Scholarly article on topic 'Herding cats – Roman to Late Antique glass groups from Bubastis, northern Egypt'

Herding cats – Roman to Late Antique glass groups from Bubastis, northern Egypt 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 / "Chemical composition" / Roman / "Late Antique" / Egypt}

Abstract of research paper on History and archaeology, author of scientific article — D. Rosenow, Th. Rehren

Abstract Eighty-seven glass fragments from Roman and Late Antique layers at Tell Basta/Bubastis in the Eastern Nile Delta were typologically evaluated and chemically analysed to determine chronological and compositional patterns of glass use at this important Egyptian city, and how this relates to larger pattern of glass production and consumption in the first half of the first millennium AD. Bubastis is situated in geographical proximity to Alexandria, an important seaport, and at the same time close to the raw glass production areas in the Wadi Natrun and Sinai peninsula. This paper reports the first substantial set of compositional data of Roman to Late Antique glass from a settlement in northern Egypt, filling an important gap in our knowledge of glass consumption pattern in the first half of the first millennium AD. The glass from Bubastis falls into several compositional groups known already from elsewhere in the Roman and Late Antique world, including antimony- and manganese-decoloured glass and two varieties of HIMT glass. Changes in glass composition over more than 500 years are in line with earlier observations concerning changes in prevalence of these glass groups. However, compositional groups known to dominate archaeological glass assemblages elsewhere, such as Roman blue/green during the earlier part of the period under study, or Levantine I in the later period, are notably absent. For the later period, this is probably due to the proximity of Tell Basta to the suspected production region of HIMT glass in northern Sinai/Egypt. By analogy, this might indicate that the earlier Roman blue/green glass has a production origin further away from the Delta than the decolourised glasses prevailing in Bubastis. A particular vessel type, small-volume thick-walled dark green unguentaria, is made of probably Egyptian plant ash glass, indicating the existence of a specialised glassmaker during the early first millennium AD.

Academic research paper on topic "Herding cats – Roman to Late Antique glass groups from Bubastis, northern Egypt"

Accepted Manuscript

Herding cats - Roman to Late Antique glass groups from Bubastis, northern Egypt D. Rosenow , Th Rehren

PII: S0305-4403(14)00160-5

DOI: 10.1016/j.jas.2014.04.025

Reference: YJASC 4058

To appear in: Journal of Archaeological Science

Received Date: 22 July 2013 Revised Date: 29 April 2014 Accepted Date: 30 April 2014

Please cite this article as: Rosenow, D, Rehren, T., Herding cats - Roman to Late Antique glass groups from Bubastis, northern Egypt, Journal of Archaeological Science (2014), doi: 10.1016/j.jas.2014.04.025.

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.

Herding cats - Roman to Late Antique glass groups from Bubastis, northern Egypt

D Rosenow, UCL Institute of Archaeology Th Rehren, UCL Qatar

Abstract

Eighty-seven glass fragments from Roman and Late Antique layers at Tell Basta/Bubastis in the Eastern Nile Delta were typologically evaluated and chemically analysed to determine chronological and compositional patterns of glass use at this important Egyptian city, and how this relates to larger pattern of glass production and consumption in the first half of the first millennium AD. Bubastis is situated in geographical proximity to Alexandria, an important seaport, and at the same time close to the raw glass production areas in the Wadi Natrun and Sinai peninsula. This paper reports the first substantial set of compositional data of Roman to Late Antique glass from a settlement in northern Egypt, filling an important gap in our knowledge of glass consumption pattern in the first half of the first millennium AD. The glass from Bubastis falls into several compositional groups known already from elsewhere in the Roman and Late Antique world, including antimony- and manganese-decoloured glass and two varieties of HIMT glass. Changes in glass composition over more than 500 years are in line with earlier observations concerning changes in prevalence of these glass groups. However, compositional groups known to dominate archaeological glass assemblages elsewhere, such as Roman blue/green during the earlier part of the period under study, or Levantine I in the later period are notably absent. For the later period, this is probably due to the proximity of Tell Basta to the suspected production region of HIMT glass in northern Sinai/Egypt. By analogy, this might indicate that the earlier Roman blue/green glass has a production origin further away from the

Delta than the decolourised glasses prevailing in Bubastis. A particular vessel type, small-volume thick-walled dark green ungüentaría, is made of probably Egyptian plant ash glass, indicating the existence of a specialised glassmaker during the early first millennium AD.

1. Introduction

The composition of Hellenistic to Byzantine glass is characterised by a surprising degree of

fundamental similarity and consistency over more than a thousand years (Sayre and Smith

1961), which may be explained at least in part by a combination of faithfully maintained

traditional recipes using tightly controlled raw materials, and partly by the self-governing

behaviour of the melt-forming soda-lime-silica system (Rehren 2000; Tanimoto and Rehren

2008). Within this broad homogeneity, however, there are well-developed and long recognized

specific compositional groups, characterised by their minor oxide and trace element contents. It

is generally assumed that the minor oxide and trace element contents of ancient glass reflect the

composition of the sand used in its production (e.g., Freestone 2006), while the soda levels are

determined by batch recipes. For the first four centuries AD several 'Roman' glass groups have

been established, mostly through the analysis of samples from Italy and the Northern provinces

(e.g. Jackson 2005; Silvestri et al. 2008; Foy et al. 2003). The most common glass there is

naturally blue/green coloured, with no intentional additives to manipulate its colour; this natural

colour is due to the iron impurities in the sand and the prevailing redox conditions in the

glassmaking furnace. It is often referred to as 'aqua', to distinguish it from glass intentionally

coloured blue or green through the addition of metal oxides. Colourless Roman glass is

characterised by the addition either of antimony or manganese oxide to counter-act the

colouring effect of iron oxide, or a combination of both oxides (Jackson 2005, Silvestri et al.

2008, Foster and Jackson 2010). Antimony-decoloured glass is typically dated earlier than

manganese-decoloured glass; substantial data sets have been published, among others, by Paynter (2006) for glass from Britain and Silvestri et al. (2008) from a ship wreck in the northern Adria. Roman blue/green glass is generally considered to be of Levantine origin (Nenna et al. 1997), while decoloured glass is linked to a production in northern Egypt (Nenna 2007). For the mid to late first millennium AD, five main compositional glass groups have been identified, mostly through analysis of glasses from the eastern Mediterranean region. These include Egyptian I and II, Levantine I and II and HIMT glass (Freestone et al. 2005). The first four groups can be associated with raw glass production centres in Egypt (Wadi Natrun and Ashmunein; Nenna 2007) and the Levant (bay of Haifa and Bet Eli'ezer; Gorin-Rosen 2000), respectively. The production region of HIMT glass cannot be located precisely, but is thought to be in northern Egypt, possibly the northern coast of the Sinai (Foy et al. 2003; Freestone et al. 2005). Levantine and HIMT glass has been discovered at numerous sites and regions throughout the Roman Empire, while published evidence for Egyptian I glass is relatively rare outside Egypt. Significantly, the major compositional groups have distinct chronological ranges, indicating that each production site only had a limited period of activity, spanning a few centuries. According to Freestone et al. (2000) and Freestone (2005), HIMT was mostly in circulation from the late fourth to the sixth centuries AD, Levantine I during the fourth to seventh centuries AD, Levantine II during the seventh to eighth centuries AD. Egyptian I was in use from an as yet unknown start date up to the eighth century AD, while Egyptian II was the predominant glass in the Levant during the eighth and ninth centuries AD (Freestone et al. 2000). Very little is known, however, about the relative proportions of these various glass groups in northern Egypt, the heartland of early glass making, restricting our ability to discuss the organisation of 1st millennium AD glass making and consumption.

The link between regionally different sand compositions and the minor oxide and trace element content of ancient glass provides a promising tool to explore the relationship between production origin and regions of glass consumption. Two competing models have been put forward, supporting either a more localised or dispersed system of raw glass production (Wedepohl and Baumann 2000; Baxter et al. 2005; Degryse and Schneider 2008; Silvestri et al. 2008; Foster and Jackson 2010) or a more centralised one (Foy and Jezegou 1997; Foy et al. 2000; Freestone et al. 2000; Picon and Vichy 2003; Paynter 2006; Nenna 2007). The first model assumes the existence of a range of regional primary glass production centres, not exclusive to the Levant or Egypt, but also including sites in e.g. Italy (Silvestri et al. 2008) or the northern provinces (Wedepohl and Baumann 2000; Jackson 2005). The centralised system on the other hand supports the idea of raw glass production on a large scale at only a small number of locations at any one time. From the 4th century AD onward there is good archaeological and compositional evidence for a strongly centralised production of glass in large scale factories on the Eastern Mediterranean shores, both in Egypt and the Levant (Gorin-Rosen 2000; Freestone et al. 2000; Picon and Vichy 2003; Nenna 2007), from where the raw glass would then have been sent as chunks to secondary glass working furnaces across the Empire for artefact production serving local or regional markets. It is less clear, however, whether this system also holds for the first three centuries AD, and this study aims to throw some light on this issue.

2. Introduction to the site

The ancient city of Bubastis (Tell Basta) is an Egyptian site of major historical and cultural

significance, with continuous occupation ranging from the Old Kingdom (2686-2160 BC) to Late

Antiquity (6th century AD). It is situated on the south-eastern edge of Zagazig in the Eastern Nile

Delta (Fig. 1), and is best known for its temple dedicated to the Egyptian cat goddess Bastet. Its

visible remains date to the Third Intermediate Period (1069-664 BC) and Late Period (664-343 BC) (all dates following Shaw 2000). Bubastis still played a significant cultic role during Early Ptolemaic times, but at some point after the late 3rd century BC the temple collapsed, probably due to an earthquake. In the aftermath, and for sure during Late Roman times, it was used as a quarry, and its significance as a cultic centre seems to have waned. However, the city continued to be well integrated into the Roman world, as indicated by the presence of numerous imported ceramic vessels.

The Tell Basta Project1 spent the last six years exploring the area east of the temple, where -following Herodotus' description of the ancient city - the settlement of Bubastis was situated. This area of approximately 40 hectares remains completely unexcavated. A survey in 2008 revealed a large number of objects dating to the Graeco-Roman Period, including numerous glass fragments of the Roman and Late Antique periods. Earlier excavations in the zone connecting the temple and the settlement (so-called Area A) brought to light remains of a Roman monument (Habachi 1957, 93-94). To the east and south, remains of domestic and semiofficial buildings of red bricks were unearthed, clearly connected to the Roman edifice. These building remains can be ascribed to a period when the temple, after its collapse, had not been in use anymore as a cult place. Only the contexts closest to the Roman limestone monument revealed glass finds, probably belonging to a period of subsequent use or reuse of the temple.

