Excavation on the "High Mound" Excavation in the Lower Town The Tombs Regional Surveys and Excavations Remote Sensing

Field School

2008 Season 2009 Season

Regional Surveys and Excavations

THE EXPANSION OF SETLLEMENT INTO THE BALIKH-EUPHRATES UPLANDS

by Michael D. Danti (2000)

In addition to excavating relevant early-EBA sites, our regional project expanded Tony Wilkinson’s (University of Chicago) survey coverage to the surrounding steppe uplands to gather data on EBA pastoral production (Zettler 1997a). The steppe uplands were historically important to herders as a source of seasonal pasture; thus, EBA sites in this region might further document and help to elucidate this landuse pattern.

Over 30 years ago, M. van Loon carried out the first reconnaissance along the Syrian upper Euphrates in connection with the Tabqa Dam Archaeological Salvage Project (van Loon 1967, Freedman 1979). T. J. Wilkinson continued this work with a systematic archaeological survey and geomorphological study of the Tell es-Sweyhat embayment in the early 1970’s and continued in 1991 and 1992 (Wilkinson n.d a, n.d b, 1994). He is currently preparing a final report. After a brief review of Wilkinson's results, I present the findings of our initial research in the Balikh-Euphrates uplands.

Early Bronze Age Settlement in the Tell es-Sweyhat Region

The period immediately preceding the EBA, the late Chalcolithic, is marked by a scarcity of archaeological sites in the Tell es-Sweyhat embayment (Wilkinson n.d b: 3). However, surrounding areas were the focus of colonization by southern Mesopotamian polities. Across the river from Sweyhat lies the Uruk site of Jebel Aruda (van Driel and van Driel-Murray 1979, 1983). Not far down-river from Sweyhat, and also on the right bank, is the site of Habuba Kabira South/Tell Qanas — a southern Mesopotamian Uruk settlement exported to northern Syria (Strommenger 1979, 1980). Another fully Uruk site, Sheikh Hassan, lies on the left bank downstream of the Sweyhat area (Boese 1995). The general pattern of Uruk sites suggests an interest in controlling the north-south routes along the right bank of the river and possibly strategic points along the left bank at fords. Conversely, the Sweyhat embayment appears to have been largely outside the range of Uruk influence.

In the Sweyhat area, late Chalcolithic occupation is limited to sparse settlement of the fringe of the floodplain and possibly two small, low tells on the Tell es-Sweyhat plain (Sites 13 and 21). This pattern of settlement is largely in keeping with earlier periods, although the shift to settlement on the plain marks an important, if not substantial, departure. The fringes of the floodplain had been sparsely settled since the late Halaf (Radi and Seeden 1980, Seeden 1982, Gustavson-Gaube 1981, Wilkinson n.d b).

Settlement in the embayment increased markedly in the early EBA. In his survey, Wilkinson found signs of occupation at sites 2, 3, 9, 19 and 21. Researchers have excavated contemporary levels at Sweyhat (Site 1), Hadidi and Tell Abd (Site 22). Excavation of Site 3, Tell Hajji Ibrahim, has demonstrated that this process started in the late 4th millennium BC, contemporary to the terminal phases of Uruk influence (Chapters 4 and 5). Surface collections at Sites 2 and 9, similar to Hajji Ibrahim in terms of mound morphology and location, indicate occupation was contemporary to Hajji Ibrahim. Our re-survey of Wilkinson’s Site 5 (Tell Nefileh) in 1995 revealed the presence of “pinched rim” jars, evidence of occupation contemporary to Hajji Ibrahim Phases B/C (early 3rd millennium BC).

Most of the small early-EBA sites (sites 2, 3, 9) were abandoned as Sweyhat started to expand in the mid 3rd millennium BC (Wilkinson 1994: 489). Sweyhat gradually expanded from a maximum of 6 ha (the main mound) to encompass an area of around 15 ha, including parts of the outer town (Zettler 1997: 48-51). Occupation continued at Site 5, and new sites such as Tell Othman (Site 20) and Sites 8 and 17 were founded (Wilkinson n.d b: 6). Across the river, Tell Hadidi (ancient Azu) also grew to around 50 ha by the late EBA (Dornemann 1979, 1988).

