We would like to thank first and foremost Prof. Timothy Harrison and the CRANE Project for hosting Prof. Abbas Al-Hussainy as a visiting professor at the Department of Near and Middle Eastern Civilizations, University of Toronto. Without this hospitality this work would not have been possible. We would also like to thank the State Board of Antiquities and Heritage (SBAH) Baghdad for providing the permission and the support for carrying out this work in Qadisiyah. A special thanks to our friends and colleagues Elizabeth Gibbon and Dr. Michael Lewis whom read and commented on early versions of this paper. Our deepest gratitude to Mr. Walid Abdel Moneim whom without his help in the field this work would have been incomplete. Finally, our warmest regards to the three anonymous reviewers whose insightful comments contributed greatly to improving this work.
1The Ubaid and Uruk/Late Chalcolithic periods in greater Mesopotamia have been subject to intensive research programs and debates over the past four decades (see Henricksen and Thuesen 1989; Stein and Rothman 1994; Rothman 2001; Postgate 2002; Carter and Philip 2010; Marro 2012; McMahon and Crawford 2014; Abu Jayyab and Baldi 2022; Baldi et al. 2022). The significance of these periods and the transition between them, stems from the large-scale social changes attested throughout them. Between the end of the Ubaid (4600/4500 BCE) and the Middle Uruk/LC3-4 (3700-3400 BCE), we see the emergence of large urban centers (Adams and Nissen 1972; Adams 1981; Ur et al. 2007) that combined mixed populations (Adams 1981; Emberling 2003) and were characterized by a degree of hierarchy and social differentiation (McMahon 2009). This was manifest in the development of formalized administrative practices (Nissen et al. 1993; Frangipane 2001, 2002; Reichel 2002; McMahon 2009, 2014), the construction of large public works (Lloyd et al. 1943; Safar et al. 1981; Liverani 2006; Reichel 2009; Kelly-Buccellati 2010; Kepinski 2011) and administrative buildings (Lloyd et al. 1943; Safar et al. 1981; Oates and Oates 1997; Frangipane 2002; Oates et al. 2007), increased specialization in craft production (Al-Quntar 2009; Al-Quntar and Abu Jayyab 2014) and the sponsorship of public feasts (D’Anna 2012). This was accompanied, in both northern and southern Mesopotamia, by a substantial increase in settlement size and the emergence of early urban conglomerations (Adams and Nissen 1972; Adams 1981; Ur et al. 2007). These developments cumulate in the unprecedented expansion of the city of Uruk during the Late Uruk period (3200 BCE), which reached up to 200 ha in extent.
2Despite the strides that have been made in understanding the urbanization process in northern Mesopotamia, our knowledge of these processes in the Uruk heartland has rarely been examined (Nissen 1993; Algaze 2008). The dearth of this knowledge could primarily be attributed to the political instability in the south that spurred research in other regions. Our knowledge of the Early Uruk period in the south, until recently (see Lippolis 2016; Lippolis et al. 2019), has been primarily restricted to settlement pattern analysis (Adams and Nissen 1972; Gibson 1972; Adams 1981; Pollock 2001) or from narrow deep soundings (Boehmer 1972; Sürenhagen 1986). The gaps in our knowledge of the Early Uruk period have largely been explored through theoretical debates on potential developmental trajectories that occurred between the end of the Ubaid and the Middle/Late Uruk periods in southern Mesopotamia (see Algaze 2008; Benati 2018).
- 1 The MI.Enlil-Arakhtum Survey Project (2013 and ongoing) covers the areas to the north and west of (...)
- 2 Upon revisiting the site on October 31st 2022, with the surrounding fields clear of plant cover, w (...)
3In this paper, we present the results of a site survey carried out in 2017 at the site of Jemdat Zabi by the MI.Enlil-Arakhtum Survey Project1 directed by Prof. Abbas al-Hussainy. Jemdat Zabi (fig. 1) is a 6 ha.2 Early Uruk site located in the district of Sumer on the northern edge of the Nippur-Adab site cluster (see Adams 1981; Pollock 2001). Significantly, this site represents an archaeologically accessible medium-sized settlement dating primarily to the Early Uruk period. Through the presentation of the surface collections and the visible architectural remains, Jemdat Zabi has the potential to expand our understanding of the Early Uruk period and set the ground for future excavations at the site. Here we will present the results of our documentation and surface collections with the aim of contextualizing the site in its temporal and spatial setting.
Fig. 1 – Location of Jemdat Zabi, with the main sites mentioned in the text.
- 3 Some Early Uruk sites listed in the map are not completely confirmed. Sites such as documented Nip (...)
4The Early Uruk (4200-3800 BCE) has been a relatively obscure period. Nonetheless, numerous sites dating to the period have been identified from surveys (Adams 1972, 1981; Adams and Nissen 1972) and limited excavations (see Boehmer 1972; Sürenhagen 1986; Calvet 1991; Lippolis et al. 2019). From five major surveys in the alluvial plain, which include Warka (Adams and Nissen 1972), the Heartland of Cities (Adams 1981), Kish and Akkad (Adams 1972; Gibson 1972), Qadis (Marchetti et al. 2019, 2020) and recently, MI.Enlil-Arakhtum, over one hundred sites with Early Uruk occupation have been documented (fig. 2).3
Fig. 2 – The distribution of Early Uruk sites in southern Mesopotamia.
