The Early Prehistory of Wadi Faynan, Southern Jordan: Archaeological Survey of Wadis Faynan, Ghuwayr and Al Bustan and Evaluation of the Pre-Pottery Neolithic A Site of WF16

By Bill Finlayson and Steven Mithen

This edited volume provides a full report on the Pre-Pottery Neolithic A site of WF16, southern Jordan. Very few sites of PPNA date have been excavated using modern methods, so this report makes a very significant contribution to our understanding of this period. Excavations have shown that the site contains a highly dynamic use of architecture, and the faunal assemblage reveals new information on the processes that lead to the domestication of the goat.

• Language: English
• Publisher: Oxbow Books
• Release date: Mar 22, 2007
• ISBN: 9781782975052

Bill Finlayson

Bill Finlayson is Director of the Council for British Research in the Levant and works on the early Neolithic of southwest Asia, primarily undertaking fieldwork in southern Jordan at a number of Pre-Pottery Neolithic sites. He has been working to promote southwest Asian Neolithic heritage as an asset for local communities and tourism.

Book preview

1. The Dana-Faynan-Ghuwayr Early Prehistory Project

Bill Finlayson and Steven Mithen

1.1 Origin of the project

The Dana-Faynan-Ghuwayr Early Prehistory Project (DFGEP) was established in 1997 and involved four seasons of survey and excavation in its study region located in southern Jordan (Figure 1.1). Four additional short-term field visits were made to complete the acquisition of data. Post-excavation studies lasted until the spring of 2004. Following a series of interim reports (Finlayson and Mithen 1998; Mithen and Finlayson 2000; Mithen et al. 2000; Finlayson et al. 2000; Finlayson et al. 2001), this monograph provides the final project report.

The project originated from a research initiative by the former British Institute at Amman for Archaeology and History (BIAAH). In 1993, William Lancaster, its then director, was commissioned to undertake a study of the resident Bedouin in Wadis Dana and Faynan in southern Jordan by the Royal Society for the Conservation of Nature (RSCN). This was part of the RSCN’s preparation for the creation of a nature reserve in Wadi Dana, a project funded by the World Bank. The RSCN had recognised the threat to the archaeological heritage in the wadis from economic development and grave robbing. These had already disturbed significant elements of the Roman/Byzantine field system and cemetery in Wadi Faynan.

BIAAH began a survey and mapping project, initially funded by the Foreign and Commonwealth Office (McQuitty 1998). This involved the production of a photogrammetric map and a detailed survey of the most threatened sites (Barnes et al. 1995). A pilot field-walking survey was also conducted (Ruben et al. 1997). An agreement was made with the RSCN to share their camp in Wadi Faynan, and funds were obtained to improve its facilities to allow the accommodation of archaeological field teams. On the basis of their preliminary work, BIAAH undertook rescue excavations at the Byzantine South Cemetery site in collaboration with Yarmouk University (Findlater et al. 1998) and sought to interest British universities into conducting further archaeological research in Wadis Dana and Faynan.

By 1996, BIAAH had been subsumed within the Council for British Research in the Levant (CBRL), and three new UK-based archaeological projects in Wadi Faynan had been established. Professor Graeme Barker, University of Leicester, had convened an interdisciplinary team of archaeologists and geographers to examine landscape development within Wadi Faynan, with a focus on the Roman/Byzantine field system (Barker et al. 1997, 1998, 1999, 2000). A second initiative was led by Dr Katherine Wright, Institute of Archaeology, University College London, entitled the Wadi Faynan Fourth and Third Millennium Project. This was concerned with the excavation of a Bronze Age settlement (WF100) located during the BIAAH survey (Barnes et al. 1995). This project sought to examine issues relating to Levantine–Egyptian links and the development of complex societies (Wright et al. 1998). Thirdly, Dr Freeman of Liverpool University had begun a building survey of the Khirbet Faynan, the principal Roman/Byzantine standing monument of the area located at the juncture between Wadis Faynan, Dana and Ghuwayr (Figure 1.2). All of these projects were focused on the history of agriculture, mining and metallurgy, and the related development of the Christian community in Faynan.

As a preliminary to the development of its nature reserve, the RSCN had commissioned the Centre for Field Archaeology, Edinburgh University, to undertake an archaeological survey in Wadi Dana, this being directed by Bill Finlayson in his then capacity as manager of the Centre (Finlayson and Baird 1995). Finlayson was then approached by Alison McQuitty, who had become Director of BIAAH in 1994, to undertake work on the early prehistory of Wadi Faynan to complement the later prehistoric and historic studies being undertaken by Barker, Wright, Freeman and their colleagues. Finlayson invited Steven Mithen to collaborate on this project, recognising that research within this region could potentially address their mutual research interests concerning Levantine Palaeolithic, Epipalaeolithic and early Neolithic archaeology. A reconnaissance of Wadi Faynan for early prehistoric settlement was conducted in Spring 1996 (Finlayson and Mithen 1998), the results of which were sufficiently positive to justify the formation of the DFGEP with a four year programme of fieldwork.

e9781782975052_i0002.jpg

Figure 1.1 Location map of study area, marking the three areas surveyed as described in Chapter 3 and location of the PPNA site of WF16.

e9781782975052_i0003.jpg

Figure 1.2 Confluence of Wadi Faynan and Wadi Ghuwayr, looking south-east towards the Jordanian plateau. The PPNA site of WF16 is located on the knoll on the right of the picture and the PPNB site of Ghuwayr 1 is visible by its excavation trench in the mid- ground of the picture on the right side of Wadi Ghuwayr.

1.2 The study area of the DFGEP and previous archaeological studies

The study area is situated at the confluence of three wadis, the Ghuwayr, the Dana, and the Shayqar, which meet to form Wadi Faynan (Figure 1.1). The area has a very diverse topography, with an altitudinal variation from c. 30 m above sea level in Wadi Faynan to the limestone uplands c. 1200 m above sea level on the edge of the Jordanian plateau, combining both Mediterranean and arid zone regions (Figures 1.2, 1.3). This area has various microenvironments, partly reflecting its topographic diversity. Rainfall in the Faynan desert zone varies from a monthly mean of 0–0.1 mm from June to September, rising to only 17 mm a month in December and January, although flash floods can occur as rainfall on the plateau of more than 200 mm a year (Figure 1.4). The topography of Wadi Faynan largely consists of a series of flat and broad wadi bottoms and terraces (Figure 1.5). When travelling eastwards up the wadis, the topography becomes very rugged as they become narrow gorges with few areas suitable for either cultivation or habitation (Figure 1.6). Before reaching the Jordanian plateau, the topography becomes varied and spectacular, with steep hills and cliffs.

Archaeological work in the Wadi Faynan area has taken place for over a century. Initial visits by pioneering scholars such as Lagrange (1898, 1900), Musil (1907), Frank (1934), Glueck (1935) and Kind (1965) established the importance of the area, particularly in relation to mining and early Christian history. More recently, survey work has been conducted by Raikes (1980), King et al. (1989), and MacDonald (1992), while the German Mining Museum at Bochum conducted a project to examine the mining history of the area (Hauptmann and Weisgerber 1987). Adjacent to Wadi Faynan, Russ Adams and Thomas Levy direct a substantial field project dedicated to later prehistory (Adams 1991; Adams and Genz 1995; Levy et al. 1999).

