White Swan Public House, Yabsley Street, Blackwall, London Borough of Tower Hamlets

White Swan Public House, Yabsley Street, Blackwall, London Borough of Tower Hamlets A Post-Excavation Assessment for St James Group Ltd By Sarah Coles, Steve Ford and Andy Taylor Thames Valley Archaeological Services Ltd Site Code YAB02 November 2003

Contents 1. Introduction 2. Archaeological background 3. The evaluation 4. Original project objectives 5. Purpose of this report 6. Excavation methodology 7. Excavation results 8. Phase by phase summary 8.1 Phase 1: Neolithic 8.2 Phase 2: Early Bronze Age 9. Nature and character of recovered material and statement of potential 9.1 Pottery by Frances Raymond 9.2 Fired clay by Steve Ford 9.3 Struck flint by Steve Ford 9.4 Burnt flint by Steve Ford 9.5 Human remains by Sian Anthony 9.6 Wood by Rowena Gale 9.7 Charred plant remains by Mark Robinson 9.8 Pollen by Michael Keith-Lucas 9.9 Radiocarbon dating 10. Summary of the significance of the data 10.1 Research questions the material will address 10.2 Intrinsic value of the material itself 11. Conclusions 12. Updated project design 13. Proposals for publication 14. Resources and timetable 15. References APPENDIX 1: Summary of contexts and finds APPENDIX 2: Pottery fabric descriptions APPENDIX 3: Summary of the flint assemblage APPENDIX 4: Charred plant remains APPENDIX 5: Outline publication synopsis APPENDIX 6: GLSMR summary i

White Swan Public House, Yabsley Street, Blackwall, Tower Hamlets A Post Excavation Assessment by Sarah Coles, Steve Ford and Andy Taylor with contributions by Sian Anthony, Rowena Gale, Michael Keith-Lucas, Frances Raymond, and Mark Robinson Report 02/54c Summary Excavations in advance of redevelopment of the former White Swan public house revealed traces of Neolithic settlement and burial located on a sand and gravel bar beneath alluvium. A large expanse of peat also buried by alluvium was located between these finds and the modern river, suggesting that the occupation was located on the foreshore of the Thames. A single grave cut into the natural sand contained a poorly preserved crouched inhumation, possibly of a woman. The burial was accompanied by a fragment of carinated bowl, a flint knife and other struck flints. A radiocarbon date obtained from an oak retaining plank within the grave was calibrated to 4220 3979 cal BC. This burial is one of the earliest within the British Isles and is the earliest known from London. A scatter of struck flint and pottery predominantly of Early Neolithic date was recovered from adjacent areas of the sand. A nearby hearth contained 9 fragments of Early Bronze Age pottery pointing to some other prehistoric activity nearby. A small number of charred plant remains recovered indicate both the collection of wild plant foods and cereal growing. Pollen analysis of the peat revealed a marked decline in woodland cover at the start of the sequence with a rise in grassland and herb species. Cereal pollen then briefly becomes a significant component of the sequence before declining to more modest levels. Radiocarbon dating is awaited to determine the chronological relationship of these environmental indicators to the nearby Neolithic occupation and burial. 1 Introduction 1.1 This report outlines the potential for further analysis arising from an archaeological investigation carried out on the site of the former White Swan Public House, Yabsley Street, Blackwall, London Borough of Tower Hamlets (TQ 3844 8040) (Fig. 1). Research aims which may be addressed by the analysis are identified. The aim is to target post-excavation resources where the information gain will be greatest, in line with current local, regional and national research priorities. A programme for analysis is proposed. 1.2 Planning permission (Application no PA/01/01323) has been granted by the London Borough of Tower Hamlets for the construction of residential apartments, subject to a condition relating to archaeology requiring the provision of an archaeological survey prior to the commencement of work. The investaigation was commissioned by Mr Andy Ainsworth of St James Group Ltd to satisfy the condition on planning consent for the redevelopment of the site 1.3 The site comprises an L-shaped parcel of land with frontage onto Blackwall Way, Yabsley Street and Preston s Road. The site lies on the eastern side of the Isle of Dogs to the south of Blackwall and approximately 100m west of the contemporary river Thames. The site lies on the Thames floodplain 1

with alluvium overlying gravel (and peat) (BGS 1981), at a height of between 2.84m and 3.7m above Ordnance Datum. 1.4 The archaeological potential of the site was highlighted by a desk-based assessment (Hardy 2002) and this potential was confirmed by a field evaluation (Taylor 2002), which revealed prehistoric finds and deposits on a raised area of sand and gravel beneath alluvium to the west with deep peat beneath alluvium to the east. 1.5 As the majority of the site was to be occupied by a basement which would cut into the archaeologically relevant levels, a formal program of archaeological excavation was requested for the site, following a specification approved by Mr Nick Truckle, the Archaeology Adviser with the Greater London Archaeology Advisory Service. This was in accordance with the Department of the Environment s Planning Policy Guidance Archaeology and Planning (PPG16 1990) and the Borough s policies on archaeology, in order to satisfy the archaeological condition placed on the planning permission. 1.6 The field work was supervised by Sarah Coles and Andy Taylor, with the assistance of Sian Anthony, Simon Cass, Danielle Colls, Jennifer Lowe, Andy Mundin, and Jo Pine. The fieldwork took place between March and July 2003. 1.7 The archive is currently held by Thames Valley Archaeological Services Ltd but it is anticipated that it will be deposited with the Museum of London in due course. The site code is YAB 02. 2 Archaeological background 2.1 The site lies in an area which, in historic times was prone to inundation. The Isle of Dogs was embanked during the medieval period and possibly earlier to reclaim the land from flooding for use as farmland, but this embankment was often breached. Early maps of the area show the presence of a Marsh Wall. The Rocque map of 1766 shows this wall meeting the medieval settlement of Blackwall and continuing beyond to the north east. This inundation has led to the deposition of alluvium with land surfaces contemporary with Roman and earlier occupation now beneath 1m or more of silty clay. The rising water levels began much before Roman times with large areas of the Thames Estuary also witnessing the formation of peat (Siddell et al. 2000). 