Results of the Winter Excavations at Berenike Egypt & Related Fieldwork in the Eastern Desert

Transcription

1 Results of the Winter Excavations at Berenike Egypt & Related Fieldwork in the Eastern Desert Steven E. Sidebotham* and Iwona Zych** Under direction of the authors and Martin Hense fieldwork continued at Berenike (Red Sea coast), Egypt in winter Survey and excavation also took place in the Eastern Desert (between the Nile and the Red Sea) at that time. At Berenike excavations documented more structures and activities in the large southwestern harbor, in the early Ptolemaic urban fortifications and water supply facilities and uncovered early Roman-era human burials, some with grave goods. Excavations in the first century CE trash dump north of the city center recorded additional animal burials (mainly cats, dogs, and a few baboons and vervet or grivet monkeys); some animals still wore iron collars, one decorated with beads. Excavations recorded the foundations of a large enigmatic edifice west of the city and remains of a room in another building near the Great Temple (formerly called the Serapis Temple). Excavations also recorded two areas of the Great Temple, which contained a number of first-third century CE inscriptions in Greek together with fragments of stelai from the Egyptian Middle Kingdom Pharaoh Amenemhat IV of the eighteenth century BCE. Fieldwork in the Eastern Desert continued a total station survey to draw a detailed plan of the Roman-era beryl/emerald mining settlement at Nugrus. Excavations of the Neolithic and later long-horned cattle cemetery in Wadi Khashab progressed. Survey work also located and dated a previously unrecorded and relatively large fourth-sixth century CE settlement in Wadi Angoriyya. 1 hanks to the generous support of a number of donors, the University of Delaware (USA) (co-director Steven TE. Sidebotham) - Polish Centre of Mediterranean Archaeology, University of Warsaw (co-director Iwona Zych) continued fieldwork at Berenike (Red Sea coast), Egypt during the winter season (Figure 1). The project started excavations on December 20, 2014 and work ended on February 9, A study season prior to excavations began on November 22. Berenike was a principal player in the Old World global economy as an important nexus for commercial and cultural exchanges between * History Department, University of Delaware, USA ** Polish Center of Mediterranean Archeology, University of Warsaw, Poland 1 Financial support provided by the following: the Honor Frost Foundation, the Office of the Dean of the College of Arts and Sciences, the Institute for Global Studies and the History Department (all of the University of Delaware), Exxon-Mobil, the Seeger Family Fund, private donors (Mr. Norman Shelly, Mr. Marshall Parsons, Ms. Elizabeth H. Devries, Mr. Thomas E. Sidebotham, Mr. Johannes Girardi, manager of the Gorgonia Beach Resort, Ms. Silva Kluitenberg, Mrs. Van Zijl-de Bruin), Het Huis van Horus Foundation, The Museum for Family History (Netherlands), Ministry of Science and Higher Education of the Republic of Poland, the Polish Centre of Mediterranean Archaeology (PCMA) University of Warsaw, including logistical assistance from the PCMA Research Center in Cairo and its director Dr. Zbigniew E. Szafrański, the British Museum, and the authors Steven E. Sidebotham and Iwona Zych. Journal of Indian Ocean Archaeology No. 12,

2 Fig. 1: Map of the southeastern area of Egypt. Areas marked in gray are discussed in the text. Drawing by M. Hense. the Mediterranean World on the one hand and the Red Sea Indian Ocean littorals on the other. The preponderance of the evidence excavated since 1994 suggests that Berenike functioned from the third century BCE until sometime before the mid-sixth century CE (Sidebotham 2011). Some fragmentary inscriptions in hieroglyphs found this season may indicate, however, much earlier activity at the site. Additionally, excavations at a fifth third millennium BCE long-horned cattle cemetery in Wadi Khashab (Figure 1) and a total station survey undertaken to draw a detailed plan of the Romanera emerald mining settlement in Wadi Nugrus (Figure 1) continued from previous seasons. Other surveys in the Eastern Desert recorded a major late Roman (fourth-sixth century CE) settlement in Wadi Angoriyya (Figure 1). There was additional survey work undertaken at a World War II RAF airfield at Hawarry (Figure 1). 2 Journal of Indian Ocean Archaeology No. 12, 2016

3 Work at Berenike The project continued a magnetic survey of Berenike (Figure 2) (cf. Herbich 2007, 2011; Zych and Herbich 2015) and also excavated in whole or in part 12 trenches during the course of the winter season (Figure 3). Participants included dozens of staff from the USA, Poland, the UK, Spain, Romania, Canada, the Netherlands, Sweden and Egypt together with (at its peak) 76 locally recruited Ababda Bedouin workmen. The magnetic survey this season elucidated the northern fringes of the site, identifying more habitable areas from historical times. The lowintensity undisturbed spaces towards the north (to either side of the white area at top left of the map on Figure 2, which marks the unexplored part under the modern work/storage magazine) indicate areas of marshy ground, similar to the modern sabkha east and south of the site, which did not encourage occupation; this specific soil deposit formed as a result of the silting processes taking place in the wadi north of the city. This siltation was well advanced by the early Roman period when rubbish was being dumped here. The extent of these activities can now be traced, but the location of the early and late Roman era shorelines remains to be determined. The high-intensity anomalies curving across this part of the site and under the camp (visible on Figure 2) are not man-made, but are natural features, possibly reflecting reef formation processes determining the original landscape in the area millennia ago. The magnetic survey confirmed that the necropolis occupying a raised spur of bedrock at the northwestern limits of the site (presumed to be of Late Roman date, but with possible earlier burials as well), had not moved down the eastern flank of this ridge. This indicates in the Late Roman period that ground north of the site must have been too wet and marshy for use. The magnetic survey has not encompassed the entire area of this necropolis; Fig. 2: Magnetic map of Berenike. Survey by D. Święch and R. Ryndziewicz. Processed by T.M. Herbich. Journal of Indian Ocean Archaeology No. 12,

