One million years of the British Pleistocene
At this point the waters of the Colne River ran through a wide floodplain, cutting a gravelly route due south towards its confluence with the mighty Thames barely 15 km downstream. The lower slopes leading to the river allowed access to the waters, to the flint and chert nodules glinting in its shallows that sufficed for knapping, and to fording points that could be used to trap fish and disadvantage prey in the hunt. The patchy stands of pine trees provided wood for fuel and for replenishing the hafts and shafts of tools and weapons. After several months of snow the region was greening up; although it was still cold, reindeer were passing through in number on the way to their spring calving grounds and, here and there, small herds of wild horse grazed on the grassy tundra. From the high ground above this place their numbers could be seen for a great distance across the floodplain and their movements studied. The familiar path of the Thames and Colne – preserved as folk knowledge despite infrequent visits to this edge of the world – had guided these people here.
They were present in small numbers – a task group you could count on one hand – charged with monitoring the reindeers’ movements and beginning the hunt that would provide meat and fat, antler for tools and calfskin and sinew for clothing. A few days in the area should provide enough to take back to the remainder of the small band left behind where the two rivers meet. The day’s hunt had been a success; an adult reindeer had been killed at the ford just downstream; its legs had been removed and carried to this place, a low slope overlooking the river. Fish traps set at this place had produced a freshwater fish and fallen pine wood had been gathered to set a hearth against the growing afternoon chill. While there was still enough light each person set to their tasks, huddling as close as they could to the hearth to benefit from its heat. Sometimes they sat, sometimes they stood, each preoccupied with swiftly and efficiently executing the tasks on which they were dependent. When the hearth had been lit, two wiped dry the nodules of flint they had selected from the chilly waters and knapped them. After removing the chalky cortex they skilfully produced a small series of long, regular blades, using timehonoured techniques to shape the cores and control the flakes and blades knapped from them. Some of these were passed on to the two members busy with completing the dissection of the reindeer legs. In addition to the sinew, meat and marrow the footpads of the animal would be saved to make boot soles. A fifth individual – the last – set to descaling the fish for immediate consumption. After eating, as night fell, the group would set to repairing their javelins and nets before retiring under lightweight bivouacs tethered to the pines. Tomorrow they would rejoin their band, paddling downriver in the skin boats drawn up out of the water and now packed with reindeer antler and meat carefully wrapped in skins. They would leave offerings to placate the waters and the spirits of this place, ensuring a successful return next year, set their paddles in the water, and take one last look back before setting their minds on the brief journey south.
They would never return. They could not know it but their world was coming to an end. Soon, the reindeer and horse would be gone from here, and the grasslands would give way to woodland and forest. First, the boreal woodlands would thicken; later a thick mat of warmth-loving forest would cover the land. Although their hunter-fisher-gatherer way of life would continue for several thousand years more, the vast open lands that had been home to the large herds and their Palaeolithic predators would be no more. The world of gravelly rivers, hills and plains would disappear.
They could not know, but this little group – one of only several scattered about in this vast and untamed world at the northern edge of humanity’s reach – had inherited the legacy of nearly one million years of intermittent visits into this land. Soon, with the reindeer, it would be completely gone.
The fanciful reconstruction above is based on interpretation of Scatter A at the Terminal Pleistocene/Early Holocene Long Blade Industry site of Three Ways Wharf in Uxbridge, Greater London (see Chapter 8). It has a little of our imagination mixed in for sure but is otherwise based on analyses and interpretations of the site’s excavators (Lewis and Rackham 2011). Here, at the Pleistocene/Holocene transition, a small group of terminal Palaeolithic hunter-gatherers camped, leaving several lithic scatters and bones of reindeer and horse. Long Blade Industry sites – which almost certainly are linked to the continental Ahrensburgian culture – are not common in Britain. It seems that there were few people of this cultural attribution in the country – probably for a very brief time – a small and barely perceptible dispersal into a vast landscape during a brief window of opportunity when climate and environment allowed. If we were to go back in time from this point to that of the earliest known hominin dispersal into Britain – currently ~750-980,000 years ago – we would find essentially the same thing; remarkably small groups of humans, in this case a different species, engaged in similar social and economic tasks registered mainly through their involvement with stones and bones.
