Palaeolithic Investigations at the Ebbsfleet Academy (formerly the Swan Valley Community School)

Francis Wenban-Smith1, David Bridgland2 and Lis Dyson3

Department of Archaeology, University of Southampton
2Department of Geography, Durham University
3Kent County Council, Heritage Conservation

Executive Summary

This case study concerns the controlled investigation of artefact-rich gravels in the nationally important Swanscombe area.

Ebbsfleet Academy was not initially identified as of high Palaeolithic potential due to a mistaken reliance on the precision of geological mapping. The southern edge of a key deposit — the Boyn Hill Terrace — was shown 100m to the north of the site, but this was subsequently found to be erroneous. The Palaeolithic aspect of the site was only fully recognised when a handaxe was recovered from gravel deposits into which Roman features had been cut, during a standard field evaluation. Thus the site’s discovery was dependent upon the general lithic knowledge and interest of a fieldworker who was carrying out an evaluation for all archaeological periods.

The site then underwent a separate Palaeolithic-focused field evaluation that established the sequence of deposits and systematically investigated their potential by controlled sieve-sampling. After the presence of Pleistocene deposits (including a horizon of richly artefact-bearing gravels) was identified across the site, a mitigation programme was implemented. This involved (a) making stratigraphic records and carrying out lithological analyses that allowed the deposits to be related to the classic sequence at Barnfield Pit, and (b) volume-controlled sieve-sampling.

Some aspects of the project are reviewed below, with a focus on the curatorial process, and full details are given in the published report (Wenban-Smith and Bridgland 2001). An essential factor in the project was that additional resources were provided for the Palaeolithic work after the recognition of the higher-than-expected Palaeolithic importance of the site.


Where: Swanscombe, Kent
Region: South-East
Palaeolithic period(s): Lower Palaeolithic (MIS 11–8)
Type of investigation: Fieldwork
Methods: Trenches & Test-Pits; Sampling & sieving; Watching Brief; Lithological analyses
Type(s) of deposit: River terrace deposits (fluvial sand and gravel)
Features of interest: Systematic and volume-controlled gravel sieving

Project stages

  • Evaluation (phase 1)
  • Test pitting/boreholes
  • Evaluation (phase 2)
  • Excavation
  • Watching brief
  • Post-excavation assessment (and reporting)
  • Post-excavation analysis (and reporting)
  • Final Report
  • Deposit with HER and museum (and Oasis)
  • Publication (academic and/or public)

The different aspects of work involved in the Ebbsfleet Academy Palaeolithic archaeological project are identified in the overview graphic above. These involved various different archaeological organisations, and took place over a period spanning more than six years (Table 1), typical for a medium-size archaeological investigation.

Table 1: Project phases and timeline

1Field evaluation (trial trenching)Canterbury Arch. Trust (CAT)SWS 97AMarch 1997CAT 1997
Field evaluation (Palaeolithic: pre-determination)CAT (with F Wenban-Smith and DR Bridgland)SWS 97BApril 1997Wenban-Smith 1997;
Wenban-Smith &
Bridgland 1997
Mitigation (Roman)Mus. of London Arch. ServiceSSF 97April 1997 to April 1998MOLA 2010
Mitigation (Palaeolithic)University College London (with FWS)SCS 97
SCS 98
April 1997 to April 1998Wenban-Smith &
Bridgland 2001;
Wenban-Smith & Bridgland 2002
2Field evaluation (Palaeolithic: post-determination)F Wenban-SmithSCS2 00April 2000Wenban-Smith 2000
Mitigation (Palaeolithic)University of Southampton (with FWS)SCS2 01May 2001 to June 2002Wenban-Smith 2003

Initial field evaluation was carried out in February–March 1997 in advance of construction (phase 1) of a new school by Kent County Council (Figure 1). This field evaluation involved the excavation of 27 conventional archaeological trial trenches, with investigation and sampling of any features revealed below the topsoil. Specific Palaeolithic work was not on the radar until a handaxe was recovered from gravel deposits at the base of a Roman feature during this evaluation.

