Introduction
The dynamics that regulated the movements of people and trade goods from the Roman to the early Islamic period (first–eighth centuries AD) in what today is the desert of southern Jordan are only partially understood. After the Roman annexation of the Nabataean Kingdom in AD 106, the pre-existing road network was integrated into the Via Nova Traiana (AD 111–114), a road that crossed the region from north to south, connecting Bosra, the capital of the province of Arabia, with the Red Sea port of Aila (modern Aqaba), passing by Petra, the former Nabatean capital (Graf Reference Graf1993; Abudanah et al. Reference Abudanah, Tarawneh, Twaissi, Wenner and Al-Salameen2016). New evidence shows that Petra prospered through the Byzantine period (fifth–seventh centuries AD) and was not, as previously thought, abandoned after the AD 551 earthquake (Al-Nasarat & Twissi Reference Al-Nasarat and Twissi2016). This change of perspective involves a reconsideration of longue durée phenomena at desert and semi-desert sites connected to Petra by the Via Nova Traiana. The survey project launched by the Universities of Aarhus (Denmark), Alberta (Canada) and Milan (Italy) in 2023 at Khirbet al-Khalde, a site located for strategic purposes in the Wadi al-Yutm, 26km north-west of Aila, aimed at clarifying the chronology of this overlooked commercial and military site (Harvey et al. Reference Harvey, Intagliata, Mokránová, Raja and Zoppi2024a, Reference Harvey, Intagliata and Raja2024b) (Figure 1). Khirbet al-Khalde is shown on the Peutinger Table (the c. 1200 copy of a fourth-century-AD map showing a representation of the inhabited world with emphasis on the road network) as Praesidio (Stein Reference Stein1940: 437) and the site comprises a fortlet, a smaller structure, which is assumed to be a caravanserai (an inn for travellers), watchtowers, a cemetery, several cisterns, as well as an aqueduct almost 1km in length (Harvey, Reference Harvey, Intagliata and Raja2024c, Reference Harvey, Intagliata and Raja2025) (Figure 2).
Map showing the location of Khirbet al-Khalde (figure by Khirbet al-Khalde Archaeological Project).
Overview of main site, viewed from the south (photograph by Khirbet al-Khalde Archaeological Project).
Previous surveys at the site identified a Nabataean building phase and a major phase of activity dating to the Roman period and presumably ending during the fifth century (Kennedy Reference Kennedy2002; Kennedy & Bewley Reference Kennedy and Bewley2004: 191). Ceramics, collected during the 2023 survey, do not provide conclusive evidence for later frequentation of the site, and thus the question of the continuity of activity after the fifth century AD has, until now, remained open. This article discusses the results of an interdisciplinary study of glass collected at Khirbet al-Khalde, with the primary aim of clarifying the presence and extent of a post-fifth-century frequentation. Glass, through comparison of known base-glass compositions and artefact typologies, can help clarify the dating of certain contexts. Although the fragmentary nature of some samples from Khirbet al-Khalde makes a clear typological classification difficult, comparing the chemical composition of the samples with known base-glass types makes possible the refinement of the existing chronology. During the Roman period, glass was predominantly made in few workshops located in Egypt and on the Levantine coast and was exchanged over long distances (Jackson Reference Jackson2005; Freestone Reference Freestone2015; Phelps et al. Reference Phelps, Freestone, Gorin-Rosen and Gratuze2016; Schibille et al. Reference Schibille, Meek, Tobias, Entwistle, Avisseau-Broustet, Da Mota and Gratuze2016; Jackson & Paynter Reference Jackson and Paynter2016; Freestone et al. Reference Freestone, Degryse, Lankton, Gratuze, Schneider, Rosenow, Phelps, Meek and Freestone2018). Thanks to the ever-growing corpus of published chemical data, ancient base-glass compositions can now be distinguished and dated more securely. Between the first century and the end of the third century AD, the glass market was dominated by two base glasses attributed to Egyptian and Levantine furnaces (Brems & Degryse Reference Brems, Degryse and Degryse2014; Freestone Reference Freestone2015; Barfod et al. Reference Barfod, Freestone, Lesher, Lichtenberger and Raja2020). Starting from the middle of the fourth century, new Egyptian and Levantine compositions appeared on the market, reflecting the development of a less centrally located glass industry (Phelps et al. Reference Phelps, Freestone, Gorin-Rosen and Gratuze2016; Schibille et al. Reference Schibille, Meek, Tobias, Entwistle, Avisseau-Broustet, Da Mota and Gratuze2016; Freestone et al. Reference Freestone, Degryse, Lankton, Gratuze, Schneider, Rosenow, Phelps, Meek and Freestone2018; Balvanović & Šmit Reference Balvanović and Šmit2022; Freestone et al. Reference Freestone, Barfod, Chen, Larson and Gorin-Rosen2023). Our study is the first major interdisciplinary investigation of glass from the Wadi al-Yutm. The identification of late base-glass compositions in Khirbet al-Khalde demonstrates that frequentation of the site continued at least until the eighth century.
