Scholarly article on topic 'Ancient glass: from kaleidoscope to crystal ball'

Ancient glass: from kaleidoscope to crystal ball Academic research paper on "History and archaeology"

CC BY
0
0
Share paper
Academic journal
Journal of Archaeological Science
OECD Field of science
Keywords
{Glass / "Late Bronze Age" / Roman / "Late Antique" / Islamic / Medieval}

Abstract of research paper on History and archaeology, author of scientific article — Th. Rehren, Ian C. Freestone

Abstract Research over the last few decades has greatly enhanced our understanding of the production and distribution of glass across time and space, resulting in an almost kaleidoscopically colourful and complex picture. We now recognise several major ‘families’ of glass composition, including plant-ash based glass in Late Bronze Age Egypt and Mesopotamia, and the Islamic World; mineral natron glass in the Greek, Roman and Byzantine Empires; mineral-based lead- and lead–barium glass in Han period China and medieval Europe; and wood-ash and ash-lime glass in medieval Europe. Other glass groups include a peculiar granite-based glass in medieval Nigeria, and probably mineral-based glass in Bronze Age southern Europe. However, despite two centuries of research, we know very little about the actual production locations and technologies for most of these glass groups, and how and where glass making was invented. The early literature reflects the comparatively limited number of individuals and research groups working on glass; only recently there is a significant broadening of the research community and expansion and refinement of the data base. This enables us now to take stock of our current understanding and identify major lacunae and areas where additional work may make the most significant contributions to our understanding of the complex picture. Hopefully this will help moving from the traditional descriptive and often fragmented opportunistic data-gathering phase (asking ‘what’, ‘where’ and ‘when’) to a more interpretative period looking with fresh eyes at the ‘why’ and ‘how’ of compositional and technical developments. This opening of the research field includes addressing the relationship of the different glass industries to the societies that used glass, and how they organised its production and distribution. A major overarching issue remains the question of the initial invention of glass, and how the idea as well as the material itself spread. Major debates should ask whether there were multiple inventions of glass making; how best to identify and interpret long-distance trade; how to ensure data compatibility and quality; and how to integrate different types of data, from archaeology through craftsmanship and typology to chemistry and optical properties.

Academic research paper on topic "Ancient glass: from kaleidoscope to crystal ball"

Accepted Manuscript

Ancient glass: from kaleidoscope to crystal ball Thilo Rehren, Ian Freestone

PII: S0305-4403(15)00059-X

DOI: 10.1016/j.jas.2015.02.021

Reference: YJASC 4351

To appear in: Journal of Archaeological Science

Received Date: 15 December 2014 Revised Date: 2 February 2015 Accepted Date: 8 February 2015

Please cite this article as: Rehren, T., Freestone, I., Ancient glass: from kaleidoscope to crystal ball, Journal of Archaeological Science (2015), doi: 10.1016/j.jas.2015.02.021.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Ancient glass: from kaleidoscope to crystal ball

Thilo Rehren and Ian Freestone

UCL Qatar and UCL Institute of Archaeology 2 UCL Institute of Archaeology

Abstract

Research over the last few decades has greatly enhanced our understanding of the production and distribution of glass across time and space, resulting in an almost kaleidoscopically colourful and complex picture. We now recognise several major 'families' of glass composition, including plant-ash based glass in Late Bronze Age Egypt and Mesopotamia, and the Islamic World; mineral natron glass in the Greek, Roman and Byzantine Empires; mineral-based lead- and lead-barium glass in Han period China and medieval Europe; and wood-ash and ash-lime glass in medieval Europe. Other glass groups include a peculiar granite-based glass in medieval Nigeria, and probably mineral-based glass in Bronze Age southern Europe. However, despite two centuries of research, we know very little about the actual production locations and technologies for most of these glass groups, and how and where glass making was invented.

The early literature reflects the comparatively limited number of individuals and research groups working on glass; only recently there is a significant broadening of the research community and expansion and refinement of the data base. This enables us now to take stock of our current understanding and identify major lacunae and areas where additional work may make the most significant contributions to our understanding of the complex picture. Hopefully this will help moving from the traditional descriptive and often fragmented opportunistic data-gathering phase (asking 'what', 'where' and 'when') to a more interpretative period looking with fresh eyes at the 'why' and 'how' of compositional and technical developments. This opening of the research field includes addressing the relationship of the different glass industries to the societies that used glass, and how they organised its production and distribution. A major overarching issue remains the question of the initial invention of glass, and how the idea as well as the material itself spread. Major debates should ask whether there were multiple inventions of glass making; how best to identify and interpret long-distance trade; how to ensure data compatibility and quality; and how to integrate different types of data, from archaeology through craftsmanship and typology to chemistry and optical properties.

1. Introduction

The scientific analysis of glass has a relatively short history, despite some very early work going back to the late 18th century (Caley 1962). The first meaningful analyses were those published in the 1950s in two series of papers by W.E.S. Turner in Glass Technology and by

W. Geilmann in Glastechnische Berichte, respectively, followed quickly in the 1960s by papers by Edward Sayre and then Robert Brill of the Corning Museum of Glass, whose seminal work dominated scientific research on early glass for nearly half a century.

Only for the last few decades though has the analysis of ancient glass attracted the attention of a wider group of scholars, resulting in both a diversity of analytical approaches and the development of a 'critical mass' of research activity that led to the current vibrancy in the field.

Three main factors were behind this: the rapid development of new analytical techniques requiring ever smaller sample volumes (Janssens 2013); the increasing archaeological interest and excavations targeting more technical sites (Nenna et al. 2000; Gorin-Rosen 2000; Eramo 2006; Pusch and Rehren 2007); and the development of interpretative models based on the series of analyses in a wider theoretical concept (e.g. Lankton and Dussubieux 2006). Here, we summarise our perspective on some of the recent thinking about ancient glass making and how the scientific analysis of glass and glassmaking waste can inform our understanding of past societies and the people who formed them.

1.1. Raw materials: fluxes

All ancient glass is based on silica as the main component, combined with a flux to lower the melting or softening point of the silica. The type of flux used in Antiquity often serves to characterise the resulting glass in our modern terminology, even though we have little evidence that this separation ever mattered in the past. The main pre-industrial glass families include plant-ash glass, mineral-natron glass, lead and lead-barium glass, and wood-ash glass, which then each split into more specific compositional groups.

Plant-ash glass is based on the ash of particular plants that typically grow in salt-rich environments (desert, on- and off-shore on the coast) and accumulate sodium and to a lesser extent potassium (Brill 1970); when the plants are burned these alkalis form a considerable part of the ash component, which reaches up to 25 % by weight of the dry matter (e.g. Barkoudah and Henderson 2006). This plant ash forms the basis of most Bronze Age glass in the Middle East, Iron Age, Roman-period and late Antique glass east of the Euphrates (in particular Sasanian glass: e.g. Mirti et al. 2009), and most of the Islamic glass from Central Asia to Portugal. It is mostly soda-based (10 to 20 wt% Na2O), with a few percent each (2 to 8 wt%) of magnesia and potash.

Mineral-natron glass is also soda-rich, but based on evaporite minerals, in particular from the Wadi Natrun in Lower Egypt (Shortland et al. 2006, Devulder et al. 2014). It is characterised by similar soda levels as the plant-ash glass, but typically has less than 1.5 wt% each magnesia and potash. Ongoing research addresses whether there are other significant natron sources beyond the Wadi Natrun that were exploited in antiquity, such as in western Turkey (Rehren et al. 2015), and why the Wadi Natrun seemingly stopped supplying natron to the eastern Mediterranean glass industry even in Egypt itself, towards the end of the first millennium AD (Whitehouse 2002; Shortland et al. 2006).

