ORIGINAL FINDS

Between December 2007 and March 2008, 88 Palaeolithic lithic artefacts (including 33 handaxes) and over 100 fragments of vertebrate fauna were recovered during aggregate dredging from licence Area 240, located 11 km from the nearest point on the East Anglian coast at Great Yarmouth (Fig. 1). Material was recovered from oversize aggregate stockpiles at SBV Flushing Wharf in Vlissingen, Netherlands. The discoveries were reported to English Heritage (now Historic England) through the Marine Aggregate Industry Protocol for Reporting Finds of Archaeological Interest (MAI Protocol).

Fig. 1. Location plan

A 3.5 × 1.0 km area on the eastern side of Area 240 was identified as the source of this material. To prevent further impact on the archaeology, the operator holding the marine licence (regulatory consent) to extract sand and gravel (Hanson Aggregates Marine Ltd) voluntarily implemented Archaeological Exclusion Zones (AEZ) covering this area (Fig. 2). The lithic artefacts were subsequently analysed by De Loecker (De Loecker Reference De Loecker2010; Tizzard et al. Reference Tizzard, Bicket and De Loecker2015) and vertebrate fauna by Glimmerveen (Tizzard et al. Reference Tizzard, Bicket and De Loecker2015).

Fig. 2. Area 240 plan

Between 2008 and 2013, Wessex Archaeology undertook the Seabed Prehistory Project to understand the palaeogeography and archaeological context of the Palaeolithic material (Wessex Archaeology 2010a; 2010b; 2011). The results were published by Tizzard et al. (Reference Tizzard, Bicket and De Loecker2015).

Investigations included assessment of geophysical and geotechnical data, palaeo-environmental assessment and analysis, and scientific dating. Techniques for targeted artefact recovery were also trialled which included sampling using clamshell grabs, still photographic survey, and beam trawling. This work was followed in 2011 by a programme of archaeological wharf monitoring of dredged aggregate (Wessex Archaeology 2011). The Project demonstrated that Quaternary deposits of the Palaeo-Yare were present across Area 240. Subsequently, aggregate dredged from licence areas covering the Palaeo-Yare catchment has been subject to archaeological sampling through wharf monitoring (Fjordr 2015; Wessex Archaeology 2015; 2021), which has recovered the further lithic artefacts assessed in this paper.

NEW DISCOVERIES: ANALYSIS OF NEW PALAEOLITHIC ARTEFACTS FROM AREA 240

The new finds comprise 88 flint artefacts and a further 14 possible flint artefacts (Table 1). Additional material is present within the collection which, while exhibiting evidence for conchoidal fracture of flint clasts, are not considered to be humanly modified.

Table 1: Palaeolithic artefacts recovered from wharf monitoring 2011–2022

* This artefact may be from a cross-contaminate sample from several aggregate areas.

Artefacts are now known from the Palaeo-Yare catchment in Licence Areas 212 (within 494), 240 and 511, with possible artefacts noted from Areas 228, 319 (within 511), 401/2 and 512 (Fig. 1). Nevertheless, most new finds (94.3%) are, like the original 2007–2008 discoveries, from Area 240.

Analysis of the new lithic collection has comprised taphonomic, techno-typological, and spatial study to consider whether a single, uniform, or multiple/diverse assemblage(s) is present, to assess the likely original lithostratigraphic context, to consider how the material relates to 2007–2008 discoveries, and to place all the archaeology within the context of the wider British Palaeolithic record. The methodology applied is provided in Appendix S1.

Techno-typology

The techno-typological characteristics of most chronologically diagnostic artefacts are indicative of an earlier (Lower or Middle) Palaeolithic date. An exception is a bipolar blade core (Fig. 3) which is consistent with a late Upper Palaeolithic (Creswellian), Terminal Upper Palaeolithic (Long Blade) or early Mesolithic date. This core has a different taphonomic history to the earlier Palaeolithic pieces: it is moderately abraded, with much battering on the aretes. The raw material used is also distinct, being a heavily rolled and battered cobble likely obtained from a beach. This blade core is notable as it is the first clear evidence for Upper Palaeolithic/Mesolithic activity within the submerged Palaeo-Yare catchment. The bulk of the chronologically diagnostic artefacts are handaxes (alongside a small number of flakes potentially reflective of handaxe working) and Levallois flakes (Table 1).

Fig. 3. Bipolar blade core, Area 240

Handaxes: The assemblage contains 17 whole, finished handaxes and a butt fragment, along with a roughout and a possible roughout abandoned during manufacture. All except four of the pieces were available for study; photographs and detailed descriptions were available for these four pieces.

