Introduction

1Our understanding of Palaeolithic human demographies, behavioural adaptation to environmental change, technological change and cultural evolution is heavily reliant on cave sedimentary archives. As they sometimes cover multiple glacial-interglacial cycles, deep cave stratigraphies provide unparalleled opportunities to study the timing, nature and duration of human occupation at a single location over thousands of years. They have provided key chronological tethering points for Palaeolithic narratives worldwide.

2In the Mediterranean region, cave records are critical for present understanding of major debates in human evolution including the migration(s) of Homo sapiens from Africa between c. 130 and c. 40 ka (e.g. Barton et al. 2009; Benazzi et al. 2011; Bosch et al. 2015; Hershkovitz et al. 2015), the demographic history of Neanderthals and their interactions with Homo sapiens (e.g. Higham et al. 2014; Hublin 2000), the Middle to Upper Palaeolithic technological transition (e.g. Marín-Arroyo et al. 2018), behavioural and technological adaptations to changing environmental conditions (e.g. Banks et al. 2013), and the emergence of modern human behaviour (e.g. Bosch et al. 2015). In these models, deteriorations or ameliorations in climate, on stadial-interstadial timescales, are often considered key drivers of change. Increased aridity, in particular, has often been implicated as a key factor in narratives of regional abandonments and the development of ‘refugia’ (e.g. Carrión et al. 2008; Delanges et al. 2013), human migrations (Garcea 2010a), population stress (Maier et al. 2016), and technological change (e.g. Straus 2015).

3Several changes in the Late Pleistocene archaeological record of the Mediterranean basin have been identified as evidence of human responses to arid or highly degraded climatic conditions. In terms of subsistence, they include changes in resource exploitation, where arid-tolerant species abruptly begin to dominate the faunal record (e.g. Campmas et al. 2015; Starkovitch & Ntinou 2017), the incorporation into the diet of reliable but energy-costly food sources such as molluscs (Lubell 1984), and increased exploitation of dryland plant foods (Barton et al. 2018). In terms of technology, changes in raw material use and/or core reduction strategies have been identified as evidence for technological adaption to environmental change (e.g. Banks et al. 2013; Pereira & Benedetti 2013), or the development of tool types that might be associated with dryland resource acquisition as has been argued in the case of the Aterian techno-complex of North Africa (Garcea 2010a; Scerri 2013a). The robustness of these narratives, though, depends heavily on precise intra-site and inter-site chronological frameworks, which in turn necessitate a thorough understanding of cave sedimentation processes.

4Our understanding of the geomorphological processes that created and transformed sediments in Mediterranean caves, and of the manner in which these processes interacted with the human record, has greatly improved over the past two decades (e.g. Goldberg & Bar-Yosef 2002; Hunt et al. 2010, 2015; Inglis et al. 2018; Woodward & Goldberg 2001). Cave sedimentation processes here were primarily characterised by episodic depositional regimes which were highly responsive to regional or indeed local changes in temperature and rainfall. This was especially the case in Marine Isotope Stages 4‑2 (dated globally to 71‑11.7 ka), when there were millennial-scale fluctuations between stadial and interstadial climatic conditions, resulting in oscillating types of sedimentation. Aridity could be expressed by aeolian sedimentation, where source material was available, or by standstills or hiatuses in sediment accumulation if source material to permit dry sediment deposition was lacking. Such hiatuses could be of millennial-scale magnitude and if they coincided with significant phases of human demographic history and technological change (e.g. Ksar Akil, Lebanon (Bosch et al. 2015; Douka et al. 2013); Yabroud II, Syria (Pastoors et al. 2008)), they present significant challenges to establishing robust chronologies of the human use of caves and of interactions between people and climate. In the rest of this paper we discuss these challenges in the context of the deep sedimentary record of the Haua Fteah cave in Cyrenaica, northeast Libya.

