Imperial College Sports Grounds and RMC Land, Harlington: The development of prehistoric and later communities in the Colne Valley and on the Heathrow Terraces

By Andrew B. Powell, Alistair Barclay, Lorraine Mepham and Chris J. Stevens

This volume brings together the results from the excavations at the former Imperial College Sports Ground, RMC Land and Land East of Wall Garden Farm, near the villages of Harlington and Sipson in the London Borough of Hillingdon. The excavations revealed parts of an archaeological landscape with a rich history of development from before 4000 BC to the post-medieval period. The opportunity to investigate two large areas of this landscape provided evidence for possible settlement continuity and shift over a period of 6000 years. Early to Middle Neolithic occupation was represented by a rectangular ditched mortuary enclosure and a large spread of pits, many containing deposits of Peterborough Ware pottery, flint and charred plant remains. A possible dispersed monument complex of three hengiform enclosures was associated with the rare remains of cremation burials radiocarbon dated to the Middle Neolithic. Limited Late Neolithic and Early Bronze Age activity was identified, which is in stark contrast to the Middle to Late Bronze Age when a formalized landscape of extensive rectangular fields, enclosures, wells and pits was established. This major reorganized land division can be traced across the two sites and over large parts of the adjacent Heathrow terraces. A small, Iron Age and Romano-British nucleated settlement was constructed, with associated enclosures flanking a trackway. There were wayside inhumations, cremation burials and middens and more widely dispersed wells and quarries. Two possible sunken-featured buildings of early Saxon date were found. There was also a small cemetery. Subsequently, a middle Saxon and medieval field system of small enclosures and wells was established.

• Language: English
• Publisher: Wessex Archaeology
• Release date: Feb 16, 2016
• ISBN: 9781874350750

Book preview

Chapter 1

Introduction

by Alistair J. Barclay, Andrew B. Powell, Chris J. Stevens and Philippa Bradley

Between 1996 and 2009, programmes of excavation were undertaken by Wessex Archaeology on two large blocks of land proposed for mineral extraction to the north of Heathrow Airport, lying between the villages of Harlington and Sipson in the London Borough of Hillingdon (Fig. 1.1). The southern block, the former Imperial College Sports Ground (ICSG), covered 23.6 ha centred on NGR 50800 17770. The northern block comprised two sites covering 13.3 ha centred on NGR 50840 17825, one on land formerly owned by Ready Mixed Concrete Ltd (RMC Land), the other, to its immediate east, referred to as Land East of Wall Garden Farm (LEWGF) (Fig. 1.2).

The excavations uncovered evidence for Neolithic, Bronze Age, Iron Age, Romano-British, Saxon and medieval activity, adding significant new information to our understanding of the development of what previous extensive excavations in the area have shown to be the archaeologically rich landscape of the Middle Thames Valley.

Background to the Project

ICSG

In August 1990, the Minerals Planning Authority (MPA), the London Borough of Hillingdon, granted Henry Streeter (Sand and Ballast) Ltd (HSL) conditional planning permission for mineral extraction in the central and eastern parts of the site (then referred to as ICSG East) (Fig. 1.2). One of the conditions of the planning permission was that the results of an archaeological evaluation, including a mitigation strategy, be submitted to and approved by the MPA.

The evaluation (site code IMP 96), comprising 97 trenches, was carried out by the Museum of London Archaeology Service (MoLAS) in 1996 and revealed a range of archaeological features concentrated towards the east of the site, spanning the Neolithic to the late Romano-British period, including enclosures, cremation burials, ditches, pits and postholes (MoLAS 1996). Subsequently, the MPA also granted conditional permission for extraction in a westward extension to the quarry (ICSG West), resulting in a further archaeological evaluation comprising 20 trenches, undertaken by Wessex Archaeology in 1999 (Wessex Archaeology 1999).

On the basis of the results of the evaluations, the MPA required the full archaeological excavation of the quarry area (site code IMC 96). This work was undertaken between September 1996 and July 2001 (Wessex Archaeology 2004a). Interim reports on the results from Phases 1–5 (see Methods, below) were published in The London Archaeologist (Crockett 2001; 2002).

RMC Land

RMC Land is operated by HSL in partnership with CEMEX UK Materials Ltd. In May 2002 the MPA granted them planning permission for gravel extraction at the site, conditional upon the implementation of a programme of archaeological work in accordance with a Written Scheme of Investigation approved by the MPA and their archaeological advisor English Heritage (Greater London Archaeological Advisory Service – GLAAS).

An archaeological evaluation of the northern part of the site, comprising 53 trenches, was undertaken in September 2000, revealing features of Neolithic, Bronze Age, Saxon and medieval date (Wessex Archaeology 2001) (Fig. 1.2). The southern part of the site could not be evaluated, but was considered to have similar archaeological potential. On the basis of the evaluation the MPA required that a full programme of archaeological excavation be undertaken (site code SIE 00). This was carried out between August 2002 and September 2006 (Wessex Archaeology 2003; 2005; 2007a).

Land East of Wall Garden Farm

In December 2008, HSL was granted planning permission for mineral extraction on Land East of Wall Garden Farm (LEWGF), immediately to the east of RMC Land. This followed an evaluation in 2007, comprising a further 10 trenches (Wessex Archaeology 2007b) (site code WGA 07) (Fig. 1.2). The excavation was carried out between June and September 2009 (Wessex Archaeology 2009a). For the sake of simplicity, this site is treated below as part of RMC Land, and is shown as such in the figures.

Figure 1.1 Locations of sites and other archaeological investigations in the Heathrow area and the Middle Thames Valley

Figure 1.2 Imperial College Sports Ground (ICSG), RMC Land and LEWGF, showing areas of archaeological investigations, area notation used in the report and adjacent archaeological sites

Site Location, Topography and Geology

ICSG is bounded to the north by Sipson Lane, to the south and south-east by the village of Harlington, to the south-west by open fields, and to the west by the M4 Heathrow Spur. RMC Land, which lies 250 m to the north of ICSG, is bounded to the north by the M4, to the south by a sports ground and Victoria Lane, and to the west by open fields (Fig. 1.2).