Comparison has been made between the new finished examples and 20 complete handaxes recovered in 2007–2008 and analysed by De Loecker (Reference De Loecker2010). Metrical data have been used to compare handaxe planform utilising methodological criteria developed by Bordes (Reference Bordes1961) and Roe (Reference Roe1964; Reference Roe1969).

Following the criteria of Roe (Reference Roe1964; Reference Roe1969), the handaxes in both collections are refined, well-made examples with a weak tendency to having pointed tips. For the new material, the data also indicate high levels of cross-sectional uniformity (this information is not available for the earlier collection). The planforms of the handaxes coalesce close to Roe’s metrical division between ovate and pointed groups. The specific ranges in handaxe planforms are illustrated in Figure 4. Proportionally, more handaxes tend towards the pointed end of the spectrum of variation within the new material.

Fig. 4. Area 240 handaxe planforms (following Roe Reference Roe1964; Reference Roe1968)

Following Bordes (Reference Bordes1961) methodological criteria, the handaxes from the two collections are highly consistent, being flat (as opposed to thick) and cordiform in planform (Fig. 5). This uniformity is particularly marked amongst the new material.

Fig. 5. Area 240 handaxe planforms (following Bordes Reference Bordes1961)

Metrical data demonstrate that both new and original finds from Area 240 are highly homogeneous in form, being refined, thin handaxes that have a consistent cordiform planform. Although the data do not demonstrate that the handaxes are a single assemblage from a single lithostratigraphic context, a highly consistent approach to handaxe manufacture is indicated.

Detailed study of the new material provides additional insights into their technology. These handaxes have been produced through extensive flaking using a soft hammer (88.9%), leaving little cortex (all but one example = <10%, with 46.7% retaining no cortex) and resulting in extensive cutting edges around most edges (77.8% all round or most edges). It is generally not possible to determine the original blank due to the intensity of flaking but, where this is apparent (n=4), this was a large flake. One piece on a flake is a large roughout which, along with a further possible example, indicates that handaxe blank selection and manufacture was being carried out in the immediate landscape within Area 240. That handaxes were at least being modified locally is further demonstrated by the presence of two thinning flakes (one soft and one likely soft hammer) and three possible examples.

A particularly striking feature of the handaxes is a consistent approach to manufacture, with initial phases of invasive removals designed to thin the blank, followed by phases of less invasive scars around the margins, which shaped the artefact. Additionally, three pieces exhibit secondary reworking and modification that post-dates initial phases of thinning and shaping. In at least one case this reflects secondary reworking of a broken handaxe.

These technological characteristics are similar to those reported for the original Area 240 handaxe finds (De Loecker Reference De Loecker2010, 10) which were extensively worked, removing most cortex, thinned, and shaped using soft hammers, and frequently preserved evidence of a phase of shaping after thinning (referred to as ‘continuous retouch’). The only notable difference from the new material is that the original finds include examples with tranchet removals to resharpen the tip (De Loecker Reference De Loecker2010, 10); no examples with tranchet blows were recorded amongst the new finds.

Levallois artefacts: The new collection includes eight definite and four probable Levallois flakes, along with a further possible example; all but one definite flake was available for direct analysis. The Levallois flakes are all large (average maximum dimension = 87.0 mm) and all but one is complete. Only the possible example retains any cortex. The flakes generally reflect centripetal preparation (66.7%, 54.3%), a feature often displayed by large Levallois flakes detached early in reduction from large cores (Scott Reference Scott2011). Many can only be classed as single removals but a significant number retain dorsal scars that reflect a series of preferential removals made during a phase of core exploitation. Generally, exploitation was through unipolar (n = 4), and less frequently centripetal (n = 2), recurrent methods. No definite Levallois cores were identified in the new collection but a core fragment may relate to Levallois reduction, with indications of a flaking surface that has been subject to centripetal preparation and lineal exploitation.

The 2007–2008 collection contained ten Levallois flakes, termed Levallois sensu stricto, whilst a further ten were termed ‘extended’ Levallois, possibly reflective of Levallois core working (De Loecker Reference De Loecker2010). It is not possible to assess whether they reflect similar or different approaches to Levallois reduction as those in the new collection. The 2007–2008 sample included two Levallois and one ‘Levallois extended’ cores (De Loecker Reference De Loecker2010); the latter would broadly conform to a simple prepared core according to our methodological criteria (Appendix S1). The Levallois methods of preparation and exploitation applied to these cores are unknown.