A Pleistocene geologist and Palaeolithic specialist were then commissioned to view the trial trenches and assess the potential. A separate Palaeolithic-focused evaluation was then carried out in April 1997. This involved 7 deep test pits dug across the site.

The Palaeolithic evaluation confirmed that Pleistocene deposits underlay the site, and that these included a gravel bed rich in flint artefacts. Therefore a mitigating programme of work for phase 1 of the construction programme was implemented between April 1997 and April 1998.

Further evaluation then took place in April 2000. This was in advance of phase 2 of the school’s construction programme, although subsequent to the granting of planning consent, which was given for both stages of the project at the outset. This was followed by further mitigating work for the phase2 construction programme.

Development Context

Work at the site was done within the framework of commercial pre-development archaeological investigation. Written Schemes of Investigation were produced for each stage of evaluation and mitigation, defining the objectives and methods of each phase of work. As a procedural formality, these were prepared and issued by Kent County Council’s Heritage Conservation section. In practice their formal issue was preceded by substantial discussion between Francis Wenban-Smith (as principal contractor and Palaeolithic specialist), Lis Dyson (as the lead KCC curator for this project), the KCC site agent and the developer (Wates Construction Ltd) as to what work was desirable and feasible within the resources available.

This consultative approach was only practicable here since KCC were acting as curator for their own development, and due to the Palaeolithic expertise of the KCC curator. A more conventional approach would have the curator determining objectives and methods for the work — formally defined in a Written Scheme of Investigation (WSI) — without consultation, or requiring (and then approving) WSIs that had been prepared by consultants or contractors on behalf of a developer.

These approaches either require the people in these roles to have Palaeolithic expertise, or (in accordance with CIfA standards) to bring in suitable external specialist expertise. The key point here is that, under all these approaches, it is still incumbent upon the curator to recognise the need for Palaeolithic work, and take appropriate steps to secure appropriate work, if the need has not already been flagged up by the developer (or their consultant).

Archaeological Context

The site is located at TQ 608739, within the south-east outskirts of Swanscombe, north-west Kent (Figure 1). Geological mapping (British Geological Survey 1977, and also the 1998 revision which was issued during the project) shows this as an area of Thanet Sand, 100m to the south of the Swanscombe 100ft/Boyn Hill terrace (formally labelled as the Orsett Heath terrace formation; Bridgland 1994).

This formation is preserved on the south side of the Lower Thames as an intermittent east–west trending band from Dartford Heath through Dartford, Stone, Greenhithe and Swanscombe to Northfleet. The deposits consist of a sequence of (predominantly fluviatile) loam, sand and gravel units laid down by the Thames in the post-Anglian interglacial period between 450,000 and 350,000 BP (late MIS 12 to early MIS 10) and can be traced along the Thames Valley (Bridgland 1994).

The formation is noted for its depth and complexity at Swanscombe, with a sequence of deposits containing archaeological, faunal and palaeoenvironmental evidence almost 20m deep, from a base level of around 22m OD to a maximum level of around 40m OD, now proved by recording at the Ebbsfleet Academy site and the adjacent primary school.

The formation in the Swanscombe area is rich in significant Palaeolithic archaeological remains, with quarrying activity at numerous locations having produced flint artefacts, mammalian fauna and other biological evidence relating to climate and environment (Wymer 1968).

The best-investigated site is Barnfield Pit (Conway et al. 1996), about 1.5km to the north-west of the Academy, where a deep sequence of mostly fluvial deposits has been found containing lithic evidence (often minimally disturbed or on undisturbed palaeo-landsurfaces) in association with a range of biological evidence including molluscan, avian, fish and mammalian remains (Table 2).

One horizon at Swanscombe, the Middle Gravel, also produced an early human skull (‘Swanscombe Man’) making it one of only two sites in England — the other being Boxgrove in West Sussex (Pitts and Roberts 1997) — with human remains from the Lower/Middle Palaeolithic.