The sites are situated on fairly level topography, between 25.5–26.0 m above Ordnance Datum (aOD) at ICSG and 26.5 m aOD at RMC Land, approximately midway between two south-flowing tributaries of the River Thames – the River Colne to the west and the River Crane to the east (Fig. 1.1). They are situated on the north side of the Middle Thames Valley, which in this area comprises a series of broad, flat gravel terraces stepping down gently from north to south, generally referred to as the Heathrow Terrace.

The basal solid geology is London Clay. The underlying drift geology is Taplow Gravel (189,000–128,000 BP, Oxygen Isotope stage 6: Bridgland 1994), one of the Pleistocene gravel terraces formed through a series of erosional and depositional episodes associated with the post-diversionary phase of the River Thames (BGS 1981). Overlying the gravel is a deposit of yellowish brown silty sand identified as the Langley Silt Complex (Brickearth), a complex deposit of probable Late Devensian date (19,000–13,000 BP) derived from a combination of wind-borne and water-borne deposition (Rose 1999, 56).

Methods

Excavation Areas and Context Numbering

At ICSG, the archaeological works were undertaken from east to west, from each of the six areas of the site to be quarried (previously referred to as Phases 1–6, following the extraction phases). At RMC Land, the work proceeded from west to east across the five extraction areas (Phases 1–5). During the excavations at both sites, however, these areas were further subdivided – into 12 blocks at ICSG (Phases 1, 2 (three blocks), 3 (two blocks), 4a (three blocks) and 4b, 5, and 6 (four blocks); and nine blocks at RMC Land (Phases 1, 2, 3 (north, middle and south – five blocks), 4 (middle-west and south – two blocks) and 5 (north and south – two blocks)). Wall Garden Farm (LEWGF) to the east formed the tenth block (Fig. 1.2a).

In order to simplify references in the text to the different parts of the two sites, the sites have been broken down into Areas – five at ICSG (Areas A–E) and four at RMC Land/LEWGF (Areas 1–4) (Fig. 1.2b), which correspond broadly to the earlier divisions but which are easier to follow (Table 1.1).

Table 1.1 Concordance of site Areas referred to in the text, and previous extraction/excavation Phases

Because ICSG and RMC Land were initially unrelated projects, each was recorded as a separate entity with its own context numbering sequence. Consequently, there is some duplication of numbers between the two sites, although care has been taken to ensure that it is clear in the text to which site any context relates. The context sequence for LEWGF is a continuation of that for RMC Land. At ICSG, the group number sequence duplicates the context number sequence; consequently group numbers from that site (but not RMC Land) have a ‘G’ prefix. Contexts from the evaluation phases of both sites have an ‘EV’ prefix.

Field Methods

At both sites, the topsoil and subsoil overburden were removed by HSL using 360° tracked excavators under archaeological supervision to the surface of undisturbed geological deposits or the level at which archaeological features could be identified. Generally this level equated to the surface of the undisturbed brickearth 0.5 m below modern ground surface.

The site and archaeological features were tied to the Ordnance Survey National Grid initially using nearby available OS triangulation points, and in the later phases of work using a GPS unit. All survey, plan and contour data was collected using an on-site Total Station, for production of digitised mapping and plotting via AutoCAD.

All archaeological features and deposits were recorded using Wessex Archaeology’s pro forma recording system. All site plans were drawn at a minimum scale of 1:100, detail plans at 1:20, and sections at 1:10 (Pl. 1.1). A full photographic record was maintained using colour transparencies, black and white negatives (on 35 mm film) and digital format.

Plate 1.1 Land East of Wall Garden Farm during excavation

Although dependant on many factors, the percentage of archaeological features to be excavated averaged 10% of all linear features (ie, ditches, gullies etc.) and 50% of discrete features (ie, pits, postholes etc.). In general, these percentages were considered as a minimum response, with a more detailed investigation undertaken of significant deposits (ie, structures, burials etc.). However, following discussions with English Heritage, the 10% level of formal excavation (ie, scaled plans and sections, detailed photographs and comprehensive context recording) was reduced at ICSG for linear features considered to represent field boundaries or other such features beyond settlement centres. This reduction was offset by an increase in rapid excavation of narrow slots through such features to chart artefact (and where possible ecofact) assemblage compositions and distributions. As a minimum response, each rapid ‘slot’ was uniquely identified and 3D recorded.

Targeted environmental sampling strategies were employed, comprising bulk samples of up to 30 litres from most sealed and dated deposits, ensuring that an appropriate range of feature types was sampled for each period. Additional samples were also taken from either sealed or dated deposits to establish the nature and/or date of such deposits. Where appropriate, soil monoliths, mollusc columns, artefact samples and magnetic susceptibility readings were also obtained.

As each phase of each excavation progressed, land parcels were ‘released’ to HSL, subject to the agreement of English Heritage. On release, each area was stripped of all remaining brickearth, to the top of the gravel, which was then commercially extracted. An intermittent monitoring of Pleistocene deposits was undertaken during the gravel extraction from ICSG Area A, in order to assess the potential for Palaeolithic material within the gravels and the brickearth/gravel interface. This was not repeated in subsequent phases, or at RMC Land.

Topographic Analysis

Topographic analysis of approximately 20 square kilometres of the Heathrow landscape was undertaken by Framework Archaeology (2006) using the results of a detailed contour survey, conducted in the closing stages of World War II before most of the area was developed. The survey of the site of the future London Airport, undertaken by Italian prisoners of war for the Air Ministry, extended north to Sipson Lane, and so included the ICSG site but not RMC Land.