Our central aim in this book is a synthesis and interpretation of the entire British Palaeolithic record in terms of the occupation, behaviour and societies of the ancient hominins who once roamed these shores. Several of the elements touched upon in the opening vignette are constant themes in this book. Rivers are critical to Palaeolithic archaeology and their ‘fluvial archive’ contains by far the richest record of hominin presence. Equally, rivers were central to the lives of Palaeolithic hominins, forming a focus of critical resources (water, plants, animals and stone), as well as conduits for movement through the landscape. Since at least MIS12 the Thames has formed a main route of dispersal into Britain, being connected at times of low sea level to the Rhine–Meuse systems of Europe, while its many tributaries formed a network for incursions into large parts of the country. Prior to MIS12, the erstwhile Bytham River served a similar function.
The landscapes and environments of the Pleistocene are both alien and familiar. As we shall see in later chapters, during the warm interglacials the vegetation of Britain had a remarkably British feel, a mosaic of woodland and grassland in which most of the component species still occur today. The animals that walked this land, though, were often rather exotic, with mammoth, bison, reindeer, and rhinoceros forming key prey species for humans and non-human carnivores such as lion and hyaena. During colder periods, however, Britain was a truly alien place, its geography unrecognisable as low sea levels connected it to Europe across the vast plains of ‘Doggerland’ (Coles 1998). Ice sheets at times extended as far as the Thames Valley, beyond which existed polar deserts with frozen ground, biting winds and minimal tundra vegetation. During the ‘non-analogue’ environments of MIS3, though, Britain, like much of northern Europe, formed part of the so-called ‘Mammoth-Steppe’ (Guthrie 1982), a rich grassland populated by vast herds of megafauna, in which trees were a rare commodity.
Located at the north-western corner of North-western Europe, Britain formed the edge of the hominin range throughout most of the Palaeolithic. ‘Here, at the very edge of their range, biological and cultural adaptations were stretched to their limits’ (Roebroeks et al. 2011, 113). When combined with the frequency and amplitude of climatic and environmental change, this made Britain a very hard place to live. It is therefore not surprising that for much of the Pleistocene Britain appears to have been unoccupied, or occupied only very briefly by small groups of humans. To paraphrase a geological axiom: the Palaeolithic record of Britain is essentially a long series of hiatuses, interrupted by a few handaxes. While the same may be said of neighbouring regions of Europe, the occupational gaps in Britain are longer and more frequent (Roebroeks et al. 2011). As such, Britain may be considered a population sink, where regular abandonment and/or extirpation necessitated the constant influx of people originating from elsewhere; even during periods of occupation populations may have been reproducing below replacement levels and thus required ‘topping-up’ from outside. This has had a profound impact on the record we find here.
This is not a book about the history of Palaeolithic archaeology in Britain. In order that we may present a coherent interpretation of the British Palaeolithic record we do not even seek to integrate discussion of the history of investigation of British Pleistocene deposits into the main text. Where we discuss key sites – usually in boxes – we note the major excavators and phases of investigation but this is not intended to constitute a comprehensive account of the history of the discipline. There are colleagues better suited to undertaking this and, indeed, excellent accounts of the British contribution to the development of Palaeolithic archaeology and of the historical investigation of British sites already exist. The reader is directed in particular to Grayson (1983) and O’Connor (2007) for general accounts contextualised in the wider scientific world; various papers in Great Prehistorians: 150 Years of Palaeolithic Research, 1859–2009 (Volume 30 of Lithics: the Journal of the Lithic Studies Society for 2009) for accounts of specific individuals; Pettitt and White (2011), White and Pettitt (2009), Weston (2008) and particularly Sommer (2007) for discussion of the work of Buckland, MacEnery and their contemporaries in the first half of the nineteenth century; Gamble and Kruszynski 2009 for discussion of the British involvement in the shattering of the ‘time barrier’ at Amiens and acceptance of deep time in that annus mirabilis of 1859; McNabb (1996b, 2007) for the history of Lower Palaeolithic archaeology with particular reference to the Clactonian and Swanscombe; Scott (2010) for the investigation of what came to be defined as the Early Middle Palaeolithic and Campbell (1977) for the Upper Palaeolithic. A very brief survey, is nonetheless warranted.