Archaeological context
Khirbet al-Khalde is presently accessible only by dirt roads, but in antiquity it was located at a strategically important point along a Nabataean trade route, which ran to Aqaba and was later integrated into the Via Nova Traiana (Abudanah et al. Reference Abudanah, Tarawneh, Twaissi, Wenner and Al-Salameen2016). The now-abandoned railway line between Ma‘an and Aqaba is located only a few metres east of the ancient site. The construction in 1975 of the railroad heavily damaged parts of the structures and probably also contributed to the high level of modern looting at the site (Harvey et al. Reference Harvey, Intagliata, Mokránová, Raja and Zoppi2024a). The main features at Khirbet al-Khalde are two large buildings: a fortlet measuring approximately 65 × 40m, and a smaller building identified as a caravanserai (Figure 2). Two possible middens were located next to these two main structures. The remains of two watchtowers are visible, respectively, on a hill just south of the main site and on a steeper hill, just to the west, on the other side of the modern railroad (Harvey et al. Reference Harvey, Intagliata and Raja2024b). A cemetery and two rock-cut cisterns can be found on the north-east side of the site. The water supply was ensured by an aqueduct, located to the south-east and originating in the hills, at ‘Ain or Moyet el-Khaldeh, 1km from the site (Harvey et al. Reference Harvey, Intagliata, Mokránová, Raja and Zoppi2024a).
In the summer of 2023, a surface survey was undertaken at Khirbet al-Khalde, to clarify the chronology of the site and its structures, as well as to assess the degree of damage over time. Fieldwork was conducted over a two-week period and included an intensive surface survey, with the collection of all visible finds for further study. A new and precise plan of the site was produced, including a mapped survey of the aqueduct. The site was divided into nine sectors (A–H & J), with the survey covering 20 000m2 (Figure 3, Table 1). The survey collected approximately 26 000 finds, including 84 glass fragments (Harvey, Reference Harvey, Intagliata and Raja2024c: 173). Apart from two fragments from sector H, all glass fragments were found in sectors A, B and C. The study of ceramics determined that only a handful of the more than 25 000 sherds may possibly date to after the fifth century AD.
Map of Khirbet al-Khalde showing the sectors explored during the survey (figure by Khirbet al-Khalde Archaeological Project).
Sectors of Khirbet al-Khalde covered by surface survey.
Materials and methods
Identification of joining pieces indicates that the 84 glass fragments from Khirbet al-Khalde stem from 80 objects. The fragments were studied with regard to their forming technology and typology. One fragment was excluded from the study, because it is clearly modern. The assemblage includes naturally coloured and colourless vessels: 21 diagnostic fragments with an additional four from the decorated walls of vessels (Figure 4), 33 non-diagnostic (Figure 5), 20 naturally coloured windowpane shards (Figure 6) and one black bracelet (KG082: Figure 4). Glass vessel fragments were classified according to the typologies of Isings (Reference Isings1956) and De Tommaso (Reference de Tommaso1990). Glass from all 79 non-modern objects was sampled for chemical analysis by LA-ICP-MS (laser-ablation inductively coupled plasma mass spectrometry). The cleaned but unprepared samples were analysed using a Thermo Fisher Scientific ELEMENT XR mass spectrometer combined with a Resonetics M50E excimer 193nm laser at the Institut de Recherche sur les Archéomatériaux Centre Ernest-Babelon in Orléans (France), following an established analytical protocol (Gratuze Reference Gratuze and Janssens2013, Reference Gratuze, Dussubieux, Golitko and Gratuze2016; Schibille et al. Reference Schibille, Klesner, Neuville, Stark, Torgoev and Mirzaakhmedov2024).