Table 2: Stratigraphic and archaeological summary of Barnfield Pit sequence, Swanscombe. 1Note there is an alternative and widely-held view (e.g. White and Bridgland 2019) that the Upper Loam is associated with an industry of twisted ovate handaxes, based on a small number in museum collections (Roe 1968a, b); however none of these have reliable provenance as to the location and context of their finding, so this supposition currently remains unsubstantiated.

PhaseMI StageDate BPStratigraphic unitHeight, mODPalaeolithic archaeology
– ?375,000
Upper Gravel
Upper Loam
c. 33–35
c. 32–33
Uncertain, no reliably provenanced material other than technologically undiagnostic debitage1
– 400,000
Upper Middle Gravel
Lower Middle Gravel
c. 28–32

c. 27–29
Mostly pointed handaxes with thick partly-trimmed butts (often large and well-made, but also small and crude); also occasional cores, debitage and ad hoc flake-tools — Acheulean [Swanscombe Skull was found at the junction between Upper and Lower Middle Gravels]
– 425,000
Lower Loam
Lower Gravel
c. 25–27
c. 23–27
Cores, debitage, ad hoc flake-tools, and very occasional crude ‘proto-handaxes’ — Clactonian

Once the presence of a gravel deposit rich in Palaeolithic remains had been established at the site (Figure 2), the overall justifications for further targeted mitigating investigation were that this was one of the key Palaeolithic areas in the UK and it remained to be established how the deposits at the site related to the Barnfield Pit sequence; and the possibility of recovering further hominin remains. The objectives of the work were:

a. to assess the palaeoenvironmental potential of the sequence, and take appropriate mitigating steps if potentially suitable deposits were encountered;

b. through controlled sieve-sampling of the gravel to provide new information on the density and distribution of finds, thus improving understanding of site formation processes, as well recovery of mammalian fossils;

c. through controlled sieve-sampling to provide well-provenanced unbiased artefact recovery, in contrast to the majority of historic collections that are undoubtedly focused on nicer-looking specimens recovered by collectors;

d. to establish the technological strategies and types of tool present in the lithic industry from the gravel, which had been identified by evaluation as the main artefact-bearing horizon;

e. to date the sequence and correlate it with the classic sequence at Barnfield Pit;

f. to develop the wider picture of the Middle Pleistocene Swanscombe landscape and integrate the rich remains from the numerous sites in the area into their landscape context.


The first Palaeolithic evaluation involved machine-excavation of 7 deep test pits through the deposits, with recording of the stratigraphic sequence and sieve-sampling of 100L samples at regular intervals down through the sequence, and through the different deposits encountered.

Test pits were distributed evenly across the site, taking account of topography so as to investigate the full vertical and spatial range of the Pleistocene sequence. Once it had been established that the sequence was thickest in the west part of the site and that artefact-bearing gravel had been truncated at its east side (Figure 3), subsequent mitigation work was focused upon targeting sieve-sampling investigations (Figure 4) at evenly-distributed locations where construction had a major impact on the gravel.

In conjunction with this, detailed records were made of the major exposures through the Pleistocene sequence caused by excavation of the so-called ‘Grand Canyon’, which was a major ground reduction and drainage axis through the site. This was excavated with vertically-stepped sides (Figure 5) to facilitate recording of the detailed fluvial bedding structures in the upper sequence (Figure 6) which was here a sand, although attributable to the Upper Middle Gravel of the Barnfield Pit sequence. And further opportunistic records of the stratigraphic sequence were made across the site, to contribute to construction of an overall 3-dimensional model of the geological geometry.

For phase 2 of the construction programme, field evaluation and mitigation were targeted at as-yet-uninvestigated areas using the same methods as for the first phase of work. These controlled investigations were complemented by a ‘Watching Brief’ on the large number of small holes being dug for foundation pads.

This approach proved productive, in that it led to the discovery of a well-preserved tusk, 2m long, of the extinct straight-tusked elephant Palaeoloxodon antiquus (Figure 7). Thus the Ebbsfleet elephant (Wenban-Smith 2013; also see the Ebbsfleet Elephant case study) has not been the only recent elephant discovery in the Swanscombe area.