Lead- and lead-barium glass have heavy metal oxide(s) as the main flux, resulting in a higher specific gravity and a higher refractive index and hence brilliance of the glass. Lead oxide was available in large quantities as a waste or by-product from the refining of silver-containing lead metal, or could have been made specifically by oxidising lead metal, or obtained in mineral form. It has been used continuously but in relatively small quantities as a component of strongly coloured glass across Europe, the Mediterranean and the Middle East since the emergence of glass use, both to make glass beads and for glazes and enamel (Shortland 2002; Henderson and Warren 1981; Tite et al. 1998; Brun and Pernot 1992). Lead-barium glass was probably based on the exploitation of a particular mineral association, and its occurrence is restricted to a few hundred years in central China (Gan 2005; Cui et al. 2011). In medieval Europe, the use of lead as a flux gave rise to simple lead-silica compositions as well as mixed lead-wood-ash glass (Mecking 2013), but these leaded glasses make up a very small part of the overall corpus for the period. Isotopic data suggest that the lead oxide in at least some of these glasses is a by-product of silver smelting (Wedepohl and Baumann 1997). The optical properties resulting from the high dispersion imparted by lead oxide to glass were exploited by English glassmakers in their development of transparent lead crystal glass in the late seventeenth century as a competitor to the famous cristallo of Venice. The compositional development of "English crystal" has been charted by Dungworth and Brain (2009), while the origins of this "invention" are significantly more complex than the impression given in the earlier literature (Moretti and Zecchin 2009).

Wood-ash glass is predominantly based on potash as the main flux, obtained from the ash of timber, shrubs or plants such as fern (Jackson and Smedley 2008; Smedley and Jackson 2002). Due to the lower overall ash content of timber compared to the plants used for plant-ash glass making, much larger quantities of wood needed to be burnt for ash-making (Crossley 1998). In addition, wood ash is less clean than plant ash, in that it contains higher amounts of impurities such as iron oxide resulting in the formation of more strongly coloured lower-quality glass. Wood-ash glass appears first in the eighth century AD north of the Alps and dominated glassmaking there by the High Middle Ages (Wedepohl 1998). The low quality led to different responses of European glassmakers; the Venetian approach was from the outset to import high-quality plant ash from the Levant, where they had a trade monopoly which underpinned the success of the Venetian glass industry (Ashtor and Cevidalli 1983). Elsewhere, cleaner raw materials such as limestone were eventually added to the wood ash resulting in more lime-rich glass, while the attempts to refine the raw wood ash, through washing, decanting and drying, contributed to the development of modern chemistry. It also gave us the term potash for the refined salt crystallising in the bottom of the pot in which the decanted solute from the washed ash was concentrated, and thence the chemical name potassium for the dominant element in potash. With increasing understanding of chemistry came the relatively rapid changes in flux composition from the later Middle Ages, leading eventually, as part of the Industrial Revolution, to the emergence in the early 19th century of glass made from artificially produced soda.

These glass categories based on their flux are not in all cases as strictly demarcated as it may appear. There are lead glasses with potash levels that indicate the addition of refined wood

123 ash, and mineral-natron glass with elevated levels of magnesia and potash that could either

124 indicate the intentional addition of an ash component, or the contamination of the glass by

125 the fuel ash during extended periods of firing (e.g. Paynter 2008), in addition to the inherent

126 contribution of these oxides from the sand. For other glasses, such as the north Italian

127 Bronze Age glasses associated with the site of Frattesina (Towle et al. 2001, Angelini et al.

128 2004), African glasses from medieval Nigeria (Lankton et al. 2006; Freestone 2006a), or the

129 alumina-rich Indian glasses (Dussubieux et al. 2010), we know little about the flux used;

130 these may well include mineral fluxes or self-fluxing sediments that have not been

131 characterised yet.

133 1.2. Raw materials: silica

134 Compared to the variability of fluxes used, the silica sources are relatively limited. For

135 relatively small-scale production, crushed and finely ground vein quartz was used, as

136 suggested by the low alumina and trace element levels in the glasses of New Kingdom Egypt

137 and Mesopotamia (Shortland et al. 2007; Rehren and Pusch 2008; and see below, 4.1.). The

138 large-scale glass production of the Roman to Byzantine and Islamic Empires, on the other

139 hand, was based on quarried sand (Brems et al. 2012). While the high quality colourless

140 glasses of the European glass industry, for example the famous Venetian cristallo depended

141 on the availability of quartz pebbles (Verita 1985), beds of flint nodules were also used, and

142 most potash-lime or wood-ash glass of the medieval industry used relatively impure river

143 sands, as suggested by high concentrations of elements such as zirconium, present as

144 detrital minerals (Wedepohl et al. 2011). For the African, Indian and Chinese glasses, we can

145 only assume that quartz-rich sand or rocks were used. A group of widely traded Indian soda-

146 rich glasses is characterised by varying levels of high alumina, surely indicating the use of a

147 sand raw material (Dussubieux et al. 2010), while the barium-lead-silica glasses of Han China

148 seem to have been based on a pure source of quartz, gathered directly from veins or as

149 pebbles (see e.g. glass compositions reported in Brill et al. 1991). These different quartz

150 sources bring with them different, and sometimes diagnostic, levels of minor and trace

151 elements which are then found in the finished glass.

152 Late Antique mineral-natron glass is among the most intensively analysed glass types; here,

153 mineralogical differences in the main sand sources lead to chemical differences in the

154 finished glass, which are recognised as different primary compositional groups, and in many

155 circumstances can assist in provenancing the sand source. These groups have been largely

156 based upon simple bivariate inter-element plots and characterise the glasses according to

157 the lime, alumina, iron oxide and titania contents of the glassmaking sands, allowing glasses

158 from primary production centres in Egypt and the Levant to be distinguished (Gratuze and

159 Barandon 1990, Freestone et al. 2000, 2002a). Statistical analysis of similar datasets reveals

160 similar groups (Foy et al. 2003; Freestone et al. 2002a, Gliozzo et al. 2013), but even so our

161 current understanding of different production groups in this period needs an extension of

162 geographical coverage, a refinement of chronology and an authoritative revision of

163 terminology leading to a common set of labels.

2. Production Technology

A key characteristic that sets glassmaking apart from metal making is the absence of durable slag that would help locate glassmaking sites, and provide clues about the raw materials and the melting conditions. Instead, glass making involved the melting together of raw materials to a chemically complex product, with limited waste in the form of non-reactive salts (but see Tanimoto and Rehren 2008) floating on top of the glass melt. These were removed one way or the other during the process and due to their solubility in water almost immediately also lost from the archaeological record. Hence only the finished product and remains of the melting installations such as crucibles, kilns and furnaces and some working waste provide information about the production processes.

Investigation of Late Bronze Age glass production debris from Egypt has shown that at this time, glass making involved two firing stages at different temperatures (Rehren and Pusch 2005; Smirniou and Rehren 2011). In contrast, there is apparently only a single firing stage in the very large scale Roman to early Islamic production of the eastern Mediterranean coastal plane, where raw materials on a scale of from ten to in excess of thirty tonnes were melted in large tanks, probably for many weeks in a firing (Gorin-Rosen 2000; Aldsworth et al. 2002). Medieval European glassmaking involved fritting as a first stage hot enough to break down the carbonate molecules of the flux but not hot enough to produce a substantial melt phase. Taking the frit to higher temperatures, typically in large refractory crucibles or pots, would then have led to the formation of a homogenous melt that could be worked directly from the pot before it solidified upon cooling to a glass.

