PREVIOUS   NEXT   CONTENTS   ISSUE   HOME 

3.2 The Late Iron Age and Roman Pottery by Edward Biddulph, Joyce Compton and Scott Martin

Cite this as: Biddulph, E., Compton, J. and Martin, T.S., 2015 The Late Iron Age and Roman Pottery, in M. Atkinson and S.J. Preston Heybridge: A Late Iron Age and Roman Settlement, Excavations at Elms Farm 1993-5, Internet Archaeology 40. http://dx.doi.org/10.11141/ia.40.1.biddulph1

Introduction

The recording, analysis and reporting of the Late Iron Age and Roman pottery was undertaken by a number of people over an extended period of time. Initial recording and assessment was carried out by Colin Wallace. Recommendations for analytical work (see UPD in paper archive at Colchester and Ipswich Museum) were taken up and developed by Scott Martin (see UPD pottery addendum in paper archive). Recording by fabric and vessel form by context, and quantification of the assemblage, was subsequently carried out by the project's three pottery researchers, Edward Biddulph, Joyce Compton and Anne Thompson. Analysis and reporting was largely undertaken by Edward Biddulph and Joyce Compton, with Scott Martin contributing study of the latest Roman material. Additional study has been carried out on particular components of the assemblage by other specialists; their contributions are acknowledged where appropriate.

The Elms Farm excavations yielded more than 282,000 sherds of Late Iron Age and Roman pottery, weighing a total of 6.4 tonnes, from 4986 contexts. Of this, 269,070 sherds, weighing 5213kg, were stratified. This constitutes one of the largest single assemblages of Late Iron Age and Roman pottery to have been studied using modern quantification techniques in Essex and, indeed, in the region. The assemblage is comparable to the 5.7 tonnes of pottery used for detailed analysis at Colchester (Symonds and Wade 1999, 1) and far in excess of the 1.6 tonnes from Chelmsford (Going 1987, 1; 1992a, 93). Spanning the mid-1st century BC to late 4th century+, analysis of this assemblage has provided the opportunity to test critically and enhance current views of chronology and supply principally derived from studies of the pottery from Camulodunum/Colchester and Chelmsford.

The exceptional quality of the assemblage provided an opportunity to investigate a range of aspects for the pottery at Heybridge. Determining a ceramic sequence for the site and the pattern of pottery supply to the settlement were fundamental to the project aims. The analysis, however, was able to go beyond aspects of chronology and supply to explore issues of site development and daily life in the settlement. Pottery was an intermittent part of the local economy, with the presence of kilns confirming that the settlement produced, as well as received, ceramic vessels. Various studies have shown that pottery was used in diverse ways, functional in both ritual and mundane settings (see Vessel function and use). These studies have not been exhaustive and the recorded pottery data provide a comprehensive resource for further research.

Methodology

The assemblage was recorded on an area-by-area basis (Areas D-R and W) using the divisions devised for excavation purposes. The stratified pottery was sorted into fabric groups and recorded by context to the basic level of sherd count and weight, in grams, following Study Group for Roman Pottery guidelines (1994). The information was recorded on pro forma sheets along with vessel class and form identifications, typological references, dating and other observations. Roman forms were classified using the Chelmsford type series (Going 1987, 13-54; Classes A-S), with the Camulodunum type series (Hawkes and Hull 1947, 215-75; updated by Bidwell and Croom 1999, 468-87) used principally for the Late Iron Age pottery. Roman forms not found in either were classified using regional typologies where possible, such as Monaghan's Upchurch/Thameside series (1987) and Young's Oxfordshire series (1977). Late Iron Age forms not present in the Camulodunum type series have been incorporated into a new site-specific typology published in this volume. Any forms not in the above can be found in the intrinsic pottery section. Unstratified pottery was recorded in less detail, with quantification in fabric groups by weight carried out for all of this material but with sherd counts and form identifications only partially undertaken.

Preliminary dating was undertaken at the time of basic recording and each group assigned to either single or multiple date ranges. The scheme, adopted purely for analytical purposes, comprised sixteen ceramic phases based upon the tripartite division of the centuries covered by the assemblage (i.e. mid-1st century BC to late 4th century+ AD). The detail of this ceramic phasing is held in the archive (x_code_ceramic_phases) and has been converted into a more robust scheme used throughout this report. The ceramic phasing scheme (CP) comprises eleven phases that follow the regional pattern as identified at Chelmsford (Going 1987, Ceramic Phases 1-8). Comparisons can also be made with the ceramic phasing scheme used at Colchester (Symonds and Wade 1999, Period Ending Groups 0-18). Context dating was refined by reference to the stratigraphic record.

Table 1: Elms Farm site phasing and comparative ceramic phasing schemes
Elms Farm Site Period Elms Farm CP Chelmsford CP Colchester PEG Date Range
1 - - - Palaeolithic-MIA
2 1 - - c. 50-15 BC
2 2 - 0 c. 15 BC-AD 20
2 3 - 3 c. AD 20-55
3 4 1 4-5 c. AD 55-80
3 5 2 5-8 c. AD 80-125
3 6 3 9-10 c. AD 125-170
4 7 4 9-10 c. AD 170-210
4 8 5 10-12 c. AD 210-260
5 9 6 13-14 c. AD 260-310
5 10 7 14-16 c. AD 310-360
6 11 8 16-18 c. AD 360-400+

All well-dated, stratigraphically sound and relatively large groups were further quantified by Estimated Vessel Equivalence (EVE) using vessel rim measurements. The values for the percentage of extant rim, expressed as a proportion of a complete rim, and the rim diameter were entered onto a separate set of pro formas along with archive drawing numbers and details such as vessel decoration. A total of 314 fully quantified and described groups (representing 6% by context count or 22% by assemblage weight of the total assemblage) were thus selected. For the purposes of analysis, a further round of criteria subsequently identified a 'pool' of the 152 best groups, which were used to form the basic understanding of pottery supply to the settlement, and as a source of data in the study of the many aspects of pottery use. Forty of the most representative are published in detail, and illustrated, as Key Pottery Groups (KPG). Despite this selection, data drawn from the entire pottery assemblage continued to be used throughout the programme of analysis. That consistently good results were obtained is testament to the effectiveness of even a basic level of recording. To enable manipulation and interrogation of this very large assemblage, all collected data were entered onto FoxPro relational databases (see the archive).

The extensive archive includes the pottery, which has been ordered and boxed into type following the recording protocol. All sets of paper pro forma pottery record sheets are included, along with selected information used during recording and analysis. There are full dating evidence sections for each context/feature that contained pottery. A total of 1973 vessels were drawn as recording progressed.

The fabrics

This section presents a detailed list of the 100 fabrics and fabric groups recorded for the Late Iron Age and Roman pottery assemblage. Of these fabrics, at least thirty-seven were not included in the Chelmsford fabric series. Many would normally be found in contexts dated to the Late Iron Age and are, therefore, unlikely to have occurred at Chelmsford, which was not established until c. AD 60/5 (Going 1987; Drury 1988, 125-8). The Roman pottery volume for Colchester (Symonds and Wade 1999) provides a comprehensive list of the pottery from the colonia, but this too is inadequate in its description of Late Iron Age wares. The Fabrics section is presented in full below since it provides the most extensive list of Late Iron Age and Roman fabrics for the region. It supersedes the numeric system devised for Chelmsford (Going 1987, 3-11) and enhances the system used at Colchester (Symonds and Wade 1999, 2-5, 12). Each fabric is identified by a common name and a mnemonic code and use of this code is standard for all ECC Field Archaeology Unit projects. The scheme provides a high level of standardisation and facilitates easier comparison between sites.

Fabrics are listed in strict alphabetical order by fabric name, followed in brackets by the standard ECC Field Archaeology Unit mnemonic code. The corresponding Chelmsford fabric codes and the National Roman Fabric Reference Collection codes (Tomber and Dore 1998), the latter prefixed NRFRC, have then been appended where applicable. These provide the principal references for fabric descriptions. Fabrics have been described in full in the few instances where no description exists in either volume. A 'thumb-nail' biography for each fabric is provided, comprising a date range specific to Elms Farm, a list of forms present in that fabric, and a brief overview of its occurrence. Most fabrics were present in sufficient quantities to gain an overall date range. In the cases of fabrics where this is equivocal, for instance, where the appearance of a fabric is restricted to a few sherds, a generally accepted published date range, not specific to Elms Farm, is provided instead. Any differences between accepted chronologies and the date range of a fabric at Elms Farm are discussed, as are any notable aspects of form and fabric. The forms encountered in a given fabric are listed below the date. Roman forms were classified using the Chelmsford type series (Going 1987, 13-54), with the Camulodunum type series (Hawkes and Hull 1947) used principally for the Late Iron Age pottery. Roman forms not found in either were classified using regional typologies, where possible. Late Iron Age forms not present in the Camulodunum type series have been incorporated into a new site-specific typology published in this volume. These forms have the prefix EF, followed by the type number. Dragendorff (1895) form codes, which are used universally for samian vessels, have been simplified here. The common abbreviations (Drag. or Dr.) have not been used; Dragendorff forms are denoted solely by the prefix 'f'. Less common form codes, however, such as those delineated by Ludowici (1927) and Curle (1911), are given in full. Finally, a brief overview of the occurrence of a given fabric is provided, with its distribution noted where appropriate. Full distributions by site area for individual fabrics are contained in the archive.

Fabrics were identified and grouped on a macroscopic basis, aided by the use of x20 magnification where necessary. Most fabrics are based on pre-defined fabric divisions and are known by their common names or have direct equivalents published elsewhere. There are, however, some exceptions. Grog-tempered fabrics are not sufficiently represented at Chelmsford for the description offered by Going (1987, 10) to cover the full range of fabric variations. The fabric group has thus been given four subdivisions at Elms Farm in order to record fully these variations. Going's 'Romanizing wares', fabric codes 34 and 45, are now identified under the umbrella code for black-surfaced wares (BSW). During the recording stage for the purpose of analysis, mortaria were assigned specific fabric codes comprising the relevant fabric code and the suffix 'M'. Because mortaria are just one form within a product range, their fabric descriptions have been incorporated with the generic fabric group, where appropriate. This allows comment to be made on the full range of forms and the scale of production of individual industries. The descriptions for the amphora fabrics have been compiled with the assistance of Dr Paul Sealey, whose contribution is gratefully acknowledged. Condensed entries for these are given here, following the alphabetic system in Tomber and Dore (1998, 82-113); fuller descriptions for all of the amphora fabrics at Elms Farm are held in the archive. A limited programme of thin-sectioning was undertaken by Dr D.F. Williams of Southampton University. Summary results of this programme have been incorporated with the relevant fabric descriptions, where appropriate and the results for pottery from the kilns can be found in the relevant section. The full report by Dr Williams forms part of the research archive. The archive also contains a full list of the thin-sectioning results for selected vessels from pyre-debris pit 15417 (KPG5).

Alice Holt grey ware (ALH)

Description: Chelmsford Fabric 43, NRFRC code ALH RE
Date: Mid- to late 4th century+
Forms: Dish B6, jar G41.1
Occurrence: The bulk of the stratified sherds came from the central settlement zone, otherwise the fabric was thinly distributed. The most common form is a storage jar. Two bead-and-flanged dishes, associated with Saxon pottery, were present in the top fill of well 14529 (KPG40). This vessel type is rare, and may represent one of the latest identifiable Roman forms found in Essex.

Amphoras, unsourced (AMISC)

Description: This is an omnibus fabric group reserved for a minority of amphora body sherds, usually small, which could not be assigned to a specific fabric.

Amphora stoppers (ABUNG)

Description: An omnibus fabric group for amphora stoppers. Generally, the stoppers have fine and soft fabrics with little in the way of inclusions. Typically, they are off-white (10YR 8/2), buff (7.5YR 8/6) or light brown (7.5YR 7/6); one is (exceptionally) light red (10R 6/8).
Form: Amphora stoppers
Occurrence: Little of this fabric group was recovered and the occurrence matched that for ABAET (Baetican amphoras, see below).

Argonne samian ware (ARSW)

Description: Chelmsford Fabric 60, NRFRC code ARG SA
Date: Late 3rd century+
Forms: Bowl f37
Occurrence: Very little of this fabric was present, amounting to a maximum of four vessels. All sherds were found in late 3rd century, or later, contexts, mostly in the northern zone.

Arretine (Italian-type) samian ware (ITSW)

Description: Chelmsford Fabric 60, NRFRC code PIS SA
Date: c. 15 BC to AD 20
Forms: Platters Conspectus 12 Conspectus B1, Cups Conspectus 13/14 Conspectus 23
Occurrence: This fabric is poorly represented, occurring mainly in contexts of early 1st century AD date. A maximum of twenty vessels was recorded, with the majority being platters of Conspectus form 12. One platter fragment, from pit 20030, has a radial stamp, and this vessel is the earliest, dating to the late 1st century BC.

Baetican amphoras (ABAET and ABAEH)

Description: Chelmsford Fabric 55, NRFRC code BAT AM 1
Date: Late 1st century BC to mid-3rd century AD
Forms: Dressel 20 Haltern 70
Occurrence: This is the most common amphora fabric, accounting for almost half of the amphoras by weight. Imports of this fabric to Heybridge began in the late 1st century BC, with a steady increase in volume through to the early to mid-3rd century. Levels tailed off thereafter, with the fabric residual in contexts later than late 3rd century. The fabric was not uniformly distributed. The top half of a Dressel 20 amphora was found inverted in pit 4585.

Black burnished ware category 1 (BB1)

Description: Chelmsford Fabric 40, NRFRC code DOR BB 1
Date: Early 3rd to mid-4th century
Forms: Dishes B1.4 B2 B5.2 B6.3, jar G9
Occurrence: This fabric was recovered from most parts of the settlement. A limited supply to Heybridge began during the mid-2nd century, in line with Colchester (Symonds and Wade 1999, 352). Supply rose dramatically from the early 3rd century, reaching a peak before the end of the century, declining thereafter. It is notable that the amount of BB1 found in late 4th century AD+ contexts is more than double that recovered from mid-4th century contexts, and at least five times that from early 4th century contexts. Given the later 4th-century decline of BB1 in Britain generally (Gillam 1976, 59), its appearance in late 4th-century contexts is entirely residual. Forms are restricted to dishes and jars, often decorated with burnished lattices or intersecting arcs.

Black burnished ware category 2 (BB2)

Description: Chelmsford Fabric 41, NRFRC codes COL BB 2, MUC BB 2, CLI BB 2 and COO BB 2
Date: Early 2nd to mid-3rd century
Forms: Dishes B1 B2 B3 B4 B5, jar G9
Occurrence: BB2 is less common than BB1 and HAB (Hadham black-surfaced ware), but it was recovered from all parts of the settlement. While the fabric first reached Heybridge during the early 2nd century, the quantity of BB2 peaked in the second half of that century, declining from the early 3rd century. Sources have not been certainly identified, but it is reasonable to presume that the bulk of BB2 was manufactured at Colchester and Mucking, the nearest known producers to Heybridge. However, some of the earliest products were undoubtedly imported along with grey and oxidised wares from sources situated along the Kentish side of the Thames Estuary (cf. Monaghan 1987; Pollard 1988). A wider range of forms compared to Chelmsford is represented here. Straight-sided dishes are common, though less so than bead-rimmed dishes. Decoration is restricted to lattices and burnished diagonal lines.

Black burnished wares, unsourced (BB)

Description: Black burnished pottery not characteristic of BB1, BB2 or HAB in terms of fabric or surface treatment has been placed in this category.
Forms: Dishes B1.3 B2/B4 B3 B6, jar G9.1
Occurrence: This group contains a small amount of pottery. Almost all is wheel-made and occasionally decorated with lines or lattices. Dishes and jars, typical products of the black burnished industries, were recovered from contexts spanning the Roman period, concentrating in the 2nd century AD. Most of the forms present are more indicative of BB2, and if not actually from known BB2 production sites in Essex or Kent, represent something of the output of perhaps a number of as yet unknown sites producing black burnished fabrics.

Black eggshell ware (BLEGG)

Description: NRFRC code GAB TN 2
Date: Late 1st century to early 2nd century AD
Forms: Beaker H1
Occurrence: This black, thin-walled fabric was present in three contexts (4138, 4183 and 17255) of late 1st to early 2nd century date. A carinated beaker with a wide, flat, decorated cordon, as yet unparalleled, was identified in addition to the everted-rimmed and round-shouldered H1 beaker. The H1 beaker is not attested in eggshell terra nigra and these sherds may be British-made.

Black-surfaced wares (BSW and BSWM)

Description: Chelmsford Fabrics 34 and 45
Date: Early/mid-1st to late 4th century+
Forms: Platters A1 A2 A4 Cam 32 Cam 33, dishes B1 B2 B3 B4 B5 B6 B7 B8, bowls C1 C6 C12 C13 C14 C16 C22 C27 C28 C29 C32 Cam 44 Cam 211 Cam 212 Cam 246 EF59 EF62, mortaria D3 D14, bowl-jars E1 E2 E3 E5 E6, jars G3 G5 G7 G8 G9 G12 G16 G17 G18 G19 G20 G21 G22 G23 G24 G25 G26 G28 G29 G30 G36 G38 G40 G42 Cam 218 Cam 220 Cam 227 Cam 231 Cam 232 Cam 249 Cam 256 Cam 258 Cam 259 Cam 260 Cam 264 Cam 266 Cam 296 EF127 EF169, beakers H1 H4 H7 H10 H14 H24 H25 H26 H27 H32 H33 H34 H35 Cam 91 Cam 117, flagon Cam 140, lids K1 K3 K4 K6, strainer M2, funnel N2, miniatures R2 R3 EF210
Occurrence: Going (1987, 7-9) provided a later 1st and 2nd century AD date range for local sand-tempered black-surfaced wares at Chelmsford. He saw these 'Romanizing' wares as a short-lived continuation of grog-tempered fabric, as the range of forms was similar and the black-surfaced fabrics contained grog in addition to sand. From the 2nd century onwards, these wares were replaced by sandy grey wares (Chelmsford Fabric 47), though black-coloured pottery continued to be produced and at Chelmsford was included in this later category. Similarly at Elms Farm, black-surfaced wares were recovered in considerable quantity throughout the Roman period, forming part of a tradition of reduced ware production in East Anglia. As the fully sandy fabric cannot be easily distinguished from the grog-and-sand-tempered fabric, both variants have been assigned to this single category. Black-surfaced ware occurred in very small quantities during the first half of the 1st century AD. Volume increased considerably during the second half of the 1st century, but fell during the mid-2nd century. From then, the fabric continued to form a major component of most assemblages, though its proportional share rarely exceeded that of the sandy grey wares. Black-surfaced ware was widespread, with most occurring in the central settlement zone, the least in the northern and southern zones. Vessels in this fabric were manufactured at Heybridge.

Buff wares, unsourced (BUF and BUFM)

Description: Chelmsford Fabric 31
Date: Mid-1st to mid-3rd century AD
Forms: Dishes B6 B7 B10, bowls C1 C22 C26, mortaria D1 D2.1 D D4 D5.3 D11 D13 D14 Cam 193, bowl-jar E2, jar G26, beakers H1 H17 Cam 117, flagons J3 Cam 148 EF193, unguentarium Cam 389B
Occurrence: Buff ware was recovered from across the settlement in contexts largely dating to the 2nd century and first half of the 3rd. The highest proportion, largely comprising mortarium sherds, was recovered from the peripheral zone. Buff ware mortaria were almost certainly manufactured at Heybridge their fabric is not easily distinguished from other buff ware fabrics. Mortaria were well represented throughout the settlement. The Area W kiln contexts almost exclusively contained hammerhead-rimmed types. A more varied range of forms was present elsewhere. The central zone also yielded large amounts of buff ware. Colchester and other parts of East Anglia may be among the sources.

Cam 114 herringbone beakers (CAMF)

Description: A white or buff (10YR 8/1) fine-grained, sand-tempered fabric with external mica-coating and internal (sometimes also external) red-brown (10R 4/8) slip, decorated with zones of herringbone-pattern barbotine.
Date: Early to mid=1st century AD (Tyers 1996, 164)
Form: Beaker Cam 114 [GB25 (Stead and Rigby 1989, 134-7)]
Occurrence: A Gallo-Belgic fabric typically found in Britain c. AD 10-40. The earliest examples have contrasting clay slips, but later vessels tend to have only the mica-coating with a paler fabric. These later vessels are the more common British site finds, although sherds with red slip occurred at Elms Farm, unfortunately residual here. Very small amounts of the fabric were recovered. Three examples were present in pits assigned to the first half of the 1st century AD, but much is residual, appearing as single body sherds in contexts dating to the mid-2nd century and later.

Campanian amphoras (ABSAN)

Description: NRFRC code CAM AM 1
Date: 1st century AD
Forms: Dressel 1 Dressel 2-4
Occurrence: This distinctive fabric had a brief floruit at Heybridge. It first occurred in quantity early in the 1st century AD and peaked later that century. Distribution across the settlement was not uniform.

Catalan amphoras (ARCAT and AWCAT)

Description (ARCAT): NRFRC code CAT AM
Description (AWCAT): Hard, cream (7.5YR 8/4) fabric with abundant, poorly sorted, rounded and subangular white inclusions <2mm across, also evident on exterior surfaces. This is the counterpart of the red fabric (ARCAT), although the mica of the standard ARCAT fabric is normally not present. Catalan fabric has been identified and defined by Williams (1981, 127-8).
Date: Late 1st century BC to mid-late 1st century AD
Form: Pascual 1 Dressel 2-4
Occurrence: Very little of this fabric was found, with most occurring in part of the southern settlement zone. Most appeared in contexts dating to the first half of the 1st century AD, with the highest incidence in the mid-1st century.

Central Gaulish colour-coated ware (CGCC)

Description: NRFRC code CNG CC 1 and CNG CC 2
Date: Mid- to late 1st century AD
Form: Beaker H (Greene 1979, fig. 17.4/18.1)
Occurrence: Three sherds were found, with both white and cream fabric variants represented. The cream-buff ware (CNG CC 2), was recovered from pit 7575, present in the form of a beaker decorated with hairpin pattern barbotine, to which a mid- to late 1st century AD date may be applied. A late 1st-century pit, 20174, produced a beaker rim in white fabric (CNG CC 1), which fits the late 1st to early 2nd century AD date that Greene (1979, 43) proposed for this fabric.

Central Gaulish fine cream-slipped ware (CGFCS)

Description: NRFRC code CNG TR, Standard Fabric 1B (Rigby 1989)
Date: Late 1st century BC to c. AD 20
Forms: Flagons Cam 165 [CL3b (Stead and Rigby 1989, 119-20)] EF200 EF201 (Rigby and Freestone 1986, Types F1 F2)
Occurrence: The micaceous fabric is distinctive, with a cream-coloured kaolin-based slip usually confined to the outer surface. At least twelve vessels were identified, although the fabric occurred, mainly as body sherds, in 118 contexts. There is a near-complete vessel in pit 9611, another in pyre-debris pit 15417, and the lower half of a flagon in pyre-related feature 2195. The rims of two flagons, EF200 and EF201, are reeded, unlike the customary dished, cornice rim of Cam 165. There are only two British parallels for the reeded rim form, the Welwyn Garden City burial (Stead 1967) and the Dorton mirror burial in Buckinghamshire (Farley 1983). Thin-sectioning of the flagon from pit 15417 (KPG5, report in archive) confirms a Central Gaulish source.

Central Gaulish glazed ware (CGGLZ)

Description: NRFRC code CNG GL 1 and CNG GL 2, Fabric 2B (Rigby 1989)
Date: Mid- to late 1st century AD
Forms: Beaker H1, flagon J (Greene 1979, fig. 40.2/3)
Occurrence: This fabric was found infrequently, only occurring in the southern and peripheral zones. Most examples are in the white fabric variant (CNG GL 1), produced in the Allier Valley. A single piece is buff (CNG GL 2), originating in Lezoux (Tomber and Dore 1998, 52-3). All sherds were recovered from stratified contexts, providing a date in the second half of the 1st century AD. Three contexts are dated more precisely to the mid-1st century. Just two forms were identified; flagon and beaker. The latter is decorated with barbotine dots (cf. Greene 1979, fig. 42.14).

Central Gaulish micaceous ware (CGMIC)

Description: NRFRC code CNG TR, Standard Fabric 1C (Rigby 1989)
Date: Late 1st century BC to c. AD 25
Forms: Jar Cam 262, beaker Cam 102 [CJ1 (Stead and Rigby1989, 120)]
Occurrence: This coarse red-brown fabric is distinctive, with a golden micaceous slip normally applied to the rims and shoulders of vessels. Sherds were recovered from fifty-seven contexts, with a minimum of twenty-two Cam 102 beakers identified. This compares well with the seventy-two vessel occurrences at Camulodunum, and the forty-nine vessels found at Skeleton Green, although finds elsewhere are restricted to single figures. Two Cam 262 jars are also represented, one in KPG14, the second, in ditch fill 11261, appears to be in non-standard fabric. Thin-sectioning was unable to confirm a Central Gaulish source for this vessel. The fabric was thinly distributed, appearing mainly in contexts dating to the early 1st century AD.

Central Gaulish Rhenish ware (CGRHN)

Description: Chelmsford Fabric 8, NRFRC code CNG BS
Date: Late 2nd to early/mid-3rd century
Forms: Cup F2, beakers H27 H28 H (Symonds 1992, Gps 9 and 14)
Occurrence: Few forms were recorded, with the assemblage generally composed of body sherds. The identified forms mirror examples found at Chelmsford. Beakers appear to be the dominant form. As at Chelmsford, the ware first occurred during the second half of the 2nd century. While the near-complete folded beaker (Archive 2417, fill 7086) from a late 3rd to mid-4th century ditch section is probably residual, this form was typically produced during the first half of the 3rd century (Symonds 1992, 26). The example here may have remained in use or been curated up to the time of its deposition.

Central Gaulish samian ware (CGSW)

Description: Chelmsford Fabric 60, NRFRC codes LMV SA, LEZ SA1 and LEZ SA2
Date: Late 1st century to late 2nd century AD
Forms: Platters f15/17 f15/17R, dishes f15/31 f18/31 f18/31R f18/31-31 f18/31R-31R f31 f31R f42 f79 f79R Ludowici Tg, bowls f29 f30 f35/36 f36 f37 f37R f38 f44 f78 f81 Curle 11 Curle 15 Curle 21 Curle 23, mortarium f45, cups f27 f33 f33a f33/46 f35 f46 f80 Ludowici Tx, jar ?f68, beakers f64 f72
Occurrence: The bulk of the samian found is Central Gaulish, with most of this originating in Lezoux. The earliest pieces, of which there are seven, are 1st century, but most of the ware from Lezoux is 2nd century. A small number of vessels, mainly the work of one potter, came from Les Martres-de-Veyre. Central Gaulish samian mainly occurred in contexts dated mid-2nd through to mid-3rd century, with high residuality apparent. The range of forms is wide, with less common jar and beaker forms also present, and the lion-spouted mortarium making an appearance.

Central Gaulish Terra Rubra (TRCG)

Description: NRFRC code CNG TR, Standard Fabric 1A (Rigby 1989)
Date: Second half of 1st century BC (Rigby and Freestone 1986, 8)
Forms: Platters Cam 1 Cam 2, dish
Occurrence: This fabric is extremely rare in Britain, representing one of the earliest continental imports along with Dressel 1 amphoras. The fabric is highly micaceous, and a red slip originally covered the surfaces, but this only survives patchily. Ten examples were recorded in contexts dated to the late 1st century BC. The majority of the vessels are platters, but there is a single thin-walled dish (Figure 283, no. 2). A red-slipped finish is thought to be reserved for platters, dishes or cups (Rigby 1989, 119), and the evidence from Elms Farm is in agreement with this observation. Thin-sectioning has confirmed a Central Gaulish source for the Heybridge sherds (see archive).

Céramique à l'éponge (CEP)

Description: Chelmsford Fabric 22, NRFRC code EPO MA
Date: Mid- to late 4th century+
Form: Bowl C (Fulford 1977, fig. 3.1)
Occurrence: Céramique à l'éponge is rarely found in Essex, or in Britain, but a higher proportion of this marbled ware than usual was recovered. Indeed, the twelve sherds recovered here, representing five or six vessels, rivals the sixteen sherds found at Mucking (Going 1987, 6). Mainly present singly in late 4th century+ contexts, the central zone produced the highest amount. A single form was identified, a flanged bowl; some body sherds could well belong to flagons.

Coarse reduced grog-tempered ware (GROGC)

Description: Chelmsford Fabric 53, NRFRC code SOB GT
Date: Late 1st century BC to mid-1st century AD
Forms: Platters Cam 23 Cam 28 Cam 31 Cam 32 A2 A4, dish B8, bowls Cam 47 Cam 230 Cam 242 Cam 243 Cam 250 Cam 253 C1 C29 C32 C33 EF28 EF32 EF34 EF35, jars Cam 218 Cam 219 Cam 220 Cam 229 Cam 232 Cam 234 Cam 249 Cam 254 Cam 255 Cam 256 Cam 257 Cam 258 Cam 259 Cam 260 Cam 263 Cam 264 Cam 266 Cam 267 Cam 270 Cam 271 Cam 272 G1 G3 G4 G5 G8 G17 G18 G19 G20 G21 G23 G30 G31 G44 G45 G45 EF92 EF97 EF99-101 EF105 EF107 EF112 EF116 EF119 EF121-3 EF128 EF130 EF133-4 EF136 EF148 EF152 EF176 EF180-5, beaker H4.1, lids K1 K3 K6 EF205, funnel N2
Occurrence: Assigned to the fabric of grog-tempered storage jars and other sherds that have lumpy coarse fabric. Jars are the most common form and of these, storage jars Cam 270 and Cam 271, and the everted rim form Cam 259, are the most numerous. Jars with stabbing on the shoulder, and with various schemes of combed decoration on the body, were also popular. A variety of platters, bowls and lids were also made, and there are three examples of the funnel N2; one with an internal strainer-plate at the junction of the spout and the bowl of the vessel (see Figure 338, no. 5). Also present is a probable tripod foot (see Figure 295, no. 13; cf. Niblett 1985, fig. 32, no. 245) and a second possible example (see Figure 295, no. 12), although this may be a lug handle. The fabric was uniformly distributed, appearing mainly in Late Iron Age contexts. Its occurrence in contexts dated mid- to late 1st century AD perhaps demonstrates the gradual change from fully grog-tempered vessels to those in storage jar fabric (STOR) (see Pattern of pottery upply).

