The HS1 Ebbsfleet Elephant, Southfleet Road, Swanscombe

Francis Wenban-Smith
Archaeology, University of Southampton

Executive Summary

Southfleet Road (Ebbsfleet elephant): managing an unexpected, but internationally significant, discovery after the site’s Palaeolithic potential was not recognised at the DBA stage.

The HS1 elephant site is located at TQ 6116 7328, to the south-east of Swanscombe, north-west Kent (Figure 1). The site was contained in a deep and complex multi-phase Pleistocene sequence (Table 1; Figure 2), preserved under the previous course of Southfleet Road. The sequence was revealed when the road was re-designed to allow a link between Ebbsfleet International and the A2.

The elephant remains comprise the remnants of the skeleton of a large adult male, and were found in association with an undisturbed scatter of refitting flint artefacts. It is uncertain whether the elephant was killed or found dead, but the associated flint scatter can confidently be interpreted as on-the-spot production of sharp-edged flint flakes and simple notched flake-tools to facilitate butchery for its meat.

The elephant skeleton can be dated to the early temperate stage of the Hoxnian interglacial (MIS 11c), around 425–400,000 BP, contemporary with the Lower Loam deposits at nearby Barnfield Pit. The site has made a significant contribution to current understanding of Lower Palaeolithic foraging strategies.

Although the elephant is its headline element, there are other important aspects of the site. A scatter of about 1800 flint artefacts was found at the same horizon a short distance to the south. The consistent character of this substantial assemblage — simple flake tools made on flakes from simple knapping strategies — has made a major contribution to confirming the early Hoxnian presence in the UK of a non-handaxe-making population using this Clactonian lithic industrial tradition.

The elephant horizon was buried by clayey and gravelly slopewash deposits (phase 7 of the site sequence; Table 1). These in turn were cut through and capped by a fluvial gravel, representing a major change in the local landscape, probably accompanied by changing climate. This latter gravel bed (phase 8; Table 1) also produced lithic artefacts. In contrast to the Clactonian industry of the underlying phase 6 (Table 1), these were dominated by well-made handaxes of various shapes, and thus attributable to an Acheulean industry. Many were in very fresh condition despite being found in a fluvial gravel context, demonstrating the potential of fluvial gravel beds to contain little-disturbed activity sites.

The review below focuses upon curatorial and methodological aspects of the site; more detailed results of the work have been published (Wenban-Smith 2013), and this monograph is also available for free download on-line, as is appropriate for any publically-funded pre-development investigation.

Table 1: Site sequence and key finds

Stratigraphic phase (formation process)Archaeological remainsPeriod, date
11 - Modern made groundPost-medieval gunflint manufacturing and derived Palaeolithic artefactsPost-medieval through to 20th century
10 - Post-Palaeolithic featuresVarious pottery and flint artefactsHolocene
9 - Brickearth bank (slopewash)Handaxes, débitageUncertain, MIS 10?
8 - Gravel (fluvial)Handaxes of various forms, large scrapers and other flake tools, occasional débitageMiddle Pleistocene, late MIS 11
7 - Mixed clay/gravel (slopewash)Cores, flakes and simple flake tools similar to Clactonian industry of phase 6Middle Pleistocene, late MIS 11
6 - Grey clay, with peaty beds (quiet water, periodically drying, with slopewash)
  • Elephant carcass with refitting lithic scatter.
  • Abundant cores, flakes, and flake tools (Clactonian industry).
  • Mammalian and molluscan fauna reflecting rich interglacial woodland
Middle Pleistocene, early MIS 11
5 - Sand (fluvial)
  • Cores, flakes, a flake tool.
  • Some faunal remains reflecting temperate open woodland
Middle Pleistocene, early MIS 11
4 - Sandy/gravelly clay (quiet water, periodically drying)
  • Ostracods reflecting similar temperate conditions as phase 6
Middle Pleistocene, early MIS 11
3 - Silt/sand with flint and chalk pebbles (slopewash, into lake margin)
  • Two flakes.
  • Ostracods reflecting similar temperate conditions as phase 6
Middle Pleistocene, early MIS 11
2 - Bedded sand/clay (quiet water?)-Earlier Pleistocene?
1 - Isolated bedrock block (upheaval/ collapse?)-Earlier Pleistocene?


