Editor Mike Pitts
The Lost Valley
Excavation on the route of a proposed road south of Salisbury, Wiltshire, has revealed what could turn out to be one of the country's most important early human sites. Ken Whittaker, Mark Beasley, Martin Bates and Francis Wenban-Smith repot
After Salisbury’s bypass was cancelled in 1997, Wiltshire County Council assembled smaller schemes to address local traffic problems. One such scheme was the Harnham relief road, to take heavy vehicles out of the historic and residential city core.
Early archaeological evaluation of the route revealed Palaeolithic flint tools in the ploughsoil near Britford, on a low hill overlooking the river Avon to the east. These were not the only significant finds: excavated remains include two large post holes, and possibly a rectangular enclosure ditch, associated with Later Neolithic Peterborough pottery, an extremely rare occurrence.
The flints were sufficiently interesting, however, that RPS Historic Environment Consultants, acting for the council’s consultant engineers (Ringway Parkman), contracted Gifford and Partners and Pre-Construct Archaeology (PCA) to investigate further.
So between October 2002 and January 2003, a team of scientific specialists was brought in under Martin Bates and Francis Wenban-Smith. PCA’s Jim Leary excavated five evaluation trenches, each up to 30 m long but less than 2 m wide.
The results were spectacular.
With no prior expectations, we found a small buried valley crossing the hill that as much as 300,000 years ago had been a tributary of the Avon. During a relatively warm period in an era that was otherwise glacially cold, hominids, probably ancestors of Neanderthals (it would be another 100,000 years before they evolved), had butchered animals there with flint tools. They might even have set fires: if so, this would be by far the oldest proven fire use in Britain.
Animal bones survive in good condition, and the flint debris left from making handaxes can be fitted back together. This is the first time anything of this particular date has been found in Britain so well preserved. There may be 750 square metres of buried ancient land surface, with the potential to give unique insights into a little understood period of Europe’s earliest history.
Cold wet grasslands
All of this was found in the deposits lining the former north/south valley, still showing as a clearly visible dip in the modern ground surface.
Lying on the chalk bedrock is coarse gravel, in turn overlain by up to 70 cm of fine sand/silt. Along at least the western valley edge this sand is covered with chalk rubble, apparently representing the slumped valley-side. All three of these contexts, the gravel, the sand and the slumped chalk, contain mint condition flint artefacts.
The gravel is likely to have been laid down by a river in cool to cold climate. The tributary was here about 60 m wide and, at a height of 72.25 m OD, about 25 m above the present level of the nearby Avon/Nadder floodplain.
The overlying sand/silt contained most of the biological evidence, which helps us picture the contemporary environment.
Molluscs are dominated by a few species characteristic of open, wet ground typical of ancient British periglacial conditions, that is of freezing cold landscapes close to permanent ice sheets. The climate was cooler than today but warmer than a full glacial, with sand and silt forming in shallow, flowing water in the form of sheet-wash or in ephemeral channels.
Rare ostracods (microscopic crustaceans) support the mollusc evidence, with species typical of small or seasonal water bodies. Tufa (limestone deposits around springs and streams) further suggests shallow pools or seeps onto the sand surface.
Other remains indicate the ground was partly vegetated. Rodent teeth are etched by plant roots, and abundant calcite granules (worm casts) show there must have been a developed soil. Parts of a small horse were found (Equus ferus), an animal adapted to grazing Ice Age cold climate grasslands.
The sand/silt layer is extensive, though in places the upper surface appears to have been washed away relatively recently. Along the western edge the sand inter-leaves with the slumped chalk: in conditions of extreme cold, chalk collapsed and flowed down slope when spring sun thawed the surface, while deeper ground stayed permanently frozen. These solifluction deposits buried a land surface on the top of the sand.
Less than a square metre of this surface has been investigated, but the high incidence of refitting flint artefacts shows these remain entirely undisturbed since being discarded. The land surface’s extent is uncertain, but there is probably a north–south strip five to ten metres wide parallel to the valley edge, buried, and protected from the effects of ploughing, by chalk solifluction sediments. This is the very ground on which walked ancient hominids and extinct animals.
The strip of soliflucted chalk is today visible on the ground as a marked bank, and chalk in the plough soil. Artefacts recovered from fieldwalking, including 11 handaxes, suggest that these undisturbed deposits extend more widely to the south west and total an area of more than two hectares.
There is good agreement between the various ways of dating these deposits. The artefacts are typical of the Lower Palaeolithic, older than about 200,000 years. Water voles offer a guide, as teeth are characteristic of a form alive some time between 400,000 and 150,000 years ago.
