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.

The paper reports on the fifth (2012) season of fieldwork of the Cyrenaican Prehistory Project. The primary focus of the season was the continuation of the excavation of the prehistoric occupation layers in the Haua Fteah cave. A small trench (Trench U) was cut into Holocene (Neolithic) sediments exposed on the south wall of Charles McBurney's Upper Trench. Below this, the excavation of Trench M was continued, on the southern side of McBurney's Middle Trench. In previous seasons we had excavated Oranian ‘Epipalaeolithic’ layers dating to c. 18,000–10,000 BP (years before the present). In 2012 the excavation continued downwards through Dabban ‘Upper Palaeolithic’ occupation layers, one of which was associated with a post-built structure and likely hearths. There are indications of an occupational hiatus separating the oldest Dabban from the youngest Levallois-Mousterian (Middle Palaeolithic or Middle Stone Age) lithic material. The Deep Sounding excavated by Charles McBurney in 1955 was cleared of backfill to its base, and its south-facing wall was recorded in detail and sampled extensively for materials for dating and palaeoenvironmental reconstruction. McBurney believed that he had reached bedrock at the base of the Deep Sounding, but a small sounding (Trench S) cut into the sediments below this level found further, albeit sparse, evidence for human occupation. Whilst the antiquity of ‘Pre-Aurignacian’ human occupation at the site still needs to be resolved, it seems likely to reach back at least to Marine Isotope Stage 5e, the beginning of the last interglacial (c. 130,000–115,000 BP). Important finds from the 2012 excavations in terms of the behavioural complexity of the human groups using the cave include a possible worked bone point from a Pre-Aurignacian layer and a granite rubbing stone in a Dabban layer from a source over 600 km from the cave.

The 1950s excavations by Charles McBurney in the great Haua Fteah cave in northeast Libya revealed a deep (14 m) sequence of human occupation going back at least 100,000 years, with evidence for the presence of both Neanderthals and Modern Humans in the Pleistocene, and for Neolithic farmers in the Holocene. In 2007 a renewed programme of archaeological and geomorphological investigation began with the objective of improving understanding of the cave's occupation sequence and, combined with fieldwork in the landscape, of the history of landscape change and human responses to it. The initial season of fieldwork removed the upper c. 4.5 m of backfill in the McBurney trench; established the robustness of the original faces and their suitability for analytical interventions; recorded detailed running sections spanning from the present day to (at least) the Last Glacial Maximum c. 20,000 years ago; and indicated the potential of the surviving archaeology to reveal not just sequence but also activities or ‘taskscapes’ at the site. The geomorphological fieldwork identified rich sequences of later Quaternary deposits (marine, colluvial, alluvial, aeolian) with the potential to provide significant results regarding the history of climate and environment in the region. Archaeological survey around the cave indicates that the variability of the surface lithic evidence appears to reflect real differences in past human behaviour and use of the landscape and not just post-depositional taphonomic processes. Fifty years after the extraordinary pioneering work of McBurney and his colleagues, the new work demonstrates the continued potential of the Haua Fteah's unique occupation sequence and the multi-period ‘human landscapes’ around it to transform understanding of early human societies in North Africa.

The second (2008) season of fieldwork of the Cyrenaican Prehistory Project has significantly advanced understanding of the Haua Fteah stratigraphy and of the archaeology and geomorphology of the landscape in which the cave is located. The excavations of the McBurney backfill have reached a total depth of 7.5 m below the present ground surface, the depth at which two human mandibles were found in the 1950s excavations. Reconnaissance at the Hagfet ed-Dabba established that the sediments associated with the Upper Palaeolithic ‘Dabban’ industry were more or less entirely removed by the McBurney excavation. Exploratory excavations in the Hagfet al-Gama, a coastal cave west of the Haua Fteah, found evidence of Mesolithic, Neolithic, and Hellenistic occupation. The initial results from the study of botanical remains, both macroscopic and microscopic, obtained in the 2007 season at the Haua Fteah confirm the potential of the site to yield a rich suite of materials to inform on climatic and environmental change, and on human activities in the cave.