Britain was central to the development of Palaeolithic archaeology, along with its closest continental neighbours. The absence of an understanding of deep time until the mid-nineteenth century meant, however, that the earliest discoveries of what we now know to be Palaeolithic materials – the handaxes discovered at Gray’s Inn Lane, London (1679) and Hoxne, Suffolk (1797) – passed largely unnoticed. Similarly, John MacEnery’s discoveries of Middle and Upper Palaeolithic tools stratified under stalagmite floors in association with extinct animals at Kent’s Cavern during the early 1820s remained unpublished during his lifetime, largely due to his deference to the ‘antediluvian’ theories of the Reverend William Buckland, his friend and mentor (White and Pettitt 2009; Pettitt and White 2011). Even Buckland’s involvement with the discovery of an Upper Palaeolithic burial, fauna, stone and bone tools at Goat’s Hole, Paviland, in 1823 and his excavation of a Pleistocene hyaena den at Kirkdale Cave in 1822, did little to shake his conviction. Although MacEnery’s findings were adequately vindicated by the publication of his notes by Edward Vivian in 1845, this came after his and Buckland’s death because Buckland’s validation – which MacEnery was waiting for – never came (White and Pettitt 2009). The world, it seems, was not ready for such an epochal shift.
This would not occur for another 35 years, and then, sadly, not in Britain. At Amiens, in April 1859, Joseph Prestwich and John Evans gave their seal of approval to Jacques Boucher de Perthes’ claims to having found early evidence of humans in association with the bones of extinct animals. A frenzy of discovery occurred over the decades that followed, during which time several of the ‘flagship’ British sites were investigated, including the caves of Creswell Crags, Swanscombe on the North Kent terraces of the Thames, the Hyaena Den at Wookey, Gough’s Cave in Cheddar, Clacton-on-Sea in Essex, and High Lodge in Suffolk, and more systematic investigations in Kent’s Cavern. By the beginning of the twentieth century thousands of new sites and findspots had been identified, most of which still remained undated and poorly understood chronologically. As the number of sites grew, in Britain and especially in France, attempts were made to place some sense of order on the record. The most influential of the resulting chrono-cultural systems was that of Gabriel de Mortillet, who introduced terms such as Acheulean, Mousterian and Magdalenian; still in use today, for better or for worse. Culture history had begun, and from this time dominated British Palaeolithic archaeology until well into the 1960s. Developments during the early decades of the twentieth century also saw the birth of the Clactonian (Chapter 4), the infamous Piltdown hoax and the re-emergence of the pointless eolith debate, with Ray Lancaster and James Reid Moir picking up the reins from Benjamin Harrison and Joseph Prestwich in the quest to prove the existence of ‘Tertiary Man’ (O’Connor 2007). These latter controversies can be seen in the wider context of the ‘scramble’ for evidence of the development of humanity. At this time everything was up for grabs; it had not yet been recognised that Africa was the cradle of humanity and, therefore, it seemed plausible that humans may have originated in Europe. The embarrassing forgery of the Piltdown remains – a recent human cranium matched with the mandible of an orang-utan, both chemically stained to appear fossil and teeth filed clumsily to occlude together (Spencer 1991) – fooled scientists of the time but its authenticity had begun to be questioned by the 1950s.