As noted, the late EBA marks the peak in settlement of the Tell es-Sweyhat plain, a pattern unmatched until the modern era. At Sweyhat, the outer town was extensively settled, the outer town fortifications were constructed and settlement continued to expand beyond the southern wall. The area of the original settlement (the main mound) became a fortified citadel (Zettler 1997: 4). The settlement was probably between 35 and 45 ha. Other sites on the Tell es-Sweyhat plain continued to be occupied and new settlements were located on or near the flood plain (Sites 8, 24, 27).

Wilkinson found little evidence of occupation on the floodplain itself except for Tell Juweif (Site 8), a 2-3 ha site of the mid to late 3rd through the first half of the 2nd millennia BC. In terms of settlement patterns, the lack of EBA sites is inconclusive since the river has reworked nearly the entire floodplain since the Roman period (Wilkinson n.d a: 13). Only a small relict section of ancient floodplain is preserved south of Tell Juweif. This settlement probably sat atop a floodplain terrace of the type reported on by Wilkinson for the floodplain site of Dibsi Faraj, located on the right bank 40 km downstream of Tell es-Sweyhat. These terraces consist of “low spreads of silt about 5-7m above the lowest level of the stream . . . subject to only sporadic floods” (1978: 216). The lack of settlement on the relict floodplain and on Tell Juweif in later urbanized periods (the Roman, Byzantine and Abbasid periods) raises the issue as to whether settlement on the floodplain was feasible or preferable to settlement on the floodplain’s fringe. Perhaps few floodplain terraces were available for settlement and the floodplain itself was too valuable for settlement since irrigated summer gardens could be located there.

Tell es-Sweyhat collapsed in the early 2nd millennium BC. The main mound was occupied during the transition from the EBA to the Middle Bronze Age. By 1800 BC, the site was completely abandoned.

As noted in Chapter 2, Wilkinson concluded that Sweyhat’s dry-farming economy relative to the city’s late-EBA population was highly susceptible to deficits in dry years (Wilkinson 1994: 498, 501). This brittle subsistence base begged the question of Sweyhat’s existence and raised the issue of how state-level institutions and their activities were supported. Compared to other agrarian states of the Fertile Crescent, Sweyhat the fortified city, situated amid an unirrigable, arid plain 4 km from the Euphrates and off any major transportation route, presents a conspicuous anomaly on the EBA landscape. This pattern of urban development was unparalleled in earlier and later times in this region.

The Tell es-Sweyhat Project has learned a great deal more concerning EBA subsistence practices from excavations at Sweyhat and Hajji Ibrahim since Wilkinson’s pioneering research. First, the hunting of wild riverine and upland steppe animals contributed substantially to the diet of the mid-to-late 3rd millennium inhabitants of Tell es-Sweyhat (Weber 1997: 141-142). Second, faunal remains from Tell es-Sweyhat and Tell Hajji Ibrahim attest to the importance of herding small ruminants, as well as cattle and pigs (Ibid.). Botanical remains strongly suggest the pasturing of animals in the surrounding steppe uplands (Miller 1997: 102-103). They also support the premise that animals were fodder-fed during part of the year, based on the occurrence of barley and charred straw and chaff in the charred remains of dung fuels. Thus, any complete model of EBA subsistence practices and their development must include pastoral production and the exploitation of resources beyond the narrow confines of the embayment.

Since a large portion of the Balikh-Euphrates had not been surveyed, we initiated a regional survey in 1993, focusing on EBA settlement and landuse.

Objectives of the Archaeological Reconnaissance

The survey's broadest objective involved the documentation of the cultural landscape of the steppe uplands throughout history with particular emphasis on the EBA. Another important goal was to seek the limits of the Sweyhat region during the late EBA, and thereby more accurately model the regional economy. Scholars have frequently cited the importance of delimiting regional boundaries (Johnson 1977, Evans and Gould 1982: 279). Wilkinson correlated Sweyhat’s agricultural territory with the embayment, describing the area as "restricted by topography", that is, the high plateau (Ibid., 489, 491). His interest in agricultural systems made survey of the high plateau and uplands unnecessary, but as already noted the plateau did not circumscribe all subsistence activities.