The paper reports on the third (2009) season of fieldwork of the Cyrenaican Prehistory Project, and on further results from the analysis of materials collected in the previous (2007 and 2008) fieldwork. Sediments in a 14 m-deep core drilled beside the McBurney trench provide an invaluable overview of the overall stratigraphic sequence, including at depths reached by the 1950s Deep Sounding but not yet investigated by the present project. Sampling of newly-exposed faces of the original excavation trench for dating (14C, ESR, OSL, U-series) and palaeoenvirommental indicators continued. Excavation was begun of sediments assigned to the early Holocene Libyco-Capsian (McBurney's Layer X), and of Pre-Aurignacian layers beside the top of the Deep Sounding. The Libyco-Capsian layers are particularly prolific in lithic debris, shells, and animal bones; preliminary analysis of the lithics suggests a development from Typical to Upper Capsian within the layers excavated in 2009. Geoarchaeological survey along the littoral to the west and east of the Haua Fteah identified complex sequences spanning most of the last interglacial-glacial cycle. Geoarchaeological survey south of the Haua Fteah characterized the major landforms of the Gebel Akhdar mountain and of the pre-desert and desert-edge zones further south, with Late Stone Age (Upper Palaeolithic and Epipalaeolithic) material being found especially on the southern side of the Gebel Akhdar, and Middle Stone Age (Middle Palaeolithic) material in the pre-desert and desert regions. The first suite of 14C dates (from charcoal samples taken in 2007) indicates the use of the Haua Fteah by Oranian hunter-gatherers during the Last Glacial Maximum and in the succeeding millennia, but not in the Younger Dryas cold/dry phase (c. 11,000–10,000 cal. BC), with Libyco-Capsian occupation resuming soon after the beginning of the Holocene c. 9000 cal. BC, suggesting that the cave, and perhaps the Gebel Akhdar in general, have a complex history as refugia for human settlement during the Pleistocene.

This paper is a sequel to an earlier study of the archaeological evidence for olive oil production in Roman Tripolitania, and combines archaeological information on olive presses with the evidence for trade in Tripolitanian oil, as manifested by the identification of Tripolitanian containers (amphorae). Attention is drawn, in particular, to the apparent correspondence between conclusions made in respect of the settlement data and of the socio-economic significance of stamps on Tripolitanian amphorae. Both sets of data support the view that the urban elite of the region, and in particular that of Lepcis Magna, had control of a very considerable quantity of surplus oil production and that this constituted a major source of individual prosperity, in part manifested by the extravagant donations of public buildings. The notable wealth of the Lepcitanian aristocracy enabled increasing numbers from amongst it to participate at a high level in the central government of the Roman Empire as senators or equestrians. The culmination of this process was the elevation of Septimius Severus as the first African Emperor in AD 193. His reign brought clear short term advantages to his home town, but after his death and the assassination of both his sons by AD 217, the fortunes of the town changed for the worst. Under Severus the town had offered the Roman State a voluntary ‘tax’ in olive oil and this together with the absorption of much land into Imperial estates must have taken a significant amount of agricultural surplus out of local control. In an Appendix, some comments are made on the problems of attempting to quantify the scale of oil production, and some broad estimates are suggested.

The Garamantes exercised a powerful hold upon the Mediterranean imagination during the early and high Empire (Daniels 1970; Desanges 1962; Merighi 1940; Romanelli 1959; Ruprechtsburger, forthcoming). Remote and beyond the southern limits of the African provinces, they represented the mythical south, only accessible after perilous desert travel; they were inextricably interwoven in people's minds with swarming serpents, fabulous fountains and precious stones, with the silent noon tide of the terrifying desert and such inhuman creatures as men with no voices or no heads, strapfoots, Goat-Pans and Satyrs (Pliny, N.H. 5.1.7).

The reality of Roman contact with them was no less exotic. The last entry of all in the Fasti Triumphales (CIL I2, p. 50) on the Arch of Augustus in the Roman Forum commemorates a triumph celebrated on 27th March 734 A.V.C. (19 BC) by L. Cornelius Balbus, ex Africa: below, the stone surface remains rough and no further names follow. But Balbus' honour was much more noteworthy for he was not even a Roman by birth (PIR2 C1331; Thomasson 1960, 11). Like his famous uncle, Caesar's Balbus, he was a full-blooded Spaniard hailing from Cadiz (Gades) — and the only foreigner ever to be accorded a triumph, the Elder Pliny informs us (N.H. 5.5.36). His rise to fame came about through a steadfast devotion to Caesar's cause, for which he was ultimately rewarded by Augustus with consular rank (Velleius 2.51.3), and became Proconsul of Africa in ?20 BC (Syme 1939, 80, 235, 325, 339, 367).