4 Fig. 3: Plan of Berenike with locations of trenches from the winter season. Drawing by M. Hense. hence its northeastern-most part may yet yield new evidence. Trenches excavated in Berenike s southwestern harbor Six of the 12 trenches excavated at Berenike in winter were in the southwestern harbor and two of these were continuations from previous seasons (Figure 3). Those in the southwestern harbor excavated during winter where last season the remains of an early Roman ship frame made of cedar wood had been discovered and, previously, cedar wood ship timbers joined using pinned mortise-and-tenon construction techniques - contained a range of structures and finds. The purpose of some structures excavated in the southwestern harbor this season remains unknown at this time. One was an edifice of uncertain function (Trench BE15-103: possibly a warehouse or work area) that lay west of an enigmatic sunken structure excavated in 2010 and 2011 (Trench BE10/11-70) (Sidebotham and Zych 2010: 16 17; Sidebotham and Zych 2012a: 36 37; Sidebotham and Zych 2012b: 136, ; Sidebotham 2014: ). Excavations also included part of the interior of a Late Roman Harbor Temple (Trench BE10/12/13/14/15-61), industrial and metal working areas (Trenches BE14/ and BE15-108) and a possible ship repair or ship dismantling facility (Trench BE15-109). There was also a small sondage (BE15-106) excavated to recover additional archeobotanical remains for flotation studies (Figure 3). Early Roman-era (first-early second century CE) trench BE (Figure 3) abutted and was immediately south of a trench where excavations in winter had documented a curved timber fashioned from cedar that proved to be the remains of an early Roman-era ship frame (in trench (BE14-98)(cf. Sidebotham and Zych 2010: 19 21; Sidebotham and Zych 2012a: 32 33; Sidebotham and Zych 2012b: ). The project anticipated documenting additional early Roman-era ship remains in BE and our excavations here did 4 Journal of Indian Ocean Archaeology No. 12, 2016

5 reveal a large timber, but it was so badly decayed that the species of wood and its function could not be conclusively identified (Figure 4, to the extreme left); though it was likely part of a dismantled ship. Excavations in this same trench also recorded a number of ropes, substantial remains of carbonized matting or basketry and branches of acacia (Figures 4 6). The latter appear to have been firewood, which is not surprising considering the heavy burning and thick ash deposits documented in this trench and others abutting to the north. However, excavations at the late Old Kingdom/Middle Kingdom/early New Kingdom (third second millennium BCE) port of Saww (in Mersa/Wadi Gawasis) ca. 23 km south of Safaga (not on the map in Figure 1) revealed ship timbers made of acacia as well as of sycamore and cedar (Bard and Fattovich 2011: 124 and note 136; 2013: 5). The nature of the remains in BE was not as pronounced as in trenches excavated immediately adjacent to the north in previous seasons, but the general impression from excavations in this row of trenches during this and previous seasons is that of a series of more or less makeshift shelters used as open storerooms for shipping paraphernalia, presumably for use on ships being repaired in the harbor. The dismantled parts of boats and ships and coils of thick mooring rope would suggest that activities like ship repair were carried out on the beach of the natural landing place, which existed in the southwestern bay, at the end of the long lagoon that stretched into the wadi mouth to the south. A sectioned and excavated area in the northwestern corner of BE revealed large bivalve shells (at least one matching pair), very likely of one of the pearl oyster Pinctada radiata species, that had been deliberately placed in the positions in which they had been excavated (cf. Schörle 2015: 48) (Figure 7). This species was originally distributed throughout the Indo-Pacific zone and was harvested for its pearls and nacreous shell, but also for its edible muscle (Zenetos, et al. 2015; Gervis and Sims 1992: 39). As the shells were found in place, it is possible that they had been harvested in this case for the meat and possibly for the pearls that had been secreted inside (Gervis and Sims 1992: 37 38). A number of other artifacts and pottery that would not have been at home in an industrial/workshop area (e.g. gold-in-glass beads) might suggest that merchants or shipping agents (possibly of Arab, East African or South Asian origin), rather than simple sailors and harbor workers, used or at least visited for business the flimsy shelters built on this side of the harbor at Berenike. Deep excavations in this same sectioned area in BE where the large bivalve shells were found recorded no activity in this part of the harbor prior to the Early Roman period (first second centuries CE). In addition, excavations in this trench and in adjacent ones in previous seasons recorded no activity in this part of the harbor after the second century CE. Excavations in industrial areas (trenches BE14/ and BE15-108) within the southwestern harbor (Figure 3), on the western side of the bay, recorded intensive production of iron and copper alloy nails, fittings and other metal objects. Excavations in trench BE15-108, at the extreme southwestern portion of the southwestern harbor (Figure 3), documented many long iron nails, fixtures and other items. These, together with evidence of heavy burning, suggested the presence of a furnace/workshop for making these metal items somewhere nearby; yet, excavations recorded no structures of any kind here. Of particular interest Journal of Indian Ocean Archaeology No. 12,

6 from BE was a bone ear plug, the first jewelry of its kind to be found at Berenike. Excavations in the other trench in this area, BE14/15-102, produced extensive quantities of terracotta crucibles with Fig. 4: Trench BE At far left is a large timber (whitish in color), to right (and above) the photo scale are ropes, matting and remains of acacia branches (to right of scale). View looking east. Scale = 20 cm. Photo by S.E. Sidebotham. Fig. 6: Detail of Figures 4-5. Ropes and matting. View looking east. Scale = 20cm. Photo by S.E. Sidebotham. Fig. 5: Detail of Figure 4. Ropes, matting and acacia branches. View looking east. Scale = 20 cm. Photo by S.E. Sidebotham. Fig. 7: Bivalve shells (likely Pinctada radiata) in a sectioned portion of the northwestern corner of trench BE View looking east. Scale = 20 cm. Photo by S.E. Sidebotham. 6 Journal of Indian Ocean Archaeology No. 12, 2016

