Libyan Studies 41 (2010) The Cyrenaican Prehistory Project 2010: the fourth season of investigations of the Haua Fteah cave and its landscape, and further results from the 2007–2009 fieldwork By Graeme Barker,1 Annita Antoniadou,2 Simon Armitage,3 Ian Brooks,4 Ian Candy,3 Kate Connell,5 Katerina Douka,6 Nicholas Drake,7 Lucy Farr,1 Evan Hill,2 Chris Hunt,2 Robyn Inglis,5 Sacha Jones,1 Christine Lane,6 Giulio Lucarini,8 John Meneely,2 Jacob Morales,9 Giuseppina Mutri,8 Amy Prendergast,5 Ryan Rabett,1 Hazel Reade,5 Tim Reynolds,10 Natalie Russell,3 David Simpson,2 Bernard Smith,2 Chris Stimpson,5 Mohammed Twati11 and Kevin White12 Abstract The paper reports on the fourth (2010) season of fieldwork of the Cyrenaican Prehistory Project, and on further results of analyses of artefacts and organic materials collected in the 2009 season. Ground-based LiDar has provided both an accurate 3D scan of the Haua Fteah cave and information on the cave’s morphometry or origins. The excavations in the cave focussed on Middle Palaeolithic or Middle Stone Age ‘Pre-Aurignacian’ layers below the base of the Middle Trench beside the McBurney Deep Sounding (Trench D) and on Final Palaeolithic ‘Oranian’ layers beside the upper part of the Middle Trench (Trench M). Although McBurney referred to the upper part of the Deep Sounding as more or less sterile, the 2010 excavations found evidence for small-scale but regular human presence in the form of stone artefacts and debitage, though given the sedimentary context the latter are unlikely to represent in situ knapping. The excavations of Trench M extended from the basal Capsian layers investigated in 2009 through Oranian layers to the transition with the Dabban Upper Palaeolithic. Some 17,000 lithic pieces have been studied from the Capsian and Oranian layers excavated in Trench M, in an area measuring less than 2 m by 1 m by 1.1 m deep, along with numerous animal bones, molluscs, and macrobotanical remains, as well as occasional shell beads. Preliminary studies of the lithics, bones, molluscs, and plant remains are revealing the changing character of late Pleistocene (Oranian) and early Holocene (Capsian) occupation in the Haua Fteah. Alongside the work in the Haua Fteah, the project continued its assessment of the Quaternary and archaeological sequences of the Cyrenaican coastland and completed a transect survey of surface lithic materials and their landform contexts from the pre-desert across the Gebel Akhdar to the coast, with a new focus on the al-Marj basin. Significant differences are emerging in patterns of Middle Palaeolithic and later hominin occupation and palaeodemography. Introduction By Graeme Barker The principal aim of the Cyrenaican Prehistory Project is to investigate the long-term relationship between environmental change and human history in the Gebel Akhdar (or ‘Green Mountain’) massif in northeast Libya (Fig. 1). The Gebel Akhdar rises steeply from the Mediterranean littoral up a series of escarpments to a broad ridge over 1000 metres above sea level, and then slopes southwards to the Sahara, so encapsulates a typical range of North African environments. The project began in 2007 and full reports on the 2007, 2008, and 2009 seasons of fieldwork have been published in Libyan Studies (Barker et al. 2007, 2008, 2009). As described in those reports, two particular questions dominate our research agenda: how old is Homo sapiens in North Africa? and how old is farming in North Africa? In both instances we are particularly interested in how people developed behaviours to cope with and adapt to climatic and environmental change, which is known to have been, on occasion, both rapid and profound. To investigate these relationships the project is combining archaeological and geoarchaeological studies across the Gebel Akhdar with excavations at a large coastal cave, the Haua Fteah. Excavations of a 14-metre deep trench in the Haua Fteah in the 1950s by Cambridge archaeologist Charles McBurney revealed the longest occupation 1 McDonald Institute for Archaeological Research, University of Cambridge, UK. 2 School of Geography, Archaeology and Palaeoecology, Queen’s University of Belfast, UK. 3 Department of Geography, Royal Holloway, University of London, UK. 4 Engineering Archaeological Services Ltd, Blaenau Ffestiniog, UK. 5 Department of Archaeology, University of Cambridge, UK. 6 Research Laboratory for Archaeology and the History of Art, Oxford, UK. 7 Department of Geography, King’s College, London, UK. 8 Dipartimento di Archeologia e Preistoria, University ‘La Sapienza’, Rome, Italy. 9 Departamento de Ciencias Historicas, University of Las Palmas, Spain. 10 Faculty of Continuing Education, Birkbeck College, University of London, UK. 11 Department of Archaeology, Omar Mukhtar University, al-Beida, Libya. 12 School of Geography and Archaeology, University of Reading, UK. 63 Graeme Barker et al. Ikctikctwocj Ejgtuc Nconwfcj gt fk"Mwh Yc pc Dtcm"Pqvcj Uwucj Ycfk gn/Sncc gn"Sncc fk" F Yc Jcll"Etgkgo Figure 1. Northeast Libya, showing the principal study area of the Cyrenaica Prehistory Project, and the principal sites and locations mentioned in the text. McBurney filled in his trench at the end of the 1955 season, and much of the effort of the present Ycfk gn/Sncc project has been targeted on the :342 :327 :328 removal of this backfill. The backfill :33; :32: :329 :33: :32; :322 :325 of the Upper Trench (which mea:339 :326 :323 :338 :324 :332 sures approximately 10 m by 10 m, :333 by 2 m deep, and broadly encomNconwfcj passed the Graeco-Roman, Neo:334 lithic, and Capsian occupations) was removed during the 2007 :335 season of fieldwork. The backfill :336 of the Middle Trench, which mea:337 sures approximately 8 m by 7 m, sequence of any archaeological site in North Africa by 6 m deep, was removed in 2008 and 2009. When (McBurney 1967). The sequence began with assemexcavated by McBurney, this trench yielded Capsian blages of Middle Palaeolithic stone tools termed ‘Prematerial in the top 0.5 m, so to about 2.5 m below the Aurignacian’ by McBurney and ended with traces of a present-day ground surface, Dabban from c. 2.5 m ritual building dating to the Graeco-Roman centuries. to c. 5.5 m, Levallois-Mousterian Middle Palaeolithic Ages earlier than 40,000 years ago (the extreme range from c. 5.5 m to c. 7.5 m, and Pre-Aurignacian at the of radiocarbon dating at the time of the excavation) bottom. At the end of the 2009 season we started to had to be estimated from assumed sedimentation remove the fill of McBurney’s Deep Sounding, which rates, and on this basis McBurney suggested that Premeasured approximately 2.5 by 1.5 m, by 6 m deep Aurignacian occupation began some 80,000 years ago and had Pre-Aurignacian material throughout its (modern understanding of Mediterranean palaeoclidepth to the base at 14 m below the ground surface. mates has suggested that the basal occupation might The walls of the McBurney trenches, once cleaned, be at least twice as old as this: Moyer 2003). He argued have been recorded in detail using the single-context that Homo sapiens arrived at the site c. 40,000 years system, and sampled intensively for dating materials ago equipped with a European-style Upper Palaeo(charcoal for 14C dating, calcrete for U-series dating, lithic stone-tool technology which he termed the and quartz-rich sediments for Optically Stimulated Dabban after the cave of Hagfet ed-Dabba, where his Luminescence or OSL dating), proxy indicators of excavations had found similar material (McBurney past climates and environments (micromorphology, 1960), though two human mandibles found in prepollen, charcoal etc.), and cultural materials (espeDabban Middle Palaeolithic layers at the Haua Fteah, cially stone tools, animal bones, and shells). In 2009 originally thought to have affinities to European Neanwe began excavation proper, with a small (2 m by derthals, are now regarded as most likely to be robust 1 m) trench, Trench M, at the top of the Middle Homo sapiens (Hublin 2000). Above the Dabban Trench to investigate Capsian occupation, and, in the layers (Layers XXIV–XVI) he found a Final Palaeolithic final days of the campaign, a 1.5 m by 1.5 m trench, or Epipalaeolithic stone-tool assemblage (in Layers Trench D, beside the top of the Deep Sounding, to XV–XI), which he termed the Oranian and dated to investigate Pre-Aurignacian occupation. c. 14,000 BP (before the present) to around the PleisAlongside the work in the Haua Fteah, other caves tocene/Holocene transition c. 12,000 BP. This was and locations investigated by McBurney in the Gebel succeeded by evidence of ‘Capsian’ Mesolithic hunterAkhdar have been re-visited, including coastal expogatherers, in the period from c. 12,000 BP to c. 7000 BP sures of ancient beach deposits and overlying sedi(Layers X and IX). The Neolithic, the assumed period ments containing prehistoric artefacts (McBurney and of initial agriculture (Layers VIII–IV), was dated to Hey 1955), and subjected to the same range of analytic. 7000–470 BP. The sequence ended above the cal procedures as the Haua Fteah sediments. SystemNeolithic with the Graeco-Roman structure. atic collections have also been made of surface lithic Cn"Octl cn/Octl 64 The Cyrenaican Prehistory Project 2010 artefacts in different zones of the Gebel Akhdar, to characterise the stone-tool technologies used in different periods and through this, how the different parts of the landscape were exploited in different periods. The objectives of the 2010 season of fieldwork were defined as follows: • the site collecting a series of high-definition overlapping scans or ‘point clouds’. In total, ten scans were collected at various locations in and around the Haua Fteah cave, collecting approximately 40,000,000 coordinates. These individual scans were joined together (registered) using Cyclone software. Within the cave there are a number of distinctive morphological elements that can be linked to active and previous geomorphological processes. to progress the new excavations in the Haua Fteah beside the Middle Trench (Trench M) and Deep Sounding (Trench D), in order to investigate (respectively) the nature of Oranian and PreAurignacian occupation and the environmental contexts in which they were situated; • to continue the investigation of coastal geomorphology and associated traces of human settlement; • to continue the survey of surface lithic material in the Gebel Akhdar begun in 2009, so as to create a transect of data from the coast to the Sahara; • to investigate the geomorphology and associated archaeology of a deep basin in the Gebel Akhdar, in order to provide a long geoarchaeological sequence for the interior of the Gebel Akhdar to compare with the coastal record emerging from the Haua Fteah and studies of the coastline. • Relatively smooth, gently curving surfaces make up the bulk of the cave and exhibit a range of diagnostic karst phenomena, including both extensive and locally concentrated flowstone deposits and stalactites where the walls arch over to form the roof of the cave. These surfaces are associated with a band of strongly karstic (generally hard, dense and well-jointed) limestone that runs around the cave at approximately mid-wall height and which dips gently from west to east. This limestone can be followed out of the cave on to the adjacent hillside, where it is characterised by a distinctive assemblage of micro-karstic phenomena including micro-rills (rillenkarren) and solution pans (kamenitzas). • There is a series of side caves and pipes associated with localised flowstone deposition and in places partial collapse of the overlying cave wall. These features are characteristically located within the karst limestone bed, although there is some indication of their concentration at mid-height within the bed where there is possibly a structural discontinuity. Below this level the limestone is less massive and possibly more granular in character (see below). • Below this level at the back of the cave is a zone that appears to be undergoing active weathering from surface flaking and limited granular disaggregation. This creates large-scale ‘honeycombing’ of the back wall in which numerous cavernous hollows (characterised by pure white surfaces) have breached an outer, case-hardened, surface typically light brown in colour. Typically this weathering zone extends to a height of 4–5 m and is located within the lower half of the karst limestone bed, although the selective nature of the weathering clearly identifies within this bed a number of more resistant strata that are left as prominent bands by the wearing back of the limestones above and below by cavernous weathering. • Above the actively weathering zone at the back of the cave is an overhanging area of loosely Ground-based LiDar survey of the Haua Fteah morphometry By Bernie Smith and John Meneely The central purpose of this preliminary investigation was the first use of ground-based laser surveying technology to produce a complete digital terrain model (DTM) of the Haua Fteah cave, for two purposes. The first was to provide a basis for future surface recording of precise locations within three-dimensional space, and for monitoring change through periodic re-survey (Meneely et al. 2008, 2009). The second was to combine this with a detailed geomorphological examination of the cave so as to better understand its origins and subsequent modification from the ongoing rock-weathering of the cave