Other artefacts: The only whole core in the new collection, aside from the late Upper Palaeolithic/early Mesolithic blade core, is an example abandoned early in reduction due to the exposure of a large flaw in the flint. The remainder of the new artefacts are flakes. These are generally large (average maximum dimension = 85.2 mm) and removed using a hard hammer (86.4%). Two flakes are detached from discoidal cores. The original 2007–2008 collection included three ‘disc’ and one ‘disc/discoidal’ core (De Loecker Reference De Loecker2010) which may also reflect discoidal core working. Three flakes in the new collection are retouched.

Taphonomy

To assess the taphonomic history of the new artefacts from Area 240, physical characteristics comprising levels of surface abrasion, edge damage, patination, staining, battering, and surface scratching have been assessed (Fig. 6). Edge damage has been excluded from the analysis as most artefacts exhibit fresh edge damage likely relating to the dredging recovery process. This masks any earlier damage that may have been present.

The collection exhibits low levels of surface abrasion, battering, and surface scratching, demonstrating that, whilst not necessarily in situ, the artefacts have not undergone extensive post-depositional reworking. Although the sample size is limited, the analysis has shown that within this broad pattern there are notable differences between handaxes and Levallois flakes.

Although there is a continuum of condition states, the Levallois material contains proportionally more slightly abraded examples than the handaxes. Additionally, the Levallois artefacts are generally more patinated and more stained (Fig. 6). This suggests that the handaxes and Levallois flakes in the new collection have had different taphonomic histories prior to recovery. The taphonomic data presented for the 2007–2008 discoveries (De Loecker Reference De Loecker2010) does not allow for similar analysis. Consequently it is unclear whether similar taphonomic differences between techno-typological artefact classes apply to the original material.

Fig. 6. Area 240 artefact condition

Spatial distribution

It is possible to consider the broad spatial distribution of the new lithic artefacts from within Area 240. For 72 pieces, data is available to establish the individual aggregate loads from specific dredge lanes from which pieces were obtained. This can be tied to GPS coordinates for the dredge plot recording the specific route taken by the dredge vessel. Centre points of dredge lanes have been used to plot to the spatial distribution of the artefacts recovered (Fig. 7). Although there are limits to the accuracy of this data, it nevertheless provides the first indication of specific recovery areas for individual artefacts within Area 240.

Fig. 7. Spatial distribution of Palaeolithic artefacts from Area 240

The data demonstrate that the new artefacts mostly originate from two sub-zones in the south of the licence area. One is immediately to the west of the Archaeological Exclusion Zones covering the area of the 2007–2008 discoveries, the other sub-zone is further to the south-west (Fig. 7). This demonstrates that multiple locations within Area 240 have produced Palaeolithic artefacts.

The distribution of techno-typological classes of Palaeolithic artefacts are also revealing. Handaxes, handaxe roughouts, and flakes from handaxe manufacture have only been recovered from the northern sub-zone, in the area adjacent to where similar material was recovered in 2007–2008 (Fig. 7). In contrast, Levallois flakes are more widespread. Although these were recovered from within the northern sub-zone, they also occur within the southern sub-zone. The late Upper Palaeolithic/early Mesolithic blade core also originates from this southern sub-zone.

THE MIDDLE PALAEOLITHIC OF THE PALAEO-YARE

Lithostratigraphic and palaeo-landscape context

Perhaps the greatest challenge in analysing the Palaeolithic archaeology from Area 240 is attributing the material to original contexts. No artefact can be directly attributed to a single, specific sedimentological unit reflecting a discrete period of deposition, all material being recovered ex situ. However, vibrocore logs and geophysical data have enabled a lithostratigraphic and palaeo-landscape framework for the Palaeo-Yare to be developed, both specifically for Area 240 (Tizzard et al. Reference Tizzard, Benjamin and De Loecker2014; Reference Tizzard, Bicket and De Loecker2015) and for the wider submerged Palaeo-Yare catchment (Wessex Archaeology 2013; Table 2). This framework can be related to the much less well understood lithostratigraphy of on-shore sediments of the Palaeo-Yare catchment, in which deposits are divided between the Holocene Bredon Formation and Pleistocene Yare Valley Formation (Arthurton et al. Reference Arthurton, Booth, Morigi, Abbott and Wood1994). The latter subsumes poorly defined separate Pleistocene fluvial sediment bodies.