2.2 The presence of this alluvium and peat, whilst limiting the opportunities for site discovery, has led to good survival of deposits often with preservation of organic remains. One particular example of this is the presence of a complex of wooden structures of Bronze Age date at Atlas Wharf, on the south-west tip of the Isle of Dogs (MoLAS 2000, 99). 2.3 Recent excavation at Express Wharf on the west side of the Isle of Dogs (site WYO01, see Fig. 1) revealed both earlier prehistoric finds (flint and pottery) and Roman occupation (Anthony and Ford 2003). One feature of the increasingly wet environment in the Neolithic, and more especially in the Bronze Age, is the construction of wooden trackways presumably to link higher, drier settled areas (gravel islands) with marshy areas between (Meddens 1996). A Neolithic trackway is recorded from Silvertown to the east (Crockett 2002). 3 The evaluation 3.1 An evaluation was carried out by Thames Valley Archaeological services in 2002 3.2 This consisted of the excavation of 4 trenches (Fig. 2) varying in length from approximately 6m to 10m. Trench 1 contained a shallow scoop, possibly a hearth, containing several sherds of degraded Bronze Age pottery and other undiagnostic prehistoric pottery. 3.3 In addition a sand capped terrace edge was identified in the trenches located closer to the river, with higher ground to the west. The trenches to the west of the site identified a possible peat-filled channel, and both the sand and peat natural levels were sealed by alluvium. 3.4 A column sample for pollen analysis was taken (see 9.8). 2

4 Original Objectives 4.1 General objectives 4.1 A number of national policy documents for archaeological research have considered the time span that the deposits represent for example, Exploring Our Past (English Heritage 1991, 36) and the English Heritage Archaeology Division Research Agenda (1997, 44). 4.1.1 The more general objectives of this project were to: 4.1.1.1 Excavate and record all archaeological deposits and features within the areas threatened by the proposed development. 4.1.1.2 Produce relative and absolute dating and phasing for deposits and features recorded on site. 4.1.1.3 Establish the character of these deposits in an attempt to define functional areas on the site such as industrial, domestic etc. 4.1.1.4 Produce information on the economy and local environment and compare and contrast this with the results of other excavations in the region. 4.2 Specific research objectives for the excavation and post excavation project aimed to answer the following questions: 4.2.1 Is there any prehistoric activity or settlement on the site, and if so, what is its date, nature and extent? 4.2.2 Are there any structural remains on the site and if so, what is their date? 4.2.3 What activities are taking place on the site? 4.2.4 What is the nature and function of the site use? Was it an occupation site permanently or seasonally used, was it for a one-off use or was it repeatedly used over many years or in different seasons? Do the deposits and finds already recorded merely reflect casual use? 4.2.5 When was the site abandoned and when was the site buried by alluvium? 4.2.6 What use was of floral and faunal resources and can these be identified and assessed from a programme of environmental sampling 4.2.7 Is there preservation of any organic remains in the vicinity if the peat-infilled channel? 4.2.8 What is the palaeoenvironmental setting of the site during, before and after its use? 5 Purpose of this report 5.1 The current report summarizes the results of the excavation, the archaeological features recorded and the finds recovered, and provides considered assessments of the potential these possess to answer research questions about the site, and how they fit into local, regional and national context. The archaeological remains are first quantified and described, to establish their quality, character and significance. These are then assessed relative to the original project objectives. The potential to address these objectives is discussed, and any new potential objectives arising from the nature of the results of the excavation are also highlighted. 5.2 The result is the production of an updated project design to match the appropriate levels of analysis and publication to the results of the excavation. A methodology and timetable are proposed and the required resources quantified. 6 Excavation Methodology 6.1 The excavation was carried out across the whole extent of the proposed basement area which occupied the majority of the site (Fig. 2) and was divided into three areas A, B and C, excavated sequentially and 3

episodically to encompass the unusually complicated engineering works required due to the presence of the Blackwall Tunnel beneath. 6.2 Made ground was removed by a 360 mechanical excavator fitted with a toothless bucket to expose the uppermost surface of archaeological deposits. This was subject to archaeological supervision and control for the archaeologically sensitive levels at the base of the alluvium in Area A, and to the base of the peat in Areas B and C. In both cases the remainder of the alluvium and peat were then removed by hand. 6.3 A full written, drawn and photographic record of the excavation was made following the procedures laid down in the TVAS Fieldwork Manual. 6.4 Finds recovered from the alluvium, peat and humic sand were 3D-recorded using a Total Station and given a unique identifying number. 6.5 Two soil samples were taken during the course of the excavation, for the recovery of environmental evidence. 7 Results 7.1 It was noted that on the eastern portions of the site, the formation level of the basement more or less coincided with the base of the alluvium and top of the peat. In these areas therefore, apart from localized deeper intrusions such as for the lift pits (and earlier evaluation trenches 3 and 4), the sand/ gravel horizons were not exposed. The main excavation was located towards the west of the site. In this location the peat thinned to become little more than a humic sand overlying clean sand, which represented the archaeologically relevant horizon. 