walls. The cave and part of the surrounding landscape were digitally documented using a Leica HDS 3000 ground-based LiDAR system. This rotating pulsebased laser system has a range of 100 m and a field of view of 360° horizontally by 270° vertically. It can measure 2000 surface coordinates per second, assigning a reflectance value to each point based on the return strength of the laser pulse from the surface. This survey method, normally referred to as ‘laser scanning’, involves traversing the laser scanner around 65 Graeme Barker et al. interlocked, angular blocks indicative of previous and potentially continuing block fall. • outflow of water from within the cave. This feature argues against the formation of the cave as an initial surface feature that progressively eroded back into the hillslope. It is possible for caverns to form in bedrock by the progressive wearing back of specific areas of cliff face to form features that are often referred to as ‘tafoni’. These are particularly common within Mediterranean and arid environments, but are typical of salt-rich environments in general and are considered by many to be diagnostic of salt weathering (Smith 2009). There appear to be maximum dimensions to tafoni as their growth is self-limiting through the eventual loss of significant thermal cycling, and their dimensions are invariably measured in metres or fractions of metres, rather than the tens of metres that characterise the Haua Fteah (Smith 1978). Although the active weathering at the back of the cave has a superficial resemblance to weathering observed within many tafoni, unlike in tafoni the observed flaking is single rather than multiple and below surface flakes the limestone remains coherent and, superficially at least, relatively unweathered. This could suggest a very slow rate of weathering consistent with the absence of frequent marked environmental cycling (heating/cooling, wetting/drying) within the stable cave interior. In the absence of these normal pre-requisites for processes such as salt weathering (at least in terms of accelerated breakdown), it is necessary to pursue alternative weathering mechanisms. One possibility could be the gradual loss of integrity as inter-granular cements are removed through seepage of groundwater. Alternatively, there may be much longer-term environmental cycling over possibly a seasonal timescale, possibly linked to the periodic flooding of the back of the cave by surface water inwash. It is intended to pursue these and other explanations for the cave’s morphometry through laboratory analyses and simulations. Samples of weathered stone and its by-products have been collected for mineralogical and structural analysis. Further evidence of the historic as well as active nature of block fall and collapse can be seen at the rear of the cave, where a large portion of the foot of the back wall comprises a single, large collapsed section of wall. This appears to be embedded within the sediment of the cave floor and is separated from the bedrock behind by a distinctive open fracture. The fact that it was able to fall away from the back wall and is now embedded within the sediment suggests that this major collapse must have in part, at least, pre-dated the latter’s deposition. The overall morphology of the cave, in combination with the surface characteristics of the cave walls and roof, is indicative of a karstic origin, and the prior existence of a subterranean cavern that has been exposed through collapse of the ‘frontal’ element of the cave roof. The prior morphology of the cavern can be clearly identified within the DTM. A horizontal section though the cave at approximately mid-wall height identifies the smooth curve of the karstic surfaces that make up the majority of the cave wall at this height, and its subsequent interruption by the collapse of a wide section of the back wall. A more detailed cross-section of the back of the cave illustrates the large section that has broken away from the cave and now lies partially buried within the cavefloor sediment. The cross section through the cave as a whole (Fig. 2) demonstrates the overall curvature of the original karstic cavern and how its curvature is now intersected by the present-day escarpment. A plausible explanation for the creation of the cave is the progressive retreat of the escarpment to the point where it either exposed the cave or weakened it to the point where it triggered collapse of the downslope section of the roof. This hypothesis requires further testing, but it is significant that the floor of the cave rises towards the lip of bedrock at the mouth of the cave, which acts as an effective barrier to any simple Figure 2. Cross-section laser scan through the Haua Fteah cave. The scan includes the profile of the main excavation trench under the overhang, and clearly shows the outline of the Upper Trench and Middle Trench (it was made early in the 2010 season, before the excavation of the top 2 m of the Deep Sounding (laser scan: J. Meneely). 66 The Cyrenaican Prehistory Project 2010 Archaeological investigations in the Haua Fteah Trenches M and D were both excavated by trowel, and all sediment was bagged and removed to the project base in Shahat for processing by water flotation. By Tim Reynolds, with Ian Brooks, Kate Connell, Ryan Rabett, Hazel Reade, David Simpson and Chris Stimpson Trench M A short season of work undertaken in the summer of 2009 to remove the last of the McBurney backfill from the Middle Trench revealed that the published dimensions of the Deep Sounding did not include an extension of the Deep Sounding made in the last week of the 1955 excavations. This caused a slight change in strategy. It was originally intended to retain a small baulk of backfill within the Middle Trench to facilitate excavation of Trench D, but this overlay part of the Deep Sounding extension, so was removed. The discovery of the full extent of the Deep Sounding made it advantageous to change the position of Trench D: the original position (the trench started in the final days of the main 2009 season) lay across a change in direction of the McBurney section edge, so Trench D was moved 0.5 m to the west to align with a step cut into the southern section edge by McBurney. This realignment provided the opportunity to create an L-shaped trench with one arm formed by the McBurney Deep Sounding and the other by the new Trench D. It has better light and easier access than the original Trench D, and because it is somewhat larger than the 2 m 7805:3 by 1 m area originally planned as it 32.229 cuts down to meet the McBurney step, a greater sample size of sediment will be examined. Trench M is located adjacent to the southwest corner of the Middle Trench. Because of logistical and safety factors, its size was reduced from c. 2 m by c. 1 m to c. 2 m by 0.6 m. The excavation techniques were consistent with those employed in 2009, apart from two thin contexts (10,007 and 10,008) which were dug stratigraphically. The 2010 excavations removed c. 0.7 m of sediment, giving a total depth of deposits excavated so far in Trench M of 1.1 m (Fig. 3). The deposits at the top of the 2010 excavation (context 10,006) are essentially a continuation of 10,005 excavated in 2009. These sealed a brown silty layer with few limestone fragments (10,007) and a charcoalrich layer (10,008), both of which sloped markedly to the east and north, suggesting that the underlying sediments may be banked against a stone block or similar obstacle outside the excavation area. Context 32.222 32.224 32.223 32.227 32.228 7805:3 32.22: 32.22; 32.232 32.233 Charcoal fragments Molluscs 2 Figure 3. Trench M at the end of the 2010 excavations: (above) composite drawing of the north-facing Section 63.1, the arrows marking the c. 40 cm step between the 2009 (upper) and 2010 (lower) sections (heights shown are metres above sea level); (below) photograph at the end of the 2010 excavations, looking south. In the photograph, the horizontal 1 m scale rests on context 10,006, and the top of the vertical 30 cm scale is beside the charcoal-rich context 10,008, which is visible sloping down to the left (east) (drawing: I. Brooks; photo: G. Barker). 67 72"eo Graeme Barker et al. 10,009, immediately below 10,008, forms a marked break in the types of materials recovered: the layers above contained considerable numbers of marine and terrestrial shells, whereas those in the lower part of the sequence contained very few, and there also appeared to be a change in the character of the lithic material, with a high percentage of larger, blade-like, flakes and core fragments being recovered than in the higher layers. Context 10,009 also contained a limestone block with a pecked groove around one end, of uncertain function, but reminiscent of the ‘tethering blocks’ used by prehistoric hunters in the Sahara to trap game (Le Quellec 1990). The lowest context excavated (10,011) was essentially a tumble layer with randomly orientated limestone blocks and flakes. It may be a concentration of rocky material at the base of a tip of material and thus be the base of context 10,009. Towards the base of 10,011 a distinct group of chert artefacts, including ten scrapers, was found in the vicinity of a small (c. 200 mm in diameter) circular patch of burnt sediment on the surface of 10,012. The cultural materials collected during the excavations, or from the hand-sorting of the heavy residues from the water flotation, are discussed later in this report, but can be broadly ascribed to the Oranian Epipalaeolithic phase of the McBurney sequence (see below, Preliminary Observations on the Capsian and Oranian Occupation in the Haua Fteah (Trench M)). Trench D A north-facing section of a remnant baulk from the McBurney excavations, which formed a step down into the deep sounding, was cleaned back and recorded (Fig. 4). Eleven stratigraphic layers were identified, described, and assigned context numbers (Fig. 5: Section 60.1). A trench measuring 1.30 m by 0.50 m was demarcated for excavation at the top of the section, with the long axis running west–east. It was divided into two units of equal size (1 to the east and 2 to the west), with these units further sub-divided into four quadrats of roughly equal size designated (from southeast to northwest) A to D. Each quadrat was hand-excavated following the stratigraphic layers revealed in section (two further contexts, 706 and 707, both stream channel-fills oriented north–south, were identified and recorded in plan). The trench was excavated to a depth of c. 85 cm, to the top of the McBurney step revealed in the summer 2009 season. Stone 695 Stone 707 696 697 698 699 51.432 51.432 Figure 4. (top left) Looking south into the Middle Trench of the Haua Fteah, showing the sampling work on the Deep Sounding in the foreground and, beyond, the excavations of Trench D (photo: G. Barker). 700 701 703 704 705 50 cm 0 68 Figure 5. (bottom left) Trench D: the north-facing Section 60.1; dashed lines mark uncertain boundaries between contexts (heights shown are metres above sea level) (drawing: R. Rabett and C. Stimpson). The Cyrenaican Prehistory Project 2010 chert pieces (Fig. 6) included a convex sidescraper made on a secondary flake and a borer made by notching on a tertiary chert flake. Neither piece would be out of place in either a Pre-Aurignacian or LevalloisoMousterian context. It is likely, given that the Trench D sediments are laid down by water-flows, that humans were sheltering deep under the overhang of the site and that the Trench D artefacts represent the washing of materials into this part of the cave rather than in situ knapping. 2 1 3 The McBurney Deep Sounding 4 5 0 6 Backfill was removed from the westernmost part of the Deep Sounding to a depth of 2.5 m, approximately half way down to the base of the McBurney excavation. The south-facing McBurney section exposed by this work was recorded, a sample column was dug for pollen and tephra, and five samples of sediment were taken for OSL dating (Fig. 7). (Zenobia Jacobs of the Wollongong team who are undertaking the 2 cm 7 Figure 6. Struck chert pieces from Trench D: 1. burin on a snapped bladelet (667); 2. burin on a flake (667); 3. transverse burin (695); 4. convex sidescraper (695); 5. borer (697); 6. fragment of flake knife (667); 7. retouched flake fragment (697) (drawings: T. Reynolds). The excavated layers appear to be predominantly the result of a series of mudflow, stream channel, and runnel deposits. Each layer consisted of variations of reddish-brown silt, generally uniform in colour and consistency within layers, although localised changes were observed. These changes were interpreted to be a result of in situ degradation of limestone rock and ephemeral water-flow events. Near the top of the excavation, in context 695, was an area of burnt limestone rocks, discrete patches of charcoal, compressed ash and staining, that may indicate the location of a hearth subsequently degraded by sedimentation and water-flow. McBurney classified the industry he found at the base of the Middle Trench and through the Deep Sounding as Pre-Aurignacian, though he referred to the sediments at the depth of our Trench D as essentially sterile. Our excavations, however, indicate small-scale but regular human presence at the time these sediments were being laid down. A few stone artefacts were found by the excavators, especially in the upper layers (695 and 696), and the flotation of the sediment bulk samples recovered micro-débitage from most of the Trench D contexts. The struck Figure 7. Looking north to Section 54, the southfacing face of part of the upper 2.5 m of the McBurney Deep Sounding. The McBurney metal labels are clearly visible on the facing and left faces. The vertical groove to the right of the ranging pole is the column sample excavated for pollen and other palaeoenvironmental indicators. The sediment on the right is backfill. Scale: 2 m (photo: G. Barker). 