Where: Southfleet Road, Ebbsfleet
Region: South-East
Palaeolithic period(s): Lower Palaeolithic (MIS 11)
Type of investigation: Excavation
Methods: Section-cleaning; Palaeoenvironmental sampling; Ground-reflecting laser survey; Machine-trenching; Hand-dug test pits; Watching brief; 3D artefact recording; Sediment block lifting of faunal remains
Type(s) of deposit: Lacustrine; Fluvial gravel; Colluvial
Features of interest: Deep and complex sequence; Elephant remains; Undisturbed lithic scatter; Multiple horizons with lithic evidence

Project stages

  • Desk based assessment
  • Test pitting/boreholes
  • Evaluation
  • 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)

Development Context

All pre-HS1 archaeological work was carried out under the heritage provisions of the Channel Tunnel Rail Link Act 1996, rather than under the normal planning controls for development. Under the Act, curatorial decisions were made by a body of Statutory Consultees, formed by Local Authority and English Heritage curators.

Archaeological Context

As spelt out below (Table 2), the site was not included in the targeted pre-HS1 archaeological programme. A late reactive case was made for a brief initial investigation on the basis that the gravel deposits revealed by ground reduction either side of the old Southfleet Road might represent the southern bank of the Lower Middle Gravel outcrop at Ebbsfleet Academy (Wenban-Smith and Bridgland 2001).

The importance of the site, and the need for further investigation, was then justified on the basis of (a) pollen evidence from presumed Hoxnian fluvial deposits relatable to the presumed Lower Middle Gravel, and (b) the presence of a palaeo-landsurface that had produced two flint flakes in mint condition.

It was only after more thorough investigation had begun that the most important aspects of the site were revealed, namely: the elephant remains, the southern lithic concentration, the tufaceous channel rich in palaeoenvironmental remains, and the Acheulean industry in the gravel capping the sequence.

Table 2: Project stages of the Southfleet Road investigation

PhaseBrief detailsDateComments
Initial recognition
  • Pleistocene gravels noted in unexcavated deposits from a hillside vantage point on the opposite east side of the Ebbsfleet, during fieldwork in advance of housing there
September 2003, and mid-November 2003The base of the gravels dipped to the north, and the initial suspicion was that the deposits might represent the south bank of the Lower Middle Gravel, as represented at Ebbsfleet Academy
Preliminary site visit
  • Closer inspection of the section exposure, which showed well-bedded fluvial gravels dipping to the north, as well as other underlying clay and sand beds that were partly obscured and inaccessible (Figure 3)
21 November 2003It was agreed after this visit that a basic record of the exposed sequence should be made, accompanied by some palaeoenvironmental sampling
Initial field recording
  • A basic record was made of the stratigraphic sequence.
  • Palaeo-landsurface identified in sequence.
  • Samples taken to assess for palaeoenvironmental remains.
  • A few artefacts found from (a) the gravels, and (b) a clay horizon equivalent to the palaeo-landsurface
December 2003The results of the environmental assessment suggested moderately abundant and well-preserved pollen in some clay-laminated sands towards the base of the sequence; this was regarded as important for the potential of directly linking a Hoxnian sequence in the Swanscombe area with Hoxnian pollen sequences in East Anglia
Follow-up fieldwork
  • Section clearance by machine and hand-cleaning, with more-detailed drawn record and sequence description.
  • Targeted pollen sampling through the clay-laminated sand.
  • Test pit evaluation of the palaeo-landsurface.
  • Controlled sieve-sampling through the gravel
February–March 2004The elephant (or rather its fragmentary bones, tusks and a molar) was discovered on 15th March during machine-clearance of the west face of the north-south spine of deposits that formed the site, exactly at the dark clay horizon that had initially been identified as a palaeo-landsurface (Figure 4)
Main excavationThe programme of fieldwork continued as planned, in conjunction with careful excavation of the elephant remains, which were found to be associated with lithic artefacts. Proper section clearing exposed (as well as the elephant) a very restricted ‘tufaceous channel’ deposit that was rich in palaeoenvironmental remains. And the test pit evaluation revealed a substantial lithic concentration in the south part of the site, which then required open-area excavation.April–August 2004The discovery of the elephant transformed the dynamic of the project, confirming the importance of the site, and a reasonable and realistic amount of time was then allocated for its thorough excavation.
Watching BriefReduction of the gravel still capping the north part of the site. Extensive network of drainage under/beside the new route of Southfleet RoadSeptember–November 2004Numerous handaxes and other flint artefacts were recovered during the gravel reduction; this was a useful complement to the volume-controlled sieving. The drainage Watching Brief was challenging, due to inaccessibility into the trenches and smearing of their sides, making it hard to record the sequence accurately