Diagnostic fragments of glass vessels collected during the survey (figure by Cristina Boschetti, Khirbet al-Khalde Archaeological Project).
Non-diagnostic fragments of glass vessels collected during the survey (figure by Cristina Boschetti, Khirbet al-Khalde Archaeological Project).
Fragments of windowpanes collected during the survey (figure by Cristina Boschetti, Khirbet al-Khalde Archaeological Project).
Typology and forming technology of the glass finds
Vessels
Identification of the forming technology used to produce the glass is the first criterion for grouping artefacts and building a typology. The vessels from Khirbet al-Khalde were made using two different forming techniques: slumping (pressing a disc of hot glass on an open mould and finishing the surface of the cold vessel by wheel polishing) and free blowing. Only 21 fragments are diagnostic parts of vessels (Figures 4 & 7). Of these, two are decoloured bowls with a flat and expanded rim produced by slumping (KG005, KG011). These are similar to a group of first- to second-century AD bowls excavated in the emporium of Myos Hormos on the Red Sea and are probably Egyptian imports (Peacock Reference Peacock, Peacock, Blue and Whitewright2011). All the other vessels are free blown. Eight fragments (KG030, 046, 051, 073, 074, 076, 079, 083) are of cracked-off and fire-rounded rims, five of which (KG030, 051, 074, 076, 079) are sufficiently large to be attributed to beakers of the first- to third-century Isings 12 or of the fourth- to eighth-century Isings 106 type (Keller Reference Keller2006; Jones Reference Jones, Oleson and Schick2013; Swan Reference Swan, Davies and Magness2015: 529). High-quality examples of these beakers are often decorated by lathe-cut lines, such as the four wall fragments found at Khirbet al-Khalde (KG006, 025, 034, 037). These high-quality beakers are interpreted as Egyptian imports (Nenna Reference Nenna2009). Only one colourless cracked-off and ground rim (KG069) can be associated with the Isings 38 beaker, a type that was popular in Jordan during the third century (Swan Reference Swan, Davies and Magness2015: 145).
Drawings of diagnostic fragments of vessels collected during the survey (figure by Miriam Romagnolo, Khirbet al-Khalde Archaeological Project).
Two fragments of decoloured, thick-walled blown vessels are engraved (KG015, 080). Engraved vessels begin to appear in the first century AD and are attributed to specialised workshops, probably located in different parts of the Mediterranean, including Egypt (Foy et al. Reference Foy2018). These glasses travelled long distances, as documented by their discovery as far away as Begram, in Afghanistan (Oliver Reference Oliver1984; Whitehouse Reference Whitehouse2001; Grose & Scott Reference Grose and Scott2017). The specimens from Khirbet al-Khalde include a fragment of the rim of a cup (KG015), which is too small for further diagnostics, and part of the rim of a dish with lathe-cut lines (KG080), which can be associated with wide-rimmed bowls excavated in Red Sea ports, including Aila (Jones Reference Jones and Janssens2000; Peacock Reference Peacock, Peacock, Blue and Whitewright2011: 68–69). The thick-walled base of a cup with foot (KG024) is most likely Egyptian, comparing well with vessels from Egypt dating to the first to second centuries AD (Kucharczyk Reference Kucharczyk2005). Finally, the bottom of one tubular unguentarium (a small flask for fragranced liquids) can be identified as de Tommaso 73 (KG077), a popular type from the middle of the first to the third centuries AD (de Tommaso Reference de Tommaso1990). Two small fragments of rims with lathe-cut lines (KG033, 050) and one vessel base (KG018) cannot be securely interpreted.