Colchester buff ware (COLB and COLBM)

Description: Chelmsford Fabric 27, NRFRC code COL WH
Date: Mid-1st to mid-3rd century
Forms: Bowls C12 C26, mortaria D1 D2 D4 D11 D13 Cam 191 Cam 192 Cam 195 EF75, beaker H13, flagons J3 J4 J6 Cam 140 Cam 154 Cam 163 Cam 167, funnel N2, costrel
Occurrence: The fabric was widely distributed across the settlement. Going detected a decline in the supply of Colchester products, other than mortaria, to Chelmsford after the mid-2nd century AD (1987, 7). This trend can be seen at Heybridge, where, up until the mid-2nd century, flagons and, to a lesser extent, bowls were more common than mortaria. During the second half of the 2nd century, mortaria production increased but at the same time supplies of flagons and bowls declined. The Colchester industry produced wall-sided gritless mortaria (Cam 191) with a pronounced bead rim and these are probably pre-Flavian in date (Hawkes and Hull 1947, 254). Some of these later examples have fine trituration grits.

Colchester colour-coated ware (COLC)

Description: Chelmsford Fabric 1, NRFRC code COL CC 2
Date: Mid-2nd to mid-3rd century
Forms: Beakers H20 H22.1 H23.1 H24.2 H25 H26 H27.2 H28 H29.1 H32 H33 H35.1
Occurrence: Given the proximity of its source, this fabric is poorly represented, forming less than one quarter of the total amount of Roman colour-coated wares. As at Chelmsford, Heybridge received Colchester colour-coated ware from the mid-2nd century AD, or perhaps slightly before. Supply continued into the 3rd century. Despite a slight increase in volume in 4th-century contexts, its appearance from the later 3rd century onwards must be residual. Forms are confined solely to beakers, though the range is wider than at Chelmsford. Beakers decorated with clay pellet roughcasting are by far the commonest and among the earliest of forms. Large folded and barbotine-decorated beakers are typical late 2nd and 3rd century products.

Colchester samian ware (COLSW)

Description: Chelmsford Fabric 60, NRFRC code COL SA
Date: Mid- to late 2nd century AD
Forms: Dishes f18/31 f18/31-31 f18/31R f31 f31R f79, bowls ?f30R f32 f36 f37 f38, mortarium f45, cups f27 f33 f40 f80
Occurrence: Colchester samian ware is well represented, with a maximum of 101 vessels, probably not unexpected given the proximity of Colchester to Heybridge. Most occurred in contexts dating to the late 2nd century, but there is also a high level of residuality. The fabric was evenly spread across the settlement, with the largest proportion in the central zone.

Colour-coated wares, unsourced (UCC)

Description: Colour-coated fabrics were assigned this code where the source could not be readily identified because sherds were either very small, overfired, or burnt and abraded.
Forms: Beakers H20 H22 H24 H33
Occurrence: Most sherds, if not all, are likely to have originated either in Colchester or the Nene Valley. Forms are confined to beakers, and the majority of sherds were recovered from later 2nd and 3rd century contexts. Both date and form ranges are typical of the industries mentioned.

Early Colchester colour-coated ware (COLCE)

Description: NRFRC code COL CC 1
Date: Mid- to late 1st century AD
Forms: Beaker H1/Cam 94
Occurrence: Sherds representing eight vessels were recovered, including the lower half of a large beaker. Greene (1979, 85) suggested that the fabric imitated Lyon ware, and this is reflected in the forms at Heybridge, which include beakers roughcast with sand. The ware is found mainly in contexts dated mid- to late 1st century AD, although new vessels are unlikely to have been produced after c. AD 55 (Symonds and Wade 1999, 233). A beaker from layer 4706 is decorated with painted vertical stripes, perhaps copying the Cam 114 herringbone beaker (CAMF).

Early shell-tempered ware (ESH)

Description: Chelmsford Fabric 50
Date: Late 1st century BC to early 2nd century AD
Forms: Bowl EF36, jars Cam 254 Cam 255 Cam 256 Cam 258 Cam 259 G3 G4 G5 G (Monaghan Class 3D4), cauldron L1
Occurrence: Vessels in this fabric were made in the Thameside area of south Essex, for instance, at Gun Hill (Drury and Rodwell 1973) and Mucking (Jones and Rodwell 1973). The fabric had its widest distribution in the Late Iron Age, although wheel-thrown ledge-rimmed jars (G5) continued into the Flavian period and north Kent storage jars occurred into the 2nd century. This is much the case at Heybridge, but here the handmade jar form Cam 254 predominates. This simple vessel form is the earliest in date, typologically, and occurred in late 1st-century BC contexts. Early shell-tempered ware was found in moderate quantities across the settlement. A storage jar oven, 10501, contained a Monaghan (1987) Class 3D4 storage jar, current until the mid- or late 2nd century (Tyers 1996, 194). It is suggested that these jars were traded as containers and this may account for the presence of a single example of this vessel type at Heybridge. Uncommon forms include a cauldron, complete with skeuomorphic rivet (see Figure 295, no. 10), and a near-complete, deep, straight-sided bowl, EF36.

East Anglian mortaria (EAM)

Description: NRFRC code EAA RE
Examples from Elms Farm have yellow-grey (2.5Y 8/4; 5Y 7/1) surfaces, but in all other respects are identical to the grey or black fabric held in the National Reference Collection.
Date: Mid-2nd to early 3rd century
Forms: Mortaria D1 D11
Occurrence: This fabric is present only in small quantities. Stratigraphy, and the forms present, predominantly the hammerhead-rimmed D11, strongly suggest mid-2nd to early 3rd century usage.

East Anglian stamped wares (EASTA)

Description: Two fabrics, while superficially dissimilar, both display stamped decoration, and are presented together here for the sake of convenience. One has a grey-brown surface (2.5Y 5/2) with darker grey-brown core (2.5Y 3/2) and orange (5YR 6/6) margins, the other a burnished orange-brown (5YR 5/6) surface and core.
Occurrence: The reduced fabric is present as a single unstratified body sherd only (see Figure 297, no. 39), decorated with vertical bands of rouletting and linked ring-and-dot stamps (cf. Rodwell 1978, fig. 7.13, no. 101). This sherd is likely to date to the late 1st and early 2nd centuries (Rodwell 1978, 268), perhaps originating in West Stow, where pottery of similar fabric and style was manufactured (West 1990, 76). Dating to the mid-4th century, the oxidised fabric is represented by a near-complete bowl (see Figure 296, no. 35) resembling Oxfordshire form C84. The bowl is decorated with rouletted arcs, vertical lines and stamped circles. Its source is uncertain, but an origin in Oxfordshire or the Hadham region is unlikely on the grounds of fabric and decorative scheme.

East Gaulish Rhenish ware (EGRHN)

Description: Chelmsford Fabric 9, NRFRC code MOS BS
Date: Later 2nd to mid-3rd century
Forms: Beakers H (Symonds 1992, Gps 32, 33, 34, 35, 37, 38, and 39), Cam 342
Occurrence: In contrast to Chelmsford, this fabric was more common than Central Gaulish Rhenish ware (CGRHN). It appeared first in contexts dating to the second half of the 2nd century AD, although quantities are low. There was no significant change until the mid-3rd century, when the amount of East Gaulish Rhenish ware increased. Despite the presence of a near-complete vessel in late 3rd century+ pit 4989, it seems likely that the fabric was imported in significant volume only during the 3rd century, and mainly during the middle of this century. The near-complete vessel suggests continued use beyond this date. Forms are confined to beakers, usually folded or long-necked, and often decorated with a variety of rouletted, painted and barbotine motifs.

East Gaulish samian ware (EGSW)

Description: Chelmsford Fabric 60, NRFRC codes BLW SA, HGB SA, MAD SA, RHZ SA and TRI SA
Date: Late 2nd to mid-3rd century
Forms: Dishes f18/31 f18/31R f31 f31R f32 f79 Ludowici Tr/Ts, bowls f30 f35/36 f36 f37 f38 f44, mortarium f45, cups f33 f40
Occurrence: Most (74%) of the East Gaulish samian originated from the Rheinzabern industry, with further small amounts coming from Trier, Blickweiler, Heiligenberg and La Madeleine. The ware from the latter dates to the 2nd century, probably accounting for the small amounts of East Gaulish samian appearing in contexts of this date. However, the fabric mostly occurred in mid-3rd century, or later, contexts, with a rise in numbers towards the end of the 4th century, though presumably residual by this time. The range of forms is restricted to dishes and bowls, although the cup form f33 is present, as always, and the lion-spouted mortarium f45 is well represented. The fabric was evenly spread across the settlement.

Fine grey wares (GRF)

Description: Chelmsford Fabric 39
Date: Mid-1st to late 4th century
Forms: Platters A1 A2 A4 Cam 28, dishes B1 B2 B3 B4 B5 B6 B7, bowls C1 C2 C3 C7 C8 C12 C15 C16 C23 C27 C29 EF57 EF64, bowl-jars E1 E2 E3 E5 E6, jars G5 G8 G9 G10 G16 G17 G18 G19 G20 G23 G24 G29 G31 G35 G37 G38 G40 G42 Cam 232 Cam 249 Cam 259, beakers H1 H2 H5 H6 H7 H8 H10 H21 H24 H25 H26 H27 H32 H33 H34 H35 H39 H41 H42 Cam 96, flagon J3, lids K3 K6
Occurrence: While never as common as the coarser sandy grey ware (GRS), fine grey ware nevertheless formed a major proportion of the reduced wares. Large quantities were recovered from the central settlement zone. Production, probably local, did not commence much before the mid-1st century AD, increasing from the late 2nd or early 3rd century, perhaps in response to the expansion of the market for fine, burnished, pottery, driven by the Hadham industry. This phenomenon was similarly noticed by Going at Chelmsford (1987, 8). The levels of fine grey ware decreased during the 4th century; the upturn at the end of the century suggesting high levels of residuality. Unsurprisingly, dishes, bowls and beakers are well represented. Jars, too, are present, but the range of forms is not as wide as that in sandy grey or black-surfaced wares. Vessels in fine grey ware were manufactured at Heybridge.

Fine red-surfaced grog-tempered ware (GROGRF)

Description: Chelmsford Fabric 53, NRFRC code SOB GT
Date: Late 1st century BC to mid-1st century AD
Forms: Platter, dish, bowl Cam 246, cup Cam 57, jars Cam 249 G2 G8 EF172, beakers Cam 85 Cam 115 Cam 116 H1 H7 EF186, flagon EF198, strainer bowl M1
Occurrence: Assigned to finer vessels with red-coloured and well-finished surfaces, usually beakers imitating terra rubra types. The fabric variation has been fully described by Thompson (1982, 22). Vessels in this fabric were relatively common, and some resembled terra rubra closely enough to be described as TR4 at Camulodunum (Hawkes and Hull 1947, 204). The most common form is the beaker, but a variety of other forms is also represented. The fabric, as with the sandier version, was not uniformly spread across the settlement.

Fine reduced grog-tempered ware (GROG)

Description: Chelmsford Fabric 53, NRFRC code SOB GT
Date: Late 1st century BC to mid-1st century AD
Forms: Platters Cam 21 Cam 22 Cam 23 Cam 24 Cam 26 Cam 28 Cam 32 Cam 31 Cam 33 A2 A4 EF1-17, dishes B7 EF25 EF26, bowls Cam 44 Cam 47 Cam 211 Cam 212 Cam 217 Cam 230 Cam 252 Cam 253 C12 C15 C29 C33 EF30 EF33 EF37-56 EF58 EF60 EF61 EF65-8 EF70 EF72, tazza-bowl Cam 210, cup Cam 57, jars Cam 202 Cam 204 Cam 218 Cam 220 Cam 229 Cam 231 Cam 232 Cam 233 Cam 234 Cam 235 Cam 249 Cam 254 Cam 255 Cam 259 Cam 260 Cam 264 Cam 267 Cam 270 Cam 271 G3 G4 G5 G8 G16 G17 G18 G19 G20 G21 G30 G38 EF76 EF79 EF81-2 EF86 EF88-91 EF93-5 EF103-4 EF106 EF109-11 EF113-5 EF117-8 EF120 EF124-5 EF129 EF131-2 EF135 EF137-47 EF149-51 EF153-68 EF170-1 EF173-5 EF177-9, beakers Cam 85 Cam 109 Cam 115 Cam 117 Cam 118 Cam 119 H1 H3 H7 EF188 EF189, flagons Cam 136 EF196 EF197, lids K1 K3 K6 EF202-4, strainer bowls M1 M2, funnel N2, miniatures EF209 EF211-2
Occurrence: This fabric code is assigned to finer grog-tempered vessels that are well finished, usually with burnished or polished surfaces. The fabric was well distributed across the settlement. Platters, bowls, jars and beakers proliferate, although other forms were also produced in some numbers. The platters are mainly copies of terra nigra types, as is the tazza-bowl Cam 210. Many of the jars and beakers are decorated with cordons and some jar and bowl forms have pedestal bases. Although no complete pedestal vessels were found, a large number of pedestal fragments were recorded, demonstrating that this type of vessel was present in some quantity. Only one flagon was tentatively identified, as these were usually produced in the red-surfaced fabric variant. Other uncommon forms are strainer-bowls, triple vase cups, and two 'cylinders', which are either handles or tripod feet from a bowl. There are a large number of spindlewhorls made from reused potsherds, accounting for over 80% of the total number of ceramic spindlewhorls found. Of interest are the examples of graffiti, many of which are complex crosses, but there are at least six that may be literate. Thin-sectioning was carried out on several vessels from pyre-debris pit 15417 (KPG5), the results of which pointed to production of grog-tempered pottery on a regional, rather than local, scale.

Gaulish amphoras (AGAUL)

Description: Chelmsford Fabric 56, NRFRC code GAL AM 1
Date: Mid-1st to 2nd century AD
Forms: Gauloise 3 Gauloise 4 Gauloise 6
Occurrence: Vessels in this fabric first reached Heybridge in the mid-1st century AD and remained current until the 2nd century, after which the material present (despite its quantity) is apparently residual. It was not uniformly distributed; the highest incidence is in the central settlement zone. Very little occurs elsewhere, and none at all appears in the peripheral zone.

Grey-green mortaria (GRM)

Description: Chelmsford Fabric 49
Date: Mid-2nd to early/mid-3rd century
Form: Mortarium D11
Occurrence: Five sherds from a total of nine came from the Area W kilns, which were backfilled during the early 3rd century. The hammerhead-rimmed D11 form is also consistent with a later 2nd and early 3rd century date. This is somewhat later than the AD 120/25 to 160/75 date that Going provided at Chelmsford (1987, table 9).

Hadham black-surfaced ware (HAB)

Description: Chelmsford Fabric 35, NRFRC code HAD RE 2
Date: Late 2nd to late 4th century+
Forms: Dishes B1 B2 B4 B3 B5 B6, bowl C8, bowl-jars E1.1 E2 E3 E5.3 E5.4, jars G9.3 G19 G40, beaker H34
Occurrence: Heybridge received this fabric from the late 2nd century AD, continuing to do so until the end of the Roman period, although apparently in declining volume from the early 4th century in marked contrast to Chelmsford. Dishes, closely followed by bowl-jars, were the main products in this ware. Some vessels were decorated with external and/or internal burnished lines. Identification by surface treatment alone may not be enough to separate Hadham wares from other black-surfaced wares produced in the Essex and Hertfordshire region. This may account for the unexpected fall off in the incidence of this fabric in the 4th century.

?Hadham fine-slipped red ware (FSR)

Description: Chelmsford Fabric 18
Date: Mid- to late 2nd century AD
Forms: Bowl C1.2
Occurrence: Twelve sherds of this fabric were recovered, although only a single form was identified, a bowl with a reeded flange from pit fill 8802. The very fine appearance of the fabric prompted Rodwell (1978, 260-2) to suggest a Hadham source. Thin-sectioning of the bowl rim sherd has indeed confirmed this source for the fabric (report in archive).

Hadham grey ware (HAR)

Description: Chelmsford Fabric 36, NRFRC code HAD RE 1
Date: Late 2nd/early 3rd to late 4th century+
Forms: Dishes B1.3 B2 B B4.2 B5.1 B6.1, bowl C8, bowl-jars E2.2 E5.4 E6.1, jars G9 G36, beakers H34 H35, flagon J3
Occurrence: As with Hadham black-surfaced ware, this fabric was recovered from all parts of the settlement. While overall quantities are lower, the range of forms is similar, again with dishes and bowl-jars predominating. Hadham grey ware first appeared during the second half of the 2nd century AD, but did not reach Heybridge in any significant quantity until the early 3rd century. Volume increased towards the end of the 3rd century, declining from the mid-4th century.

Hadham oxidised ware (HAX and HAXM)

Description: Chelmsford Fabric 4, NRFRC code HAD OX
Date: Mid-3rd to late 4th century+
Forms: Dishes B1 B10, bowl C8, mortarium D12, bowl-jars E3 E5.2 E6.1, jars G24 G26 G32 G38, beaker H17, flagons J8 J9 J (Symonds and Wade 1999, fig. 5.56, no. 157)
Occurrence: Large quantities of this fabric were present, except in the peripheral settlement zone where the fabric was scarce. While a later 3rd century date has been posited for its widespread distribution across parts of Essex (Going 1999a, 297), Hadham oxidised ware reached Heybridge earlier, almost certainly during the mid-3rd century. It continued to arrive in the settlement with no notable change in volume until the end of the Roman period. Forms include shallow dishes, flanged bowls, wide-mouthed bowl-jars and flagons. So-called 'Romano-Saxon' decoration (Roberts 1982), comprising stamped and bossed motifs, was present on a number of vessels, all likely to date to the second half of the 4th century.

Hadham white-slipped grey ware (HAWG)

Description: A grey (5Y 5/1) fabric with red-brown (5YR 5/6) margins and cream (2.5Y 8/4) exterior slip. As with all Hadham wares, the fabric is typified by inclusions of white quartz and black iron-rich grains, visible microscopically. Date: Early 2nd to late 4th century+
Forms: Beaker H1, flagon J3
Occurrence: This fabric is the least common of the Hadham wares, although it may be under-represented owing to poor survival of the slip. A large portion of a flagon was recovered from layer 6053, otherwise the fabric was thinly distributed.

Hadham white-slipped oxidised ware (HAWO and HAWOM)

Description: Chelmsford Fabric 14
Date: Early 2nd to 3rd century
Forms: Mortarium D5.3, jar G26, beaker H17.1, flagons J3.3 J6
Occurrence: This fabric was widespread across the settlement. Flagons were common, along with frill-rimmed jars. Face-jars, found rarely in any fabric, can also be assigned to this ware. Going suggests that the white-slipped oxidised fabric was principally a mid-2nd century product (1987, 5; 1999a, 297). The evidence here tends to support this view, although the fabric also appears in later contexts, suggesting continued, but smaller scale production, at least as far as the 3rd century.

Highgate grey ware (HGG)

Description: Chelmsford Fabric 37, NRFRC code HGW RE C
Date: Late 1st to late 2nd century AD (Davies et al. 1994, 82)
Forms: Beakers H1 H34
Occurrence: Sparsely represented, the fabric appeared first in a late 1st century AD pit, and then in features dating to the second half of the 2nd century. The H34 folded beaker in pit 16189, a form not attested in London or Chelmsford, is unusual in this fabric, and may represent the latest period of production.

'Hollow foot' amphoras (AKAPT)

Description: Chelmsford Fabric 59, NRFRC code P&W AM 47
Date: Late 4th century+
Form: Kapitän 2
Occurrence: The fabric is rare in Britain, only occurring from the later 3rd century until c. AD 400. It is represented at Heybridge by just two sherds from the same vessel, and from the same context, in well 5806.

Imported buff ware gritless mortaria (IBUFM)

Description: Fine, powdery, very pale buff (10YR 8/2) fabric, with occasional quartz inclusions and flecks of mica at the surfaces.
Date: Late 1st century BC to early 1st century AD
Forms: Mortarium EF74 (Cam 191-type; Partridge 1981, fig. 79.7)
Occurrence: Four examples of this wall-sided gritless mortarium were recorded. The form is probably modelled on stone mortars, so the earliest examples do not have a bead rim at the top of the collar (Hartley 1981, 196). The vessel interior is smooth with no trituration grits. The exterior is frequently rilled. The example from pyre debris pit 15417 has a very small bead and appears to be one of the earliest in the series, likely to have been imported from Italy (K. Hartley pers. comm.), although production in the Rhineland is also possible. Thin-sectioning (report in archive) has confirmed a continental origin for three of the vessels (those from pits 4285, 14225 and 15417), the fourth, from ditch 11100, is more likely to have been locally made.

Imported mica-dusted fine wares (IMIC)

Description: Chelmsford Fabric 11
Date: Mid-1st to early 2nd century AD
Forms: Bowl C (Marsh 1978, fig. 148, no. 14.3), beakers H1 H (Marsh 1978, fig. 6.10, no. 22.6), balsamarium Cam 196
Occurrence: This fabric group appeared principally in mid- to late 1st century AD contexts; a small sherd was recovered from a 2nd-century deposit, but is possibly residual. A total of six examples was found including a very fine thin-walled, everted-rimmed beaker, a product typical of the Braives workshop, and a bowl with paint or mica-coating over the rim. Identification of the bowl, from layer 5603, is uncertain. That it may be a 'Roanne' bowl imported from Gaul (cf. Tuffreau-Libre 1992, 50) remains a strong possibility (P. Tyers pers. comm.). The balsamarium foot (Cam 196) recovered from pit 11344 is an exception in this fabric group. The form is a fairly common occurrence in continental burials of the 1st century BC and is therefore a probable Gaulish import dating to the Late Iron Age.

Imported thorn beakers (THORN)

Description: Hard, fairly sandy fabric, pale orange-brown (7.5YR 6/6) in colour with a grey core. Flecks of mica are visible on the surfaces.
Date: Late 1st century BC
Forms: Beaker H (Greene 1979, fig. 1.3)
Occurrence: Vessels in this fabric are thin-walled, wheel-thrown and decorated with a series of short, barbotine, vertical ridges. Two vessels are represented, the first has the remains of a cream slip externally, the second is in finer fabric, unslipped and with finely applied barbotine 'thorns'. One of the precursors of an Early Roman fine beaker tradition, these beakers are considered by Greene (1979, 4) to be Italian in origin and Republican in date, although the dating extends into the Augustan period. Vessels have been found in southern France (Tyers, in litt.), although site finds in Britain appear to be extremely rare. The sherds recovered are from two contexts dated to the late 1st century BC (see Vessels of intrinsic interest, Figure 295, nos 7 and 8).

Italian amphoras (AITAL, AITAB-AITAJ)

Description: NRFRC code CAM AM 2
Date: Late 1st century BC
Form: Dressel 1 Dressel 2-4
Occurrence: This amphora fabric, and its nine variants, forms the second largest amphora group. Full details for each of the variants can be found in the archive. Although imports to Britain began in the first half of the 1st century BC, the fabric did not occur in any quantity at Heybridge until the last quarter of that century. Most of the fabric was found, residually, in deposits of early 1st century AD date. It occurred in all areas, although not uniformly spread throughout. Parts of three burnt Dressel 1 amphoras were found in pyre debris pit 15417, and other burnt sherds were recovered from several pyre sites.

Late shell-tempered ware (LSH)

Description: Chelmsford Fabric 51, NRFRC code ROB SH
Date: Mid- to late 4th century+
Forms: Dishes B1 B5.3 B6, jars G27.1 G27.2
Occurrence: At Heybridge, this fabric first appeared during the mid-4th century, but was never common until the late 4th century. It is possible that small amounts of this fabric arrived at a slightly earlier date, but more often than not, any sherds in contexts earlier than c. AD 340/50 can be shown to be intrusive. The sources of this fabric are probably much the same as Going (1987, 10) has suggested; Harrold in Bedfordshire, the lower Nene Valley, or Lakenheath. The range of forms is broadly similar to that at Chelmsford and Colchester. Jars are by far the most common vessel class; by comparison, dishes are rare.

London-Essex stamped ware (LESTA)

Description: Chelmsford Fabric 19
Date: Late 1st to early 2nd century AD
Forms: Bowls C12.1 C23.3
Occurrence: This oxidised fabric was found principally in contexts dated late 1st to mid-2nd century. As the fabric is typically no later than early 2nd century, it is likely to be residual in later features. Forms are limited to two bowl types, both based on samian f30; decoration comprises a variety of stamped ring and lattice motifs. Very little of the fabric was recovered, though of note is a large part of a single C23 vessel found in ditch 10159 (KPG22). A possible Hadham source, based on distribution, was suggested by Rodwell (1978, 243) and Going (1999a, 297). Thin-sectioning of sherds from four contexts has confirmed a Hadham source for the fabric (report in archive).

London-type ware (LOND)

Description: Chelmsford Fabric 33, NRFRC code LON FR
Pottery of uncertain origin but similar to the London style in terms of decoration or fabric is also included here. In these cases, the fabric, always finely made, may be oxidised, having red-brown (5YR 5/4) surfaces and core, or reduced with dark grey or black surfaces (10YR 8/1, 2.5Y 3/0) and light (5Y 4/1) or dark (5Y 2.5/1) grey core. Sand tempering is usual, though one sherd, decorated with incised lines and compass-scribed circles, is grog-tempered.
Date: Late 1st to mid-2nd century AD
Forms: Bowls C4 C10.2 C12 C16 C (Monaghan 1987, Class 4H1.1), beaker H1
Occurrence: This fabric largely occurred in contexts dating to the late 1st to mid-2nd centuries. It appeared occasionally in late 2nd to early 3rd century contexts, but is likely to be residual by this time. Quantities are generally small. Forms include bowls based on samian f30, and globular beakers. London and North Kent are both likely sources, while the grog-tempered example was possibly made locally.

Lower German (Soller) mortaria (SOLM)

Description: NRFRC code SOL WH
Date: Undated
Form: Mortarium D2
Occurrence: This fabric code refers to specific vessels produced in the workshops of Verecundus at Soller. These mortaria are distinctive heavy vessels with a broad flange, very often stamped with the maker's name. Six large sherds were recovered from two contexts, 5340 and 5385, probably representing a single vessel. While the fabric is undated, as the pieces in one context are residual and unstratified in the other, the form was produced from the mid-2nd to mid-3rd centuries (Richardson 1986, 111).

Lower Rhineland colour-coated ware (LRC)

Description: Chelmsford Fabric 6, NRFRC code KOL CC
Date: Early to mid-2nd century AD
Forms: Beakers H20.1 H20.2 H23 H24
Occurrence: This fabric occurred principally in 2nd century AD contexts. In terms of weight, around twice as much is present in contexts dated to the second half of 2nd century as is present in contexts dated to the first half. This may represent something of a mid-2nd century peak in importation. Parts of the central and southern zones produced the highest amounts, comprising large sherds in good condition. The roughcast beaker was undoubtedly a popular form, although beakers decorated with a variety of barbotine animal and plant motifs are by no means unusual.

Lyon colour-coated ware (LYN)

Description: Chelmsford Fabric 5, NRFRC code LYO CC
Date: Mid- to late 1st century AD
Form: Beaker H (Greene 1979, fig. 14.1)
Occurrence: The fabric, recovered from just four contexts, is uncommon, more so than its imitator, early Colchester colour-coated ware (COLCE). It appeared in two stratigraphically reliable pits, 13824 and 18019, dated late 1st to early 2nd century. As this date extends beyond the pre-Flavian date suggested by Greene (1979, 18), both occurrences are probably residual. Although no rims were present, the sherds, some roughcast with sand, are likely to belong to beakers. This is in contrast to Chelmsford, where the sole form identified was the F1 cup (Going 1987, 5).

Mancetter-Hartshill mortaria (MHM)

Description: NRFRC code MAH WH
Date: Late 2nd to 4th century
Forms: Mortaria D11 (Gillam 280) D (Gillam 284)
Occurrence: This fabric, absent at Chelmsford and present only in small quantities at Colchester, is similarly rare here. It appeared first in features dating to the second half of the 2nd century, though most examples were found in late 4th century contexts. These examples are likely to be residual, given the apparent decline in importation after the mid-4th century (Bidwell and Speak 1994, 210). Two forms were identified, both with hammerhead rims and probably dating to the 4th century, although the rim of one example (Gillam 284; Figure 297, no. 42) is decorated with red paint, a 3rd-century phenomenon (Tyers 1996, 123).

Mayen ware/Eifelkeramik (MEK)

Description: Chelmsford Fabric 54, NRFRC code MAY CO
Date: Late 4th century+
Form: Bowl C (Fulford and Bird 1975, fig. 1.8)
Occurrence: Sherds were recovered from four stratigraphically reliable contexts, 5207, 10683, 15023 and 24198. Fulford and Bird (1975, 179) place the importation of this fabric within a 4th to early 5th century date range. Two sherds came from late 4th century+ contexts, lending some support to this dating. The remaining stratified sherds were found in 3rd-century contexts, and should be considered intrusive. A single form, a bowl (see Figure 297, no. 37), was identified. At Colchester, forms included dishes, bowls and lid-seated jars, and lids and flagons were also attested (Symonds and Wade 1999, 463-7).

Micaceous Terra Nigra (TNM)

Description: NRCFC code CNG TN, Fabric 2A (Rigby 1989)
Date: First quarter of 1st century AD (Rigby 1989, 120)
Forms: Platters Cam 1 Cam 4 EF18 - EF22, bowls Cam 51 Cam 52B, beaker EF192
Occurrence: Vessels in this fabric are among the earliest continental imports and were Central Gaulish products. Forms are limited to platters, cups and bowls, although new platter forms (EF18-22), and a beaker (EF192), were identified (see The vessels). Cam 1 platters are the most commonly found vessel type in this fabric, reflected at Elms Farm with at least thirty-four examples. A substantial part of one platter appeared in pyre-debris pit 15417. A large number of vessels were found in part of the southern settlement zone, including three of the new platter forms, but the fabric was thinly distributed elsewhere. The Cam 210 tazza-bowl is represented only by body sherds, and carinated body sherds from layer 13576 may be from a Cam 50 bowl. Of interest are two spindlewhorls made from platter base sherds and a graffito inside the footring of a further platter (KPG8).

Miscellaneous fine white- or cream-slipped red-buff wares (MWSRF)

Description: Chelmsford Fabric 16
Date: Mid-1st to early 3rd century AD
Forms: Bowl C12, flagon J1
Occurrence: White-slipped sherds whose origins could not be determined were placed in this category. The largest amount was in contexts dating to the late 1st and 2nd centuries, although smaller quantities were found in 3rd century and later contexts. Two forms were identified: a bowl based on samian f30, and a mid- to late 1st-century collared flagon. While fabrics could not be definitely sourced, Colchester is undoubtedly a likely candidate for at least some of them.

