected fragment (312/29028) may be part of a pulled thread, but this is not certain. Three glass production waste fragments date to the period 900–950. A small lump of translucent clear bluish-green raw glass (project 434/99144) has a conchoidal fracture and was struck off a larger chunk of raw glass. It is of a similar quality and colour to a funnel beaker fragment with optic blown oblique ribs (project 434/10380). A heavily weathered fragment with a triangular section (project 434/99154) is similar to fragment 312/29028 and may be a transverse breaking splinter. A heavily weathered fragment with two irregular, heataffected surfaces (project 434/99154) could be a partially melted chip of raw glass. An intriguing fragment (project 312/29048) dating to 950–1050 is difficult to interpret. It consists of two layers of translucent bright bluish-green and deep turquoise glass covered by a very thin film of opaque red glass. The fragment has two irregular surfaces which are heat-affected, probably as a result of being in a high-temperature glass workshop environment. The turquoise colour is very similar to that of a contemporary fragment of very thin flat glass, either window glass or a glass inlay with very fine grozing, from the same area (project 312/29063). A fragment of a deep turquoise quarry dating between 900–950 267 All information about the excavations has been kindly provided by Emile Mittendorff, Project leader Archaeology, Deventer. 268 Cf. Raux et al. 2015, Fig. 3F. 58 — 269 Isings 1957, 113–114. comes from another location (project 434 434/99139). A knocked-off rim of a late Roman yellowgreen cup of Isings type 96a269 (project 434/ 99116) dated to 900–950 may be cullet intended to be melted down for the production of vessel or possibly window glass. It is not unusual to find old glass among early medieval glass production waste (see for instance MaastrichtJodenstraat Section 3.2.1). An alternative explanation is that Roman glass was accidentally mixed in with Roman pottery (Samian ware) and tegulae fragments which are regularly found in Deventer in 10th–12th century contexts. It is thought that the tegulae and possibly (part of) the pottery had been transported to Deventer together with the Roman tuff that was reused to build stone houses. Part of the tegulae have stamps proving they were made in Xanten. There are as yet no indications for Roman habitation in Deventer. 59 — 4 The materials, analytical techniques and methodology 4.1 Introduction This chapter provides a brief introduction to the research samples studied and describes the analytical techniques used and methodology applied in the scientific analysis of the samples. The rationale of how the analytical data collected were used for the interpretation and comparison with previously published data is also explained. 4.2 An overview of the sites and glass samples A detailed description of the evidence for the Dutch early medieval glass-working is given in Chapter 3. One of the earliest sites providing glass for this project is Gennep in the province of Limburg. It is a 5th century AD Frankish settlement probably founded around 400 AD located on a high river terrace overlooking the confluence of the rivers Meuse and Niers. It is located between the late Roman fortress of Cuijk and the burgus of Asperden on the Niers to the east.270 It did not yield any evidence for glass working. Some 200 glass vessels were found at the site which were mainly table ware, mostly drinking vessels. The samples analysed were all typical Frankish glass vessels consisting of bowls and cones.271 They have provided critical compositional data for an early phase of the Dutch middle ages with which to compare other early medieval glasses. Excavations on nine Merovingian sites mostly on the west bank of the river Meuse in Maastricht has produced some of the most comprehensive evidence for Early Medieval glassworking yet found in Europe. The best evidence for a glass industry was found during excavations at the Jodenstraat (MAJO) site in Maastricht (see Section 3.2.1). Evidence of glass bead making, including 38 fragments of crucibles containing opaque yellow and white glass were found with more good evidence from the Mabro site in Maastricht (see Section 3.2.2). Crucibles containing glass were sampled along Table 4.1 Photographs of crucibles from Maastricht, Jodenstraat (MAJO) together with their sample numbers. Sample MAJO 1 Sample number Photo number 20 figure appendix IV.11 MAJO 2 21 figure appendix IV.12 MAJO 3 22 figure appendix IV.13 MAJO 4 23 figure appendix IV.14 MAJO 5 (inside) 24 figure appendix IV.15 MAJO 5 (outside) 24 figure appendix IV.16 MAJO 6 25 figure appendix IV.17 MAJO 7 26 figure appendix IV.18 MAJO 8 29 figure appendix IV.19 MAJO 9 30 figure appendix IV.20 MAJO 10 39-40 figure appendix IV.21 MAJO 11 41 figure appendix IV.22 MAJO 12 42 figure appendix IV.23 MAJO 13 43 figure appendix IV.24 MAJO 14 44 figure appendix IV.25 MAJO 15 45 figure appendix IV.26 MAJO 16 46 figure appendix IV.27 MAJO 17 47 figure appendix IV.28 MAJO 18 50 figure appendix IV.29 MAJO 19 51 figure appendix IV.30 MAJO 20 52 figure appendix IV.31 MAJO 21 53 figure appendix IV.32 MAJO 22 54 figure appendix IV.33 MAJO 23 58 figure appendix IV.34 MAJO 24 60 figure appendix IV.35 MAJO 25 61 figure appendix IV.36 MAJO 26 68 figure appendix IV.37 MAJO 27 73-74 figure appendix IV.38 with material from rods, bead fragments, splinters, drops and punty glass. Samples from both Jodenstraat and Mabro sites were analysed. Tables 4.1 and 4.2 are list of sampled crucibles from Jodenstraat and Mabro respectively, together with their photograph numbers provided here. 270 Brüggler 1994. 271 Sablerolles 1992; 1993. 60 — Table 4.2 Photographs of crucibles from Maastricht, Mabro together with their sample numbers. Sample Sample number Table 4.3 Photographs of crucibles from Utrecht, Domplein together with their sample numbers. Photo number Sample Sample number Photo number MABRO 1 7 figure appendix IV.1 DOM 1 31 figure appendix IV.39 MABRO 2 8 figure appendix IV.2 DOM 2 32 figure appendix IV.40 MABRO 3 9 figure appendix IV.3 DOM 3 33 figure appendix IV.41 MABRO 4 10 figure appendix IV.4 DOM 4 34 figure appendix IV.42 MABRO 5 11 figure appendix IV.5 DOM 5 35 figure appendix IV.43 MABRO 6 12 figure appendix IV.6 DOM 6 36 figure appendix IV.44 MABRO 7 13 figure appendix IV.7 MABRO 8 14 figure appendix IV.8 MABRO 9 15 figure appendix IV.9 MABRO 10 16 figure appendix IV.10 The Wijnaldum-Tjitsma (henceforth Wijnaldum) terp which has its heyday c. 550-650 AD when the area surrounding Wijnaldum, northern Westergo, had developed into the centre of a kingdom that covered the entire terp region of the northern Netherlands (see Section 3.12). Many beads, a tessera, vessel fragments, a rod and a possible furnace or thick tray fragment with opaque yellow glass adhering, were analysed. Archaeological investigations of the protourban site at Utrecht dating to c. 700-10th century AD also produced 17 crucible fragments with glass adhering (probably 8th-9th century Carolingian pots) (see Section 3.5.1). These were found at the Domplein site and sampled for this project (a list if the crucible samples together with photo numbers are given in Table 4.3). In addition, scanty evidence for glass working was found at the Utrecht Oudwijkerdwarsstraat site dating to the 7th- first half of the 8th century AD. This glass was also sampled (Table 4.4). Excavations at the Carolingian site of Susteren-Salvatorplein (henceforth Susteren), a monastic site, produced two crucible fragments along with polychrome beads, windows and vessels (see Section 3.3). Excavations of the famous emporium of Wijk bij Duurstede at the Hoogstraat and vicus sites (henceforth Dorestad) dating to between c. 600 and 900 AD produced a wide range of glass artefacts. Those selected for scientific analysis were mainly fragments of funnel beakers, bowls and bell beakers but also tesserae, linen smoothers. The latest site which produced glass included in this project was Deventer-Stadhuiskwartier (henceforth Deventer, see Section 3.14). The site dates to between c. 850 at the earliest and the 10th-11th centuries AD. Site excavations produced raw chunks of glass as well as vessel glass, glass beads and window glass. The quantitative major and minor chemical composition and trace element chemical composition have been determined for each glass sample studied in this work. Altogether 279 glass objects have been sampled in this project. Neodymium and strontium isotopic compositions were determined for 20 glass samples which were selected based on their chemical characteristics. The major, minor and trace element compositions of our samples constitute the primary data for this project. The compositional group and/or formula group for each sample have been identified using certain compositional and isotopic characteristics. The technical details of the analytical methods used to produce chemical and isotopic data are elaborated below. Table 4.4 Photographs of glassworking evidence from Utrecht, Oudwijkerdwarsstraat. Sample Sample number Photo number OUDWIJ 1 78 figure appendix IV.45 OUDWIJ 2 79 figure appendix IV.46 61 — 4.3 Electron probe microanalysis (EPMA) for major and minor chemical composition 4.4 Quantitative major and minor chemical compositions of our samples were determined on the JEOL JXA-8200 electron microprobe housed in Nanoscale and Microscale Research Centre, University of Nottingham. Fragments of each ceramic shard were mounted in crosssection in epoxy resin blocks and polished to a 0.25 mm diamond paste finish so as to reveal a fresh flat analytical surface. The blocks were carbon coated to prevent surface charging and distortion of the electron beam during analysis. The EPMA system is equipped with four wavelength-dispersive X-ray spectrometers with LIF, TAP, PETJ and LIFH crystals, a single energy dispersive X-ray spectrometer and both secondary and backscattered detectors. A defocused electron beam with a diameter of 40 µm was used so as to prevent volatilization of light elements such as sodium. The probe was run at an accelerating voltage of 15 kV and a beam current of 20 nA. The system was calibrated with a mixture of mineral and oxide standards. A ‘Phi-rho-z’ correction program was used to quantify the results. The Corning B glass standard was routinely used as a secondary standard to check for accuracy and precision and to monitor any drift in the instrument. The analytical precision and accuracy achieved by using the Corning B standard are listed in Table 4.5 Laser ablation inductively coupled plasma mass spectrometry (LAICP-MS) analysis for trace element compositions The trace elemental compositions of our glass samples were determined using the LA-ICP-MS instrument consisting of either a NewWave UP193FX excimer (193 nm) or Elemental Scientific Lasers imageGEO (193nm) laser system and an Agilent 7500cs series instrument housed in the Analytical Geochemical Laboratories of British Geological Survey. The same samples used by EPMA major and minor chemical analysis are analysed for their trace element compositions. Prior to analysis the carbon coating of the samples was removed and the samples were cleaned by rubbing a tissue soaked in dilute acid over the surface for a few seconds. The sample was placed in a two volume ablation cell with a 0.8 L min−1 He flow. In addition to the sample block, NIST glass standards SRM610 and 612 as well as USGS glasses standards GSD-1G and BCR-2G were placed in the chamber. The UP193FX laser was fired for 40s at 10 Hz using a beam diameter of 70 μm; whereas the imageGEO was fired at 20Hz or 10s using a square 50 X 50 μm beam. Fluence and irradiance as measured by the internal monitor were typically 3 J/cm2 and 0.85 GW/cm2 respectively for both laser systems. With the UP193FX laser prior to introduction into the ICP-MS the He flow was mixed, via a Y-junction, with 0.85 L min−1 Ar and 0.04 L min−1 N2 gas flows supplied by a Cetac Aridus desolvating nebulizer. The Aridus allowed introduction of ICP-MS tuning solutions and optimization of the Aridus sweep gas (nominal 4 L min−1 Ar). During solid analysis by the laser, the Aridus only aspirated air. The imageGEO system mixed the argon gas as above but added the N2 Table 4.5 The recommended composition for the Corning B standard compared to average analytical results (n = 44) and associated standard deviations and errors using the electron microprobe. SiO2 Al2O3 Na2O K2O CaO 4.36 17 Measured 62.43 4.65 16.83 1.04 8.75 1 0.17 0.6 0.09 0.26 1.4 6.7 1 4 2.2 12.4 Error % 8.56 FeO 61.55 Standard deviation (n=44) 1 TiO2 Quoted 0.089 MnO MgO 0.31 0.25 0.1 0.3 0.02 0.05 3.2 CoO CuO P2O5 Sb2O5 1.03 0.05 2.66 0.82 0.46 0.26 1.05 0.06 2.39 0.85 0.51 0.02 0.05 0.04 0.12 0.07 0.04 4 1.9 20 10.2 3.7 10.9 62 — gas flow internally. Tuning was by rastering the laser beam over the glass standards. The ICP-MS instrument was set for a dwell time of 7 ms for each of the 47 isotopes of interest to give one time-slice Data were collected in a continuous time resolved analysis (TRA) fashion as a repetitive series of time-slices. Prior to laser firing a period of at least 120 s of ‘gas blank’ was collected, then three ablations being made on the SRM610; three ablations on GSD-1G; 3 ablations on the SRM610; three ablations on the BCR-2G, 3three ablations on up to eight samples and finally three ablations on the SRM610; three ablations on GSD-1G. The SRM610 and GSD-1G were used to calibrate the system whilst the SRM612 and BCR-2G were used as a quality control (QC) materials. The full quality control report of our trace element analysis is listed in Table 4.6. Calibrations and data reduction were performed using Elemental Scientific Lasers Iolite4 software, with data compilation in Microscoft Excel 2016. The nature of laser ablation means that there is some variability in ablation volume and transport efficiency with different materials (matrix effects). Therefore, accepted practice is to normalize results to an internal standard element; in the current study Si was chosen for this purpose with its concentration being known in the NIST glasses and provided by the EPMA data for the study glasses. The 26 trace element pattern has been used in many recent publications to identify pristine natron glass of different compositional types. We have adopted this approach here in the discussion of our results below. 4.5 272 Birck 1986. Thermal ionization mass spectrometry (TIMS) analysis to determine Nd and Sr isotopic compositions For isotopic analysis, a small glass fragment was first sampled and transferred to a clean (class 100, laminar flow) working area for further preparation. In the clean laboratory, the samples were cleaned ultrasonically in Milli-Q water, dried on a hotplate and then weighed into pre-cleaned Teflon beakers. For Sr isotopic analysis, the samples were spiked with 84Sr tracer solution and dissolved in Teflon distilled 8M HNO3 and Ultrapure 29M HF. Samples were converted to chloride form using Teflon distilled 6M HCl. The samples were then taken up in calibrated 2.5M HCl and centrifuged. Strontium was collected using Eichrom AG50 X8 resin columns. Each sample was then loaded on to a single Re filament with TaF, following the method of Birck.272 The 87Sr/86Sr and Strontium elemental concentrations were determined by Thermal Ionization Mass spectroscopy (TIMS) using a Thermo Triton multi-collector mass spectrometer at the National Environmental Isotope Facility of the British Geological Survey. The international standard for 87Sr/86Sr, NBS987, loaded in the same way, gave a value of 0.710259 ± 0.000018 (n = 21, 2σ) during the analysis of these samples, and sample data was normalized to the accepted value for this standard of 0.710250. Procedural blank values were in the region of 100 pg. For Nd isotopic analysis, fractions were dissolved in 1 ml of 2% HNO3 prior to analysis on a Thermo-Electron Neptune mass spectrometer, using a Cetac Aridus II desolvating nebulizer. 0.010 L min-1 of nitrogen were introduced via the nebulizer in addition to argon in order to minimize oxide formation. The instrument was operated in static multi-collection mode, with cups set to monitor 142Ce, 143Nd, 144Nd, 145Nd, 146Nd, 147Sm, 149Sm, and 150Nd. 1% dilutions of each sample were tested prior to analysis, and samples diluted to c. 20 ppb. Jet sample cones and X-skimmer cones were used, giving a typical signal of c. 800–1000 V/ppm Nd. Correction for 144Sm on the 144Nd peak was made using a ratio for 147Sm/144Nd derived from multiple analyses of SpecPur© samarium solution. This correction was insignificant due to the efficiency of the column separation. Data are reported relative to 146Nd/144Nd = 0.7219. The Nd standard solution JND-i was analysed during each analytical session and sample 143Nd/144Nd ratios are reported relative to a value of 0.512115 for this standard. 63 — Table 4.6 Summary of quality control (QC) data for analysis of glass samples. Reference material: SRM612 Element Li Number of analyses=101 Measured isotope Nuber of analytical sessions=3 Expected concentration (mg/kg) 7 39.5 Mean concentration (mg/kg) Standard deviation 40.2 RSD% Error% 1.5 3.8 -2 B 11 37.3 34.3 2.7 7.2 9 Na 23 99780 103858 1867 2 -4 Mg 24 58 68 4 7 -15 Al 27 11102 11167 295 3 -1 P 28 73 46.6 152 208 56 K 31 53 62.3 4 8 -15 Ca 39 84382 85002 2298 3 -1 Ti 42 39.8 44 4 9.9 -9 V 47 37.8 38.8 1.5 3.9 -3 Cr 51 34.8 36.4 1.7 4.8 -4 Mn 52 38 38.7 1 4 -3 Fe 55 46 51 3 7 -11 Co 56 34.3 35.5 0.9 2.7 -3 Ni 59 37.7 38.8 1.4 3.7 -3 Cu 60 36.7 37.8 1.2 3.3 -3 Zn 63 37.6 39.1 2.2 6 -4 As 66 33.2 35.7 2.2 6.5 -7 Rb 75 31.3 31.4 0.8 2.5 0 Sr 85 76.9 78.4 2.9 3.7 -2 Y 88 38.4 38.3 1.1 3 0 Zr 89 38.5 37.9 1.1 2.9 2 Nb 90 38.3 38.9 1.2 3.2 -2 Mo 93 35.6 37.4 1.6 4.6 -5 Sn 95 37.6 38.6 1.6 4.1 -3 Sb 120 33.6 34.7 1 3.1 -3 Cs 121 41.2 42.7 1.1 2.6 -3 Ba 133 38.7 39.3 1 2.5 -1 La 138 35.7 36 0.9 2.6 -1 Ce 139 37.7 38.4 1.2 3.3 -2 Pr 140 37.3 37.9 1.2 3.3 -2 Nd 141 34.9 35.5 1.4 4 -2 Sm 146 37.3 37.7 1.5 4 -1 Eu 147 35 35.6 1.3 3.6 -2 Gd 153 37.7 37.3 1.4 3.7 1 Tb 157 36.6 37.6 0.9 2.6 -3 Dy 159 35.6 35.5 1.1 3.1 0 Ho 163 37.8 38.3 1 2.6 -1 Er 165 38.4 38 1 2.7 1 Tm 166 36.6 36.8 1.1 2.9 -1 Yb 169 38 39.2 1.2 3.2 -3 Lu 172 36.6 37 0.9 2.5 -1 Hf 175 37 36.7 1.2 3.4 1 Ta 178 36.9 37.6 1 2.6 -2 Pb 208 38.2 38.6 1.1 2.9 -1 Th 232 37.5 37.8 1.2 3.1 -1 U 238 36 37.4 1.2 3.4 -4 64 — 4.6 How analytical data is used in this study The type of glass whether natron, plant ash or wood ash, can be easily identified by major and minor oxide contents such as Na2O, K2O, CaO and MgO. The majority of the glass studied here is natron glass; wood ash glass and plant ash glass only account for a small fraction of the samples. One of the main aims of this study is to categorize the majority which are natron glasses according to different compositional types (see Section 2.4.1). These are related to their provenance, so we can gain an insight into raw glass supply in the early medieval Netherlands. Because a large number of samples are involved in this study, the job of categorizing natron glass samples into compositional groups has mainly been achieved by using three plots, Al2O3/SiO2 against TiO2/Al2O3, Pb against Sb, and a 26 trace element pattern. Firstly the Al2O3/SiO2 against TiO2/Al2O3 plot is used to provide a preliminarily classification of natron glasses into compositional groups: TiO2, Al2O3 and SiO2 essentially represent the heavy mineral, feldspar and quartz contents of the sands used for making the glass,273 which can reflect their provenance very well (Figure 4.1).274 The Pb against Sb plot is used to show the levels of impurities brought in by recycling of the glass samples. In ‘pristine’ (non-recycled) natron glass the levels of a few correlated elements such as Pb, Sb and Cu are very low, but for recycled glass the levels of these elements are much higher. Pb and Sb (both in ppm) are the most consistent demonstrators among these elements, so they have been chosen to distinguish ‘pristine’ glass samples from recycled glass samples. The criterion for the identification of a ‘pristine’ glass is that the Pb and Sb contents are both under 1000 ppm, following previous conventions.275 The compositional groups of ‘pristine’ glass samples identified using the two previously mentioned plots are then confirmed by using the 26 trace element patterns of the samples. Although it has been found that the rare earth element patterns of all natron glasses tend to be very similar, when lighter trace elements (excluding some elements which may have been introduced with the colourant, such as transition metals) are included, the patterns of the four main different compositional groups, HIMT sensu stricto, Foy 2, Egyptian II and Levantine,276 can be distinguished very well (Figure 4.2).277 The 26 trace elements used in this study are V, Cr, Rb, Sr, Y, Zr, Nb, Cs, Ba, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, Hf, Th, U, and their normalized concentrations compared to that for the upper continental crust.278. 273 Freestone et al. 2018. 274 Data sources: Foy et al. 2003 (HIMT 275 276 277 278 sensu stricto, Levantine I), Schibille, Sterrett-Krause & Freestone 2016 (Foy 2), Schibille et al. 2019 (Egyptian I and Egyptian II), Freestone et al. 2015 (Levantine II), Silvestri, Molin & Salviulo 2008 (Roman Mn and Roman Sb). Foster & Jackson 2009. The trace element patterns of Levantine I glass and Levantine II glass are very similar. Thus only the pattern of Levantine I glass is shown here to demonstrate its difference with that of other compositional groups. It is difficult to distinguish Egyptian I glass and HIMT sensu stricto glass by their trace element patterns. This it is not too much of a problem here since Egyptian I glass is not a significant compositional group in northwestern Europe. Schibille, Sterrett-Krause & Freestone 2016; Bertini, Henderson & Chenery 2020. Kamber et al. 2005. Figure 4.1 Al2O3/SiO2 against TiO2/Al2O3 plot showing the compositional differences between major groups of natron glasses. 65 — Figure 4.2 The 26 trace element patterns of four major compositional groups of natron glass. The trace elements are V, Cr, Rb, Sr, Y, Zr, Nb, Cs, Ba, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, Hf, Th and U 5 The analytical results and discussion 5.1 Introduction Glass samples from nine sites were studied, including two sites in both Maastricht and Utrecht. In this chapter the chemical compositional features of glass samples from each site will be elaborated separately. The major and minor chemical compositions of analysed samples are given in Appendix II. The trace element data of analysed samples are given in Appendix III. 5.2 Glass samples from Maastricht (Jodenstraat and Mabro sites) Apart from two glass vessel fragments and three window glass fragments, which may be glass cullet, all the rest of the 51 glass samples from Jodenstraat are the remains of on-site bead making, and they all came from one pit filled between late 6th century and early 7th century AD.279 The glass samples from the bead-making context can be roughly divided into three groups: translucent naturally coloured and cobalt blue glass waste, highly coloured opaque glass, and vitreous materials attached to crucibles (Appendix I). 67 — 5.2.1 Naturally coloured and cobalt blue bead-making glass waste The bead-making glass waste from Jodenstraat includes glass drops, glass rods, punty glass, glass fragments and glass attached to crucibles. Because of the shape and state of the glass samples found in the bead-making context, and the similarity of their chemical compositions with that of the glass beads, it has been suggested that bead making involved firstly colouration of the naturally coloured base glass followed by further procedures to make the glass into beads.280 If this suggestion stands, the bead-making glass waste should reflect the chemical features of the base glass used by the bead makers. Their low K2O and MgO and elevated Fe2O3, TiO2 and MnO contents suggest that they can all be categorized as the so-called HIMT natron glass dominating northwest Europe during the 4th–7th century. HIMT glass is a general name for early medieval natron glass with elevated Fe2O3, TiO2 and MnO contents (see Section 2.4.1). There are some further compositional groupings that can be demonstrated by the ratios of TiO2, Al2O3 and SiO2, which represent essentially the heavy mineral, feldspar and quartz contents of the sands used for making the glass. In Figure 5.1 we can see that the naturally coloured and blue glass waste from Jodenstraat forms one tight group, and it agrees very well with the well understood Foy 2 compositional group of natron glass (compare with Figure 4.1), including one of the crucible fragments with pale green glass attached (Joden 29) which is least Figure 5.1 A plot of Al2O3/SiO2 versus TiO2/Al2O3 for Jodenstraat naturally coloured bead-making waste and highly coloured opaque glass and beads. 279 Sablerolles, Henderson & Dijkman 1997. 280 Sablerolles, Henderson & Dijkman 1997. 68 — Figure 5.2 The 26 trace element patterns for Jodenstraat naturally coloured and blue bead-making waste and Jodenstraat highly coloured opaque glass and beads compared to that of Foy 2 glass. contaminated. The latter shows no evidence that lead or tin had been added (yet). The slightly elevated PbO contents of some naturally coloured glass waste may have been caused by contamination during the production process since lead was an important raw material ingredient in bead making of this period in northwestern Europe. The elevated PbO content of early medieval glass could also indicate glass recycling.281 However, the Sb contents, which are correlated with PbO in recycled early medieval glass, and often used together with PbO as an indicator of glass recycling, are all very low at ≤1000 ppm level. 281 Foster & Jackson 2009. Therefore we are more inclined to believe that this glass waste is ‘pristine’ glass rather than recycled, and that the elevated PbO was introduced during the production procedure. The 26 trace element pattern in the naturally coloured and blue glass waste from Jodenstraat agrees very well with that for Foy 2 glass published previously (Figure 5.2) confirming that they are Foy 2 glass. Amongst the bead making waste there is a group of five blue samples: four blue fragments (Joden 37-40) and one translucent blue coloured bead fragment (Joden 61). In the Al2O3/SiO2 against TiO2/Al2O3 plot, these five samples locate in the area of Foy 3.2, one of the two subgroups of Foy 2, the other subgroup being Foy 2.1 (see Section 2.4.1). The flat tapered bead fragment (Joden 61) does not contain elevated PbO, nor is it opacified with SnO2 like the many other highly coloured beads found at Jodenstraat. The four cobalt blue fragments (Joden 37-40) and the cobalt blue bead (Joden 61) have the same composition which suggests this type of non-tin opacified blue beads was made directly from ‘pristine’ cobalt blue coloured raw glass. 5.2.2 Highly coloured opaque glass The highly coloured glasses found at Jodenstraat are beads, rods and drops, potentially all linked to the process of bead making. These highly coloured glass samples are in four basic colours: yellow, white, red and greenish-blue. They all have elevated PbO contents ranging from a little over 2.0 wt% to over 40 wt% PbO, which demonstrates that lead was used as one of the key raw materials in bead making in addition to natron glass. As shown in figure 5.3, the average PbO contents in these yellow, white, red and greenish-blue glasses are not quite the same especially for yellow glass: its PbO contents are much higher than that of the other three colours. The reason for the different PbO concentrations in yellow glass and in other colours is addressed in detail in Section 5.10.3 below. Figure 5.3 The PbO contents of highly coloured opaque glasses from Jodenstraat (G blue = greenish-blue). 69 — a b c d e f Figure 5.4 Backscattered SEM images of yellow (top left: 5.4a), green (top right: 5.4b), white (middle left: 5.4c), red (middle right: 5.4d) coloured glass from Jodenstraat and crystalline inclusions: SnO2 opacifiers found in white, green and red colour glass (bottom left: 5.4.e), ground fayalitic slag surrounded by 0 valence micron sized particles in red glass (bottom right: 5.4f ). From the backscattered SEM images of the highly coloured opaque glass samples from Jodenstraat (Figure 5.4), we can see that they all contain some highlighted (pale grey) inclusions of variable sizes. The quantitative EPMA analyses of these phases show that they differ compositionally. The typical chemical compositions of the three crystalline phases of different highly coloured glasses are listed in Table 5.1. According to their quantitative chemical compositions and backscattered SEM images, the crystalline phases in yellow glass can be identified as lead tin yellow II (PbSn(Si)O3) with a PbO/(SnO+SiO2) ratio of about 2:1 (Figure 5.4). Lead tin yellow II was widely used as the colourant in yellow coloured beads in early medieval Europe.282 Three types of crystalline inclusions are found in the glass matrix of the red coloured glass (Figure 5.4): firstly a phase containing SnO2 of 50–70 wt% and variable SiO2 and PbO contents; secondly, a high iron phase, which was also observed in similar medieval red colour beads from England and identified as ground fayalitic slag;283 finally, micron sized particles of 0 valence metallic copper, which are scattered evenly in the glass matrix. 282 Heck, Rehren & Hoffmann 2003. 283 Peake & Freestone 2012. 70 — Table 5.1 The typical chemical compositions of the three crystalline phases found in Jodenstraat highly coloured glasses. SiO2 (wt%) 284 Bandiera et al. 2020. Lead tin yellow II SnO2 opacifiers Fayalitic slag 8–18 15–30 10–25 Al2O3 (wt%) 1–3 2–3 <1 Na2O (wt%) 3–5 5–8 1–5 K2O (wt%) - - CaO (wt%) - 1 - SnO2 (wt%) 15–24 50–70 - PbO (wt%) 60–65 2–15 <1 FeO (wt%) - - 65–90 According to previous studies we can conclude that the phase containing SnO2, the dominant composition, acts as the opacifier in the glass matrix. The red glass is mainly coloured by micron sized particles of 0 valence copper, like many other types of red glass and ceramic glazes,284 and the ground fayalitic slag would have acted as an internal reducing agent during the formation of the copper particles. The crystalline inclusions found in the white glass and the greenish-blue glass beads are the same (Figure 5.4), and they also contain the first type of inclusion found in the red glass, mainly SnO2 at 50–70%, with variable SiO2 and PbO % levels, acting as the opacifier in the glass matrix. The amount of SnO2 found in the white glass is much higher than that found in the red and greenish-blue glass, producing its opacified white colour. The colour of the greenish-blue glass is caused by a copper based colourant (CuO), as indicated by high copper contents, which would have dissolved in the glass matrix so it cannot be observed as a separate phase in the SEM images. Jodenstraat are also very low at ≤1000 ppm, the same as found in naturally coloured and blue bead-making waste found on the site. These results indicate that the base glass of Jodenstraat highly coloured opaque glass is also ‘pristine’ Foy 2 glass and that the bead-making debris formed during the production process was produced using this same base glass. The observation that Jodenstraat highly coloured opaque glasses generally have higher Al2O3 contents than the bead-making waste, demonstrated by their clustering to the right of colourless bead-making waste in Al2O3/SiO2 against TiO2/Al2O3 plot (Figure 5.1), can be attributed to the addition of the lead tin yellow II colourant and tin opacifiers during the colouring process: the lead tin yellow II colourant and tin opacifiers have higher Al2O3/ SiO2 than the base glass (Table 5.1). In the plot of Al2O3/SiO2 against TiO2/Al2O3 (Figure 5.1), Jodenstraat highly coloured opaque glass samples form a tight cluster in the area of the Foy 2 compositional group (with Figure 4.1 as reference), located slightly to the right of the cluster of Jodenstraat colourless bead-making waste. In Fig. 5.2 it can be seen that the 26 trace element pattern of the averaged composition of these highly coloured glass samples is identical with that of Jodenstraat colourless bead-making waste and the Foy 2 pattern published previously. Additionally, the Sb contents of all highly coloured opaque glass samples from From twelve crucibles retrieved from the beadmaking context of Jodenstraat, three types of vitreous materials attached to them have been examined scientifically. They are naturally coloured natron glass (two samples), naturally coloured glass mixed with bright yellow residues (eight samples), and a white melt with a light yellow tinge (two samples). The description of the vitreous residues attached to each crucible is listed in Table 5.2. Naturally coloured natron glass has been found attached to crucibles Joden 21 and Joden 29 from the bead-making context of 5.2.3 Vitreous and semi-vitreous materials attached to the crucibles 71 — Jodenstraat, and their chemical compositions have been discussed along with other naturally coloured glass in Section 5.2.1: they conform to a Foy 2 composition. The naturally coloured glasses mixed with bright yellow residues in 8 crucibles turned out to be very pure lead oxide-silica glass. The Na2O concentrations are very low in these lead glasses (Appendix II), which shows that natron glass was not involved in the procedure that produced these lead glasses and the yellow/white residue mixture. Yellow residues attached to crucibles from early medieval northwest European sites have been studied before285 and the analytical results for yellow residues attached to crucibles from Jodenstraat are the same, namely lead tin oxide. Chemical compositions and SEM images show that the main phase of the bright yellow lead tin residue is lead tin yellow II (PbSn(Si)O3) (Figure 5.5a), where the (SnO+SiO2) to PbO weight ratio is close to 1:2. It has been suggested that these crucibles containing lead tin yellow residues are evidence of on-site production of the yellow colourant which was then added to base glass during the manufacture of yellow beads.286 The chemical compositions and SEM morphologies of the lead tin yellow II crystallites attached to the Jodenstraat crucibles are very similar to the lead tin yellow II found in the yellow glass beads and yellow bead-making debris from Jodenstraat. Therefore we also suggest that these crucibles are remains of onsite lead tin yellow colourant production, and that the lead tin yellow II produced in the crucibles would have been used directly to colour the base glass to create a yellow colour. More details regarding the procedures of on-site lead tin yellow II production and how the crucibles were used during the process are addressed in Section 5.10.2 below. A white melt with a light yellow tinge has been found attached to two crucibles (Joden 23 and Joden 30). Their chemical compositions show that this white melt also contains SnO2, PbO and SiO2 as the main components, but that their weight ratios are quite different from that of lead tin yellow II (Appendix II). SnO2 is the dominant component, ranging from 50 wt% to 70 wt% in different areas of the white melt; the PbO and SiO2 contents are variable. The SEM morphology of the tin white crystallites is also quite different from that of lead tin yellow in that no lead silica glass surrounds the tin oxide in the former whereas it does in the latter (Figure 5.5b). a b Table 5.2 Description of the vitreous residues attached to Jodenstraat crucibles Sample number Vitreous residue description Maastricht-Jodenstraat 19 lead yellow ii surrounded by pure lead glass Maastricht-Jodenstraat 20 lead yellow ii surrounded by pure lead glass Maastricht-Jodenstraat 21 natron glass Maastricht-Jodenstraat 22 pure lead glass with small yellow spots Maastricht-Jodenstraat 23 white melt Maastricht-Jodenstraat 24 pure lead glass with small yellow spots Maastricht-Jodenstraat 25 pure lead glass with small yellow spots Maastricht-Jodenstraat 26 pure lead glass with small yellow spots Maastricht-Jodenstraat 27 lead yellow ii surrounded by pure lead glass Maastricht-Jodenstraat 28 lead yellow ii surrounded by pure lead glass Maastricht-Jodenstraat 29 natron glass Maastricht-Jodenstraat 30 white melt Figure 5.5 Backscattered images of yellow residue (left: 5.5a) and white melt (right: 5.5b) found attached to crucibles from Jodenstraat. 285 Henderson & Ivans 1992; Heck, Rehren & Hoffmann 2003; Peake & Freestone 2014. 286 Heck, Rehren & Hoffmann 2003; Peake and Freestone 2014. 72 — Moreover, the chemical compositions and SEM morphologies of this tin white melt are very similar to that of the tin opacifiers in white, red and greenish-blue beads. Therefore, we think these white melts attached to crucibles are evidence for on-site production of tin opacifiers in early medieval northwestern Europe, and they constitute the first such evidence to be reported. More details of the separate production of the tin white opacifier at Jodenstraat is addressed in Section 5.10.3 below. un-recycled (‘pristine’) glass. The 26 trace element patterns of Joden 45 and Joden 46 confirm that they have the same patterns as Foy 2 and the HIMT sensu stricto respectively published previously, but the trace element pattern of Joden 44 shows some clear differences from that of the Foy 2 glass pattern which may be related to the unusually high MnO content (1.9%) in the sample (Figure 5.7a). 5.2.5 Crucibles from the Mabro site, Maastricht 5.2.4 Glass artefacts 287 Silvestri, Molin & Salviulo 2008. 288 Pactat et al. 2017. Five glass samples from Jodenstraat, three window glass fragments (Joden 44–46) and two glass vessel fragments (Joden 60, 68), may be cullet. The elevated MnO, Fe2O3 and TiO2 contents suggest that they are HIMT natron glass. Their Al2O3/SiO2 and TiO2/Al2O3 ratios show that the window glass (Joden 46) belongs to the HIMT sensu stricto compositional group, samples 44, 45 and 60 distribute in the area of the Foy 2 compositional group and sample 68 could be a piece of Roman glass (Figure 5.6a with Figure 4.1 as reference). The form of Joden 68 also suggests that it is a Roman vessel fragment (thick naturally coloured ribbed green glass), but it does not have high Mn or high Sb contents normally found in decolourized and green Roman glasses.287 The Sb and Pb contents of the five glass samples show that apart from Joden 60, which has Sb and Pb contents over 1000 ppm, the balance have low Sb and Pb contents <1000 ppm (Figure 5.6b). Therefore, this result shows that Joden 60 was made using recycled glass while the other four samples were made from Ten crucibles with vitreous residues attached from the Mabro site in Maastricht, which is located very close to Jodenstraat, were also analysed here. Lead tin yellow II surrounded by pure lead silica glass has been found in two crucibles, Mabro 12 and Mabro 14. ‘Mixed alkali’ glass was found attached to crucible Mabro 7, a high potassium oxide glass containing grey mainly angular unmelted silica grains was found attached to crucible Mabro 9 (Figure 5.8), though both are contaminated with 7.2 and 7.9% aluminium oxide, and high iron and titanium oxide. In Figure 5.8 the body of the glass is the homogenous pale grey layer on the right-hand side. The glass in crucible 9 contains 13% K2O but low levels of MgO and P2O5 so it is unlikely to be evidence for working wood ash glass. Moreover, it contains 7.54% Al2O3 and 3.49% Na2O. The presence of angular unmelted SiO2 grains and high Al2O3 suggests that what remains is the interaction layer with the crucible fabric. The mixed alkali glass in crucible 5 contains levels of MgO and phosphorus pentoxide that are significantly lower than in typical mixed Figure 5.6 Plots of Al2O3/SiO2 against TiO2/Al2O3 (a) and Sb against Pb (b) for Jodenstraat glass artefacts. 73 — a b Figure 5.7 The 26 trace element patterns for Joden 44–45 compared to that of relevant natron glass types published previously (top: 5.7a) and the same for Joden 46 (bottom: 5.7b) (Joden= Jodenstraat, Maastricht). Figure 5.8 Backscattered SEM image of glassy residue attached to crucible Mabro 9. alkali glasses, such as those from Méru in France.288 It is therefore more likely to be a contaminated natron glass. Soda glass with an unusual chemical composition was found attached to five crucibles: Mabro 8 and Mabro 11 contain high Al2O3 contents at 9.87% and 12.48% respectively as well as high iron and titanium interpreted as contamination by interaction with the crucible fabric. They may therefore be contaminated natron glasses. Mabro 13, Mabro 15 and Mabro 16 have higher CaO contents at 6.74%, 7.52% and 5.82%, approaching normal levels for natron glass. Although Mabro 13 and 15 contain relatively high Cl levels at 0.52% and 0.38% they contain normal levels of alumina, iron and titanium. All three contain elevated K2O levels of up to 2.6% but these are not paired with elevated magnesia. The elevated K2O levels may be due to contamination. No vitreous phase was found in Mabro 10. Analysis of the ‘frit-like’ material on the rim of crucible 11 dating the late 4th-early 5th century 74 — shows that it is a fuel ash slag with no detected CaO, high Cl (0.32%), high K2O (4.56%) but relatively low MgO (1.15%), high Fe2O3 (2.9%) and high MnO (0.98%). Paynter has shown that melting a natron glass in wood fired furnace can lead to fuel ash slags and contaminated glass of highly variable compositions.289 5.3 289 Paynter 2008. 290 The trace element patterns of HIMT glass and Egyptian I glass are very similar, so we are unable to further categorize the two samples which plot in the overlapping area of HIMT glass and Egyptian I glass, Ge 44 and Ge 45, into a more specific compositional group using trace elements. 291 Freestone 2015. Glass samples from Gennep The 28 vessel glass fragments from Gennep are the earliest glass studied here: they are tightly dated to between the late 4th century and mid-6th century AD. Their low K2O and MgO contents and elevated MnO, TiO2 and Fe2O3 contents suggest that they are all a type of HIMT natron glass. The TiO2/Al2O3 and Al2O3/SiO2 ratios of Gennep glass show that seven of them locate in the area of the HIMT sensu stricto compositional group (Ge 44 and Ge 45 fall in the overlapping area of HIMT sensu stricto and Egyptian I compositional groups) and 21 of them cluster closely in the area of the Foy 2 compositional group (Figure 5.9a with Figure 4.1 as reference). Pb and Sb concentrations are both <1000 ppm and are used here as the criterion to provide a preliminary distinction between ‘pristine’ glass from recycled glass. According to this criterion, 12 out of the total 28 samples can be regarded as ‘pristine’ glass: six of them belong to the HIMT sensu stricto group and the other six belong to the Foy 2 group (Figure 5.9b). The 26 trace element patterns of the six ‘pristine’ HIMT sensu stricto glass samples are essentially identical and their average pattern is very similar to that of HIMT glass reported previously thus confirming their HIMT Figure 5.9 Plots of Al2O3/SiO2 against TiO2/Al2O3 (a) and Sb against Pb (b) for Gennep glass vessels. compositional type (Figure 5.10a).290 The 26 trace element pattern of the average composition of the six ‘pristine’ Foy 2 glass is also identical with the Foy 2 glass pattern reported previously (Figure 5.10b). It has been suggested that the recycling of Roman tesserae and coloured vessel glass to supplement the dwindling supply of natron glass in northwestern Europe started approximately in the early 8th century, which is signified by the elevated Sb and Pb contents of glass in this period.291 However, our Gennep vessel glass samples are securely dated to late 4th to mid-6th century AD, and we already see elevated Sb and Pb contents for the majority of them (17 out of 29). 