Three fragments (two rims and one base) belong to types developed from the fourth century onwards. The first is the rim of a jug (KG008); although small, the fragment retains a thick filament applied around the neck, a form documented in Jordan between the fourth and eighth centuries (Meyer Reference Meyer1988: 195–97; Swan Reference Swan, Davies and Magness2015: 149; Jackson-Tal Reference Jackson-Tal, Lichtenberger and Raja2021: 17). Two other fragments can be dated between the sixth and the seventh/eighth centuries: a tubular base-ring with rounded high push-up base (KG063) and the trefoil-folded rim of a pitcher (KG068) with parallels in Gerasa and Yotvata, Jordan, and in Khirbat al-Karak and Mezad Tamar, Syro-Palestine (Meyer Reference Meyer1988; Swan Reference Swan, Davies and Magness2015: 149–50; Jackson-Tal Reference Jackson-Tal, Lichtenberger and Raja2021).
Windowpanes
Early glass windows are documented from the middle of the first century AD, when they were formed by manually stretching a gob of hot glass, while forming technology advanced at the beginning of the second century, with the introduction of blowing (Max Reference Max2020; Allen et al. Reference Allen, Taylor and Hill2023). The 20 windowpane fragments from Khirbet al-Khalde are very thin (1.3–3mm) and were thus all produced by blowing; in some cases, the fragments retain the remains of the plaster used to fix them to their frames, indicating they were used in the structures (Figure 6). The date of the introduction of blown windows therefore provides a terminus post quem for the specimens from Khirbet al-Khalde.
Body ornaments
The fragment of a bracelet (KG082) is the only glass body ornament collected during the survey (Figure 4). It has a flat interior, sub-circular section and pinched ribs on the outside. Similar black-glass bangles, together with black beads, pendants and finger rings were popular across the Roman Empire during the third and fourth centuries (Cosyns Reference Cosyns2011). Published chemical analyses of these ornaments include both Egyptian and Syro-Palestinian compositions (Cosyns Reference Cosyns2011; Boschetti et al. Reference Boschetti, Gratuze and Schibille2020, Reference Boschetti, Gratuze, Cavalieri, Lenzi and Schibille2021a).
Chemical composition
The full chemical compositions of the 79 samples analysed for this study are presented in the online supplementary material (OSM Table S1). With the exception of one sample, all the fragments are natron glasses, with low magnesium and potassium oxide concentrations (<1.5wt%), in accordance with the dominant technological trend of the Roman period (Sayre & Smith Reference Sayre and Smith1961; Shortland et al. Reference Shortland, Schachner, Freestone and Tite2006). The wall of blown vessel KG002 is the only plant-ash glass identified, with higher magnesium (2wt%), potassium (1.6wt%) and phosphorus oxides (0.6wt%). The ratios of alumina and silica (Al2O3/SiO2) versus titanium and alumina (TiO2/Al2O3) and of yttrium and zirconium (Y/Zr) versus cerium and zirconium (Ce/Zr) are good markers to distinguish between the different silica raw materials used by ancient glassmakers (Figure 8a & b). It should also be noted that Roman glass was commonly decolourised during glassmaking. The Egyptian workshops privileged antimony, while manganese was preferred in Syro-Palestine (Jackson Reference Jackson2005; Gliozzo Reference Gliozzo2017; Paynter & Jackson Reference Paynter and Jackson2019). The result was a glass with concentrations of antimony (Sb) or manganese (Mn), respectively, above the levels naturally occurring in silica raw materials (Sb>30ppm or Sb2O3>0.004%, Mn>250ppm or MnO>0.03%) (Schibille Reference Schibille2022: 33).
Elemental analysis of glass fragments: TiO2/Al2O3 and Al2O3/SiO2 (a) and Ce/Zr and Y/Zr (b) ratios reflect the silica source used to make the glasses. The simultaneous presence of both Mn and Sb above background levels (c) is indicative of mixing Sb and Mn-decoloured glass (figure by Nadine Schibille).
