1. Introduction

Project background

The Elms Farm site (NGR TL847082) was located on the western periphery of Heybridge, immediately to the north-west of Maldon, in Essex (Figure 1 [JPG]). The site lay at the head of the Blackwater Estuary, immediately north of the flood plain of the Rivers Chelmer and Blackwater. It is at this point that the rivers run closest together (Wickenden 1986, 7). Prior to excavation, the site area was farmland, with the well-drained upper terrace being under arable cultivation while the lower terrace was exclusively pasture; a division that reflected the natural drainage and soils of these two areas.

Figure 1: Site location
Figure 2: Overall site plan

The Elms Farm excavations were part of a total development area of c. 29 hectares (71.6 acres), which extended north and eastward from the Chelmer and Blackwater Navigation as far as Langford, Crescent and Holloway Roads (Figure 2 [JPG]). Some 18 hectares were subject to varying degrees of archaeological investigation by the Essex County Council Field Archaeology Unit (ECC FAU), in two phases of work, in advance of residential development by Bovis Homes Ltd.

Volume 1, Section 1 (Atkinson and Preston 2015) and Wickenden's gazetteer of sites and finds from Heybridge (Wickenden 1986, 53-61) provides a summary of the features and finds recorded prior to the excavations described in this volume. Of the known remains, a single series of cropmarks were actually within the development area. Prior to these investigations, the Essex Heritage Conservation Record (EHCR; formerly Essex Sites and Monuments Record or ESMR) contained some 38 references of relevance to the project. Fourteen of these occurred in close proximity to the site and have direct relevance to it. These are discussed in Wickenden's gazetteer of sites and finds from Heybridge (1986, 53-61) and only the briefest of overviews is provided here.

Figure 743
Figure 743: Aerial photograph from 1975 showing linear features and pits in the area of the Northern Hinterland of the site (NMR photo TL8408/3)

The cropmarks covered approximately 8 hectares (HER 7801), the majority of which occupied the area of the 1993 site (Figure 743). The linear cropmarks were interpreted as probable Late Iron Age and Romano-British field systems. A large number of pit-like cropmarks were clustered in the south-eastern corner of this complex; these appeared to be roughly rectangular, to share a common alignment, and to overlie the linear features. All were subsequently investigated during the 1993 excavation, within what was subsequently referenced as the Northern Hinterland (Area W).

The various sites and casual findspots of prehistoric, Late Iron Age, Roman and Saxon date record the discoveries made during gravel extraction and small-scale development since the late 19th century. Although there are references to the finding of material as early as 1839, the most significant early discovery was made in 1887 during the construction of a branch of the Great Eastern Railway line and creation of a gravel pit (now a pond), at Langford Junction, to the north of the roundabout on Figure 1. Local antiquarian E.A. Fitch recorded that large amounts of Late Iron Age and Roman pottery, as well as hundreds of coins and other metalwork, were discovered during this work (Fitch 1905). It was largely on the strength of Fitch's work that Heybridge acquired a reputation as a port involved in continental trade. Late Iron Age and Roman burials were also discovered at various locations to the east of the site in the late 19th and early 20th centuries. These included Late Iron Age cremations from the Heybridge Cemetery and inhumations in lead and stone coffins from The Towers, Heybridge (Wickenden 1986, fig. 1).

Formal archaeological excavation in the vicinity of the site only took place after 1967, generally in advance of housing developments to the north and north-east of the present site. The most significant of these was Drury's Crescent Road excavation in 1971-2 (Drury and Wickenden 1982; Wickenden 1986), immediately to the north of the 1994-5 excavation. The Crescent Road site spanned the Late Bronze Age to early Saxon periods with evidence of Roman occupation from the 1st to 4th centuries AD, elements of which were interpreted as a street frontage. The early Saxon occupation was more substantial, represented by five Grubenhäuser and a possible post-built structure, all dated to the early 5th century AD. It was from Drury's work that Heybridge acquired the label of a 'small town' (Rodwell 1975). It was also cited as an example of probable later Roman-early Saxon settlement continuity.