7 remains of copper-alloy metal residue adhering to them (Figure 8). The construction of furnaces here allowed prevailing northerly and northwesterly winds to carry the smoke and fumes towards the south and away from the city center except on those few days of contrary winds blowing from the south or west. Rubbish deposits appeared in some of the top levels in trench BE14/ and, apparently, had little to do with the industries recorded in the lower/ earlier strata. An intaglio of a mounted horseman had been found during the winter season and this season excavations documented another intaglio. Carved on an oval shaped cabochon likely made of carnelian (length 2.54 cm x width 1.87 cm x thickness 0.42 cm), this specimen had engraved on it a draped female figure standing left (Figure 9); stylistically it seems to date to the late Hellenistic period (second first century BCE). The female figure depicted on the cabochon was likely a deity. The block-like featureless head and the disproportionately big and apparently tri-forked right hand, as if a flowering bud had replaced the fingers, as well as the extraordinary perspective for showing the other, left hand, all seem to suggest more than just a naturalistic representation of a deity. The earliest levels reached in trench BE14/ were late Ptolemaic/early Roman (late first century BCE) and earlier Ptolemaic data are expected when excavations continue in this area in the future. That our excavations recorded such artifacts as carved intaglios in trench BE14/ and, only a meter or so away from trench BE09-55, which produced a rare molded glass beaker with relief almond-shaped decoration (cf. Kucharczyk 2015: 98, entry 20 & 99, Fig. 9-12, no. 20), suggests that some other non-industrial high status activities lay close by within or near to the southwestern harbor. The identification of one or more buildings portions of which appeared in trench BE (west of trenches BE15-61 and BE10/11-70: cf. Fig. 9: Intaglio from a ring or necklace, likely made of carnelian. Second/first century BCE. Photo by S.E. Sidebotham. Fig. 8: Crucible fragments from trench BE14/ Scale = 10 cm. Photo by S.E. Sidebotham. Journal of Indian Ocean Archaeology No. 12,

8 Sidebotham 2014: ) (Figure 3) in the southwestern harbor remains uncertain. The walls here were Late Roman (fourth fifth century CE), built of the same fossilized coral heads seen in contemporary structures elsewhere on the site. They would have formed part of a complex west of the Late Roman Harbor Temple/Lotus Temple serving, perhaps, as work areas or storage facilities. By that late date the temple and affiliated structures constituted facilities in the abandoned and silted up harbor. Of particular interest was the find of 40 fragments of garnets imported from India. Since garnets were exports from South Asia, these semiprecious stones are one indication of a vibrant trade with that part of the world in Late Roman times, something already known from the results of excavations conducted at Berenike in earlier seasons. They may also point to the nature of the cult and/or the people who visited the Late Roman Harbor Temple/Lotus Temple. Trench BE15-61 (Figure 3) lay inside this late Roman temple, which had already been excavated extensively since 2010 (Sidebotham and Zych 2010: 15 19; Sidebotham 2011: ; Sidebotham and Zych 2012a: 33 36; Sidebotham and Zych 2012b: ; Rądkowska, et al. 2013; Sidebotham 2014: ). Excavations this season reached pre-temple levels. The nature of the activities that preceded the building of the late temple remains uncertain, although there is some reason to believe that it, too, was cult-related. Analysis of the magnetic map of the area suggested the presence of an island or peninsula, on which the Late Roman temple was constructed, reusing some of the old architecture and artifacts recovered from nearby ruins. Careful exploration of a huge piece of coral that had been a sacred object in the late temple revealed layers that resembled the stratigraphy of an active spring. Should this theory be verified by future specialist examination, it could explain the sanctity of the shaft with the coral head that lasted for some time, but which fell out of use before the final phases of activity in the temple itself. Additional work here in the future should document more about this period of history in this part of the harbor. Excavations elsewhere in Berenike Outside the southwestern harbor excavations recorded remains of dismantled ships or other types of wooden objects constructed using mortise-andtenon techniques in a number of trenches. Possible timber ship parts or remains of a wooden box appeared in the late fill of trench BE at the northern part of the site (Figure 3), in a room whose function could not be ascertained this winter and in which work will continue in the future. The room thus far excavated in trench BE had walls built of white gypsum/anhydrite ashlars, some at least of which comprised recycled blocks from Ptolemaic or Early Roman contexts (Figure 10). The eastern end of this room had an entrance that stepped down into it, while the western end contained a square alcove. The southern part of the alcove at its eastern end still preserved remains of stone revetment decoration (Figures black circles). On either side of the long slab that formed the sill of the alcove were sockets containing remains of wood that may have formed part of some kind of ornamental or structural frame built around the opening. Inside the rectilinear shaped alcove was a timber used to construct a box (cf. Bard and Fattovich 2013: 6 7); it was made of carob wood (Ceratonia siliqua) and preserved at least one large iron nail and several dowel holes (Figures 11 13). Carob is a hardwood derived from a Mediterranean species 8 Journal of Indian Ocean Archaeology No. 12, 2016

9 Fig. 10: Trench BE looking southwest. Circle around marble revetment. Photo by S.E. Sidebotham. Fig. 11: Niche on western side of a structure of unknown function in trench BE Note carob wood beam indicated by black arrow. Circle around marble revetment. View looking southeast. Scale = 50 cm. Photo by S.E. Sidebotham. Journal of Indian Ocean Archaeology No. 12,

10 Fig. 12: Detail of carob wood beam seen in Figure 11. Note one of the dowel holes immediately above the scale. Scale = 5 cm. Photo by S.E. Sidebotham. Fig. 13: Detail of carob wood beam seen in Figure 11. Note the head of a large iron nail. Scale = 5 cm. Photo by S.E. Sidebotham. Fig. 14: Small coarse ware jar near the late entrance into the room in trench BE View looking northeast. Scale = 20 cm. Photo by S.E. Sidebotham. 10 Journal of Indian Ocean Archaeology No. 12, 2016