Another objective was to examine water sources in the uplands. The lower hilly uplands skirting the high plateau had the potential for runoff-exploitation, so we sought to determine whether water harvesting or runoff agriculture were practiced in antiquity (Eger 1987). Today, embayment farmers highly regard dry-farmed fields in this part of the jebel, which produced a better crop in 1993 than dry-farmed embayment fields. This may partly be the result of localized fluctuations in rainfall visible on distribution maps (Figure 6.1a) (cf. de Brichambaut and Wallén 1963: 47). Additionally, one explanation for the differential yields of 1993 may be that certain areas of the uplands adjacent to the high plateau receive large amounts of runoff and water-born silts brought down from the high plateau by dendritic wadi systems. Settlements and fields located along major wadi channels can utilize the concentrated runoff of a large catchment, reducing the effects of localized rainfall fluctuations (Kennedy 1995: 277). As shown on the SPOT satellite image (Figure 6.2), the wadis channeling runoff to the Sweyhat embayment originate only a short distance to the east in the high plateau. Most uplands runoff flows away from the embayment to the east and southeast for 40 km, these wadis joining to form the wadi el-Fayyed, which turns abruptly south to join the Euphrates opposite Hammam (Figure 6.1). Fifteen km east of the Tell es-Sweyhat embayment, these wadis begin to level out after descending the high plateau. At this point, runoff flow slows. This area is clearly visible on satellite imagery as a patchwork of dry-farmed and pump-irrigated fields (Figure 6.2). Prior to sedenterization, this region was important for its wells and was one of the first areas resettled in the early 20th century, indicating a high potential for agropastoral economies. If a similar land-use pattern existed in the EBA, it might radically alter estimates of the agropastoral productivity of sites along the Euphrates and Balikh.

Methods

Preliminary research in 1994 and 1995 involving the analysis of maps and satellite imagery supported the theory that the uplands were important to agropastoralists. French maps of the area record the major landforms, a few tells and water sources used by 19th-century, and earlier, nonsedentary groups (Insitut Geographique National 1941 and 1951, hereafter IGN). They also plot some of the villages established in the 19th and early 20th centuries under resettlement programs of the Ottoman and French Mandate governments (Charles 1942: 43-44, 64; Lewis 1987: 27-37, 154-65). Overlaying these maps with panchromatic SPOT (Satellite Pour l'Observation de la Terre) imagery (10m2/pixel resolution) in a geographic information system revealed a high degree of locational correlation between tells, premodern wells and 18th and early-19th century villages. These villages predate the widespread introduction of cash crops, pump irrigation, mechanized agriculture and government-subsidized agricultural development (Al Ashram 1990: 178-80). The similarity between settlement location in the 19th and early 20th centuries, reflective of pre-modern agropastoral strategies, and antiquity implied long-term continuities in upland landuse and strengthened the validity of analogies drawn from ethnohistoric, ethnographic and governmental studies.

I defined the project's study universe as the area of unsurveyed upland steppe between the Balikh and Euphrates (Figure 6.1). Berthold Einwag reconnoitered the northern half of the Balikh –Euphrates uplands in 1993 (Einwag 1993).

I divided the survey area into a grid system of five-km quadrats aligned along latitudinal and longitudinal divisions. I grouped survey quadrats into three strata (zones) based on environmental distinctions for the purpose of performing a stratified random sample (Dunnell and Dancey 1983: 276) (Figure 6.1). I used these distinctions to balance field coverage between regions with different hypothetical land-use potentials. The high plateau constitutes Zone 1 and would have provided seasonal rangeland. Zone 2 takes in the area of potential runoff utilization running parallel to the high plateau, and so has a higher potential for permanent settlement and cultivation. Zone 3 consisted of the lower hilly uplands where extensive rangeland and some dry farming might have been practiced in antiquity.