Small-scale excavations and a watching brief were undertaken in 1983 by the Maldon Archaeology Group at Boucherne's Farm, some 200m to the north-east of the 1994 area (Wickenden 1986, 60). This site produced evidence of both Late Iron Age and Romano-British occupation in the form of boundary ditches and a concentration of unstratified pottery. More recently, an evaluation to the south of Holloway Road, on the northern perimeter of the development area, was undertaken by the Cotswold Archaeological Trust Ltd in March 1993 (Timby 1993). Trial trenches revealed a number of Middle Iron Age pits sealed by a deposit of gravel, thought to be of Roman date, and Roman ditches, an ensemble that would seem comparable to those across the northern part of Areas A2 and A3 of the Elms Farm site. It was concluded that the vicinity had not been significantly disturbed since the Roman period. Further excavation was not undertaken as the ground surface was to be raised prior to construction. Cotswold Archaeological Trust also carried out an area excavation at Langford Road, immediately to the north-east of the 1993 site (Langton and Holbrook 1997). In addition to features of prehistoric date, it contained evidence of Late Iron Age, Roman and possibly Saxon occupation activity.


The fieldwork stages of the Elms Farm project are described in greater detail in the digital archive (Essex County Council 2015) and are only very briefly summarised here.

Geophysical survey

Geophysical surveys were undertaken in 1993/1994 and 1997. The 1993/4 surveys by Geophysical Surveys of Bradford covered approximately 13ha within the Elms Farm development area. A total of four fields, were surveyed by gradiometry. The survey located a large number of anomalies of archaeological significance, including ditches that echoed the cropmark alignments to its north, possible roads and areas of large-scale pitting, as well as a former stream channel.

In 1997, following completion of the Elms Farm excavations, the Field Archaeology Unit extended the magnetometer survey to the south-west of the excavated area, on the other side of the B1018/A414 and close to Langford Junction. This established that the Late Iron Age and Roman settlement extended beyond the excavation limits in this area. The high density of magnetic anomalies included more linear features (on a similar alignment to the excavated/crop-mark examples), an extension to the main east-west road across the site, oval and rectangular enclosures and a large number of pits. Most intriguing of all is a pair of concentric circular ditches, reminiscent of the temple excavated on the main site, albeit with a smaller 18-20m diameter.

Trial trenching

While no intrusive evaluation work was undertaken prior to the excavation of the 1993 site, a limited programme of trial trenching was carried out in areas peripheral to the 1993 site and across the 1994 site, the latter in order to verify and amplify the results of the magnetometer survey undertaken by Geophysical Surveys of Bradford in 1993/4.

The trenching of the 1993 site peripheries was undertaken in two areas beyond the known cropmark complex of the main field; within the land-take of the by-pass extension and of a gravel borrow pit located to the west of the B1018/A414 (Figure 1). Those trenches in the intended position of the gravel pit contained no archaeological remains, while those to the north in the bypass land-take revealed only a single ditch of early Roman date. This ditch almost certainly related to features later found on the Langford Road site (Langton and Holbrook 1997).

Within the 1994 site, a total of eight trenches (Tr.1 - 8) of varying size were positioned with reference to the geophysical plot to assess further the nature of tentatively identified aspects such as roads, roadsides and plot interiors. Archaeological remains were found in all of the trenches and, where appropriate, are included on plans of the area excavations and discussed in the site narrative of this report. While these eight trial trenches (amounting to only some 0.6% of the c. 13ha 1994 development area) showed that the geophysical survey was correct in its broad interpretations of the magnetic anomalies recorded, this work was of insufficient scale to reveal the sheer density, depth and complexity of the settlement remains. However, three of the trial-trenches (1, 2 and 8) were located in areas in which no subsequent area excavation was undertaken. These trenches therefore constitute our only insight into the nature of the archaeological remains in these vicinities.