Emphasis on fieldwork has been constant throughout the history of the British Palaeolithic. The large-scale excavation of caves was largely restricted to the nineteenth and early twentieth centuries. Sadly, this too often took the form of wholesale clearing of vast sedimentary deposits, archaeology and palaeontology from which were selectively retained, lost, distributed across the world and rarely published comprehensively. Cave and rockshelter excavations of the latter half of the twentieth century were typically smaller in scale and usually comprised very small trenches, in many cases thankfully, as they remain unpublished: certain British scholars put trenches through critically important deposits in a number of caves and failed to undertake even the most basic analysis and publication. In worst case scenarios we know they excavated in flagship caves such as Kent’s Cavern but don’t even know the location of their trenches.
By contrast, the monitoring and excavation of productive pits and quarries has profitably continued for the best part of one 150 years. The rate of discovery of archaeological sites in these contexts has diminished over time, as extraction procedures have become increasingly mechanised and urban expansion has made many Pleistocene deposits inaccessible, although these sites continue to provide world-class information. During the second half of the twentieth century, the excavations of John Wymer, Mark Roberts, Francis Wenban-Smith, David Bridgland, Nick Barton, John Gowlett, Danielle Schreve, and the British Museum/AHOB (led by Nick Ashton, Simon Parfitt and Simon Lewis) have been particularly important in drawing together the understanding of the British Quaternary discussed in Chapters 2, 3 and 4. With a few key exceptions, for example Pakefield and Happisburgh (Chapter 2), Lynford (Chapter 6), Glaston (Chapter 7), Three Ways Wharf at Uxbridge (Chapter 8), and to an extent Boxgrove (Chapter 4), most excavations of the past 50 years have concentrated on re-investigating old sites to answer specific chronological, environmental or cultural issues. Sadly, many of the very early discoveries at such sites were, with the best of contemporary intentions, poorly and sometimes completely excavated, meaning that they are now effectively lost to us (Roe 1981). Britain may have taken an early lead in the race to study the Palaeolithic but unfortunately quickly ran out of steam: now is an exciting time to be working in the British Palaeolithic, but we sorely need more new sites, something that only renewed and extensive survey will achieve.
Alongside fieldwork, more sophisticated methods of artefact analysis have been developed in Britain since the 1960s, resulting in classificatory schemes still in use today such as the handaxe classifications of Roe (1968a); and in considerable advances in our understanding of assemblage variation in the Lower Palaeolithic (e.g. McNabb 1992) and Early Middle Palaeolithic (White et al. 2006; Scott 2010) and the establishment of a chronology for the Late Middle and Upper Palaeolithic (e.g. Campbell 1977; Barton and Roberts 1996; Barton et al. 2003 and the numerous publications of Roger Jacobi). In recent years popular (but nevertheless weighty) accounts of the British Palaeolithic have been published (Barton 1997; Stringer 2006), although it is perhaps no surprise that the relatively brief coverage of the Palaeolithic in general accounts of British prehistory (e.g. Pryor 2004; Darvill 2010), however useful, fail to do justice to a period which represents 98.5–99% of British Prehistory.
In the same way that this is not a book about the history of the British Palaeolithic, it is also not a book about Palaeolithic chronology, although we are, of course, entirely reliant upon the ability to hang sites as precisely in time as we can in order to make sense of the record. The host of relative and absolute dating techniques on which Quaternary specialists rely have been developed since the mid-twentieth century (Walker 2005), and it may be said that we have entered a period of maturity in which we can now place our confidence in the reliability of several techniques. We simply make some points here about our use of Quaternary time.
It is not a straightforward task to marshall the variety of complex relative and chronometric dating techniques and seriation schemes into one consistent whole for the purposes of clarity, but that is what we have attempted here. Overall we refer to the Marine Isotope Stage (MIS, alternatively Oxygen Isotope Stage, OIS) system, which has become the global standard among Quaternary scientists (e.g. Shackleton 1987). If the recent dating of Happisburgh is upheld (see Chapter 2) then hominins have visited Britain intermittently over a period spanning at least 25 of such stages.