The Haua Fteah 1950s and recent excavations

5The Haua Fteah, located at 22^o3'5″E and 32^o53'70″N, is a large limestone cave overlooking the present-day coast (figure 9.1). The north-facing entrance is about 1 km from the coast and is c. 50 m wide and c. 20 m high. The interior roofed area measures c. 80 m across. The cave is situated on the northern edge of the Gebel Akhdar or Green Mountain, a massif rising sharply from the sea to almost 1000 m above sea level that extends for c. 300 km east from the modern city of Benghazi and c. 100 km from the coast to the Sahara. The Gebel Akhdar receives over 200 mm of rainfall a year, the limit for dry farming without irrigation, and as a result, much of its landscape today is semi-arid Mediterranean like that of Crete or the mainland of southern Greece. To its east are more than 700 km of flat desertic littoral as far as the Nile delta. With the unending gravel plains and sand dunes of the Sahara to the south, the Gebel Akhdar is an island of green separated from the rest of North Africa by hundreds of kilometres of arid steppe or waterless desert.

figure 9.1

6In three seasons of excavation (1951, 1952, 1955), a team led by Dr Charles McBurney of the University of Cambridge excavated a 14 m-deep stepped trench in the middle of the roofed area of the Haua Fteah, revealing a sequence of cultural deposits that has since been a key reference sequence for North African prehistory (McBurney 1967). Based on the typology and technology of the lithic material, he defined seven cultural phases of occupation (table 9.1). The earliest phase (A), at c. 14.0‑7.5 m below the present ground surface (from the base of his trench to Layer XXXV), was a Middle Stone Age industry based on flakes and blades which he termed the “Pre-Aurignacian” because he thought it resembled Pre-Aurignacian and Amudian assemblages in Southwest Asia. This was overlain (Phase B, Layers XXXIV-XXV at c. 7.5‑5.0 m depth) by “Levalloiso-Mousterian” industries, so called from their broad comparability with these industries in the Levant and Europe. Two human mandibles were found in these layers (McBurney et al. 1953). Phase C (Layers XXV‑XVI, at c. 5.0‑3.0 m depth) was an industry characterized by blade technology and an Upper Palaeolithic tool inventory that he termed “Dabban” after the Cyrenaican cave of Hagfet ed-Dabba where he had found similar material (McBurney 1960). This was succeeded by a microlithic late or final Upper Palaeolithic industry (Phase D, Layers XV-XI c. 3.0‑2.0 m) and termed the “Eastern Oranian” or “Iberomaurusian” from its similarities with assemblages in the Maghreb (northwest Africa). Phase E (Layers X and IX at c. 2.0‑1.5 m depth) was a Mesolithic-type microlithic industry with parallels to Capsian assemblages in the Maghreb and hence classified as the “Libyco-Capsian”. Above this, at c. 1.5‑1.0 m depth, was the final prehistoric occupation (Phase F, Layers VIII-IV), with Neolithic pottery and domestic sheep and goats, which was termed “Neolithic of Libyco-Capsian tradition” because of similarities in stone tools (the frequency of backed pieces, for example) with the preceding Libyco-Capsian. The prehistoric sequence was capped (Phase G, Layers III‑I at 1.0‑0 m) by a substantial boulder-supported structure dating to the Graeco-Roman period covered by burnt animal dung and other evidence of animal enclosures dating to recent centuries.

7The ages of the cultural phases in the upper part of the sequence were established with reasonable confidence by a series of radiocarbon or ¹⁴C measurements (table 9.1). At that time the method had a limit of around 40,000 years, the calculated age of the beginning of Phase C, the Dabban. The ages of the earlier cultural phases, the Levalloiso-Mousterian and Pre-Aurignacian, could only be estimated from calculations of rates of sedimentation: these provided McBurney with an estimate for the earliest Pre-Aurignacian occupation of around 80,000 years ago. The cultural sequence was set within a framework of major phases of environmental/climatic change established by pioneering studies of sediment granulometry, the mammalian fauna (Higgs 1967) and measurements of oxygen isotopes in marine shells (Emiliani 1955; Emiliani et al. 1963).