Prior to fieldwork, I defined potential sites, modern land-use patterns and other regions of interest (ROI's) using IGN maps, SPOT imagery, and information from local informants. I digitized the IGN map of the study area and entered it into a GIS program as a complementary layer to the SPOT imagery. I georeferenced this database using benchmarks provided by SPOT and GPS readings recorded in 1994 and 1995. This digital map enabled us to determine the location of ROI’s to within ca. 100 m prior to fieldwork. I planned to conduct fieldwork in every sampled quadrat regardless of whether it contained ROI's in order to gather data on conditions where sites were absent (Cherry 1983: 387). Furthermore, I knew that many sites were too ephemeral to appear on the satellite imagery or modern maps.

To estimate coverage within each quadrat, I recorded GPS points at fixed intervals during fieldwork and plotted them on the survey map. Fieldwork involved a combination of walking and working from automobiles. Three target areas, one quadrat per stratum, were chosen randomly for intensive walking survey to assess the effects of survey done solely from automobiles (cf. Stein and Wattenmaker 1990: 12).

Total survey coverage in all the sampled quadrats proved unfeasible in the time available. I fell far short of covering all the quadrats chosen for sample, but those the team did survey were covered intensively. The survey team completed intensive walking survey in only the Zone 1 and 2 quadrats. In the high plateau quadrat, the survey team found many ephemeral un-mounded sites such as sheepfolds and rock shelters constructed for shepherds, as well as many cairn burials. Thus, I decided to intensively survey three additional quadrats adjacent to the Sweyhat embayment.

The Regional Reconnaissance (1996, 1997)

In 1996 and 1997, the project conducted a preliminary archaeological reconnaissance of the uplands. A larger survey was planned, but unforeseen developments have delayed this work. These two short seasons doubled as a feasibility study, providing valuable logistical information and methodological tests for a future large-scale survey.

I was able to test initial hypotheses on upland settlement and land use. The reconnaissance demonstrated 1) that ancient pastoral sites were recoverable; 2) that the uplands were settled during the EBA; 3) that the uplands are currently used intensively for herding and dry farming; and 4) that recent, and possibly ancient, water harvesting facilities are present.

Once located, I recorded the exact location of ROI's with the GPS and then took photographs. The survey team then carried out surface collections. Due to the low density of surface artifacts and the small size of upland sites, the survey team achieved 100% surface coverage in all cases. Site size was estimated by pacing mounded areas and/or distributions of sherd scatters. After each day of fieldwork, I downloaded all computer data to the main computer system at the field lab. I recorded all artifacts according to the system employed for the Tell es-Sweyhat excavations (see Chapter 4 and 5). I entered this information into the Sweyhat project's relational database. I had originally planned to plot densities of surface artifacts to identify potential areas of specialized production, to delimit separate chronological phases at multicomponent sites and to measure changes in site size over time (Whallon 1979, 1980). This proved difficult due to time constraints and the presence of modern villages on or near most of the sites. Moreover, the low artifact densities did not merit such an expenditure of time given the limited results. I also intended to conduct off site survey to gather data on agricultural territories. This also proved impracticable since in our initial trials we found no off-site sherd scatters surrounding EBA upland sites.

In all, we recorded 28 sites dating from the early 3rd millennium to the 19th and early 20th centuries AD (Figure 6.1). The majority of sites fell within either the EBA or the Roman to early Islamic periods, paralleling the peak periods of settlement in the Sweyhat embayment. We discovered two pastoral emplacements (sheepfolds, corrals) of unknown date (Sites 5 and 19). We also found five pastoral camps of similar date with stone outlines from semi-permanent or seasonally inhabited structures, outbuildings and animal pens (Sites 1-3, 8). We recorded eight individual tombs and tomb groups of the tumuli type similar to those at Tell el-'Abd (Bounni 1979: 57-61), Mumbaqat (Orthmann and Kuhne 1974: 66-67) and Tell Juweif (Heinrich, et al. 1969: 33) (Sites 4, 6, 7, 9-11). We found 15 tells, ranging from estimates of 0.10 to 15+ ha (Sites 13, 14, 15b, c-18, 21-28). Four sites had wells, one of which, Tell ej-Jaber (Site 12) is a large well with the ruins of a substantial stone-block well house or watchtower (possibly contemporary to Qalaat Jabr AD 1128 - 1516). The team also surveyed an isolated watchtower or tower tomb (Site 20) atop the jebel overlooking Tell Jerniyeh, visited by Gertrude Bell in 1909 (Bell 1910: 517). We discovered a potential area of runoff agriculture/water harvesting near the modern village of Bir Abu Chaaf. A wadi had been diverted into a low-lying region to form a small lake. The small mounds of Mjeibna (Site 16) and the village of the same name are situated nearby. Three other tells found during survey (Sites 13-15) lay in this region, one of which, Tell Jedi (Site 14), dates to the EBA. All three of these tells were located on wadi branches, and two, Bir Abu Jedi (Site 12) and Bir Mjeibna (Site 15c), had derelict camel wells.