More precise dates come from two scientific techniques, Optically Stimulated Luminescence (OSL), which uses sand grains, and Amino Acid dating of mollusc shells. Both of these suggest 250,000 years ago. This is near the end of a period of extreme cold, known as Marine Isotope Stage 8, which began 300,000 years ago.
The excavations were small, yet large numbers of artefacts were found: 44 handaxes, 18 roughed-out axes abandoned at some point in manufacture, eight hammerstones or percussors for working flint, and about 2500 waste flakes. Worked flints are densest in the chalk deposits, where their concentration is exceptionally high at around 2000 per cubic metre.
There are also thousands of smaller pieces, less than 2 cm across, recovered by sieving. No typological or technological contrasts were noted between different site phases. The surface developed on the alluvial sands is likely to have been little disturbed, but refitting flakes and microscopic chips also confirm the presence of undisturbed material in the chalk solifluction deposits and within the entire extent of the fluvial sands.
There is no sign amongst the large quantities of manufacturing debris of any structured approach to flake production such as Levalloisian (a distinctive ‘prepared core’ technique, typically associated with Neanderthals and their predecessors). Although only flint hammers have been found, flakes show that soft hammers (in materials such as bone or antler) were also used. The handaxes are mostly pointed, including well-made ‘ficron’ and ‘cordate’ forms, in some cases partly unworked with the outer cortex retained as a convenient handle. There are also some flake tools. Apart from the handaxes made on flake blanks, these appear to be waste flakes with a strong and sharp edge that have had other parts blunted to help handling and use as a knife.
One item shows what at least some of these flint tools were used for. The shaft of a cattle long bone preserves three hammerstone impact scars where it had been hit to extract the marrow. There is a small chip of flint firmly embedded in the bone’s surface, perhaps a piece of the hammer or anvil.
From large samples of sieved sand came tiny pieces of wood charcoal, possible burnt resin or pitch and burnt bone fragments. If these could be shown to be from fires constructed by the ancient hominids, it would add another level of significance to the site: use of fire this early is notoriously difficult to prove.
There are two other possible explanations. The burnt remains may not be that old, having been brought down from above by burrowing earthworms. However most of the scientists think it is more likely the burning is as old as the artefacts. This will be tested with radiocarbon dating. The other possibility is that the fires, though ancient, were natural. Only further excavation and the discovery of constructed hearths would clinch the case for artificial fires.
There is on Harnham Hill an exceptionally rare survival of the landscape 250,000-300,000 years ago, usually eroded in subsequent cold and warm climatic stages that saw major environmental change.
The lost valley allows us to picture conditions when ice sheets covered parts of the British Isles, and steppe or tundra existed where now forests might grow. Sea levels dropped across the globe due to the amount of water locked up as ice, exposing wide areas offshore as dry land, connecting Britain to the rest of Europe.
Because much of the evidence is literally microscopic, it is possible, with good deposits like these at Harnham, to learn a great deal about an ancient environment from even very small excavations. Larger exposures would be needed, however, to understand what the hominids were doing.
Why were they attracted to this narrow tributary valley? There is good flint for making tools; the abundant debris confirms that handaxes were regularly knapped there. Cutting tools were essential for a range of tasks, such as butchering. The valley was occupied by grazing animals adapted to a cold grassland environment, but whether the tools were manufactured here because the valley was also important for the herding and hunting of large mammals is unknown. We can but speculate on how the form of the valley might have influenced hunting strategies, or what was the role of fire, which may by this time have been a basic aspect of human life. The evaluation suggests it may be possible to address such questions.
Whilst there are significant differences, the Harnham discovery bears comparison with the famous site at Boxgrove, West Sussex. Hominid remains (Homo heidelbergensis) and perfectly preserved artefact scatters associated with butchered animals occur there on temporary land surfaces in a lagoon or tidal flat during a warm climatic stage 500,000 years ago. The two sites demonstrate the potential in southern Britain for localised deposits to represent a wide range of ancient environments and span a long time scale. Rare as they are, it is these sites that offer the real promise of getting close to our earliest human predecessors.
Whittaker is Director of Archaeology at RPS Historic Environment Consultants; Beasley is Senior Archaeologist at Gifford and Partners; Bates is Lecturer at the University of Wales, Lampeter; Wenban-Smith is Senior Research Fellow at Southampton University. Other scientists in the project include David Bridgland (gravels), Matthew Collins (Amino Acids), David Keen (molluscs), Simon Parfitt (animal bones), Ed Rhodes (OSL) and John Whittaker (ostracods)
CBA web:Jan/Feb 2005