The Desert Migrations Project is a new interdisciplinary and multi-dimensional collaborative project between the Society for Libyan Studies and the Department of Antiquities. The geographical focus of the study is the Fazzan region of southwest Libya and in thematic terms we aim to address the theme of migration in the broadest sense, encompassing the movement of people, ideas/knowledge and material culture into and out of Fazzan, along with evidence of shifting climatic and ecological boundaries over time. The report describes the principal sub-strands of the project's first season in January 2007, with some account of research questions, methods employed and some preliminary results. Three main sub-projects are reported on. The first concerns the improved understanding of long-term climatic and environmental changes derived from a detailed palaeoenvironmental study of palaeolake sediments. This geo-science work runs alongside and feeds directly into both archaeological sub-projects, the first relating to prehistoric activity and mobility around and between a series of palaeolakes during wetter climatic cycles; the second to the excavation of burials in the Wadi al-Ajal, exploring the changing relationship between material culture, identity and ethnicity across time, from prehistory to the early Islamic period (the span of the main cemetery zones). In addition, some rock art research and a survey of historic period sites was undertaken in the Wadi ash-Shati and Ubari sand sea.

Two seasons of work have now been conducted by British and French survey teams, in conjunction with members of the Libyan Antiquities Department, under the charge of Dr. Abdullah Shaiboub. The objectives of the survey are to locate, survey and analyse the extensive remains of the ancient agricultural settlements that can be found in the wadis of the hinterlands of Tripolitania and the Sirtica. Within the framework established by the Department in cooperation with Unesco lies the archaeological aim of recording the evidence for periods when extensive areas of the pre-desert were, for whatever reasons, cultivated in ways that are not similarly practised today. In the longer term the programme is designed to locate those areas where modern farming might be re-established. Archaeology is thus brought into line with the aims of the modern world.

For the purposes of this report we intend to concentrate on the period which we call the Romano/Libyan in which the great majority of those farming settlements flourished. The prehistoric evidence is in any case mainly of the palaeolithic period, on which there is a separate section.

The preferred zone of settlement in Tripolitania has traditionally been the well watered coastal plain and the adjacent limestone hills of the Tarhuna Gebel as far south as the town of Beni Ulid, for these regions have more than 200 mm of rain a year, regarded as the threshold for settled farming without irrigation. Prehistoric settlement concentrated here, and mixed farming has probably characterised this zone from the fourth millennium b.c. In the Roman period the coastal cities like Sabratha and Leptis Magna were supported by prosperous farms on the plain and in the Gebel. In the Islamic period, too, the same region was densely settled.

Excavations at Zinchecra, a hill fort settlement of the Garamantes in Fezzan, southern Libya, have recovered a rich assemblage of desiccated and carbonised plant remains. The archaeobotanical analysis of this assemblage has produced a unique insight into the state of agriculture in the Sahara during the first half of the first millennium BC. Three cereal crops and three fruit crops have been identified, as well as salad plants and aromatic herbs. The use of wild plant resources has also been attested. The assemblage is dated by eleven radiocarbon dates to 900–400 cal BC. A well-developed agricultural regime was present, despite the harsh climatic conditions.

Three seasons of intensive survey in the central Tripolitanian pre-desert, primarily in the regions of the Wadis Sofeggin and Zem Zem, have provided a new basis for the study of pre-desert cultures. Not only have hundreds of new sites been discovered, but the combination of environmental and geomorphological studies with archaeological survey is now providing important information about the underlying mechanisms which supported them. An increasing settlement density and social complexity can be identified in the Romano–Libyan period together with the development of intensive farming and elaborate irrigation systems. There is little evidence for significant climatic change to account for this increased settlement intensity or for its subsequent decline. The explanation for this must probably be sought in a complex relationship between social organisation, population change, and the agricultural system, and its impact on the environment.

Reconnaissance survey in the Murzuq area, some 150 km south-east of Jarma, was carried out as part of the 2011 field programme of the Desert Migrations Project, with separate funding from the Leverhulme Trust for this element of work entitled the ‘Peopling the Desert Project’. This survey was designed to provide field verification of details of settlement systems identified and mapped from high-resolution satellite images in an area of c. 600 km2 immediately east of the oasis town of Murzuq. Examination of high-resolution QuickBird and Ikonos satellite imagery has permitted identification of a large dossier of more than 200 sites (fortified buildings known as qsur, other settlements, cemeteries, wells, fields/gardens and linear irrigation works called foggaras). The majority of these sites have never been previously noted or mapped and the date of the sites was unknown at the outset, though they clearly pertained to the historic periods. While further study of the finds and scientific dating evidence is required, the initial results of the brief field visit have major implications for our understanding of Garamantian and early Islamic settlement in south-eastern Fazzan.

