Acknowledgements

We warmly thank first of all the Directorate of Antiquities of Iraqi Kurdistan, its former director, Mala Awat (Z. Abubaker), and its current director, K. Ali, who have provided us with the authorisation to carry out all our field missions.
In particular, we would like to thank the Soulaimaniah Antiquities Department and its director, K. Rasheed, who has always been highly supportive, for resolving all administrative, logistical and political problems. We also wish to thank his team, in particular our colleagues and friends who worked with us during eight campaigns as well as during the preparation of this article: A. Ameen, K. Raeuf and S.A. Saber.

We would also like to thank the newly created Directorate of Antiquities of Raparin and its director B. Baiz for his local diplomatic assistance.
We thank all the Kurdish archaeologists and scientists who have worked with us, including R. Sofy, J. Jameel, Z. Baiz, I. Mohammed of Nurridin and Ali of Qala Dizah, and all the drivers working at the Directorate of Antiquities who assisted us far more as guides than as drivers: Kak Aziz, Kak Abdulrahman, Kak Salekh, Kak Moussa and Kak Bahzaid.
Without the support of the Ministry of Foreign Affairs and the Commission for Excavations, these archaeological activities could not have taken place. We would like to acknowledge the former director of the Commission for Excavations, J.-F. Jarrige, and the present director, N. Grimal, for the confidence they have placed in a young team of researchers without permanent jobs. We would like to thank IFPO for its unfailing support for this project, in particular, M. Griesheimer and F. Alpi.
Finally, we thank the French Embassy and the French Consulate in Erbil, without which we would never have been able to maintain the mission during times of crisis. In particular, we would like to thank D. Gauer, F. Tissot, A. Guépratte, D. Mas, O. Chatelais, R. Oudjani, T. D’Anna, T. Vergon and J.-N. Baléo.
We also wish to acknowledge the institutions and research teams that have supported us financially and administratively: the CNRS through UMR 7041 (VEPMO and AnTeT research groups), the ANR “L’altérité culturelle du continent asiatique”, the Museum national d’Histoire naturelle, CNRS research units through UMR 7209 (AASPE) and UMR 7194 (HNHP). We are also grateful to the doctoral school ED112 at the University of Paris 1 Panthéon-Sorbonne. Finally, we must not forget INRAP, which allowed us to have two PAS contracts for M. Lemée to join us.
The mission would not have survived without all the researchers, doctoral students and students who were with us in the field: E. Fouache, E. Boeda, J.-J. Herr, R. Palermo, M. Ahmed, D. Arhan, Z. Hashemi, C. Caze, L. Arbez, C. Douché, I. Mesfin and C. Padovani.

1With the collaboration of C. Caze, R. al-Debs, C. Douché, A. Havé, A. Louchet, C. Verdellet and L. Walika

2Manuscript received on 7 March 2019, accepted on 8 September 2019

3The south-western foothills of the Zagros range, in Iraqi Kurdistan, have long been largely unexplored because it has been impossible for archaeologists to carry out fieldwork research in this area for more than half a century. The first excavations carried out in the 1960s and 1970s revealed the crucial importance of the region for the prehistoric periods. Many important sites, such as Shanidar (Solecki 1963), Jarmo (Braidwood et al. 1983) and Shemshara (Mortensen 1970) were discovered and excavated.

4Although for several decades this territory has been considered a peripheral area, it is, undoubtedly, a region that is fundamental for the understanding of the first Mesopotamian cultures. In fact, this area spawned the emergence, development and spread of major cultural entities, from the Neanderthal hunter-gatherers to the first complex societies characterised by the very large villages of the Late Chalcolithic period.

5The reopening of this region to archaeologists in 2011 has made it possible to resume research in this vast and little-known territory. Currently, there are more than 60 archaeological missions working in Iraqi Kurdistan. The surveys and excavations have produced a great deal of new data that provide a larger picture of these first prehistoric societies (fig. 1).

6Since 2012, the French archaeological mission in the Governorate of Soulaimaniah (MAFGS) has identified more than 366 archaeological sites in the provinces of Rania, Peshdar and Dukan (fig. 2), among which a third have been dated to the prehistoric period. The distribution of these prehistoric settlements, combined with rescue excavations carried out on six sites, is providing us with a first picture of the early cultures that developed along the south-western flanks of the Zagros. Local cultural entities, characterised by features already identified in other areas of Northern and Southern Mesopotamia, were established in a wide territory of foothills beyond the Mesopotamian steppes but retained close connections with the steppes and with the Iranian plateau.

The French archaeological mission in the Governorate of Soulaimaniah: goals and methods

7The history of archaeological research in Northern Iraq dates back to the 19^th century, during which the first archaeological discoveries were made by travellers and diplomats. However, it was during the 1940s that the Iraqi Directorate of Antiquities and Heritage conducted more extensive and systematic investigations to document and date archaeological sites throughout Iraq (Directorate General of Antiquities 1970). Later, in the 1950s and 1960s, the construction of large dams led to salvage surveys and rescue excavations to document the majority of archaeological sites that would be flooded (Husaini 1962; Abu al-Soof 1964, 1970). From 1960 to 2011, archaeological research was conducted in a piecemeal fashion because the geopolitical situation prevented work from being carried out regularly. It is for this reason that the Atlas of archaeological sitesin Iraq from the 1940s still provides the basis for our resumption of extensive survey work.

Goals of the project

9The MAFGS survey concentrates on the Governorate of Soulaimaniah, the eastern governorate of Iraqi Kurdistan. The survey project has two main objectives. The first is to implement a clear and exhaustive archaeological map that will be an essential instrument for the protection of the heritage of this region. The lack of knowledge with regard to ancient sites and their locations was, indeed, endangering the sites because the area is undergoing rapid economic development.

10The second objective is aimed at understanding the settlement patterns in the south-western foothills of the Zagros from the Lower Palaeolithic to the present day, and to reconstruct the territorial organisation specific to each culture. By defining the evolution of settlement patterns over time with detailed surveys and well-targeted excavations, it is now possible to provide a preliminary description of the development of human presence in this region. Obviously, the development of these cultural spaces depended upon economic, social, political and religious factors, whose specific characteristics and overall extent are yet to be defined locally, as well as their relations with other cultural entities.

Remote sensing and field survey methods

11The archaeological surveys have been the primary source of data for the MAFGS project. The survey of the Governorate of Soulaimaniah required a broad approach; thus, it combines the study of the archaeological sites themselves with landscape analysis in order to create the datasets allowing us to address major regional-scale questions about Mesopotamian history, as did previous research in the South (Adams 1981).