1 The Tell Basta Project is a German-British-Egyptian Joint Mission, funded by the Egypt Exploration Society and directed by Eva Lange.

The amount of glass discovered in deposits further away is negligible, and probably represents finds that have been accidentally moved during the last centuries; due to their uncertain archaeological origin these are not included in this study.

2.1 Glass at Bubastis

About 2500 glass fragments have so far been recorded at Tell Basta. The pieces studied here originate primarily from three contexts: the unexplored area east of the temple, the area connecting the temple and the settlement (Area A), and the entrance court of the sanctuary (Fig. 2). All glass fragments discovered during the 2008 survey are surface finds and have no specific archaeological contexts.

The overwhelming majority of glass finds, approximately 90%, are from 10 grid squares (10 x 10m) in Area A, most of them coming from the uppermost layers covering or surrounding the limestone monument and red-brick buildings. They are associated with other artefacts and pottery dating from the 3rd century BC to Late Antiquity. About 10% of the finds have been excavated from the entrance court of the temple of Bastet. These fragments derive also from heavily disturbed contexts, with some associated ceramic finds in deeper deposits dating from the New Kingdom to the 5th century AD.

All glass finds from Tell Basta were studied, recorded drawn and typologically compared to

published parallels; the results of this will be published elsewhere. Fragments include pieces of

lamps, beakers, bowls, plates, cups, jugs, bottles, jars, flasks, goblets, oval dishes, and small and

large containers. In addition, intensively coloured bracelets, beads and counters/gaming pieces

were recovered. The majority of the assemblage belongs to vessel types representing utilitarian

ware for daily use. This is consistent with the ceramic evidence from Area A, indicating that at

this time the temple of Bastet was no longer used as an active place of worship. Luxury glass is

thus scarce, with just a few fragments of millefiori glass dishes, facet-cut colourless glass or indented beakers. Only very few pieces can be related to secondary production processes, such as wasters, moils or chunks.

A few fragments represent mould cast vessels such as cast ribbed bowls, plates or bowls made of millefiori glass, or rims and bases from cast bowls and plates. However, free blown glass is by far the most dominant; some vessels are mould blown. The most usual colour within the Bubastis glass is yellowish green to deep olive green to brown, in varying shades and intensities, particularly for the typologically later material (Fig. 3). Among the earlier finds, however, many sherds are pale bluish-greenish ('aqua') to colourless. Some pieces are amber, a few finds are blue, purple or red. Due to the moist environment of the Egyptian Nile delta, corrosion is affecting the majority of glasses, and more so the earlier finds (Fig. 4); the formation of dark brown or whitish crusts can obscure the original colour and transparency of the glass. Some glass vessels are decorated by single wheel-engraved lines or bands and/or ornaments of applied blobs of blue glass, pinched elements, indents, single applied threads of the same or a different colour below the rim, or incised horizontal lines. One fragment displays facet-cut circular impressions.

The dating of the glass vessels used for this study is based on typology, since most glass was retrieved from disturbed contexts. According to parallels with dated finds from Roman and Late Antique Egypt, such as Mons Porphyrites (Bailey 2007), Kom el-Dikka (Kucharczyk 2004, 2006, 2010), Quseir al-Qadim/Myos Hormos (Meyer 1992, Peacock 2011), Bagawat (Nenna 2010, Hill/Nenna 2001), Ismant el-Kharab (Marchini 1993), several sites in the Eastern Desert (Brun 2003a, 2003b, 2011) Karanis (Harden 1936), Kom el-Nana (Faiers 2013) or Tebtynis (Nenna 2000, Foy 2001), the Bubastis corpus roughly covers a time between the first century BC to the sixth

century AD. The majority of the material dates to the second to fourth centuries AD.

3. Materials and methods

The main aims of this project are to learn more about the economic position of Bubastis during the Roman and Late Antique periods, and to improve our understanding of the distribution of specific glass compositions in space and time, particularly for Egypt. The project also aims to test the utility of portable XRF analysis to assign glass fragments to specific compositional groups, in order to be able to analyse large assemblages such as this one on site to quantify the relative importance of each glass group over time, while minimising the need for more invasive and time-consuming laboratory-based analyses. Detailed results of this will be presented elsewhere, as this topic is outside the remit of this paper.

3.1 The analysed assemblage

Eighty-seven glass vessel fragments were selected for quantitative analysis, including mould cast, free-blown and mould blown vessels (Table 1). Due to their state of preservation, a few samples could not be typologically identified. In this case, sampling was motivated by the colour of the glass. The majority of the samples are of pale greenish colour, yellow-greenish, green or colourless. Two samples are intentionally coloured blue, four samples are light blue. There is one sample each of pale purple, red-brown, burgundy and brownish pink colour. The low capacity ungüentaría are made of dark green or bluish-green 'emerald' glass.

Fragments selected for analysis are thought to reflect the whole range of glass vessel types,

manufacturing methods and decoration techniques, over the entire period of time when glass

finds are attested in Bubastis. With less than five percent of all fragments analysed, and covering

a period of more than 500 years, this is necessarily only a pilot study, and the relative

proportions of samples reported here are not representative of the types and compositions

constituting the total glass assemblage. In particular the early glasses are over-represented in the analysed corpus, with nearly all cast glass and a significant proportion of the visually identified antimony-decoloured glass analysed. In contrast, the vast abundance of late glass, typically of olive green to brown colour, is under-represented among the analyses, even though these pieces make up about two thirds of all analysed material.

3.2 Data generation and handling

EPMA was done on polished cross sections using a JEOL JXA 8100 with three spectrometers. The instrument was operated at 20kV with a beam current of 6 nA and count times of 10 seconds on the peak position and 5 seconds on the background positions. Soda loss during analysis was prevented by scanning the beam over the area visible at 800 times magnification. Table 2 reports the results of measurements of Corning A and B reference glasses analysed together with the Bubastis samples. For most oxides the measured values are within 5% of the published values; however, alumina and phosphorus oxide were consistently analysed lower than the published values, while antimony oxide was measured higher by about one third (20 and 40% in Corning A and B, respectively) of the published value (Brill 1999). No adjustment has been made for these systematic deviations in the reported data in Table 3. Concentrations of antimony oxide need to be treated with caution, and values below 0.3 wt% are not reported; these may well reflect analytical error rather than real presence of this compound.

The glasses were sorted into compositional groups based on minor oxide concentrations. Typical

values for published glass groups of the oxides of aluminium, calcium, titanium, manganese and

antimony informed a first allocation of the newly-analysed samples to the known groups; this

was then further refined by checking the levels of the remaining minor oxides for consistency

with those typically found in the established compositional groups, re-allocating samples as

necessary to obtain a subjective best fit. All but one sample were thus allocated to specific compositional groups.

4. Results

Eighty-three of the eighty-seven samples are mineral-natron based soda-lime-silica glasses (Table 3), while four fragments appear to be made from plant ash glass. Among the mineral natron glasses, four main groups dominate: Manganese-decoloured (6), antimony-decoloured (19, including one coloured blue by cobalt), weak HIMT (29) and strong HIMT (28). A single pale-coloured stemmed goblet cannot be attributed to any of these groups, but stands compositionally alone. The main groups are presented below in chronological order as listed in Table 1, with the plant-ash based glass group set between the antimony- and manganese-decoloured and the two HIMT groups, respectively.

The six manganese-decoloured glasses are all but one from cast vessels. They have between 0.5

and 1.3 wt% manganese oxide. Compared to the antimony-decoloured glass, they have higher

calcium oxide, alumina and phosphorus, and significantly lower levels of iron oxide and titania.

The antimony-decoloured glass has between half and one percent antimony oxide, and relatively

low levels of calcium oxide (5 to 8.5 wt%) and alumina (typically around 2 wt%, reaching up to

2.5 wt%). Potash, magnesia and iron oxide are all around half of a percent, and titania from 0.05

to 0.11 wt%. This closely matches data published by Paynter (2006), Silvestri (2008) and Schibille

(2011) for contemporary antimony-decoloured glass found in Britain, northern Italy and Albania.

Four glasses have higher potash (1.2 to 1.8 wt% K2O), magnesia (1.4 to 3.4 wt% MgO), and very

high phosphorus oxide (0.4 to 1.1 wt% P2O5) compared to both the early and later glasses. These

elevated values are the reason for interpreting the glass as plant-ash based, in line with

arguments developed first by Brill (1970) for Egyptian Late Bronze Age glass. Alternatively, the

elevated levels could originate from contamination by fuel ash during extended periods of heating (Paynter 2008; Rehren et al. 2010: 75-76, Schibille 2011: 2940); further research is necessary to understand this issue better.

The 'weak HIMT' group has from 0.5 to 1.1 wt% iron oxide, from 0.1 to 0.2 wt% titania, and typically between 0.5 and 2 wt% manganese oxide. Its calcium oxide content ranges from about 6 to about 9 wt%, and alumina ranges from 2 to 3 wt%. Potash concentrations are around one half of a percent, while magnesia is as high as iron oxide - around one percent by weight. This compositional pattern clearly differs from typical HIMT glass; in particular, the calcium oxide levels are too high by comparison, and show a slight positive correlation with alumina (Fig. 5) not normally seen in HIMT glass. Despite a basic similarity in composition, the colour of some of these samples does not always match the olive green tint of typical HIMT glass. The final large group among the analysed samples is made from unambiguous HIMT glass. Iron oxide in this 'strong' HIMT group ranges from 1 to more than 3 wt%, manganese oxide from 1.5 to 2.5 wt%, and titania from 0.3 to nearly 0.8 wt%. Calcium oxide levels are relatively narrowly set between 5 and 6.5 wt%, while alumina ranges from 2.3 to more than 3 wt%. Potash is present at just under half a percent, while magnesia levels fall closely around 1 percent; all these values are fully compatible with published HIMT analyses from other assemblages elsewhere (Mirti et al. 1993; Freestone 1994; see particularly the discussion of weak and strong HIMT glass from Britain and France in Foster and Jackson 2009: 193-4).

One sample did not meet our criteria to be allocated to one of the groups above; it is compositionally closest to the Sb-decoloured glass, but has no antimony above the detection limit, and deviates also in its content in lime (too low) and alumina (too high). We therefore left it unassigned and labelled it ukn for unknown.

5. Discussion

The compositional groups identified among the Tell Basta samples are discussed below in chronological order. The manganese-decoloured, antimony-decoloured and the plant-ash based glasses fall almost all into the first to third centuries AD. The weak HIMT glasses potentially overlap with these early groups and continue into the fifth century AD, while the strong HIMT glasses are all from the fourth century or later. It has to be stressed here that the dating of individual samples is done purely on typological grounds, often with rather long run times of some types stretching over several centuries, and not on stratigraphic or other evidence that would date specific finds more narrowly.