First- to third-century AD glass
A first group of vessels (Roman-Sb, n = 16), with relatively low alumina, calcium, transition metals and rare-earth elements and elevated antimony is consistent with the glass made in Egypt between the first century and the end of the third century AD (Jackson Reference Jackson2005; Gliozzo Reference Gliozzo2017; Paynter & Jackson Reference Paynter and Jackson2019; Schibille Reference Schibille2022). These samples include one of the two slumped bowls (KG005), the high foot (KG024), the two thick-walled and engraved vessels (KG015-080) and the four walls of beakers with lathe-cut lines (KG006, 025, 034, 067). A second group (Roman-Mn, n = 10), consisting of eight fragments of vessels and two windows (KG027, 039), shares the same chronology as the vessels from the first group, but were made with glass produced on the Levantine coast (Jackson Reference Jackson2005; Gliozzo Reference Gliozzo2017; Paynter & Jackson Reference Paynter and Jackson2019; Schibille Reference Schibille2022). Compared to the Egyptian glasses, these have slightly higher alumina and calcium, lower soda and manganese levels elevated above typical silica sources (MnO >0.03wt%). One vessel wall (KG020) is the exception in the Roman-Mn group; the very low manganese content is sometimes documented in non-decoloured (natural aqua-coloured) Roman glasses produced in the Levant (Jackson & Paynter Reference Jackson and Paynter2016).
The presence of both antimony and manganese above natural levels is an indication of the mixing and recycling of Roman-Mn and Roman-Sb glass (Freestone Reference Freestone2015) (Figure 8c, Table S1). Half of the glass fragments collected during the survey have mixed Roman compositions (Roman-MnSb, n = 21 vessels, n = 18 windows). Increased phosphorus and potassium oxide content confirms recycling, as this contamination results from repeated exposure to the vapours of the furnace (Schibille & Freestone Reference Schibille and Freestone2013). The slumped cup KG011 contains no manganese and was presumably produced by mixing a Roman-Sb glass with the Roman natural aqua-coloured glass from Syro-Palestine. Finally, the fragment of bracelet (KG082) appears black due to high concentrations of iron and manganese, and also has elevated concentrations of heavy elements. It can be associated with Egyptian black glass production of the third to fourth centuries (Boschetti et al. Reference Boschetti, Gratuze, Cavalieri, Lenzi and Schibille2021a).
Fourth- to eighth-century AD glass
Eleven natron-glass vessels, including the three diagnostic fragments identified as later specimens through typology and one plant-ash non-diagnostic fragment of vessels (KG002), exhibit the new base-glass compositions that appeared between the middle of the fourth and the eighth centuries (Figure 8, Table S1). Some of the glass is recycled, as is frequently documented in the Early Middle Ages. A first group of samples have Egyptian compositions. The vessel wall made with plant ash (KG002) is a Magby (Magnesium-Byzantine) glass of the low-manganese type, mixed with some Roman-Sb glass. Magby is a high-magnesium Egyptian glass with higher heavy element content than earlier Roman glass and was made between the middle of the sixth and the end of the seventh century (Schibille et al. Reference Schibille, Meek, Tobias, Entwistle, Avisseau-Broustet, Da Mota and Gratuze2016; De Juan Ares et al. Reference De Juan Ares, Vigil-Escalera Guirado, Cáceres Gutiérrez and Schibille2019; Schibille Reference Schibille2022: 42–45). This sample is the first instance of Magby glass identified in Jordan. The rim of the jug KG008 appears, at first glance, to be a Foy 2.1 glass with some Roman-Sb glass. Foy 2.1 was made in Egypt from the second half of the fifth century until the seventh century and has an elevated heavy element content that could explain the trace element patterns observed in sample KG008. However, Foy 2.1 glass usually has high manganese levels (Cholakova et al. Reference Cholakova, Rehren and Freestone2016; Schibille et al. Reference Schibille, Sterrett-Krause and Freestone2017), which KG008 does not. Three vessel-wall fragments (KG060, 061, 062) are extremely close in composition and are most likely part of the same vessel. All have a mixed composition and may be the result of mixing either Foy 2.1 or HIMT (high iron, manganese and titanium glass) with some Roman-Sb glass (see group ‘HIMT rec’ in Figure 8). HIMT was produced in Egypt between the fourth and fifth centuries AD and is characterised by elevated levels of iron, manganese and titanium (Freestone et al. Reference Freestone, Degryse, Lankton, Gratuze, Schneider, Rosenow, Phelps, Meek and Freestone2018).