Under the terms of the CTRL 1996 Act, it required intervention by the Secretary of State, advised by the Statutory Consultees, to allow sufficient time for the excavation to be completed. The basis for gaining support for this decision was the demonstrable national importance of the site, when set against English Heritage’s (1998) criteria for the importance of Palaeolithic sites (Table 3).

There is a widely held misconception that the only nationally important Palaeolithic sites are those where undisturbed remains are present. However, it is often overlooked that, as spelt out in the 1998 guidance, any of these criteria are sufficient for a site to be regarded as of national importance and meriting pre-development investigations.

Table 3: Assessment of site importance with regard to English Heritage (1998) guidance, as argued in May 2004

Human boneNo-
Primary undisturbed contextYesLandsurface in strip from lake-bank through to lake-margin sediments
Period/area rareYesDeposits with Clactonian material are rare in Britain. If the site dates, as thought, to before the Swanscombe Lower Middle Gravel, then archaeologically rich deposits of this period are also very rare
Organic artefactsNo-
Well-preserved associated bio-evidenceYesExtremely rich tufaceous horizon immediately under the main landsurface that contains wide range of well-preserved small vertebrate and molluscan evidence
Evidence of lifestyleYesIsolated skeletons and partial skeletons of megafauna with associated refitting knapping scatters
Stratigraphic relationships between different archaeological horizonsYesThe upper horizons at the site contain abundant handaxes (Acheulean). The lower horizons contain abundant Clactonian material. This superposition has only been found at two other sites, and is very significant in confirming the reality of this patterning and the distinction of these two cultural traditions, which have been disputed by some.
Artistic evidenceNo-
Evidence of hearths or structuresNo-
Resource exploitationYesThe situation of the site at the margin of a clay-lined lake may be central to what is going on. The lake itself may be considered a resource, for water, for attracting game, and possibly as a tool to aid their capture. Flint raw material was also clearly locally available, although little is currently known about the local palaeo-landscape.
Artefacts are particularly abundantYesConcentrations of mint condition artefacts are very high at the southern end of the site, just above the bank of the lake, which would presumably have made a more amenable location for occupation and knapping than the wet lake-margin itself

Methodology & Research Questions

As spelt out above (Table 2), the planning and methods for the site’s investigation evolved reactively as work progressed and discoveries were made. Thus, an initial plan for a 2-day investigation morphed into a 10-month field project. This is not therefore an ideal example of careful forward planning and smooth progression of a project along its anticipated path.

However, it is a good example of how a heritage-committed team on all sides of a project (HS1 including its engineers, archaeologists and advisors; the Statutory Consultees; the contractor [Oxford Archaeology] and the Palaeolithic specialist) worked together in a flexible manner to investigate and then deal properly with a major Palaeolithic discovery in a location earmarked for imminent ground removal. The methods applied, and the rationale behind them, are also instructive, a few of which are reviewed below. The methods used are summarised in Box A.

The initial phase of work involved section-cleaning and palaeoenvironmental sampling of deposits through the sequence. This allowed identification of a palaeo-landsurface and a rapid turn-around of the environmental samples allowed an initial identification of deposits with palaeoenvironmental potential.

From a methodological point of view the height of the section meant that in many places its main stratigraphic boundaries could only be safely recorded from a distance with a ground-reflecting laser beam. Safe direct recording was eventually achieved by machining a sloping face with a horizontal step halfway up it. If this is done, one then needs to record the sloping face by measuring parallel to it, and to survey various reference points that are marked on the drawing, to allow subsequent 3D and true-vertical modelling. In general stepped faces with vertical drops are better for recording stratigraphic sequences, if practicable.