Table 2: Quaternary lithostratigraphic framework

The lithostratigraphic and palaeo-landscape framework for the Palaeo-Yare summarised in Table 2 provides a base dataset for considering the lithostratigraphic context of the Middle Palaeolithic archaeology from Area 240. Chronology for this framework is provided by luminescence dating of sediments and through AMS radiocarbon dating of organic material. These dates have been tied to the Marine Isotope Stage (MIS) record. All available direct dates for the Palaeo-Yare deposits are summarised in Table 3.

Table 3: Published radiometric dates for sediments and vertebrate fauna from area 240

References: 1. Wessex Archaeology 2011; 2. Wessex Archaeology 2008; 3. Limpenny et al. Reference Limpenny, Barrio Froján, Cotterill, Foster-Smith, Pearce, Tizzard, Limpenny, Long, Walmsley, Kirby, Baker, Meadows, Rees, Hill, Wilson, Leivers, Churchley, Russell, Birchenough, Green and Law2011; 4. Tizzard et al. Reference Tizzard, Bicket and De Loecker2015; 5. Arthurton et al. Reference Arthurton, Booth, Morigi, Abbott and Wood1994; 6. Hazell pers. comm. (in Tizzard et al. Reference Tizzard, Bicket and De Loecker2015).

Tizzard et al. (Reference Tizzard, Bicket and De Loecker2015) assessed the possible lithostratigraphic contexts of the 2007–2008 Palaeolithic artefacts. This concluded that Units 3b and/or 5 were the likely original context(s). Three scenarios to account for their taphonomic histories and the techno-typological characteristics of the artefacts were provided (Tizzard et al. Reference Tizzard, Bicket and De Loecker2015; Fig. 8). These are:

1. 1. The handaxes and some of the Levallois material have been reworked through Units 3b, 5 and, potentially, 8, with some of the Levallois material being contemporary with, and originating from, Unit 3b (Fig. 8A).

2. 2. Both the handaxes and the Levallois material are contemporary with, and originates from, Unit 3b, with potential for some material from both groups to have been reworked into later deposits (Fig. 8B).

3. 3. The Levallois material is contemporary with, and originates from, Unit 3b and the handaxes are contemporary with, and originate from, Unit 5, with paths of reworking of some of the material through later deposits (Fig. 8C).

Fig. 8. Proposed scenarios for depositional histories of Area 240 Palaeolithic artefacts

Of these possible scenarios Tizzard et al. (Reference Tizzard, Bicket and De Loecker2015) concluded that scenario 2 was the most likely and that the handaxes and Levallois artefacts were broadly contemporary in age. This conclusion was based on both groups of artefacts being in similar physical condition states, indicative of shared taphonomic histories, and the presence of artefacts relating to Levallois flaking reflecting the chronology of Unit 3b, this technique being typical of MIS 8–7–6 early Middle Palaeolithic assemblages in the Thames Valley (Scott Reference Scott2011).

The new Palaeolithic artefact collection from Area 240 allows this interpretation to be revisited utilising a dataset with greater spatial information. Analysis of the physical condition of the new material has demonstrated that, whilst different techno-typological classes of artefacts do exhibit a similar continuum of surface modifications, quantitative differences are apparent between handaxes and Levallois material, demonstrating that these two groups have different taphonomic histories.

These different taphonomic histories suggest that they are not likely to reflect a single contemporary phase of activity. Additionally, the spatial distribution of artefacts in the new collection can be mapped on to the lateral distribution of different units of the Palaeo-Yare lithostratigraphy (Figs 9–10). This demonstrates that handaxes have only been recovered from locations where both Unit 3b and Unit 5 have been identified (Fig. 9), while Levallois artefacts have been recovered from areas where both Unit 3b and 5 are mapped, and from areas where only Unit 3b is known to occur (Fig. 10). This combined evidence raises the possibility that the Levallois material in the new collection is from Unit 3b, while the handaxes may be from Unit 5

Fig. 9. Spatial distribution of Palaeolithic handaxes from Area 240

Fig. 10. Spatial distribution of Palaeolithic Levallois artefacts from Area 240

Dates for Unit 3b (Table 3) generally cluster around MIS 8–7–6 (Tizzard et al. Reference Tizzard, Bicket and De Loecker2015; Marshall et al. Reference Marshall, Bayliss, Grant, Bridgland, Duller, Housley, Matthews, Outram, Penkman, Pike, Schreve and Xuan2020). No direct dates are available for Unit 5. However, as they have been observed to stratigraphically overlie Unit 3b, while vertebrate remains obtained alongside artefacts recovered in 2007–2008 have produced MIS 3 radiocarbon ages and are most likely to be from Unit 5 sediments, or post-MIS 6 sediments within the top of Unit 3b. One caveat with the radiocarbon dates is that radiocarbon ages from the North Sea can under-estimate dates when carried out following the standard dating methods applied to this material (Briant & Bateman Reference Briant and Bateman2009; Hijma et al. Reference Hijma, Cohen, Roebroeks, Westerhoff and Busschers2012).