Trial Trenches 1 and 2

Both were narrow machine-cut trenches measuring approximately 2 x 20m and located in the northernmost field of the 1994 site (Area B). Aligned north-south, both were positioned across the line of the relict watercourse that ran through the area (Figure 3). Blue-grey water-lain clay deposits were encountered in Trench 1, but no archaeological features were discerned. A single Early Saxon pottery sherd was recovered from its surface, though further machining to a depth of 0.5m revealed no further remains. At the time, this clay was concluded to be a natural water-lain deposit. In Trench 2 dark grey silty clay and gravel was exposed. It contained Roman tile and pottery sherds throughout the 0.8m of homogeneous waterlogged deposits removed by machine, but no features were visible within them.

Trial Trench 8

Located in the southernmost field of the 1994 site (Area C), Trench 8 was a 10m square in what was the lowest-lying part of the site (Figure 3). Beneath the topsoil a 0.1-0.25m thick layer of clean, artefact-free, blue-grey clay was removed to expose a homogeneous deposit of dark grey silty clay and gravel. Part of this was machine-excavated to a depth of c. 1m, at which point a large animal leg-bone (probably of a cow) was exposed. No further features or artefacts were observed within this waterlogged deposit.

The excavations

Site areas

Figure 3
Figure 3: Site areas and zones

During the course of pre-excavation planning, fieldwork and post-excavation analysis, a hierarchy of different codes was evolved to aid reference to general and specific areas of the site (Figure 3). The developer's scheme of works imposed the first three levels of subdivision, which basically laid out the sequence of earth-moving, stockpiling and development. Stages I and III more or less coincide with the two phases of archaeological area excavation. Stage I contained the 1993 excavation and the trial trenching to its north. Stage II contained the trial trenching undertaken in advance of gravel extraction. Stage III contained the 1994/5 excavation. The largest of these, Stage III, was subdivided by necessity into general Development Areas A to C; Area A was then further subdivided into Areas A1 to A4.

For the purposes of archaeological excavation, particularly as it became clear that the sampling strategy needed to be revised, individual excavated areas within A1 to A4 were given an alphabetic code (Areas D to R, omitting the letter O). This was later extended during post-excavation analysis to the 1993 site, which was labelled Area W.

The 1993 excavations

The central field, an area of some 3.8ha, at the northern end of the development on the upper gravel terrace was selected for area excavation (referred to throughout this volume as the Northern Hinterland or Area W). Truncation had clearly occurred as a result of modern ploughing and so all top- and sub-soil was removed down to the natural gravel. Within the stripped area, sampling of all features was achieved, except where flooding prevented work in the south-eastern corner of the site. Discrete features such as pits and post-holes were excavated to a minimum of 50% and larger features such as ditches were sectioned. Features deemed to be of particular interest, notably a prehistoric ring-ditch, pottery kilns, and cremations, were excavated to a higher degree, often to 100%.

The 1993 excavations revealed a substantial part of the hinterland/infields located immediately to the north of the main Late Iron Age and Roman settlement area. Within these there was evidence for crop-processing, pottery manufacturing, funerary and burial practice, and possible reuse in the early Saxon period.

The 1994-5 excavations

The excavations in 1994-5 consisted of the large-scale excavation of much of the lower gravel terrace portion of the development area (Figure 3). The evidence from the geophysics and trial-trenching had already established that this area comprised the eastern half of an extensive occupation site, which may have stretched as far west as the marshes bordering the River Chelmer. However, the full complexity of the archaeological remains subsequently revealed was not apparent at this stage. The excavations were to reveal the eastern part of what could perhaps be termed a 'large market village', which had begun in the late pre-Roman Iron Age and continued into the Early Saxon period. The excavated area focused on a temple complex, centrally placed within a road and track network and flanked by numerous domestic and industrial features (roundhouses, ditches, pits, pottery kilns, crop-drying ovens etc.).