11 Fig. 15: Portion of a stone temple pool/offering table/libation table from Trench BE Scale = 10 cm. Photo by S.E. Sidebotham. used by humans over the past 4000 years, which can grow to heights ranging from 6 to 15 meters (Gale, et al. 2000: 338). This is the first time our project has documented carob wood used in this capacity. Other remains of teak wood, too fragmentary to determine their original use, but possibly from one or more dismantled ships, lay in the southeastern corner of this trench. These remains may reflect late squatter occupation; excavations also documented other artifacts indicative of a maritime-based economy: a brailing ring and copper-alloy fish hooks. The squatters also reused stone blocks to rearrange the space on the floor of the room and in a corner thus formed, just inside the entrance, they concealed a small and very worn coarse ware jar that lacked a lid (Figure 14). This jar s contents, consisting of a piece of charcoal, two sherds and a chip of stone, may have been a foundation deposit with substitute offerings inside, a concept with a long (e.g. Biblical story of the sacrifice of Isaac) and widespread tradition in the ancient world (including Christianity). Large blocks either thrown or tumbled into the room dominated the interior (Figure 10). Many were so large that they could not be removed, at least with the equipment available this season. One of the more interesting and smaller shaped stones was a portion of a large temple pool/ offering table/libation table, which had been deliberately cut down (Figure 15). Excavations have documented many examples similar to this one elsewhere in Berenike from fourth fifth century CE contexts, some of them also reused (Sidebotham and Zych 2010: 17 18; Sidebotham and Zych 2012b: 145) These libation tables appear in substantial numbers in Nubia and Meroë (areas farther south along the Nile stretching from southern Egypt and into Sudan) throughout much of their history (Kuentz 1981; Cunningham- Bryant 2012; Sidebotham 2014: , 608). Excavations at the entrance to the Great Temple (previously called the Serapis Temple/so-called Serapis Temple) (cf. Sidebotham 2014: ) in trench BE (Figures 16 18) also recorded large quantities of cedar wood and some acacia (Figures 19 20), likely the remains of roofing; much of the former had undoubtedly been recycled from dismantled ships. Also excavated in trench BE were numerous inscriptions in Greek dating from the first to at least the third centuries CE. Two of the monumental texts recorded secretaries. One was a dedication to Isis the Greatest Nurse Goddess in 49 CE (Figures 16A and 21 22). The other noted a secretary, in charge of an aromatics warehouse, who dedicated his inscription together with a statue to a prominent citizen of Berenike in 112/113 CE (Figures 16B and 23 24) (Ast and Bagnall 2015). None of the individuals recorded Journal of Indian Ocean Archaeology No. 12,

12 Fig. 16: Plan of the Great Temple locating trenches BE and 112 and the largest of the inscriptions (A and B inside trench BE15-111). Drawing by M. Hense. Fig. 17: Trenches BE and 112 (portions of the Great Temple). BE is to the left and BE is to the right. View looking southeast. Photo by S.E. Sidebotham. 12 Journal of Indian Ocean Archaeology No. 12, 2016

13 Fig. 18: Trenches BE and 112 (portions of the Great Temple). BE is in the foreground. View looking southwest. Photo by S.E. Sidebotham. Fig. 19: northwestern corner of trench BE Note marble revetment protruding from wall on left and broken cedar wood beams below photo scale. View looking north northwest. Scale = 20 cm. Photo by S.E. Sidebotham. Journal of Indian Ocean Archaeology No. 12,

14 Fig. 20: northeastern corner of trench BE Note underside of inscription fragment (above and left of the scale) shown in Figures (below) and large fragments of cedar wood beams. View looking north northeast. Scale = 20 cm. Photo by S.E. Sidebotham. Fig. 21: Inscription dedicated by a secretary to Isis the Greatest Nurse Goddess in 49 CE (See Figure 16A for placement in temple). View looking west. Scale = 50 cm. Photo by S.E. Sidebotham. 14 Journal of Indian Ocean Archaeology No. 12, 2016

15 Fig. 22: Copy from a squeeze of inscription (in Figure 21) dedicated by a secretary to Isis the Greatest Nurse Goddess in 49 CE (See Figure 16A for placement in temple). Scale = 30 cm. Drawing by R. Geerts. Fig. 23: Inscription (records a statue) dedicated by a secretary in charge of an aromatics warehouse at Berenike to a prominent citizen of the city in 112/113 CE (See Figure 16B for placement in temple). Scale = 50 cm. View looking north. Photo by S.E. Sidebotham. Fig. 24: Copy from a squeeze of inscription (in Figure 23) (records a statue) dedicated by a secretary in charge of an aromatics warehouse at Berenike to a prominent citizen of the city in 112/113 CE (See Figure 16B for placement in temple). Scale = 30 cm. Drawing by R. Geerts. Journal of Indian Ocean Archaeology No. 12,

16 Fig. 25: Two of the many inscription fragments (this pair possibly from the same text) from the Great Temple. Carved in Greek during the joint reigns of Septimius Severus and Caracalla. Late second early third century CE. Scale on left = 10 cm; scale on right = 5 cm. Photos by S.E. Sidebotham. Fig. 27: Stele in Egyptian pharaonic style of Ptolemaic date depicts (from left to right) images of the deities Hathor/Isis, Osiris, Harpocrates standing on a podium and Min/Pan. Scale = 10 cm Photo by S.E. Sidebotham. Fig. 26: One of the many inscription fragments from the Great Temple. Carved in Greek during the reign of Septimius Severus ( CE). Scale = 10 cm. Photo by S.E. Sidebotham. 16 Journal of Indian Ocean Archaeology No. 12, 2016