7.2 Because of the episodic nature of the investigation, the excavation strategy developed for each phase. In Area A (Fig. 3), following the removal of made ground, the alluvium was stripped to expose the humic sand. This layer was hand cleaned and a small number of struck flints, burnt flint and pottery were retrieved from this, and the interface of it with the overlying alluvium. It did not however, reveal any cut features other than the shallow scoop (50) recorded in the evaluation. This layer was then removed by hand to expose the sand beneath. Further finds were recovered. However, it was not until approximately 100mm of the sand was removed that the cut feature (100) and the majority of the flint and pottery finds were exposed (Fig. 4). 7.3 For the remaining two areas, (Fig. 3, B and C) the humic sand was exposed, scanned for finds and removed by machine to expose the sand. The sand was then hand cleaned and excavated down to 100mm as before. A post medieval well (101) was recorded in Trench B. 8 Phase by Phase summary 8.1 Two phases of activity were identified on the site, being Neolithic and Early Bronze Age in date. 8.1.1 The phasing is broadly based on the artefactual evidence. 8.2 Phase 1: Neolithic 8.2.1 Grave 100 8.2.1.1 Upon excavation cut feature (100) was revealed to be a grave (Fig. 5). This was nearly rectangular in plan 1.62m by 0.8m wide (max) and 0.2m deep. It had a bowl-shaped profile with a flat base (Fig. 6). It was filled with a grey/black sand, mottled yellow with charcoal flecks and some burnt flint. There was a large amount of empty space at the feet. The grave contained a single inhumation which was accompanied by flint and pottery grave goods. 8.2.1.2 The northern side of the grave was partly defined by a single length of oak plank (155). It was initially considered that this represented the presence of a coffin. However, despite the good preservation of this substantial piece of timber, no other wood, or even staining was revealed, indicating that this absence is unlikely to be a result of differential survival. A more appropriate interpretation is that the wood represents some sort of retaining structure, perhaps holding back collapse of the soft sand. 4

8.2.1.3 A sample from the wood provided a radiocarbon determination of 5252 ± 28 BP which calibrates to between 4220 3979 cal BC at the 2 sigma level. 8.2.1.4 Orientated east-west, the body faced south to be on its left side The body was flexed but not tightly although the head was facing towards the knees. The spine touched the northern edge of the grave and the piece of wood that lined the grave, the arms were flexed and the feet together. The skeleton was in an extremely poor state of preservation, the acidic conditions had destroyed the majority of the bone leaving just darkened sand stains in place of most of the major limbs, the spine, the mandible and teeth. The poor survival is due to the loose, well aerated acidic sand that led to rapid decomposition (Lyman 1994) and fluctuating amounts of groundwater leaching out the mineral content of the bone. The areas best preserved were portions of the body where bone was densely packed reducing the surface area of bone opposed to the sand and areas with less cancellous bone. 8.2.1.5 Nineteen sherds of pottery and 8 flints were recovered from the grave. One of the flakes and the spall had been burnt. A large rim fragment of a carinated bowl (Figs. 3, 7) appeared to have been placed above the head. Similarly, a flint knife was positioned at the foot of the grave and also appears to have been placed. It is less clear if the remaining sherds and flints are placed, or are accidental incorporations within the grave. 8.3 Phase 2: Early Bronze Age 8.3.1 Hearth 50 8.3.1.1 In the north-west corner of evaluation trench 1 was a shallow scoop (50) (Fig. 2). This was 0.7m across and 0.05m deep. It contained much charcoal, several pieces of fired clay, burnt flint and 9 sherds of degraded Early Bronze Age pottery. It is probable that the charcoal patch was a hearth, though no firereddening of the natural sand was present. 9 Character of material and statement of potential 9.1 Pottery by Frances Raymond 9.1.1 A small assemblage of prehistoric pottery, comprising 220 sherds weighing 1163 grams, was recovered from the site. Much of this material is derived from an early Neolithic carinated bowl found in the sand (context 152) adjacent to a grave (100). The more complete of the featured sherds, including a rim fragment accompanying the burial, are from vessels of a similar style, while the remainder of the sherds from the site are likely to be broadly contemporary. The only exceptions are a few fragments of grog tempered pottery from a shallow scoop in Evaluation Trench 1 which are probably of early Bronze Age origin. 9.1.2 The assemblage has been analysed in accordance with the guidelines of the Prehistoric Ceramics Research Group (PCRG 1997). A full record of all recorded traits including fabric, form, surface treatment, colour, wall thickness, sherd size and abrasion is available in the project archive. 9.1.3 The Assemblage from the grave 9.1.3.1 The grave fill produced 19 sherds in variable condition with a total weight of 79 grams, derived from at least four different vessels. The assemblage includes three rims, the most complete of which is from an early Neolithic open carinated bowl. This is dark grey (10YR4/1), has burnished surfaces and is made from a sandy fabric tempered with fine flint (FS/2). The sherd is relatively large (50g), in good condition and was found behind the skull where it gave the impression of having been placed deliberately. A second beaded rim, weighing two grams and three thin-walled body sherds (weighing 4g) with identical surface finishes, and made from the same fabric were also found in the grave fill. 9.1.3.2 It is quite possible that this pottery is derived from the open carinated bowl from the sand adjacent to the grave, although none of the fragments refit with this vessel. The surface finish, colour range and fabric of the sherds from both contexts are strikingly similar. Furthermore, the more complete vessel from the sand has a rim which is simple and rounded in some places (similar to no. 4) and beaded in others (similar to no. 1). 9.1.3.3 The third rim from the grave fill is small (1g) and its outer surface is missing. Although the form of the vessel is uncertain, it is made from a distinctive sandy, shell filled fabric (VS/1). Six body sherds, weighing 13g and of variable wall thickness (5 to 9 mm) are also characterized by this same ware. Some 5