Surprisingly, given the extent of archaeological activity, very little was known about early prehistoric settlement in the Wadi Faynan region prior to the DFGEP project. Survey work by the Centre for Field Archaeology (CFA), Edinburgh University, had identified and sampled numerous archaeological sites within the Dana Reserve, part of which fell within the Faynan area (Finlayson and Baird 1995). Some artefact scatters were thought to be Middle Palaeolithic in date but the earliest dated settlement in the region was the Pre-Pottery Neolithic B (PPNB) site of Ghuwayr 1, at that time only partially excavated by Dr Mohammed Najjar (Simmons and Najjar 1996, 1998) (Figures 1.7, 1.8), followed by the Pottery Neolithic site of Tell Wadi Feinan (Najjar et al. 1990) (Figure 1.7). Up until April 1996, no major concentration of pre-PPNB material had been identified within the region covered by the DFGEP. The archaeology in the immediate vicinity of the confluence of the Wadis Dana, Ghuwayr and Shayqar is dominated by the Khirbet Faynan (Figures 1.3, 1.7, 1.9), a dense tumble of sandstone masonry within which there are substantial standing walls, which is associated with an aqueduct across the Wadi Shayqar, a reservoir, mill, cemetery, extensive deposits of slag and a field system that extends across the Wadi Faynan, the latter being the principle subject of Barker et al.’s (1997, 1998, 1999, 2000) project.

e9781782975052_i0004.jpg

Figure 1.3 Looking west along Wadi Faynan to the Wadi Araba. The Roman/Byzantine settlement of the Khirbet Faynan is located on the right side of the wadi floor in the mid-ground of the picture, and the PPNA site of WF16 on the left hand side in the foreground.

Despite the absence of any evidence for pre-PPNB settlement, the study area appeared to provide considerable potential for addressing issues in Palaeolithic, Epipalaeolithic and early Neolithic archaeology of the Levant. Surveys elsewhere in Jordan and in similar landscapes to the study area had identified sites from all of these periods, such as in the Azraq basin (Garrard et al. 1985, 1987, 1988, 1994), the Badia (Betts 1991), Wadi Hammeh (Edwards 1988, 1990, 1991), Wadi Hasa (Olszewski and Coinman 1998), Wadi Hismeh (Henry 1988), and Wadi Ziqlab (Banning et al. 1989). Consequently, there was every expectation that a rich early prehistoric record remained to be discovered within Wadis Faynan, Ghuwayr and Dana. The DFGEP aimed to document this record with the possibility of addressing key issues concerning Palaeolithic, Epipalaeolithic and early Neolithic archaeology.

The manner in which the project developed required the last of these to become our major focus of interest owing to the discovery of a Pre-Pottery Neolithic A (PPNA) site, designated as WF16, during the reconnaissance survey of 1996. A sampling exercise in 1997 identified WF16 as having stratified deposits and organic preservation. As a consequence, the majority of the project’s limited resources were placed into evaluating this single site rather than into the field survey within the study area. That was, nevertheless, extensive and recovered a considerable sample of Lower and Middle Palaeolithic artefacts, but sparse evidence for Upper Palaeolithic and Epipalaeolithic settlement in the area. We will, therefore, briefly summarise the key research issues concerning these periods as they relate to the design and results of the survey.

e9781782975052_i0005.jpg

Figure 1.4 Wadi Faynan in flood, April 2002.

e9781782975052_i0006.jpg

Figure 1.5 Gravel terraces in Wadi Faynan, looking south-east towards Wadi Ghuwayr.

e9781782975052_i0007.jpg

Figure 1.6 Wadi Ghuwayr, showing a typical gravel terrace on the meander of the wadi and the steep, rugged terrain.

1.3 Research issues in early prehistoric Levantine archaeology

1.3.1 The Palaeolithic: artefact variability and chronological succession

Palaeolithic sites are relatively rare in Jordan when compared to those of later periods or other regions of the Levant. An examination of the site distribution maps in MacDonald et al. (2000; see also Olszewski 2001; Rollefson 2001) shows a general low frequency of sites from the Palaeolithic through to the Neolithic on the east bank of the Jordan river and in the Wadi Araba. The majority of Jordanian Palaeolithic sites are located on its plateau, in other highland areas in southern Jordan, or in the eastern deserts and oases. While relatively few stratified Lower and Middle Palaeolithic sites have been discovered, numerous surface artefact scatters are known. The scarcity of cave or rock shelter sites may account for the absence of hominin remains. Upper Palaeolithic sites are more numerous than those of earlier periods but have a similar distribution (Olszewski 2001).

e9781782975052_i0008.jpg

Figure 1.7 Wadis Faynan, Ghuwayr, Dana and Shayqar, showing location of the PPNA site of WF16, PPNB site of Ghuwayr 1 and the Pottery Neolithic site of Tell Wadi Feinan, the Nabatean/Roman/Byzantine settlement of Khirbet Faynan and associated aqueduct, cemetery, reservoir, and mill.

In 1997 there were two key issues of particular interest to the DFGEP co-directors that could potentially have been addressed within the study area. The first was the relationship between artefact form, technology and raw material availability. While artefact studies have remained largely typologically oriented in Levantine research, that undertaken elsewhere in the Old World has placed a greater emphasis on the impact of raw material availability on stone technology. This has been especially the case for the study of bifaces. White (1998), for instance, argued that the variation in form of bifaces from southern England was a direct reflection of the size of raw material nodules available; when these were sufficiently large ovates would be produced, otherwise hominins had to compromise and produce pointed artefacts. He found some support for his thesis by comparing artefacts made from secondary and primary raw material sources, the latter being assumed to provide larger and better quality nodules. McPherron (1995) proposed an alternative model, arguing that as the resharpening of bifaces reduces their length, this would cause a shift from pointed to ovate forms.

The DFGEP study area provided an opportunity to examine this and related issues in the context of the Levant. The 1996 reconnaissance survey by Finlayson and Mithen had recognised that the gravels constituting much of the Quaternary deposits contained heavily rolled nodules of flint, while seams of high quality flint were known to exist within the limestone geology of the Wadi al-Bustan region on the plateau (Figure 1.1). Whether or not scatters of bifaces also existed within the study area, and whether sufficient identification to a raw material source was possible to enable research on the impact of raw material on biface technology, remained unclear. Similar questions about raw materials, artefact form and stone tool technology have also been asked about Middle and Upper Palaeolithic assemblages and could also have been potentially addressed within the study area.

e9781782975052_i0009.jpg

Figure 1.8 The PPNB site of Ghuwayr 1 in 2001 (photo: Alan Simmons), marking the location of WF16.

e9781782975052_i0010.jpg

Figure 1.9 Arch at the western extent of the Khirbet Faynan, looking west towards the Wadi Araba with the Roman/ Byzantine field system in the near distance. Nicholas Mithen (aged 13) as scale (April 2003).