5.4 Glass samples from Wijnaldum Apart from two pieces of evidence for bead production, one greenish-white glass rod splinter (WIJ 41) and one turquoise punty glass (WIJ 42), the rest of the 40 glass samples from Wijnaldum can be separated into three groups: highly coloured opaque glass beads, colourless glass beads (four out of five were metal foil and the other one was trail-decorated) and glass vessels. 5.4.1 Highly coloured opaque glass beads The highly coloured glass beads are in three basic colours: opaque yellow, opaque white and opaque red. Wijnaldum beads share a similar 75 — a b Figure 5.10 The average 26 trace element patterns of Gennep ‘pristine’ HIMT sensu stricto glass and ‘pristine’ Foy 2 glass compared to that of relevant natron glass types published previously (top: 5.10a) and the same for Gennep ‘pristine’ Foy 2 (bottom: 5.10b). date to Maastricht Jodenstraat beads, and the chemical compositions of glass beads are also very similar. First of all, the highly coloured glass beads from Wijnaldum all have elevated PbO contents: 23–56% in yellow beads, 3–23% in red beads and 1–5% in white beads. Secondly, they share the same compositional feature of low Sb contents as found in Jodenstraat opaque beads: all highly coloured beads from Wijnaldum have a Sb content of <1000 ppm. Moreover, the colouring mechanisms of Wijnaldum beads are the same as for Jodenstraat beads. The Wijnaldum yellow beads are coloured by lead tin yellow II; the Wijnaldum white beads and red beads are opacified by tin oxide; and the Wijnaldum red beads are coloured by micron sized copper particles with iron-rich fayalitic slag acting as an internal reducing agent (Figure 5.11). The glass working tray covered with a contaminated opaque yellow vitreous layer was analysed previously (see section 2.4.2). 5.4.2 Colourless glass beads The five colourless glass beads are a special group (WIJ 35–39), not seen in Maastricht. Unlike highly coloured opaque glass beads, they do not have elevated PbO contents. Since their Sb and Pb contents are all under 1000 ppm, it suggests that ‘pristine’ rather than recycled glass was used to make them. Among the five colourless glass beads, WIJ 35 (a gold foil bead) and WIJ 37 76 — a b c d Figure 5.11 The backscattered images of a Wijnaldum yellow bead (top left: 5.11a), white bead (top right: 5.11b), red bead (middle left: 5.11c), tin opacifiers found in white beads and red beads (middle right: 5.11d) and fayalitic slag surrounded by 0 valence copper micron particles found in red beads (bottom: 5.11e). e (a silver foil bead) have completely different chemical compositions from glass vessels and other glass beads from Wijnaldum. Their lower Na2O and CaO contents and higher K2O and MgO contents compared to natron glass indicate that they are plant ash glasses, they date to between 775 and 900 AD according to their context dates and this correlates with the introduction of plant ash glasses in western Asia by Islamic glassmakers. The low MgO and K2O contents and elevated Fe2O3, TiO2 and MnO contents of the other three colourless glass beads (a gold foil bead and a silver foil bead dating to 450-550, and a colourless bead with red streaks dated to 750850) suggest they were made from a type of Figure 5.12 Plots of Al2O3/SiO2 against TiO2/Al2O3 (a) and Sb against Pb (b) for Wijnaldum glass samples. 77 — a b c Figure 5.13 The 26 trace element pattern for colourless Wij 36, 38 and 39 compared to that of relevant Foy 2 glass published previously (top: 5.13a), for Wij 10 and 15 compared to that of relevant HIMT and Egyptian II glasses published previously (middle: 5.13b) and Wij 42 compared to that of relevant Levantine II glasses published previously (bottom: 5.13c) (Wij = Wijnaldum). HIMT natron glass; their Al2O3/SiO2 and TiO2/ Al2O3 ratios show that they belong to the Foy 2 compositional group (Figure 5.12a with Figure 4.1 as reference). The Sb and Pb contents of the three Foy 2 beads are all very low at <200 ppm, so this suggests that they were made from ‘pristine’ Foy 2 glass rather than from recycled glass (Figure 5.12b) and two are early examples of this kind of glass. The 26 trace element pattern of the average composition of the three beads confirms this suggestion (Figure 5.13a). 5.4.3 Vessel glass The nine vessel glass samples from Wijnaldum are dated between the mid-5th century and late 9th century. The low MgO and K2O contents of these samples show that they are all made from natron glass. Their Al2O3/SiO2 and TiO2/Al2O3 ratios show that WIJ 10 and WIJ 15 plot at the overlapping zone of HIMT and Egyptian II compositional groups while the other seven 78 — samples belong to the Foy 2 compositional group (Figure 5.12a with Figure 4.1 as reference). Their Pb and Sb levels are both <1000 ppm and can be used as a criterion to distinguish pristine glass from recycled natron glass in a preliminary way. We can see that WIJ 10 and WIJ 15 can both be regarded as ‘pristine’ glass, while all Foy 2 glass samples are recycled glass (Figure 5.12b). The 26 trace element patterns for WIJ 10 and WIJ 15 resolve the inconclusive identification of the two samples using major chemical compositions which suggested that they are Egyptian II glass or HIMT sensu stricto glass. The trace element patterns of WIJ 10 and WIJ 15 show a close resemblance to that of Egyptian II glass with clear compositional differences from that of HIMT sensu stricto glass (Figure 5.13 middle). Therefore we can confirm that WIJ 10, a blue green funnel beaker, and WIJ 15, a dark blue funnel beaker with an incalmo rim, were made from ‘pristine’ Egyptian II glass. The dates of WIJ 10 (800–850 AD) and WIJ 15 (770–900 AD) are also consistent with the suggested dates for when Egyptian II glass was in circulatation in the 8th–9th centuries AD, a time when the supply of pristine Levantine glass was drying up.292 5.4.4 Bead production materials 292 P helps et al. 2016. The high Na2O content and low K2O and MgO contents in the greenish-white rod splinter (WIJ 41) and turquoise punty glass (WIJ 42) suggest that they are both natron glass. However, their different Fe2O3 and TiO2 contents and Al2O3/SiO2, TiO2/Al2O3 ratios suggest that they do not belong to the same compositional group. WIJ 42 has quite low Fe2O3 and TiO2 contents of 0.45% and 0.08% respectively, a feature that differs from the dominating HIMT natron glass discussed here, and the Al2O3/SiO2 and TiO2/Al2O3 ratios suggest it can be categorized as having a Levantine II (8th–9th century) natron glass composition (Figure 5.12a with Figure 4.1 as reference); the 26 trace element pattern for WIJ 42 confirms this identification (Figure 5.13c). The established date for this compositional group agrees with the date provided from the archaeological context (750–800 AD). The high CuO content in WIJ 42 shows that the turquoise colour was caused by copper. The greenish-white glass rod splinter WIJ 41 has elevated TiO2, Fe2O3 and MnO contents like most of the natron glass studied in this work. The Al2O3/SiO2 and TiO2/ Al2O3 ratios of WIJ 41 plot at the overlapping area of HIMT sensu stricto and the Egyptian II compositional groups (Figure 5.12a with Figure 4.1 as reference). The elevated Sb and Pb contents of WIJ 41, with 1650 ppm and 4933 ppm respectively, suggest that it was made from recycled glass (Figure 5.12b). 5.5 Glass samples from Utrecht Nine samples from two sites in Utrecht were analysed. Three glass fragments derive from Utrecht Oudwijkerdwarsstraat which dates to the 7th to mid-8th century AD. Six crucible samples (Utr 31–36) are from Utrecht Domplein which dates to between the mid-8th to late 9th century AD. The three glass fragments from Utrecht Oudwijkerdwarsstraat (Utr 77–79) are soda lime glasses. Utr 77, a fragment with a yellow tinge and Utr 79, a piece of green debris from glass working are natron glasses. Utr 78 has a modern composition and therefore will not be discussed further. In the plot of Al2O3/SiO2 against TiO2/Al2O3 (Figure 5.14a with Figure 4.1 as reference), Utr 77 plots in the area of the Foy 2 compositional group, while Utr 79 plots in the area of Roman glass. Utr 79 contains high CaO and slightly high Cr2O3. The Sb content of Utr 77 is low at 70 ppm, but its Pb content is high at 4034 ppm, suggesting it is a recycled glass. Both Sb and Pb contents of Utr 79 are low at <1 ppm and 92 ppm respectively, consistent with the Sb and Pb contents of non-Sb-decolourized Roman glass (Figure 5.14 right). Glass attached to six crucible fragments from Utrecht Domplein was also investigated. There is a thin very pale green, appearing colourless, glass layer attached to crucibles Utr 31 and Utr 32. There is evidence from the chemical compositions that the glass had interacted with the body of the crucible: they contain 12.6% and 4.65% Al2O3 respectively. Utr 31 especially has levels of Fe2O3 (3.17%) and TiO2 (0.64%) very likely to be the result of these elements migrating into the glass from the crucible fabric at high temperatures. Utr 31 was probably 79 — Figure 5.14 Plots of Al2O3/SiO2 against TiO2/Al2O3 (a) and Sb against Pb (b) for Utr 77 and Utr 79 (Utr = Utrecht). originally a soda glass; a high level of K2O (4.73%) (yet only 0.57% MgO) is probably also due to contamination. A high level of antimony (0.45% Sb2O3) is difficult to explain but could be due to glass recycling. Utr 32 contains 1.66% MgO, 1.9% K2O, 15.54% Na2O and 5.75% CaO, all characteristics of a plant ash glass. However, a very low level of P2O5 (0.06%) is unusual for a plant ash glass. The thick layer of green and red striped glass in crucible Utr 34 from Utrecht Domplein also warrants more detailed discussion. It contains 1.45% MgO and 1.42% K2O, levels that are probably consistent with a natron glass, as well as a probable uncontaminated level of Al2O3 (2.82%). However it also contains 0.56% P2O5 as well as 4.91% PbO, 0.9% SnO2, 0.615% CuO and 1.5% Fe2O3. The copper in the red glass would be in a reduced form (Cu2O) and also the iron; the lead, tin and iron were probably introduced as part of the colouring process along with the copper. This composition is similar to red coloured glass from Maastricht Jodenstraat and Wijnaldum. It also contains 0.728% MnO and 0.12% CoO. The levels of MgO and K2O are below 1.5% and fall within the values for a natron glass, yet the phosphorus level would be more in line with a plant ash glass and may indicate a degree of contamination. Red streaks of decoration are sometimes found in early medieval vessel glass. It can be assumed that it was produced deliberately by mixing in small amounts of red glass or red glass colorant. Number 33 is too contaminated to be able to discern the original chemical composition of the glass, with an Al2O3 level of 36.89%. The greenish glasses in crucibles 35 and 36 also contain a mismatch between the potassium and magnesium oxide levels as well as elevated antimony. Number 36 contains 5.4% K2O and 8.21% Na2O so it is tempting to suggest this might be a mixed alkali glass. However, the relatively low level of MgO (0.84%) suggests that the high K is due to contamination of the glass. Both 35 and 36 contain high levels of Al2O3 at 7.41% and 7.68%, so again this shows that contamination has occurred. They contain 1.1% and 0.39% CuO probably originally added as colourants to a natron glasses. 5.6 Glass samples from Wijk bij Duurstede (Dorestad) Apart from one raw glass chip, five tesserae fragments, one opaque yellow glass rod and two linen smoother fragments, the other 55 glass samples from Dorestad are all glass vessel fragments dated to between the mid-8th and mid-9th centuries, and do not derive from a glass-working context. 5.6.1 Vessel glass Apart from two wood ash glass samples (Dor 103 and Dor 136), a yellow-green palm funnel and an iridescent yellow-green funnel beaker base the low K2O and MgO contents and elevated Fe2O3, MnO and TiO2 contents of the rest of the 52 vessel glass samples from Dorestad suggest they are all made from natron glass. Their Al2O3/SiO2 and TiO2/Al2O3 ratios show that apart from Dor 122, a pale green funnel beaker, which can be categorized as Egyptian II glass, the other 51 samples cluster closely in the area the Foy 2 80 — Figure 5.15 Plots of Al2O3/SiO2 against TiO2/Al2O3 (a) and Sb against Pb (b) for Dorestad vessel glass samples. compositional group (Figure 5.15a with Figure 4.1 as reference). Apart from Dor 122, all the rest of the natron glass vessels have either Sb contents >1000 ppm or Pb contents >1000 ppm and mostly both (Figure 5.15b). This suggests these glass vessels were generally made from recycled glass rather than ‘pristine’ glass. Dor 122 is the clear exception in this assemblage. Not only does it belong to a different glass compositional group (Egyptian II), its low Sb and Pb contents, 285 ppm and 461 ppm respectively, also suggest it could have been made from ‘pristine’ glass (Figure 5.15b): the 26 trace element pattern for Dor 122 confirms that it is Egyptian II glass (Figure 5.16). The date of Dor 122 (9th century AD) is also consistent with the suggested date for when Egyptian II glass was in circulation, in the 8th–9th centuries AD. The two wood ash vessel glass samples (Dor 103, a palm funnel and Dor 136, a funnel beaker base) have quite similar chemical compositions. 293 Wedepohl and Simon 2010 294 Krüger & Wedepohl 2003. Their CaO contents are very high at over 13%, their K2O and MgO contents are significantly higher than that found in natron glass at over 7% and their Na2O contents are low at below 2%. Such compositions are quite typical of relatively early wood ash glass, even if the K2O levels are quite low for such glass.293 5.6.2 Other glass The two dark green linen smoothers (Dor 150, 151) have a very peculiar composition. They have high lead contents at around 22% and high Al2O3 contents at around 7%. Their high CaO, P2O5, MgO and K2O contents also suggest they could have been made from wood ash as a main ingredient. Over a hundred linen smoothers were found at an important medieval Viking city, Hedeby in northern Germany.294 Analytical work shows that they are generally of two glass compositions: wood ash glass and wood Figure 5.16 The 26 trace element pattern for Dor 122 compared to that of relevant Egyptian II natron glasses published previously (Dor=Dorestad). 81 — ash-lead glass. Our linen smoother samples are wood ash-lead glass, similar to those from Hedeby. Although the presence of the two linen smoothers could result from Viking trade a clear compositional match with slag resulting from the refinement of lead-silver ores at the Carolingian mine of Melle demonstrates that, along with many other linen smoothers found in France, Ireland, Germany, Norway Denmark and Belgium, the origin of the vitreous slag used to make them was Melle.295 It has been suggested that the glass working carried out in the important trading entrepôt of Dorestad was based on remelting tesserae, imported raw glass and glass rods (see Section 3.4). A raw glass chip from the site has a typical Foy 2 natron glass chemical composition, the same as nearly all of the vessel glass from the site. The elevated Sb and Pb contents of this raw glass chip, 4311 ppm and 2844 ppm respectively show that it was recycled glass. The five tesserae samples are in two colours, turquoise (three samples) and blue (two samples). From their major and minor chemical compositions, we can see that they are typical natron glass chemical compositions with high Na2O, low K2O and low MgO; the turquoise samples are coloured by copper and the blue samples are coloured by cobalt. Three of the five tesserae (Dor 145, 146, 147) have very high Sb contents of over 10000 ppm, a concentration significantly higher than that found in recycled Foy 2 glass. This compositional feature confirms that they are Roman glass with high Sb due to the presence of calcium antimonate opacifying crystals. The opaque yellow glass rod (Dor 149) has a similar chemical composition to opaque yellow glass from Maastricht, Jodenstraat and Wijnaldum. It has high Na2O and PbO contents and low K2O and MgO contents: it is also coloured by lead tin yellow II. Therefore it would have been made using the same procedure as the opaque yellow glass from Jodenstraat and Wijnaldum: natron glass was used as the base glass, and it was coloured by lead tin yellow II. The Al2O3/SiO2, TiO2/Al2O3 ratios and low Sb content of Dor 149 suggest that the base glass used could also have been ‘pristine’ Foy 2 glass. 5.7 Glass samples from Susteren The glass samples from Susteren fall into four categories: trail decorated glass beads (six samples), window glass (ten samples), glass attached to crucibles (two samples) and vessel glass (eleven samples). 5.7.1 Trail decorated glass beads The trail decorated glass beads (Sust 1–6) do not occur on the other sites studied here. The bodies of the trail-decorated glass beads are mainly a green colour of different shades, and the decorated glass trails are red, yellow and white. The matrix glass of all six trail-decorated glass beads was analysed (Sust 1–6 body); five coloured glass trails were analysed too (Sust 2–6 trail). The chemical compositions of the bodies shows that four of them are made from natron glass and two (Sust 3 and 4) were made from plant ash glass. In the Al2O3/SiO2 against TiO2/Al2O3 plot, three natron glass bead bodies (Sust 1, Sust 2 and Sust 6) distribute in the area of the Foy 2 compositional group while the Sust 5 body composition locates in the area of Egyptian II glass compositional group (Figure 5.17a with Figure 4.1 as reference). The elevated copper, antimony, lead and tin contents of four body glass samples suggest they were made from recycled glass (Figure 5.17a). There are two other glass samples from Susteren with compositional features like Sust 5: Sust 14 (window glass) and Sust 22 (glass vessel). They are recycled glass with high Sb and Pb contents over 1000 ppm, and they plot together in the area of the Egyptian II compositional group in the Al2O3/SiO2 against TiO2/Al2O3 plot. The chemical compositions of bead Sust 3 and Sust 4 bodies are clearly different from that of the other four trail-decorated bead bodies as they are soda-lime-silica glass with high K2O and MgO contents and are therefore plant ash glasses. More details regarding the possible origins of these glass beads are given in Section 5.10.7. As for the coloured glass trails, the red colour (Sust 2) contains high levels of CuO (1.74%) and Fe2O3 (4.59%) in a natron glass 295 Gratuze et al. 2003; Pactat et al. 2017. 82 — matrix. This suggests that the colour could be a result of 0 valence copper micron sized metallic particles and the iron inclusions, perhaps introduced in a slag, would have acted as an internal reducing agent.296 The opaque yellow colour of Sust 3 trail decorated glass bead has a similar chemical composition to the yellow glass beads from Maastricht Jodenstraat and Wijnaldum: it is coloured with lead tin yellow II.297 The colouring mechanism of opaque white Sust 4 and Sust 6 is different from any other opaque white glass beads studied in this work as they are not coloured by tin-based opacifiers but by calcium antimonate crystals. This suggests the opaque white glass trails are recycled Roman opaque white tesserae. Opaque yellow Sust 5 is also different from other opaque yellow glass studied here: it is coloured by lead antimonite rather than lead tin yellow II. This also suggests that the yellow glass used for the trail may have been recycled Roman tesserae.298 5.7.2 Window glass 296 297 298 299 Peake & Freestone 2012. Henderson 2023. Henderson 2023. Henderson 2023. As for the window glass samples, apart from Sust 11, which is a wood ash glass, the remaining nine window glass samples are all natron glass. In the Al2O3/SiO2 against TiO2/Al2O3 plot of these natron window glass samples (Figure 5.17a with Figure 4.1 as reference), Sust 16, which contains a high Al2O3 content at 3.6%, does not plot in the area of any recognized natron glass compositional group, and Sust 14 plots in the area of the Egyptian II glass compositional group, while the remaining seven samples all distribute in the area of the Foy 2 compositional group. All natron window glass samples contain elevated Sb and Pb contents (Figure 5.17b), suggesting they are recycled rather than pristine glass. We also noticed that Sust 7 and Sust 8 contain extra high Sb2O3 contents (1.58% and 1.70% respectively), which are much higher than that found in common recycled Foy 2 glass. The wood ash glass Sust 11 has very high P2O5 (2.5%), MgO (5.6%), CaO (11.0%) and K2O (14.2%) levels, which is a typical of northern European wood ash glass dating to the late 8th century at the earliest. 5.7.3 Glass attached to crucibles The two glass samples attached to crucibles (Sust 17, Sust 18) are natron glass but their composition is different from common natron glass. Sust 17 contains high Sb2O3 (1.5%) and K2O (5.2%) contents, but less than 1 wt% MgO. This peculiar composition may suggest the addition of recycled Roman tesserae and contamination from furnace wood ash. On the other hand, Sust 18 contains high Al2O3 (7.74%) and K2O (3.3%), which suggests that the glass has been contaminated by interaction with the crucible fabric.299 5.7.4 Vessel glass Eleven vessel glass samples from Susteren have been analysed. Their chemical compositions show that apart from two samples, Sust 19 (the pale green tubular base of a funnel beaker) and Sust 28 (a pale green funnel beaker fragment decorated with a green and white reticella rod), the others are natron glass. The Al2O3/SiO2 and TiO2/Al2O3 ratios of the nine natron glass samples indicate that Sust 26 (a nearly colourless funnel Figure 5.17 Plots of Al2O3/SiO2 against TiO2/Al2O3 (a) and Sb against Pb (b) for Susteren natron glass samples. 83 — Figure 5.18 The 26 trace element pattern of Sust 26 compared to that of the relevant HIMT natron glass type. beaker with a dark blue incalmo rim) belongs to the HIMT sensu stricto compositional group and Sust 22 (a blue-green funnel base) plots in the area of the Egyptian II compositional group, while the other seven samples are of a Foy 2 composition (Figure 5.17a with Figure 4.1 as reference). The Sb and Pb concentrations show that Sust 26 was made from ‘pristine’ glass while the other eight samples were made from recycled glass (Figure 5.17 right). The 26 trace element pattern of Sust 26 matches with that of pristine HIMT sensu stricto reported previously (Figure 5.18). Therefore this confirms that Sust 26 was made from ‘pristine’ HIMT sensu stricto glass. Sust 19 is a highly weathered funnel base which contains Na2O and K2O at 8.87% and 8.5% respectively. It is therefore a mixed alkali glass. Sust 28 is a plant ash glass sample containing 0.3% P2O5, 4.2% MgO, 2.5% K2O and 7.0% CaO. The trace element composition of Sust 28 has shed light on its provenance; this will be addressed along with other plant ash samples identified in this work in the discussion below. 5.8 Glass samples from Deventer Forty one samples from Deventer-Stadhuiskwartier were analysed. Apart from two monochrome glass beads (Dev 2, 8), the rest of the samples are glass vessels, window glass and raw glass fragments. Four compositional types of glass were identified. The largest number are wood ash glass (19 samples), thirteen samples are natron glass, four samples are mixed alkali glass and one sample is plant ash glass. Four samples were too weathered to be worthy of analysis (DEV 16, 17, 23, 34). Deventer samples are elaborated according to their compositional types in the following. 5.8.1 Wood ash glass The 19 wood ash glass samples are all vessel and window glass. The sample numbers are: DEV 1,6,7,8, 14,15, 21, 24, 26, 27, 28, 29, 30a, 32, 33, 35, 36, 38 and 39. The main fluxing agent in wood ash glass is K2O, with CaO in variable amounts. These wood ash glasses generally have high MgO contents over 3.5%, high P2O5 contents over 2.0% and low Na2O contents of below 3.0%. Apart from sample Dev 24, which has a CaO/K2O ratio of 5.8 which is the only example of a high lime (23.6% CaO), low alkali (2.9% Na2O, 4.1% K2O) glass (HLLA), the rest of the 19 samples have CaO/K2O ratios from 0.7 to 2.0 with an average at 1.1. Deventer 24 is a pale blue-green window fragment: its chemical composition is much more typical of medieval and post-mediveal glass300 so a later production date for it can not be ruled out (however see Section 6.5). The chemical compositions of early medieval wood ash glass are not especially well defined and as can be seen from previously published results they are quite variable.301 Analytical research on later (13th century) wood ash glasses provide some broad compositional trends but the origins in northern France, ‘Rhenish’ (based on borders of c. 1300 AD which includes the Low Countries) and central Europe 300 Van Wersch et al. 2018. 301 Wedepohl & Simon 2010. 84 — Figure 5.19 The average 26 trace element patterns of wood ash glass samples from Deventer compared to that of relevant natron glass studies published previously. do not provide especially clear distinctions and in many cases the ‘types’ overlap.302 Moreover a comparison with earlier Deventer wood ash glasses offers a somewhat confused picture and such a comparison is, in any case, inappropriate given the difference in dates associated with the different socio-economic contexts of production. The 26 trace element pattern of the average composition of these wood ash glass is plotted in Figure 5.19 along with that of natron glasses of selected compositional groups (Foy 2 and Egyptian II) reported previously. We can see that the trace element pattern of wood ash glass is quite different from that of natron glass, especially the clearly elevated Rb, Cs and Ba contents. 5.8.2 Natron glass 302 Adlington et al. 2019. Among the thirteen natron glass samples (DEV 2, 3, 9, 10, 11, 12, 13, 19, 20, 25, 31, 40, 41), three samples, surprisingly for such a late site, have features suggesting that they are of a Roman provenance: Dev 9 (an amber coloured fragment), Dev 13 (a funnel beaker with a very high Mn content suggesting Mn decolourisation) and Dev 20 (a typical Roman vessel shape, Isings 96a). All date to between 900 and 950 AD. At the same time, the three samples also distribute in the Roman glass area in the Al2O3/SiO2 against TiO2/ Al2O3 plot and have low Sb and Pb concentrations <1000 ppm (Figure 5.20). Therefore they can definitely be identified as glass produced in the Roman tradition. The Al2O3/SiO2 and TiO2/Al2O3 ratios of the other ten natron glass samples show that Dev 11 (the rim of a possible funnel beaker), Dev 31 (possible window glass) and Dev 40 (a fragment of dark and light green layered glass) all plot in the area of Egyptian II glass; they date to 900– 925, 890–925 and 950–1050 AD respectively according to the find contexts. The rest belong to the Foy 2 compositional group (Figure 5.20a with Figure 4.1 as reference) and date to between 850 and 1050. The Sb and Pb concentrations of the ten samples suggest that Figure 5.20 Plots of Al2O3/SiO2 against TiO2/Al2O3 (a) and Sb against Pb (b) for Deventer natron glass samples. 85 — Figure 5.21 The 26 trace element patterns for Deventer 11, 31 and 40 compared to that for relevant Egyptian II natron glass published previously. the Egyptian II subgroup (Dev 11, Dev 31 and Dev 40) are all ‘pristine’ glass while only one Foy 2 monochrome blue glass bead (Dev 2) is a ‘pristine’ glass (Figure 5.20b) dating to 850–900 AD. The 26 trace element pattern of Dev 2 confirms that it was made from ‘pristine’ Foy 2 glass, while the patterns for Dev 11, Dev 31 and Dev 40 confirm that they were made from ‘pristine’ Egyptian II glass (Figure 5.21). The rest of the Foy 2 glasses are recycled. Two mixed alkali glasses have elevated PbO contents of between 2.4 and 2.8% lead oxide. This could be interpreted as an addition of lead glass, introduced during the recycling process and which should not be regarded as a surprise if a ‘potluck glass working’ strategy was used during glass recycling. 5.8.3 Mixed alkali glass The one plant ash glass identified from the Deventer assemblage (Dev 37) is a green possible beaker fragment with a little weathering, from a context dating to between 950 and 1050. It has a typical plant ash chemical composition. It is a soda-lime-silica glass with high MgO and K2O contents at 1.8% and 4.77% respectively. Discussion of its possible origin is given in Section 5.10.7. In contexts dating to 850-900 AD, six out of eight samples are wood ash or mixed alkali glasses, with the remaining two being recycled Foy 2. In contexts dating to between 900 and 950 AD there are ten wood ash glasses, three Roman (natron) glasses, two recycled Foy 2 glasses and two Egyptian II glasses. Two glasses dating to between 950 and 1000 are wood ash; between 950 and 1050 four glasses are of the wood ash type, one is plant ash, one recycled Foy 2 and one Egyptian II. Given that there is no evidence for a wood ash glass industry in the Netherlands possible ways that wood ash glass was imported, would have been as part of the Viking trade network or from the various sites further north. The four mixed alkali glass samples from Deventer (DEV 4, 5, 18, 22) are an interesting group, which are not found amongst glasses reported elsewhere in this study and are a reflection of their production date. The samples tested derive from archaeological contexts dated to between 850 and 950 AD. Their high MgO contents (>1.8%) suggest the potassium content was introduced in the form of plant ash/wood ash rather than potassium-containing minerals. Their Na2O/K2O values are variable, between 0.5 and 1.7. Glass with similar features has also been found in 8th–10th century sites in Germany, France and the Netherlands (see Section 2.4.1). It has been suggested that this type of mixed alkali glass was produced in order to extend stocks of soda glass,303 and the procedure can be interpreted either as the mixing together of wood ash and natron glasses or, less likely, by the addition of an increasing amount of wood ash in a mixture of glass cullet.304 5.8.4 Plant ash glass 303 Krueger & Wedephol 2003. 304 Pactat et al. 2017. 86 — 5.9 Nd-Sr isotope analysis Twenty samples selected from Maastricht (Jodenstraat), Wijnaldum and Dorestad were analysed for their strontium and neodymium isotopic compositions, and eighteen valid results were obtained. In this study 143Nd/144Nd ratios are represented in parts per 104 deviation from the present-day value of a model evolution of Nd isotopes in a chondritic Earth (Chondritic Uniform Reservoir, CHUR)305 according to the following equation: ε Nd= (143Nd⁄144Nd ) Sample (143Nd⁄144Nd ) CHUR -1 ×104, with 143Nd/144NdCHUR =0.512638 The strontium and neodymium isotopic compositions of our 18 samples and their chemical compositional groups are listed in Table 5.3 and plotted in Figure 5.22. The two wood ash glass samples, Dor 136 and Dor 150, both lead-rich linen smoothers, have very different Nd–Sr isotopic signatures from the other samples. Their Nd isotopic signatures reflect the Nd isotopic signatures of the sands used for making them, and their Sr isotopic signatures reflect the bio-available Sr isotopic signatures in the calcium-rich raw materials used to make them. The Sr from wood ash glass would have mostly been introduced in the wood ash used as the flux. The Nd–Sr isotopic signatures of Dor 136 and Dor 150 have a similar range to that of 14th–15th century forest glass produced in Staffordshire, England.306 However there are no other Nd–Sr isotopic datasets of wood ash glasses by geographic region currently available so we cannot suggest provenances based on Nd–Sr results for two wood ash glass samples at this stage. The only plant ash sample, WIJ 37, from the assemblage clusters together with natron glass samples in figure 5.22 but it has a low 87Sr/86Sr signature. Comparing the Nd–Sr isotopic signatures of WIJ 37, a silver foil colourless bead, with that of other available plant ash glass data (Figure 5.23), we found that it overlaps with 3rd–7th century Sasanian glass found in Veh Ardašīr, an old Sasanian administrative centre 40 km to the southeast of modern Baghdad.307 Although the 3rd–7th century Sasanian glass samples from Veh Table 5.3 The Nd–Sr isotopic compositions and compositional groups of samples analysed (WIJ=Wijnaldum; DO=Dorestad). Sample number 305 Depaolo & Wasserburg 1976. 306 Meek, Henderson & Evans 2012. 307 Ganio et al. 2013. Compositional group 143Nd/144Nd 87Sr/86Sr εNd Maastricht-Jodenstraat 47 ‘pristine’ Foy 2 0.512 -5.6 0.709 Maastricht-Jodenstraat 49 ‘pristine’ Foy 2 0.512 -7.6 0.709 Maastricht-Jodenstraat 58 ‘pristine’ Foy 2 0.512 -5.4 0.709 Maastricht-Jodenstraat 63 ‘pristine’ Foy 2 0.512 -5.4 0.709 Maastricht-Jodenstraat 68 Roman glass 0.512 -5.2 0.709 Maastricht-Jodenstraat 73 ‘pristine’ Foy 2 0.512 -6 0.709 Maastricht-Jodenstraat 76 ‘pristine’ Foy 2 0.512 -5.6 0.709 WIJ 13 recycled Foy 2 0.512 -5 0.709 WIJ 27 not identified 0.512 -6.8 0.709 WIJ 36 ‘pristine’ Foy 2 0.512 -5.4 0.709 WIJ 37 plant ash glass 0.512 -6.4 0.709 DOR 111 recycled Foy 2 0.512 -6 0.709 DOR 113 recycled Foy 2 0.512 -5.6 0.709 DOR 115 recycled Foy 2 0.512 -6 0.709 DOR 128 recycled Foy 2 0.512 -5 0.709 DOR 136 wood ash glass 0.512 -9.9 0.715 DOR 147 Roman tesserae 0.512 -6.2 0.709 DOR 150 wood ash glass 0.512 -10.7 0.711 87 — Figure 5.22 A Nd–Sr isotopic plot of the glasses analysed; WIJ 37, Joden 49, DOR 136 and DOR 150 are labelled separately (WIJ= Wijnaldum; DOR = Dorestad; Joden = Jodenstraat, Maastricht). Figure 5.23 A Nd–Sr isotopic plot of WIJ 37 compared with that of plant ash glasses from the Middle East published previously (WIJ = Wijnaldum). Data source: 3rd–7th century Sasanian glass Nd–Sr isotopic data is from Ganio et al. (2013), Tyre raw furnace glass from Degryse et al. (2010), al-Raqqa 9th century and 11th–12th century vessel glass data from Henderson, Evans & Barkoudah (2009), the rest from Henderson, Ma & Evans (2020). Figure 5.24 A Nd–Sr isotopic plot of the natron glass samples analysed. Joden 49 is labelled separately (Joden = Jodenstraat, Maastricht). The ellipses are drawn according to Degryse & Schneider (2008). Ardašīr are too early for the late 9th century date of WIJ 37, this observation still sheds some light on the possible geographical origin of this plant ash glass sample. This is addressed in detail along with other plant ash samples identified in this study in Section 5.10.7. We have suggested that Joden 68 was a Roman ribbed vessel fragment: this is confirmed as a Roman natron glass from its major and minor chemical elemental composition; Dor 147 is a 88 — Figure 5.25 The 26 trace element patterns for Joden 68 and Dor 147 compared to that for Levantine I natron glass published previously (Joden= Jodenstraat, Maastricht; DOR = Dorestad). Roman glass tessera. We did not try to make further provenance identification based on their trace element data since the geographical production centre of this Roman glass type is not yet clear. Joden 68 has a 87Sr/86Sr of 0.7090 and εNd of -5.03, and Dor 147 has a 87Sr/86Sr of 0.70892 and εNd of -6.20, typical Nd–Sr signatures of glass produced in Levant.308 Following this lead, we compared the trace element patterns of Joden 68 and Dor 147 with that of early medieval Levantine glass, and we found they both have similar trace element patterns to Levantine I glass (Figure 5.25). This observation indicates that Joden 68 and Dor 147 were made with a similar sand source as that used for making Levantine I glass. Therefore we suggest that Joden 68 and Dor 147 are Roman glasses produced in the Levant. 308 Degryse & Schneider 2008. 309 Brems et al. 2018. The Nd–Sr isotopic compositions of the thirteen natron glass samples mostly distribute in the typical range for natron glass Nd–Sr signatures, εNd at between -5 and -7, and 87Sr/86Sr at between 0.7085 and 0.7093,309 except for one outlier: Joden 49, a drop of green glass with soil contamination (Figure 5.24). The εNd of Joden 49 is -7.3, lower than the normal range of natron glass -7 ≤ εNd ≤ -3, and this may be related to contamination. Twelve out of the thirteen natron samples have been identified as belonging to the Foy 2 compositional group: seven ‘pristine’ Foy 2 glass and five recycled Foy 2 glass (Table 5.3), and the 87Sr/86Sr ratios of the twelve samples stretch in a rather wide range between 0.7085 and 0.7093. Six out of the seven ‘pristine’ Foy 2 glasses (except for the outlier Joden 49) actually cluster closely at 87Sr/86Sr of 0.7086, a typical value for glass produced in Egypt, the commonly suggested origin of Foy 2 glass. Five of these samples are the working debris from bead production in Maastricht (a glass strand, a drop and three rods) and a gold foil bead from Wijnaldum. The 87Sr/86Sr values for five recycled Foy 2 glass samples, a pale green funnel beaker from Wijnaldum as well as two funnel beakers, one palm funnel and a gold leaf decorated funnel from Dorestad stretch between 0.7088 and 0.7093. This is higher than the typical Egyptian range. These recycled ‘Foy 2’ glasses are all dated to the late 8th to mid-9th century. We suspect the Sr isotopic signatures of recycled Foy 2 glass samples would have changed from the ‘pristine’ Foy 2 values during the recycling process, which would have involved mixing different types of glass. The mixing of glass types with very high 87Sr/86Sr values (such as wood ash glass), into a mainly natron glass recycling batch, would have caused the elevation of 87Sr/86Sr values in the recycled natron glass compared to ‘pristine’ natron glass. Evidence of recycling/mixing wood ash glass along with natron glass has also been noted in the trace element patterns of recycled Foy 2 glass dated to the late 8th century and later. This is addressed in Section 5.10.4. 89 — 5.10 Discussion bead-making artefacts found at the Maastricht Jodenstraat site were made from the same glass as the base glass used for making the opaque glass beads: it is all ‘pristine’ Foy 2 glass. 5.10.1 The base glass used for bead making at Jodenstraat It has been suggested that a two-step procedure was used for the manufacture of highly coloured opaque glass beads in early medieval northwestern Europe: first, naturally coloured or colourless base glass was coloured and made into rods or strands. Then coloured glass rods or strands were softened and formed into beads.310 Apart from the highly coloured opaque glass objects, the site of Jodenstraat in Maastricht also yielded translucent naturally coloured and cobalt blue bead-making waste, which is ideal to test this two-step production proposition and understand what kind of glass was used as the base glass. In this work we found three types of evidence that prove that the naturally coloured translucent glass-working debris is of the same compositional type as the base glass used to make opaque glass and thus that the two-step production proposition stands. Firstly, the Al2O3/SiO2 and TiO2/Al2O3 ratios of the coloured opaque glass-working remains and translucent glass samples cluster together and they can all be categorized as being of the Foy 2 type. From the Al2O3/SiO2 versus TiO2/Al2O3 plot we can see that, on average, highly coloured glasses have a slightly higher Al2O3/SiO2 ratio than naturally coloured glass debris: the cluster of highly coloured glasses plot slightly to the right of the cluster of naturally coloured glassworking debris (Figure 5.1). This can be attributed to the higher Al2O3/SiO2 ratios of the lead tin yellow colourant and tin opacifiers that were added to the highly coloured glass matrix (Table 5.1). Secondly, the Sb contents of the naturally coloured glass-working debris and highly coloured glass samples are all very low, suggesting they were both made from ‘pristine’ glass. Finally, the 26 trace element patterns of the average composition of the naturally coloured bead-making artefacts and the average composition of the highly coloured glass are identical and the same as that of ‘pristine’ Foy 2 glass previously published. Therefore we can suggest strongly that the naturally coloured 5.10.2 The use of crucibles in on-site lead tin yellow colourant production in early medieval northwestern Europe Early medieval lead tin yellow residues attached to crucibles from northwest European sites have been studied before.311 The shapes of the crucibles containing lead tin yellow residues are mostly shallow and with a wide opening that resembles a tray. It has been suggested this type of crucible was not used for metallurgical processes but was specifically used for making the lead tin colourant for yellow coloured beads,312 which were very popular in early medieval Europe. We agree with the suggestion that the crucibles with yellow residues were used for making the yellow colourant for bead making, since the main phase in the yellow residue is lead tin yellow II (PbSn(Si)O3), the common colouring agent of yellow beads at the time. However, from the study of the rich material remains related to bead making found at Jodenstraat Maastricht, we have managed to provide some new insights into how lead tin yellow was made in these tray shaped crucibles. It has been suggested that this type of crucible may have been used for calcining lead and tin, a chemical reaction which would result in lead tin yellow I, and that lead tin yellow II identified in the crucible may have been formed by the reaction between the siliceous crucible body and lead tin yellow I.313 Heck and colleagues also suggested that the colourant produced in the crucible was to be mixed, in a ratio of one to one by volume, with natron glass to make the yellow glass for beads.314 Peake and Freestone315 studied yellow beads from Tarbat Ness and Eriswell and reviewed the results of Henderson and Ivens316 and Heck and colleagues317 for the yellow residues attached to crucibles, and concluded that the reaction which occurred in the crucible, resulting in the lead tin yellow II colourant, involved silica as a raw material. The thorough replication experiments to produce lead tin yellow colourant by Rooksby318 and Matin and colleagues319 demonstrated that the 310 Sablerolles, Henderson & Dijkman 1997. 311 Henderson & Ivens 1992; Heck, Rehren 312 313 314 315 316 317 318 319 & Hoffmann 2003; Peake & Freestone 2014. Heck, Rehren & Hoffmann 2003. Heck, Rehren & Hoffmann 2003. Heck, Rehren & Hoffmann 2003. Peake & Freestone 2014. Henderson & Ivens 1992. Heck, Rehren & Hoffmann 2003. Rooksby 1964. Matin, Tite & Watson 2018. 90 — a b c d Figure 5.26 Backscattered SEM image (top left: 5.26a) for Joden 19 showing the unmelted silica grains. Elemental lead maps for Joden 19 lead (Pb) (top right: 5.26b), silica (Si) (bottom left: 5.26c) and tin (Sn) (bottom right: 5.26d) (Joden = Jodenstraat, Maastricht). yellow colourant for glass- and glaze-making would have been produced in a two-step procedure, as recorded in ancient Islamic literature. Firstly, a lead tin calx, which contains lead tin yellow I (PbSn2O4) as the main component, is produced by calcining lead and tin together in a stoichiometric ratio of Pb:Sn over 3.5. Then lead tin calx is mixed with silica and the mixture is heated to over 800°C for the lead tin yellow colourant (lead tin yellow II) used in glass- and glaze-making to be produced. During the second step, variable amounts of SiO2 substitute for SnO2 in PbSn2O4 (lead tin yellow I) which causes a crystalline conversion to PbSn(Si)O3 (lead tin yellow II). No lead tin yellow I phase has been identified in the yellow tin residues in this study, and the chemical composition of lead tin yellow II attached to crucibles is in good agreement with that found in the yellow glass and yellow beads, especially with respect to the Sn to Si ratio, which could be variable depending on how much Sn was replaced by Si during the reaction involving lead tin yellow I and silica. Therefore we suggest that the lead tin yellow colourant production procedure used by early medieval northwestern European bead makers would have been very similar to that recorded in later Islamic literature. This was a two-step procedure: the tray shape crucibles with lead tin yellow residues attached would have been used during the second step when heating silica with lead tin yellow I for making lead tin yellow II occurred. 