5Based on the survey work carried out in the southern alluvium, there appears to be distinct clustering that occurs during the Early Uruk period (Adams 1981: 66; Pollock 2001). Three occupation clusters could be identified here, separated by near empty boundary spaces in-between them (fig. 2). These clusters were identified from south to north as Warka, Nippur-Adab and Kish. Interestingly, based on settlement hierarchy and variation in location, each of these settlement clusters appear to be structurally and politically distinct (Pollock 2001: 217).
6In the Warka cluster, the settlement system clearly revolved around the city of Uruk. Uruk, during the Early Uruk period, reached over 70 ha in extent. This substantially exceeded all of its neighbours in size and, most likely, socio-political importance (see Adams and Nissen 1972; Pollock 2001; Algaze 2008; Algaze and Pournelle 2014), placing it at the top of the hierarchy in the south. The location of Uruk most likely played an important role in its growth. Its position on a bird foot delta overlooking a web of marshlands and river channels (Pournelle 2013: 17) not only allowed Uruk and its inhabitants to exploit the rich surrounding marshland resources but also take full advantage of the extensive waterscape, which played a major role in facilitating movement and allowing Uruk to draw on labour and resources from distant and diverse ecotones (Algaze 2008, 2018; Algaze and Pournelle 2014).
7The Kish cluster, furthest to the north, shows a stark contrast to the Warka cluster. The area comprises a series of relatively small sites, with the largest sites (AS219, Tell Abu Hurayabah and Ishan Hamid) not exceeding 20 ha. in extent. The settlements are distributed in a clear linear pattern, implying that these sites were located along north-south waterways (Adams 1981: 67). This suggests that the area was characterized by relatively stable river channels allowing settlements to continuously prosper along levees for an extended period from the Ubaid to the Middle Uruk period. This area remains the most obscure in regards to Early Uruk settlements, as it is not confirmed that the sites listed here (fig. 2) actually had Early Uruk occupation, where Adams listed the sites as Early/Middle Uruk with the majority abandoned during the Late Uruk period (Adams 1981: 66).
- 4 There has been an increased emphasis on the importance of marshland resources in Mesopotamian socie (...)
8The Nippur-Adab cluster, where Jemdat Zabi is located, is positioned between the Warka cluster to its south and the Kish cluster to its north (fig. 2). The alluvium here was characterized by the anastomosing and intermingled flows of branches of the Euphrates River system (Pournelle 2013: 14). This condition led to great diversity in seasonal ecosystems. The area here combines seasonal marshes interspersed with steppe zones along with areas of seasonal cultivation in the north (see Adams 1981). This mosaic created an area with a wide variety of affordances and resources, including marshland faunal and floral resources, access to steppe pasture and levee agriculture (Pournelle and Algaze 2014: 25). The ease of water transport4 further facilitated access to these diverse ecosystems.
9Beyond the differences we see in the ecosystem, this area had a completely different settlement structure from what we see to the north and south. In stark contrast to the Warka and Kish clusters, the Nippur-Adad cluster had a number of large settlements concentrated along its southern flank. The southern arc of the area from Tell al Hayyad to Abu Salabikh contained several large sites and a paucity of small villages. These sites included Tell al Hayyad (50 ha), Tell Dlehem (35 ha) and perhaps Nippur further north (estimated at 40 ha during the Early Uruk period). According to Adams (1981: 66), these larger sites overlooked sizeable areas of surrounding lands. These areas were larger than could possibly be cultivated by the sites’ inhabitants and consistent agricultural yields were not possible due to the absence of stable water channels. This has led Adams to suggest that these sites were specialized in a larger system of relations that incorporated the entire Nippur-Adab cluster, with larger sites regulating and facilitating reciprocal exchanges between different resource zones within the cluster (Adams 1981: 66).
10While culturally unified, southern Mesopotamia during the Uruk period cannot be regarded as a homogeneous entity with an overarching political, economic or social organization (Pollock 2001: 217; Algaze 2008: 114-115). Analysing spatial data, Pollock shows clearly that two distinct socioeconomic systems – one in Warka and another in Nippur-Adab – existed during the Early Uruk period (Pollock 2001: 218). Pollock (2001: 207-209) argued that the Nippur-Adab area during the earlier Uruk period was not as politically integrated as Warka, with households responsible for fulfilling their own subsistence needs and producing most of their craft needs. In contrast, the Warka cluster had Uruk at the epicenter, acting as its political, economic, and religious hub (Algaze 2008 114-115; Pollock 2001: 218). Under these circumstances, Uruk would have relied primarily on tribute extraction from nearby sites and redistribution of resources. In the Nippur-Adab cluster, larger settlements (especially Tell al Hayyad) had less power to influence smaller communities within the cluster. This lower degree of integration may have fostered more stable relations within the Nippur-Adab area (Pollock 2001: 218).
- 5 Neither the Akkad survey (Adams 1972) nor the MI.Enlil-Arakhtum (Al-Hussany in prep.) surveys have (...)
11The site of Jemdat Zabi provides us with an opportunity to explore social, economic, and political organization within the Nippur-Adab cluster during the Early Uruk period. Jemdat Zabi is the northernmost town along the Nippur-Adab frontier5 and the largest at slightly over 6 ha. From the visible architecture on the site's surface and the materials' distribution, we can begin to understand the spatial organization at Jemdat Zabi and situate the site in the broader debates surrounding the period.
- 6 Sites have traditionally been named after the site owner, features that characterize the site at f (...)