A second question regarding Palaeolithic archaeology within the study area concerned cultural succession and chronology. As in other regions of the Old World, it is conceivable that there is a chronological pattern to changes in artefact form and assemblage composition within the key periods of the Lower, Middle and Upper Palaeolithic. Identifying such chronological change and the date of the transition between these periods is notoriously difficult when the bulk of one’s data comes from surface scatters of artefacts. With regard to the Levant, there are two particularly important issues, both of which could feasibly have been addressed by research in the DFGEP study area: initial colonization and the Middle/Upper Palaeolithic transition.

It is now widely believed that the date of hominin dispersal from Africa may have been as early as 2 million years ago (Straus and Bar-Yosef 2001). The Jordan Valley is a continuation of the African Rift Valley and is likely to have provided one of two key dispersal routes from Africa, the other being across the Afar ‘gap’ into modern-day Saudi Arabia (Mithen and Reed 2002). The earliest dated site in the Jordan Valley is ‘Ubediya at c. 1.5 million years ago, but it is conceivable that considerably earlier artefact scatters and even hominin fossils may exist within the region.

Specific field strategies to locate such early material are difficult to devise and implement in southern Jordan. But the study region of the DFGEP provided the potential of establishing a long Quaternary sequence with absolute dates from the application of luminescence techniques. Consequently there was the possibility of devising a chronological sequence for Palaeolithic artefacts, or at least placing some chronological constraints on the dates for the earliest traces of hominin activity by establishing the terminus post quem and ante quem for artefact scatters by dating the sediment horizons which they were either above or below.

One of the most significant challenges facing the interpretation of surface collections arises from the likelihood of palimpsests from multiple periods of activity. In stratified Levantine sites, bifaces are found exclusively in the Lower Palaeolithic (Copeland 1998). In contrast, artefacts produced using the Levallois technique are found in both Lower and Middle Palaeolithic contexts (Bar-Yosef 1998, although there is some debate regarding this issue; see Copeland 1995). It is, therefore, difficult to distinguish between surface scatters that derive only from the Lower Palaeolithic and those that derive from a combination of Lower and Middle Palaeolithic activity. Similarly the presence of blades cannot be used as a chronological indicator as these exist in the Lower, Middle and Upper Palaeolithic industries of the Levant (Bar-Yosef and Kuhn 1999). In general, as our knowledge of Palaeolithic chipped stone assemblage from the Levant increases, it is becoming increasingly difficult to assign assemblage to cultural periods on the basis of their gross technological characteristics (Copeland and Hours 1989).

1.3.2 The Middle/Upper Palaeolithic transition

Until the last 20 years our knowledge of the Upper Palaeolithic in Jordan was very limited and relied heavily on data from the Mediterranean coast (Henry 1986). A series of survey projects in the Wadi Hasa (MacDonald 1988), southern Jordan (Henry 1979, 1988 and 1995; Schyle and Uerpmann 1988), the Jordan Valley (Edwards et al. 1988; Muheisen 1988) and eastern Jordan (Garrard and Price 1977; Garrard et al. 1986, 1988; Betts 1983, 1984, 1985) has altered this picture. Such work has changed our understanding of settlement patterns, the nature and chronological succession of cultural traditions, and the transition from the Middle to the Upper Palaeolithic.

The Levantine Aurignacian was initially identified in cave and rockshelter sites in Lebanon, Syria, and northern Israel, and was seen to have strong similarities to the European Aurignacian. As information has been gained from beyond the Mediterranean area it has become increasingly clear that there is a great deal of technological and typological variation within the Levantine Aurignacian (Coinman 1998). This realization of variability was enhanced by the identification of the Ahmarian (Neuville 1951) in open air sites in the Negev and subsequently in central and southern Jordan (Coinman 1993, 1997, Coinman and Henry 1995; Olszewski et al. 1990, 1994). The Ahmarian has been divided into Early (38–30,000 BP) and Late (23–17,000 BP) phases. The presence of such variability has led to a replacement of unilinear evolutionary explanations with a multilinear model. However, the Upper Palaeolithic assemblages from Jordan show considerable variation, and it now appears that the meaning of variability is becoming increasingly unclear (Coinman 1998, 55). It may be more related to site function and duration of occupation than to cultural succession.

There appears to be a local transition from the Middle Palaeolithic (Mousterian) to the Upper Palaeolithic, specifically to the Early Ahmarian. This has been documented at Wadi Aghar (J433) (Coinman and Henry 1995) and Tor Sadaf (WHNBS 8) (Coinman and Fox 2000). Both sites show transitional technological features, similar to those at Boker Tachtit (Marks 1983) and Ksar Akil (Ohnuma and Bergman 1990), especially with regard to core technology and debitage production. A particularly characteristic product are elongated blades with converging dorsal facets. The transition from the production of Levallois points to el-Wad points at Tor Sadaf, where there is a change from true Levallois points to Levallois-like points produced using non-Levallois reduction techniques and then to the production of el-Wad points on blades, is particularly striking (Coinman and Fox 2000). However, where material arises from surface collections, rather than stratified excavations, it may be difficult to determine whether material is Mousterian or early Upper Palaeolithic.

Coinman and Henry (1995) suggest that the transition to the Ahmarian indicates an innovation by the indigenous Levantine Mousterian groups who expanded their exploitation of the arid zone. They also proposed that the heavier toolkit of this industry (the so-called B and C variants) may be an adaptation to woodland environments, and a direct precursor to the heavier version of the Ahmarian. In contrast, Bar-Yosef (1993) has suggested that the Levantine Aurignacian is the product of a new population moving into the area.

1.3.3 Epipalaeolithic: industrial variability and settlement pattern

The Epipalaeolithic period in the Levant (c. 20,000–10,000 BP) is a period of considerable environmental change during which human communities had to continually adapt to changing distributions of animals and plants. The vast majority of archaeological sites are no more than scatters of chipped stone artefacts. These are complemented, however, by a number of sites during the final millennia of the late Pleistocene that have impressive architectural remains, human burials, art objects and a diverse material culture. Bar-Yosef’s (1970) original scheme for the cultural variability of Epipalaeolithic assemblages has been substantially elaborated resulting in several industries: the Kebaran, the Mushabian, the Geometric Kebaran, the Ramonian, the Natufian and the Harifian (see Fellner 1995; Goring-Morris 1995). The Natufian is often divided into an earlier and later phase. Many of the other industries are also further subdivided, principally on the basis of relative frequencies of microlith types.

These industries have been principally defined by research undertaken on the western side of the Wadi Araba. Consequently, their applicability to assemblages in Jordan is questionable. Epipalaeolithic sites are relatively common in Jordan compared to those of earlier periods, with Olszewski (2001) suggesting a minimum number of nine recognizable lithic industries and phases (Olszewski 2001, 48). In 1997, no Epipalaeolithic sites were known in the DFGEP study area, the closest being those sampled in Wadi Hasa (Clark et al. 1994). The nearest Natufian site was that at Beidha (Byrd 1989), interpreted as a seasonal hunter-gatherer encampment.