12The MAFGS survey methods are mainly based on the methods of British landscape archaeology established by T.J. Wilkinson (1982, 1989, 2003; Wilkinson and Tucker 1995) in Syrian and Iraqi Jazirah, in which the study of the sites (intra-sites) and the “off-site” landscapes (inter-sites) are combined for a general understanding of the territory. The intra-site study is based on a statistical analysis of a reasoned collection of material in order to understand the site’s internal organisation. The inter-site study is based on photography and satellite imagery to establish structures, roads and canals, which allow us to understand how sites are structured between them, forming a territory (Altaweel et al. 2012; Ur et al. 2013; Morandi Bonacossi and Iamoni 2015; Altaweel and Marsh 2016; Gavagnin et al. 2016; Giraud 2016, 2018; Mühl and Fassbinder 2016; Iamoni 2018; Mühl et al. 2018; Ur 2018).

13The MAFGS method combines the study of archives, old maps and published research with using satellite images from GIS. Systematic interviews with the inhabitants encountered in the villages were also carried out. The remote sensing of recent (QuickBird, GeoEye) and old (CORONA) satellite images enabled the sites or other archaeological features (canals, paths, roads) to be systematically located and digitalised in the GIS.

14Archaeological features are recognisable on images by the particular signature (Philip 2002; Ur 2010: 43-46), of their anthropogenic soils (differences of colour in the image). This way, we reach a 75% positive correlation between the possible sites visible on the images and the field assessment. However, interviewing the locals was indispensable to identify other sites where the images were unexploitable, such as in mountains or hilly landscapes.

15Each identified archaeological site was then divided into sub-areas according to their dimensions and morphology, with careful surface collection carried out in each area. At the qualitative level, sampling is a compromise between rapidity, effectiveness and accuracy. On a quantitative level, meticulous and broad collection enables maximal recording of the potential of the sites. All the sherds were described, photographed and sorted to analyse precisely the diagnostic elements: rims, bases, handles and decoration. So far, 11,344 diagnostic sherds have been collected as a typological reference for the chronological framework. All the sherds have been recorded, drawn and photographed, but only two-thirds have yet been dated.

16The use of this methodology has allowed us to multiply by five the number of sites known for the Rania and Peshdar plains.

Excavations

17Test trenches at six sites were excavated in order to define a local lithic and ceramic chrono-typology. These new stratified typologies (in progress) will help us to refine the dating of the local sites. Furthermore, they allow us to understand the development of human occupation from the Palaeolithic to the Late Chalcolithic in the region.

18These sites (fig. 2) were chosen according to their potential for establishing chrono-stratigraphy, and because of their small size. The last-but-not-least criterion was the risks threatening sites. More than half of the excavated sites were largely looted (Cave Mewe, no. 268), or threatened by urbanisation (Boskin, no. 60) or intensive agriculture (Halawezha/Bijian, no. 187 and Dargrdal, no. 266). The last site was chosen for its chronological importance and its stratigraphic sequence, unique in this regional context (Sarsyan, no. 77).

The Rania and Peshdar plains: now and in the past

19The MAFGS survey began in the north-eastern part of the Governorate of Soulaimaniah, around Lake Dukan. Administra-tively, this area belongs to the districts of Rania and Peshdar and the sub-districts of Bingird and Khadran in the Dukan district. Geographically, this is a coherent spatial unit that will allow a global reflection on the evolution of its occupation.

The Rania and Peshdar plains: geography, hydrology, climate and biogeography

21The hydrographic network of the region is mainly organised around the Lesser Zab. The river runs along the Peshdar Plain before flowing into Dukan Lake, via the gorge. In the Peshdar area, it receives contributions from several tributaries located at intervals from the mountains to the Iranian and Iraqi borders. These rivers literally divide the plain into vast north-south-facing terraces before joining the Lesser Zab that digs its bed from east to west. In the Rania Plain, the tributaries come from the east and west and currently flow directly into the lake. These last wadis wander in wide beds at the bottom of the
valleys leading to the Governorate of Erbil.

22The study area is located in the large semi-arid climatic zone (Walter and Lieth 1960) but the presence of mountainous massifs induces higher humidity with cold and wet winters and hot and dry summers. Data from Dukan’s weather station show that the mountainous location of the region gives it high thermal contrasts during the year and high rainfall (743.8 mm per year), mainly during the winter and spring seasons (Karim et al. 2014). However, as summers are particularly warm, the vegetation is quickly under water stress, but the evapotranspiration effect is compensated for by numerous rivers and springs, leading to the practice of dry farming (Walliser 2010). Moreover, given climatic and geological conditions, soils are particularly well developed, with a large organic matter content and a low salinity rate (Muhaimeed et al. 2014).

Palaeoenvironmental proxies in the long term

25The region’s palaeoenvironment is very little known. We will very quickly outline here a broad and more local paleoclimatic history.

26Regional climate evidence comes mainly via lake core analysis (pollen, sediment, chemicals and diatoms) from Iran (Van Zeist and Woldring 1978; Stevens et al. 2001), from Lake Zeribar and Lake Mirabad and from Lake Van and Lake Nar in Turkey (Roberts et al. 2001; Wicke et al. 2003; Dean et al. 2015). The records have a low temporal resolution (Marsh et al. 2018). Very recently, two speleothem analyses were carried out in Iraqi Kurdistan. The Geijkar Cave stalagmites (Piramagron Mountains) cover the period 430 BC-350 AD and that from 930 AD (Flohr et al. 2017). Too recent for our study, the positive results show drought peaks in an already rather hot and dry context. Some stalagmites from Shalaii Cave provide palaeoclimate proxies from 8500 BC to the present time (Marsh et al. 2018; Altaweel et al. 2019). These preliminary data are particularly promising. They are consistent with known regional data but provide a more local and, therefore, accurate view of Holocene climate fluctuations.

27In the Quaternary, the Pleistocene is marked by a series of glacial and interglacial phases. During the glaciations, temperatures drop and aridity increases. These periods are interrupted by interglacial periods, characterised by an increase in temperature and precipitation and involving significant river activity and the formation of terraces. The vegetation cover in the uplands and plains would probably have been steppes (Blanchet et al. 1997; Altaweel et al. 2012: 5-6).

28The Holocene, around 10,000 BP, after a final series of interstadial fluctuations of the Last Glacial Maximum (21,000-15,000 BP), the Tardiglacial (15,000-11,000 BP) and the Younger Dryas (12,800-11,700 BP), manifest themselves in a new climate regime throughout the Near East characterised by dry summers and wet winters (Blanchet et al. 1997). The vegetation becomes shrubby, with oaks and pistachio trees, and also herbaceous vegetation with wild cereals. Shalaii’s speleothems, lake cores and other environmental proxies from Southern and Northern Iraq (Marsh et al. 2018; Altaweel et al. 2019) show local climatic variations within the Holocene. The Early Holocene (9700-4000 BC) was characterised by wetter conditions with increased rainfall (Altaweel et al. 2019: 15-16) and the rivers incised the Pleistocene and Early Holocene terraces (Altaweel et al. 2012: 6). The end of the mid-Holocene (3000 BC) witnessed the shift to drier conditions and the climate remained stable for the rest of the Holocene until today; however, there are always drier or wetter phases (Altaweel et al. 2019: 16).