69 Graeme Barker et al. main OSL dating of the Haua Fteah sequence was prevented from joining the excavation because of the Iceland volcanic eruption, and samples were taken for her by the Royal Holloway team undertaking OSL dating in the landscape, Simon Armitage and Natalie Russell.) the Research Laboratory for Archaeology, University of Oxford as part of the RESET project funded by the Natural Environment Research Council. Micromorphology: preliminary results By Robyn Inglis Tephrochronology sampling Micromorphological analysis of sediment structure, supported by detailed geochemical and physical analyses, is a powerful technique to investigate sediment sequences within cave and rockshelters, which result from the complex interplay between anthropogenic and ‘natural’ influences (Bailey and Woodward 1997; Goldberg and Bar-Yosef 1998; Goldberg and Sherwood 2006; Stein 2001). The micromorphological investigations at the Haua Fteah are focussed on the characterisation of the site-formation processes of the Middle Palaeolithic layers and their transition to the Upper Palaeolithic as identified by McBurney (1967). The high-resolution information provided by this technique augments a broader-scale programme of geochemical and physical analyses of sediments spanning the entirety of the Middle and Upper Trenches, a sequence likely to span (more or less) the last 100,000 years. This nested approach allows the detailed characterisation of key layers and the linkage of specific micromorphological features with chemical and physical characteristics which can be extrapolated to the overall sequence, enabling the identification of long-term environmental trends, such as Oxygen Isotope Stages. Analysis of the thin sections and other sediment samples taken in the 2008 and 2009 seasons are ongoing, but a number of trends are emerging. The particle-size distributions of the non-carbonate fraction from the sample columns, for example, show striking similarities throughout the entire sequence, from both layers defined as inwash and layers formerly assumed to be of windblown origin. This implies that the fraction may be part of sediment which, although originally deposited in Cyrenaica by dust transport from the Sahara, was also subject to intensive local reworking and recycling, eventually deposited in the cave by a number of different processes which cannot be differentiated solely through particle size. Sediment samples collected from across the Gebel Akhdar, both ancient and modern, are being investigated to test this hypothesis of sediment recycling. Micromorphology indicates that the sediments in the basal c. 70 cm of the Middle Trench (contexts 528–521, more or less equivalent to McBurney layers XXXV–XXXIII), were potentially dominated by periods of inwash, with By Christine Lane Tephrochronology is the use of layers of volcanic ash (tephra) in stratified sediments as event horizons, which are potentially able to provide both absolute dates and stratigraphic tie points between sequences over wide areas. Volcanic ash entrained within the atmosphere can be transported thousands of kilometres from the eruption centre. Deposition of tephra after an eruption is effectively instantaneous, providing an isochronous marker horizon that may be traceable over extremely wide areas. The thickness and grain-size of a tephra layer decrease away from source, with ‘cryptotephra’, deposits invisible to the human eye, only detectable through microscopic analysis in the laboratory. Subsequent geochemical analysis of a tephra deposit allows identification of the volcanic source and in most cases, correlation to an eruption event with an associated absolute age, which can then be directly transferred into the host sediment sequence. Preliminary investigations on samples from the 2009 season indicated that at least four cryptotephra layers are present within the Haua Fteah stratigraphy, with the largest layer identified in the Oranian deposits of the west-facing section of the Middle Trench, and also in sediments from Trench M. Initial geochemical analyses suggest that these tephra layers originate from eruptions of the Campanian volcanic region, in southern Italy, and the Hellenic Arc, in Greece. This research therefore provides the first evidence for Mediterranean tephra layers in Libya and highlights the potential of tephrochronology as an important tool in dating archaeological and environmental sequences both within Libya and across North Africa and southern Europe. Given these promising results, further samples were collected during the 2010 season to secure the positions and confirm the identification of the cryptotephra layers in the Haua Fteah. Continuous 2 cm sediment samples were taken across each of the cryptotephra horizons in the west-facing section of the Middle Trench and the 2009 tephra sampling column from Trench M was continued as excavations deepened. Samples will be processed at 70 The Cyrenaican Prehistory Project 2010 Molluscs in the Haua Fteah significant hiatuses in deposition marked by periods of occupation. The next 50 cm (contexts 520–509, more or less equivalent to McBurney layers XXXII– XXX), were dominated by an increase in limestone gravel content, potentially representing a period of increased physical weathering in the cave, with the fine fraction derived from aeolian deposition and reworking, both indicating relatively slow deposition of sediments and potential environmental degradation. This phase was capped by a large lens (contexts 513 and 536, regarded by McBurney as within Layer XXX) that appears to represent a major anthropogenic burning event. Micromorphology indicates that this and other burning events visible in the sections in fact consist of complex series of ash and charcoal layers that have been subject to varying degrees of reworking. Semivesicular silicaceous material within some of these layers, probably resulting from diagenesis of burnt plant material, is likely to be the ‘slag’ reported by McBurney from burnt layers. Lenses of phytoliths within the burnt layers offer considerable potential for detailed study of fuel materials. The sediments investigated by micromorphology above the burning episode (contexts 508–442, McBurney Layers XXIX–XXV) are characterised by few definitive signatures of human impact and a succession of reddish inwash layers and paler more gravelly and aeolian deposition layers, punctuated by periods of calcite deposition from internal precipitation of water dripping from the roof of the cave. This complex story of deposition, stabilisation and remobilisation of sediments is likely to be related to rapid environmental changes and suggests that McBurney’s identification of inter-stratified assemblages at the Middle/Upper Palaeolithic transition as likely to be an artefact of sediment re-deposition rather than alternating occupations by people with Middle and Upper Palaeolithic lithic industries (McBurney 1967, 125) could well hold true. Alongside the completion of the micromorphological sampling programme, comparative material was sampled from a large slump of material which had washed into the trench as a result of winter storms, forming cones of laminated colluvial material in the northern corners of the Middle Trench, capped by a 2–3 cm-thick cracked crust. These cones, and the surface gullies in the unconsolidated cave surface outside the trench, highlight the massive energy required to move sediment en-masse from the surrounding landscape into the cave, and the potential alteration of the surface archaeology and truncation of underlying sediments that could accompany such inwash events. By Chris Hunt and Evan Hill The brief summary by Hey (1967) of molluscs found in the Haua Fteah sequence includes species lists for the main cultural periods, but with little indication of abundance, taphonomy, palaeoecology, or human use. Numerous research projects have subsequently indicated the value of molluscs as ecological indicators and as sources of ancient human nutrition and ornament. Hence the present project is recovering both large specimens in the hand excavation and all mollusc fragments larger than 2 mm and all smaller diagnostic specimens larger than 0.5 mm from the flotation residues. Systematic samples of molluscs have now been recovered from sample columns in the Upper and Middle Trenches and from the excavation of Trenches M and D. Some preliminary observations are provided here. Molluscs are rare in the few Trench D (Pre-Aurignacian) contexts studied so far, but include relatively frequent Rumina decollata, plus some Helix melanostoma, Theba pisana and Eobania sp., suggesting relatively open but vegetated environments. Marine molluscs include rather fragmented and often burnt Patella and Osilinus, with minute whole specimens of Conus mediterraneus and Natica spp. In Trench M, marine molluscs, mostly Osilinus but also some Patella, are present throughout the Oranian and Capsian layers (though comparatively rare early in the Oranian) and other marine taxa are also present intermittently. Noteworthy finds are beads made from scaphopod molluscs and from the gastropod Columbella rustica. Land molluscs from the Oranian and Capsian layers – abundant Helix melanostoma and Theba pisana, plus some Eobania sp. and rare other species – suggest open rather dry environments. In the sample column in the Upper Trench, marine molluscs (especially Osilinus turbinatus followed by Patella spp.) are common in the Capsian and Neolithic layers, with Patella increasing in frequency in the Neolithic; land molluscs are dominated by Helix melanostoma, which may have been used for food, but other species indicative of relatively open, dry, environments are intermittently present. Marine molluscs are rare in the Graeco-Roman and Historic layers. This preliminary analysis suggests that marine molluscs were routinely part of the human diet throughout the layers studied so far, as well as being used as ornament in the Oranian and Capsian phases. Helix melanostoma and possibly other landsnail species were also routinely part of the diet. The limited information from terrestrial mollusc palaeoecology is consistent with rather open but vegetated 71 Graeme Barker et al. landscapes throughout, though the number of contexts studied, particularly in the lower part of the sequence, is too low for more secure conclusions to be reached at this stage. contexts (10,007–10,011) by GM, and the material from 10,006, at the Capsian/Oranian transition, studied by both. Given the remarkable density of material from Trench M, the study concentrated on the finds from just one of the eight 40 cm by 40 cm quadrats: square 2A in 10,003 and 10,004, and square 2B in the rest of the sequence (the quadrats selected by the excavators as typical of the spatial characteristics of the context). In total some 17,000 lithic pieces were examined from these quadrats, analysed in terms of their technological and typological characteristics. The dominant material throughout the sequence is the light brown and grey (Munsell 7.5YR 6/4) Eocene chert found in tabular and semi-tabular bands in the vicinity of the cave, but a finer translucent chert, pinkish-grey in colour (7.5YR 7/2), was also used in the upper Capsian contexts (10,001, 10,002, 10,003). Preliminary observations on the Capsian, Oranian, and final Dabban occupation in the Haua Fteah (Trench M) The lithic assemblages By Giulio Lucarini and Giuseppina Mutri A preliminary study of the lithics collected from Trench M in 2009 and 2010 was carried out at the end of the 2010 season, the upper (Capsian) contexts (10,001–10,005) by GL, the lower (mainly Oranian) Figure 8. Lithics from Trench M, contexts 10,001–10,005 (Capsian) and 10,006 (transition from Oranian, see text for discussion): 1. opposed platform core (10,004); 2. multiple platform core (10,005); 3. core side (10,005); 4. crested blade (10,001); 5. crested blade (10,005); 6. endscraper (10,003); 7. denticulated endscraper (10,005); 8. endscraper (10,001); 9. notch (10,005); 10. arched backed blade passing to segment (10,005); 11. scalene blade (10,005); 12. double truncation on backed blade (10,005); 13. backed flake (10,003); 14. pointed backed bladelet (10,001); 15. reversed backed blade (10,004); 16. fragment of bladelet with Ouchtata retouch (10,001); 17. burin on backed flake (10,005); 18. drill on backed blade (10,005). Oranian lithics in Trench M (context 10,006): 19. opposed platform core; 20. 90° core; 21. rectilinear backed blade; 22. arched backed blade; 23. truncated backed bladelet; 24. reversed backed blade; 25. Gibbeux backed blade; 26. scalene blade; 27. backed blade; 28. truncated backed blade; 29. Krukowski microburin (drawings: G. Lucarini). 