On the basis of previous excavation evidence from the Crescent Road area and on that of the 1993 Elms Farm site, along with the geophysical survey results and the trial-trench findings, an excavation strategy had been formulated for the 1994 excavation. This aimed to undertake sampling of all archaeological features and deposits. However, owing to the high quantity of archaeological deposits identified in the 1994/5 season, the excavation strategy needed to be revised at an early date and remained under revision throughout the excavation.

All areas were mechanically stripped of topsoil and a thin skim of subsoil, under archaeological supervision. The presence of overhead electricity cables required clearance of 5m either side so that stripping could not be undertaken in these areas.

Following the topsoil strip of Area A1, much of the area was conveniently subdivided by a system of metalled roads. This was used as the basis for further subdivision into Areas G to K. The north-western part of Area A1 was less well defined and its subdivision into Areas D to F was initially influenced by the need to investigate as a priority the land-take of a contractor's haul-road along the northern edge of the site. Some minor adjustment to the extent of these latter areas has taken place during post-excavation work, where landscape features such as major boundary ditches have been adopted in preference to the arbitrary divisions imposed in the field.

In the absence of recognisable landscape units to use as the basis of the subdivision of Area A2 and A4, a series of alternate 20m strips, aligned north-south, were investigated as Areas L to Q. The aim was to investigate 10% of all linear features and 50% of all cut features within the strips. In Area A2 excavation was hindered by severe and persistent flooding.

Owing to Areas B and C being particularly prone to flooding, and the fact that the ground surface of these areas was scheduled to be built up with gravel, these areas were not extensively investigated. With minimal disturbance from construction works anticipated, only the first stage of stripping and pre-excavation recording was undertaken in these areas. Although Area A3 was subject to topsoil stripping, no excavation was undertaken following its pre-excavation planning (see Figure 2). Investigation of Area B was limited to two small open areas linked by a machine-excavated trench across the ancient watercourse that ran between them.

The above approach resulted in approximately 18% of the entire development area being subjected to detailed excavation. In the 1994-5 excavations some 34% of the machine-stripped area were subject to detailed excavation. Conditions on site prevented both excavation of the full depth of stratigraphic sequences and sampling every single discrete feature. Thus it is estimated that the true percentage of completed sampling is some way below the 34%. This variable level of sampling from area to area must be borne in mind when interpreting the data set.

Metal detecting

Metal detecting of the topsoil was undertaken by both ECC FAU staff and local enthusiasts on the 1994 excavations. The metal detecting of Area A1 proved to be very productive. As a result of this, a more systematic metal-detecting strategy evolved, based on a 20m-grid square. While the metal-detected finds form a very large assemblage, they are of varying provenance and their location varies from the level of general area (e.g. Area A1, A2, etc.) to 20m grid and, occasionally, eight-figure OS grid referenced. Thus, few are suitable for inclusion in detailed considerations of artefact distributions but have made a significant contribution to the general study of status, function and economy (e.g. the coins). The site also suffered from illegal metal detecting at night.


The post-excavation for a site as large and complex as Elms Farm has proved challenging. In total there were some 17,000 contexts, along with 26,000 finds entries (by individual 'registered find' or as a single entry per category per context for 'bulk finds'), and 1350 soil samples.

For ease of reference, individual features and their layers and fills were 'grouped' together in the post-excavation process and assigned Group numbers, and it is these numbers that have generally been used both on the site plans and in the text.

The decision was taken relatively late during the post-excavation process to rework the whole framework of the site narrative. This entailed the move away from the description and analysis of the site by each Excavation Area (i.e. Areas A-R and W) and toward that based on archaeologically tangible settlement units of enclosure systems (ES) and open areas (OA) (see Figure 3).

However, most of the specialist publication reports had been drafted by 2000 and they therefore refer to Excavation Areas rather than the enclosure/open area system. These have since been partially edited and adjusted to reflect subsequent changes in site narrative presentation, etc. In addition, there has been no substantive updating of analyses to take account of more recent discoveries and advances in understanding with regard to individual finds-types.