Formation of the Rania and Peshdar plains

30The region is mainly included in the folded sedimentary zone of the Zagros (Jassim and Goff 2006). This relief was created when the Paleo-Tehys and Neo-Tehys oceans closed as the Arabian and Eurasian plates collided during the Upper Miocene and Lower Pliocene. Thus, the folded landforms surrounding these plains are reliefs inherited from the differential erosion of Miocene and Pliocene sedimentary folds. The limestone, which is mainly from the Cretaceous period, has a karst structure. Many limestone pavements are visible, which explains the presence of caves. The southern landforms surrounding the Rania Plain consist of inverted landforms, anticline valleys and syncline ridges.

31The region, which has been geographically considered as two large plains, appears to be actually one plain formed in depressions filled by alluvial cones, river terraces and slope deposits, that is, colluvium formed from the Pleistocene to the Holocene.

Morphogenesis of the Rania Plain

32The morphogenesis of the area to the east of the Rania Plain has been subjected to climatic pulsations that have modified the nature of sediment inputs, mainly from large dejection cones. During calm phases, the plain was covered with clay, whereas during agitated phases, pebbles (of varying sizes) spread over a huge area. The plain does not seem to have an identifiable terrace or, perhaps, it has been “smoothed” by the influence of dejection cones or the significant impact that humans have had on it.

34This assumption is particularly valid for the eastern sector where conglomeratic deposits are based on clays with a clear limit. A period of biostasis materialised by the deposition of a clay level would have been disrupted and the clay covered by a massive spreading of pebbles (testimony to soil erosion and slope instability). We can identify the following successive phases:

• The clay deposition during a phase of biostasis and hot episodes (Late Tertiary), on which a hydrographic network is built;
• Still, in the biostasis phase, the hydrographic network cuts the clay level;
• The rhexistasis phase in Villafranchien, with a spreading of stones that completely covers the clay level;
• The current situation, in which the environment is said to be in “relative biostasis” after a succession of different cold phases (Quaternary glaciations).

Morphogenesis of the Peshdar Plain

35In the depression of the Peshdar Plain, different stages seem to have created levels of clearly identifiable terraces. Three main levels, shaped between 500,000 BP and 15,000 BP, have resulted from climatic pulsations that have generated significant sediment inputs. A Holocene terrace currently in formation seems to be emerging as the next low level. The data from the Bora Plain (Altaweel and Marsh 2016: 23-24) to the east of the Peshdar Plain help us to understand the stages of terrace shaping in the Peshdar Plain. The Bora Plain is a small alluvial plain lying on a tectonic zone where, “the main geological units consist mainly of Cretaceous period limestone with some conglomerates and sandstones especially in the foothill areas closest to the plain, and with Quaternary period alluvium deposited along the channel of the Lesser Zab. The Bora channels originating from the Zagros Mountains were likely originally carved out by the meandering of the Lesser Zab, with deposition of alluvium from both the Zab and the feeder channels originating from the Zagros Mountains” (Altaweel and Marsh 2016).

Prospects for further work

36Although the shaping of the surface formations of these two plains has resulted in different landforms and compositions, the main force in action in both cases is the climate. However, further and more refined studies are still needed for an even more complete understanding of the functioning of sediment inputs from these depressions. The implementation of drilling or coring would allow us to learn more about stratigraphy and confirm or modify our assumptions.

37Finally, although the paleoenvironmental evolution of the study area is mainly related to climate action, we are entitled to ask ourselves what influence the tectonic component may have in this region, given that the massif of Zagros is relatively young and that its over-reaction is still active today.

Archaeological data acquired by the MAFGS: 2012-2016

38The MAFGS project located and studied 366 sites throughout the region. For the moment, only two-thirds of them have been dated. Of these 244 dated sites, a small number have been occupied from the Palaeolithic to the Late Chalcolithic and six of them were excavated (fig. 2-3).

Palaeolithic periods

39The material recovered by the surveys reveals the extreme richness of the lithic assemblages. Fieldwork has recovered 75 of these. Although chrono-cultural determination is often made difficult by the lack of local chrono-stratigraphic contexts and distinctive typo-technical elements, typo-technological analyses of some lithic assemblages provide some chrono-cultural attributions: Middle Palaeolithic and Epipalaeolithic, with many assemblages dating to the first period. Sixteen sites could be dated to the Middle Palaeolithic (fig. 3-4).

Palaeolithic settlements

40The Palaeolithic lithic material was mainly located in lower and hilly zones and was discovered in or close to caves, rock shelters and several open-air sites in the lower and mid-mountain zones. There are no open-air sites located in the plains, which could partly be explained by the formation of terraces from the Pliocene to the Holocene.

41The lithic assemblages typical of the Lower Palaeolithic have not yet been identified. Upper Palaeolithic assemblages with Aurignacian/Baradostian sequences, broadly present in the Zagros region (Olszewski and Dibble 1994; Otte and Kozłowski 2007; Otte et al. 2007; Bordes and Shidrang 2009; Shidrang et al. 2016; Shidrang 2018), are so far absent. However, with regard to the Middle Palaeolithic, the assemblages are relatively well represented.

42We discovered several Middle Palaeolithic sites in the open-air locations of Ser Girdy (no. 78), Grtk Qala (no. 85), Hassana Rash Qalat (no. 87) and Geremka (no. 88), located in a valley southeast of the city of Qalat Dizah (Peshdar district) and characterised by the presence of Levallois reduction and products, including some points made in limestone. The others sites are in the mountains, either in caves or in rock shelters, as in the Besara Valley (no. 10, 91-92), the Dolmaran Valley (no. 130, 167), the Marbab Valley (no. 24-25) and inside Maqooq Mount (Sarsyan, no. 74, 77).

43The most significant assemblage belongs to the rock shelter site (Sarsyan, no. 77). Its lithic assemblage presents clear affinities with the Zagros Mousterian industries (Dibble 1993; Olszewski and Dibble 1993; Lindly 2005; Jaubert et al. 2006; Biglari 2007; Bazgir et al. 2014).

Sarsyan: a complete stratigraphy from the Middle Palaeolithic

45The different layers and deposits that were identified produced more than a thousand finds, mainly lithics and bones, throughout the trench. Archaeological remains from SU 1-5 mainly come from the underlying layer, SU 6, along with mixed material, such as historical and modern pottery, from more recent periods. The lithic artefacts of these layers as well as the one found on the surface present the same typo-technological traits as the underlying layer, SU 6 (see table 1).