The survey data were recorded in imperial measurements (feet and inches) on a 1943 Ordnance Survey map, with elevations measurements taken every 30 m. These data were scanned, georeferenced and digitised, and given X, Y and Z coordinates in AutoCad. They were then processed using Surfer to produce a 3D contour map, as well as colour models to highlight minor topographic variations. This confirmed the predominantly flat nature of the land at ICSG, and revealed no traces of the Neolithic monuments recorded during the excavation (below). Given the extensive late prehistoric and medieval field systems on the site it is likely that these had already been plough-levelled prior to post-medieval and modern (pre-World War II) cultivation. The only features on the site revealed by the topographic analysis were modern field boundary ditches.

Archaeological Background

Much of the landscape immediately surrounding the sites has been subject to archaeological intervention during the last twenty years (Fig. 1.1), principally in advance of, and during, gravel extraction. Archaeological sites include Frogs Ditch Farm (FDF 79) to the north, Wall Garden Farm (WGF 79–84 and WGD 95) and Little Harlington Fields (LHF 91) at the north-west, and Cranford Lane (CLH 89, CFL 94) to the east of Harlington.

The land between ICSG and RMC Land was subject to a watching brief during gravel extraction in 1986 (ICSG 86), during which a number of north–south aligned gullies were recorded, along with a scoop containing burnt flint, and a large oval feature containing fragments of wood from its lower fills and possibly Late Bronze Age pottery from its upper fills (MoLAS 1993, 31).

Table 1.2 Main archaeological periods represented

More recently, extensive excavations have been undertaken in advance of development at and around Heathrow Airport, at Perry Oaks (Framework Archaeology 2006) and Terminal 5 (Framework Archaeology 2010), where the recorded deposits include Neolithic monuments, Bronze Age to post-medieval field systems, Iron Age and Romano-British settlements and Bronze Age and Romano-British inhumation and cremation burials.

Together these investigations have revealed an archaeologically rich landscape, with evidence for nearly continuous occupation and settlement from at least the Neolithic to the medieval period (Table 1.2). The archaeology of the Heathrow area is also considered in various papers in Cotton and Field (2004) (in particular Lewis and Welsh 2004; Bradley 2004; Cotton 2004; Jones and Ayres 2004).

Pre-Neolithic

The earliest archaeology from the immediate area is of Lower and Middle Palaeolithic date, and comprises mostly artefacts recorded as part of The English Rivers Palaeolithic Survey (TERPS) (Wessex Archaeology 1996a; Wymer 1999). The majority of these objects have been recovered from the Lynch Hill Gravel to the north of the two sites (eg, 426 hand-axes were recovered from Bowyers Pit, Hillingdon off Stockley Road). A much smaller number, in a more rolled condition, have been recovered from the Taplow Gravel that underlies the sites, perhaps indicating that this deposit is reworked Lynch Hill Gravel (Phil Harding pers. comm.). Two Middle Palaeolithic find spots on the Taplow Terrace in the vicinity of the sites, comprising two hand axes and two Levallois flakes, are likely to be derived from the Langley Silts and the Lynch Hill Terrace, where in situ Levallois flints have been recorded (eg, at Creffield Road). Much later Late Upper Palaeolithic material, comprising in situ flint scatters and animal bones, was excavated within Colne Valley Silts at Three Ways Wharf, Uxbridge, some 8 km to the north-west of the site (Lewis 1991; Lewis with Rackham 2011). Investigations at Kingsmead Quarry, Horton have produced a redeposited collection of similar material (Chaffey et al. forthcoming).

In situ Mesolithic finds have been recorded mainly in association with river valley silts, such as the scatter of flints and animal bones representing a probable butchery site close to the Upper Palaeolithic occupation at Three Ways Wharf (Lewis 1991; Lewis with Rackham 2011). It is likely that there was comparable activity at other locations along the valleys of the Colne and Crane. Recent work at Perry Oaks suggests Mesolithic activity pre-dating the Stanwell bank barrow (or cursus) (Framework Archaeology 2005; 2006).

Neolithic to Early Bronze Age

The initial phase of the Neolithic (c. 4000–3650 BC) is well represented in the surrounding area and includes the remains of timber long-houses at Kingsmead Quarry, Horton and Cranford Lane (Chaffey and Brook 2012; Nick Elsden pers. comm.) and various pit and midden deposits from sites along the adjacent stretch of the River Thames (eg, the Eton Rowing Course and Cannon Hill, Maidenhead: Allen et al. 2004; Allen et al. 2013; Bradley et al. 1975–6).

During the Neolithic a number of substantial monuments were constructed in this part of the Middle Thames Valley, identifying this area as a major centre of ceremonial and ritual activity to rival other complexes found upriver at Dorchester-on-Thames and Avebury (Whittle et al. 1992; Whittle 1993). Four causewayed enclosures are known from the surrounding area; the nearest was excavated at Yeoveney Lodge, Staines, and another possible one has been identified at East Bedfont (Robertson-Mackay 1987; Oswald et al. 2001, 112 and 152), while two further examples are known from near Eton (Allen et al. 2004). A number of U-shaped enclosures and an oval barrow are known from the Colne Valley terraces (eg, Perry Oaks and Horton: Framework Archaeology 2005; 2006; Preston 2003). At Staines Road, Shepperton a small penannular ditched enclosure was associated with human and animal remains, worked flint, redeposited Carinated Bowl and Mortlake ware (Jones 2008).

The most impressive monument is the Stanwell bank barrow, which is aligned approximately north–south and extends for a distance of some 3.5 km along the east side of the Colne Valley, delineating for much of its length the Taplow Terrace/Colne Valley boundary. It has recently been excavated at Perry Oaks, to the west of Heathrow Airport (Fig. 1.1), where it postdates an earlier ‘avenue’ of timber posts (Framework Archaeology 2010, 53–4). Other monuments, interpreted variously as long barrows or subsidiary cursus monuments, have been recorded as cropmarks adjacent to it (Field and Cotton 1987, fig. 4.5), two of which have been partially excavated (Framework Archaeology 2010, 67–9 and fig. 2.23).