As work progressed, and new deposits were discovered, further rapid-turnaround palaeoenvironmental sampling was carried out for small vertebrates, pollen, molluscs and ostracods. This proved invaluable in highlighting deposits with palaeoenvironmental potential while there was still time to target them for more intensive sampling.

There was some dispute early in the project about the correlation of the sequence in the west and east faces of the north–south spine of deposits that formed the site. It was eventually resolved that this needed to be addressed by properly recording both faces, and also digging four transverse stepped trenches, A–D (Figure 5). These steps revealed the full complexity of the sequence, with the asymmetric syncline of the ‘Skateboard ramp’ (Figure 6), and established the palaeo-landsurface as only surviving in the southern part of the site.

The palaeo-landsurface was initially uncovered by machine, revealing its synclinal dip between Trenches C and D, where it was also deeply buried by clays/gravels. It was evaluated by nine hand-dug test pits in the vicinity of Trench D (Figure 7, Trenches I–IX). This successfully identified the area of lithic concentration south of Trench D, all of which was then excavated by hand.

All artefacts had their precise 3D positions recorded with an on-site Total Station. This was an essential record for subsequent interpretation of the site, and for its 3D digital modelling, which allowed the artefacts to be viewed in relation to the stratigraphy.

Unfortunately, the lithic concentration was contained in clay, which hardened to the texture of a sun-baked brick while the clay was exposed and evaluated, before the go-ahead was given for full excavation. This meant that much of the recovery had to be done with mattock rather than trowel, which (a) probably led to incomplete recovery of smaller artefacts, and (b) led to some damage when artefacts were hit by a mattock. Although regrettable, the latter is often inevitable, but it is important that excavation damage is noted in the site records in a no-blame culture, to avoid subsequent difficulty in distinguishing recent damage from ancient use-wear, secondary working or other forms of ancient damage, when studying the lithics.

Difficulty was also encountered with dealing with fragile faunal remains encased within the clay. Part of the problem was that faunal remains were rare, within a large body of clay. A judgement had to be made, therefore, on how this could be investigated. The only practicable way was to remove the clay as carefully as possible with a mechanical excavator, and then to switch to hand excavation when individual large pieces or concentrations of faunal remains were encountered.

The other part of the problem was that faunal remains were fragile within the clay. Thus, a faunal specialist was on site continually throughout the project, and many remains were lifted in clay blocks cased in plaster-of-Paris for subsequent off-site extraction.

Box A: Methods used at Southfleet Road (see text for details)

  • Section cleaning
  • Palaeoenvironmental sampling: spot-samples for pollen, molluscs and ostracods, larger samples for small vertebrates, ranging from 30L for initial evaluation to hundreds of litres for proper recovery
  • Ground-reflecting laser survey
  • Machine-trenching (with sample-sieving) & hand-dug test pits
  • Watching brief
  • 3D artefact recording (total station)
  • Faunal remains lifted in sediment blocks & extracted off-site.

Results and Significance

The main result of the project was the rare and iconic discovery of the undisturbed remains of a 400,000 year-old elephant butchery event (Figure 8), contextualised at the edge of the marshy base of the Ebbsfleet Valley within the wider luxuriant forested landscape of the peak temperate part of the Hoxnian interglacial.

At a more academic level, the site was instrumental in confirming the distinct status of the Clactonian as the industry of the initial Hoxnian inhabitants when Great Britain was recolonised after the local extinction of the Anglian (MIS 12) ice age.

Despite the contrasts between the Clactonian industry of phase 6 and the Acheulean industry of phase 8 (Table 1), there are key shared elements that support in situ cultural development as opposed to a second wave of migration. The site also indicates that elephant (and other megafaunal) exploitation may have been an important aspect of a successful hominin adaptation facilitating northward range-expansion in Europe.

At a community level, the site has stimulated local awareness and appreciation of the rich Palaeolithic heritage of the Swanscombe area, complementing and reinforcing the Swanscombe Skull Site National Nature Reserve.