17 in these two texts is otherwise attested at Berenike. Three other epigraphic fragments were late second to early third century in date (Septimius Severus and Caracalla: reigned CE) (Figures 25-26) (readings of Greek texts by R.S. Bagnall and R. Ast). Another fragment, the top of a stele, was Pharaonic in appearance, but probably of Ptolemaic date. It bore (from left to right) images of the deities Hathor/ Isis, Osiris, Harpocrates standing on a podium and Min/Pan (Figure 27). Of potentially great importance were fragments of several sandstone stelae of Pharaonic date. A portion of the top left hand corner of one bore the cartouche of the Middle Kingdom pharaoh Amenemhat IV (reigned ca /1763 BCE) (Figures 28-29) (initial reading by R. Geerts). Another one (Figures 30 31) recorded year seven of an unknown pharaoh. The latter fragment may be part of the same inscription as that of Amenemhat IV (Hense, et al. 2015). Alternatively, Fig. 28: Upper left portion of an inscription on a sandstone stele in hieroglyphs. The cartouche records the pharaoh Amenemhat IV (reigned ca /1763 BCE). Each black and white increment on the scale = 1 cm. Photo by S.E. Sidebotham. Fig. 29: Copy from a squeeze of a fragment of a limestone inscription (in Figure 28) recording the Middle Kingdom pharaoh Amenemhat IV. Scale = 15 cm. Drawing by R. Geerts. Fig. 30: Lower portion of an inscription on a sandstone stele in hieroglyphs. Perhaps part of a single text along with the fragment mentioning Amenemhat IV or from a different text from the Middle Kingdom or Second Intermediate Period. Scale = 10 cm. Photo by S.E. Sidebotham. Journal of Indian Ocean Archaeology No. 12,

18 Fig. 31: Copy from a squeeze of the lower portion of an inscription (in Figure 30) on a sandstone stele in hieroglyphs. Perhaps part of a single text along with the fragment mentioning Amenemhat IV or from a different text from the Middle Kingdom or Second Intermediate Period. Scale = 15 cm. Drawing by R. Geerts. another scholar dates the fragment lacking an attribution to the late Middle Kingdom/early Second Intermediate Period (mid-eighteenth/midseventeenth century BCE) (personal communication from Z.E. Szafrański, based on examination of photographic documentation) noting close parallels in clothing worn by the human figure in the fragment with texts from the thirteenth dynasty Pharaoh Sobekhotep IV found at the Nile emporium of Edfu. Texts appearing on two cargo boxes and an ostracon documented from Saww (Mersa/Wadi Gawasis: see above) (cf. Bard and Fattovich 2007, 2011, 2013) record an expedition sent by sea from that roadstead to Punt during year 8 of the reign of Amenemhat IV (Bard and Fattovich 2007a: 242; Tallet 2009: 697; Mahfouz 2010; Bard and Fattovich 2011: 110 and note 27; 111 Table 2; 119). The bottom of the stele excavated this season (in Figures 30 31) records year seven of a pharaoh whose name is lost. Yet, it may be part of the same text preserving the cartouche of Amenemhat IV. That excavations documented the stone with the cartouche from the southern balk towards its western end and the fragment indicating year 7 some distance away in the northeastern corner of the trench may suggest, however, that the two belong to different texts. These two fragments may point to activity at Berenike some 1,500 years earlier than previously believed. We plan future excavations in this area with the hope of documenting additional information from the pharaonic era. Excavations in trench BE also documented portions of carved and painted wall decorations on the frame of the door connecting what appeared to be an open courtyard (in trench BE15-111) with an interior room (in trench BE15-112) of the Great Temple (See Figure 16). These door frame decorations included incised and painted lotus and papyrus plants (Figure 32) on the southern side of the door; Wilkinson had noted similar decorations inside the temple on his visit in 1826 (Meredith 1957: & Fig. 4). There was a standing figure of a female deity (Figure 33) on the northern side of the same portal. Huge fragments of thick marble revetment, both light gray/white and light green in color, also littered trench BE15-111; these would have decorated either walls or floors, or both. Excavations also recorded two fragments of white marble statuary (Figure 34) from what originally would have been a large simulacrum and a fragment 18 Journal of Indian Ocean Archaeology No. 12, 2016

19 of an ostrich eggshell painted with an elaborate pattern of rosettes (Figure 35). The eggshell falls within a tradition of decorated ostrich eggs occurring in cult contexts in other shrines dating from fourth fifth century CE Berenike (Rądkowska et al. 2013: 219 and Fig. 10; Sidebotham 2000: 140 and Fig. 2-96; cf. Wilburn 2015). It remains to be determined how the Pharaonic-era stelae fragments came to rest amidst the tumble of fragmentary and damaged pieces of Roman statuary, inscriptions and architectural elements scattered in front of the entrance to the presumably already destroyed Great Fig. 32: Lotus and papyrus decoration of the southern side of an entrance into the Great Temple (i.e. from trench BE into trench BE15-112). View looking west. Scale = 20 cm. Photo by S.E. Sidebotham. Fig. 34: Fragments of white marble statuary from trench BE Scale = 5cm. Photo by S.E. Sidebotham. Fig. 33: Standing figure (indistinct) likely of a goddess in Egyptian style. Decoration on the northern side of an entrance into the Great Temple (i.e. from trench BE into trench BE15-112). View looking west. Each black & white increment on the scale = 10 cm. Photo by S.E. Sidebotham. Journal of Indian Ocean Archaeology No. 12,