Two of the most important areas for Epipalaeolithic assemblages in Jordan are the Azraq basin (Garrard and Byrd 1992), and southern Jordan where a very early Epipalaeolithic industry, the Qalkhan, has been defined (Henry 1982). Qalkhan assemblages have been reported in the Petra area, approximately 30 km from Wadi Faynan (Schyle and Uerpmann 1988). Henry (1988, 1996) has proposed a new industry, the Hamran, which he considers to be a southern Jordanian variant of the Kebaran and Geometric Kebaran. The Hamran is divided into Early, Middle, Late and Final phases, which appear to match chronological variations in the Kebaran/ Geometric Kebaran. Henry has also suggested the use of the term Madamaghan, for what appears to be the southern Jordanian variant of the Mushabian industry (Henry 1988), although Garrard and Byrd (1992) have suggested that this may in fact be related to the early phase of the Epipalaeolithic identified at Azraq.

What the archaeologically-defined Epipalaeolithic cultures actually represent in terms of human communities remains unclear. The implicit assumption in the work of Bar-Yosef (1970) and Goring-Morris (1995), among others, is that they are direct reflections of ethnic groups who used the style of their chipped stone artefacts as a means of self identity. Another possibility is they reflect culturally transmitted patterns of tool making within families and groups that have no explicit social function. When Neeley and Barton (1994) proposed a functional explanation for variability in microlithic form and assemblage composition they received a hostile response (e.g. Kaufman 1995; Fellner 1995; Goring-Morris 1996; Henry 1996; Phillips 1996). Although there were flaws in Neeley and Barton’s specific explanatory model, they were quite correct to challenge the underlying assumption that the employed typological classifications reflect prehistoric ethnic identity as revealed through style. Moreover, Pirie (2004) argues that the particular research goals and methods adopted by researchers have had a greater influence over the definition of cultural entities than has previously been recognized.

Our own experience with variability in microlith form and assemblage composition (Finlayson 1990; Finlayson and Mithen 1997, 2000) supports this challenge as microwear and raw material studies have revealed a complex relationship between technology, form and function. The latter is often hard, if not impossible, to establish, when organic remains and contextual information for chipped stone artefacts are absent from archaeological sites. Nevertheless, if systematic patterning exists within a landscape between palaeoenvironmental variables and assemblage composition, a functional rather than cultural explanation should be favoured. We examined this in a previous field project involving the acquisition of multiple microlith-dominated assemblages across a varied landscape in western Scotland (Mithen 2000) and found that such patterning could be identified. The study area of the DFGEP appeared ideal to repeat such a study owing to the marked variation in topography and distribution of resources such as water and raw materials. Consequently, it was anticipated that field survey would recover a number of microlith-dominated, Epipalaeolithic assemblages that could be assessed in light of existing industrial schemes and interpreted with regard to palaeoenvironmental variables. This proved not to be the case (see Chapters 3 and 23).

Where and when the shifts from entirely mobile hunter-gatherers to semi- and then fully sedentary hunter-gatherers/cultivators/farmers occurred remains unclear and subject to considerable debate. Goring-Morris (1995, 166) separates the majority of the Epipalaeolithic from the Natufian, which seems to appear quite abruptly and fully fledged, and is interpreted by some as representing sedentary hunter-gatherers. There appears to be a chronological correlation between the appearance of sites with substantial architecture, burials and a diverse material culture including artefacts for intensive plant processing, and the late glacial interstadial (c. 14,700–12,800 cal. BP). That climatic phase most likely created an abundance of plant and animal foods owing to increased rainfall. Bar-Yosef and Belfer Cohen (1989) argue that Early Natufian sedentism was the ‘point of no return’ with regard to the origin of farming. The identification of hunter-gatherer sedentism is, however, extremely difficult, and some reject the notion that this is reflected in the Early Natufian settlements (e.g. Edwards 1989).

One scenario for the final phase of the Levantine Epipalaeolithic is that the favourable environments of the late glacial stadial came to an abrupt end with the onset of the Younger Dryas (12,800–11,600 cal. BP). This led to the complete or partial abandonment of the Early Natufian settlements and a return to a mobile hunter-gatherer lifestyle as reflected in the rather sparse record of the late Natufian, prior to the emergence of farming communities at the start of the Holocene (for a review see Mithen 2003).

The problem that archaeologists face, however, is that such scenarios are largely based on evidence from the Mediterranean coastal region which is then generalized across the whole of the Levant. Garrard et al. (1988) found a quite different sequence of developments in the Azraq basin where domesticates appear to have been introduced into hunter-gatherer economies. As the DFGEP study region lies between the Mediterranean coastal region and the arid eastern desert, an important question is whether its own developments matched one or other of these other regions, or provided its own unique sequence of change. Also, it remains unclear whether the environmental impact of the Younger Dryas in the southern reaches of the Levant was equivalent to that in the Mediterranean zone and the northern Levant.

1.3.4 The Pre-Pottery Neolithic A: origins and diversity

One of our attractions to the DFGEP study area in 1996 was the on-going excavation of Ghuwayr 1, a PPNB settlement (Simmons and Najjar 1996). This was the oldest known settlement in the area and it raised the general question regarding the origin of such PPNB settlements in southern Jordan: do they reflect the immigration of people from the north or a development by resident PPNA/Epipalaeolithic communities? Ghuwayr 1 was just one of several impressive PPNB sites in southern Jordan, the others including Beidha (Kirkbride 1968), Basta (Nissen 1990) and Ba’ja (Gebel and Bienert 1997). In light of the extreme rarity and peripheral nature of PPNA and Natufian sites in this region, as it appeared in 1997, the notion of immigration into the area seemed most likely. Architectural transitions from round to rectangular dwellings are evident in the northern regions of the Levant, notably at Mureybet, Syria (Cauvin 1977; and as reported during the course of our project at Jerf al Ahmar, Stordeur et al. 1997), and this is where the earliest dates for PPNB settlements also occur.

The issue of the origin of the PPNB in the southern Levant required examination by survey within the DFGEP study area to establish whether any PPNA settlement(s) had been in existence, and if so, whether any specific cultural links to Ghuwayr 1 could be identified. The presence of at least one PPNA site, designated as WF16, was identified by the 1996 reconnaissance survey. Other than this, the only other PPNA site in the vicinity of the study area known in 1996 was Dhra’ (Kuijt 1995), 60 kilometres to the north of Wadi Faynan.

Within the PPNA of the Levant, two phases are frequently identified: the Khiamian and the Sultanian (Crowfoot Payne 1976). These cultural entities are claimed to differ with regard to the composition of their chipped stone assemblages, with the Sultanian having lower frequencies of microliths and the presence of bifacial artefacts and ‘Hagdud truncations’, and are traditionally believed to form a chronological succession. The reality of this division is subject to debate. It has been claimed that there is a single PPNA entity, with some assemblages becoming mixed with earlier Natufian deposits, which has inflated the proportion of microliths. Others argue that the division is real but reflects contemporary variation within the PPNA rather than a chronological succession (e.g. Garfinkel and Nadel 1989; Nadel 1990; Garfinkel 1996; Kuijt 1997). Some scholars see the distinction as important to their models of the transition to farming, viewing the Sultanian as the first Neolithic, and the Khiamian as the final Epipalaeolithic (e.g. Cauvin 2000, Valla 1995).