72 The Cyrenaican Prehistory Project 2010 In the Capsian contexts (Fig. 8: 1–18; and see also Figs 10 and 11 in Barker et al. 2009), cores are represented by seventeen specimens, mainly of opposed and 90° platform types, followed by single and unpatterned multiple platform cores of reduced dimensions and mainly regular prismatic shape. The platform surfaces are characterised by tiny or continuous traces of preparation. The removals on the flaking surfaces indicate that the cores were intensively exploited, mainly for bladelets. The frequency of core-trimming elements (mainly crested blades and core sides), and the low incidence of primary flakes and blades, confirms the intensity of core exploitation. Debitage makes up between 87% and 93% of the assemblage, with numerous burnt chunks and chips of chert, and fire-cracked pieces, occurring especially in 10,002–10,004. Flakes (mainly single and opposed platform types, followed by flakes with one with 90° or multiple platforms) are common, but their frequency declines with depth, from 31% of the debitage in 10,001 to 13% in 10,006. The frequency of blades (>25 mm in length) and bladelets also declines with depth, from 20% of the assemblage in 10,001. Blades are more common than bladelets. They are highly standardised, and mainly used as blanks for backed tools. The percentage of chips and chunks increases with depth. Retouched tools, analysed using Tixier’s (1963) typology for the Maghreb Epipalaeolithic, comprise around 8% of the assemblage in the upper Capsian layers, increasing to almost 13% in 10,005, but falling to 6% in the transitional context 10,006. The most frequent types are backed elements, the commonest of these (16–29%) being simple (rectilinear) backed blades/bladelets, followed by arched backed types, including both the normal and the ‘passing to segment’ type. The typical Capsian scalene blade/bladelet (with the short side manufactured both on proximal or distal end) is well represented in the lower Capsian context (10,005) but also present in low frequencies in 10,006, the transitional Oranian layer. A second type indicating continuity from the Oranian assemblage is the reversed backed blade/bladelet and flake, often characterised by an abrupt or semi-abrupt (60°) slightly invasive retouch on the ventral face and with detachments reducing in size towards the distal end. This is most frequent in 10,006 (23%) and 10,005 (15%), declining to 6% in 10,001. Another class of tool found across the Capsian/Oranian transition is the blade/bladelet with Ouchtata retouch (12% in 10,006, 7% in 10,005). Simple and Krukowski microburins are also found in 10,006 and 10,005. Truncation on blades/bladelets and backed elements is present in 10,006 and common in 10,005 and 10,004. The main element differentiating the upper Oranian and lower Capsian layers, however, is the appearance in the latter of denticulates and notches on flakes and blades (and more rarely on backed elements), forms almost completely absent from 10,006; comprising 6% and 2% respectively in 10,005, their incidence increases to 16% and 10% in 10,001. Amongst these is a particular type defined as a ‘saw’, characterised by a continuous and non-invasive scaled retouch, a typical element in the Maghreb Capsian. The frequency of cores in the Oranian contexts (Fig. 8: 19–29, and Fig. 9) is low but constant (10,006: 0.4%; 10,007: 0.2%; 10,008: 0.3%; 10,009: 0.5%; 10,011: 0.9%) and there is a strong presence of residual cores (10,006: 66%; 10,008: 75%; 10,009: 60%; 10,011: 33%). This factor, along with the almost total absence of primary and secondary blanks and scarcity of core-trimming elements, suggests that the first stage of core reduction probably took place elsewhere. Opposed platform cores are the most common, but there are also single platform cores, 90° platform cores and, in the lowest layers, some discoidal and Levallois cores. Blanks are dominated by flakes (10,006: 17%; 10,007: 13%; 10,008: 25%; 10,009: 16%; 10,011: 17%), but bladelets are also represented (10,006: 8%; 10,007: 4%; 10,008: 0.4%; 10,009: 9%; 10,011: 3%). Blades are scarce except in 10,011, and the bladelets in 10,009 and 10,011 are also larger. Chips and chunks indicative of manufacturing activities are abundant throughout the Oranian layers. The retouched tools are mainly represented by a wide range of backed tools, as is normal in Oranian assemblages. In the upper layers their blanks are usually bladelets, though small flakes were also used for some tools such as reversed backed bladelets and arch-backed bladelets passing to segments. In the lower layers (10,009 and especially 10,011) backed tools were mostly made from blades. The backed tools are dominated by simple rectilinear backed bladelets (10,006: 27%; 10,007: 47%; 10,008: 38%; 10,009: 28%; 10,011: 28%) and reversed backed bladelets (10,006: 21%; 10,007: 17%; 10,008: 10%; 10,009: 13%; 10,011: 8%). Arch-backed flakes in the lower layers (10,009: 2%; 10,011: 5%) are replaced in the higher layers by arch-backed bladelets (10,008: 5%; 10,007: 4%; 10,006: 10%), along with arch-backed elements passing to segment (10,006: 5%; 10,007: 3%) and real segments (10,006: 2%; 10,008: 2%), indicating the development of a geometric component in the upper Oranian layers. Ouchtata bladelets and denticulate backed bladelets are also well represented, microburins and Krukowski microburins are present, and other retouched tools include burins, continuously retouched pieces, and notches. 73 Graeme Barker et al. Figure 9. Lithics from Trench M, contexts 10,006–10,009 (Oranian) and 10,011 (transition from Dabban) (see text for discussion of the transition): 1. double truncated bladelet (10,011 spit 2); 2. backed blade (10,011 spit 2); 3. crested blade (10,009 spit 7); 4. retouched blade (10,009 spit 7); 5. backed bladelet (10,009 spit 5); 6. truncated bladelet (10,009 spit 7); 7. denticulated endscraper 10,009 spit 6); 8. reversed backed bladelet (10,009 spit 6); 9. denticulated backed bladelet (10,009 spit 6); 10. denticulated endscraper (10,011 spit 2); 11. rounded endscraper on a blade blank (10,011 spit 2); 12. rounded endscraper on a flake blank (10,011 spit 2); 13. discoidal core (10,011 spit 2); 14. backed blade (10,011 spit 2); 15. scalene (10,008 spit 2); 16. backed bladelet (10,008 spit 2); 17. arc backed bladelet (10,008 spit 2); 18. reversed backed bladelet (10,008 spit 2); 19. blade with Ouchtata retouch (10,008 spit 2); 20. arc backed bladelet (10,007); 21. piquant trièdre (10,007); 22. backed flake (10,007); 23. backed blade with notch (10,007); 24. reversed backed flake (10,008 spit 2) (drawings: G. Mutri). 5 8 7 6 2 3 1 4 4 9 12 10 0 2 cm 11 11 16 15 14 13 19 17 20 21 22 24 18 23 The fauna Context 10,011 is strikingly different from the layers above it in the frequency of endscrapers, which are absent in 10,009–10,006. They are well made, on both big flakes and blades, with a regular semi-circular front and a semi-abrupt retouch that is sometimes denticulate. Together with the changes in debitage mentioned above (more blades, bigger bladelets) the numbers of endscrapers suggest that 10,011 represents a distinct change in lithic tradition, namely that it is a final horizon of the Dabban Upper Palaeolithic. It is noteworthy, however, that the transition is not abrupt: the Dabban layers excavated by McBurney include many backed bladelets that form a classic part of Oranian assemblages; chamfered pieces typical of the Dabban are absent from 10,011; and new elements appear gradually, the increase in the dimensions of the bladelet elements and the disappearance of more geometric tools starting in 10,009. By Ryan Rabett The faunal data from contexts 10,000–10,004 in Trench M (McBurney Capsian Layers IX and X) indicate that there was a peak in midden formation during the latter part of the period. During this time low quantities of fish and crab remains begin to appear in the record beside long-term staples such as Barbary sheep (Ammotragus lervia), gazelle (Gazella sp.) and hartebeest (Alcelaphus sp.) (Table 1). If Trench M is typical of the cave as a whole, occupation during the earlier part of the Capsian phase appears to have been more ephemeral, with lower frequencies of bone being deposited. A recent study of the exploitation of Barbary sheep at the site suggests that during the Capsian phase most were being taken during the summer months (Wall 2004), raising the possibility that the use of the cave may have been episodic, potentially part of a more far-ranging 74 The Cyrenaican Prehistory Project 2010 Table 1. Animal bones recovered from Trench M in the Haua Fteah in 2009 and 2010: preliminary species representation in the Capsian (contexts 10,000–10,004) and Oranian (10,005–10,009) phases of occupation. NISP: number of identifiable specimens; NSP: number of specimens. Class/Order Family Taxon Common name Trench M Contexts (10,000–10,004) Contexts (10,005–10,009) NISP NISP % % Chondrichthyes - Fish Fish 2 3.03 0 0 Decapoda - Decapoda Crab 5 7.57 0 0 Testudines Testudinidae/ Geoemydidae Testudinidae/Geoemydidae spp. Tortoise/hard-shell turtle 11 16.66 27 35.52 Testudo graeca cyrenaica Mediterranean spur-thighed tortoise 0 0 1 1.31 Testudinidae Testudo sp. Tortoise 13 19.70 2 2.63 Squamata - Snake/lizard Snake/lizard 5 7.57 3 3.94 Galliformes Phasianidae cf. Alectoris sp. Partridge 0 0 1 1.31 3.94 Columbiformes Columbidae Columbidae spp. Pigeon/dove 1 1.51 3 Falconiformes Accipitridae Accipitridae spp. Hawk 1 1.51 0 0 Strigiformes Strigidae cf. Athene sp. Owl 1 1.51 0 0 Aves - Bird Bird 7 10.60 1 1.31 Carnivora Felidae Felis sp. (cf. Felis lubica) Wild cat 1 1.51 0 0 Artiodactyla Bovidae Rodentia Muridae Bovidae spp. Bovid 4 6 1 1.31 Bos primigenius Aurochs 0 0 1 1.31 Alcelaphus sp. Hartebeest 1 1.51 2 2.63 Gazella sp. (cf. Gazella dorcas) Gazelle 3 4.54 5 6.57 Ammotragus lervia Barbary sheep 10 15.15 20 38.15 Gerbillus sp. Gerbil Total NISP 1 1.51 0 0 66 99.88 76 99.93 NSP % Ungulate 173 Large mammal Intermediate mammal NSP % 4.36 139 5.94 145 3.69 95 4.06 9 0.22 4 0.17 Unidentiiable 3635 91.74 2102 89.82 Total NSP 3962 100.01 2340 99.99 Total (NISP & NSP) 4028 system of hunting across the Gebel Akhdar. The limited data for bone element fusion available from the Trench M sample indicate that young and sub-adult individuals were being routinely taken, older animals perhaps less so. The presence of juvenile and young Ammotragus does lend support to the use of the cave in the summer/autumn months, though cannot rule out its use at other times of the year. With the other ungulates, too, there are few older individuals in the sample, suggesting that foragers were employing a selective hunting strategy. Towards the base of the deposits excavated in 2009, particularly in context 10,005 at the transition from the Oranian to the Capsian in the lithic material, the intensity of site use appears again to rise. Although little Oranian material has yet been studied, the proportion of remains attributable to Ammotragus appears to be 2492 considerably higher than in the Capsian. The cause of this shift might relate to changes in game availability and hunting preference and/or to changing strategies in carcass processing, resulting in the survival of larger numbers of identifiable elements. Although little weight can be placed on the occurrence of a single specimen, the presence of aurochs (Bos primigenius) in the Oranian contexts and its absence from the Capsian, if confirmed, is noteworthy. Across its wide range, from North Africa to China, this species is thought to have favoured woodlandbordered grasslands, low-lying floodplains and river margins (e.g. Hall 2008; Zong 1984), and Dobson and Wright (2000) note its frequency in the Maghreb in the early Holocene in response to ameliorating (wetter, and more vegetated) conditions. Higgs (1967) observed that large bovines were more common in 75 Graeme Barker et al. 0 Bone surface traces attributable to trampling (e.g. see Olsen and Shipman 1988), such as abbreviated superficial striations close to the edge of a fragment, are limited enough in the study sample to suggest that occupation, at least in this part of the cave, was probably never intensive. The common occurrence of root-etching on bone fragments from the Capsian levels provides some support for this, implying that there was a persistent, if minimal, covering of vegetation present in this period, whereas such evidence diminishes significantly during the Oranian phase. The reason for this change is currently unclear. It is possible that changes in the spatial arrangement of site occupation during the Oranian phase may have kept plant growth down in the part of the cave we are excavating, but the low occurrence of bone-surface modification consistent with trampling suggests that increased levels of foot-traffic are an unlikely cause. 