This substantial volume presents the results of long-term archaeological research carried out in the Eastern Desert of Egypt, starting in 1987 and continuing until 2015. During these nearly three decades, multidisciplinary investigations were developed under the indefatigable direction of Steven E. Sidebotham (University of Delaware), in cooperation with the University of Michigan in the early years (1986–93), and later with the University of Leiden and the Polish Center for Mediterranean Archaeology. In the mid-1980s, the construction of new roads, the availability of off-road vehicles, and the unlimited access suddenly open to archaeological teams brought about international projects in the Upper Egyptian Eastern Desert. While Sidebotham's team surveyed the ancient desert roads and settlements, extensive excavations by international teams under the aegis of the French Institute of Oriental Archaeology (IFAO) were carried out, first at the Mons Claudianus porphyry quarries,Footnote ¹ and later at several Roman forts or praesidia along the Koptos–Qoseir and Koptos–Berenike roads.Footnote ² The combination of these two different approaches – survey and excavation – has produced in the past years a considerable amount of new data, as reflected in the 57-page bibliography of this volume. These investigations cast new light on the pivotal position of the Eastern Desert in the flow of goods between Asia and the Mediterranean world during the Graeco-Roman period and later. The archaeological exploration still continues, with – among other projects – the excavations led by Sidebotham himself in Berenike, and the French mission (MAFDO) operating since the mid-1990s and currently managing the European Research Council (ERC) Desert Networks project, headed by Bérangère Redon.Footnote ³ In 2016, a number of scholars working in the Eastern Desert met in Paris, at the Collège de France, to present the current state of their research.Footnote ⁴ For nearly 20 years, a wide scientific community has convened every two years for international conferences about the Red Sea.Footnote ⁵ As our archaeological knowledge of the deserts and coasts is expanding at a tremendous pace not only in Egypt but also on the shores of East Africa, the Arabian peninsula, and India, new avenues of research are wide open for historians of the ancient world in the longue durée and on a global scale. In such a context, the long-awaited publication of the Eastern Desert survey appears as a timely contribution to this active research stream.

The first chapter is a dense introduction, written by S. Sidebotham with contributions by J. Gates-Foster, who joined the survey to study the pottery from the early 2000s. They first briefly contextualize the survey by emphasizing its connection with the excavations of Berenike since 1994 (1–3). Indeed, from prehistory, the southern Eastern Desert connected the Nile Valley with the Red Sea, and its history and archaeology cannot be understood without keeping in mind that it belongs to a network that includes Nile cities, Red Sea ports, and maritime roads. On the other hand, the numerous and diverse mineral resources of mines and quarries had always been a powerful incentive for the Egyptians to equip and use desert roads. The authors then describe the survey methods and goals (3–5). In a desert environment, understanding the modes of travel and transport (7–13) is an essential prerequisite for the archaeology of roads, in Egypt and elsewhere.Footnote ⁶ An even more essential concern is the water supply for residents, travelers, and animals (13–18): wells often feature as the focal point of Graeco-Roman fortlets, and cisterns are numerous in and around the forts and settlements. After briefly introducing maritime roads (18–19), the authors describe different types of desert roads (19–23), rarely paved but for the most part simply cleared and lined by windrows, of uneven width, and marked by cairns of piled stones instead of milestones (a specific feature of this region in the Roman world). But the most conspicuous markers of the roads are the small forts, some of which are Ptolemaic foundations reoccupied or developed in Roman times, along the roads or near the mines and quarries (23–27). These fortlets belong to a well-known type, built of local materials (cobblestones, small boulders) on a generally rectangular plan, often enclosing a well and defended, for the larger ones, by towers flanking the gate and reinforcing the angles and outer faces.Footnote ⁷ These praesidia,Footnote ⁸ housing small garrisons and some civilians, offered convenient stops and water supply for the caravans (extramural buildings for those are sometimes located nearby).