2.1. Firing technology

In order to produce a glass which was transparent and clear, there was a need to keep it clean and as free of contaminants as possible. Mixing the fuel with the charge, common practice in metal smelting, was not an option, and modes of indirect heating were employed, typically involving the use of luminous flames and hence timber rather than charcoal as fuel (Crossley 1998), and the construction of reverberatory furnaces whose domed roofs project the heat back onto the batch at their sole (e.g. Gorin-Rosen 2000, Nenna 2015; see Kock and Sode 1996; Sode and Kock 2001 for an ethnographic example). Depending upon the type of firing technology, different approaches were used to minimise contamination of the molten glass by the container material. In Late Bronze Age Egypt, a parting layer of lime was applied to the interior of the crucible to isolate the glass (Turner 1954; Rehren 1997), while in late and post-medieval Europe, refractory kaolinitic clays were selected which withstood very high temperatures (Eramo 2006; Paynter 2012) as well as the chemically aggressive melt charge. In the Roman period, the low surface area to volume ratios of the large tanks used for glass making and frequently also for working, meant that the risk of spoiling the glass by over-reaction with the container material was much lower. For glass working, smaller kilns were often used holding a number of pots or crucibles (Gaitzsch et al. 2000: 102, 110-113; Jackson et al. 2003; Paynter 2008; Schibille and

205 Freestone 2013); again, the heating of crucibles and the glass batch within them was

206 indirect, with the firing technology and airflow more akin to pottery kilns than to

207 metallurgical furnaces.

209 2.2. Manipulating colours

210 Ancient glass typically had levels of iron oxide (c. 0.5%) an order of magnitude higher than

211 modern window glass and the reduction of the relatively strong aqua-blue colour imparted

212 by Fe2+ appears to have been a major concern of the glassmakers. This was typically attained

213 by adding manganese or antimony oxides to the batch, which oxidised the iron to the ferric

214 state, which imparts a pale yellow tint to glass (Sayre 1963). The oxidation states of glass

215 have only recently begun to be determined, but significant work has been undertaken on

216 Roman glass, following pioneering work by Schreurs and Brill (1984; e.g. Bingham and

217 Jackson 2008, Meulebroek et al. 2011, Zoleo et al. 2014). Without additives, glass obtained

218 directly from the primary furnaces was relatively reduced, but the addition of manganese or

219 antimony oxide led to the oxidation of around 90% of the iron (Arletti et al. 2013). However,

220 analysis of common samples is required to determine the extent to which the data obtained

221 using the various methodologies are intercomparable.

222 Over three millennia, the most important compounds in the colouration of glass, either

223 solely or in combination, include the oxides of only seven metals, namely copper, cobalt, tin,

224 antimony, lead, manganese, and iron, to achieve the full range of glass colours from water-

225 clear decoloured glass to the intense opaque and transparent colours used to imitate all

226 precious stones known in antiquity, most notably lapis lazuli, turquoise, amethyst, obsidian,

227 jade, alabaster, carnelian, rock crystal, and emerald. The production technologies of the

228 various colours have been investigated to varying degrees (eg. Sayre 1963; Shortland 2002,

229 Shortland et al. 2006, Tite et al. 2008; Barber et al. 2009; Freestone and Stapleton 2014) and

230 there has been a large amount of work on individual assemblages, building a pattern which

231 will allow us to understand change from one compound or method of coloration to another

232 over time and space and its underlying reasons. Outstanding issues include questions of raw

233 material identification and access, in particular for such materials as cobalt or manganese

234 which were not known in their own right (cf. Gratuze et al. 1996 for cobalt in medieval and

235 early modern Europe); reasons for change from one compound or method of coloration to

236 another over time and space; as are the practicalities of the addition of these compounds

237 either during the primary production, or the secondary working stage (Lahlil et al. 2010,

238 2011). Furthermore, the locations of the workshops that produced large quantities of

239 coloured glasses are still not well understood, whether for the large wall mosaics of

240 antiquity or the stained glass windows of the medieval period.

242 3. Organising glass making

243 The production of glass requires a combination of particular chemical skills, access to

244 specific and often exotic or ill-defined raw materials, large quantities of fuel, and a market

for the finished product. Several models have been developed to describe the organisation of these factors. For the Bronze Age, Rehren has proposed a model comprising of several elite-attached workshops producing coloured glass ingots which were then provided to also elite-attached glass studios processing glass from different producers, in different colours (Pusch and Rehren 2007: 162, Fig. 1). This model incorporates the textual and archaeological evidence for long-distance exchange of coloured glass ingots linking Egypt and Mesopotamia with the Levant and Greece, but still does not explain why there is so little evidence (if any) for Egyptian glass in Mesopotamia, and for Mesopotamian glass in Egypt, when the pictorial sources (Thutmose III) and the textual evidence (Amarna letters) indicate a relatively regular flow of glass at least from Mesopotamia into Egypt (Rehren 2014).

Fig. 1: Schematic of the Late Bronze Age glass industry in Mesopotamia and Egypt, where most workshops were able to make copper-coloured light blue glass, using copper in Mesopotamia and bronze in Egypt. More specialised colours were only made in specific workshops and then exchanged so that each glass studio had the full range of glass colours available, without necessarily being able to make each colour.

For the Late Roman and Byzantine Empires, Freestone has developed a model that is based on the spatial separation of primary production of glass in a few large-scale production centres and countless secondary working sites processing the primary glass as well as recycled cullet into new artefacts (Fig. 2). The link here between the primary and secondary workshops is provided by irregular fist- to head-sized chunks of fresh glass that have been found across the Roman Empire (Fig. 3). Open questions include the socio-economic organisation of the large production sites, the possible existence of specialised workshops producing coloured glass, the relationship between glass tesserae production for mosaicists and the normal glass industry, among others.

Fig. 2: Model of the Late Antique glass industry, with a limited number of glass factories producing glass from imported natron and regional sand. This primary glass was then shipped to individual glass houses across the Empire who would work the glass into artefacts, serving their regional or local markets (from Freestone et al. 2002b).

Fig. 3: Glass chunks ready for re-melting and blowing into vessels, stored in a pottery vessel in the 6th-7th century workshop at Beth Shean, Israel. (Reconstruction displayed in the Israel Museum, Jerusalem.)

In contrast, the medieval European (Fig. 4a) as well as the later Islamic glass industries (Fig. 4b) both seem to have functioned around integrated workshops, combining the production of glass from the raw materials to the forming of finished artefacts for the local or regional market.

286 Fig. 4a: Depiction of a medieval glass furnace, showing the quarrying of sand in a landscape with

287 managed woodlands (background), the actual glass furnace in the foreground with vessels being

288 blown (front), and the inspection of finished vessels as they are retrieved from the annealing

289 chamber (front left). Mandeville manuscript, illustration c 15th century. © The British Library Board,

290 Add. 24189, f.16.

292 Fig. 4b: Procession of glass workers with a working furnace on a float. From a manuscript dated to

293 the reign of the Ottoman sultan Murat III, late 16th century AD (Kügükerman 1985, 114).