46A first consideration of the lithic remains (still under study) found in all the layers was aimed at assessing the typo-technological traits of the lithic industry. As the quantity of material is large, the procedure focused on significant technical elements. Waste, chips, debris, most damaged fragments and undetermined pieces were not included in the analysis. The result of the lithic sorting carried out here was 178 pieces (table 1). The lithic industry is composed of numerous chert and limestone artefacts and the raw material was locally available in the form of different sized blocks. The state of preservation is relatively good, indicating minor post-depositional disturbance. The production results from two main debitage: Levallois for the main intended artefacts; and non-Levallois. There is also discoïdal and lamellar in small quantities (table 1). The lithic industry consists of two major categories: convergent, including Levallois points; and triangular and wide/elongated flakes. Convergent artefacts, mainly Levallois, are well represented and range from 26 mm to 56 mm long (table 1; fig. 6). Their contours are variable (elongated with small, wide-based points, symmetric and asymmetric pieces) and their profiles are straight or curved. These convergent pieces are unretouched or retouched on at least one lateral edge. For the most part, the assemblage also includes Levallois and non-Levallois flakes, wide and more or less elongated (table 1; fig. 6). Some of them are retouched. The lithic industry also contains some unretouched blades and bladelets. In the case of the unidirectional bladelets, the cores were formed by a bladelet reduction strategy. Mousterian points, retouched points and typo-scrapers (including side scrapers, double scrapers and convergent scrapers) dominate the typological group. A smaller number of retouched pieces and notches are present, along with one dihedral burin. The presence of some Upper Palaeolithic elements (bladelet reduction strategy, dihedral burin) may be intentional or, more likely, they are from the dismantled upper layer.

471. Retouched Levallois flake (side scraper); 2-3. Levallois flakes; 4. Levallois point; 5-7. Mousterian points; 8-9. Retouched points; 10-11. Fractured points; 12. Déjeté convergent flake; 13-16. Elongated flakes; 17-18. Levallois cores; 19. Biface

48The Sarsyan site provides a Mousterian lithic assemblage characterised by many Mousterian points and many typo-scrapers comparable with the Zagros Mousterian industries (Solecki and Solecki 1974, 1993; Dibble 1984, 1993; Dibble and Holdaway 1990; Biglari 2001; Jaubert et al. 2009; Beshkani 2018).

49The Zagros Mousterian industry is globally characterised by small retouched artefacts dominated by scrapers, including convergents, side and double scrapers, Mousterian points and elongated Levallois flaking.

50The Sarsyan rock shelter has proved to be an important Middle Palaeolithic site in this regional context. Indeed, this site represents the only Middle Palaeolithic sequence preserved (at least 50 cm deep) for the surveyed area. Obviously, we now need to excavate extensively to fully understand the distribution of the Middle Palaeolithic artefacts throughout the cave.

Early Neolithic

Distribution of Early Neolithic sites

52Evidence of Early Neolithic sites (fig. 7) in the Rania and Peshdar plains is difficult to identify on the basis of surface collections. With the MAFGS method used, only one site, Halawezha/Bijian, was probably identified as a PPN site. The other two sites, Qalat Saïd Ahmadan (Tsuneki et al. 2015, 2016) and Shemshara (Mortensen 1970; Matthews et al. 2016), were dated from this period following the excavation of their oldest levels.

53Concerning Shemshara, Mortensen’s excavations, as early as 1970, have established levels of occupation dating to the Neolithic, more precisely to the Hassuna period (Mortensen 1970). The resumption of excavation of these earliest levels (Matthews et al. 2016) clearly indicates a PPN occupation, with a radiocarbon sample of ca. 7300-7200 BC (Matthews et al. 2016) lying on the last layer of occupation excavated in trenches 1 and 2. The lithic assemblage is mainly composed of tools and flakes in obsidian and chert, specifically, several Cayönü tools and some sickle blades and diagonal-ended bladelets in both materials (Matthews et al. 2016).

54Similarly, at Qalat Saïd Ahmadan, the last levels before the natural substrate, SU 5-6 of trench B, are dated from 7600-7300 BC (radiocarbon date). The lithic of these layers is far richer and more varied than the other upper layer. There is a great variety of flint and obsidian tools including sickle
elements, scrapers, serrated blades and points.

55During the surveys carried out by the MAFGS, only Halawezha/Bijian (fig. 8) was spotted in fortuitous circumstances. This can be explained in different ways. First, the method used by the MAFGS does not lend itself well to the search for this type of site, which is too intangible with thin and highly fragmented lithic material. In addition, CORONA imagery does not give a clear view of this type of site. Furthermore, the history of the morphogenesis of the plain has barely stabilised: the alluvial terraces of the Early Holocene are still being set up. Finally, in the case of reoccupied sites, PPN levels can be buried below metres of more recently occupied layers, as at Qalat Saïd Ahmadan and Shemshara.

56Nevertheless, 75 lithic assemblages have been collected in the region. Half of this material is still under study and we can very well imagine that other sites from this period exist. On the surface, Halawezha/Bijian showed an abundance of ground stone tools, some features of lithic industry and an absence of pottery, suggesting to us an Early Neolithic occupation in the Bora Plain (Peshdar Plain) on this open-air site. Its location is on one of the middle terraces bordering the Lesser Zab, which indicates a very old occupation of this small plain (Altaweel and Marsh 2016).

Halawezha/Bijian site: evidence for Early Neolithic

58In trench 1, a semi-circular structure 2 m wide (fig. 8), consisting of large river pebbles around 0.2 m to 0.3 m in size, extended beyond the trench. Very close and to the west, a small hearth along with a few non-diagnostic sherds was identified. A flat layer of similar pebbles was interpreted as an area of circulation or a floor (courtyard?) directly related to the other features. Very close to these features, in the eastern extremity of the trench, an obsidian blade core (fig. 9.13) was discovered in the clay but was not associated with any archaeological feature.

CAD F. Pichon

The Pottery Neolithic: 7^th-5^th millennium BC

60For the period from the 7^th to the 5^th millennium BC, 27 sites provided material. More precisely, 9 were occupied in the Hassuna period, 3 in the Samarra period, 4 in the Halaf period and 10 in the Ubaid period (fig. 3).