The authors raise the thorny issue of the number of persons living inside the fortlets (27–30). After calculating the interior area of each enceinte (43–44, Table 1.3), they try to estimate their maximum population (39–42, Table 1.2). The approach is somewhat questionable, given that the method applied should rather take into account the roofed floor space and not the whole enclosed area. For lack of extensive excavations and written evidence, very little is known about the buildings, their chronology, or the nature – military and civilian – of the occupation. Under much more favorable conditions, for the praesidia of the Myos Hormos road, H. Cuvigny pointed out the unreliability of such calculation methods.Footnote ⁹ Moreover, for a site such as Samut North (site 44), which has been recently excavated, the estimation of the archaeologists (100–126 persons in the fort), based on the identification of several rooms as dormitories,Footnote ¹⁰ differs widely from the 45–62 men proposed by Sidebotham (41). The existence of military patrols, attested by a number of documents but poorly defined, is also considered (32–38) and obviously could complicate even more the demographic estimates. The chronological phases of the road sites (30–32), as indicated by surface-collected material and some light excavation, confirm what was previously known through other archaeological data: the main occupation periods are Early to Middle Ptolemaic (mid-3rd to 2nd c. BCE), Early Roman (1st c. BCE to early 2nd c. CE) with a “lesser acme” of the late 2nd to early 3rd c. CE, and a last floruit between the end of the 4th and the early 6th c. CE. The chapter is rounded out by several tables, some of which were already mentioned above. The last one (44–49, Table 1.4) is a convenient list of the 69 sites of the gazetteer forming the third chapter, with their current and ancient names, coordinates, dates, and function. To this list must be added number 70 of the gazetteer, only very recently identified on the Berenike–Edfu road (Wadi Angoriyya, a small Christian settlement; see 4, fig. 1.1 and 283–85.

The second chapter (51–72), by James Harrell, clearly and concisely provides useful information about the geology, mines, and quarries of the southern Eastern Desert. The most exploited resources are the gold mines, some of which were being worked in predynastic times. If many of them already existed in the New Kingdom (ca. 1539–1077), the Ptolemies operated all the available mining sites at the same time, and on a larger scale.Footnote ¹¹ Roman remains are found on some, but not all, of the Ptolemaic gold sites, and in the Late Roman period, three new gold mines were open (53). But the Romans were also interested in gemstones (emeralds in particular) and ornamental stones such as the pegmatitic diorite of Jabal Umm Naqqat and the marble of Jabal Rokham. Steatite (soapstone) is also well attested through Roman times on both desert and coastal sites.

Chapter 3, the gazetteer of sites (73–285, by Sidebotham and Wright, with contributions by J. Gates-Foster and R. Tomber) is an essential contribution of the volume. The presentation follows first the Berenike-Edfu / Apollinopolis Magna road along its entire length (sites 1–50 and 70), and then the Berenike-Koptos road from the branch of Ed-Dweig (site 35, ancient Phalacro[n]?) to the valley (site 69, Matula). Each site description begins with important details, among which are the dates of the survey work and the pottery and other artifacts found by survey. The chapter, like the rest of the book, is lavishly illustrated; the photographs and the maps and plans (drawn and edited, often from mere sketches, by Jean-Louis Rivard, in a very clear and elegant style) offer the reader quite valuable help in following the descriptions because the sites frequently comprise a number of different, often shallow, features. Old maps drawn by early travelers, whenever possible, are included. The bulk of survey fieldwork was already completed in the 1990s and early 2000s, with only occasional verification visits in more recent years. While this publication was maturing, the archaeological activity in the Eastern Desert intensified, and a number of relevant books and reports appeared. The authors faced the challenge of absorbing and incorporating into their own work not only the already rich corpus of ancient and modern sources, but also this massive amount of fresh data. From this point of view, as well, the book is successful and offers the reader a valuable synthesis of the state of research on important sites, mainly forts, which have been excavated by the Mission Archéologique Française du Désert Oriental (MAFDO), such as the Samut gold mines (site 44, now published), or Girf/Xeron (site 52) and Abu Q(u)raya / Dios / Iovis (site 56), which is still awaiting full publication.

Returning to the survey methodology, the sites were generally visited several times and, in some cases, rapidly excavated. The material recovered through survey is almost exclusively pottery (see Chapter 4), a strong and valuable dating clue. In addition, random surface finds that may be significant are occasionally mentioned – for example, the prutah minted in Jerusalem under the procurator Porcius Festus during Nero's reign and found in the fortlet of Matula / Wekalat al-Mefarik, the first station when traveling from Koptos (281, site 69). This find strongly suggests that the fort was operated as early as the 1st c. CE under the Julio-Claudians, and that it was not built, like many others fortlets of the Myos Hormos road, under the Flavians.

While focused on the roads and connections between the major sites and their satellites, the team systematically recorded the main archaeological features of each site. Before the survey, some sites were known only by name (136, site 31). Others were utterly unknown, such as the unfortified station of Rawd al-Buram (238–48, site 54), which features light buildings but a tank of sizable dimensions and 12 tethering lines for animals. A Ptolemaic foundation at the intersection of the Edfu–Berenike and Edfu–Marsa Nakari roads,Footnote ¹² this station was probably occupied continuously until Late Roman times – and maybe later, given its proximity to steatite mines operated in the medieval era. Another noticeable discovery is the southern building at Dakhbaj/Daghbag (255–56, site 58). This site, probably the Compasi of ancient sources, was well known for its Roman praesidium and gold mines, with mining settlements ranging from the 2nd millennium BCE to the Ptolemaic period. Tentatively dated to the Early Roman (possibly Flavian) period, this newly discovered southern enclosure – located 5 km south of the main praesidium and the mines, and right in the middle of the wadi – has been almost entirely washed away by the floods, and it went unnoticed until now. The proposed dating raises the question, why were two large simultaneous enclosures built 5 km apart? Unfortunately, the major site was also heavily damaged in the years before being excavated, so the relative chronology and role of each precinct must remain undefined. Another discovery, in Abu Hagalig North (107–09, site 12), is a Late Roman enclosure, unfortified and apparently briefly occupied.