8New excavations by an international team began at the site in 2007 as part of a survey and excavation project led by GB termed the Cyrenaican Prehistory Project (CPP) (Barker et al. 2007, 2008, 2009, 2010, 2012; Farr et al. 2014; Rabett et al. 2013). The excavations had to be suspended for a year during the civil unrest that culminated in Muammur Ghaddafi’s death in October 2011, were resumed in 2012 and 2013, suspended again in 2014 and 2015 with the deterioration of security, and finally completed in 2016 by members of the Libyan Department of Antiquities who had been part of the CPP work since its inception. An Advanced Investigator Grant awarded to GB from the European Research Council (“TRANSNAP: Cultural Transformations and Environmental Transitions in North African Prehistory”) transformed the scale of the project in 2009‑2015 (see Acknowledgements).

9At the end of McBurney’s excavations, his 14 m-deep trench consisted of three stepped units (figure 9.1): an Upper Trench (the new project’s terminology, not McBurney’s) measuring approximately 10 × 11 m surface area × 2 m deep; a Middle Trench measuring about 7 × 6 × 5.5 m; and a Deep Sounding that was published by McBurney (1967) as being 2.5 × 1.5 × 6.5 m although on re-excavation it measured 3.8 × 1.6 × 6.5 m. He had backfilled the trench at the end of the excavation and over the several campaigns the backfill was emptied out and the original trench walls, which were remarkably well preserved with many of the McBurney aluminium labels still in situ, were cleaned and studied (Hunt et al. 2010; Inglis et al. 2018). Over 1 m of unexcavated sediment was located below the McBurney Deep Sounding. A series of 30 × 30 cm column samples were excavated down the cleaned faces to collect a continuous sequence of sediment samples for dating and palaeoecological analysis, and two 2 × 1 m trenches were excavated: Trench M on the southern side of the Middle Trench and Trench D down the southern side of the Deep Sounding (figure 9.1). Climatic and environmental conditions during the site’s occupation were investigated by sediment micromorphology, pollen, charcoal, molluscs, vertebrate fauna, and isotope analyses of animal teeth and marine and terrestrial molluscs, with several of these classes of data also informing on subsistence activities. A number of dating techniques were employed including over 200 radiocarbon dates on charcoal and shell, 26 OSL dates on quartz and feldspar, eight ESR dates on animal teeth, and 20 AAR dates on shell (Douka et al. 2014; Hill 2016; Jacobs et al. 2017). The sequence was bracketed by three identified volcanic tephras. The dating programme demonstrated that the base of the sedimentary sequence dated to c. 140‑150 ka —almost twice as old as McBurney had estimated (table 9.1) (Jacobs et al. 2017).

The Pleistocene sedimentary sequence

10The newly exposed stratigraphy has been divided into five major sedimentary facies on the basis of field observations, bulk sediment analyses and micromorphology (Hunt et al. 2010; Inglis et al. 2018). These have been fully described elsewhere (Douka et al. 2014; Hunt et al. 2010; Inglis et al. 2018) and their principal characteristics are summarised here and in figure 9.2. Figure 9.3 illustrates Facies 2‑5 in the Middle Trench. Facies 1 formed in the Upper Trench in the Holocene and is not described here.