Marine Isotope Stages reflect relatively long periods of time – typically ~15,000 years (MIS2) to ~60,000 (MIS11). They are subsumed in or define the old units of ‘glacials’ and ‘interglacials’. As such they form the largest definable units of the highly complex climatic and environmental instability of the Pleistocene. Within these, however, several nested scales of change are also observed, and it has become necessary to subdivide each Marine Isotope Stage further. The ice cores, deep sea cores, and terrestrial records preserve evidence of these climatic and environmental fluctuations on the millennial and sub-millennial scale. At the level of hominin dispersals and behavioural change it is probably these scales that provided the adaptive pressures that determined whether Palaeolithic societies survived and ultimately propagated the changes that are visible in the archaeological record.
Two conventions have been established for the naming of these isotopic substages. MIS 11, for example, has been divided into substages based on an alphanumeric system, that is, 11c, 11b and 11a (e.g. Tzedakis et al. 2001). In other records, however, (e.g. MD900963, Bassinot et al. 1994) a more complex pattern can be seen with additional warm–cold oscillations (Figure 2.5). Therefore, an alternative system identifies negative and positive isotopic events, which are numbered using a decimal system (Imbrie et al. 1984; Bassinot et al. 1994; Desprat et al. 2005). This has the advantage of allowing additional isotopic events to be incorporated as they are discovered. The two conventions differ because the first denotes periods of time, whereas the second identifies specific isotopic events, and therefore the terminology is not directly interchangable. We discuss issues where they appear in the text.
From an environmental point of view, the Hoxnian Interglacial (MIS11) and the Ipswichian Interglacial (MIS5e) have been divided into pollen subzones, the identification of which has proved critical to our understanding of exactly when hominins were present in Britain. Alongside these, micro- and macro-faunal, coleopteran and molluscan biostratigraphy are critical to the division of time and correlation of sites. In chronometric terms, palaeomagnetism and amino-acid racemisation techniques have proven invaluable in the seriation of sites, but do not produce dates. Correlating such seriated sites with Marine Isotope Stages is nowadays possible with high degrees of confidence, but it is not without its problems, as will be seen, for example, in Chapter 2. Non-radiocarbon dating methods such as thermoluminescence (TL), optically-stimulated-luminescence (OSL) and Uranium-series underpin our chronology and, for the Middle Pleistocene and earlier stages of the Upper Pleistocene, are associated with measurement imprecision consistent with that of the other techniques. The powerful combination of all these methods has resulted in the impressive chronological control of British Middle Pleistocene sites we rely on in Chapters 2, 3, 4 and 5.
Non-radiocarbon dating techniques noted above, particularly Uranium-series, TL and OSL, while of critical use from MIS3 backwards, are associated with relatively large measurement errors (imprecision) which render them of limited use for structuring late MIS3 and MIS2 archaeology in time, at least where radiocarbon measurements are available. As a result we rely in Chapters 6, 7 and 8 almost entirely on radiocarbon for our chronological framework. We will not rehearse in detail here the usual issues relating to radiocarbon accuracy and precision, but make some simple points which, we hope, justify why we use calendrical (calibrated) radiocarbon dates in the way we do. Correction for radiocarbon inaccuracy has been available back to ~50 ka (14C) BP for the last decade, notably in the form of the CALPAL curve (see below). INTCAL09 now calibrates back to a little beyond ~44.5–45 ka 14C BP, that is ~48 ka BP in calendrical terms (Reimer et al. 2009). The result of calibrating measurements using these curves has revealed how considerably radiocarbon measurements underestimate real time in this period; the cause being the complex factors relating to the influx and production of 14C in the Earth’s atmosphere, itself, it seems, governed to a large extent by changes in the Earth’s magnetic field. Here is one example of such age underestimation which we revisit in Chapter 7: a tooth of the scimitar-toothed cat Homotherium latidens dredged from the North Sea close to the Brown Bank that has been directly dated to 28100 ± 220 (14C) BP (UtC-11000, tooth) and 27650 ± 280 (14C) BP (UtC-11065, mandibular bone). These calibrate to ~31–32 ka BP, revealing that the uncalibrated radiocarbon measurements on the dentary underestimate its real age by four to five thousand years.