The survey also discovered less conspicuous remains described as “de facto stops,” unrecorded in previous literature. In Rawd Lekhaya/Legaya (238, site 53), on the road to Koptos about 4 km northwest of the praesidium of Xeron Pelagos, only faint traces of buildings (239, fig. 3.212), invisible on satellite views, suggest ancient (Roman?) occupation. At the beginning of the 20th c., a well still existed there, and it is surmised that water was also available in antiquity. Pottery sherds trimmed into Y- or T-shapes, found at both this site and several others, are interpreted as devices used to tighten tent ropes, offering evidence for the presence of (undatable) tent camps.

One of the main purposes of the forts built by the Greeks and Romans along the desert roads was to enclose and protect water resources for their own use, and most wells (hydreumata, the Greek term for the fortlets) were located in the center of the enceintes. In addition, the forts contain masonry basins or cisterns, which could be also partly filled by rainwater. It seems a habit of the Ptolemaic builders to equip their fortlets with twin tanks, a feature which can therefore be used as a dating clue.Footnote ¹³ In less hospitable places, access to the water was artificially procured by construction of hafayir (pl. of Arabic hafir), shallow circular tanks built to retain rain or flash-flood water (99–101, site 8, Wadi Khashir, dated to the Roman period; see also 267–69, site 62, Wadi Minayh el-Hir). Other devices for diverting and collecting water run-off are “dams” or conduits built between the mountains, and basins or cisterns, such as in Abu Hagalig South (107). Those features are often poorly preserved and difficult to recognize from surface observations, but the few occurrences described by the survey might help archaeologists to identify them at other sites.

Other stops are marked by graves and/or cairns, basic stone structures often in a dilapidated condition and sometimes difficult to distinguish from each other without excavation.Footnote ¹⁴ Such cairns may lie at the juncture of several wadis/roads (sites 6, 14), but the intersection of major roads was often equipped with substantial architecture, such as the praesidium of (Al-)Dweig (171–77, site 35, ancient Phalacron?) at the juncture between the Edfu and Koptos tracks to and from Berenike. In this area, despite erasure through continuous use of the roads, the survey succeeded in tracing several segments of ancient tracks (175, fig. 3.145); although the authors doubt that the praesidium was built precisely at this intersection (175), their map suggests the opposite. At Umm Qaria / Umm Ushra (site 32), thanks to sherd scatters left by temporary camps, the team recognized two different roads for the Ptolemaic and Roman periods (167, fig 3.133). It is suggested that this change was due to the switch from donkey to camel in Roman times (169). In other cases, roads might have moved simply to follow the variable water resources.Footnote ¹⁵

Mines and quarries were constantly operated throughout the ages, but the foci varied in successive periods: hard stones for statuary and architecture under the pharaohs, gold for the Ptolemies, gemstones and ornamental stones for the Romans – with none of these preferences excluding the search for other minerals. Mining sites usually abound in remains of all kinds, scattered across wide areas. This is the case particularly for the beryl mines of Mons Smaragdus, a region of around 30 × 20 km (131, fig. 3.82) including 15 of the sites described by the survey. The northernmost site, Gabal Zabara, was apparently connected directly with the Red Sea (132), so it is not included here. All over this area, significant buildings, settlements, rock-cut temples, cemeteries, skopeloi, and mines are impressively well preserved. One of the sites included in the survey (123–30, site 19, Kab Marfu'a) possibly served as a control point for the whole mining region, and it has yielded a corpus of particularly remarkable pottery (see infra, Chapter 5). The whole region certainly deserves further studies, all the more as modern destruction increases. Indeed, some buildings of Siqait (136–45, site 22) were excavated by Sidebotham and the survey team, who also carried out light conservation work, and the mines are currently under exploration by a Barcelona team headed by Joan Oller Guzmán.Footnote ¹⁶ Much remains to be done on a regional scale to better trace the network of sites and the way each of them operated through time and space.