Miscellaneous Late Iron Age coarse wares (MICW)

Description: This code is reserved for the fabrics of coarse sand-tempered vessels, which are usually handmade, occasionally wheel-finished. The colour varies from brown through to black, with much variation in individual examples. Inclusions can be varied, but consist mainly of ill-sorted sand, occasional rounded quartz, flint chips, some burnt, and sparse red ?iron oxide. The fabric can also include variable amounts of grog and the remains of vegetable matter.
Date: Mid- to late 1st century BC
Forms: Platter, bowls EF29 EF31, jars Cam 229 Cam 259 Cam 260 Cam 263 EF77 EF78 EF80 EF83-5 EF87 EF96 EF98 EF102 EF108, miniature EF208
Occurrence: The fabric continued handmade Middle Iron Age pottery traditions, with refinements such as wheel-finishing appearing towards the end of the 1st century BC. Handmade jars from Heybridge can be closely paralleled at Kelvedon (Rodwell 1988), where they are dated c. 50-20 BC. Wheel-finished vessels are closer to Camulodunum forms and can be decorated with cordons or shoulder grooves. The fabric was unevenly distributed, with the northern settlement zone providing the most examples and the central zone very few. The highest incidence is in contexts dated to the late 1st century BC, although the fabric occurred into the 1st century AD, probably residually.

Miscellaneous oxidised wares (RED and REDM)

Description: Chelmsford Fabric 21
Date: Roman
Forms: Dishes B1 B6 B10, bowls C8 C12 C17 C22, tazza-bowl Cam 198, mortaria D1 D4 D13, jars G5 G26 G37, beakers H1 H7 H20 H21 H24 H35, flagons J2 J3 J5 J10, lid K5
Occurrence: This category comprises sand-tempered oxidised pottery that could not be sourced with any certainty. Some vessels may be Colchester or Hadham products, and a number of forms are typical of these industries. Most examples were undoubtedly manufactured locally. Pottery was recovered from contexts dating from the Late Iron Age to the late 4th century. Most sherds, however, were found in late 1st to late 3rd century contexts.

Miscellaneous slipped red wares (MSR)

Description: Chelmsford Fabric 17
This category includes a fabric with orange (7.5YR 6/6) surfaces, grey core (7.5YR 4/0) and red (2.5YR 5/8) slip. There are inclusions of fine quartz, and occasional mica is visible on the surface.
Date: Roman
Form: Bowl C7
Occurrence: The fabric occurred mainly as small or abraded body sherds, and found in contexts ranging in date from the later 1st to the late 4th century. A hemispherical flanged bowl dated late 2nd or early 3rd century was identified. While both fabric and form are superficially similar to Oxfordshire red colour-coated and Hadham oxidised wares, the piece may have greater affinity with fabrics encountered in East Anglia (e.g. Caister Fabric RBSL; Darling and Gurney 1993, 161). The upper surface of the flange is decorated with a white-painted zigzag motif (Figure 296, no. 17).

Miscellaneous white- or cream-slipped sandy red wares (MWSRS and MWSRSM)

Description: Chelmsford Fabric 15
Date: Mid-1st to late 3rd century AD
Forms: Tazza-bowl Cam 198, mortaria D11 D13, jars G26 Cam 207, flagons J3 J4, triple vase S2 Occurrence: This fabric group comprises coarse white-slipped sherds of uncertain origin. Much of it was recovered from 2nd and early 3rd-century contexts and, following a sharp decline after the late 3rd century, from contexts dated to the end of the 4th century. The forms mainly comprise mortaria, frill-rimmed jars and flagons, and were typical products of the Colchester and Hadham industries. The sources for some examples might be found there, while others may have been of more local manufacture.

Miscellaneous white-slipped fine and sandy grey wares (MWSGF and MWSGS)

Description: Both fabrics are grey (10YR 5/1) with white (10YR 7/2) slip on exterior surface. The coarse variant (MWSGS) has inclusions of frequent quartz with mica and occasional flint. The fine fabric (MWSGF) is similarly tempered, but is smooth or powdery to the touch.
Date: Late 2nd to late 4th century
Forms: Bowl C16.1, flagon
Occurrence: Just sixteen sherds were recovered, mainly comprising body sherds possibly from flagons. Cordoned neck sherds and a four-ribbed handle, probably from a flagon copying double-handled whiteware flagons (Cam 161 and Cam 163) were recovered from pit 11316. A bowl with a reeded rim was also identified (pit 10018).

Nene Valley colour-coated ware (NVC and NVCM)

Description: Chelmsford Fabric 2, NRFRC code LNV CC
Date: Early 3rd to late 4th century
Forms: Dishes B1.2 B2 B B5.1 B6.1 B6.2 B10, bowls C8 C11 C18 C (Howe et al. 1980, NV97), mortarium D12, bowl-jar E4, jar G (Perrin 1999, fig. 65.275), beakers H21 H23 H24 H25 H27 H28 H32 H33 H34.2 H39 H41 H42.1, flagon J9, lid K7
Occurrence: The colour-coated products of the Nene Valley pottery industry were a major feature at Heybridge. Indeed, the quantity recovered is more than double that of Colchester colour-coated ware, although this may be explained partly by the production in Colchester ware of forms lighter in weight (Symonds and Wade 1999, 278). The fabric was recovered from all parts of the settlement, and first appeared in small quantities during the first half of the 3rd century or at the end of the 2nd. The main period of importation, however, occurred after the late 3rd century, reaching a peak by the second half of the 4th century. A far wider range of forms than those supplied by Colchester was provided in this fabric. Dishes, which were used mainly during the 4th century, are well represented. Jars, not readily associated with the dinner table and therefore not required in a colour-coated fabric, were less popular. Beakers, especially, often carry barbotine or painted decorative motifs.

Nene Valley grey ware (NVG)

Description: A white (10YR 8/1) fabric with moderate and well-sorted clear or white quartz. Surfaces are dark grey (2.5YR 4/0), streaky in appearance.
Date: Mid-3rd to mid-4th century
Occurrence: This is the least common of all the Nene Valley fabrics. The ware first appeared during the first half of the 3rd century, suggesting that the fabric was reaching Heybridge along with other Nene Valley products. The fabric occurred infrequently up to the later 4th century, but is probably residual after the early 4th century, when production declined (Perrin 1999, 78).

Nene Valley painted ware (NVP)

Description: NRFRC code LNV PA
Date: Late 2nd to late 4th century
Forms: Bowl C (Howe et al. 1980, NV97, NV99), jar G (Howe et al. 1980, NV95), flagon J (Howe et al. 1980, NV96)
Occurrence: This fabric, absent at Chelmsford, was found infrequently at Elms Farm. At Colchester, the fabric occurred first between c. AD 275 and 300 (Symonds and Wade 1999, 289). Its initial appearance at Heybridge was very much earlier. Two bowls were found in late 2nd to early 3rd century features, while a near-complete narrow-necked jar or flask was recovered from early to mid-3rd century pit 14758. However, most sherds were from contexts of the late 3rd century onwards.

Nene Valley self-coloured mortaria (NVM)

Description: Chelmsford Fabric 24, NRFRC code LNV WH
Date: Mid-3rd to mid-4th century (Going 1987, 6)
Forms: Mortaria D14.1 D14.2 D (Hartley and Perrin 1999, fig. 79, type M71) D (Perrin and Hartley 1996, fig. 116, type M128)
Occurrence: The fabric was uniformly distributed and occurred in contexts dated mid-3rd to late 4th centuries. Much of it is residual in late 4th century+ contexts. Forms include grooved-flange types common to Chelmsford. Wall-sided and hammerhead-rimmed forms with grooved flanges manufactured in the 3rd century were also found, producing a range of forms corresponding more closely with the range from Colchester.

North-eastern Gaulish mortaria (NEGM)

Description: NRFRC code NOG WH 4
Date: Mid- to late 1st century AD
Form: Mortarium D1
Occurrence: Three sherds from separate vessels were recovered (layer 5603, slot fill 5843 and pit fill 20108). Two are stamped with the potter's name, 'Q. Valerius Veranius', who produced mortaria from c. AD 65 to AD 100 (Hartley 1999, 197). The third is stamped Q. Valerius Se-- (see Mortarium stamps). It is interesting that the three sherds found in this fabric each have maker's stamps; undoubtedly, plain rim and body sherds have been assigned to other fabrics, either Colchester or unsourced buff wares. The similarity between north Gaulish mortaria and those produced in south-east England has led to discussions on their provenance (Hartley 1977, 5-17; 1999, 197).

North-Essex stamped ware (NESTA)

Description: Chelmsford Fabric 20
Date: Late 1st to early 2nd century AD
Occurrence: The presence of this reduced ware was confined to a single stamped body sherd recovered from post-hole 17024, appropriately dating to the late 1st century or first half of the 2nd century.

North Gaulish white fine wares (NGWF and NWGFS)

Description: NRFRC codes NOG WH 1 and NOG WH 3
Date: First half of 1st century AD
Forms: Beakers Cam 113 EF195, flagons Cam 154 Cam 161
Occurrence: The same forms appear in both fabric variations; each contains sand that is more noticeable in the fine sandy fabric, NGWFS. White fine ware vessels generally have a burnished or polished rim and exterior and zones of rouletting. The beaker form Cam 113 occurred frequently in Britain, and this is also much the case in Essex and thus at Heybridge. Joining body sherds in pit 19104 (KPG10) come from a rouletted beaker which has a vertical barbotine line applied over the rouletting. Other forms are restricted to flagons, those represented are the two-handled Cam 161 and the ring-necked Cam 154, although, generally, only handles and body sherds have been identified. The fabric was thinly distributed, but there are higher incidences in the southern settlement zone. Both fabric variants appear occasionally in contexts dated to the late 1st century BC, but were more common in the first half of the 1st century AD.

North Kent grey ware (NKG)

Description: Chelmsford Fabric 32, NRFRC code UPC FR
Date: Late 1st to late 2nd century AD
Forms: Platter A4, dish B10, bowls C1 C2 C22 C (Monaghan 1987, Types 2H2, 4H2, 5B6, 5B0), cup F (Monaghan 1987, Type 6C1), jars G19 G40, beakers H1 H5 H6.1 H6.2 H6.3 H7 H10.1 H13 H26 H (Monaghan 1987, Type 2A2)
Occurrence: Grey ware was a major product of the North Kent industries, and this fabric occurred at Heybridge from the late 1st century AD until the end of the 2nd century. The mid-2nd century was something of a boom period for the industry (Monaghan 1987, 219), and quantities reaching Heybridge dramatically increased at this time. The fabric was widely distributed, with the central zone producing particularly high amounts. A far fuller range of products than is present at Chelmsford is represented here. In addition to the ubiquitous barbotine-dot decorated poppy-head beakers, platters or shallow dishes, and bowls were recovered, along with jars and flasks.

North Kent oxidised ware (NKO)

Description: A fine and soft sand-tempered fabric with frequent silver mica visible on orange-brown (2.5YR 5/6-8 to 6/6-8) surfaces, dark brown or grey (2.5YR 2.5/0 to 3/0) core.
Date: Late 1st to mid/late 2nd century AD
Form: Bowls C6 C12
Occurrence: In contrast to the reduced ware, the oxidised fabric occurred in small quantities. It was recovered from most excavated areas, recorded in late 1st and 2nd century contexts and residually in those of the 3rd century.

?North Kent white-slipped oxidised ware (NKWO)

Description: Fabric as NKO with additional white (10YR 8/2) slip on exterior surface.
Date: Late 1st to early 2nd century AD
Occurrence: Recorded as undiagnostic body sherds only, this fabric is identical to North Kent oxidised ware and, with the exception of a darker core, Hoo ware (Davies et al. 1994, 38). A North Kent origin therefore seems likely.

Oxfordshire parchment ware (OXP)

Description: Chelmsford Fabric 30, NRFRC code OXF PA
Date: Late 4th century+
Forms: Bowl C (1977, Types P17, P24 and P27), bowl-jar E (1977, Type P32), ?beaker H
Occurrence: As at Chelmsford, Oxfordshire parchment ware is far from common. Apart from a possible beaker, all of the forms are either bowls or bowl-jars. The carinated bowl P24 was the most numerous. This is long-lived, current from the mid-3rd century onwards (1977, 87) but, like most Oxfordshire products, only present in late 4th century contexts. The distribution of this fabric was mainly confined to the central settlement zone and part of the northern zone.

Oxfordshire red colour-coated ware (OXRC and OXRCM)

Description: Chelmsford Fabric 3, NRFRC code OXF RS
Date: Mid- to late 4th century+
Forms: Dish B10.2, bowls C8 C (1977, Types C40 C41 C45 C46 C47 C48 C49 C52 C55 C59 C61 C68 C69 C70 C71 C75 C79 C81 C83 C84 C93), mortaria D9.1 D12.2, bowl-jar E4.1, cup F (1977, Type C110), beaker H (1977, Types C25 C37 C38), flagon J (1977, Type C11)
Occurrence: Measured by weight, this is the most common of the latest Roman fabrics, mostly occurring in contexts of late 4th century date and widely distributed across the settlement. The main concentrations were in the central zone. Bowls are the most common class, several of which (1977, Types C46, C52, C61, C79 and C84) were only developed from the mid-4th century onwards. Beakers, a flagon and a cup were also identified. It is also noteworthy that the overall range of forms appears to be far more diverse compared to Chelmsford (Going 1987, 3). Forms that are exclusively 3rd century have yet to be identified in Essex, although one of the bowl forms is dated by Young (1977, Type C59) to the first half of the 4th century.

Oxfordshire white-slipped red ware (OXSW and OXSWM)

Description: Chelmsford Fabric 13, NRFRC code OXF WS
Date: Mid-3rd to late 4th century+
Forms: Bowl C (1977, Type WC3), mortaria D3 D6 D7
Occurrence: As for the red-slipped ware, this fabric was far more common at Heybridge than at Chelmsford, with the main concentration again occurring in the central zone. Nearly all of the identifiable forms are mortaria, which correspond with Young (1977) Types WC4, WC5, WC6 and WC7. At Chelmsford, the fabric was only observed in mid- to late 4th century contexts (Going 1987, 5). At Colchester, the fabric first appeared in contexts dating to c. AD 300. The presence of WC4 and WC5-type mortaria indicates that the white-slipped ware was reaching Heybridge from the mid-3rd century onwards. Continued supply into the period c. AD 350-400+ is indicated by the presence of WC6-type mortaria. Since the bulk of the sherds were recovered from contexts belonging to the late 4th century, it would seem that the main period of importation was from the mid-4th century onwards.

Oxfordshire white ware (OXW and OXWM)

Description: Chelmsford Fabric 25, NRFRC code OXF WH
Date: Mid-3rd to late 4th century+
Forms: Mortaria D5.1 D5.3 D6 D7 D9.1 D (1977, Types M17, M18, M22), flagon J (1977, Types W15, W25)
Occurrence: The fabric was found over much of the settlement. As at Chelmsford (Going 1987, 6), Oxfordshire white ware reached Heybridge from the mid-3rd century onwards. However, most occurred in contexts dated to the second half of the 4th century. Forms comprise mortaria, commonly Young's M22, and include types also present at Chelmsford and Colchester. Flagons were also founh3d at Heybridge; these are not present at either Chelmsford or Colchester.

Pompeian-red wares (PR)

Descriptions: NRFRC code CAM PR 1, Fabric 1 (Peacock 1977a), NRFRC code CNG PR 3, Fabric 3 (Peacock 1977a), Fabric 2C (Rigby 1989), Fabric 5 (Peacock 1977a), NRFRC code IMP PR 6, Fabric 6 (Peacock 1977a)
Date: Late 1st century BC to 2nd century AD
Forms: Platters Cam 17 EF23 EF24, dish B (Peacock 1977a, fig. 3.14), lid
Occurrence: Pompeian-red ware was a widespread, but thinly distributed, pottery type in Britain, and was also extensively copied. Most British site finds are confined to plain-rimmed platters and lids (Cam 17). Peacock (1977a) identified seven fabrics, four of which occurred at Heybridge. The most common with seven examples is Fabric 1, the characteristic 'black sand' fabric from Campania. Forms present are platters; a single Cam 17 and two examples with a bead rim, EF23, an early type, probably Augustan. There are five occurrences of Fabric 3; vessels in this fabric continued to be imported into the 2nd century AD (Tyers 1996, 158). Interestingly, both forms identified, a dish (cf. Peacock 1977a, fig. 3.14) and a platter with a mica-coated, reeded collar, EF24, are unusual. Fabric 5 originated in Colchester, and is represented in the assemblage by a lid and a possible platter. Pompeian-red ware occurred sporadically, with most examples appearing in part of the southern zone. A large section from one bead-rimmed platter came from pyre debris pit 15417 (KPG5), its fabric (Fabric 1) still easily identifiable despite the high degree of burning.

Portchester D (Tilford/Overwey) ware (PORD)

Description: NRFRC code OVW WH
Date: Late 4th century+
Forms: Dishes B3 (Fulford 1975b, Type 109) B6 (Fulford 1975b, Type 87), jar G27 (Fulford 1975b, Type 137)
Occurrence: Portchester D ware seems to be more common at Heybridge than at Chelmsford or Colchester. The bulk of the fabric was recovered from contexts securely dated to the later 4th century, and there is very little evidence for the fabric reaching Heybridge before c. AD 360/70. Most stratified sherds occurred in the central settlement zone. This relatively large amount is to be expected given the scale of late activity in this part of the settlement. Forms are limited to bead-and-flanged or straight-sided dishes, and jars with rilled bodies.

Pulborough samian ware (PULSW)

Description: Chelmsford Fabric 60, NRFRC code PUL SA
Date: Second half of the 2nd century AD
Forms: Cups f27 f33
Occurrence: Very little of this British-made fabric was found, amounting to a maximum of three vessels. Three sherds, possibly from the same cup, came from dump layers 5602, 5607 and 5610, and a fourth came from layer 17196. Thin-sectioning (report in archive) has confirmed an origin at Pulborough.

Red-surfaced grog-tempered ware (GROGRS)

Description: Chelmsford Fabric 53, NRFRC code SOB GT
Date: Late 1st century BC to mid-1st century AD
Forms: Platter Cam 32, dish B10, bowls Cam 43 Cam 211 Cam 246 C29 EF69 EF71 EF73, jars Cam 218 Cam 220 Cam 221 Cam 249 Cam 256 G19, beakers Cam 92 Cam 115 Cam 116 Cam 118 EF187, flagon Cam 168
Occurrence: The fabric code is assigned to grog-tempered vessels with red-coloured surfaces, usually bowls, beakers or flagons, although the fabric is not restricted to these forms. Dishes, bowls, jars and beakers are well represented. Other vessel types appeared as single examples, although where the fabric has been identified it generally occurred as body sherds. Flagons are usually recognised only by their handles and this may explain their apparent low incidence. The fabric was not uniformly distributed, with larger amounts in the northern and peripheral zones.

Rettendon-type ware (RET)

Description: Chelmsford Fabric 48
Date: Late 3rd to late 4th century
Forms: Dish B6.2, bowl C1, bowl-jars E2 E6.1, jars G21 G24.1 G24.2 G24.3 G28 G34 G36 G37
Occurrence: This flint-tempered fabric, as at Chelmsford (Going 1987, 10), first appeared in significant amounts during the late 3rd century AD. It occurred in earlier 3rd-century contexts in very small quantities, but is intrusive in such cases. The bulk of this fabric was recovered from 4th-century features, with the emphasis very much on the second half of the 4th century. While production of the ware in the Moulsham Street kilns at Chelmsford ended by the mid-4th century (Going 1987, 10), the appearance of Rettendon ware at Heybridge after this date indicates continued production elsewhere in Essex. Indeed, manufacture of this distinctive flint-tempered fabric is attested at Elms Farm in the 4th century. The fabric was recovered from all parts of the settlement.

Romano-British mica-dusted wares (MIC)

Description: Chelmsford Fabric 12, NRFRC code ROB MD
Date: Late 1st to early 3rd century AD
Forms: Platter A (Davies et al. 1994, fig. 115.739), bowls C2 C16 C19.2 Cam 41 C (Symonds and Wade 1999, fig. 5.22, no. 144), beakers H1 H2 H26.1, flagon J5, lamp
Occurrence: This fabric, coarser than imported mica-dusted ware (IMIC), was thinly distributed, except for part of the southern zone where it was found in comparatively high quantities. At Heybridge, the fabric initially appeared earlier than at Colchester, where it is dated AD 110 to AD 225 (Symonds and Wade 1999, 245). It occurred first in contexts dated to the mid- to late 1st century AD, but never became common until the end of the 1st and beginning of the 2nd centuries. The ware continued to be deposited up to the early 3rd century, though not in significant volume after the mid-2nd century. Bowls with reeded rims are well represented, as are everted-rimmed beakers. A handle from a bowl or patera (see Vesssls of intrinsic interest, Figure 296, no. 19), and an open lamp or 'crusy' (from pit 15229) were also made in this fabric. The latter is a copy of bronze lamps of Claudio-Neronian date (Greene 1993, 40-1), more commonly found on military sites.

Salazones amphoras (ASALA)

Description: NRFRC codes CAD AM and P&W AM 16
Date: Early 1st century AD to early 2nd century AD
Form: Beltrán 1 (Dressel 7-11)
Occurrence: This fabric was relatively common, although its distribution across the settlement was not uniform. It first appeared in quantity in the first half of the 1st century AD. Quantities remained high until the early 2nd century, although the fabric is probably residual at Heybridge from the later 1st century. The sole form identified is the Beltrán 1, which was superseded by Beltrán 2 during the later 1st century AD.

Samian wares, unsourced (TSG)

Description: Chelmsford Fabric 60
Forms: Platter ?f18, cup f33
Occurrence: This fabric code was assigned to the few sherds that could not be identified to specific fabric type, mainly due to the small size of the sherds concerned.

Sandy grey wares (GRS, GRSM and GRSWSM)

Description: Chelmsford Fabric 47
Date: Early/mid-1st to late 4th century+
Forms: Platters A1 A2 A4 Cam 28 Cam 32, dishes B1 B2 B3 B4 B5 B6 B7 B8 B9, bowls C1 C3 C7 C16 C19 C28 C29 C31 C33, mortaria D3 D4 D7, bowl-jars E2 E4 E5 E6, jars G3 G5 G6 G8 G9 G10 G16 G17 G18 G19 G20 G21 G22 G23 G24 G25 G26 G28 G29 G31 G34 G35 G36 G38 G40 G42 G43 G44 G45 Cam 204 Cam 207 Cam 231 Cam 232 Cam 241 Cam 259 Cam 291, beakers H1 H6 H7 H8 H13 H21 H26 H30 H32 H33 H34 H35 H39 Cam 117 Cam 119, flagons J2 Cam 365 Cam 370, lids K3 K6, cauldron L2
Occurrence: Sandy grey wares appeared during the first half of the 1st century AD, but in relatively small quantities, only occurring in any significant volume after the late 1st century. Sandy grey wares were found in decreasing quantities during the 3rd century and first half of the 4th, after which quantities noticeably rose, although much of the fabric may well be residual. A wide range of forms was produced, with jars predominating. Late Iron Age or Early Roman forms, more frequently produced in grog-tempered fabrics, such as pedestal jars and platters, were made occasionally in sandy grey ware. Flagons and mortaria were rarely produced, perhaps never featuring in many potters' repertoires. Single mortarium sherds occurred in nine contexts. One mortarium sherd, from pit 20193, is white-slipped and it is conceivable that the others were once slipped, too. The bulk of the sandy grey wares were locally made. Vessels in this fabric were manufactured at Heybridge and products from the kilns include bowl-jars and lid-seated jars. The fabric was well represented, with the highest incidence in the central zone.

Silty ware (SILT)

Description: Soft, smooth fabric with few inclusions, light reddish-brown (5YR 6/6) in colour with occasional mica, especially on internal surfaces. Some sherds have a cream (10YR 7/4) slip, although this has not survived well.
Date: Mid-1st century AD
Forms: Beaker H (Rigby 1989, fig. 68, type 6C) EF190 EF191
Occurrence: This fabric code refers to vessels first identified at Verulamium, made in a fine fabric and copying Gallo-Belgic forms (Rigby 1989, 195). The vessels at Heybridge also appear to be copying these forms, in particular Cam 113-type butt beakers, but the fabric, although fine, is not necessarily the Silty Ware identified at Verulamium. Thin-sectioning, however, has revealed that the fabric of these sherds compares well with Verulamium Silty Ware, and perhaps the nuances within the fabric show regional influence (report in archive). Decoration consists of cordons and zones of incised lines in imitation of rouletting. The fabric occurred infrequently in the central and southern zones.

South-East English glazed ware (GLZE)

Description: Chelmsford Fabric 10, NRFRC code SOB GL
Date: Late 1st to early 2nd century AD
Form: Bowl C12
Occurrence: The fabric appeared mainly in contexts dating to after the mid-2nd century AD, and must be residual in all these cases. It was present in just one feature (pit 20066) contemporary with its conventional production date range of late 1st to early 2nd century (Arthur 1978, 300-1). Forms were confined to bowls imitating samian f30, decorated body sherds of which were present.

South Gaulish samian ware (SGSW)

Description: Chelmsford Fabric 60, NRFRC codes LGF SA and MON SA
Date: Mid-1st to early 2nd century AD
Forms: Platters f15/17 f15/17R f16 f18 f18R f19 Ritterling 1, dishes f18/31 f23 f42, bowls f29 f30 f35/36 f36 f37 Curle 11 Ritterling 12, cups f24 f27 f27g f33 f33a f35 ?f46 Ritterling 5 Ritterling 8 Ritterling 9, beaker f67, inkwell Ritterling 13
Occurrence: The bulk of the south Gaulish samian originated in the factories at La Graufesenque, in keeping with the rest of Britain. The few remaining pieces were from Banassac and Montans. The range of forms is large, with a higher than usual number of the globular beaker f67, and an uncommon form, the inkwell, completing the assemblage. Most of the ware appears in contexts of late 1st century and early 2nd century date, with some residuality apparent. South Gaulish samian was evenly distributed across the settlement.

Storage jar fabric (STOR)

Description: Chelmsford Fabric 44
Date: Mid-1st to late 4th century+
Forms: Platters A2 A4 Cam 28, dishes B7 B8, bowls C28 C29 C33, jars G3 G5 G22 G23 G36 G37 G42 G43 G44 G45 Cam 252 Cam 259 Cam 260 Cam 270 Cam 271 Cam 272 EF126 EF178, lids K3 K5, cauldron L1
Occurrence: This fabric was widely distributed, with very large amounts in the central zone. A number of storage jar ovens account for much of this localised large amount. Apart from jars, a range of platters, dishes and bowls was manufactured, most of which derive from Late Iron Age grog-tempered prototypes and whose presence did not extend far beyond the early 2nd century AD.

Terra Nigra (TN)

Description: NRCFC code GAB TN 1
Date: First half of 1st century AD
Forms: Platters Cam 2 Cam 5 Cam 8 Cam 9 Cam 12 Cam 13 Cam 16, cup Cam 56, beaker Cam 120
Occurrence: Vessels in terra nigra were among the latest Gallo-Belgic imports, gradually superseding terra rubra from the Tiberian period onwards. Platters, usually Cam 16, can occur in contexts as late as AD85 (Rigby 1989, 123), although only a single example of Cam 16 was recorded. The form with the highest incidence is the long-lived platter Cam 2 with at least forty-two examples. Platters are by far the most common vessel class, but examples of the cup Cam 56 also occurred, and the Cam 120 carinated beaker was represented by body sherds. This beaker form commonly occurred in Britain from c. AD 60 (Tyers 1996, 166). Three vessels carry makers' stamps, which usually only appear on the later forms. The fabric was thinly distributed across the settlement, with larger amounts occurring in the southern zone.

Terra Rubra (TR)

Description: NRCFC code GAB TR 1A, GAB TR 1B, GAB TR 1C, GAB TR 2, GAB TR 3
Date: Late 1st century BC to mid-1st century AD
Forms: Platters Cam 2 Cam 3 Cam 4 Cam 5 Cam 8 Cam 12/13, dish EF27, cup Cam 56, beakers Cam 78/79 Cam 79 Cam 82 Cam 84 Cam 91 Cam 112 EF193 EF194 (Holwerda 1941, no. 104)
Occurrence: Production of terra rubra probably commenced as early as c. 20 BC, influenced by Arretine pottery types, and was replaced gradually by terra nigra during the first half of the 1st century AD. The fabric has been divided into three main variants, initially by Hawkes and Hull (1947), subsequently modified (Rigby 1973), and more recently for the King Harry Lane site (Rigby 1989, 121-6). These variants are chronologically significant, although most of the terra rubra from Elms Farm has not been so precisely recorded. The fabric occurred thinly across the settlement, with higher concentrations in the southern zone. There is a greater range of forms in this fabric than for terra nigra. Butt beakers (Cam 112) and platters are well represented, although girth and pedestal beakers are few, with just four types recorded. Girth and pedestal beakers are typologically early forms and are thus uncommon British site finds. One butt beaker (EF194) has round bosses applied over the rouletted decoration. Vessels with applied bosses are not common, but several examples have been found in southern Britain.

Unidentifiable pottery (UPOT)

Description: Pottery assigned to this category was generally either very burnt or abraded, or very small, thus making identifications impossible.

Verulamium region buff ware (VRB)

Description: Chelmsford Fabric 29
Date: Late 1st century AD
Form: Beaker H1.6
Occurrence: The fabric was recovered in small quantities in contexts that provide a later 1st century AD date for vessel use and deposition. As at Chelmsford, a single form was recorded, the ring-and-dot beaker. This form was also manufactured at Colchester (Bidwell and Croom 1999, 472).

Verulamium region coarse white-slipped ware (VCWS)

Description: A sandy pink-red (5YR 7/6) fabric with pale yellow or cream (10YR 7/4) slip, thinly applied. Texture and inclusions are identical to VRW.
Date: Early 2nd to mid-3rd century AD
Occurrence: This fabric was recovered from four contexts, two of which are cleaning layers in the peripheral zone. The remaining contexts, layer 13568 and gully fill 13734, dating to the 2nd and 3rd centuries, yielded eight sherds, including a ?flagon handle.

Verulamium region grey ware (VRGR)

Description: A hard grey-white (10YR 8/1) fabric with frequent sand inclusions producing a hackly fracture. The fabric is identical to VRW, save for a dark grey (2.5Y 4/0) exterior surface.
Date: Undated
Occurrence: The fabric was present as body sherds only, possibly from a jar or flagon. It occurred residually in Period 4 pit 16088. Symonds and Wade (1999, 357) suggest that grey-surfaced Verulamium products tended to be among the earliest to be manufactured, a proposal that cannot be substantiated here.

Verulamium region white ware (VRW and VRWM)

Description: Chelmsford Fabric 26, NRFRC code VER WH
Date: Mid-1st to mid-2nd century AD
Forms: Bowls C16 C (Davies et al. 1994, fig. 44.245), mortaria D1 Cam 195A, flagons J3 J6 J11
Occurrence: Products from the Verulamium region, principally mortaria and flagons, reached Heybridge from the mid-1st century AD until the mid-2nd century. The fabric continued to be deposited up to the mid-3rd century, although given the collapse of the Brockley Hill industry in c. AD160 (Going 1987, 19), its appearance in these instances is surely residual. The fabric was recovered from all parts of the settlement. The mortarium was clearly the most popular of Verulamium products, accounting for over half of the total weight. Heybridge provided only a minor market for the industry, with the volume of imports never approaching that attained by the Colchester industry.