Distribution and typology of sites: Hassuna to Ubaid period

61With regard to the 7^th and early 6^th millennia (fig. 7), the pottery style of the Pottery Neolithic, or “Early Chalcolithic”, characterised by the so-called Hassuna and Samarra “cultures”, has been identified in the MAFGS region based on a small quantity of combed and painted sherds from a number of sites: Quruja (no. 16), Tle Tell (no. 26), Qalat Saïd Ahmadan (no. 32; Tsuneki et al. 2015, 2016), Du Grdan (no. 48), Dinka (no. 53), Boskin (no. 60), Shemshara (no. 90; Mortensen 1970, Matthews et al. 2016), Dari Zewe (no. 173) and Dargrdal (no. 266). Despite the limited amount of collected samples, the region appears to have been characterised by reciprocal influences between the different traditions, especially with regard to the painted wares. There are clear parallels with Hakemi Use and Yarim Tepe (Merpert and Munchaev 1993; Tekin 2005).

62During the Late Pottery Neolithic, Halaf pottery is also in little evidence at four sites: Qalat Saïd Ahmadan (no. 32), Girdy Dema (no. 58), Boskin (no. 60) and Xwchaw Poke Saru (no. 112). Its morpho-stylistic typology is close to the collections from Nineveh, Sabi Abyad, Tell Hassan, Tell Halaf and Yarim Tepe III (Von Oppenheim 1943; Fiorina 1987; Merpert and Munchaev 1993; Gut 1995; Nieuwenhuyse 2007), even if the Early Halaf (or Hassuna III) influence is conspicuously absent from the surveyed area (as with the SSP so far; Mühl and Nieuwenhuyse 2016; Nieuwenhuyse 2018: 49).

64For the Ubaid period, we distinguish nine sites: Malan Girdy (no. 6), Ibrahim Katshal (no. 13), Aliyawa (no. 20), Blil (no. 21), Saxima (no. 22) Qalat Saïd Ahmadan (no. 32), Pirota Sour (no. 37), Boskin (no. 60) and Dargrdal (no. 266).

65The Ubaid ceramics fall outside of the Pottery Neolithic proper and represent rather a Middle Chalcolithic marker. They have a quite ordinary look, with varying shapes: hemispherical; bowls with an inturned rim; middle-sized jars with beaded or flaring rims. Even if black-on-buff painted decorations belong to a widely shared Ubaid repertoire, in the Zagros Piedmont some decorations (chequered motifs, or lozenges decorated by a central point) are very much indebted to previous Halaf traditions. At the same time, as already observed in the Shahrizor Plain and at Logardan (Altaweel et al. 2012; Baldi 2016), Ubaid pottery from the Rania and Peshdar plains is much more closely related to sites in Central Mesopotamia and in the Hamrin Valley (Jasim 1985) than to the Zammar and Mosul regions in the north (Tobler 1950; Gavagnin et al. 2016).

66We distinguished two types of shape for sites belonging to this part of the Pottery Neolithic period. The first type is represented by mounds with extensive surfaces (1-3 ha) having a gentle slope and low elevation (about 5 m), located in the plain near a wadi or canal. On the CORONA images their signatures are dark, indicating a more humid and organic subsoil (remains of architecture with a light material structure or mudbrick), whereas whitish signatures indicate a dry subsoil and more densely built structures (probably stone architecture). These characteristics correspond to a number of sites: Ibrahim Katshal (no. 13), Saxima (no. 22), Dari Zewe (no. 173), Du Grdan 2 (no. 48), Dargrdal (no. 266) and Xwchaw Poke Saru (no. 112). They do not particularly correspond to a specific period within this Neolithic period. These are small villages that were occupied for a shorter period than other types of site.

67Concerning the second type, the sites correspond to small tells with a high mound associated with a lower town. These tells are attributed to the Chalcolithic, but at their bases we found earlier layers of Pottery Neolithic, such as at Boskin (no. 60), Qalat Saïd Ahmadan (no. 32), Tle Tell (no. 26), Shemshara (no. 90), Malan Girdy (no. 6), Pirota Sour (no. 37), Quruja (no. 16), Dinka (no. 53) and Blil (no. 21).

Dargrdal: evidence for Hassuna and Ubaid 3

68The Dargrdal corresponds to type 1 (fig. 10). It is a small hill measuring 3 ha.

70However, human occupation is obvious in the entire trench because several pottery sherds were found in each of the smaller trenches.

71As the bedrock was not reached, the description of the SU begins at the top and extends to the deepest level reached.

72Between 0.2 m (at the top of the hill) and 0.7 m (at the bottom) under the topsoil, a dark orange-brown soft clay sediment containing carbonate nodules was uncovered (SU 2, SD I-II and V-VI; SU 16, SD III-IV). There are many indications of bioturbation here, and this may have concealed some archaeological features. Only one hearth has been identified in this level (FP 13 in SD I). The presence of terracotta nodules in the fill could indicate the existence of a cover.

73In SD III and IV, only SU 19 (under SU 16), in which many rolled sherds and fractured lithics were found, indicates the occupation of the area. The compactness of this area, as well as the rolled appearance, high fragmentation and quantity of the sherds, suggest that it is a level of human circulation, such as a courtyard, rather than a level of debris. In any case, no domestic or artisanal activities appear to have taken place in this part of the trench. The pottery sherds belong exclusively to the Ubaid 3 period. SU 20-21 and 35 consist of very compact brown sediment in which no archaeological material or features were found.

74In SD I, below the carbonaceous level SU 2, and cut off by FP 13, a second oven was discovered (FP 14) in SU 4, which was composed of yellow-brown loose clay sediment. This level is rich in pottery fragments and includes numerous snail shells. Beneath SU 4, SU 24, a yellowish clay sediment containing many inclusions of carbonate nodules, is more compact. Several noteworthy artefacts were found in this level: a stone mortar, pottery, flint fragments, fauna and terracotta fragments. The presence of a posthole (PH 22) and a small pit (SU 25) indicate that this is another stratigraphic layer, although the difference compared with SU 4 is slight. In these two layers, SU 4 and 24, the pottery fragments consist of heterogeneous material dated to the Hassuna and Ubaid 3 periods (fig. 11). In conjunction with the mixed material, this indicates a gap in occupation that is difficult to detect in the stratigraphy.

CAD J.S. Baldi, A. Amin, A. Havé, M.-A. Pot, M. Mura

75SU 26, beneath SU 24, consists of loose, yellow-brown clay sediment containing frequent inclusions of carbonate nodules. Some sherds were found in low concentration but are not associated with any archaeological feature. However, they all belong to the Hassuna period (fig. 11). SU 26 is about 0.2 m thick and covers SU 29, in which no material or archaeological structures were found. The excavation stopped at this level.