Review of all lines of evidence indicates that the new Palaeolithic artefacts from Area 240, as well as those previously recovered in 2007–2008, contain three distinct groups of artefacts:

• a group containing Levallois artefacts, likely from Unit 3b with deposits dated to MIS 8–7–6;

• a group containing handaxes that may be from Devensian deposits (MIS 5d–3); and

• late Upper Palaeolithic/Early Mesolithic material, represented by a single blade core.

PALAEO-YARE

In addition to the small number of artefacts from wharf monitoring of aggregate from locations beyond Area 240 (Table 1), over 1000 artefacts have recently been documented dredged from the offshore Palaeo-Yare catchment and redeposited as part of the Bacton to Walcott Coastal Management Scheme (Davies et al. Reference Davis, Ashton, Bynoe, Craven, Ferguson, Gardner, Grimmer, Harris, Hatch, Johnson and Leonard2023). Although there is some uncertainty regarding which licence area, or areas, that these are from (multiple areas were dredged to provide the aggregate) data indicate that they may be largely from Areas 511 and 512. These areas are to the south and west of Area 240 (Fig. 1).

Although most artefacts in the Bacton to Walcott collections are undiagnostic hard hammer flakes, they include handaxes (n = 34) and soft hammer flakes (n = 28), possibly from handaxe thinning, and Levallois cores (n = 12) and Levallois flakes (n = 71). Some of this material displays notable similarities with the Area 240 artefacts.

Many of the handaxes are similar to those from Area 240, being cordiform/ovate with comparable measurements of elongation and planform (Davies et al. Reference Davis, Ashton, Bynoe, Craven, Ferguson, Gardner, Grimmer, Harris, Hatch, Johnson and Leonard2023; Appendix S1), extensively thinned and shaped, and on flake blanks. Although detailed comparison between the methods of working of the Area 240 and Bacton to Walcott handaxes is not possible based on the available data, the latter are described as extensively shaped, which suggests a similar focus on the imposition form. The cordiform/ovate handaxes from Bacton to Walcott occur alongside others that are morphologically different from the Area 240 examples being less refined and more elongated (Davies et al. Reference Davis, Ashton, Bynoe, Craven, Ferguson, Gardner, Grimmer, Harris, Hatch, Johnson and Leonard2023; Appendix S1).

The data suggest that the Bacton to Walcott handaxes include examples which may belong to the same repertoire of handaxe manufacture as Area 240 (the cordiforms/ovates) but which are from different locations in the Palaeo-Yare catchment. The outliers from Bacton to Walcott may be part of a wider continuum of similar handaxes or indicate a distinct approach to their manufacture.

The lack of Levallois cores from Area 240 prevents meaningful comparison with the Bacton to Walcott examples but there are similarities in the Levallois flakes. As with those from Area 240, they are large (average maximum dimension = 94.2 mm) and indicative of centripetal preparation (54.3%), with many examples part of unipolar, recurrent exploitation sequences (36.6%). These similarities may simply reflect the common exploitation of large cores but also imply similar technological approaches to Levallois core working which may reflect a uniformity in landscape-use practices. For example, the exploitation of raw material sources that were close to where large flint nodules were available and/or transport of large Levallois products.

As with Area 240, similar differences in condition between the handaxes and Levallois material from Bacton to Walcott have been noted (Davies et al. Reference Davis, Ashton, Bynoe, Craven, Ferguson, Gardner, Grimmer, Harris, Hatch, Johnson and Leonard2023) but interestingly this pattern of difference is inverse. Most of the Bacton to Walcott artefacts are at least slightly rolled but handaxes tend to be more abraded than Levallois pieces. These observations suggest that the handaxes and Levallois artefacts from both Area 240 and Bacton to Walcott have distinct taphonomic histories but that the specific patterns differ for equivalent groups of artefacts.