of these fragments are in fresh condition, while others are lightly to moderately abraded. The surfaces of most of the better preserved sherds have been smoothed, while the exterior of one example has been burnished. 9.1.3.4 The remaining seven fragments of pottery from the grave fill are made from predominantly sandy wares. These include two heavily abraded sherds (4g), in a fabric filled with coarse sand and a few fragments of shell (S/1); and five sherds (5g) in a very fine sandy ware (S/2). The latter are thin-walled and the better preserved are burnished on both surfaces or have burnished exteriors. 9.1.4 The pottery from the sand 9.1.4.1 The sand produced 66 scattered sherds (126g) and approximately half of an early Neolithic open carinated bowl. This portion of the vessel was found intact adjacent to the grave, but on lifting separated into numerous fragments along ancient stress fractures (125 sherds, 939g). Like the sherds from the grave, the bowl is thin-walled, burnished on both surfaces and made from a sandy fabric filled with fine flint (FS/2). It is predominantly grey to dark grey in colour (10YR5/1, 4/1 and 3/1) with some dark greyish brown areas (10YR4/2). 9.1.4.2 Twenty-four of the scattered body sherds, (51g) and mostly in good condition, share similar attributes with the carinated bowl including the fabric. These were either found close to this vessel (No. 30); formed part of a cluster to the south (Nos. 61, 62 and 65); or were distributed in an apparently random pattern to the north across Areas B and C (Nos. 1002, 1018, 1031, 1052, 1059 and 1077). 9.1.4.3 The pottery from the sand also includes 16 sherds (28g), in the other three fabrics represented in the grave (VS/1, S/1 and S/2). These are generally in poorer condition than the grave assemblage, being either moderately or heavily abraded. There is little patterning in the distribution of the sandy wares, which occur in Area A (Nos. 25, 29 and 44) and Areas B and C (Nos. 1015 and 1070), but there is a small cluster of sherds made from the shell tempered fabric in Area B (Nos. 1015, 1020 and 1023). 9.1.4.4 The rest of the assemblage from the sand incorporates eight sherds in variable condition made from three contrasting flint tempered wares and 18 fragments (11g), which are too small or heavily abraded for fabric characterisation. At least one of the identifiable wares, filled with fine flint (FS/1), is demonstrably Neolithic having been used for a rim and shoulder (8g) which may be part of the same carinated bowl. This has a dark grey (10YR4/1), burnished exterior and the rim has a hole drilled after firing possibly denoting a repair. Two small burnished body fragments (1g) made from the same fabric are also represented. The rim and body fragments were found in close proximity in Area C (Nos. 1071 and 1077), while the carinated sherd came from some 20 metres to the north in Area B (No. 1027). 9.1.4.5 The other four flint tempered sherds (27 g) are featureless body fragments providing no evidence for style. They are made from coarser fabrics (FMS/1 and FS/3) and are from vessels with relatively thick walls (8 to 10mm). The fabrics display characteristics consistent with an early Neolithic date and while it is quite probable that they belong to this period, they could equally have been produced during subsequent phases of prehistory. The distribution of these coarser wares is scattered across all three areas (Find Nos. 25, 57, 1024 and 1078). 9.1.5 The sherds from the buried soil and evaluation trench 1 9.1.5.1 A single moderately abraded body sherd (2g) was found in the humic soil (context 151) overlying the sand. This is made from the same fine flint tempered ware (FS/1) as one of the early Neolithic carinated bowls. 9.1.5.2 An additional nine sherds of pottery (17g) were recovered from scoop 50. These are in poor condition and are almost certainly shattered fragments from a single sherd. The grog tempered fabric is most typical of the early Bronze Age (GS/1), while the wall thickness (11mm) points to its derivation from an urn, rather than a beaker. 9.1.6 The fabric descriptions 9.1.6.1 The prehistoric assemblage includes eight distinctive fabrics, each identified by a unique code. This is composed of the initial letters of the principal inclusion types, followed by a number which distinguishes between fabrics sharing the same inclusions, but in contrasting frequencies and/or size ranges. See Appendix 2. All of the wares are soft and the fractures of most are hackly, apart from one of the sandy fabrics (S/2) and the shell filled ware (VS/1) which are laminated. 6

9.1.7 The Significance of the ceramic assemblage 9.1.7.1 The early date from the grave is entirely consistent with the character of the pottery. Plain carinated bowls have long been recognised as typical components of assemblages described variously as part of the Grimston/Lyles Hill series (Smith 1974) or Eastern style (Whittle 1977). Initial appraisals of the radiocarbon chronology suggested that this stylistic repertoire emerged towards the beginning of the Neolithic and continued largely unchanged over a period of approximately two millennia (Green and Sofranoff 1985; Smith 1974; Wainwright 1972; and Whittle 1977). 9.1.7.2 A more recent critique of the dates, coupled with a reconsideration of depositional context, has cast serious doubt on the anomalous evidence for the survival of Grimston/Lyles Hill assemblages into the late Neolithic (Hearne 1988). This same research has led to the identification of a distinctive class of undecorated carinated bowls characterized by fine fabrics and surface finishes; open and shallow, bipartite profiles; marked carinations; and by simple everted and beaded rims (Hearne 1988). These vessels have a widespread distribution across Britain and Ireland, tend to occur either singly or in small groups and are associated with radiocarbon dates clustered towards the beginning of the early Neolithic, suggesting chronological priority over the middle Neolithic decorated assemblages (Hearne 1988). 