A fourth season of work was carried out in the Spring of 1984. The centrepiece of the programme was the investigation of the well-preserved farm in the Wadi el-Amud (Lamout). The main farm buildings were excavated, faunal and botanical samples were collected systematically from within and between them, the field systems and sluices were examined in detail and investigations begun on the geomorphological and hydrological context. This report presents the basic description of the archaeological data resulting from excavation and survey; a second report will present the results of the various laboratory studies which were generated by the field work.

Traditional historiography views the Garamantes from the perspective of the Roman empire, as a peripheral people located in a remote desert. Recent field research is focussing on an inner Saharan perspective, locating the Garamantes at the very centre of trans-Saharan caravan networks. The emergence of the Garamantian state is explained in terms of the ‘longue durée’ and of ‘wide horizons’, in connection with the desertification process that precipitated changes in land use, settlement patterns, and technological availability. Toward the mid-first millennium BC the former pastoral exploitation of the environment changed to one based on oases and caravan traffic, with irrigation agriculture (supplied by foggaras) and long-distance trade (based on the dromedary). Summing up, the case of the Garamantes is presented as a clear case of tentative adaptation to climatic change: some features of the adaptation were short-lived, or required further improvement, while other features remained as definitive acquisitions in the exploitation of the arid environment.

Interpretations of ancient wall-technologies in the Libyan pre-desert are briefly reviewed. The forms, patterns, distributions and geological/geomorphic/hydrologic relationships of walls in a series of study areas are described and interpreted with the aid of a new, non-genetic, ‘wall-technology’ classification. The remarkable hydrological and geomorphic insights of their constructors are clear. Several wall types are shown to have been primarily concerned with functions other than water control, although this aspect is usually dominant. In some cases the location of the walls appears to have been likely to exacerbate the perennial problems of soil erosion and gullying, in others walls appear to have been constructed specifically to control soil erosion. These data have implications for reconstructing past land use and evaluating the degree of success or failure experienced in particular situations.

Excavations at Jarma, associated with the ancient capital of the Garamantes, Garama, in the Fazzan, southern Libya, have recovered a long sequence of charred and desiccated plant remains. The archaeobotanical analysis of the samples have provided an insight into the crop plants cultivated by the Garamantian period occupants of Jarma, and the nature of crop processing activities, plant use and disposal. A broad range of cereal, fruit and fibre crops have been identified which demonstrate both winter and summer season cultivation in the region from the end of the first millennium BC. A sophisticated arable regime relying on foggara irrigation was practised.

Since time immemorial attractive colours have held man's attention. Occasionally he has assigned to them symbolic significance, often as religious or status insignia. The various shades of reds and purples have been especially chosen; some see in them the colour of blood, the sun or fire. The presence of ochre and haematite in Palaeolithic burials may be the first instance of the high regard for these colours.

Probably the most famous red–purple colour associated with religion, status and general elitism is the shell-based purple-dye of the ancient Mediterranean, known variously as Royal purple, Tyrian purple or Imperial purple. The recent discovery of crushed purple-dye producing shells at Berenice provided a welcome addition to this field, a study which Alexander Dedekind, in the 1890s, termed “Pourprologie”.

Before examining the Berenice shells and their context a brief survey will be made of the world-wide usage of shell purple-dye, its Mediterranean origin and distribution, the basic method of production and the contemporary Mediterranean history of the Berenice finds.

Purple-dye is obtained from the hypobranchial gland found in the mantle cavity of living Murex and Thais (or related Purpura and Nucella) marine snails. The purpose of this gland is unclear, though it may be the reason such shells are avoided as food by marine life (and are thus not a useful fish bait). The dye might also be ejected by the mollusc as a defence mechanism, much like the black ink of the octopus or squid.