5 mm Figure 10. Cut marks on the medial epicondyle of an unfused ungulate metapodial (HFT-29189) from Trench M context 10,008. Biometric measurements on this specimen suggest that it was probably Barbary sheep (Ammotragus lervia) (photo: R. Rabett). The macrobotanical remains By Jacob Morales In total 810 litres of sediment from Capsian contexts 10,000–10,005 in Trench M were processed by flotation, and the resultant flots were examined by JM for macrobotanical elements such as seeds and fruits, with the wood charcoal then being analysed by Caroline Cartwright (British Museum). All of the plant remains retrieved from the flots were charred. The presence of plant remains is consistent though quite low (0.27 items per litre), but preservation is excellent. The methodology used in seed identification follows Morales and van der Veen (2008) and Morales and Barker (2009). The list of identified plants is shown in Table 2. The main plant remains in the Capsian contexts, present in 41 of the 47 analysed samples, are seeds and seed scales (a wooden structure that protects the seeds) of Aleppo pine (Pinus halepensis Mill.). Seeds from P. halepensis are edible and have been a traditional food in North Africa and Mediterranean Europe (Boulos 1983; Rivera and Obón 1991). The pine cone containing the seeds and seed scales opens during hot summer days to spread the seeds, but this process can be accelerated by throwing the cone into a fire. The latter facilitates the collection of the seeds and makes them more palatable, but it also increases the chances of preservation of seed scales and seeds through charring. The presence of burnt seed-scale fragments in almost all of the samples suggests that Capsian foragers may have employed this practice. At the same time seeds can be stored inside the cones for almost a year if the cones are collected before they are ripe, and this strategy may also have been the Oranian than the Capsian levels in the Haua Fteah, and in the sample recovered from the 2009 Trench M excavations, the proportion of large dense bone fragments (as revealed through measurement of unidentified bone fragment frequency and size to weight) is higher in the Oranian (and possibly early Capsian) layers than in the later Capsian occupation, suggesting an increased availability of large ungulates. Chelonian remains are comparatively high during both periods and include tortoise (Testudo graeca cyrenacia), common in the local landscape today. The distinctiveness of chelonian carapace and plastron dermal plate elements makes them easy to identify even when highly fragmented, unlike most mammalian bone, so the high incidence of this species may at least partly reflect recovery bias. However, the possibility that chelonians were included as a staple in forager diet should not yet be discounted. Evidence of butchery in the form of cut marks has so far been identified on 43 pieces of bone from Trench M (Fig. 10). Taken together with data on body part representation, this suggests that whole carcasses up to the size of Alcelaphus were being brought back to the site to be disarticulated. Large bovid remains, however, are predominantly represented by head elements only, as also noted by Klein and Scott (1986) in their analysis of the fauna excavated by McBurney. It can be inferred that carcass transportation was affected by game size and/or the distance away from the cave where it was killed. 76 The Cyrenaican Prehistory Project 2010 Table 2. Provisional identification of the charred macrobotanical remains from Trench M contexts 10,000–10,005 (Capsian). Context 10,000 10,001 10,001 10,001 10,002 10,003 10,004 10,005 10,005 10,005 Spit 1 1 2 3 1 1 1 1 2 3 No. samples 5 4 8 7 1 1 2 3 8 8 total 47 Volume (litres) 98 64 140 113 1 8 48 22 177 139 810 Crop plants Hordeum vulgare, grain 7 7 Hordeum vulgare, rachis 1 1 Triticum sp., grain 1 1 Cereal node 1 1 Wild species Amaranthus sp. 1 Chenopodium murale 1 Cistus sp. 5 1 Cupressaceae leaf fragment 7 5 Fumaria sp. 1 1 1 2 3 1 7 1 16 1 Galium sp. Indeterminate pulse 2 Juniperus phoenicea, seed 6 Juniperus phoenicea, fruit fragment 3 1 1 2 Neslia paniculata 1 Phalaris sp. 1 Pistacia lentiscus, seed 1 1 6 2 3 3 15 1 1 2 8 1 2 2 1 1 1 2 1 1 1 3 2 1 Pinus halepensis, seed Pinus halepensis, seed scales 1 1 Lens-Vicia sp. Myrtus communis Pistacia lentiscus, fruit 1 1 1 3 Plantago sp. 9 1 8 2 4 13 4 6 16 58 1 1 Poaceae 6 6 cf. Quercus sp. 1 1 Scorpiurus muricatus 2 2 Small-seeded legume 6 Vicia sp. 2 1 1 1 Vicia-Lathyrus sp. 1 2 Indeterminate seed/ fragments 3 3 Total no. seeds per litre of sediment 9 1 1 1 2 3 3 1 6 10 14 45 1 6 3 1 3 10 59 20 21 19 5 3 5 14 30 45 221 0.60 0.31 0.15 0.16 5 0.37 0.10 0.63 0.25 0.32 0.27 employed. Nowadays there are no natural forests of P. halepensis growing in the vicinity of the site. It is assumed that P. halepensis is not native to Cyrenaica and was introduced in the early 20th century by Italian colonisers (Charco 2001), but the presence of Pinus sp. charcoal in Oranian levels in the Haua Fteah suggests that pines were in fact available in the past, and the identification of seeds and seed scales of P. halepensis confirms the ancient presence of this species in Cyrenaica. Seeds from wild pulses or large-seeded legumes are the second most important group of wild food plants identified in the Capsian levels. This group of plants includes taxa such as wild vetch (Vicia sp.), wild grass pea (Lathyrus sp.), and wild lentil (Lens sp.). These are native Mediterranean plants that produce edible seeds, and their remains have been identified in Epipaleolithic and Mesolithic occupations in caves in Spain (Aura et al. 2005) and France (Vaquer and Ruas 2009). They are not as abundant as pine in the 77 Graeme Barker et al. facies (Barker et al. 2009: 69–72, 92–93). This understanding was tested and expanded in the 2010 field season by the exploration and recording of coastal deposits at sites 1831–1851 between the western side of the Derna urban area and the village of Chersa, completing the project’s exploration of coastal sites between Susah and Derna. Selected coastal Quaternary sequences were sampled in detail for artefacts, OSL, U/Th and radiocarbon dating, and for analysis of palaeoenvironments using molluscs and thin-section analysis. The stratigraphy of the coastal sections recorded in 2010 is broadly similar to that seen further west, with complex and archaeologically-rich sequences: a basal raised beach is overlain by palaeosoils and gravels, this sequence is cut by a second raised beach, and the latter is typically overlain by aeolianites and then alluvial fan gravels. In the Wadi Derna, the Hajj Creiem section and sites CPP1816 and CPP1817 were sampled for U/ Th and OSL dating. The Quaternary sequences at these sites consist of a thick tufa dam behind which shallow-water lacustrine sediments, up to 40 m in thickness, are stacked. These sequences reflect sediment deposition within surface water bodies that appear to have been perennial, so likely to represent important resources for early humans. Human activity during the accumulation of these deposits is represented by the presence of in situ lithics at various levels. As the time period represented by these sequences is a key research question, a coupled dating strategy was applied in Wadi Derna and at Hajj Creiem, with samples for OSL and U/Th dating being taken from the top and bottom of the exposed sequences. The combination of two independent dating techniques should increase the reliability of the derived chronology, as well as providing maximum and minimum ages for the artefacts associated with the sediments. Holocene environments are an important component of the CPP programme and coring of lagoonal sediments near Giargiarumah and a doline fill west of the Haua Fteah in previous seasons has yielded sequences in facies unsuitable for complete pollen preservation through the sampled interval. In the 2010 season, therefore, two dolines at Brak Notah Capsian deposits in the Haua Fteah, but this may be a consequence of pulses having fewer opportunities to be charred in food activities because they are generally eaten raw or in soups. Certainly they are very rich in protein and could make an important contribution to the forager diet. The third important group of wild plants identified in the samples consists of Mediterranean shrubs or small trees such as juniper (Juniperus phoenicea L.), the mastic tree (Pistacia lentiscus L.), and myrtle (Myrtus communis L.). These are all common in the vicinity of the site today, and produce edible fruits and seeds so they may have been gathered for food, though wood charcoals from these species have been identified in the same levels (Cartwright and Hunt 2008), so it is possible that some seeds or fruits were introduced to the site as part of fuel wood. One seed of P. halepensis and one seed of wild vetch (Vicia sp.) from context 10,005 have been dated by 14C and have yielded dates that confirm the antiquity of their exploitation (Table 3). Seven seeds and one rachis fragment of morphologicallydomesticated barley (Hordeum vulgare) and one seed of wheat (Triticum durum) have been in contexts 10,000 and 10,003. A few remains of cultivated plants were also detected in late Capsian levels in Haua Fteah excavated during 2007 (Morales and van der Veen 2008) and at Hagfet al-Gama, excavated during the 2008 season (Morales and Barker 2009), but radiocarbon dating of four barley grains and one wheat rachis fragment from the two sites demonstrates that they are recent intrusions. Clearly the cultivated plants in contexts 10,000 and 10,003 also need to be dated directly. Geoarchaeological investigations on the Cyrenaican littoral By Annita Antoniadou, Simon Armitage, Ian Candy, Chris Hunt, Natalie Russell, Evan Hill and Graeme Barker Research during previous seasons led to a provisional understanding of the Quaternary and archaeological sequences of the Cyrenaican coastland and a model of sedimentation during typical Quaternary climate Table 3. Radiocarbon dates on charred plant remains recovered by the Cyrenaican Prehistory Project. Site Context Sample no. Species Lab. no Date bp Date cal. BC/AD Date cal. BP Haua Fteah 10,000 3003 Hordeum vulgare UBA-15024 83±25 AD 1691–1924 Haua Fteah 10,000 3001 Hordeum vulgare UBA-15023 241±26 AD 1528–1955 422–4 11 11 Triticum durum UBA-15026 1069±27 AD 896–1021 1055–930 Haua Fteah 10,005 3111 Vicia sp. UBA-15027 12,202±38 12,262–11,931 BC 14,211–13,380 Haua Fteah 10,005 3117 Pinus halapensis UBA-15028 13,282±40 16,743–13,655 BC 16,743–15,604 Hagfet al-Gama 78 260–27 The Cyrenaican Prehistory Project 2010 c. 11 km west of Susah were visited and an initial coring programme was implemented in the south doline. Both dolines are water-filled, providing conditions favourable for high-quality organic preservation in the infill-sediments. Two parallel cores of 1.47 m and 1.85 m were obtained from brown Phragmites peat in the fringing reedswamp, and two parallel cores of 2.86 m and 2.95 m were obtained from algal muds in 2 m of water at the edge of the reedswamp. 0 2 cm 2 1 3 Geoarchaeological investigations on the Gebel Akhdar By Sacha Jones, Kevin White, Robyn Inglis and Kate Connell 4 Prior to the 2010 season of fieldwork, the geoarchaeological survey of prehistoric occupation evidence across the Cyrenaican landscape has concentrated on the coast and, in 2009, on the wadi systems that drain southwards from the Gebel Akhdar to terminate in pre-desert or desert palaeolakes. The focus of the 2010 work was to link these previous studies, in order to provide a broad transect of data from the coast to the desert edge. This was achieved by survey down the c. 9 km length of the Wadi el-Qlaa from its headwaters north of Lamludah down the escarpments and structural benches of the northern slopes of the Gebel Akhdar to where it debouches on the coastal plain, and a second transect south of Lamludah, from the Wadi el-Qlaa headwaters across the steep environmental gradient characterising the transition from the Gebel Akhdar to the pre-desert (Fig. 1). Survey was also undertaken in the rugged terrain immediately to the south of the Haua Fteah, west of Wadi el-Mahbul, in order to assess the archaeological characteristics of the landscape within daily walking distance of the Haua Fteah. Samples of lithics (and sometimes pottery) were collected from nine ‘sites’ or concentrations of surface material (Fig. 11); some preliminary observations on selected sites are given here. The Wadi el-Qlaa drains a catchment of approximately 1800 ha, crossing three main structural benches resulting from neotectonic activity. The channel system first heads along the upper surface, averaging 650 m above mean sea level (amsl), then cuts through a steep escarpment to an intermediate surface averaging 520 m amsl, then cuts through another major escarpment down to the coastal plain surface c. 100 m amsl. There are remnants of a major tufa dam that blocked the wadi system at the top of the escarpment between the intermediate surface and the coastal plain. Dense maquis vegetation in many parts of the transect made survey conditions difficult and ground visibility poor. 5 6 Figure 11. Lithics from the geoarchaeological survey of the Gebel Akhdar, all likely to be of late Pleistocene (Dabban, Oranian, Capsian) age: 1. bladelet core (CPP8111); 2. backed bladelet fragment (CPP811); 3. pointed backed flake (CPP8113); 4. side-retouched bladelet fragment with Ouchtata retouch (CPP8113); 5. chamfered blade (CPP8117); 6. bifacial foliate (CPP8117) (drawings: S. Jones). The highest density of material was found at CPP8111, an extensive lithic scatter on the undulating hilltop and hillsides marking the headwaters. The small blade cores, blades, and backed blades (made from limestone and, especially, chert) indicate a later Pleistocene age, possibly equivalent to the Dabban and/or Oranian phases at the Haua Fteah. CPP8104 and CPP8103, on hillslopes adjacent to the wadi channel, and CPP8100, CPP8101, and CPP8102 to the northwest of the headwaters, are all low-density lithic scatters including microblades on low-quality chert, likely to be Holocene in age. Isolated sherds of Roman pottery occur in and around patches of cultivable soil built up behind simple check