5.1 Early decoloured glass

The early date of the six manganese-decoloured glasses is noteworthy, as is the fact that five of

the six are from mould-cast vessels (see Table 1, Mn 01-06, with primary typological reference).

Two fragments belong to cast ribbed bowls, one represents the handle of a skyphos and two

fragments are rims of cast hemispherical grooved bowls (see fig. 6.3). The blown fragment

belongs to a beaker with wheel-cut horizontal grooves. The early use of manganese-decoloured

glass is consistent with the occurrence of the same glass among Hellenistic assemblages in the

eastern Mediterranean (e.g. Connolly et al. 2012), while elsewhere in Europe, manganese-

decoloured glass only appears much later, as in Southern France (Foy et al. 2000: 54-56,

corresponding groupe 3) or Italy (Silvestri et al. 2008), where manganese seems to have replaced

antimony as a decolourant only at the end of the second/start of the third century AD, or in

Roman Britain, where the use of manganese as a decolourant also does not seem to start before

the fourth century AD (Foster and Jackson 2010). One of the six analysed fragments (Mn 06) is

dark blue coloured by cobalt (0.05 wt%) and copper oxides; this sample has also much higher

iron oxide content than the others in this group. This is consistent with the observation from contemporary Pergamon (Rehren et al. forthcoming), where dark blue glasses are also coloured by a combination of cobalt, copper and iron oxides. This colorant has some similarity with the cobalt-blue colorant used earlier in New Kingdom Egypt (Smirniou and Rehren 2013), and may indicate a continuity of its exploration well into the first millennium AD.

Antimony-decoloured glass predominates at Roman Bubastis. It has been used for mould cast vessels, vessels with a folded/tubular base ring (see fig. 6.1), an applied foot ring (see fig. 6.2) or pinched feet elements, a wall fragment with cut circular facettes, the rim of an aryballos, possibly an indented beaker with a thick base, a small container and a bottle or flask with flaring rim and applied thread, all of which - apart from one waster (Sb 12) - have typological parallels in the first three centuries AD (see Table 1, Sb 01-19, with primary typological reference). Comparable analytical data exists from the eastern Mediterranean, e.g. from Petra (Schibille et al. 2012) or Pergamon (Rehren et al. forthcoming) where the majority of glass decoloured by antimony can be dated to the first and second centuries AD. Elsewhere, as in Thamusida (Morocco: Gliozzo et al. 2013) and Roman Britain, these glasses are dating to the first to third (Jackson 2005, Paynter 2006), or even to the fourth centuries AD (Foster and Jackson 2010). Significantly, antimony-decoloured glass has been identified by Picon et al. (2008) at the Wadi Natrun primary production installations in Zakik, Beni Salame and Bir Hooker, which date to the first and second centuries AD. In particular their group wnc has close similarities to the Bubastis glass; it has, however, significantly higher amount of soda compared to the Bubastis material which can therefore not be linked directly to the Wadi Natrun production sites. Elsewhere in the Roman world, decoloured glass, particularly by antimony, is often seen in more high-class products, and being made of purer raw materials and the presumably expensive

antimony (Jackson 2005, Paynter 2006, Nenna 2007; Silvestri et al. 2008, Foster/Jackson 2010).

This does not apply for the Bubastis material, where antimony-decoloured glass seems to predominate heavily over manganese-decoloured glass. It is remarkable that the analysed assemblage does not contain any naturally-coloured aqua or blue-green glass, despite the fact that during the first three centuries AD in many parts of the Roman Empire this 'Roman blue/green' glass was most commonly used for vessels (Paynter 2006, Silvestri 2008).

5.2 Plant-ash glass

Particular mention has to be made of the plant-ash based glasses (see Table 1, PA 01-04, with

primary typological reference). All dark green-turquois translucent low-capacity, thick walled

ungüentaría fall into this group; three of these are reported here, while one sherd represents a

ridged handle, probably from a relatively large and also thick walled transport or storage

container. Large quantities of such ungüentaría are also known from other Roman-period sites in

Egypt (such as in Tuna, M. Flossmann pers. com.), and we hope to analyse these in the near

future. The only published comparable contemporary glass vessels from Egypt and apparently

made of plant ash glass are those deriving from Wadi Natrun (Picon et al. 2008) and several kohl

flacons and unguentaria, dating to the second and third centuries AD, mentioned among the

glass finds in the Louvre collection (Arveiller-Dulong, V., and Nenna, M.-D., 20052, Nenna et al.

2005). Further examples of this composition have been published from Britain and France

(Jackson et al. 2006) as well as Italy (Gallo et al. 2013) and Albania (Schibille 2011).

Plant ash glass is very rare in the eastern Mediterranean during the first millennium BC and the

first half of the first millennium AD. Having dominated Egyptian and Mesopotamian glassmaking

during the second millennium BC, it is replaced in Egypt and the eastern Mediterranean around

1000 BC by mineral-natron based glass (Schlick-Nolte and Werthmann 2003). However, plant-ash

based glassmaking persisted in the Sasanian Empire, to the east of the Euphrates, from where it

may have found its way back into the west following the collapse of natron supply in the eighth or ninth century AD (Whitehouse 2002; Shortland et al. 2006). Islamic-period glassmaking appears to be concentrated in Syro-Palestine, with little if any Egyptian plant-ash glass making known from this period (Freestone et al. 2009). Against this traditional narrative, Picon et al. (2008) in their study of Roman glass from the Wadi Natrun report four plant-ash glass finds which they link to a local Egyptian production. A comparison of their analyses with ours, however, shows significant differences. The Wadi Natrun plant ash glasses are very rich in lime (10 to 16 wt%), alumina (4 to 7 wt%) and iron oxide (2-3 wt%), and very low in soda (9.5-12.5 wt%). This composition is very unusual and does not resemble the plant ash glass analyses from Bubastis, or elsewhere. We therefore do not link the Bubastis samples to a production from Wadi Natrun.

The geographical origin of the glass used to produce these vessels is of considerable interest, as the presence of several such vessels in Tell Basta could imply that the Roman town was still engaged in long-distance trade if the glass were indeed coming from east of the Euphrates. To discuss this, we can look at both the composition of the flux and that of the sand used to make this glass. There appears to be a tendency that the levels of potash and magnesia in plant ash glasses follow a broad trend of increasing levels from Syria eastwards (Freestone 2006). Using this criterion suggests that the relevant values found in the Tell Basta assemblage are far too low to assume an import of these vessels from the Sasanian Empire; they are also lower than the levels found in most Islamic glass from Syria (Fig. 7). The closest match in their alkali composition for PA 02-04 is with the relatively low-potash New Kingdom cobalt-blue glasses found both in Egypt and Mycenaean Greece (Smirniou and Rehren 2013) for which production evidence has been found in Amarna (Smirniou and Rehren 2011). This suggests that also these Roman-period plant ash glasses might have been made in Egypt. The elevated concentrations of sand-derived

components in these glasses, particularly iron oxide, titania and manganese oxide, are also more consistent with glasses typically linked to an Egyptian origin than a Levantine one. However, the phosphorus content of these glasses exceeds that of most other glasses from the Bronze Age and Classical Antiquity, reaching more than one percent by weight in one sample. It is hoped that our ongoing research on vessels of this type across Egypt will shed more light on the chemical composition and geographic and chronological distribution of this intriguing glass group. For now, we note that only these particular vessels were made consistently using plant ash glass rather than mineral-natron based glass. Their existence strongly indicates that production of plant-ash glass persisted in Egypt a millennium after the introduction of mineral-natron based glassmaking, and at least half a millennium before the re-introduction of plant ash-based glassmaking in the Islamic period, as already observed by Picon and co-workers (2008). It may be significant that particularly low-capacity unguentaria and kohl flacons seem to fall into this glass category, raising the question whether plant ash glass was produced specifically for vessels carrying low-volume high-value goods.

5.3 HIMT glass

HIMT glass is widespread between the fourth and sixth centuries AD throughout the whole

Mediterranean and Europe. After its introduction it soon dominated the Roman and Late

Antique glass industry, and at least outside the Levant became more popular than the

contemporary Levantine glasses (Freestone et al. 2002b). It is therefore not surprising that

almost two thirds of the samples reported in Table 1 are HIMT glass, divided equally between a

'weak HIMT' and a 'strong HIMT' group. The weak HIMT group (see Table 1, wH 01-29, with

primary typological reference) consists of an oval dish, a wall fragment of an mosaic glass vessel,

an aryballos, a stemmed goblet, a stemmed bowl, an indented beaker, fragments of flasks (see

fig. 6.7), base fragments of vessels with pinched feet, base fragments of vessels with single applied base rings (see fig. 6.6) and multiple applied base rings, lamps with a pointed base, a conical hollow base or manufactured with a solid stem (see fig. 6.5), vessel bases with a high footring, flaring rims of bottles or jugs with an applied thread and rims of cups with tubular or up-going rims (see figs 6.4 and 6.8). Some of these are only faintly-coloured, appearing almost colourless when thin-walled, and at least one third of the analysed objects in this group can be dated by typology to the first three centuries AD. However, the majority of this group, and all of the strong HIMT glasses, are later finds, starting mostly in the fourth century AD, and are of dark green to brown colour (see Table 1, H 01-28, with primary typological reference). Identifiable objects include conical lamps or hemispherical bowls with cracked-off rims and pointed bases or a solid blob, a vessel base with multiple applied base rings, cups or bowls with various rim shapes (see figs 6.9, 6.10, 6.11 and 6.12), one oval dish, vessel bases with a high footring or a folded/tubular base, a stemmed bowl, four wall fragments of mould blown vessels, one wall fragment of a lamp with an applied blue blob, two fragments of ridged handles, possibly deriving from transport or storage containers and the neck of a flask.

Despite its importance, HIMT glass has not been well defined compositionally (Foster and Jackson 2009: 193). It is generally accepted that in addition to the eponymous higher concentrations in iron, manganese and titania it also has elevated levels of magnesia as well as zirconium, chromium, barium and other trace elements, and typically a good positive correlation between alumina, iron oxide, and most of the other characteristic elements. In contrast, lime levels are normally relatively constant and scatter around 6 wt% CaO, regardless of alumina concentrations (e.g. Fig. 5 in Freestone et al. 2002). The increase in recent years in published data for HIMT glasses has resulted in the identification of considerable compositional variability within this group (see e.g. Gallo et al. 2014, Schibille 2011: group WD2), including the presence

of HIMT glass without manganese (dubbed HIT glass: Rehren and Cholakova 2010), and of ever more extreme concentrations of some of the characteristic elements. On the other hand, considerable uncertainty exists regarding the lower end of acceptable HIMT compositions or, in other words, how little iron, manganese and titania can a glass have and still be called HIMT? Here is not the place to discuss this, but suffice it to say that Foster and Jackson (2009) consider a group of glass with an average of 0.6 wt% FeO, 0.1 wt% TiO2 and 1 wt% MnO still as (weak) HIMT.