12The site of Jemdat Zabi (fig. 1) is located in Al-Qadisiya Governorate (UTM 38s 504921 3577194). The site was initially found by Mr. Ahmed Ali an employee of the Sumer office of the State Board of Antiquities and Heritage (SBAH) of Iraq. While on his daily work in the agricultural land north of the town of Sumer, he collected some painted pottery sherds and clay sickles from the site. Following its discovery, the site was documented in 2017 as part of the survey work carried out under the umbrella of the MI.Enlil-Arakhtum survey project in the district of Sumer by an Iraqi archaeological team. The site was named Jemdat Zabi after Zabi al-Shri al-Jamawi Al-Boujmaa, a member of the al-Saaed clans from the Zubaid tribes and the person who lived on the eastern side of the site6.
13Jemdat Zabi is oval in plan, with the highest elevation about one meter from the surrounding plain. The site is bound on the east by the Makia canal and on the west by the Maasumah canal, both modern and originating from the Shatt al-Dagharah. The surrounding areas to the north, south and west of the site are agricultural lands. A group of houses were built on the southeastern part of the mound in the 1940s by the farmer Zabi Al Sheri. The site was not registered in the registry of archaeological sites in Iraq during past surveys carried out by the SBAH and was not registered during Adams’s work in the Akkad survey (Adams 1972). The absence of the site from past records is perhaps due to its low visibility relative to the surrounding plain. The increased siltation resulting from its location between two canals and the height of the crops during the growing season may have played a role in further masking the site.
14Regarding collection strategy and documentation, we used similar methods to those applied by the QADIS survey team see (Marchetti et al. 2019). After the site was reported, its exact location was pinpointed using Google satellite imagery, where its rough extent was recorded. On the ground, the site was mapped using a DJI Phantom 4 drone, creating a 3D model of the site and a topographic contour map (fig. 3a-b).
Fig. 3 – a. Drone orthophoto of Jemdat Zabi; b. Topographic map of Jemdat Zabi.
- 7 A recent visit to the site (October 31st 2022) allowed us to observe many more artifacts on the su (...)
15For collection purposes, the site was divided into six collection units (fig. 4) A, B, C, D, E and HO, to facilitate an intensive survey collection. The area on which the farmer's houses were built was registered under the name of (HO). Large collection units were favoured over a smaller grid collection due primarily to the low density of artifacts observed on the site’s surface (n: 340 artifacts in total). The paucity of artifacts was perhaps attributed to the minimal anthropogenic disturbance at the site, as Jemdat Zabi has not been recently ploughed or used for agriculture.7 Each unit was collected completely by a single surveyor, with pottery separated from other finds (clay and stone objects). Besides ceramic slag, 340 diagnostic artifacts were collected in total from the six collection areas (table 1); among these were 201 pieces of diagnostic pottery sherds. In our collection, we elected not to collect body sherds unless they were painted. The finds were later washed, photographed, drawn and digitized in the lab.
Fig. 4 – Collection units at Jemdat Zabi.
Table 1 – Distribution of objects across the various collection areas at Jemdat Zabi.
Area |
A |
B |
C |
D |
E |
HO |
Total |
Rims |
18 |
16 |
32 |
27 |
38 |
5 |
136 |
Bases |
6 |
6 |
11 |
11 |
3 |
0 |
37 |
Spouts |
0 |
1 |
2 |
0 |
1 |
1 |
5 |
Dec. Sherds |
4/3 |
4/3 |
3/1 |
9/4 |
3/1 |
0 |
23/12* |
Stone objects |
1 |
2 |
2 |
5 |
9 |
0 |
19 |
Clay Sickles |
41 |
20 |
0 |
24 |
37 |
0 |
122 |
Clay Mullers |
2 |
1 |
2 |
1 |
0 |
1 |
7 |
Clay Picks |
2 |
1 |
1 |
0 |
0 |
0 |
4 |
Misc. Clay Obj. |
0 |
0 |
0 |
1 |
3 |
0 |
4 |
Total Diag. |
71 |
47 |
51 |
73 |
91 |
7 |
340 |
Cer. Slag (kg) |
0 |
0 |
8 |
58 |
2 |
0 |
68 |
(*) total decorated/decorated rim sherds.
16The density of artifacts varied from unit to unit, with the highest collections coming from unit E (91 diagnostic artifacts) and the lowest coming from unit HO (seven artifacts). The study of the artifacts suggests a single period occupation dating to the Early Uruk (LC1/2) period, with no other periods represented from the collection. Another significant characteristic of the site is the visibility of architectural features on the surface in the western part of the mound. This became apparent after processing and viewing the drone imagery (figs. 3a, 5).
Fig. 5 – Architecture visible on the surface of Jemdat Zabi.
Table 2 – Distribution of vessel types across the various collection areas from Jemdat Zabi.