The ability of specialists to remove contaminating sources of carbon from dating samples and thus isolate only the carbon relevant to the actual age of the sample has also had a significant effect on chronometric accuracy. Recent improvements in pretreatment methods, notably ultrafiltration, seem to be far more efficient at removing contaminating sources of carbon and thus of producing more reliable (i.e. accurate) age estimations. Although the technique was not invented at Oxford – and was indeed practised in other laboratories before it was adopted there – it has perhaps become particularly associated with this laboratory’s work on the chronology of the Late Middle and Upper Palaeolithic (e.g. Higham 2011). The redating of ultrafiltrated carbon from samples originally pretreated using non-ultrafiltration methods has often resulted in new (and one assumes more reliable) measurements that are either younger or older than the original results. It would be fair to say, however, that new results typically produce older ages. There is, therefore, a strong tendency for more recently produced measurements to be older, and thus shift back in time our chronologies for the Late Middle and Upper Palaeolithic, while at the same time eliminating chronometric noise.
All dates we use in the book may be considered to be ‘calendrical’, that is we present calibrated radiocarbon dates. In order to correct the uncalibrated radiocarbon measurements pertinent to Chapters 6, 7 and 8 we have used the CALPAL curve, a splined, multi-component curve based on high-precision U/Th and radiocarbon data from Hulu Cave synchronised with palaeoclimatic data from the Greenland ice cores (Weninger and Jöris 2004, 2008). The reason we use this rather than the INTCAL09 curve is familiarity and loyalty: the effort expended into developing CALPAL over the last 15 years or more made this available long before INTCAL was extended back beyond ~25 ka BP although the two datasets are very similar. Reimer et al. (2009, 1112) suggest that where calibrated dates are used original radiocarbon measurements on which they are based should also be cited. We do this where we think it is necessary, but in the interests of space do not make a habit of it. We cite references to the publications in which the original radiocarbon measurements were presented and thus, where we do not present original measurements in tables or text, readers, should they wish, may follow a trail back to original sources and check the accuracy of our calibration. In any case we do not attempt any correlations of dated material (between sites, or with climate, for example) that require high degrees of precision, and even towards the end of the Pleistocene there is still a large degree of imprecision; single radiocarbon measurements around 12,000 BP produced in recent years using ultrafiltration pretreatment methods – which may be considered to be the most precise currently available – typically have errors in the order of 50 (14C) years (see for example the results on samples from Gough’s Cave – Jacobi and Higham 2009); sets of such measurements from contexts that one might assume to be chronometrically contemporary (e.g. assemblages such as Gough’s) typically produce age ranges of three to four centuries, and even Bayesian analyses – which perhaps specialists put a little too much faith in – result in modelled ranges of around two centuries for samples that are assumed to be contemporary or which reflect single events. This is some achievement for which the radiocarbon community should be justifiably proud, but also a degree of imprecision with which we are probably stuck and thus that, in our opinion, merits our use of calibrated dates. When we quote a date ‘~14.5 ka BP’ it should be assumed that there is a spread of uncertainty of around a century either side at 2σ. Where a set of dates have been produced for a given assemblage we state the range over which measurements overlap at 2σ; it will be seen in the text that, for the Late Glacial, this typically results in ranges of two, three, four or more centuries (Chapter 8). Needless to say, the further one goes back the greater the imprecision; measurements at around five half lives of radiocarbon, for example ~30 ka (14C) BP – pertinent to the arrival in Britain of the first Homo sapiens groups – typically possess age ranges of some seven centuries (Chapter 7), and at around seven half lives/~42 ka (14C) BP – pertinent to late Neanderthals – around two millennia.