White wares, unsourced (UWW)

Description: White fabrics that could not be placed easily in specific white or buff ware categories were assigned this code.
Form: Jar G26
Occurrence: This group mainly comprised body sherds. Small rim and handle sherds, most probably belonging to flagons, are also present. The single form identified is a jar with a frilled rim. While this 3rd and 4th century form is typical of the Hadham industry (Going 1987, 25-6), a local source may perhaps be more appropriately applied here on fabric grounds.

Wine amphoras (AWINA - AWINE)

Description: This fabric group encompasses Dressel 2-4 fabrics that are not shared with Dressel 1 (AITAL, ABSAN, see above). Full descriptions for each fabric variant can be found in the archive.
Date: Late 1st century BC to late 1st century AD
Form: Dressel 2-4
Occurrence: Very little of the assemblage was assigned to this fabric group. Most occurred in contexts of early 1st century AD date in part of the southern settlement zone.

The pottery sequence: ceramic phases and key pottery groups

This section incorporates thirty-nine pottery groups spanning eleven ceramic phases, with one further group representing the immediate post-Roman period. These groups, encompassing seventy-eight contexts, were selected from an initial list of some 300 fully quantified contexts and are among the best in ceramic terms at Elms Farm. They have secure and narrow date-ranges, display low residuality (albeit increasing in the Late Roman phases), and are largely free of contamination. The groups are generally large, usually exceed a total weight of 2kg, and the pottery is in good condition. A further requirement for their inclusion here is that the pottery is illustrated, although Key Pottery Group 29 is the exception. A number of roles are served. An unrivalled collection of fully quantified groups is presented, and these form the basis for detailed comparison with the pottery from other sites. The uniform treatment accorded to both Late Iron Age and Roman groups has facilitated a consistent approach to subsequent analysis. Ordered chronologically, the groups provide 'snapshots' of shifting supply patterns and vessel use through time. Both typical and atypical groups within each ceramic phase are presented here. Typical groups are perhaps best described as conforming most closely to the 'average' for supply or assemblage composition in a given ceramic phase, as described in the next section. Conversely, atypical groups diverge from the average and are illustrated to highlight the diversity of pottery deposits.

Each of the Key Pottery Groups (KPG) has been assigned to a ceramic phase. Within each phase, the groups are ordered to reflect chronological progression. All of the groups follow a consistent format. A table accompanies each, in which the quantified data (sherd count, sherd weight, and EVE) is presented. This is followed by a summary of assemblage composition, dating evidence, and justification for selection. Comments on condition and fragmentation are based on recorded observations and average sherd weights. Most assemblages comprise the pottery from a single context; the remainder are composites of multiple fills within a feature. A summary is provided in Table 2 below. The prefix EF, used to designate Late Iron Age forms which are not attested at Camulodunum, refers to vessels catalogued in the Elms Farm typology. Full quantification of the Key Pottery Groups entailed measuring the extant section of the rim for all the vessels present. This measurement is normally known as Estimated Vessel Equivalence (EVE). The study at Elms Farm has used rim measurements only; bases are not included in the statistics.

The groups presented in this section provide some of the principal, but by no means all, data on which subsequent analysis was based. They have been of most use in drawing out trends in pottery supply and use, which form the basis of the Pottery Supply section. To ensure reliability when conducting this analysis, a pool of other fully quantified, well-dated groups was additionally used, expanding the dataset to 106 groups (183 contexts). These have been excluded from the pottery sequence because the selected KPGs adequately reflect pottery supply and use, or, more simply, because they are not illustrated. Full data for the wider pool can be found in the archive.

A feature of the Key Pottery Group is low residuality. The issue of residuality, itself a major component of the pottery analysis along with other aspects of context formation processes, cannot be dealt with satisfactorily by these groups alone. Studies of residuality, deposition and intra-site comparisons form part of the research archive.

Table 2: The Key Pottery Groups
Ceramic phase Chelm. CP Colch. PEG Key group Feature Contexts Area Date range
1 - - 1 ditch 25094 19116, 19145 P c. 50-30 bc
2 ditch 25252 6875, 6907, 6957 H c. 50-25 bc
3 pit 8786 8785 P c. 30 bc
4 pit 11342 11329 N c. 25-10 bc
2 - 0 5 pit 15417 15416, 15418, 15420, 15490 M c. 10 bc-ad 5
6 pit 11344 11343 N c. 5 bc-ad 10
7 pit 9611 9585 , 9610 D c. AD 1-10
8 pit 11316 11269, 11301 N c. AD 5-20
9 pit 8282 8271 E c. AD 5-20
10 pit 19104 19105, 19107, 19109, 19110, 19111 P c. AD 1-25
3 - 3 11 pit 8026 8003 , 8014, 8018 E c. AD 20-40
12 pit 9230 9231 D c. AD 20-45
13 pit 7167 7168, 7178, 7179 G c. AD 25-45
14 pit 11723 11720 N c. AD 45-55
4 1 4-5 15 ditch 25018 9214 D c. AD 50/55-60
16 pit 9218 9217, 9370 D c. AD 55-65
17 pit 20008 20009 L c. AD 70
18 pit 24013 24014 M c. AD 70-80
5 2 5-8 19 pit 15773 24258 M c. AD 80-100
20 pit 6201 6203 H c. AD 100-120
21 pit 5147 5146 J c. AD 120/25
6 3 9-10 22 ditch 25245 10182 F c. AD 120/5-150
23 pit 7118 7119, 7116 G c. AD 155/60
24 pit 9029 9028, 9064 D c. AD 140-160
7 4 9-10 25 pit 7122 7123 G c. AD 170-200
26 well 6280 16083 H c. AD 190-200
27 stoke-hole 1589 all deposits (20) W c. AD 190-210
8 5 10-12 28 pit 6182 6178 H c. AD 200/10-230
29 pit 16088 16073 H c. AD 210-250
30 pit 10062 10061 E c. AD 250/60
9 6 13-14 31 pit 11303 11302 N c. AD 250/70
32 pit 6267 6268 H c. AD 260-300
33 ditch 25270 12026, 12029 R c. AD 290-310
10 7 14-16 34 pit 14125 4315 K c. AD 310-330
35 pit 8745 8766 P c. AD 310-350
36 pit 10067 10017 E c. AD 350-400
11 8 17-18 37 pit 5209 5210 J c. AD 350-400
38 gully 25079 15056 M c. AD 360/70-400
39 well 5806 5763 I c. AD 370-400+
post Roman - - 40 pit 14529 14528, 14558, 14613 L c. AD 400+

Ceramic Phase 1: c. 50-15 BC

Thirteen groups, with a total weight of 51.5kg, and 21.51 EVE, have been assigned to this ceramic phase, of which four (21kg, 11.39 EVE) are presented below. Most assemblages comprise mainly coarse wares, with the occasional presence of Dressel 1 amphora and Central Gaulish imports. The majority of the groups assigned to the phase came from the southern zone of the settlement. Independent dating evidence occurs in the form of two coins, a Class I/II potin (SF5645) in ditch segment 16018, and a potin fragment (SF6841) in ditch segment 8208. Quantified groups of this date are few but comparisons may be drawn with the pottery types from the Airport Catering Site, Stansted (Going 2004) and Ditch 350 at Kelvedon (Rodwell 1988). Residuality in all groups is negligible, though earlier prehistoric pottery occurs in ditch segment 19144 in the form of three flint-tempered body sherds weighing 30g.

Key Pottery Group 1. Area P: Ditch 25094, segments 19115/19144, fills 19116, 19145, OA3, Group 46 (Figure 229)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE EVE % Forms
ESH 1 8 8 <1% 0.05 6% Jar Cam 254
GROG 57 486 9 24% 0.27 32% Jars EF76 Cam 219 Cam 259
GROGC 31 784 25 39% 0.03 4% Jar Cam 255
MICW 37 744 21 37% 0.49 58% Jars EF77 EF80 EF83 EF85
Total 126 2020     0.84    

This group is representative of the assemblages recovered from contexts dating to the transitional period between the middle and Late Iron Ages, although relatively few of these features were identified. A feasible date for this assemblage is c. 50-30 BC. This is the smallest Ceramic Phase 1 group and is composed entirely of jars, mainly handmade, of which sand-tempered coarse wares form a major component. Imports, including amphoras, are absent. The finer grog-tempered jars are fragmentary.

Key Pottery Group 2. Area H: Ditch 25252, segment 16018, fills 6875, 6907, 6957, OA5, Group 63 (Figure 230)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
AITAB 1 314 314 7%     Dressel 1
GROGRS 4 16 4 <1%      
GROG 124 1536 12 36% 1.14 70% Bowl EF41, jars EF79 EF104 EF106 Cam 229 G3
GROGC 111 2049 18 48% 0.35 21% Jar EF107
MICW 24 346 14 8% 0.14 9% Bowl EF31, jar EF98
Total 264 4261     1.63    

Handmade jars again dominate this assemblage, although here the sand-tempered coarse wares form a much smaller proportion, giving way to grog-tempered fabrics. This is a trait that is common for the remainder of the ceramic phase. Dressel 1 amphora is present, but there are no other imports. The coin evidence suggests a date early in the ceramic phase, between perhaps 50 BC and 25 BC. Four small sherds (34g) of Roman grey ware present in the upper fills are clearly intrusive and have been excluded from the table.

Key Pottery Group 3. Area P: Pit 8786, fill 8785, OA3, Group 48 (Figure 231)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
BUF 1 6 6 <1%      
GROG 292 4120 14 51% 3.40 61% Bowl Cam 231, jars EF81 EF82 EF86 EF89 EF110 EF138 EF139
GROGC 64 2170 34 27% 0.13 2% Jar Cam 270
MICW 78 1697 22 21% 1.88 34% Jars, handmade (MIA style)
NGWFS 3 34 11 <1% 0.08 1% Beaker Cam 113
TNM 1 30 30 <1% 0.12 2% Platter Cam 1
Total 439 8057     5.61    

This assemblage demonstrates a greater diversity of form and fabric. There are platters, bowls and beakers present alongside the jars. The latter are still mainly handmade, but now grog-tempered pottery forms the larger component. A small number of imports are present, accounting for just over 3% by EVE. The Central Gaulish Cam 1 platter is among the earliest imports to Britain, and the North Gaulish Cam 113 beaker is typologically early in form. The tiny sherd of buff ware may well be an import, possibly from an amphora. A date of c. 30 BC is proposed for the group. Substantial parts of several vessels are present, allowing for their reconstruction.

Key Pottery Group 4. Area N: Pit 11342, fill 11329, OA3, Group 58 (Figure 232)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
AITAD 2 84 42 1%     Dressel 1
AMISC 2 64 32 1%      
CGFCS 19 126 7 2% 0.08 2% Flagon Cam 165
GROG 180 3228 18 48% 1.46 44% Bowl Cam 211, jars Cam 220 Cam 249 Cam 256 Cam 264 G19
GROGC 100 2708 27 40% 1.08 33% Jars EF119 Cam 255 Cam 266
GROGRS 21 505 24 7% 0.69 21% Jar Cam 256, flagon handle
THORN 1 8 8 <1%     Beaker base
TRCG 1 22 22 <1%     Platter base
Total 326 6745     3.31

Pottery Group 4 is typical of the later part of the Ceramic Phase 1 date range, perhaps 25-10 BC. It comprises a variety of grog-tempered forms, including the handle from a flagon, although jars still predominate. Sand-tempered coarse ware is no longer present in this particular group, and forms a very small proportion for the remainder of the phase in others. Imports are few, mainly Dressel 1 amphoras, while Central Gaulish vessels are more prevalent. The latter were imported into southern Britain from c. 25 BC (Rigby 1986b, 270). Increased diversity of form is apparent, typical of assemblages towards the final decades of the 1st century BC and later. The average sherd weight is generally high, but the flagon is fragmentary, demonstrating the fragile, thin-walled nature of these Central Gaulish vessels.

Ceramic Phase 2: c. 15 BC-AD 20

Fourteen groups, with a total weight of 155.8kg and 52.68 EVE, have been assigned to this ceramic phase. Six of these (101.5kg, 34.81 EVE) are presented below to illustrate the pottery which is characteristic of the phase, although the pottery from pit 15417 is, in many ways, atypical. Ceramic Phase 2 groups came mainly from the northern and southern settlement zones. Few assemblages of similar date have been quantified, but groups B50, C5 and C27 from Puckeridge-Braughing in Hertfordshire (Witherington and Trow 1988) provide comparable data. Assigned to this phase, though not presented here, is a cremation burial assemblage (8177) accompanied by fragments of an iron brooch-and-ring ensemble (SF7019) (see The brooches), which has been dated broadly late 1st century BC to early 1st century AD. In addition, two Colchester brooches (SF2393, SF378), dated c. AD 10-70 (see The brooches) occur in features assigned to this phase, one from pit 7415 and the other from pit 4026. Supporting dating evidence is provided by the presence of Arretine ware. Platter floor sherds from pit 9611 are dated to 15 BC-AD 20 and platter rim sherds from pits 7060 and 10288 are both dated to c. 15 BC-AD 10. Residuality is apparent in the later part of the phase, in particular the Dressel 1 amphora and sand-tempered pottery, and occasional earlier prehistoric sherds. The fragmentary nature of some of the grog-tempered pottery may also indicate residuality, although in this fabric this is harder to assess.

Key Pottery Group 5. Area M: Pit 15417, fills 15416, 15418, 15420, 15490, OA2, Group 33 (Figures 233, 234, 235)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
AITAB 139 10002 72 17%     Dressel 1
AITAC 152 11391M 75 19% 0.79 8% Dressel 1
AITAD 528 20897 35 36% 0.68 6% Dressel 1
CGFCS 37 436 12 1% 0.18 2% Flagon Cam 165
ESH 50 930 19 2% 0.76 7% Jar Cam 255
GROG 356 11095 31 19% 5.09 48% Bowl Cam 253, jars Cam 204 Cam 218 Cam 220 Cam 221 Cam 259, beakers Cam 115 Cam 118, flagon EF196
GROGC 7 344 49 1%      
IBUFM 10 464 46 1% 0.01 <1% Mortarium Cam 191-type (EF74)
PR 25 630 25 1% 0.48 5% Platter EF23
TNM 2 128 64 <1% 0.28 3% Platter Cam 1
TR 173 2026 12 4% 1.93 18% Platters Cam 2 Cam 5, beaker Cam 112
MICW 1 26 26 <1%      
NGWF 4 100 25 <1% 0.30 3% Beaker Cam 113
Total 1484 58469     10.50    

This pottery, although almost certainly pyre debris, demonstrates the diversity of fabric and form associated with assemblages deposited at the beginning of Phase 2. This large collection of pottery can probably be dated to the very end of the 1st century BC or beginning of the 1st century AD. Grog-tempered pottery predominates at 48% by EVE, and there are substantial parts of three Dressel 1 amphoras present plus two other vessel types imported from Italy. These, a Pompeian-red ware platter and a wall-sided gritless mortarium, are both typologically early forms. There are also imports from Central Gaul, and Gallo-Belgic ware is now in evidence in the form of terra rubra. The sherd of sand-tempered coarse ware is residual in this feature. Although large and conjoining sherds from many of the vessels are present, the pottery has been burnt to varying degrees, resulting in distortion, discolouration and much splitting along heat-fractures.

Key Pottery Group 6. Area N: Pit 11344, fill 11343, OA3, Group 59 (Figure 236)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ARCAT 1 449 449 5%     Pascual 1
AWINE 1 82 82 1%      
CGFCS 9 42 5 <1%     Flagon handle
CGMIC 36 242 7 3% 0.47 11% Beaker Cam 102
GROG 197 2710 14 29% 2.26 51% Platter Cam 21, dish EF26, jars G19 Cam 219 Cam 249, beaker Cam 115
GROGC 214 5501 26 58% 1.29 29% Jars EF116 Cam 255 Cam 271 G44
IMIC 1 14 14 <1%     Balsamarium foot (Cam 196)
TN 3 52 17 1%      
TNM 4 50 13 1% 0.03 1% Bowl Cam 51
TR 49 320 7 3% 0.36 8% Platter Cam 5, beaker Cam 112
Total 515 9462     4.41

This assemblage again exhibits characteristics that continue from Ceramic Phase 1, demonstrating the continuance of pottery types, such as Central Gaulish ware, into the first decades of the 1st century AD. Wine amphoras Pascual 1 and Dressel 2-4 have made an appearance and the wider range of Gallo-Belgic wares now includes terra nigra. Imports account for 13% of the assemblage by weight, but grog-tempered pottery predominates, as expected. The forms present include bowls and beakers along with the platters and jars. Many of these vessels are copies of imported forms. The finer vessels are fragmentary. The assemblage dates to the end of the 1st century BC into the first decade of the 1st century AD.

Key Pottery Group 7. Area D: Pit 9611, fills 9585, 9610, OA6, Group 75 (Figure 237)
Fabrics Sherd no. Weight (g) Average sherd wt % Weighth EVE % EVE Forms
AITAL 3 391 130 2%     Dressel 1
AWINC 4 120 30 1%     Dressel 2-4
BSW 4 132 33 1%      
BUF 1 8 8 <1%      
CGFCS 36 895 25 6% 0.54 7% Flagon Cam 165
CGMIC 5 44 9 <1% 0.25 3% Beaker Cam 102
GROG 256 3690 14 23% 3.16 40% Platters Cam 21 Cam 22, bowls Cam 230 Cam 214, jars EF103 EF153 EF154 Cam 249 Cam 229 G19 G20, lid
GROGC 318 9763 31 61% 2.23 28% Jars Cam 218 Cam 219 Cam 255 Cam 259 G4, funnel N2
GROGRF 14 74 5 <1%      
GROGRS 2 26 13 <1% 0.23 3% Jar G19
GRS 2 10 5 <1%      
ITSW 3 8 3 <1%     Platter
NGWF 4 16 4 <1%      
NGWFS 5 31 6 <1%      
RED 4 24 6 <1%      
STOR 1 32 32 <1%      
TN 3 44 15 <1%      
TNM 5 374 75 2% 0.57 7% Platter Cam 1
TR 22 238 11 1% 0.85 11% Platters Cam 2 Cam 4, cup Cam 56, beakers Cam 82 Cam 112
PREHIST 1 20 20 <1% 0.02 <1% Small bucket urn
Total 693 15940     7.85

This large assemblage falls in the middle of Ceramic Phase 2, perhaps spanning the first decade of the 1st century AD. Grog-tempered pottery again predominates at 85% by weight, and the range of imports has expanded to include North Gaulish fine white ware and Arretine. Central Gaulish vessels are still in evidence, and there is Dressel 2-4 amphora present along with the Dressel 1, although the latter must be residual by this time. This group contains a wide range of fabrics with a wider range of forms, including a specialised form, a funnel, in grog-tempered ware. Many grog-tempered forms are, again, copies of imports. The few Roman sherds present (fabrics BSW, GRS and STOR) are from the top fill and probably intrusive.

Key Pottery Group 8. Area N: Pit 11316, fills 11269, 11301, OA25, Group 227 (Figure 238)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
AITAD 3 114 38 1%     Dressel 1
ARCAT 2 227 114 3%     Pascual 1
AWINC 2 109 55 1%     Dressel 2-4
CGFCS 2 22 11 <1% 0.19 4% Flagon Cam 165
CGMIC 31 274 9 4% 0.57 12% Beaker Cam 102
GROG 230 2847 12 37% 1.75 38% Bowl Cam 210, jars EF117 Cam 204 Cam 260, beakers EF189 Cam 85 Cam 115, flagon handle
GROGC 160 3414 21 45% 1.62 35% Jars Cam 254 Cam 255 Cam 256 Cam 257 Cam 259
GROGRS 10 158 16 2% 0.06 1% Bowl Cam 43
NGWF 4 14 4 <1%      
PR 1 30 30 <1% 0.06 1% Platter Cam 17
TN 1 36 36 <1% 0.09 2% Platter Cam 2
TNM 11 266 24 3% 0.10 2% Platters Cam 1 Cam 4, bowl Cam 51
TR 13 108 8 1% 0.18 4% Beaker Cam 82
Total 470 7619     4.62    

This group illustrates the diversity of fabric and form typical of the mid- to later part of the ceramic phase (c. AD 5-20). Grog-tempered pottery continues to form the major component at 84% by weight. Platters appear in Gallo-Belgic fabrics only; grog-tempered copies are confined to beaker and bowl forms. The Cam 17 platter in Pompeian-red ware is a type more commonly imported from Italy from c. AD 40. The finer vessels are fragmentary, and the Dressel 1 amphora is residual.

Key Pottery Group 9. Area E: Pit 8282, fill 8271, OA30, Group 298 (Figure 239)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
AITAL 1 27 27 <1% 0.13 2% Dressel 2-4
AMISC 1 18 18 <1%      
ASALA 4 616 154 8%     Beltrán 1
CGFCS 1 32 32 <1%     Flagon handle
CGMIC 4 14 4 <1%      
GROG 128 1914 15 24% 4.12 63% Platters EF6 Cam 21 Cam 31, jars Cam 218 Cam 221 Cam 232 Cam 264, strainer bowl M1
GROGC 77 5059 66> 63% 2.21 34% Bowl EF28, jars EF133 Cam 255 Cam 259 Cam 270
GROGRF 4 10 3 <1% 0.01 <1% Beaker Cam 116
NGWF 12 136 11 2%      
STOR 2 138 69 2%     Jar Cam 270
TR 2 34 17 <1% 0.06 1% Platter Cam 5
Total 236 7998     6.53

Although consisting mainly of grog-tempered ware (87% by weight), KPG9 is, again, characteristic of the mid- to later end of the date range. Several grog-tempered forms are types which continue into the mid-1st century AD. Salazon amphoras have made an appearance, although the range of imports is now less extensive. There is nothing obviously residual in the group.

Key Pottery Group 10. Area P: Pit 19104, fills 19105, 19107, 19109, 19110, 19111, OA3, Group 47 (Figure 240)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABSAN 1 12 12 1%     Dressel 2-4
BUF 3 30 10 1%      
CGFCS 5 87 17 4%      
GROG 79 1322 17 63% 0.61 68% Bowls EF46 Cam 211, jars EF135 Cam 218 Cam 249 Cam 256
GROGC 18 462 26 22% 0.03 3% Jar Cam 255
MICW 6 114 19 5% 0.09 10% Jar EF78
NGWFS 12 42 4 2%     Beaker EF195
TNM 1 8 8 <1% 0.02 2% Platter Cam 1
TR 1 6 6 <1% 0.15 17% Beaker Cam 112
Total 126 2083     0.90

Although relatively small, with a narrower range of fabrics, this assemblage is typical of Ceramic Phase 2, and is dated broadly to the first quarter of the 1st century AD. It is included here because the feature cuts ditch segment 19115 (Ceramic Phase 1). The assemblage, therefore, forms part of a ceramic sequence with the pottery from KPG1. The sand-tempered coarse ware (Figure 240, no. 8) is residual from this earlier feature.

Ceramic Phase 3: c. AD 20-55

Eleven groups, with a total weight of 60.3kg and 31.22 EVE, have been assigned to this ceramic phase, four of which (30.5kg, 14.76 EVE) are presented below. Phase 3 groups came mainly from the northern and southern zones. The phase is characterised by the gradual appearance of samian and various Roman fabrics, accompanied by a decline in other imported fine wares and amphoras. Two groups of pottery, Group 1 from ditch 1124 and Group 2 from slots 662/4, at Ivy Chimneys, Witham (Turner-Walker and Wallace 1999) provide broad comparisons. Independent dating evidence occurs in the form of a Langton Down brooch (SF7820) from pit 24181 and a coin of Cunobelin (SF2385) from pit 7167, both current in the first half of the 1st century AD. Additional dating is provided by the presence of Neronian samian in contexts assigned to the later part of the phase and the appearance of Verulamium region white ware and Central Gaulish glazed ware, both mid-1st century AD types. Grog-tempered pottery continues to form a major component, averaging 85%, of all assemblages assigned to this phase. Residuality is apparent in the presence of prehistoric pottery, handmade vessels and many of the Central Gaulish products.

Key Pottery Group 11. Area E: Pit 8026, fills 8003, 8014, 8018, OA31, Group 296 (Figure 241)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABAET 1 77 77 1%     Dressel 20
AITAL 1 11 11 <1%     Dressel 2-4
ASALA 3 482 161 5%      
CGFCS 1 4 4 <1%      
CGMIC 1 8 8 <1%      
ESH 6 224 37 2% 0.62 10% Jar Cam 254
GROG 160 5712 36 58% 3.31 55% Platters EF16 Cam 32, bowls EF72 Cam 241/6, cup Cam 212, jars EF146 EF179 Cam 218 Cam 219 Cam 249 Cam 221, beakers Cam 116 Cam 118 Cam 119
GROGC 48 2931 61 30% 1.41 24% Jars Cam 259 Cam 258 Cam 271 G5 G44
GROGRF 3 132 44 1% 0.12 2% Bowl
GRS 1 10 10 <1% 0.10 2% Beaker H7
NGWF 2 76 38 1% 0.08 1% Beaker Cam 113
SGSW 2 2 1 <1%      
TN 5 180 36 2% 0.15 3% Platter Cam 2
TR 4 24 6 <1% 0.18 3% Beaker Cam 112
Total 238 9873     5.97

Pottery Group 11 is typical of the earliest part of Ceramic Phase 3, c. AD 20-40, and is dominated by grog-tempered pottery in a wide variety of forms. Imports are present in the form of Gallo-Belgic and Central Gaulish ware, although the latter is represented by small body sherds only. Salazon amphora sherds are present and Dressel 20 amphora has made an appearance, whereas Dressel 2-4 wine amphora is reduced to a single small body sherd. There are a few sherds in Roman fabrics, including a tiny sherd of south Gaulish samian, and Romanised forms in grog-tempered wares are increasingly evident. The average sherd weight is generally high, although the finer fabrics tend to be fragmentary.

Key Pottery Group 12. Area D: Pit 9230, fill 9231, OA29, Group 288 (Figure 242)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
AITAL 2 36 18 1%     Dressel 2-4
BSW 4 48 12 1%      
GROG 265 3306 12 66% 1.73 73% Bowls Cam 212 Cam 244, jars Cam 219 Cam 229 Cam 249, beaker Cam 115, lid K6
GROGC 44 1396 32 28%      
GROGRS 8 142 18 2% 0.29 12% Bowl EF73, flagon
GRS 1 6 6 <1%      
RED 1 4 4 <1%      
TR 25 76 3 2% 0.35 15% Beaker Cam 112
PREHIST 3 18 6 <1%      
Total 353 5032     2.37    

Grog-tempered pottery in a variety of forms again predominates at 96% by weight. The group is dated broadly to c. AD 20-45. There is a narrow range of fabrics present; unusually, terra rubra is the sole imported fine ware type represented. Amphoras are again scarce and Roman fabrics are still very much in the minority. The pottery is fragmentary, but there is very little that is obviously residual, apart from three small flint-tempered body sherds of prehistoric date.

Key Pottery Group 13. Area G: Pit 7167, fills 7168, 7178, 7179, OA36, Group 313 (Figure 243)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABAET 4 108 27 1%     Dressel 20
AITAL 1 199 199 2%     Dressel 1
AMISC 1 22 22 <1%      
ASALA 1 66 66 1%      
BSW 33 280 8 3%      
BUF 8 16 2 <1%      
CGFCS 15 116 8 1%      
GROG 427 5850 14 59% 2.78 82% Platters Cam 21 Cam 33, bowls EF38 EF42 Cam 210 Cam 252, jars EF129 Cam 218 Cam 229 Cam 254 Cam 259 G17 G19, beakers EF188 H7
GROGC 39 1780 46 18% 0.12 4% Jar Cam 270
GRS 28 160 6 2%      
MICW 10 160 16 2% 0.12 4% Jar
NGWF 4 24 6 <1%      
NGWFS 2 6 3 <1%      
STOR 26 1095 42 11%      
TN 1 22 22 <1%      
TNM 5 34 7 <1% 0.06 2% Platter EF20, beaker EF192
TR 8 48 6 <1% 0.30 9% Platter Cam 8, cup Cam 56, beaker Cam 112
Total 613 9986     3.38    

Dressel 20 and salazon amphoras are present in this assemblage, but the sherd of Dressel 1 wine amphora is residual. Although grog-tempered pottery continues to predominate, Roman fabrics are beginning to form an increased part of the assemblage. Imports are again represented mainly by body sherds, while the forms present in Gallo-Belgic ware are the latest types to be produced. The newly identified forms in micaceous terra nigra may reflect the continuance of this industry beyond the first two decades AD. The group is dated to c. AD 25-45. Besides the Dressel 1, only the handmade jars are likely to be residual.

Key Pottery Group 14. Area N: Pit 11723, fill 11720, OA26, Group 142 (Figure 244)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABAET 6 229 38 4%     Dressel 20
AWINE 1 30 30 1%     Dressel 2-4
BSW 24 228 10 4%      
CGFCS 1 4 4 <1%      
CGGLZ 1 2 2 <1%      
CGMIC 3 16 5 <1% 0.11 4% Jar Cam 262
COLB 1 4 4 <1%      
GROG 171 2361 14 42% 2.06 68% Platters Cam 21 A2, bowl Cam 212, jars Cam 249 Cam 258 G19
GROGC 82 2158 26 38% 0.40 13% Jar Cam 260
GROGRF 3 30 10 1% 0.10 3% Beaker Cam 116
GROGRS 3 76 25 1% 0.08 3% Bowl EF71
GRS 18 246 14 4% 0.23 8% Jar G23, lid
NGWF 3 10 3 <1%      
NGWFS 2 14 7 <1%     Beaker Cam 113
RED 2 1 1 <1%      
SGSW 4 36 9 1% 0.06 2% Platter f18R, bowl f29
STOR 6 168 28 3%      
TNM 2 30 15 1%      
Total 333 5643     3.04

This group is characteristic of pottery assemblages at the later end of Ceramic Phase 3. Most of the identifiable forms are mid-1st century types and the assemblage is dated to c. AD 45-55. Much of the imported fine ware is likely to be residual, except for the North Gaulish beakers and the samian. Roman fabrics are more in evidence, forming more than 12% of the assemblage by weight. Samian and Central Gaulish glazed wares are both present, but in general, the quantity of imports is in decline. There is Dressel 20 amphora, but wine amphoras are again represented by a single small sherd, and there is a smaller range of forms than previously in grog-tempered wares. The finer wares are again fragmentary. It should perhaps be noted that the small sherds of samian have been dated to AD 60-80.