dams across narrow wadi channels. To the north, CPP8108 is a low-density lithic scatter on a hilltop beside the increasingly incised wadi channel. Faceted flakes and long blades were recovered, but not cores, and microblade technologies are absent. CPP8107, on a 79 Graeme Barker et al. plateau west of the Wadi el Qlah, yielded large cores and flakes on limestone and chert of several colours and varieties. The paucity of cores in the latter stages of reduction and of retouched flakes suggests that the site preserves evidence of the primary phases of core reduction. At the northern end of the wadi near the tufa dam, CPP1805 consists of in situ lithics in a palaeosol underlying a horizontally bedded and heavily indurated calcrete duricrust (a hard layer formed by the precipitation of calcium carbonate) within the thick tufa sequence. The sedimentary sequence suggests that a swamp or shallow lake formed behind the tufa dam, similar to the situation at Hajj Creiem. Below the dam, lithics manufactured on poor-quality chert and high-quality blue-grey chert, together with potsherds, were found on the surface of 30 m-thick river terraces formed of coarse poorly-sorted gravels incised by the wadi channel. The transect south of Lamludah lies in the rainshadow on the southern side of the Gebel Akhdar and is rapidly more arid and less vegetated with distance southwards. Four sites were identified along the wadis here that drain to the south or west: (from north to south) CPP8112, CPP8113, CPP8114, and CPP8115. Most of the lithics at all of these sites are manufactured on a blue-grey chert, patinating to cream, a type of chert encountered by the survey at several sites discovered elsewhere on the Gebel Akhdar in 2009 and 2010. Only a few of the lithics were typologically diagnostic, including microblade cores, a Neolithic (?) bifacial point, and backed blades. It is likely that these lithics are terminal Pleistocene and Holocene in age. In the transect from the Haua Fteah into the hills to the southwest, an interesting cluster of sites was located about nine kilometres from the Haua Fteah, on the hilltop and slopes on the western side of the Wadi el-Mahbul. CPP8118, on the hilltop, consisted of a low density but extensive scatter of lithics of Middle Palaeolithic and later date formed on low-quality chert and high-quality limestone, both available locally. CPP8119, on the lower slopes, consisted of weathered patinated artefacts, mostly on limestone, the high frequency of radial flaking and faceting and a Levallois blade core indicating Middle Palaeolithic affinities. CPP8120 is likely to be a northern extension of the CPP8119 scatter. The location of the CPP8119 and CPP8120 sites suggests that Middle Palaeolithic artefacts are likely to be preserved buried in this locality, undisturbed by ploughing. A pattern has started to emerge regarding the nature of the surface lithic assemblages mapped in the 2009 and 2010 reconnaissance surveys, although this will need to be tested in the future by systematic surface collection at key sites. There is clear evidence for Middle Palaeolithic occupation along the littoral zone (notably at the Haua Fteah and Hajj Creiem, of course), and along the edges of palaeolakes such as al- Marj (discussed in the next section) and in the predesert. In contrast, there is a notable absence of Middle Palaeolithic occupation evidence across the Gebel Akhdar, particularly in areas at high elevations. There is no geomorphological reason why artefacts of this age would not be exposed given that the sediments in the upland zone are thin or virtually absent, so it would appear that Middle Palaeolithic use of this part of the landscape was temporally and spatially restricted, giving it low archaeological visibility. The infrequency of Middle Palaeolithic material in the uplands contrasts with the low density but more or less continuous presence here of later Pleistocene and Holocene artefacts characterised by blade and microblade technologies, with backed blades as a notable feature. On the basis of these preliminary observations, there appear to have been substantial differences in patterns of Middle Palaeolithic and later hominin occupation and palaeodemography in the Cyrenaican landscape during the Pleistocene and Holocene periods. Geoarchaeological investigations in the al-Marj basin By Lucy Farr, Nick Drake, Natalie Russell, Simon Armitage, Sacha Jones, Chris Hunt and Ian Candy Lake basins are critical repositories of palaeoenvironmental records, so the Shuttle Radar Topography Mission (SRTM) digital elevation model (DEM) and remotely-sensed satellite imagery (Landsat TM) have been used to locate potential lake basins by identifying large closed basins in the Gebel Akhdar where water would be ponded during past humid periods. One of the most promising of these is the al-Marj basin c. 125 km west of the Haua Fteah. Lineaments clearly evident on the DEM (Fig. 12) indicate a fault-bounded basin. The basin bedrock is largely composed of Eocene limestone, with slivers of Miocene and Cretaceous limestone exposed along a fault escarpment on the southern margin of the depression. A number of substantial ephemeral rivers (wadis) drain the highlands to the south, east, and west of the basin, with smaller rivers draining the low escarpment to the north. Towards the centre of the depression the channels become unconfined, forming low-angle alluvial fans that feed water into the centre of the depression where there are seasonallyflooded wetlands subject to periodic flooding and 80 The Cyrenaican Prehistory Project 2010 Figure 12.(top right) Shuttle Radar Topography Mission digital elevation model (DEM) of the al-Marj basin, showing the location of al-Marj city, the canal, the wetland zone, and sample sites. Faults show up in the DEM as linear features that mark abrupt changes in altitude. Figure 13. (bottom right) The Libyan Technical Consultancy Company drilling the al-Marj core using an Acker drilling rig with a Denison coring apparatus. Looking west (photo: G. Barker). o"cun -762 EM9 EM10 EM4 EM11 EM12 al-Marj Î392 Ygvncpf ephemeral lake formation during exceptional rainfall events. In the late 1970s an 8.5 km-long canal up to 3 m deep was excavated across the northern part of the basin to protect farmland from flooding by the river that drains the northeast corner of the basin, extending from the point where the river becomes unconfined at the headwaters of the alluvial fan to the wetland in the base of the depression. At the beginning of the 2010 season we arranged for a core to be drilled by the Libyan Technical Consultancy Company beside the canal near the centre of the basin (Fig. 12: site EM12), to establish the latter’s sedimentary history. LTTC used an Acker drilling rig with a Denison coring apparatus (Fig. 13), with casing employed down to a depth of 4.5 m to prevent caving. The deposits were highly compact from the surface and became extremely hard with increasing depth, causing technical problems below 5.0 m. As continuous coring was impossible below 10.25 m because of the hardness of the deposits, core samples were taken below that depth at c. 2 m intervals until the termination of drilling at 29.75 m. The sedimentary log obtained by visual inspection and hand sediment testing of the bases of the core segments indicates a sequence of distal alluvial fan and ephemeral lacustrine or playa environments as the lake variously expanded and contracted in response to climatic shifts. It is intended to establish an age model for the core using a range Ucorng"ukvg Ftckpcig"fkvej1ecpcn of chronological techniques and to investigate its record of changes in lake level and duration through changes in sediments, degrees of pedogenis, diatom content, and other biological climate proxies. The canal has exposed a large number of sedimentary units, many of which contain archaeological artefacts. Sediment logs were recorded and scaled sections drawn at four main locations (EM4, EM9, EM10, and EM11) using standard sediment recording techniques, Bulk sediment samples were taken from each main stratigraphic unit for geochemical characterisation to provide indications of the provenance of the source material. Samples for micromorphological analysis were taken from sediment units that appeared to contain palaeosol development indicative of periods of landscape stability, for indicators of prevailing climatic conditions. In situ 81 Graeme Barker et al. lithic artefacts were recorded on the section drawings before removal. The field studies demonstrate the evolution of a complex lacustrine and alluvial depositional environment that was subjected to considerable pedogenesis. Four sedimentary units can be recognised from north to south down the canal, each characterised by one of the four exposures recorded and sampled in detail. At EM9 fluvial and slope-wash deposits overlie colluvium and weathered bedrock. The latter consists of Eocene limestones inter-bedded with tabular chert. At EM4 (Figs 14 and 15) silty clays are overlain by a gravel unit in turn overlain by palaeosols. At EM13 the limestone bedrock is overlain by what are possibly lacustrine or marginal marine limestones, marls, and clays, containing fossil molluscs. A basal palaeosol at EM11 contains fine gravels overlain by a main gravel unit which is in turn overlain by three palaeosols of different colours (the top and bottom ones are red, the middle one is brown). Pairs of OSL dating samples were taken directly above and below the lithic-rich gravels at EM4, EM10 and EM11 with the aim of providing a maximum and minimum age of the sediments and further samples were taken from any fine sediment lenses within the gravels, as at EM4, to try to provide a higher-resolution dating model. Two OSL samples were also taken at EM4 and EM11 directly below the upper palaeosol unit to provide a maximum age for the deposition of this horizon. Figure 14. (right) Looking southwest across the al-Marj drainage ditch to the EM4 section (to the left of the two standing figures) (photo: G. Barker). Figure 15. (below) Section EM4 in the al-Marj drainage ditch, with simplified sedimentary descriptions (drawing: L. Farr). 15 Poorly-sorted, clast-supported gravel 14 Silty clay Poorly-sorted, clast-supported gravel Clayey silt OSL 3 17 Very poorly-sorted, clast-supported gravel 18 19 Silty clay Poorly-sorted, clast-supported, angular gravel 20 Well-sorted, clast-supported, medium-fine gravel 21 Poorly-sorted, clast-supported, very coarse gravel OSL 9 23 Well-sorted, clast-supported, fine gravel 24 Lithic artefact 25 26 OSL sample 27 Tabular chert 28 Gravel Coarse sand/fine gravel 29 Very poorly-sorted, clast-supported gravel OSL 2 Silt 30 Clay Clayey silt 0 22 1m 82 OSL 1 16 The Cyrenaican Prehistory Project 2010 Figure 16. Lithics from al-Marj: 1. handaxe (EM20); 2. handaxe (EM102); 3. Levallois core (EM11); 4. Levallois core (EM122); 5. Levallois core (EM131) (drawings: S. Jones). 1 Lithic artefacts were encountered both in the sections exposed by the construction of the canal and at 31 locations as surface finds, the latter on dry areas of the canal floor, on the floor of narrow side gullies, and on the spoil heaps formed during the construction of the canal. Preliminary observations of this material (Fig. 16) suggest that the lithic artefacts encountered in the canal cutting belong predominantly or exclusively to the Middle Palaeolithic or Middle Stone Age period, and possibly to the Lower Palaeolithic. The latter is suggested by the discovery of numerous bifacial hand-axes of varying size and shape in the basin’s gravel deposits, the former by the presence of Levallois cores and flakes comprising the general category of ‘prepared core technologies’. Notable for their apparent absence are cleavers, tanged points of the Aterian industry, and blade, microblade, and microlithic technologies characteristic of the late Pleistocene. Further study is necessary to confirm if this represents a true absence of lithic types or simply a sampling error. A second issue concerns the stratigraphic relationship between handaxes and the products of Levallois technology, as the former are characteristic of the Acheulean Lower Palaeolithic but can also be a notable feature of the Middle Palaeolithic. Hence it is currently uncertain whether there is an Acheulean followed by a Middle Palaeolithic presence in the al-Marj basin, or solely the latter. Virtually all the in situ lithic artefacts in the canal section occur in gravel deposits. Some of the artefacts are unpatinated and unabraded, others heavily patinated and rounded. It is possible that there was variation in the extent to which artefacts were transported and reworked into the gravels. Some artefacts may have been exposed on the land surface for longer periods of time than others prior to burial. It is also possible that artefacts of different periods are now preserved in the same contexts within the gravel units through intermixing. An aim of future investigations in the al- Marj basin is to locate areas where artefacts may be present in primary contexts. 