91 — Silica grains are observed in one of our samples (Joden 19) in support of this suggestion (Figure 5.26), also observed in the Early Christian Irish evidence from Dunmisk.320 5.10.3 The separate production of a tinbased white opacifier at Maastricht, Jodenstraat Apart from crucibles containing lead tin yellow residues, two crucibles containing tin white residues (Joden 23 and Joden 30) were found at Jodenstraat. Our study shows that the main phase of the white residues is SnO2 (50–70 wt%) in the presence of variable amounts of PbO and SiO2 (see Section 5.2.3). Because the Sn:Pb ratio in the tin white residues is very different from that of lead tin yellow residues and very similar to the tin-based opacifiers found in the white, red and greenish-blue glass and beads, we can strongly suggest that the tin-based opacifiers were intentionally produced as a separate material on site at Jodenstraat and that they were probably used directly in the production of the white, red and greenish-blue glass for bead making. Highly coloured glass opacified by tin-based opacifiers re-emerged around the same time when lead tin yellow glass became mainstream in northwest Europe in c. 3rd century AD.321 Although early medieval northwestern European lead tin yellow colourant production has been studied on a few occasions, the two crucibles containing tin-based opacifiers used to make white glass are the first reported evidence confirmed by scientific analysis that tin opacified white glass was produced separately in early medieval northwestern Europe. In a previous study of bead-making materials from Jodenstraat322 it was suggested that the white, red and greenish-blue glass rods found at the site for producing beads of the same colours may have been imported from other sources since no crucibles containing glassy residues of these three colours were found at the site. Two extra pieces of evidence identified here suggest that the white, red and greenish-blue glass rods for producing beads of the same colours were also produced on site, rather than being imported from elsewhere. Firstly the base glass of all four different coloured glasses and beads as well as the colourless glass from working the glass were made from a very similar ‘pristine’ Foy 2 glass (see Section 5.10.1), and secondly the opacifiers used in the white, red and greenish-blue glass were probably also produced on site. The contrasting PbO contents (Figure 5.3) between yellow glass and the white, red and greenish-blue glasses is probably also related to the use of lead tin yellow colourant in the yellow glass and tin-based opacifiers in white, red and greenish-blue glass. When the lead tin yellow colourant was produced, a pure lead silica glass formed around lead tin yellow II in the lead tin yellow colourant as we can see in Figure 5.5. When in the third stage the colourant is added to the melted natron base glass, the pure lead silica glass would have melted and mixed into it and the lead content of the resulting glass would have increased greatly as a result. On the other hand, the production of the tinbased white opacifier would have involved far less lead; no pure lead silica glass formed around the tin opacifier.323 So when the tin opacifier was mixed into melted natron base glass, lower levels of lead would have been introduced. This is a reasonable explanation for the source of the lead content in highly coloured opaque glasses and beads, and why there is a big difference in the PbO contents between yellow glass and glass of the other three colours (Figure 5.3). Moreover, the methods used to make the lead tin yellow colourant and tin oxide opacifier were different. As we can see in the SEM backscattered images of the four highly coloured opaque glasses (Figure 5.4), the bright phase in yellow glass (lead tin yellow II crystals) consists of very small particles of a similar size and they are distributed homogeneously in the glass matrix, while the tin oxide crystals in white, red and green glass have heterogeneous crystal sizes and a lot of them are in rather big aggregates. This observation suggests that the lead tin yellow colourant (lead tin yellow II) may have been mixed with the base glass while they were both in melted or semi-melted state, and that is why the lead tin yellow II consists of small particles that are distributed homogeneously in the glass matrix. On the other hand, the white tin oxide opacifier may have been added in a powdered state after it was produced on site in a large volume. 320 321 322 323 Henderson & Ivens 1992. Matin 2019. Sablerolles, Henderson & Dijkman 1997. Matin 2019. 92 — 5.10.4 The other chemical characteristics of the glass and its archaeological implications 324 325 326 327 328 329 330 Foy et al. 2003. Freestone et al. 2018. Foy et al. 2003. Foy et al. 2003. Ares et al. 2019. Bertini, Henderson & Chenery 2020. Freestone 2015. Glass fragments from non-industrial contexts account for the majority of the samples studied. They include glass vessels, window glass and some raw glass fragments. Altogether 160 of such glass samples have been chemically analysed – 28 from Gennep, five from Jodenstraat Maastricht, nine from Wijnaldum, three from Utrecht, 56 from Dorestad, 21 from Susteren and 38 from Deventer. The majority (95) of these samples have been identified as recycled Foy 2 glass, corresponding to the Foy 2.2 group of Foy et al.324 Nine HIMT sensu stricto glasses have been identified from the collection, and seven of them are from the site with the earliest date: Gennep (late 4th to mid-6th century). Eight ‘pristine’ Foy 2 glasses have been identified: they are from Gennep (six samples) and Maastricht Jodenstraat (two samples). They date to between the late 4th and early 7th centuries. Six ‘pristine’ Egyptian II glasses have been identified: two from Wijnaldum, one from Dorestad and three from Deventer. Their dates provided by the archaeological contexts in which they were found are as follows: three date to 8th–9th centuries AD, two date to the early 10th century and one dates to between 950 and 1050 AD. They are consistent with or slightly later than the suggested date when Egyptian II glass was widely circulated, in the 8th and 9th centuries AD. Twenty seven wood ash/mixed alkali glass samples have been identified. They are from three sites: Dorestad (two samples), Susteren (two samples) and Deventer (24 samples) and their context dates are consistent with the suggested date when wood ash glass started to be made and was circulating in Europe from the late 8th century onwards. Five Roman glasses have been identified: they are from Jodenstraat (Joden 68), Utrecht (Utr 79) and Deventer (Dev 9, 13, 20). Four plant ash samples have been identified: Sust 3, 4 and 28, and Dev 37. The remaining eight samples are five compositional outliers (Utr 78, Dor 95, Sust 14, 16, 22) and three highly weathered samples (Dev 16, 17, 23). We can see that the identified compositional types of these mainly vessel and window glass fragments are consistent with our current understanding of the dates when the different types of glass emerged, peaked and disappeared. It has been concluded that HIMT glass in the widest sense (including HIMT sensu stricto, Foy 2 etc.) was in use from the middle of the fourth century until the seventh century.325 However, in this study we have found that glass made from recycled glass with Foy 2 glass compositional features (referred to as recycled Foy 2 glass in this study, corresponding to Foy 2.2 group)326 was still the dominant group between early 8th century and late 9th century (in Dorestad and Wijnaldum vessel glasses). It remained in circulation till possibly as late as the first half of the 10th century: four out of 38 glasses analysed from Deventer dated to 850–950 AD have been identified as recycled Foy 2 glass- with a single example, perhaps redeposited, in a context dating to 950-1000. This means that recycled glass with Foy 2 compositional features was still in use centuries after the ‘pristine’ Foy 2 glass supply ended in northwestern Europe. The chemical composition of recycled Foy 2 glass samples, have some differences from that of ‘pristine’ Foy 2 glass, one of which is that they nearly all contain elevated Sb and Pb contents. This recycled version of Foy 2 glass (especially for Foy 2.1 subgroup) is labelled as Foy 2.2 subgroup in the work of Foy and colleagues,327 and is known from a very limited number of assemblages in France, Italy and Spain that are typically dated to the end of the seventh and the eighth centuries AD.328 The 7th–11th century glass assemblage from Comacchio, northern Italy, is compositionally very similar to the samples studied here, as recycled glass with Foy 2 compositional features account for the majority of the glasses: the authors of that paper labelled such recycled glass ‘intermediate’.329 This observation is in line with the suggestion that around the early 8th century, Roman tesserae became a ‘new’ glass source in northwestern Europe; it was recycled along with other glass, supplementing the dwindling natron glass supply.330 From the 26 trace element patterns of the average compositions of recycled Foy 2 glasses from Gennep, Wijnaldum, Dorestad, Susteren and Deventer we can see that they all retain the basic signature of ‘pristine’ Foy 2 glass (Figure 5.27). However, we have also noticed that the 26 trace element patterns of recycled Foy 2 93 — a b c Figure 5.27 The average 26 trace element patterns of recycled Foy 2 glasses from Gennep (5.27a), Wijnaldum (5.27b), Wijk bij Duurstede (Dorestad) (5.27c). 94 — d e f Figure 5.27 (continued) The average 26 trace element patterns of recycled Foy 2 glasses from Susteren (5.27d), Deventer (5.27e) and ‘intermediate’ glass from Bertini, Henderson & Chenery (2020) (5.27f ). 95 — glasses from sites of the Carolingian period (Dorestad, Susteren) and slightly later at Deventer clearly have elevated Cs and Rb contents. Recycled Foy 2 glass from Gennep and Wijnaldum, which mostly predate the 8th century, do not show these features so strongly, nor do the recycled ‘intermediate’ group studied by Bertini and colleagues.331 This could suggest that glass of a new compositional type started to be involved in glass recycling as a minor component in the Netherlands from around the start of the Carolingian period. We suspect that this ‘new’ glass could have been wood ash glass since elevated Ba, Rb and Cs contents are the clear discriminating characteristics between wood ash and natron glass as defined by 26 trace element patterns with higher concentrations in wood ash glasses (Figure 5.19). This suggestion is also supported by the elevated Sr isotopic compositions of recycled Foy 2 glass samples dated to late 8th to mid-9th century from Wijnaldum and Dorestad, and in line with the understanding that wood ash glass started to be manufactured in Europe from the late 8th century in northwestern Europe. A further compositional characteristic of both Merovingian glass from Gennep and Carolingian glass from Dorestad is that potassium is correlated with both Rb and Li. An intriguing characteristic of these correlations (not shown here) is that Rb levels in the Gennep ‘pristine’ Foy 2 glass are mainly below 12ppm whereas the concentrations in Dorestad recycled Foy 2 glasses are mainly between 12ppm and 20ppm with a small number containing levels upto 28ppm. This is another clear marker of the increased degree of recycling in the later (Carolingian) glasses and the same thing is true for Li concentrations. Some Dorestad glasses contain between 23 and 40ppm Li. These same correlations found in 7th century and later Foy 2.1 glasses have been attributed to evidence of sitespecific contamination from e.g. muscovite in the crucibles used for working the glass at Tolmo de Minateda, Spain. 332 Our results suggest that such contaminants including very similar concentrations of Rb and Li as in the Tolmo glasses may not be characteristic of local production in our case. A more likely interpretation is that the increase in concentrations of potassium, rubidium, lithium and cesium in Carolingian glasses compared to earlier glasses may be attributable to muscovites associated with sands used to make some of the glasses that were mixed as part of the Carolingian recycling processes involving Foy 2 glass. 5.10.5 A comparison of 7th–11th century vessel glass from Comacchio with early medieval Dutch glass and the suggested supply of raw glass in the two areas We have mentioned that the recycled Foy 2 glass in this study bears similar compositional features to the ‘intermediate glass’ identified by Bertini and colleagues333; they are both recycled glasses and their major and minor chemical compositions are very similar to that of ‘pristine’ Foy 2 glass. Just as the recycled Foy 2 glass was the dominating compositional group of vessel glasses from early medieval Dutch sites, the recycled intermediate glass was also the dominating compositional group for glass of 7th–11th century dates found at Comacchio, northern Italy (53 out of 77). However, in the two studies only a very limited number of ‘pristine’ Foy 2 glasses have been identified. This phenomenon suggests that ‘pristine’ Foy 2 glass would have been the main raw glass imported to the two areas for glass working prior to or in the early part of the period. However, after this supply waned, local glass working in the two areas had to rely more on recycling contemporary cullet and old Roman tesserae. The chemical compositions of glass from Comacchio and from the early medieval Netherlands also reflect some differences in the supplies of raw glass. First of all, Levantine glass was one of the important sources of raw glass in Comacchio, possibly after the ‘pristine’ Foy 2 glass supply waned in the area, but not in the Netherlands. Altogether 17 pristine Levantine glass samples were identified from the total of the 77 items analysed from Comacchio.334 This is a proportion that is much higher than the number (four) of ‘pristine’ Foy 2 samples found at the site. Besides, the chemical compositional features of mixing Foy 2 glass with Levantine glass were also noted in the intermediate glass group, which suggests that apart from Foy 2 glass, Levantine glass was another source of glass contributing to the recycling process. 331 332 333 334 Bertini, Henderson & Chenery 2020. Schibille et al 2022. Bertini, Henderson & Chenery 2020. Bertini, Henderson & Chenery 2020. 96 — 335 336 337 338 339 340 341 342 Schibille et al. 2022. Pactat 2021, Fig 7. Cagno et al. 2012. Genga et al. 2008. Casellato et al. 2003. Pactat 2021. Phelps et al. 2106. Krueger & Wedepohl 2003. In contrast, apart from one punty glass fragment from Wijnaldum (WIJ 42), no pristine early medieval Levantine glass was found amongst the samples analysed here, and no mixing with Levantine glass can be suggested from the chemical compositions of the predominantly recycled Foy 2 glass found here. Analysis of glass from another contemporary site in the southern Mediterranean, Tolmo de Minateda (Spain), also revealed the presence of a higher proportion of Levantine I glass (33 out of 253 samples).335 Therefore, we can be sure that pristine Levantine natron glass did not arrive in the Netherlands in the relatively large quantities that reached northern Italy and Spain. Secondly, no wood ash glass was identified amongst the 7th–11th century Comacchio glasses analysed. By the 9th–11th centuries, wood ash glass accounts for more than half of the samples tested from Deventer with the proportion of wood ash glasses found in contexts dating to after 900 increasing somewhat, perhaps as a result of the Viking trade network. Wood ash glass was almost entirely dominant in the manufacture of a range of glass vessels by the 10th century in France with minimal recycled natron glass.336 Therefore this shows that wood ash glass found in northern, western and central Europe may not have been available in Comacchio and other sites in northern Italy337 during the period, just as Levantine glass may only have had very limited availability in the early medieval Netherlands. No wood ash glass was found at the 9th to 13th century glass from the site of Siponto in southern Italy338 and only plant ash glass was being manufactured by the 13th14th century at the northern Italian site of Germagnana.339 Moreover, six pieces of Egyptian II glass have been identified amongst Wijnaldum, Dorestad and Deventer glasses. The three Deventer examples all contain lower levels of Li, R, Cs and Ba than detected in contemporary recycled Foy 2 glasses discussed above reflecting their ‘pristine’ nature. Although Egyptian II do not account for a high proportion of the total glasses analysed, they are the only ‘pristine’ natron glass found from the 8th century and later with only an ‘anecdotal’ occurrence in France.340 This could mean that amongst the limited amount of ‘pristine’ natron glass that arrived in the Netherlands during this period, Egyptian II glass was quite important. Moreover, no Egyptian II glass was identified amongst 7th–11th century glass vessels from Comacchio even though a higher proportion of Egyptian II glass was found in the Levant after the 9th century.341. This therefore reflects an important contrast in the availability of raw glass in northern Italy and the Netherlands and it also partly reflects a collapse in Mediterranean trade. 5.10.6 Wood ash glass and mixed alkali glass Wood ash glass and mixed alkali glass are treated as one group here because they were both produced using wood ash as one important raw material. All wood ash glass and mixed alkali glass identified came from Carolingian sites: two wood ash glasses from Dorestad, one wood ash glass and one mixed alkali glass from Susteren, and twenty four wood ash and mixed alkali glasses from Deventer. Their dates are all later than the end of the 8th century, which is in line with the date that wood ash glass technology emerged in northwestern Europe.342 We have also noted above that wood ash glass may already have been involved in glass recycling in the Netherlands as a minor component during the Carolingian period as suggested by the clearly elevated Rb and Cs contents of the recycled glass from Dorestad, Susteren and Deventer (see Section 5.10.4). However, the proportions of wood ash in all the glasses studied from the three sites are quite different. In Dorestad and Susteren, wood ash glass and mixed alkali glass only account for a very small fraction of early medieval glass, while in Deventer wood ash glass and mixed alkali glass make up two thirds of the total vessel glass from the site. This difference can largely be attributed to the dates of glass samples from the three sites. The glass from Dorestad and Susteren dates to between the 7th and 9th centuries, while the glass from Deventer dates to between the mid-9th and mid-11th centuries. This suggests that the supply of wood ash and mixed alkali raw glass was quite limited in the Netherlands during the late 8th century and early 9th century and that they are likely to have formed a significant part of the raw glass supply from the mid9th century onwards. 97 — Figure 5.28 Plots of Cr/La against 1000Zr/Ti ratios (top) and Li/K against Cs/K ratios (bottom) of plant ash glasses from Deventer and Susteren (labelled) and Wijnaldum compared to data published previously. Data source: Henderson et al. 2016. 5.10.7 Plant ash glass Altogether only six plant ash glasses were identified from all the glass analysed here. They are four glass beads, WIJ 35 and WIJ 37 (a gold foil and a silver foil bead), Sust 3 and Sust 4 (both trail decorated conical beads) and two glass vessels, Sust 28, a pale green funnel beaker and Dev 37, a possible beaker fragment. It is difficult to give clear provenance identification for plant ash glass by using major and minor chemical compositions alone. However, by plotting certain key trace element values343 (Cr/La against 1000Zr/Ti and Li/K against Cs/K) found in plant ash glass samples along with a large dataset of plant ash glasses of different origins from the Middle East,344 we can provide some clues about the possible origins of our samples (Figure 5.28). From the two plots using key trace element ratios we can see that three samples (WIJ 35, WIJ 37 and Sust 28) cluster together in both plots. Sust 3 clusters together with WIJ 35, WIJ 37 and Sust 28 in the Cr/La versus 1000Zr/Ti plot and with Sust 4 in the Li/K versus Cs/K plot. Dev 37 clearly has a rather different trace element composition from the remaining five samples. In terms of the origins of the plant ash samples, we can see WIJ 35, WIJ 37 and Sust 28 basically cluster closely with samples from sites located in the eastern zone of Western Asia, in Iraq and Iran: Nishapur, Ctesiphon and Samarra. The Nd– Sr isotopic ratio of WIJ 37 also clusters with Sasanian glass samples from Veh Ardašīr, 40 km to the southeast of modern Baghdad, Iraq (see Section 5.9) providing a geological provenance so this agrees with the trace element results. Therefore we can suggest that WIJ 35, WIJ 37 and Sust 28 originated from the eastern zone of Western Asia. Dev 37 clusters closely with samples from Damascus in both plots and 343 The approach was taken by Henderson et al. 2016. 344 Henderson et al. 2016. 98 — therefore derives from the Levantine region. Moreover, it can be suggested that Sust 3 was made in the eastern zone and Sust 4 possibly northern Syria, though this is not entirely clear. 5.11 Summary The analytical results for the early medieval glass samples studied have provided new insights, especially into two aspects of the glass used and manufactured in the Netherlands, namely the highly coloured opaque monochrome glass beads produced in 6th-7th century AD Netherlands and changes in the glass supply in the period between the late 4th century and mid-11th century AD. The analysis of the materials used to make coloured beads from the Jodenstraat site in Maastricht has confirmed the previously suggested two-step manufacture mode for coloured bead making (see Section 5.10.1) and on-site lead-tin yellow colorant production at the site (see Section 5.10.2). Our new understanding from analysing these materials is that ‘pristine’ Foy 2 glass was used as the base glass for making beads at Jodenstraat (see Section 5.10.1) and that the tin opacifiers, to make both yellow and white glass, used in making coloured beads were also produced on-site (see Section 5.10.3). The majority of glass analysed was imported, and our aim in analysing it was mainly to identify its chemical compositional group. We have managed to group the glass according to its compositional features for the majority of the glass analysed and we have also identified shifts in glass supply in the Early Medieval Netherlands according to changes in glass compositional groups over time (see Section 5.10.4 and Section 5.10.5). The political and economic factors probably responsible for the shifts in glass supply to the Early Medieval Netherlands are addressed in Chapter 6, along with evidence from other historical studies. 99 — 6 Synthesis and conclusions 6.1 The Early Merovingian period (450-550 AD) McCormick345 considers that the study of textiles, relics and coins, but also glass and ceramics, are a good way to disentangle the complexities of the early medieval economy. Our scientific results for glass help to unravel some aspects of the early medieval economy in the Netherlands and build on the established picture for the production and supply of early European glass. The early Merovingian period dating to between 450 and 550 AD saw important economic, social and cultural developments under the Pippinids in the Meuse valley. The glass in use was essentially a continuation of the Roman tradition but with a reduction in the range of production techniques for vessels. It mostly involved translucent blue-green nearly colourless hues for vessels and there was a continued import of glass beads from the Mediterranean. There is some evidence that glass beads were made in Cologne. Bowls, cups, bottles and cone beakers sometimes with trail decoration, including the famous Kempston type,346 date to this period. There is evidence for a workshop at Huy in the Meuse valley where glass was worked but no evidence for the types of objects/vessels made there has been found.347 There is also evidence for Rhenish production348 with possible evidence in Cologne349 and concentrations of early Frankish vessels around Mayen (Eifel) might suggest a production centre in the area.350 Furthermore a possible production site for 5th century Helle bowls has been found in western Germany at Asperden351, quite close to Gennep, where some were discovered.352 It is possible that glass vessels were made in the proto-urban centre of Maastricht at this time.353 So the glass vessels found at Maastricht and certainly those found at Gennep studied here would have been imported from one or more of these places. The glass that we have analysed dating to this phase mainly derived from Gennep, with two vessels from Wijnaldum. By determining major, minor and trace levels of element oxides in glass vessels from Gennep dating to between the late 4th and mid 6th centuries we have found the use of (pristine) HIMT sensu stricto, Foy 2, and quite early examples of recycled Foy 2. HIMT glass was probably made in Egypt354 from the mid 4th to the 5th centuries and Foy 2 also probably had an Egyptian origin from the second half of the 5th and the 6th centuries. Therefore, raw furnace glass would have been imported to centres on the Rhine and Meuse to be remelted and worked/blown into vessels. Glasses would also have been recycled on these secondary production sites though no crucibles in which evidence for such glass mixing have been found. There are some interesting relationships between compositional type, colour and vessel type for Gennep vessel glasses. Four cone beakers are made from pristine HIMT, three of which are decorated with spiral coils below the rims. Seven other cones were made from recycled Foy 2 glass. Whereas the Foy 2 glass is very pale green ‘colourless’ or pale green, the HIMT glass cones are olive-green, yellow-green and amberbrown due especially to higher levels of iron and manganese. Pristine Foy 2 was used to make four bowls with vertical loops below an outfolded rim, similar to the decoration on Kempston cones; pristine Foy 2 was not used to make cone beakers in our data set. It appears therefore, that a specific supply of pristine Foy 2 glass was used to make the bowls we have analysed from Gennep; single examples were made from pristine HIMT and recycled Foy 2 glass. Of the five bowls from Gennep decorated with an opaque white feather pattern that we analysed, four are of recycled Foy 2 glass and the fifth made from pristine Foy 2. Both cone beakers and bowls were made with nearly colourless (very pale green) recycled Foy 2 glass, including the two cone beakers from Wijnaldum dating to this phase, one a Kempston cone. Therefore, the majority of cone beakers at this site/time were made from pristine pale green Foy 2 glass or more deeply coloured yellow-green and brown pristine HIMT glass. In contrast the majority of ‘colourless’ bowls decorated with feather patterns and some ‘colourless’ cone beakers were made from recycled Foy 2 glass. Perhaps unsurprisingly this suggests that colour was an important consideration when it came to pale green or nearly colourless drinking vessels made from Foy 2 glass, where the colour of the liquid could be observed depending on whether beer or wine was being consumed. Nevertheless, the use of strongly coloured olive-green, yellow-green and amber-brown Egyptian HIMT glass to make cone beakers suggests something else. It is possible 345 346 347 348 349 350 351 352 McCormick 2001, 281. Evison 1972. Van Wersch 2013. Koch 1987. Dodt, Kronz & Simon 2021. Sablerolles 1993. Brüggler 1994. Sablerolles 1992, Sablerolles 1993, Henderson 2000, 68-70. 353 Van Lith & Sablerolles 1995. 354 Arles et al. 2019. 100 — 355 Biblioteca Apostolica Vaticana, Rome, 356 357 358 359 360 361 362 363 364 Ms. I. Cod. Reg. Lat. 438, 25r (https:// digi.vatlib.it/view/MSS_Reg. lat.438/0053). Bibliothèque nationale de France, Paris, Département des Manuscrits, Lat. 8085 fo 61v. https:// portail.biblissima.fr/ark:/43093/ mdata69093fcda745577d3ff9d21597dc3fc4c51ca346. Special Collections University Library Leiden, Codex Burmanni Q 3, folio 120v. (https://disc.leidenuniv.nl/view/ item/1935754?solr_nav%5Bid%5D=22d3ef11949c6ef5312f&solr_ nav%5Bpage%5D=0&solr_nav%5Boffset%5D=0#page/138/mode/1up). See for instance Arbman 1937, 41-44, fig. 5a-b; Gaut 2011, 194, 255-265. Aunay, et al.2020, 298 (Paris, BnF. Ms. Lat. 1, f327v). Later 2010; Van Winkelhof 2021. Van Winkelhoff 2021. Henderson, Sode & Sablerolles 2019. Pion 2014. Boschetti, Gratuze & Schibille 2020. The Rural Riches project funded by the European Research Council (ERC) of the European Union, and directed by Professor Frans Theuws is investigating the sources and distribution of such beads. Matin 2019. that glassblowers used whatever came to hand but we have nevertheless found evidence for a relationship between vessel type and colour so probably some vessels were made in batches using raw glass of the same colour. Unstable Merovingian vessels such as bell beakers, palm cups and deep palm cups as well as Carolingian funnel-shaped and pointed conical beakers are usually interpreted as drinking vessels which had to be emptied before placing them upside down. Indeed, illuminations in several Carolingian manuscripts show the use of 9th century glass cones as wine glasses.355 In Viking period graves and settlements the association between glass funnels and cones and fine ceramic tablewares such as Tating jugs and Badorf pitchers also indicates a drinking function.356 However, an illustration in the First Bible of Charles the Bald, also known as the Vivian Bible, after Count Vivian of Tours who commissioned the bible in 845, shows the use of individually suspended glass cones as lamps.357 It appears that these Carolingian glass vessels were multi-functional. The same is probably true for unstable Merovingian glasses with rounded bases. Experiments with bell beakers, palm cups and deep palm cups have led researchers to believe that all these vessel types could have been used as lamps.358 Van Winkelhoff found that mould-blown and optic blown vertically ribbed glass vessels were especially effective in that they emitted bright and clear “sunlike” patterns and that small mould-blown palm cups work especially well as lamps when hung at a low level above the ground.359 She concluded that such usage is especially relevant in the context of a grave lamp, or votive lamp, hung as a visual reminder of the deceased at the grave. Although raw pristine HIMT glass imported from Egypt would undoubtedly have passed through intermediaries before being blown into vessels, perhaps its exotic origin was still known and was socially and/or ritually significant depending on whether it was used in a domestic context or in a burial. Six beads and a greenishwhite glass breaking splinter from Wijnaldum date to this phase. Two beads are colourless with gold and silver-foil respectively. They appear to have been made using pristine Foy 2 glass, adding to the evidence for a much higher proportion of pristine glass in circulation in the early Merovingian period than in the Carolingian period. Out of the six beads a single opaque yellow one is opacified with lead stannate so dates to before the evidence of the production of such glass in Maastricht; the breaking splinter is opacified with calcium antimonate in the Roman tradition found in tesserae.360 6.2 The Middle Merovingian period (550– 650 AD) In our second period dating to between 550 and 650 AD the production of glass vessels became somewhat diminished, with a reduction of vessel types. For some reason beads were no longer imported from the Mediterranean and this seems to have led to the birth of Merovingian bead production,361 which saw massive numbers of glass beads being manufactured with very similar designs. These were distributed across Europe between the Anglo-Saxon realms, the Merovingian territories as well as the Frankish kingdom of Italy362 as late as the end of the 7th century. So, much imported glass would have been used to make beads.363 In the Netherlands, glass beads, especially opaque yellow ones, were being manufactured in Maastricht, with especially good evidence for their production dating to between the late 6th and early 7th century at Jodenstraat, Maastricht, located well within the Merovingian empire. Comprehesive evidence for bead making has also been found at the central places of Rijnsburg Abdijterrein on the Rhine delta and WijnaldumTjitsma, a northern terp site. Both of these two sites are located well outside the border of the Merovingian empire so there was clearly a demand for fashionable beads outside the empire; a local ruler may well have invited a bead maker to the sites. The bead making evidence at all three sites dates to the last quarter of the 6th and first half of the 7th centuries AD. We have found evidence for the manufacture of lead tin yellow (lead tin yellow II) for colouring glass in a number of crucibles at Jodenstraat, Maastricht. The occurrence of silica crystals associated with this yellow pigment, as well as pure lead-silica glass in eight crucibles, is clear evidence for the second step in the production of lead tin yellow II when it was added to the lead tin calx.364 The first step in the production of lead tin yellow II – the production of lead tin calx 101 — – would have happened elsewhere (perhaps even in Maastricht), but not in the crucibles we have examined. The same evidence for the manufacture of lead tin yellow glass in a tray, dating to between 575 and 625, was found at Wijnaldum. The yellow pigment was made in what was originally 6th–7th century Merovingian wheel-thrown ovoid-shaped domestic coarseware storage pots which would have had a constricted opening and an everted rim. The top half of the pot was taken off for the production of the yellow pigment. This production process for opaque yellow glass is described in later Islamic literature, the earliest being Abu’l Qasim Kashani dating to 1301.365 No lead-silica glass is in use for the manufacture of vessels or, by itself, for the manufacture of beads at the time. We have found direct scientific evidence linking the crucible yellow glass to the yellow beads and yellow rods at Maastricht so they are clearly part of the production process there. The yellow beads and rods are united compositionally because a pristine Foy 2 base glass was used to make them. Translucent glass vessel fragments and wasters from Maastricht Jodenstraat were also made with pristine Foy 2 natron glass. Had the Roman vessel glass fragments found at Jodenstraat been used as a base glass instead this would have been detected, but this is not the case. Although similar evidence for the production of opaque yellow lead tin colourant has been found elsewhere,366 including at Wijnaldum (discussed here), the Maastricht evidence constitutes the best evidence for its production in terms of its scale and for its use for making beads in northwestern Europe. Quite why there was this demand for yellow glass beads is an intriguing question. No evidence for the production of opaque yellow glass in crucibles has been found in Carolingian contexts in the Netherlands. Further new evidence for the production of opaque glass is for the manufacture of a tin white opacifier, found in two crucibles from Maastricht, Jodenstraat. This is the first evidence for this from northwestern Europe. We have also demonstrated that it was used to make the white, red and greenish glass beads at Jodenstraat, including their characteristic microstructures and the use of the same base natron glass, Foy 2. No crucibles in which the white pigment was mixed with base glass have been found. Whole pots, for which there is production evidence from Maastricht itself, were used for working translucent glass from both the Jodenstraat and Mabro sites in Maastricht, and also in Utrecht. Detailed investigation of the ‘frit-like’ material observed on the rim of a crucible dating to the late 4th-early 5th century from the Mabro site in which translucent greenish glass was reheated, which would be tentative evidence for primary glass production367, has instead been shown to have a variable composition and to be fuel-ash slag. Seventeen of the beads from Wijnaldum date to this phase. Like the Maastricht Jodenstraat beads they were made from highly coloured opaque yellow, white and red glass. Their leadrich chemical compositions, the colourants and the opacifiers used, as well as the evidence for the use of a pristine base glass, are all very similar to Maastricht beads. Two later (8th–9th century) gold and silver-foil decorated colourless plant-ash glass beads from Wijnaldum are discussed below. Bead makers were evidently located in proto-urban or urban centres in the Meuse valley. This is in contrast to the situation further north where the relatively small scale of production suggests that bead makers travelled to centres like Wijnaldum, Rijnsburg and perhaps Valkenburg-De Woerd. Evidence for the manufacture of very popular Merovingian bead types with crossed swag decoration have been found at Rijnsburg, one of the types being found as far south as Italy. Callmer368 has suggested that glass bead production was regarded as having a magical aspect at the time. If magic was considered as important, perhaps this is one reason why beads were no longer imported. 6.3 The Late Merovingian period (650 – 750 AD) By 650–750 AD fewer glass beads were made; there was a restricted range of rather poorly made vessel forms, such as palm cups and deep palm cups usually made with poor quality bubbly glass full of inclusions, reflecting the high level of recycling. A further reduction in imported beads was a catalyst for more beads to be manufactured, especially in Scandinavia. The artefacts that we 365 Allan 1973; Matin 2019. 366 Henderson & Ivens 1992; Peake & Freestone 2014. 367 Sablerolles, Henderson & Dijkman 1997. 368 Callmer 2003. 102 — have studied from this period are four opaque (yellow, red, orange and white) glass beads from Wijnaldum though these beads were probably made earlier than their context dates. 6.4 369 370 371 372 373 Lassaunière et al. 2016. Van Versche et al. 2015. Pactat et al. 2017. Pactat et al. 2017. Foy et al. 2003. The Carolingian period (750 – c. 850 AD) The Carolingian dynasty (750–887 AD) saw a renaissance in vessel production, especially with the use of new decorative techniques particularly in northern France, but also with clear evidence for the manufacture of pale green beakers from the Rhenish area with an expansion in the scale of production. It is also possible that highly coloured vessels were made in monasteries in the Netherlands. The vessels produced in France included globular jars and reticella decorated beakers: there is evidence for the production of lead-tin yellow pigment and its use in reticella rods to decorate glass vessels from the early 8th century site of Hamage, northern France.369 This must have created different markets for both simple and more highly decorated vessels. Although the glass-working evidence for the Carolingian period is scantier, a contrast with the Merovingian period is that it occurs on a wider range of site types, including the emporium of Dorestad and the ecclesiastical centres of Susteren-Salvatorplein and Utrecht-Domplein. Wearing glass beads became less fashionable in the Frankish heartlands though they were still worn in the northern periphery of the empire. From the end of the 8th century onwards Islamic glass beads were imported via Viking trade networks and occur in settlements along the Rhine and in cemeteries north of the Rhine. The availability of Islamic glass beads may well have impacted on the manufacture of Frankish glass beads. By this time the evidence for glass recycling had increased, with a much higher proportion of weak HIMT/recycled Foy 2 glass in circulation, with few examples of HIMT or other pristine glass types. Our analyses of fifty-five samples of palm cups, palm funnels (including a gold-foil palm funnel), bell beakers, funnel beakers and a bowl from Dorestad, as well as a Kempston cone, a bowl and four funnel beakers from Wijnaldum, show that, with a few exceptions, those who made these vessels relied on a supply of recycled (Foy 2) natron glass. The same is true for the Carolingian vessel glass from Susteren. Therefore we have assembled very strong evidence for recycled Foy 2 glass being the dominant glass type in the 8th and 9th century Netherlands, with pristine glass, especially Foy 2, having almost gone out of use. The few exceptions are the use of Egyptian II glass used to make a possible pale green funnel beaker from Dorestad, and the latest dated funnel beakers from Wijnaldum (770–900): one is a dark blue incalmo rim, the other a bluegreen colour. A single pale green funnel from Susteren was made with pristine HIMT. The only pristine Levantine (II) glass found in this study is a turquoise punty dating to 750–800 from Wijnaldum. So even when recycling was such a dominant practice at this time, a small amount of pristine glass was in circulation and would have been imported in raw form and made into funnel beakers at production centres. This was a period of technological transition when the first European wood-ash glass was being manufactured and used to make vessels such as those from Baume-les-Messieurs,370 and the possible production of mixed-alkali glass by extending natron glass with wood-ash (glass) at Méru, France.371 Reflecting this period of technological transition, two wood-ash-lead linen smoothers (partly made using slag from silver smelting)372 and two yellow-green woodash palm funnels derive from Dorestad, and a single piece of wood-ash blue-green window glass from Susteren. A single mixed-alkali pale green funnel was also found at Susteren. Raw wood-ash and mixed-alkali glass used to make the vessels would have been made more locally. By using trace element analysis we have demonstrated that there was an increase in the levels of, for example, potassium and phosphorous oxides over time. Even though these are initially at low levels in the 8th–9th centuries, we have suggested that this indicates that a small proportion of wood-ash was being mixed into the (recycled) Foy 2 glass (referred to by Foy et al. as Foy 2.2)373 from the late 8th century and into the Carolingian empire. This is supported by the occurrence of elevated trace levels of cesium, barium, rubidium and strontium in Foy 2 glass after the 8th century, all characteristics of woodash glass. We have also demonstrated that this is the case with the first neodymium and strontium 103 — isotope analyses of recycled Foy 2 glass and observed that a probable explanation for an increase in the strontium isotope ratio (when compared to pristine Foy 2 glass) can be attributed to the mixing in of a small proportion of woodash glass. These compositional and isotopic results reflect the emergence of the first European-made (wood-ash) glass technology. The main evidence for early wood-ash glass production zones, based on the occurrence of dated glass objects, is in France and Germany. Previous research has pointed to potential evidence for the addition of wood-ash glass to natron glass in Anglo-Saxon vessels from Jarrow,374 and Ares et al.375 have noted that a ‘plant ash’ component must explain the elevated levels of potassium, magnesium and phosphorus oxides above those found in natron glasses. Nevertheless, elevated levels of these elements could also be introduced into glasses with the addition of a small proportion of fuel-ash slag, also ultimately with a wood-ash component, as seen in the analysis of material attached to a crucible rim from Maastricht Mabro (crucible 9). It is however more than a coincidence that increasing levels of potassium and phosphorus, and especially elevated concentrations of cesium, rubidium, barium and strontium are correlated with a time when wood-ash glasses were being introduced in Europe, so this is a far better explanation. Window glass with a full wood-ash composition has been found at the monastery of Baume-les-Messieurs, Jura in France376 dating to the late 8th century. It is possible that the presence of small proportions of wood-ash (glass) in recycled Dutch early medieval natron glass resulted partly from wood-ash glass production further south in Belgium and France – and we suggest that a source to the east in Germany as mentioned in Chapter 2.4.1 is more likely. There is no early medieval wood-ash glass making evidence from the Netherlands. We have compared our results with those from Comacchio in northern Italy377 and observed some intriguing differences. In Comacchio a far higher proportion of pristine Levantine glasses was identified (17/77). This compares with a single example amongst our data: a punty glass from Wijnaldum. The mixture of Levantine and Foy 2 glass from Comacchio led to an ‘intermediate’ group being recognized. No examples of woodash glass were found at Comacchio, nor evidence for its mixture with natron glass. We are therefore the first to define these contrasting production spheres using the characteristics of recycled glasses: a southern European one exemplified by Comacchio glass and Spannish glass from Tolmo de Minateda378, with far greater evidence for the use of Levantine glass, and a northwestern European Dutch one with a reliance on an admixture of wood-ash glass and no apparent evidence for mixing with Levantine glass. Our approach could be used to define recycling traditions, reflecting trade links, in other parts of Europe. The manufacture of the first European (wood-ash) glass would have partly been driven by the demand for windows in monasteries and churches: by the 10th century it is therefore no coincidence that wood-ash glass was being worked at the ecclesiastical site of La Milesse, Sarthe in France. In our study nine out of the eleven window glasses analysed from the monastic site of Susteren are of a natron composition – seven are recycled Foy 2, and the single example of Egyptian II noted above. This is further evidence for very different glass supply in the two areas. From the early 9th century another type of glass was made, from ashed halophytic plants and sand in Islamic cosmopolitan centres in western Asia. Although the Sassanids had made plant-ash glass earlier there is no current evidence that it was used to make glass objects in western Europe. Trace element analysis has shown that the manufacture of plant-ash glass by the Muslims formed part of a decentralized production system379 and that it is possible to link quite securely the provenance of Islamic plant-ash glasses to production centres or zones. A gold-foil bead (dated 775–850 AD) and a silver-foil bead (dated 875–900 AD) from Wijnaldum are plant-ash glasses imported from western Asia; the bodies of two rather unique trail-decorated conical beads and a blue-green funnel from Susteren are also made of plant-ash glass. The Susteren beads and beaker would have been made in the west using raw plant-ash glass imported from western Asia. A probable funnel beaker fragment of a west Asian plant ash composition from Deventer was found in a context dated to between 950 and 1050 AD. Trace element analysis shows that the metal foil beads from Wijnaldum were probably made in 374 375 376 377 378 379 Freestone & Hughes 2006. Ares et al. 2019. Van Wersch et al. 2015. Bertini, Henderson & Chenery 2020. Schibille et al. 2022. Henderson et al. 2016. 