Ceramic Phase 4: c. AD 55-80

Fifteen groups, mainly from the central and southern settlement zones, with a total weight of 118kg and vessel rim equivalence (EVE) of 86.5, have been assigned to Ceramic Phase 4. Four of these groups, totalling 40.5kg and 32 EVE, are presented below. Ceramic Phase 4 is a period of continuity and change: sand replaces grog as the predominant tempering agent by the end of the phase, but many of the forms are little altered from Phase 3. Chelmsford provides a number of comparable assemblages; namely groups 1-3 (Going 1987, table 3) and the pottery from ditch K205 (Going 1992a, 96).

Key Pottery Group 15. Area D: Ditch 25018, segment 9213, fill 9214, OA29, Group 764 (Figure 245)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ASALA 1 159 159 2%     Salazon
BSWM 29 536 18 8% 0.53 21% Jar G20
BUF 1 2 2 <1%      
COLB 9 44 5 1%     Flagon
ESH 1 18 18 <1%      
GRF 5 78 16 1% 0.19 7% Bowl EF57, jar G19
GROG 126 1680 13 26% 0.92 36% Platter A2, bowl C19, jars G19 G20 Cam 232
GROGC 30 1165 39 18% 0.48 18% Jars G45 Cam 259
GRS 10 182 18 3%      
RED 3 16 5 <1%      
STOR 70 2650 38 41% 0.46 18% Jars EF126 EF178 G3
Total 285 6530     2.58

Pottery Group 15 is typical of the early part of Ceramic Phase 4. Measured by EVE, grog-tempered pottery predominates, though less so than in Phase 3; conversely, black-surfaced ware and storage jar fabrics form a larger proportion. A greater range of wheel-thrown sand-tempered pottery is also present in sandy and fine grey wares and buff ware, though still in relatively small proportions, while imports typical of the Late Iron Age are noticeably absent. A date of c. AD 50/55-60 is supported by the presence of the salazon amphora, as well as the remaining forms. The condition of the pottery is generally good, and the group is coherent with no obvious residual element.

Key Pottery Group 16. Area D: Pit 9218, fills 9217, 9370, OA31, Group 768 (Figure 246)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
BSW 73 858 12 5% 0.90 7% Jar G3, beaker H1
BUF 12 39 3 <1%      
COLB 96 297 3 1% 0.43 3% Bowl, flagon
GRF 72 274 4 1% 1.40 10% Bowl EF64, beaker H1
GROG 580 5430 9 26% 6.15 44% Platters Cam 21 A2, jars EF131 Cam 249 Cam 258 G3 G19 G20, beakers H1 H7, funnel N2
GROGC 240 5887 25 28% 3.19 23% Bowl C33, jars Cam 249 Cam 259 G45, lid K3
GROGRS 4 28 7 <1% 0.11 1% Jar Cam 249
GRS 26 418 16 2% 0.57 4% Jar G20
SGSW 31 324 10 2% 0.24 2% Platter f15/17, bowls f29 f30, cups f27 Ritt.8
STOR 210 7470 36 35% 0.73 5% Jars Cam 270B Cam 271 G44
TN 2 15 8 <1% 0.06 <1% Platter Cam 13
TR 12 48 4 <1% 0.11 1% Beaker Cam 112
TRCG 1 4 4 <1%      
Total 1359 21092     13.89

Grog-tempered pottery continues to dominate, but wheel-thrown, sand-tempered forms and fabrics are now common. Gallo-Belgic imports are barely represented and the forms present are among the latest to be produced. South Gaulish samian ware is present, with the range of forms falling within a date band of AD 45-80. The assemblage as a whole can probably be dated more closely to AD 55-65 because of the presence of two Colchester B brooches (SF1553, SF3281). Jars, beakers and platters are still the most common vessel classes, but bowls, cups and flagons are also much in evidence. There is also a grog-tempered ware funnel. The pottery is fragmentary and the small Central Gaulish sherd is residual.

Key Pottery Group 17. Area L: Pit 20008, fill 20009, OA27, Group 708 (Figure 247)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
BSW 202 5200 26 49% 6.66 53% Platter A2, bowls C27 Cam 214, jars G17 G19 G20, beaker H1, lid K6
COLB 24 346 14 3% 1.00 8% Flagon J3
GRF 19 1590 84 15% 3.11 25% Bowl C12, jars G20 G40
GROG 36 432 12 4% 0.33 3% Platter A2, jar G, lid K3
GROGC 15 620 41 6% 0.03 <1% Jar
GRS 50 1925 39 18% 1.15 9% Jars G17 G23
HGG 1 8 8 <1%      
IMIC 3 8 3 <1%     Beaker H1
NGWF 1 20 20 <1%      
SGSW 2 12 6 <1% 0.14 1% Platter f18
STOR 6 525 88 5% 0.06 <1%  
Total 359 10686     12.48    

This assemblage, dating to c. AD 70, demonstrates both typical and atypical aspects of the later part of this ceramic phase. Typically, grog-tempered pottery now forms a minor component, while wheel-thrown, sand-tempered pottery predominates. The assemblage is atypical in that it comprises eleven complete or near-complete vessels, with some of the remaining pottery being sufficiently preserved to gain complete profiles of another three vessels. The very low quantity of sandy grey ware is also unusual, and perhaps reinforces the view that most or all of the assemblage was deposited under special conditions, and cannot be regarded as a mundane rubbish deposit. There were a number of sherds in this feature that joined sherds in contemporaneous pit 20010; the mica-dusted beaker sherd from this feature is included for illustration here (Figure 247, no. 15).

Key Pottery Group 18. Area M: Pit 24013, fill 24014, OA26, Group 692 (Figure 248)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABAET 1 12 12 1%     Dressel 20
BSW 55 556 10 25% 0.34 11% Platter A4, dish B8, jar, lid K3
BUF 6 12 2 1% 0.11 4%  
BUFM 2 60 30 3% 0.03 1% Mortarium D1
COLB 2 20 10 1%      
GRF 11 174 16 8% 0.57 19% Cup, jar G20
GROG 23 102 4 5% 0.28 9% Jar
GROGC 5 148 30 7% 0.03 1%  
GRS 33 526 16 23% 1.18 39% Platter A2, jar G3, beaker H1, lid K6
SGSW 2 14 7 1% 0.39 13% Bowl f35
STOR 4 156 39 7%      
VRWM 2 394 197 18% 0.08 3% Mortarium D1
Total 146 2174     3.01

A large quantity of pottery was collected from pit 24013, and can be placed at the later end of Ceramic Phase 4. Grog-tempered pottery is present, but only in small quantities, while wheel-thrown, sand-tempered fabrics are predominant. The range of forms is wide. The Verulamium region white ware mortarium, with its deep and hooked flange, is one of the earliest products of that industry, dating to the pre-Flavian and early Flavian periods (Davies et al. 1994, 47). This, plus the samian bowl, which is dated no earlier than c. AD 70, provides an AD 70-80 date range for the deposition of the whole assemblage. As with KPG16, the condition of the group is poor, though fairly consistent. There is little that can easily be dismissed as residual, though the grog-tempered pottery could well be.

Ceramic Phase 5: c. AD 80-125

Seventeen groups from the central and southern zones, with a total weight of 864kg and vessel rim equivalence (EVE) of 55.19 have been assigned to Ceramic Phase 5. Three of these groups, totalling 200kg and 21 EVE, are presented below. Phase 5 covers a period of c. 45 years, but, despite this, there are few distinctions between individual pottery assemblages. Locally produced sand-tempered pottery predominates in all groups, and is accompanied by much smaller proportions of regional and imported wares. Groups 4 and 5 from Chelmsford provide comparable assemblages (Going 1987, table 3).

Key Pottery Group 19. Area M: Pit 15773, fill 24258, OA26, Group 688 (Figure 249)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
BSW 29 615 21 30% 1.18 55% Jars G16 G17 G19 G20 G23
COLB 5 98 20 5%      
GRF 6 72 12 3%     Beaker
GROG 3 36 12 2%      
GROGC 5 146 29 7%     Jar
GRS 18 515 29 25% 0.63 30% Platter A2, jars G17 G20
RED 3 88 29 4% 0.27 13% Flagon J3
SGSW 1 10 10 <1%     Platter f15/17 or f18
STOR 5 240 48 12% 0.05 2%  
VRW 6 254 42 12%     Flagon
TotaL 81 2074     2.13

This assemblage dates to the early part of this ceramic phase, probably no later than c. AD 100. As expected for a group of this date, black-surfaced ware predominates with sandy grey ware forming a significant contribution, while the quantities of grog-tempered wares are small. The range of forms, mainly comprising high-shouldered jars, is typical. There is seemingly little or no obvious difference in the coarse wares between this group and those assigned to the end of Phase 4. The samian platter, however, dating to the late 1st century AD, provides the strongest evidence for pushing the date of deposition beyond AD 80. A coin of Domitian (SF7934) from an overlying fill does not conflict with this date. The condition of the assemblage is generally good, with nothing apart from the grog-tempered pottery that is obviously residual.

Key Pottery Group 20. Area H: Pit 6201, fill 6203, OA23, Group 530 (Figure 250)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ASALA 2 150 75 3%     Salazon
BSW 130 1230 9 24% 2.01 46% Platter A2, jars G17 G20 G23, beaker H1
BUF 23 76 3 1%    
CGSW 1 58 58 1%     Platter f18/31
GRF 1 2 2 <1%      
GROGC 1 22 22 <1%      
GRS 90 2205 25 42% 2.06 48% Jars G8 G17 G23, lid K
NKG 1 4 4 <1%      
RED 3 16 5 <1% 0.15 3% Bowl C
SGSW 3 28 9 <1%     Platter f15/17
STOR 28 1540 55 29% 0.11 3% Jar G44
Total 283 5331     4.33    

This group is dated to the early 2nd century, probably no later than c. AD 120. Composition is similar to that from earlier dated Roman period assemblages. Once again the samian evidence is fundamental, with the Central Gaulish platter providing an AD 100-120 date. The condition of the group is somewhat variable. Fabrics such as buff ware, North Kent grey ware and black-surfaced ware comprise small sherds, while sandy grey ware has a noticeably higher average sherd weight. Obviously residual pottery includes the grog-tempered ware, the amphora and South Gaulish samian platter, all of which are dated no later than AD 70.

Key Pottery Group 21. Area J: Pit 5147, fill 5146, OA23, Group 409 (Figures 251, 252)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
BB2 1 12 12 <1% 0.05 <1% Dish B2
BSW 537 6825 13 54% 8.77 61% Platter A2, dishes B1 B7, bowls C1 C6 C12, jars G16 G19 G20 G23, beaker H1, lid K3
BUF 8 54 7 <1%      
BUFM 5 580 116 5% 0.09 1% Mortarium D1
CGSW 7 84 12 1% 0.29 2% Dishes f18/31R f18/31 or f31, bowl f37, cup
COLB 46 615 13 5%     Flagon handle
GRF 19 256 13 2% 0.72 5% Platter A2, jar, beaker H1
GROG 4 40 10   0.08 1% Jar
GRS 125 2780 22 22% 3.54 25% Dish B7, jars G20 G23 G24
LOND 3 16 5        
NKG 5 136 27 1% 0.34 2% Platter A4
NKO 1 20 20 <1% 0.11 1% Bowl
SGSW 13 220 17 2% 0.37 3% Platters f15/17 f18, dish f36, bowl f30, cup f27g
STOR 8 860 108 7%      
VRW 1 70 70 1%     Flagon handle
Total 783 12568     14.36    

Pottery Group 21, dating to c. AD 120/25, is the latest within this ceramic phase. It retains elements that date to the early 2nd century or earlier - platters, high-shouldered jars and North Kent wares - and yet includes pottery types, specifically BB2 and Central Gaulish samian, that take the date to the beginning of Phase 6. The BB2 dish has lattice decoration and a rim that is triangular in profile. Dishes of this description (Cam 37A) are dated to the Trajanic/Hadrianic period (Bidwell and Croom 1999, 469). None of the Central Gaulish samian pre-dates AD 120. The condition of the assemblage is fairly uniform. As for KPG20, there is less sandy grey ware than black-surfaced ware but, considering its higher average sherd weight, the sandy grey ware seems to be better preserved. The grog-tempered pottery is residual, as is some or all of the south Gaulish samian. The context also included 4g of intrusive late Roman Oxford ware derived from overlying features. This has been excluded from the tables.

Ceramic Phase 6: c. AD 125-170

Twelve groups, mainly from the northern settlement zone and with a total weight of 65.8kg and vessel rim equivalence (EVE) of 50.4, were assigned to Ceramic Phase 6. Four of these, totalling 52kg and 39 EVE, are presented below. As in Phase 5, assemblages comprise locally produced pottery, with varying, but always much smaller, proportions of regional and imported wares. This remains very much the pattern until the end of the Roman period. The groups below are comparable to a number of assemblages from Chelmsford, namely groups 6-8 (Going 1987, table 3) and the pottery from pit K90.2 (Going 1992a, 99-104).

Key Pottery Group 22. Area F: Ditch 25245, segment 10159, fill 10182, OA31, Group 361 (Figures 253, 254)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
BB2 30 509 17 2% 1.16 6% Dishes B2 B3
BSW 363 4267 12 20% 7.57 38% Platter A4, dishes B4 B7 B9, jars G16 G18 G23 G40, beaker H7, lid K3, miniature
BUF 10 42 4 <1%      
CGSW 16 148 9 1% 0.60 3% Dishes f31 f42, bowls f37 f78, cup f33
COLC 26 68 3 <1% 0.74 4% Beaker H20
GRF 36 276 8 1% 0.53 3% Dish B10, jar G29, beakers H6 H34/H35
GROG 21 260 12 1%     Jar
GRS 403 6185 15 30% 7.52 37% Dish B4, bowl C16, jars G20 G22 G23 G24 G29 G42, lid K3
LESTA 8 256 32 1% 0.41 2% Bowl C23
NKG 36 384 11 2% 0.55 3% Platter A4, beaker H6
RED 2 4 2        
SGSW 16 148 9 1% 0.22 1% Platter f15/17 or f18, dish f42, bowl f37
STOR 206 8530 41 40% 0.53 3% Jars G42 G44 G45
VRW 1 8 8 <1%      
VRWM 1 212 212 1%      
Total 1175 21297     19.83    

Ditch segment 10159 contained a large amount of pottery dating to c. AD 125-150. The assemblage includes a range of pottery typical of an early to mid-2nd century context, comprising bead-rimmed dishes, poppy-headed beakers, and flasks or narrow-necked jars. The presence of Colchester colour-coated ware, however, which was not widely distributed before AD 150/60, pushes the date of deposition towards the end of the offered date range. The folded H34/H35 beaker is unusual in this deposit; the type first appears in Chelmsford c. AD180 (Going 1987, 31). The condition of the assemblage is mixed, with a variable average sherd weight. The grog-tempered pottery and some of the South Gaulish samian, such as the f15/17 platter, are residual. The assemblage has been interpreted as a possible ritual deposit.

Key Pottery Group 23. Area G: Pit 7118, fills 7119, 7166, OA34, Group 852 (Figure 255)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABAET 5 644 129 10%     Dressel 20
BB2 10 390 39 6% 0.85 14% Dishes B1B2 B4
BSW 68 1244 18 19% 2.09 35% Dish B4, jars G5 G9 G20 G29 G40
CGSW 18 184 10 3% 0.66 11% Dishes f18/31 f18/31R or 31R f31, bowl f37, cup f33
COLB 1 62 62 1%     Flagon neck
COLBM 5 228 46 4% 0.29 5% Mortaria D1 D13
COLC 4 14 4 <1% 0.28 5% Beaker H20
COLSW 6 44 7 1% 0.18 3% Dish f18/31 or f31
GRF 3 48 16 1%      
GROGC 1 54 54 1% 0.06 1% Jar
GRS 34 706 21 11% 0.73 12% Dish B2/B4, jar G23, lid
HAWO 2 34 17 1%      
MWSRS 1 6 6 <1%      
NKG 9 56 6 1% 0.19 3% Beaker H6
RED 9 30 3 <1% 0.16 3% Beaker ?H20
STOR 41 2720 66 41% 0.46 8% Jar G42 G45
Total 217 6464     5.95    

Pit 7119 contained three fills, the pottery from two of which was quantified by EVE and has been amalgamated here. The assemblage provides a mid-2nd century date for filling, with the Central Gaulish samian recovered from the top fill, dating the final episode of deposition to around AD 155/60 or a little later. Mid-2nd century pottery groups at Elms Farm generally display a greater range of fabrics, particularly Colchester wares, than earlier dated contexts and KPG23 is no exception. While black-surfaced ware predominates, sandy grey ware forms a lower proportion measured by EVE than is usual. Conversely, BB2 forms a higher proportion. Some jar types, such as the high-shouldered G20, were residual at Chelmsford by the mid-2nd century, and are likely to be here also, as is the grog-tempered pottery.

Key Pottery Group 24. Area D: Pit 9029, fills 9028, 9064, OA31, Group 783 (Figure 256)
FabricsSherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABAET 2 376 188 2% Dressel 20
BB1 1 6 6 <1%
BB2 9 300 33 2% 0.22 3% Dish B4
BSW 145 1296 9 7% 2.43 31% Dish B4, jars G19 G23 G29
BUF 5 6 1 <1%
CGSW 13 112 9 1% 0.50 6% Cup f33, dishes f19/31R or f31R f31, bowl f37
COLB 4 76 19 <1%
COLBM 4 128 32 1%
COLC 6 80 13 <1% 0.09 1% Beaker H20
COLSW 5 36 7 <1% 0.22 3% Cup f27
EGSW 1 64 64 <1% Dish f32
GRF 7 60 9 <1% Dish B2/B4
GROG 13 274 21 1% 0.07 1% Bowl
GRS 190 2754 19 14% 2.15 28% Dishes B3 B7, bowl C1, jars G5 G19 G22 G23, beaker
GRSWSM 1 132 132 1% 0.08 1% Mortarium D1
MIC 1 6 6 <1%
NKG 12 158 13 1% 0.71 9% Jar G40, beaker H5
STOR 151 13510 89 70% 1.28 17% Jars G36 G44
Total 570 193747.75

This assemblage, a composite of two fills from pit 9029, is dated c. AD 140 to c. AD 160. The date of deposition is likely to be at the later end of this range, or slightly beyond. The East Gaulish samian dish from the bottom fill is one of the latest pieces and dates from AD 160 onwards. The range of pottery present, such as bead-rimmed dishes and the roughcast colour-coated beaker, is very typical of this period. The large quantity of storage jar fabric is, however, unexpected. Assemblage condition is mixed. Fabrics such as black-surfaced ware and fine grey ware are represented by small sherds, and there is a quantity of clearly residual pottery (10% by EVE), notably the grog-tempered pottery and the high-shouldered G19 coarse ware jars.

Ceramic Phase 7: c. AD 170-210

Six groups recovered from across the settlement, with a total weight of 46kg and rim equivalence (EVE) of 31.32, were assigned to Ceramic Phase 7. Three of these, totalling 35kg and 22.5 EVE, are presented below. These groups are comparable to Great Dunmow Group 464/820 (Going and Ford 1988, 61-6) and groups 9-16 from Chelmsford (Going 1987, table 3).

Key Pottery Group 25. Area G: Pit 7122, fill 7123, OA34, Group 868 (Figure 257)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
BSW 105 2390 23 16% 2.00 23% Dish B1, jars G19 G23
BUF 2 60 30 <1% 0.09 1% Bowl Cam 198
CGSW 2 20 10 <1% 0.06 1% Dish f18/31 or 31, cup f33
COLB 2 6 3 <1%
COLBM 21 1845 88 12% 1.11 13% Mortaria D1 D2 D13
COLC 16 224 14 2% 0.45 5% Beaker H20
COLSW 1 4 4<1% Bowl f37
GROG 4 56 14 <1% 0.13 2% Jar G
GROGC 1 52 52 <1%
GRS 286 8815 31 58% 3.62 42% Jars G5 G17 G22 G23 G24 G25, lid K6
LOND 1 10 10 <1% Bowl C
MIC 1 20 20 <1% 0.12 2% Bowl Cam 41
MWSRS 2 24 12 <1% 1.00 12% Flagon J4
STOR 20 1775 89 12% 0.07 1% Jar G44
Total 464 15301 8.65

This group is dated to the late 2nd century AD. The offered date is supported by the presence of types produced at Heybridge during the late 2nd or early 3rd century, namely the D13 wall-sided mortarium, the G5 ledge-rimmed jar, and G25 jar (see Pottery production). Black-surfaced ware, always so dominant in previous groups, is relegated to second place after sandy grey ware, while, as expected, Colchester mortaria are present in greater quantity than in Ceramic Phase 6. The condition of the assemblage is generally good. Sherds are large, although there are perhaps fewer than expected rim sherds for a group of this size. The grog-tempered wares, London-type ware, and the mica-dusted bowl are all residual.

Key Pottery Group 26. Area H: Well 6280, fill 16083, OA19, Group 531 (Figure 258)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
BB2 9 412 46 8% 0.83 17% Dish B2
BSW 93 1552 17 32% 2.21 45% Dishes B2 B3, jars G9 G23
CGSW 13 745 57 15% 1.04 22% Dishes f31 f31R, bowl f37, cup f33
COLC 12 149 12 3% 0.48 10% Beakers H24 H35
EGSW 1 22 22 <1% Bowl f37
GRF 10 46 5 1% 0.03 1% Flagon J
GRS 13 154 12 3% 0.14 3% Jar G
MIC 2 5 3 <1% 0.05 1% Beaker H
MWSRF 3 42 14 1% Flagon
STOR 55 1801 33 37% 0.03 1% Jar G44
VRW 1 6 6 <1%
Total 212 4934 4.81

Pottery Group 26 is likely to have been deposited at the end of the 2nd century AD, possibly during the final decade. The bag-shaped H24 beaker, with barbotine decoration in the form of tendrils, is dated to after AD 190 at Chelmsford (Going 1987, 30); similar products from Colchester were first produced a little earlier (Bidwell and Croom 1999, 486). The assemblage includes six complete or near-complete vessels and so differs from similarly dated groups in this respect. This has resulted in several atypical elements immediately noticeable here. The proportion of sandy grey ware is far lower than usual, and there is, conversely, a higher proportion of fine wares. Composition, too, is unusual, with dishes and beakers, rather than jars, predominating. The condition of the pottery, apart from the complete vessels, is poor, with fabrics such as fine grey ware, mica-dusted ware and Verulamium region white ware represented by small sherds. These may have been deposited in a separate episode to the complete or near-complete vessels: a samian dish was one such vessel, which is not the sort of vessel typically used to lift water from a well. It is possible that this vessel, and others like it, formed part of a structured deposit.

Key Pottery Group 27. Area W: Stoke-hole 1589, all deposits, OA38, Group 910 (Figure 259)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
BSW 40 343 9 2% 1.16 13% Jars G3 G5, flagon
BUF 55 259 5 2% 0.08 1% Jar G26
BUFM 178 5579 31 37% 2.59 29% Mortaria D3 D11
COLC 2 10 5 <1%
GRF 142 737 5 5% 0.35 4% Dish B1, jar
GRM 5 146 29 1% 0.03 <1% Mortarium D11
GROG 13 130 10 1%
GROGC 7 124 18 1%
GRS 610 3710 6 24% 3.51 39% Dish B2/B4, jars G5 G25
HAB 6 48 8 <1% 0.09 1% Dish B2/B4
HAX 1 28 28 <1%
MWSRS 5 20 4 <1%
NVC 1 6 6 <1%
RED 1 4 4 <1%
STOR 3 48 16 <1%
UCC 50 245 5 2% 0.11 1% Beaker
UPOT 667 3725 6 25% 1.13 13% Dish B2/B4, beaker, flagon
Total 1786 15162 9.05

Stoke-hole 1589 was serving kilns 1223 and 1618 during the late 2nd and early 3rd centuries AD. The group is an amalgam of the twenty contexts that comprised deposits within the stoke-hole. The pottery recovered was very likely deposited during the early 3rd century, and is therefore placed at the very end of Ceramic Phase 7. Buff ware mortaria and grey wares, being the principal fabrics of products fired in these kilns, predictably form the largest proportions within the assemblage. Black-surfaced ware also makes a significant contribution. Fine wares are present in small quantities. The condition of the group is generally poor. Much of the assemblage is burnt or very abraded and individual fabrics sometimes cannot be identified. The group is fragmented overall, although there is little besides the black-surfaced ware G3 jar and the grog-tempered pottery that must be residual.

Ceramic Phase 8: c. AD 210-260

Recovered from the northern, southern and central zones of the settlement, five groups with a total weight of 28kg and rim equivalence (EVE) of 20 were assigned to Ceramic Phase 8. Three of these, totalling 13kg and 11 EVE, are presented below.

Key Pottery Group 28. Area H: Pit 6182, fill 6178 OA19, Group 553 (Figure 260)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
BSW 32 420 13 18% 0.60 21% Dish B4, jar G5, beaker H34
BUF 1 24 24 1% 0.08 3%
CGRHN 7 30 4 1% 0.14 5% Beaker
CGSW 4 102 26 4% 0.14 5% Platter f79 or Ludowici Tg, dish f18/31 or f31, bowl f37
COLC 1 2 2 <1%
EGRHN 1 1 1 <1% Beaker
GRF 16 184 12 8% 0.06 2% Dish B4, beaker
GRS 146 1303 9 55% 1.66 58% Dish B3, jars G5 Cam 218, beaker
MWSRS 1 4 4 <1%
NVC 1 2 2 <1%
SGSW 1 2 2 <1% 0.01 <1% Platter f15/17
STOR 6 310 52 13% 0.19 7% Jar G44
Total 2172384 2.88

This assemblage, dated to the early 3rd century AD is, in terms of the range of forms and general proportions of fabrics, similar in character to Ceramic Phase 7. The principal difference between them lies in the fine wares. There is a greater range here with the presence, albeit in small quantities, of Rhenish wares and Nene Valley and Colchester colour-coated wares. The condition of the assemblage is mixed, though generally poor. There is little that need be residual, except the 1st-early 2nd century Cam 218 jar and South Gaulish samian.

Key Pottery Group 29. Area H: Pit 16088, fill 16073, OA19, Group 559 (not illustrated)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABAET 2 95 48 1% Dressel 20
AGAUL 1 80 80 1% Gauloise 4
BSW 86 1770 21 22% 1.12 16% Dish B2/B4, jars G5 G24
BUF 5 84 17 1% Face-pot
BUFM 2 286 143 3% 0.23 3% Mortarium D11
CGSW 20 206 10 2% 0.73 11% Platter f18/31, dishes f31 f31R, bowls f36 f37, cups f33 f46
EGSW 1 14 14 <1%
GRF 30 540 18 6% 1.06 15% Dishes B2 B3 B4, jar G9
GROG 12 228 19 3% 0.24 3% Lid
GROGC 4 114 29 1%
GRS 101 1681 17 21% 2.38 35% Dish B2/B4, bowl-jar E5, jars G5 G24 G25, beakers H24 H34
HAWO 1 32 32 <1% 1.00 15% Flagon
HAXM 3 30 10 <1%
LESTA 1 10 10 <1% 0.11 2% Bowl
NVC 3 16 5 <1%
RED 17 200 12 2%
STOR 46 2960 64 35% Jar
VRGR 9 146 16 2%
Total 344 8492 6.87

Pottery Group 29 is dated to the first half of the 3rd century AD. The sandy grey ware is particularly typical of the period - the vessels present in this fabric are types produced into the 3rd century at Heybridge. Residual grog-tempered pottery is present, as ever. The stamped London-Essex bowl and Verulamium grey ware vessel are also residual. Assemblage condition is generally good; sherd size is reasonably uniform. Three sherds of Saxon pottery are intrusive and have been discounted.

Key Pottery Group 30. Area E: Pit 10062, fill 10061, OA31, Group 811 (Figure 261)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
AITAL 2 824 412 32% Dressel 1
AMISC 1 33 33 1%
BSW 21 662 32 26% 0.87 76% Dishes B1 B3 B4 B5, beaker H34, lid K2
EAM 1 68 68 3%
EGSW 1 18 18 1% Dish or bowl
GROG 1 2 2 <1%
GRS 20 442 22 17% 0.20 18% Jars G22 G25 G45
HAX 2 18 9 1%
NVC 2 18 9 1% Beaker H32
NVM 1 12 12 <1% 0.07 6% Mortarium
RED 1 38 38 1% Bowl
STOR 3 442 147 17%
Total 56 2577 1.14

This small group is dated to the mid-3rd century AD. The presence of bead-rimmed dishes (B4) in association with incipient bead-and-flanged dishes (B5) and Nene Valley mortaria is typical of contexts dated to c. AD 260. Black-surfaced and sandy grey wares remain predominant. Fine wares are also present, but typically there is less diversity of fabric here compared to an early 3rd century group. For example, there are no imported fine wares, other than samian, or Colchester products. Despite the low number of rims, the condition of the pottery in this context is good. Sherd size is generally large, while some of the dish profiles can be reconstructed. Residual pottery is restricted to a single grog-tempered ware sherd.

Ceramic Phase 9: c. AD 260-310

Recovered from across the whole of the settlement area, ten groups with a total weight of 49kg and rim equivalence (EVE) of 38 were assigned to Ceramic Phase 9. Four of these, totalling 27kg and 23 EVE and characteristic of the phase, are presented below.

Key Pottery Group 31. Area N: Pit 11303, fill 11302, OA50, Group 671 (Figure 262)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABAET 2 242 121 2% Dressel 20
AWINC 2 42 21 <1% Dressel 2-4
BB1 4 80 20 1% 0.07 1% Dish B6
BB2 7 186 27 1% 0.50 5% Dish B3
BSW 383 2568 7 20% 2.25 24% Dishes B3 B4, bowl, bowl-jars E2 E5, jars G5 G24, beaker H34
BUFM 10 434 43 3% 0.11 1% Mortaria D3 D11
CGFCS 1 6 6 <1%
CGSW 13 112 9 1% 0.13 2% Platter f79R, dish f18/31 or 31, mortarium f45, cup f33, beaker f72
EGSW 13 260 20 2% 0.11 1% Dish f31R, bowls f30 f37
GRF 401 3197 8 25% 1.54 16% Dishes B2/B4 B3, jars G5 G24, beaker
GROGC 16 390 24 3% Jar
GRS 185 2660 14 21% 3.05 32% Bowl-jar E2, jars G5 G24 G25 G42, beaker
HAB 4 98 25 1% 0.33 3% Dish B3
HAR 27 435 16 3% 0.88 9% Dishes B2 B3, jar, beaker
HAWO 3 42 14 <1% Jar
MWSRS 4 26 7 <1% 0.31 3% Flagon J3
NKG 1 6 6 <1% 0.06 1% Jar
NVC 3 18 6 <1% Lid K7
NVM 2 122 61 1% 0.12 2% Mortarium D14
RED 13 44 3 <1%
RET 2 10 5 <1%
STOR 51 1793 35 14% 0.03 <1% Jar G44
Total 1147 12771 9.49

This large group is dated to the early part of Ceramic Phase 9, probably extending only a little beyond AD 260. The assemblage retains characteristics associated with early to mid-3rd century contexts: buff ware mortaria, bead-rimmed dishes, and ledge-rimmed jars. Other elements, such as the Nene Valley mortarium, Rettendon ware and BB1, undoubtedly push the date of deposition into the later 3rd century. As ever, black-surfaced and sandy grey wares dominate the assemblage. Of the remaining fabrics, only fine grey ware makes a significant contribution. The condition of the group is generally good, but fragmentary, and there is a relatively large proportion of residual pottery (around 8% by weight), including the grog-tempered ware, Central Gaulish samian, North Kent grey ware, and the white-slipped ring-necked flagon (J3). The group also contains an abraded sherd of later 4th century Alice Holt grey ware (excluded from the table), which must be intrusive.