295 In both regions there is a period of transition from the Late Antique mineral-natron based

296 glass industry and its physical remains to the newly-established indigenous industries; in

297 Europe, this dependency on recycling pre-medieval glass was initially stronger due to the

298 necessity to develop a completely new glassmaking tradition, based on wood ash. However,

299 once wood-ash based glass making had been established, the huge demands for glass from

300 first ecclesiastical and later urban customers drove a massive growth in glass making. The

301 need for fire wood and wood for ash production put increasing pressure on European

302 woodlands, in competition with other major wood-consuming industries such as mining and

303 metallurgy; ship-building; urban construction and fire wood; and game hunting by the

304 nobility. In response, large quantities of wood ash were traded from the forested regions of

305 northern Russia and the Baltic into central Europe (e.g. Clow and Clow 1956: 87, 98), and

306 woodland management strategies such as coppicing and pollarding were implemented to

307 increase timber and firewood production (Crossley 1998).

308 In the East, in contrast, plant-ash based glass making had persisted throughout the Roman

309 and Byzantine periods in the Sasanian Empire and Central Asia, and it is generally assumed

310 that this was the route to its relatively easily re-adoption in the Levant and Egypt. The

311 availability of high-quality raw materials and a social environment favourable to the

312 development of the arts and crafts under the early Abbasid rulers led to a flourishing art of

313 glass making. Here, there is a remarkable chemical differentiation between glass made in

314 different parts of the Islamic world (Fig. 5; Freestone 2006b), most probably reflecting the

315 different soil chemistries from which the halophytic plants metabolise their inorganic

316 constituents. Barkoudah and Henderson (2006) provide some insight into the variability of

317 plant ashes from different regions, but more work is clearly needed before the glass

318 composition can be used for anything more than the broadest of geographical assignations

319 (e.g. Rosenow and Rehren 2014: Fig. 7).

321 Fig. 5: Potash and magnesia contents of early Islamic glass of the Syrian type (Banias, Tyre and Serce

322 Limani) and Sasanian glass from Mesopotamia. The Raqqa Type 4 glass is early Islamic, showing the

323 continuity of the Sasanian tradition of glassmaking. The presence of natron glass (lower left)

324 indicates a significant trade in material or vessels from the Roman West into the Sasanian Empire.

325 For data sources see Freestone (2006b); Raqqa data from Henderson et al. (2004).

Major research lacunae exist with regard to the production of Han Chinese glass. Very little can be said about the raw materials used for its production, and nothing about the places of production and the organisation of the industry, if ever it was one and not just an elite-attached production. Remarkable, however, is the way in which Chinese glass sets out to imitate the particular optical quality of jade (sensu lato), by achieving an opalescent quality unknown to Western glasses, which in contrast mimic the strong colours of precious stones cherished there. This, in our opinion, strengthens again the idea that ancient glass was very much a substitute for, or at least inspired by, precious stones (Nicholson 2012; Shortland 2012).

The primary production of glass in sub-Saharan Africa has only recently been recognised (Freestone 2006; Lankton et al. 2006), and raises interesting questions not only regarding the raw materials used (Ige et al. 2015), but also about the mechanisms of knowledge transfer or inspiration for invention in a society that values glass highly for its bead-making ability, but lacked the centuries-old experience in glass making shared by the northern societies that provided much of the glass to Africa.

4. Tracing the glass cycle

The ability of molten glass to incorporate a very wide range of oxides is of major importance, for a number of reasons. The coloration of glass, through the addition of small amounts of transition and base metal oxides, has been briefly mentioned above. Similarly, small amounts of minerals other than quartz which are associated with it contribute to the trace element signature of the finished product. Finally, working and recycling of glass further influences the composition of the glass, providing opportunities as well as challenges for the interpretation of trace element pattern.

4.1. Provenancing

The ability to distinguish production groups based on their raw material characteristics has inevitably led to attempts to provenance glass materials to their primary production centres. As the sand carries many of the trace elements which are useful in this respect, provenancing glass is essentially reduced to provenancing sand. Exceptions include for example post-medieval European glass, where the availability of very clean vein quartz and the diversity in ash compositions used for glassmaking during and after the Renaissance means that glasses attributable to specific regions are characterised by different alkalis (Gratuze and Janssens 2004). For the major industries of the Late Bronze Age and the first millennium AD, trace element and isotopic studies have proven very successful. For example, Shortland and co-workers (2007) have shown that Late Bronze Age Egyptian glass may be distinguished from Mesopotamian glass by the elements Cr, Zr, La and Ti, thereby resolving a number of issues around their origin and confirming that in the Eighteenth Dynasty, most glass used in Egypt was made from Egyptian raw materials, while LBA glass

found in Europe came partly from Egypt and partly from Mesopotamia (e.g. Jackson and Nicholson 2010; Varberg et al. 2015). The isotopes of neodymium have proved very useful in provenancing sand-based glasses of the Mediterranean. The 144Nd/143Nd ratio of a sand essentially reflects the geological age of the rocks of its source terrain, and young rocks tend to have higher 144Nd/143Nd tha n older rocks. The Nile carries a sediment load derived from the young volcanic rocks of East Africa and these sands are carried around the coast of the Mediterranean between the Delta and the Bay of Haifa. The sands of this region therefore have a distinctive isotopic signature which distinguishes them from most other potential sources in the Mediterranean and Europe. In a series of publications, Degryse and coworkers have shown that not all natron glass has Nd isotopes characteristic of eastern Mediterranean sources, possibly indicating the existence of other, so far unknown glass producing regions elsewhere in the Late Roman world (Degryse 2014).

Trace elements are increasingly coming into play also in the investigation of other glass types. For example Indian beads have been shown to fall into high uranium and low uranium types, of as yet unclear origins (Lankton and Dussubieux 2006), while Renaissance glass in Venetian style made in the Netherlands or Britain differs from that made in Venice (Janssens et al. 2013). We still have progress to make in our understanding of trace and isotopic data, however. According to Wedepohl and Baumann (2000) fourth century Roman glass from the Hambacher Forest in north-western Germany has a compositional profile characteristic of local production, but its composition and chronology is remarkably similar to HIMT glass, found across the Roman Empire, and probably made in Egypt (Nenna 2014, Foster and Jackson 2009).

4.2. Recycling

Intimately linked to the organisation of glass making and glass use is the issue of glass recycling. In particular in areas removed from primary glassmaking centres the recycling of broken glass would have played a significant role in sustaining provision of newly-formed artefacts, and a substantial body of evidence, both archaeological and textual, points to the importance of this in the Roman (Keller 2004; Silvestri 2008; Silvestri et al. 2008) and Islamic worlds (e.g. the cullet carried by the Serge Limani wreck; see Bass et al. 2009). Less clear is how to recognise the extent of recycling in the composition of the glass; early work by Jackson (1996) identified elevated trace amounts of the transition metals that were used for colouring purposes as good indicators. This approach clearly shows that by the 9th to 10th centuries A.D., a substantial proportion of the glass in use in western Europe comprised recycled coloured glass, probably removed as mosaic tesserae from Roman public buildings (Fig. 6; e.g. Schibille and Freestone 2013). However, not all recycled glass would include coloured cullet, and what constitutes elevated levels anyway? Additional indicators, such as increased concentrations of fuel ash components, can contribute to the discussion (Paynter 2008; Rehren et al. 2010), but ash can contaminate glass already during the primary stage. Similarly, while mixing lines have been demonstrated based upon isotopic compositions (Degryse et al. 2006) or the concentration of antimony Silvestri (2008), much foundation work still needs to be done before this aspect can be better understood.

409 Fig. 6: Lead and copper contents of natron glass from Anglo-Saxon Hamwic (Southampton, 9th-10th

410 century; data of Hunter and Heyworth 1998) compared with Roman London (2nd century; data of

411 Freestone et al. in press). Note the elevated lead and copper contents in the early medieval glass,

412 indicating the incorporation of substantial quantities of recycled coloured glass.