104 — 380 381 382 383 384 385 386 387 388 389 Kronz et al. 2016. Thedéen 2009. Philippsen et al. 2021. Langbroek 2021b. Henderson & Holand 1992. Gaut & Henderson 2011. Pactat 2021. Schibille et al. 2020. Neri, Gratuze & Schibille 2018. Pers. Comm Professor Kronz. the eastern zone of Western Asia, in Iraq/Iran. The glass used to make the two Susteren beads probably derived from Iran/Iraq, and possibly northern Syria, respectively. One of the Susteren beads (Sust 4) is decorated with opaque white glass produced in the Roman tradition (calcium antimonate crystals) used almost universally in Roman glass tesserae. Therefore the bead combines an intriguing combination of western European and west Asian traditions. The raw plant-ash used to make the Deventer vessel (fragment) derived from the Levant. The import of raw plant-ash glass probably formed part of the Viking trade network via centres like Hedeby. Though only plant ash glass beads have been reported so far, the occurrence of mixed plant ash and lead glass at Hedeby provides indirect evidence for the import of raw plant ash glass.380 The occurrence of other Islamic artefacts made from plant-ash glass is further evidence of such a trade network. Examples are the import of large numbers of early Islamic glass beads to Scandinavia from the late 8th century and later in Viking-age contexts such as in Gotland burials,381 from precision-dated excavations at Ribe, Denmark382 and millefiori decorated glass beads from Dorestad.383 No examples were found at Borg in Norway384 but five plant ash glass beads were found at Kaupang.385 Early Islamic glass beads are found along and to the north of the Dutch Rhine, along the German Wadden sea and on the Baltic coast. However minimal numbers have been found along the German Rhine or in Belgium and France. An exception is the occurrence of 9th century small glass bottles made from plant-ash glass imported from Islamic lands,386 which were probably used for the import of specific western Asian liquids. The relative rarity of Islamic plant ash glass in France at this time may be because the demand was lower, due to the greater availability of wood-ash and mixedalkali glass. However, a better explanation is that France and Belgium did not form part of the Viking trade network that existed to the north. Further south, in Umayyad Spain, plant-ash glass was imported from western Asia in the 8th and 9th centuries; 387 Islamic glass beads dominated amongst those found at Illyricum, Albania.388 6.5 The late phase, including the Ottonian period (c. 850 – c. 1000 AD) The last phase (850–1000) was a time when very few glass beads were produced in the Netherlands apart from evidence of small-scale production at Dorestad. It includes the Ottonian Dynasty (919-1024 AD). Beads were imported from Scandinavia and continued to be imported from Islamic western Asia. This phase is represented in our research by glass from Deventer. The site produced a wide range of glass compositional types: nineteen wood-ash glasses, one pristine and five recycled Foy 2 glasses, four mixed-alkali glasses, three Egyptian II (natron) glasses, three Roman (natron) glasses and one plant-ash glass. A single example of high lime -low alkali window glass is potentially a very early example of what is generally considered a much later technology though it appears there are other early examples from east of the Rhine.389 Funnel beakers and conical beakers were the dominant vessel forms in the second half of the 9th and the 10th centuries. By this time the vessels from Deventer were made out of wood-ash glass, recycled Foy 2 glass, plant-ash glass and Egyptian II glass. This shows that they were blowing these vessels from whatever was available; unless the different sources of natron glass were known (which is a possibility for pristine glass) the different origins of pale green recycled natron glass – potentially with different working properties - would normally be unknown. The wood ash and natron glasses would certainly have had different working properties: thickwalled wood ash glass vessels started to replace the thin walled Carolingian beakers after c. 900 AD. A single chunk of raw Egyptian II and three chunks of raw wood-ash glass were found on the site. This could constitute tentative evidence for a glass industry there, or it could simply mean the glass was being traded through the site. The use of fresh high quality Egyptian II glass for the manufacture of funnel beakers is in contrast to the preceding phase when a very high proportion of recycled glass was in use, for example to make the funnel beakers found at Dorestad. Apart from the four definite examples of wood-ash funnel beakers from Deventer (two dating to the late 9th century), there is an 105 — 8th–9th century palm funnel from Dorestad,390 as noted above, with another nine examples from Hedeby.391 As is to be expected, many were very poorly preserved. The balance of the Hedeby funnel beakers analysed were twenty natron (Foy 2) and one mixed-alkali glass. In contexts dating to the 9th and 10th centuries at Deventer 60% of the glasses were wood-ash or mixed alkali glass. These would have been manufactured as part of a decentralized production system.392 Some of the earliest full wood-ash glasses have been found at Stavelot in Belgium and at Baume-les-Messieurs, Jura in France. It also seems to be the case that France was a centre for the production of mixed-alkali glass (and perhaps wood- ash glass), one probable location being Méru, France. In contrast to Deventer, by the 10th century almost all glass found in France was of the wood-ash type.393 Currently we do not know where the Deventer mixed-alkali or wood-ash glasses were made. Although France is one possible source for Dutch wood-ash glass a more likely one was the Viking trade network including through Hedeby, perhaps from northern Germany, for example, where funnel beakers and a crucible containing wood-ash glass have been found.394 The higher proportion of non wood-ash glass found in 10th century Deventer, including pristine Egyptian II glass – and possible working debris- also contrasts with the situation in France, suggesting that a separate supply route from Egypt was involved, probably including several intermediaries. 390 Henderson 2012. 391 Kronz et al. 2016. 392 Meek, Henderson & Evans 2012; Van Wersch et al. 2015; Adlington et al. 2019. 393 Pactat 2021. 394 Kronz et al. 2016. 107 — 7 Answering the research questions In the introduction to this book a series of research questions were posed. In this chapter we discuss the extent to which we have been able to answer them. site. Nevertheless it would seem that the cobaltrich colorant was added directly to the pristine base glass at some point. 2. 1. What raw materials were used in the local production of simple, monochrome Early Medieval glass beads? We have been able to identify the base glass used in the manufacture of monochrome yellow, white, red, and blue beads as a type of natron glass, Foy 2 (see Section 5.10.1). No base glass was fused from raw materials in the early medieval Netherlands, so it needed to be imported (see below). The majority of the material examined scientifically that is relevant to this research question forms part of some of the most comprehensive evidence for the manufacture of such beads in early medieval northwest Europe, from Merovingian Maastricht at the Jodenstraat site dating to the late 6th-early 7th century. The evidence consists of beads, broken beads, rods, drops - and crucibles with evidence for the manufacture of the opaque yellow and white pigments attached to them. The 26 trace element signature of the glasses, determined by laser ablation inductively coupled plasma mass spectrometry, showed that they are unified compositionally by being made with the same base glass: ‘pristine’ Foy 2. The other raw materials which both coloured and opacified the glasses used to make the monochrome beads are the colorants. Tin-based pigments were made at Maastricht which were then used to make bright opaque yellow and white glass. The tin based opacifier was also found in red and greenish-blue coloured glass beads which, in all likelihood, were also made in Maastricht. A thick possible furnace or tray fragment from Wynaldum also has a thick layer of opaque yellow material attached to: it provides probable evidence for the manufacture of leadtin yellow II pigment. Monochrome beads would have been made there too. A separate source of lead would have been involved. Given that pristine Foy 3.2 glass (a sub group of Foy 2) was used to make the translucent blue bead and the four blue glass fragments found at Maastricht it is a possibility that the cobalt colorant was added to the base glass in Maastricht though, unlike the evidence for the manufacture of lead tin oxide II, there is no archaeological evidence for the manufacture of cobalt blue glass at the Where were these raw materials obtained from? Glass that has a Foy 2 trace element signature is very likely to have derived from Egypt. 395 It was made with silica sources that are characterised by the presence of minerals that introduced elevated elements such as titanium, manganese and iron. It is difficult to be certain where the tin used to make the opaque yellow and white glasses was derived from. Possible sources are in Cornwall in the UK396 and Turkey.397 Lead is a far more common mineral with possible sources in northern Spain, central and southern England, the Saxon-Bohemian metalliferous mountains (including the Erzgebirge) and Harz mountains in Germany398, northern Italy and the Taurus mountains in Turkey. Lead isotope analysis has the potential to determine in an increasingly precise way the source(s) of the lead used when used in an appropriate way such as used for European iron age glass399 and in ancient metal research.400 The lead source used in early medieval Dutch glass would be introduced either as an impurity401 or deliberately as part of a colorant, such as lead-tin yellow II. Therefore, lead isotope analysis could potentially source the lead raw material, but not the glass, as discussed below. Sub-questions here are: i. What substances were used to make the different colours of glass in the artefacts tested? As noted above the monochrome opaque beads from Maastricht, Jodenstraat were mainly coloured with lead-tin oxide II and tin oxide. Tin oxide was also combined with copper and cobalt to produce opaque red and blue glass respectively. Examples of yellow glass opacified with lead-tin oxide II have also been identified in glasses found at Wijnaldum and Dorestad. Detailed analysis of the opaque red glass from Maastricht, Jodenstraat revealed the presence of micron-sized copper droplets, iron-rich fayalitic slag402 and a crystalline phase containing high tin associated with lead and silica: the colorant is copper. The presence of micron sized copper droplets and fayalitic slag was also found in red beads from Wijnaldum. 395 396 397 398 399 400 401 402 Foy et al. 2003. Meharg et al 2012. Yener et al. 2015. Wedepohl & Baumann 1997. Huisman et al. 2017. Artioli et al. 2020; Standish et al. 2021. Henderson et al. 2005b. Peake & Freestone 2012. 108 — The pale green vessel glasses analysed were coloured mainly with a combination of manganese and iron oxides: if HIMT or one of its recycled variants was used as the base glass, elevated manganese (and iron) would have modified the final colour, partly depending on the melting atmosphere in the furnace. For example HIMT glass cones form Gennep are olive-green, yellow-green and amber-brown due to relatively high levels of iron and manganese. Other vessel glasses were coloured with cobalt producing a blue colour and ferrous iron has produced an amber colour. Deliberately added colorants/ opacifiers for glasses in both Merovingian (Wijnaldum and Maastricht) and Carolingian glasses (e.g. Susteren glass beads) are very similar: elevated Fe oxide (up to around 5 weight %) and CuO2 (up to around 1.7%) in red glass, high PbO and SnO2 in opaque yellow (probably in the form of PbSnO3 crystals) as well as a combination of Pb and Sb which responsible for opaque yellow glass (in the form of Pb2Sb2O7 crystals). Opaque turquoise and blue tesserae from Dorestad are coloured by copper and cobalt respectively. Three of the five are opacified with calcium antimonate. An opaque yellow glass rod from Dorestad (DOR149) has a similar chemical composition to opaque yellow glass from Maastricht, Jodenstraat and Wijnaldum. It has relatively high Na2O (10.6%) and PbO (10.7%) contents and low K2O and MgO contents: it is also coloured by lead tin yellow II. Therefore, it would have been made using the same procedure as the opaque yellow glass from Jodenstraat and Wijnaldum. The Al2O3/SiO2, TiO2/Al2O3 ratios and low Sb content of suggest that the base glass used could also have been ‘pristine’ Foy 2 glass. ii. 403 Foy et al. 2003. 404 Nenna 2014. What compositional groups can be distinguished in the glasses based on chemical analyses? The compositional groups that we have identified amongst the glass samples that we have chemically analysed are all known from the literature (see Section 2.4.1 for a full discussion of the glass types and associated literature). By chemically and isotopically analysing recycled Foy 2 (natron) glasses which was the dominant composition between the early 8th century and late 9th century we have been able to provide a new explanation for some of the impurity levels detected in the glass in a new way (see Sections 5.9 and Section 5.10.4). We have identified the following compositional types: Natron (soda-lime) glasses: • ‘Roman’ • High iron, manganese and iron (HIMT) • Foy 2.1 • Foy 2.2 (recycled glass) • Foy 3.2 • Egyptian II • Levantine II Plant ash (soda-lime) glass Mixed-alkali (sodium and potassium) glass High potassium glass iii. What does this tell us about dating of primary glass production of these groups? There is no evidence for primary glass production in the early medieval Netherlands. It is known that glass of a ‘Roman’ composition was made between the 1st and 4rd centuries AD on the Levantine coast and in Egypt. Glass of the HIMT composition was made between the 4th and 5th centuries and it is probable that weak HIMT (HIMT-2 = Foy 2) was also made from around the mid 4th century though the recycled variants of HIMT/Foy 2 glasses have been found in much later contexts (see below) so will probably have been recycled multiple times. Glass of Foy 2.1 and 3.2 compositions were probably made from around the 6th century, the recycled Foy 2 (Foy 2.2) has been found in later contexts. Pristine Levantine II glass was made from the 8th century AD; Egyptian II was made between 760/780 and 870 AD. Five Roman (vessel) glasses have been identified, from Jodenstraat Utrecht and Deventer. Our analyses suggest a Levantine source for this relic glass. A single Levantine II sample has been identified amongst our samples dating to between 750 and 850. Nintety-five mainly vessel glasses are of a recycled Foy 2 glass composition (corresponding to the Foy 2.2 group of Foy et al.).403 Highly coloured opaque beads and translucent beads from Maastricht, Jodenstraat were made with pristine Foy 2 glass, as were the highly coloured opaque glass beads from Wijnaldum. Nine HIMT sensu stricto glasses have been identified404, six being from Gennep, the site with the earliest date (late 4th to mid-6th century) from which we have obtained samples. Eight ‘pristine’ Foy 2 glasses have been 109 — identified: six from Gennep and 2 from Maastricht, Jodenstraat, dating to between the late 4th and early 7th centuries AD. Six ‘pristine’ Egyptian II glasses have been found: two from Wijnaldum, one from Dorestad and three from Deventer. Using context dates, three from Deventer date to the 8th–9th centuries AD, two date to the early 10th century and one dates to between 950 and 1050 AD. They are consistent with or slightly later than the suggested date when Egyptian II glass was widely circulated, in the 8th and 9th centuries AD. Previously it had been suggested that HIMT glass in the widest sense (including HIMT sensu stricto and Foy 2) was in use from the middle of the fourth century until the seventh century.405 However, our results show that recycled Foy 2 glass (referred to as ‘recycled Foy 2’ glass in this study, corresponding to Foy 2.2 group)406 was still the dominant compositional group between early 8th century and late 9th century in Dorestad and Wijnaldum vessel glasses. Moreover, it remained in circulation till possibly as late as the first half of the 10th century according to our results from Deventer. This shows that recycled Foy 2 glass was in use centuries after the ‘pristine’ Foy 2 glass supply dried up in northwestern Europe. The recycled version of Foy 2 glass is labelled as Foy 2.2 subgroup in the work of Foy and colleagues,407 and is known from a very limited number of assemblages in France, Italy and Spain that are typically dated to the end of the seventh and into the eighth centuries AD.408 Twenty six wood ash/mixed alkali glasses have been found in this study. Two samples derived from Dorestad, two from Susteren and twenty-three from Deventer. These dates are consistent with the suggested date from when wood ash glass started to be made and was circulating in Europe from the late 8th century onwards. Six plant ash samples have been identified; all date to post-9th century AD the time when plant ash glass started to be the dominant technology in western Asia. Islamic plant ash glasses started to appear further west and east after this time. Plant ash glass was used to make single funnel beakers from Susteren and Deventer, the bodies of two decorated traildecorated beads from Susteren and it was formed into single examples of gold and silver foil beads found at Wijnaldum, imported from western Asia. Raw plant ash glass was therefore imported from western Asia to the west where it was remelted to form funnel beakers and some (trail-decorated) beads. iv. What do the isotope ratios (Sr, Nd) obtained from the glasses of selected compositional types tell us about the their origin and dating? Thirteen natron glass samples have a typical range of Nd–Sr signatures for such glass with εNd between -5 and -7, and 87Sr/86Sr between 0.7085 and 0.7093,409 (Figure 5.24); one sample was contaminated (see below). The twelve clean natron samples are of the Foy 2 type: seven are ‘pristine’ Foy 2 glass and five are recycled Foy 2. Their 87Sr/86Sr ratios stretch in a rather wide range between 0.7085 and 0.7093. Six pristine’ Foy 2 glasses form a cluster at 87Sr/86Sr of 0.7086, a typical value for glass produced in Egypt, the commonly suggested origin of Foy 2 glass. Five ‘pristine’ glasses are from bead making at Jodenstraat, Maastricht (late 6th- early 7th century AD), the 6th being a gold foil bead from Wijnaldum. The 87Sr/86Sr of 0.7088 and 0.7093 for recycled Foy 2 (vessel) glasses all dating between the late 8th to mid 9th centuries AD is higher than the typical Egyptian range of values: it is likely that the 87Sr/86Sr values have been modified by the recycling process. We suggest that the mixing of wood ash glass, with very high 87Sr/86Sr values, with natron glass is the explanation; this agrees with the results from trace element analysis. We determined the Nd–Sr isotopic signatures for two wood ash-lead linen smoothers from Dorestad with very high 87Sr/86Sr. A single plant ash glass, an Islamic silver foil glass bead has the lowest 87Sr/86Sr signature out of the 18 samples analysed and may have a provenance near Baghdad in Iraq. A single ribbed glass vessel and a glass tesserae are of ‘Roman’ chemical compositions and both have the anticipated typical Levantine natron isotopic signatures with 87Sr/86Sr values of 0.7090 and 0.7092 respectively. v. What networks inside and outside the Netherlands were used in obtaining glass, including the colourants used? In all cases it is difficult to ascertain how many intermediaries were involved during the process of obtaining glass, whether it was unworked ‘pristine’ raw glass, recycled raw glass or fully formed objects. The colorants used deliberately 405 406 407 408 409 Freestone et al. 2018. Foy et al. 2003. Foy et al. 2003. Ares et al. 2019. Brems et al. 2018. 110 — (and their sources) which formed part of networks of interaction are discussed under question 2. The source of many of the Carolingian beakers found in the Netherlands is considered to be Germany/Cologne. There is clear evidence that opaque yellow (and other brightly coloured) glass beads were made at both Maastricht and Wijnaldum, but there is still a possibility that beads found at Wijnaldum originated from Maastricht, for example. Lead may well have been obtained from Germany too; the tin source may potentially have been Cornwall in the UK or, less likely, Turkey. Egypt was clearly the ultimate source for Egyptian II glass and pristine Foy 2 glass. Our isotopic results support an Egyptian provenance which, up to now, has mainly been suggested using the results of chemical analyses. There is little doubt that pristine HIMT glass was originally made in Egypt too, as recently confirmed using Nd and Sr isotope analysis, helping to distinguish it from Levantine glass.410 There is a single example of an early medieval pristine Levantine glass in our study, underlining that the primary source of pristine glass for early medieval glass working was ultimately Egypt. Plant ash glass would have been obtained from western Asia and using trace element analysis we have determined that northern Syria, Iraq/ Iran and the Levant are likely sources for the plant ash glasses identified in this study. Plant ash glasses were mass produced in early Islamic cosmopolitan centres such as Damascus, Baghdad and Samarra411 and started to appear on European sites in any number as a result of Viking trade networks. In the absence of glass making furnaces in the Netherlands, Belgium and northern France are possible sources of wood ash glass but a more likely source is Germany which formed part of the Viking trade network, including Hedeby in northern Germany. The probable evidence for mixed-alkali glass production in France suggests that this is one possible source for the type of glass in the early medieval Netherlands though again Germany may also have been a source. 410 Gliozzo et al. 2019. 411 Henderson 2022. vi. To what extent were the raw materials or semi-finished products derived from primary production sites, and to what extent did they derive from systematic recycling of glass, including Roman? In ‘pristine’ (non-recycled) natron glass the levels of a few correlated elements such as Pb, Sb and Cu, Ba, Rb, Cs are low or very low, but for recycled glass the levels of these elements are higher. Elevated levels of Pb and Sb most consistently demonstrate recycling has occurred, so they have been chosen here as one of the means of distinguishing ‘pristine’ glass samples from recycled glass samples. The criterion for the identification of a ‘pristine’ glass is that the Pb and Sb contents are both under 1000 ppm, following previous conventions. We have noted that there is a higher proportion of ‘pristine’ (unrecycled) natron glass imported as a raw material, however indirectly, from primary production sites found on Merovingian sites than on Carolingian sites. For example ‘pristine’ Foy 2 glasses from the late 6th-early 7th century Maastricht Jodenstraat site were used as the base glass to make highly coloured yellow, white and red opaque glass beads there and perhaps the translucent glass beads too. The same was found for the few semi-finished products from the site. The late 4th to mid 6th century glass vessels analysed from Gennep are pristine HIMT sensu stricto or ‘pristine’ Foy 2 glass. Nevertheless, elevated levels of Sb and Pb suggest there was an admixture of highly coloured Roman vessel glass or glass tesserae to some glass as early as this. In contrast if we use the results of 55 vessel glasses from Dorestad dating to the Carolingian period we have found that, apart from two wood ash glasses, and an Egyptian II glass, the remaining glasses are of the Foy 2 composition. These Foy 2 glasses contain at least 1000 ppm of Pb or 1000 ppm of Sb, or both, as well as elevated levels of Rb and Cs, characteristics of recycled glasses. By the c. 850 and into the 10th-11th centuries the site of Deventer provides an intriguing contrast to earlier periods, reflecting an important period of technological transition. From the glass that is sufficiently unweathered to provide a valid analysis nineteen out of thirtyeight samples are of a wood ash composition, twelve are natron glass, four are mixed-alkali and one is a plant ash glass. Therefore, wood 111 — ash and plant ash glasses are unrecycled; the mixed-alkali glasses are the product of recycling. The twelve natron glasses surprisingly include three of a Roman composition, three of a ‘pristine’ Egyptian II composition, one ‘pristine Foy II glass and five recycled Foy 2 glasses. Therefore the majority of glasses analysed from Deventer are pristine and have not been recycled. It would seem therefore, that the lowest proportion of recycled glass was imported and used in the late 6th-early 7th century and between c. 850 and 10th-11th centuries, partly in the latter case because decentralized primary production of wood ash glass had developed. Between these dates there was clearly a dependence on using recycled glass for making the majority of vessels found at Dorestad. The research also provides building blocks for two NOaA questions: What are the nature, manifestations, extent and context of craft specialization? (NOaA 2.0 question 67) The description of glass craft specialization is discussed in detail in Chapter 3, some of which is ephemeral. Therefore only the most significant evidence is discussed here. Monochrome glass bead production was a craft specialization in the early medieval Netherlands. The most comprehensive evidence in Europe to date for the manufacture, especially of opaque yellow glass beads, has been excavated from the Jodenstraat 30 site in Maastricht where a rubbish pit filled with the debris from glass bead making was found.412 The pit also contained waste from copper-alloy-working and amberworking.413 Based on the pottery finds, the pit was filled sometime in the late 6th to early 7th centuries. The debris from bead making consists of 750 glass objects which represent the full range of waste from glass bead production. The production waste was divided into eight main groups: glass rods (n=369), ‘punty’ glass from a beadmaker’s tool (n=36), glass threads with and without tweezer marks (n=17), glass drops (n=39), finished and failed beads (n=123), crucibles (n=38, EMN=17), cullet or scrap glass (n=20), glassy slags/fuel ash slags (n=53) and non-diagnostic fragments (n=55) which include (small lumps of) melted glass and fragments that are too small to classify. All waste categories are dominated by opaque yellow glass (apart from scrap glass and glassy slags). Almost all beads are wound and have tapering perforations showing they were made by winding melted glass around a mandrel, a bead-making tool with a conical point. Such a tool may have been found at the Rijksarchief site. The crucibles associated with bead production consist of 38 fragments from at least 17 coarse-ware cooking pots (Wölbwandtöpfe). In 15 cases, only the lower halves of the pots were used to melt highly coloured opaque yellow glass. Two crucible bases with opaque white pigment are also present. Drops of translucent greenish glass among the waste products suggest this glass colour was worked on or near the site. Excavations at the Maastricht Mabro site produced twelve fragments of crucibles with glass deposits; eleven of these date to the 6th7th centuries. The colours of the glass in the crucibles are colourless or pale green or opaque yellow. A number of wound beads have been found: they are either monochrome or decorated with trails in contrasting colours. The site has not been published, so it is impossible to state which beads are likely to be local products, but given the dates for the crucibles, those beads dating to roughly the 2nd half of the 6th and first half of the 7th century are the most likely candidates: a small globular bead, a bi-globular bead and a possible cylindrical bead of opaque yellow glass, a medium-sized globular bead of opaque white glass and a short cylindrical bead of opaque red glass. There are also two polychrome, trailed beads: an opaque white globular bead with translucent light blue narrow crossing trails and an opaque white disc-shaped bead with translucent dark blue crossing trails. Excavations of the Maastricht-Rijksarchief site are discussed by Hulst.414 Dozens of rubbish pits were full of the evidence for glass working, antler working and iron working and dated by (late) 5th and 6th century pottery. The evidence of glass working consists of 55 fragments of early Merovingian vessel glass, two fragments of glass rods, 25 beads, drops of glass, melted glass, glassy slags and fragment of a glass crucible.415 Furthermore, one fragment of a glass crucible, glassy slags (included some attached to a furnace floor or thick tray fragment) were found. A forged iron rod which is round in section at one end and square in section at the other match the perforations of beads found on the site; it is an example of a very unusual bead- 412 Sablerolles, Henderson & Dijkman 1997. 413 For waste from amber-working, see Dijkman 2013. 414 Hulst 1992. 415 Hulst 1992. 112 — 416 Sablerolles 2023. (basispublicatie chapter 29). 417 Pottery identification by Jan de Koning. 418 Nyst 2003, 13. 419 Preiß 2010, 125. making mandrel around which glass filaments would have been wound. Evidence for the manufacture of opaque yellow glass was also found at Wynaldum in the form of a thick fragment covered with a layer of opaque yellow material of between 1.0-1.3 cm thick. The yellow substance would probably have been used to make beads once formed into glass. Two small flattened opaque yellow and white beads were found and are probably local products. This evidence was found amongst waste produced by a blacksmith/bronze-caster. The dump dates to the last quarter of the 6th and the first quarter of the 7th century and is contemporary with the glass-working evidence from the Jodenstraat site in Maastricht. A tessera and a piece of punty glass glass were found in Carolingian contexts. There is a possibility that they relate to bead production. Brightly coloured Merovingian monochrome beads would presumably have been used by local populations but would also have been exported. As mentioned in Chapter 6 there was evidently a demand especially for bright yellow glass beads from at least the middle Merovingian period; such glass must have had social significance but it is difficult to suggest what it might have been. The five fragments of glass production waste have been found at the monastic site of Susteren-Abdijterrein suggest that glass was worked there during the early medieval period.416 The finds are two fragments of glass crucibles (one with a layer of cobalt blue glass and the other with a layer of bluish green and colourless glass attached), a partially melted Roman tessera, a glass fragment from glassblowing tool and a possible fragment of opaque yellow raw glass. The crucible fragments derive from a context dated to 800–1300 which contains 60% redeposited Carolingian material. One crucible fragment, which contains bluish-green glass, is probably made of Carolingian Badorf ware.417 The second crucible fragment which contains cobalt blue blue glass was made of possible grey Meuse Valley ware: it may have been used to make dark blue window quarries on the site, the first such evidence from the Netherlands. A partially melted, opaque dark blue Roman tessera may have formed part of the manufacture. A translucent dark bluish-green glass fragment with thick walls is covered on its concave inside with dark grey iron scale from a glassblowing tool. This fragment is the only direct evidence for glassblowing in the Netherlands since the Roman period. Excavations at Wijk bij DuurstedeVeilingterrein and Frankenweg/Zandweg revealed evidence for Carolingian bone- working (and those at the adjoining Parkeerplaats Albert Heijn (PPAH) site produced traces of metal working and loom weights).418 Some thirty-six fragments of glass-working waste were found, including eight tesserae, a large fragment of a glass crucible of a late Merovingian form found in a pit with two blue tesserae, a regular and an irregular drop of translucent pale green glass, a small dark sphere, six melted lumps of translucent pale green vessel glass and a melted fragment of ‘black’, deep olive-green glass. The crucible contains almost colourless glass of c. 1 mm thickness with a thicker patch of opaque white glass which is probably a melted tessera. Preiß points out that defects in the translucent glass probably indicate locations where other (crushed?) tesserae had been attached.419 The crucible may be linked to bead-making, but it also may be linked to vessel or window glass production at the site. Excavations at Wijk bij Duurstede – Veilingterrein and Frankenweg/Zandweg revealed evidence for glass production along with iron smithing, brass production, weaving (wool) and amber working; glass and amber working evidence were sometimes found in close proximity. The largest category of glass working evidence is tesserae, almost all a blue/ green colour. Fragments of translucent bluishgreen and dark blue glass probably result from breaking up glass ingots, cakes or raw glass chunks. There are dark blue drops, two square sectioned opaque yellow glass rods and opaque white glass which were probably used for beadmaking. A lump of opaque yellow glass has part of a composite opaque yellow glass rod and yellow punty glass from a glassworker’s tool melted onto it. An opaque green object is waste from glass bead production; there is also a twisted bi-chrome (opaque white and bluegreen) rod fragment. Such rods were used or decorating Carolingian glass vessels. A thick opaque yellow glass fragment has embedded iron oxide scale in it as well as ceramic fragments. It may have been used in bead production or decorating 8th century vessels. A ‘bright blue’ tessera and clay covered with a thick layer of ‘blobby greenish glass’ were found on the 113 — Veilingterrein site.420 Moreover two small spheres of whitish translucent glass were found.421 The finds from Dorestad therefore provide evidence for local bead production; vessels may also have been made there. The relatively small amount of glass production debris derives from pits, wells and ditches; wet sieving was not carried out universally. It is therefore difficult to decide whether the evidence indicates that production was on a small-scale, at household level, or on a larger scale for export. Excavations at Utrecht Domplein produced twelve crucible fragments with a layer of glass attached to their insides. The glass is either apparently colourless, green with red streaks of glass running through it or pale green. Red marbled translucent blue-green/bluish-green glass was popular for making late Merovingian and Carolingian glass vessels. Excavations at Utrecht Oudwijkerdwarsstraat produced an irregular drop of bluish-green glass as well as some crushed fragments. The discovery of fourteen hundred amber fragments shows that some craft activity occurred on quite a large scale. The evidence from Merovingian Leidsche Rijn-Leeuwesteyn Noord consists of a single crucible fragment with a layer of blue-green glass with marbled opaque red streaks perhaps used to make windows. The Merovingian Rijnsburg-Abdijterrein site produced some useful evidence for glass bead production consisting of finished, unfinished and failed beads, glass rods and three pieces of punty glass from a beadmaker’s tool; eight crucible fragments and two lumps of fired clay covered with translucent greenish glass could have been part of a glass-working furnace floor. A single crucible fragment with green glass attached may be contemporary. Nearly half of the waste from Rijnsburg is opaque yellow. The crucible fragments have yellow and colourless glass attached to them. This material was not available for this research project. It is possible that colourless glass was modified on site using lead-tin yellow. The chemical composition of the yellow glass in the crucible, the beads and the rods are similar and therefore likely to have been made on site. The bead types made at Rijnsburg were monochrome opaque yellow globular, bi- and tri-globular beads and bi-globular red beads. Trailed beads include bi-globular beads of red glass with both white crossing trails and a white spiral, tri-globular beads of opaque red glass with opaque white crossing trails and white beads with translucent blue crossing trails were also possibly made at the site. It is likely that the production phase occurred in the 7th century partly based on dated bead typologies (see Section 3.8). Two hundred and one (unstratified) glass tesserae from the terp at Wierum have been interpreted as a supply of ‘raw’ glass for making beads in the early medieval period.422 The terp probably dates to the 8th/9th centuries. Most tesserae have rounded profiles so appear to have been heated. Five fragments of highly coloured early Roman vessel glass, one fragment of Roman or early medieval vessel glass, three opaque green nearly colourless and translucent dark blue plano-convex drops of glass and thirteen irregular drops/melted lumps of nearly colourless, pale green and pale bluegreen glass have also been found. This evidence from Wierum may have resulted from a travelling beadmaker visiting terp sites such as Wijnaldum in the northern coastal region, which was most easily accessible by boat from the central riverine area, with Dorestad at its centre. The latest site to be considered in our research project is Deventer- Stadhuiskwartier. The evidence for the glass industry is scanty, dates to between c. 850 and c. 1050 and derives from waste pits or cesspits. Two pits yielded production waste of iron smithing, boneworking and textile production; glassy slag dates to 850–900. Glass working evidence dating to between 900 and 925 consists of a heavily weathered chip of glass with a conchoidal fracture, a heavily weathered triangular fragment and a small, heat-affected fragment which may be part of a pulled thread. Waste from the 900– 950 phase consist of a small lump of translucent clear bluish-green raw glass, a heavily weathered fragment with a triangular section and a heavily weathered fragment with two irregular, heataffected surfaces. An unusual fragment dating to 950-1050 consists of a translucent bright bluishgreen and deep turquoise glass layer covered by a very thin film of opaque red glass. The turquoise colour is very similar to a contemporary fragment of very thin flat glass, either window glass or a glass inlay, found in the same area. It is evident that early medieval glass working in the Netherlands was frequently 420 Van Doesburg 2004. 