Area |
A |
B |
C |
D |
E |
HO |
Total |
BRB |
0 |
0 |
14 |
0 |
12 |
2 |
28 |
Uruk bottles |
4 |
4 |
2 |
5 |
1 |
0 |
16 |
Proto-BRB |
2 |
5 |
7 |
0 |
3 |
1 |
18 |
Dec. Sherds |
4 |
4 |
2 |
9 |
3 |
0 |
23/12* |
Corrugated pot |
3 |
4 |
4 |
2 |
2 |
0 |
15 |
Cord-imp. pot |
2 |
0 |
2 |
1 |
3 |
0 |
8 |
Simple bowls |
0 |
1 |
1 |
1 |
1 |
1 |
5 |
Deep bowls |
1 |
2 |
1 |
1 |
0 |
0 |
5 |
Band rim bowl |
0 |
0 |
3 |
0 |
0 |
0 |
3 |
Deep band rim |
0 |
0 |
1 |
2 |
1 |
0 |
4 |
Shouldered jar |
0 |
0 |
3 |
0 |
0 |
1 |
4 |
Flaring rim Jar |
1 |
0 |
5 |
0 |
3 |
0 |
9 |
Straight neck Jar |
1 |
0 |
1 |
0 |
1 |
0 |
3 |
Corrugate Bowl |
0 |
0 |
2 |
0 |
0 |
0 |
2 |
17The accessibility of the Early Uruk levels at Jemdat Zabi, just below the surface, coupled with its manageable size gives us a unique opportunity for investigating early Mesopotamian momentum towards urbanism and a prelude to the large-scale settlements that emerged during the Middle and Late Uruk periods.
- 8 The aerial orthophoto of Jemdat Zabi was taken in March 2017 following a period of light rains. Th (...)
18From the aerial imagery, we can make out roughly four large buildings visible on the surface8. These buildings were designated A, B, C and D. Other features visible on the surface include parts of two possible perimeter walls and portions of other smaller walls and features (fig. 5). The architectural features were primarily concentrated in the west in survey units C and E with some features extending into unit A. The clearest or most complete plan is that of building A. Here we have traces of a monumental rectangular building, oriented northeast to southwest, measuring roughly 20 × 25 m in size. Two buildings on either side of and parallel to building A could also be observed; buildings C to the southeast and B to the northwest. Building B appears to be slightly larger than building A; however, without carrying out excavations at the site, we cannot be certain since its full extent was not visible on the surface. A fourth incomplete building, building D, could be seen to the east of the three above-mentioned structures and could prove to have some connection with building C. Several other walls appeared across the western portion of the site. These walls were not as thick as the walls of the three buildings (A, B, C) and a clear building plan could not be traced around them.
19A thick buttressed wall appears to enclose buildings A and B. We were only able to trace portions of this feature’s south and east sides. The wall seems to represent an enclosure wall for the two large buildings and may have been connected to another, larger wall that enclosed the central part of the site.
- 9 It is believed that the site could have extended to around 11 ha. based on a recent visit (see foo (...)
20The architecture allowed us to define a number of paths or streets between the buildings. A major path seems to be defined between the interior of the buttressed enclosure wall and the eastern walls of buildings A and B (fig. 5). This street was roughly 2-3 m wide and visibly extended about 50-60 m at least. Another path, albeit less substantial, could be seen between the southern part of the enclosure wall and building C to the south. The architecture detected at the site suggests that there was a concentration of large buildings along the western and northwestern portions of the site. From what we can see from the visible architecture, the layout seems to be highly organized and perhaps centrally planned, with buildings of roughly equal sizes defined by streets and pathways and demarcating a central open space from the rest of the settlement. If this was the case and the buildings observed on the surface were public buildings rather than households, then the area of habitation at the site may have been extremely limited. If so, then either the site had a specialized administrative/religious function serving the region’s population, or habitation extended into the lower fields that we were unable to access due to agricultural coverage9.
21The best parallel to Jemdat Zabi is the contemporary, recently excavated Tell 7 (TB7) of the site of Tlul al Baqarat (Lippolis 2016; Lippolis et al. 2019). TB7 is close in size to Jemdat Zabi (roughly 7.5 ha in extent) and therefore providing a good comparison. The central area of the mound seems to have acted as an administrative/religious hub for the settlement with three soundings (S1, S2 and S4) producing evidence of niched and buttressed architecture, a portion of a platform, decorative wall cones and remains of painted plaster, all features associated with Early-Middle Uruk public and religious architecture. In sounding S3 along the slope of the mound to the northwest, the excavators uncovered what appears to be a large household (building A) with over 20 rooms (Lippolis et al. 2019: 139). The household was occupied during two phases (3 and 2), both of which date to the Early Uruk period. The inhabitants of this household seemed to have been involved in a number of domestic and craft activities, as seen by the presence of tannurs, agricultural tools, grain storage areas and ceramic kilns. From the finds at TB7, we can propose that the lower areas surrounding Jemdat Zabi could produce evidence of domestic household activities.
22From what could be observed at Jemdat Zabi and in addition to the finds at TB7, we can see that some smaller sites or ‘towns’ seemed to have followed some of the practices attested during the preceding Ubaid period. During the Ubaid period, site size does not necessarily correlate to the presence or absence of public architecture. We see public architecture at larger sites such as Eridu (40 ha) and at smaller sites such as Uqair (6 ha) and Tepe Gawra level XIII (3 ha). That being said, there is a stark difference in the overall size range and number of settlements between the Ubaid period and the early Uruk (see Adams and Nissen 1972: 11). In the Nippur-Adab region, there seems to be an overall increase in settlement density during the Early Uruk period, this is accompanied by a clearer settlement size hierarchy, where several sites, such as Tell al-Hayyad, HC1237 and HC1172 reached 40, 35 and 25 ha. respectively. With that in mind, the presence of public architecture at relatively smaller sites (i.e., Jemdat Zabi and TB7) during the Early Uruk has different implications to its presence at Ubaid sites in the northern alluvium, which have less size differentiation between them. This observation supports Pollock’s assertion that during the earlier Uruk, settlements in the Nippur-Adab area were not as highly integrated as in the Warka region and that larger communities there had less power to influence smaller communities (Pollock 2001: 208-209).