While such sites imply Neolithic settlement in the area, direct evidence for settlement and domestic activity is less easy to identify. The strongest indications come from Runnymede Bridge to the south-west, where flint and pottery characteristic of Neolithic domestic settlements were associated with hearths and post-built structures (Stuart Needham pers. comm.) and from Kingsmead Quarry, Horton and Cranford Lane (see above).

Isolated pits, many containing mixed assemblages of diagnostic finds, such as flint tools, waste flakes and fragments of stone axes, are relatively common in the area (Holgate 1988, map 17). Many contained Peterborough Ware (3500–2850 BC), but have tended to occur in small numbers, as at Heathrow (Grimes 1960) or Sipson Lane, Harmondsworth (Cotton et al. 1986, 36). Grooved Ware (2900–2400 BC) is locally, and to some extent regionally, much rarer (Barclay 1999). The main finds have been at Prospect Park, near Harmondsworth to the west (Laidlaw and Mepham 1996), Holloway Lane in the same parish (Cotton et al. 1986, 37) and from the Framework Archaeology excavations at Heathrow (Matt Leivers pers. comm.; Framework Archaeology 2010, 40).

There is other less tangible evidence for human habitation of the area, including river finds (axe and mace-heads, pottery, animal and human bone), surface scatters of lithic material and stray finds (eg, stone axes) (Holgate 1988; Allen et al. 2004; Barclay 2011).

Early Bronze Age remains, including Beaker pottery (c. 2450–1700 BC), are scarce in the area, with the notable exception of a large pit containing the dismembered body of an aurochs (wild ox), associated with six barbed and tanged flint arrowheads, on Holloway Lane, Harmondsworth (Brown and Cotton 2000, 86). Another significant find is that of a flat axe of bronze from Harlington (Cotton et al. 1986, fig. 30). Elsewhere scatters of pottery and flintwork have been found preserved beneath alluvium at Runnymede (Needham and Spence 1996) and the Eton Rowing Course Project (Allen et al. 2004; Allen et al. 2013). There are also important Beaker finds from the river Thames (Barclay 2011; Jon Cotton pers. comm.).

A number of ring ditches visible as cropmarks may represent the remains of Early Bronze Age round barrows (although some may date to the Neolithic). It is argued that an array of ring ditches running east from the south end of the Stanwell bank barrow may be an Early Bronze Age barrow cemetery (Cotton et al. 1986, fig. 28). Where Early Bronze Age burials occur, they are almost invariably cremations, with little trace of the sometimes elaborately furnished inhumations which characterise the period in much of Britain (MoLAS 2000, 85–6). In the Middle Thames Valley barrows tend to occur either as isolated monuments or in small clusters, which contrasts with the Upper Thames where large cemeteries containing over 20 monuments are quite common (Garwood with Hey and Barclay 2011, fig. 14.24).

Middle–Late Bronze Age and Iron Age

In contrast, the Middle Bronze Age is characterised by evidence for domestic as well as burial activity. Excavations close to the sites within the A4/M4 corridor have revealed pits and postholes containing assemblages of domestic material, including Deverel-Rimbury pottery, daub, loomweights and flints. Contemporary ditches and enclosures have been interpreted as stock enclosures and field boundaries. A number of urned cremation burials have been dated to this period.

There is evidence for a consolidation of settlement in the area during the Late Bronze Age. This includes high status centres like Runnymede Bridge (Needham 1987) and smaller undefended settlements characterised by huts and associated field systems, such as those found at Cranford Lane (Nick Elsden pers. comm.).

The settlement and development of landscape in the Iron Age represents a broad continuation of patterns established in the Late Bronze Age, with both defended enclosures and unenclosed settlements recorded in the area. Undefended settlements, such as the one at Mayfield Farm, comprised roundhouses, pits and field boundaries.

Caesar’s Camp, an earthwork that was surveyed by William Stukeley in 1723 and subsequently excavated by Professor W. F. Grimes in the 1940s (before the construction of Heathrow Runway 1), revealed a defended enclosure and at least 11 Middle Iron Age (400–100 BC) roundhouses. Some of the roundhouses were overlain by the enclosing bank, suggesting that the site had originally been an undefended, and more dispersed, settlement. Within the enclosure there were numerous four-post structures, a characteristic feature of the period thought to represent grain stores and/or driers, as well as a square shrine (Grimes and Close-Brooks 1993). Other similar enclosures have been recorded in the area (eg, Fern Hill and Staines Moor).

Over the years extensive Iron Age settlement has been recorded at Perry Oaks, Terminal 5 and Heathrow (Canham 1978; Framework Archaeology 2005; 2006). There is evidence for the re-use and development of the earlier field system, for the maintenance of old and the creation of new waterholes, and for a shift in settlement focus. There would appear to be a shift in agricultural regime, from arable- to pastoral-based. By the Middle Iron Age the field system was no longer maintained. In the Late Iron Age there was a notable reduction in the level of activity, although these sites were not totally abandoned.

Romano-British to Medieval

There is evidence for two broad categories of Romano-British activity within the wider area. One comprises the semi-urban roadside settlements or posting stations discovered at Staines and Brentford, some 6 km away, that occur on one of the principal roads (running west to Silchester, Bath and Exeter) that radiate out from Londinium (Bird 1987). Both these sites were situated on the River Thames near what were important river crossings. Away to the north-east of Harlington is the similar settlement of Brockley Hill, again placed on a major route leading out of Londinium. The other encompasses the more widespread evidence for small-scale rural settlements and farmsteads on the fringes of Londinium. Their pattern and distribution in the early Romano-British period appear to represent a continuation of the Late Iron Age pattern of undefended settlements. Evidence is generally in the form of field boundaries, pits, droveways and wells. Structural and burial evidence, however, are seldom recorded, possibly because these were relatively ephemeral and hence did not survive the post-Roman agricultural impact on the landscape.