9.1.7.3 The diagnostic ceramics from Yabsley Street display all the defining attributes of these early vessels. Further corroboration is provided by the direct association between the carinated bowl rim, apparently a placed mortuary deposit, and the late 5th to early 4th millennium BC radiocarbon date, which complements the chronological evidence from other sites (cf. Hearne 1988). 9.1.7.4 The more fragmentary pottery from the grave is made exclusively from fine fabrics and the majority of better preserved sherds are thin-walled and burnished. Although there is no evidence for the form of the vessels from which this latter material is derived, the attributes are also consistent with a date early in the Neolithic. Unlike the larger rim fragment, these sherds might have been incorporated accidentally in the grave fill from the surrounding sand, perhaps pointing to an earlier phase of activity in the area preceding the burial. 9.1.7.5 The rim from the grave may be part of the vessel found nearby on the sand. Even if this is not the case, the similarities between the two bowls are so marked that they are most likely to have been the work of a single potter. This suggests that the more complete vessel was placed adjacent to the grave during the mortuary ritual, or after a short interval in commemoration of that event. The scattered distribution of sherds across the sand which may be from this bowl might reflect deliberate breakage or could equally have been dispersed by natural agencies. 9.1.7.6 The character of the rest of the assemblage from the sand cannot be used to establish the likely duration of activity. Eighty-three percent of the sherds in identifiable wares are either made from fabrics used for carinated bowls or represented in the grave. Some could have been in place prior to the burial and all have an early Neolithic origin. The remaining coarser wares could easily be contemporary, but this must remain uncertain. 9.1.7.7 In general terms, the deposit at Yabsley Street recalls the evidence from other sites. While carinated bowls do not appear to have been restricted to a particular type of context, they do occur in mortuary settings (Hearne 1988) and in waterlogged places including rivers (Thomas 1991). While it is very unusual to find such a large portion of an early Neolithic vessel outside of a negative feature in a context likely to represent a contemporary land surface, carinated bowl sherds do occur in the buried soils below some southern long barrows (Hearne 1988). Such evidence may well afford a rare glimpse of practices which generally leave few surviving traces. 9.1.7.8 Full analysis has now been completed on the pottery therefore no further work is required. This report will form the basis of the publication text. Suitable sherds have been illustrated (Fig. 7). 9.2 Fired clay by Steve Ford 9.1.1 Eighteen fragments of fired clay (155g) with a sandy fabric were recovered from context 50. Some of the pieces appear too large to be a product of accidental firing and may therefore be parts of an object such as a loomweight, though no distinctive features such as perforations were present. Two other fragments (38g) came from the surface of the sand. 9.1.2 No further work is required 7

9.3 Struck Flint by Steve Ford 9.3.1 A small collection comprising just 103 struck flints was recovered during the course of the excavation and previous evaluation. Appendix 3 summarizes the composition of the collection as a whole. From the presence of remaining cortex, all pieces appear to have been derived from the local gravel. It is possible though that the larger pieces such as the polished flint axe may have been made on less flawed material derived direct from a chalkland source. A number of the pieces show a degree of patination. 9.3.2 The majority of the finds were not recovered from any cut features or other deposits. The finds were most commonly encountered on the surface of the natural sand, embedded within the natural sand, or within the humic level overlying the sand. A core was recovered from the peat to the east of the main area of interest. The grave (100) produced seven struck flints comprising 6 flakes, a spall and a knife. Three of these pieces located close together on the eastern side of the grave were burnt. The knife was located at the foot (southern end) of the grave (Fig. 3). 9.3.3 The purpose of the following analysis is to characterise the nature of the flint assemblages to define both their chronological attributes and to determine the broad range of flint using activities that may have taken place. Regrettably the collection is neither large enough nor securely stratified to merit detailed metrical analysis but some comment is in order. 9.3.4 Flakes and blades 9.3.4.1 The 69 flakes and 9 blades (narrow flakes) recovered were assigned by eye. It is clear that the proportion of narrow flakes is relatively high suggesting a significant component of mesolithic, or more likely earlier neolithic material within the collection. 9.3.5 Retouched pieces 9.3.5.1 Ten pieces are considered as being deliberately retouched and can be defined as formal tool types. A further 4 of the flakes and blades are possibly retouched and/or presenting utilisation damage. As this interpretation was not clearcut, these pieces are not differentiated from the other flakes and blades. Whilst the collection total is modest, the proportion of retouched material at about 11% (excluding spalls) is high (cf Healey and Robertson-Mackay 1983, 21). In contrast the number of spalls (9) is low. These three attributes suggest that use of material, probably in a domestic setting, is the dominant activity rather than procurement or production of material. The latter activities generate significant quantities of both usable and unusable material including spalls. 