With the discovery of the PPNA site of WF16 during the 1996 reconnaissance survey, the possibility arose of addressing the Khiamian/Sultanian issue. This required that chipped stone assemblages were acquired from distinct and well dated contexts from different spatial areas of the site to allow a detailed study of within site inter-assemblage variation.

The Khiamian/Sultanian issue relates to a broader research question within early Neolithic and Epipalaeolithic archaeology of the Levant, and indeed within the whole of prehistoric archaeology, that concerning the significance of typological variability. The functional interpretation that Neeley and Barton (1994) proposed for the Epipalaeolithic industrial variants could equally be applied to the Khiamian/Sultanian distinction, or indeed on a finer level within the archaeology of the PPNA within regard to specific artefact types.

Pointed artefacts proliferate within the early Neolithic and are often placed into categories described as ‘awls’, ‘borers’ and ‘projectile points’, on morphological criteria. Such terms imply function, but whether a strong association exists between form and function remained untested in the PPNA. Previous studies by the project co-directors on microliths from Mesolithic sites in western Scotland had indicated that such relationships were complex, and assumptions about microliths functioning as projectile points were often incorrect (Finlayson 1990; Finlayson and Mithen 1997, 2000).

The discovery of the PPNA site of WF16 during the 1996 reconnaissance survey provided the potential to explore these issues by two methods. First, it was hoped that artefacts could be recovered and interpreted in light of contextual information, especially their association with animal bones, plant remains and types of deposits (midden, floors, etc.). Second, as the surface material from WF16 appeared quite fresh, it was anticipated that artefacts from sealed contexts might be suitable for microwear analysis. Consequently there was the possibility of undertaking a microwear study of pointed and other artefact classes, building upon the methods established in our previous studies.

A third key research area regarding the PPNA that we hoped to address in the DFGEP concerned settlement patterns and economy within this period. While PPNA sites are rare within the Levant, those with sufficient animal and plant remains that enable questions about economic practices to be addressed are even more scarce. Opinion varies as to whether the PPNA is a period of hunting and gathering or farming, with numerous variants within such categories, notably the cultivation of wild cereals. Related to this issue is the extent to which the PPNA settlements represent permanent ‘villages’, or seasonally occupied campsites. While the monumental architecture at Jericho certainly suggests the former, that site remains unique. Kuijt (1994) proposed that PPNA settlement in the southern Levant formed a hierarchical pattern of settlement, and his ideas seemed ready for re-evaluation.

With the discovery of one PPNA settlement in Wadi Faynan during the 1996 reconnaissance survey, the possibility existed that further traces of PPNA activity could be located within the study area, which might be interpreted as either ‘off-site’ activities from WF16, or independent settlements in their own right. Moreover, although the state of preservation of WF16 was unknown in 1996, the possibility existed that palaeoenvironmental data could be acquired by excavation to address issues of economy and the seasonality of occupation.

1.4 The project programme

Following the preliminary reconnaissance in 1996, we designed a programme of four seasons of survey and trial excavations (1997 to 2000). This was followed by four supplementary and shorter fieldwork seasons (2001 to 2004) to acquire additional data.

In light of the research issues discussed above, and the projects being undertaken by Barker, Wright, Freeman and their colleagues, the DFGEP project’s research objectives were defined as follows:

To document the archaeological record for early prehistoric settlement (i.e. pre-PPNB) within the study region.

To reconstruct the palaeoenvironmental context for early prehistoric settlement.

To interpret the variation in that early prehistoric archaeological record across space and through time, at both an intra-site and inter-site level.

As we have already noted, the 1996 reconnaissance survey identified a dense scatter of chipped stone covering a knoll close to the junction between Wadis Faynan and Ghuwayr. This was established as a relatively well preserved PPNA settlement in 1997, and in light of the significance of this discovery a single site came to dominate the project to a greater extent than had been anticipated for what was instigated as primarily a survey project. Specific research objectives regarding WF16 were defined as:

To undertake an archaeological evaluation of the site with regard to identifying levels of preservation, and documenting the character of architectural and artefactual assemblages, as a means to establish the viability and likely returns from a dedicated excavation project that would take place after the DFGEP was completed.

Secure sufficient assemblages of chipped stone and other artefacts to address key issues relating to the PPNA. The following were identified as preeminent:

–The origin of the PPNA with regard to its continuity with Epipalaeolithic cultural entities, notably the Natufian, and environmental change during the Late Pleistocene.

–The cultural variation within the PPNA, with particular attention to whether the Sultanian and Khiamian are distinct entities, and if so, whether they represent successive cultural phases or functional variants within the PPNA.

–The economic basis of the PPNA with the study region: was this hunting-gathering, plant cultivation, herding, or fully-fledged farming?

–The chronological, economic and cultural relationship between the PPNA and PPNB within the study region, this being possible owing to the presence of the PPNB site of Ghuwayr 1 within 500 metres of WF16.

The project began as a joint Universities of Edinburgh and Reading exercise under the direction of Bill Finlayson and Steven Mithen. Dr Richard Tipping of the University of Stirling joined the project to undertake a geomorphological survey to reconstruct Pleistocene and Early Holocene environmental change and the formation of the current landscape. This proved to be a particularly complex undertaking leading to his lengthy and detailed report in Chapter 2, which provides the basis for the reconstruction of the environment in which the settlement of WF16 was occupied during the PPNA. As a means to extend that reconstruction, a survey of modern vegetation within the region was undertaken towards the end of the project by Mithen, Emberson and Lancaster (Chapter 22), which also served to facilitate interpretation of the plant remains recovered from the excavation.

In 1997 Anne Pirie was invited to join the project as the principal lithic analyst, at that time engaged in doctoral studies at the University of Durham into the significance of variation in Epipalaeolithic chipped stone assemblages of the Levant. She undertook the majority of the cataloguing and analysis of the artefacts acquired by the field survey described in Chapter 3, with further work undertaken by Bill Finlayson and Sam Smith, a graduate student at the University of Reading. Anne Pirie also catalogued and interpreted the WF16 chipped stone (Chapter 8). Two further University of Reading graduate students, Suzie Pinkett and Eileen Clegg, made contributions towards documenting and interpreting the rock art (Chapter 4) and bifaces (Chapter 5) located during the survey.

Following the discovery of WF16 in 1996 and a test excavation in 1997, the two most substantial field seasons were undertaken in 1998 and 1999, with additional fieldwork at the site in 2001. Chapter 6 describes the results of the test excavations at WF16 in terms of stratigraphic sequences and architectural structures. A programme of environmental sampling was devised and implemented by Amanda Kennedy, a graduate of University College London, who also undertook the analysis of plant macrofossils (Chapter 20). Denise Carruthers, a graduate student at the University of Edinburgh, also joined the project to analyse animal bones, supported by Sam Dennis of the University of Reading (Chapter 15). A series of specialists were recruited to catalogue and interpret specific classes of material: Ruth Shaffrey (Oxford Archaeology Unit) for coarse stone artefacts (Chapter 11); Pat Critchley (Institute of Archaeology, London) for stone beads (Chapter 12); Ruby Cerón-Carrasco (University of Edinburgh) for shells and shell artefacts (Chapter 13); Karen Wicks (University of Reading) for a bitumen deposit (Chapter 14); Yvonne Edwards and Louise Martin (Institute of Archaeology, London) for microfauna (Chapter 16); Kevin Rielly (Museum of London) for bird bones (Chapter 17); Charlotte Roberts (University of Durham) for human bones (Chapter 18); and Phil Austin (Institute of Archaeology, London) for wood charcoal (Chapter 19); Emma Jenkins and Arlene Rosen (Institute of Archaeology, London) for phytoliths (Chapter 21); Bill Finlayson catalogued bone artefacts (Chapter 10); Dr Tim Astin and Carol Mansfield (University of Reading) conducted a geophysical survey of the PPNA site (Chapter 23).