Evidence for this rural and agricultural landscape has been found at Wall Garden Farm and Holloway Lane. At Wall Garden Farm, on the north side of Sipson Lane, a pair of Romano-British corn-driers and a timber-lined well were discovered along with associated features and finds (MoLAS 1993). The remains were predominantly 1st century AD, although the well was infilled in the 3rd century.

Romano-British settlement in the area is concentrated either within the 1st and early 2nd centuries, or within the 3rd and 4th centuries. There was an apparent break in occupation, with little evidence for continuity between the two periods, possibly reflecting a general economic decline in the province during the 2nd and early 3rd centuries.

No villas are know from the immediate area, although possible buildings have been found at Manor Farm, Harmondsworth, and further afield at Rickmansworth and Ruislip (Bird 1987, 66).

Archaeological and documentary evidence indicates Saxon settlement centres at both Sipson and Harmondsworth, the archaeological evidence consisting mainly of buildings and structures. There are few indications of the associated rural settlement, and again burial evidence is scarce. However, the place names Harlington, Hayes and Harmondsworth all feature in charters of 8th–10th-century date.

At Prospect Park, near Harmondsworth, a group of at least 11 sunken-featured buildings was recorded, probably associated with a rectilinear postbuilt structure with an ‘apsidal’ end, of middle Saxon (AD 650–850) date. A number of ditches and gullies, and cereal residues, provide limited evidence for associated agricultural activity (Farwell et al. 1999).

Excavation evidence suggests that these settlements probably shifted, with buildings being gradually abandoned and new ones built. It is likely that the middle Saxon settlement around Harmondsworth formed the origins of the later manor.

Figure 1.3 Imperial College Sports Ground (ICSG), principal phases of activity

Figure 1.4 RMC Land, principal phases of activity

Medieval settlement in the area is clearly shown by a number of existing villages of late Saxon and medieval origin, such as Harlington, Harmondsworth and Sipson. These small, nucleated villages situated on the roads leading west from London indicate the apparently prosperous agricultural settlement of the capital’s rural hinterland. Fieldwalking at Harlington has recovered large quantities of medieval pottery, though as yet no direct evidence for settlement. Manors are known from both Harlington and Harmondsworth.

Post-Medieval and Modern

The pattern of settlement and agricultural landuse appears to have changed little during the post-medieval period and the area remained very much the rural hinterland of urban London. The area became subject to Parliamentary inclosure in the early 19th century. The high grade/high yield status of the soil for market gardening purposes has long been recognised, with the area latterly referred to by planning authorities as the ‘A4/M4 Horticultural Belt’.

Even with the onset of the Industrial Revolution, with the construction of canals and railways and the industrialisation of rivers such as the Crane, the area remained predominantly rural in character. It escaped the urbanisation experienced by much of London during the Victorian period, and it was only with the growth in post-World War II housing, the construction of Heathrow Airport and the increasing demand for sand and gravel to supply the construction industry, that the area was widely developed. Even so, villages like Sipson, Harlington and Harmondsworth remain more or less discrete settlements, reflecting their early origins.

Project Research Themes

A number of broad research themes, based on prior knowledge of the West London area, were defined for both sites (Wessex Archaeology 1996b, 11–3; 2004b, 9–12), to be addressed by the excavations and the subsequent analyses. Although no significant evidence relevant to the Palaeolithic or Mesolithic was recovered, the excavations made substantial contributions to three themes relating to the development of the landscape from the Neolithic to the post-medieval period (Figs 1.3–4), leading to the formulation in the updated project design of a series of research questions to be addressed by the analysis (Wessex Archaeology 2008). These questions guided and structured the post-excavation analysis and culminated in the present publication:

•        How far can the vegetation and landuse of the sites and their surroundings at different periods be characterised, and the subsistence of their occupants defined?

•        What relation was there between the topography and the human use of what is superficially a flat terrain?

•        What was the spatial organisation of the ICSG and the RMC Land sites in successive periods, and to what extent did new elements relate to pre-existing monuments and other structures?

•        What was the timespan of the cremation burials on both sites, in particular, when did they start to be made? How far did they extend into the Late Bronze Age? What can the accompanying charred pyre material and tinder tell us about cremation practices, both within periods and over time?

•        How do the routeways across the site, especially the Iron Age and Romano-British trackway, relate to wider patterns of communication?

Radiocarbon Dating

Radiocarbon measurements have been calculated using the calibration curve of Reimer et al. (2009), and the computer program OxCal v4.2 (Bronk Ramsey 1995; 1998; 2001; 2009). The calibrated date ranges cited in the text are those for 95% confidence and are quoted in the form recommended by Mook (1986), with the end points rounded outwards to 10 years. The ranges quoted in italics are posterior density estimates derived from mathematical modelling of given archaeological problems (see below). The ranges in plain type have been calculated according to the maximum intercept method (Stuiver and Reimer 1986). All other ranges are derived from the probability method (Stuiver and Reimer 1993).

Location of Archives

The finds and archives for ICSG, RMC Land and LEWGF will be deposited with the London Archaeological Archive Research Centre (LAARC) under the project codes IMC 96, SIE 00 and WGA 07 respectively.

Chapter 2

Hunters, Herders and First Farmers

by Alistair J. Barclay, Andrew B. Powell, Chris J. Stevens and Philippa Bradley

Introduction

Prior to the widespread organisation of the landscape in the Middle Bronze Age, with its extensive field system, trackways and waterholes pointing to agricultural production as a dominant social concern (see below), evidence for the economic basis for Mesolithic, Neolithic and Early Bronze Age society is much harder to discern. In fact, the varied evidence from these periods suggests, at least at face value, that other, non-subsistence aspects of life were as, if not more, important.