76To conclude, the Dargrdal site was occupied at least from the Neolithic, during the Hassuna period. Following this, Ubaid 3 people settled in the same place. The mixture of the pottery material in SU 24 and SU 4 in SD I can be explained by the erosion of the earlier Hassuna site, which occurred during the Ubaid occupation. However, it is possible that in another part of the hill, especially at its base, intermediate levels between Hassuna and Ubaid 3 exist that are covered by colluvium levels. Indeed, it is not certain that SU 20-21 and 35 in SD III-IV and SU 29 in SD I are equivalent because Hassuna levels or the natural soil in SD III-IV were not reached. On the top of the hill, intermediate levels between Hassuna and Ubaid 3 could have been eroded. In any case, this gap in chronology needs to be explained.

Test trench at Boskin: a complete stratigraphy from the Hassuna period to the Late Chalcolithic 2

CAD J.S. Baldi, M.-A. Pot, L. Walika

78The site is badly damaged; the remains of the mound measure 80 m long, 140 m wide and 10 m high. A 3 m x 10 m test trench was opened on the eastern side of the hill, but the width of the trench was progressively adapted to the topography of the slope for a total surface area of about 70 m², excavated on eight levels (fig. 12).

79Level 8 consists of an extremely hard clay bed (SU 38) covered by a layer of sandy alluvial soil (SU 37) 20-35 cm thick. Both layers are completely sterile and represent the non-archaeological base of the site. The first traces of human presence were recognised in a thin burnt layer (SU 36), disturbed by a large and deep surface pit (SU 34) cut on the slope of the tell through levels 6-7 and the upper deposit of level 8.

80Level 7 displays clear traces of the first major development of the site with extensive architectural works. A huge mudbrick terrace wall (SU 32) was covered by a floor 4 cm thick paved with Hassuna sherds (SU 33).

81In level 6, above some burnt ashy layers that mark the separation from the previous stage, it has been possible to identify the first of a series of workshops for the firing of pottery. A large circular potter’s kiln (SU 27-K7), located in the southern sector of the trench, was determined to have been a two-storey updraught structure whose lower part, the firing chamber was conserved and visible in section. The surrounding work area consisted of a floor with Halaf sherds, pieces of slag and some potter’s tools such as obsidian blades, bone polishers and spherical pebbles.

82In level 5, the firing chambers of the kilns (SU 22-K3, SU 23-K4, SU 24-K5 and SU 26-K6) were not cut through earlier layers of fill, but rather built up within a huge mudbrick platform (SU 25), which was probably used both as a work bench and as a space for drying vessels.

83Level 4 is represented by two later kilns (SU 18-K1 and SU 19-K2), cut through the destruction layer (SU 021, containing sherds, ashes and crushed mudbrick) of the firing chamber of the level 5 kilns. The surrounding workshop area was occupied by a mudbrick platform (SU 20).

84Level 3 marks a reorganisation of the site. Built in mudbrick upon a burnt layer (SU 17), two well-preserved walls are part of one coherent Ubaid architectural complex.

85In level 2, after an abandonment of this sector of the site, the entire area was levelled and reorganised. The main structure of this new architectural phase is a huge stone terrace wall (SU 012), lying on a pebble floor 8 cm thick.

86In level 1, the first constructions built upon the stone retaining wall of level 2 were erased by some later Late Chalcolithic 2 cist graves. Four burials of this type (05-T1, 06-T2, 07-T3 and 011-T4) have been identified in level 1. Except for one structure (T4), which is north-south oriented and contains stones taken from the terrace wall (SU 12), the other cist graves were lined with stones and roughly oriented east-west.

87In accordance with the stratigraphy, four distinct pottery phases (fig. 13) have been identified:

• Levels 8-7 (Hassuna–Samarra traditions);
• Levels 6-4 (Halaf and Early Ubaid);
• Levels 3 (Ubaid 3 phase);
• Levels 2-1 (Late Chalcolithic 2).

The Chalcolithic

88With regard to the Chalcolithic, the number of occupied sites is higher and 59 sites have been recorded (fig. 14): 6 from the Halaf-Ubaid transition-Late Chalcolithic 1, 21 for the Late Chalcolithic 1, 8 for the Late Chalcolithic 1-2 transition and 24 for the Late Chalcolithic 3-4 transitions.

Distribution of Chalcolithic sites

89The late and post-Ubaid phase is particularly interesting because, until recently, the Late Chalcolithic ceramic typology of Northeast Mesopotamia during the Late Chalcolithic 1 (4600-4200 BC) was identified with the assemblages from the Mosul region and the Syrian Jazirah, at Yarim Tepe III, Gawra, Hamoukar and Tell Arpachiyah (Abu Jayyab 2012). The dramatic simplification of the decorative motifs, as well as the fewer and fewer quantities of painted pottery, reveal the decline of Ubaid traditions. The pottery was fired in poorly oxidised conditions and became increasingly homogeneous, with serially produced Coba bowls, the first samples of potters’ marks and chaff-faced wares (Baldi and Abu Jayyab 2012). In this phase, it is possible for the first time to detect a tendency towards the regionalisation of pottery in Northern Mesopotamia (Baldi 2016: fig. 1). At al-Hawa, Hamoukar and in the Khabur basin, this process depended upon the reorganisation of a system of villages that were fairly homogeneous in size and devoid of spatial hierarchies into a more territorially ranked model, with some areas coming under the influence of major centres (Wilkinson and Tucker 1995; Ur 2010). In the Rania Plain, the regionalisation process can be recognised because of some peculiarities in the pottery: there is no trace of sprigware or painted animal motifs (i.e. birds or scorpions). Moreover, the only type of Coba bowl is the V-shaped wide flower pot, although at Gawra, another rounded type has also been attested (Baldi 2012). All these aspects are also shared by the Shahrizor area at Tell Begun and Gurga Chiya (Nieuwenhuyse et al. 2016: fig. 26.11; Wengrow et al. 2016: fig. 12.26-27). In this sense, the Late Chalcolithic 1 continues the late prehistoric close connection between the Hamrin region, the Halabja region and the Rania Plain.

90The Late Chalcolithic 2 (4200-3800 BC) is characterised by the so-called “Gawra material culture” throughout Northern Mesopotamia. The various meso-regions tend to merge into two macro-regions (Baldi 2016: fig. 2-3). The Khabur basin, the Upper Tigris and Northern Iraq belong to one single area. The assemblage around Lake Dukan provides evidence for this Northeast Mesopotamian koine, as seen at Gawra XIA-XI, Brak CH13, Gawra A from the Nineveh period, Khirbet Hatara, Musharifa, Begun and Gurga Chiya (Numoto 1987; Oates 1987; Gut 1995; Fiorina 2001; Rothman 2002; Nieuwenhuyse et al. 2016; Wengrow et al. 2016). Chaff-faced materials have become ubiquitous, as have some common types of artefact (jars with interior-angled rims and bowls with inwardly turned bevelled rims) whose presence continues at the beginning of the Late Chalcolithic 3.