23Area A is an oval-shaped unit located at the site’s summit. In total, 71 diagnostic artifacts were collected in this area. The finds consist of pottery fragments, including rims, body sherds, bases and stone and clay tools. In total, 25 diagnostic sherds were recovered, mainly consisting of rim sherds (18), in addition to six bases and one decorated body sherd. A few decorated sherds were recovered from this area. These include three diagnostic jars decorated with horizontal streaks, two of which are decorated with two horizontal lines in black; in one of them, the two lines are connected with five horizontal streaks. Similar decorated wares were prevalent during the Early Uruk period (Adams and Nissen 1972: fig. 40.4).
24In total, 41 fragments of clay sickles were recovered from this area; all of them were well manufactured and likely fired at a high temperature based on their green colour. A fused stack of clay sickle wasters was also found in the area, suggesting that the sickles were manufactured locally rather than at a nearby site. Other clay objects recovered in this area were two clay mullers and two picks made of clay.
25Area B covers the northeastern corner of the site. A total of 47 diagnostic artifacts were collected in Area B, 27 of which were diagnostic ceramic sherds (see table 1). This assemblage shows affinities with the Early Uruk assemblages observed in Adams and Nissen’s survey in the vicinity of Uruk (Adams and Nissen 1972). In addition to the ceramic assemblage, 20 fragments of clay sickles of various shapes, sizes and colours were found. Finally, a single clay muller and a single clay pick were also found in this area.
26Area C covers the northwestern corner of the site. A total of 51 diagnostic artifacts were recovered from this area, of which 47 were diagnostic ceramic sherds. What distinguishes this area is the prevalence of bevelled rim bowls (BRBs) and the complete absence of clay sickles. BRBs accounted for 14 of the diagnostic sherds found. They were hand or mold made, from a coarse clay with mainly vegetal temper with colours ranging between yellow and green. Additionally, 19 rims were found, including a jar with a broken spout at the top of the shoulder (fig. 6H). Worth noting here is that two clay mullers were also recovered from this area, in addition to a complete flint blade.
Fig. 6 – Vessels collected from the surface of Jemdat Zabi.
Fig. 7 – Vessels collected from the surface of Jemdat Zabi.
Fig. 8 – Decorated vessels collected from the surface of Jemdat Zabi.
27Area D covers the southeast corner of the site. In total, 73 diagnostic artifacts were recovered from the area. The finds here were similar to other areas, the most common of which was pottery. What distinguishes the pottery of area D is the higher number of decorated forms both in terms of rims and body sherds. Nine of the decorated pottery sherds were painted with straight and wavy lines. The most represented motifs included parallel lines connected by vertical lines.
2824 fragments of dark-coloured clay sickles were found. Most fragments were green, except for one that was yellow in colour. Similar to area A, we found a fused stack of clay sickle wasters due to over-firing at high temperatures without spacers. Worth noting here is that the only spindle whorl recovered during the site survey came from area D.
29Area E is located in the southwest corner of the site. In total, 60 diagnostic artifacts were recovered from this area during the survey. We primarily recovered ceramic sherds and stone objects from this area. In total, 42 diagnostic ceramic sherds were recovered, 38 of which were rim sherds. Forms include pots, jars, cups and bowls. Similar to area C, a high proportion of BRBs were found in area E, 12 fragments in total.
30Clay sickles were also found in this area. In total, 37 broken sickles were found. These sickles were of different shapes and colours. Other notable clay objects from Area E include a fragment of a clay axe (fig. 9F) and a clay ladle (fig. 9E).
Fig. 9 – Clay and stone objects collected from the surface of Jemdat Zabi.
31What distinguishes this area is the relatively higher concentration of stone objects compared to other areas. Worth noting here is the presence of an alabaster mace head (fig. 9M), a fragment of a flint core (fig. 9J) and two potential fishnet weights, in addition to other miscellaneous stone objects.
32Area HO marks the location of a modern house compound along the southern slope of the site. Due to the presence of modern architecture, this area, understandably, produced the least number of finds. Nevertheless, we were able to recover seven diagnostic artifacts, all of which were pottery fragments, except for a fragment of a clay muller. Five of the ceramic sherds were rim fragments while one was a spout. Two of these sherds were fragments of BRBs. A short spout of the type that prevailed in the Early Uruk period was also recovered in this area.
33Identification of Early Uruk sites in survey work could be attributed to Adams and Nissen’s (1972) analysis of the Eanna sounding ceramic materials from Uruk. There they were able to identify a number of fossil types useful for survey work. More recent excavations (Di Michele 2016; Lippolis et al. 2019; Bruno 2020) and quantitative analytical work (Finkbeiner 2001; Bruno 2020; Baldi 2022; Volpi 2022) have shed more light on Early Uruk assemblages.
34Overall, we collected 201 diagnostic pottery sherds. Of these, 23 had traces of painted decoration, 27 had applique decoration and two were incised. Fourteen broad vessel types could be distinguished (see table 2). The most common are two types of mass-produced bowls: BRBs and wide flower pots with a footed base (fig. 6A-B). These were followed by Uruk bottles (fig. 6C), corrugated pots and bowls (figs. 6D, 7H-I), pots with rope applique (fig. 6E) and band rim bowls (both deep and shallow) (fig. 6D-G). Beyond these common types, we recovered a number of different flaring rim jars, shallow simple bowls, shouldered jars and straight neck jars (figs. 6, 7). It’s worth noting here that a flaring rim jar with a spout was recovered in area E (fig. 6H), similar to examples found in the Eanna Temple sounding at Uruk level XIII (Sürenhagen 1986: fig. 327) and at Tlul al Baqarat (Lippolis et al. 2019: fig. 13.3). In area C, one example of flaring rim jar cooking pot with a scraped body was found. It is difficult to attribute a periodization to this vessel as it seemed to have been used during the majority of the Uruk period.