421 Langbroek 2021b, 64, table 7, findnrs Zandweg WD 754.2.63b, WD 754.2.63b. 422 Crocco et al. 2021. 114 — 423 Verhulst 2002, 72-84. associated in industrial areas on the sites with a range of other industries including iron smithing, copper-alloy production and brass production, as well as amber working and weaving. It is likely that the same fuels (yet to be determined) would have been used for glass and metal production, depending on the maximum temperature required. The artisans involved in glass working may have taken part in other activities associated with shared aspects of other high temperature industries, such as obtaining fuel, making crucibles and building kilns/furnaces or separate groups were involved in such activities. During the Carolingian period such industrial organisations, that were involved in several different production activities on particular sites, have been found, for example, at San Vincenzo al Voltorno, Augsburg and Corvey.423 Where do non-local raw materials of utilitarian objects come from? (NOaA 2.0 question 139) There is evidence that much of the base glass used to make utilitarian beads and vessels originated in Egypt; by the Carolingian period most of it was recycled to the extent that its source is indeterminate but most still has evidence that the ultimate source of almost all ‘pristine’ unrecycled glass was Egypt. The small number of plant ash glasses used to make utilitarian objects in the early medieval Netherlands probably derived from Iraq/Iran, Syria and the Levant. The lead and perhaps the cobalt may have originated in Germany, but this is to be confirmed - and there are other possible sources. The possible sources of tin are more restricted: Cornwall in the UK is one such source; Turkey is another one. 115 — Bibliography Adlington, L.W., I.C. Freestone, J.J. Kunicki-Goldfinger, T. Ayers, H. Gilderdale Scott & A. Eavis 2019: Regional patterns in medieval European glass composition as a provenancing tool, Journal of Archaeological Science, 110, 104991. Aldsworth, F., G. Haggerty, S. Jennings & D. Whitehouse 2002: Medieval Glassmaking at Tyre, Lebanon, Journal of Glass Studies 44: 49–66. Alfonsi, H. & Gandolfo P. 1997: L’épave Sanguinaire A, Cahiers d’Archéologie subaquatique 13, 35-74. Allan, J.W., 1973: Abū’l-Qāsim’s treatise on ceramics, Iran, 11, 111–120. Andersen, J.H. & T. Sode 2010: The Glass Bead Material, in: M. Bencard & H.B. Madsen (eds), Ribe Excavations 1970–76, vol. 6., Ribe Esbjerg: Sydjysk Universitetsforlag (Ribe Excavations, Jutland Archaeological Society Publications), 17–59. Artioli, G., C. Canovaro, P. Nimis & I. Angelini 2020: LIA of Prehistoric Metals in the Central Mediterranean Area: A Review, Archaeometry, 62, 53-85. Aunay, C., A.A. Berthon, B. Gratuze, M. Guérit, J. Motteau & I. Pactat 2020: Le verre creux du VIIIe au Xe siècle dans la vallée de la Loire moyenne et de la Vienne. Essai typo-chronologique et archéométrique, in: Inès Pactat & C. Munier (eds), Le verre du VIIIe au XVIe siècle en Europe occidentale, Les Cahiers de la MSHE Ledoux, Presses universitaires de FrancheComté, 293–314. Bandiera, M., M. Verità, P. Lehuédé & M. Vilarigues 2020: The technology of copper-based red glass sectilia from the 2nd century AD Lucius Verus villa in Rome, Minerals, 10(10), 875. Arbman, H., 1937: Schweden und das Karolingische Reich. Studien zu den Handelsverbindungen des 9. Jahrhunderts. Kungl, Stockholm (Vitterhets Historie och Antikvitets Akademiens Handlingar Del 43). Barber, D.J, I. Freestone & K.M. Moulding 2009: Ancient Copper Red Glasses: investigation and analysis by microbeam techniques, Chapter 11 in: A.J. Shortland, I.C. Freestone & T. Rehren (eds), Mine to Microscope, Oxford: Oxbow Books & the David Brown Book Company. Ares, J.D.J., A.V.E. Guirado, Y.C. Gutiérrez & N. Schibille 2019: Changes in the supply of eastern Mediterranean glasses to Visigothic Spain, Journal of Archaeological Science, 107, 23–31. Bardet, A.C., 1995: Pottery traded to Dorestad: some explanatory archaeometrical analyses of early medieval Rhenish wares, Berichten van de Rijksdienst voor Oudheidkundig Bodemonderzoek, 41, 187–251. Bass, G.F., 1984: The nature of the Serçe Limani glass, Journal of Glass Studies 26, 64–9. Bass, G.F., B. Lledo, S. Matthews & R.H. Brill 2009: Serçe Limani. Volume 2: The glass of an eleventhcentury shipwreck, Rachel Foundation Nautical Archaeology Series, College Station, Texas A & M University Press. Boschetti, C., C. Leonelli, R. Rosa, M. Romagnoli, M. Valero Tévar & N. Schibille 2020: Antimony or Opacified White Glass Tesserae, Heritage, 3, 549–560. https://doi. org/10.3390/heritage3020032 Baumgartner, E. & I. Krüger 1988: Phoenix aus Sand und Asche. Glas des Mittelalters, München. Boschetti, C., B. Gratuze & N. Schibille 2020: Commercial and social significance of glass beads in Migration-Period Italy: the cemetery of Campo Marchione, Oxford Journal of Archaeology, 39, issue 3, 319–342. Bertini, C., J. Henderson & S. Chenery 2020: Seventh to eleventh century CE glass from Northern Italy: between continuity and innovation, Archaeological and Anthropological Sciences, 12:120. https://doi. org/10.1007/s12520-02001048-8 Besteman, J.C., J.M. Bos, D.A. Gerrets, H.A. Heidinga & J. de Koning (eds) 1999: The Excavations at Wijnaldum. Reports on Frisia in Roman and Medieval times, Volume I, Rotterdam/ Brookfield. Birck, J.L., 1986: Precision of K-Rb-Sr Isotopic Analysis - Application to Rb-Sr Chronology, Chemical Geology, 56, 73-83. Boschetti, C., J. Henderson, J. Evans & C. Leonelli 2016: Mosaic tesserae from Italy and the production of Mediterranean coloured glass (4th century BC–4th century AD). Part I: Chemical composition and technology, Journal of Archaeological Science, Reports, 7, 303–311. Brems, D., M. Ganio, K. Latruwe, L. Balcaen, M. Carremans, D. Gimeno, A. Silvestri, F. Vanhaecke, P. Muchez & P. Degryse 2013a: Isotopes on the beach, part 1: strontium isotope ratios as a provenance indicator for lime raw materials used in Roman glass-making, Archaeometry 55, 214-234. Brems, D., M. Ganio, K. Latruwe, L. Balcaen, M. Carremans, D. Gimeno, A. Silvestri, F. Vanhaecke, P. Muchez, & P. Degryse 2013b: Isotopes on the beach, part 2: neodymium isotopic analysis for the provenancing of Roman glass-making, Archaeometry, 449-464. Brems, D., I.C. Freestone, Y. Gorin-Rosen, R. Scott, V. Devulder, F. Vanhaecke & P. Degryse 2018: Characterisation of Byzantine and early Islamic primary tank furnace glass, Journal of Archaeological Science: Reports, 20, 722–735. 116 — Brill, R.H., 2005: Chemical Analyses of Some Sasanian Glasses from Iraq, in: David Whitehouse, Sasanian and Post Sasanian Glass in The Corning Museum of Glass, Corning: the museum, 2005, 65–88. Brill, R.H., 2006: Scientific Investigations of Some Glasses from Jarrow and Monkwearmouth, in: R. Cramp, Wearmouth and Jarrow monastic sites, volume 2, English Heritage. Brüggler, M., 1994: Burgus und Glaswerkstatt der Spätantike bei Goch-Asperden, Berichte. Grabung–Forschung–Präsentation, Köln. Bult, E.J., J. van Doesburg & D.P. Hallewas 1990: De opgravingscampagne in de Vroeg-Middeleeuwse nederzetting op de Woerd bij Valkenburg (Z.H.) in 1987 en 1988, in: E.J. Bult & D.P. Hallewas (eds), Graven bij Valkenburg III, het archeologisch onderzoek in 1987 en 1988, Delft, 147–166. Cabart, H., I. Pactat & B. Gratuze 2017: Les verres du Haut Moyen âge issus des fouilles du monasterium Habendum (Vosges, France), Annales du 20e Congrès de l’Association Internationale pour l’Histoire du verre, 346–353. Cagno, S., M.B. Badano, F. Mathis, D. Strivay & K. Janssens 2012: Study of medieval glass fragments from Savona (Italy) and their relation with the glass produced in Altare, Journal of Archaeological Science, 39 (7), 2191–2197. Callmer J., 1977: Trade beads and bead trade in Scandinavia ca. 800– 1000 AD, Acta Archaeologica Lundensia, 4 , 11, Habelt, Bonn. Callmer, J., 2003: Wayland. An essay on craft production in the Early and High Middle Ages in Scandinavia, in: L. Larsson & B. Hårdh (eds), Centrality-Regionality: the social structure of southern Sweden during the Iron Age, Stockholm. Acta Arch. Ludensia Ser. 8°, 40, 337-361. Callmer, J. & J. Henderson 1991: Glassworking at Åhus, S. Sweden (eighth century AD), Laborativ Arkeologi, 5, 143–154. Casellato, U., F. Fenzi, P. Guerriero, S. Sitran, P.A. Vigato, U. Russo, M. Galgani, M. Mendera & A Manasse 2003: Medieval and renaissance glass technology in Valdelsa (Florence). Part 1: raw materials, sands and non-vitreous finds, Journal of Cultural Heritage, 4, (4), 337-353. Ceglia, Andrea, P. Cosyns, K. Nys, H. Terryn, H. Thienpont & W. Meulebroeck 2015: Late antique glass distribution and consumption in Cyprus: a chemical study, Journal of Archaeological Science, 61, 213–222. Ceglia, Andrea, P. Cosyns, N. Schibille & W. Meulebroeck 2019: Unravelling provenance and recycling of Late Antique glass with trace elements, Archaeological and Anthropological Sciences, 11, 279–291. Conte, S., T. Chinni, R. Arletti & M. Vandini 2014: Butrint (Albania) between eastern and western Mediterranean glass production: EMPA and LA-ICP-MS of late antique and early medieval finds, Journal of Archaeological Science, 49, 6–20. Corbella, F., 2017: The Lent Merovingian glass beads µ-Computed Tomography, XRF and SEM analysis of complex glass beads, Unpublished Internship report at TU Delft for École Polytechnique, Université Paris-Saclay, s.l., s.n., 64 pp. Crocco, R., H. Huisman, Y. Sablerolles, J. Henderson, B. van Os & A. Nieuwhof 2021: Hunting colours: origin and reuse of glass tesserae from the Wierum terp, Archaeological and Anthropological Sciences, 13(9), 1–22. https://doi. org/10.1007/s12520–021– 01391–4 De Bruin, J., 2018: Living in Oegstgeest 575–725 AD, in: M. Kars, R. van Oosten, M.A. Roxburgh & A. Verhoeven, Rural riches & royal rags? Studies on medieval and modern archaeology, presented to Frans Theuws, Zwolle: SPA-Uitgevers, 20–25. De Bruin, J., C. Bakels & F. Theuws (eds) 2021: Oegstgeest. A riverine settlement in the early medieval world system, Bonn: Habelt Verlag. De Koning, J., 2012: Het aardewerk, in: J. Dijkstra (ed.), Het domein van de boer en de ambachtsman. Een opgraving op het terrein van de voormalige fruitveiling te Wijk bij Duurstede: een deel van Dorestad en de villa Wijk archeologisch onderzocht, ADC monografie 12, Amersfoort, 117–235. De Koning, J., 2015: Onder het stuifzand. Overstoven vroegmiddeleeuwse nederzettingen bij Bloemendaal de opgravingscampagnes Groot Olmen 2005: 2006 en 2007, Zaandijk. De Sigoyer, S.D.B., C. Peters, S. Mathieu & C. Fontaine 2005: Vestiges de fours de verriers d’epoque Mérovingien à Huy aux Ruelles (Belgique): Aperçu des trouvailles, Bulletin de l’Association Française pour l’Archéologie du Verre, 2005, 29–33. Degryse, P. & J. Schneider 2008: Pliny the Elder and Sr–Nd isotopes: tracing the provenance of raw materials for Roman glass production, Journal of Archaeological Science, 35(7), 1993–2000. Degryse, P., I. Freestone, S. Jennings & J. Schneider 2010: Technology and provenance study of Levantine plant ash glass using Sr–Nd isotope analysis, Römisch-Germanisches Zentralmuseum. Dekówna, M., 1978: Les verres de Haithabu. Rapport préliminaire, Annales du 7e Congrès de l’Association Internationale pour l’Histoire du Verre, Liège, 167-188. 117 — Dekówna, M., 1980: Methods of examining ancient glasses, in: J. Schild (ed.), Unconventional Archaeology, Wroclaw: Ossolineium, Wydawnictwo Polskiej Akademii Nauk, 213–233. Dell’Acqua, F., 1997: Ninthcentury window glass from the monastery of San Vincenzo al Volturno (Molise, Italy), Journal of glass studies, 33–41. DePaolo, D.J. & G.J. Wasserburg 1976: Nd isotopic variations and petrogenetic models, Geophysical Research Letters, 3(5), 249–252. Dijkman, W., 2013: Artisanal Activities in Merovingian Maastricht, Medieval and Modern Matters, 4, 23–39. Dijkstra, J. (ed.) 2012: Het domein van de boer en de ambachtsman. Een opgraving op het terrein van de voormalige fruitveiling te Wijk bij Duurstede: een deel van Dorestad en de villa Wijk archeologisch onderzocht, Amersfoort (ADC-monografie 12). Dijkstra, M.P.F. 2011: Rondom de mondingen van de Rijn en Maas. Landschap en bewoning tussen de 3e en de 9e eeuw in Zuid-Holland, in het bijzonder de Oude Rijnstreek, (PhD thesis University of Amsterdam). Dijkstra, M., Y. Sablerolles & J. Henderson 2011: A traveller’s tale. Merovingian glass bead production at Rijnsburg, the Netherlands, in: C. Theune, F. Biermann, R. Struwe & G.H. Jeute, Zwischen Fjorde und Steppe, Festschrift für Johann Callmer zum 65. Geburtstag, Rahden/Westf.: Leidorf, 175–199 (Internationale Archäologie; Studia honoraria; Band 31). Dodt, M., A. Kronz & K. Simon 2021: Production of early medieval glass in Cologne, in: A. Willemsen and H. Kik (eds), Dorestad and its networks, Proceedings of the third ‘Dorestad Congress’ held at the National Museum of Antiquities, Leiden, the Netherlands 12–15 June 2019, 179–191. Evison, V.I., 1972: Glass Cone Beakers of the “Kempston” Type, Journal of Glass Studies, 48–66. Fontaine, S. D. & D. Foy 2007: L’ épave Ouest-Embiez 1, Var; le commerce maritime du verre brut et manufacturé en Méditerranée occidentale dans l’Antiquité, Revue Archéologique de Narbonnaise, 40, 235-268. Foster, H.E. & C.M. Jackson 2009: The composition of ‘naturally coloured’ late Roman vessel glass from Britain and the implications for models of glass production and supply, Journal of Archaeological Science, 36(2), 189–204. Foy, D. & M.-D. Nenna 2001: Tout feu tout sable: mille ans de verre antique dans le Midi de la France, Edisud, Marseilles. Foy, D., M. Picon, M. Vichy & V. Thirion-Merle 2003: Caractérisation des verres de la fin de l’Antiquité en Méditerranée Occidentale: l’émergence de nouveaux courants commerciaux, in: D. Foy & M.-D. Nenna (eds), Échanges et commerce du verre dans le monde antique, Actes du colloque de l’Association française pour l’archéologie du verre, Aix-en-Provence, juin 2001, 41–85, Éditions Monique Mergoil, Montagnac. Freestone, I.C., 2015: The recycling and reuse of Roman glass: analytical approaches, Journal of glass studies, 29–40. Freestone, I.C., Y. Gorin-Rosen & M.J. Hughes 2000: Primary glass from Israel and the production of glass in late antiquity and the early islamic period, in: N. Marie-Dominique (ed.), La route du verre. Ateliers primaires et secondaires du second millénaire av. J.-C. au Moyen Âge, Maison de l’Orient et de la Méditerranée, Jean Pouilloux, Lyon, 65–83. Freestone, I.C. & M.J. Hughes 2006: The origins of Jarrow glass, in: R. Cramp, Wearmouth and Jarrow Monastic sites, Vol. 2, English Heritage, 147–155. Freestone, I.C., R.E. Jackson-Tal, O. Tal & I. Taxel 2015: Glass production at an early Islamic workshop in Tel Aviv, Journal of Archaeological Science 62, 45-54. Freestone, I.C., P. Degryse, J. Lankton, B. Gratuze & J. Schneider 2018: HIMT, glass composition and commodity branding in the primary glass industry, in: D. Rosenow, M. Phelps, A. Meek & I. Freestone (eds), Things that Travelled: Mediterranean Glass in the First Millennium AD, UCL Press, 159–190. https://doi. org/10.2307/j.ctt21c4tb3.14 Gai, S., 2005: Vitres et vitraux du palais impérial de Charlemagne à Paderborn, in: D. Foy (ed.), De transparentes spéculations. Vitres de l’Antiquité et du haut Moyen Âge (OccidentOrient), Exposition temporaire en liaison avec les 20e rencontres de l’AFAV sur le thème du verre plat, Bavay, Musée-site d’archéologie, 83–85. Gallo, F., A. Silvestri, P. Degryse, M. Ganio, A. Longinelli & G. Molin 2015: Roman and late-Roman glass from north-eastern Italy: The isotopic perspective to provenance its raw materials, Journal of Archaeological Science, 62, 55–65. Gam, T., 1990: Prehistoric Glass Technology – Experiments and Analyses, Journal of Danish Archaeology, 9, 203–213. Ganio, M., M. Gulmini, K. Latruwe, F. Vanhaecke & P. Degryse 2013: Sasanian glass from Veh Ardašīr investigated by strontium and neodymium isotopic analysis, Journal of archaeological science, 40(12), 4264–4270. 118 — Gaut, B., 2011: Vessel Glass and Evidence of Glassworking in Skre, D. (ed.). Things from the town: artefacts and inhabitants in Viking-age Kaupang, Kaupang Excavation Project publication series volume 3, Aahus University Press, 169-258. Gaut, B. & J. Henderson 2011: Compositional analysis. Appendix 9.1 in Skre, D. (ed.). Things from the town: artefacts and inhabitants in Viking-age Kaupang, Kaupang Excavation Project publication series volume 3, Aahus University Press, 262-271. Genga, N., M. Siciliano, A. Tepore, A. Mangone, A. Traini & C. Laganara 2008: An archaeometric approach about the study of medieval glass from Siponto (Foggia, Italy), Microchemical Journal, 90, (1), 56-62. Gliozzo, E., E. Braschi, F. Giannetti, A. Langone & M. Turchiano 2019: New geochemical and isotopic insights into the Late Antique Apulian glass and the HIMT1 and HIMT2 glass productions— the glass vessels from San Giusto (Foggia, Italy) and the diagrams for provenance studies, Archaeological and Anthropological Sciences, 11, 141–170. Gorin-Rosen Y., 2000: The ancient glass industry in Israel—summary of the finds and new discoveries, in: M.-D. Nenna (ed.), La route du verre Ateliers primaires et secondaires du second millénaire av. J.C. au Moyen Âge, Maison de l’Orient Méditerranéen, Jean Pouilloux, Lyon, 49–63. Gratuze, B. & J.-N. Barrandon 1990: Islamic glass weights and stamps, Archaeometry, 32, 155–162. Gratuze, B., D. Foy, J. Lancelot & F. Téreygeol 2003: Les “lissoirs” carolingiens en verre au plomb: mise en evidence de la valorisation des scories issues du traitement des galènes argentifères de Melle (Deux-Sèvres), in: D. Foy & M.-D. Nenna (eds), Échanges et commerce du verre dans le monde antique, Actes du colloque de l’AFAV, Aix-en-Provence et Marseille, 7–9 June 2001, Montagnac, France: Éditions Monique Mergoil, 101–108. Haevernick, T.E., 1979: Karolingisches Glas aus St Dionysius in Esslingen, Forschungen und Berichte der Archäologie des Mittelalters in Baden-Württemberg, 6, 157–171. Heaser, S., 2018: Anglo-Saxon Glass Beadmakers, s.l. 19 pp. Heck, M., T. Rehren & P. Hoffmann 2003: The production of lead–tin yellow at Merovingian Schleitheim (Switzerland), Archaeometry, 45(1), 33–44. Hemminga, M. & T. Hamburg 2006: Een Merovingische nederzetting op de oever van de Oude Rijn. Opgraving (DO) en Inventariserend Veldonderzoek (IVO) Oegstgeest - Rijnfront zuid 2004, Leiden (Archol Rapport 69). Hemminga, M., T. Hamburg, M. Dijkstra, C. Cavallo, S. Knippenberg, S.M.E. van Lith, C.C. Bakels & C. Vermeeren 2008: Vroegmiddeleeuwse nederzettingssporen te Oegstgeest. Een Inventariserend Veldonderzoek en Opgraving langs de Oude Rijn, Leiden (Archol Rapport 102). Henderson, J., 1985: The raw materials of early glass production, Oxford Journal of Archaeology, 4(3), 267–291. Henderson, J., 1985: The raw materials of early glass production, Oxford Journal of Archaeology, 4, 267-291. Henderson, J., 1989: The scientific analysis of ancient glass and its archaeological interpretation, in: J. Henderson (ed.), Scientific analysis in archaeology and its interpretation, Oxford University Committee on Archaeology Monograph no. 19, UCLA Institute of Archaeology Research Tools 5, Oxford: Oxbow Books, 30–62. Henderson, J., 1991a: Chemical characterisation of Roman glass vessels, enamels and tesserae, in: P.B. Vandiver, J. Druzik & G.S. Wheeler (eds), Materials issues in art and archaeology II, Research Society Symposium Proceedings, vol. 185, Pittsburgh, PA: Materials Research Society, 601–607. Henderson, J., 1991b: The glass, in: P. Armstrong, D. Tomlinson & D.H. Evans (eds), Excavations at Lurk Lane Beverley 1979–82, Sheffield Excavation Reports 1, Sheffield, England: J.R. Collis, 124–30. Henderson, J., 1993: Aspects of Early Medieval glass production in Britain, Annales du 12e Congrès de l’Association Internationale pour l’Histoire du verre, Vienna 26–31 August 1991, Amsterdam: AIHV, 247–259. Henderson, J., 1999: Scientific analysis of the glass and glassbearing artefacts: Technique, raw materials used and archaeological interpretation, in: J.C. Besteman, J.M. Bos, D.A. Gerrets, H.A. Heidinga & J. de Koning, The Excavations at Wijnaldum, reports on Frisia in Roman and Medieval times, volume 1, Rotterdam: Balkema, 287–297. Henderson, J., 2000: The science and archaeology of materials, Routledge: London and New York. Henderson, J., 2002: Tradition and experiment in 1st millennium ad glass production – the emergence of early Islamic glass technology in late antiquity, Accounts of Chemical Research, 35, 594–602. Henderson, J., 2012: De Glasvondsten, in: Het domain van de boer en de ambachtsman. Een opgraving op het terrein van de voormalige fruitveiling, te Wijk bij Duurstede: een deel van Dorestad en de villa Wijk archeologisch onderzocht, Amersfoort (ADC Monografie 12), 293–354. Henderson, J., 2013: Ancient glass, an interdisciplinary exploration, New York and Cambridge: Cambridge University Press. 119 — Henderson, J., 2022: The politics of production, glass provenance and social context on the early Islamic Silk Roads, Journal of Islamic Archaeology 8.2, 203-237. Henderson, J., 2023: Susteren glass: scientific evidence for raw material use, supply and provenance, Stoepker, H. (ed.), 2023: Sporen van Susteren, archeologische vondsten uit een Karolingisch klooster en een adellijk vrouwenstift, de basispublicatie, Venlo, 1357-1382. https://doi.org/10. 17026/dans-xsf-nypw Henderson, J., K. Challis, S. O’Hara, S. McLoughlin, S. Gardner & G. Priestnall 2005a: Experiment and innovation: early Islamic industry at al-Raqqa, Syria, Antiquity, 79, 1–15. Henderson, J., J.A. Evans, H.J. Sloane, M.J. Leng, & C. Doherty 2005b: The use of oxygen, strontium and lead isotopes to provenance ancient glasses in the Middle East, Journal of Archaeological Science, 32, 665-673. Henderson, J., S. Chenery, E. Faber & J. Kröger 2016: The use of Electron Probe Microanalysis and Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry for the investigation of 8th–14th century plant ash glasses from the Middle East, Microchemical Journal, 128, 134–152. Henderson J., S. Chenery, E.W. Faber & J. Kröger 2021: Political and technological changes, glass provenance and a new glass production model along the west Asian Silk Road, in: Florian Klimscha (ed.), Berlin Studies of the Ancient World, 67, Berlin: edition Topoi. Henderson, J. & I. Holand 1992: The Glass from Borg, an early Medieval Chieftain’s Farm in Northern Norway, Medieval Archaeology, XXXVI, 39–58. Henderson, J. & R. Ivens 1992: Dunmisk and glass-making in Early Christian Ireland, Antiquity, 66, 52–64. Henderson, J., J. Evans & Y. Barkoudah 2009: The roots of provenance: glass, plants and isotopes in the Islamic Middle East, Antiquity, 83, 414–429. Henderson, J., H. Ma & J. Evans 2020: Glass production for the Silk Road? Provenance and trade of Islamic glasses using isotopic and chemical analyses in a geological context, Journal of Archaeological Science, 119, 105–164. doi.org/10.1016/j. jas.2020:105164 Henderson, J. & Y. Sablerolles 2020: An overview of Dutch Early Medieval glassworking, published chemical and isotopic analyses of glass beads and vessels, raw, material provenance of beads and vessels, changes in raw material use over time and a plan for future scientific research. Henderson, J., T. Sode & Y. Sablerolles 2019: Early medieval tesserae from Scandinavia and the Netherlands: a case for re-use and recycling, in: L. Van Wersch, L. Verslype, S. Strivay & F. Theuws (eds), Early medieval tesserae in Northwestern Europe, Habelt, Bonn, 68–95. Henton, A., 2020: Aux portes de l’abbaye d’Elnone – SaintAmand. Premiers résultats de la fouille préventive de la Grand-Place de Saint-Amandles-Eaux (Nord, Hauts-deFrance), 43e Colloque Archaeologia Mediaevalis, Mars 2020: Namur, Belgique, 54–58. hal-03229653 Huisman, D.J., J. Van der Laan, G.R.Davies, J.H.van Os, N. Roymans, B.Fermin & M. Karwowski 2017: Purple haze: Combined geochemical and Pb-Sr isotope constraints on colourants in Celtic glass, Journal of Archaeological Science 81, 59-78 Huisman, H., M. Aarts, M. Kars, F. Mulder, D. NganTillard & B. van Os, 2019, Maken en handelen: Merovingische kralen uit het Sittard-Kemperkoul grafveld geanalyseerd, Paleoaktueel 30: 65-74 Hulst, R.A., 1992: Opgravingen Rijksarchief te Maastricht: Merovingische afvalkuilen en laat-middeleeuwse verdedigingswerken, Archeologie in Limburg, 42, 89–94. Isings, C., 1957: Roman glass from dated finds, Groningen: Archaeologica Traiectina, 2. Isings, C., 1978: Glas, Spiegel Historiael, 13, 4 (Dorestadnummer), 260–262. Isings, C., 2009: Glas, in: M. Nokkert, A.C. Aarts & H.L. Wynia (eds), Vroegmiddeleeuwse bewoning langs de A2. Een nederzetting uit de zevende en achtste eeuw in Leidsche Rijn, Utrecht (Basisrapportage Archeologie 26 Gemeente Utrecht), 246–251. Isings, C., 2015: Glass, in: W.A. van Es & W.J.H. Verwers, Excavations at Dorestad 4, The settlement on the river Bank Area, Amersfoort: Nederlandse Oudheden 18, 422–446. Isings, C., G. Rauws, H. Lägers & R. de Kam 2009: Schitterend! Twintig eeuwen glas uit Utrechtse bodem, Utrecht. Jackson, C.M., C.A. Booth & J.W. Smedley 2005: Glass by design? Raw materials, recipes and compositional data, Archaeometry, 47: 781–95. Jankowiak, M., 2021: Dirham flows into northern and eastern Europe and the rhythms of the slave trade with the Islamic world, in: J. Gruszczyński, M. Jankowiak & J. Shepard (eds), Viking-Age Trade: Silver, Slaves and Gotland, Routledge: Abingdon and New York. https://doi.org/10.4324 /9781315231808 Jezeer, W. & S. Jongma 2002: Valkenburg-De Woerd. Werkput 510 – De Geul. Een studie naar vroegmiddeleeuws aardewerk uit Valkenburg (ZH), Amsterdam (unpublished thesis, University of Amsterdam). 120 — Kamber, B.S., A. Greig & K.D. Collerson 2005: A new estimate for the composition of weathered young upper continental crust from alluvial sediments, Queensland, Australia, Geochimica et Cosmochimica Acta, 69(4), 1041–1058. Kaspers, A., 2020: WijnaldumTjitsma revisited. Testing the potential value of fieldsurveying terp sites, in: A. Nieuwhof (ed.), The excavations at Wijnaldum, Volume 2: Handmade and wheel-thrown pottery of the first millennium AD, Groningen: Groningen Archaeological Studies 38, University of Groningen & Barkhuis Publishing, 193–240. Kato, N., Y. Nakai & Y. Shindo 2009: Change in chemical composition of early Islamic glass excavated in Raya, Sinai Peninsula, Egypt: on-site analyses using a portable X-ray fluorescence spectrometer, Journal of Archaeological Science, 36, 1698–1707. Keller, C., 2012: Karolingerzeitliche Keramikproduktion am rheinischen Vorgebirge, in: L. Grunwald, H. Pantermehl & R. Schreg (eds), Hochmittelalterliche Keramik am Rhein. Eine Quelle für Produktion und Alltag des 9. bis 12. Jahrhunderts, Mainz: RGZM-Tagungen 13, 209–224. Kind, T., A. Kronz & K.H. Wedepohl 2002: Karolingerzeitliches Glas und verschiedene Handwerksindizien aus dem Kloster Fulda. Aufarbeitung der Altfunde Joseph Vonderaus von 1898–99, in: Zeitschrift für Archäologie des Mittelalters, 31, 2003, 2004: 61–93. Koch, U., 1977: Das Reihengräberfeld bei Schretzheim, Berlin: Mann. Koch, U., 1987: Der runde Berg bei Urach VI, Die Glas- und Edelsteinfunde aus den Plangrabungen 1967–1983, Heidelberger Akademie der Wissenschaften, Komm. Alamannische Altertumskunde 12, Sigmaringen. Kouwatli, I., H.H. Curvers, B. Stuart, Y. Sablerolles, J. Henderson & P. Reynolds 2008: A pottery and glass production site in Beirut (BEY 015), Bulletin d’Archéologie et d’Architecture Libanaise, 10, 103–129. Kronz, A., V. Hilberg, K. Simon & K.H. Wedepohl 2016: Glas aus Haithabu, Zeitschrift für Archäologie des Mittelalters, 43, 39–58. Krüger, I. & K.H. Wedepohl 2003: Composition and shapes of glass of the early medieval period (8th to 10th century AD) in Central Europe, Échanges et commerce du verre dans le monde antique, 93–100. Lahlil, S., I. Biron, M. Cotte, J. Susini & N. Menguy 2010: Synthesis of calcium antimonate nano-crystals by the 18th dynasty Egyptian glassmakers, Applied Physics A, 98(1), 1–8. Langbroek, M., 2021a: Beads and beadmaking in the early medieval settlement at Oegstgeest, in: J. De Bruin, C. Bakels & F. Theuws (eds), Oegstgeest. A riverine settlement in the early medieval world system, Bonn, 2021, 278–293. Langbroek, M.B., 2021b: Beads from Dorestad, in: A. Willemsen & H. Kik (eds), Dorestad and its networks, Proceedings of the third ‘Dorestad Congress’ held at the National Museum of Antiquities, Leiden, the Netherlands 12–15 June 2019: 55–99. Lassaunière, G., I. Pactat, B. Gratuze & É. Louis 2016: L’artisanat du verre au haut Moyen Âge dans le nord de la France (Nord et Pas-de-Calais), Bulletin de l’Association Française pour l’Archéologie du Verre, 76–82. Later, C., 2010: Von Trinkgläsern und Hängelampen, Gedanken zur Funktion frühmittelalterlicher Sturzbecher, in: G. Suhr & K. Hösch (eds), Bajuwarenhof Kirchheim – Projekt für lebendige Archäologie des frühen Mittelalters, Jahresschrift 2009, München 2010, 65-76. Louis, E., 2015: Les indices d’artisanat dans et autour du monastère de Hamage (Nord), Bulletin du centre d’études médiévales d’Auxerre, BUCEMA [on-line], Hors-série n° 8, mis en ligne le 28 janvier 2015: consulté le 16 décembre 2021. https://doi.org/10.4000/ cem.13684. Lund Feveile, L., 2006: Hulsglasskår fra markedspladsen i Ribe, ASR 9, Posthuset, in: C. Feveile (ed.), Ribe Studier. DetÆldste Ribe. Udgravninger på nordsiden af Ribe Å 1984-2000, Bind 1.1 (Jysk Arkologisk Selskabs skrifter 51). Moesgård, 195-278. Lundström, P., 1981: De kommo vida ... Vikingars hamn vid Paviken på Gotland, Stockholm (Sjöhistoriska museets rapportserie 15). Magendans, J.R. & J.A. Waasdorp 1989: Franken aan de Frankenslag. Een vroegmiddeleeuwse nederzetting in ’s-Gravenhage, ’s-Gravenhage (VOM-reeks 1989-2). Matin, M., 2019: Tin-based opacifiers in archaeological glass and ceramic glazes: a review and new perspectives, Archaeological and Anthropological Sciences, 11(4), 1155–1167. Matin, M., M. Tite & O. Watson 2018: On the origins of tin-opacified ceramic glazes: New evidence from early Islamic Egypt, the Levant, Mesopotamia, Iran, and Central Asia. Journal of Archaeological Science, 97, 42–66. Matthes, C., 1998: Die Glasperlen des merowingerzeitlichen Gräberfeldes Griesheim, Magisterarbeit Berlin. Meharg. A. A., K. J.Edwards, J. E. Scholfield, A. Raab, J. Feldmann, A. Moran, C. L. Bryant, B.Thornton & J.J. C. Dawson 2012: First comprehensive peat depositional records for tin, lead and copper associated with the antiquity of Europe’s largest cassiterite deposits, Journal of Archaeological Science, 39, 717-727. 121 — McCormick, M., 2001: Origins of the European economy. Communications and commerce, AD 300–900, New York: Cambridge University Press. Meek, A., J. Henderson & J. Evans 2012: The isotopic analysis of English Forest glass from the Weald and Staffordshire, Journal of Analytical Atomic Spectroscopy, 27, 786–795. Miedema, M., 1983: Vijfentwintig eeuwen bewoning in het terpenland ten noordwesten van Groningen, (PhD dissertation, Vrije Universiteit, Amsterdam). Mirti, P., M. Pace, M.M. Negro Ponzi & M. Aceto 2008: ICP-MS analysis of glass fragments of Parthian and Sasanian epoch from Seleucia and Veh Ardašīr (central Iraq), Archaeometry, 50, 429–50. Mirti, P., M. Pace, M. Malandreno & M. Negro Ponzi 2009: Sasanian glass from Veh Ardašīr: new evidences by ICP-MS analysis, Journal of Archaeological Science, 36, 1061–1069. Nenna, M.-D., M. Vichy & M. Picon 1997: L’atelier de verrier de Lyon du Ier siècle après J.-C., et l’origine des verres “Romains”, Revue d’archéométrie, 21 (1997), 81–87. Nenna, M.-D., 2014: Egyptian glass abroad. HIMT glass and its markets, in: J. Bayley, C. Jackson, D. Keller & J. Price (eds) Neighbours and successors of Rome. Traditions of glass production in use in Europe and the Middle East in the later first millennium AD, Oxbow Books, Oxford, 178–193. Nenna, M.-D., 2015: Primary glass workshops in GraecoRoman Egypt: preliminary report on the excavations of the site of Beni Salama, Wadi Natrun (2003: 2005–9), in: Justine Bayley, I. Freestone & C. Jackson (eds), Glass of the Roman world, Oxbow Books, Oxford, 1–22. Neri, E., B. Gratuze & N. Schibille 2018: The trade of glass beads in early medieval Illyricum: towards an Islamic monopoly, Archaeological and Anthropological sciences 11, 1107-1122. https://doi. org/10.1007/s12520-0170583-5. Nicolay, J.A.W., 2014: The splendour of power. Early medieval kingship and the use of gold and silver in the southern North Sea area (5th to 7th century AD), Groningen: Groningen Archaeological Studies 28: Barkhuis & University of Groningen Library. Nieuwhof, A., 2006: De wierde Wierum (provincie Groningen). Een Archeologisch Steilkantonderzoek, Groningen: Groningen Archaeological Studies 3. Nieuwhof, A. (ed.) 2020: The Excavations at Wijnaldum, Vol. 2: Handmade and wheel-thrown Pottery of the first Millennium AD, Groningen: Groningen Archaeological Studies 34. Nokkert, M., A.C. Aarts & H.L. Wynia 2009: Vroegmiddeleeuwse bewoning langs de A2. Een nederzetting uit de zevende en achtste eeuw in Leidsche Rijn, Utrecht (Basisrapportage archeologie 26 Gemeente Utrecht). Norde, E.H.L.D., 2019: Nederzettingsresten uit de vroege middeleeuwen in het plangebied Leeuwesteyn Noord in Leidsche Rijn, gemeente Utrecht, Archeologisch onderzoek: een opgraving, Weesp: RAAPRapport 3855. Nyst, C.L., 2003: Karolingisch Glas van Dorestad. Een glasinventarisatie van vier opgravingen , (unpublished BA thesis, University of Amsterdam). Pactat, I., M. Guérit, L. Simon, B. Gratuze, S. Raux & C. Aunay 2017: Evolution of glass recipes during the Early Middle Ages in France: analytical evidence of multiple solutions adapted to local contexts, in: S. Sophie & A. Pury-Gysel (eds), Annales du 20e congrès de l’Association Internationale pour l’Histoire du Verre: Fribourg/Romont, 7–11 septembre 2015: Association Internationale pour l’Histoire du Verre, 334–340 Pactat, I., 2021: L’activité verrière en France entre le viiie et le xie siècle. Résilience et mutations d’une production artisanale, Thèse de doctorat en archéologie, sous la direction de Philippe Barral et de Danièle Foy, soutenue le 3 septembre 2020 à l’Université de Bourgogne Franche-Comté, Laboratoire Chronoenvironnement/MSHE C.-N. Ledoux. Bulletin du centre d’études médiévales d’Auxerre| BUCEMA, (25.1). https://doi. org/10.4000/cem.18365 Paynter, S., 2008: Experiments in the reconstruction of Roman wood-fired glassworking furnaces: waste products and their formation processes, Journal of Glass Studies, 50, 271–290. Peake, J.R. & I.C. Freestone 2012: Cross-craft interactions between metal and glass working: slag additions to early Anglo-Saxon red glass, in: W, Meulebroeck, K. Nys, D. Vanclooster & H. Thienpont (eds) Integrated approaches to the study of historical glass, IA1S 12, Proceedings of SPIE vol. 8422, 1-12, International Society for Optics and Photonics. Peake, J.R. & I.C. Freestone 2014: Opaque yellow glass production in the early medieval period: new evidence, Neighbours and successors of Rome: traditions of glass production and use in Europe and the Middle East in the later 1st millennium AD. Oxford: Oxbow Books, 15–21. 122 — Phelps, M., I.C. Freestone, Y. Gorin-Rosen & B. Gratuze 2016: Natron glass production and supply in the late antique and early medieval Near East: The effect of the ByzantineIslamic transition, Journal of archaeological science, 75: 57–71. Philippsen, B., C. Feveile, J. Olsen, J. & S. M. Sindbæk 2021: Single-year radiocarbon dating anchors Viking Age trade cycles in time, Nature. https://doi.org/10.1038/s41586021-04240-5 Pion, C., 2014: Les perles mérovingiennes. TypoChronologie, fabrications et fonctions, Unpublished PhD Thesis, Université libre de Bruxelles, Brussels. Pol, A., 1999: Munten, in: J. Plumier-Torfs, S. PlumierTorfs, M. Regnard & W. Dijkman, Mosa nostra. De Maasvallei van Verdun tot Maastricht in de Merovingische periode (5e-8e eeuw), Namur: Carnet du Patrimoine 28, 11. Preiß, F., 2010: Tesserae and glass drops, in: A. Willemsen & H. Kik (eds), Dorestad in an international framework. New research on centres of trade and coinage in Carolingian times, Turnhout: Proceedings of the first ‘Dorestad Congress’ held at the Museum of National Antiquities Leiden, The Netherlands, June 24–27, 123–134. Raux, S., B. Gratuze, J.-Y. Langlois & E. Coffineau 2015: Indices d’une production verrière du Xe siècle à La Milesse (Sarthe), Bulletin de l’Association Française pour l’Archéologie du Verre, Actes des 29e Rencontres, Paris (Nov. 2014), 66–70. Redknap, M., 1988: Medieval pottery production at Mayen: recent advances, current problems, in: D.R.M. Gaimster, M. Redknap & H.-H. Wegner (eds), Medieval and later pottery from the Rhineland and its markets, Oxford: BAR Int. Series 440, 3–37. Risom, T., 2013: Perlemageren fra Ribe, Historien, Materialerne, Teknikkerne, Ribe. Rooksby, H.P., 1964: Yellow cubic lead-tin oxide opacifier in ancient glasses, Physics and Chemistry of Glasses, Section B of J. Soc. Glass Technol., 5, 20–25. Rossi, R., 2009: Il vetro grezzo e le altre materie prime del reli tto romano di Mljet (Meleda), Croazia Quaderni Friulani di Archeologia, 19, 193-202. Sablerolles, Y., 1992: Het glas van Gennep. De glasvondsten van een nederzetting uit de volksverhuizingstijk te Gennep (Zuid-Limburg). Doctorate thesis, University of Amsterdam. Sablerolles, Y., 1993: A Dark-Age glass complex from a Frankish settlement at Gennep (Dutch Limburg), Proceedings of the International Association for the history of glass, Vienna 1991, Amsterdam: AIHV, 197-206. Sablerolles, Y., 1999: Beads of glass, faience, amber, baked clay and metal, including the production waste from glass and amber bead making, in: J.C. Besteman, J.M. Bos, D.A. Gerrets, H.A. Heidinga & J. de Koning (eds), The excavations at Wijnaldum. Reports on Frisia in Roman and Medieval Times I, Rotterdam: Brookfield, 253–287. Sablerolles, Y., 2019: Een smeltkroes voor glas uit de vroege middeleeuwen, in: E.H.L.D. Norde (ed.), Nederzettingsresten uit de vroege middeleeuwen in het plangebied Leeuwesteyn Noord in Leidsche Rijn, deel 1, Gemeente Utrecht, Archeologisch onderzoek: een opgraving (RAAP Rapport 3855), 136–138. Sablerolles, Y., 2023: Glasproductie in het klooster van Susteren, Stoepker, H. (ed.), 2023: Sporen van Susteren, archeologische vondsten uit een Karolingisch klooster en een adellijk vrouwenstift, de basispublicatie, Venlo, 1249-1279. https://doi. org/10.17026/dans-xsf-nypw Sablerolles, Y., J. Henderson & W. Dijkman 1997: Early medieval glass bead making in Maastricht (Jodenstraat 30), the Netherlands. An archaeological and scientific investigation, in: U. von Freeden & A. Wieczorek (eds), Perlen. Archäologie, Techniken, Analysen, Bonn: Akten Internat. Perlensymposium Mannheim 1994: Koll. Vor- und Frühgeschichte 1, 295–313. Sablerolles, Y. & J. Henderson 2012: De glasvondsten, in: J. Dijkstra (ed.), Het domein van de boer en de ambachtsman. Een opgraving op het terrein van de voormalige fruitveiling te Wijk bij Duurstede: een deel van Dorestad en de villa Wijk archeologisch onderzocht, Amersfoort, (ADC monografie 12), 293–350. Sablerolles, Y. & J. de Koning 2015: Kralen in het zand: de glasvondsten, in: J. de Koning, Onder het stuifzand – Overstoven vroegmiddeleeuwse nederzettingen bij Bloemendaal. De opgravingscampagnes Groot Olmen 2005: 2006: en 2007: Inclusief een uitgebreide re-interpretatie van de opgraving Wijk aan Zee-Hoogovens uit de jaren zestig van de vorige eeuw, Zaandijk, 307–320. Sanke, M., K.H. Wedepohl & A. Kronz 2002: Karolingerzeitliches Glas aus dem Kloster Lorsch, Zeitschrift für Archäologie des Mittelalters, 30, 37–75. Schibille, N. & I.C. Freestone 2013: Composition, Production and Procurement of Glass at San Vincenzo al Volturno: An Early Medieval Monastic Complex in Southern Italy, PLoS One 8(10), e76479. https://doi.org/10.1371/journal. pone.0076479 Schibille, N., A. SterrettKrause & I.C. Freestone 2016. Glass groups, glass supply and recycling in Late Roman Carthage, Archaeological and Anthropological Sciences, 9, 1223–1241. 123 — Schibille, N., A. Meek, B. Tobias, C. Entwistle, M. Avisseau-Broustet, H. Da Mota & B. Gratuze 2016: Comprehensive chemical characterisation of Byzantine glass weights. PLoS One 11(12), p.e0168289. Schibille, N., B. Gratuze, E. Ollivier & É. Blondeau 2019: Chronology of early Islamic glass compositions from Egypt, Journal of Archaeological Science, 104, 10–18. Schibille, N., J. de J. Ares, M.T.C. García & C. Guerrot 2020: Ex novo development of lead glassmaking in early Umayyad Spain, Proceedings of the National Academy of Sciences 117 (28) 16243-16249; https://doi.org/10.1073/ pnas.2003440117 Schibille,N., V. Amorós Ruiz, J. De Juan Ares & S. Gutiérrez Lloret 2022: Rare alkali elements as markers of local glass working in medieval Tolmo de Minateda (Spain), ChemPlusChem 87, e202200147. Shortland, A.J., 2004: Evaporites of the Wadi Natrun: Seasonal and annual variation and its implication for ancient exploitation, Archaeometry, 46, 4: 497–517. Sier, M.M., J. van Doesburg & W.J.H. Verwers (eds) 2004: Wijk bij Duurstede-Frankenweg/ Zandweg, Amersfoort (ADC-rapport 282). Silvestri, A., G. Molin & G. Salviulo 2008: The colourless glass of Iulia Felix, Journal of archaeological science, 35(2), 331–341. Silvestri, A., S. Tonietto, G. Molin & P. Guerriero 2012: The palaeo-Christian glass mosaic of St. Prosdocimus (Padova, Italy): archaeometric characterisation of tesserae with antimony- or phosphorusbased opacifiers, Journal of archaeological science, 39(7), 2177–2190. Simpson, S.J., 2014: Sasanian glass: an overview, in: D. Keller, J. Price, & C. Jackson (eds), Neighbours and Successors of Rome. Traditions of Glass Production and Use in Europe and the Middle East in the Later 1st Millennium AD, Oxford: Oxbow Books, 200–231. Siu, I., J. Henderson, D. Qin, Y. Ding, J. Cui & and H. Ma 2020: New light on plant ash glass found in Africa: Evidence for Indian Ocean Silk Road trade using major, minor, trace element and lead isotope analysis of glass from the 15th—16th century AD from Malindi and Mambrui, Kenya, PLoS One, 15, p. e0237612. Smith, T., J. Henderson & E.W. Faber 2016: Early Byzantine glass supply and consumption: the case of Dichin, Bulgaria, in: Guxi Gan, Julian Henderson & Qinghui Li (eds), Recent research in the scientific investigations of ancient glass, World Scientific, 207–232. Sode, T., 2004: Glass bead making technology, in: M. Bencard, A.K. Rasmussen & H.B. Madsen (eds), Ribe excavations 1970–76 (Vol. 5), Moesgard: Jutland Archaeological Society, 83–102. Standish, C.D., S.W. Merkel, Y.-T. Hsieh, & J. Kershaw 2021: Simultaneous lead isotope ratio and gold-lead-bismuth concentration analysis of silver by laser ablation MC-ICP-MS, Journal of Archaeological Science 125, 105299. Steppuhn, P., 1998: Die Glasfunde von Haithabu, Berichte über die Ausgrabungen in Haithabu 32, Neumünster. Stoepker, H., 2021: Het klooster van Susteren (714–1802). Archeologisch onderzoek van een Karolingische abdij en een adellijk vrouwenstift, Amersfoort (Nederlandse Archeologische Rapporten 73). Szöke, B.M., K.H. Wedepohl & A. Kronz 2004: Silver-stained windows at Carolingian Zalavar, Mosaburg (southwestern Hungary), Journal of Glass Studies 46, 85–104 Tal, O., E.E. Jackson-Tal & I.C. Freestone 2004: New evidence for glass production at Lat Byzantine Apollonia-arsuf, Israel, Journal of Glass Studies, 51, 51–65. Thedéen, S., 2009: Who’s that girl? The cultural construction of girlhood and the transition to womanhood in Viking Age Gotland, Childhood in the Past 1(1), 78–93. Van Dockum, S.G., 1997: Parkeerplaats AH, in: Archeologische Kroniek Provincie Utrecht 1992–1993, Utrecht, 120–121. Van Doesburg, J., 2004: Glas, in: M.M. Sier, J. van Doesburg & W.J.H. Verwers (eds), Wijk bij DuurstedeFrankenweg/Zandweg, Amersfoort (ADC-rapport 282), 42. Van Es, W.A. & W.J.H. Verwers 1980: Excavations at Dorestad 1. The harbour: Hoogstraat 1, Amersfoort: Nederlandse Oudheden 9. Van Lith, S.M.E. & Y. Sablerolles 1995: Verres du IVe et du Ve siècle des sites d’habitat aux Pays-Bas, in: F. Foy (ed.) Le verre de l’antiquité tardive et du haut moyen âge, Association Française pour l’Archéologie du Verre, huitième rencontre, Guiry-enVexin, 18-19 Novembre 1993, 41-50. Van Os, B.J.H., R.M. Vogelzang, J.W. de Kort, D.J. Huisman, M. Kars, D.J.M. Ngan-Tillard, W. Verwaal & E. Meijvogel 2014: Kralen van glas en barnsteen, in: R.C.G.M. Lauwerier & J.W. de Kort (eds), Merovingers in een villa 2, Romeinse villa en Merovingisch grafveld Borgharen – Pasestraat Onderzoek, Amersfoort (Rapportage Archeologische Monumentenzorg 222). Van Wersch, L.F. 2013: Merovingian ceramic and glass in the early towns of the middle Meuse valley, Medieval and modern matters 4, 133-151. 