Key Pottery Group 32. Area H: Pit 6267, fill 6268, OA19, Group 561 (Figure 263)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABAET 1 78 78 1% Dressel 20
BSW 32 855 27 11% 1.15 26% Dishes B1 B2/B4 B6, jar G5, beaker
BUF 4 66 17 1%
BUFM 7 553 79 7% 0.42 10% Mortaria D3 D11
CGSW 6 78 13 1% 0.24 5% Dishes f18/31R or f31R f31, cup f33
COLBM 3 480 160 2% 0.24 5% Mortarium D11
COLC 326 9 <1% Beaker
GRF 21 378 18 5% 0.49 11% Dish B2/B4, beaker H33
GRS 69 1245 18 16% 1.48 34% Dish B3, bowl-jar E2, jars G5 G9 G24 G42, beaker H34
NVC 3 36 12 <1% Beaker
NVM 1 48 48 1%
OXWM 1 124 124 2% 0.09 2% Mortarium D5
RED 6 114 19 1% 0.08 2% Jar G26
RET 2 32 16 <1% 0.11 3% Jar G24
STOR 40 3490 87 46% 0.09 2% Jar G44
Total 199 7603 4.39

The pottery from fill of pit 6267 is dated to the late 3rd century. Although this group and KPG31 are broadly similar in terms of assemblage composition, KPG32 is likely to have been deposited later. This assertion is made on the basis of the bead-rimmed dishes. Here, they are represented by small abraded sherds and are probably residual. Other residual pottery includes the Central Gaulish samian, the Colchester colour-coated beaker, and probably the buff ware mortaria. The assemblage comprises relatively large sherds, and overall condition is good.

Key Pottery Group 33. Area R: Ditch 25270, segment 12027, fills 12026, 12029, OA58, Group 968 (Figure 264)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
BB1 14 575 41 9% 0.40 5% Dishes B1 B6
BB2 1 50 50 1% 0.07 1% Dish B4
BSW 188 2316 12 35% 2.81 34% Jar G23
CGSW 3 52 17 1% Dish f31R, bowl Curle 21
COLC 2 8 4 <1% Beaker
EGRHN 1 6 6 <1% Beaker
EGSW 1 6 6 <1% Bowl f37
GRF 8 208 263% 0.13 2% Dish B6, jar G24, beaker H32/H33
GRS 200 2474 12 37% 2 38% Dishes B1 B2 B5, jars G5 G9 G23, beaker H39
HAB 32 132 4 2% 0.08 1% Flask/beaker
NVC 38 474 12 7% 1.39 17% Beakers H28 H41
RET 14 332 24 5% 0.26 3% Jar G24
Total 502 6633 8.36

This group incorporates two fills from a ditch segment and is dated to the late 3rd or very early 4th century AD. There are no marked differences between KPG33 and previous groups. A later date, however, is suggested by the presence of the folded beaker (H39) in sandy grey ware and the Nene Valley colour-coated H41 beaker. The former is dated exclusively to the 4th century at Chelmsford (Going 1987, 31). In addition, the Hadham black-surfaced ware beaker (Figure 264, no.18) is a copy of a 4th-century Rhenish form. The condition of the pottery is mixed, although the smallest sherds belong to residual pottery, such as the East Gaulish Rhenish and Colchester colour-coated wares. Intrusive pottery, namely a sherd of Oxfordshire red colour-coated ware, has been omitted.

Ceramic Phase 10: c. AD 310-360

Seven groups, mainly from the southern settlement zone and with a total weight of 45kg and vessel rim equivalence (EVE) of 37, were assigned to Ceramic Phase 10. Three of these, totalling 13kg and 14 EVE and characteristic of the phase, are presented below. The Pit 10067 assemblage is comparable to a pit group from Chigborough Farm (Horsley and Wallace 1998, table 7), and the others to assemblages from Chelmsford, namely groups 17-18 (Going 1987, table 3).

Key Pottery Group 34. Area K: Pit 14125, fill 4315, OA28, Group 743 (Figure 265)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABAET 2 391 196 14%     Dressel 20
BB1 4 42 11 1% 0.09 4% Jar G9
BSW 18 370 21 12% 0.45 19% Dish B6, jar
COLBM 1 120 120 4%      
GRF 8 188 24 6% 0.41 18% Dishes B1 B2/B4, bowl-jar E5, beaker
GROG 7 112 16 4% 0.15 7% Platter Cam 21, jar
GROGC 1 10 10 <1% 0.04 2% Jar Cam 259
GRS 32 410 13 14% 0.58 25% Jars G5 G9
HAB 9 162 18 5% 0.32 14% Dish B1, bowl-jar E1, jar
STOR 12 1220 102 40% 0.25 11% Jar G42
TN 1 8 8 <1%     Platter
Total 95 3033     2.29   

This group is dated to the early 4th century AD. Groups of this date and those dated to the late 3rd century are not readily distinguishable from each other. The E1 bowl-jar, an early to mid-4th century form at Chelmsford (Going 1987, 21), and the absence of pottery dated exclusively to the mid-4th century or later, provide the strongest case for the assemblage being deposited during the early part of Phase 10. Locally produced reduced coarse wares remain dominant, though the bead-rimmed dish in fine grey ware suggests that some of this may well be residual. Notably, Hadham black-surfaced ware is also strongly represented. There is a clear residual element to this assemblage. It includes the grog-tempered wares and terra nigra, as well as the Colchester mortarium fabric and the amphora. The condition of the group is generally good.

Key Pottery Group 35. Area P: Pit 8745, fill 8766, OA50, Group 675 (Figure 266)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABAET 2 292 146 10%     Dressel 20
BB1 4 174 44 6% 0.30 9% Dish B6, jar G9
BSW 28 527 19 18% 0.78 25% Dishes B1 B2/B4 B6, bowl-jar, jar G12
EGSW 1 6 6 <1% 0.03 1% Mortarium f45
GRF 12 250 21 8% 0.70 23% Dish B6, bowl-jars E1E5
GROG 3 116 39 4%      
GROGC 2 110 55 4%      
GRS 76 1230 16 42% 1.24 39% Dish B4, jars G21 G35, beaker H24
HAR 3 46 15 2% 0.11 3% Bowl C8
HAWO 1 6 6 <1%      
HAX 1 22 22 1%      
MSR 1 32 32 <1%      
MWSRS 1 2 2 <1%      
NVC 3 28 9 <1%      
RET 1 12 12 <1%      
STOR 4 102 26 <1%      
Total 143 2955     3.16    

Pottery Group 35 dates to the first half of the 4th century AD. Fine grey ware, as appears to be the case in the majority of groups in Ceramic Phases 9 and 10, contributes significantly to the assemblage, and is present in almost equal proportions to black-surfaced and sandy grey wares. There is a greater variety of Hadham wares here, but, as usual, these fabrics make little impact in terms of overall quantities. Dishes are more prolific than jars, although some of them, specifically the bead-rimmed examples, are residual. Other residual pottery includes the ubiquitous grog-tempered pottery, the samian and the amphora. The pottery is in relatively good condition, and most sherds are fairly large and unabraded.

Key Pottery Group 36. Area E: Pit 10067, fill 10017, OA31, Group 837 (Figure 267)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ABAET 2 174 87 2%     Dressel 20
BB1 3 36 12 1% 0.09 1% Dishes B1 B6
BB2 2 50 25 1% 0.24 3% Dishes B1 B4
BSW 171 1660 10 23% 1.57 18% Dishes B1 B3 B4 B6, bowl-jar E2, jars G9 G26 G37
BUFM 2 52 26 1% 0.11 1% Mortaria D11 D13
CGRHN 1 1 1 <1%      
CGSW 17 90 5 1% 0.20 2% Dishes f18/31R f31 f79, bowls f30 or 37 Curle 15 or 23, cup f33
COLB 3 14 5 <1%      
COLBM 6 208 35 3% 0.15 2% Mortaria D11 D13 Cam 195
COLC 2 6 3 <1% 0.06 1% Beaker H
EGSW 3 48 16 1%     Dish f31R, mortarium f45
GRF 11 80 7 1% 0.14 2% Jar G9, beaker H
GROG 1 1 1 <1%      
GRS 513 3610 7 51% 5.54 62% Dishes B1 B2/B4, jars G5 G18 G21 G24 G25 G31 G36 G40, beaker H35
HAWO 2 20 10 <1%      
HAX 11 68 6 1% 0.11 1% Bowl C8
LSH 2 16 8 <1% 0.07 1% Jar G27
NVC 24 111 5 2% 0.48 6% Beakers H28 H32 H39 H41, lid K7
NVP 2 2 1 <1%      
OXRC 3 6 2 <1%     Bowl C8
RED 6 39 7 1%      
RET 6 86 14 1%      
STOR 16 715 45 10%     Jar G44
VRW 1 2 2 <1%      
Total 810 7095     8.76    

The pottery, such as the small amount of late shell-tempered ware, Oxfordshire red colour-coated ware and Hadham oxidised ware, provides a mid-4th century date for deposition. Locally made reduced coarse wares continue to predominate at more than 80% by weight. Aside from these, there is a wide range of fabrics, although these are never present in great amounts. Nene Valley colour-coated ware forms the largest group. KPG36 provides evidence for the initial appearance of late shell-tempered ware at Elms Farm. Dishes and jars remain prolific, though beakers are well represented in Nene Valley colour-coated ware. Much of the pottery is noticeably fragmentary. There is a high level of residuality, including the Colchester products, Verulamium white ware, grog-tempered pottery, samian ware, Central Gaulish Rhenish ware and Dressel 20 amphora.

Ceramic Phase 11: c. AD 360/70-400+

Twelve groups, with a total weight of 56kg and vessel rim equivalence (EVE) of 48.21, were assigned to Ceramic Phase 11. Three of these, from the central and southern settlement zones and totalling 12.5kg and 13.08 EVE, are presented below. These groups comprise the most securely dated accumulations of pottery assigned to this final ceramic phase and thus provide the most dependable evidence for understanding pottery supply and use in this period. The pottery from these three key groups is broadly comparable with assemblages from Chelmsford (Going 1987), Great Dunmow (Going and Ford 1988; Wallace 1997), Great Holts Farm, Boreham (Martin 2003) and Great Sampford (Martin 1998). Independent dating evidence in the form of late Roman coins comes from several CP11 key contexts, KPG37 in particular.

Key Pottery Group 37. Area J: Pit 5209, fill 5210, OA23, Group 442 (Figure 268)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
ALH 2 44 22 1%      
BB1 1 6 6 <1%      
BSW 63 1075 17 15% 1.44 20% Dishes B1 B3 B4 B5 B6, jar G
BUF 5 70 14 1% 0.08 1% Dish B6
CGSW 2 16 8 <1% 0.08 1% Cup f27
COLC 1 2 2 <1%      
EGSW 2 90 45 1%     Dish f31R
GRF 53 852 16 12% 1.36 19% Dishes B1 B2 B4 B3 B6, jar G, beaker ?H
GRS 140 1653 12 22% 1.83 25% Dishes B1 B6, jar G24
HAR 2 30 15 <1%      
HAX 18 144 8 2% 0.11 2% Bowl C, bowl-jar E
LSH 20 220 11 3% 0.41 6% Jar G27
NVC 28 530 19 7% 0.86 12% Dish B1, jar G, beaker H-folded, lid K7
NVM 15 420 28 6% 0.41 6% Mortarium D14
OXRC 27 218 8 3% 0.25 3% Bowls C8 C (1977, forms C48, C55)
OXRCM 1 6 6 <1%      
OXSWM 3 20 7 <1%      
OXWM 10 306 31 4% 0.20 3% Mortarium D (1977, form M22)
RED 5 44 9 1%     Beaker H-folded
RET 25 382 15 5% 0.17 2% Jar G24
STOR 17 1265 74 17% 0.12 2% Jar G44
VRW 1 12 12 <1%     
Total 441 7405        

This group produced a wide range of forms and fabrics, many of which are attributable to the period c. AD 360/70 onwards. 'Latest' Roman material, however, comprises a relatively small portion of the pottery, with Oxfordshire red colour-coated and late shell-tempered wares forming just 6% of the assemblage by weight. A small amount of Alice Holt grey ware is also present. Fine and sandy grey wares represent 34% of the assemblage by weight, while flint-tempered Rettendon wares account for a further 5%. When measured by weight, however, fabrics that are necessarily residual account for just over 2% of the total assemblage and are confined to BB1, Colchester colour-coated ware, Verulamium region white ware, and Central and East Gaulish samian. Several bead-rimmed dishes are present, which suggests that the level of residuality is higher than in the other late groups. Overall, the condition of the group seems to be quite good with few really abraded sherds. Seventeen 4th-century coins were also present, including one of Valentinian II (SF1882, 388-392) and a second which is Theodosian (SF1881, 388-408).

Key Pottery Group 38. Area M: Gully 25079, segment 15055, fill 15056, OA46, Group 469 (Figure 269)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
BB1 1 2 2 <1%      
BSW 28 160 6 6% 0.47 16% Dishes B1 B6, jar G
BUF 1 1 1 <1%      
GRF 41 442 11 16% 0.76 26% Dishes B1 B6, bowl-jar E5.4, beaker H
GRS 60 588 10 21% 0.23 8% Dishes B4 B6, jar G
HAR 5 14 3 <1%      
HAX 3 12 4 <1% 0.02 1% Bowl-jar ?E
LSH 19 170 9 6% 0.27 9% Jar G27.2
NVC 22 107 5 4% 0.26 9% Dish B, bowl ?C, beakers H42 H-funnel-neck
OXRC 28 372 13 13% 0.07 2% Bowl C (1977, forms C40 C45 C75), beaker H
OXW 2 190 95 7% 0.75 26% Flagon J(1977, form W25)
OXWM 1 164 164 6%      
RED 1 1 1 <1%      
RET 16 364 23 13% 0.10 3% Jar G24
STOR 2 238 119 8%      
Total 230 2825     2.93    

This relatively small group produced a range of forms and fabrics, attributable to the period c. AD 360/70 onwards. Latest Roman material comprises a significant portion of the pottery, with Oxfordshire red colour-coated and late shell-tempered wares together forming important assemblage components (19% by weight). Fine and sandy grey wares represent 37% by weight of the assemblage, while flint-tempered Rettendon wares account for a further 13%. This group comprises a narrow range of fabrics, but very little is clearly residual. The sole residual vessel form identified is a sandy grey ware bead-rimmed dish. Much of the Oxfordshire red colour-coated ware is very abraded, implying that this fabric may have had a different depositional history compared to the rest of the pottery.

Key Pottery Group 39. Area I: Well 5806, fill 5763, OA18, Group 639 (Figure 270)
Fabrics Sherd no. Weight(g) Average sherd wt % Weight EVE % EVE Forms
BSW 5 46 9 2% 0.06 2% Dish B2/B4
BUF 4 27 7 1%      
GRF 7 86 12 4% 0.35 12% Dish B6, jar G
GROG 1 8 8        
GRS 10 166 17 7% 0.16 6% Dish B2/B4, jar G28
HAB 2 27 14 1% 0.12 4% Dish B1.3
HAX 18 112 6 5% 0.06 2% Jar G
LSH 69 495 7 22% 0.95 34% Dish B1, jar G27
NVC 24 147 6 6% 0.25 9% Dish B1.2, bowl C
OXP 1 43 43 2% 0.07 2% Bowl C (1977 form P17)
OXRC 55 434 8 19% 0.81 29% Bowls C8 C25 C (1977, forms C69 C79), bowl-jars E4 E
OXRCM 1 6 6       Mortarium D12
OXSW 1 12 12        
OXSWM 4 94 24 4%     Mortarium D (1977, form WC7)
OXWM 1 24 24 1%      
STOR 6 564 94 25%      
Total 209 2291     2.83    

Pottery Group 39 is securely dated to the late 4th to early 5th centuries on grounds of both the pottery present and stratigraphy. While the high levels of late shell-tempered ware are closely paralleled in the 'late shrine' group at Great Dunmow (Going and Ford 1988, table 2), the level of Oxfordshire red colour-coated ware is exceptional. The range of Oxfordshire products is very wide, contrasting sharply with the paucity of grey ware forms. Fabrics dating exclusively to the later 4th century account for 47% of the total assemblage measured by weight. This all points strongly to a date well within this final Roman phase. Obviously residual pottery is barely represented in this group at less than 2% by weight, although the poor condition and low showing of the sandy grey wares imply higher levels of residuality. Much of the pottery is fragmentary. A coin of Valens (SF3310, 367-75) was recovered from the context.

Ceramic Phase: Post-Roman

In addition to the three groups described above, the fills within pit 14529 contained considerable amounts of Late Roman pottery. This material, however, is associated with substantial amounts of handmade Early Saxon pottery. Consequently, the group is probably post-Roman in date and is presented here as a means of investigating the relationship between the 'latest' Roman and Early Saxon material cultures. KPG40 is potentially of regional significance, as few groups of this type have been published supported by quantified data. The group may also be compared with other stratified groups of this nature from Essex. These comprise the pottery from a number of Early Saxon sunken-featured buildings (Drury and Wickenden 1982, 20-5) and the Early Saxon fill of pit 520 at Downhouse Farm, West Hanningfield (Tester and Martin Unpub.). The pottery from the Saxon settlement at West Stow, Suffolk, is also of use in this respect (Plouviez 1985). All of these groups provide important insights into the manner in which Roman pottery was deposited in contexts of 5th to 6th century date and the relationship with Early Saxon handmade wares of this period.

Key Pottery Group 40. Area L: Pit 14529, fills 14528, 14558, 14613, OA50, Group 722 (Figure 271)
Fabrics Sherd no. Weight (g) Average sherd wt % Weight EVE % EVE Forms
SAXON 115 2328 20 26% 1.47 17%  
ALH 7 272 39 3% 0.28 3% Dish B6, jar G
BB1 1 4 4 <1%      
BB2 1 14 14 <1% 0.08 1% Dish B2/B4
BSW 98 890 9 10% 1.02 12% Dishes B1 B3 B2/B4 B6, jar G, beaker ?H14, strainer M
BUF 4 46 12 1%      
CGSW 4 14 4 <1% 0.09 1%  
COLB 3 78 26 1%     Flagon ?J4
EGSW 1 8 8 <1% 0.03 <1%
GRF 42 462 11 5% 0.59 7% Dishes B1 B6 B, jar G, beaker H
GROG 16 284 18 3% 0.01 <1% Jar G
GRS 119 1150 10 13% 2.02 24% Dishes B1 B6, bowl-jar E5.2, jars G5 G24 G42 G
HAB 3 50 17 1% 0.08 1% Dish B2/B4
HAR 4 88 22 1% 0.11 1% Dish B6
HAWO 1 36 36 <1%     Jar G
HAX 10 254 25 3% 0.06 1% Dish B10 (1977, form C49), flagon ?J
LSH 31 358 12 4% 0.22 3% Jar G27
NVC 19 283 15 3% 0.52 6% Dishes B1 B6, bowl C8, bowl-jar E6.1, beaker H
OXP 1 7 10 <1%      
OXRC 37 740 20 8% 1.52 18% Bowls C8.1 C (1977, forms C45 C52 C70 C75 C83 C100)
OXSWM 4 116 29 1% 0.11 1% Mortarium D (1977, form WC7)
OXWM 1 56 56 1% 0.08 1% Mortarium D (1977, form M17)
RED 3 54 18 1%      
RET 7 290 41 3% 0.14 2% Jar G24
SGSW 2 6 3 <1% 0.06 1%  
STOR 15 1037 69 12% 0.12 1% Jar G
Total 545 8902     8.61    

Deposition of this group is likely to have occurred well beyond the Roman period. The assemblage is characterised by a large amount of Early Saxon pottery, which accounts for 26% of the total assemblage. In addition, a 5th-century cruciform brooch (SF6238) was found in fill 14558. The wide range of Roman pottery types present (430 sherds, weighing 6.6kg) includes Late Roman pottery typical of Ceramic Phase 11 horizons. Measured by weight, Oxfordshire products account for 10% of the total assemblage, while late shell-tempered ware represents a further 4%. A notable presence is the B6-type dish in Alice Holt grey ware, which is probably one of the latest Roman ceramic types to appear in Essex. Unfortunately, the presence of such a large amount of Saxon material suggests that all of the Roman pottery in the group is residual.