2 4 0 5 cm 3 5 Other ecological and artefactual studies Ecological sampling for land snails outside the Haua Fteah cave By Evan Hill Land and marine molluscs are an important potential dating material in the Haua Fteah, but radiocarbon dating of molluscs is known to be problematical because they are not at radiocarbon equilibrium with their environment. Modern samples are therefore necessary in order to understand the fractionation of carbon isotopes by species commonly found in the region. An initial sampling programme was therefore carried out to locate live examples of the land snails known to be common in the cave fill (e.g. Helix melanostoma, Theba pisana, Eobania vermiculata, Rumina decollata) for calibration of radiocarbon dating and to refine ecological data. Six sites designated HF Eco01–06 were sampled at c. 20 m intervals directly downslope from the Haua Fteah. The ground underneath the most common shrubs at each location was excavated during dry weather and immediately succeeding a substantial rainstorm. Very few live specimens were uncovered on the rocky ground at Eco01. Excavation of Salvia plants here revealed no live or aestivated snails, but several Theba pisana were found in leaf-litter-filled rock crevices after the rainstorm. A number of aestivated specimens of Eobania vermiculata and Helix melanostoma were obtained from the soil and leaf litter surrounding Pistacia lentiscus at Eco03 and Rhus tripartita at Eco02, Eco03, and Eco05. No live or aesti- 83 Graeme Barker et al. vated species were found on or around the shallow and highly-compacted root soils around Juniperus phonenicea, nor under Sarcopoterium spinosum sampled at Eco05 and Eco06. Fourteen aestivated specimens of Helix melanostoma and Eobania vermiculata were found in dry-weather conditions in the soil and leaf litter alongside the roots of Ceratonia siliqua bushes at Eco04, Eco05 and Eco06. A further live example of Helix melanostoma was collected at Eco04 in wet weather. Overall, Helix melanostoma and Eobania were found to be particularly densely clustered in the leaf litter and soil under Ceratonia siliqua (carob) and to a lesser extent under Rhus tripartite and Pistacia lentiscus: aestivated clusters of these species were found in dry weather at Eco05 and Eco06. During wet-weather sampling, by contrast, numbers of aestivated specimens were rare and shells of recently predated specimens common, suggesting that predators quickly consume these species as they emerge into the landscape after rainfall. Apart from their implications for future sampling for the radiocarbon study, the results suggest that prehistoric foragers at the Haua Fteah could have collected land snails either on the surface immediately after rain (most likely in the autumn to spring rainy season) or by digging up aestivated specimens during the summer. The fact that aestivated molluscs occur in clusters under only a few types of bush suggests that this latter strategy might have been highly effective. sequence at the site offers the potential of establishing a detailed framework for the character, duration, and scale of climate change against which changing patterns of human behaviour can be compared. Of particular interest are bones and teeth of Barbary sheep (Ammotragus lervia), as this species appears to be abundant throughout much of the archaeological sequence, and shells of the marine molluscs Patella caerulea and Osilinus turbinatus and the terrestrial mollusc Helix melanostoma. In order to obtain robust palaeoenvironmental proxy data from the geochemical analyses of faunal remains, it is necessary to determine whether the proxy species precipitate carbonate in isotopic equilibrium with their surrounding environments. Studies of modern species allow the primary controls on geochemical variation in the faunal carbonate to be determined. During the 2010 field season, therefore, a programme of modern sampling was initiated, of teeth from carcasses of modern sheep and goats grazing in the vicinity of Haua Fteah and elsewhere in the study area (modern Ammotragus are being sourced from outside Libya as the species is now extinct there), and of live specimens of Patella caerulea and Osilinus turbinatus. Samples of the latter will be collected and analysed each month and compared with monthly collected sea-surface temperature and salinity measurements. Results from these isotopic studies will be compared with modern climatic data in order to establish a modern analogy, to provide the basis for the quantitative interpretation of data obtained from the analysis of the archaeological material. Isotopic studies of palaeoclimate By Amy Prendergast and Hazel Reade Residue and use-wear analysis Human presence in North Africa does not appear to have been constant throughout the late Pleistocene, with periods of abandonment and reoccupation appearing to correlate with changes in climate (Cremaschi et al. 1998), but the relationship between climate and human presence remains to be clearly demonstrated. The palaeoclimate of the eastern Mediterranean region has received comparatively little attention (Bar-Matthews et al. 1999) and North Africa has been identified as a significant gap in the coverage of both palaeoclimate and modern isotopic climate data (Roberts et al. 2010). A programme of palaeoclimatic research using isotopic analysis of faunal material (bones, teeth, and shell) from the Haua Fteah was therefore initiated in 2009. The oxygen-isotope signature contained within teeth and bones can be used as proxies for terrestrial palaeoclimate, and isotopic values within mollusc shells as proxies for sea-surface temperature, rainfall, and palaeovegetation. The presence of these faunal remains within the archaeological By Kate Connell Residue and use-wear analysis aims to determine the function of stone tools. Over the course of a threeyear PhD the method will be applied to a selection of lithics from the Haua Fteah. A pilot study was carried out in the months leading up to the 2010 field season on lithics collected in previous seasons to determine the likelihood of finding preserved residues on newly excavated material. In the test group many artefacts exhibited characteristic wear associated with use, together with some diagnostic residue traces. Given these positive results, the main focus of the project has been identified as the behavioural implications of technological change, such as the explosion in the frequency of backed pieces in the Oranian phase, a notable technological feature of the Haua Fteah sequence that has yet to be explained. As elsewhere in North Africa, the Oranian was marked by a near 84 The Cyrenaican Prehistory Project 2010 total dominance of backed pieces in the retouched category at the expense of scrapers and burins, both of which are present in higher numbers in the cultural phases before and afterwards. It is not known whether this technological shift represents very different behaviour or preferential selection of backed pieces for tasks previously executed with other tool forms. All of the hand-collected backed pieces from contexts 10,006 to 10,009 in Trench M were examined under a low-power microscope before washing, and are being compared with artefacts recovered from the flotation residues, to assess the effect of the flotation process on residue survival. Trench, reaches these depths, but it is already apparent that the hominin use of the cave at this time was highly variable and that it developed within dramatically changing environmental conditions. The sediments in the basal c. 70 cm (equivalent to McBurney’s Layers XXXV–XXXIII) appear to have been laid down through a series of large mudflows. An increase in gravel content and wind-blown sediments in the next 50 cm (McBurney’s Layers XXXII–XXX) implies increasing aridity and more degraded landscapes. Above a major anthropogenic burning event regarded by McBurney as within Layer XXX, the sedimentary record of the upper Middle Palaeolithic layers (XXIX–XXV) represents a complex sequence of deposition and re-working indicative of rapid environmental fluctuations, associated with few definite signs of human presence. The newly excavated sequence explored by Trench M in 2009 and 2010 extends from the middle of the (Mesolithic) Capsian through the (Late Palaeolithic or Epipalaeolithic) Oranian to the top of the (Upper Palaeolithic) Dabban. The Oranian layers defined by McBurney (Layers XIV–XI) were dated by him to c. 14,000–10,000 years ago on the basis of four 14 C determinations. In the 2009 report we published nine new 14C dates on charcoal from contexts equivalent to Layers XIV–XI, which span from 13,360±55 bp or 16,840–15,904 cal. BP (OxA-19189) to 11,160±45 bp or 13,220–12,850 cal. BP (OxA-19030) (Barker et al. 2009: 90). To these we can add four further dates from the Oxford laboratory. Two are on charcoal, of 11,855±50 bp or 13,845–13,485 cal. BP (OxA-22138) from context 183 (around McBurney’s Layer XII) and 10,900±45 bp or 12,920–12,630 cal. BP (OxA-22137) from context 180 (McBurney’s Layer XI). The other two were obtained by Katerina Douka on perforated shells (Columbella rustica) from the McBurney Excavation Archive in the Museum of Archaeology and Anthropology at the University of Cambridge, one from Layers XIV–XV at the Oranian/Dabban transition, dated twice to 19,015±70 bp (OxA-21087) and 19,040±65 bp (OxA-21088) or 22,480–22,010 cal. BP, and one of 10,957±39 bp or 12,610–12,360 cal. BP (OxA-21086) of uncertain provenance but likely to be around the Oranian/Capsian transition from the label with it (‘red soil at 8'–8'.6" depth’). The calibrated shell dates have the global mean reservoir correction R (c. 400 years), as well as the local variation R for the Mediterranean (58±85 14C years), subtracted from them during calibration with Marine09 (Reimer and McCormac 2002; Reimer et al. 2009). The new dates suggest that the Oranian phase represents the human use of the cave between the Last Glacial Maximum c. 20,000 cal. BP and the Pleistocene/Holocene transition c. 11,500 cal. BP. Discussion By Graeme Barker, Chris Hunt and Tim Reynolds The 2010 excavations in the Haua Fteah and the ongoing programme of study of materials from the earlier seasons have considerably enhanced our understanding of the sedimentary sequence and its palaeoenvironmental correlations, and the changing character of the human use of the cave. The programme of micromorphology and sediment analysis by Robyn Inglis is emphasising the importance of local reworking and recycling of sediment being washed or blown into the cave. The lowest layers investigated so far, in the walls of the upper 2 m of the McBurney Deep Sounding and in the adjacent Trench D, appear to be predominantly formed by ephemeral water-flows as well as in situ degradation of limestone rock. The indications from a programme of OSL dating (to be published separately) are that these sediments are in the order of 100,000 years old. Although McBurney referred to the upper part of the Deep Sounding (within his Pre-Aurignacian phase) as more or less sterile, there is clear evidence for small-scale but regular hominin presence in the form of stone artefacts and debitage, and patches of burnt rocks and charcoal. Another indicator of hominin activity is the occasional presence of the marine molluscs Patella and Osilinus, including burnt specimens, and numbers of the land snail Helix melanostoma, likely to have been collected for food. Given the sedimentary context it seems likely that hominins visiting the cave at this time sheltered further back under the overhang, their occupation debris collecting in the excavation zone through water-flow. The lower half (c. 2 m) of the Middle Trench represents McBurney’s Layers XXXV–XXV, assigned by him to the Middle Palaeolithic. The nature of Middle Palaeolithic occupation will only be explored in detail when Trench M, being excavated alongside the Middle 85 Graeme Barker et al. The preliminary study of the prolific lithic material from Trench M (some 17,000 pieces from one of eight quadrats) by Giuseppina Mutri and Giulio Lucarini indicates that the material from the basal context excavated so far (10,011) is a mixture of Dabban and Oranian elements, with a marked frequency of endscrapers and more blades and bigger bladelets compared with the higher layers. There is some continuity between the Oranian and Capsian assemblages such as the presence in low frequencies in the Oranian layers of the scalene blade or bladelet common in the Capsian layers, the opposite trend in the frequency of the reversed backed blade/bladelet and flake, and the presence in both assemblages of blades/bladelets with Ouchtata retouch and Krukowski microburins. The major trend differentiating the two is the appearance in the Capsian layers of denticulated and notched tools. The extent to which these technological changes may be indicative of behavioural changes are the subject of PhD research of microwear and residues on the stone tools by Kate Connell, but there are other indicators of subsistence change in the faunal and botanical material from Trench M, though these have not yet been studied in any detail. The Oranian layers consist mostly of silts with limestone fragments but 10,008 in the middle is notably charcoal-rich and from this layer upwards shells of marine and terrestrial molluscs increase markedly in frequency into the Capsian layers (10,005–10,001), where they are uniformly prolific. As in the Pre-Aurignacian layers in Trench D (though in far greater densities), the main species collected for food appear to have been the marine molluscs Patella and Osilinus, and the land snail Helix melanostoma, the latter (on the basis of modern sampling around the cave) probably collected on the surface immediately after rain or by digging