124 — Van Wersch, L., F. Mathis & P. Hoffsummer 2009: Étude des verres du Haut Moyen Âge découverts dans l’église des saints Hermès et Alexandre à Theux, in: Vitrail, Verre et Archéologie entre le Ve et le XIIe siècle, Actes de la table ronde réunie au Centre d’Études Médiévales d’Auxerre, 15–16 juin 2006, Paris, 161–171. Van Wersch, L., I. Biron, B. Neuray, F. Mathis, G. Chêne, D. Strivay & C. Sapin 2014: Les vitraux alto-médiévaux de Stavelot (Belgique), ArchéoSciences 38, 219–234. Van Wersch, L., C. Loisel, F. Mathis, D. Strivay & S. Bully 2015: Analyses of Early Medieval Stained Window Glass from the Monastery of Baume-lesMessieurs (Jura, France), Archaeometry, 58, 930–946. Van Wersch, L., B. Gratuze, F. Mathis, M. Bonnin, D. Strivay, H. Da Mota Rocha & C. Sapin 2018: Glass tiles from SaintSauveur (Burgundy, France), Journal of Glass Studies, 60, 163-181. Van Winkelhoff, A.M., 2021: Light in the Dark Ages: a conceptual approach to the role of glass vessels in the Merovingian burial rite, (RMAthesis Leiden University, Leiden). Wedepohl, K.H., 2008: Mittelalterliches HolzascheGlas, in: H. Flachenecker, G. Himmelsbach & P. Steppuhn (eds), Glashüttenlandschaft Europa, Regensburg: Schnell and Steiner. Verhulst, A., 2002: The Carolingian Economy, Cambridge Medieval Texbooks, Cambridge: Cambridge University Press. Wedepohl, K. H. & A. Baumann 1997: Isotope composition of Medieval lead glasses reflecting early silver production in Central Europe, Mineralium Deposita 32, 292–295. Vollgraff, C.W. & G. van Hoorn 1934: Opgravingen op het Domplein te Utrecht, Wetenschappelijke Verslagen II, De opgravingen in juni en juli 1933: Haarlem. Vrielynck, O., F. Mathis & C. Pion 2018: Vers une typochronologie des perles mérovingiennes en Gaule du Nord, Namur. Wedepohl, K.H., 2003: Flachund Gefäßglas der Karolingerzeit, Nachrichten der Akademie der Wissenschaften Göttingen, II. Mathematische–Physikalische, Klasse, 3, 115–156. Wedepohl, K.H. & K. Simon 2010: The chemical composition of medieval wood ash glass from Central Europe, Geochemistry, 70(1), 89–97. Wedepohl, K.H., W. Winkelmann & G. Hartmann 1997: Glasfunde aus der karolingischen Pfalz in Paderborn und die frühe Holzasche-Glasherstellung, Ausgrabungen und Funde in Westfalen Lippe, 9A, 41–53. Willmott, H. & K. Welham 2013: Late seventh-century glassmaking at Glastonbury Abbey, Journal of Glass Studies 55, 71–83. Willmott, H. & K. Welham 2015: Saxon Glass Furnaces, in: R. Gilchrist & C. Greene (eds) Glastonbury Abbey: archaeological investigations 1904–79, Society of Antiquaries of London, London, 218–238. Wynia, H., 2013: Graven op het Domplein, in: DompleinMagazine, Special I: Het Domplein in Utrecht, Utrecht, 9–15. Yener, K.A., F. Kulakoğlu, E. Yazgan, R. Kontani, Y.S. Hayakawa, J.W. Lehner, G. Dardeniz, G. Öztürk, M. Johnson, R. Kaptan & A. Hacar 2015: New tin mines and production sites near Kültepe in Turkey: a thirdmillennium BC highland production model, Antiquity, 89, 596-612. Appendices Appendix I sample list Appendix II major and minor chemical compositions of samples analysed by electron probe microanalysis Appendix III trace element chemical compositions of samples analysed by LA-ICP-MS Appendix IV photos of the samples from Maastricht and Utrecht 125 — 126 — Appendix I sample list The list of the samples from the different sites used in this study: Maastricht (Jodenstraat and Mabro sites), Gennep, Wijnaldum, Utrecht, Wijk bij Duurstede (Dorestad), Susteren and Deventer. Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Maastricht-Jodenstraat 1 1-1-7 1/7 pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 2 1-1-7 1/7 pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 3 1-1-7 1/7 pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 4 01-01-2007 1/7 pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 5 1-1-7 1/7 pit 580/90 610/20 production waste thread Maastricht-Jodenstraat 6 1-1-7 1/7 pit 580/90 610/20 production waste rod Maastricht-Mabro 7 3-4-0 0 pit - - production waste crucible Maastricht-Mabro 8 3-OA-55 3/55 pit - - production waste crucible Maastricht-Mabro 9 1-3-51 1/51 pit - - production waste crucible? Maastricht-Mabro 10 3-OA-1 3/1 pit - - production waste crucible Maastricht-Mabro 11 1-5-OA 1/OA pit - - production waste crucible Maastricht-Mabro 12 3-AA'-400 3/400 pit - - production waste crucible Maastricht-Mabro 13 2-2-18 2/18 pit - - production waste crucible Maastricht-Mabro 14 3-5-24 3/24 pit - - production waste crucible Maastricht-Mabro 15 3-5-24 3/24 pit - - production waste crucible Maastricht-Mabro 16 3-4-12 3/12 pit - - production waste crucible Maastricht-Jodenstraat 17 1-1-7 1/7 pit 580/90 610/20 production waste crucible Maastricht-Jodenstraat 18 1-1-7 1/7 pit 580/90 610/20 production waste crucible Maastricht-Jodenstraat 19 1-1-7 1/7 pit 580/90 610/20 production waste crucible 127 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication section - rod, red with weathered surface - opaque - x 2 Sablerolles et al. 19997 section - rod, red with weathered surface - opaque - x 2 Sablerolles et al. 19997 section - white twisted rod - opaque - - 2 Sablerolles et al. 19997 section - white rod - opaque - - 2 Sablerolles et al. 19997 section - white rod (thin, c 2-3 mm) - opaque - - 2 Sablerolles et al. 19997 section - yellow rod > 5 mm - opaque - - 2 Sablerolles et al. 19997 rim - crucible rim fragment with white material (frit?) on both sides - opaque - - 2 Sablerolles et al. 19997 body - crucible base fragment with colourless glass on both sides, white on the inside - opaque/ translucent - - 2 Sablerolles et al. 19997 ? - small crucible fragment with green and weathered opaque yellow glass - translucent/ opaque - x 2 Sablerolles et al. 19997 rim - crucible rim with colourless glass on inside - transparent - - 2 Sablerolles et al. 19997 rim - crucible rim fragment with thick white material (frit?) on both sides; colourless glass on the inside - opaque/ transparent - - 2 Sablerolles et al. 19997 base - crucible base fragment with opaque yellow glass on inside. colourless vitrification on the outside. - opaque/ transparent - - 2 Sablerolles et al. 19997 ceramic - small fragment of red ceramic with (natural?) green glass on both sides - translucent - - 2 Sablerolles et al. 19997 base - red ceramic pot base with yellow (outside) and yellowcolourless (inside) glass - opaque/ transparent - - 2 Sablerolles et al. 19997 ceramic - small ceramic fragment with deep translucent glass on inside - transparent - - 2 Sablerolles et al. 19997 base - thick red ceramic base with deep translucent green glass on both sides, partially weathered glass on both sides - transparent - x 2 Sablerolles et al. 19997 base - crucible base of reddish-grey ceramic with opaque yellow glass on inside - opaque - - 2 Sablerolles et al. 19997 base - crucible base of grey ceramic, with weathered opaque yellow glass on inside. - opaque - x 2 Sablerolles et al. 19997 base - crucible base of red ceramic, weathered opaque yellow glass on inside - opaque - x 2 Sablerolles et al. 19997 128 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Maastricht-Jodenstraat 20 1-1-7 1/7 pit 580/90 610/20 production waste crucible Maastricht-Jodenstraat 21 1-1-7 1/7 pit 580/90 610/20 production waste crucible Maastricht-Jodenstraat 22 1-1-7 1/7 pit 580/90 610/20 production waste crucible Maastricht-Jodenstraat 23 1-1-7 1/7 pit 580/90 610/20 production waste crucible Maastricht-Jodenstraat 24 1-1-7 1/7 pit 580/90 610/20 production waste crucible Maastricht-Jodenstraat 25 1-1-7 1/7 pit 580/90 610/20 production waste crucible Maastricht-Jodenstraat 26 1-1-7 1/7 pit 580/90 610/20 production waste crucible Maastricht-Jodenstraat 27 1-1-7 1/7 pit 580/90 610/20 production waste crucible Maastricht-Jodenstraat 28 1-1-7 1/7 pit 580/90 610/20 production waste crucible Maastricht-Jodenstraat 29 1-2-3 1/3 pit 580/90 610/20 production waste crucible Maastricht-Jodenstraat 30 1-1-7 1/7 pit 580/90 610/20 production waste brick/tegula? Maastricht-Jodenstraat 37 1-1-7 1/7 pit 580/90 610/20 production waste drop Maastricht-Jodenstraat 38 1-1-7 1/7 pit 580/90 610/20 production waste drop Maastricht-Jodenstraat 39 1-1-7 1/7 pit 580/90 610/20 production waste ingot? Maastricht-Jodenstraat 40 1-1-7 1/7 pit 580/90 610/20 production waste ingot? Maastricht-Jodenstraat 41 1-1-7 1/7 pit 580/90 610/20 production waste punty? Maastricht-Jodenstraat 42 1-1-7 1/7 pit 580/90 610/20 production waste punty? Maastricht-Jodenstraat 43 1-1-7 1/7 pit 580/90 610/20 production waste punty? Maastricht-Jodenstraat 44 1-2-3 1/3 pit 580/90 610/20 window flat quite thick window glass; one side worked with grozing (sp.) iron? Maastricht-Jodenstraat 45 1-2-3 1/3 pit 580/90 610/20 window thin window glass; two sides been worked with grozing (sp.) iron? 129 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication base - crucible base of red ceramic, layer of opaque yellow glass on inside - opaque - - 2 Sablerolles et al. 19997 base - crucible base of thin grey ceramic, colourless glass - translucent - - 2 Sablerolles et al. 19997 base - crucible base with opaque yellow glass and white material - opaque - - 2 Sablerolles et al. 19997 base - crucible base with opaque yellow glass and white material - opaque/ translucent - - 2 Sablerolles et al. 19997 base - crucible base of red ceramic, opaque yellow and brownish red vitrification - opaque - - 2 Sablerolles et al. 19997 base - crucible base of red ceramic, opaque yellow and redishbrown vitrification - opaque - - 2 Sablerolles et al. 19997 base - crucible base of red ceramic, opaque yellow glass - opaque - - 2 Sablerolles et al. 19997 base - crucible base of red ceramic, opaque yellow glass - opaque - - 2 Sablerolles et al. 19997 base - crucible base of red ceramic, opaque yellow glass - opaque - - 2 Sablerolles et al. 19997 base - crucible base of red ceramic, deep translucent green glass, esp. thick on bottom - translucent - - 2 Sablerolles et al. 19997 - - possible furnace brick fragment with weathered opaque yellow glass - opaque - x 2 Sablerolles et al. 19997 complete - blue glass drop, c. 1 cm - translucent - - 2 Sablerolles et al. 19997 complete - blue glass drop, c. 1 cm - translucent - - 2 Sablerolles et al. 19997 crushed - small (crushed?) blue glass fragments - translucent - - 2 Sablerolles et al. 19997 crushed - small (crushed?) blue glass fragments - translucent - - 2 Sablerolles et al. 19997 undiagnostic - fragment of opaque red glass - opaque - - 2 Sablerolles et al. 19997 undiagnostic - fragment of opaque red glass - opaque - - 2 Sablerolles et al. 19997 undiagnostic - fragment of blue-green glass - translucent - x 2 Sablerolles et al. 19997 - - flat quite thick translucent yellowish window glass; one side worked with grozing (sp.) iron? - translucent - - 2 Sablerolles et al. 19997 - - thin amber window glass; two sides been worked with grozing iron - translucent - - 2 Sablerolles et al. 19997 130 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Maastricht-Jodenstraat 46 1-2-3 1/3 pit 580/90 610/20 window moderately thin window glass. One side rounded. Maastricht-Jodenstraat 47 1-1-7 1/7 pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 48 1-1-7 1/7 pit 580/90 610/20 production waste drop Maastricht-Jodenstraat 49 1-1-7 1/7 pit 580/90 610/20 production waste drop Maastricht-Jodenstraat 50 1-1-7 1/7 pit 580/90 610/20 production waste drop Maastricht-Jodenstraat 51 1-1-7 1/7 pit 580/90 610/20 production waste drop Maastricht-Jodenstraat 52 1-1-7 1/7 pit 580/90 610/20 production waste drop Maastricht-Jodenstraat 53 1-1-7 1/7 pit 580/90 610/20 production waste drop Maastricht-Jodenstraat 54 1-1-7 1/7 pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 55 1-1-7 1/7 pit 580/90 610/20 production waste drop Maastricht-Jodenstraat 56 1-1-7 1/7 pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 57 1-1-3 1/3 pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 58 ? - pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 59 ? - pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 60 1-2-5 1/5 pit 580/90 610/20 vessel cone Maastricht-Jodenstraat 61 1-2-5 1/5 pit 580/90 610/20 bead bead Maastricht-Jodenstraat 62 1-1-7 1/7 pit 580/90 610/20 production waste drop Maastricht-Jodenstraat 63 1-1-7 1/7 pit 580/90 610/20 production waste drop Maastricht-Jodenstraat 64 1-1-7 1/7 pit 580/90 610/20 production waste ? Maastricht-Jodenstraat 65 1-1-7 1/7 pit 580/90 610/20 production waste ? Maastricht-Jodenstraat 66 1-1-7 1/7 pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 67 1-1-7 1/7 pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 68 1-1-7 1/7 pit 580/90 610/20 vessel ribbed bowl Maastricht-Jodenstraat 69 1-1-7 1/7 pit 580/90 610/20 production waste punty 131 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication - - moderately thin pale green window glass. one side rounded. - translucent - - 2 Sablerolles et al. 19997 section - thin opaque olive green rod - opaque - - 2 Sablerolles et al. 19997 - - opaque green drop - opaque - - 2 Sablerolles et al. 19997 - - irregular green drop with soil fused to it - translucent - - 2 Sablerolles et al. 19997 - - drop of weathered opaque yellow glass - opaque - - 2 Sablerolles et al. 19997 - - irregular drop of reddish glass - opaque - - 2 Sablerolles et al. 19997 - - irregular drop of deep translucent (‘black’) glass - translucent - x 2 Sablerolles et al. 19997 - - stretched pale opaque blue piece of rod? - opaque - - 2 Sablerolles et al. 19997 section - thin opaque red rod with weathered exterior - opaque - x 2 Sablerolles et al. 19997 - - irregular red drop, weathered surface - opaque - - 2 Sablerolles et al. 19997 section - double rod; yellow with slightly greenish tint - opaque - - 2 Sablerolles et al. 19997 section - blue-green rod - opaque - - 2 Sablerolles et al. 19997 section - twisted opaque white rod - opaque - - 2 Sablerolles et al. 19997 section - white rod - opaque - - 2 Sablerolles et al. 19997 base - base of translucent green cone - translucent - - 2 Sablerolles et al. 19997 half - fragmented tapering cobalt blue bead - opaque - - 2 Sablerolles et al. 19997 drop - drop; naturally coloured yellowish - translucent - - 2 Sablerolles et al. 19997 drop - irregular drop; naturally coloured yellowish - translucent - - 2 Sablerolles et al. 19997 undiagnostic - thin opaque yellow fragment - opaque - - 2 Sablerolles et al. 19997 undiagnostic - thin opaque yellow fragment - opaque - - 2 Sablerolles et al. 19997 section - rod, red with weathered surface - opaque - x 2 Sablerolles et al. 19997 section - rod, red with weathered surface - opaque - x 2 Sablerolles et al. 19997 ribbed bowl - quite thick blue green glass fragment with a rib (roman?) - translucent - - 2 Sablerolles et al. 19997 - - “punty glass” yellow thinwalled fragment - opaque - - 2 Sablerolles et al. 19997 132 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Maastricht-Jodenstraat 70 1-1-7 1/7 pit 580/90 610/20 production waste punty Maastricht-Jodenstraat 71 1-1-7 1/7 pit 580/90 610/20 production waste punty Maastricht-Jodenstraat 72 1-1-7 01-jul pit 580/90 610/20 production waste punty Maastricht-Jodenstraat 73 ? ? pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 74 ? ? pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 75 ? ? pit 580/90 610/20 production waste rod Maastricht-Jodenstraat 76 ? ? pit 580/90 610/20 production waste rod Gennep GE 41 4481 35/2 sunken hut 400 550 vessel cone Gennep GE 42 4060 28/8 sunken hut 400 550 vessel cone Gennep GE 43 1749 13/34 sunken hut 400 550 vessel cone Gennep GE 44 2115 20/37 sunken hut 400 550 vessel cone Gennep GE 45 3079 19/26 sunken hut 400 550 vessel cone Gennep GE 46 2527 6/10 sunken hut 400 550 vessel cone Gennep GE 47 2557 7/120 sunken hut 400 550 vessel cone Gennep GE 48 1313 8/54 sunken hut 400 550 vessel cone Gennep GE 49 2278 6/1 sunken hut 400 550 vessel cone Gennep GE 50 1399 8/35 sunken hut 400 550 vessel cone Gennep GE 51a 2795 10/38 well 400 550 vessel cone Gennep GE 51b 1397 8/35 sunken hut 400 550 vessel cone Gennep GE 52 - - - 400 550 vessel cone Gennep GE 53 2598 10/1 sunken hut 400 550 vessel bowl Gennep GE 54 4254 27/13 sunken hut 400 550 vessel bowl 133 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication - - “punty glass” yellow thinwalled fragment - opaque - - 2 Sablerolles et al. 19997 - - “punty glass” yellow thinwalled fragment - opaque - - 2 Sablerolles et al. 19997 - - “punty glass” yellow thinwalled fragment - opaque - - 2 Sablerolles et al. 19997 section - turquoise green rod fragment - opaque - - 2 Sablerolles et al. 19997 section - turquoise green rod fragment - opaque - - 2 Sablerolles et al. 19997 section - green rod fragment - opaque - - 2 Sablerolles et al. 19997 section - green rod fragment - opaque - - 2 Sablerolles et al. 19997 base Koch 1987 III? - yellow-green translucent - - 1 Sablerolles 1992, cat. 121.1 base Koch 1987 III I - olive brown - - - 1 Sablerolles 1992, cat. 85.1 base Koch 1987 III? - olive green - - - 1 Sablerolles 1992, cat. 134.2? rim Koch 1987 III H - olive green - selfcoloured spiral - 1 Sablerolles 1992, cat. 81.6 rim Koch 1987 III? - light bluegreen - selfcoloured spiral - 1 Sablerolles 1992, cat. 127.1 rim Koch 1987 III I - yellow-green - selfcoloured spiral - 1 Sablerolles 1992, cat. 86.3 rim Koch 1987 III? - pale yellow/ colourless - selfcoloured spiral - 1 Sablerolles 1992, cat. 124.1 rim Koch 1987 III? - pale green/ colourless - white spiral - 1 Sablerolles 1992, cat. 123.1 rim Koch 1987 III I? - pale green/ colourless - selfcoloured spiral - 1 Sablerolles 1992, cat. 155.2 rim Isings 1957 106b2? - yellow-green - brown spiral - Roman Sablerolles1992, cat. 2.3 rim Isings 1957 106c2? - yellow-green - brown spiral - Roman Sablerolles 1992, cat. 5.1 body Isings 1957 106c2? - yellow-green - brown arcades - Roman Sablerolles 1992, cat. 5.2 rim Koch 1987 III? - light yellowgreen - selfcoloured spiral - 1 Sablerolles 1992, cat. - rim Koch 1987 IV K - light yellowgreen/ colourless - - - 1 Sablerolles 1992, cat. 141.1 rim Koch 1987 IV M - light greenyellow - selfcoloured spiral - 1 Sablerolles 1992, cat. 17 134 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Gennep GE 55 - - - 400 550 vessel bowl Gennep GE 56 1915 16/38 sunken hut 400 550 vessel bowl Gennep GE 57 - - - 400 550 vessel bowl Gennep GE 58 - - - 400 550 vessel bowl Gennep GE 59 - - - 400 550 vessel bowl Gennep GE 60 4549 35/2 sunken hut 400 550 vessel bowl Gennep GE 61 4549 35/2 sunken hut 400 550 vessel bowl Gennep GE 62 4464 35/2 sunken hut 400 550 vessel bowl Gennep GE 63 4563 35/2 sunken hut 400 550 vessel bowl Gennep GE 64 4549 35/2 sunken hut 400 550 vessel bowl Gennep GE 65 4901 41/3 post hole sunken hut 400 550 vessel bowl Gennep GE 66 4257 32/20 pit 400 550 vessel bowl Gennep GE 67 1316 8/54 sunken hut 400 550 vessel bowl Gennep GE 68 1512 8/8 sunken hut 400 550 vessel bowl Gennep GE 69 1500 8/8 sunken hut 400 550 vessel bottle Wijnaldum WIJ 1 9782 2351 occupation surface 450 500 vessel cone (Kempston) Wijnaldum WIJ 2 6794 1346 waste deposit? 450 500 vessel bowl Wijnaldum WIJ 3 10802 2565 sunken hut/ waste deposit 600 700 vessel bowl Wijnaldum WIJ 4 6242 801 ditch 450 500 bead small glob Wijnaldum WIJ 5 1356 625 area with metal waste/ waste deposit 575 625 bead small glob Wijnaldum WIJ 6 1428 608 well 9 575 625 bead small glob 135 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication rim Koch 1987 IV M - pale green - selfcoloured spiral - 1 Sablerolles 1992, cat. - rim Koch 1987 IV M - pale bluegreen - selfcoloured spiral - 1 Sablerolles 1992, cat. 172.1 rim Koch 1987 IV M - pale green/ colourless - selfcoloured spiral - 1 Sablerolles 1992, cat. 173.2 body Koch 1987 IV M - pale green - selfcoloured spiral - 1 Sablerolles1992, cat. - base Koch 1987 IV M - light yellowgreen - selfcoloured spiral - 1 Sablerolles 1992, cat. - rim Koch 1987 IV B? - light yellow/ colourless - - - 1 Sablerolles 1992, cat. 136.2 base Koch 1987 IV B? - light yellow/ colourless - - - 1 Sablerolles 1992, cat. 136.3 body Koch 1987 IV K - light yellowgreen/ colourless - white festoons - 1 Sablerolles 1992, cat. 147.1 rim Koch 1987 IV L - light greenblue/ colourless - white feather - 1 Sablerolles 1992, cat. 151.5 rim Koch 1987 IV L - light greenblue - white feather - 1 Sablerolles 1992, cat. 151.5 rim Koch 1987 IV L - light greenyellow/ colourless - white feather - 1 Sablerolles 1992, cat. 154.1 body Koch 1987 IV L - light bluegreen - white feather - 1 Sablerolles 1992, cat. 150.1 base Koch 1987 IV ? - light bluegreen - - - 1 Sablerolles 1992, cat. 162 body Koch 1987 IV ? - light bluegreen - - - 1 Sablerolles 1992, cat. 161 body Isings 1957 101 - yellow-green - red streaks - Roman Sablerolles 1992, cat. 1.3 body Koch 1987 III N - pale bluish green transparent selfcoloured loops x 1 Sablerolles 1999 VESSEL cat. 7 rim Koch 1987 IV L - pale bluish green transparent white feather x 1 Sablerolles 1999 VESSEL fig. 1.9 rim Koch 1987 IV - light bluish green transparent white spiral x 1 Sablerolles 1999 VESSEL cat. 10 complete Pion 2014 B1.1-2b - yellow opaque - - 1 Sablerolles 1999, BEAD fig. 5.13 - Pion 2014 B1.1-2a - yellow opaque - - 2 Sablerolles 1999 BEAD cat. 14 - Pion 2014 B1.1-2b - yellow opaque - - 2 Sablerolles 1999 BEAD cat. 15 136 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Wijnaldum WIJ 7 1462 1167 sod layer 600 700 bead small glob Wijnaldum WIJ 8 3901 574 ditch 575 625 bead small glob Wijnaldum WIJ 9 3901 574 ditch 575 625 bead small glob Wijnaldum WIJ 10 3695 2605 truncated layer/waste deposit 800 850 vessel funnel Wijnaldum WIJ 11 7359 1636 truncated layer/waste deposit 800 900 vessel funnel Wijnaldum WIJ 12 7507 3296 truncated layer/ occupation surface 775 850 vessel funnel Wijnaldum WIJ 13 7877 3358 truncated layer/ ooccupation surface 800 850 vessel funnel Wijnaldum WIJ 14 6704 1233 ditch 550 600 bead small glob Wijnaldum WIJ 15 10906 340 occupation surface 770 900 vessel funnel Wijnaldum WIJ 16 1526 1114 sod layer 700 750 vessel jar? Wijnaldum WIJ 17 5812 2098 occupation surface/waste deposit 450 550 bead small glob Wijnaldum WIJ 18 6712 1330 occupation surface 450 550 bead small annular Wijnaldum WIJ 19 5534 1233 ditch 550 600 bead small glob Wijnaldum WIJ 20 7448 1233 ditch 550 600 bead small glob Wijnaldum WIJ 21 2655 625 met/wd 575 625 bead small glob Wijnaldum WIJ 22 5632 975 cultivation layer 550 560 bead small glob Wijnaldum WIJ 23 6704 6704 ditch 550 600 bead small glob Wijnaldum WIJ 24 1461 548 occupation surface/waste desposit 550 600 bead short cylindrical Wijnaldum WIJ 25 5666 1232 ditch 550 650 bead short cylindrical Wijnaldum WIJ 26 7448 1233 ditch 550 600 bead short cylindrical Wijnaldum WIJ 27 2454 558 cultivation layer 500 550 bead short cylindrical 137 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication - Pion 2014 B1.1-2a - yellow opaque - - 2 or 3 Sablerolles 1999 BEAD cat. 16 - Pion 2014 B1.1-2a - yellow opaque - - 2 Sablerolles 1999 BEAD cat. 18 - Pion 2014 B1.1-2a - yellow opaque - - 2 Sablerolles 1999 BEAD cat. 19 rim Lund Feveil 2006, rim type e - light (blue-) green translucent - x 4 Sablerolles 1999 VESSEL fig. 1.20 rim Lund Feveil 2006, rim type e - almost colourless transparent - x 4 Sablerolles 1999 VESSEL fig. 1.21 rim Lund Feveil 2006, rim type d - yellowish green translucent - x 4 Sablerolles 1999 VESSEL fig. 1.22 rim Lund Feveil 2006, rim type e - almost colourless transparent - x 4 Sablerolles1999 BEAD fig. 1.23 - Pion 2014 B1.1-02b - yellow opaque - - 2 Sablerolles 1999 BEAD cat. 25 rim Lund Feveil 2006, rim type g - dark blue translucent incalmo rim x 4 Sablerolles 1999 VESSEL fig. 1.26 rim - - blue-green translucent yellow spiral, white arcade - 3/4 Sablerolles 1999 VESSEL fig. 1.30 - Pion 2014 B1.1-3b - red opaque - - 2 Sablerolles 1999 BEAD cat. 46 complete Pion 2014 B1.1-3b - red opaque - - 2 Sablerolles 1999 BEAD fig. 5.47 - Pion 2014 B1.1-3b - red opaque - - 2 Sablerolles 1999 BEAD cat. 49 - Pion 2014 B1.1-3b - red opaque - - 2 Sablerolles1999 BEAD cat. 53 - Pion 2014 B.1.1-4a - white opaque - - 2 Sablerolles 1999 BEAD cat. 56 - Pion 2014 B.1.1-4a - white opaque - - 2 Sablerolles 1999 BEAD cat. 57 complete Pion 2014 B1.1-4a - white opaque - - 2 Sablerolles 1999, fig. 5.58 complete Pion 2014 B1.4-1a - yellow opaque - - 2 Sablerolles 1999, fig. 5.66 - Pion 2014 B1.4-3a - white opaque - - 2 Sablerolles 1999 BEAD cat. 73 - Pion 2014 B1.4-3a - white opaque - - 2 Sablerolles 1999 BEAD cat. 78 - Pion 2014 B1.4-2a - red opaque - - 2 Sablerolles 1999 BEAD cat. 80 138 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Wijnaldum WIJ 28 10824 1384 truncated layer/waste deposit 650 750 bead short cylindrical Wijnaldum WIJ 29 11090 2546 ditch 650 750 bead short cylindrical Wijnaldum WIJ 30 6884 575 well 8 770 850 bead short cylindrical Wijnaldum WIJ 31 7448 1233 ditch 550 600 bead short cylindrical Wijnaldum WIJ 32 1024 1079 sod layer 640 750 bead biglobular Wijnaldum WIJ 33 6704(2) 1233 ditch 550 600 bead biglobular Wijnaldum WIJ 34 6562(1) 2064 sod layer 500 550 bead irregular spiral Wijnaldum WIJ 35 3316(2) 3532 pit? 775 850 bead segmented, ‘gold’ foil Wijnaldum WIJ 36 6562(1) 2064 sod layer 500 550 bead segmented, gold foil Wijnaldum WIJ 37 3326(1) 3542 ditch/ occupation surface? 875 900 bead segmented, silver foil Wijnaldum WIJ 38 9737(1) 2341 occupation surface/sod layer 450 500 bead segmented, silver foil Wijnaldum WIJ 39 10608(1) 100 well 6 750 850 bead segmented, layered Wijnaldum WIJ 40 10786 514 ditch 750 770 production waste tessera Wijnaldum WIJ 41 3829 696 occupation suface 425 500 production waste rod Wijnaldum WIJ 42 4601 2817 truncated layer/ occupation surface 750 800 production waste punty glass Wijnaldum WIJ 43 2943 625 area with metal waste 575 625 production waste furnace? Utrecht-Domplein 31 1933-77-36 ? ? - - production waste crucible Utrecht-Domplein 32 1933-77-53 ? ? - - production waste crucible Utrecht-Domplein 33 1933-zn3 ? ? - - production waste crucible Utrecht-Domplein 34 1933-234 ? ? - - production waste crucible 139 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication - Pion 2014 B1.4-2a - red opaque - - 3 Sablerolles 1999 BEAD cat. 81 - Pion 2014 B1.4-2a - red opaque - - 3 Sablerolles 1999 BEAD cat. 82 - Callmer 1977 A135? - red opaque - - 4 Sablerolles 1999 BEAD cat. 83 - Pion 2014 B1.4-2a - red opaque - - 2 Sablerolles 1999 BEAD cat. 84 - Pion 2014 B1.2-1b - yellow opaque - - 3 Sablerolles 1999 BEAD cat. 97 - Pion 2014B1.21b - yellow opaque - - 2 Sablerolles 1999 BEAD cat. 99 - Pion 2014 B1.8-01 - black/dark blue opaque - - 1 Sablerolles 1999 BEAD cat. 102 complete? Callmer 1977 E140? - yellowish? transparent - - 4 Sablerolles 1999 BEAD, fig. 5.115 - Pion 2014 A4.1-1 - colourless transparent - - 1 Sablerolles 1999 BEAD cat. 116 complete Callmer 1977 E140? - colourless transparent - - 4 Sablerolles 1999 BEAD, fig. 5.118 - Pion 2014 A4.2-1 - colourless transparent - - 1 Sablerolles1999 BEAD cat. 119 complete Pion 2014 A3.1-7 - red on colourless opaque - - 4 Sablerolles1999 BEAD, fig. 5.122 - - - yellow opaque - - Roman Sablerolles 1999, fig. 4, cat. 216 - - - greenish white opaque - - 1 Sablerolles 1999, fig. 4, cat. 217 - - - turquoise opaque - - 4 Sablerolles 1999, fig. 4, cat. 218 - - - yellow opaque - x 2 Sablerolles 1999, fig. 3, cat. 219 body? - crucible fragment (one of two), thin-walled grey ceramic with a thin layer of green glass - translucent - - 4 - body? - crucible fragment, thinwalled beige ceramic with a thin (cracked) layer of green glass. - translucent - - 4 - base - crucible base fragment, thick grey with red glass, overlain by a think layer of cracked green glass. - opaque/ translucent - - 4 Vollgraff & van Hoorn 1934 rim - crucible rim fragment, thick grey - pink fabric with a thick layer of striped red and green glass. weathered areas and a white glassy material under the rim. - opaque/ translucent - x 4 Vollgraff & van Hoorn 1934 140 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Utrecht-Domplein 35 1933-zn2 ? ? - - production waste crucible Utrecht-Domplein 36 1933-77-84 ? ? - - production waste crucible UtrechtOudwijkerdwarsstraat 77 6-1-170 170 pit - - production waste undiagnostic UtrechtOudwijkerdwarsstraat 78 5-1-135 135 pit - - production waste drop UtrechtOudwijkerdwarsstraat 79 5-1-135 135 pit - - vessel sherd Wijk bij Duurstede (Dorestad) LM 16 - - - - - vessel lamp base Wijk bij Duurstede (Dorestad) LM 17 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) LM18 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) LM 19 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) LM 20 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) LM 21 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) LM 22 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) LM 23 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) LM 24 - - - - - vessel bowl Wijk bij Duurstede (Dorestad) LM 25 - - - - - vessel possible unguentarium Wijk bij Duurstede (Dorestad) LM 26 - - - - - vessel possible unguentarium/ bowl Wijk bij Duurstede (Dorestad) LM 27 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) LM 28 - - - - - vessel bell beaker Wijk bij Duurstede (Dorestad) LM 29 - - - - - vessel bell beaker Wijk bij Duurstede (Dorestad) LM 30 - - - - - vessel rim?beaker Wijk bij Duurstede (Dorestad) LM 31 - - - - - vessel rim?beaker Wijk bij Duurstede (Dorestad) LM 32 - - - - - vessel jar Wijk bij Duurstede (Dorestad) LM 33 - - - - - vessel trail decorated rim 141 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication rim - crucible rim, grey fabric with green glass and weathered surface. - translucent - x 4 Vollgraff & van Hoorn 1934 rim - crucible rim fragment (1 of 4), thin pitted pink-grey ceramic, green and red glass attached. - translucent/ opaque - - 4 - undiagnostic - fragments (crushed?) - translucent - - 3 - - - irregular drop, pale green modern - translucent - - 3 - sherd - small green sherd - translucent - - 3 - base - - green translucent - - 4 - body - - green translucent - - 4 - body - - green translucent - - 4 - base - - green translucent - - 4 - body - - green translucent - - 4 - body - - green translucent - - 4 - body - - green translucent - - 4 - body - - olive green translucent - - 4 - body Isings 1957, type 24? - green translucent - - Roman - body Isings 1957, type 10 - green translucent - - Roman - body Isings 1957, type 10 or 20 - pale blue translucent - - Roman - body - - green translucent - - 4 - body - - green translucent - - 3 - body - - green translucent - - 3 - rim - - green translucent - - 4? - rim - - turquoise translucent - - 4? - body - - green translucent - - 3 - rim - - turquoise translucent - - 4 - 142 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Wijk bij Duurstede (Dorestad) DOR 53 - - - - - vessel gold glass decorated ?beaker Wijk bij Duurstede (Dorestad) DOR 61 - - - - - vessel red trailed beaker Wijk bij Duurstede (Dorestad) DOR 66 - - - - - vessel blue rimmed beaker Wijk bij Duurstede (Dorestad) DOR 90 - - - - - vessel blue rimmed beaker Wijk bij Duurstede (Dorestad) DOR 91 - - - - - vessel sub-sample: body of 90 Wijk bij Duurstede (Dorestad) DOR 96 - - - - - vessel beaker Wijk bij Duurstede (Dorestad) DOR 97 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 98 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 100 - - - - - vessel jar Wijk bij Duurstede (Dorestad) DOR 101 - - - - - vessel palm cup Wijk bij Duurstede (Dorestad) DOR 102 - - - - - vessel palm funnel series Wijk bij Duurstede (Dorestad) DOR 103 - - - - - vessel palm funnel series Wijk bij Duurstede (Dorestad) DOR 104 - - - - - vessel palm cup Wijk bij Duurstede (Dorestad) DOR 105 - - - - - vessel base Wijk bij Duurstede (Dorestad) DOR 106 - - - - - vessel palm cup or funnel Wijk bij Duurstede (Dorestad) DOR 107 - - - - - vessel palm cup Wijk bij Duurstede (Dorestad) DOR 108 - - - - - vessel palm cup Wijk bij Duurstede (Dorestad) DOR 109 - - - - - vessel palm cup Wijk bij Duurstede (Dorestad) DOR 110 - - - - - vessel palm funnel Wijk bij Duurstede (Dorestad) DOR 111 - - - - - vessel palm funnel Wijk bij Duurstede (Dorestad) DOR 112 - - - - - vessel palm funnel Wijk bij Duurstede (Dorestad) DOR 113 - - - - - vessel gold foil decoarted palm funnel Wijk bij Duurstede (Dorestad) DOR 115 - - - - - vessel funnel beaker base Wijk bij Duurstede (Dorestad) DOR 116 - - - - - vessel funnel beaker 143 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication body - - pale green translucent - - 4 - body - - pale green or colourless translucent - - 4 - rim - - pale green translucent - - 4 - rim - - blue translucent - - 4 - body - - pale green translucent - - 4 - body - - pale green translucent - - 4? - body - - pale green translucent - - 4 - body - - pale green translucent - - 4 - body - - pale green translucent - - 3 - body - - pale green translucent - - 3 - body - - mid green translucent - - 4 - body - - yellow-green translucent - - 4 - body - - mid green translucent - - 3 - base - - red and colourless opaque and transparent - - 4 - body - - pale green translucent - - 3 - body - - pale green translucent - - 3 - body - - pale green translucent - - 3 - body - - pale green translucent - - 3 - body - - pale green translucent - - 3 - body - - pale green translucent - - 3 - body - - pale green translucent - - 3 - body - - pale green translucent - - 4? - base - - yellow-green iridescent translucent - - 4 - body - - mid green translucent - - 4 - 144 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Wijk bij Duurstede (Dorestad) DOR 117 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 118 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 119 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 120 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 121 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 122 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 123 - - - - - vessel funnel beaker with bulge Wijk bij Duurstede (Dorestad) DOR 124 - - - - - vessel funnel beaker with bulge Wijk bij Duurstede (Dorestad) DOR 125 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 126 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 127 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 128 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 129a - - - - - vessel vessel with applied blue thread Wijk bij Duurstede (Dorestad) DOR 129b - - - - - vessel sub-sample thread decorating 129a Wijk bij Duurstede (Dorestad) DOR 130a - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 130b - - - - - vessel vessel? Wijk bij Duurstede (Dorestad) DOR 131 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 132 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 133 - - - - - vessel ?funnel beaker Wijk bij Duurstede (Dorestad) DOR 134 - - - - - vessel vessel? Wijk bij Duurstede (Dorestad) DOR 135 - - - - - vessel beaker trail below tim Wijk bij Duurstede (Dorestad) DOR 136 - - - - - vessel funnel beaker base Wijk bij Duurstede (Dorestad) DOR 137 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 138 - - - - - vessel funnel beaker 145 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication body - - acqua translucent - - 4 - body - - pale green translucent - - 4 - body - - pale green translucent - - 4 - body - - pale green translucent - - 4 - body - - pale green translucent - - 4 - body - - pale green translucent - - 4 - body - - pale green translucent - - 4 - body - - pale green translucent - - 4 - body - - colourless transparent - - 4 - body - - pale green translucent - - 4 - body - - colourless transparent - - 4 - body - - colourless transparent - - 4 - rim - - pale green translucent - - 4 - rim - - cobalt blue translucent - - 4 - body - - blue translucent - - 4 - body - - cobalt blue translucent - - 4? - body - - pale green translucent - - 4 - body - - aqua mid green translucent - - 4 - body - - pale green translucent - - 4 - body - - mid green translucent - - 4? - rim - - pale green translucent - - 4 - base - - yellow-green translucent - - 4 - body - - pale green translucent - - 4 - body - - mid green translucent - - 4 - 146 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Wijk bij Duurstede (Dorestad) DOR 139 - - - - - vessel bowl Wijk bij Duurstede (Dorestad) DOR 140 - - - - - vessel funnel beaker Wijk bij Duurstede (Dorestad) DOR 141 - - - - - vessel funnel beaker applied cable Wijk bij Duurstede (Dorestad) DOR 142 - - - - - vessel funnel beaker trail decoarted Wijk bij Duurstede (Dorestad) DOR 143 - - - - - raw chip raw chip Wijk bij Duurstede (Dorestad) DOR 144 - - - - - tessera tessera Wijk bij Duurstede (Dorestad) DOR 145 - - - - - tessera tessera Wijk bij Duurstede (Dorestad) DOR 146 - - - - - tessera tessera Wijk bij Duurstede (Dorestad) DOR 147 - - - - - tessera tessera Wijk bij Duurstede (Dorestad) DOR 148 - - - - - tessera tessera Wijk bij Duurstede (Dorestad) DOR 149 - - - - - rod rod Wijk bij Duurstede (Dorestad) DOR 150 - - - - - linen smoother linen smoother Wijk bij Duurstede (Dorestad) DOR 151 - - - - - linen smoother linen smoother Susteren SUST 1 V12-053-GL-09 S12/067 water course 4310 800 1200 bead annular Susteren SUST 2 V07-216-GL-01 S07/148 cistern 600 900 bead biconical Susteren SUST 3 V09-205-GL-01 S09/200 posthole 700 1000 bead conical Susteren SUST 4 V01-304-GL-01 S01/212 water course 4200 1000 1300 bead conical Susteren SUST 5 V08-190-GL-17 S08/171 water course 4302 700 1000 bead conical Susteren SUST 6 V04-245GL-04 S04/244 water course 4250 1000 1350 bead cylindrical Susteren SUST 7 V09-129-GL-01 S09/100 grave 58 900 1100 window irregular Susteren SUST 8 V04-194-GL-01 S04/199 water course 4400 700 1200 window triangle? Susteren SUST 9 V08-190-GL-10 S08/171 water course 4302 700 1000 window trapezium Susteren SUST 10 V09-273-GL-01 S09/179 grave 67 800 900 window leaf? 147 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication body - - mid green translucent - - 3 or 4 - body - - yellow green translucent - - 4 - rim - - pale green translucent - - 4 - rim - - pale green translucent - - 4 - chip - - cobalt blue translucent - - 4? - whole - - cobalt blue translucent - - 4? - whole - - cobalt blue translucent - - 4? - whole - - turquoise opaque - - 4? - whole - - turquoise opaque - - 4? - whole - - opaque mid green opaque - - 4? - incomplete - - opaque yellow opaque - - 4? - incomplete - - dark green translucent - - 4 - incomplete - - dark green translucent - - 4 - half Koch 1977 Group O - black translucent 3 white zigzags, 4 blue trails - 1 - complete Callmer 1977 B546? - bluish green translucent 3 yellow zigzags, 4 red trails - 4? - complete ? - blue-green translucent yellow feather, 2 red trails x 4? - half ? - bluish green translucent white feather, 2 yellow bands x 4? - half ? - greenish translucent yellow festoons x 4? - fragment ? - greenish translucent white and orange festoons, yellow bands x 4? - complete? - - dark blue translucent - x 4? - complete? - - dark blue translucent - - 4? - almost complete - - dark green translucent - x 4? - almost complete - - colourless transparent - - 4? - 148 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Susteren SUST 11 V04-133-GL-01 S04/162 water course 4400 700 1200 window rectangle? Susteren SUST 12 V08-138-GL-01 S08/155 water course 4302 700 1000 window undiagnostic Susteren SUST 13 V06-158-GL-01 S06/150 water course 4200 700 1300 window semi-circle Susteren SUST 14 V08-214-GL-01 S08/236 water course 4301 700 1000 window undiagnostic Susteren SUST 15 V08-190-GL-12 S08/171 water course 4302 700 1000 window undiagnostic Susteren SUST 16 V09-190-GL-11 S08/171 water course 4302 700 1000 window undiagnostic Susteren SUST 17 V12-053-GL-10 S12/067 water course 4310 800 1200 production waste crucible Susteren SUST 18 V12-053-GL-11 S12/067 water course 4310 800 1200 production waste crucible Susteren SUST 19 V08-190GL-02 S08/171 water course 4302 700 1000 vessel funnel Susteren SUST 20 V08-190-GL-07 S08/171 water course 4302 700 1000 vessel bowl Susteren SUST 21 V08-190-GL-03 S08/171 water course 4302 700 1000 vessel funnel Susteren SUST 22 V05-194-GL-01 S05/219 water course 4400 700 1000 vessel (palm)funnel Susteren SUST 23 V04-232-GL-01 S04/199 water course 4400 700 1200 vessel funnel Susteren SUST 24 V04-166-GL-01 S04/171 water course 4302 700 1000 vessel bowl Susteren SUST 25 V12-053-GL-01 S12/067 water course 4310 800 1200 vessel funnel Susteren SUST 26 V08-219-GL-01 S08/218 pit 800 900/1000 vessel funnel Susteren SUST 27 V12-053-GL-02 S12/053 water course 4310 800 1200 vessel (palm)funnel Susteren SUST 28 V07-148-GL-01 S07/023 pit 900 1000 vessel funnel Susteren SUST 29 V07-148-GL-02 S07/023 pit 900 1000 vessel funnel Deventer DEV 1 434/16203 15050 cesspit 278 850 900 production waste slag, hollow Deventer DEV 2 434/10479 11011 floor level 24 850 900 bead globular Deventer DEV 3 434/99302 10239 cesspit 228 850 900 vessel funnel Deventer DEV 4 434/99721 12507 cesspit 142 850 900 vessel undiagnostic Deventer DEV 5 434/21098 12507 cesspit 142 850 900 vessel undiagnostic Deventer DEV 6 434/99025 290 waste pit 56 850 900 vessel funnel 149 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication almost complete - - bluish green translucent - xx 4? - fragment - - bluish green translucent - - 4? - complete - - bluish green translucent - x 4? - fragment - - bluish green translucent - x 4? - fragment - - dark bluegreen translucent - - 4? - fragment - - dark bluegreen translucent - - 4? - fragment - - dark blue translucent - xx 4? - fragment - - light (bluish) green translucent - - 4? - base - - bluish green translucent - xx 4 - body Lund Feveile 2006, rim type a - almost colourless transparent yellow reticella, spiral x 4 - rim Lund Feveile 2006, rim type e - bluish green transparent - - 4 - base - - bluish green translucent - - 3/4 - base - - light bluegreen translucent - - 4 - body - - dark blue translucent yellow spiral - 4 - rim Lund Feveile 2006, rim type d - yellow-green translucent - - 4 - rim Lund Feveile 2006, rim type g - almost colourless transparent blue incalmo rim - 4 - base - - blue-green transparent - - 3/4 - body - - bluish green transparent white reticella - 4 - rim Lund Feveile 2006, rim type e - light bluish green translucent - - 4 - complete - - grey-green opaque - - 5 - complete - - dark blue translucent - - 5 - body - - bluish green translucent - - 5 - body - - colourless - x 5 - body - - colourless transparent - x 5 - body - - bluish green translucent optic blown ribs - 5 - 150 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Deventer DEV 7 434/99302 10239 cesspit 228 850 900 vessel funnel? Deventer DEV 8 312/29024 21911 wastepit 60 900 925 bead globular Deventer DEV 9 312/29064 22458 cesspit 74 900 925 undiagnostic undiagnostic Deventer DEV 10 312/29089 22638 cesspit 74 900 925 vessel funnel? Deventer DEV 11 312/29089 22638 cesspit 74 900 925 vessel funnel? Deventer DEV 12 312/29090 22638 cesspit 74 900 925 vessel funnel? Deventer DEV 13 434/12394 12765 cesspit 235 900 900 vessel funnel/conical beaker? Deventer DEV 14 312/29028 21911 wastepit 60 900 925 vessel undiagnostic Deventer DEV 15 312/29028 21911 wastepit 60 900 925 production waste raw glass Deventer DEV 16 312/29090 22638 cesspit 74 900 925 window window Deventer DEV 17 312/29057 22457 cesspit 74 900 925 production waste raw glass Deventer DEV 18 312/29028 21911 wastepit 60 900 925 production waste? undiagnostic (trail?) Deventer DEV 19 434/10380 10879 layer 19 900 950 vessel funnel Deventer DEV 20 434/99116 2301 wastepit 107 900 950 vessel cup Deventer DEV 21 434/99144 2538 wastepit 174 900 950 production waste raw glass Deventer DEV 22 434/99289 6623 house 9 900 950 vessel beaker? Deventer DEV 23 434/99289 6623 house 9 900 950 vessel undiagnostic Deventer DEV 24 434/99598 7160 cesspit 130 900 950 window window Deventer DEV 25 434/99578 7120 cesspit 128 925 950 vessel funnel/beaker? Deventer DEV 26 434/99139 2535 cesspit 172 900 950 window window? Deventer DEV 27 434/99154 2583 cesspit 116 900 950 production waste raw glass Deventer DEV 28 434/10638 11039 cesspit 256 890 925 vessel funnel? Deventer DEV 29 434/99154 2583 cesspit 116 900 950 production waste raw glass Deventer DEV 30a 434/7682 7214 cesspit 133 900 950 vessel bottle? Deventer DEV 30b 434/7682 7214 cesspit 133 900 950 vessel bottle? Deventer DEV 31 312/20975 20356 wastepit 30 900 950 window? window Deventer DEV 32 312/22981 20765 wastepit 39 950 1000 window window Deventer DEV 33 312/22985 20765 wastepit 39 950 1000 window window Deventer DEV 34 312/22986 20765 wastepit 39 950 1000 bead? undiagnostic Deventer DEV 35 434/99423 14377 cesspit 327 950 1050 vessel undiagnostic Deventer DEV 36 434/99923 15616 cesspit 286 950 1050 vessel undiagnostic 151 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication rim - - light green translucent - x 5 - complete - - light bluegreen translucent - x 5 - splinter - - yellowbrown translucent - - 5 - rim - - blue-green translucent opaque white spiral - 5 - rim - - blue-green translucent - - 5 - rim - - light bluish green translucent - - 5 - body - - bluish green translucent - - 5 - body/base - - undiagnostic ? - xxx 5 - chip? - - undiagnostic ? - xxx 5 - fragment - - - ? xxx 5 - chip? - - undiagnostic ? - xxx 5 - fragment - - bluish green translucent - - 5 - body - - bluish green translucent optic blown ribs - 5 - rim Isings 1957, type 96a - yellow-green translucent - x 5 - chunk - - bluish green translucent - - 5 - body - - colourless transparent selfcoloured trail x 5 - body, curved - - pale pink translucent - xx crizzled 5 - fragment - - pale bluegreen translucent - xx 5 - body - - colourless transparent 2 white trails - 5 - fragment - - deep turquoise translucent - xx 5 - chip - - light green translucent - xx 5 - body - - yellowish green translucent - xx 5 - chip - - bluish green translucent - xx 5 - neck? - - pale green translucent - xx 5 - neck? - - red purple streak translucent - xx 5 - fragment - - blue-green translucent - - 5 - fragment - - green? ? - xx 5 - fragment - - green? ? - xxx 5 - fragment - - amber translucent - - 5 - rim, thick - - greenish translucent - xxx 5 - body - - greenish translucent - xx 5 - 152 — Appendix I sample list Site Sample Find number Feature number Feature type Begin date feature (AD) End date feature (AD) Category object Form/object Deventer DEV 37 434/12362 12915 house 12 950 1050 vessel funnel? Deventer DEV 38 434/99923 15616 cesspit 286 950 1050 window window Deventer DEV 39 312/20678 20306 wastepit 34 950 1050 window window Deventer DEV 40a 312/29048 22270 cesspit 80 950 1050 production waste? melted, 2 layers Deventer DEV 40b 312/29048 22270 cesspit 80 950 1050 production waste? melted, 2 layers Deventer DEV 41 312/29063 22425 layer 8 950 1050 window? window/inlay? 