figure 9.2

figure 9.3

Description

• Facies 5 is located from Trench S at the base of the stratigraphy c. 15.0 m below the present ground surface through the Deep Sounding to the base of the Middle Trench at c. 6.5-6.0 m depth. It consists of reddish silty clay layers of variable depth interpreted (based on micromorphological, bulk sediment and field data) as redeposited mass movements of soil material mixed within the cave with soil from the surrounding hillsides, or as the deposition and recycling of sediment already within the dry shelter through aeolian processes. It developed between c. 140‑150 ka as exposed at the base of Trench S and c. 68 ka at the base of the Middle Trench, so began towards the end of MIS 6 (dated globally to 191‑130 ka), mostly formed in MIS 5 (globally dated to ca.130‑71 ka) and ended early in MIS 4 (globally dated to 71‑57 ka).
• Facies 4 (at c. 6.0‑5.5 m depth) consists of silty greyish sediments and angular limestone gravel interpreted as éboulis created by increased physical weathering of the bedrock. Facies 4 has a modelled age of c. 68‑47 ka (Douka et al. 2014), and therefore has broad contemporaneity with MIS 4.
• Facies 3 (at c. 4.0‑5.5 m depth) consists of dense reddish clay-rich fabrics alternating with lighter-coloured silty clay fabrics, interpreted as reflecting colluvial inwash and aeolian deposition respectively (Inglis et al. 2018). At the top of the facies, those corresponding to McBurney’s Layer XXV show an increase in clay-rich layers interpreted as inwash, layered between sporadic occurrences of increasingly thick éboulis lenses. The modelled age of Facies 3 is 48‑34 ka, indicating that it developed entirely within MIS 3, which is globally dated to 57‑29 ka.
• Facies 2 (at c. 2.0‑4.0 m depth) consists of silty greyish sediments and éboulis similar to but thicker than those of Facies 4, containing a series of thick, red clay lenses, interpreted as the transport of well-developed soil material from the surrounding hillsides into the shelter during inwash events, potentially caused by landscape degradation. Within Trench M, Facies 2 includes a large truncation feature, which may have resulted from tectonic adjustment or from a high energy inwash event. The modelled age of Facies 2 is 35‑12 ka, spanning the latter part of MIS 3 and all of MIS 2 (globally dated from 29 ka to 11.7 ka, the Pleistocene/Holocene boundary).

Climatic interpretations

11Climatic interpretations based on the sedimentary environments in the Haua Fteah are constrained, like in other Mediterranean caves (Frumkin et al. 2016), by cave morphology, aspect, and nature of the local source material. Whilst the shallow hillside the cave is located in means it is well-placed to capture and preserve soil eroded during periods of landscape instability, its north-facing aspect in the lee of the Gebel Akhdar shields it from Saharan dust transport and deposition, one of the major sedimentological markers for aridity in the region. The likely dominant environment within the cave over time was been low-energy, localised aeolian reworking of local fine material; an environment that could have existed in both stable, warm environments, and arid conditions when a lack of rainfall meant that soil material from a denuded landscape was not transported into the shelter (Inglis et al. 2018). At times, rapid, mass mudflows transported large amounts of material in short periods of time, in conditions of landscape instability; these deposits have enhanced clay content, are more compact and deeper red in colour and likely contain re-worked well-developed soil components. Weathering of limestone, and production of éboulis, likely primarily through freeze-thaw processes, can be used as a marker of cooler conditions in the Gebel Akhdar. Whilst the latter two processes can be seen as positive indicators of environmental conditions outside of the cave, the accumulation of sediments built up in the Haua Fteah by low-energy reworking can only be robustly viewed as an absence of the conditions which produced the mudflows and éboulis.

12Mudflow events occurred most noticeably in the Deep Sounding at c. 13.0‑12.0 m depth where they can be correlated with MIS 5e (Jacobs et al. 2017: 84) and in stratigraphic association with Facies 3 deposits in the Middle Trench at c. 4.5 m depth, for which OSL sample HF6 provides a date of 47.9 ± 6.0 ka. Units which are predominately silt but with some sand content are more yellow in colour and may result from minor mud-flows of lesser developed soil cover and/or aeolian derived material. It is notable that silt units within Facies 3 and 2, in the MIS 3 and MIS 2 part of the sequence, become increasingly yellower, possibly reflecting progressively deteriorating climate (figure 9.3).