20 decorated with incised stars (Figure 36). These had likely been painted, but any pigment has since disappeared; earlier European visitors, who had partly cleared the temple, had also noted ceiling decoration in the form of stars (Meredith 1957: 62). In addition, excavations in BE recorded three joining fragments of a small stone inscription dated to the reign of the Roman emperor Septimius Severus ( CE) (Figure 26). Finally, one should mention a number of cowry shells found on the door threshold from the late period. The shells must have once been strung and hung in the apparently blocked doorway. A very similar situation Fig. 35: Fragment of a painted ostrich eggshell from Trench BE Scale = 5 cm. Photo by S.E. Sidebotham. Temple. They may be evidence of a general leveling of the town site around the temple in the middle of the fourth century CE during the revival of the emporium, furthering already collected data to this end from other trenches explored by the project in the central and eastern parts of the site. Further excavations in and around the temple should help to answer this question. Huge collapsed architectural blocks filled much of BE (Figures 17-18), part of an interior room of the Great Temple. Joined to the putative outer courtyard excavated in trench BE by a doorway partially blocked in antiquity (and on the eastern face of which were decorations noted above: see Figures 32-33), excavations in trench BE also documented portions of the structure s ceiling Fig. 36: Portion of ceiling decorated with incised stars from the Great Temple from trench BE Scale = 10 cm. Photo by S.E. Sidebotham. had been observed in the entrance to the Late Roman Harbor Temple/Lotus Temple in the southwestern harbor (Rądkowska et al. 2013: and Fig. 14). These shells may have served as apotropaic talismans or for prognostication purposes (Cf. Rądkowska, et al. 2013: ; Sidebotham 2014: 607). Clearly, the elaborate decorative features, rich dedications and prominent location of the Great 20 Journal of Indian Ocean Archaeology No. 12, 2016

21 Temple made it one of the premier public buildings in Berenike at least in the Early and Middle Roman periods (first early third century CE) and perhaps earlier. That two large dedications were made to Isis strongly suggests that she was also worshipped in this temple in addition to one or more other deities. Excavations in the Great Temple have yet to establish an initial date of construction. Nevertheless, it, the sunken feature excavated in trench BE (noted above) and remains in adjacent trenches are probably contemporary and functionally related. It is possible that they were built during the reigns of Ptolemy VI VIII ( BCE), eras of extensive temple construction elsewhere in Egypt [cf. Hölbl 2000: 18 20, 26 (Pls ), 27 (Pl. 36); Hölbl 2001: ]. The discovery of an inscription of Ptolemy VIII Physcon (Euergetes II) (reigned BCE) inside the Great Temple in the nineteenth century (Meredith 1957: 69 misidentified the text as that of Ptolemy VII) may provide a more precise date for its foundation and for the construction of the sunken feature. Strabo (Geography 2.3.4) drawing on Poseidonius reported towards the end of his reign that Ptolemy VIII sponsored Eudoxus of Cyzicus to explore the Red Sea and Indian Ocean (cf. Habicht 2013). So the discovery of an inscription of that monarch at Berenike is not surprising, but corroborates what literary sources recorded about official interest in those regions at that time. In the Ptolemaic area at the western side of the site (primarily dating to the third first centuries BCE) excavations continued in the early military defenses of the city (trenches BE14-97/BE15-104) (Figure 3). During his visit in 1826 J.G. Wilkinson had suggested the possibility of urban defenses (Meredith 1957: 57). These remains at Berenike (together with those found in trenches BE11-83/85/86: a robbed tower, and BE13-90/93: a robbed portion of city wall) (cf. Sidebotham and Zych 2012a: 31 32) are the only archaeologically attested Ptolemaic-era city defenses known anywhere in Egypt outside of Alexandria. A large rectangular shaped shaft cut into bedrock immediately behind (south of) a section of these early city walls in BE14-97/BE15-104, along with a heavy stone counterweight with a socket, likely for a wooden beam, suggested the possible existence of a drawbridge here. Alternatively, this large pierced stone, very similar in appearance to an anchor, may have been used as a counterweight in some type of shadoof system to raise or lower items (including, possibly, containers of water) from the shaft. Excavations in the shaft descended about 2.5 meters where five tunnels appeared (Figures 37 39). One of these, stretching 6 7 meters towards the east before turning northeast, was evidently hydraulic in nature. The total length of this tunnel could not be traced this season as a result of its blockage by sand towards its eastern end and due to a lack of time. In some ways this tunnel resembles in appearance, if Fig. 37: Overall view of trenches BE14-97/BE Black arrows indicate early Roman graves (described below in the text) and shaft containing five tunnels from the early Ptolemaic period. View looking southeast. Photo by S.E. Sidebotham. Journal of Indian Ocean Archaeology No. 12,

22 Fig. 39: Trenches BE14-97/BE shaft cut into bedrock with five tunnels. Tunnel immediately to the right of the Bedouin is hydraulic. Note foot/hand holds above the hydraulic tunnel and on the (southern) wall to the right. View looking east. Photo by S.E. Sidebotham. Fig. 38: Trenches BE14-97/BE shaft cut into bedrock with five tunnels. Tunnel at top right is hydraulic. Note foot/hand holds above the hydraulic tunnel and on the (southern) wall to the right. View looking east. Photo by S.E. Sidebotham. not in function or length, qanats/foggaras found in southern Libya, Egypt s Western Desert, elsewhere in the Middle East and in Central Asia (cf. English 1968; Wilson 2008: ; Grewe 2008: ). Four other tunnels were also incompletely excavated in this shaft on its northern and southern sides. So, their dimensions and functions could not be determined due to lack of time; additional fieldwork here in the future should allow a better understanding of all of these tunnels. There were handholds/footholds for climbing up to the surface cut into the rock walls in the southeastern corner of the shaft. However, the entire complex the vertical rectangular shaft with tunnels radiating from it resembles an early Hellenistic one in (Berenike) Euesperides (Benghazi), in Libya; similar installations also appear in late Roman Republican times and during the early Roman Principate (Wilson 2008: 288). In this area of the Ptolemaic city defenses and hydraulic facilities (trenches BE14-97/BE15-104) excavations also recorded three human burials: two male and one female. All had grave goods, which allowed dating the interments to the first century CE (Figures 40 44). The graves did not follow any single orientation, apparently making Fig. 40: Skeleton of an adult male wearing an iron ring (indicated by black arrow). First century CE. Scale = 50 cm. Photo by S.E. Sidebotham. 22 Journal of Indian Ocean Archaeology No. 12, 2016