Such a conclusion ignores the fact that Neolithic agriculture, based perhaps on a mobile, pastoral way of life, is likely to have low archaeological visibility (see Stevens below). Nonetheless, the varied, widespread and substantial evidence for activity during the Neolithic, particularly the Middle Neolithic, is characterised almost exclusively by the construction of monuments, some containing cremation burials, and the excavation of pits for purposes that remain unclear but which cannot be simply equated with domestic or economic activity.

The Middle Thames Valley, stretching from the Goring Gap to the present tidal reaches of the Thames at Teddington, is recognised as a core area of Early Neolithic activity. This area is centred on the lower reaches of the Colne Valley (Figs 1.1, 2.1), although other foci have been found further upriver between Eton and Maidenhead and at Sonning near Reading. The lower Colne Valley sits adjacent to, and east of, an area rich in Late Mesolithic sites (Holgate 1988; Barclay 2007, 333 and fig. 15.1). The earliest evidence for the change to a Neolithic lifestyle comes from a number of midden, pit and house sites that are often associated with assemblages of Carinated Bowl pottery. One of the earliest sites is that of Kingsmead Quarry, Horton where four possible buildings have been excavated (Chaffey et al. forthcoming). However, despite the apparent rarity of such structures in southern England, no less than six have been identified in or close to the Colne Valley. A structure very similar to some of the ones at Horton is known from the excavations at Cranford Lane (Nick Elsden pers. comm.) just outside the watershed of the Colne, while another structure was identified at Gorhambury in the upper reaches of the valley (Neal et al. 1990). These structures are likely to have been built during the early centuries of the 4th millennium BC (Barclay 2007; Hey and Barclay 2007). Of slightly later date is a series of monuments, including the Staines causewayed enclosure and a second possible site at East Bedfont, the Heathrow (Stanwell) bank barrow monument complex and a series of mortuary monuments of a variety of forms.

The same area has a wealth of pit deposits that span most of the Neolithic period (Holgate 1988; Cotton 2004; Lamdin-Whymark 2008). Early Neolithic pits are generally rare in comparison to ones associated with Peterborough Ware, in particular in the Mortlake substyle. Many pits associated with Grooved Ware have been found in the last 20 years and in certain areas they are more frequent than those associated with Peterborough Ware (eg, Kingsmead Quarry, Horton: Chaffey et al. forthcoming). This may reflect that different areas of the landscape were used at certain times, perhaps as areas of settlement shifted. With the notable exception of the Stanwell bank barrow, monument building was not widespread and may have been all but absent during the 3rd millennium BC as there appears to be no classic henge tradition in the Middle Thames Valley (Bradley 1984, 65 and fig. 3.6). This could indicate that society had a different structure to that of the adjacent regions of the Upper Thames Valley and Wessex.

In the later part of the 3rd millennium BC the evidence for the adoption of Beaker culture and values is rare. For whatever reasons the rite of single inhumation burial did not emerge and likewise the practice of pit digging with the discard of occupation material did not continue. Why these practices did not follow the same trajectory as other regions is a moot point. However, their absence may simply equate to a discontinuity in practice rather than a real absence of people. Beaker material has been found in small quantities, and some significant finds including pots, metalwork, worked stone and flint daggers have been recovered from the River Thames (Barclay 2011). For whatever reasons the practice of monument building was abandoned in the earlier 3rd millennium BC and this could have made the need for selective formal burial in the later centuries unnecessary.

Some barrows are known from the area although their numbers are relatively small when compared to areas of the Upper Thames and the Upper Kennet (in particular around Avebury) (Garwood with Hey and Barclay 2011, fig. 14.24). Isolated barrows are generally common, while barrow cemeteries are generally rare, small and seldom comprise more than about four or five monuments. Where these do occur they are often associated with Collared Urn pottery. As well as burial contexts this type of pottery has been recovered from pits and occupation deposits, as at Kingsmead Quarry (Chaffey et al. forthcoming), the Eton Rowing Course (Allen et al. 2013), Taplow (Allen et al. 2009) and Cippenham (Ford and Taylor 2004). There is little in the way of more tangible evidence for permanent settlement before about 1600 cal BC generally.

Figure 2.1 Neolithic and Early Bronze Age sites referred to in the text

Environment and Landscape

by Chris J. Stevens

Work from the wider area suggests that during the Late Mesolithic the dominant vegetation was a mixed woodland of oak, elm and lime, with alder-willow carr and reed marsh along the floodplain of the Thames and its tributaries (Scaife 2000; Keith-Lucas 2000; Branch and Green 2004).

Pollen data from a single pit at Perry Oaks suggests a similarly wooded landscape in the Early or Middle Neolithic, consisting of oak and hazel with some pine, birch, ash, lime and elm (Lewis and Brown 2006), although it seems likely that some significant clearance, including that associated with the Stanwell bank barrow, had taken place in the Early Neolithic (Framework Archaeology 2010, 59). The Thames floodplain itself was shown to be dominated by alder, with small clearances in which Rosaceae shrub (hawthorn, sloe, bramble etc) was able to establish itself, along with grasses and other weed flora (Wiltshire 2006).

At ICSG/RMC Land the small quantities of environmental information for the landscape prior to the Middle Bronze Age come mainly from the assemblages of wood charcoal and charred plant remains. Analysis of the charcoal reveals a similar picture of the Middle Neolithic woodland, with oak dominating, but with hazel and the hawthorn group (which includes other species such as bramble and sloe, as well as hawthorn) being reasonably well represented (Challinor, Chapter 10). There is also evidence for hazel in the charred plant remains, with the shells of hazelnuts, collected from the wild, dominating many of the samples.

The extent of such woodland is more difficult to ascertain. The Middle Neolithic monuments uncovered at ICSG (see below) are likely to have been constructed, at the very least, in relatively large clearings, as suggested in the reconstruction of the site at this time (see Cover). Further, if a tuber of onion-couch grass (Arrhenatherum elatius subsp. bulbosus), recovered from a deposit of possible pyre debris in the double ring ditch monument, is associated with the creation of a firebreak around a pyre in long grassland (Stevens, Chapter 10), this would suggest the existence of reasonably extensive areas of long grassland in the area by this time.