13Signatures of deteriorated climate are also evident in Facies 4 in the lower part of the Middle Trench at c. 6 m depth, where physically weathered gravels (figure 9.4: E), dated to MIS 4, are capped with a well-defined carbonate crust (figure 9.4: G) and a large combustion feature with the heavily weathered surface (see figure 9.3). Above the combustion feature, a sedimentary hiatus of c. 15,000 years is confirmed by a combination of OSL and cryptostratigraphic analyses (Douka et al. 2014). Repeated sub-annual‑millennial-scale hiatuses in sedimentation are apparent within the silt-dominated sediments of Facies 5 and Facies 3. During excavation these were noted as being characterised by extremely abrupt sedimentary contacts (figure 9.4: F), and desiccated (D), bioturbated (B) and concreted (A) surfaces. Cultural material such as lithics and animal bone was found predominately on the surfaces of the silt units, and not contained within the sediments themselves.

figure 9.4

14In the upper part of the Middle Trench, the compacted clast to matrix-supported, angular to sub-angular, physically-weathered gravels (éboulis) (figure 9.4: C) formed in the cooler climate of MIS 2. These gravels are interspersed with orange silts that represent mudflow events and periods of landscape instability. Sedimentary contacts between the silts and gravel deposits are very sharp, perhaps indicating a degree of depositional hiatus or aeolian winnowing occurring in between phases of sediment deposition. During excavation, stone tools were found mainly on the surfaces of gravel units, but sometimes within the deposits themselves. This vertical distribution of cultural material may have resulted from the overall slow accretion rates of the éboulis deposits and the continued use of the cave by human groups during times of degraded climate, although minor reworking of the deposits and cultural material cannot be discounted. Some of the gravel units have weathered upper surfaces, with associated lithics, postholes and anthropogenic combustion features (potentially hearths) indicating human activity occurring on stable surfaces during stadial phases.

15We observed a distinct stratigraphic truncation towards the top of the Middle Trench at c. 3 m depth, forming an unconformity between the stratigraphic units on the east-facing and west-facing sides of the trench that is also visible in the McBurney section drawings (McBurney 1967: fig. I.5). The feature was lined with Biancavilla tephra from Mount Etna in Sicily, dated there to 14,180 ± 260 BP or 17,920‑16,810 cal BP (Kieffer 1979), calibrated using IntCal09 (Reimer et al. 2009). The process or processes that caused this abrupt stratigraphic truncation is/are unclear, although a collapse through tectonic adjustment in the karstic landmass is a possibility, given the context of rising sea levels in the Mediterranean basin and the proximity of the Haua Fteah to major plate margins. The deposits directly above the Biancavilla tephra are well-defined, loose, clast-supported debris avalanche deposits which contain large quantities of Oranian cultural material.

Sedimentary hiatuses and human behaviour

16While the sedimentary hiatuses observed in the Haua Fteah’s MIS 4‑2 sedimentary sequence in the Middle Trench are variable in their formation processes and duration, four broad scenarios can be defined (figure 9.5). Type A consists of singular hiatuses of long (millennial-scale) duration where no éboulis-type or dry-process deposits were accreting. The main one observable is in the latter part of the Facies 4 sediments, and has a minimum duration of c. 13,000 years (Douka et al. 2014). Type B consists of repeated hiatuses of sub-annual to centennial-scale duration within sediments dominated by fine grained deposits such as red-yellow silts and clay interspersed with physically weathered gravel deposits. These characteristically occur between mud flow and aeolian deposition within Facies 3. Type C is a more pronounced version of Type B and consists of hiatuses of likely decadal to millennial-scale duration within predominately ‘cold phase’ mud-flow events and éboulis formation in the upper part of Facies 3, or within éboulis in Facies 2 sediments. Type D consists of sedimentary truncation events of variable chronostratigraphic implication, including the one associated with the Biancavilla tephra at c. 17,000 cal BP. All four of these hiatus types are associated with signatures of human occupation such as spreads of stone tools, butchered animal bone and combustion features. The repeated associations between hiatus surfaces and cultural material imply that the formation of the sedimentological and archaeological record was highly episodic. The implication is that cultural material on hiatus surfaces is liable to represent palimpsests of accumulation potentially containing and mixing material related to different occupations and different activities (Bailey & Galanidou 2009).

figure 9.5

17The long depositional hiatus identified here as Type A, dating to the latter part of MIS 4, appears to occur immediately on top of the heavily-weathered burning unit described by McBurney as ‘an immense hearth’ (McBurney 1967: 121). This burning unit lies directly on top of a thin, laterally extensive, carbonate precipitate (figure 9.4: G), which probably represents some degree of sedimentary stasis. The burning event occurs towards the base of McBurney’s Layer XXX and on top of Layers XXXI and XXXII. Described by him as ‘small scree’ (McBurney 1967: 117), layers XXXI and XXXII almost certainly reflect the climatic downturn associated with the start of MIS 4.