We adopt the concept of weak HIMT here, even though Foster and Jackson's (2009) weak HIMT

group is compositionally different from the Tell Basta weak HIMT. In some aspects, the weak

HIMT glass from Tell Basta forms a continuum with the strong HIMT group (Figs. 8, 9). However,

it differs from typical HIMT glass in its higher calcium oxide content (Fig. 5). We interpret the

existence of weak and strong HIMT glass to indicate the use of two only broadly similar sand

sources, possibly in geographical proximity, but probably producing glass independently of each

other as indicated by the chronological and compositional differences between the two groups.

Compositionally closest to the weak HIMT group, including the higher lime levels and despite

some subtle differences in the alumina and iron oxide ratios, is a set of glasses from northern

Europe (Saxon I: Freestone et al. 2008), dating from 400 to 550 AD. Our identification of glass of

this composition as a major group in Egypt supports Freestone's assumption that the Saxon glass

was an import, and that its elevated levels of HIMT indicators did not result from the repeated

recycling of earlier Roman glass following the end of Roman rule in northern Europe. On current

evidence one can suspect that it is of Egyptian rather than Levantine origin; however, more work

is clearly needed to better understand the HIMT glass family in all its compositional complexity.

Types found among this glass group include vessels made of mosaic glass, indented beakers,

vessels with pinched feet, oval dishes, stemmed goblets, and an aryballos. Their dating covers a

relatively wide time span, ranging from the first to the seventh centuries AD, and includes a number of intentionally coloured samples.

The strong HIMT group matches compositionally published data for HIMT glass; noteworthy here is the emergence of a small sub-group characterised by excess iron oxide relative to titania and alumina compared to the bulk of this group (Figs. 8, 9); this has been observed elsewhere before (Rehren and Cholakova 2010) and again underlines the compositional complexity of this glass group, as well as its super-regional importance for much of Late Antiquity.

5.5 Absence of Egypt I and II glass compositions

In the introduction we mentioned the glass groups Egypt I and II; Egypt II is chronologically outside the frame of our study, and its absence from our data hence not surprising. In contrast, the absence of Egypt I glass was unexpected. It is linked to production in the Wadi Natrun since it shares some characteristics with the composition of glass finds from primary glassmaking installations there published by Picon et al. (2008), such as very low levels of lime and rather high soda levels. These primary glassmaking furnaces are dated to the first two centuries of the first millennium AD, contemporary with much of the glass from Bubastis. It is remarkable then that none of the glasses analysed to date from Tell Basta show this very characteristic Wadi Natrun signature, despite the relative proximity of the two sites.

6. Conclusion

The aim of this paper was to provide a first insight into glass supply and consumption at an Egyptian town between the first century BC and the end of the sixth century AD. Situated in the

Eastern Nile Delta, Bubastis can be expected to have been well integrated into the Roman trade - as indicated by the contemporary ceramic finds from the city. On the basis of compositional analysis, the glass discovered here falls into five main groups, four of which are well known from elsewhere. In the first three centuries of the first millennium AD, manganese and antimony-decoloured glass compositions dominate, while a previously little known high-phosphorus plant-ash glass was used for unguentaria. The later part of the assemblage consists of two different HIMT glass groups, one relatively low and one rather high in iron, manganese and titanium oxides. With the exception possibly of the manganese-decoloured glass, contributing less than 10% of the analysed sample and certainly even less of the entire assemblage, none of the material appears to be imported from outside Egypt, painting a picture in contrast to what the ceramic indicates.

Three observations are of particular interest and underscore the wider significance of this data set.

First: Manganese-decoloured and antimony-decoloured glass is evident here as elsewhere, with nearly identical base glass compositions and levels of additives as seen elsewhere in the Roman Empire. For Bubastis, this confirms that the town was integrated into the wider trade network of the Roman and Late Antique world. The data is consistent with a model of a limited number of primary glass producers serving super-regional markets, spanning the entire Roman world, and beyond, from Britain and France to Turkey and Egypt. There is, however, no evidence so far for the presence of the faintly-coloured Roman blue/green glass, which during this time is so widespread in the Northern Provinces. In contrast, in Bubastis the antimony-decoloured glass is not a glass chosen only for high-status objects, but appears during the early period to be the

predominant glass type, apart from a few possibly imported and relatively early samples of

manganese-decoloured glass. The absence of Roman b/g glass could indicate that this particular group was not produced in Egypt, while the Syro-Palestine area has been mentioned repeatedly in this context (e.g. Nenna et al. 2000, Foy et al. 2003, Gliozzo 2013). The existence of plant-ash glass in this early period, probably made regionally, is intriguing, particularly with its close association to a particular vessel type.

Second: Only HIMT glasses were used at Bubastis from the fourth century AD onward, with two co-existing sub-groups recognisable by their different levels of diagnostic elements. No evidence for the use of Levantine glass has been found, despite those glasses dominating archaeological assemblages in current-day Israel and Jordan, where in turn HIMT glass is rather rare (Kato et al. 2009; Rehren et al. 2010). This suggests a strong element of regional preference in glass consumption, most likely based on proximity to the primary producer. This, in turn, could indicate that the earlier Roman blue/green glass was also made in the Levant, and hence not used in Bubastis. The chronological and compositional difference between the two HIMT groups, and the difference in composition between these two and some other HIMT assemblages reported elsewhere, indicates that there is significant systematic variation within the HIMT group. This could indicate that there were a number of contemporary and consecutive glass-making installations active using similar but different sand sources, within the broader HIMT definitions. Also, weak HIMT glass is compositionally very close to Anglo-Saxon I (Freestone et al. 2008), Frankish German (Wedepohl et al. 1997) and Merovingian French (Velde 1990) glass dating to the 5th and 6th centuries AD, confirming that fresh raw glass from Egypt was still reaching Europe after the collapse following the departure of the Roman army. It is hoped that a comprehensive study of the HIMT glass family and its sub-groups will shed more light on this, even though trying to order the increasing numbers of analyses seem to make the picture more complex rather than clearer - a bit like herding cats.

Third: The complete absence at Bubastis of low-lime glass compositions typically linked to the Wadi Natrun, such as the primary production remains reported by Picon et al. (2008) or Freestone's Egyptian I glass based on analyses by Gratuze and Barrandon (1990) is remarkable, given that the city is located relatively close to the Wadi Natrun. This underscores how little we really know about this glass group, and its significance relative to the other, better-known groups.

The material presented here currently comprises the only available substantial data set for glass compositions from Roman and late Antique Egypt, making it difficult to generalise our observations beyond the statements just made. Clearly, more analytical data is required to support a more refined discussion about glass supply and consumption in Roman and Late Antique Egypt; some of this is currently underway as part of the Marie Curie project Glass in Late Antiquity - Science and Society, of which this is the first outcome.

Acknowledgements

A debt of gratitude is due to the Egyptian Ministry of State for Antiquities, for permitting the

work reported here to be undertaken. We are grateful to Professor Ian Freestone, who first

pointed out the close similarity of our 'weak HIMT' glass composition to the Anglo-Saxon I

glasses analysed by him, and for inspiring and informative discussions throughout - even where

we did not always follow his suggestions. Marie-Dominique Nenna provided help with the dating

of the weak HIMT group fragments, for which we are most thankful. Several funding bodies

supported the underlying excavations of the Tell Basta Project. Starting in 2009, the Egypt

Exploration Society generously funded our fieldwork at site. The glass related research is

supported by a current Marie Curie Intra-European Fellowship within the 7th European

Community Programme (project number 298127, FP7-PEOPLE-2011-IEF, to D.R.). Initial funding for this project was provided by the Fritz Thyssen Foundation, establishing the collaboration between the two authors at the UCL Institute of Archaeology. Comments by three anonymous reviewers are greatly appreciated; any remaining errors are ours.

References

Arveiller-Dulong, V., and Nenna, M.-D., 20052. Les verres antiques du musée du Louvre II, Vaisselle et contenants du Ier siècle au début du VIIe siècle après J.-C., Paris.

Baxter, M. J., Cool, H. E. M., and Jackson, C. M., 2005. Further studies in the compositional variability of colourless Romano-British vessel glass, Archaeometry 47, 47-68.

Bailey, D., 2007. Glass, in: Maxfield, V. A., and Peacock, D. (Eds.), The Roman Imperial Quarries Survey and Excavations at Mons Porphyrites 1994-1998, vol. II. The Excavations, London, 233266.

Brun, J.-P., 2003a. Les objets en verre, in: Cuvigny, H. (Ed.), La route de Myos Hormos, L'armée romaine dans le désert oriental d'Égypte, Cairo, 515-537.

Brun, J.-P., 2003b. Le verre dans le désert Oriental d'Égypte. Contextes datés, in: Foy, D., and Nenna, M.-D. (Eds.), Échanges et commerce du verre dans le monde antique, Montagnac, 377-387.

Brun, J.-P., 2011. Les objets en verre, in: Cuvigny, H. (Ed.), Didymoi. Une garnison romaine dans le désert Oriental d'Égypte, 1. Les fouilles et le matériel, Cairo, 215-241.

Connolly, P., Rehren, Th., Doulgeri-Intzesiloglou, A., and Arachoviti, P., 2012. The Hellenistic glass of Pherai, Thessaly, Annales du 18e Congrès de l'Association Internationale pour l'Histoire du Verre: Thessaloniki 2009, 91-97.

Degryse, P., and Schneider, J., 2008. Pliny the Elder and Sr-Nd isotopes: tracing the provenance of raw materials for Roman glass production, Journal of Archaeological Science 35, 19932000.

Faiers, J., 2013. Late Roman Glassware and Pottery from Amarna and Related Studies, Egypt Exploration Society Memoirs, 102, 5-38.

Foster, H. E., and 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.

Foster, H. E., and Jackson, C. M., 2010. The composition of late Romano-British colourless vessel glass: glass production and consumption, Journal of Archaeological Science 37, 3068-80.

Foy, D., 2001. Secteur Nord de Tebtynis (Fayyoum), Le verre byzantin et islamique, in : Annales Islamologiques 35, 465-489.