From a curatorial point of view, it was advised by the specialist and the curator that there was no benefit in undertaking the difficult work of recovering the tusk since it was not affected by the construction work, and that it was best left where it was, having been photographed and identified. However, the school governors then of their own will raised the cost of £5000 for the tusk to be lifted and conserved, and for it to be retained and displayed in local schools, where it has subsequently inspired connection with, and understanding of, the rich Pleistocene history of the Swanscombe area (Figure 8).

Results and Significance

The results of the fieldwork were successful in providing one of the best records we currently have of the changing artefact density across and through a wide exposure of (what was confirmed as) the Swanscombe Lower Middle Gravel.

The work showed that, even in a rich deposit such as this, artefacts may often not be recovered from a 100L sample, that artefacts can be distributed throughout the thickness of a gravel body, and that artefact concentration can vary widely at different depths within a gravel body, and spatially across a site. Artefact concentrations were typically in the range of 10–30 artefacts per m3, with a maximum concentration of 100 per m3, and artefacts were slightly more abundant in the top part of the gravel, although whether this is a pattern that can be expected more widely in this deposit, or in other gravel bodies, would require further systematic investigations.

At a wider scale, the site was important in improving understanding of the wider Middle Pleistocene landscape in the Swanscombe area in the Hoxnian interglacial (MIS 11), at the time of formation of the middle part (phase II: see Table 2) of the Barnfield Pit sequence (Figure 9). It established the continuation of archaeologically rich gravels from the main Thames channel further south than known.

Besides providing a better understanding of the Hoxnian palaeo-landscape in which to situate the various archaeological remains of this period that proliferate in the Swanscombe area — in particular: Barnfield Pit, Craylands Lane Pit and Dierden’s Pit (Wymer 1968) — the work was important in flagging up the wider, and previously unrecognised, extent of important Palaeolithic deposits that might be impacted by the extensive development in the Ebbsfleet area.

This thus provided the essential background for the subsequent discovery of the HS1 Ebbsfleet elephant site (Wenban-Smith 2013; see also the Ebbsfleet Elephant case study), which was initially identified as gravel beds further to the south-east, thought to possibly represent the south bank of the Lower Middle Gravel outcrop at the Swan Valley School.

The recording of Upper Loam deposits up to 40m OD during the concurrent investigation at the adjacent Sweyne County Primary School also contributes to the debate on the relationship between the Swanscombe terrace deposits and those at Dartford Heath, strengthening the interpretation that they are equivalent.

From a wider curatorial point of view, the tusk of Palaeoloxodon antiquus had no direct connection with human activity, and thus it could be regarded by some as outside the remit of pre-development investigations. However, it is just as important to investigate the faunal and palaeoenvironmental context of human activity as to investigate the direct evidence.

It is appropriate to target work at deposits with faunal (or other palaeoenvironmental) remains that are contemporary with wider Palaeolithic presence, even if they do not themselves contain artefactual (or other) evidence of human activity — and not forgetting that larger mammalian remains may have evidence of human behaviour such as cut-marks or deliberate breakage. It is also, in the rare event of prior confidence on the date of a deposit, worth targeting investigation at deposits from stages of the Pleistocene when Britain is currently thought to be unoccupied, or else prior conceptions will never be changed: as recently as the 1980s it was thought that Britain was uninhabited until after the Anglian glaciation about 425,000 BP, whereas we now know of occupation in Norfolk as long ago as 700,000 BP, and perhaps earlier (Parfitt et al. 2010; see also the Happisburgh case study).

Finally, regardless of the presumed date of deposits and the absence or otherwise of faunal and artefactual contents, Ice Age megafauna and landscape stories exert a hold over the imagination, as well as explaining the physical structure of the present landscape. There is therefore a strong argument for putting ‘the Ice Age’ to the forefront as a primary research theme in its own right, and for carrying out pre-development investigations in relation to impact on the Ice Age heritage of the landscape as an adjunct to work justified on the basis of evidence of human activity.