91Compared with the slow increase in population and number of sites between the 7^th and the early 5^th millennium, the beginning of the Late Chalcolithic, around the middle of the 5^th millennium BC, represents a major transition. The landscape was much more densely populated, with an important number of new settlements (e.g. Waranga Saru, no. 15; Sofian Kawlan, no. 43; Dinka, no. 53). Nevertheless, virtually all the Ubaid sites continued to be occupied, and settlements appear to have been more and more clustered in the plains. The sites had become characterised by a tell with a citadel and a lower town.

93The end of the Chalcolithic is represented by the Late Chalcolithic 5, characterised by the peak of the Uruk colonial presence in Northwestern Mesopotamia and by the decline of the local pottery traditions (Helwing 2002, 2005; Oates 2002). The materials used to make ceramics demonstrate this very clearly. Chaff-faced wares represent a minority of the assemblage and many local shapes were made with mineral-
tempered materials from Southern Mesopotamia. In addition to artefacts from Southern Uruk in evidence since the Late Chalcolithic 3-4 (pierced lugs, bevelled-rim bowls, inverted-rim jars with incised shoulders; Gut 1995: pl. 59-68), some Late Uruk ceramic features (such as drooping spouts and reserved slip decoration) are typical of this phase (Sürenhagen 1978: pl. 12.76, 17, 102).

95Overall, Late Chalcolithic 1-5 pottery indicates different trends of regionalisation between the end of a cultural koine (Ubaid) and the end of another broad cultural community (Uruk; Stein and Özbal 2007). If the Ubaid phase marks the beginning of a clustered pattern for settlements in the plain, the Uruk phase in Southern Mesopotamia appears to have had a limited regional impact. The large number of Late Chalcolithic 1-2 sites contrasts with their decrease during the Late Chalcolithic 3-5, as well as with the scarcity of material from Uruk in Southern Mesopotamia. As recently suggested for the area around Halabja (Altaweel et al. 2012: 24), the southern colonial presence in the Zagros Piedmont shows a rather dispersed settlement pattern in this area, probably crossed by roads to Iran (Tomé et al. 2016: 431), but not occupied by important southern colonies.

Mewe Cave excavations

96The cave, situated in a narrow mountain valley in the Peshdar Plain, is elongated, oriented east-west and measures 43 m long by 38 m wide. It has an overall height of approximately 5 m (Mashkour et al. 2017). It is divided by a huge central stalactite merged with the rock upheaval in the centre of the cave. This has produced two rooms, one in the front and one in the rear, joined by a corridor 1-2 m wide. Room 1, near the entrance, is sub-rectangular and measures 9 m x 5 m. Room 2 is elongated and measures 26 m x 3 m x 8 m (fig. 15). Three large stalactites are visible in this part.

97The looting of the cave was extensive, so we used the holes created by the looting to make clear sections in order to understand the stratigraphic layers of the cave. The sedimentation is homogenous in all parts of the cave. Considering the existing slope in the cave, we shall begin the description from the rear of the cave, moving towards the entrance.

98Room 2 has a stratigraphy of 2 m comprising, from the substratum to the top:

• A natural sedimentation of between 5 cm and 20 cm, containing no artefacts, that lies over the bedrock (SU 23, SD IV; SU 39, SD II; fig. 15).
• This sedimentation is covered by a red clay layer of approximately 20 cm that contains many sherds (SU 22, SD IV; SU 34, SD II).
• This lay, in turn, is covered by a dark brown layer of approximately 20 cm, (SU 21, SD IV; SU 29, SD II). In some places, it is pierced by pits (SU 24-25, SD IV; SU 34-35, SD II). These pits have produced a relatively important number of artefacts: fragments of storage jars attributed to the Bronze Age or the Late Chalcolithic; animal bones (sheep/goat, Ovis aries/Capra hircus); rodents (Muridae); and molluscs (Hellicidae – Eobania vermiculata; SU 24, SD IV).
• Finally, there is a series of heated, ashy, chalky layers in approximately one metre of deposit, forming several similar sequences. These are most probably related to the cleansing of the cave by burning when it was used as a sheepfold up to recent times.

99Room 1 contains a maximum of 60 cm of sedimentation:

• A brownish-grey ashy layer approximately 20 cm deep containing very few artefacts (SU 5-7, SD I), above the rock in the northern part of the investigated area.
• A stratified ashy sedimentation (SU 4, SD I), covering the previous component and the bedrock in the southern part of the investigated area. In this part, the sediment is trapped in the rock crevices. It contained pottery sherds, five lithic artefacts (flakes, core fragment and blade) and a Chalcolithic/Neolithic ground stone made of an exogenous sandstone.
• A sequence similar to the fourth stratigraphic sequence described above for investigated area 2, in the room 2.
• Finally, a thick layer of contemporary dung (between 10 and 25 cm) covering the whole.

100Most of the pottery fragments consist of storage pottery that belongs to the Late Chalcolithic/Uruk phase and to the Bronze Age, with a few pieces belonging to Late Antiquity. It should also be noted that a Byzantine coin was found in room 1. This indicates the cave continued to be sporadically occupied in the latter part of Late Antiquity.

101Concerning the Chalcolithic period, the Mewe ceramic materials (fig. 16) indicate that over a relatively restricted period, at the beginning of the 4^th millennium BC, the cave was occupied by culturally different human groups. They used the second room as a storage area.

• People from Northern Mesopotamia, using local Late Chalcolithic 2 ceramics.
• Uruk people from Southern Mesopotamia, using Early Uruk vessels.

The MAFGS region and its connection with neighbouring areas from Middle Palaeolithic to Late Chalcolithic

102The first analyses of the data collected allow us to begin to highlight the evolution of the occupation of the Rania and Peshdar plains. Subsequently, further analyses (dating, spatial analyses) will make it possible to establish broader territorial organisational models, enabling a more precise understanding of the history of occupation in the Southern Zagros and in Upper Mesopotamia over the long term.

103For the prehistoric periods, the figure 3 shows the trends in the evolution of occupation in the Rania and Peshdar plains. We are only talking about trends here because we are in the preliminary stages of our analyses. To go further, we will have to finish dating 4600 sherds collected last season and find a way to weight the “black hole” that is Dokan Lake. Indeed, since the lake flooded a large part of the Rania Plain, we have a spatial hole in the data. The “lake data” are mainly bibliographic information and, possibly, obsolete surveys from the 1940s and 1960s focusing on the most prominent and well-known sites. No systematic survey was undertaken during this period. Future analyses will, therefore, have to take this issue into account and find a way to weight the effect of the lake for the analysis of a territorial organisation.