Deposits assigned to Units 3a, 3b, 7, and 8 of the Palaeo-Yare lithostratigraphy are mapped within Licence Areas 511 and 512 (Wessex Archaeology 2013; 2020). This may imply that all the Middle Palaeolithic artefacts from Bacton to Walcott are from Unit 3b. However, other units could occur, particularly where the licence areas abut Area 240 and more localised occurrences of Units 5 and 6 are recorded (Fig. 9).

There are also indications from vibrocore data of complexity in deposits mapped as Unit 3b in Areas 512 and 513. A vibrocore (VC29_2) in Area 511 was assessed as part of the East Coast Regional Environmental Characterisation (Limpenny et al. Reference Limpenny, Barrio Froján, Cotterill, Foster-Smith, Pearce, Tizzard, Limpenny, Long, Walmsley, Kirby, Baker, Meadows, Rees, Hill, Wilson, Leivers, Churchley, Russell, Birchenough, Green and Law2011). These comprised a basal shelly sand with an interbedded sandy clayey silt further up the profile that contained microfossils (foraminifera and ostracods) indicative of probably cold estuarine/shallow marine conditions. These deposits were overlain by sand interbedded with finer silt clay laminae. Optically Stimulated Luminescence (OSL) dates within MIS 7–6 were obtained for the shelly sand and interbedded sandy clayey, whilst a date of 57±5.6 ka (MIS 3) was obtained from the upper sand. This younger date for the uppermost deposits was dismissed on the grounds that the sediments could have been subaerially exposed in the Devensian, with a fall in sea level indicated by oxidisation. Whether this is indeed the case is not known, but it is worth noting the sample was taken 0.41 m below the top of these sediments, which were truncated and overlain by recent marine deposits. This evidence, at the very least, implies the presence of mid-Devensian terrestrial landscapes in this area.

The combined archaeological evidence from Area 240 and Bacton to Walcott illustrate that Middle Palaeolithic archaeology occurs in different locations across the offshore Palaeo-Yare and that there is a re-occurring pattern of differences in taphonomic histories between handaxes and Levallois artefacts. Additionally, similar and repeated approaches to handaxe production are evident (refined, extensively worked and shaped cordiform handaxes on flake blanks). The Bacton to Walcott Levallois artefacts and handaxes may reflect artefact assemblages from different Unit 3b deposits dated to MIS 8–7–6, but a division between early Middle Palaeolithic Levallois artefacts and cordiform/ovate handaxes from younger, potentially Devensian, contexts, as suggested for Area 240, is also possible.

Archaeologically the closest point of reference for the submerged Middle Palaeolithic record of the Palaeo-Yare is the now onshore Yare catchment. Unfortunately, terrestrial outcrops of Pleistocene Yare deposits have been minimally investigated and extant archaeology from them is poorly understood.

Aside from chance individual surface finds, known Lower/Middle Palaeolithic archaeology divides between two groups. The first consists of fresh to minimally abraded flakes and handaxes from fluvial deposits of the River Yare, situated above the current floodplain. The principal examples are large artefact assemblages from Whitlingham Sewage Farm, Kirby Bedon (Sainty Reference Sainty1927; Wymer Reference Wymer1985; Reference Wymer1999) and Mill Gravel Pit, Keswick (Sainty Reference Sainty1933; Wymer Reference Wymer1985; Reference Wymer1999). The ages of these sites are unknown but the handaxe assemblages from both are distinctive and similar, comprising elongated pointed handaxes with large numbers of cleavers (Wymer Reference Wymer1985). No similar handaxes or any cleavers have been recovered from the offshore Palaeo-Yare.

The second group of artefacts consists of small numbers of fresh, minimally disturbed finds, principally of handaxes, from low river terraces on the edge of the current floodplain of the Yare catchment. Examples include a fresh handaxe from amongst more abraded artefacts from Yare deposits at Carrow Works, Norwich (Sainty Reference Sainty1933; Wymer Reference Wymer1985; Reference Wymer1999) and a mint handaxe from a low terrace of the Upper Waveney at Homersfield (Wymer Reference Wymer1985; Reference Wymer1999). These may be younger than the Whitlingham and Keswick assemblages and could relate chronologically to Palaeo-Yare Middle Palaeolithic archaeology, but there is no evidence to directly link them.

BEYOND THE PALAEO-YARE

The suggestion that the Middle Palaeolithic artefacts from Area 240 include two chronologically distinct groupings, an earlier one containing Levallois artefacts and a later one containing handaxes, can be further interrogated through consideration of the Middle Palaeolithic archaeological record from the wider region.