23 Fig. 41: Detail of hand of skeleton in Figure 40 showing iron ring. Scale = 5 cm. Photo by S.E. Sidebotham. Fig. 42: Skeleton of an old man with grave goods in the form of a string(s) of beads (some of the beads were imported from India) and iron ring threaded on the string as well (indicated by black arrows). Dark coloring of the bones suggests a disintegrated organic substance, possibly a shroud or garment. First century CE. Scale = 50 cm. Photo by S.E. Sidebotham. Fig. 43: Detail of skeleton in Figure 42 showing beads (indicated by black arrow). Scale = 10 cm. Photo by S.E. Sidebotham Journal of Indian Ocean Archaeology No. 12,

24 Fig. 44: Skeleton of tall adult female with part of an amphora covering her skull and a regular truncated biconical stone on her abdomen. First century CE. Scale = 50 cm. Photo by S.E. Sidebotham. use of any softer ground that those interring the dead could find. The presence of these burials indicates that the earlier Ptolemaic hydraulic facilities had ceased to operate in that capacity by the early Roman occupation of the city. Excavations in an early Roman era (first century CE) animal cemetery continued from previous seasons and included scores of burials. Most of the remains were cats and kittens, but there were also a few dogs and puppies, a baboon and some vervet or grivet monkeys. A few still wore collars made of iron (Figure 45). One of these appears to have been decorated with beads (Figure 46). Many of the bodies had been either covered with broken pieces of pottery or placed inside jars and amphorae. Finally, following the excavation season, study of Corona satellite imagery taken in the late 1960s/early 1970s revealed a large architectural complex northwest of the known northern limits of the city and west of the late Roman necropolis, which had been partially excavated in 2001 (Sidebotham and Wendrich : 35, 36, Fig. 24). These structures are not visible on Google Earth imagery. White gypsum/anhydrite ashlar blocks formed the foundations of a long, narrow building with three internal rectilinear shaped podia (trench BE and extensions) in this area (Figures 3 and 47). Unfortunately, excavations revealed nothing of the structure s function since it had been almost completely cleared of artifacts either by the Fig. 45: Skeleton of a cat. Note iron collar (indicated by the black arrow). First century CE. Scale = 5 cm. Photo by S.E. Sidebotham. 24 Journal of Indian Ocean Archaeology No. 12, 2016

25 Fig. 46: Iron collar decorated with beads from the first century CE animal cemetery. Scale = 5 cm. Photo by S.E. Sidebotham. latest users or perhaps by periodic flooding of the adjacent wadi. Analysis of the paltry quantity of finds mostly sherds did, however, indicate that its latest use was during the Augustan era (30 BCE CE 14). A distribution study of the finds in particular contexts also documented fragmented, but joining pottery vessels (two cooking pots and a fish plate), as well as concentrations of burnt sheep bones and ash, an offering of a giant clam and burnt acacia wood deposited in specific places inside the structure. There was also a burial of a dog on one of the westernmost cross-walls of this edifice. Survey in the environs of Berenike A brief survey of the area of modern military bunkers immediately northwest of Berenike identified a number of ancient graves. These were of the typical Fig. 47: Foundations of a building of unknown function found northwest and outside of the city limits/west of the late Roman-era necropolis partially excavated in Dated to the Augustan era (30 BCE-CE 14). View looking northeast. Photo by S.E. Sidebotham. ring cairn variety that is ubiquitous throughout the Eastern Desert. Most had been looted. The surface sherds and other small finds indicated burials ranging from the Early to Middle Roman periods (first third centuries CE). One of the finds was a tiny intaglio (length 0.75 cm x width 0.63 cm x thickness 0.20 cm) carved with a human or eros figure milking a goat (Figures 48 49). There is a very close parallel for this specimen from the Roman Mediterranean port of Caesarea Maritima, Israel (Hamburger 1968: 22 and Pl. VIII, no. 149: J. Rądkowska kindly provided this citation). The Berenike specimen came from near a looted grave with pottery of Early Roman date. Southwest of Berenike (at N/ E) the survey documented a narrow pass (Figure 50) on the route leading from the large and Journal of Indian Ocean Archaeology No. 12,

26 Figs : Early Roman era intaglio and clay impression of a human or eros figure milking a goat. Photos by S.E. Sidebotham. Fig. 50: Pass on the route linking the two praesidia in Wadi Kalalat with Berenike. View looking west. Photo by S.E. Sidebotham. small forts (praesidia) in Wadi Kalalat that was littered with Early and Late Roman era sherds, including those made in Egypt as well as imported from elsewhere in the Mediterranean. Clearly, this pass lay along one route linking these nearby installations with Berenike. Both forts in Wadi Kalalat would have provided protection for the emporium and also probably supplied some of the potable water consumed in the city. Fieldwork in other areas of the Eastern Desert Under a separate grant from the Polish National Science Center, Piotr and Marta Osypiński continued excavation of a necropolis in Wadi Khashab (at N/ E) that they began in 2012 (Sidebotham and Zych 2012: 41& 42, Figs ) (Figures 1 and 51 53). The site lies at the confluence of two great wadis in the Eastern Desert, one leading to the Red Sea and the other to the Nile Valley debouching near Kom Ombo. The installation in Wadi Khashab was likely a Late Neolithic (fifth millennium BCE and later) ceremonial center corresponding chronologically and functionally with Nabta Playa and other sites in the deserts farther south on either side of the Nile (Brass, et al. 2003, 26 Journal of Indian Ocean Archaeology No. 12, 2016