Pollen analysis covering the Neolithic to the Early Bronze Age in and around London has frequently indicated localised patches of clearance followed by regeneration of secondary woodland (Scaife 2000, 112–3). Similarly, the relatively high numbers of hawthorn-type charcoal in the assemblages from ICSG and RMC Land (Challinor, Chapter 10) also suggests the presence of thorny scrub either at the edge of woodland or regenerating within previously cleared areas. The charcoal assemblage from an Early Bronze Age cremation grave at ICSG points to long, wet rough grassland possibly in relict woodland, with some indication of patches of overgrown shrub, probably brambles and/or dog rose; the grave contained no oak charcoal, but had a predominance of hazel and hawthorn type/group.

Economy

by Chris J. Stevens

The environmental evidence for the economy in this period is sparse. No charred material was recovered from Early Neolithic features, but the Middle Neolithic features did produce ample evidence for hazelnuts, along with a single sloe, these probably representing only a fraction of a much wider range of wild foods that were collected.

Although cereal agriculture is at least demonstrated for the Early Neolithic from Kingsmead Quarry Horton only 8 km to the west (Chaffey and Brook 2012; Pelling forthcoming), there is still no conclusive evidence, despite the many environmental samples taken from ICSG and RMC Land (and also from Heathrow Terminal 5: Framework Archaeology 2010), for cereal agriculture in the immediate area during the Middle Neolithic to Early Bronze Age. While several Early and Middle Neolithic features did produce cereal remains, the fact that these were of hulled barley and free-threshing wheat suggested that the remains were intrusive and of more recent date, a suspicion confirmed by radiocarbon dating (Barclay and Stevens, Chapter 11: NZA-32684 from the long enclosure G3001, NZA-32687 and NZA-36738 from pits 5783 (Group I) and 11024 (Group Q)). None of the cereal remains, therefore, can be attributed safely to the Middle or Late Neolithic, and there is no secure evidence for cereals until the Early to Middle Bronze Age transition (c. 1600 BC).

Animal bone was similarly sparse (Grimm, Chapter 9). A single rib of cattle was recorded from an Early Neolithic feature at ICSG, while only a few fragments of cattle, along with a single metapodial of sheep/goat and two pig teeth came from Middle Neolithic pits spread across both sites. A pit or shaft of possible Early Bronze Age date contained both antler and a cattle horncore (of either domesticated cattle or aurochs), and a few other features of Late Neolithic to Early Bronze Age date also contained some remains of cattle. This pattern reflects the generally poor representation and preservation of animal bone in most of the surrounding sites; for example, only a few remains of cattle and possibly red deer were recorded from the Heathrow Terminal 5 excavations (Knight 2006).

Plate 2.1 The excavation of feature G2004, viewed from the south-east

Pre-Neolithic

There is very little evidence for human habitation prior to the first recorded traces of Neolithic activity on the two sites. The earliest evidence consists of two Palaeolithic flakes (ONs 13046 and 13049) that were recovered from the natural gravel (10810) on ICSG. Palaeolithic material has been recovered from the vicinity of the site (Wymer 1999) both from the Lynch Hill and Taplow Gravels (see Bradley, Chapter 7). Similarly, very little diagnostic Mesolithic flintwork was identified from either site. A microburin from a tree-throw hole on ICSG (10313), which also contained 10 flakes together with later pottery, and three microliths from RMC Land (tree-throw hole 3872; Late Bronze Age/Early Iron Age pit 2266; and 1100, a flint scatter of apparently mixed date). The microliths, two of simple edge-blunted type and the third a geometric type, are of Late Mesolithic date (Fig. 7.1, 1–3). A small number of blades and bladelike flakes recovered from both sites may also be contemporary. Other probable Mesolithic flintwork was recovered as redeposited material mostly from RMC Land.

The Late Mesolithic flintwork indicates at the very least that people were passing through the area during the 8th–5th millennia BC. It is possible that any contemporaneous occupation sites are situated closer to the River Colne. Holgate noted that sites of broad Late Mesolithic date tend to cluster further up the Colne Valley catchment or just outside its lower western reaches close to the River Thames (1988, 99–104 and map 9).

Early Neolithic

Within the overall distribution of Neolithic features in ICSG/RMC Land (Fig. 2.2), the evidence for Early Neolithic activity is concentrated in the southern part of ICSG (Areas C–D) (Fig. 2.3), where three features (G2004) within an area 15 m across produced 317 sherds (4496 g) of Early Neolithic pottery, amounting to 96% (by weight) of all the Early Neolithic pottery from both sites. It is possible that some of this evidence belongs to an early phase of the Neolithic (before c. 3650 BC) when farming practices were first introduced to Britain (Barclay 2007, 333 and table 15.1).

Feature G2004

Most of the pottery (255 sherds, 4089 g) came from a large irregular feature (G2004) measuring up to 4.5 m by 8.8 m wide, and up to 1.7 m deep (Fig. 2.4). This feature, which was located within an area of more widely disturbed ground that contained a number of tree-throw holes and other possibly natural features, comprised a sequence of cuts and recuts. These were excavated in a series of 2 m wide slots through the disturbed ground (Pl. 2.1).

It is clear that the two cuts that contained the bulk of the Early Neolithic finds (30064 and 30666) had cut through the fills of earlier features whose form it was possible to only partly determine. At the southern end of G2004, features 30080 and 30081, possibly a single feature up to 8 m long and 0.7–1.1 m deep with irregular sides and an uneven base (possibly a hollow left from a fallen tree trunk), contained a series of lower (gravel-rich) and upper (brickearth-derived) fills, one of the latter producing a single Early Neolithic sherd (3 g). These fills were cut by feature 30064, which was at least 4.2 m long and of similarly irregular profile, but cutting no deeper into the underlying natural than the earlier features. It contained a series of fills with varying quantities of brickearth and gravel components resulting largely from collapse of the sides and natural silting. Together these produced 255 sherds (2912 g) of Early Neolithic pottery representing at least three vessels (Fig. 6.1, 1–3). Most of the sherds came from the basal fill, the rest being distributed throughout the profile. From an upper fill, were also recovered six pieces of struck flint, and burnt flint (20 g).