35BRBs were the most common vessel type recovered during the survey. BRBs were typical throughout the Near East during the entire duration of the Uruk period (see Roaf 1990: 65). A fused stack of BRBs was found in area E which, together with the higher concentration of BRBs in that area, suggests that they were produced somewhere along the western part of the mound. Wide flower pots with a footed base on the other hand seem to be the earliest form of mass-produced bowls in southern Mesopotamia, with their distribution limited to the southern Mesopotamian alluvium during the early Uruk period. These vessels have been found at the Eanna sounding at Uruk level XII (Sürenhagen 1986: fig. 316), the Uruk Steingebäude sounding K XVII layers 4-2 (Boehmer 1972: fig. 387-388), Tell el ‘Uwaili (Calvet 1991), Eridu levels VII-VI (Safar et al. 1981) and from the recent excavations at TB7 at Tlul al Baqarat (Lippolis et al. 2019: fig. 12.3-12.9). This type was also recovered from the Early Uruk sites WS046, WS178 and WS386 from the Warka survey (Adams and Nissen 1972).
36Corrugated and rope applique pots are common southern Uruk types that have been found primarily during the Early Uruk period. Corrugated pots have been found at Tlul al Baqarat (Di Michele 2016: 101; Lippolis et al. 2019: fig. 14.5-14.8), level XII of the Eanna sounding at Uruk (Sürenhagen 1986: fig. 231), the Uruk Steingebäude sounding K XVII layers 4-2 (Boehmer 1972: figs. 510.11, 555), Tell el ‘Uwaili couche C (Calvet 1991: figs. 87, 108-109) and the Early Uruk site of WS178 (Adams and Nissen 1972). Pots with rope applique were also present. In general, they seem to have similar forms to the ridged or corrugated pots, however they use the application of a finger-impressed coil of clay as a decorative feature instead of the corrugated ridge. Similar examples have been found at other Early Uruk sites such as Uruk itself, level XII of the Eanna sounding (Sürenhagen 1986: figs. 317-318, 323-325), the Uruk Steingebäude sounding K XVII layers 4-2 (Boehmer 1972: fig. 367), Tell el ‘Uwaili couche C (Calvet 1991: fig. 119), Tlul al Baqarat (Lippolis et al. 2019: fig. 15), in addition to sites WS022, WS178, WS218 and WS386 from the Warka Survey (Adams and Nissen 1972).
37Another common pottery type from our survey was the pointed-base Uruk bottle. While some of the other types mentioned could be found during later periods, albeit with variation in form, these bottles are found primarily during the Early Uruk period. For this reason, in addition to its high frequency of occurrence, this vessel type was one of the main type fossils used to identify Early Uruk settlements in the Warka survey (Adams and Nissen 1972). Parallels could be found at sites WS022, WS042, WS178, WS218 and WS386. In excavated contexts they were found in level XIII of the Eanna sounding (Sürenhagen 1986: fig. 315), Tell el ‘Uwaili couche C (Calvet 1991: fig. 105) and Tlul al Baqarat (Lippolis et al. 2019: fig. 16.10).
38Finally, two examples of simple rim bowls with wavy incised lines on the interior were recovered from Jemdat Zabi (fig. 7F-7G). These bowls tend to be attributed to the Ubaid 5/Early Uruk period. Examples of this type were found at the Uruk Steingebäude sounding K XVII layers 4-2 (Boehmer 1972: fig. 427), Tell el ‘Uwaili couche C (Calvet 1991: figs. 79-80) and WS275 (Adams and Nissen 1972: fig. 63.12).
39In addition to these common forms, a few painted sherds were found during the survey (n:22). Painted wares began to decrease during the late Ubaid period and by the Middle Uruk, they were almost completely absent except for a few residual sherds (see Sürenhagen 1986). At Jemdat Zabi (fig. 8), the decorative patterns and decorated vessel forms (n:12), find strong parallels with vessels in excavated Early Uruk levels, such as at Uruk, Tlul al Baqarat, Tell el ‘Uwaili.
40The most common decoration pattern observed was thick parallel painted bands (fig. 8F-8G). In fact, all sherds had either one, two or three bands. Of course, since we are dealing with sherds, we cannot be certain that the composition was exclusively made up of bands. In cases where we had other decorative motifs, they were placed in a register enclosed by bands. The motifs found were vertical lines or blobs (n:2) connecting the bands (fig. 8D-8E), two wavy lines (n:4) running parallel to the bands (fig. 8A-8C) and a zigzag line framed by bands (n:2). Two unique pieces were also recovered (fig. 8H). Both these pieces had cross-hatched diamonds, one combined with a framed zigzag pattern. Worth mentioning here was the presence of two incised pieces from area E. One of the pieces was a body sherd with cross-hatched incisions (fig. 6F), while the other was a rim lug with roulette pattern (fig. 6G).