414 Even in primary furnaces, the glass has been shown to be inhomogeneous (Freestone et al.

415 2000) and this means that consignments of glass chunks sent to different workshops for

416 remelting and fabrication will have varied in average composition from batch to batch. The

417 recycling of old glass will add to the variability between batches. It has been shown that

418 individual batches of vessels may be recognised through analysis of their major elements

419 (Price et al. 2005; Freestone et al. 2009), while trace element analysis of material in a single

420 workshop may identify the products of different campaigns or individual melts (Freestone et

421 al. in press). This opens a way to a finer understanding of consumer assemblages and of a

422 hierarchy of secondary workshops (e.g. Cholakova in press).

424 5. Looking to the future

425 The kaleidoscopic plethora of isolated case studies has begun to give way to the

426 development of larger pictures, as seen for instance in the development of organisational

427 models, and a better understanding of long-distance movement of glass for instance in

428 Central Asia (Nakai and Shindo 2013), SE Asia (Lankton and Dussubieux 2006), or into North

429 America (Hancock et al. 1999). Just as the proverbial crystal ball provides a seductive but not

430 necessarily accurate view of the future, heavily biased by the preconceptions of the gazer,

431 so our emerging picture of the ancient glass industry although clearer, remains vague and

432 potentially erroneous.

433 Specifically, there are still large geographical and chronological gaps to fill in our knowledge

434 of prevailing glass compositions; such as in the early Iron Age to Hellenistic periods in the

435 Mediterranean; or in much of Central, East and Southeast Asia. Beyond this filling of gaps,

436 however, we believe that major debates would be helpful in a number of key areas. Firstly,

437 how and why was glass invented? Persistent reports of glass finds dated to the 3rd

438 millennium BC (Peltenburg 1987) point to a precursor tradition of glass making before the

439 emergence of regular glass production in the mid-2nd millennium BC in Mesopotamia. Even

440 after this, there were multiple events of innovation that helped spread glass further, such as

441 the re-emergence of a glass industry in the Mediterranean Iron Age after the collapse of the

442 Late Bronze Age world system, or the emergence of new and seemingly indigenous glass

443 making industries in Han China, medieval Europe, India and Nigeria. To what extent were

444 these inspired by the knowledge of imported glass, and how did they develop their own,

445 often fundamentally different, recipes? Linked to this question is the ongoing need to map

446 the long-distance movement of glass, both as finished artefacts and in raw or ingot form, in

447 order to understand the social and economic drivers of these movements. Furthermore, the

chronologies of change are imprecise and need refining, as they are inadequate to understand the relationships between material change or choice, and wider social context.

Secondly, the increasing availability of analytical equipment poses new challenges. The use of hand-held, portable XRF instruments leads to an increase of the quantity of data, but major work is still required to ensure its quality. This does not only concern the analytical capability and proper calibration of the instruments themselves, but also a much clearer understanding of the limitations that come from the surface quality of the analysed objects, and the increasing pressure for non-invasive analysis. Despite its limitations, pXRF has huge potential to revolutionise glass analyses in certain settings (e.g. Dungworth 2012). On the other end of the spectrum we see a major step change in analytical resolution through the routine application of LA-ICPMS with its sub-ppm sensitivity and often astounding accuracy, opening up a window onto trace element signatures the meaning of which we still need to fully explore (Gratuze 2013). Similar fundamental progress is being made with new stable isotope studies, most notably 144Nd/143Nd in conjunction with the established Sr and Pb systems (Degryse 2014). Here too we still need to develop a better understanding of the potential of these systems.

Last, but not least, there is an enduring need to better integrate the growing data across the disciplines, integrating the analytical data into meaningful archaeological and anthropological research frameworks and combining it with similarly refined typological and craft studies. In view of inter-regional distribution of massive amounts of chemically relatively homogenous glass it is the identification of workshop traits or particular craft styles that can offer increasing resolution to the hierarchy of glass workshops, from massive primary production centres to elite workshops producing prestige items and large-scale specialist units for mass production, to small shops recycling and working, possibly even making, all sorts of glass depending on their availability and the local demand for objects. Tracing this flow of glass from the production sites to the smallest workshop is another promising topic of research. The role of local landowners and merchants in organising the production and distribution of glass as a major commodity is poorly understood; but even on the level of individual artefacts glass plays a role in daily commerce, be it as a measure for volumes (Fig. 7), or as a tamper-proof material, such as for the Islamic stamped glass weights, or as beads taking on the role of currency in non-monetary societies. Studying this, of course, requires the close integration of scientific approaches into archaeological, anthropological, social and economic research frameworks, as well as the full publication of the archaeological, typological and analytical data for each project. We are very much looking forward to the progress made in this respect by the next generation of scholars.

Fig. 7: Series of seven square bottles from Pompeii (1st c AD), presumably used to contain known measures of a liquid commodity such as oil or wine. The capacity doubles with each step from the smallest to the largest. ©De Agostini/The British Library Board.

489 Acknowledgements

490 We owe a huge debt of gratitude to numerous people who have inspired, supported and

491 challenged us over the last decades, as students, senior colleagues, personal friends and

492 anonymous reviewers. We acknowledge a plethora of institutional support, through access

493 to collections and equipment, core budgets and competitive grants, which has enabled us

494 and our colleagues to study such a fascinating material. Ultimately, all our work is based on

495 public funding, and concerns material which is part of our common cultural heritage; the

496 results of this should be freely available to all. We are grateful to the British Library Board

497 for granting permission to use Figs. 4a and 7 in this article.

499 References

500 Aldsworth, F., Haggarty, G., Jennings, S., Whitehouse, D. 2002. Medieval glass making at

501 Tyre. Journal of Glass Studies 44, 49-66

502 Angelini, I., Artioli, G., Bellintani, P., Diella, V., Gemmi, M., Polla, A., Rossi, A. (2004).

503 Chemical analyses of Bronze Age glasses from Frattesina di Rovigo, northern Italy.

504 Journal of Archaeological Science, 31, 1175-1184.

505 Arletti, R., Quartieri, S., Freestone, I. C. (2013). A XANES study of chromophores in

506 archaeological glass. Applied Physics A 111, 99-108.

507 Ashtor, E. and Cevidalli, G. 1983. Levantine alkali ashes and European Industries, Journal of

508 European Economic History 12, 475-522

509 Barber, D., Freestone, I., Moulding, K. (2009). Ancient copper red glasses: investigation and

510 analysis by microbeam techniques. In: Shortland, A., Freestone, I. & Rehren, Th. (eds):

511 From Mine to Microscope - Advances in the Study of Ancient Technology. Oxford, 115512 127.

513 Barkoudah, Y., Henderson, J. (2006). Plant ashes from Syria and the manufacture of ancient

514 glass: ethnographic and scientific aspects. Journal of Glass Studies, 48, 297-321.

515 Bass, G. F., Brill, R., Lleddo, B., & Matthews, S. (2009) Serce Limani, Volume 2: The glass

516 of an eleventh century shipwreck. Texas: Texas A&M University Press

517 Bingham, P. A., & Jackson, C. M. (2008). Roman blue-green bottle glass: chemical-optical

518 analysis and high temperature viscosity modelling. Journal of Archaeological Science, 35,

519 302-309.