9.3.6 Knife 9.3.6.1 The single knife from the site was recovered from the foot of the grave (100, 152), (Fig. 3). The knife has a triangular cross section with the apex towards the left hand edge. This edge has been heavily retouched, and almost bruised. The other edge has been invasively retouched, presumably using a soft hammer or pressure flaking in part. The latter retouch is present on both sides of the knife but is not coincident on either side. Rather it appear that the retouch has been used to straighten the edge. The retouch is of two distinct periods as lightly patinated surfaces have been flaked. It either represents the use of an old flake or possibly the reshaping of an old knife. There is a very slight trace of a use gloss on the cutting edge. 9.3.7 Retouched flakes 9.3.7.1 Three flakes had been retouched but were not shaped into formal tool types. 9.3.8 Arrowheads 9.3.8.1 A petit tranchet derivative arrowhead was recovered. It is a distinctive type most frequently encountered in either Mesolithic or later Neolithic contexts, but which does on occur in Early Neolithic assemblages (Green 1980). Its presence here is not out of context with the Early Neolithic material forming the bulk of the evidence but might suggest some other activity in addition. This has already been confirmed from the association of feature 50 which is associated with EBA pottery. Some of the other flintwork, therefore could also belong to this later period though it does need to be remembered that arrowheads are distributed widely across the landscape when lost on hunting expeditions. 9.3.8.2 A second fragmentary flint is invasively retouched such that it might also have originally been an arrowhead but it is now not possible to determine type. 9.3.9 Polished flint axe 8

9.3.9.1 A small broken polished flint axehead was recovered. The piece was narrow butted and made from a grey flint with large cherty inclusions. The piece appears to have been used as an axe as the blade in plan is asymmetrical but with a symmetrical profile. The manufacturing flake scars had not been fully polished away. Following its breakage, the piece had been used as a core as it had been reflaked from the break. 9.3.10 Discussion 9.3.10.1 The analysis of the flint assemblages from the site has provided three components with which to understand some aspects of the development of the site. 9.3.10.2 The dating of the collection, although mostly not stratified as a closed group, appears to be homogenous and is entirely in keeping with the associated Early Neolithic pottery. Perhaps inevitably, one or two pieces are possibly of earlier or later date, but there are no items present to suggest that this observation reflects more than the necessary caution when considering the integrity of assemblages of durable flint. 9.3.10.3 The collection as a whole appears to reflect a use assemblage, that is, one that does not include procurement or manufacture to any great extent. The latter activities usually produce much larger quantities of material which includes significant proportions of unused and unusable flakes, large numbers of cortical flakes and spalls, micro debitage, and if the flint source is poor and frost damaged, much irregular workshop waste. The proportions of retouched material would also be low. The collection here has produced only a modest quantity of material overall, does not include significant proportion of knapping debris and has a high retouched component. 9.3.10.4 A small number of assemblages of various periods have been documented where the flintwork repertoire is dominated by distinctive tool types indicating some form of specialist activity. Such sites are exemplified by the assemblage at the late Neolithic site at Rackham, Sussex (Holden and Bradley 1975) where the numbers of scrapers is exceptional, or at the early Neolithic site at Bishopstone, Sussex (Bell 1977) where serrated flakes are well represented in several of the pit groups. The collection here whilst containing a relatively high proportion of retouched material, is not dominated by specific types and thus one can conclude that it originates from non-specific general occupation-related activities. 9.3.10.5 Full analysis has been carried out on the struck flint, therefore no further work is required. Suitable pieces have already been illustrated for publication (Fig. 8) 9.4 Burnt flint by Steve Ford 9.4.1 A modest quantity of burnt flint was recovered during the fieldwork. This material was scattered across the excavated areas with no marked concentrations. Some 15 of the struck flints had also been burnt. 12 fragments (9g) came from the fill (153) of the grave (100). 9.4.2 No further work is required 9.5 Human remains by Sian Anthony 9.5.1 The single skeleton was lifted in blocks. During transportation the blocks were disturbed and cracked despite being well supported. The blocks comprised the skull, vertebrae, right humerus, pelvis and left hand, lower legs and feet. During excavation the position of the vertebrae and arms were misinterpreted. Each block was carefully lifted and excavated using wooden tools and de-ionised water occasionally to avoid fragmentation of the soil blocks. The residue from each block was sieved using a 2mm mesh. The blocks were cleaned and photographed during the process. Upon examination the blocks containing the vertebrae and right humerus contained only darkly stained sand, no bone was remaining. The remains were identified and recorded according to Buikstra and Ubelaker (1994). 9.5.2 The preservation is very poor with mostly friable elements, much was degraded, black and soft which crumbled when exposed, no cancellous bone survived, elements directly on the ground surface (the left side of the body) did not survive. Pieces of the right side of the skull were in slightly better condition with pieces of the right orbit, zygomatic and zygomaxillare identifiable. 9.5.3 Sexing indicators are minimal and so do not lead to a secure identification, part of the glabella and orbital margins are present and although the occipital was much degraded the original shape could be ascertained as being gracile and smooth. Sex indicators are limited to just three the brow ridge, one 9