The coastal alluvial fan sequences of Cyrenaica are important archives of environmental change data, but hitherto relatively little has been known about their formative processes and rates. The Wadi Zewana coastal fan near Tolmeita was studied and a range of dating techniques (U-Th, ESR and OSL) applied to selected components of the stratigraphy. The sequence spans the last two global glacial periods separated by an Interglacial. Cemented alluvial fan gravel units yielded U-Th leachate-residue ages of 201 ± 18 ka, 179 ± 15 ka and 138 ± 8 ka respectively. The fan toe units are interdigitated with bioclastic beach rock deposits dated to 150 ± 10.9 ka corresponding to an Interglacial high stand in sea level and marine recession sequence featuring transgressive lag gravels, beach sand and cemented aeolian dunes dated to 121 ± 8 ka. Within the Wadi Zewana catchment a complex cut and fill history is evidenced. Aggradation phases dated to 76 ± 4 ka, 42.1 ± 5.1 ka and 12.5 ± 1.5 ka are broadly coincident with global glacials and stadials, whilst during the Last Interglacial and successive interstadials the drainage system underwent entrenchment, manifested on the coastal plain as telescopic fan segmentation and associated fan head trenching.

The ‘Burials and Identity’ team of the Desert Migrations Project carried out two main excavations in the 2009 season, at the monumental Garamantian cemeteries of TAG001 and TAG012, by the Taqallit headland. In addition, a detailed survey was made of cemeteries and other sites on the west side of the Taqallit headland, to set the two main cemetery excavations in context. A total of over 2,100 individual burials was recorded in this small area of a few square kilometres. This cemetery survey was combined with further research on the well-preserved foggara systems in this area, which originate at the escarpment among the cemeteries and run in a north-westerly direction towards the valley centre, where some additional Garamantian settlement sites were also located. The foggara research also involved excavation at four locations to try to elucidate issues relating to the dating of these.

A total of 22 burials was investigated at TAG001, an imposing cemetery of stone-built stepped tombs that had been badly damaged by illegal bulldozing in the 1990s. Although these had been subjected to robbing at some point in the past, many preserved considerable parts of the skeletons buried within and some surprisingly complete artifact groups. Of particular importance are a series of Garamantian necklaces in ostrich eggshell, carnelian and glass beads, which we were able to lift in perfect sequence and restring. At TAG012, about 2 km north of the Taqallit headland, we excavated an area of a mudbrick cemetery, exposing 12 square/rectangular tombs. Two further burials were excavated at the dispersed cemetery TAG006, in both cases involving tombs that had an interesting stratigraphical relationship with foggara spoil mounds.

This project began in 1979. Its objectives are to locate, survey and analyse the extensive remains of ancient agricultural settlement that can be found in the wadis of the hinterlands of Tripolitania, Sirtica and Cyrenaica. Within this framework, established by the Department of Antiquities of the Libyan Arab Republic and UNESCO, is the archaeological aim of recording the evidence for periods when extensive areas of the pre-desert were, for whatever reason, cultivated in ways that are not immediately possible today, and in the longer term of assessing those areas where modern farming might be re-established. Archaeology is thus brought into direct line with the aims of the modern world. This paper presents some preliminary comments on the findings of the 1979 survey, which (as we conclude) have raised questions about ancient settlement and land use rather than provided answers; we hope that our full report on the project in Libya Antiqua will be able to offer a reasoned assessment of these issues, understanding which is surely critical for modern agricultural planning.

The first season of fieldwork by the UNESCO-sponsored Anglo-Libyan team under the auspices of the Department of Antiquities, Tripoli, occupied the period from late August to early October. The succeeding months have been taken up with the assimilation, processing and plotting of the initial results. The group was led on the British side by Dr Graeme Barker, of the University of Sheffield, and Professor G. D. B. Jones, of the University of Manchester. Their grateful thanks are due to Dr Salaheddin, at the time President of the Department of Antiquities, and his successor Dr Abdullah Shaiboub, who was at the time in Britain and took an active part in the preparations for the 1979 programme. Mr Abdul Wahab, Controller of Antiquities for Tripolitania took part in the original preliminary survey conducted by Mr Charles Daniels, Chairman of the Society for Libyan Studies, and Professor Jones late in May 1979. Mr Omar Marjub, Controller of Antiquities for Leptis Magna, kindly provided the base for the team at Homs and helped in a variety of ways, both at Leptis and during his visit with Dr Salaheddin in the field. The day-to-day work in the field was organised by Mr Abdullah Ahmed and Mr Muhammed Shitewi, Assistant Controller of Leptis Magna, to whom the Directors are deeply grateful. In Tripoli, the British Ambassador and his wife, Mr and Mrs A. Williams actively fostered the work on the project, not least through their hospitality to members of the team. A particular debt of thanks is due to Olwen Brogan for her constant advice, encouragement and active presence in the field, as well as to Miss Tina Watson.