The 3rd-Millennium BC Upland Sites

The following section provides a more detailed account of the four EBA sites we found in the uplands, ordered by inferred occupation periods.

Site 17 / Khirbet Taha (Early EBA)

This site is situated in the lower rolling uplands south of the Tell es-Sweyhat embayment near the village of Khirbet Taha (Figure 6.1). The tell actually lies in the small village of Beer Dahar. Both villages are inhabited by sedentarized “Weldeh of the Jezireh” of the Bu Jabr subgroup (Oppenheim 1939: 212) and are currently recognized simply as Bu Jabr. The mound is 4 m high and, based on the sherd scatter, may have a small lower tell surrounding it. I estimate total site area at 3 ha by pacing. The mound is covered in modern burials marked by stone mounds. We found evidence of an ephemeral late occupation, although the majority of the ceramic diagnostics date to the early 3rd millennium BC. This was the only upland site the survey team found with unequivocal early-EBA occupation.

Surface sherds from Khirbet Taha indicate a date of occupation contemporary to Tell es-Sweyhat Period 1 and Tell Hajji Ibrahim Phases A2 and B, or roughly 3050 to 2900 BC. It is possible the site has earlier Hajji Ibrahim Phase A1 and later Hajji Ibrahim Phase C (early Sweyhat Period 2) occupation levels. Phase A1 material has few obvious ceramic diagnostics, making it difficult to recognize in surface assemblages. Moreover, the earliest levels of occupation at Khirbet Taha, as at Hajji Ibrahim, are almost certainly sealed by the later occupation and thus virtually invisible to archaeological survey. Hajji Ibrahim Phase C diagnostics are still poorly understood and the excavated assemblage is too similar to Phase B for ready distinction. The survey team found no mid-to-late 3rd millennium BC remains.

Except for a few late sherds not shown here, the assemblage is exclusively Plain Simple Ware (Chapter 5). We found a single sherd with incised design (Figure 6.3: A) and some vessels have incised lines below the rim (Figure 6.3: F). The open forms include a carinated bowl with thinned rim (Figure 6.3: C) similar to one from Hajji Ibrahim A2. A cyma-recta-related cup (Figure 6.3: D) is identical to a Hajji Ibrahim Phase B example and one from Tell es-Sweyhat Period 1 (Armstrong and Zettler 1997: Appendix 2.1, c). A vat (Figure 6.3: H) is similar to one from Hajji Ibrahim Phase A2. Another form (Figure 6.3: I) has parallels in Hajji Ibrahim Phase B and late Phase 2 at Tell es-Sweyhat.

Jars have strong parallels with those from Hajji Ibrahim Phases A2 through C. Rim forms are limited in range, but we found examples with simple rims and interior indentations (Figure 6.3: J, K) and a jar with pinched rim (Figure 6.4: A), providing a strong link to Hajji Ibrahim Phases B and C and Tell es-Sweyhat Periods 1 (Armstrong and Zettler 1997: Appendix 2.1, i, j). Another large jar (Figure 6.4: B) has a rim form similar to ones from Hajji Ibrahim Phase A2.