Association between the Levallois artefacts and Unit 3b sediments dated to MIS 8–7–6 would be consistent with the Middle Palaeolithic archaeological record from terrestrial contexts in southern England, particularly in the Thames Valley (Scott Reference Scott2011). Additionally, Levallois material of a similar date has also been identified from now submerged contexts of the lower reaches of Thames–Medway systems from aggregate licence Area 447 (Bynoe et al. Reference Bynoe, Grant and Dix2022).

The Palaeolithic artefacts from Area 447 are from aggregate redeposited as part of the Clacton-on-Sea to Holland-on-Sea, Essex beach replenishment programme. The archaeology includes 121 Levallois flakes and 21 Levallois cores which, generally, have slightly rolled edges, with some more moderately rolled pieces (Bynoe et al. Reference Bynoe, Grant and Dix2022). This condition is indicative of some, but not extensive, reworking of the material within their original lithostratigraphic context. The Levallois artefacts are suggested to be from sands and clays forming part of sequences of freshwater/estuarine terrestrial deposits, cut by later Pleistocene fluvial channels (Bynoe et al. Reference Bynoe, Grant and Dix2022). The freshwater/estuarine deposits have been dated through luminescence to MIS 7/early MIS 6.

Evidence for extensive early Middle Palaeolithic hominin activity associated with now submerged contexts around Britain is also documented in the English Channel region. A prime example of this is the large archaeological assemblage from several stratigraphic units at La Cotte de St Brelade, Jersey, which reflects repeated periods of early Middle Palaeolithic activity focused in now submerged MIS 7 and early MIS 6 landscapes (Shaw et al. Reference Shaw, Bates, Conneller, Gamble, Julien, McNabb, Pope and Scott2016; Bates et al. Reference Bates, Shaw, Bates, Pope and Scott2023; Shaw & Scott Reference Shaw and Scott2023).

This emergent picture of hominins being present in both the upper, now terrestrial, and the lower, now submerged, catchments of river systems during the early Middle Palaeolithic is highly significant for Palaeolithic settlement histories. It is particularly important for debates relating to hominin demographics and presence and absence (White & Schreve Reference White and Schreve2000; Ashton & Lewis Reference Ashton and Lewis2002; Ashton et al. Reference Ashton, Lewis, Hosfield, Ashton, Lewis and Stringer2011; Roebroeks et al. Reference Roebroeks, Hublin, MacDonald, Ashton, Lewis and Stringer2011).

Conclusions based on assessments that do not consider periods of lowered sea levels, when terrestrial landscapes extended across now submerged areas, are questionable. This is particularly so given that the low-lying landscapes associated with the lower reaches of river systems and the estuaries would include resource-rich environments that would have been particularly attractive to hominin populations. During such periods, hominin activity could be expected to be concentrated in these now submerged areas.

The new evidence suggests that the handaxes from Area 240 may be younger than the Levallois artefacts, with Unit 5 of the Palaeo-Yare lithostratigraphy suggested as a possible context. This implies a late Middle Palaeolithic date within the period from after MIS 5d to MIS 3.

The late Middle Palaeolithic archaeological record of Britain is poorly understood. For the most part identification of late Middle Palaeolithic activity is based on findspots of individual techno-typologically distinctive handaxes, generally with limited contextual information. Such handaxes are termed bout coupés. Definitions of bout coupés vary and are principally based on typological criteria. A strict definition of ‘classic’ bout coupés was provided by Tyldesley (Reference Tyldesley1987) who defined such handaxes as refined, fully bifacial, medium sized, cordiform or rectangular with a symmetrical planform, with a straight or slightly convex butt edge, slightly convex sides, and a rounded tip. Both the butt and the tip are well worked, principally through soft-hammer removals, and they possess a cutting edge around the full circumference of the piece.

White and Jacobi (Reference White and Jacobi2002) have reviewed the lithostratigraphic and chronological context of ‘classic’ bout coupés handaxes. This demonstrated that examples with secure contextual information and chronological information are from deposits dating from within the period MIS 5d to MIS 3. White and Jacobi (Reference White and Jacobi2002) further suggest a specific association with MIS 3. Although none of the handaxes from Area 240 can be classed as ‘classic’ bout coupés as defined by Tyldesley (Reference Tyldesley1987), several approach this planform (Fig. 11) and most share approaches to their manufacture with ‘classic’ bout coupés.