Pottery Sequence: The Catalogue

Key Pottery Group 1. Ditch 25094, Fills 19116 19145, Area P, OA3, Group 46
NumberFabric Code FormFigure 229: Key Pottery Group 1
1 ESH Jar Cam 254Figure 229
2 MICW Jar EF85
3 MICW Jar EF83
4 MICW Jar EF80
5 MICW Jar EF77
6 GROG Jar EF76
7 GROG ?Pedestal jar
8 GROG Jar Cam 259
Key Pottery Group 2. Ditch 25252, Fills 6875 6907 6957, Area H, OA5, Group 63
Number Fabric Code FormFigure 230: Key Pottery Group 2
1 MICW Bowl EF31Figure 230
2 GROG Bowl EF41
3 GROG Jar EF79
4 GROGC Jar EF107
5 MICW Jar EF98
6 GROG Jar EF106
7 GROG Jar EF104
8 GROG Jar EF94
9 GROG Pedestal base Cam 203
Key Pottery Group 3. Pit 8786, Fill 8785, Area P, OA3, Group 48
Number Fabric Code FormFigure 231: Key Pottery Group 3
1 MICW Pedestal-foot jar, MIA type Figure 231
2 MICW Conical jar, MIA type
3 GROG Jar EF89
4 GROG Jar EF86
5 GROG Jar EF110
6 GROG Jar EF81
7 GROG Jar EF82
8 GROG Jar EF138
9 GROG Jar EF139
10 GROGC Storage jar
11 TNM Platter Cam 1
12 NGWF Butt beaker Cam 113
Key Pottery Group 4. Pit 11342, Fill 11329, Area N, OA3, Group 58
Number Fabric Code FormFigure 232: Key Pottery Group 4
1 GROG Bowl Cam 212 Figure 232
2 GROGC Jar Cam 255
3 GROG Jar Cam 249
4 GROGC Jar EF119
5 GROGRS Jar Cam 256
6 GROGRS Jar Cam 256
7 GROGC Jar Cam 266
8 GROG Jar EF104
9 GROG Jar EF155
10 CGFCS Flagon Cam 165
Key Pottery Group 5. Pit 15417, Fills 15416 15418 15420 15490, Area M, OA2, Group 33
Number Fabric Code FormFigures 233/234/235: Key Pottery Group 5
1 TNM Platter Cam 1Figure 233
Figure 234
Figure 235
2 TR Platter Cam 2
3 TR Platter Cam 5A
4 TR Platter Cam 5B
5 PR Platter EF23
6 GROG Bowl Cam 253
7 IBUFM Mortarium EF74
8 ESH Jar Cam 255
9 GROG Jar Cam 259
10 GROG Jar Cam 204
11 GROG Jar Cam 218
12 GROG Jar Cam 221
13 GROG Jar Cam 221B
14 GROG Jar Cam 220
15 NGWF Butt beaker Cam 113
16 TR Butt beaker Cam 112
17 GROG Butt beaker Cam 118
18 GROG ?Butt beaker Cam 115
19 GROG Butt beaker Cam 115 base
20 GROG Butt beaker Cam 115
21 GROG Butt beaker Cam 115
22 CGFCS Flagon Cam 165
23 GROG Flagon EF196
24 AITAD Dressel 1 amphora
25 AITAC Dressel 1 amphora
26 AITAB Dressel 1 amphora
27 AITABDressel 1 amphora spike
Key Pottery Group 6. Pit 11344, Fill 11343, Area N, OA3, Group 59
Number Fabric Code FormFigure 236: Key Pottery Group 6
1 TR Platter Cam 5Figure 236
2 GROG Platter Cam 21
3 TNM Bowl Cam 51
4 GROG Dish EF26
5 GROG Jar Cam 219
6 GROGC Jar Cam 249
7 GROGC Jar Cam 255
8 GROGC Jar EF116
9 GROG Jar Cam 249
10 GROG Jar Cam 218
11 GROGC Storage jar Cam 271
12 CGMIC Beaker Cam 102
13 CGMIC Beaker Cam 102
14 TR Butt beaker Cam 112
15 GROG Butt beaker Cam 115
Key Pottery Group 7. Pit 9611, Fills 9585 9610, Area D, OA6, Group 75
Number Fabric Code FormFigure 237: Key Pottery Group 7
1 TNM Platter Cam 1Figure 237
2 TNM Platter Cam 1
3 TNM Platter Cam 1
4 TR Platter Cam 4
5 GROG Platter Cam 21
6 GROG Platter Cam 22
7 GROG Bowl Cam 214
8 GROG Bowl Cam 230
9 TR Cup Cam 56
10 GROGC Jar Cam 249
11 GROGC Jar G4
12 GROG Jar Cam 229A
13 GROG Jar EF153
14 GROG Jar EF154
15 GROGC Jar Cam 219
16 GROGC Jar Cam 218
17 GROGRS Jar G19.1
18 GROG Jar G19.1
19 GROG Jar G20
20 CGMIC Beaker Cam 102
21 TR Girth beaker Cam 82
22 TR Butt beaker Cam 112
23 GROG Jar EF103
24 CGFCS Flagon Cam 165
25 GROGC Funnel N2.1
Key Pottery Group 8. Pit 11316, Fills 11269 11301, Area N, OA25, Group 227
Number Fabric Code FormFigure 238: Key Pottery Group 8
1 TNM Platter Cam 1 with graffitoFigure 238
2 TNM Platter Cam 4
3 TN Platter Cam 2
4 PR Platter Cam 17
5 GROGRS Bowl Cam 43
6 GROG Bowl Cam 210
7 GROG Jar Cam 204
8 GROGC Jar Cam 254
9 GROGC Jar Cam 255
10 GROGC Jar Cam 257
11 GROG Jar EF117
12 GROGC Jar Cam 256
13 GROGC Jar Cam 259
14 CGMIC Beaker Cam 102
15 TR Girth beaker Cam 82
16 GROG Girth beaker Cam 85
17 GROG Butt beaker Cam 115
18 GROG Butt beaker EF189
19 CGFCS Flagon Cam 165
Key Pottery Group 9. Pit 8282, Fill 8271, Area E, OA30, Group 298
Number Fabric Code FormFigure 239: Key Pottery Group 9
1 GROG Platter Cam 21Figure 239
2 GROG Platter EF6
3 GROG Platter Cam 21
4 GROGC Bowl EF28
5 GROGC Jar Cam 255
6 GROGC Jar EF133
7 GROG Jar Cam 229
8 GROG Jar Cam 218/220
9 GROG Jar Cam 221
Key Pottery Group 10. Pit 19104, Fills 19105 19107 19109 19110 19111, Area P, OA3, Group 47
Number Fabric Code FormFigure 240: Key Pottery Group 10
1 GROG Bowl Cam 211Figure 240
2 GROG Bowl EF46
3 GROG Jar EF135
4 GROG Jar Cam 218
5 GROG Jar EF94
6 TR Butt beaker Cam 112
7 NGWFS Bossed butt beaker EF195
8 MICW Jar EF78 (residual)
Key Pottery Group 11. Pit 8026, Fills 8003 8014 8018, Area E, OA31, Group 296
Number Fabric Code FormFigure 241: Key Pottery Group 11
1 GROG Platter EF16Figure 241
2 GROGRF Bowl EF72
3 GROG Bowl Cam 241/246
4 GROG Cup Cam 212
5 ESH Jar Cam 254
6 GROG Jar EF134
7 GROG Jar EF146
8 GROG Jar EF179
9 GROG Jar Cam 259
10 GROGC Jar Cam 256
11 GROGC Jar Cam 259
12 GROG Jar Cam 219
13 GROG Jar Cam 218
14 GROG Jar Cam 220
15 GROG Butt beaker Cam 115
Key Pottery Group 12. Pit 9230, Fill 9231, Area D, OA29, Group 288
Number Fabric Code FormFigure 242: Key Pottery Group 12
1 GROG Bowl Cam 212Figure 242
2 GROGRS Bowl EF73
3 GROG Jar Cam 225
4 GROG Jar Cam 256
5 GROG Jar Cam 249
6 GROG Jar Cam 229A
7 TR Butt beaker Cam 112
8 GROGRS Flagon handle
9 GROG Lid K6
Key Pottery Group 13. Pit 7167, Fills 7168, 7178, 7179, Area G, OA36, Group 313
Number Fabric Code FormFigure 243: Key Pottery Group 13
1 TNM Platter EF20Figure 243
2 GROG Platter Cam 21
3 GROG Platter Cam 21
4 GROG Platter Cam 33
5 GROG Bowl Cam 210
6 GROG Bowl EF38
7 GROGBowl Cam 252
8 GROG Bowl EF42
9 GROG Jar EF129
10 GROG Jar Cam 254
11 GROG Jar Cam 259
12 GROG Jar Cam 259
13 GROG Jar Cam 259
14 GROG Jar Cam 229
15 GROG Jar G17
16 GROG Jar Cam 218
17 GROG Jar Cam 204
18 GROG Jar G19
19 GROG Jar G19
20 TNM Butt beaker EF192
21 TR Butt beaker Cam 112
22 GROG Butt beaker H7
23 GROG Beaker EF188
Key Pottery Group 14. Pit 11723, Fill 11720, Area N, OA26, Group 142
Number Fabric Code FormFigure 244: Key Pottery Group 14
1 GROG Platter Cam 21Figure 244
2 GROG Platter Cam 21
3 GROG Platter Cam 21
4 GROG Platter A2.4
5 GROG Bowl Cam 212
6 GROGRS Bowl EF71
7 GROG Jar Cam 249
8 GROG Jar Cam 256
9 GROG Jar Cam 258
10 GROGC Jar Cam 260
11 GROG Jar G19
12 GROGRF Butt beaker Cam 116
13 CGMIC Jar Cam 262 (residual)
Key Pottery Group 15. Ditch 25018, Fill 9214, Area D, OA29, Group 764
Number Fabric Code FormFigure 245: Key Pottery Group 15
1 GROG Platter A2Figure 245
2 GRF Bowl EF57
3 BSW Jar G19
4 BSW Jar G20.1
5 GROGC Storage jar G45
6 GROG Jar Cam 232
7 GROGC Jar Cam 259
8 STOR Jar EF178
9 STOR Jar EF126
10 STOR Storage jar
Key Pottery Group 16. Pit 9218, Fills 9217 9370, Area D, OA31, Group 768
Number Fabric Code FormFigure 246: Key Pottery Group 16
1 SGSW Platter f15/17Figure 246
2 GROG Platter Cam 21
3 GROG Platter Cam 21
4 GROG Platter A2
5 COLB Bowl (Bidwell and Croom 1999, Cam 327)
6 GRF Bowl EF64
7 STOR Bowl C33
8 STOR Bowl C33
9 GROGC Jar EF135
10 BSW Bowl EF57
11 GROG Jar EF131
12 GROGC Jar Cam 259
13 GROGC Jar Cam 259
14 GROG Jar Cam 249
15 GROG Jar Cam 258
16 GROGC Jar Cam 259
17 GROGC Jar Cam 259
18 BSW Jar G3
19 STOR Storage jar Cam 270B
20 STOR Storage jar Cam 271
21 GROG Jar G19.1
22 GROG Jar G20
23 GROG Jar G20
24 GRS Jar G20
25 STOR Storage jar
26 GRF Beaker H1
27 GROG Beaker H1
28 TR Butt beaker Cam 112
29 GROGC Lid
30 GROGC Lid K3
31 GROG Funnel N2
Key Pottery Group 17. Pit 20008, Fill 20009, Area L, OA27, Group 708
Number Fabric Code FormFigure 247: Key Pottery Group 17
1 BSW Platter A2 with piece removed from rimFigure 247
2 BSW Bowl Cam 214
3 GRF Bowl C12
4 BSW Bowl C27
5 BSW Jar G17 with two notches on the rim
6 GRS Jar G17/Cam 232
7 BSW Jar G19.4 with holes in base and walls
8 BSW Jar G20
9 GRF Jar G20
10 BSW Jar G20.1
11 GRS Jar G23
12 GRF Jar G40 with piece removed from rim
13 STOR Storage jar G44
14 BSW Beaker Cam 108 with piece removed from rim
15 IMIC Beaker H1
16 GRF Butt beaker H7.1
17 COLB Flagon J
18 BSW Lid K6 with piece removed from rim
Key Pottery Group 18. Pit 24013, Fill 24014, Area M, OA26, Group 692
Number Fabric Code FormFigure 248: Key Pottery Group 18
1 GRS Platter A2.2Figure 248
2 BSW Platter A4
3 BSW Dish B8
4 BUFM Mortarium D1
5 VRWM Mortarium D1
6 SGSW Cup f35
7 GRF Cup samian f27 copy
8 GRF Jar G20
9 GRS Jar G3
10 GRS Jar G3
11 GRS Beaker H1
12 GRS Lid K6
Key Pottery Group 19. Pit 15773, Fill 24258, Area M, OA26, Group 688
Number Fabric Code FormFigure 249: Key Pottery Group 19
1 GRS Platter A2.3Figure 249
2 GRS Jar G17
3 BSW Jar G17
4 BSW Jar G19
5 BSW Jar G20.1
6 BSW Jar G23
7 RED Flagon J3.1
Key Pottery Group 20. Pit 6201, Fill 6203, Area H, OA23, Group 530
Number Fabric Code FormFigure 250: Key Pottery Group 20
1 BSW Platter A2Figure 250
2 GRS Jar G8
3 GRS Jar G17
4 BSW Jar G20
5 GRS Jar G23
6 STOR Storage jar G44
7 BSW Beaker H1
8 GRS Lid
Key Pottery Group 21. Pit 5147, Fill 5146, Area J, OA23, Group 409
Number Fabric Code FormFigures 251/252: Key Pottery Group 21
1 SGSW Platter f15/17Figure 251
Figure 252
2 SGSW Platter f18
3 GRF Platter A2
4 BSW Platter A2
5 BSW Platter A2
6 BSW Platter A2
7 NKG Platter A4
8 BSW Dish B1
9 BB2 Dish B2
10 GRS Dish B7
11 GRS Dish B7
12 BSW Dish B7
13 BSW Dish B7
14 CGSW Bowl f37
15 BSW Bowl with thickened bead rim
16 BSW Bowl C1
17 BSW Bowl C12
18 NKO Bowl (Monaghan 1987, type 4I)
19 BSW Bowl with everted rim and lattice decoration
20 BUFM Mortarium D1
21 SGSW Cup f27
22 BSW Jar G16
23 BSW Jar G16
24 BSW Jar G16
25 BSW Jar G19
26 GRS Jar G20
27 BSW Jar G20
28 GRS Jar G23
29 GRS Jar G23
30 GRS Jar G23
31 BSW Jar G23
32 GRS Necked jar with globular body
33 GRF Beaker H1
34 BSW Beaker H1
35 BSW Beaker H1
36 BSW Beaker H4
37 BSW Lid K3
38 BSWMiniature
Key Pottery Group 22. Ditch 25245, Fill 10182, Area F, OA31, Group 361
Number Fabric Code FormFigures 253/254: Key Pottery Group 22
1 NKG Platter A4Figure 253
Figure 254
2 BSW Platter A4.5
3 BB2 Dish B2.3
4 BB2 Dish B2.3
5 BB2 Dish B2
6 BB2 Dish B
7 BB2 Dish B4
8 BB2 Dish B4
9 BSW Dish B4
10 GRS Dish B4
11 GRS Dish B4
12 BSW Dish B7
13 BSW Dish B7
14 BSW Dish B9
15 GRF Dish B10
16 BSW Dish B2-type, decorated
17 SGSW Bowl f37
18 GRS Bowl C16
19 LESTA Bowl C2
20 GRS Jar G16
21 BSW Jar G17
22 BSW Jar G3
23 GRS Jar G20
24 GRS Jar G20
25 GRS Jar G22
26 GRS Jar G22
27 BSW Jar G23
28 BSW Jar G23
29 GRS Jar G23
30 GRS Jar G23
31 GRF Jar G29
32 GRS Jar G29
33 BSW Jar G40
34 STOR Storage jar G45
35 GRS Bowl-jar E6 variant
36 GRS Jar G42
37 BSW Jar base with graffito
38 COLC Beaker H20
39 BSW Lid K3
40 BSW Miniature
Key Pottery Group 23. Pit 7118, Fills 7119 7166, Area G, OA34, Group 852
Number Fabric Code FormFigure 255: Key Pottery Group 23
1 CGSW Dish base with X graffitoFigure 255
2 BB2 Dish B1
3 BB2 Dish B2.4
4 GRS Dish B2/B4
5 BSW Dish B4.1
6 BSW Dish B4.1
7 BB2 Dish B4.2
8 COLBM Mortarium D13.1
9 BSW Jar G20
10 GRS Jar G23
11 BSW Jar G29.2
12 BSW Jar G40
13 STOR Storage jar G45
14 BSW Jar G5.2
15 RED Beaker H20
16 COLC Beaker H20.2
17 GRS Lid
Key Pottery Group 24. Pit 9029, Fills 9028 9064, Area D, OA31, Group 783
Number Fabric Code FormFigure 256: Key Pottery Group 24
1 BSW Dish B4Figure 256
2 GRS Dish B3
3 BSW Dish B4
4 BB2 Dish B4.2
5 GRS Dish B7
6 GRS Bowl C1.1
7 GROG Bowl C33
8 GRSWSM Mortarium D1
9 GRS Jar G19
10 GRS Jar G20
11 GRS Jar G23
12 BSW Jar G23
13 BSW Jar G29.1
14 STOR Jar G36
15 NKG Beaker (Monaghan 1987, type 2H1)
16 BSW Base with X graffito
Key Pottery Group 25. Pit 7122, Fill 7123, Area G, OA34, Group 868
Number Fabric Code FormFigure 257: Key Pottery Group 25
1 BSW Dish B1Figure 257
2 LOND Bowl C22 (Monaghan 1987, type 4H1)
3 BUF Tazza-bowl Cam 198
4 MIC Bowl C16
5 COLBM Mortarium D1.3 with herringbone stamp
6 COLBM Mortarium D1.3 with stamp of Martinus
7 COLBM Mortarium D13
8 BSW Jar G29 (Monaghan 1987, type 2G2)
9 GRS Jar G5.5
10 GRS Jar G17
11 GRS Jar G17
12 BSW Jar G19.5
13 GRS Jar G22
14 BSW Jar G23
15 GRS Jar G24
16 GRS Jar G25
17 COLC Beaker H20.2
18 MWSRS Flagon J4
19 GRS Lid K6
Key Pottery Group 26. Well 6280, Fill 16083, Area H, OA19, Group 531
Number Fabric Code FormFigure 258: Key Pottery Group 26
1 CGSW Dish f31R, rivetedFigure 258
2 CGSW Dish f31R
3 BSW Dish B2
4 BB2 Dish B2
5 BSW Dish B3
6 BSW Jar G23
7 BSW Jar G9.3
8 MIC Beaker H20
9 COLC Beaker H24.1
10 COLC Beaker H35
Key Pottery Group 27. Kiln 1223/1618, Stoke-hole 1589, All deposits, Area W, OA38, Group 910
Number Fabric Code FormFigure 259: Key Pottery Group 27
1 UPOT Dish B2/B4Figure 259
2 BUFM Mortarium D3
3 BUFM Mortarium D3
4 BUFM Mortarium D3
5 BUFM Mortarium D3
6 BUFM Mortarium D11
7 BUFM Mortarium D11.1
8 BUFM Mortarium D11.1
9 BUFM Mortarium D11.1
10 BUFM Mortarium D11.1
11 GRS Jar G5.5
12 GRS Jar G5.5
13 GRS Jar G5.5
14 GRS Jar G5
15 GRS Jar G10.1
16 UPOT Flagon J9
Key Pottery Group 28 Pit 6182, Fill 6178, Area H, OA19, Group 553
Number Fabric Code FormFigure 260: Key Pottery Group 28
1 GRS Dish BFigure 260
2 BSW Dish B4
3 EGSW Bowl f38
4 BSW Jar G5.5
5 GRS Jar G5.5
6 GRS Jar Cam 218
7 STOR Storage jar G44.2
8 GRF Beaker base with graffito
9 CGRHN Beaker (Symonds 1992, fig. 9, Gp 9)
10 BSW Beaker H34
Key Pottery Group 30 Pit 10062, Fill 10061, Area E, OA31, Group 811
Number Fabric Code FormFigure 261: Key Pottery Group 30
1 BSW Dish B1.3Figure 261
2 BSW Dish B4.2
3 BSW Dish B
4 BSW Dish B5.1
5 NVM Mortarium (Perrin and Hartley 1996, fig. 116, M128)
6 GRS Jar G22
7 BSW Lid K2
Key Pottery Group 31 Pit 11303, Fill 11302, Area N, OA50, Group 671
Number Fabric Code FormFigure 262: Key Pottery Group 31
1 GRF Dish B2Figure 262
2 BSW Dish B2/B4
3 BB2 Dish B3
4 HAR Dish B
5 BSW Dish B4
6 BB1 Dish B6.3
7 EGSW Bowl f37
8 CGSW Mortarium f45
9 BUFM Mortarium D3.3
10 NVM Mortarium D14
11 BSW Bowl-jar E5
12 GRS Jar G5
13 BSW Jar G5.5
14 STOR Storage jar G44
15 HAWO Pedestal base Cam 296
16 MWSRS Flagon Cam 365
Key Pottery Group 32: Pit 6267, Fill 6268, Area H, OA19, Group 561
Number Fabric Code FormFigure 263: Key Pottery Group 32
1 BSW Dish B1Figure 263
2 BSW Dish B2/B4
3 GRS Dish B
4 BUFM Mortarium D3
5 OXWM Mortarium D5
6 BUFM Mortarium D11
7 GRS Bowl-jar E2
8 BSW Jar G5.5
9 GRS Jar G9
10 GRS Jar G24
11 RET Jar G24.1
12 GRF Beaker H33
Key Pottery Group 33: Ditch 25270, Fills 12026 12029, Area R, OA58, Group 968
Number Fabric Code FormFigure 264: Key Pottery Group 33
1 BB1 Dish B1.4Figure 264
2 BSW Dish B1
3 GRS Dish B1
4 BB2 Dish B4
5 GRS Dish B5.1
6 BB1 Dish B6.3
7 GRS Jar G5.5
8 GRS Jar G9.3
9 RET Jar G23
10 GRS Jar G23
11 GRS Jar G23
12 BSW Jar G38
13 HAB Beaker copying Symonds (1992) Gp 67
14 NVC Beaker H28
15 GRS Beaker H39
16 NVC Beaker H41
17 NVC Beaker H41
18 GRF Beaker base with X graffito
Key Pottery Group 34. Pit 14125, F1ll 4315, Area K, OA28, Group 743
Number Fabric Code FormFigure 265: Key Pottery Group 34
1 BSW Dish B6.2Figure 265
2 HAB Bowl-jar E1.1
3 GRF Bowl-jar E5.4
Key Pottery Group 35. Pit 8745, Fill 8766, Area P, OA50, Group 675
Number Fabric Code FormFigure 266: Key Pottery Group 35
1 BSW Dish B1Figure 266
2 BSW Dish B2/B4
3 BSW Dish B2/B4
4 GRS Dish B4
5 BB1 Dish B6.3
6 BB1 Dish B6
7 GRF Dish B6
8 BSW Dish B6.2
9 BSW Dish B6.2
10 HAR Bowl C8
11 GRF Bowl-jar E5.4
12 BSW Bowl-jar E6
13 BB1 Jar G9
14 BSW Jar G12
15 GRS Jar G21
16 GRF Jar with flange-like rim and looped decoration
17 GRS Jar with rolled-over rim
18 GRF Beaker
Key Pottery Group 36. Pit 10067, Fill 10017, Area E, OA31, Group 837
Number Fabric Code FormFigure 267: Key Pottery Group 36
1 BB1 Dish B6.3 Figure 267
2 HAX Bowl C8
3 BSW Bowl-jar E2.3
4 LSH Jar G27
5 GRS Jar G37
6 GRS Jar G40
7 GRS Jar G5.4
8 NVC Beaker H32
9 NVC Beaker H32
Key Pottery Group 37. Pit 5209, Fill 5210, Area J, OA23, Group 442
Number Fabric Code FormFigure 268: Key Pottery Group 37
1 BSW Dish B1 Figure 268
2 BSW Dish B1
3 BSW Dish B1
4 BSW Dish B1
5 GRF Dish B1
6 GRF Dish B1
7 NVC Dish B1.2
8 BSW Dish B
9 BSW Dish B4
10 GRF Dish B4
11 BSW Dish B5.1
12 BSW Dish B6.1
13 BUF Dish B6.1
14 BSW Dish B6.2
15 GRS Dish B6.2
16 GRS Dish B6.2
17 GRF Dish B6.2
18 GRF Dish B6.2
19 OXRC Bowl (1977, type C55)
20 NVM Mortarium D14
21 NVM Mortarium (Hartley and Perrin 1999, fig. 77, M9)
22 OXWM Mortarium (Young 1977, type M22) with reeded flange
23 OXWM Mortarium (Young 1977, type M22)
24 LSH Jar G27.1
25 LSH Jar G27.2
26 LSH JarG27.2
27 LSH Jar G27.2
28 NVC Jar G28 (Howe et al. 1980, no. 70)
29 NVC Narrow-necked jar
30 GRF Funnel-necked beaker
31 GRF Everted-rimmed beaker
32 NVC Lid K7
Key Pottery Group 38. Gully 25079, Fill 15056, Area M, OA46, Group 469
Number Fabric Code FormFigure 269: Key Pottery Group 38
1 BSW Dish B1Figure 269
2 BSW Dish B1
3 GRF Dish B1
4 BSW Dish B6.2
5 GRF Dish B6.2
6 GRS Dish B6.2
7 OXRC Bowl (1977, type C40)
8 NVC Bowl (Perrin 1999, fig. 64, nos 248-50)
9 GRF Bowl-jar E7
10 RET Jar G24.2
11 LSH Jar G27.2
12 NVC Beaker H41
13 GRF Beaker
14 OXW Flagon (1977, type W25)
Key Pottery Group 39. Well 5806, Fill 5763, Area I. OA18, Group 639
Number Fabric Code FormFigure 270: Key Pottery Group 39
1 LSH Dish B1Figure 270
2 LSH Dish B1
3 NVC Dish B1.2
4 HAB Dish B1.3
5 GRF Dish B6.2
6 OXP Bowl (1977, type P17)
7 OXRC Bowl (1977, type C69)
8 OXRC Bowl C8 (1977, type C51)
9 OXRC Bowl-jar E4 (1977, type C75)
10 OXRC ?Bowl-jar (1977, ?type C73)
11 NVC Bowl, probably Howe et al. (1980), no. 82
12 LSH Jar G27
13 LSH Jar G27
14 GRS Jar G28
15 GRF Jar G42
Key Pottery Group 40. Pit 14529, Fills 14528 14558 14613, Area L, OA50, Group 722
Number Fabric Code FormFigure 271: Key Pottery Group 40
1 NVC Dish B1Figure 271
2 GRF Dish B1.3
3 BSW Dish B1.2
4 HAR Dish B6
5 ALH Dish B6
6 GRF Dish B6
7 OXRC Bowl (Young 1977, type C77)
8 OXRC Bowl (Young 1977, type C74)
9 OXRC Bowl (Young 1977, type C70)
10 OXRC Bowl C8 (Young 1977, type C52)
11 OXRC Bowl C8 (Young 1977, type C52)
12 OXRCM Mortarium D6 (Young 1977, type C100)
13 OXSWM Mortarium D (Young 1977, type WC7)
14 OXWM Mortarium D (Young 1977, type M21)
15 NVC Bowl-jar E6.1
16 GRS Jar G24.1
17 LSH Jar G27.2
18 GRS Jar G24

The Pattern of Pottery Supply to Heybridge

This section uses the data extracted from the Key Pottery Groups presented in the Pottery Sequence section, along with the expanded pool of securely dated groups listed in the archive. The trends in the pattern of pottery supply and assemblage composition are derived from the data and presented in Ceramic Phase order. Quantities depicted on the maps are based on the EVE total for each fabric present in a given ceramic phase, expressed as a percentage of the EVE total for all fabrics. Where fabrics are in amounts too small to be recorded by EVE, square symbols denote their presence. A discussion of supply patterns and changes through time follows the maps. Chelmsford and Colchester remain key points of reference, though other sites, mainly located within the region, are introduced for comparison. A discussion of assemblage composition for each ceramic phase follows. Again, changes through time are highlighted and comparisons are made between Heybridge and other sites. Quantified groups from any period in Essex are comparatively few and are mainly late Roman in date. The groups presented in this report add significantly to the overall number, together with those from the settlement at Great Holts Farm, Boreham (Martin 2003).

Context numbers given in the text usually refer to pottery groups from the expanded dataset; otherwise they can be found in the stratigraphic narrative (Section 2). It should be noted that assemblages comprising the expanded dataset were used up to and including Ceramic Phase 10. For CP11, and for the post-Roman phase, information has been extracted from the four relevant Key Pottery Groups only (KPGs 37-40). Versions of most form types mentioned are illustrated in the Pottery Sequence section. Otherwise, for the Roman period, forms can be found in the Chelmsford type series (Going 1987, 13-54), or the updated Camulodunum/Colchester series (Bidwell and Croom 1999, 468-87), which is also useful for the Late Iron Age. Occasional reference is made to regional typologies, such as Monaghan's Upchurch/Thameside (1987) and Young's Oxfordshire (1977) series. Attention is drawn to the typology of Late Iron Age forms, comprising types found at Heybridge but not attested at Camulodunum.

To avoid repetition, the word 'phase' is often used instead of 'ceramic phase'. These are one and the same. Naturally, the division of this section into ceramic phases results in a somewhat disjointed discussion. However, a more cohesive overview, drawing together the trends presented here and bringing in further interpretation, is taken up in the related synthetic monograph (Atkinson and Preston 2015).

Key for ceramic phase maps
Blocked-in sectors on the pie charts represent the %EVE for a given fabric
■ indicates fabric presence only, where EVE could not be measured
□ (open or white square) represents Heybridge

Ceramic Phase 1: c. 50-15 BC

Figure 272
Figure 272: Ceramic Phase 1 pottery supply map (c. 50-15 BC)

Assemblages assigned to the first part of Ceramic Phase 1 include a significant element of locally made, sand-tempered pottery, characteristic of the middle to Late Iron Age transition, although grog-tempered fabrics generally form the major component. Dressel 1 amphoras occur sporadically, their numbers increasing towards the end of the phase. Also late in the phase, other imports are apparent in the form of Central Gaulish micaceous flagons and platters. Ceramic Phase 1 heralds a change from local middle Iron Age traditions, marked by the introduction of grog-tempered pottery, closely followed by the establishment of wheel-throwing techniques and the increasing importation of amphoras. This pattern, which is in keeping with settlements elsewhere in south-eastern Britain, is amply demonstrated by Key Pottery Groups 1 to 4.

Drury's discussion of the early and middle Iron Age pottery from Little Waltham (1978, 51-63) forms the basis for the study of Iron Age pottery in central Essex. At Little Waltham, 14km to the west of Heybridge, assemblages are dominated throughout by early/middle Iron Age, coarse sand-tempered pottery (Drury's Fabric H). Grog-tempered pottery (Fabric E) did not appear until c. 50-25 BC, forming just 9% of the total overall (Drury 1978, table 10), whereas at Elms Farm this fabric comprises an average of 80% by weight in Ceramic Phase 1 assemblages. The reason for this may be largely one of chronology, as the settlement at Little Waltham was in decline in the later 1st century BC, in contrast to Heybridge. It is unlikely that new material was thus being introduced at Little Waltham at this time.

The pottery from Heybridge is more comparable with pottery found at Kelvedon (Rodwell 1988), 11km distant. There, three main fabrics were identified, a coarse sand-tempered fabric, equivalent to Drury's Fabric H, a second tempered with grog, and another intermediate between the two, but with the addition of occasional flint grits. The forms present in these fabrics ranged from handmade vessels in sand-tempered fabrics, through to grog-tempered, cordoned, wheel-thrown bowls and jars (Rodwell 1988, 102). A large group (22.7kg) from Ditch 350, representative of all of these forms and fabrics, was assigned a mid- to late 1st century BC date (c. 50-20 BC). Grog-tempered pottery forms 77% of this ditch assemblage. At Heybridge, the proportions of forms and fabrics are entirely consonant with those at Kelvedon.

A similar assemblage was excavated from Ditch AF1 at nearby Woodham Walter (Buckley et al. 1987), where the proportion of wheel-thrown grog-tempered and handmade sand-tempered wares appears to be roughly equal. Unfortunately, the pottery was not fully quantified. Rodwell proposed a date towards the middle of the 1st century BC for its deposition (1987a, 38).

The middle Iron Age pottery from Elms Farm has been discussed by Brown (2001, 64), but precise dating within the middle Iron Age is difficult in the absence of supporting scientific dating (e.g. dendrochronology). All pottery made in the middle Iron Age tradition found in Late Iron Age contexts was assigned a single fabric code, MICW, during recording. Most assemblages assigned to Ceramic Phase 1 contain this mainly sand-tempered fabric, in an amount averaging 11% of the total by weight. Only in one feature, ditch 25174, does sand-tempered coarse ware greatly exceed this figure, although the quantity of pottery recovered is small (1.19kg). The potentially early date for this group, perhaps c. 50 BC, is supported by the stratigraphy.

The similarity between the pottery assigned to the first part of Ceramic Phase 1 and that from Ditch 350 at Kelvedon indicates that a date of c. 50-20 BC can also be confidently assigned to these assemblages at Elms Farm. Independent dating evidence in the form of two coins lends weight to the assigned date. A Class I/II potin (SF5645) in the bottom fill of ditch segment 16018 (KPG2) supports a date in the 1st century BC, although most of the pottery present in this feature consists of handmade grog-tempered ware. The second coin, a Class II potin (SF6871), is associated with handmade grog-tempered pottery only and also supports a date of c. 50-20 BC.

Early shell-tempered ware appears in four Ceramic Phase 1 contexts, although the quantities are very low, amounting to an average of 3% by weight. This fabric is more common in the Thameside area of south Essex and north Kent, where it was probably produced, forming only a small component of central Essex assemblages. The pottery from Elms Farm conforms, with the fabric maintaining a low proportion in all ceramic phases.

The level of imports in Ceramic Phase 1 is low throughout, mostly appearing towards the end of the phase. Italian amphoras average 4% by weight. Imported fine wares, forming an average of 3% by weight, consist of Central Gaulish micaceous wares and North Gaulish white ware. Central Gaulish imports appeared in Britain slightly earlier than Gallo-Belgic, which were more common from c. 15 BC (Rigby 1989, 121). The evidence at Elms Farm confirms that importation of Central Gaulish micaceous wares occurred at least a decade before the more widespread Gallo-Belgic types. Central Gaulish ware forms a major component, compared with other imports, from the middle of this phase until the end of Ceramic Phase 2. North Gaulish white ware vessels were sporadic arrivals before the 1st century AD, but their presence in Ceramic Phase 1 groups confirms earlier importation at Heybridge.

Imported pottery on other sites in the vicinity is scarce, especially that dated early in the Late Iron Age. A possible Cam 165 flagon, probably from the New Cemetery, Heybridge, was donated to the Colchester and Essex museum in 1922 (Wickenden 1986, 55), but all other imports of this date found in the vicinity comprise amphoras (Wallace 1998, table 8). At Woodham Walter, a cream-slipped, two-handled flagon was found in Ditch CF101, dated mid-1st century AD, and described by Rodwell (1987a, 32; fig. 25.172) as a Gallo-Belgic copy. The vessel, pace Wallace (1998, 155), is probably not a Central Gaulish import; the form, as illustrated, and the fabric description indicate that the flagon is a later local copy of a Gallo-Belgic form.

Assemblage composition

The range of vessel forms in Ceramic Phase 1 is very restricted, mainly comprising jars and bowls in coarse fabrics. Vessels from both KPG1 and KPG2 are generally handmade with little difference in style between those in sand-tempered fabric and grog-tempered examples. Bowl types are largely indistinguishable from jars and together these predominate at 85% by vessel rim equivalence (EVE). Forms are simple with everted or bead rims and any decoration is confined to combing on the lower body or lines of stabbing along the shoulder. Wheel-thrown pottery seems to occur in slightly later groups, perhaps first appearing in the decade c. 40-30 BC, although this is difficult to date with precision. The range of forms is again restricted to jars and bowls, either plain or with rippled or cordoned shoulders. The few identifiable bowl types, at 6% by EVE, are mainly those with one or more cordons at the mid-body constriction (Cam 211), reminiscent of vessels made in wood or shale. The only identified vessel form in early shell-tempered ware is the club-rimmed jar Cam 254; all Ceramic Phase 1 vessels in this fabric are handmade.

Platters, cups, beakers and flagons appear towards the end of the ceramic phase, but are few in number and mostly imported. There are also two examples of the carinated and pedestalled tazza-bowl Cam 210. Platters form 2% of the assemblage by EVE and comprise the micaceous Cam 1, grog-tempered copies of these, Cam 21, and later in the phase, a single grog-tempered Cam 28. Cups are grog-tempered and restricted to three examples of the carinated Cam 212, which are smaller versions of the Cam 211 bowl. Flagons are rare and consist of two examples of the Central Gaulish Cam 165, plus a grog-tempered handle, which is probably from a flagon. Grog-tempered flagons are an uncommon form, with a total of twenty-two examples identified, mainly by their handles. Fifteen come from stratified contexts, ranging in date from late 1st century BC to late 1st century AD. Beakers are also uncommon, consisting of a single imported Cam 113 butt beaker and sherds from a 'thorn' beaker (Figure 295, no. 7), a rare British site find. The few grog-tempered beakers noted could just be small, 'hand-sized' examples of jars. There is one grog-tempered lid. At the end of Ceramic Phase 1 grog-tempered pottery is beginning to show the diversity of form that characterises Ceramic Phase 2. Amphoras form a small part of Phase 1 assemblages with examples in just four contexts, comprising neck and body sherds from Italian Dressel 1 wine amphoras only.

Ceramic Phase 2: c. 15 BC-AD 20

Figure 273
Figure 273: Ceramic Phase 2 pottery supply map (c. 15 BC-AD 20)

Grog-tempered pottery continues to dominate in Ceramic Phase 2, averaging 89% by weight. Sand-tempered coarse wares have virtually disappeared and are probably residual in most contexts assigned to this phase. Early shell-tempered ware also forms a minor component and is present in only three contexts. Assemblages display a greater diversity of fabric and form, as imports proliferated and the range of grog-tempered copies rapidly expanded. Amphoras have increased both in numbers and types present, with Catalan and salazon vessels from Spain making an appearance alongside Italian wine amphoras. Central Gaulish micaceous ware is present in many Ceramic Phase 2 contexts, showing a large increase on the quantities in Ceramic Phase 1. Gallo-Belgic and North Gaulish fine wares are present in smaller numbers. Assemblages assigned to this phase have begun to exhibit the results of increased trade with Gaul as the Augustan period progressed.

The earliest Key Pottery Group (KPG5) of this ceramic phase is typical in terms of the types of pottery present but atypical in terms of relative assemblage composition. The group differs in the proportion and quantity of imports present. The main deposit, 15416, is considered to be pyre-related debris. The pottery, therefore, is likely to have undergone selection both before, and after, destruction on the pyre. Nevertheless, it illustrates the range of imports available during Ceramic Phase 2. The imported wares include Central and North Gaulish wares as found in Phase 1, with the addition of Gallo-Belgic ware in the form of terra rubra. The assemblage is dominated by the presence of substantial parts of three Dressel 1 amphoras, accounting for 72% of the total by weight. Grog-tempered pottery forms just under 20%, although this figure increases to 71% if the disproportionate weight of the amphoras is removed from the totals. The proportion of grog-tempered pottery is lower than the average, because of the variety of imported fine wares which form 23% by weight (amphora weight excluded). Early shell-tempered pottery accounts for 2% of the total by weight (6% without the amphora) and is represented by a wheel-thrown, near-complete Cam 255 jar. The presence of this jar is notable, as other vessels in this fabric are handmade at this time. Wheel-thrown forms in early shell-tempered ware occur in higher numbers in contexts of the later 1st century AD for the rest of the Elms Farm assemblage.