Site 14 Tell al-Hassan/Tell Jedi (Mid to Late 3rd Millennium BC)

Tell al-Hassan, recorded as Tell Jedi on the 1943 Levant Series "Djerablous" map, is situated atop a bluff along a deep north-south running wadi. The modern village of al-Hassan lies across the wadi to the east. The villagers informed us that in the past many of the surrounding locations were called Tell Jedi. The archaeological deposits appear to be shallow and run along the wadi approximately 250 m. I tentatively estimate the surface artifact scatter to cover 5-10 ha. Large numbers of rectilinear stone footings were visible on the surface and probably date to the mid to late third millennium BC (Tell es-Sweyhat Periods late 2 through 4). The survey team collected a total of 45 diagnostic sherds; all but one was EBA plain simple ware.

Particularly diagnostic are the cups and bowls with beaded rim (Figure 6.5: D-J). These have parallels with Tell es-Sweyhat Periods late 2 and 3 and Tell Kebir (Porter 1995: Figure 6: 1, 19). These forms do not postdate the Early Bronze/Middle Bronze transitional period (Sweyhat Periods 5 and 6). Incurving bowls with beaded, banded or thickened rim are quite common in medium (Figure 6.6: A-D) and large sizes (Figures 6.6: F, G, 6.7: A-G). These forms occur at Sweyhat in Periods late 2 through 4.

Small jars typically have flaring simple or thickened rims (Figure 6.8: A-C), as well as band rims with indentations (Figure 6.8: D, E). The band rim forms have close parallels at Tell es-Sweyhat Area IV, dated to the late 3rd millennium BC (Holland 1977: Figure 7: 5). One jar with out-turned "doubled" rim is particularly diagnostic of the late 3rd/ early 2nd millennium at Sweyhat (Figure 6.8: F). A similar form from Tell Kebir Late Building One is shown as a base (Porter 1995b: Fig. 6, 26); our example could equally be from a base or stand. High-necked jars with simple rim are also common (Figure 6.8: G-I). We also found an example of a collared-rim cup or jar (Figure 6.8: L), which is dateable to the transition between the EBA and Middle Bronze Age (Cooper 1998: Figure 2: c). Similar vessels were found at Sweyhat and Tell Kebir (Holland 1976: Figure 9: 9, 20; Porter 1995: Figure 6: 5, 17).

The typical medium and large jar was neckless or had a short neck and flaring thickened or molded rim. Medium and large storage vessels were also present (Figures 6.9 to 6.11).

The survey team found no evidence of earlier EBA occupation or EB/MB transitional material (Sweyhat 5 and 6), indicating that settlement was of short duration and coterminous to the peak in urbanization along the Euphrates and Balikh.

Site 22 / Joub esh-Shayeer (Mid to Late 3rd Millennium BC)

This low mounded site lies in the village of the same name atop a ca. 15 m high rocky outcropping on the east side of a deep, winding north-south wadi — the Wadi el-Fayyed. We distinguished several small mounds, estimated at 0.25 ha each, on the edge of the village and many more may lay under the modern occupation. Some looting was evident. The eastern side of the rock outcropping falls off steeply to the wadi. Near the edge of this cliff is a 1-by-1-m square shaft descending straight down into the rock. This opening is surrounded by four large, flat limestone slabs. I could not determine the actual depth, but according to local villagers, this feature gives the place its name "joub", a type of well sunk into rock (Nutahara 1982). The shaft might be part of an ancient qanat system given it topographic and geological surroundings (Lightfoot 1996). However, the survey team was unable to spend time searching for other shafts on the plain below. Most of the identifiable surface sherds from the site were of the mid to late 3rd millennium BC and obviously not associated with a joub or qanat. However, we found a few sherds diagnostic of Byzantine wares, which may help to date the feature.

We found one fragment of a late-3rd millennium colander bowl (Figure 6.12: A). Simple cups were present (Figure 6.12: G-K), as well as incurving beaded rim cups (Figures 6.12: L-W, 6.13: A). Medium and large bowls with incurving, thickened rims were also ubiquitous (Figure 6.13: E-I). Small and medium jars typically had flaring, thickened (Figure 6.14: A-C) or banded rims with indentations (Figure 6.14: D-H). Rounding out the mid-to-late 3rd millennium forms is a single example of a “screw top” jar (Figure 6.14: J), which begin in Tell es-Sweyhat Period 3 and continue in Period 4 (cf. unpublished TSw numbers 93.0490.11, 93.0490.3, 93.0623.4). Medium to large jars typically have short necks and straight or slightly flaring thickened rims (Figures 6.14: K-M, 6.15: A-D).