Foy, D., and Jezegou, M.P., 1997. Une epave chargee de !ingots et de vaisselle de verre, un temoignage exceptionnel du commerce et de la technologie du verre en Mediterranee antique. Verre 3.3, 65-70.

Foy, D., Vichy, M., and Picon, M., 2000. Lingots de verre en Mediterranee occidentale (IlIe siècle av.J.-C. - Vlle siècle ap.J.-C.), Annales du 14e Congrès de l'Association Internationale pour l'Histoire du Verre: Milano, 1998, 51-57.

Foy, D., Picon, M., Vichy, M., and 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., and Nenna, M.-D. (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 et Marseille, 7-9 juin 2001, Éditions Monique Mergoil, Montagnac, 41-86.

Freestone, I.C., 1994. Appendix: Chemical analysis of 'raw' glass fragments, in: Hurst, H.R. (Ed.), Excavations at Carthage, Vol. II, 1. The Circular Harbour, North Side. The Site and Finds other than Pottery. British Academy Monographs in Archaeology, No. 4. Oxford University Press, Oxford, 290.

Freestone. I., 2006. Glass production in Late Antiquity and the Early Islamic period: a

geochemical perspective, in: Maggetti, M., and Messiga, B. (Eds.), Geomaterials in Cultural Heritage. Geological Society, London, Special Publications, 257, 201-216.

Freestone, I., Gorin-Rosen, Y., and Hughes, M.J., 2000. Primary glass from Israel and the

production of glass in Late Antiquity and the Early Islamic period, in: Nenna, M.-D. (Ed.), La Route du Verre: ateliers primaires et secondaires de verriers du second millénaire av. J.-C. au Moyen Âge, Travaux de la Maison de l'Orient 33, 65-82.

Freestone, I. C., Ponting, M., and Hughes, M. J., 2002a. The origins of Byzantine glass from Maroni Petrera, Cyprus, Archaeometry 44, 257-72.

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

glassmaking in the Eastern Mediterranean: production and distribution of primary glass, in: Kordas, G. (Ed.), Hyalos - Vitrum - Glass. History Technology and Conservation of glass and vitreous materials in the Hellenic World. 1st International conference Rhodes - Greece 1-4 April 2001, Athens, 167-174.

Freestone, I.C., Wolf, S., and Thirlwall, M., 2005. The production of HIMT glass: elemental and isotopic evidence, Annales du 16e Congrès de l'Association Internationale pour l'Histoire du Verre: London 2003, 153-7.

Freestone, I.C., Hughes, M. J., and Stapleton, C. P., 2008. Composition and production of Anglo Saxon glass, in: Evison, V. I. (Ed.), Catalogue of Anglo-Saxon glass vessels in the British Museum, London, The British Museum, 29-46.

Freestone, I.C., Wolf, S., and Thirlwall, S., 2009. Isotopic composition of glass from the Levant and the south-eastern Mediterranean Region, in: Degryse, P., Henderson, J., and Hodgins, G. (Eds.), Isotopes in vitreous materials (Studies in Archaeological Sciences), 31-52.

Gallo, F., Marcante, A., Silvestri, A., and Molin, G., 2014. The glass of the "Casa delle Bestie Ferite": a first systematic archaeometric study on Late Roman vessels from Aquileia, Journal of Archaeological Science 41, 7-20.

Gliozzo, E., Santagostino Barbone, A., and D' Acapito, F., 2013. Waste glass, vessels and window panes from Thamusida (Morocco), Grouping natron-based blue-green and colourless Roman glasses, Archaeometry 55, 609-639.

Gorin-Rosen, Y. 2000, The Ancient Glass Industry in Israel: Summary of the Finds and New Discoveries, in: Nenna, M.-D. (Ed.), La Route du Verre: ateliers primaires et secondaires de verriers du second millénaire av. J.-C. au Moyen Âge, Travaux de la Maison de l'Orient 33, 4963.

Gratuze, B., and Barrandon, J.-N., 1990. Islamic glass weights and stamps: analysis using nuclear techniques, Archaeometry 32, 155-62.

Habachi, L., 1957. Tell Basta: Tell Basta, Supplément aux Annales du Service des Antiquités de l'Égypte: Cahier N° 22, Cairo.

Harden, D. B., 1936. Roman Glass from Karanis found by the University of Michigan Archaeological Expedition in Egypt, 1924-1929, Michigan.

Hill, M., and Nenna, M.-D., 2003. Glass from Ain et-Turba and Bagawat Necropolis in the Kharga oasis (Egypt), Annales du 15e Congrès de l'Association Internationale pour l'Histoire du Verre: New York-Corning 2001, 88-92.

Isings, C., 1957. Roman glass from Dated Finds. Archaeologica Traiectina, Groningen.

Jackson, C. M., 2005. Making colourless glass in the Roman period, Archaeometry 47, 763-80.

Jackson, C., Price, J., and Lemke, C., 2009. Glass production in the 1st century AD: Insights into Glass Technology, Annales du 17e Congrès de l'Association Internationale pour l'Histoire du Verre: Antwerp 2006, 150-155.

Jennings, S., 2006. Vessel Glass from Beirut BEY 006, 007 and 045, Berytus, 48-49.

Kato, N., Nakai, I., and Shindo, Y., 2009. Change in chemical composition of early Islamic glass excavated in Raya, Sinai Peninsula, Egypt: On-site analysis using a portable X-ray fluorescence spectrometer, Journal of Archaeological Science 36, 1698-1707.

Kucharczyk, R., 2004. Two Roman Toilet Bottles from Kom el-Dikka, in: Polish Archaeology in the Mediterranean 15 (Reports 2003), 43-45.

Kucharczyk, R., 2006. Late Roman/ Early Byzantine Glass from the Auditoria on Kom el-Dikka in Alexandria, in: Polish Archaeology in the Mediterranean 17 (Reports 2005), 45-50.

Kucharczyk, R., 2010. Glass from Area F on Kom el-Dikka (Alexandria), in: Polish Archaeology in the Mediterranean 19 (Reports 2007), 56-70.

Marchini, C., 1999. Glass from the 1993 Excavations at Ismant el-Kharab, in: Hope, C., and Mills, A. (Eds.), Dakhleh Oasis Project: Preliminary Reports on the 1992-93 and 1993-94 Field Seasons, Oxford, 75-82.

Meyer, C., 1992. Glass from Quseir al Qadim and the Indian Ocean Trade, Chicago.

Mirti, P., Casoli, A., and Appolonia, L., 1993. Scientific analysis of Roman glass from Augusta Praetoria, Archaeometry 35, 225-40.

Nenna, M.-D., Vichy, M., and Picon, M, 1997. L' Atelier de verrier de Lyon, du Ier siècle après J.-C., et l'origine des verres « Romains», Revue d'Archéométrie 21, 81-87.

Nenna, M.-D., 2000. Ateliers de production et sites de consommation en Égypte, Ve s. av. J.-C. VIIe s. apr. J.-C., in : Annales du 14e congrès de l'AIHV : Venise-Milan 1998, 20-24.

Nenna, M.-D., 2007. Production et commerce du verre à l'époque impériale : nouvelles découvertes et problématiques, Facta 1, 125-147.

Nenna, M.-D., Picon, M., Thirion-Merle, V., and Vichy, M., 2005. Ateliers primaires du Wadi Natrun: nouvelles decouvertes, Annales du 16e Congrès de l'Association Internationale pour l'Histoire du Verre: London 2003, 59-63.

Nenna, M.-D., 2010. The Glass Vessel Decor of a Funerary Chapel in the Bagawat Necropolis, Kharga Oasis, Egypt, in: Drauschke, J., and Keller, D., (Eds.), Glass in Byzantium - production, usage, analyses, RGZM, Mainz, 199-214.

Nicholson, P. T., 2000. Roman Glass from Berenike (Egypt), Some New Work, Annales du 14e congrès de l'AIHV : Venise-Milan 1998, 151-155.

Paynter, S., 2006. Analyses of colourless Roman glass from Binchester, County Durham, Journal of Archaeological Science 33, 1037-57.

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.

Peacock, D., 2011. Glass, in: Peacock, D., and Blue, L. (Eds.), Myos Hormos - Quseir al-Qadim: Roman and Islamic Ports on the Red Sea, vol. 2: Finds From excavations 1999-2003, Oxford, 57-78.

Picon, M., and Vichy, M., 2003. D'orient en occident: l'origine du verre a l'epoque romaine et durant le haut Moyen Age, in: Foy, D., and Nenna, M.-D. (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 et Marseille, 7-9 juin 2001, Éditions Monique Mergoil, Montagnac, 17-32.

Picon, M., Thirion-Merle, V. and Vichy, M., 2008. Les verres au natron et les verres aux cendres du Wadi Natrun (Egypte). Bulletin de l'Association Francaise pour l'Archeologie du Verre 2008, 36-41.

Rehren, Th., 2000. Rationales in Old World base glass compositions, Journal of Archaeological Science 27, 1225-1234.

Rehren, Th. and Cholakova, A., 2010. The early Byzantine HIMT glass from Dichin, Northern Bulgaria. Interdisciplinary Studies 22/23, 81-96.

Rehren, Th., Marii, F., Schibille, N., Swann, C. and Stanford, L., 2010. Glass supply and circulation in early Byzantine southern Jordan, in: Drauschke, J., and Keller, D., (Eds.), Glass in Byzantium - production, usage, analyses, RGZM, Mainz, 199-214.

Rehren, Th., Connolly, P., Milevska, B., Schibille, N. and Schwarzer, H., forthcoming. Changes in glass composition in Pergamon (Turkey) from Hellenistic to late Byzantine times.

Sayre, E. V., and Smith, R. W., 1961. Compositional categories of ancient glass, Science 133, 18246.

Shaw, I., 20032. The Oxford H istory of Ancient Egypt, Oxford.

Schibille, N., Degryse, P., O'Hea, M., Izmer, A., Vanhaecke, F., McKenzie, J., 2012. Late Roman glass from the 'Great Temple' at Petra and Khirbet et-Tannur, Jordan - Technology and provenance, Archaeometry 54, 6, 997-1022.

Schibille, N., 2011. Supply routes and consumption of glass in first millennium CE Butrint (Albania), Journal of Archaeological Science 38, 2939-2948.

Schlick-Nolte, B. and R. Werthmann, 2003. Glass vessels from the burial of Nesikhons, Journal of Glass Studies 45, 11-34.

Shortland, A. J., Schachner, L., Freestone, I. C., and Tite, M. S., 2006. Natron as a flux in the early vitreous materials industry: sources, beginnings and reasons for decline, Journal of Archaeological Science 33, 521-530.