520 Brems, D., Degryse, P., Hasendoncks, F., Gimeno, D., Silvestri, A., Vassilieva, E., Luypaers,

521 S. & Honings, J. (2012). Western Mediterranean sand deposits as a raw material for

522 Roman glass production. Journal of Archaeological Science 39, 2897-2907.

523 Brill, R., 1970. The chemical interpretation of the texts. In: Oppenheim, L., Brill, R., Barag,

524 D., von Saldern, A. (Eds.), Glass and Glassmaking in Ancient Mesopotamia. Corning

525 Museum of Glass. Corning, New York, pp. 105-128.

526 Brill, R.H., Tong, S.S.C., Dohrenwend, D. (1991) Chemical analyses of some early Chinese

527 glasses. In Scientific Research in Early Chinese Glass (R.H. Brill and J.H. Martin, eds.)

528 New York, Corning Museum of Glass, 31-58.

Brun, N., Pernot, M. 1992. The opaque red glass of Celtic enamels from Continental Europe. Archaeometry 34, 235-252.

Caley, E.R. (1962). Analysis of ancient glasses 1790-1957; A comprehensive and critical survey. Corning Museum of Glass.

Cholakova, A., (in press). Networks of distribution at the margins of the Empire - late antique glass vessels from the Lower Danube region. - In: V. Kassianidou and M. Dikomitou-Eliadou (eds) The NARNIA Project: Integrating Approaches to ancient material studies. Nicosia, Cyprus.

Clow, A., Clow, N. 1956. The timber famine and the development of technology. Annals of Science 12, 85-102.

Crossley, D., 1998, The English glassmaker and his search for raw materials in the 16th and 17th centuries. In: P. McCray & D. Kingery (eds.), The Prehistory and History of Glassmaking Technology, 167-179.

Degryse, P. 2014 Glass Making in the Greco-Roman World: Results of the ARCHGLASS Project. Leuven University Press: Studies in Archaeological Sciences.

Degryse, P., Schneider, J., Haack, U., Lauwers, V., Poblome, J., Waelkens, M., Muchez, P. (2006). Evidence for glass 'recycling'using Pb and Sr isotopic ratios and Sr-mixing lines: the case of early Byzantine Sagalassos. Journal of Archaeological Science, 33(4), 494-501.

Devulder, V., Vanhaecke, F., Shortland, A., Mattingly, D., Jackson, C., Degryse, P. (2014). Boron isotopic composition as a provenance indicator for the flux raw material in Roman natron glass. Journal of Archaeological Science, 46, 107-113.

Dungworth, D. (2012). Historic Window Glass-The Use of Chemical Analysis to Date Manufacture. Journal of Architectural Conservation, 18, 7-25.

Dungworth, D., Brain, C., 2009 Late 17th-Century Crystal Glass: An Analytical Investigation. Journal of Glass Studies 51, 111-137.

Dussubieux, L., Gratuze, B., Blet-Lemarquand, M. (2010). Mineral soda alumina glass: occurence and meaning. Journal of Archaeological Science 37, 1646-1655.

Eramo, G., 2006, The glass-melting crucibles of Derriere Sairoche (1699-1714 AD, Ct. Bern, Switzerland): a petrological approach. Journal Archaeological Science 33, 440-452.

Foster, H., Jackson, C.M. 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, 189-204.

Foy, D., Picon, M., Vichy, M., Thirion-Merle, V., 2003. Caractérisation des verres de la fin de l'Antiquité en Méditerranée occidentale: l'émergence de nouveaux courants commerciaux, in: Foy, D., Nenna, M.-D. (Eds.), Échanges et commerce du verre dans le monde antique. Monographies Instrumentum 24, Éditions Monique Mergoil, Montagnac, pp. 41-85.

Freestone, I.C. 2006a. An indigenous technology? A commentary on Lankton et al. "Early Primary Glass Production in Southern Nigeria." Journal of African Archaeology 4: 139141.

569 Freestone I.C. (2006b) Glass production in Late Antiquity and the Early Islamic period: a

570 geochemical perspective. Geomaterials in Cultural Heritage M. Maggetti and B. Messiga

571 (eds) Geological Society of London Special Publication 257, 201-216.

572 Freestone I.C., Stapleton C. P. (2014) Composition, technology and production of coloured

573 glasses from Roman mosaic vessels. In Bayley J., Jackson C. M. and Freestone I.C. (eds.)

574 Glass of the Roman World. Oxbow, pp. 62-77.

575 Freestone I.C., Gorin-Rosen Y., Hughes M.J. (2000) Composition of primary glass from

576 Israel in Ateliers primaires et secondaires de verriers du second millinaire av. J.-C. au

577 Moyen-Age, M.-D. Nenna (ed.), Travaux de la Maison de l'Orient Méditerranéen no. 33,

578 65-84.

579 Freestone, I.C., Greenwood, R., Gorin-Rosen, Y., 2002a. Byzantine and early Islamic

580 glassmaking in the Eastern Mediterranean: production and distribution of primary glass,

581 in: Kordas, G. (Ed.), Hyalos - Vitrum - Glass, Athens, pp. 167-174

582 Freestone, I.C.,Ponting, M., Hughes, M., 2002b: The origins of Byzantine glass from Maroni

583 Petrera, Cyprus. Archaeometry 44, 257-272.

584 Freestone I.C., Price J., Cartwright C. (2009) The batch: its recognition and significance,

585 Annales 17th Congress AIHV pp. 130-135.

586 Freestone, I.C., Kunicki-Goldfinger, J.J., MacDonald, I., Gutjahr, M., Pike, A. (in press)

587 Composition, Technology and Origin of the Glass. In: Wardle, A., (ed.) Roman life on the

588 margins: excavations at 35 Basinghall Street, City of London. Museum of London

589 Archaeology.

590 Gaitzsch, W., Follmann-Schulz, A.-B., Wedepohl, K.H., Hartmann, G., Tegtmeier, U. 2000.

591 Spätrömische Glashütten im Hambacher Forst - Produktionsort der ECVA-Fasskrüge.

592 Bonner Jahrbücher 200, 83-241.

593 Gan, F.X., 2005. Evolution of the chemical composition and the origin of making technology

594 of Chinese ancient glass. In: Gan, F.X. (Ed.), Development of Chinese Ancient Glass.

595 Shanghai Scientific and Technical Publishers, pp. 220-240 (in Chinese).

596 Gliozzo, E., Santagostino Barbone, A., D'Acapito, F., 2013. Waste glass, vessels and

597 window-panes from Thamusida (Morocco): grouping natron-based blue-green and

598 colourless Roman glasses. Archaeometry 55, 609-639.

599 Gorin-Rosen, Y., 2000. The ancient glass industry in Israel: summary of finds and new

600 discoveries, in: Nenna, M.-D. (Ed.), La Route du verre. Ateliers primaires et secondaires

601 du second millenaire av. J.-C. au Moyen Age, Travaux de la Maison de l'Orient 33, Lyon,

602 pp. 49-63.

603 Gratuze, B. 2013. Glass Characterisation Using Laser Ablation Inductively Coupled Plasma

604 Mass Spectrometry Methods. In: K. Janssens (ed), Modern Methods for Analysing

605 Archaeological and Historical Glass, John Wiley and Sons, 201-234.

606 Gratuze, B., Barrandon, J.-N., 1990, Islamic glass weights and stamps: analysis using nuclear

607 techniques, Archaeometry, 32, 155-62.

608 Gratuze, B., Janssens, K. (2004) Provenance analysis of glass artefacts. Pp. 663-712 In

609 Janssens K., R. Van Grieken (2004) Non-destructive microanalysis of cultural heritage

610 materials.

611 Gratuze, B., Soulier, I., Blet, M., Vallauri, L. (1996). De l'origine du cobalt: du verre à la

612 céramique. Revue d'archéométrie, (20), 77-94.