In addition to employing these specialists the project was supported by two PhD studentships. Sam Smith (funded by the Arts and Humanities Research Board, UK) undertook a use wear study of pointed artefacts from WF16 and the 2001–02 excavations at the PPNA site of Dhra’ undertaken by Finlayson et al. (2003). His report in this volume (Chapter 9) focuses on the WF16 sample and addresses one of the key research questions identified above–the relationship between artefact form and function. Eleanor Roe (funded by the University of Reading) undertook a microstratigraphic analysis of the deposits at WF16, providing the first application of such methods to a PPNA site and making a key contribution to site interpretation (Chapter 7).

The interpretation of radiocarbon dates for early Holocene sites should always be taken with care. Hence we devote Chapter 24 to establishing the absolute and relative chronology of settlement at WF16.

Chapter 25 of this volume seeks to synthesize all of the above lines of evidence to interpret the archaeological remains at WF16, addressing the PPNA research objectives defined above.

While the DFGEP fieldwork and post-excavation studies were being undertaken, two new field projects focusing on PPNA sites in southern Jordan were initiated. Bill Finlayson and Ian Kuijt began an extensive programme of excavation at the site of Dhra’ (Finlayson et al. 2003), while Phillip Edwards undertook excavations at the PPNA site of Zahrat adh-Dhra’ 2 (ZAD 2), a mere two kilometres northwest of Dhra’ (Edwards and Higham 2001; Edwards et al. 2002; Sayej 2004). Such intensity of work on the PPNA in Jordan is unprecedented, and the preliminary reports from these sites have refined and extended the research questions about the PPNA that were posed when the DFGEP was initiated. Consequently, Chapter 25 of this volume also addresses the contribution that WF16 has made to our understanding of the PPNA of the Jordan Valley and beyond, integrating our findings with those from the new work at Dhra’ and Zahrat adh-Dhra’ 2, along with the previous studies of the PPNA at such classic sites as Netiv Hagdud and Jericho.

PART ONE

ARCHAEOLOGICAL SURVEY

2. Long-term landscape evolution of the Wadis Dana, Faynan and Ghuwayr

Richard Tipping

2.1 Purpose and objectives of the study

This study has two objectives. First, it is to establish the relative and absolute age of sedimentary deposits in the Dana-Faynan-Ghuwayr Early Prehistory Project (DFGEP) study area to facilitate the archaeological survey (Chapter 3). Second, it is to use geomorphological and sedimentological data to reconstruct landscape and climate change contemporaneous with the local transition from hunter-gatherer (Late Epipalaeolithic) to earliest farming communities (Pre-pottery Neolithic B, PPNB) in Wadi Faynan. As such, this geomorphological survey aimed to provide data to complement that derived from the excavation of WF16 as described within this volume, and to place that site in its long-term palaeoenvironmental context.

Sediment sourcing was seen as a critical element of this study to enable an understanding of the balance between alluvial (riverine) and colluvial (slope) processes through time, and to aid in palaeohydrological reconstruction. Routine stone count data (Bridgland 1986) were collected to explore this, which also provided a source of information to explore more fully the procurement of raw materials by past communities, and notably of course the accessing of flint (Finlayson et al. 2000).

2.2 The Dana-Ghuwayr catchment at the head of Wadi Faynan

2.2.1 The Wadi Araba Rift Valley Floor and the Wadi Faynan

Most visitors enter the Dana-Ghuwayr catchment from the west, from the Wadi Araba and Wadi Faynan (Figure 2.1). This is part of one of the world’s major tectonic features, a rift valley or sunken block that extends from Africa through the Gulf of Aqaba, the Wadi Araba and the Dead Sea, and north into Lebanon (Figure 1.1), a major suture between the African and Arabian tectonic plates. The Wadi Araba Fault is a major north-northeast /south-south-west fault system with many subparallel faults, showing predominant lateral movement of around 100 km, but also with vertical displacements. The fault system was formed within the Cenozoic, in the Tertiary and Quaternary periods, and is active at the present day. The Jordan River flows south through the Wadi Araba, within an arid (<80 mm mean annual rainfall) desert landscape in which aeolian sand dunes are common. These currently mobile landforms drape very large alluvial fans of Pleistocene age derived from the mountains to the east.

The Wadi Araba is generally at a substantially higher altitude than the Dead Sea rift, lying between 80–100 m above sea level (asl), not reaching the 430 m below sea level of the Dead Sea itself. The Dead Sea basin to the north contains sediments of former large lakes, the Samra and Lisan Formations. The highest levels of Lake Lisan, the precursor to the present Dead Sea, are thought to have reached to 180 m below sea level in the late Pleistocene period (Neev and Emery 1967, 1995; Begin 1985; Frumkin et al. 1991; Yechieli et al. 1993). Nevertheless, in southern Jordan the Wadi Araba contains comparable calcareous marls, also called the Lisan Marls. A precise chronology has still to be established for the Lisan Marls in the Wadi Faynan area, but they overlie and cut into Pleistocene alluvial fans and are not thought to have been actively forming in the Holocene period (Bender 1974; Rabb’a 1994). The sediments must record climatically controlled fluctuations in lake deposition in this period. Dissected outcrops of these are preserved around Al-Qurayqira (Figure 1.1) at the entrance to Wadi Faynan, to altitudes of 60 to 70 m asl, within 10 km of the lowest point of the surveyed area west of the Khirbet Faynan (Figure 2.2), and a lake in the Wadi Araba would periodically have controlled the base level of the fluvial system draining from the east.

At Al-Qurayqira the Wadi Faynan is some 5 km wide, the valley floor covered in blown sand, Holocene and active fluvial gravels. Al-Qurayqira is the only permanent settlement eastward until the villages at the head of the catchment. Here, small irrigated fields are supplied by pumping groundwater and by pipe from Wadi Ghuwayr, but the valley floor carries only winter storm stream flows. Low cliffs confine the valley floor to a 1 km wide corridor, and from Al-Qurayqira over 5 km the valley floor rises eastward at a low gradient to 200 m asl. Very prominent gently sloping fluvial terraces border the northern edge of the Wadi Dana (Figures 2.3, 2.4), and the most easterly of the major alluvial fans descends from the 550 m asl mountains of the Jabal Mubarak to the south (Figures 2.2, 2.3, 2.5). There is an absence of soil and any expanse of vegetation.

e9781782975052_i0011.jpg

Figure 2.1 A view from Khirbet Faynan across lower Wadi Ghuwayr in the foreground and into the narrow gorge of the upper Wadi Ghuwayr at the left of the photograph. To the extreme right the track leads into Wadi Shayqar across the fan sediments and surface of the Murad Fanglomerate. These sandstone derived fan sediments in the middle distance are mid-brown in colour. Immediately south (right) of the entrance to upper Wadi Ghuwayr are pale cream outcrops of Ghuwayr Gravel, and the larger of these is the archaeological site of WF16. These contrast with the surrounding dark colluvial fills of the Picnic Fanglomerate, and with the black, bare and heavily gullied slopes of the Ghuwayr Volcanic Series behind. On the skyline are pale brown Ram and Kurnub Sandstone Formations of the Jabal Mubarak.