18The two Homo sapiens mandible fossils were recovered from Layer XXXIII, an underlying silt unit containing stone tools of Levalloiso-Mousterian type. A modelled age estimate of 73‑65 ka places the human fossils very broadly in MIS 4, but the overlying éboulis-type deposits and the dated depositional hiatus position the mandibles stratigraphically towards the earlier part of this age estimation. The mandibles attest to the occupation of the cave by anatomically modern humans at the end of MIS 5a or beginning of MIS 4. The geoarchaeological association of the fossils suggests that this occupation occurred prior to a major climatic downturn. Stone tools and faunal remains attest the continuation of occupation, presumably also by anatomically modern humans, through a subsequent phase of degraded climate (the éboulis Layers XXXI and XXXII). The stone tools in these units are broadly Levalloiso-Mousterian in type, although McBurney thought the assemblage also had ‘Aterian’ attributes. The faunal assemblages from Layers XXXI and XXXII are dominated by Barbary Sheep, Ammotragus lervia, in contrast with the wider faunal range in the layers below these that included larger antelopes needing good access to water (Higgs 1967). Though the argument cannot be pressed given the size of the samples, it is possible that these changes in the lithic and faunal material in the Haua Fteah reflect technological and subsistence adaptions to aridification and environmental degradation during the earlier part of MIS 4.

19The large burning event in Layer XXX that caps the éboulis, and its associated lithic material, are the last solid evidence of human occupation in the cave for some considerable time. McBurney also remarked on the absence of material from the overlying Layer XXIX (McBurney 1967: 121). Without other cave stratigraphies of this time period in the Gebel Akhdar it is impossible to know whether the apparent cessation of human occupation in the cave between c. 65 ka and c. 50 ka was a local phenomenon or part of a more widespread depopulation of the massif at this time.

20Moving up the sequence, Facies 3 consists of layers of fine red-yellow silt interspersed with physically-weathered gravel deposits, which become more pronounced towards the top of the facies (Types B and C hiatuses). Layer XXV, described by McBurney as a ‘scree’ or éboulis-type deposit, marked for him the Levalloiso-Mousterian to Early Dabban transition. In our chronostratigraphic model (Douka et al. 2014), the transition to the Dabban is estimated at c. 40 ka (65.4%) or 43.5 ka (95.4%). This places it within the first series of major climatic downturns (Greenland Stadials 11‑9) that are recorded in northern hemispheric ice core records (figure 9.5). In Europe, population replacement models are frequently cited to explain the Middle to Upper Palaeolithic transition, given the assumed associations of Neanderthals with the former and of modern humans with the latter. McBurney argued the same in the case of the Levalloiso-Mousterian/Dabban transition in the Haua Fteah, though we now know that both industries were made by Homo sapiens. An alternative hypothesis for the Gebel Akhdar is that the techno-cultural shift may have had its genesis in technological adaptations to deteriorating climatic conditions and increasing aridity.

21Further up the Middle Trench, a gap in the OSL chronology is evident between samples HF3 (31.9 ± 3.5 ka) and HF2 (18.3 ± 1.4 ka) on either side of Type C and Type D hiatuses (figure 9.5). The minimum span of this gap is c. 10,000 years. Layered éboulis-type sedimentation is evident between the OSL sampling locations, and the gap in the OSL age estimates indicates slow sedimentation rates, consistent with éboulis-type depositional processes. During the Type C hiatus a stable surface developed on the cave floor, characterised by an increase in lithic density, a thin layer of anthropogenically-derived ash, and a series of cut features filled with ash and charcoal fragments interpreted as post-holes (figure 9.6). A radiocarbon date from charcoal in one of these fills provides an age estimate of 19,680 ± 140 BP, or 24,076‑23,331 cal BP (95.4%), placing this Late Dabban occupation in Greenland Stadial 3 (GS3).