Finally, seven samples have late antique Syro-Palestinian compositions. The pitcher KG068 is a pristine Bet Eli’ezer base glass with low soda, high alumina and moderate lime levels, made in Syro-Palestine during the eighth century (Phelps et al. Reference Phelps, Freestone, Gorin-Rosen and Gratuze2016). This sample defines the latest period of frequentation at Khirbet al-Khalde. The same date can be assigned to six other vessel fragments, including five cracked-off and hot-polished rims of beakers (KG073, 074, 076, 079, 083) (group ‘Roman Mn & Levantine’ in Figure 8). The recycled composition of these samples, with low soda (12.14 < Na2O < 13.41) and manganese above the background level of sands, is most likely a mixture of Bet Eli’ezer and Roman-Mn glass (Phelps et al. Reference Phelps, Freestone, Gorin-Rosen and Gratuze2016; Freestone Reference Freestone and Tal2020).
Definition of the chronology of Khirbet al-Khalde
Windowpanes and the second-century AD Roman renovation
The combined technical and chemical study of windowpanes helps refine the chronological framework of the Roman renovation of the buildings at Khirbet al-Khalde. The almost exclusive identification of recycled compositions for windowpanes, obtained by mixing Roman-Mn with Roman-Sb glass (Figure 8, Table S1), is particularly revealing. Egyptian Roman-Sb glass is rare before the end of the first century AD, becoming popular at the beginning of the second century (Foy et al. Reference Foy2018; Boschetti et al. Reference Boschetti, Furlan, Schibille, Raja and Bonetto2024). We can therefore assume that the windows were made after Roman-Sb glass became more easily available and blowing was introduced as a technique for making windows (i.e. from the early second century AD onwards). This date is coherent with the date presently attributed to the Roman renovation of the buildings (Kennedy Reference Kennedy2002; Harvey et al. Reference Harvey, Intagliata and Raja2024b) and with the period of construction of the Via Nova Traiana between AD 111 and 114.
Main occupation of Khirbet al-Khalde: first to fourth centuries AD
A total of 46 vessels (79%) are made of pristine or recycled Roman-Mn and Roman-Sb glass (Figure 8, Table S1). These vessels can be attributed to the Roman occupation of the site, which lasted from the late first/early second century to the third or fourth century AD. The vessel forms found at the site include a combination of objects used by the community living in Khirbet al-Khalde and possibly of vessels transported by caravans and discarded because they were damaged during the journey. The use of glass tableware and beakers at a frontier site is not surprising, as the custom of drinking wine from glass beakers was deeply rooted in Roman society. Glass drinking sets are often found in Roman forts, even in remote locations (Swan Reference Swan, Davies and Magness2015: 160). If only the fragments of vessels made from pristine glass are considered, we can observe that Roman-Sb glass dominates the corpus of Roman glass from Khirbet al-Khalde (62%). The finest vessels from the glass corpus (e.g. the beaker with lathe-cut lines and the engraved vessels) all have Roman-Sb compositions and reflect the importance of Egypt for the supply of fine glassware. The glass bracelet also belongs to the main period of occupation at the site. Body ornaments are not always documented at forts in Jordan and, when present, are interpreted as markers of a civilian presence (Swan Reference Swan, Davies and Magness2015: 159–60). Similar bracelets are found systematically in the graves of juvenile or adult females (Cosyns Reference Cosyns2011: 219). Jewellery cannot be used uncritically to indicate gender, but the bracelet from Khirbet al-Khalde may hint at the presence of women at the site.