Key Insights

  1. From a retrospective curatorial point of view, the primary problem with this site (and with so many others) is that its high Palaeolithic potential was not highlighted at the desk-based assessment stage. It is vital that all who carry out initial Statements of Heritage Potential or Desk-based Assessments give proper consideration to the Palaeolithic potential of a site.
  2. It was also fortuitous following its discovery late in the cycle of archaeological investigation that suitable resources were made available for its targeted Palaeolithic investigation. In this instance the Kent curator had Palaeolithic expertise, and the developer was Kent County Council, so KCC’s heritage team were dealing with their colleagues as the developer — otherwise it would have been much harder to divert suitable resources to carry out the scale of Palaeolithic investigations that took place.
  3. If present in reasonable abundance, lithic artefacts from terrace gravels should not be categorised as ‘disturbed’, and therefore condemned as of low potential and not meriting evaluation and mitigation. Firstly, such lithic material is usually not especially disturbed from a grand Pleistocene/Palaeolithic perspective. And secondly, even if disturbed to a certain extent, such material still has potential to contribute to priority research questions such as the big picture of hominin presence and patterns of cultural change through the Palaeolithic.
  4. It is important to consider, and check for, Palaeolithic presence, during general field evaluations; this site was originally identified due to recognition of a handaxe in the base of a Roman feature (by a field archaeologist without any particular Palaeolithic expertise), and it was only after this that a Palaeolithic/Pleistocene programme was implemented.

    Unanticipated discoveries such as this are a recurring theme in the recent recognition of several major Palaeolithic sites (e.g. Harnham; Bates et al. 2014). This flags up the importance of general archaeological fieldworkers having lithic expertise, and being encouraged to bear in mind the possibility of Palaeolithic remains from the ‘natural’ or ‘bedrock’ into which post-Palaeolithic features are cut.
  5. Geological mapping of Pleistocene deposits cannot be relied upon as accurate, for either their absence, or their presence; in retrospect, the general proximity of the Boyn Hill Gravel and important sites such as Barnfield Pit should have been sufficient to flag the site as of high Palaeolithic potential, and thus requiring specific Palaeolithic field evaluation, despite geological mapping suggesting the absence of Pleistocene deposits.

    There were also (it later became clear) several hitherto-unrecognised references to Palaeolithic finds and Pleistocene exposures in the vicinity that were picked up when the Kent Historic Environment Record (HER) underwent a systematic Palaeolithic/ Pleistocene enhancement in 2015–2017. It is important that HERs have the best available information, since this is the primary resource for assessing the impact of a proposed development, so there would be great benefit in similar enhancements across other parts of the country where there is a Palaeolithic/Pleistocene resource (see also the Worcestershire HER case study).
  6. Interpretations of deposits in boreholes or test pits can be wrong; in this project, a deposit that later proved to be the sand-rich Upper Middle Gravel, and that overlay the artefact-rich Lower Middle Gravel, was misattributed as Thanet Sand at an early stage of work, although this initial on-site misattribution was rapidly corrected by deeper machine excavation.

    It is always useful to dig a bit deeper to (a) confirm a pre-Quaternary deposit interpretation, and (b) examine as much as possible of a Pleistocene sequence, and to reach its base if possible, to help in its interpretation.
  7. It is essential to carry out regional programmes of volume-controlled sieve sampling, and only once there is a sufficient background of these can any judgements be made as to which gravel bodies are likely, or unlikely, to contain artefact concentrations, and as to whether any particular part (such as its valley-side edge) is of higher potential.
  8. Archaeological investigation is almost always dependent upon curatorial insistence, and its scope and methods usually signed off by curatorial authorities, so it is important to build curatorial awareness of, and confidence about, the Palaeolithic.


Bates M.R., Wenban-Smith, F.F., Bello, S.M., Bridgland, D.R., Buck, L.T., Collins, M.J., Keen, D.H., Leary, J., Parfitt, S.A., Penkman, K., Rhodes, E.J., Ryssaert, C. and Whittaker, J. 2014. 'Late persistence of the Acheulian in southern Britain in an MIS 8 interstadial: evidence from Harnham, Wiltshire'. Quaternary Science Reviews 101: 159–176

British Geological Survey. 1977. 'Dartford: England and Wales Sheet 271, Solid and Drift Geology, 1:50,000 (Revised 1988)'. British Geological Survey: Keyworth, Nottingham.