153 — Key weathering Key archaeological periods x=slightly weathered xx = moderately weathered xxx= badly weathered 1= 450-550 AD 2= 550-650 AD 3= 650-750 AD 4= 750-850 AD 5= 850-1000 AD Type of fragment Typology (Isings, Koch, Callmer, Ribe or Pion) Description object Colour Transparancy Decoration Weathering Archeological period Publication body - - light bluegreen translucent - x 5 - rectangle? - - bluish green translucent - x 5 - fragment - - greenish translucent - xx 5 - chip - - red opaque - - 5 - chip - - turquoise translucent - - 5 - fragment - - turquoise translucent - - 5 - 154 — Appendix II Element oxide Na2O major and minor chemical compositions of samples analysed by electron probe microanalysis MgO Al2O3 SiO2 P2O5 SO3 Cl K2O Gennep glass samples GE 41 13.69 0.86 2.68 67.89 0.08 0.33 - 0.76 GE 42 15.23 0.85 2.97 68.84 - 0.4 - 0.54 GE 43 15.18 0.77 2.94 69.55 0.03 0.37 - 0.43 GE 44 15.38 1 3.37 71.06 - 0.32 - 0.44 GE 45 14.74 0.98 3.1 68.21 0.05 0.31 - 0.55 GE 46 14.98 0.78 2.87 68.99 - 0.35 - 0.44 GE 47 15.91 0.62 2.7 74.73 - 0.33 - 0.69 GE 48 13.63 0.77 2.91 71.17 0.05 0.37 - 0.72 GE 49 14.34 0.85 2.91 70.65 - 0.36 - 0.76 GE 51 13.93 0.94 2.94 70.37 0.02 0.34 - 0.72 GE 52 14.44 0.89 2.4 70.83 0.01 0.33 - 0.56 GE 53 15.1 0.73 2.7 71.4 - 0.4 - 0.68 GE 54 15.85 1.11 2.81 67.76 0.01 0.42 - 0.61 GE 55 16.15 1.11 2.9 67.59 0.05 0.42 - 0.6 GE 56 16.36 1.22 2.8 70.4 - 0.31 - 0.41 GE57 14.04 0.92 2.84 68.36 - 0.41 - 0.86 GE 58 13.97 1 3.01 68.04 0.03 0.42 - 0.8 GE 59 15.86 0.99 2.77 68.66 0.01 0.44 - 0.68 GE 60 16.12 1.19 2.93 66.78 0.04 0.46 - 0.69 GE 61 13.78 0.69 2.74 72.39 0.01 0.26 - 0.72 GE 62 14.39 0.78 2.88 70.89 0.06 0.29 - 0.71 GE 63 13.6 0.68 2.85 72.72 0.01 0.23 - 0.73 GE 64 14.15 0.76 2.96 70.94 0.04 0.26 - 0.81 GE 65 15.05 0.74 2.86 71.65 - 0.3 - 0.61 GE 66 14.19 0.82 2.99 72.26 0.01 0.26 - 0.57 GE 67 14.57 0.83 2.85 72.54 0.02 0.27 - 0.61 GE 68 12.36 0.69 2.93 73.37 - 0.25 - 0.74 GE 69 14.95 1.04 3.02 71.02 - 0.25 - 0.42 11.77 1.15 2.43 53.26 0.13 0.31 0.63 0.62 Joden 2 12.43 1.01 2.44 60.89 0.15 0.24 0.64 0.72 Joden 3 16.42 1.21 2.56 67.05 0.09 0.35 0.79 0.57 Joden 4 14.1 1.13 2.67 64.19 0.15 0.33 0.78 0.44 Maastricht-Jodenstraat (MAJO) glass samples Joden 1 Joden 5 14.49 1.24 2.79 62.07 0.18 0.27 0.8 0.47 Joden 6 10.49 0.92 2.23 58.09 0.19 0.23 0.67 0.48 Joden 37 16.04 0.72 2.25 69.97 0.05 0.34 0.96 0.55 Joden 38 17.03 0.88 2.35 70.36 0.1 0.39 0.84 0.44 Joden 39 16.54 0.76 2.35 70.14 0.05 0.31 0.95 0.46 Joden 40 16.87 0.73 2.26 71.34 0.02 0.3 1 0.3 Joden 41 15.49 1.14 2.6 62.42 0.18 0.22 0.7 0.84 Joden 42 14.15 0.72 2.38 61.92 0.1 0.23 0.84 0.66 Joden 43 16.14 1.16 2.58 61.94 0.2 0.25 0.66 0.87 155 — CaO TiO2 MnO FeO CoO Sb2O5 SnO2 CuO PbO Total 7.42 0.12 0.84 1.49 0.01 0.66 - 0.53 2.99 100.41 7.43 0.32 1.53 2.14 - 0.19 0.02 0.13 0.35 101.04 6.94 0.31 2.26 2.04 0.03 0.02 0.03 - 0.02 100.97 6.44 0.32 1.59 1.99 0.04 0.04 - 0.03 0.06 102.16 7.75 0.31 2.4 1.99 - 0.06 - 0 0.11 100.6 6.76 0.31 2.25 1.61 - 0.03 - - 0.02 99.47 7.6 0.11 0.93 0.6 - 0.35 0.02 0.03 0.04 104.74 8.2 0.16 0.53 0.92 0.04 0.18 ‘- 0.03 0.21 99.9 8.58 0.2 1.03 0.95 - 0.15 0 0.12 0.38 101.37 7.33 0.19 1.06 1.26 - 0.54 0.01 0.41 1.37 101.48 7.07 0.16 0.86 1.14 0.01 0.36 - 0.29 1.51 100.96 8.18 0.14 1.52 0.69 0.02 0.08 0.04 0.06 0.09 101.9 8.84 0.14 1.49 0.68 - 0.08 - - - 99.88 8.9 0.18 1.51 0.78 - 0.09 - - 0.02 100.32 6.98 0.25 1.78 0.86 0 0.01 0.01 - 0 101.44 9.2 0.15 1.4 0.92 - 0.08 - - 0.08 99.35 9.23 0.18 1.34 1.09 0.03 0.08 - - 0.02 99.28 9.06 0.15 1.3 0.85 - 0.07 - 0.04 0.01 100.96 9.46 0.15 1.54 0.86 0 0.08 - - 0.03 100.37 7.79 0.09 0.66 0.73 - 0.27 - 0.02 0.06 100.28 7.67 0.15 0.98 0.78 - 0.19 0.01 0.02 0.14 99.96 7.6 0.11 0.9 0.6 - 0.24 - - 0.09 100.43 7.84 0.13 0.99 0.84 0.02 0.14 0.01 0.04 0.2 100.17 7.05 0.16 1.09 0.8 0.02 0.23 - 0.08 0.39 101.1 7.67 0.15 0.95 0.8 0.02 0.15 0.01 0.01 0.09 101.01 7.44 0.17 1.11 0.75 - 0.14 - 0.06 0.1 101.5 8.28 0.1 0.74 0.64 - 0.15 0 0.02 0.07 100.45 6.32 0.42 1.96 1.23 - 0.02 0.01 0.03 0.04 100.78 5.11 0.11 1.21 2.66 0.01 - 0.65 0.45 7.13 87.61 6.53 0.12 1.18 2.51 0.01 - 0.63 0.58 6.13 96.22 7.07 0.12 1.5 0.79 - - 0.53 - 0.37 99.42 6.41 0.16 1.45 1.24 - - 3.08 0.2 4.8 101.12 6.26 0.19 1.63 1.39 - - 2.86 - 4.69 99.32 5.25 0.09 1.36 0.71 - - 1.72 - 19.16 101.58 5.55 0.1 0.05 0.86 0.01 - 0.18 0.02 0.67 98.31 6.18 0.12 0.13 0.57 0.01 - - 0.02 0.27 99.69 5.62 0.11 0.04 0.56 - 0.07 - - 0.3 98.27 5.42 0.11 0.06 0.68 0.02 - - - 0.53 99.64 6.9 0.14 1.27 1.3 - 0.04 0.14 1.54 3.13 98.03 6.28 0.1 0.69 1.55 - 0.2 0.17 1.7 7.32 99 6.74 0.14 1.45 0.86 0.01 - 0.59 2.77 0.71 97.08 156 — Appendix II major and minor chemical compositions of samples analysed by electron probe microanalysis Element oxide Na2O MgO Joden 44 19.39 Joden 45 16.58 Joden 46 17.46 Joden 47 11.07 Joden 48 15.82 Joden 49 Joden 50 Joden 51 Al2O3 1.45 SiO2 P2O5 SO3 Cl K2O 2.62 63.2 0.18 0.28 0.88 0.78 0.95 2.65 0.99 2.68 66.86 0.22 0.22 0.87 0.65 65.6 0.05 0.21 0.97 0.43 0.83 2.17 47.47 1.1 2.68 63.17 0.17 0.17 0.44 0.54 0.14 0.24 0.78 0.88 16.84 1.22 2.65 64.27 0.17 0.31 0.83 0.97 12.05 1.18 2.25 56.27 0.13 0.3 0.67 7.99 12.65 1.17 2.38 54.17 0.16 0.28 0.67 0.64 Joden 52 16 1.18 2.37 58.7 0.19 0.36 0.66 0.79 Joden 53 14.31 0.72 2.57 70.63 0.07 0.2 0.88 0.4 Joden 54 12.14 1.08 2.52 56.33 0.2 0.29 0.71 0.69 Joden 55 9.66 0.74 2 38.02 0.14 0.17 0.37 1.69 Joden 56 10.53 0.96 2.09 44.95 0.12 0.21 0.56 0.46 Joden 57 13.1 1 2.39 60.96 0.17 0.26 0.75 0.64 Joden 58 14.88 1.2 2.45 60.2 0.19 0.34 0.78 0.49 Joden 59 15.05 1.11 2.53 61.74 0.18 0.29 0.78 0.72 Joden 60 16.81 0.8 2.72 67.87 0.21 0.14 0.77 0.84 Joden 61 15.49 0.72 2.38 69.34 0.14 0.26 0.92 0.58 Joden 62 16.14 1.29 2.64 67.06 0.13 0.31 0.73 0.93 Joden 63 16.76 1.55 2.67 64.02 0.18 0.34 0.76 1.77 Joden 64 9.8 0.72 1.89 39.52 0.06 0.15 0.46 0.32 Joden 65 9.96 0.86 1.91 37.42 - 0.09 0.42 0.41 Joden 66 11.7 1.02 2.49 55.59 0.17 0.25 0.64 0.74 Joden 67 14.48 1.04 2.42 61.18 0.15 0.23 0.69 0.78 Joden 68 15.95 0.52 2.53 69.67 0.14 0.17 1.04 0.58 Joden 69 10.01 0.83 2.17 47.3 0.2 0.22 0.49 0.37 Joden 70 11.24 1.06 2.36 56.05 0.16 0.25 0.61 0.53 Joden 71 10.41 0.9 2.29 46.31 0.19 0.19 0.57 0.58 Joden 72 10.8 0.88 2.18 50.67 0.15 0.24 0.69 0.5 Joden 73 14.28 1.28 2.38 60.58 0.14 0.31 0.76 0.74 Joden 74 14.8 1.26 2.4 59.42 0.28 0.34 0.77 0.74 Joden 75 13.08 0.9 2.38 58.4 0.15 0.23 0.67 0.66 Joden 76 14.26 1.3 2.44 58.18 0.2 0.31 0.69 0.73 Maastricht-Jodenstraat (MAJO) crucibles Joden 19 (lead glass) 0.31 0.35 4.07 24.57 0.1 0.69 - 1.01 Joden19 (yellow residue) 0.62 0.54 8.15 27.46 0.14 0 0.01 1.62 Joden 20 (lead glass) 0.41 0.68 2.53 22.74 0.03 0.07 0.1 1.03 Joden 20 (yellow residue) 0.406 0.675 2.533 22.737 0.1 - 0.029 1.034 Joden 21 (natron glass) 12.29 0.91 10.56 69.24 0.13 0.07 0.18 2.49 Joden 22 (lead glass) 0.26 0.59 4.39 26.65 0.21 0.17 0.04 0.85 Joden 23 (white melt) 0.11 0.39 2.47 12.45 0.04 0.04 - - 157 — CaO TiO2 7.61 MnO FeO CoO Sb2O5 SnO2 0.16 1.9 0.66 6.91 0.14 1.02 5.85 0.34 1.67 5.14 0.15 0.98 0.96 0.01 - 6.29 0.16 1.35 0.79 0.02 0.03 7.43 0.15 1.55 0.84 0.07 0.1 4.11 0.15 1.31 0.99 - - 5.02 0.11 1.22 2.19 0.05 - CuO PbO Total - 0 - - 0.05 99.16 0.95 - 0.08 - 0.08 0.02 98.21 1.08 0.03 0.04 - 0.12 - 97.52 3.71 - 23.75 97.55 0.31 1.84 3.14 98.72 - 0.08 0.04 97.54 1.16 0.12 2.48 91.15 0.81 0.75 8.63 90.9 6.59 0.12 1.36 0.78 0.05 - 0.17 4.16 0.27 93.74 4.86 0.14 0.05 0.74 0.04 0.03 0.05 0.17 0.62 96.48 5.74 0.13 1.24 3.94 - - 1.5 1.1 7.69 95.31 3.16 0.1 0.87 1.96 0.06 - 1.02 0.36 36.54 96.86 5.5 0.11 1.13 0.9 0.02 - 3.13 0.07 26.27 96.99 7.42 0.14 1.22 0.72 0.05 - 0.24 2.64 3.71 95.42 5.96 0.12 1.59 0.7 0.02 - 5.68 0.07 2.21 96.89 6.53 0.13 1.56 0.8 0.07 - 2.12 0.08 3.13 96.81 6.63 0.15 0.68 0.8 0.01 0.21 - 0.11 0.35 99.11 5.89 0.09 0.41 0.74 0.04 1.72 - 0.15 0.33 99.19 7.53 0.16 1.76 0.73 0.02 0.06 - 0 - 99.48 7.96 0.16 1.8 0.78 - 0.09 - - - 98.85 3.05 0.09 0.87 0.7 - - 3.04 0 42.07 102.75 2.82 0.09 0.59 0.56 0.01 - 1.71 - 27.16 84 5.69 0.14 1.44 2.7 - - 1.57 1.42 10.78 96.35 6.65 0.17 1.29 2.55 - 0.1 0.13 2.33 3.27 97.44 6.92 0.07 0.46 0.42 0.03 0.12 - 0.08 0.08 98.8 4.27 0.09 0.86 1.07 0.01 - 3.8 0.01 24.46 96.15 5.58 0.14 1.1 0.89 - - 1.77 0.01 17.23 98.98 4.98 0.08 1.01 1.25 0.07 - 2.81 0.09 26.72 98.44 4.95 0.12 1.37 0.7 - - 1.46 0.04 24.48 99.22 8.23 0.13 1.58 0.84 0.06 - 0.28 2.57 3.96 98.13 7.23 0.14 1.64 0.89 0.08 - 1.8 2.89 4.2 98.88 8.83 0.11 1.09 0.84 0.03 - 0.67 2.59 5.86 96.48 6.88 0.12 1.47 0.82 - 0.03 0.39 2.47 3.9 94.19 0.68 0.26 0.01 1.55 - - 0.49 0.06 61.53 95.66 3.84 0.51 0.04 2.43 0.03 - 9.48 - 42.98 97.84 0.86 0.27 0.05 2.48 - - 0.54 - 62.48 94.26 0.86 0.272 0.051 0.483 0.028 - 9.536 - 62.477 101.22 2.17 0.48 0.33 3.17 0.02 0.03 - - - 102.07 1.64 0.24 0.05 1.5 0 0.06 1 0.1 56 93.74 0.54 0.11 0.01 0.87 - - 61.58 0.01 22.83 101.45 158 — Appendix II major and minor chemical compositions of samples analysed by electron probe microanalysis Element oxide Na2O MgO Al2O3 SiO2 P2O5 Joden 25 (lead glass) 0.15 0.67 6.24 28.15 Joden 27 (lead glass) 0.28 0.44 3.44 Joden 27 (yellow residue) 0.05 0.05 1.05 Joden 28 (lead glass) 0.31 0.34 Joden 28 (yellow residue) 0.09 0.15 Joden 29 (natron glass) 17.08 1.1 3.24 0.49 0.36 2.18 Joden 30 (white melt) SO3 Cl 0.03 - 35.43 0.17 11.12 0.25 4.26 25.04 1.71 15.39 K2O 0.05 0.84 0.01 0.03 1.38 0.01 0.04 - 0.07 0.7 0.05 0.95 0.03 - 0.05 - 67.71 0.16 0.24 0.65 1.02 19.78 0.02 - 0.01 - Maastricht-Mabro crucibles Mabro 7 9.87 1.33 7.24 65.88 0.38 0.06 0.04 7.17 Mabro 8 15.79 0.72 9.87 69.38 0.08 0.13 0.38 0.85 Mabro 9 3.49 0.98 7.54 67.21 0.18 0.03 - 13 Mabro 10 12.43 1.49 2.74 65.69 0.24 0.11 0.01 2.99 Mabro 11 11.2 0.82 12.48 62.9 0.16 0.05 0.01 2.99 Mabro 12 0.02 0.25 0.65 15.64 0.06 - 0.06 0.84 Mabro 13 13.46 1.21 3.57 69.98 0.21 0.18 0.52 2.36 Mabro 14 0.49 0.59 6.27 28.43 0.09 - 0.05 0.86 Mabro 15 11.64 1.65 2.76 69.88 0.27 0.12 0.05 2.62 Mabro 16 16.67 0.99 3.48 68.99 0.1 0.19 0.38 1.27 WIJ1 16.97 0.92 2.57 69.16 0.1 0.26 0.79 0.73 WIJ2 17.69 0.95 2.62 68.26 0.1 0.23 0.92 0.74 WIJ3 16.44 0.94 2.69 68.43 0.13 0.28 0.78 0.85 WIJ4 10.24 0.23 1.03 40.69 0.01 0.13 0.72 0.19 WIJ5 10.86 0.89 2.39 37.21 0.11 0.15 0.41 0.57 WIJ6 10.76 0.73 2.18 39.23 0.11 0.14 0.48 0.45 WIJ 7 9.29 1.21 1.86 38.89 0.24 0.15 0.22 0.95 WIJ 8 8.72 0.38 2.35 40.37 0.09 0.01 0.48 0.4 Wijnaldum glass samples WIJ 9 6.96 0.3 2.24 35.31 0.04 0.01 0.42 0.31 WIJ 10 14.93 0.71 2.47 68.16 0.11 0.12 1 0.61 WIJ 11 16.48 0.95 2.57 69.18 0.16 0.21 0.73 1.14 WIJ 12 15.99 0.88 2.65 70.57 0.11 0.22 0.99 0.77 WIJ 13 16.46 0.74 2.55 71.35 0.09 0.18 0.98 0.62 WIJ 14 7.05 0.46 1.97 26.42 0.07 0.11 0.37 0.39 WIJ 15 14.65 0.64 2.57 69.7 0.14 0.1 1 0.41 WIJ 16 low lead 15.72 0.84 2.79 69.24 0.16 0.19 0.83 0.91 WIJ 16 high lead 10.98 0.48 1.9 44.92 0.09 0.12 0.52 0.51 WIJ 17 6.58 0.58 4.13 48.65 0.34 0.03 0.27 1.23 WIJ 18 13.37 1.07 2.79 59.73 0.2 0.22 0.76 0.91 WIJ 19 13.5 1.28 2.78 58.87 0.21 0.3 0.88 2.69 WIJ 20 14.87 1.45 2.76 61 0.29 0.26 0.74 1.11 WIJ 21 15.01 1.42 2.74 64.76 0.23 0.29 0.92 0.79 159 — CaO TiO2 0.34 MnO 0.23 FeO 0 CoO Sb2O5 SnO2 2.2 0.02 - CuO PbO 0.47 0.12 Total 55.03 94.55 1.36 0.21 0.03 1.64 0.02 - 0.44 0.09 51.62 96.58 0.3 0.25 0.02 0.52 0 - 23.03 0.04 65.34 102.06 0.77 0.24 - 1.61 - - 0.75 - 60.24 95.33 0.25 0.16 0.03 0.46 0.01 - 17.85 0.09 59.97 96.24 7.23 0.18 1.34 0.91 0.01 0.05 - 0.21 0.1 101.22 0.76 0.12 0.02 0.78 0.01 - 55.81 0.49 22.5 103.33 4.95 0.33 0.55 2.65 0 0.2 - - - 100.66 2.44 0.38 0.09 2.22 0.01 0.02 - 0.08 0.05 102.48 - 0.52 0.1 2.71 0.15 0.56 - 0.01 - 96.46 7.58 0.18 0.38 1.17 0.01 0.17 - 0.11 - 95.29 2.52 1.36 0.7 4.79 - 0.11 - 0.02 0.09 100.19 0.47 0.29 0.09 1.16 0.01 - 12.75 0.27 59.24 91.8 6.74 0.23 0.39 0.99 0 0.15 - 0.36 0.08 100.43 0.66 0.31 0.06 2.48 0.01 - 1.1 0.1 53.77 95.26 7.52 0.14 0.43 1.16 - 0.11 - 0.03 0.05 98.42 5.82 0.25 1.05 1.17 - 0.04 - - 0.05 100.45 7.15 0.15 1.04 0.85 0.02 0.13 0.01 0.1 0.15 101.1 7.07 0.16 1.18 0.9 0.02 0.14 0.03 0.05 0.17 101.22 7.79 0.18 1.07 0.99 0.01 0.1 0.06 0.14 0.18 101.05 2.76 0.04 0.01 0.3 0 - 4.85 - 40.26 101.47 3.41 0.11 0.33 1.32 0.01 - 3.69 0.05 38.82 100.32 3.39 0.12 0.33 0.91 0 - 5.12 0.06 36.76 100.77 5.72 0.15 1.1 1.27 0.03 - 3.11 - 36.64 100.83 3.76 0.08 0.01 0.53 0.02 - 5.11 0.09 36.84 99.23 2.76 0.06 0.02 0.43 0.03 - 6.18 0.06 44.35 99.47 9 0.27 0.2 0.82 0.02 0.07 - 0.13 0.02 98.64 6.93 0.17 0.83 0.77 0.02 0.31 - 0.14 0.17 100.76 6.6 0.18 0.98 0.79 0.01 0.28 - 0.1 0.23 101.34 7.02 0.11 0.93 0.64 0 0.39 - 0 0.1 102.18 1.81 0.1 0.04 0.83 0.01 - 4.86 0.15 56.58 101.23 9.48 0.29 0.23 1.21 0.13 0.02 0 0.11 0.11 100.81 7.38 0.1 0.54 1.05 0.02 0.35 0 0.38 0.37 100.86 4.29 0.07 0.41 0.75 0.02 - 2.38 0.15 33.93 101.53 2.69 0.29 1.72 7.46 0 - 0.14 1.81 23.21 99.12 5.99 0.16 1.12 4.36 0.03 - 0.79 1.99 6.62 100.11 5.49 0.17 0.27 3.84 0.04 - 1 1.02 7.61 99.96 7.23 0.18 1.25 5.63 0.02 - 0.7 0.57 2.89 100.93 6.55 0.14 1.29 2.7 0.01 - 0.98 - 1.49 99.32 160 — Appendix II major and minor chemical compositions of samples analysed by electron probe microanalysis Element oxide Na2O MgO WIJ 22 14.07 WIJ 23 16.18 WIJ 24 11.23 WIJ 25 15.05 WIJ 26 17.07 WIJ 27 13.38 WIJ 28 14.09 WIJ 29 13.16 Al2O3 0.69 SiO2 P2O5 SO3 Cl K2O 2.5 67.49 0.13 0.18 0.95 0.51 1.57 2.49 68.22 0.25 0.27 0.81 1.03 0.98 2.49 51.57 0.15 0.2 0.7 0.56 1.25 2.39 62.26 0.22 0.27 0.79 1.5 1.25 2.67 63.85 0.17 0.38 0.83 0.78 1.59 3.7 62.36 0.2 0.26 0.77 1.01 1.32 2.97 59.2 0.26 0.33 0.76 0.95 2.14 2.12 52.46 0.77 0.47 0.63 1.71 WIJ 30 14.47 1.33 2.29 62.44 0.22 0.21 0.78 0.88 WIJ 31 15.09 1.43 3.59 62.15 0.18 0.28 0.81 0.97 WIJ 32 11.1 0.64 2.08 45.68 0.1 0.21 0.66 0.51 WIJ 33 9.9 0.55 1.89 33.09 0.09 0.14 0.41 0.38 WIJ 34 0.24 0.19 6.26 35.54 0.5 0.01 0.01 1.66 WIJ 35 12.67 5.45 1.38 70.23 0.11 0.07 0.57 2.42 WIJ 36 17.07 1.31 2.72 65.1 0.17 0.28 0.91 0.71 WIJ 37 12.18 5.43 1.19 70.77 0.09 0.23 0.64 2.33 WIJ 38 16.01 1.11 2.65 69.88 0.07 0.31 0.55 0.62 WIJ 39 low lead 16.76 1.24 2.79 66.28 0.2 0.29 0.53 1.05 WIJ 39 high lead 11.07 1.2 2.78 58.87 0.4 0.22 0.54 1.19 WIJ 40 11.19 0.52 2.19 63.51 0.06 0.23 0.88 0.47 WIJ 41 16.97 1 2.82 67.28 0.15 0.23 0.93 0.68 WIJ 42 15.21 0.86 3.15 66.88 0.06 0.13 0.92 0.39 LM 25 16.85 0.66 2.55 70.62 0.1 0.2 0.99 0.75 LM 26 16.87 0.72 2.77 66.61 0.17 0.27 0.86 0.74 LM 27 15.02 0.67 2.34 65.33 0.14 0.23 0.61 1.04 LM 28 16.71 0.71 2.49 67.94 0.08 0.26 0.91 0.73 LM 29 15.44 0.8 2.42 68.03 0.14 0.24 0.87 0.93 LM 30 15.37 1.15 2.37 68.57 0.05 1.14 - 1.4 Wijk bij Duurstede (Dorestad) glass samples LM 31 16.21 0.79 2.39 68.51 0.06 1.12 - 1.14 LM 33 15.54 0.92 2.75 67.2 0.02 1.17 0.01 1.03 DOR 53 17.01 0.73 2.63 69.39 0.13 0.18 0.85 0.89 DOR 61 16.44 0.79 2.85 68.86 0.16 0.17 0.79 0.87 DOR 66 17.72 0.67 2.34 71.22 0.03 1.23 - 0.54 DOR 90 15.95 0.77 2.61 68.89 0.01 1.19 0.04 1.1 DOR 91 18.09 0.61 2.48 70.9 0.02 0.28 - 0 DOR 95 1.3 1.53 12.35 64.2 0.25 0.01 - 4.32 DOR 100 15.62 0.64 2.9 70.82 0.03 1.12 0.01 1.09 DOR 101 16.4 0.69 2.66 71.25 0.07 1.25 0.01 0.83 DOR 102 15.7 0.66 2.61 71.92 - 1.25 - 0.91 DOR 103 1.48 7.14 2.37 61.68 0.01 0.99 0.01 8.56 161 — CaO TiO2 MnO FeO CoO Sb2O5 SnO2 - CuO PbO Total 0.58 0.08 0.56 0.85 0.01 3.44 0.02 3.19 95.25 0.32 0.18 0.63 0.97 0.02 0.3 0.14 1.29 1.43 0.02 - 1.7 0.02 1.06 95.7 - 1.56 - 23.37 95.99 6.18 0.16 1.24 2.04 - - 1.51 - 5.12 99.99 6.42 0.1 1.27 1.78 0.01 - 0.54 0.14 3.62 100.87 6.05 0.17 0.55 3.81 0.02 - 0.11 1.57 4.22 99.77 6.38 0.14 0.61 4.09 0.01 - 0.27 2.45 5.64 99.46 8.44 0.17 0.22 3.32 0.01 1.57 0.63 11.39 1.75 100.95 5.85 0.21 0.89 3.9 0.01 0.03 0.03 2.22 3.9 99.66 6.09 0.21 0.51 3.62 - 0.01 - 1.12 4.3 100.36 3.61 0.07 0.21 0.86 - - 2.41 - 32.37 100.51 2.22 0.13 0.09 0.83 - - 3.26 0.28 46.8 100.04 0.22 0.41 4.24 9.09 0.04 0.01 0.17 0.25 40.15 98.97 6.64 0.1 0.64 0.6 0.01 0.1 - - 0.05 101.05 9.19 0.16 1.85 1.29 0.01 0.05 - 0.11 - 100.93 6.21 0.02 0.64 0.42 - 0.08 - 0.01 0.02 100.25 7.97 0.14 0.49 0.86 0.02 0.04 - 0.02 - 100.72 8.65 0.11 1.09 0.94 0.04 0.04 - 0.16 0.29 100.46 6.1 0.22 1.1 1.56 0.02 - 3.25 1.85 8.29 98.67 3.86 0.11 0.52 1.81 - 1.33 - - 13.27 99.94 6.65 0.29 1.13 1.35 0 0.26 - 0.06 0.36 100.16 8.89 0.08 0.03 0.45 0.03 0.01 0.01 2.26 1.61 100.96 6.61 0.12 0.66 0.59 - 0.48 - - - 101.18 8.66 0.13 0.59 0.73 0.02 0.29 - 0.34 0.32 100.08 7.03 0.14 - 0.01 0 0.29 0.02 0 0.39 93.26 6.64 0.11 - - 0.01 0.32 0.01 - 0.27 97.17 7.81 0.15 0.01 0.01 0 0.16 - - 0.26 97.26 7.44 0.11 - 0.73 0.05 0.64 - 0.11 - 99.11 6.54 0.11 0.06 0.54 0.01 0.21 - 1.25 - 98.93 7.07 0.21 - 0.87 0.01 0.17 - 2.24 - 99.2 7.25 0.15 0.67 0.67 - 0.41 - 0.07 0.24 101.26 7.46 0.15 0.61 0.84 - 0.33 0.02 0.16 0.69 101.19 5.94 0.1 0.02 0.51 - 0.52 - 0.04 1.26 102.13 7.28 0.15 - 0.79 0.02 0.22 0 0.37 - 99.41 5.99 0.12 - 0.6 0.03 0.58 0.03 0.02 1.19 100.93 2.78 0.55 0.06 11.94 0.02 0.2 - 0.1 0.02 99.65 7.2 0.14 - 0.89 0.04 0.3 - 0.16 0.31 101.27 6.65 0.12 - 0.85 0.03 0.62 - 0.33 0.49 102.26 6.82 0.13 - 0.91 - 0.37 - 0.11 0.03 101.42 13.65 0.18 - 0.71 0.04 0.33 - - - 97.16 162 — Appendix II major and minor chemical compositions of samples analysed by electron probe microanalysis Element oxide Na2O MgO DOR 104 15.86 DOR 105 colourless DOR 105 op red Al2O3 SiO2 P2O5 0.71 2.86 71.87 15.94 1.11 2.66 68.89 15 0.64 2.73 68.24 SO3 - Cl K2O 1.21 0.01 1.01 0.03 1.19 0.01 1.11 0.06 1.18 - 0.91 DOR 106 16.28 0.62 2.82 71.63 0.01 1.26 0.02 0.81 DOR 107 16 0.82 2.8 71.4 - 1.19 0.01 0.95 DOR 108 16.47 0.73 2.69 72.34 - 1.14 - 0.7 DOR 109 15.92 0.75 2.8 72.3 - 1.14 0.02 0.91 DOR 110 16.56 0.81 2.78 68.65 0.15 0.21 0.85 1.02 DOR 111 14.73 1.11 3.07 65.81 0.22 0.19 0.64 1.08 DOR 112 16.35 0.82 2.63 69.85 0.15 0.2 0.93 1.02 DOR 113 16.25 0.86 2.7 69.08 0.13 0.27 0.88 0.93 DOR 115 16.57 1.33 2.21 66.95 0.21 0.28 0.81 1.43 DOR 116 16.55 0.84 2.71 67.86 0.13 0.28 0.89 0.75 DOR 117 14.8 0.98 2.77 69.81 0.19 0.21 0.82 1.08 DOR 118 16.19 0.81 2.85 69.74 0.15 0.17 0.86 0.94 DOR 119 15.78 0.81 2.75 68.22 0.14 0.18 0.69 0.85 DOR 120 16.51 0.85 2.6 68 0.17 0.19 0.84 1.11 DOR 121 17.27 0.86 2.63 69.12 0.1 0.24 0.97 0.78 DOR 122 15.13 1.01 2.3 67.69 0.11 0.23 1 1 DOR 123a 16.92 0.84 2.91 67.41 0.09 0.22 1 0.61 DOR 123b 17.28 0.85 2.69 68.44 0.08 0.21 0.97 0.69 DOR 124 15.07 0.71 2.98 68.85 0.22 0.15 0.72 0.86 DOR 125 17.6 0.77 2.38 71.25 0.09 0.2 1.09 0.49 DOR 126 16.11 0.7 2.48 70.13 0.08 0.24 1.15 0.57 DOR 127 17.21 0.7 2.67 69.84 0.05 0.16 0.99 0.62 DOR 128 17.17 0.69 2.51 70.15 0.08 0.17 1.1 0.6 DOR 129 16.51 0.87 2.83 66.66 0.15 0.17 0.86 0.97 DOR 130a 14.07 0.94 2.81 69.25 0.19 0.19 0.72 0.93 DOR 130b 16.2 0.81 3.44 69.36 0.19 0.19 0.49 1 DOR 131 16.97 0.72 2.77 69.13 0.11 0.19 0.72 0.62 DOR 132 16.29 0.87 2.94 67.68 0.24 0.17 0.71 1.31 DOR 133 15.4 0.74 2.8 68.69 0.19 0.24 0.69 0.98 DOR 134 15.33 0.81 2.63 68.6 0.1 0.23 1.04 0.69 DOR 135 16.54 1.02 2.94 69.59 0.25 0.2 0.35 1.14 DOR 136 1.21 8.09 1.07 63.76 2.06 0.08 0.03 7.51 DOR 137 15.96 0.8 2.74 68.3 0.19 0.21 0.78 0.95 DOR 138 14.71 0.86 2.91 69.39 0.25 0.19 0.78 0.87 DOR 139 15.84 0.72 2.82 69.19 0.16 0.17 0.78 1.02 DOR 141 15.74 0.87 2.8 68.08 0.22 0.19 0.81 1.1 DOR 142 15.95 0.85 2.77 68.54 0.15 0.21 0.66 0.85 DOR 143 17.47 0.9 2.7 67.55 0.1 0.28 1.06 0.54 163 — CaO TiO2 MnO FeO CoO 7 0.12 - 0.67 7.29 0.16 - 6.8 0.14 - Sb2O5 SnO2 CuO PbO 0.01 0.29 - 0.02 0.72 - 0.28 0.01 2.14 0.01 0.2 - Total 0.27 101.9 0.15 - 99.54 1.11 0.38 99.53 6.93 0.12 - 0.84 0.03 0.12 0 0.09 0.63 102.22 6.81 0.14 0.02 0.84 - 0.39 - 0.2 0.77 102.34 6.97 0.14 - 0.66 0.02 0.27 - 0.06 1.09 103.28 7.16 0.12 - 0.8 0.02 0.22 0.01 0.05 - 102.22 7.25 0.11 0.56 0.83 0 0.25 - 0.06 0.41 100.48 7.03 0.2 0.63 1.86 0.02 0.08 0.25 0.55 3.31 100.79 7.27 0.15 0.57 0.74 0.01 0.21 - 0.16 0.35 101.41 7.29 0.16 0.77 0.79 - 0.35 - 0.08 0.39 100.91 7.89 0.15 1.18 0.98 0.01 0.19 - 0.1 0.14 100.43 7.6 0.17 1 0.78 0.01 0.22 - 0.21 0.38 100.39 7.6 0.18 0.77 0.86 0.02 0.34 - 0.08 0.35 100.87 7.18 0.09 0.73 0.72 - 0.26 - - 0.33 101.01 7.12 0.1 0.53 0.72 0.01 0.3 - 0.29 0.63 99.11 7.78 0.14 0.71 0.75 - 0.27 - 0.06 0.31 100.28 6.96 0.15 0.89 0.73 0.02 0.25 - 0.1 0.29 101.37 9.49 0.25 1.27 0.82 - 0.11 - - 0.09 100.5 7.27 0.2 1.16 0.84 0.04 0.2 - 0.1 0.11 99.9 7.65 0.15 1.08 0.85 - 0.22 - - 0.02 101.18 6.78 0.18 0.42 2.49 0.02 0.34 - 0.03 0.9 100.7 6.22 0.13 0.49 0.5 - 0.41 - - 0.08 101.7 6.21 0.13 0.57 0.55 0.02 0.59 - 0.07 0.16 99.76 7.02 0.11 0.87 0.42 0.01 0.3 - 0.02 - 100.99 6.81 0.09 0.67 0.5 - 0.35 - - - 100.89 6.9 0.1 0.61 0.82 0.02 0.42 - 0.34 1.16 99.37 6.95 0.11 0.63 0.82 0 0.37 - 0.27 0.29 98.55 6.56 0.2 0.51 1.14 0.04 1.07 - - 0.37 101.57 6.58 0.15 0.38 0.74 0.01 0.63 - 0.03 0.19 99.95 7.97 0.15 0.58 0.89 0.01 0.23 0.03 0.25 0.46 100.77 7.11 0.1 0.58 0.74 - 0.39 0.12 0.34 1.03 100.13 6.92 0.13 0.63 0.98 - 0.68 - 0.53 1.36 100.66 7.74 0.13 0.77 0.73 - 0.17 - - 0.03 101.6 13.21 0.07 0.93 0.52 - 0.33 - - 0.09 98.95 7.17 0.14 0.65 0.78 0.03 0.44 - 0.08 0.43 99.63 7.26 0.15 0.63 0.85 0.02 0.33 - 0.12 0.85 100.15 7.4 0.12 0.43 0.91 0.02 0.27 0.21 0.01 0.12 100.18 7.37 0.15 0.66 0.85 0.03 0.41 0.05 0.19 1.05 100.56 7.1 0.15 0.51 0.88 - 0.74 - 0.33 1.07 100.76 7.13 0.16 0.7 1.09 0.03 0.74 - 0.29 0.09 100.83 164 — Appendix II major and minor chemical compositions of samples analysed by electron probe microanalysis Element oxide DOR 144 Na2O MgO Al2O3 15.19 0.56 DOR 145 14.25 DOR 146 16.97 DOR 147 DOR 148 DOR 149 SiO2 P2O5 SO3 Cl K2O 2.33 71.88 0.08 0.2 1.11 0.5 0.65 2.56 69.38 0.08 0.29 0.88 0.61 0.42 2.02 67.79 0.05 0.36 1.38 0.4 15.32 0.68 2.62 66.99 0.09 0.31 0.73 0.57 11.36 0.54 2.54 62.79 0.19 0.18 0.8 0.62 10.79 0.44 1.99 41.2 0.05 0.1 0.49 0.47 DOR 149 mineral 1.52 0.03 0.81 9.1 - - 0.14 - DOR 150 1.49 1.87 7.54 43.81 1.29 0.02 0.02 3.66 DOR 151 1.2 1.82 6.38 43.41 1.12 0.01 0.01 3.7 Deventer glass samples DEV 1 3.1 1.1 7.81 64.21 0.25 0.1 0.11 9.97 DEV 2 16.95 0.73 2.56 69.18 0.06 0.34 0.92 0.53 DEV 3 16.19 0.72 2.81 70.23 0.16 0.25 0.96 0.92 DEV 4 7.2 1.91 1.38 63.86 0.15 0.16 0.38 13.77 DEV 5 10.25 1.81 1.26 64.4 0.47 0.13 0.5 10.44 DEV 6 2.83 5.98 2.99 59.23 2.28 0.05 0.55 12.68 DEV 7 3.25 6.04 3.48 59.99 2.29 0.14 0.58 11.47 DEV 8 0.4 4.16 0.95 60.66 2.07 0.08 0.08 11.62 DEV 9 12.87 1.56 1.93 73.24 0.04 0.06 0.03 0.58 DEV 10 14.55 0.97 2.84 69.44 0.24 0.12 0.9 1.57 DEV 11 14.97 0.69 2.89 70.25 0.18 0.14 0.94 0.61 DEV 12 16.06 0.84 2.74 69.76 0.12 0.21 0.89 0.9 DEV 13 16.76 0.95 2.23 70.38 0.04 0.19 1.19 1.39 DEV 14 0.99 4.39 2.05 59.15 2.97 0.15 0.37 11.85 DEV 15 1.11 3.91 3.29 53.43 4.38 0.53 0.18 17.01 DEV 16 weathered 0.55 0.15 5.85 75.71 - 0.13 0.03 5.38 DEV 17 weathered 0.05 0.03 1.47 82.61 - 0.01 0.04 1.46 DEV 18 8.27 3.48 2.17 60.23 2.47 0.15 0.46 9.17 DEV 19 16.8 0.94 2.76 69.01 0.12 0.21 0.85 1.1 DEV 20 15.98 1.03 2.5 71.68 0.08 0.22 1.22 0.75 DEV 21 0.54 3.54 2.6 59.36 1.87 0.06 0.04 9.52 DEV 22 7.84 1.91 1.24 63.73 0.18 0.19 0.38 14.54 DEV 23 weathered 2.7 0.88 1.08 78.46 1.22 0.13 0.05 2.97 DEV 24 2.92 3.82 1.61 56.1 3.89 0.15 0.39 4.1 DEV 25 13.63 1.84 1.22 69.37 0.17 0.19 0.43 8.03 DEV 26 1.47 3.59 1.24 55.88 2.76 0.16 0.15 17.79 DEV 27 1.84 5.2 3.19 54.63 3.64 0.17 0.17 13.09 DEV 28 3.29 6.25 2.99 58.73 2.35 0.03 0.49 12.73 DEV 29 0.4 3.78 2.24 56.24 2.18 0.1 0.02 10.59 DEV 30a 1.26 3.75 2.82 50.15 1.92 0.21 0.26 18.18 DEV 30b 0.28 0.64 6.05 86.49 - 0.03 0.08 1.57 165 — CaO TiO2 5.96 MnO 0.09 FeO 0.59 CoO Sb2O5 0.76 0.15 SnO2 CuO PbO Total 1.16 - 0.2 - 100.76 6.55 0.12 0.37 0.75 0.02 2.61 - 0.09 0.62 99.83 5.77 0.06 0.02 0.43 0.01 2.89 - 2.5 0.14 101.2 6.9 0.1 0.55 0.54 - 1.98 - 0.94 1 99.31 6.91 0.11 0.97 0.5 - 0.59 - 1.79 10.66 100.56 3.78 0.14 0.31 0.49 0.01 - 3.01 0.29 39.12 102.66 0.99 0.04 0.08 0.36 - - 26.72 0.26 62.64 102.68 12.28 0.22 0.18 2.3 0 0.66 - - 21.95 97.29 13.75 0.24 0.22 2.27 0.02 0.69 - - 22.12 96.95 10.95 0.47 0.03 2.29 0.02 0.54 - - - 100.95 6.95 0.1 0.87 0.96 0.06 0.05 - - 0.11 100.38 6.75 0.17 0.59 0.66 0 0.32 - 0.08 0.18 100.99 5.77 0.11 1.08 0.49 0.01 0.77 - 0.11 2.83 99.98 8.6 0.07 1.24 0.47 0 0.54 - - 0.08 100.24 11.09 0.24 0.75 0.76 - 0.54 - 0.06 0.01 100.03 10.59 0.29 0.65 0.9 - 0.59 - 0.2 - 100.45 17.98 0.11 0.72 0.31 0.02 0.56 - - - 99.72 10.43 0.03 0.01 0.31 - 0.05 - 0.04 0.03 101.2 8.46 0.21 0.48 0.94 0.01 0.13 - 0.13 0.05 101.04 9.02 0.25 0.17 0.88 - 0.05 - 0.17 0.01 101.22 6.73 0.16 0.91 0.73 - 0.24 - 0.01 0.07 100.37 6.51 0.09 1.09 0.98 - 0.02 - 0.01 0.01 101.83 15.24 0.29 0.72 1 - 0.54 - 0.12 - 99.82 13.73 0.2 0.38 0.93 0.02 0.79 - 0.15 0.03 100.06 1.93 0.24 0.04 0.44 - 0.19 - - 0.06 90.7 0.81 0.15 0.01 0.12 0.02 0.2 - 0.01 - 86.99 11.5 0.17 0.64 0.96 0.02 0.5 - 0.07 0.09 100.33 7.06 0.16 1.09 0.99 0.01 0.19 - - 0.21 101.48 6.08 0.08 1.08 1.02 0.02 0.04 - 0.05 0.02 101.84 17.78 0.36 0.87 0.85 0 0.61 - - - 98 5.13 0.08 0.49 0.23 - 0.74 - - 2.42 99.09 1.33 0.11 0.28 0.36 0.02 0.12 - 0.1 0.09 89.9 23.59 0.14 1.23 0.55 0.04 0.26 - - 0.01 98.79 4.36 0.12 0.49 0.26 0.01 0.37 - - 0.01 100.49 11.95 0.07 0.42 0.42 - 0.86 - 2.46 0.06 99.28 14.31 0.3 0.79 1.25 - 0.54 - 0.23 - 99.35 10.79 0.18 0.72 0.84 0.02 0.62 - 0.02 0.13 100.18 21.15 0.27 0.74 0.71 - 0.54 - - - 98.97 17.56 0.12 0.95 0.61 0.02 0.93 - - - 98.72 0.99 0.1 0.01 0.26 0.13 0.06 - 0.05 0.07 96.8 166 — Appendix II major and minor chemical compositions of samples analysed by electron probe microanalysis Element oxide Na2O MgO DEV 31 14.97 DEV 32 0.89 DEV 33 0.84 DEV 34 weathered Al2O3 0.59 SiO2 P2O5 SO3 0.15 Cl 0.1 K2O 2.6 70.13 1.11 0.71 5.4 0.74 54.94 2.77 0.21 0.11 12.71 4.78 0.88 57.22 2.08 0.06 0.54 17.03 - - - - - - - - 2.09 5.14 0.98 52.27 3.08 0.19 0.23 13.77 DEV 36 1.18 3.72 1.24 56.07 3.15 0.2 0.13 17.51 DEV 37 18.06 1.8 1.91 63.98 0.53 0.04 1.25 4.77 DEV 38 1.38 3.52 1.02 58.08 3.2 0.23 0.27 16.1 DEV 39 0.74 4.46 2.41 56.75 3.12 0.04 0.27 14.09 DEV 40a 14.95 0.62 2.61 67.97 0.12 0.16 0.87 0.82 DEV 40b 14.93 0.67 2.7 70.23 0.14 0.13 0.91 0.71 DEV 41 14.55 1.08 2.62 67.19 0.41 0.2 0.92 1.97 Sust 1 (bead body) 16.71 0.83 2.76 63.08 0.18 0.1 0.82 0.83 Sust 2 (bead body) 18.01 0.81 2.75 68.22 0.08 0.18 0.88 0.84 Sust 2 (decoration) 16.31 0.74 2.61 62.49 0.2 0.19 0.81 0.95 Sust 3 (bead body) 15.85 3.4 2.79 65.12 0.25 0.27 0.53 1.9 Sust 3 (decoration) 8.98 0.39 1.5 36.15 0.08 0.03 0.44 0.51 Sust 4 (bead body) 16.88 5.75 2.06 64.47 0.23 0.18 0.68 1.86 DEV 35 Susteren glass samples Sust 4 (decoration) 16.81 1.36 2.72 67.7 0.15 0.15 0.85 1.57 Sust 5 (bead body) 20.03 0.78 2.49 66.7 0.1 0.23 1.17 0.68 Sust 5 (decoration) 18.5 0.65 2.52 67.29 0.1 0.6 1.15 0.7 Sust 6 (bead body) 16.96 0.75 3.05 69.97 0.06 0.16 0.92 1.21 Sust 6 (decoration) 17.67 0.85 2.79 67.06 0.13 0.22 0.91 1 Sust 7 18.89 0.71 2.47 67.15 0.08 0.32 1.05 0.64 Sust 8 18.99 0.78 2.44 66.01 0.09 0.37 1.08 0.58 Sust 9 17.24 0.84 2.95 64.99 0.1 0.16 0.62 0.78 Sust 10 19.29 0.83 2.6 65.33 0.09 0.29 0.84 0.66 Sust 11 1.9 5.6 2.46 59.06 2.53 0.06 0.47 14.21 Sust 12 17.99 0.99 2.72 66.44 0.21 0.22 0.98 0.84 Sust 13 16.98 1.98 2.66 64.89 0.21 0.22 0.8 1.48 Sust 14 16.18 1.17 2.64 65.83 0.31 0.19 0.82 1.27 Sust 15 18.29 1.07 2.66 62.45 0.15 0.28 1.23 0.74 Sust 16 15.07 1.41 3.6 66.99 0.31 0.16 0.74 1.54 Sust 17 14.26 0.66 2.33 65.14 0.18 0.24 0.67 5.2 Sust 18 13.96 0.88 7.74 63.34 0.31 0.04 0.03 3.3 Sust 19 8.87 3.69 2.23 61.74 1.5 0.16 0.47 8.5 Sust 20 19.04 0.86 2.58 64.85 0.1 0.35 1.26 0.64 Sust 21 16.48 1.05 2.68 67.22 0.19 0.18 0.83 1.1 Sust 22 16.56 1.08 2.69 67.45 0.27 0.27 0.81 1.24 Sust 23 18.06 1.01 2.6 66.25 0.37 0.22 0.96 0.84 167 — CaO TiO2 MnO 9.33 0.3 20.26 0.09 14.37 0.1 FeO 0.12 CoO Sb2O5 SnO2 CuO PbO Total 0.86 0.02 - - 0.19 0.07 101.25 0.53 0.31 0.01 0.64 - 0.07 - 99.68 0.57 0.41 - 0.71 - - - 99.58 - - - - - - - - - 0 17.69 0.11 2.36 0.44 0.01 0.5 - - 0.09 98.94 14.31 0.11 0.39 0.44 0 0.72 - 0.1 - 99.27 7.55 0.12 0.08 0.46 0.01 0.21 - - - 100.77 14.59 0.05 0.35 0.31 - 0.66 - 0.06 - 99.81 15.19 0.26 0.4 0.83 - 0.61 - 0.08 0.03 99.28 8.89 0.23 0.18 0.88 0 0.06 - 2.29 0.13 100.78 8.95 0.27 0.17 0.88 0.02 0.07 - 0.09 0.29 101.14 7.01 0.11 0.53 0.56 - 0.24 - 2.57 0.02 99.97 7.08 0.17 0.63 5.09 0.01 0.27 0.23 0.26 1.35 100.4 7.1 0.12 0.62 0.86 0 0.29 0.12 0.27 0.71 101.86 6.74 0.13 0.57 4.59 0.01 0.34 0.28 1.74 1.69 100.38 7.03 0.18 0.54 1.29 0.1 0.2 0.32 0.14 0.24 100.14 3.5 0.12 0.43 0.54 0 0 2.72 0.13 44.29 99.8 7.28 0.1 0.61 0.63 0 0 0 0.12 0.12 100.95 7.68 0.12 0.67 0.95 0.01 0.25 0.14 0.16 0.44 101.73 6.66 0.22 0.84 0.63 0.02 0.34 0 0.03 0.09 101.01 6.28 0.09 0.44 0.69 0 1.79 0 0.16 0.68 101.62 7.15 0.13 0.25 0.38 0 0 0.19 0.39 0.21 100.96 7.97 0.13 0.71 0.82 0.01 0.32 0 0.17 0.22 101.76 6.56 0.11 0.32 0.88 0.07 1.58 0 0.3 0.22 101.35 6.55 0.08 0.37 0.94 0 1.7 0 0.27 0.3 100.53 6.28 0.18 0.54 1.38 0.02 0.48 0.22 3.06 1.3 101.12 7.07 0.13 0.75 0.44 0 0.35 0 0 0.2 98.89 11.02 0.34 0.59 1.01 0 0 0 0.04 0 99.28 7.13 0.2 0.81 0.91 0 0.36 0.14 0.31 0.84 101.09 7.39 0.15 0.55 0.89 0 0.3 0.14 0.28 0.57 99.48 8.89 0.24 0.66 1.05 0.03 0.16 0 0.26 0.36 100.05 6.46 0.18 0.31 1 0 0.69 0.2 0.6 2.73 99.01 7.49 0.2 0.51 1.31 0 0.17 0 0.17 0.58 100.25 7.08 0.1 0.43 1.04 0.07 1.5 0.08 0.25 0.68 99.91 5.81 1.07 0.47 1.35 0.02 0.17 0.06 0.79 0.05 99.37 9.73 0.16 0.7 0.89 0.01 0.25 0.09 0.26 0.46 99.69 6.94 0.11 0.87 0.92 0.01 0.33 0 0.29 0.65 99.79 8.39 0.16 0.6 1.1 0.01 0.18 0 0.2 0.42 100.8 8.28 0.23 0.71 1.01 0.02 0.09 0.13 0.35 0.71 101.89 7.46 0.1 0.43 0.9 0 0.38 0.13 0.28 0.75 100.74 168 — Appendix II major and minor chemical compositions of samples analysed by electron probe microanalysis Element oxide Sust 24 Na2O MgO Al2O3 SiO2 P2O5 17.92 0.9 2.8 66.41 Sust 25 16.61 0.76 2.62 Sust 26 19.09 1.16 2.87 Sust 27 16.62 1.2 Sust 28 15.38 4.2 Sust 29 17.79 0.84 16.45 0.92 SO3 Cl K2O 0.22 0.24 0.89 0.85 67.02 0.08 0.28 0.82 0.62 65.79 0.08 0.24 1.03 0.68 2.61 66.22 0.28 0.17 0.84 1.23 2.03 64.61 0.27 0.15 0.66 2.48 2.73 66.47 0.15 0.26 0.92 0.76 2.65 68.09 0.18 0.35 1.1 0.39 Utrecht glass samples Utr 77 Utr 78 modern - - - - - - - - 13.5 0.41 1.7 69.87 0.03 0.21 0.03 0.39 Utr 31 8.77 0.57 12.6 66.05 0.03 0.01 0.1 4.73 Utr 32 15.54 1.66 4.65 67.07 0.06 - 0.3 1.9 Utr 33 0.37 0.44 36.89 54.29 - 0.04 0.04 2.26 Utr 79 Utrecht crucibles Utr 34 14.39 1.02 2.82 63.59 0.56 0.15 0.2 1.45 Utr 35 13.08 0.76 7.41 69.7 - 0.05 0.24 2.24 Utr 36 8.21 0.84 7.68 66.9 - 0.16 0.07 5.4 169 — CaO TiO2 MnO FeO CoO Sb2O5 SnO2 7.5 0.18 0.61 0.88 0.03 1.97 7.28 0.09 0.34 0.86 0.09 6.27 0.4 1.79 1.26 0.01 7.88 0.16 0.65 0.88 0 0.25 7.1 0.16 0.6 0.7 0.03 0.11 7.24 0.18 0.74 0.95 0 0.5 6.29 0.11 0.08 0.74 - CuO PbO Total 0 0.24 0.61 102.25 0.24 0 0.24 0.85 98.79 0 0 0 0.08 100.74 0 0.24 0.59 99.82 0 0.06 1.64 100.16 0.14 0.53 1.4 101.6 - - 0.06 0.57 97.98 - - - - - - - - - - 10.39 0.03 0.03 0.12 0.08 0.04 - 0.06 0.07 96.97 2.48 0.64 0.14 3.17 0.11 0.45 0.02 0.03 0.09 99.98 5.78 0.21 0.22 1.53 0.14 0.11 0.02 0.01 - 99.19 0.34 2.22 0.02 2.88 0 0.12 0.02 0.01 0 99.93 6.82 0.15 0.73 1.52 0.13 0 - 0.62 4.91 99.07 4.87 0.59 0.37 1.36 0.06 0.29 0.03 1.12 - 102.16 3.29 0.45 0.3 1.59 0.03 0.5 - 0.39 - 95.81 170 — Trace elements Appendix III trace element chemical compositions of samples analysed by LA-ICP-MS Li B V Cr Co Ni Cu Zn As Rb Sr Y Zr Nb Mo Sn Sb Gennep glass samples GE41 7.4 187.9 27 22 54 27.9 5177.3 258.9 59.5 9.1 484 7.5 88.8 3.3 2.4 2546 4570 GE42 8 351.5 78 133 22.1 27.6 1194.5 31.6 19.4 7.3 522 12.2 181 4.5 2.4 308 1081 GE43 5.1 107 74 40 11.5 22 114.6 32.6 11 6.1 500 10.1 142.7 3.8 3.7 10 24 GE44 6.5 199.3 74 40 12 25.5 231.8 37.5 11.7 6.8 501 10.6 139.1 3.8 3.6 125 163 175 GE45 7.2 151.3 76 47 16 25.3 358.6 44.7 15.8 7.4 552 11.5 185.5 4.6 3.1 67 GE46 4.9 114.7 75 40 11.9 21.9 120.6 37.2 10.8 6 504 10.9 151.8 3.9 3.7 25 68 GE47 7.2 201.3 26 19 37.1 18.2 3408.8 105.9 44.5 8.9 536 7.6 86.1 2.8 2.5 2639 3559 GE48 9.3 158.1 16 12 6.7 7.7 131.3 37.2 10.6 9.2 450 6.5 52.7 1.8 1.3 72 1741 GE49 11.2 165.5 25 18 14.4 11.7 474.3 69.3 10.1 10.7 542 7.5 81 2.6 2.1 608 1021 GE51 10.9 156.4 26 18 14.8 12.3 724.9 69.3 10.7 9.2 575 7.3 82.3 2.8 2.2 571 872 GE52 13.5 187 26 21 58.9 30.5 6044.5 293.6 64.9 11.9 508 8.4 93.3 3.6 2.6 1640 4274 GE53 17.7 142.2 33 17 13.9 13.6 382.5 49.5 6.1 8.8 697 7.7 83.6 2.7 3.6 220 322 GE54 9 208.2 31 18 7.7 16.7 66.5 35.5 5 8.2 734 8.5 99.5 3.2 3.2 445 331 GE55 7.3 203.2 32 17 7.2 11.6 66.9 24.4 4.7 7.5 746 8.4 95.9 3.2 3.6 25 300 GE56 5.7 236.4 34 30 8.7 14.2 34.9 27.8 2.5 5.6 562 8.3 141.4 3.7 4.6 4 1 GE57 11 183.6 32 22 7.5 11 83 25.2 5.2 9.5 693 8.3 89.4 3 2.6 456 267 GE58 11 177.2 31 17 7.8 11.3 85.3 37.8 5.4 10.1 686 8.3 91.2 3.2 2.3 395 340 GE59 9.2 194.8 27 15 6.6 16 217.4 121.2 22.1 22.9 773 8.2 89.4 2.8 3.2 24611 237 GE60 8.1 208.7 33 17 7 12.2 73.4 21.7 4.6 7.4 770 8.4 97 3.3 3.2 9 314 GE61 12.2 174.8 19 14 7.4 9.4 135.8 32.7 11.9 11.3 503 6.9 62.6 2.2 1.5 203 1998 GE62 15.1 165.8 25 18 12.7 11.8 397 76.7 11.2 10.4 545 7.6 77.9 2.6 2.4 704 1469 GE63 11.6 158.2 23 18 8.7 10.5 209.7 40.9 8.1 11 522 7.9 83.2 2.6 2.3 300 1123 GE64 16.4 180.7 26 17 11.7 12.4 242 104.4 9.2 14.2 577 7.8 77.6 2.7 2.3 1315 1148 GE65 12.4 163.3 27 18 17.1 12.3 1042.8 64 13.2 9.1 530 7.6 82.7 2.6 2.7 337 1474 GE66 8.8 159.8 19 14 9 10 337.9 39.1 10.4 8.8 488 6.9 56.8 2 2.7 76 1007 GE67 10.6 169 27 20 14.1 12 456.6 64.3 9.7 8.8 490 7.5 83.2 2.6 3 150 1042 GE68 8.4 170.2 20 14 10.1 10.3 217 43.1 6.4 10.4 506 7.3 59.9 2.1 1.7 254 845 GE69 7.3 173.9 39 46 12 15.2 83.6 33.7 4.6 6 514 9.6 199.7 4.9 5.2 34 8 Maastricht-Jodenstraat (MAJO) glass samples Joden 1 4.5 179.3 26.38 27.36 9.4 41.4 15637.3 4231.5 51.3 18.1 479 10.2 113.3 3.6 3.7 11459 118 Joden 2 7.4 145.9 27.23 15.95 10.9 37.7 15536.7 2119.5 43.5 10.6 615 7.3 81.4 2.8 4.7 6426 206 Joden 3 9 169.1 31.11 16.8 8.4 16.7 90.1 36.6 5.6 9.1 745 7.8 85.2 3.1 4.3 18692 165 Joden 4 8 155.1 50.54 18.22 11.7 22.6 91.6 36 9.9 8.2 649 9.4 97.4 3 3.5 52964 38 Joden 5 6.7 164 30.69 15.05 9.9 23.1 201.4 45.2 11.7 9 631 7 77.9 2.9 5.3 53150 84 Joden 6 11.6 138.8 32.97 19.4 8.7 22.1 302.2 74.7 30.1 17 714 8.9 91.3 3.5 4.2 78067 216 Joden 37 4.5 161.2 13.79 14.41 259.2 82.9 357 23 12.2 14.4 436 6.3 78.5 2.2 3.3 9097 3 Joden 38 4.9 189.7 15.62 12.67 204.7 69.3 823.9 17.6 9.1 6.6 541 6.2 73.8 2.2 2.6 81 6 Joden 39 4.4 163.1 15.09 14.03 311.2 89.9 1070.8 18 9 7.1 491 6.7 78.3 2.2 3.7 33 4 Joden 40 4.3 174.5 14.93 14.02 308.6 97 1059.1 19.3 12.7 7.4 491 6.2 76.3 2.1 4.1 55 4 Joden 41 8.4 166.3 21.8 16.27 35.4 48.2 20139.9 1286.1 65.9 12.5 521 7.9 66.8 2.3 1.7 2182 1513 Joden 42 9.3 153.8 20.55 12.56 32.4 47.9 20130.9 1804.3 72.1 12.9 486 7.7 63.1 2.2 1.9 2413 1648 Joden 43 9.9 158.2 30.69 15.57 13.9 74.5 31204.5 3655.4 53.3 10.4 691 7.8 84.4 3 4.9 8394 252 Joden 44 4.6 121.5 21.88 9.93 4.3 14.7 38.8 18.5 2.1 5.4 572 5 54.8 2 4.4 6 34 Joden 45 15.2 153.9 28.82 18.13 10.7 16.7 152.4 53.1 7.4 10.6 568 7.8 83.1 2.7 3.2 190 684 Joden 46 15.4 157.4 36.32 36.73 9.3 12.6 138.6 57.1 5.5 6.8 470 8.6 157.3 4 3.1 24 357 Joden 47 11.3 166 37.74 17.88 10.8 25.8 277.7 68.8 13.8 12.3 740 9.2 98.3 3.5 4.9 50418 247 Joden 48 10.1 170.8 33.58 16.66 8.9 18.8 1287.7 44.4 3.9 10.5 755 7.6 92.2 3.1 5.3 215 189 Joden 49 9.2 404.3 52.39 28.12 7.6 268.3 5825 2077.4 2.4 41.7 164 20.2 286.3 8.4 0.3 195 66 171 — Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U 0.27 327 7.7 13.4 1.7 6.9 1.5 0.4 1.2 0.2 1.2 0.3 0.7 0.1 0.7 0.1 2 0.2 25821 1.3 1.2 0.13 353 11.9 15 2.6 10.8 2.2 0.6 2.4 0.4 2.2 0.4 1.1 0.2 1.3 0.1 4.2 0.3 3265 1.8 1.5 0.09 385 9.9 14.3 2.3 9.9 2 0.5 1.8 0.3 1.8 0.4 1.1 0.2 1.1 0.1 3.4 0.2 125 1.5 1.3 0.08 386 10.7 14.8 2.4 10 2.2 0.6 1.8 0.3 1.7 0.4 1.1 0.1 1 0.2 3.4 0.2 647 1.5 1.3 0.08 464 10.9 16.2 2.6 11.4 2.5 0.6 2.4 0.3 2 0.4 1.1 0.2 1.3 0.2 4.3 0.3 922 1.8 1.4 0.08 396 10.1 14.7 2.4 9.7 1.9 0.6 2 0.3 1.8 0.4 1.1 0.2 1.1 0.2 3.6 0.2 192 1.6 1.3 0.2 321 7.8 13.7 1.8 7.4 1.6 0.4 1.3 0.2 1.3 0.2 0.8 0.1 0.6 0.1 2.2 0.2 18040 1.3 1.3 0.18 262 6.5 11 1.5 5.7 1.3 0.4 1.2 0.2 1 0.2 0.6 0.1 0.5 0.1 1.3 0.1 441 1.1 1 0.21 352 7.3 13.4 1.7 7 1.4 0.4 1.3 0.2 1.2 0.3 0.8 0.1 0.7 0.1 2.1 0.2 2036 1.3 1.2 1.2 0.18 360 7.1 13.4 1.7 7.3 1.4 0.4 1.3 0.2 1.2 0.2 0.8 0.1 0.6 0.1 2.1 0.2 3371 1.3 0.34 367 8.7 14.5 1.9 7.8 1.7 0.4 1.5 0.2 1.3 0.3 0.8 0.1 0.9 0.1 2.3 0.2 16358 1.6 1.2 0.12 439 7.4 12.9 1.8 7.2 1.5 0.3 1.5 0.2 1.4 0.3 0.8 0.1 0.7 0.1 2 0.2 1407 1.3 1.4 0.14 359 8.2 14.7 1.9 7.6 1.8 0.4 1.6 0.2 1.5 0.3 0.9 0.1 0.8 0.1 2.6 0.2 140 1.7 1.4 0.13 361 8.1 14 1.9 7.9 1.6 0.4 1.6 0.2 1.4 0.3 0.8 0.1 0.8 0.1 2.5 0.2 121 1.5 1.6 0.07 415 7.9 14.5 1.9 7.4 1.4 0.4 1.5 0.2 1.4 0.3 0.8 0.1 0.9 0.1 3.4 0.2 16 1.6 1.4 0.17 339 8.4 14.3 2 7.7 1.8 0.4 1.5 0.2 1.5 0.2 0.9 0.1 0.7 0.1 2.2 0.1 262 1.4 1.4 0.18 330 8.3 14.6 2 7.6 1.6 0.5 1.5 0.2 1.5 0.3 0.9 0.1 0.7 0.1 2.3 0.2 356 1.5 1.4 0.22 383 7.7 13.5 1.8 7.7 1.6 0.4 1.7 0.2 1.4 0.3 0.8 0.1 0.7 0.1 2.2 0.2 25043 1.3 1.4 0.11 368 8.3 14.7 1.9 8.3 1.7 0.4 1.9 0.2 1.4 0.3 0.8 0.1 0.9 0.1 2.3 0.2 113 1.5 1.6 0.19 276 6.6 12.4 1.5 6.4 1.3 0.4 1.3 0.2 1.1 0.2 0.7 0.1 0.6 0.1 1.5 0.1 498 1.2 1.1 0.24 367 7.6 13.7 1.7 7.2 1.5 0.4 1.4 0.2 1.3 0.3 0.7 0.1 0.6 0.1 1.9 0.2 1581 1.3 1.1 0.18 342 7.4 12.8 1.7 7.2 1.5 0.4 1.4 0.2 1.3 0.2 0.7 0.1 0.7 0.1 2.1 0.2 713 1.3 1.1 0.27 367 7.7 13.3 1.8 7.1 1.4 0.4 1.5 0.2 1.3 0.3 0.7 0.1 0.7 0.1 1.8 0.1 1361 1.3 1.2 1.3 0.22 425 7.6 12.9 1.7 7.2 1.5 0.4 1.4 0.2 1.2 0.3 0.9 0.1 0.7 0.1 2 0.2 3237 1.3 0.16 348 6.9 12.3 1.6 6.2 1.4 0.4 1.3 0.2 1.2 0.3 0.7 0.1 0.6 0.1 1.4 0.1 803 1.1 1 0.19 387 7.2 13.1 1.7 6.6 1.4 0.4 1.3 0.2 1.2 0.2 0.7 0.1 0.7 0.1 2.1 0.2 1267 1.2 1.1 0.16 303 6.9 12.6 1.6 6.5 1.3 0.4 1.2 0.2 1.3 0.2 0.7 0.1 0.7 0.1 1.6 0.1 700 1.1 1.1 0.07 453 9.1 16.2 2.2 9.1 1.9 0.5 1.9 0.2 1.7 0.3 1.1 0.1 0.9 0.1 4.8 0.3 211 2 1.5 0.26 386 9.9 16.9 2.3 9.5 1.7 0.4 1.8 0.3 1.6 0.3 1 0.1 0.9 0.1 2.8 0.2 88534 1.9 1 0.26 274 7.5 14 1.8 7 1.5 0.4 1.3 0.2 1.3 0.3 0.8 0.1 0.7 0.1 2.1 0.2 61933 1.4 1.1 0.24 339 8 14.1 1.8 7.8 1.6 0.5 1.5 0.2 1.4 0.3 0.9 0.1 0.8 0.1 1.9 0.2 2319 1.4 1.2 0.31 350 9.3 13.7 2.1 8.4 1.7 0.5 1.8 0.3 1.7 0.3 0.9 0.1 0.9 0.1 2.3 0.2 47497 1.5 1.1 0.46 268 7.1 12.7 1.6 6.6 1.3 0.4 1.4 0.2 1.2 0.2 0.7 0.1 0.6 0.1 2 0.2 316647 1.3 1.3 1.05 239 9.5 17.5 2.2 8.9 1.8 0.4 1.8 0.3 1.5 0.3 0.9 0.1 0.9 0.1 2.4 0.2 780143 2.2 1.1 0.06 151 6.1 11.1 1.4 5.7 1.2 0.4 1.3 0.2 1.1 0.2 0.7 0.1 0.6 0.1 2 0.1 5767 1.1 1.1 0.08 159 5.9 10.8 1.5 5.8 1.3 0.3 1.2 0.2 1 0.2 0.6 0.1 0.6 0.1 1.7 0.2 3529 1 1.2 0.1 155 6.3 11.3 1.4 5.9 1.2 0.3 1.1 0.2 1.2 0.2 0.7 0.1 0.7 0.1 2.1 0.1 5276 1.1 1 0.07 155 6.3 11.3 1.4 6.3 1.2 0.4 1 0.2 1 0.3 0.8 0.1 0.6 0.1 1.9 0.1 5394 1.1 1 0.39 298 7.8 13.1 1.7 6.9 1.4 0.5 1.3 0.2 1.3 0.2 0.9 0.1 0.6 0.1 1.5 0.2 75113 1.3 1 0.54 270 7.3 13.1 1.7 7.3 1.4 0.4 1.3 0.2 1.3 0.3 0.8 0.1 0.8 0.1 1.6 0.1 84741 1.4 1.2 0.22 297 7.8 14.2 1.8 7.8 1.7 0.4 1.6 0.2 1.2 0.3 0.8 0.1 0.8 0.1 2.2 0.2 7289 1.3 1.1 0.08 217 5 9.1 1.2 4.7 1 0.2 1 0.1 0.9 0.2 0.6 0.1 0.4 0.1 1.4 0.1 40 0.9 0.8 0.24 346 7.5 13.2 1.8 7.3 1.8 0.4 1.4 0.2 1.3 0.2 0.7 0.1 0.7 0.1 2.1 0.2 338 1.3 1.1 0.08 1529 8.1 14.4 2 7.2 1.4 0.6 1.5 0.2 1.6 0.3 1 0.1 1 0.1 3.9 0.3 178 1.6 1.2 0.55 314 9.6 15.6 2.2 9.1 1.8 0.4 1.9 0.3 1.6 0.3 0.9 0.1 0.9 0.1 2.4 0.2 407431 1.7 1.2 0.18 343 8 14.2 2 7.7 1.6 0.4 1.6 0.2 1.3 0.2 0.8 0.1 0.7 0.1 2.3 0.2 835 1.5 1.2 1.69 234 19.9 38.5 4.3 16.3 3.6 0.7 3.6 0.5 3.8 0.7 2.5 0.3 1.9 0.4 7.9 0.6 297 5.9 2 172 — Appendix III trace element chemical compositions of samples analysed by LA-ICP-MS Trace elements Li B V Cr Co Ni Cu Zn As Rb Sr Y Zr Nb Mo Sn Sb Joden 50 5.1 138.1 41.79 18.01 10.6 22.8 94.2 71.2 12.9 72.6 482 7.8 85.6 2.9 9.3 124317 122 Joden 51 6.8 139.1 25.17 13.94 12 38.4 7657.8 1092 29.2 8.4 608 5.8 65.9 2.4 4.6 6446 188 Joden 52 7.2 149.9 28.97 14.81 12.1 75.3 42945.3 5559.9 74.4 8.6 666 6.9 84.1 2.8 5.1 1830 159 Joden 53 4 160.5 17.07 15.45 178.8 47.2 805.3 22.5 7.6 6.7 365 5.9 84.2 2.5 2.5 3897 3 Joden 54 7.9 143.4 34.27 15.11 24.7 42.3 9609.1 1721.7 32.7 10 548 7 70.7 2.7 3.3 11562 170 Joden 55 8.1 88.9 21.38 13.91 9.3 68.3 5565.2 6530.8 85.4 10.9 426 5.5 58.4 2.3 2.7 10115 92 Joden 56 6.6 114.8 23.68 12.84 6.2 15.2 126 35 10 8.6 561 6.1 65.8 2.4 4 30841 169 Joden 57 6 159.1 25.4 14.58 20.7 88.3 26978.6 1474.6 43.2 8.6 602 7.1 74.6 2.7 4.4 5930 110 Joden 58 5.6 165.1 32.57 14.07 9.3 21.5 57.6 38.4 9.7 7.9 660 6.5 70 2.5 5.5 41978 60 Joden 59 7.5 157.7 31.1 14.33 9.5 16.9 112.2 36.8 7.7 9.1 659 7.2 77.2 2.6 4.3 27321 212 Joden 60 0.1 237 0.34 29.13 1.3 14.9 82 68.6 0.9 0.2 7 0.2 0.7 - - 376 11 Joden 61 12.5 195.2 18.66 11.61 378.8 22.6 2057.6 59.8 43.4 13.1 441 6.4 60.7 2.2 2.1 124 11536 Joden 62 11.1 160.3 32.81 17.08 7.1 16.1 51.1 29 5.1 8.7 788 7.9 100.7 3.2 5.4 10 99 Joden 63 6.3 166.7 30.21 16.38 7.5 16.1 7.3 26.6 3.3 10.8 860 7.5 84.6 2.9 5.6 684 179 Joden 64 5.2 99.1 21.46 12.13 4.6 15.1 121.4 32.3 17.1 8.5 401 5 59.9 2.3 3.3 39547 98 Joden 65 5.6 93.9 18.73 10.93 5.2 10.9 136.2 36.5 7.7 7.1 481 4.7 51.3 1.9 3.3 23702 121 Joden 66 7.3 161 33.09 17.2 14.2 53.5 14623.8 2929.4 46.4 10.4 655 7.2 81.4 3 4.9 14412 113 Joden 67 10.6 150.4 