Silvestri, A., 2008. The coloured glass of lulia Felix, Journal of Archaeological Science 35, 1489 -1501.

Silvestri, A., Molin, G., and Salviulo, G., 2008. The colourless glass of lulia Felix, Journal of Archaeological Science 35, 331-41.

Smirniou, M. and Rehren, Th., 2011. Direct evidence of primary glass production in Late Bronze Age Amarna, Egypt. Archaeometry 53, 58-80.

Smirniou, M. and Rehren, Th., 2013. Shades of blue - cobalt-copper coloured blue glass from New Kingdom Egypt and the Mycenaean World: a matter of production or colorant source? Journal of Archaeological Science 40, 4785-4792.

Tanimoto, S. and Rehren, Th., 2008. Interactions between silicate and salt melts in LBA glassmaking, Journal of Archaeological Science 35, 2566-2573.

Velde, B., 1990. Alumina and calcium oxide contents of glass found in western and northern Europe, first to ninth centuries, Oxford Journal of Archaeology 9, 105-117.

Wedepohl, K. H., and Baumann, A., 2000. The use of marine molluskan shells for Roman glass and local raw glass production in the Eifel area (Western Germany), Naturwissenschaften 87, 129-32.

Wedepohl, K. H., Pirling, R., and Hartmann, G., 1997. Römische und fränkische Gläser aus dem Gräberfeld von Krefeld-Gellep, Bonner Jahrbücher, 177-189.

Whitehouse, D., 2002. The transition from natron to plant ash in the Levant, Journal of Glass Studies 44, 193-196.

Captions

Fig. 1: The Egyptian Nile Delta showing the position of Bubastis, the Wadi Natrun, Alexandria,

Cairo and Sinai peninsula

Fig. 2: The ancient site of Bubastis

Fig. 3: Base fragment of an oval dish demonstrating the average colour of HIMT glass from Bubastis

Fig. 4: Rim fragment demonstrating the average colour of Sb decol glass from Bubastis Fig. 5: Plot of the four natron groups from Bubastis. Note the good match of HIMT glass from Bubastis with published HIMT analyses, and the absence of glass with a Wadi Natrun signature. Levantine I glass is also not represented. Graph based on Freestone (2005)

Fig. 6: Glass vessel fragments from Bubastis. Scale 1:2 (Drawings: Daniela Rosenow, digitalisation: Mandy Mamedow)

Fig. 7: Comparison of the four plant ash glasses from Bubastis to typical compositional fields of other glasses. The best match is with Egyptian LBA glasses (data from Smirniou and Rehren 2013, and references therein). Graph based on Freestone (2006)

Fig. 8: Weak HIMT glass plots between the decoloured glasses and HIMT glass. HIMT glass is split into two groups, with six samples having higher iron oxide relative to titania than the majority of the HIMT glass

Fig 9: Both weak HIMT and HIMT show a positive correlation between alumina and iron oxide, in contrast to Levantine I glass. The correlation is more pronounced for HIMT glass. Note also the group of six HIMT samples with excess iron oxide. The two decoloured glasses with high iron oxide content are coloured by cobalt, which is associated with increased iron oxide.

Table 1: Catalogue of analysed samples

Table 2: Comparison of published compositions for Corning A and B (Brill 1999: 544) and the average values of 7 measurements of Corning A and B during the course of the analysis of the Tell Basta samples. The precision of the analyses is indicated by the standard deviation among the seven individual analyses for each of the Corning glasses, while the accuracy is expressed by the deviation of the analysed value Ca from the published composition Cp. This 8 rel% value is calculated using the formula (Ca - Cp)/Cp»100.

Table 3: EPMA analyses of glass samples from Bubastis, data in weight percent. Tin, cobalt and lead were analysed for, but not found at levels above 300 mg/g (500 mg/g for lead).

Table 1 Catalogue of analysed samples Mn decoloured

Mn 01 TB3a- -Z/3.SCH.1-G009 skyphos handle mould cast colourless-light purple 1 Isings type 55 1. AD

Mn 02 TB2a-O PQ/2006-G 001 ribbed bowl wall mould cast aqua 7 Isings type 3a 50 BC-130 AD

Mn 03 TB1a-T/8-G001 bowl rim mould cast colourless 4 Meyer 1992, pl. 2.26 1 ./2. AD

Mn 04 TB2b-W/3.SCH.1-G008 beaker rim free blown colourless 5 Peacock 2011, 67, fig . 7.7.77 1 ./2. AD

Mn 05 TB2b-X/3.SCH.1-G004 bowl rim mould cast colourless-light green 4 Jennings 2006, 35, fic |. 2.6.5 50 BC-50 AD

Mn 06 TB3b-X/4.SCH.1-G025 ribbed bowl rim mould cast dark blue 7 Isings type 3b 50 BC-130 AD

Sb decoloured technique

Sb 01 TB2b-X/4.SCH.1-G033 bowl or plate base mould cast colourless-light green 48 Isings type 80 late 1.-mid 3. AD

Sb 02 TB1 a-D/12.2-G001 flask? base with folded (tubular) ring free blown colourless-light green 103 Isings type 104a mid 1 .-mid 3. AD

Sb 03 TB3a- -Z/3.SCH.1-G056 plate base free blown colourless-light green 42 Bailey 2007, 254, fi g. 8.14.15 late 1 .-late 2. AD

Sb 04 TB1 b-X/2.AbH-G022 bowl rim mould cast colourless 3 Peacock 2011, 69, ■ fig. 7.9.106 late 1.-175 AD

Sb 05 TB3a- -Z/3.BEF.1-G003 bowl base free blown colourless 103 Brun 2003b, 384, fi g. 8.6 2. half 2. AD

Sb 06 TBXIV-0PQ-G010 plate base mould cast colourless-light green 48 Bailey 2007, 238, fi g. 8.2.19 1 ./2. AD?

Sb 07 TB3a-X/4.SCH.1-G023 facet-cut beaker? wall free blown light green-colourless 1 Peacock 2011, 65, ■ fig. 7.5.62 1 ./2. AD?

Sb 08 TB3a-Y/3.SCH.1-G060 beaker? base free blown colourless-light green 37 Bailey 2007, 247, fi g. 8.8.71 late 1.-2. AD

Sb 09 TB3a-X/2.TS.SCH.1-G012 flask? base with folded (tubular) ring free blown colourless-light green 103 Isings type 104a 1 ./2. AD?

Sb 10 TB2b-X/3.SCH.1-G029 beaker, goblet, sprinkler? base with pinched feet free blown colourless 15 Brun 2011, 239, fig. 271.136 2.-6. AD

Sb 11 TB3a- -Z/3.SCH.1-G050 aryballos rim free blown colourless 5 Bailey 2007, 256, fi g. 8.15.118 1 ./2. AD?

Sb 12 TB3b-Y/4.SCH.1-G042 waster base free blown light green-colourless unknown

Sb 13 TB1a-T/8-G018 flask? base with folded (tubular) ring free blown colourless 103 Isings type 104a 1 ./2. AD

Sb 14 TBXIV-D/7-G001 cup? base with folded (tubular) ring free blown colourless 103 Isings type 37 1 ./2. AD

Sb 15 TB2a-X/2.AbH-G007 plate or bowl? base free blown (?) colourless-light green 42 Peacock 2011, 74, ■ fig. 7.13.158 1.-3. AD

Sb 16 TB2a-X/3.SCH.3-G007 bottle/flask rim with applied thread free blown colourless 72 Isings type 102b 3. AD

Sb 17 TB1 b-W/2.SCH.1-G008 bowl base mould cast colourless-light green 48 Isings type 80 1.-3. AD

Sb 18 TB2b-X/3.SCH.1-G011 small container base free blown colourless-light green 10 Bailey 2007, 259, fi g. 8.17.142 1.-4. AD

Sb 19 TB1a-Survey-G024 small container rim free blown blue 33 Bailey 2007, 258, fi g. 8.16.135 1 ./2. AD

Plant ash

PA 01 TB3b- -Z/3.SCH.1-G001 mid-capacity unguentarium base free blown green-turquois 2 Bailey 2007, 259, fi g. 8.17.141 1 ./2. AD

PA 02 TBXX-G20093 mid-capacity unguentarium base free blown green 2 Bailey 2007, 259, fi g. 8.17.138 1 ./2. AD

PA 03 TB1 a-Survey-G017 low-capacity unguentarium base and body free blown dark green 4 Bailey 2007, 261, fi g. 8.18.155 1 ./2. AD

PA 04 TB1 b-W/2.SCH.1-G006 storage or transport container ridged handle free blown turquois 6 Bailey 2007, 262, fi g. 8.19, 165 1.-5. AD

weak HIMT

wH 01 TB3a-Y/3.SCH.1-G024 beaker, jug, goblet, flask? base with applied rings free blown colourless 39 Sternini 1999, 99, fi g. 9.119 4./5. AD

wH 02 TB2a-X/2.SCH.2-G019 bottle, jug, flask, beaker? rim with applied thread free blown wall (sampled part) light green, 72 Harden 1936, pl. XIX, 739 ab 3. AD

ring blue

wH 03 TB3a- -Z/3.SCH.1-G073 bottle, jug, flask, beaker? base, pinched feet free blown colourless 15 Harden 1936, pl. XIX, 682 2.-6. AD

wH 04 TB3a-Survey-G003 bottle, beaker or flask? base, pinched feet, wall indented free blown green 15 Harden 1936, pl. XIX, 682 1.-4. AD

wH 05 TBXX-G20087a lamp base (pointed) free blown light green 6 Isings type 106d from 4. AD

wH 06 TB1 b-W/2.SCH.1 -G019 beaker, jug, goblet, flask? base with applied rings

wH 07 TB3a- -Z/3.SCH.1-G002 lamp or beaker base (conical hollow)

wH 08 TB3a-Y/3.SCH.1-G002 bowl rim (tubular)

wH 09 TB1 a-Survey-G012 lamp base (with solid stem)

wH 10 TBXX-G 20049 lamp base (pointed)

wH 11 TB2a-X/2.AbH-G037 beaker, jug, goblet, flask? base with applied rings

wH 12 TB3a-Z/3.SCH.1-G027 small container base

wH 13 TB2a-X/3.SCH.5.BEF.1-G00' I plate or bowl base (high footring)

wH 14 TB2b-W/3.SCH.2-G005 bottle, jug, flask? rim with applied thread

wH 15 TB2b-W/2.AbH-G010 aryballos handle

wH 16 TB3a-Y/3.SCH.1-G008 bottle or jug? base with applied rings

wH 17 TB3a-Y/3.SCH.1-G042 bottle, flask or jug? base with applied ring

wH 18 TBXIV-S/2-G013 goblet base and stem

wH 19 TB3a- -Z/3.SCH.1-G001 goblet, beaker or flask? base with applied ring

wH 20 TB1 b-OPQ-G009 oval dish rim, tubular

wH 21 TB1 b-OPQ-G006 bottle, flask, beaker or jug? base, pinched feet

wH 22 TB3a-Z/3.SCH.1-G024 dish made of mosaic glass wall

wH 23 TBXIV-G14011 flask/toilet bottle? neck and rim

wH 24 TBXIV-OPQ-G005 beaker, jug, goblet? base with applied ring

wH 25 TB2a-X/2.AbH-G020 flask/bottle base

wH 26 TB1 b-W/2.SCH.5-G006 base indented beaker base

wH 27 TB3b-X/4.Steg-G011 jug, flask, bowl? base (high footring)

wH 28 TB3a-X/4.SCH.1-G027 bowl stem

wH 29 TB3a- -Z/3.SCH 1-G028 cup or bowl rim (inturned)