2.2.2 The Rift Valley Escarpment

At the Khirbet Faynan, the Wadi Faynan bifurcates into northern and southern tributary valleys, the Wadi Dana and the Wadi Ghuwayr (Figure 2.2). The sense of confinement increases as these valleys narrow to rocky gorges (siqs) 1.5 km east of the Khirbet Faynan (Figure 2.1). This reach of the wadi system is the major ecotone in the region, for humans and animals. Four-wheel drive vehicles become redundant, and even in more traditional transport, camels need to be substituted for donkeys. Tracks in the mountains are often single file routes high on cliff walls, leading to the Eastern Highlands above. The terrain is, as an understatement, rugged, with no soil, precious little vegetation, little scope for chemical weathering of rock, and physical shattering (freeze-thaw) instead producing angular, unstable rock fragments on very steep slopes. Vertical cliffs abound.

The mountains rise steeply from the Wadi Araba, rising in little more than 10 km from 100 m asl to over 1500 m asl (Figure 2.2). The mountains represent the eastern edge of the rift valley, and are intensely affected by faulting, which controls much of the drainage. Within the valleys draining into the Wadis Dana and Ghuwayr, the major structural controls are the east-north-east/west-south-west Dana Fault and the east/west trending Salawan Fault, rather than the broadly north/south systems which parallel the rift valley itself (Figure 2.3). In the survey area, the Dana Fault shows vertical lowering of rocks to the north by up to 150 m (Barjous 1992). To the south, the Salawan Fault downthrows strata on the south side of the fault by around 900 m, but with lateral movement of around 7 km, displacing rocks westward on the south side of the fault. Defined by and between these two faults is a raised block called the Dana Horst (Figure 2.3) in which a thin sliver of the oldest rocks in the catchment are exposed. These east/ west trending faults are probably earlier than the more common but less significant north/south trending faults, and were rejuvenated in the Late Pliocene period and into the Middle Pleistocene period.

e9781782975052_i0012.jpg

Figure 2.2 The catchment outline, topography, and settlements of the Dana-Ghuwayr system. The contours are in metres at 200 m intervals.

The Dana Fault controls entirely the straight and deeply incised Wadi Dana (Figure 2.3). The catchment of the Wadi Dana parallels this straight course as the well jointed rocks and minor faults control short but very steep tributary wadis descending, often at right angles, from the plateaux at 700–1100 m asl. The valley floor is narrow but does preserve a continuum of Pleistocene and later slope and fluvial features.

In contrast, the Wadi Ghuwayr is incised within the uplifted blocks of the Dana Horst, and winds into the core of the mountains with no clear structural control until it intersects the Salawan Fault 3.5 km east of Khirbet Faynan (Figure 2.3). In the lower Wadi Ghuwayr, downstream of the Salawan fault, the valley sides are steep and precipitous, but ‘fossil’ fluvial and colluvial features are comparatively well preserved. At the Salawan Fault the character of this wadi system changes abruptly. One major tributary, the Wadi al Hamra, is controlled east/ west by the Salawan fault, rising in a straight line confined within cliffs and steep slopes of volcanic rocks, sandstone and limestones to the top of the catchment at Shajarat al Tayar (Figures 2.2, 2.3). Old fluvial terrace surfaces are present but rare. This valley provides the most direct route, for rivers and humans, to flint outcrops in the headwaters. However, the route is blocked by one of a number of major landslips in the Ghuwayr catchment (Figure 2.3), the Wadri al Hamra landslip. This and major landslips at al Mansura and in Wadi al-Bustan have failed on underlying soft and easily deformed sediments. Barjous (1992) argued that these landslips were relatively undissected, and most probably formed under more humid climates either within the later Pleistocene or the Holocene periods. This suggestion is re-examined in Section 2.7.

e9781782975052_i0013.jpg

Figure 2.3 The distribution of major rock types (see Section 2.3), major fault systems, major landslips, Pleistocene, later and currently active sediment fills in the Dana–Ghuwayr catchment. The contours are in metres at 200 m intervals.

e9781782975052_i0014.jpg

Figure 2.4 From Khirbet Faynan, the few trees in the landscape are in the foreground, growing on moister late Holocene sediments examined by Barker et al. (1999). In the distance is lower Wadi Dana, and at the foot of the sandstone hills of the Jabal Finan are a series of prominent fluvial terrace sediments beneath the Ghuwayr, Khaled and Shayqar Surfaces.

e9781782975052_i0015.jpg

Figure 2.5 A view from Khirbet Faynan south to the extensive surface of the dissected surface of the Murad Fanglomerate descending from Jubal Mubarak.

The Wadi Ghuwayr itself is a second major tributary, draining a very extensive catchment to the south of the Salawan Fault (Figure 2.3). From its confluence with the Wadi al Hamra to its headwaters, it initially flows south in an astonishing siq (bedrock gorge) cut into Umm Ishrin Sandstone two metres wide but with vertical cliffs hundreds of metres high. No ‘fossil’ fluvial features are preserved beyond this point until the headwaters high above. Springs flow from these sandstone outcrops and provide permanent and perennial water. At the present day, this reach of the Ghuwayr, from the Salawan Fault downstream 4 km to the exit of the gorge, is the only wadi in the region to carry water other than seasonally, and the groundwater source of this water suggests this to have been an enduring and significant environmental factor in the past.

The Wadi Ghuwayr emerges in its headwaters into a still heavily dissected but more open landscape, as it erodes into substantially less resistant calcareous rocks of the Ajlun and Balqa Groups (Barjous 1992: Figure 2.3). Slopes are still steep, but the predominantly fine grained rocks weather to uniform V-shaped slopes which are easier to move on, although too steep for soils to be retained. The ease with which these softer rocks, exposed on the south of the Salawan Fault, are eroded is the major reason for the eccentricity of the catchment shape (Figure 2.2) and the development of the sub-catchment of the Wadi al-Bustan. The drainage network is particularly intricate in the southern headwaters, controlled by a series of faulted basins and blocks and by synclinal folding of rocks within Wadi al-Bustan (Barjous 1992). This fold structure has also allowed the storage of very great thicknesses of Pleistocene sediment, the Dana Conglomerate (Figure 2.3), the only significant accumulation of sediment of this age in the catchment above Wadi Faynan.