Figure 2.2 Distribution of Neolithic and Early Bronze Age features at ICSG and RMC Land

There was a similar sequence at the northern end of G2004, where feature 30666 also appears to have cut through an earlier, sterile feature (30681, visible in section). Feature 30666 was at least 3 m wide and 1.7 m deep, and produced 208 sherds (3641 g) of Early Neolithic pottery, eight pieces of struck flint and a small quantity of cattle bone (5 g). Here too the pottery was distributed through the fill sequence, although most of it (2727 g) came from a layer of backfilled gravel (30661) in the upper half of the profile.

The function of these cuts is unclear, and there must be some uncertainty as to their date. Their size and irregular form are similar to Middle Bronze Age feature 30814 (see below) which cut the eastern edge of feature 30666, and which contained 29 sherds (181 g) of Middle Bronze Age pottery (as well as a further 24 Early Neolithic sherds (129 g), presumably redeposited). Feature 30814 was cut in turn by a ditch (G2198) of the later prehistoric field system, which also defines the eastern edge of the area of disturbed ground. While it is possible that G2004 is also of Middle Bronze Age date, with the Early Neolithic pottery deriving from an earlier feature that it cut, possibly a tree-throw hole similar to a nearby feature (G2005) which contained further sherds (see below), the relatively high quantity of Early Neolithic pottery from this feature makes it stand out, and a Neolithic date, therefore, cannot be ruled out.

Figure 2.3 Neolithic and Early Bronze Age features and pit groups at ICSG

Figure 2.4 Feature G2004 (ICSG): plan, and section of feature 30666

Tree-throw Holes

A tree-throw hole (G2005) 46 m south of feature G2004 (Fig. 2.3) produced 42 sherds (132 g) of Early Neolithic pottery, 58 struck flints and burnt flint (14 g), with a further five sherds (5 g) from the fill of a ditch terminal that cut it. A single sherd (2 g) and burnt flint (64 g) were also recovered from a sample from tree-throw hole G3067, 20 m to the east of G2004 (Fig. 2.3). A tree-throw hole (17072) on the southern edge of Area C contained a further 15 sherds (86 g), and nine pieces of struck flint including flakes from a polished implement made of a grey, very cherty flint distinct from many of the polished axe fragments found in the Middle Neolithic pits (Bradley, Chapter 7).

A further 11 sherds (32 g) of Early Neolithic pottery were recovered from two features associated with a penannular ditched monument (G3002, see below) in the centre of the site (Fig. 2.3). Seven of the sherds (along with 37 sherds of Late Bronze Age/Early Iron Age (LBA/EIA) pottery) were recovered from the monument’s ditch (G152) (see Fig. 2.8), although this has been tentatively assigned a Middle Neolithic date on the basis of its similarity to monument G2008 (see below). The ditch was cut by an irregular feature (G151), measuring 3.4 m by 5.4 m and 0.5 m deep, possibly another tree-throw hole, from which a further four Early Neolithic sherds were recovered (as well as nine LBA/EIA sherds from the upper of its three fills).

At RMC Land a single Early Neolithic sherd (13 g) was recovered from tree-throw hole 4478, which lay within the four Middle Neolithic pits of pit group C (below, Fig. 2.18).

As with the Late Mesolithic, little can be said about the earliest Neolithic activity on the two sites other than to note its presence, which appears to indicate that this was an area where people passed through, rather than one they occupied for any length of time. The pottery recovered from the features that comprise G2004 is likely to derive from a small temporary settlement, and is of a type that should belong to the 38th or 37th centuries BC (see Barclay 2007, 335 and table 15.1). Similar pottery comes from the ring ditch at Staines Road Farm, Shepperton (Jones 2008), where it was probably redeposited, and from a timber-built house at Kingsmead Quarry, Horton (Chaffey et al. forthcoming). It is probably not as early as assemblages from a natural shaft at Cannon Hill, Maidenhead (Bradley et al. 1975–6), perhaps of 40th or 39th-century BC date, and from the earliest deposits at the Eton Rowing Course (Barclay 2013; Allen et al. 2004; Allen et al. 2013). However, combined with the evidence from Shepperton and Horton it could indicate that farming practices spread across the Colne Valley catchment, probably with the movement of people, from the 38th century BC onwards. The similarity in ground plan between the house structures at Horton and Cranford Lane (Chaffey et al. forthcoming; Nick Elsden pers. comm.) can be taken to suggest a similar date. The actual location of the Cranford Lane house site is only 1 km east of the activity at ICSG.

Early to Middle Neolithic

Introduction

There are three main elements to the archaeology in this period (Figs 2.2 and 2.3). The first is a rectangular ditched enclosure (G3001) located at the north-east of ICSG (Area A) (Pl. 2.2). The ditch produced a small quantity of Middle Neolithic Peterborough Ware pottery, and comparison with similar monuments in the region suggests a mid- or late 4th millennium BC date (below). The second element comprises three circular monuments arranged in a north–south line in ICSG (Area D). The most southerly (G2007), a double ring ditch with a central cremation burial and further burials from within the ditches, produced a series of close radiocarbon dates from the end of the 4th millennium BC. Comparable dates were obtained from two further burials within a penannular ditched monument (G2008) to its immediate north. A second penannular ditched monument (G3002) lay some 200 m to the north, but contained no evidence for burials. The third element consists of some 90 pits, most containing varying quantities of Peterborough Ware and struck flint. These occurred both in groups