H 01 TB3b-V/3.SCH.1-G159 bowl or beaker? wall

H 02 TB2b-X/3. BEF. 1-G005 lamp base (pointed)

H 03 TB2b-X/3.BEF.1-G010 ???? vrall

H 04 TB2b-X/3.SCH.1-G040 bowl or beaker? wall

H 05 TB2a-M/1.SCH.1-G015 ??? wall (with cut decoration)

H 06 TB2b-X/3.SCH.3-G009 conical lamp or beaker rim

H 07 TB1a-W/2.AbH-G007 stemmed bowl stem

H 08 TB2b-X/3.SCH.1-G016 hemispherical bowl or cup rim

H 09 TBXX-G 20041 oval dish base

H 10 TB2b-W/3.SCH.3-G008 bowl or drinking vessel base (high footring)

H 11 TBXX-G 20008b lamp base (with twisted blob)

H 12 TB2a-W/3.SCH.4-G001 bowl or flask? base (high footring, wavy)

H 13 TBXX-G 20026 lamp or beaker base (conical hollow)

H 14 TB3a-Y/3.SCH.1-G039 bowl rim (horizontal)

H 15 TB3a-Y/3.SCH.1-G033 bowl rim (edge going up)

H 16 TB1 a-V/2. AbH-G189 transport or storage container handle (ridged)

free blown light green-colourless 39 Tatton-Brown 1984, 206, fig. 68 .103 4./5. AD

free blown light green 14 Harden 1936, pi. XVI, 457 from 4. AD

free blown light green-colourless 25 Marchini 1999, 80, flg. 3.b 1.12. AD

free blown light olive green 9 Jennings 2006, 146, fig. 6.20.11 -13 from 4. AD

free blown yellowish green 6 Isings type 106d from 4. AD

free blown unknown, corroded 39 Tatton-Brown 1984, 206, fig. 68 .103 4./5. AD

free blown light green 20 Harden 1936, pi. XX, 799 1.-3. AD

free blown colourless-light green 120 Harden 1936, pi. XII.83/130 from 4. AD

free blown light green 72 Harden 1936, pi. XIX, 712 from 3. AD

free blown unknown, corroded 5 Isings type 61 late 1.-7. AD

free blown light olive green 39 Tatton-Brown 1984, 206, fig. 68 .103 4./5. AD

free blown wall light green, ring blue 37 Keller 2006, Tafel 21.g 4./5. AD

free blown light green 6 Harden 1936, pi. XVI, 482 5.-7. AD

free blown wall colourless-light green, 37 Bailey 1998, pi. 93.Y72 4./5. AD

base ring blue (sampled part)

free blown light blue 17 Isings type 97b 3.-5. AD

free blown light green 15 Harden 1936, pi. XIX, 682 2.-6. AD

cast green (sampled) and yellow 3 1.-5. AD

free blown colourless-light green 5 Nenna 2010, 210, fig. 10.34 1.-3. AD

free blown wall colourless, base ring 37 HIII/Nenna 2001, 91, fig. 4.4 4./5. AD

blue (sampled part)

free blown colourless 70 Isings type 133 1.-4. AD

free blown unknown, corroded 25 Harden 1936, pi. XV, 376 1.-4. AD

free blown reddish brown 120 Harden 1936, pi. XIV.274 4.-7. AD

free blown purple-red 17 Harden 1936, pi. XV, 360 4./5. AD

free blown light blue 35 Nenna 2000, 23, fig. 9.4 from 4. AD

mould blown unknown, corroded 33

free blown light green-colourless 6

mould blown colourless-light green 33

mould blown green 25

colourless 2

free blown green 90

free blown green 17

free blown light green-colourless 20

free blown olive green 17

free blown light green 120

free blown light green 6

free blown olive green 2

free blown green 14

free blown olive green 4

free blown light green 10

light green 6

Harden 1936, pi. XIII, 217 4./5.AD

Isings type 106d from 4. AD

2.-4.AD

Harden 1936, pi. XIII, 189, pi. XV, 409 unknown

unknown

Isings type 106d from 4. AD

Harden 1936, pi. XV, 358 4./5. AD

Isings type 96 from 4. AD

Isings type 97b 3.-5. AD

Harden 1936, pi. XV, 360 from 4. AD

Kucharczyk 2006, 48, fig. 1.4 from 4. AD

Harden 1936, pi. XIX, 672 from 4. AD

Harden 1936, pi. XVI, 457 from 4. AD

Harden 1936, pi. XII, 130 from 4. AD

Nenna 2000, 23, fig. 9.2 from 4. AD

H 17 TB2b-OPQ-G014 bowl rim

H 18 TB3b-V/3.SCH.1-G172 bowl rim

H 19 TB2a-X/2.SCH.1-G054 conical lamp or beaker rim

H 20 TB3a-Y/3.SCH.1-G093 beaker/jug/flask? base with applied rings

H 21 TB3a-Y/3.SCH.1-G017 cup or bowl rim (strongly everted)

H 22 TB1 a-T/8-G003 cup or bow rim (horizontal)

H 23 TB1 b-W2.SCH.3-G 006 flask neck

H 24 TBXX-G 20008c bowl base with folded (tubular)

H 25 TB1 b-X/2.AbH-G119 bottle wall

H 26 TB1 a-V/2-G005 amphora? handle (ridged)

H 27 TB3a-X/2.TS.SCH.3-G001 conical lamp or beaker? wall

H 28 TB2a-X/2.AbH-G047 bowl rim (tubular)

TB1 a-D/11 .SCH.3-G004

stemmed goblet

optical blown olive green 4 Harden 1936, pi. XXIV, 256 from 4. AD

free blown pinkish brown 1 see drawing mid 4.-mid 5. AD

free blown green 90 Isings type 106d from 4. AD

free blown green 39 Tatton-Brown 1984, 206, fig. 68.103 4./5. AD

free blown light olive green 15 Tatton-Brown 1994, 283, fig. 15.1.5 from 4. AD

free blown green 4 Weinberg 1988, 52, fig. 4-12.95 4.-6. AD

free blown light green-yellowish 26 Isings type 133 unknown

free blown yellowish green 103 Harden 1936, pi. XIV, 245 4. AD

mould blown olive green 25 Harden 1936, pi. XIX, 700 and 701 4. AD

green 6 4.-5. AD

free blown wall light green (sampled part), 18 Kucharczyk 2006, 48, fig. 1.15 from 4. AD

blue blob

free blown olive green 7 Harden 1936, pi. XII, 89 4.-7. AD

free blown aqua

6 Sternini 1999, 95, fig. 6.67

from 4. AD

Table 2: Comparison of Corning A and B composition as published (top row) and as measured (averai

SiO2 Na2O CaO K2O MgO Al2O3 Fe2O3 TiO2

Cor A publ 66.56 14.30 5.03 2.87 2.66 1.00 1.09 0.79

Cor A aver 67.08 14.16 4.92 2.78 2.59 0.91 1.03 0.78

StdDev 0.61 0.11 0.03 0.03 0.03 0.02 0.02 0.01

5 rel% 0.8 -1.0 -2.2 -3.1 -2.6 -9.0 -5.5 -1.3

Cor B publ 61.55 17.00 8.56 1.00 1.03 4.36 0.34 0.09

Cor B aver 62.18 17.08 8.43 0.99 1.00 4.15 0.33 0.09

StdDev 0.47 0.29 0.07 0.03 0.01 0.10 0.03 0.01

5 rel% 1.0 0.5 -1.5 -1.0 -2.9 -4.8 -2.9 1.1

ge and standard deviation of 7 separate measurements done during the analysis of the Bubastis sample;

Sb2O5 MnO CuO P2O5 Cl SO3nalytical Total

1.76 1.00 1.17 0.13 99.53

2.12 0.97 1.19 0.09 0.09 0.15 99.87

0.04 0.00 0.04 0.01 0.01 0.01 0.67

20.5 -3.0 1.7 -30.8

0.46 0.25 2.66 0.82 0.20 0.54 99.98

0.64 0.21 2.70 0.69 0.17 0.54 100.04

0.01 0.03 0.05 0.03 0.01 0.02 0.83

39.1 -16.0 1.5 -15.9 -15.0 0.0

butnotfoundatlevelsabove 300 mg/g(500 mg/gM)UPT

EPMA analyses of glass samples from Bubastis, data in weight percent. Tin and lead were analysed for

Mn 01 Mn 02 Mn 03 Mn 04 Mn 05 Mn 06

66.6 69.2 66.8

68.5 67.7

8.10 7.28 8.68 7.62 7.43

0.52 0.65 0.53 0.45 0.87 0.59

0.53 0.42 0.65 0.60

2.03 2.13 2.13 2.49 2.25 2.44

0.27 0.28 0.33 0.32

0.05 0.04 0.05 0.06 0.05 0.04

<0.3 <0.3 <0.3 <0.3 <0.3 <0.3

0.86 0.46

1.13 1.30

1.14 1.11

<0.03 0.03 0.03 <0.03 <0.03 0.07

0.07 0.06 0.09 0.10

0.68 0.96 1.13 0.92 0.82 0.83

0.32 0.14 0.18 0.21 0.19 0.29

96.8 98.4 97.0 97.7

Average StDev

0.60 0.05

2.25 0.15

0.05 0.00

1.00 0530

Sb 01 Sb 02 Sb 03 Sb 04 Sb 05 Sb 06 Sb 07 Sb 08 Sb 09 Sb 10 Sb 11 Sb 12 Sb 13 Sb 14 Sb 15 Sb 16 Sb 17 Sb 18 Sb 19