figure 9.6

22Faunal evidence from the equivalent McBurney layer (Layer XVII) is almost exclusively (n = 76) Ammotragus lervia, with no Bos or Gazella represented. In-progress microwear studies of Dabban artefacts indicate that the toolkit now included hafted projectiles (G. Mutri, pers. comm.) that were presumably more effective for killing fast-moving prey animals such as Barbary sheep in the increasingly open steppe conditions. The use wear and organic residues on a grinding stone dated to around 31,000 years ago demonstrate its use for crushing nuts and grinding the seeds of wild grasses such as goat grass (Barton et al. 2018), indicating that lower ranked food resources were increasingly targeted as aridity developed. Techniques were also developed for storing pine nuts and dried fruits (Morales 2010). In short, it is clear that the human population using the cave during the harsh conditions of Greenland Stadial 3 was highly adaptable and resourceful. Overlying stratigraphic layers continue to yield lithic and faunal material, implying that similar technologies and subsistence strategies enabled human populations in the region to survive through the subsequent climatic stress of the Last Glacial Maximum.

Conclusion

23Detailed geoarchaeological investigations of the MIS 4‑2 deposits at the Haua Fteah have begun to elucidate the timing, nature and duration of geomorphological site formation processes and signatures of anthropogenic occupations. The combination of fine-scale stratigraphic analyses combined with a multi-method dating strategy (e.g. ¹⁴C of charcoal and shell, OSL, cryptotephra, ESR dating of bone) has revealed the highly episodic nature of sedimentation, in particular revealing a succession of long-duration sedimentary hiatuses that are critical for understanding the chronology and character of the human use of the cave.

24The implication of the MIS 4‑2 stratigraphy in the Haua Fteah is that sedimentary hiatuses can impact on the archaeological record of hominin-occupied caves in four main areas. First, where lithic material is deposited on a cave floor surface without rapid accretion of deposits, palimpsests of activity are likely to form. Second, where spit-excavation methodologies incorporate sedimentary units (and sedimentary hiatuses) of different ages, and where single context excavation methods do not account for episodic depositional processes and stabilised cave surfaces, the archaeological material collected from sedimentary contacts with the overlying and/or underlying stratigraphic unit may well be of different ages. Third, where human occupation occurs during a sedimentary hiatus phase, it is likely to have a non-synchronous relationship to the sedimentary units above and below, and absolute dating and palaeoenvironmental proxies that use bulk sediment sampling methods cross-cutting sediment hiatuses may not directly relate to human occupation phases. Finally, the various hiatus processes that we can discern in the Haua Fteah sedimentary record are critical for understanding the associated archaeological record. Similar relationships between cave sedimentation processes and human occupation histories should be widely expected in other Mediterranean caves. The implication of the new work in the Haua Fteah is that understanding the effects on the archaeological record of stratigraphic hiatuses caused by degraded environmental conditions or events is likely to be critical in terms of building robust models of technological, subsistence and demographic change from cave archaeologies.

Acknowledgements

25We thank in particular the Department of Antiquities of Libya for its permission to undertake the fieldwork at the Haua Fteah and their sustained commitment to supporting the project. The project was initially termed the Cyrenaican Prehistory Project, with the excavation permit arranged by and funding provided by the Society for Libyan Studies, whose support is gratefully acknowledged. From 2009, it was also termed the TRANS-NAP project from the acronym of the European Research Council grant to the author (Advanced Investigator Grant 230421: “Cultural Transformations and Environmental Transitions in North African Prehistory”), which provided the major funding for the project that is also gratefully acknowledged. Funding was also provided by the British Academy, the Radiocarbon Facility of the Natural Environment Research Council, and the Society for Libyan Studies. That the evidence discussed in this paper derives from the efforts of a large multi-disciplinary team will be apparent from the project publications cited here.