The chronology scheme, which had to be a broad period-based structure robust enough to facilitate the bringing together of a very large quantity of complex, diverse, often unrelated and spatially distant remains, has, however, remained a constant construct throughout the post-excavation analysis. While pottery analysis employed a more specific ceramic pottery dating scheme within this basic framework, to facilitate a greater detail of analysis of supply and consumption, all stratigraphic analysis and other specialist study has been carried out using the following site period scheme.

Phase Broad Period Date Range
1 Prehistoric Palaeolithic to Middle Iron Age
2 Late Pre-Roman Iron Age and Transition Mid-1st BC to mid-1st Cent AD
3 Early Roman Later 1st to mid-2nd Cent AD
4 Mid-Roman Later 2nd to mid-3rd Cent AD
5 Late Roman Late 3rd to mid-4th Cent AD
6 Latest Roman-Early Saxon Late 4th to 5th Cent AD
7 Later (Post-Abandonment) Post 5th Cent to Now (Present Day)

Guide to the dissemination strategy

The dissemination strategy for the Elms Farm project comprises three components; Volume 1, Volume 2 and the archive.

Volume 1, a monograph in the East Anglian Archaeology series (Atkinson and Preston 2015), is a synthetic and interpretative overview of what the authors feel to be the most important aspects of the Late Iron Age to Early Saxon settlement, presenting them as a series of thematic chapters that draw their supporting evidence from the descriptive and analytical reports that comprise Volume 2.

Volume 2 (this volume) is published as a digital monograph report. However, despite this being a digital volume it must be borne in mind that the project (fieldwork undertaken 1993-5) dates to the beginning of the digital age and the full use of digital recording methodologies on archaeological sites largely post-dates the fieldwork stages of this project. As a consequence, Volume 2 follows a predominantly conventional monograph publication format. While it contains more description, analysis and basic level interpretation than Volume 1, the evidence presented is also, of necessity, highly selective. Description of excavated features and deposits and of artefact and environmental assemblages are not comprehensive, but focus upon the key aspects that contribute to the interpretation and understanding of the settlement. The site narrative largely omits unphased or otherwise uninterpreted features along with virtually all unexcavated remains. Other, often reasonably well-dated and phased, features such as relatively unremarkable pits have not been described other than in the most collective and summary terms. However, all features do appear on the relevant area and phase plans. Feature fills, unless specifically relevant, are not generally described (nearly all were some variant of grey-brown sandy silt, with a low gravel or small pebble component). Only those features specifically discussed in the text are labelled on the period plans.

The project archive comprises two elements. Firstly there is the conventional archive comprising original primary records and finds, which are deposited with the Colchester and Ipswich Museum at Colchester. Secondly there is the digital archive, which is deposited with the Archaeology Data Service at York. Again, it must be borne in mind that the project dates to the beginning of the digital age and the full use of digital recording methodologies on archaeological sites largely post-dates the fieldwork stages of this project. The digital archive contains the original digital data for the site in the form of a series of relational database tables (including context descriptions, basic finds identification and quantification, etc.), shapefiles of the digitised site plans and those elements of the specialist analytical reports that are not presented in Volume 2 (largely appendices and supporting data sets). During the post-excavation processes, further databases were created specifically for the Roman building materials, and the LIA/Roman pottery, whose analysis required information in fields not provided in the main Finds database, which are also made accessible.


Internet Archaeology is an open access journal based in the Department of Archaeology, University of York. Except where otherwise noted, content from this work may be used under the terms of the Creative Commons Attribution 3.0 (CC BY) Unported licence, which permits unrestricted use, distribution, and reproduction in any medium, provided that attribution to the author(s), the title of the work, the Internet Archaeology journal and the relevant URL/DOI are given.

Terms and Conditions | Legal Statements | Privacy Policy | Cookies Policy | Citing Internet Archaeology

Internet Archaeology content is preserved for the long term with the Archaeology Data Service. Help sustain and support open access publication by donating to our Open Access Archaeology Fund.