31.43 16.77 13.5 65.5 26901 4409.2 60.3 11.8 648 8.1 84.9 2.8 4.6 1585 215 Joden 68 3.6 133.6 12.57 32.33 15.6 9.5 154.7 18.7 2.1 9.1 389 6.1 45.5 1.5 1.5 34 592 Joden 69 7.2 105.3 43.54 15.87 10.1 23.2 183.8 39.1 11.4 8.5 473 7.8 67.1 2.5 2 104859 109 Joden 70 8.6 137.9 30.49 15.75 8.7 19.7 182.4 52.5 6.7 9.6 631 7.3 78.6 2.8 4 24800 153 Joden 71 6.8 100.4 40.71 12.9 8.7 21.7 238.1 36 11.6 8.9 450 7.2 62.3 2.3 2.4 39902 86 Joden 72 6.2 136 24.71 13.51 8.2 18.2 116.7 37.4 10.6 9.1 496 6.1 65.6 2.4 4.4 37442 70 Joden 73 6.5 154 28.16 15.2 19.3 96.5 28045.7 1778.3 50.9 7.9 737 6.9 86.6 2.7 5.3 18725 183 Joden 74 6 149.3 26.78 15.54 19.3 89.8 27398.5 1726.3 47.4 7.5 717 6.8 80.7 2.7 5 5641 154 Joden 75 8.5 136.6 34.06 14.53 12.2 62.7 28328.3 3250.1 63.5 9.3 582 7.2 70.6 2.4 3.2 50978 294 Joden 76 6 150.4 27.77 14.63 18.9 90.7 26781 1682 46 7.5 719 7.2 80.1 2.6 5.2 5800 154 Wijnaldum glass samples WIJ1 10.1 163.4 28.18 15.89 9.8 11.2 230.1 54.6 7.1 9.6 631 8.6 79.7 2.5 2.1 402 626 WIJ2 19 174.3 27.85 18.31 11.3 11.7 576.3 70.2 8.8 10.4 598 8.3 89.6 2.7 2.3 542 989 WIJ3 15 168.2 28.67 17.59 12.1 12.4 573.4 144.8 9.9 11.5 625 8.9 86.4 3 2 997 803 WIJ4 2 104.8 4.39 4.29 1.2 7.9 224.2 19.2 45 4 215 3.3 24 0.8 0.1 23001 3049 WIJ5 7.6 97.8 32 14.76 9.6 23.1 155.4 42.3 12.6 15.2 363 9.2 59.6 2.4 1 43112 117 WIJ6 6.2 95.2 28.12 17.75 9.6 20.9 214.4 32 23.6 7.7 343 8.2 71.5 3.2 1.1 59853 112 173 WIJ 7 8.4 93.3 32.74 18.13 18.1 35.6 199.4 83.6 8 13.2 565 7.8 74.4 2.8 4.3 43807 WIJ 8 6.8 49.2 12.74 12.58 2.2 12.7 190.3 15.2 2.6 11.1 240 5.2 29.1 1.5 0.4 37252 1 WIJ 9 8 41 10.88 11.2 2.1 13.4 263.5 29.2 4.3 12 212 4.7 25.6 1.6 0.3 46669 1 WIJ 10 9.2 90.9 21.33 26.39 8.7 8.3 375.9 56.6 3.2 8.2 235 7.1 184.7 3.9 0.5 82 156 WIJ 11 55.6 145.4 20.82 17 8.8 10.7 698.3 181.8 14.6 16.3 440 7 75.2 2.5 1.7 221 1802 WIJ 12 8.1 164.2 23.24 21 6.5 9 123 38.3 13.3 9.8 458 7.5 95.5 2.6 2.2 35 1679 WIJ 13 8.8 163.8 22.45 17 6.5 9.1 74.6 29.9 11.3 8.9 485 7.2 73.6 2 2 19 2258 WIJ 14 8 60.9 17.15 17 4.8 29.6 290.8 54.8 18 15.8 136 4.7 57.6 3.4 0.3 63145 53 WIJ 15 4 98.5 20.28 26 822.5 25.1 934.3 2256.1 11.2 4.2 177 6.6 169.1 3.9 2.6 97 7 WIJ 16 6.5 76.7 9.91 9 7.6 6.8 1172.2 42.8 9.7 8.3 283 4.7 38.9 1.2 0.9 359 828 WIJ 17 20.2 76.1 50.3 22 51.6 118.8 19996.3 2905.8 89.5 46.2 287 14 197.5 6.7 3.6 2095 145 WIJ 18 13.1 151.6 52.76 22 25.4 48.4 18941.6 2536.3 65.1 14.4 528 10.9 99.5 3.5 3.4 7748 551 173 — Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U 0.8 304 8 12.9 1.9 7.7 1.4 0.4 1.5 0.2 1.3 0.3 0.8 0.1 0.6 0.1 2 0.2 23851 1.3 0.16 238 6.2 11.2 1.4 5.8 1.1 0.3 1.1 0.2 1 0.2 0.6 0.1 0.6 0.1 1.6 0.2 78875 1.1 0.9 1 0.15 280 6.9 12.3 1.6 6 1.6 0.3 1.3 0.2 1.2 0.2 0.7 0.1 0.6 0.1 2 0.1 3115 1.3 1.1 0.08 159 6.4 12 1.6 5.8 1.2 0.3 1.2 0.2 1.1 0.2 0.6 0.1 0.7 0.1 2.1 0.2 4798 1.1 1 0.23 276 7.6 12.8 1.7 6.9 1.3 0.4 1.3 0.2 1.2 0.3 0.7 0.1 0.7 0.1 1.7 0.1 71856 1.3 1.1 0.57 259 7.1 11.2 1.4 6 1.2 0.2 0.9 0.2 1 0.2 0.5 0.1 0.6 0.1 1.5 0.1 354763 1.3 1.7 0.39 186 6.3 11.2 1.5 6.6 1.2 0.3 1.1 0.2 1 0.2 0.7 0.1 0.5 0.1 1.6 0.1 277719 1.3 0.9 0.18 275 7.1 12.7 1.7 6.9 1.6 0.4 1.6 0.2 1.2 0.3 0.6 0.1 0.8 0.1 1.8 0.2 37801 1.3 1 0.18 317 6.8 11.4 1.5 6.2 1.3 0.3 1.3 0.2 1.2 0.2 0.8 0.1 0.7 0.1 1.7 0.1 22283 1.1 1.2 0.19 270 6.8 12.4 1.7 6.7 1.3 0.4 1.4 0.2 1.1 0.3 0.7 0.1 0.7 0.1 2 0.2 28303 1.2 1.3 0.11 9 0.3 2.4 0 0.1 0 0 0.1 0 0 0 0 0 0 0 0.1 0 2484 0 0 0.41 261 6.9 11.7 1.5 5.8 1.3 0.3 1.2 0.2 1.1 0.2 0.6 0.1 0.6 0.1 1.5 0.1 2011 1.2 1.1 0.13 293 7.7 14.4 1.8 7.8 1.5 0.5 1.6 0.2 1.3 0.3 0.9 0.1 0.8 0.1 2.4 0.2 88 1.4 1.2 0.1 258 7.3 13.3 1.7 7.1 1.4 0.4 1.5 0.2 1.3 0.3 0.7 0.1 0.7 0.1 2.2 0.2 205 1.3 1.2 0.64 167 5.2 8.8 1.2 4.6 1.1 0.2 0.8 0.1 0.9 0.2 0.5 0.1 0.5 0.1 1.5 0.2 468522 1.1 0.7 0.34 150 4.9 8.9 1.2 4.6 0.9 0.2 0.9 0.1 0.8 0.2 0.5 0.1 0.5 0.1 1.4 0.1 221741 0.9 0.7 0.3 291 7.6 14.2 1.8 7.3 1.5 0.4 1.6 0.2 1.4 0.3 0.7 0.1 0.7 0.1 2.1 0.5 123336 1.5 1.3 0.24 318 7.8 15.2 1.9 7.7 1.7 0.4 1.5 0.2 1.4 0.3 0.8 0.1 0.8 0.1 2.1 0.2 38984 1.5 1.1 0.15 224 5.9 10.6 1.4 5.1 0.9 0.3 1.1 0.2 0.9 0.2 0.6 0.1 0.5 0.1 1.2 0.1 182 1 0.9 0.6 261 7.8 12.3 1.8 7.4 1.4 0.3 1.5 0.2 1.3 0.3 0.8 0.1 0.7 0.1 1.8 0.2 423852 1.3 0.9 0.34 249 7.6 13.4 1.7 7.3 1.5 0.3 1.4 0.2 1.4 0.2 0.7 0.1 0.6 0.1 2 0.2 206637 1.3 0.9 0.43 265 7.1 10.8 1.6 6.3 1.1 0.3 1.5 0.2 1.3 0.3 0.8 0.1 0.6 0.1 1.6 0.1 288546 1.1 0.8 0.44 221 6.5 11.8 1.4 6.1 1.3 0.3 1.2 0.2 1.1 0.2 0.6 0.1 0.5 0.1 1.5 0.1 294751 1.2 1.1 0.22 254 7.2 13.3 1.7 6.9 1.5 0.3 1.3 0.2 1.4 0.2 0.7 0.1 0.7 0.1 2.2 0.2 40403 1.3 1.1 0.18 252 7 12.5 1.6 6.5 1.2 0.3 1.3 0.2 1.1 0.2 0.7 0.1 0.6 0.1 1.9 0.2 38669 1.3 1 0.5 283 7.9 12.9 1.8 7.3 1.5 0.4 1.5 0.2 1.4 0.3 0.7 0.1 0.8 0.1 1.7 0.2 59580 1.3 0.9 0.24 249 7.2 12.7 1.6 6.6 1.4 0.4 1.2 0.2 1.2 0.2 0.7 0.1 0.7 0.1 2 0.2 38566 1.3 1.1 1.2 0.14 320 8.1 13.8 1.9 7.6 1.7 0.5 1.6 0.2 1.4 0.3 0.7 0.1 0.9 0.1 2 0.2 1149 1.3 0.14 399 8.2 13.9 2 7.8 1.5 0.4 1.5 0.2 1.3 0.3 0.8 0.1 0.9 0.1 2.3 0.2 1409 1.4 1.1 0.21 357 8.4 14.8 2 8.3 1.7 0.5 1.4 0.2 1.3 0.2 0.8 0.1 0.8 0.1 2.4 0.2 2525 1.4 1.2 0.1 85 3.1 5.4 0.7 3 0.8 0.2 0.6 0.1 0.5 0.1 0.3 0.1 0.3 0 0.7 0 328951 0.4 0.5 0.83 177 9.6 15.4 2.3 8.7 1.8 0.4 2 0.3 1.6 0.3 1 0.1 0.9 0.1 1.6 0.2 496383 1.7 0.9 0.29 150 7.8 12.2 1.7 7.1 1.5 0.3 1.2 0.2 1.4 0.3 0.9 0.1 0.9 0.1 1.9 0.2 477936 1.6 0.9 0.66 222 8.8 13.4 2.1 8 1.5 0.4 1.6 0.2 1.5 0.3 0.8 0.1 0.8 0.1 1.9 0.2 508491 1.6 1 0.62 156 5.3 10.3 1.2 5.2 1 0.3 0.9 0.1 0.9 0.2 0.5 0.1 0.4 0 0.8 0.1 440657 1.1 1 0.68 139 4.9 9.8 1.2 4.6 1 0.3 0.9 0.1 0.7 0.2 0.5 0.1 0.4 0.1 0.7 0.1 545641 1.2 0.5 0.1 199 7.3 13.9 1.8 7 1.6 0.4 1.3 0.2 1.2 0.2 0.7 0.1 0.7 0.1 4.5 0.3 584 1.6 1.1 0.25 316 7.4 13 1.6 7 1.3 0.3 1.3 0.2 1.2 0.3 0.7 0.1 0.6 0.1 1.9 0.1 1698 1.3 0.9 0.16 397 7.3 13 1.7 6.8 1.5 0.4 1.2 0.2 1.2 0.2 0.7 0.1 0.6 0.1 2.4 0.1 725 1.3 1 0.13 348 6.7 12 1.6 6.6 1.2 0.4 1.4 0.2 1.2 0.2 0.7 0.1 0.7 0.1 1.9 0.1 241 1.1 1 0.82 74 6.5 12 1.5 5.1 1.2 0.2 0.9 0.1 0.9 0.2 0.5 0.1 0.4 0.1 1.6 1 707917 2.2 0.7 0.09 169 6.8 13.1 1.7 6.8 1.1 0.3 1.4 0.2 1 0.2 0.7 0.1 0.7 0.1 3.9 0.3 354 1.4 1.2 0.18 163 4.2 7.9 1 3.7 0.8 0.2 0.8 0.1 0.6 0.2 0.5 0.1 0.4 0.1 1 0.1 1445 0.8 0.6 2.34 405 17.5 61.5 4.3 16.3 3.1 0.5 2.8 0.4 2.6 0.5 1.4 0.2 1.5 0.2 5.4 0.5 270987 7.3 2 0.48 462 11.1 17.4 2.7 11.5 2.1 0.5 1.9 0.3 1.9 0.4 1.2 0.2 1.2 0.2 2.7 0.2 73142 2.1 1.2 174 — Appendix III trace element chemical compositions of samples analysed by LA-ICP-MS Trace elements Li B V Cr Co Ni Cu Zn As Rb Sr Y Zr Nb Mo Sn Sb WIJ 19 9.6 155.5 28.19 17 23.8 38.2 8869.3 2160.2 42.2 12.2 487 15 85.6 2.9 0.9 12300 59 WIJ 20 8.1 180.2 75.45 19 28.4 44.1 7731.9 711.9 34.9 11.1 621 13.2 99.1 3.4 2.7 7222 68 WIJ 21 6.3 157.7 61.66 17 20.2 35.3 112.1 51.2 15.8 8.2 638 12.4 91.2 3.2 2.7 20435 56 WIJ 22 5.7 168.8 21.93 12 8.6 12.9 206.2 56.2 15.6 8.5 449 6.9 53.1 1.9 1.2 32949 1976 WIJ 23 5.8 165.5 27.84 21 9.5 23.3 46.6 29.1 5.5 7.5 616 7.1 101.9 3.2 1.3 120874 95 WIJ 24 7.4 139.9 45.88 16 20.5 43.1 289.3 63.5 22.1 9.7 602 10.6 89.9 3.2 4.2 33842 130 WIJ 25 8.7 157.5 45.22 15 23.5 35.1 97.9 50.5 17.2 8.6 670 10.1 81.9 2.7 3.6 37943 248 WIJ 26 7.1 185.7 41.15 15 18.4 34 94.2 46 12.4 7.7 684 10.7 87.8 2.9 2.9 31212 63 WIJ 27 9.3 162.9 45.98 20 16.1 46.2 12288.1 2793 40.3 17.2 564 17.3 90 3.5 1.8 1600 157 WIJ 28 8.4 150 40.79 18 14.8 64.6 20931.5 11130.1 68.3 12.9 593 13.4 92.4 3.4 1.8 6040 265 15816 WIJ 29 6.1 183.4 19.29 14 23 86.6 111970.5 1077.6 159.6 7.2 805 5.5 57.2 3.1 2.2 10200 WIJ 30 5.7 165.5 29.89 21 8.7 53.2 21945.1 8389.6 47.7 7.3 502 7.7 103.3 3 1.4 822 72 WIJ 31 8.1 175.3 41.19 18 22.3 45.1 11701 8078.2 134.4 17.2 544 15.6 86.9 3.5 1.3 984 142 WIJ 32 6.8 120.9 19.43 16 6.3 17.5 124.6 39.6 23.4 11.4 325 6.9 77 2.6 0.6 39588 66 WIJ 33 8.2 77.1 17.88 18 5.9 24.2 4382.9 1309.7 97.2 11.4 204 6 76.7 2.6 0.5 67882 36 WIJ 34 34.9 21.4 79.96 6 88.4 30.5 2231.8 177.9 51.9 79.4 49 26 187 9.3 1.3 1548 74 WIJ 35 22.9 70.6 10.21 39 55.5 24.6 570.1 33 4.8 13.8 356 4.3 58.8 1.5 1.7 18 0 WIJ 36 6.8 182.4 43.01 17 7.9 23.1 66.5 31.4 8.3 7.1 800 9.5 86.5 3.2 5.5 7 170 WIJ 37 20.4 68.3 10.25 21 3.9 13 22.1 152.4 1.4 13.2 370 3.1 32.5 1.2 2 1 0 WIJ 38 5.9 159.2 16.92 13 3.6 5.6 16.6 11.8 2.9 6.1 536 6.8 73.7 2.5 0.7 2 51 WIJ 39 6.2 149.3 36.77 19 125.2 248.9 12078.9 6260.3 277.6 8.7 654 7.8 89.1 3.5 6 21208 298 WIJ 40 4.9 202.6 17.46 13 3.9 7.7 80 27 25 11.5 419 5.7 77.7 2.4 0.8 483 10121 WIJ 41 10.7 164.8 34.66 25 20 16.6 1208.3 126.5 19.2 9.7 466 8.4 106.8 3.1 2.2 513 1650 WIJ 42 3.5 48.3 6.99 11 3 22 21236.9 1192.3 37.7 7.1 480 7.1 40.7 1.6 0.1 5913 50 7.6 274.8 29.4 20.4 10.7 450 6.5 66.9 2.1 1.7 51 3575 Wijk bij Duurstede (Dorestad) glass samples LM 25 5.9 203.3 18 13 6.1 LM 26 17.1 165 23 18 28.6 12.7 1143.8 107.5 11.4 20.6 467 7.3 75.3 2.5 2.1 295 1785 LM 27 12.8 144.2 19 14 13 25.7 33870.8 7164.7 202.2 17.6 408 6.2 66.1 2.2 1.8 558 2057 1734 LM 28 8.4 149.2 19 15 8.7 8.5 412.8 97.5 12.4 10.4 425 6.3 70.1 2.2 1.4 96 LM 29 12.5 124.9 20 17 16.9 12.4 2112 104.1 12.7 14.1 376 6.5 86.3 2.6 1.6 275 975 LM 30 18.7 130.4 17 15 580.5 25.6 1833.1 883.4 24.2 18.7 430 6.1 73.2 2.2 2.5 157 5703 LM 31 13.6 148 19 14 28 28.2 55184.7 15768.8 605.7 28.3 398 6.9 58.2 2.2 1.6 333 1859 LM 33 36 136.8 20 17 20.2 21.7 25129.6 2969.8 111.6 17.2 318 6.1 79.8 2.8 1.3 1994 1130 DOR 53 24 159.1 17.7 13.82 13 9.9 1708.3 90.6 14.1 14.1 469 6.6 59.5 2 1.7 230 2737 DOR 61 9.6 466.3 2.48 20.05 27.1 2.7 60.8 3693.2 58 70.3 25 3.6 138.9 1.2 0.7 18 495 DOR 66 15.2 145.1 18.73 15.28 22.3 11 1308.1 82.8 13 13 461 7 63.4 2.2 1.9 906 2122 DOR 90 16.8 153.8 22 17 17.4 13.6 3662.3 378.5 25 17.9 434 6.8 77.6 2.4 2 417 1352 DOR 91 5.6 199.8 17 13 6.6 8.6 199.6 30.3 22.6 10.1 450 6.3 64.9 2.1 1.8 57 3610 DOR 95 51 85.8 88.14 83.52 24.5 69 26.5 81.6 6.8 142.7 189 28.1 180.5 15.5 1.3 3 0 DOR 100 28.3 159.2 21 16 27.1 14.8 1789.2 115.7 13.7 18.3 463 7.5 76.6 2.6 1.9 977 2214 DOR 101 10.5 185.5 21 14 29.4 13.5 3542.6 101.6 28.6 15.3 470 6.6 65.2 2.4 1.8 579 4680 DOR 102 17.6 171.8 20 14 33.1 14.7 1332.6 92.1 18.1 16.6 465 6.7 62.4 2.3 1.9 674 2941 DOR 103 14.6 230.2 13 10 8.3 12 69.1 289 7.9 144.6 519 7.6 77.3 3.5 0.9 286 18 DOR 104 16.5 154.1 23 15 18.9 14 1765 691.8 14 16.7 482 7.4 68.2 2.6 2 495 1804 DOR 105 15.8 138.3 19 15 9.5 15.6 11995.6 3754 52.1 14.6 412 6.6 65.1 2.3 1.5 1614 1542 DOR 106 17.3 156.5 16 12 152 14.5 1005.1 62.5 9.1 11.5 482 6.9 51.7 1.9 2.6 1261 1451 175 — Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U 0.38 197 12.5 20.1 3 11.5 2.3 0.6 2.5 0.4 2.5 0.5 1.4 0.2 1.4 0.2 2 0.2 76928 1.6 1.3 0.23 247 13.6 16.3 3.2 13.3 2.6 0.7 2.9 0.4 2.3 0.5 1.4 0.2 1.3 0.2 2.3 0.2 30393 1.6 1.3 0.23 245 13.4 15 2.9 12.3 2.6 0.6 2.5 0.4 2.3 0.4 1.5 0.2 1.3 0.2 2.2 0.2 18301 1.5 1.3 0.18 215 7 11 1.7 6.8 1.6 0.3 1.1 0.2 1.2 0.2 0.7 0.1 0.6 0.1 1.4 0.1 17846 1 1.1 0.19 299 7.5 13.5 1.8 6.8 1.1 0.4 1.1 0.2 1.2 0.3 0.7 0.1 0.7 0.1 2.3 0.2 15019 1.4 1.3 0.38 306 11.4 14.6 2.4 9.6 1.9 0.5 2.2 0.3 1.8 0.4 1 0.1 1.2 0.2 2.3 0.2 269522 1.6 1.3 0.2 279 11 14 2.3 10.2 1.7 0.5 2.1 0.3 1.8 0.4 1.1 0.1 0.8 0.1 2.1 0.2 51806 1.3 1.3 0.11 288 11.3 14.7 2.5 10.9 2.2 0.6 2.2 0.3 1.9 0.4 1.2 0.2 1.1 0.1 2.1 0.2 38168 1.5 1.6 1.6 0.8 272 15.3 24.8 3.7 15.2 3.2 0.9 3.5 0.5 2.9 0.6 1.7 0.2 1.5 0.2 2.2 0.2 45749 2.3 0.43 296 11.7 18.3 2.7 11.8 2.6 0.6 2.6 0.4 2.4 0.4 1.3 0.2 1.1 0.2 2.4 0.2 58770 1.9 1.5 0.07 230 6.1 11.8 1.4 6.5 1.2 0.3 1.1 0.2 1.1 0.2 0.5 0.1 0.6 0.1 1.5 0.2 16703 1.4 0.7 0.17 325 7.2 12.9 1.7 7 1.3 0.3 1.3 0.2 1.3 0.3 0.8 0.1 0.9 0.1 2.7 0.2 43264 1.4 1.4 0.76 253 14 23.2 3.4 14.4 3.2 0.6 3.1 0.5 2.7 0.5 1.5 0.2 1.4 0.2 2.1 0.2 46027 2.1 1.5 0.52 143 7 12.9 1.6 7.4 1.5 0.4 1.4 0.2 1.3 0.3 0.7 0.1 0.7 0.1 2 0.2 416164 1.6 1 0.78 110 6.4 12.3 1.6 6.1 1.6 0.3 1.3 0.2 1 0.2 0.6 0.1 0.6 0.1 2.1 0.2 653466 1.6 0.8 5.6 1561 41.8 234.4 10.9 39.7 7.7 1.9 6.9 1 5.7 1 2.8 0.4 2.3 0.5 4.4 0.8 429979 12 3.3 0.13 125 4.5 7.7 1.1 3.7 0.4 0.1 0.7 0.1 0.7 0.1 0.5 0.1 0.4 0.1 1.3 0.1 344 1.2 0.5 0.1 263 9.4 14.1 2.2 9 1.9 0.5 2.1 0.3 1.7 0.4 1.2 0.1 0.8 0.1 2.2 0.2 74 1.5 1.2 0.23 121 3.2 6 0.7 3 0.4 0.1 0.7 0.1 0.4 0.1 0.4 0 0.4 0 1 0.1 2 0.9 0.6 0.03 186 6.9 11.7 1.5 7 1.2 0.3 1.2 0.2 1.1 0.2 0.6 0.1 0.6 0.1 1.8 0.1 23 1.2 1.2 0.14 325 7.6 13.8 2 7.1 1.4 0.5 1.3 0.2 1.6 0.2 0.7 0.1 0.7 0.1 2.1 0.3 71968 1.4 1.1 0.25 159 6.4 11.7 1.4 6.4 1.3 0.3 1 0.1 1.1 0.2 0.6 0.1 0.6 0.1 1.9 0.1 71267 1.5 1.1 0.22 362 8.5 14 2 8.3 1.8 0.4 1.5 0.2 1.3 0.3 1 0.1 0.7 0.1 2.4 0.2 4933 1.3 1.1 0.11 232 6.3 13.2 1.5 6.3 1.3 0.4 1.3 0.1 1.1 0.3 0.7 0.1 0.5 0.1 1.1 0.1 17115 0.8 0.6 0.16 264 6.8 11.8 1.5 6.2 1.2 0.3 1 0.2 1.1 0.2 0.7 0.1 0.5 0.1 1.7 0.1 1033 1.2 1.1 0.45 384 8.1 14.3 1.8 7.4 1.5 0.4 1.2 0.2 1.2 0.2 0.7 0.1 0.7 0.1 2 0.2 2119 1.3 1.1 0.39 284 6.5 12 1.6 6.3 1.2 0.4 1.1 0.2 1.1 0.2 0.6 0.1 0.5 0.1 1.8 0.1 3722 1.1 1 0.21 268 6.3 11.6 1.5 6.2 1.1 0.3 1.2 0.2 1.1 0.2 0.6 0.1 0.6 0.1 1.8 0.2 959 1.1 1 0.27 274 6.9 13 1.7 6.4 1.6 0.4 1.2 0.2 1.1 0.2 0.6 0.1 0.6 0.1 2.2 0.1 2182 1.3 1 0.31 237 6.7 12.4 1.5 6.1 1.3 0.3 1.2 0.2 1.2 0.2 0.6 0.1 0.6 0.1 1.6 0.1 1685 1.2 1 0.61 257 7.2 13.3 1.6 6.9 1.3 0.4 1.4 0.2 1.2 0.2 0.6 0.1 0.6 0.1 1.7 0.1 4782 1.3 1 2 260 6.7 13.2 1.6 6.3 1.3 0.4 1 0.2 1.1 0.2 0.7 0.1 0.5 0.1 2 0.2 4907 1.3 1.1 0.35 300 6.6 11.9 1.6 6.1 1.3 0.3 1.3 0.2 1.2 0.2 0.7 0.1 0.6 0.1 1.6 0.1 1775 1.2 1 4.6 0.29 114 3.4 6.9 0.8 2.8 0.5 0.1 0.6 0.1 0.5 0.1 0.4 0.1 0.4 0.1 3.8 0.1 5355 1.8 0.27 289 6.9 12.5 1.7 6.5 1.1 0.4 1.2 0.2 1.2 0.2 0.7 0.1 0.8 0.1 1.6 0.1 7806 1.3 1 0.31 336 7.3 13.4 1.7 6.8 1.2 0.4 1.4 0.2 1.1 0.2 0.7 0.1 0.6 0.1 2 0.2 2824 1.3 1 0.2 267 6.8 11.9 1.5 6 1.2 0.3 1.1 0.2 1 0.2 0.6 0.1 0.6 0.1 1.5 0.1 1130 1.1 1 10.25 603 34.5 69.7 8.3 30.7 6.3 1.5 5.8 0.8 5.1 1 2.7 0.4 2.6 0.4 4.9 1.1 19 11.2 2.8 0.4 328 7.6 14 1.8 7.4 1.5 0.4 1.3 0.2 1.3 0.2 0.8 0.1 0.7 0.1 2.1 0.1 6199 1.4 1.1 0.47 337 7.7 14.6 1.7 7.4 1.5 0.4 1.6 0.2 1.3 0.2 0.7 0.1 0.6 0.1 1.5 0.1 6311 1.4 1 0.35 305 7.1 12.9 1.6 7.2 1.4 0.4 1.2 0.2 1.1 0.2 0.7 0.1 0.5 0.1 1.6 0.1 5839 1.3 1 3.54 1506 71.5 93.2 8.8 27.5 2.6 0.5 2.2 0.3 1.4 0.2 0.7 0.1 0.6 0.1 2.2 0.3 1192 2.8 1 1.1 0.37 324 7.8 13.9 1.8 7.6 1.3 0.4 1.3 0.2 1.1 0.2 0.8 0.1 0.7 0.1 1.8 0.1 3801 1.4 0.27 310 6.7 12.9 1.7 6.6 1.4 0.3 1.2 0.2 1.2 0.3 0.7 0.1 0.6 0.1 1.7 0.2 11967 1.3 1 0.2 266 6.7 12.2 1.6 6.2 1.3 0.4 1.3 0.2 1.2 0.2 0.6 0.1 0.6 0.1 1.3 0.1 9092 1 1.1 176 — Appendix III trace element chemical compositions of samples analysed by LA-ICP-MS Trace elements Li B V Cr Co Ni Cu Zn As Rb Sr Y Zr Nb Mo Sn Sb DOR 107 15.3 149.8 20 15 24.6 12.4 2092 120 18.2 15.9 451 7 67 2.5 1.7 752 DOR 108 12.6 164.1 20 14 12.2 10.5 345.4 66.8 9.2 10.4 468 6.7 62.1 2.2 2.2 50 2788 1336 DOR 109 21.5 146.5 17 14 15.5 10.4 669.5 76.9 9.3 12.9 463 6.9 56.5 2.1 1.5 271 1426 DOR 110 17.6 144.5 17.71 14.35 15.8 10.9 851.1 83.4 9.8 14.4 458 6.6 54.6 2.2 1.5 531 1796 DOR 111 14.7 139.6 23.41 18.74 36 39.8 4143.2 668.7 68 16.8 472 7.7 77.9 3.5 1.7 2441 1804 DOR 112 16.4 152.1 19.05 15.41 23.3 12.6 1613.8 111.5 15.6 13.5 464 7.1 66.6 2.3 1.5 732 2187 DOR 113 17.6 152.4 21.56 17.15 18.3 13.8 881 92 15.3 15.5 482 7.5 72 2.3 2 512 2115 DOR 115 39.9 152.2 23.55 19.11 9.1 12.3 634.5 126.6 9.6 24.9 479 7.2 85.6 2.6 2 306 1313 DOR 116 21.7 144.7 25.15 20.71 18.6 13 861.5 61.8 16 11.7 475 7.5 79.3 2.5 2.3 495 1313 DOR 117 30.1 154.2 21.16 18.55 28.7 13.1 1149.1 122.7 16.1 17.7 478 7.8 74.4 2.3 2 430 1973 DOR 118 9.6 150.4 15.37 14.09 9.3 8.5 460.5 46.6 10.2 13 446 6.4 56.2 1.9 1.4 116 1523 DOR 119 15.4 141.8 18.77 14.63 31.9 12.5 1215.8 100.6 17.5 13.9 455 6.7 61.6 2.2 1.8 746 2566 DOR 120 13 136.6 22.1 18.95 15.5 10.8 1130.2 202.8 17.8 17.1 410 6.9 82 2.3 1.7 604 1082 DOR 121 14.3 161 22.41 16.88 13.6 11.2 580.3 57.3 13.3 11.3 464 7 73.3 2.4 2.4 223 1934 DOR 122 17.9 99.9 20.57 24.32 6.7 10 273.7 97 4 8.8 267 6.8 153.2 3.5 0.8 28 285 DOR 123a 6.1 146.4 27.59 22.58 15.1 13.1 608.7 47.5 13.2 8.3 474 7.7 90 2.6 2.7 136 1456 DOR 123b 8.7 153.3 26.53 21.88 8.7 12.2 564.6 42.7 10.8 11.8 494 7.8 86.8 2.4 2.8 46 1200 DOR 124 7.3 141.2 20.48 13.55 7.4 9.7 240.7 41.2 8.1 13.8 436 7.4 52.4 1.9 2.2 34 1128 DOR 125 5.4 173.6 15.38 11.48 5 7.5 225.5 30.1 24 10 464 6.3 57.4 1.8 0.9 48 2856 DOR 126 4.8 196.4 17.37 13.52 5.6 7.5 216.6 29.9 22.4 12.4 422 6.5 69.1 2.1 1.6 51 4506 DOR 127 7.9 155.6 18.86 12.44 5.7 8.6 105.5 28.2 12.4 8.4 468 6.7 48.4 1.7 1.5 31 1853 DOR 128 9.7 176.7 18.05 11.65 5.8 7.9 128.5 28.6 15.4 8.5 443 6.4 52.8 1.6 1.7 40 2250 DOR 129 14.6 176.9 18.23 14.19 23.9 14 2142 130.5 20.6 14.4 429 6.4 60 2.2 1.6 903 2840 DOR 130a 17.7 192.8 16.53 12.69 7.9 8 417.3 48.5 12.8 15.4 372 6.4 61.5 1.9 1 115 2686 DOR 130b 16.3 191.7 20.18 15.53 305.1 28.4 1656 114.4 26.8 16.7 404 6.4 59.5 3.1 2.9 243 6909 DOR 131 8.1 198.5 11.05 10.1 7.2 5.8 673.5 44.2 23.6 8.5 398 5.7 46.5 1.7 0.5 164 3303 DOR 132 30.4 145.6 18.79 14.78 23.4 13.5 1045 104.5 9.8 18.6 441 8.2 66.9 2.4 1.6 1399 1418 DOR 133 13.9 169.3 18.33 13.75 19.8 13.4 1909.5 130.6 21.6 14.4 447 6.9 61.9 2 1.5 819 2486 DOR 134 6.2 191.7 19.53 13.34 22.8 19.4 6169.7 271 47 13.8 443 6.4 60.8 2.1 1.8 885 4673 DOR 135 14.7 167.2 20.54 16.2 18.2 12.4 1454 122.4 13.6 15.4 450 7.2 69.8 2.3 1.6 496 1658 DOR 136 8.9 350.6 6.87 8.9 8.5 5.6 45 240.3 16 132.1 443 6.5 69.6 2.6 0.4 2 16 DOR 137 16.4 165.9 19.38 14.69 21.3 11.6 1573.3 98.1 16.8 16.7 469 6.6 65.4 2.2 1.7 508 2935 DOR 138 15.1 161.9 19.26 15.45 22.7 13.7 2174.1 115.8 16.1 14.1 460 7.1 65.7 2.3 1.6 997 2447 DOR 139 29.6 168.7 19.4 15.6 5.9 9.2 372.2 84.3 13.1 19.1 395 7.4 72.5 2.3 1.2 220 2481 DOR 141 14.6 148.2 19.37 15.81 27.7 13.8 1490.6 108.9 16.6 15.9 471 6.9 64.4 2.3 1.6 1045 2746 DOR 142 12.5 166.1 18.76 14.43 31.9 12.2 4528.1 160.9 30.1 13.5 421 6.1 62.5 2.1 1.4 1076 5916 DOR 143 4.6 167.3 19.55 16.85 370.2 68.4 2440.5 99.4 27.5 9 451 6.6 70.8 2.2 5 276 4311 DOR 144 4.1 153.3 16.09 10.52 831.3 31.3 954 27.3 8.7 7.1 373 5.6 42.7 1.5 2 8 4735 DOR 145 4.3 160.8 15.54 11.71 336.5 20.6 1168.7 65 29.4 12.3 447 6.5 58.3 2 1.3 111 21027 DOR 146 5 139.9 9.01 8.26 11.8 8.9 24074.2 1279.3 102.4 6.4 383 5.6 35.8 1.6 0.2 960 15011 DOR 147 5.1 118 25.49 11.83 6.3 12.9 10658.5 65.9 37.9 14.2 480 6.7 50.4 1.9 1.9 105 13203 DOR 148 7 87.9 23.95 12.8 8.5 20.3 18352.8 79.7 32.5 11.2 536 7.4 44.4 1.8 3.3 772 5691 DOR 149 12.3 67.8 16.78 12.47 20.4 16.5 1582.8 81.7 20.9 13.1 305 5.4 50.6 2.3 1.2 36161 2250 DOR 150 31.7 75.9 35.26 32.64 8.3 15.4 331.4 1700.2 29 132.9 823 14.9 107.7 6.2 0.8 33 4175 DOR 151 66.4 88 36.72 31.51 8.4 17.3 171.2 1933.9 25.7 128.3 1203 17.3 143.7 8.8 0.8 9 4493 177 — Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U 0.33 289 7 13.6 1.6 7 1.4 0.4 1.2 0.2 1.2 0.2 0.7 0.1 0.6 0.1 1.8 0.1 7015 1.4 0.2 322 6.9 12.4 1.6 6.3 1.3 0.4 1.2 0.2 1.1 0.2 0.6 0.1 0.6 0.1 1.6 0.1 627 1.1 1.1 1 0.29 267 6.8 13.5 1.7 6.8 1.4 0.4 1.2 0.2 1.2 0.2 0.7 0.1 0.6 0.1 1.4 0.1 1405 1.2 1 0.27 268 6.7 12 1.7 6.6 1.2 0.4 1.2 0.2 1.1 0.2 0.6 0.1 0.6 0.1 1.4 0.1 3937 1.1 1 0.57 309 7.9 14.3 1.8 6.9 1.5 0.4 1.5 0.2 1.3 0.2 0.7 0.1 0.6 0.1 1.9 0.2 27064 1.6 1.1 0.3 290 7.2 12.4 1.5 6.8 1.4 0.5 1.3 0.2 1.2 0.2 0.7 0.1 0.7 0.1 1.6 0.2 5615 1.3 1.1 0.32 320 7.7 13.2 1.8 7.4 1.6 0.4 1.3 0.2 1.3 0.3 0.8 0.1 0.7 0.1 1.9 0.2 3616 1.3 1.1 0.4 405 7.7 13.5 1.8 7.2 1.4 0.4 1.2 0.2 1.3 0.3 0.6 0.1 0.7 0.1 2 0.2 1307 1.2 1 1 0.15 358 7.2 12.9 1.7 6.9 1.4 0.4 1.3 0.2 1.1 0.3 0.7 0.1 0.8 0.1 2 0.1 4322 1.2 0.34 370 7.3 13.7 1.7 7.2 1.2 0.4 1.5 0.2 1.2 0.3 0.7 0.1 0.7 0.1 2 0.1 3351 1.4 1 0.22 269 6.4 12 1.7 6.2 1.2 0.4 1.4 0.2 1.2 0.2 0.6 0.1 0.6 0.1 1.5 0.1 754 1 0.9 0.32 290 7.2 12.9 1.7 6.5 1.5 0.4 1.4 0.2 1 0.2 0.7 0.1 0.5 0.1 1.6 0.1 5727 1.3 1.1 0.29 299 7.5 13.4 1.7 6.7 1.4 0.3 1.2 0.2 1.1 0.2 0.6 0.1 0.6 0.1 2 0.2 2914 1.2 1.1 0.27 334 7.2 12.3 1.6 6.7 1.4 0.3 1.2 0.2 1.2 0.3 0.7 0.1 0.7 0.1 1.7 0.1 2149 1.1 0.9 0.08 277 7.1 13.5 1.6 6.8 1.2 0.4 1.5 0.2 1.1 0.2 0.6 0.1 0.8 0.1 3.6 0.2 461 1.5 1.2 0.17 408 7.5 12.8 1.7 6.7 1.2 0.4 1.5 0.2 1.3 0.3 0.7 0.1 0.8 0.1 2.1 0.1 1325 1.2 1 0.39 392 7.3 13.2 1.7 6.8 1.6 0.4 1.4 0.2 1.3 0.3 0.7 0.1 0.8 0.1 2.1 0.2 473 1.2 1 0.24 271 7.2 12.7 1.7 6.8 1.3 0.4 1.3 0.2 1.4 0.2 0.7 0.1 0.7 0.1 1.4 0.1 538 1.2 1 0.16 238 6.2 11.5 1.5 6.3 1.2 0.3 1.2 0.1 1.1 0.2 0.6 0.1 0.5 0.1 1.4 0.1 680 1.1 1.1 0.24 294 6.5 12.2 1.5 6.2 1.1 0.4 1.1 0.2 1.1 0.2 0.5 0.1 0.6 0.1 1.7 0.1 1915 1.4 1.2 0.22 300 6.1 11.3 1.5 6 1.1 0.4 1.3 0.2 1.1 0.2 0.7 0.1 0.7 0.1 1.1 0.1 509 0.9 0.9 0.22 253 6.3 11.1 1.4 5.9 1.2 0.3 1.1 0.2 1 0.2 0.5 0.1 0.5 0.1 1.4 0.1 422 0.9 0.9 0.31 262 6.3 11.7 1.5 6.1 1.1 0.3 1.2 0.2 1.1 0.2 0.6 0.1 0.6 0.1 1.5 0.2 8405 1.2 0.9 0.36 221 6.8 12.4 1.5 6.4 1.3 0.3 1.2 0.2 1.1 0.2 0.7 0.1 0.6 0.1 1.6 0.1 1437 1.3 1 0.34 230 7.3 13.1 1.7 6.6 1.2 0.3 1.1 0.2 1.1 0.2 0.7 0.1 0.6 0.1 1.6 0.2 2794 1.4 1 0.15 190 5.6 9.7 1.3 5.6 1.3 0.3 0.9 0.2 0.9 0.2 0.6 0.1 0.4 0.1 1.2 0.1 1315 0.9 0.8 0.48 292 8 14.4 1.9 7.5 1.4 0.3 1.6 0.2 1.4 0.3 0.8 0.1 0.7 0.1 1.7 0.2 4020 1.4 1 0.38 335 6.4 12.1 1.5 6.6 1.5 0.3 1 0.2 1.1 0.2 0.6 0.1 0.6 0.1 1.6 0.1 9054 1.2 1 0.33 267 6.9 12.4 1.5 6.5 1.5 0.3 1.3 0.2 1.2 0.2 0.7 0.1 0.5 0.1 1.5 0.1 10898 1.6 1 0.31 324 6.9 12.4 1.6 6.5 1.1 0.4 1.3 0.2 1.3 0.3 0.6 0.1 0.6 0.1 1.8 0.1 3840 1.2 1 2.24 1099 68 83 8.1 23.2 2.1 0.4 1.7 0.2 1 0.2 0.5 0.1 0.5 0.1 1.7 0.2 701 1.5 0.6 0.34 312 7.2 13 1.7 6.8 1.1 0.3 1.4 0.2 1.1 0.2 0.6 0.1 0.5 0.1 1.8 0.2 4250 1.4 1.1 0.31 302 7.1 13.3 1.7 6.7 1.3 0.4 1.4 0.2 1.3 0.2 0.7 0.1 0.5 0.1 1.6 0.1 8335 1.4 1 0.39 240 7.5 14.4 1.9 7.2 1.7 0.4 1.4 0.2 1.1 0.3 0.7 0.1 0.7 0.1 1.8 0.1 485 1.5 1.1 0.37 304 7.4 13.1 1.8 6.8 1.3 0.4 1.3 0.2 1.2 0.2 0.6 0.1 0.5 0.1 1.6 0.1 9728 1.3 1 0.31 259 6.7 12.2 1.5 6.3 1.3 0.3 1.2 0.2 1.1 0.2 0.6 0.1 0.6 0.1 1.6 0.2 12180 1.4 1.1 0.21 279 6.5 12 1.6 6.7 1.4 0.4 1.3 0.2 1.1 0.2 0.7 0.1 0.6 0.1 1.7 0.2 2844 1.1 1 0.1 212 5.4 10 1.2 5.5 1.2 0.4 1 0.1 1 0.2 0.6 0.1 0.5 0.1 0.9 0.1 177 0.9 1 0.32 223 7.1 13.3 1.7 6.6 1.4 0.4 1.4 0.2 1.1 0.2 0.5 0.1 0.6 0.1 1.3 0.1 5671 1.6 1.1 0.1 174 6.4 9.9 1.4 5.9 1.1 0.3 1.1 0.1 0.9 0.1 0.5 0.1 0.5 0.1 1 0.1 1335 1.1 0.9 0.29 272 7.5 13.3 1.7 6.5 1.5 0.4 1 0.2 1.2 0.2 0.6 0.1 0.6 0.1 1.4 0.1 2305 1.6 1.1 0.36 309 7.9 13.6 1.7 7.7 1.5 0.4 1.5 0.2 1.2 0.2 0.9 0.1 0.7 0.1 1.2 0.1 121308 1.3 1.1 0.42 206 5.9 10.3 1.3 4.9 1.1 0.3 1.2 0.2 0.9 0.2 0.6 0.1 0.5 0.1 1.3 0.2 488078 1.3 0.9 8.43 25021 25.9 41.3 5.4 20.8 3.7 1.2 2.9 0.4 2.4 0.5 1.4 0.2 1.3 0.2 2.7 0.6 263859 6 6 8.58 22165 34.2 58.8 7.1 26.5 5 1.6 4.2 0.5 3.1 0.6 1.6 0.2 1.5 0.2 4.4 1 288731 12.8 7.4 178 — Appendix III trace element chemical compositions of samples analysed by LA-ICP-MS Trace elements Li B V Cr Co Ni Cu Zn As Rb Sr Y Zr Nb Mo Sn Sb Deventer glass samples DEV 1 52.5 415.5 59.57 79.22 12.8 41.8 29.8 68.5 18.6 190.5 402 18.4 180 12.5 1.6 2 1 DEV 2 5.9 114 22.93 14.77 411.8 44.1 724.6 26.7 5.1 7 580 7.1 76.3 2.5 3.7 17 5 DEV 3 10.3 155 17.24 12.94 9.4 8.3 844.9 294.1 20.6 15.2 429 6.5 62.4 2 1.6 127 1669 DEV 4 10.5 43.3 12.72 11.98 19.3 19.7 85.8 76.1 29.1 31.3 382 3.3 56.2 2.5 1.7 1 1050 DEV 5 17.7 118.7 11.71 6.42 14.5 27.8 303.8 127 63.5 65.1 466 4.9 33.4 1.3 2.3 13 2 DEV 6 14.5 262.3 15.7 17.37 6.9 9.8 45.8 329.1 2.8 179.1 410 10.4 140.7 4.8 0.3 1 4 DEV 7 12.9 226.4 19.46 24.05 5.3 11.6 44.5 260.2 3 191.7 377 11.4 192 6 0.2 1 3 DEV 8 13.8 215.6 7.21 8.92 6 8.1 56.7 332.9 1.1 119.5 682 3.9 71.3 2.7 1.7 1 1 DEV 9 6.9 44.4 5.55 168.08 2.9 4.5 8.9 27.9 24.6 10.9 182 2.6 107.8 1.5 0.2 6 3 DEV 10 31.4 137.6 19.35 19.95 12.1 11.1 1017.9 108.2 5.5 23.5 335 6.5 113.8 2.8 1.1 323 759 DEV 11 6.2 89.5 21 24.08 10.1 8 487.9 82.6 6.4 7.7 253 6.9 158.4 3.8 0.4 121 117 DEV 12 11.4 155.8 21.24 17.56 8.7 9.7 508.5 45.9 12.7 13 429 6.9 73.8 2.4 2 74 1314 DEV 13 49.9 174.2 21.57 14.67 8.5 31.8 112.3 411.4 74.2 295 916 10.6 149.5 4.4 1.7 90 45 DEV 14 28.3 190.4 19.9 18.97 7.3 20.8 412.1 161.3 2 206.3 269 11.1 189.7 5.4 0.5 29 0 DEV 15 18.1 123.5 17.77 26.31 4.1 24.8 121.7 442.1 0 174 281 9 145.8 4.6 3.4 10 0 DEV 16 weathered 0 118.2 5.27 20.08 6.5 23.7 382.9 274.7 6.5 65.9 113 9.8 143 3.7 0.3 35 1 DEV 17 weathered - - - - - - - - - - - - - - - - - DEV 18 52.3 169.2 16.22 17.27 10.9 18.2 1619.7 191.8 6.3 124.5 416 8.9 122.9 3.9 2 261 595 DEV 19 20.5 143.2 22.87 19.78 16.2 13.5 677.5 72.9 11.1 14.5 427 7.1 91.9 2.8 2.4 389 1091 DEV 20 12.2 182.1 33.84 14.79 5.2 15.1 46.2 28.7 5.5 6 456 7.8 64.6 2.1 2.1 2 38 DEV 21 10.9 191.1 19.34 27.27 6.1 17.2 57 257.7 1.8 96.1 606 9.8 192.5 6.9 1.8 1 0 DEV 22 14 58.3 7.32 10.66 7.8 7.4 40.8 68.8 14.4 34.1 371 3.6 54.6 2.3 1 1 194 DEV 23 weathered 13.7 46.1 7.96 8.16 4.5 4.4 34 43.1 5.1 16.3 102 3.1 88.5 5.7 0.5 2 34 DEV 24 24.3 245.4 10.25 10.85 42.5 46.2 87.1 462.9 2.2 38.9 782 4.9 70.4 3.7 0.6 1 0 DEV 25 10.8 63.1 8.6 7.29 3.2 6.9 95.3 87.2 6.9 22.2 260 3.1 66.7 2.3 1.1 3 121 DEV 26 35.4 120.1 6.71 6.54 1.5 9.5 21901.8 169 7 362.4 455 2.7 70.4 1.4 1 55 2 DEV 27 15.3 178.3 22.89 26.95 7.1 37.6 75.2 244.8 1.7 233.4 373 11.8 190.9 6.1 4.1 5 1 DEV 28 13 260.9 15.65 17.05 7 11.1 45.8 324 2 181.2 410 9.5 124.6 4.3 0.2 1 0 0 DEV 29 28 172.3 17.78 21.22 7.7 19.6 59.2 295.6 1.1 109.5 892 7.5 142.2 5.5 1.7 1 DEV 30a 16.2 173 8.8 6.08 49.1 38.6 85.4 321 48.3 354.7 1058 6.6 161.6 3.3 0.5 7 1 DEV 30b 0.2 18.2 15.31 9.69 40.7 40.3 115.8 716.8 41.7 112.9 240 11.3 273.5 5.3 2.3 11 2 DEV 31 50.1 92.2 19.82 25.1 7.7 9 372.5 48.9 2.3 10.6 238 6.4 155.6 3.6 0.3 111 130 DEV 32 9.3 228.3 6.21 10.14 4.7 16.4 85.7 206.8 0.6 190.3 360 3.5 89.6 2 2.9 11 0 DEV 33 8.1 251.9 7.93 13.57 5.1 14.8 69.4 164.4 0.8 170.6 218 3.5 65.3 1.7 2.5 2 0 29.4 1951.6 14.89 2391.84 0 14.9 7.4 1669.2 89.8 1.3 1 0 0 0 0 126 36 DEV 35 15.8 263.3 4.19 6.02 3.3 10.6 59.5 268.5 1.4 431.8 993 2.8 71.7 1.1 0.8 1 0 DEV 36 43.7 166.1 7.79 10.58 3.3 12 780.8 213.6 2.3 331.5 417 3.2 76.4 1.7 2.7 37 1 DEV 37 11.7 107.1 9.99 9.01 2.7 7.7 25.5 76.5 4.7 10.2 264 5.7 129.9 2.5 0.9 5 3 DEV 34 weathered DEV 38 19.9 160.3 5.31 5.98 1 4.6 158.3 187.8 2.1 306.4 583 1.8 56.7 1.3 0.9 5 1 DEV 39 22.8 168.4 17.97 19.81 5.7 14.7 49.2 196.2 1.4 167.7 339 9.3 127.6 5 8.2 1 0 DEV 40a 19 87.6 18.93 23.8 8.1 11.5 21560.9 80.5 5.3 12.6 235 6.1 145.3 3.4 0.5 156 175 DEV 40b 10 89.1 20.87 25.05 8.9 8.4 837.6 71.3 5.1 11.2 269 6.8 157.3 3.7 0.3 258 137 12.4 127.1 15.87 15.39 12.7 15.4 21451.4 113 35.3 36.3 366 6.6 96 2.8 1.1 201 928 DEV 41 179 — Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U 2.5 370 27.2 67.8 6.2 23.6 4.5 1.1 4.8 0.6 3.6 0.8 2 0.3 1.8 0.3 5.2 0.7 12 10.9 2.6 0.09 280 6.8 11.6 1.6 6.3 1.4 0.4 1.2 0.2 1 0.2 0.6 0.1 0.7 0.1 1.9 0.2 442 1.1 1.1 0.22 287 7.6 12.6 1.7 6.7 1.4 0.3 1.3 0.2 1.2 0.2 0.6 0.1 0.6 0.1 1.5 0.1 1323 1.2 1 0.38 960 5.1 9.6 1.1 4.2 0.8 0.3 0.8 0.1 0.6 0.1 0.4 0.1 0.3 0.1 1.4 0.2 31684 0.9 3 0.39 582 4.8 6.6 1 4.3 1.2 0.2 0.9 0.1 0.8 0.1 0.5 0.1 0.4 0.1 0.9 0.1 70 1.1 2 1.24 958 45.1 58.7 6.5 20.6 3 0.5 2 0.3 1.8 0.4 1.1 0.1 1.1 0.1 3.7 0.3 5 3 0.9 1.22 902 38.6 50.8 5.5 18.9 2.7 0.5 2.5 0.4 2.1 0.5 1.3 0.2 1.2 0.2 5.4 0.4 8 4 1.1 0.99 1647 10.4 16.9 1.6 5.3 0.7 0.2 0.6 0.1 0.6 0.1 0.5 0.1 0.4 0.1 1.9 0.2 3 1.1 0.4 0.22 251 3.9 10.8 0.7 2.6 0.4 0.1 0.5 0.1 0.3 0.1 0.3 0 0.2 0.1 2.8 0.1 60 0.8 0.6 0.53 341 10.3 16.3 1.8 7.9 1.5 0.3 1.3 0.2 1.1 0.2 0.8 0.1 0.6 0.1 2.9 0.2 1349 1.6 1 0.23 182 7.4 13.5 1.7 6.5 1.3 0.4 1.1 0.2 1.1 0.3 0.8 0.1 0.7 0.1 3.8 0.3 842 1.5 1.1 0.19 392 7.6 12.7 1.7 6.7 1.4 0.3 1.1 0.2 1.2 0.3 0.6 0.1 0.5 0.1 1.7 0.1 650 1.2 1 1.82 4829 12.2 24.9 2.9 10.9 2.2 0.8 1.8 0.3 1.6 0.4 1.1 0.2 1.1 0.1 3.9 0.3 394 4.5 2.5 1.1 1.2 1075 39.5 54.3 5.4 17.5 2.7 0.4 2.1 0.3 1.8 0.4 1.3 0.2 1.3 0.1 4.8 0.4 16 4.2 1.93 1082 13.8 24.1 2.9 10.6 1.7 0.5 1.9 0.2 1.7 0.3 1 0.2 1 0.1 3.7 0.4 88 3.5 1 0.45 359 16 30.7 3.6 12.6 2.2 0.6 1.9 0.3 2 0.5 1.5 0.1 0.8 0.1 3.4 0.4 29 5.3 0.6 - - - - - - - - - - - - - - - - - - - - 0.9 1.19 863 29.7 34.8 4.5 14.8 2 0.4 2 0.3 1.5 0.3 1 0.1 0.9 0.1 3.3 0.3 984 2.4 0.26 428 7.6 13.2 1.8 6.7 1.3 0.4 1.5 0.2 1.3 0.3 0.8 0.1 0.7 0.1 2.3 0.2 2035 1.3 1 0.06 235 7.6 11.5 1.7 7.1 1.4 0.3 1.5 0.2 1.4 0.3 0.8 0.1 0.8 0.1 1.5 0.1 37 1.1 0.9 1.6 2243 20.7 36.2 3.7 12.8 1.8 0.5 1.7 0.2 1.6 0.4 1 0.2 1.1 0.2 4.9 0.5 6 3.2 1.1 0.33 419 4.4 7.7 0.9 3.9 0.8 0.2 0.7 0.1 0.5 0.1 0.3 0 0.4 0.1 1.3 0.1 28453 0.9 0.9 0.26 190 6.1 11.3 1.3 4.7 0.8 0.2 0.6 0.1 0.6 0.1 0.4 0 0.4 0.1 2 0.3 70 1.6 0.8 0.34 3273 10.2 14.8 1.6 5.7 1 0.4 0.9 0.2 0.8 0.2 0.4 0.1 0.4 0 1.9 0.2 4 1.8 0.6 0.52 321 4.4 7.5 0.9 3.5 0.6 0.2 0.5 0.1 0.4 0.1 0.3 0.1 0.3 0 1.8 0.2 138 1.3 1 0.3 3.55 1273 3.6 7 0.8 2.8 0.5 0.2 0.4 0.1 0.4 0.1 0.3 0 0.3 0 1.7 0.1 521 1 1.55 1243 36.2 39.5 5.8 19 2.6 0.6 2 0.3 2.1 0.5 1.2 0.2 1.2 0.2 5.1 0.4 26 3.4 1.1 1.22 939 45.7 57.7 6.5 22 2.8 0.6 2.5 0.3 1.5 0.3 1 0.1 0.8 0.2 3.3 0.3 4 2.7 0.9 0.8 1.28 1761 14.3 26.2 2.7 8.7 1.3 0.4 1.2 0.2 1.2 0.3 0.8 0.1 0.8 0.2 3.9 0.3 4 2.4 1.52 5423 12.2 32.3 2.8 11.5 2.2 0.8 1.6 0.2 1.2 0.3 0.7 0.1 0.6 0.1 4.3 0.2 38 4.6 1.8 0.48 3473 20.8 55.4 4.8 17.6 3.1 0.9 3.2 0.4 2.1 0.4 1 0.2 1.1 0.1 7.9 0.4 68 7.4 2.8 0.22 189 6.6 13.1 1.5 6.6 1.3 0.4 1.4 0.1 1 0.2 0.7 0.1 0.6 0.1 3.8 0.2 795 1.4 1.1 1 1124 21.4 17.9 2.3 7.5 0.8 0.2 0.6 0.1 0.6 0.1 0.3 0.1 0.3 0.1 2.5 0.1 3 1.1 0.3 0.65 663 60.5 40.1 6.6 20.6 1.7 0.2 1.1 0.1 0.6 0.1 0.3 0 0.3 0.1 1.5 0.1 3 1 0.3 0 0 0 0.4 0 0 0 0 0 0 0 0 0 0 0 0 0 0.5 3 0.8 0 4.31 2718 6.2 9.5 1 3.5 0.4 0.3 0.4 0.1 0.5 0.1 0.3 0 0.3 0 1.9 0.1 4 0.9 0.3 2.75 1332 14.5 15.4 1.9 7.2 0.6 0.2 0.7 0.1 0.6 0.1 0.4 0.1 0.4 0.1 2.1 0.1 155 1.2 0.4 0.16 106 9.3 16.5 2.1 7.3 1.6 0.3 1.4 0.2 1.1 0.2 0.6 0.1 0.5 0.1 3.2 0.2 58 2.4 0.7 3.31 1350 3.2 6.1 0.6 2.1 0.4 0.2 0.3 0.1 0.3 0.1 0.2 0 0.3 0 1.3 0.1 50 0.8 0.3 1.16 605 24.5 44.7 4.3 15.1 2.2 0.4 2 0.3 1.5 0.3 0.9 0.1 0.9 0.1 3.2 0.3 3 2.6 0.8 0.29 186 7 13.2 1.6 6.2 1.2 0.3 1.3 0.2 1.2 0.2 0.7 0.1 0.7 0.1 3.4 0.3 931 1.4 1.1 0.23 203 7.2 13.9 1.6 6.5 1.3 0.3 1.3 0.2 1.1 0.2 0.8 0.1 0.6 0.1 3.8 0.2 1681 1.6 1.2 0.34 350 9.6 16 1.9 7.2 1.3 0.3 1.2 0.2 1.2 0.2 0.7 0.1 0.7 0.1 2.4 0.2 1913 1.5 1 180 — Appendix III trace element chemical compositions of samples analysed by LA-ICP-MS Trace elements Li B V Cr Co Ni Cu Zn As Rb Sr Y Zr Nb Mo Sn Sb Susteren glass samples Sust 1 (bead body) 6.5 139 11.6 11.3 138.7 24.5 9010 1638 22.3 8.7 461 6.6 47.3 1.7 1.2 1947 3471 Sust 2 (bead body) 10 150.7 17.8 14.1 16.3 12.3 1879.3 99.3 15 12.3 417 7.2 65.1 2.2 1.8 684 1988 Sust 2 (decoration) - - - - - - - - - - - - - - - - - Sust 3 (bead body) 17.5 120.7 18.7 46.4 974 34 1180 3990 23.5 13.8 506 5.7 98.6 2.4 2.3 3223 505 Sust 3 (decoration) 14.6 147 25.2 21.5 20.1 28.3 1580.7 114.2 19.1 14.2 432 8.4 91.7 3.7 2.2 96733 3357 Sust 4 (bead body) 16.1 142.6 19.3 19.7 50 13.7 1279.7 247.6 12.4 14.7 445 7 74.1 2.2 1.7 570 1635 Sust 4 (decoration) - - - - - - - - - - - - - - - - - Sust 5 (bead body) - - - - - - - - - - - - - - - - - Sust 5 (decoration) 10 166.9 21.7 18.7 13.5 13.1 947.3 81.2 16.1 12.6 428 7.6 80.6 2.6 2.1 9195 2488 Sust 6 (bead body) 15.8 159.3 20.2 16.5 14.8 11.7 1569 94.1 13.4 14.9 471 7.4 75.6 2.4 2 625 2127 Sust 6 (decoration) - - - - - - - - - - - - - - - - - Sust 7 6.4 187.4 13.9 11.9 327 28.2 2173.3 64.8 33.4 9.2 404 6.1 57.9 1.9 2.4 148 10020 Sust 8 5.5 183.7 13.2 11.2 322.3 25.7 1887.7 57.1 32.9 7.5 399 6 52.8 1.7 2.3 131 10253 Sust 9 11.6 169.7 22.9 17.8 51.9 24 22100 339.7 28.6 17.9 406 7.2 74.9 3.4 2.2 1690 3597 Sust 10 10.6 163 17 14.2 5.1 7 117.7 45.9 13.9 11.5 432 6.6 68.5 2 1.7 162 2492 Sust 11 14.2 189.9 25.4 24 11.3 16.1 82.9 257.9 2 398.3 303 12.2 226.6 7.6 0.4 11 4 Sust 12 12.6 132.1 22.7 21.5 22.9 13.4 1867.3 168.3 13.2 18.1 411 6.5 93.1 3.1 1.6 944 1072 Sust 13 14.2 138.9 19 25.1 35.9 15.5 1888.3 127.3 15.9 15.3 451 5.9 72.7 2.2 1.8 484 2078 Sust 14 16.6 132.2 21.2 21 22.2 12.8 2057 114.4 11 18.1 386 7.4 99.2 2.9 1.7 394 1198 Sust 15 6.6 154.1 18.4 13.8 53.9 14.5 6090 118.6 29 13.4 394 6.8 60.8 2.3 1.5 1082 6810 Sust 16 22.2 138.3 22.2 25.2 24.9 16 1159.7 106 13.5 24.5 402 8.3 97 3.2 1.6 557 1682 Sust 17 14.6 171.5 15.5 12.3 419 35.4 2139 94.6 29 61.1 431 5.9 51.5 1.9 2.9 498 9703 Sust 18 17.7 141 74.4 38 14.2 21.1 4906.7 190.8 8.2 106.6 290 11.2 169.6 19.2 2.3 668 1276 Sust 19 12.1 179.2 16.5 14.2 27 13.3 2070 190 13.5 152.3 396 7.1 90.5 3.3 1.1 628 1990 Sust 20 5 156.3 19.2 12.7 13.9 11.6 2463.3 66.3 21.8 12.2 443 5.9 53.4 2 1.9 279 2423 Sust 21 15.4 127.9 22.3 21 22.9 13.8 1722.3 136.1 11.4 16.2 407 7.1 90.6 2.7 1.5 642 1221 Sust 22 9.4 161 18.6 14.8 30.1 12.2 2264 116.3 17.9 12.3 460 6.4 58.8 2.2 1.6 597 2990 Sust 23 11.5 147.2 19.3 15.3 22.6 12.7 1597.7 119.3 15 13.2 455 7.1 63.6 2.2 1.7 670 2095 Sust 24 6.5 154.5 13.3 11.4 513.3 32.2 1585.7 64.8 31.1 11.9 421 6.4 45.5 1.7 2.4 559 13767 Sust 25 30.7 160.7 40.7 41 11 16.3 111.7 44.1 5.4 15.5 495 9.4 161.4 4.4 4 61 260 Sust 26 12.1 160.5 22.8 18 20.6 14.2 2322 127 20.1 12.3 461 6.7 71.5 2.4 2.2 673 2499 Sust 27 13.7 144.1 20.4 18.8 23.7 13.5 1282.3 107.7 11 19.1 451 7.2 71.7 2.3 1.9 563 1723 Sust 28 22.1 115.6 17.3 46.4 25.7 19.4 415.3 126.1 25.4 13.9 503 4.9 92.9 2.1 1.9 252 426 Sust 29 9.3 165.4 21.5 16.4 19.3 15.8 3163.3 187.2 41.2 13.6 456 6.6 66.2 2.5 2 876 3287 2.3 151.7 14.98 14.55 246.3 80.4 904 22.3 8.5 6.3 406 7.3 72.9 2.7 3.3 404 70 - - - - - - - - - - - - - - - - - 9.8 105 12.93 863.06 0.8 13.2 3.6 11.8 15.3 12.2 58 8 115.9 1.6 0.4 2 0 Utrecht glass samples Utr 77 Utr 78 modern Utr 79 181 — Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U 0.12 224 6.5 11.3 1.4 5.8 1.3 0.3 1.1 0.2 1.1 0.2 0.7 0.1 0.6 0.1 1.2 - 6687 0.9 0.7 0.28 274 6.8 12 1.5 6.2 1.2 0.3 1.3 0.2 1.2 0.3 0.7 0.1 0.6 0.1 1.7 - 5260 1.2 0.9 - - - - - - - - - - - - - - - - - - - - - 0.35 201 7.6 14.4 1.7 6.3 1.2 0.3 1.1 0.2 1.1 0.2 0.6 0.1 0.5 0.1 2.6 - 2246 1.7 0.9 0.77 430 8.7 16 2 7.8 1.6 0.4 1.5 0.2 1.4 0.3 0.9 0.1 0.8 0.1 2.5 - 868333 1.8 1.1 0.3 293 7.3 12.8 1.6 6.4 1.3 0.3 1.3 0.2 1.1 0.2 0.7 0.1 0.6 0.1 1.9 - 3800 1.4 1 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 0.35 357 7.2 12.6 1.6 6.4 1.4 0.4 1.3 0.2 1.2 0.3 0.7 0.1 0.6 0.1 1.9 - 175333 1.3 1.1 0.31 312 - - - - - - - - - - - - - - - - - - - - - 7.3 13.3 1.7 7 1.2 0.4 1.3 0.2 1.3 0.2 0.7 0.1 0.7 0.1 1.9 - 5003 1.3 1.1 0.15 205 6.7 11 1.4 5.9 1.2 0.3 1.1 0.1 1.1 0.2 0.6 0.1 0.5 0.1 1.5 - 1862 1.1 1 0.14 194 6.5 10.9 1.4 5.4 1.1 0.3 1.1 0.2 1 0.2 0.6 0.1 0.5 0.1 1.4 - 2150 1.1 1 1.1 0.83 255 8.2 14.9 1.7 7 1.5 0.4 1.3 0.2 1.1 0.2 0.7 0.1 0.6 0.1 1.9 - 11550 1.8 0.28 252 6.4 11.5 1.5 5.8 1.2 0.3 1.3 0.2 1.1 0.2 0.6 0.1 0.6 0.1 1.7 - 614 1.1 1 1.75 1107 37.3 64.3 5.9 19 2.9 0.5 2.4 0.3 2.1 0.4 1.2 0.2 1.2 0.2 5.8 - 28 5 1.5 1.1 0.37 358 8.3 14.9 1.9 7 1.3 0.4 1.3 0.2 1.1 0.2 0.7 0.1 0.7 0.1 2.3 - 4760 1.4 0.35 291 7.1 13.3 1.6 6.4 1.2 0.3 1.1 0.2 1 0.2 0.6 0.1 0.5 0.1 1.9 - 4397 1.3 1 0.32 349 8.5 14.9 1.8 7.3 1.3 0.4 1.3 0.2 1.1 0.3 0.7 0.1 0.7 0.1 2.5 - 3657 1.4 1.1 0.45 235 7.8 13.5 1.6 6.7 1.3 0.3 1.3 0.2 1.1 0.2 0.6 0.1 0.6 0.1 1.5 - 12260 1.3 0.9 0.87 310 9.5 18.5 2.2 8.7 1.7 0.4 1.6 0.2 1.4 0.3 0.9 0.1 0.8 0.1 2.5 - 5157 2.2 1.1 0.54 276 6.7 11.9 1.5 6 1.2 0.3 1 0.2 1 0.2 0.6 0.1 0.5 0.1 1.3 - 6350 1.1 1 2.97 357 13.8 27 2.7 8.9 1.7 0.5 1.6 0.3 1.8 0.4 1.3 0.2 1.3 0.2 4.6 - 302 5.8 2.9 1.68 803 28.5 41.2 4.2 13.4 1.9 0.4 1.4 0.2 1.2 0.3 0.7 0.1 0.6 0.1 2.3 - 4973 2.3 1 0.3 320 6.7 12.2 1.5 6.1 1.2 0.3 1.1 0.2 1.1 0.2 0.6 0.1 0.5 0.1 1.3 - 4217 1.2 1 0.38 318 8 14 1.8 7 1.4 0.4 1.4 0.2 1.1 0.3 0.7 0.1 0.7 0.1 2.2 - 3250 1.4 1.1 0.3 275 7 12.4 1.6 6.3 1.4 0.4 1.2 0.2 1.1 0.2 0.6 0.1 0.5 0.1 1.5 - 6760 1.2 1 0.35 288 7 12.9 1.7 6.4 1.3 0.3 1.2 0.2 1.2 0.2 0.6 0.1 0.6 0.1 1.7 - 4950 1.3 1.1 0.26 226 6.9 10.8 1.4 5.7 1.2 0.3 1.2 0.2 1 0.2 0.6 0.1 0.5 0.1 1.2 - 6077 1 0.9 0.34 1736 8.9 16.3 2.1 8.5 1.8 0.5 1.6 0.3 1.6 0.3 1 0.1 0.9 0.1 4 - 315 1.8 1.3 0.32 351 7.2 13 1.6 6.4 1.4 0.4 1.2 0.2 1.1 0.2 0.7 0.1 0.6 0.1 1.8 - 5890 1.3 1.1 0.43 345 7.6 13.4 1.7 6.8 1.5 0.3 1.3 0.2 1.2 0.2 0.6 0.1 0.6 0.1 1.8 - 3856 1.3 1 0.25 207 6.9 13.3 1.6 6.2 1.1 0.3 0.9 0.1 0.8 0.2 0.5 0.1 0.5 0.1 2.4 - 10567 1.4 0.8 0.37 299 7.2 12.8 1.6 6.4 1.3 0.4 1.3 0.2 1.1 0.2 0.7 0.1 0.6 0.1 1.7 - 11330 1.3 1 0.06 183 6.7 12.1 1.5 6.1 1.3 0.3 1 0.2 1.1 0.2 0.6 0.1 0.6 0.1 1.9 0.2 4035 1.2 1.3 - - - - - - - - - - - - - - - - - - - - - 0.36 318 5.5 12.3 1.3 5.2 1.3 0.3 1.5 0.2 1.4 0.2 0.9 0.1 0.8 0.1 2.9 0.1 93 2.9 0.6 Appendix IV photos of the samples from Maastricht and Utrecht 1:1 0 1:2 5cm Figure appendix IV.1 Photo MABRO 1, Mabro, Maastricht, rim fragment of crucible with white ‘frit-like’ material adhering. 0 Figure appendix IV.4 Photo MABRO 4, Mabro,Maastricht, rim fragment of crucible with colourless glass on inside. 1:2 0 1:2 10cm Figure appendix IV.2 Photo MABRO 2, Mabro, Maastricht, Crucible fragment with colourless and white/yellow material attached. 1:2 0 10cm 10cm Figure appendix IV.3 Photo MABRO 3, Mabro, Maastricht, small crucible fragment with green and weathered opaque yellow pigment. 0 10cm Figure appendix IV.5 Photo MABRO 5, Mabro,Maastricht, rim fragment of crucible with colourless vitreous material on inside white ‘frit-like’ material adhering to both sides. 183 — 184 — 1:1 1:2 0 10cm Figure appendix IV.6. Photo MABRO 6, Mabro, Maastricht, base fragment of crucible with opaque yellow pigment adhering. 0 5cm Figure appendix IV.9 Photo MABRO 9, Mabro,Maastricht, small crucible fragment with deep translucent glass on the inside. 1:1 0 1:1 5cm Figure appendix IV.7 Photo MABRO 7, Mabro,Maastricht, small red fragment of crucible with natural green glass adhering to both sides. 1:2 0 10cm Figure appendix IV.8 Photo MABRO 8, Mabro,Maastricht, red fragment of crucible base with opaque yellow outside and colourless-green glass adhering on the inside. 0 5cm Figure appendix IV.10 Photo MABRO 10, Mabro,Maastricht, red base fragment of crucible with green vitrification on lower side. 185 — 1:2 0 1:2 10cm Figure appendix IV.11 Photo MAJO 1, Jodenstraat, Maastricht, red base fragment of crucible with opaque yellow pigment adhering (MAJO= Jodenstraat, Maastricht). 0 10cm Figure appendix IV.14. Photo MAJO 4, Jodenstraat, Maastricht, base fragment of crucible with white residue adhering (MAJO= Jodenstraat, Maastricht). 1:2 0 10cm Figure appendix IV.12. Photo MAJO 2, Jodenstraat, Maastricht, crucible base of grey fabric with weathered opaque yellow pigment and translucent residue adhering (MAJO= Jodenstraat, Maastricht). 1:2 0 10cm Figure appendix IV.13 Photo MAJO 3, Jodenstraat, Maastricht, base fragment of crucible with weathered opaque yellow pigment adhering (MAJO= Jodenstraat, Maastricht). Figure appendix IV.15 Photo MAJO 5 (inside), Jodenstraat, Maastricht, base of crucible with opaque yellow glass adhering (MAJO= Jodenstraat, Maastricht). 186 — 1:2 0 10cm Figure appendix IV.16 Photo MAJO 5 (outside), Jodenstraat, Maastricht, base of crucible. 1:2 0 10cm Figure appendix IV.19 Photo MAJO 8, Jodenstraat, Maastricht, base of crucible with dark translucent glass adhering (MAJO= Jodenstraat, Maastricht). 1:2 0 10cm Figure appendix IV.17 Photo MAJO 6, Jodenstraat, Maastricht, base fragment of crucible with opaque yellow glass on inside (MAJO= Jodenstraat, Maastricht). 1:2 0 10cm Figure appendix IV.20 Photo MAJO 9, Jodenstraat, Maastricht, possible brick fragment with opaque yellow and white residue adhering (MAJO= Jodenstraat, Maastricht). 1:2 0 10cm Figure appendix IV.18 Photo MAJO 7, Jodenstraat, Maastricht, base of crucible with opaque yellow glass adhering (MAJO= Jodenstraat, Maastricht). 187 — 1:1 0 2:1 5cm Figure appendix IV.21 Photo MAJO 10, Jodenstraat, Maastricht, blue fragments of glass (MAJO= Jodenstraat, Maastricht). 2,5cm Figure appendix IV.22 Photo MAJO 11, Jodenstraat, Maastricht, scrap of red glass (MAJO= Jodenstraat, Maastricht). 2:1 0 2,5cm Figure appendix IV.24 Photo MAJO 13, Jodenstraat, Maastricht, scrap of green glass (MAJO= Jodenstraat, Maastricht). 1:1 2:1 0 0 2,5cm Figure appendix IV.23 Photo MAJO 12, Jodenstraat, Maastricht, scrap of red glass (MAJO= Jodenstraat, Maastricht). 0 5cm Figure appendix IV.25 Photo MAJO 14, Jodenstraat, Maastricht, yellow-green window glass (MAJO= Jodenstraat, Maastricht). 188 — 1:1 0 2:1 5cm Figure appendix IV.26 Photo MAJO 15, Jodenstraat, Maastricht, yellow-green window glass (MAJO= Jodenstraat, Maastricht). 5cm Figure appendix IV.27 Photo MAJO 16, Jodenstraat, Maastricht, pale yellowgreen window glass (MAJO= Jodenstraat, Maastricht). 1:1 0 2,5cm Figure appendix IV.29 Photo MAJO 18, Jodenstraat, Maastricht, weathered yellow drop (MAJO= Jodenstraat, Maastricht). 2:1 1:1 0 0 5cm Figure appendix IV.28 Photo MAJO 17, Jodenstraat, Maastricht, thin green glass rod (MAJO= Jodenstraat, Maastricht). 0 2,5cm Figure appendix IV.30 Photo MAJO 19, Jodenstraat, Maastricht, red glass drop (MAJO= Jodenstraat, Maastricht). 189 — 2:1 0 2,5cm Figure appendix IV.31 Photo MAJO 20, Jodenstraat, Maastricht, dark green glass drop (MAJO= Jodenstraat, Maastricht). 1:1 0 5cm Figure appendix IV.34 Photo MAJO 23, Jodenstraat, Maastricht, twisted opaque white rod (MAJO= Jodenstraat, Maastricht). 2:1 0 2,5cm Figure appendix IV.32 Photo MAJO 21, Jodenstraat, Maastricht, milky blue pulled rod (MAJO= Jodenstraat, Maastricht). 1:1 0 5cm Figure appendix IV.35 Photo MAJO 24, Jodenstraat, Maastricht, green beaker base (MAJO= Jodenstraat, Maastricht). 1:1 0 5cm Figure appendix IV.33 Photo MAJO 22, Jodenstraat, Maastricht, thin red rod (MAJO= Jodenstraat, Maastricht). 190 — 2:1 0 1:1 2,5cm Figure appendix IV.36 Photo MAJO 25, Jodenstraat, Maastricht, blue punty glass (MAJO= Jodenstraat, Maastricht). 0 Figure appendix IV.39 Photo DOM 1, Domplein, Utrecht, Body fragment of crucible with colourless glass adhering (DOM=Domplein, Utrecht). 1:1 0 1:1 5cm Figure appendix IV.37 Photo MAJO 26, Jodenstraat, Maastricht, blue-green flat ribbed fragment (MAJO= Jodenstraat, Maastricht). 2:1 0 5cm 2,5cm Figure appendix IV.38 Photo MAJO 27, Jodenstraat, Maastricht, green glass rod fragments (MAJO= Jodenstraat, Maastricht). 0 5cm Figure appendix IV.40 Photo DOM 2, Domplein, Utrecht 2 Body fragment of crucible with colourless glass adhering (DOM=Domplein, Utrecht). 191 — 1:1 1:2 0 10cm Figure appendix IV.41 Photo DOM 3, Domplein, Utrecht 3 Base fragment of crucible with pale green and red glass adhering (DOM=Domplein, Utrecht). 5cm Figure appendix IV.42 Photo DOM 4, Domplein, Utrecht 4 Rim fragment of crucible with pale green and red glass adhering (DOM=Domplein, Utrecht). Figure appendix IV.44 Photo DOM 6, Domplein, Utrecht 6 Body fragment of crucible with pale green and red glass adhering (DOM=Domplein, Utrecht). 0 2,5cm Figure appendix IV.45 Photo OUDWIJ 1 (sample 78), A lump of melted pale green glass from Utrecht (OUDWIJ=Oudwijkerdwarsstraat, Utrecht). 2:1 1:2 0 5cm 2:1 1:1 0 0 10cm Figure appendix IV.43 Photo DOM 5, Domplein, Utrecht 5 Rim fragment of crucible with green glass adhering (DOM=Domplein, Utrecht). 0 2,5cm Figure appendix IV.46 Photo OUDWIJ 2 (sample 79), A chunk of pale green glass from Utrecht (OUDWIJ=Oudwijkerdwarsstraat, Utrecht). This monograph brings together for the first time comprehensive combined archaeological, technological and scientific investigations of early medieval glass production in the Netherlands. The relationships between scientific results, archaeological contexts, sample dates, object types, colour, changes in glass technologies over time, as well as the social, economic and political factors affecting glass supply, and glass production, are discussed. We have selected samples from nine key sites, dating to between the late 4th and 11th centuries. Trace element and isotopic results for early medieval glasses have provided new and significant insights. They show that most glass in use was recycled and that there is a greater proportion of imported ‘pristine’ Egyptian glass in the Merovingian period than in the Carolingian period. A small proportion of wood ash glass was added to imported Carolingian glass found in the Netherlands; in contemporary northern Italian and Spanish glasses Levantine glass was added as part of the recycling process instead. We highlight the international importance of evidence for the production of yellow and white glass tin-based colorants in Maastricht and their use to make monochrome beads there. A wider range of glass technologies was in use after the 9th century following an important technological transition. This scientific report is intended for archaeologists, as well as for other professionals and amateur enthusiasts involved in archaeology. The Cultural Heritage Agency of the Netherlands provides knowledge and advice to give the future a past.