orbit and the occipital area which all suggest a female. This may supported by the overall gracile nature of the body however this is not indisputable and no measurements could be taken to support this. 9.5.4 The general size of the skeletal elements (and the affected black coloured sand) suggests an individual who had reached skeletal maturity rather than a juvenile, the only ageing indicator is a small fragment of saggital suture that is open. There are no surviving epiphyses or pathological changes to indicate an older date, so an approximate estimation of age would be adult, possibly a young adult. The fragmented nature of the surviving elements hinders any assessment of health status, pathologies or injuries. 9.5.5 Analysis of the bone was undertaken by Dr Mike Richards, University of Bradford for surviving collagen for stable isotope analysis but two attempts were unsuccessful in retrieving any collagen. 9.5.6 No DNA recovery can be attempted as the skeleton is too degraded for analysis. Acidic conditions tend to reduce the limitations of DNA retrieval (Brown 2000). 9.5.7 Full analysis has been carried out on the human remains and no further work is required (or possible). 9.6 Wood by Rowena Gale 9.6.1 A plank of wood 0.6m long, 0.12m wide and 0.05m thick associated with the burial was examined and identified to species. The wood was firm but liable to fragment easily. A small piece was removed and prepared for examination using standard methods (Gale and Cutler 2000). Thin sections were examined using transmitted light on a Nikon Labophoto 2 microscope at magnifications up to x400. The anatomical structure was matched to reference slides of modern wood. The species was oak (Quercus sp.). 9.6.2 The wood examined suggested that the plank had been converted (split) tangentially. It was not possible to assess the age of the tree. Tyloses were absent from the spring vessels in the wood examined suggesting an origin from sapwood. No areas of charring were observed. 9.6.3 No further work is required 9.7 Charred plant remains by Mark Robinson 9.7.1 Just two archaeological features were revealed during the excavation both of which produced charred plant remains. Bulk samples were taken from them and floated onto a 0.25 mm mesh to recover charred plant remains. All that was found in an 18 litre sample from Context 50 (Early Bronze Age hearth) was charcoal of Quercus sp. (oak). Seven samples totalling 155 litres were floated from various localities within the fill of the early Neolithic grave (100, 153) and charred remains in addition to oak were found as detailed in Appendix 4. 9.7.2 The remains either represent items amongst kindling used to ignite the oak wood or were burnt foodprocessing remains. The latter is thought more likely because of the presence of hazel-nutshell fragments and the absence of hawthorn-type twigs alongside the hawthorn stones. 9.7.3 Assemblages of Neolithic food plant remains from Britain often include both cereals and the remains of wild, edible plants. Indeed, fragments of charred hazel nutshell usually greatly outnumber charred cereal grains on Neolithic sites (Moffett et al. 1989; Robinson 2000). The assemblage from Yabsley Street is slightly atypical in that the most numerous remains of edible wild plants were stones of hawthorn berries, but they have been found elsewhere. Mallow seeds are also edible and it is possible that they too had been collected. The discovery of a glume of Triticum dicoccum (emmer wheat) is of interest because records of Neolithic cereal chaff are rare, although grain of emmer is known from other early Neolithic sites in Britain. 9.8 Pollen Analysis by Michael Keith-Lucas 9.8.1 A provisional pollen diagram has been presented for the peat deposits sampled during the evaluation (column 1) which lies just to the east of the excavation area (Fig. 9). The most notable feature is the peak of cereals at 90cm which rises to over 10% of the count. Plantago lanceolata also starts at the 90cm level. Other features of note are a decline in alder and a rise in pine and a lime decline at 80cm. 9.8.2 Four samples have been submitted for radiocarbon dating from 113cm (basal deposits), 100cm, 90cm 10

(cereal peak) and 80cm and which are expected to bracket the Neolithic occupation of the adjacent terrace. 9.9 Radiocarbon Dating 9.9.1 A single sample of wood from within the grave was submitted to the University of Kiel for radiocarbon dating. Details of the methodology are in the archive. Calibrated age is given according to CALIB rev 4.3 (Stuiver et al. 1998). In summary the results were: KIA20157 Oak wood (155) partly lining grave 100 Radiocarbon Age BP 5252 ± 28 Calibrated Ages cal BC 4041, 4015, 4004 One sigma cal BC 4217 4201; 4156 4153; 4139 4130; 4047 4035; 4025 3986 Two sigma cal BC 4220 4196; 4162 4122; 4109 4096; 4077 4057; 4053 3979 9.9.2 Further dates from the pollen analysis column are awaited. 10 Conclusion 10.1 The results of this small excavation are noteworthy for several reasons; these can be summarized as the nature of earlier Neolithic occupation, the nature of earlier Neolithic settlement and environment in the lower Thames Valley, and the further evidence for early Neolithic burial practice. 10.2 The Thames Valley is one of the major topographic zones for evidence of earlier Neolithic settlement within the British Isles. The middle and lower sections of the valley are well represented by monumental sites as at Orsett, Essex and Yeoveney Lodge, Staines (Hedges and Buckley 1978; Robertson-Mackay 1987), stray finds of flint and stone axes (Chappell 1987) and finds of pottery and leaf-shaped arrowheads (MoLAS 2000, 68). Much less evidence is recorded which can be readily interpreted as representing occupation with sites such as Runnymede Bridge being exceptional (Needham and Stig-Sorensen 1988). Groups of pits, isolated pits and flint scatters are the norm (Ford et al. 2003, 153; Holgate 1988; Holgate and Start 1985) but are still uncommon (MoLAS 2000, 68 and map 3; Siddell et al. 2002). More recently, indirect evidence from alluvial settings has been obtained with a later Neolithic wooden trackway being located and pollen analysis indicating cereal cultivation at Silvertown, London (Crocket et al. 2002). The Yabsley Street finds therefore would be noteworthy if only for their rarity value, but it is also the context and chronology which are of interest. 