69.9 69.2

70.8 68.5

69.5 67.7 66.1

68.3 66.3 66.0 65.9 71.5 71.3

5.11 5.96 5.67 6.44 5.67 5.90 6.92 6.23 6.58 8.27 6.47 6.87 7.83 7.34 8.55 8.42 8.22

0.42 0.45 0.41 0.46 0.46 0.68 0.55 0.65 0.54 0.39 0.51 0.54 0.54 0.68 0.53 0.58

0.48 0.50 0.43 0.50 0.49 0.60 0.59 0.60 0.58 0.61 0.67 0.58 0.79 0.76 0.71 1.05 0.43

2.08 2.04 2.15 2.21 2.09 2.23 2.31 2.30 2.50

0.37 0.36 0.40 0.41 0.52 0.49 0.48 0.50 0.53

0.62 0.58 0.62

0.04 0.06 0.07 0.08 0.08 0.07 0.06 0.06 0.07 0.09 0.10 0.10 0.10 0.10 0.10 0.11 0.11 0.12 0.06

<0.03 <0.03 <0.03 0.04 <0.03 0.03 <0.03 0.07 <0.03 <0.03 0.03 0.03 <0.03 <0.03 <0.03 <0.03 <0.03 0.03 0.05

<0.03 <0.03 <0.03 <0.03 <0.03 <0.03 0.05 <0.03 <0.03 <0.03 <0.03 <0.03 <0.03 <0.03 <0.03 <0.03 <0.03 0.03

0.03 0.03 0.03 0.03 0.02 0.04 0.04 0.07 0.05 0.03 0.05 0.06 0.04 0.04 0.05 0.05 0.04 0.03 0.06

1.23 1.11 1.06 0.96 1.01

1.04 1.08 1.02

1.05 0.93 1.03 1.07 0.98 0.87 0.83 0.89 1.02 0.97 0.46

0.24 0.26 0.21 0.27 0.20 0.24 0.33 0.25 0.31 0.35 0.30 0.33 0.32 0.29 0.34 0.34

99.7 97.6 97.0 98.2

99.9 96.9 95.9

97.4 99.8

Average StDev

0.03 0.05

0.05 0.t5

0502 0502

0.05 0.05

0.05 0.05

PA 01 PA 02 PA 03 PA 04

62.3 64.7

6.59 6.05 7.00 8.85

1.34 1.75

1.44 2.09

3.39 1.76

2.53 1.97

2.30 1.85

1.05 1.17 1.09

0.12 0.16 0.16

0.47 1.55 1.05 0.22

<0.03 <0.03 <0.03 0.03

0.73 0.59

0.29 96.8

0.21 100.2

0.26 98.3

0.20 97.4

Average StDev

1.13 3.052

1.03 0.t5

O.U 0503

0.05! 0.05

0.0 31. M

wH 01 wH 02 wH 03 wH 04 wH 05 wH 06 wH 07 wH 08 wH 09 wH 10

wH 12 wH 13

wH 15 wH 16 wH 17 wH 18 wH 19 wH 20 wH 21 wH 22 wH 23 wH 24 wH 25 wH 26 wH 27 wH 28 wH 29

5.77 5.32 6.04 7.45 6.86 7.57

6.25 7.93

7.47 8.79

7.26 5.93 7.25

5.62 7.57 8.28 7.19 7.56 5.88 8.52 6.50 8.06

4.63 8.88 9.73 6.55

0.38 0.40 0.46 0.52 0.50 0.57

0.55 0.49 0.52 0.49 0.68 0.65 0.43 0.69 0.52 0.81 0.57 0.69 0.54 0.47 0.55

0.65 0.56

0.68 0.86 0.82 0.76 0.81 0.87 0.86 0.97

1.13 0.98 0.82 0.84 0.68 0.98 1.02

1.14 1.04 0.98 1.02 0.79 1.37 0.85 0.97 1.17 1.20 0.95

1.77 1.99 2.10 2.04 1.97 2.32 2.23

2.44 2.24 2.20 1.96

2.19 2.40 2.30 2.66

2.39 2.53 2.62 2.53 2.96 2.82 2.82 2.26 2.43 2.34

0.49 0.66 0.56 0.60 0.58 0.76 0.81 0.76 0.70 0.98

0.95 0.88

0.88 1.00

0.09 0.12 0.09 0.09 0.10

0.12 0.12 0.12 0.12 0.13 0.13 0.14 0.14 0.15 0.15 0.15 0.16 0.17 0.17 0.18 0.18 0.19 0.20 0.22 0.25 0.26

<0.3 <0.3 0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3

<0.3 <0.3 <0.3

1.16 1.26 0.54 0.45 1.15 0.49 0.94

1.65 1.69 1.59 0.27 1.33 0.83 0.92 0.23 0.14 0.01 1.17 0.58 0.60 2.01 2.43 0.89

<0.03 <0.03 0.03 0.03 <0.03 <0.03 <0.03 <0.03 <0.03

<0.03 <0.03 <0.03 <0.03 0.03 2.78 <0.03 5.88 <0.03 <0.03 2.40 <0.03 0.55 0.03 <0.03 0.03 <0.03 <0.03

0.02 0.02 0.02 0.07 0.04 0.05 0.05 0.10 0.07 0.14 0.10 0.07 0.07 0.13 0.07 0.07 0.06 0.09 0.08 0.16 0.09 0.20 0.03 0.04 0.08

0.04 0.07 0.10

1.01 1.20

1.07 1.27 0.97 1.27 1.23 0.84 1.04 0.89

1.08 1.11 0.86 1.12 0.80 0.78 1.00

0.99 1.14

1.03 1.02 0.99 1.05

1.04 0.95 0.89

0.24 0.17 0.23 0.21 0.40 0.30 0.27 0.42 0.32 0.38 0.20 0.28 0.33 0.24 0.35 0.32 0.26 0.36 0.34 0.36 0.37 0.31 0.29 0.15 0.28 0.32 0.22 0.28 0.40

98.8 99.7

98.6 101.1

99.8 99.0

98.6 100.2

98.0 99.3

98.8 99.0 97.2

Average StDev

O.U O.U

2532 052!

h 33 051!

0530 0507

H 01 H 02 H 03 H 04 H 05 H 06 H 07 H 08 H 09 H 10 H 11 H 12 H 13 H 14 H 15 H 16 H 17 H 18 H 19 H 20 H 21 H 22 H 23 H 24 H 25 H 26 H 27 H 28

63.9 64.9

67.8 63.8

63.0 64.8

66.1 66.2 66.6

63.9 62.9 64.7

63.3 62.7 66.2

63.4 65.7 65.9

CaO 5.95 6.05 6.08 5.60 5.32 5.40

5.50 4.89 5.43

5.51 5.56 6.05 5.74 6.65 5.60 6.48 6.34 5.29 4.97 6.00

6.26 4.93 5.79 5.37 5.46

K2O 0.44 0.41 0.42 0.28 0.40 0.39 0.36 0.33 0.34 0.44 0.42 0.43 0.36 0.37 0.43 0.55 0.45 0.49 0.42 0.51 0.40 0.47 0.49 0.40 0.44 0.41 0.45 0.39

0.99 1.04 1.07 0.83 0.81 0.99 0.97 0.78

0.96 0.95

Al2O3 2.34 2.33 2.40 2.45 2.57 2.53 2.69 2.65 2.61 2.69

2.69 2.57 2.67 2.64

3.00 2.63 2.62 2.45 3.13

2.78 2.82 2.96

2.98 2.88

1.07 1.12

1.16 1.31 1.37 1.37 1.43

1.56 1.58 1.58 1.71 1.71 1.73

1.79 2.2 2.87

3.01 3.01 3.37 3.76

0.30 0.33 0.32 0.41 0.40 0.37 0.51 0.54 0.47 0.46 0.45 0.64 0.53 0.41 0.62 0.43 0.53 0.31 0.72 0.78 0.60 0.53 0.52 0.58 0.31 0.55 0.67

Sb2O5 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3 <0.3

MnO 1.85 1.80 1.91 2.07 1.98 1.52 2.30 2.25 2.25 2.45 2.05 2.54 2.17

1.95 2.74 2.13 2.77 1.86

1.96 2.13

7 477 2.87 0.90 1.50 0.96 2.831

0.05 <0.03 <0.03 <0.03 <0.03 <0.03 <0.03 <0.03 0.03 <0.03 <0.03 <0.03 0.03 <0.03 0.03

<0.03 <0.03 <0.03 <0.03 <0.03 <0.03 <0.03 0.03 <0.03 <0.03

0.05 0.04 0.05 0.04 0.05 0.05 0.03 0.05 0.04 0.03 0.05 0.04 0.06 0.06 0.04 0.09 0.05 0.08 0.04 0.05 0.06 0.05 0.09 0.09 0.11 0.00 0.14 0.19

Cl 1.17 1.11 1.14 1.23 1.03 1.06 1.08 1.07 1.07 0.92 1.03 0.84 1.14 0.97 0.81 0.87 0.87 0.98 1.00 0.86 1.06 0.99 0.83 0.86 1.07 1.01

SO3Analytical Total

0.25 0.24 0.28 0.15 0.19 0.22 0.25 0.27 0.25 0.21 0.22 0.26 0.23 0.24 0.27 0.31 0.24 0.32 0.18 0.29 0.23 0.23 0.45 0.29 0.26 0.24 0.26 0.18

100.0 96.9

100.1 98.9

97.8 97.3

97.9 98.5

97.2 98.9 97.9

99.6 99.5

98.7 99.7

0.99 0.25 98.50

0.11 0.06 1.02

ukn 72.0 16.3 4.95 0.60 0.55 2.82 0.47 0.07 <0.3 0.01 0.02 0.06 0.90 0.13 98.9

TIT r cm

II M I I i 1 I ■ I I I 1 II

Bubastis glass

♦ Sb decol ■ Mn decol w HIMT HIMT

( X ^ X

Wadi Natrun

3.0 AI2O3

0 5 10 cm

Plant Ash

Natron

0.0 1.0 2.0 3.0 4.0 5.0 6.0

Bubastis glass

w HIMT ■ HIMT Sb decol ^

Mn decol

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

A rrFPTFD MANUSCRIPT

Bubastis glass

w HIMT HIMT X Decoloured

Levantine I

* X* X X

2.0 2.5 3.0

We analysed 87 glass fragments from the Egyptian city Bubastis (1st.-6th centuries AD). They fall into 5 compositional groups: Sb dec, Mn dec, HIMT, weak HIMT and plant ash. Certain groups are absent, suggesting a strong local element in later glass consumption. By analogy Roman b/g glass might have been produced further away from the Nile delta Ungüentaría were made of plant ash glass attesting a continuous production in Egypt.