Bridgland, D.R. 1994. 'Quaternary of the Thames'. Chapman & Hall: London.

CAT. 1997. 'Swanscombe Thameside Community School: Archaeological Evaluation'. Unpublished report submitted to Kent County Council.

Conway, B.W., McNabb, J. and Ashton, N. (eds.) 1996. 'Excavations at Barnfield Pit, Swanscombe, 1968–72'. British Museum Press (Occasional Paper 94): London.

MOLA. 2010. 'A Romano-British site at Swanscombe, Kent'.

Parfitt, S.A., Ashton, N.M., Lewis, S.G., Abel, R.L., Coope, G.R., Field, M.H., Gale, R., Hoare, P.G., Larkin, N.R., Lewis, M., Karloukovski, V., Maher, B., Peglar, S.M., Preece, R.C., Whittaker, J.E. and Stringer, C.B. 2010. 'Early Pleistocene human occupation at the edge of the boreal zone in northwest Europe'. Nature 466: 229–233.

Pitts, M. and Roberts, M.B. 1997. 'Fairweather Eden: Life in Britain Half a Million Years Ago as Revealed by the Excavations at Boxgrove'. Century: London.

Roe, D.A. 1968a. 'A Gazetteer of British Lower and Middle Palaeolithic sites'. Council for British Archaeology (Research Report 8): London.

Roe, D.A. 1968b. 'British Lower and Middle Palaeolithic handaxe groups'. Proceedings of the Prehistoric Society 34: 1–82.

Wenban-Smith, F.F. 1997. 'Swanscombe Community School, Palaeolithic Archaeological Evaluation'. Unpublished report submitted to Kent County Council.

Wenban-Smith, F.F. 2000. 'Swanscombe Community School, Phase 2: Field Evaluation Report'. Unpublished report submitted to Kent County Council.

Wenban-Smith, F.F. 2002. 'Swanscombe Surgery, Palaeolithic Archaeology: Assessment Report and Updated Project Design'. Unpublished report submitted to Kent County Council.

Wenban-Smith, F.F. 2003. 'Swan Valley Community School, Swanscombe (Phase 2): Palaeolithic Archaeology, Assessment Report and Updated Project Design'. Unpublished report (ref BLUE: SCS2, dated 9th August 2003) submitted to Kent County Council.

Wenban-Smith, F.F. 2013. 'The Ebbsfleet Elephant: Excavations at Southfleet Road, Swanscombe in Advance of High Speed 1, 2003-4'. Oxford Archaeology Monograph No. 20. Oxford Archaeology: Oxford.

Wenban-Smith, F.F. and Bridgland, D.R. 1997. 'Newly discovered Pleistocene deposits at Swanscombe: an interim report'. Lithics 17–18: 3–8.

Wenban-Smith, F.F. and Bridgland, D.R. 2001. 'Palaeolithic archaeology at the Swan Valley Community School, Swanscombe, Kent'. Proceedings of the Prehistoric Society 67: 219–259.

Wenban-Smith, F.F. and Bridgland, D.R. 2002. 'Swan Valley Community School, Swanscombe (Phase I): Palaeolithic Archaeology, Final Report'. Unpublished report submitted to Kent County Council.

White M.J. and Bridgland, D.R. 2019. 'Palaeolithic artefacts from the English rivers: the human occupation of Britain and characteristic artefact assemblage types'. In: D.R. Bridgland, R.M. Briant, P. Allen, E.J. Brown and T.S. White (eds.) The Quaternary Fluvial Archives of the Major English Rivers: Field Guide: 19–24. Quaternary Research Association: London.

Wymer, J.J. 1968. 'Lower Palaeolithic Archaeology in Britain as represented by the Thames Valley'. John Baker: London.