106In the following discussion, we will review chronologically the relationships that the local cultures of the Rania and Peshdar plains maintained with both close and distant neighbouring regions.

Palaeolithic periods

108As already mentioned, if we have unmistakable evidence from the Sarsyan site of an ancient occupation of the region at least since the Late Middle Palaeolithic, there is no evidence yet of Upper Palaeolithic occupation, although such sequences are widely represented throughout the Zagros (Hole and Flannery 1967; Smith 1986; Olszewski 1993a, 1993b; Olszewski and Dibble 1994, 2006; Otte and Kozłowski 2004, 2007; Jaubert et al. 2006; Bordes and Shidrang 2009; Shidrang et al. 2016; Shidrang 2018).

Early Neolithic periods

111Nevertheless, the sites are preferably located in Pleistocene and Early Holocene terraces bordering rivers or tributaries. This settlement patterns corresponds to the typical organisation of Early Neolithic settlements in the Middle Euphrates area in the Syrian steppes (Geyer and Besançon 1997), in the Northern Levant (Geyer et al. 2019: 106) and in Iranian fertile plains such as the Sheikh-Abad and Jani (Matthews et al. 2013). A more intensive survey of these terraces should be carried out to uncover any unidentified archaeological sites from this period.

112In terms of lithic industry, the assemblages at each site seem to belong to the same group. The general traits of the assemblages evoke the M’lefaatian chipped stone tradition, as defined by Kozłowski (1999) and are present in the region at some Neolithic sites, such as M’lefaat, Shemshara (Matthews et al. 2016) and Bestantsur (Matthews et al. 2016, 2018). Cayönü tools are one of the most important diagnostic aids found in some Neolithic sites in Southeast Anatolia, Upper Mesopotamia and the Central Zagros (Nishiaki and Darabi 2018) and are dated to the later 8^th and 7^th millennia (calibrated BC; Matthews et al. 2016, 2018). Finally, the provenance analysis of chert pieces and imported obsidian pieces could be promising for the understanding of the Early Neolithic community and its long-distance network for obtaining raw materials.

7^th-5^th millennium BC

113In the following period (7^th-5^th millennium BC), when pottery appeared in Mesopotamia, the Neolithic villages thrived. Understanding the boundaries and interactions between the cultures of Hassuna, Samarra and, later, Halaf and Ubaid, is complex because of the scarcity of data for this area.

116In the mid and late 6^th millennium BC, the pottery landscape consists essentially of the Ubaid 3 repertoire and presents close parallels with Central Mesopotamia; the morpho-stylistic typology is, therefore, closer to Tell Abada I (Jasim 1985) than to Tepe Gawra XVI-XIII (Tobler 1950). Indeed, this trend continues a cultural tendency already observed, beginning with the first appearance of the Ubaid traditions, when Halaf-Ubaid transitional assemblages from the Zagros Piedmont in Northeast Mesopotamian were characterised by decorative patterns very similar to the central Mesopotamian Choga Mami transitional style (see above).

117Finally, with regard to the Hassuna and Samarra “cultures”, their co-existence at several sites such as Boskin and Dargrdal is not surprising. Indeed, Samarra decorated pottery was considered earlier to have been a “luxury component” of Hassuna assemblages (Leslie 1952). After the excavations at Choga Mami (Oates 1969, 1972), it has been examined as a culture in itself, with many samples assumed to be imported products appearing outside the supposed Samarra “heartland”, for example, at Shemshara (Mortensen 1970: 119). However, the hybridisations observed at Boskin and Dargrdal also imply contacts with Early Halaf and Ubaid traditions, some samples displaying hybrid traits both with clay bodies and stylistic decoration. Although hybrid traits combining Samarra and Early Ubaid have been recorded in Central Iraq and the Halabja area for the so-called “Choga Mami transitional style”, they were previously unknown in Northern Mesopotamia and do not represent simply a clue for a phenomenon of cultural contact between co-existing entities, but are evidence for local technical continuity, given that Samarra, Halaf and Ubaid pottery continued to be fabricated with the same clay bodies over a long period. Moreover, hybridisations observed at Boskin and Dargrdal raise questions concerning the relationships and chronology of the Samarra, Halaf and Ubaid “cultures”, which may have had simultaneous relationships with local traditions of the Zagros area (such as the so-called “Dalma ware”; Henrickson 1986).

119Finally, between the 7^th and early 6^th millennia, the sites are quite scattered and they are slow to increase in number (fig. 7, lower). This appears to indicate a gradual but continuous demographic growth in association with the proliferation of small sites. With regard to their shape, the sites belong to the first type. Concerning the Late Pottery Neolithic and the Halaf-Ubaid period, the increase in number and size of the sites should be interpreted as a process of slow but radical change in the economy and subsistence strategies, with less and less mobile populations and, therefore, more visible evidence in the archaeological record.

Late Chalcolithic

121For the Chalcolithic, and the Ubaid and Uruk periods, Boskin, Dargrdal and Mewe Cave thus provide a small encyclopaedia of the pottery of Late Prehistory in the Rania and Peshdar plains. The techno-stylistic interweaving between different traditions, as well as the co-existence of culturally distinct traditions on small sites, suggest the importance, from very ancient times, of communication routes crossing the mountains towards Iran. In the same way, the existence of a “corridor of the Zagros” running along the piedmont is evident, so that pottery features so far considered typical of Central Mesopotamia (such as the Choga Mami mixture of Samarra and Ubaid) were, in fact, widespread even in areas such as the Rania Plain, which is located several hundred kilometres to the north.

122The large majority of Late Chalcolithic settlements are located in the plain and close to small streams. It is still too early to suggest a precise reconstruction of land management, but despite the absence of a clear territorial hierarchy, throughout the Chalcolithic there is a clear trend towards a decrease in the homogeneity of the sites, both in terms of their size and distribution. The progressive formation of clusters of villages between the end of the 5^th millennium and the beginning of the 4^th millennium indicates an increase in population and, presumably, in the complexity of the management of common resources. Later, even if in the Rania and Peshdar plains several local sites show traces of a Southern Uruk presence, the decrease in the number of settlements is a phenomenon that has yet to be explained. In any case, this tendency affects the entire area east of the Tigris, as also demonstrated by the results of the LoNAP mission (see Morandi et al. in this volume). It would be tempting to interpret this apparent depopulation as a discontinuous evolution of social complexity. In fact, after a trend towards the creation of rural agglomerations between the Ubaid period and the Late Chalcolithic 2, it could be that the population tended to concentrate mainly in large proto-urban centres and in nearby villages, with few settlements to control in the more distant countryside.

123In the future, research in the Zagros Piedmont will attempt to answer questions about the Prehistory of Iraqi Kurdistan as a tale of emerging social complexity in a region that was a main cultural crossroads.