27 Fig. 51: Long-horned cattle cemetery in Wadi Khashab (fifth millennium BCE and later) prior to excavation (in 2010). Photo by S.E. Sidebotham. Brass 2007; cf. Murray 1926). This was a period and location of early pastoral societies at the dawn of civilization in the Nile Valley at about that time (the date should be confirmed by C 14 dating now being processed in the IFAO laboratory in Cairo). Inside a wall of stone megaliths ca. 19 m N S x 21 m E W x maximum height of 1.40 m was a human grave, plundered in antiquity, leaving some bones and remnants of the ornaments that once adorned the body. Arranged around it, as if a herd following its human caretaker, were graves of cattle and sheep. These included two bulls lying on their right sides and two cows, one of them pregnant, lying on their left sides. The heads of the cattle were positioned in a very specific way so that the long horns characteristic of this breed would have protruded out of the ground and would have been visible to passersby. The sheep were buried tightly curled with heads twisted onto their backs. There was evidence, in at least one case, of ceremonial butchering; it appeared that the entrails of that animal had been placed on a piece of wood. In the half of the enclosure that had been excavated (Figures 52 53), there was also grave of a human infant; the body had been furnished with a bead bracelet and stringed shells around the neck. The cattle bones, examined by archaeozoologist M. Osypińska, have already provided important data for understanding evolutionary processes taking place in cattle in Africa in the Neolithic period and the animal population in this particular region. The cattle and sheep buried in Wadi Khashab had limbs much longer than their modern counterparts. This indicates an evolutionary change in a population specifically bred to move over long distances. Moreover, the cattle proved to be hump-backed, laying to rest the long discussion about the origin of humped cattle in Africa. Until this discovery, one theory had suggested that this physical feature appeared in Africa from India no earlier than the Ptolemaic period (starting in the late fourth century BCE) (MacDonald 2000; Osypińska 2012: 710). Evidence from Wadi Khashab this season indicates that this was an indigenous characteristic that evolved independently in Africa millennia earlier. Journal of Indian Ocean Archaeology No. 12,

28 Fig. 52: Long-horned cattle cemetery in Wadi Khashab during excavations in Scale = 2 m. Photo by P. Osypiński. Fig. 53: Long-horned cattle cemetery in Wadi Khashab (detail of Figure 52). Scale = 2 m. Photo by P. Osypiński. Nomads with their herds seem to have visited this ceremonial center regularly. Likely they were moving up and down the wadis, the important direction being probably toward the Nile as this is the orientation of the buried animals. Someone plundered the site at the beginning of the third millennium BCE, about the time of the emergence of the Egyptian state in the Nile Valley. That only the main human interment and two of the cattle graves were plundered (in that part of the site that was excavated) may indicate that the looters did not find anything they considered valuable; our excavations documented the fragments of a single ceramic bowl that suggests the date for this event. Continued survey in the Eastern Desert resulted in the discovery of a large site hitherto unknown to western scholars. First reported by Karen van Opstal, a Dutch resident of the Red Sea coast, to Sidebotham 28 Journal of Indian Ocean Archaeology No. 12, 2016

29 Fig. 54: View of enigmatic late Roman-era settlement in Wadi Angoriyya. Photo by S.E. Sidebotham. Fig. 55: View of enigmatic late Roman-era settlement in Wadi Angoriyya. Note figure circled in black for scale. Photo by S.E. Sidebotham. in fall 2014, the Delaware Warsaw survey visited these ruins twice this season in December 2014 and in January This large settlement in Wadi Angoriyya (at N/ E) comprises about 100 buildings (Figures 1 and 54 55). Dating to the Late Roman period based on surface pottery (fourth-sixth centuries CE) the community likely belonged to Christians escaping problems along the Nile at that time. Our survey and those of others conducted in the Eastern Desert have recorded about a dozen settlements similar in appearance and date (Sidebotham 2011: , 276). Somewhat northwest of the settlement in Wadi Angoriyya the survey recorded petroglyphs of three large long horned cattle and several smaller ones as well as dalu (signs indicating the presence of a well) (at N/ E) (Figure 56) (cf. Fuchs 1989: , , 137; Redford and Redford 1989: 9, fig. 1; 11, fig. 6; 26, fig. 40; 30 31, fig. 51; 33, fig. 58; 34 35, fig. 61; Rohl 2000: 29, 35, 43, 73, 75, 82 83, 89, 91, 95, 115, 118, 121, 146). The relative proximity (ca. 35 km in a straight line) of the longhorned cattle cemetery in Wadi Khashab to these petroglyphs may indicate a functional relationship Journal of Indian Ocean Archaeology No. 12,

30 Fig. 56: Petroglyphs of long-horned cattle and dalu (signs indicating the presence of a well). Scale = 50 cm. Photo by S.E. Sidebotham. between the two. If so, then these images only a few kilometers northwest of the settlement in Wadi Angoriyya should like the cattle cemetery itself also be dated to sometime in the fifth millennium BCE or later. The parallels noted above for this rock art are, however, undatable or have been attributed from the pre-dynastic to the pharaonic period. The project also continued to draw a detailed Fig. 57: One of the well-preserved Roman-era buildings sitting on a platform at the emerald mining settlement at Nugrus. View looking northwest. Photo by S.E. Sidebotham. 30 Journal of Indian Ocean Archaeology No. 12, 2016

31 Fig. 58: Interior of one of the Roman-era structures at the emerald mining settlement at Nugrus. View looking north northwest. Photo by S.E. Sidebotham. Fig. 59: Interior of one of the Roman-era structures at the emerald mining settlement at Nugrus. View looking northwest. Photo by S.E. Sidebotham. plan of the Roman-era beryl/emerald mining settlement at Nugrus ( N/ E) (Figures 1 and 57 59) using a total station (cf. Sidebotham, et al. 2004: 23 26). This survey, conducted and co-directed by J.-L. Rivard, is part of another longer term endeavor that aims to study and document in detail about nine known ancient communities in a region the Romans called Mons Smaragdus (Emerald Mountain), which comprises approximately 300 km 2 of the desert northwest of Berenike. This intensively mined region was the only known source of beryls/emeralds anywhere inside the Roman Empire and was exploited from before the Roman period and into Islamic times (Sidebotham, et al. 2004; Foster, et al. 2007). The project plans to publish a separate volume on the results of this long-term project. Another site the survey team visited was the World War II RAF airfield at Hawarry not far from the coast west southwest of the modern Red Journal of Indian Ocean Archaeology No. 12,