Click to view the full image and then zoom in for more detail

Key Insights

  • As with many other important Palaeolithic discoveries, the site’s Palaeolithic potential was not recognised at the desk-based assessment stage, and therefore it was overlooked until its discovery. The focus of archaeological investigation in the Ebbsfleet Valley was on the footprint of HS1 and Ebbsfleet International station, and the impact of the road link to the A2 was not given attention.

    My initial (1992) assessment of Pleistocene deposits and Palaeolithic potential in and around the Ebbsfleet Valley correctly identified the elephant site as undisturbed ground. However, the spot was identified (incorrectly as it was later shown) in geological mapping (British Geological Survey 1977) as Head deposits overlying Thanet Sand, and on this basis I did not identify it as of Palaeolithic potential.

    This highlights the potential for geological mapping to be misleading, and for potentially important Pleistocene deposits not to appear. It is vital that all who carry out initial Statements of Heritage Potential or Desk-based Assessments give informed consideration to the Palaeolithic potential of a site, taking account of the possibility of misleading geological mapping; and also that all impacts associated with major projects (e.g. services, transport infrastructure) are taken into account, as well as the headline development.
  • After initial recognition of the site as of potential interest, credit is due to the HS1 archaeological team for their flexibility in finding the resources and adjusting the construction programme to allow for its investigation. However, this was not an entirely smooth process at the early stages of the project, and was driven forward by the curatorial Steering Group and the Statutory Consultees.

    The lessons here are (a) that strong and Palaeolithic-aware curators are essential, and (b) that important Palaeolithic discoveries are by their nature rare and hard to predict, so contingency reserves should be maintained for large projects to allow for unforeseen discoveries.
  • The elephant horizon was overlain (and partly truncated) by a fluvial gravel that produced a rich assemblage of more than 180 flint artefacts, including more than 30 handaxes. Many of these were in fresh or mint condition, and therefore represent activity at (or very near) the site. However, the initial exposure of the gravel appeared barren of artefacts.

    Prolonged searching of the exposed gravel produced two flakes from talus that were suspected (but not proven) to originate from it. Subsequent targeted evaluation of the gravel produced lithic finds at low densities of between 0 and 10 per m3 from a series of 100L samples through the main gravel body, with only its lowest bed producing richer recovery (60/m3, including one handaxe). Larger samples of 250L through the main gravel body were similarly unpromising, with an average of 6 artefacts per m3, from 17 samples. It was only as a result of full section cleaning and a careful watching brief during machine reduction that a reasonably large artefact sample was recovered from the gravel.

    This highlights (a) the potential of a gravel body to contain less disturbed assemblages (also see the Ebbsfleet Academy case study for the importance of less undisturbed finds from gravel), (b) the difficulty of carrying out appropriate evaluation to recognise the potential of a gravel body, and (c) the importance of large-scale sieving to recover a reasonable volume-controlled artefact sample, complemented by other methods to recover enough artefacts to gain an overall impression of the technology/typology of an industry.
  • The wide exposure (about 100m long) through the Pleistocene sequence at the site (Figure 2) demonstrated the potential for particular beds to rapidly thicken, thin and disappear; and for some critical beds (such as the tufaceous channel rich in palaeoenvironmental remains) to only manifest as very restricted features. The elephant horizon was a clay bed >2m thick in the centre of the site, but this thinned to a vestigial trace about 1cm thick to the north, and was then truncated altogether by gravel deposits.

    This highlights the difficulty of accurately modelling the nature, distribution and potential of Pleistocene sequences from spatially separated test pits or boreholes. Some deposits may have extensive horizontal continuity, but many may not.


This is not the place for full acknowledgement of the very numerous people whose hard work and support made the project possible, for which see the published monograph.

The work was funded by the HS1 programme, and carried out in conjunction with Oxford Archaeology as the principal archaeological contractor, with consistent support from all at HS1 and the curatorial Steering Group.


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

English Heritage. 1998. 'Identifying and Protecting Palaeolithic Remains: Archaeological Guidance for Planning Authorities and Developers'. English Heritage: London.

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.

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.