613 Hancock, R., Aufreiter, S., Kenyon, I., Latta, M. 1999. White Glass Beads from the Auger

614 Site, Southern Ontario, Canada. Journal of Archaeological Science 26, 907-912.

615 Henderson, J., Warren, S.E. 1981. X-ray fluorescence analyses of Iron Age glass: beads from

616 Meare and Glastonbury Lake villages. Archaeometry 23, 83-94.

617 Henderson, J., McLoughlin, S. D., McPhail, D. S. (2004). Radical changes in Islamic glass

618 technology: evidence for conservatism and experimentation with new glass recipes from

619 early and middle Islamic Raqqa, Syria. Archaeometry, 46(3), 439-468.

620 Hunter, J R., Heyworth, M P. (1998) The Hamwic Glass. York: Council for British

621 Archaeology Research Report 116.

622 Ige, A., Ogundiran, A., Rehren, Th., Dussubieux, L., Kusimba, C. (submitted): Trace element

623 data demonstrate granite-derived origin of Yoruba glass in Osun Grove (Osogbo), Nigeria.

624 Jackson, C.M., 1996, From Roman to early medieval glasses. Many happy returns or a new

625 birth?, in Annales du 13e Congrès de l'Association Internationale pour l'Histoire du

626 Verre, 289-301, AIHV, Lochem.

627 Jackson, C.M., Joyner, L., Booth, C.A., Day, P.M., Wager, E.C.W., Kilikoglou, V. (2003).

628 Roman glass-making at Coppergate, York? Analytical evidence for the nature of

629 production. Archaeometry, 45, 435-456.

630 Jackson, C.M., Nicholson, P. 2010. The provenance of some glass ingots from the Uluburun

631 shipwreck. Journal of Archaeological Science 37, 295-301.

632 Jackson, C.M., Smedley, J.W. 2008. Theophilus and the use of beech ash as a glassmaking

633 alkali. In: M. Martinon-Torres, M., Rehren, Th. (eds) Archaeology, History and Science -

634 Integrating Approaches to Ancient Materials, Walnut Creek, Left Coast Press, 117-130.

635 Janssens, K. (ed) 2013. Modern Methods for Analysing Archaeological and Historical Glass,

636 Wiley and Sons.

637 Janssens, K., Cagno, S., De Raedt, I., Degryse, P. (2013). Transfer of Glass Manufacturing

638 Technology in the Sixteenth and Seventeenth Centuries from Southern to Northern

639 Europe: Using Trace Element Patterns to Reveal the Spread from Venice via Antwerp to

640 London. Modern Methods for Analysing Archaeological and Historical Glass, Volume I,

641 537-562.

642 Keller, D. 2004. Social and economic aspects of glass recycling. In Bruhn J, Croxford B and

643 Grigoropoulos D TRAC 2014, Proceedings of the Fourteenth Annual Theorectical Roman

644 Archaeology Conference, Oxford: Oxbow, 65-78.

645 Kock, J., Sode, T., 1996. Glass, glassbeads and glassmakers in Northern India. Thot Print.

646 Kuçukerman, O. 1985, The Art of Glass and traditional Turkish glassware. Istanbul: Turkiye

647 Is Bankasi Kultur Yayinlari.

648 Lahlil, S., Biron, I., Cotte, M., & Susini, J. (2010). New insight on the in situ crystallization

649 of calcium antimonate opacified glass during the Roman period. Applied Physics A,

650 100(3), 683-692.

Lahlil, S., Cotte, M., Biron, I., Szlachetko, J., Menguy, N., Susini, J. (2011). Synthesizing lead antimonate in ancient and modern opaque glass. Journal of Analytical Atomic Spectrometry, 26(5), 1040-1050.

Lankton, J., Dussubieux, L. (2006). Early glass in Asian maritime trade: a review and an interpretation of compositional analyses. Journal of Glass Studies, 48, 121-144.

Lankton, W. J., Ige, O. A.,Rehren, Th. 2006. Early primary glass production in southern Nigeria. Journal of African Archaeology 4: 111-138.

Mecking, O., 2013. Medieval lead glass in Central Europe. Archaeometry 55, 640-662.

Meulebroeck, W., Cosyns, P., Baert, K., Wouters, H., Cagno, S., Janssens, K., Terryn, H., Nys, K., Thienpont, H. (2011), Optical spectroscopy as a rapid and low-cost tool for the first-line analysis of glass artefacts: a step-by-step plan for Roman green glass. Journal of Archaeological Science 38, 2387-2398.

Mirti, P., Pace, M., Malandrino, M., Negro Ponzi, M. 2009. Sasanian glass from Veh Ardasir: new evidences by ICP-MS analysis. Journal of Archaeological Science 36, 1061-1069.

Moretti, C., Zecchin, P., 2009. English lead crystal and Ravenscroft's formulation: additional information from Venetian sources. Annales du 17e Congrès de l'AIHV, 431-434.

Nakai, I., Shindo, Y. 2013. Glass trade between the Middle East and Asia. In: Janssens, K. (ed) 2013, Modern Methods for Analysing Archaeological and Historical Glass, Wiley and Sons, 445-457.

Nenna, M.-D., 2014. Egyptian glass abroad: HIMT glass and its market. In: D. Keller, J. Price and 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. Oxbow Books, 177-193.

Nenna, M.-D. 2015 Primary glass workshops in Graeco-Roman Egypt: preliminary report on the excavations of the site of Beni Salama, Wadi Natrun (2003, 2005-9). In Bayley J., Jackson C. M. and Freestone I.C. (eds.) Glass of the Roman World. Oxbow Books.

Nenna, M.-D., Picon, M., Vichy, M. 2000. Ateliers primaries et secondaires en Egypte à l'époque gréco-romaine. In: Nenna, M.-D. (Ed.), La Route du verre. Ateliers primaires et secondaires du second millénaire av. J.-C. au Moyen Âge. Lyon, Maison de l'Orient et de la Méditerranée, 97-112.

Nicholson, P.T. (2012). "Stone... That Flows": Faience and Glass as Man-Made Stones in Egypt. Journal of Glass Studies 54, 11-23.

Paynter, S., 2008: Experiments in the reconstruction of Roman wood-fired glassworking furnaces: waste products and their formation processes. Journal of Glass Studies 50, 271290.

Paynter, S. 2012. The importance of pots: the role of refractories in the development of the English glass industry during the 16th/17th centuries. Annales 18th Congress AIHV, 419424.

Peltenburg, E.J., 1987. Early faience: recent studies, origins and relations with glass. In Bimson M. and Freestone, I.C. (eds) Early Vitreous Materials. London: British Museum Occasional Paper 56, 5-29.

Price, J., Freestone, I.C., Cartwright, C.R., 2005. 'All in a day's work?' The colourless cylindrical glass cups found at Stonea revisited. In N Crummy ed., Image, Craft and the

694 Classical World. Essays in honour of Donald Bailey and Catherine Johns. Collection

695 'Monographies Instrumentum' no 29. Montagnac, éditions Monique Mergoil, 163-169.

696 Pusch, E., Rehren, Th., 2007. Rubinglas für den Pharao. Forschungen in der Ramses-Stadt,

697 Band 6. Gerstenberg Verlag, Hildesheim. Vol 1, Text.

698 Rehren, Th., 1997. Ramesside glass colouring crucibles. Archaeometry 39, 355-368.

699 Rehren, Th., 2014. Glass Production and Consumption between Egypt, Mesopotamia and the

700 Aegean. In: P. Pfälzner, H. Niehr, E. Pernicka, S. Lange and T. Köster (eds),

701 Contextualising Grave Inventories in the Ancient Near East. (Qatna Studien Supplementa

702 3, Harrassowitz, Wiesbaden, 217-223.