41Similar simple patterns can be observed at Tlul al Baqarat (Lippolis et al. 2019: fig. 17.4-17.5) and Tell el ‘Uwaili couche C (Calvet 1991: figs. 128, 130, 132, 162). Also, the majority of rim sherds with painted decoration were from corrugated pots (fig. 8A, 8C, 8F), which we do not see at other sites with the exception of one example with a painted band from Tell el ‘Uwaili couche C (Calvet 1991: fig. 175).
42The surface ceramics from Jemdat Zabi show clear parallels with Ubaid 5/Early Uruk assemblages from across southern Mesopotamia. Based on the Eanna sequence at Uruk, Finkbeiner (2001: 152), shows – using ordered presence absence matrix – a clear separation between the Ubaid and early Uruk during level XII occupation. In this level, types such as the ridged and rope applique pots and BRBs begin to appear. While forms seem to be clearly distinct between levels XVII-XIII and XII-IX, distinctions in decoration seems to be quantitative rather than qualitative (see Nissen 1993), with a smaller percentage of decorated sherds from level XII onwards (Finkbeiner 2001: 153). Moreover, incised and excised decoration increases during early Uruk levels (Finkbeiner 2001: chart 3). For the purpose of obtaining a higher temporal resolution and refining their survey results, Adams and Nissen (1972: 100) built their understanding of the Early Uruk from Eanna sequence levels XIV-VIII/VII. Types such as Uruk bottles, wide flower pots, applique decoration and shoulder spouts were the primary fossil types in distinguishing the Early Uruk period. More recently, Volpi (2022) and Baldi (2022) reached similar results using cross occurrence seriation tables. Volpi (2022: figs. 11, 12) and Baldi (2022: table 2) each have isolated types that could be attributed to the Ubaid 5/Early Uruk period such as banded rim bowls, wide flower pots with a footed base and deep urns with applique decoration. They saw that these types among others occur alongside types that carry on from the Ubaid 4 and earlier.
43Clay sickles were by far the most common clay objects found at the site. Despite their widespread distribution, little has been published about clay sickles (Abu Jayyab et al. 2020: 190). Clay sickles are highly fired with a hardened cutting edge (fig. 9A). Use-wear studies from Tell el ‘Uwaili suggest that clay sickles were used for harvesting cereals and processing tough stemmed plants like reeds, rushes and sedges, fronds, and fruits (Anderson-Gerfaud 1983: 180; Benco 1992; Pollock et al. 1996). With the near absence of flint blades at Jemdat Zabi, the clay sickles could have been harvest tools intended for use in multiple tasks.
44The majority of the clay sickles were recovered from the south and eastern parts of the mound in areas A, B, D and E (fig. 10a). In addition to the higher concentration of clay sickles, the east side of the mound also produced evidence of sickle firing with stacks of clay sickle wasters found in areas A and D (fig. 11a). The southeastern location of firing installations (especially since the sickles required longer firing) makes sense when we consider that the prevailing winds in the region are western/northwestern. This is supported by the high concentration of ceramic slag in area D, where a detailed collection of slag may hint at the presence of kilns in the center of the area (fig. 11b). Finally, it is worth noting that sickles for both left and right-handed users were found at the site. This was discerned based on the position of the cutting-edge bevel relative to the curve of the sickle.
Fig. 10 – Density of artifacts collected from the surface.
a. Clay Sickles; b. Bevelled rim bowls. c. Painted wares; d. Uruk bottles.
Fig. 11 – Distribution of wasters (a), ceramic slag (kg; b).
45Other, albeit less frequently recovered, clay objects included clay mullers (fig. 9C), clay picks (fig. 9H), clay ladles (fig. 9E), clay cones or cylinders (fig. 9D), a spindle whorl (fig. 9G) and a fragment of a highly fired clay axe (fig. 9F). Objects such as clay mullers and sickles were initially thought to have been characteristic of the Ubaid period. However, several scholars have shown that clay sickles were also part of Uruk period assemblages and even seen in later periods (see Adams and Nissen 1972; Matthews 2002: fig. 48; Abu Jayyab et al. 2020: 190), with clay mullers also found during the LC1 period in northern Mesopotamia (Stein and Fisher 2020: 134) and in the south (see Lippolis et al. 2019). Parallels to the ladles at Jemdat Zabi can be seen at Middle Uruk ‘colonies’ such as Chogha Mish (Delougaz and Kantor 1996: plate 82H-82L), Tell er-Ramadi (Abu Jayyab et al. 2020: fig. 9.16) and Qraya (Reimer 1988). Parallels to the clay cylinders were found in abundance at the nearby Early Uruk site of Tell al Baqarat (Lippolis et al. 2019: fig. 5b).
46Stone objects were generally rare from the surface collection of the site. As seen from the amount of fired clay objects, stone was generally substituted by clay (see Algaze 2008). Nevertheless, there seems to be some imported stones that arrived at the site. A majority of the stone objects were found in the southern part of the site in Areas D and E. A total of three pieces of flint were recovered: a well-made flint blade (fig. 9I) and fragment of a flint core (fig. 9J) and two flint bifaces (fig. 9K-9L). Other objects include two fishnet weights and what appears to be an alabaster mace head (fig.11M).