10.3 The radiocarbon determination obtained from the grave places the use of the site very early within the onset of the British Neolithic and to this effect, several of the distinctive components of the Neolithic background highlighted above post-date this occupation. In particular, the monuments, and deposits associated with decorated pottery styles, several of which have radiocarbon determinations, can be considered as belonging to a mature, established phase of the Neolithic; the trackway and pollen evidence for cereal cultivation at Silvertown, which lies only 2.5km to the east within a similar stratigraphic setting, reflects much later C14 dated activity (Crockett et al. 2002). For this period, to put the site within a wider context, late Mesolithic use of the Thames margins also requires consideration. The close association of Mesolithic occupation and riparian locations in the major river valleys has long been well documented (Wymer 1977). Mesolithic occupation has also recently been documented for locations on the south side of the Thames in Southwark and Lambeth (Siddell et al. 2002). Yet, as for the earliest Neolithic, documented sites of late Mesolithic date are few (Siddell et al. 2002, 80) and there is nothing to suggest even an approximate contemporaenity with the Yabsley Street Neolithic site. 10.4 Discussion of the social and chronological context of the site therefore lies within the framework of the much debated Mesolithic-Neolithic transition and the introduction of both the physical and conceptual components of the Neolithic from Continental Europe. It is some years now since the origins of the British Neolithic have been debated with vigour. Colonization by advanced agriculturalists as the mechanism for cultural change (an invasion hypothesis) is now regarded as a much over-simplified explanation. The process of this change is considered to be much more complex and the distinction between early farmers and hunter/gatherer/fishers is not nearly so clear cut. Recent reviews of the economic evidence for the Neolithic have placed a much greater emphasis on wild plant remains than agricultural produce (Moffett et al. 1989; Robinson 2000). The technological limitations of cross channel contact are also likely to be of much less importance than previously considered (cf. Case 11

1969). A greater emphasis on the adoption by indigenous communities of a Neolithic social package rather than pure economic necessity has much to commend it (Thomas 1988). 10.5 The location of the site on the very foreshore of the Thames (or one channel of a braided system) appears to support the link between the earlier Neolithic and riverine resources and that the fishing/gathering/hunting component of the subsistence economy is likely to be significant. It is regrettable that the preservation of the burial deposit was so poor that isotope analysis to determine dietary preferences could not be carried out for that individual and that there was no preservation of other faunal remains. The evidence from analysis of the small numbers of charred plant remains from within the grave though, indicates both wild and cultivated plants with hazelnut and hawthorn pips accompanied by a little cereal, and chaff from emmer wheat. In this respect also, the evidence from the pollen analysis of the adjacent peat deposits may be of particular significance once the chronology of the significant events in the profile are dated. It seems mostly likely from the pollen diagram that cereal growing is taking place. The relatively high representation of this heavy pollen which is poorly dispersed by wind is especially noteworthy and taken at face value strongly suggests that arable farming is occurring on land in the near vicinity. However, it has to be considered possible that such a significant concentration is the product of a localized dump of cereal ears or that natural grasses which produce large pollen grains are locally abundant. It remains therefore to be seen if this event does reflect intensive local cereal growing and relates in time to the occupation of the site, or to a later period. 10.6 The discovery of the burial deposit is also noteworthy as individual inhumations of this period, without any accompanying monumental structures are rare as most recorded burial is collective within long barrows or chambered tombs (Bristow 1998). Rather more individual burials are recorded within circular or oval monuments (Kinnes 1979). In the Thames Valley two examples of individual burial are recorded. From within the interior of the causewayed enclosure at Yeoveney Lodge, Staines an oval pit contained a female, around 35 years old lying on her left side and flexed. The individual was loosely arranged with no grave goods. No dating evidence was recovered (Robertson-Mackay 1987). A single inhumation of an older female with a complete Abingdon Ware vessel with antler and animal bone in a flat grave was also recorded at Pangbourne, Berkshire (Piggott 1929). A more distantly located burial is of note because of its location on the beach in a similar setting to Yapsley Street. A grave was found at Croyde, Devon associated with struck flints, which was thought to be of Neolithic date (Young 1908). Other examples of certain and probable Neolithic date in the South East of England are recorded from Bedfordshire, Surrey, Sussex and Kent (Bristow 1998). 10.7 Continental examples are more common. For example the Aisne Valley in the Paris Basin has a larger variety (Hachem et al. 1998). Menneville, a Linienbandkeramik site contained 30 inhumations divided into three locations in the site: outside the enclosure, in the settlement by structures and in the enclosure ditch. The location outside of the enclosure is more comparable to the British examples with seven single burials flexed and ochred with the head at east and facing south. Grave goods included pottery, flint and shell. 10.8 The site is thus an important addition to current knowledge and deserves to be published in full. 11 Updated Project Design 11.1 Production of this assessment report has already included completion of all of the analytical work necessary to produce the publication report. Remaining work required is limited to incorporation of the radicarbon dates for the pollen sequence (analysis of which is already underway) and the preparation of a suitable publication text. 12 Proposals for Publication 12.1 The site should be published as a short article in an academic journal such as Proceedings of the Prehistoric Society. Publication will focus on the identified important aspects of the site. An outline publication synopsis forms Appendix 5. 12.2 A summary note will also be submitted to London Archaeologist. 12