Fig. 11. Cordiform handaxes from Area 240

Beyond individual finds of bout coupé handaxes, there are very few well contextualised archaeological British assemblages dated to the late Middle Palaeolithic. The principal example is that from Lynford Quarry, near Munford, Norfolk (Boismier et al. Reference Boismier, Gamble and Coward2012). Here, Palaeolithic artefacts were recovered from organic sediments within a palaeo-channel. Vertebrate fauna dominated by woolly mammoth (Mammuthus primigenius) was also recovered from these deposits (Schreve et al. Reference Schreve, Brothwell and Stuart2012).

The artefacts and vertebrate fauna were interpreted as having been incorporated into the channel fill through periodic bank collapse, with artefacts recovered from three units (B-i, B-ii, and B-iii). While not in situ, their fresh condition indicates that the artefacts were discarded on nearby channel edges which have been periodically incorporated into the channel fill (White Reference White2012). OSL dates suggest that the channel sequence dates to between 65 and 57 ka (MIS 4–3) (Schwenninger & Rhodes Reference Schwenninger and Rhodes2012).

In total, 2720 artefacts, including 41 complete and six broken handaxes, are provenanced to the Lynford palaeo-channel. The Lynford handaxes, the largest assemblage of which was from unit B-ii (n = 38), provide the best contextualised late Middle Palaeolithic handaxe assemblage from Britain. Data provided by White (Reference White2012) enables techno-typological comparison between them and those from Area 240.

Figure 12 compares the planform of the handaxes from Area 240 (De Loecker Reference De Loecker2010) and the current study) and from Lynford B-ii (White Reference White2012). This demonstrates that the handaxes from these assemblages share similar and highly consistent planforms, generally being cordiform in shape and transcending Roe’s (Reference Roe1964; Reference Roe1969) divide between points and ovates. As White (Reference White2012) notes, this range in planform is similar to those provide by Roe (Reference Roe1968) for handaxes from Great Pan Farm, Isle of Wight, and from the Oldbury area, Kent, collections which may also be dominated by late Middle Palaeolithic handaxes.

Fig. 12. Area 240 and Lynford Quarry, Norfolk, handaxe planform (following Roe Reference Roe1964; Reference Roe1968)

Table 4 compares key metrical attributes between the assemblages from Area 240 and Lynford and clearly demonstrates similarities in levels of refinement, elongation, tip section, and cross-section. They demonstrate that the handaxes in all three are generally highly refined and uniform in cross section.

Table 4: Quantitative variables for handaxe collections from area 240 and lynford b-ii

Comparison between the technological features of Area 240 handaxes and those from Lynford (White Reference White2012) demonstrates further similarities. These include the use of flakes as blanks, a lack of cortex, extensive thinning, final phases of less invasive scars around the margin which shaped the handaxes (termed ‘edge retouch’ by White (Reference White2012)), and a high proportion of handaxes with working edges around the entire circumference. Additionally, the Lynford handaxes are notable for their high levels of recycling with examples of reconfiguring of broken and fragmentary handaxes, use as a core for laminar removals, the addition of scraper retouch, and the imposition of notches on handaxe margins. Recycling practices, particularly the reconfiguring of broken handaxes, was also observed on several of the Area 240 examples. This pattern is characteristic of tools that have been curated, carried around as personal items, and subjected to multiple phases of reworking, rather than tools produced for a specific activity, retained for limited periods, and discarded.

The evidence demonstrates that the Area 240 handaxes share many characteristics with the late Middle Palaeolithic handaxes from Lynford. Characteristics are evident that reflect a focus on extensive thinning and shaping of handaxes on flake blanks to produce refined products with a clear focus on the imposition of particular forms that have undergone multiple phases of use and maintenance.

The handaxes from Area 240 sit alongside those from Lynford as technologically characteristic of Late Middle Palaeolithic handaxes associated with Devensian (MIS 5d–3) contexts. This suggests the Area 240 examples are unlikely to be from Unit 3b sediments dated to MIS 8–7–6, but are from younger deposits, either sediments within Unit 3b, post-dating MIS 6, or from Unit 5. The Area 240 handaxes, and potentially those from Bacton to Walcott (Davies et al. Reference Davis, Ashton, Bynoe, Craven, Ferguson, Gardner, Grimmer, Harris, Hatch, Johnson and Leonard2023), demonstrate the lower, submerged reaches of the Palaeo-Yare were a major focus for Devensian Late Middle Palaeolithic human activity. The fact that this is a poorly understood period, often considered associated with an impoverished archaeological record, may at least partially reflect that it was these now submerged Devensian landscapes were those most extensively occupied during this period.