703 Rehren, Th., Pusch, E., 2005. Late Bronze Age Egyptian glass production at Qantir-

704 Piramesses. Science 308, 1756-1759.

705 Rehren, Th., Pusch, E., 2008. Crushed rock and molten salt? Some aspects of the primary

706 glass production at Qantir / Pi-Ramesse. In: Jackson, C.M. & Wager, E. (eds), Vitreous

707 Materials in the Late Bronze Age Aegean: A Window to the East Mediterranean World.

708 Oxbow Books, Oxford, 14-33.

709 Rehren, Th., Connolly, P., Schibille, N., Schwarzer, H., 2015. Changes in glass consumption

710 in Pergamon (Turkey) from Hellenistic to late Byzantine and Islamic times. Journal of

711 Archaeological Science (accepted)

712 Rehren, Th., Marii, F., Schibille, N., Stanford, L., Swan, C., 2010. Glass supply and

713 circulation in early Byzantine southern Jordan. In: J. Drauschke and D. Keller, eds., Glass

714 in Byzantium - production, usage, analyses. RGZM, Mainz, 65-81.

715 Rosenow, D., Rehren, Th. 2014. Herding cats - Roman to Late Antique glass groups from

716 Bubastis, northern Egypt. Journal of Archaeological Science 49, 170-184.

717 Sayre, E.V., 1963. The intentional use of antimony and manganese in ancient glasses. In F. R.

718 Matson and G. E. Rindone (eds.) Advances in Glass Technology, Part 2 New York:

719 Plenum Press, 263-282.

720 Schibille, N., Freestone, I.C. (2013) Composition, Production and Procurement of Glass at

721 San Vincenzo al Volturno: An Early Medieval Monastic Complex in Southern Italy. PLoS

722 ONE 8(10): e76479. doi:10.1371/journal.pone.0076479

723 Schreurs, J., Brill, R., 1984. Iron and sulphur related colors in ancient glasses. Archaeometry

724 26, 199-209.

725 Shortland, A. J., 2002. The use and origin of antimonate colourants in early Egyptian glass.

726 Archaeometry 44, 517-530.

727 Shortland, A.J., 2012. Lapis Lazuli from the Kiln: Glass and Glassmaking in the Late Bronze

728 Age. Leuven University Press.

729 Shortland, A.J., Tite, M.S., Ewart, I., 2006. Ancient Exploitation and Use of Cobalt Alums

730 from the Western Oases of Egypt, Archaeometry, 48, 153-168.

731 Shortland, A.J., Rogers, N., Eremin, K., 2007. Trace element discriminants between Egyptian

732 and Mesopotamian late Bronze Age glasses. Journal of Archaeological Science, 34, 781733 789.

734 Shortland, A.J., Schachner, L., Freestone, I. Tite, M., 2006. Natron as a flux in the early

735 vitreous materials industry: sources, beginnings and reasons for decline. Journal of

736 Archaeological Science 33, 521-530.

737 Silvestri, A. (2008). The coloured glass of IuliaFelix. Journal of Archaeological Science 35,

738 1489-1501.

739 Silvestri, A., Molin, G., Salviulo, G. (2008). The colourless glass of Iulia Felix. Journal of

740 Archaeological Science 35, 331-341.

741 Simpson, St.J. 2014. Sasanian glass: an overview. In: D. Keller, J. Price and C. Jackson (eds),

742 Neighbours and successors of Rome - Traditions of glass production and use in Europe

743 and the Middle East in the later 1 st millennium AD. Oxbow Books, 200-231.

744 Smedley, J.W., Jackson, C.M. 2002. Medieval and post medieval glass technology: A review

745 of the use of bracken in glassmaking. Glass Technology 43, 221-224.

746 Smirniou, M., Rehren, Th., 2011. Direct evidence of primary glass production in Late Bronze

747 Age Amarna, Egypt. Archaeometry 53, 58-80

748 Sode, T., Kock, J., 2001. Traditional raw glass production in northern India: the final stage of

749 an ancient technology, Journal of Glass Studies 43, 155-69.

750 Tanimoto, S., Rehren, Th., 2008. Interactions between silicate and salt melts in LBA

751 glassmaking. Journal of Archaeological Science 35, 2566-2573.

752 Tite, M., Pradell, T., Shortland, A. (2008). Discovery, production and use of tin-based

753 opacifiers in glasses, enamels and glazes from the late iron age onwards: a reassessment.

754 Archaeometry, 50(1), 67-84.

755 Tite, M., Freestone, I., Mason, R., Molera, J., Vendrell-Saz, M., Wood, N. 1998. Lead glazes

756 in Antiquity-methods of production and reasons for use. Archaeometry 40, 241-260.

757 Towle, A., Henderson, J., Bellintani, P., Gambacurta, G., 2001. Frattesina and Adria: report

758 of scientific analysis of early glass from the Veneto. Padusa XXXVII: 7-68.

759 Turner, W.E.S. 1954, Studies in ancient glass and glassmaking processes Part 1. Crucibles

760 and melting temperatures employed in ancient Egypt at about 1370 BC, Journal of the

761 Society of Glass Technology 38, 436-44.

762 Varberg, J., Gratuze, B., Kaul, F., 2015. Between Egypt, Mesopotamia and Scandinavia: Late

763 Bronze Age Glass Beads Found in Denmark. Journal of Archaeological Science 54, 168764 181.

765 Verita, M., 1985. L'invenzione del cristallo muranese: una verifica analitica delle fonti

766 storiche. Rivista Stazione Sperimentale del Vetro, 15, 17-29.

767 Wedepohl, K.H., 1998, Mittelalterliches Glas in Mitteleuropa: Zusammensetzung,

768 Herstellung, Rohstoffe, Vandenhoeck & Ruprecht, Göttingen.

769 Wedepohl, K.H., Baumann, A., 1997. Isotope composition of medieval lead glasses reflecting

770 early silver production in Central Europe, Mineralium Deposita 32, 292-295.

771 Wedepohl, K.H., Baumann, A. (2000) The use of marine molluskan shells for Roman glass

772 and local raw glass production in the Eifel area (Western Germany). Naturwissenschaften

773 87, 129-132.

774 Wedepohl, K. H., Simon, K., Kronz, A. (2011). Data on 61 chemical elements for the

775 characterization of three major glass compositions in Late Antiquity and the Middle Ages.

776 Archaeometry, 53(1), 81-102.

777 Whitehouse, D., 2002. The transition from natron to plant ash in the Levant. Journal of Glass

778 Studies 44, 193-196.

779 Zoleo, A., Brustolon, M., Barbon, A., Silvestri, A., Molin, G., Tonietto, S. (2014). Fe (III)

780 and Mn (II) EPR quantitation in glass fragments from the palaeo-Christian mosaic of St.

781 Prosdocimus (Padova, NE Italy): Archaeometric and colour correlations. Journal of

782 Cultural Heritage. doi:10.1016/j.culher.2014.07.005

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

Wt % MgO

□ Banias O Tyre

A Serce Limani • Sasanian (Brill) ■ Sasanian (BM) ▲ Sasanian (Mirti) A Raqqa Type 4

100000

£ 1000 Q_

Q_ 100

• Hamwic A Roman London

10 100 1000 Cu ppm

Highlights for

Ancient Glass: from kaleidoscope to crystal ball

• Major progress has been made in characterising ancient glass

• Larger pattern emerges based on numerous case studies

• Important research lacunae still persist