As already mentioned, most of the material used for characterizing and dating the sites is pottery. In Chapter 4 (287–375), J. Gates-Foster embarks on the difficult task of summarizing the results of the pottery studies. Even more than for the site gazetteer, the quality of information is highly variable: the pottery was collected, mostly on the surface, by different persons over a number of years (288 and 290–92, Table 4.1). Moreover, several collections are still under study by different scholars,Footnote ¹⁷ so only 27 sites are presented here (294, Table 4.3). The study benefited from well-published parallels from the urban sites of Koptos and Berenike, which provide valuable tools for dating and analyzing the assemblages.Footnote ¹⁸ Table 4.8 (308–9) summarizes the dates of occupation observed on every site. Gates-Foster reached interesting new conclusions about the Hellenistic period, especially regarding the transitions from Egyptian dynasties at the one end to Romans at the other. First, it appears that some sites were occupied continuously from the Late Period (6th–4th c. BCE), and the early Ptolemies followed – more closely than previously thought – in the footsteps of the last Egyptian (Saïte-Persian) dynasties (288–89), especially on mining sites (304–5). Second, marl clay is predominant in the fabric of the Late Period to Early Hellenistic pottery, and the proportion of Aswan products is higher than previously recognized (293–97). This suggests that commercial roads along the Nile and/or into the desert were more active than previously thought. Third, some features of the assemblages, when compared to Upper Egyptian urban sites, clearly reflect specific conditions of consumption in the desert sites (298, Table 4.5): for example, vessels used for transporting or serving liquids dominate the assemblages, and, in some cases, form the only category (301, Table 4.7); the distribution of cooking ware suggests that at some sites, no cooking facilities existed at all (299–300). Conversely, the keg, a barrel-shaped short-necked transport vessel originating from the Western Desert,Footnote ¹⁹ appears as a marker of desert material culture all over Egypt; found at every site, it seems to have been produced also in Aswan. The peak of Ptolemaic sites lasted until the early to mid-2nd c. BCE, but afterward, they were only sparsely occupied or visited until the Roman revival at the turn of the Common Era. The treatment of the Roman pottery is far more cursory because the corpus is well known through other sites (306–11). The number of imports, very high in Early Roman times, decreased drastically thereafter. A catalog of Hellenistic to Early Roman pieces from 25 sites, and a collection of significant pieces of Roman pottery, round off the chapter.

The question of Roman pottery imports is best illustrated by the sherds from Kab Marfu'a (site 19, in the Mons Smaragdus region), presented by Roberta Tomber (377–87, Chapter 5). The site was occupied from the 1st to the 5th c. CE. Wine amphorae (Dressel 30) are dominant, especially between the mid-2nd and 4th c. CE, with a wide range of origins: Gaul, Baetica, Cilicia, Mauretania Caesariensis, Africa Proconsularis, plus a few unprovenanced (but certainly imported) containers. Ruling out the presence of foreign communities, Tomber rather explains these imports as (maybe one-off) state consignments (380).

The final chapter (389–413, Chapter 6), by Hans Barnard, outlines the long history of native desert dwellers. Egyptian sources documenting these people are scarce, and it is only from the 1st c. CE onward that a specific pottery – the Eastern Desert Ware (398, table 6.1) – betrays the presence of kindred groups (Blemmyes?) roaming the desert. The ethnographic parallel with the modern Beja bedouins may help us better understand the way the ancient desert dwellers lived and moved. At the end of the book are a copious bibliography (415–72) and very useful and detailed index of proper names and main topics.

To conclude, this publication must be praised as a major achievement of the survey teams in the field and, even more, of the two editors, who succeeded in presenting the results of research carried out by many scholars over a considerable time in a convenient, homogeneous, and accessible format. It has already become a reference handbook for researchers working in the Eastern Desert – all its original visual material and archaeological data have been immediately integrated into the database of the ERC Desert Networks project, recently published online (https://desertnetworks.huma-num.fr). Notwithstanding some discrepancies in the toponymy or zoning of the sites,Footnote ²⁰ it will be particularly easy, useful, and profitable for the readers of this book to use its gazetteer together with the database, which also includes data from still-unpublished works of the French mission on some of the surveyed sites (e.g., for Abu Garahish, sites 40–41, excavated in 2016, or Abbad, site 50, excavated in 2017–2018). The satellite maps of the database complement the survey maps and photos for better visualization of the respective locations of and distances between the archaeological features. Last, the online gazetteer also records occupations prior to the Graeco-Roman period, allowing a longer diachronic view of the human presence in the desert. By combining different approaches and working in close contact, the international scholarly community involved in the southern Eastern Desert is cooperating to produce valuable tools for the numerous specialists interested in this region.

Sidebotham and Gates-Foster advance our knowledge of the Eastern Desert with a significant step forward. It is now hoped that the call launched to initiate or continue investigations in some key areas and sites will be heard before the remains are completely destroyed.