up aestivated specimens. The ecologies of the terrestrial mollusc fauna are consistent with rather dry, rather open, but still vegetated landscapes around the cave in both the Oranian and Capsian phases, though given the small sample size this interpretation will need to be confirmed. The preliminary study of the fauna by Ryan Rabett suggests that a similar range of fauna was being targeted selectively by both Oranian and Capsian hunters, notably Barbary sheep (Ammotragus lervia), gazelle (Gazella sp.), and hartebeest (Alcephalus sp.), with tortoise (Testudo graeca cyrenaica) perhaps another staple in forager diet, but there are indicators of a greater availability of large ungulates, and of a greater intensity of Barbary sheep predation, in the late Pleistocene Oranian compared with in the early Holocene Capsian. A noteworthy find is the probable trapping stone in the Oranian context 10,009. The presence of juvenile and young Barbary sheep could imply the greater use of the cave for hunting this species in the summer/autumn months, as Wall (2004) suggested. The plant remains from the Oranian levels have not been studied yet, but the preliminary study by Jacob Morales of the 2009 material from Trench M indicates that Capsian foragers collected pine cones for their seeds (charring the cones to make the extraction of the seeds easier), seeds of wild pulses and large-seeded legumes, and the fruits of juniper, the mastic tree, and myrtle. One seed of Pinus halapensis and one seed of wild vetch from the Capsian/Oranian transition have been directed dated by 14C (Table 3). What remains unclear is the level of intensity of the use of the cave in the later Pleistocene and early Holocene. The little evidence for bone trampling, for example, suggests that the use of the part of the cave represented by Trench M was not very intensive, especially in the early Holocene, whilst root surface etching on bones from the Capsian layers indicates a degree of vegetation growth unlikely if there was regular foot traffic. Despite the extraordinary amount of lithic material collected from the very small area of the cave represented by Trench M, it seems likely that the Oranian and especially the Capsian use of the cave was highly episodic, rather as is appearing to be the case in the Pre-Aurignacian and Middle Palaeolithic phases on the evidence of micromorphology. Similar variability in the use of the Cyrenaican landscape is also emerging from the geoarchaeological work outside the cave. The reconnaissance surveys of lithic material across the Gebel Akhdar are providing evidence of shifting patterns of hominin occupation, with indicators of the preferential use of the coastal zone and interior lake basins by Middle Palaeolithic hominins (and possibly in the Lower Palaeolithic on the evidence of the al-Marj basin), and low-density but continuous hominin presence in the Gebel Akhdar uplands in the later Pleistocene and early Holocene. Of course these statements about the use of the Haua Fteah cave and of the wider Cyrenaican landscape can only be tentative generalisations given the present state of knowledge. A major goal for the Cyrenaican Prehistory Project is to develop a detailed regional climatic and environmental history for the later Pleistocene and Holocene, and a detailed understanding of the character and chronology of hominin activity in the landscape over the same timescale, in order to be able to assess the extent to which there are correlations between the two histories and if so, their significance. 86 The Cyrenaican Prehistory Project 2010 Acknowledgements We would like to acknowledge in particular the permission of the Department of Antiquities of Libya to undertake the project; the financial support of the European Research Council and the Society for Libyan Studies for the fieldwork, Queen’s University Belfast for the 14C dates published here, and of the Natural Environment Research Council (RESET project) in funding Christine Lane’s participation; the invaluable logistical support provided by staff of the Department of Antiquities in Tripoli, Shahat, and Susah; and the further considerable help of Ruth Codd of AMEC. We would also like to thank Mohammed Shein and Adly Badi of the Libyan Technical Consultancy Company for their highly professional drilling operation at al-Marj, and the advice of Fathi Salloum of the Department of Geology, Gar Yunis University. The personal support of the President of the Department of Antiquities Dr Saleh Agab and the Controller of Antiquities for Cyrene Ahmed Saber has been vital to the success of the project. Further invaluable support for the project was provided by Paul Bennett (Society for Libyan Studies), Ahmed Rhaba and Abdulgader al-Muzeini (Department of Antiquities, Cyrene), Mustapha Turjman (Department of Antiquities, Tripoli), Ibreike Quinhe (Inspector of Antiquities, Apollonia), Mohammed Atatullah (Department of Antiquities, al-Marj), and Ahmed Buzaian and Abdullah Rhebi (Department of Archaeology, Gar Yunis University, Benghazi). The help of the Department of Archaeology’s representative Saad el-Lefel is also gratefully acknowledged. Special thanks are due to the team’s cook Amsaad el-Awami and his assistant Adi el-Awami, drivers Abullah Marasi, Abdulhamid Ali and Hamza al-Muzeini, custodian Abdul Hak Ahamid, residue sorters Hamdi al-Muzeini, Mohammed al-Muzeini, and Ziad, and landowner Salem al-Hassy. The 2010 team consisted of: Graeme Barker, Annita Antoniadou, Simon Armitage, Ian Brooks, Ian Candy, Kate Connell, Nicholas Drake, Lucy Farr, Evan Hill, Chris Hunt, Robyn Inglis, Sacha Jones, Christine Lane, Giulio Lucarini, John Meneely, Jacob Morales, Giuseppina Mutri, Holly Parton, Amy Prendergast, Ryan Rabett, Hazel Reade, Tim Reynolds, Natalie Russell, David Simpson, Bernard Smith, Chris Stimpson, Mohammed Twati, and Kevin White. References R., Reynolds, T., Simpson, D., Twati, M. and White, K., 2009. The Cyrenaican Prehistory Project 2009: the third season of investigations of the Haua Fteah cave and its landscape, and further results from the 2007–2008 fieldwork. Libyan Studies 40: 55–94. Bar-Matthews, M., Ayalon, A., Kaufman, A. and Wasserburg, G.J. 1999. The eastern Mediterranean paleoclimate as a reflection of regional events: Soreq cave, Israel. Earth and Planetary Science Letters 166: 85–95. Boulos, L. 1983. Medicinal Plants of North Africa. Algonac (MI): Reference Publications. Cartwright, C. and Hunt, C. 2008. Preliminary observations on the wood charcoal. In Barker et al. 2008: 205–8. Charco, J. 2001. Guía de los Árboles y Arbustos del Norte de África. Madrid: AECI. Cremaschi, M., Di Lernia, S. and Garcea, E. 1998. Some insights on the Aterian in the Libyan Sahara: chronology, environment, and archaeology. Geochimica et Cosmochimica Acta 15: 261–86. Dobson, M. and Wright, A. 2000. Faunal relationships and zoogeographical affinities of mammals in North-west Africa. Journal of Biogeography 27.2: 417–24. Goldberg, P. and Bar-Yosef, O. 1998. Site formation processes in Kebara and Hayonim caves and their significance in Levantine prehistoric caves. In Akazawa, T., Aoki, K. and Bar-Yosef, O. (eds), Neanderthals and Modern Humans in Western Asia. Plenum Press, New York: 107–26. Aura, J.E., Carrió, Y., Estrelles, E. and Pérez-Jorda, G. 2005. Plant economy of hunter-gatherer groups at the end of the last Ice Age: plant macroremains from the cave of Santa Maira (Alacant, Spain) c. 12,000–9000 BP. Vegetation History and Archaeobotany 14: 542–50. Bailey, G.N. and Woodward, J.C. 1997. The Klithi deposits: sedimentology, stratigraphy and chronology. In Bailey, G.N. (ed.), Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece. McDonald Institute for Archaeological Research, Cambridge: 61–94. Barker, G., Hunt, C. and Reynolds, T. 2007. The Haua Fteah, Cyrenaica (northeast Libya): renewed investigations of the cave and its landscape, 2007. Libyan Studies 38: 93–114. Barker, G., Basell, L., Brooks, I., Burn, L., Cartwright, C., Cole, F., Davison, J., Farr, L., Grün, R., Hamilton, R., Hunt, C., Inglis, R., Jacobs, Z., Leitch, V., Morales, J., Morley, I., Morley, M., Pawley, S., Pryor, A., Reynolds, T., el-Rishi, H., Roberts, R., Simpson, D., Twati, M. and van der Veen, M., 2008. The Cyrenaican Prehistory Project 2008: the second season of investigations of the Haua Fteah cave and its landscape, and further results from the initial (2007) fieldwork. Libyan Studies 39: 175–221. Barker, G., Antoniadou, A., Barton, H., Brooks, I., Candy, I., Drake, N., Farr, L., Hunt, C., Ibrahim, A.A., Inglis, R., Jones, S., Morales, J., Morley, I., Mutri, G., Rabett, 87 Graeme Barker et al. in the Middle and Early Upper Palaeolithic Industries at the Haua Fteah, Libya. Unpublished PhD thesis, University of Cambridge. Olsen, S.L. and Shipman, P. 1988. Surface modification on bone: trampling versus butchery. Journal of Archaeological Science 15: 535–53. Reimer, P.J. and McCormac, F.G. 2002. Marine radiocarbon reservoir corrections for the Mediterranean and Aegean Seas. Radiocarbon 44: 159–66. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Bronk Ramsey, C., Buck, C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., McCormac, F.G., Manning, S.W., Reimer, R.W., Richards, D.A., Southon, J.R., Talamo, S., Turney, C.S.M., van der Plicht, J. and Weyhenmeyer, C.E. 2009. IntCal09 and Marine09 radiocarbon age calibration curves, 0–50,000 years cal BP. Radiocarbon 51: 1111–50. Rivera, D. and Obón, C. 1991. La Guía de INCAFO de las Plantas Útiles y Venenosas de la Península Ibérica y Baleares (Excluidas Medicinales). Madrid: INCAFO. Roberts, C., Zanchetta, G. and Jones, M. 2010. Oxygen isotopes as tracers of Mediterranean climate variability: an introduction. Global and Planetary Change 71: 135–40. Smith, B.J. 1978. The origins and geomorphic implications of cliff foot recesses and tafoni on limestone hamadas in the northwest Sahara. Zeitschrift für Geomorphologie 22: 21–43. Smith, B.J. 2009. Weathering processes and forms. In Abrahams, A. and Parsons, A. (eds), Geomorphology of Desert Environments: 69–100. (2nd edition). Springer, Berlin. Stein, J.K. 2001. A review of site formation processes and their relevance to geoarchaeology in earth sciences and archaeology. In Goldberg, P., Holliday, V.T. and Ferring, C.R. (eds), Earth Sciences and Archaeology. Kluwer Academic/Plenum Publishers, New York: 37–54. Tixier, J. 1963. Typologie de l’Epipaléolithique du Maghreb. Paris: Mémoires du C.R.A.P.E. (II). Vaquer, J. and Ruas, M.-P. 2009. La grotte de L’Adeurador Félines-Minervois (Hérault): occupations humaines et environnement du Tardiglaciaire à l’Holocène. In De Méditerranée et d’Ailleurs. Mélanges Offerts à Jean Guilaine: 761–92. Toulouse: Archives d’Écologie Préhistorique. Wall, C. 2004. Coastal Ungulates: The Seasonal Use of Ibex and Barbary Sheep Among Mediterranean Hominins. Unpublished PhD thesis, University of Cambridge. Zong, G. 1984. A record of Bos primigenius from the Quaternary of the Aba Tibetan Autonomous Region. Vertebrata PalAsiatica 22.3: 239–45. Goldberg, P. and Sherwood, S.C. 2006. Deciphering human prehistory through the geoarcheological study of cave sediments. Evolutionary Anthropology 15: 20–36. Hall, S.J.G. 2008. A comparative analysis of the habitat of the extinct aurochs and other prehistoric mammals in Britain. Ecography 31: 187–90. Hey, R.W. 1967. Appendix 4A. Land-snails. In McBurney, 1967: 358. Higgs, E.S. 1967. Environment and chronology — the evidence from mammalian fauna. In McBurney 1967: 16–44. Hublin, J.-J. 2000. Modern–Non-modern hominid interactions: a Mediterranean perspective. In Bar-Yosef, O. and Pilbeam, D. (eds), The Geography of Neanderthals and Modern Humans in Europe and the Greater Mediterranean. Peabody Museum of Archaeology and Ethnology, Peabody Museum Bulletin 8, Cambridge (MA): 157–82. Inglis, R. 2009. Micromorphology. In Barker et al. 2009: 64–5. Klein, R. and Scott, K. 1986. Re-analysis of faunal assemblages from the Haua Fteah and other Late Quaternary archaeological sites in Cyrenaican Libya. Journal of Archaeological Science 13: 515–42. Le Quellec, J.-L. 1990. Pierres de Ben Barur et ‘Radnetzen’ au Fezzan (Libye). L’Anthropologie 94: 115–26. McBurney, C.B.M. 1960. The Stone Age of Northern Africa. Penguin Books, London. McBurney, C.B.M. 1967. The Haua Fteah (Cyrenaica) and the Stone Age of the South-East Mediterranean. Cambridge University Press, Cambridge. McBurney, C.B.M. and Hey, R.W. 1955. Prehistory and Pleistocene Geology in Cyrenaican Libya. Cambridge University Press, Cambridge. Meneely, J.D., Smith, B.J., Viles, H.A. and Gomez-Heras, M. 2008. In-situ monitoring of limestone buildings in Oxford. In Tiano, P. and Pardini, C. (eds), In-situ Monitoring of Monumental Surfaces. Centro Nazionale di Ricerche, Florence: 327–34. Meneely, J.D., Smith, B.J., Curran, J. and Ruffell, A. 2009. Developing a non-destructive scientific toolkit to monitor monuments and sites. Proceedings of ICOMOS Scientific Symposium (Changing World, Changing Views of Heritage: the Impact of Global Change on Cultural Heritage – Technological Change). Available at: http://www.international. icomos.org/adcom/malta2009/symposium.htm. Morales, J. and Barker, G. 2009. The macrobotanical remains from the 2008 season. In Barker et al. 2009: 83–6. Morales, J. and van der Veen, M. 2008. Preliminary observations on the macrobotanical remains. In Barker et al. 2007: 208–13. Moyer, C. 2003. The Organisation of Lithic Technology 88