Site 23 / Tell Shayeer (Mid to Late EBA)

Approximately 4 km NNW of Joub esh-Shayeer following the course of the Wadi el-Fayyed lies the site and modern village of Tell Shayeer. The site is comprised of a 5 m high conical mound and a topographically-indistinct lower tell surrounding it. The lower area is currently under cultivation. I estimate the mounded area at 2-3 ha. There is a small modern cemetery atop the mound. On the lower portions of the main mound and in the cultivated lower site, the survey team collected abundant mid-to-late 3rd millennium sherds and a single terracotta animal figurine head. We found a few sherds of a late occupation on the upper tell.

Small open forms include a number of simple cups (Figure 6.16: E-L), club rim bowl (Figure 6.16: M) and shallow simple rim bowls (Figure 6.16: N-P). Incurving beaded-rim cups occur in high proportions (Figure 6.17: A-M). Medium to large bowls have simple (Figure 6.17: O-Q) and incurving band (Figure 6.18: A-B) and club rims (Figure 6.18: C-G).

Small jars typically have out-turned simple and thickened rims (Figure 6.19: A-G) or banded rims with indentations (Figure 6.19: I, J). Medium to large jars form two basic categories. The first group is medium to large neckless or short-necked jars with thickened flaring rims (Figure 6.20: A-G). We did not find the second group at the other mid-to-late 3rd millennium sites, which was made up of a nearly hole-mouth type with flattened (Figure 6.21: A-D) or thickened rims (Figure 6.21: E-G). Two jars with thickened rims and interior rim indentations may belong to this category, but may have a much higher shoulder than the preceding group (Figure 6.21: H, I).

Summary

The overall EBA settlement history, while highly tentative, demonstrates that upland settlement parallels developments on the Tell es-Sweyhat plain. The early EBA settlements are in the portions of the uplands closest to the river (Khirbet Taha). Sites of this period are likely under-represented due to overlying occupation.

In the mid to late 3rd millennium, the number of settlements increases. These upland sites tend to be larger (5-10 ha) than those of the early 3rd millennium. Prominent bluffs near major wadis seemed to be the preferred location for settlement. This is likely due to the availability of seasonal water supplies and localized geological fluctuations making the water table shallower, which would have made wells easier to dig. Moreover, the tracts of land near the wadis would have been relatively productive for dry farming.

One hypothesis to test by excavation would be whether these sites have storage facilities akin to those at Tell Hajji Ibrahim. This might indicate that these sites served as pastoral stations where shelter, fodder and water were available throughout the year. Such settlements would help to distribute livestock in winter, when the availability of rangeland grazing was at a low point. Such a system would make effective use of pockets of arable land in the uplands. While the sites are located in rugged terrain, far away from the cities of the Euphrates and Balikh, the need for the costly transport of cereal surpluses could be obviated by using them for the pastoral economy.

These EBA settlements likely had wells given their location, the presence of substantial permanent settlement and existence of more recent pre-modern wells. Thus, they would have increased the range pastoralists could penetrate the uplands. Moreover, they would extend the length of time the innermost rangeland was accessible further into the summer as seasonal water sources dried up.

The reconnaissance did not cover much of the area where one would expect to find pastoral satellite sites of the Tell es-Sweyhat embayment, specifically the grid squares north if Sites 15 and 17 (Figure 6.1a). Possible candidates, even though no obvious EBA surface sherds were found, were the mounded sites of Beer Mjeibna and Mjeibna (Sites 15 and 16). The survey team collected surface sherds of a later occupation of an undetermined period from these site, but EBA levels could easily lie below. These sites would be excellent candidates for stratigraphic sounding.

 

Return to The Jezireh Regional Project

All Content © 2006-2010 Michael Danti
All rights reserved.