2.2.3 The catchment headwaters

Ringing the Dana catchment are rugged peaks of sandstone, but to the east of the Dana Horst and south of the Salawan Fault younger sediments present gentler terrain (Figures 2.2, 2.3). The climate at the head of the catchment contrasts sharply with that at Al-Qurayqira only 17 km to the west. Mean annual rainfall at Ash-Shawbak, south of Wadi al-Bustan, is about 300 mm. Seasonal contrasts in temperature are much greater than in Wadi Araba, between an average 27°C in August and below 0°C between December to February, and snow falls around 10 days per year. Farming is a more secure occupation in the Highlands, and small villages like Al Mansura and Dana ring the watershed (Figure 2.2). Many slopes on the soft calcareous sediments in the headwaters are terraced in attempts at soil conservation, though nearly all terraces are currently degraded. The fall from headwater villages to these formerly cultivated areas is still a significant impedance to human movement, however, not least because the plateau is capped by Pliocene and Pleistocene porphyritic basalt which can form near vertical cliffs. The most recent basaltic eruption is of mid-Pleistocene age, sealing Acheulian flint tool assemblages (Bender 1974). The basalts extend east of the Dana-Ghuwayr watershed, lying conformably on gently south east dipping Plio-Pleistocene fluvial gravels, and are covered in turn by later Pleistocene gravels.

2.3 Rock types contributing to fluvial and slope sediments

Understanding the sources of rocks collected in fluvial and slope sediments in the lower reaches of the Wadis Dana and Ghuwayr requires an understanding of where in the catchment the rocks are found, and how identifiable as sometimes small fragments the rocks are. Figure 2.3 shows the distribution of distinctive rock types in the catchments, taken from the mapping of Barjous (1992) and Rabb’a (1994). It is not the intention to define in detail the geological record in the catchment, but to characterise the rocks and indicate their sources.

2.3.1 The Hunayk Porphyric Granite

This plutonic igneous rock outcrops only over 2 km within the fault controlled Wadi Dana (Figure 2.3). It is a granite with a fine to medium grained groundmass but with large phenocrysts of feldspar.

2.3.2 The Ghuwayr volcanic suite

This distinctive volcanic suite is found only within the Dana Horst (Figure 2.3), largely confined to the Ghuwayr catchment. It forms bedrock lining both sides of this wadi from east of Khirbet Faynan until Wadi Ghuwayr intersects the Salawan Fault, and then forms the west side of Wadi al Hamra until the landslip in that valley. All five lava types (Barjous 1992) are grouped together as green-black, weathering to dark brown, amygdaloidal and porphyritic basaltic andesite. It readily shatters to very angular fragments on slopes and builds steep to very steep talus slopes but not cliffs.

2.3.3 The Fidan Syenogranite

Within the Dana Horst and also on the north side of Wadi Dana, but contributing to both Wadis Dana and Ghuwayr by short steep tributaries (Figure 2.3), this plutonic granite is recognised as a red fine to medium grained red granite, but equigranular in contrast to the Hunayk Granite.

2.3.4 Ram and Kurnub Sandstones

These sandstones form very large areas of the catchments of both the Dana and Ghuwayr. The Ram Sandstone Group is of Lower Palaeozoic age, and two units are mapped (Barjous 1992), the Umm Ishrin and Disi Sandstones. The Kurnub Sandstone overlies these unconformably, and is of Lower Cretaeceous age. All sandstones have similar hand characteristics: yellow-brown, red-brown, grey-brown and purple-brown haematite-rich medium- to very coarse-grained quart-zitic, arkosic and subarkosic sandstones, and so in Figure 2.3 the outcrops of all these sandstones, regardless of age, are shown as one sandstone unit.

This rock type is found to north and south of Wadi Faynan west of the Khirbet, and forms the alluvial fans deriving from Jabal Mubarak. Within Wadi Ghuwayr it provides material to left bank tributaries, including the major channel draining Wadi Shayqar from the Jabal Mubarak, until the Salawan Fault when the unit is laterally displaced to form the cliffs on the west side of Wadi al Hamra close to the watershed (Figure 2.3). Sandstone is found extensively on either side of Wadi Dana.

2.3.5 Ajlun Limestones

Grey to light-brown massive and nodular, crystalline and commonly fossiliferous dolomitic limestones characterise these Cretaceous rocks. Chert occurs in thin bands but is not sufficiently variable to be traced precisely to source outcrops. These limestones are grouped together (the Na’ur, Fuhays, Hummar, Shu’ayb and Wadi as Sir Formations of Barjous (1992)) in Figure 2.3. These limestones also include softer, powdery marls and calcareous siltstones, but as pebbles these were not considered separable from younger, Tertiary sediments, the Balqa Limestones (below). In Wadi Ghuwayr, this group of rocks occurs high in the catchment, above 800 m asl, well above the lower wadi floor and derived solely from tributaries draining the headwaters. In Wadi Faynan west of Khirbet Faynan major downfaulting has lowered these limestones to the valley floor at and below 200 m asl, but right bank tributaries can introduce this material only downstream of all localities studied in this analysis, and so this outcrop is not a likely source to the fluvial system considered in this report. Outcrops in the upper part of the Dana system are comparable in setting but less extensive than in the Ghuwayr catchment, but may formerly have been more extensive, reduced now through headwall retreat of gullies.

2.3.6 Balqa Limestones

These sediments are characterised by chalk, although micritic limestones and marls are found in the Ajlun Group. Chert is not unique to these sediments, but is particularly common in the Umm Rijam Chert Limestone (Barjous 1992). These sediments, lying above the Ajlun Group, are confined entirely to the headwaters of Wadi Ghuwayr, and are exceptionally rare and poorly developed in Wadi Dana (Figure 2.3).

2.4 The surveyed area

Figure 2.2 outlines the area within this large, complex and extreme landscape that was analysed in detail. The lower wadi floors, commencing around 5 km east of Al-Qurayqira, were selected for analysis because of the need to provide a landscape context for the newly discovered archaeological site of WF16 near to the confluence of the Wadis Ghuwayr and Faynan and for the adjacent PPNB site of Ghuwayr I excavated by Simmons and Najjar (1996). The mapped extent of Pleistocene sediments close to the confluence of the Wadis Dana and Ghuwayr (Barjous 1992; Rabb’a 1994; Barker et al. 1997) is much greater than within the gorge sections of the wadis, and allowed greater scope for geomorphological interpretation and for guided archaeological field survey. The work was then extended into the lower gorge of Wadi Ghuwayr, which represents an ecologically contrasting landscape.

2.5 Methods employed

2.5.1 Geomorphological mapping

Identification of superficial features formed in unconsolidated sediments was undertaken by field walking and mapping at a scale of approximately 1:10000 onto enlargements of 1:25000 topographic maps, with contour intervals at 5.0 m intervals. Coverage of these was not comprehensive and recourse was occasionally made to mapping from less detailed 1:50000 base maps. Mapping was aided by transcription in the field onto aerial photographs at a scale of 1:25000. The principal features mapped were fluvial terrace surfaces, not necessarily flat because of post-depositional modification (colluvial deposition or erosion) but always confirmed by sediment logging of underlying fills, alluvial fans and talus slopes.

The geomorphological study mapped the valley floors and lower slopes of Wadi Dana east to the