47From our work at Jemdat Zabi we can make a few preliminary, albeit important observations. The site seems to be an important site that dates to the Early Uruk period, perhaps sometime between 4200-3800 BCE. Although some vessels found at Jemdat Zabi, such as BRBs, corrugated pots and band rim bowls, were present during later periods (Middle Uruk) and the painted wares could be from an earlier Ubaid stratum at the site. Assemblages from other sites show that the pottery recovered from the surface at Jemdat Zabi represents a coherent assemblage with clear parallels found in the Early Uruk levels at the sites of Uruk (XII-IX), Tlul al Baqarat (TB7) and Tell el ‘Uwaili (couches C-G). Moreover, if the surface assemblage represented a longer or a later Uruk period occupation, we would expect to see a larger variation and certain common types from the Middle and Late Uruk periods, which we do not see represented in our survey assemblage. Examples include string-cut bowls, jars with droopy spouts, four lug jars and reserved slip jars, none of which were detected during the survey. This supports the idea proposed by Nissen that the development from the Ubaid to the Uruk period in the south was a gradual process detectable through quantitative rather than qualitative analyses (Nissen 1993: 125).
48Despite its small size, Jemdat Zabi seemed to have commanded an important position within the landscape. Its location on the northern frontier of the Nippur-Adab settlement system (fig. 2) along a 25 km settlement gap between it and other Early Uruk sites in the Kish area suggests that the site may have played an intermediary role between the two settlement systems. Moreover, this location probably provided the site with unhindered access to pasture, agricultural fields and marshland resources. The importance of this geographical position is reflected in the level of organization gleaned from the monumental architecture attested on the surface. This may suggest that Jemdat Zabi could have served as a religious or administrative center for groups in the vicinity. This also shows that smaller towns in the Nippur-Adab area had more autonomy and agency within the settlement system (see Pollock 2001: 218). If this is the case, the site represents a significant starting point for exploring the urban landscape of Mesopotamia during the formative Early Uruk period.
- 10 Due to the disturbed nature of area HO and the limited finds from there, the presence or absence o (...)
49The surface observations and survey collections from Jemdat Zabi appear to hint at functional differentiation within the site, which may imply that more central organization and planning took place there. Functional variation is hinted at by the segregation of pyrotechnical and administrative activities. What could be discerned from the observed distribution patterns is that firing was concentrated in the southern and southeastern portion of the site (fig. 11) as evidenced by the fused clay sickle stacks (in areas A and D) and BRB stack (area E). The location of firing was likely restricted at the site due to the region’s prevailing western/northwestern winds. Clay sickles were found in all areas of the site except for areas C and HO10. BRBs were clearly concentrated in the western portion of the site in areas C and E (fig. 10B). This contrasts with the concentration of painted wares and Uruk bottles found in the east part of the site (fig. 10C-10D). Though it is not clear if this spatial patterning is meaningful, what seems to be clear is that BRBs were mainly associated with the area of the monumental buildings (areas C and E). BRBs were usually associated with redistributive central institutions (Johnson 1973; Nissen 1988; Pollock 2012), functioning as disposable receptacles for either dry rations, ready-to-consume foods (Pollock 2012: 156), or leavened bread (Chazan and Lehner 1990; Potts 2009; Goulder 2010). The absence of clay sickles in area C and the presence of BRBs and storage jars (see table 2), suggests that this area could have been focused on some form of storage and redistribution with perhaps a limited focus on pyrotechnical or agriculture production.
- 11 We were reluctant to compare the survey results from Jemdat Zabi to those of Abu Salabikh (Pollock (...)
50The differentiation of functional spaces at Jemdat Zabi and indeed at Tlul al Baqarat TB7,11 hints at a level of specialization in administrative and craft production activities at these sites. That being said, we cannot be certain if the two sites were organized in a similar manner. At TB7 we can surmise that the central mound was reserved for administrative/religious functions based on the presence of large-scale public architecture. Alongside the specialized administrative/religious function, it seems that large households were the loci of production, as evident from the activities that took place at building A. The excavators of TB7 mention the presence of 3,480 pottery slag fragments collected from the site’s surface without presenting their spatial distribution (Lippolis et al. 2019: 135). It would be interesting to see whether pyrotechnical activities took place across different loci at the site or if they had a limited distribution. In other words, how prevalent was household production across the site. Similar household centered modes of production and organization are seen further north at earlier Ubaid sites, such as Tell Abada (Jasim 1985, 2022), or contemporary sites in northern Mesopotamia such as at Hamoukar levels 3-1 (Abu Jayyab 2019) and at Tepe Gawra levels XII-IX (Rothman 2002).
- 12 The deflation that has impacted the site recently (see footnote 7) has exposed large ash lenses in (...)
51In the case of Jemdat Zabi surface evidence of pyrotechnical activities was overwhelmingly concentrated in one area (in the center of area D) perhaps suggesting the presence of specialized workshop/s at the site as opposed to individual households controlling pyrotechnical activities.12 This mode of organization is not unusual and could be seen at earlier Ubaid period sites further south such as Tell el ‘Uwaili (Vallet et al. 2020), al Ubaid and Eridu (Moore 2002), where there was a clear demarcation of pyrotechnical zones at these sites, with what appears to be a ceramic workshop excavated at Tell el ‘Uwaili (Vallet et al. 2020: 25).
52Divergent modes of organization between Jemdat Zabi and TB7 are still not confirmed and a broader understanding would require excavations at Jemdat Zabi; the initial observations set up some interesting question for future exploration. For example, could the divergent modes of organization be a marker of a Tigridian vs a Euphrates mode of social organization? What was the degree of political integration of these small/medium sized towns in an Early Uruk hierarchical system? Indeed, future excavations at the site will allow us to see how reliable our surface observations are and if such questions do merit further exploration.