Late activity at Khirbet al-Khalde: fourth to eighth centuries AD
The discovery of 12 pieces of glassware (21%) with pristine or recycled post-fourth-century chemical compositions, including sixth- and eighth-century compositions, provides the first clear evidence for activity at Khirbet al-Khalde continuing into the eighth century, when the Levantine Bet Eli’ezer glass was produced (Figure 8, Table S1). This small number of glasses with later compositions may reflect the continuity of trade through the Wadi al-Yutm, yet it also raises the possibility that a small community of people continued living at the site after the end of the Roman occupation.
Glass provenance and patterns of supply and recycling in Jordan
Our study provides the first combined chemical and typological analysis of pre-fourth-century AD glass from Wadi al-Yutm, expanding the available data, which are almost exclusively from later glasses. The provenance of these earlier glasses from Khirbet al-Khalde is inconsistent with the conclusions of previous studies. The identification of a dominant presence of Egyptian glass in either pristine or recycled form is in contrast with observations from Gerasa, Petra, Umm el-Jimal and Capitolias where, in each case, almost exclusively Syro-Palestinian compositions were found (Abd-Allah Reference Abd-Allah2010; Schibille et al. Reference Schibille, Degryse, O’Hea, Izmer, Vanhaecke and McKenzie2012; Al-Bashaireh et al. Reference Al-Bashaireh, Al-Mustafa, Freestone and Al-Housan2016; Boschetti et al. Reference Boschetti, Lichtenberger, Raja, Wootton and Schibille2021b; Barfod et al. Reference Barfod, Freestone, Jackson-Tal, Lichtenberger and Raja2022). The consistent presence of Egyptian glass at Khirbet al-Khalde probably reflects a change in glass supply after the fourth century and the existence of a network of glass working and recycling workshops in urban centres across the region. Such a network would explain why the recycled glasses from Khirbet al-Khalde differ from the recycled compositions previously documented in the region. Though we cannot yet definitively identify the location of any recycling workshops in southern Jordan, the city of Aila—where both Egyptian and Syro-Palestinian glass was certainly available—is a good candidate.
Conclusions
The combined typological and chemical analysis of glass from Khirbet al-Khalde provides the first clear evidence that activity at the site continued until at least the eighth century AD, contributing to our understanding of the longue durée of movement along the southern section of the Via Nova Traiana. Yet, the significance of these results goes beyond the reconstruction of site chronology, challenging the current model for the first-millennium Jordanian glass economy, which considers Syro-Palestine the main supplier of glass in the region. An in-depth investigation of other first- to fourth-century glass assemblages from Jordan will clarify this issue, but it is becoming increasingly clear that the glass economy of the region was more nuanced and less centrally organised than previously thought.
Funding statement
The Khirbet al-Khalde Archaeological Project is funded by the Augustinus Foundation, the Carlsberg Foundation, the Danish National Research Foundation (DNRF119) through the grant for Centre for Urban Network Evolutions (all the forementioned grants are held by Rubina Raja), and an American Center of Research grant (S. Thomas Parker Memorial Fund) held by Harvey. The directors thank the Department of Antiquities of Jordan and ACOR for support.
Online supplementary material (OSM)
To view supplementary material for this article, please visit https://doi.org/10.15184/aqy.2025.10248 and select the supplementary materials tab.
Author contributions: CRediT Taxonomy
Cristina Boschetti: Conceptualization-Lead, Data curation-Equal, Formal analysis-Equal, Investigation-Supporting, Methodology-Equal, Visualization-Equal, Writing - original draft-Lead, Writing - review & editing-Equal. Nadine Schibille: Conceptualization-Supporting, Data curation-Equal, Formal analysis-Equal, Funding acquisition-Lead, Investigation-Lead, Methodology-Equal, Resources-Lead, Visualization-Equal, Writing - original draft-Supporting, Writing - review & editing-Equal. Craig Harvey: Conceptualization-Supporting, Funding acquisition-Lead, Project administration-Lead, Writing - review & editing-Supporting. Emanuele Intagliata: Conceptualization-Supporting, Funding acquisition-Lead, Project administration-Lead, Writing - review & editing-Supporting. Rubina Raja: Conceptualization-Lead, Funding acquisition-Lead, Project administration-Lead, Writing - original draft-Equal, Writing - review & editing-Equal.
Open access