Other than the amphoras in KPG5, Italian imports comprise a bead-rimmed Pompeian-red ware platter and a wall-sided, gritless, buff ware mortarium; both are typologically early in form. Italian Pompeian-red ware normally occurred in Britain during the mid- to late 1st century AD, but was present on continental sites, including Haltern and sites in the Aisne Valley (S. Willis pers. comm.), in the Augustan period. The mortarium form was also present at Haltern (Loeschcke 1909, 242, type 59), with much variation in detail; some vessels do not have a bead rim and most lack internal grooving. These two details are present on most British site finds (Hartley 1981, 196), but are absent on the mortarium from Elms Farm.

All other Key Pottery Groups assigned to Ceramic Phase 2 are more typical, although, in general, most Phase 2 assemblages are characterised by a higher level of continental imports than those in any other ceramic phase. In spite of these imports, grog-tempered pottery continues to dominate with amounts averaging 91% of the total by weight. The vessels in this fabric now exhibit much diversity of form and decoration, consistent with trends elsewhere in the Essex-Hertfordshire region, probably as a result of experimentation and greater familiarity with wheel-throwing techniques. Copies of imported vessel types abound. Other coarse wares form a negligible component, and of these, early shell-tempered ware accounts for less than 1% by weight.

The imported pottery types include amphoras, Central Gaulish micaceous ware, Gallo-Belgic ware and North Gaulish fine white wares, all present in some quantity and much variety. Other imports occur in much smaller numbers; these include Pompeian-red and Arretine wares, both never common in Britain. Pompeian-red ware, in particular, is rare in contexts dated earlier than c. AD 40, as its distribution is more commonly associated with the Roman military. Its earlier importation into Britain is implied by Peacock (1977a, 158), and there are early 1st century AD examples at Camulodunum (Hawkes and Hull 1947, 221). Arretine ware is present at Elms Farm in relatively higher numbers than is common at other regional sites with imported wares, but still forms less than 1% by weight in stratified contexts. Findspots of Arretine ware in Essex, outside of Camulodunum, are rare. Two of these are in the vicinity. A platter cut down to form the lid for a cremation urn came from the New Cemetery, Heybridge (Kenrick 1986, 53), and a platter sherd was found at Slough House Farm (Dannell 1998, 143). A further platter sherd, stamped, was found in Ditch 424 at Kelvedon, dated to the Tiberian period (Rodwell 1988, 123).

There is a wider range of Italian wine amphoras in this ceramic phase, augmented by Catalan and salazon vessels. Amphoras account for an average of 6% by weight, although those used for transporting wine comprise the largest share. This is normally the case for amphora assemblages of this period in Britain and also in Picardy (S. Willis pers. comm.). Central Gaulish micaceous ware is present in most assemblages and forms the largest component of the imported wares, after amphoras, at just over 2% by weight. The range of forms now includes tazza-bowls and ledge-rimmed beakers, along with platters and flagons. The red-slipped micaceous fabric variation no longer occurs. Gallo-Belgic imports are more numerous and the range has expanded to include forms in terra nigra. This, and terra rubra, accounts for a combined average of just over 1% of the assemblage by weight. North Gaulish fine white ware has marginally increased its presence to 1%, although the rouletted butt beaker Cam 113 is the main vessel type represented. This form is by far the most commonly found import from this source, and is widely distributed in Essex. North Gaulish white ware flagons are also present, identified mainly by their handles.

Small amounts of Gallo-Belgic wares have been found on sites in the vicinity, although none was recovered from the recent excavations in Heybridge, at either Crescent Road (Wickenden 1986) or Langford Road (Langton and Holbrook 1997). Terra nigra was found on two nearby sites in the Blackwater Estuary area, at Slough House Farm and Howell's Farm (Horsley and Wallace 1998, 143, 146). The base and lower body from a terra rubra butt beaker were found at Woodham Walter (Rodwell 1987a, 26, fig. 19.97), where joining sherds were found in three contexts dated to the mid-1st century AD. Finds of North Gaulish white ware in the area are more numerous, reflecting, perhaps, the continuing importation of this fabric into the later 1st century AD. White ware beakers occur at Colchester in contexts as late as Neronian to early Flavian (Bidwell and Croom 1999, 472). A sherd from a flagon was recovered at the Maldon Landfill site (Wallis 1991, 169), and several vessels, mainly Cam 161 two-handled flagons, were found at both Slough House Farm and Howell's Farm (Horsley and Wallace 1998, 143,146). A single Cam 113 butt beaker was also found at Slough House Farm.

Assemblage composition

Although there is more vessel diversity in Ceramic Phase 2 than previously, jars still dominate at 56% by vessel rim equivalence (EVE). Plain bead-rimmed jars remain much in evidence, but vessels with rippled and cordoned shoulders are more prevalent. Pedestal vessels, such as Cam 204, are also more common. Early shell-tempered jars are few in number, mostly the handmade Cam 254, although there is an example of Cam 255 in Key Pottery Group 5. This is not a common form, and at Camulodunum the Cam 255 is rarer than the ubiquitous shell-tempered Cam 254, where 33 occurrences were recorded against a figure of 238 for Cam 254 (Hawkes and Hull 1947, 268). Storage jars are represented mainly by the Cam 270 and Cam 271 in coarse fabric. These are the most common storage jar forms until gradually superseded by harder-fired storage jar types, such as G44, during Ceramic Phase 4.

Beakers form 17% of the assemblage by EVE, with Gallo-Belgic butt beakers and their grog-tempered copies accounting for the larger share. There is one example of the terra rubra girth beaker, Cam 82. Girth beakers are rare at Elms Farm, and are not commonly-found imports in Britain. The reason for this is largely chronological; Cam 82 girth beakers had a short production span, which ended during the Tiberian period (Rigby 1989, 132). Only eight occurrences of Cam 82 were recorded at Camulodunum, where the type was quickly superseded by Cam 84. Also present at Heybridge, in small numbers, is the ledge-rimmed beaker, Cam 102, in mica-coated Central Gaulish micaceous ware. Later in the phase examples of the Cam 114 beaker appear. These white ware vessels are decorated with herringbone-pattern barbotine, have mica-coating on the rim and are occasionally also red-painted. Their source has not been properly identified, but is likely to be Gallia Belgica (Rigby 1989, 134), and the vessels found in Britain are usually in contexts dated c. AD 10-40.

Open vessel forms, platters and bowls, have increased both in number and variety, accounting for 19% of the total by EVE. The range of bowls, at over 5% by EVE, comprises the carinated Cam 211, tazza-bowl Cam 210, various hemispherical types, and the bobbin-shaped Cam 51 in micaceous terra nigra. Of all vessel classes, platters perhaps display the most diversity in this ceramic phase. They are present in some numbers, in contrast to the previous phase, and this indicates a change in eating habits, or rather, a change in food presentation. This change may also be inferred from the increase in importance of beakers in this ceramic phase. The earliest examples of platters are those in Central Gaulish micaceous ware (mostly Cam 1), and Arretine, although the latter are few in number and largely represented by the shallow Conspectus form 12 (Ettlinger et al. 1990). The later terra rubra and terra nigra vessels are the most numerous of the imported types. Of these, Cam 2 and Cam 5 platters are the most frequent. Later in the phase, the Cam 17 platter in Pompeian-red ware has made an appearance; this is the classic Italian platter in this fabric, probably used for baking. Grog-tempered platters are also numerous, and the number in this fabric is approximately double the number of imported vessels. Many are copies of these imported forms, such as Cam 21 (copying Cam 2), but the quality of manufacture and finish of copied platters is highly variable and consequently there is less standardisation of form.

Other vessel types are few in number; flagons, lids and cups account for 4% of the total by EVE. Flagons are mainly represented by the cream-slipped, micaceous Cam 165, but there is an example of Cam 161 in North Gaulish white ware. There are also two local copies; one is grog-tempered, a copy of Cam 165, the other is in white-slipped grey fabric, probably imitating double-handled North Gaulish flagons Cam 161 or 163. There are three cups; two examples of the grog-tempered, cordoned Cam 212, plus one Cam 56 in terra rubra, and there are just two lids in grog-tempered fabric. Pottery cups and lids appear to be rare at Heybridge in all phases.

Amphoras account for 4% of the total by EVE, although normally their presence is marked by an abundance of body sherds. The range of forms has expanded to include Italian Dressel 2-4 wine vessels, which fully superseded Dressel 1 by the end of the ceramic phase. Amphora types from southern Spain include Pascual 1 wine amphoras, plus salazon vessels used for transporting fish products such as garum. At the end of the phase, the Dressel 20 olive oil amphora has appeared, again marked only by the presence of body sherds. The significance of this wide range is taken up elsewhere.

Ceramic Phase 3: c. AD 20-55

Figure 274
Figure 274: Ceramic Phase 3 pottery supply map (c. AD 20-55)

This ceramic phase is characterised by a steady decline in the types of imported wares, an increase in the standardisation of pottery types and a gradual rise in the number of so-called fully-Romanised forms and fabrics. Grog-tempered pottery continues to dominate, but black-surfaced wares, sandy greywares and storage jar fabrics have made an appearance. Imported fabrics such as samian and Central Gaulish glazed ware have superseded Gallo-Belgic, Arretine and micaceous Central Gaulish wares, although the prominence of imports is lower than in the previous ceramic phase. There is a marginally smaller and more uniform range of forms than previously. Amphoras have also declined in quantity; wine amphoras are few, though Dressel 20 and salazon vessels have increased their share.

The proportion of grog-tempered pottery remains steady, averaging 88% of the total by weight. Early shell-tempered ware has increased slightly to 3%, returning to the level that occurred in Ceramic Phase 1. Known production sites include Gun Hill (Drury and Rodwell 1973) and Mucking (Jones and Rodwell 1973) in Essex and the Black Shore area at Cliffe (Monaghan 1987, 33) in Kent. Production of wheel-thrown early shell-tempered ware at these sites occurred towards the mid-1st century AD, but vessels found at Elms Farm remained handmade until beyond the end of Ceramic Phase 3. Sandier coarse wares, in the form of black-surfaced wares, sandy grey wares and storage jar fabrics, account for 7% of the total by weight; their numbers steadily increased through the phase from less than 1% by weight at the beginning to 19% by the end (see Pottery sequence). The presence of these wares can be shown to be intrusive at the beginning of CP3, at less than 1% by weight. Sandy coarse wares only became apparent following the introduction of kiln-firing in Britain. Grog-tempered and other local Late Iron Age pottery is thought to have been traditionally made using 'clamps', which leave no trace in the archaeological record. Exactly when kiln-fired pottery was first introduced is difficult to determine, but, at Heybridge, this is likely to have occurred in the decade AD 35-45, probably in line with settlements elsewhere in the region, such as Colchester and Verulamium. At Heybridge, sandy coarse wares formed a small part of assemblages at the end of Phase 3, but were fully dominant by the beginning of Phase 5. This point is taken up in the Ceramic Phase 5 discussion.

Storage jar fabric occurred in three groups, only very gradually superseding coarse grog-tempered ware. Black-surfaced wares, at 4% by weight, and sandy grey wares, at 2%, are regularly represented throughout, although only in appreciable amounts towards the end of the phase. Colchester buff ware first appears in negligible quantities also at the end of the phase. Verulamium region white ware appears in later groups, in the form of mortarium flange fragments, averaging just over 2% by weight. The end of Ceramic Phase 3 saw non-local Romano-British products showing a definite presence for the first time.

In contrast to Ceramic Phase 2, imports are less numerous. Central Gaulish ware accounts for less than 1% of the total by weight, and much of this is likely to be residual. Gallo-Belgic and North Gaulish ware quantities also remain low at less than 2% by weight. South Gaulish samian and Central Gaulish glazed ware are the only new imports, occurring in very small amounts at the end of the phase. Neither appeared in Britain in any quantity until the mid-1st century AD, although the lack of Tiberian samian at Heybridge is notable. Only a single example was recorded. Settlements receiving imported wares during the early 1st century AD, such as Camulodunum, normally also acquire Tiberian samian (see Hawkes and Hull 1947, 175). Amphoras form a low proportion of the assemblage, averaging 3% by weight. Italian wine vessels are present, but south Spanish Dressel 20 and salazones have the larger share, with the latter predominating at 2%.

Although the number of imports appears to decline markedly, compared with other sites in the vicinity the level and variety remains remarkable. A recent survey of sites that have imports of Late Iron Age date in the Heybridge area (Wallace 1998, 155, table 8) has indicated that such find-spots are sparse. Most of these occurrences are of amphoras. Gallo-Belgic and Gaulish wares appear as occasional examples amounting to only a handful of vessels.

Assemblage composition

Jars remain the dominant vessel category at 60% by vessel rim equivalence (EVE), and many of the forms are more standardised. Examples of these are the Cam 218 and Cam 220, which are antecedents of Going's G16/17 and G19/20 respectively. The latter continue as late as Ceramic Phase 6, but are recognised as Cam forms in the early part of Phase 3, and as G16-20 in the later part, as the shoulder cordons become less pronounced. Jars with shoulder cordons, plus plain bead-rimmed vessels, are the most common in this ceramic phase, and the club-rimmed Cam 254 is again the sole type identified in early shell-tempered ware.

Platters are much in evidence at 13% by EVE and, similarly, the number of bowls remains constant at 5% by EVE. The most common platter form remains Cam 21, but numbers of Cam 28 increased steadily through the phase. The latter continue into the Roman period, identified in black-surfaced ware as Going's form A2. A samian platter, f18, appeared in an assemblage placed at the end of the phase. Gallo-Belgic platters have declined in numbers, and are restricted to the Cam 2 in terra nigra. The range of bowl forms has expanded to include the incurved Cam 252 and hemispherical forms, as well as the carinated Cam 211. A specialised form is also present; a grog-tempered, spouted strainer-bowl (see .4.2). Towards the end of the phase, the samian bowl f29 occurred; the two examples present, plus the f18, see above, are dated to the Neronian period.

Beakers are somewhat in decline at 13% by EVE, consisting solely of butt beakers - Cam 112 in terra rubra, Cam 113 in North Gaulish white ware, and their respective grog-tempered copies. Cups and lids are still much in the minority, although lids have increased their presence to 3% by EVE. Cups continue to be represented by the grog-tempered Cam 212 and by the Cam 56 in terra rubra. The later Cam 58 is completely absent at Heybridge, although an example of this cup form was found at nearby Slough House Farm (Horsley and Wallace 1998, 143). Flagons have a much-reduced presence, dropping to less than 1% by EVE, and comprising the white ware Cam 161 and grog-tempered copies, one of which is a Hofheim type (J1). By the end of the phase, dishes are apparent, although just two types were identified; the B7 which has an out-turned rim, and the flanged B10, both in grog-tempered fabric. The latter has more in common with the North Kent form 5B2, which had continental prototypes (Monaghan 1987, 138). Mortaria form a very low proportion, accounting for less than 1% by EVE. This is consistent with the evidence from other non-military sites of this period. The form identified is the D1-type mortarium from the Verulamium industry. This was among the first industries to be established following the conquest, and it seems that the products from Verulamium reached Heybridge before those from nearby Colchester. This is in contrast to Chelmsford, where mortaria were almost always supplied by Colchester until well into the 3rd century (Going 1987, 108-12), and where mortaria are rare, in any fabric, until at least AD 160.

Amphoras maintain their presence, but occur in the key pottery groups as body sherds only. Dressel 20 olive oil containers are now slightly in the majority overall, but salazon and Dressel 2-4 vessels are still also present.

Ceramic Phase 4: c. AD 55-80

Figure 275
Figure 275: Ceramic Phase 4 pottery supply map (c. AD 55-80)

Local coarse wares continue to dominate assemblages, accounting for over 90% as measured by EVE. Sandy grey ware forms 12% of the combined Phase 4 assemblage, a substantial enough proportion to confirm increasing levels of production during the Neronian and early Flavian periods. Black-surfaced and grog-tempered wares contribute roughly equal amounts, forming 31% and 36% respectively. This would seem to be in sharp contrast to Chelmsford, where grog-tempered pottery was represented by a mere handful of body sherds. However, none of the groups assigned to Chelmsford Phase 1 (Going 1987, table 8) pre-dates AD 65. Viewed against these, assemblages at Elms Farm dated to the last decade of Phase 4 (for example KPG18, yielding comparatively low proportions of grog-tempered pottery) demonstrate few differences. The relatively minor appearance of grog-tempered pottery in later Phase 4 contexts and its virtual absence at Chelmsford suggest that the fabric ceased to be produced c. AD 65/70, but could have remained in circulation at Heybridge for a further decade. As at Chelmsford (Going 1992a, 96), much of the reduced coarse ware at Heybridge was manufactured locally. Some at least may have derived from the larger potteries at Ardleigh and Colchester.

Besides the major coarse wares, what little pottery there is came from a variety of sources, both regional and continental. The more prolific of these were Colchester and South Gaul, respectively. But even these two barely contribute a total of 6% EVE to the assemblage. Heybridge continued to receive buff ware from Colchester, and early Colchester colour-coated fine ware is also present. New supplies of this fine ware were unlikely much after AD 55; at Colchester the fabric was recovered from deposits dating up to AD 55 (Symonds and Wade 1999, 233). That production had ceased by this time is seemingly confirmed by the absence of the fabric at Chelmsford.

Of the remaining Romano-British regional suppliers, only the Verulamium region industry provided any competition to Colchester, but unsurprisingly, given their relative distances from Heybridge, the market always favoured Colchester. White ware continued to come from Verulamium, now in the form of flagons as well as mortaria. Fine buff ware ring-and-dot beakers were probably reaching Heybridge from Verulamium also, although contexts in which the fabric appears cannot be assigned to this phase with certainty. The fabric was nevertheless present at Chelmsford and Colchester at this time (Going 1987, table 9; Bidwell and Croom 1999, 471-2). Heybridge may have provided a very minor market for the Verulamium industry; its wares were recovered from Crescent Road in small quantities from unstratified contexts only (Wickenden 1986, table 4). Early shell-tempered ware from south Essex or north Kent continued to arrive at the settlement, although the amount declined from previous levels to just 1%. Almost 5% was recovered from Chelmsford (Going 1987, table 9). Fine grey ware from north Kent was also reaching Heybridge, albeit in negligible quantities.

Heybridge received a number of continental fine wares during this phase. South Gaulish samian ware amounts to 3% of the overall assemblage, up from Phase 3. Other continental imports include Lyon ware from Central Gaul, white ware from North Gaul, and mica-dusted ware, terra rubra and terra nigra from Gallia Belgica. Terra rubra died out early in the phase, but terra nigra can still be found in contexts dating to as late as AD 85 (Rigby 1989, 123). Measured by weight, none provided more than 1% of the assemblage; the supply from North Gaul and Gallia Belgica had all but dissipated, and there is little evidence from either Heybridge or Colchester to indicate supply beyond the late Neronian/early Flavian periods. Dressel 20 and salazon amphoras from southern Spain are still present. The former accounts for 1% of the assemblage by weight and is the commoner of the two amphora types.

Assemblage composition

Wide-mouthed eating and food serving/preparation vessels - platters, dishes, bowls and mortaria - account for 14% of the Phase 4 assemblage, measured by EVE. As at Chelmsford, platters are the best represented of these at 6%. Most platters (A2-A4) were produced locally in reduced coarse wares and based on Gallo-Belgic prototypes. There was a much smaller number of samian platters, either f15/17 or f18, and one example of the Gallo-Belgic platter (Cam 13) in terra nigra. Dishes account for just 1%. Platters and dishes may well have been serving the same functions, as there is little to choose between the two in terms of size and shape.

Mortaria are poorly represented. The form was imported from both Colchester and Verulamium. While this would seem to indicate that the inhabitants of Heybridge only cautiously embraced specific culinary habits, robust, wide-mouthed bowls (such as C28 and C33) may well have served mortaria-associated functions (for example mixing and pounding) quite adequately. Bowls were produced locally, though one form (C14) was based on samian f29 or f30.

There was a clear preference towards beakers, rather than cups. The latter are found nearly always in samian ware (f27 and f35). Beakers are confined to two types; the globular H1 beaker and the butt-beaker H7, both ultimately derived from Gallo-Belgic prototypes, one of which, the Cam 112 beaker, was also present in terra rubra. Flagons form 3% of the assemblage and are typically ring-necked.

Jars, present in a range of coarse ware fabrics, provided the largest single category of vessel class by far, accounting for 68% of the assemblage. The reasons for this are twofold. Jars served more functions; most obviously, cooking, transport and storage, and a greater number of jars would be required at any one time to fulfil all requirements. Secondly, these functions perhaps could not be carried out easily using vessels made of cheaper, but perishable materials, such as wood. There may be methodological reasons, too. Labels, such as 'jar' or 'beaker', are assigned to vessels largely on the basis of shape. There is always a risk that these labels, or classes, with their inherent functional assumptions, act like interpretative straitjackets and the possibility that vessels of the same types could serve different functions is ignored. In the case of jars, obvious functions (cooking, transportation and storage) may be over-emphasised.

Jars comprise mainly high-shouldered types (G3, G17-G20, G23), all continuing from Phase 3 and very much in keeping with Chelmsford, despite the difference in the types and proportions of coarse wares. Pit 20009 provided a complete example of a narrow-necked flask or bottle (G40). The form is unusual in this phase and unknown in earlier phases; it is commonly found in 2nd-century burials, but did not appear at Chelmsford or Witham until the 4th century (Going 1987, 27; Turner-Walker and Wallace 1999, table 20). The early shell-tempered Cam 254 jar is replaced by the previously absent shell-tempered Cam 258.

Lids form 4% of the assemblage, representing a slight increase from Phase 3. None could be matched with specific jars; indeed, it is possible that lids were made without reference to the specifications of any individual jar, but were instead general purpose and used simply as covers, rather than well-fitting jar lids.

Ceramic Phase 5: c. AD 80-125

Figure 276
Figure 276: Ceramic Phase 5 pottery supply map(c. AD 80-125)

Ceramic Phase 5 heralds significant change from Phase 4. Locally made wares (essentially black-surfaced ware, fine and sandy grey wares, grog-tempered ware and storage jar fabric) again dominate the assemblage, forming 92% measured by EVE, but the difference rests with the proportions of these fabrics. Most dramatically, grog-tempered pottery has declined from over 35% to just 2%. Conversely, sandy grey ware and black-surfaced ware have increased their proportions; the latter now accounts for 47%, while the former has increased its share to 29%. Although the proportion of fine grey ware remains reasonably steady, that of storage jar fabric has almost doubled from 25% by weight in Phase 4 to 42% in Phase 5. Grog-tempered pottery was no longer being supplied by this time; potters were now exclusively supplying sand-tempered products. As expected, grog-tempered pottery is absent from later 1st and early 2nd century contexts at Crescent Road, Heybridge (Wickenden 1986, table 4).

The pattern is only partially repeated across the region. That Chelmsford was receiving black-surfaced ware and sandy grey ware (with the former predominating), but no grog-tempered ware (Going 1987, table 9) is unsurprising. In sharp contrast, Chigborough Farm, only 2km to the east of Heybridge, produced grog-tempered pottery, albeit fragmented, along with storage jar fabric (Horsley and Wallace 1998, 146-8). At Ivy Chimneys, Witham, grog-tempered pottery continued to form the largest proportion of pottery until the end of the 1st century AD. Until then, the amount of sand-tempered reduced coarse wares present was negligible (Turner-Walker and Wallace 1999, 163).

Clearly, these settlements were not receiving pottery from a single large supplier and were not dependent on a major producer, such as Colchester. The mechanisms controlling the production and supply of coarse wares are not clearly understood and, given the differences in the coarse wares even at neighbouring places, namely Heybridge and Chigborough Farm, may be complex. Some possibilities may be suggested, however. Coarse wares were supplied at a very local level, perhaps via the 'village potter', whose products were solely confined to the settlement. Alternatively, pottery was supplied at a wider, but still restricted, level, with the region divided into a number of small, possibly overlapping, zones based around a number of small-scale concerns. Heybridge perhaps lay at the edge of one zone, with Chigborough Farm and Witham placed within another. This perhaps best fits a model that Hodder provides, based on the organisation of 13th-century rural potters (Hodder 1974, 355). In any case, the eventual widespread adoption of sand-tempering reflects the gradual displacement of potters familiar with the grog-tempering tradition. This accounts for the differing periods at which grog-tempered pottery was superseded by sand-tempered reduced wares. At Heybridge, sand-tempered wares fully replaced grog-tempered wares by c. AD 70/80. Evidently, this process occurred somewhat later at Witham and Chigborough Farm.

Other British coarse ware industries contributed a total of little more than 1%, measured by EVE. Verulamium and Colchester continued to supply pottery, with the latter enjoying a larger market, as is the case at Chelmsford (Going 1987, table 9). Verulamium also supplied a very small amount of fine ware, in the form of ring-and-dot beakers. Early shell-tempered pottery from south Essex or north Kent is also present, again in negligible quantities. The end of the phase marks the beginning of the supply of BB2.

Fine wares account for 13% of the assemblage by EVE. Almost half of this (6%) was locally produced, presumably made by the same workshops that produced coarse wares. A greater range of other fine wares was present here than in Ceramic Phase 4, reflecting the diversification of the pottery industries. North Kent potters increased the supply of grey ware, while introducing oxidised products. London potters were supplying reduced and mica-dusted ware products and Highgate potters supplied fine grey ware. In general terms, Heybridge and Chelmsford share a similar pattern of supply. There is a dearth of anything other than local reduced coarse wares at Ivy Chimneys and Chigborough Farm until the 2nd century.

Other sources of pottery include Hadham, forming less than 1% by weight, which supplied flagons in a white-slipped oxidised fabric. The early Hadham industry may have been responsible for oxidised stamped ware during this time (Rodwell 1978, 238-41), though the evidence for production of this ware at Hadham in the absence of wasters or even substantial sherds is unproven. Lead-glazed ware from south-eastern England was reaching Heybridge at this time.

La Graufesenque in southern Gaul continued to supply all the samian ware to Heybridge, amounting to 2% measured by EVE, to be replaced by the Central Gaulish industry at Les Martres-de-Veyre in the early 2nd century. This follows an established pattern across Britain (Tyers 1996, 112-13). Central Gaulish potters at Lezoux supplied colour-coated products into the late 1st century AD. Black eggshell ware from northern Gaul (Tomber and Dore 1998, 16) is also present in this phase, represented by two beakers. Amphoras, principally Dressel 20 from south Spain, but also including Gaulish amphoras, were imported and provide 1% of the assemblage by weight.

Assemblage composition

Platters and bowls form 18% of the assemblage, as measured by EVE, with platters accounting for 12%. A quarter of these are samian (f15/17, f18 and f18/31). Gallo-Belgic style platters (A2-A4), however, account for the majority of the class. New dish forms, namely the plain-rimmed B1 and the bead-rimmed B2, were introduced at the end of the phase. Along with the well-established B7 dish, the class forms 4% of the assemblage. Again, there is no significant difference in size between platters and dishes.

Along with platters and dishes, bowls increase their share from Ceramic Phase 4. There is a variety of bowls present, from the finely made C12, based on samian f30, to the large, squat, wide-mouthed vessels normally found in storage jar fabrics. One small bowl (Cam 41), which was possibly made in London (Davies et al. 1994, 136), is mica-dusted and undoubtedly designed to imitate metal prototypes. Mortaria continued to reach the settlement from Colchester and Verulamium in negligible quantities. Chelmsford experienced a slight increase in the volume of mortaria at this time (Going 1987, table 10), although the class is poorly represented even there.

The range of beaker and flagon types is little altered from Phase 4. Together, these account for 13% of the assemblage. The globular H1 beaker and the butt-beaker H7 remain the principal beaker forms. Poppy-headed beakers (H5-6) are largely absent, even at the end of Phase 5. The type was present at Chelmsford by this time, however (Going 1987, 28), and was produced at Ardleigh (Going 1999b, 149), in North Kent (Monaghan 1987, 218) and Highgate (Davies et al. 1994, 83). At these sites, the poppy-headed beaker had superseded both the H1 and H7 by the early 2nd century. The market at Heybridge favoured producers who continued to make the H1 and H7 beaker forms. Ardleigh may be ruled out as a supplier for this reason, certainly during the early 2nd century, although the production of the H1 beaker (Cam 108) continued at Colchester (Bidwell and Croom 1999, 472). Flagons are restricted to the ring-necked J3. Chelmsford experienced a decline in these classes around this time or slightly later (Going 1987, table 10). Flagons and beakers do not feature at Ivy Chimneys until the 2nd century (Turner-Walker and Wallace 1999, table 20), while flagons, at least, were always rare at Chigborough Farm and Ardleigh. The regional decline in beakers, or outright scarcity, may suggest that, by choice, people preferred to drink out of cups, small jars and vessels made from perishable materials. As in Phase 4, the samian industries almost exclusively supplied cups, namely f27 and f35.

Ceramic Phase 5 marks the early appearance of the bowl-jar. The class is restricted to the high-shouldered E6.2, produced in black-surfaced ware and present in a late 1st century context. Again, jars dominate the assemblage. The greater visibility of residual pottery in this phase probably accounts for the slight decrease of the class to 65%. Many of the jar forms prevalent in Phase 4 - G3, G8, G16, G17-G20, and G23 - remain so in Phase 5. The cordoned G17 and high-shouldered G23 are among the commonest jar types. It is interesting to note that although the G17 (Cam 218) and G23 (Cam 266) were manufactured at Ardleigh up to the mid-2nd century (Going 1999b, 149), other common Elms Farm jar types - the G19-20 among them - are missing from the Ardleigh repertoire. While Ardleigh is unlikely to have supplied much pottery to Heybridge, the presence of the G17 and G23 jar types at both sites demonstrates just how ubiquitous these forms were across the region.

At 2%, lids now take a smaller share of the assemblage than in Ceramic Phase 4. Clearly, purpose-made ceramic lids were exceptional, the norm perhaps being stone or ceramic caps, or covers and bungs made out of organic materials.

Ceramic Phase 6: c. AD 125-170

Figure 277
Figure 277: Ceramic Phase 6 pottery supply map (c. AD 125-170)

The proportions of traded wares expanded at the expense of locally produced wares. Collectively, black-surfaced ware, fine and sandy grey wares and storage jar fabric take a reduced 70% share of the market measured by EVE. Fine grey ware has fallen from 6% in Ceramic Phase 5 to 3% in Phase 6; black-surfaced ware has dropped to 34%. The proportions of sandy grey ware and storage jar fabric are virtually unchanged. At the same time, after its early appearance at the end of Ceramic Phase 5, BB2 now takes a larger (6%) share of the market. North Kent grey ware also increased its share from 1% to 5%, reflecting the growth in production at this time (Monaghan 1987, 219). It is tempting to link the rise of these fabrics to the decline of black-surfaced and fine grey wares; the first group replacing aspects of the second. Why this occurred is unclear, though aggressive marketing, such as flooding the market with cheaper and more desirable products, can be cited as a reasonable cause.

The lack of contemporaneous, fully quantified, pottery at other sites in the region makes comparison difficult. Trends that can be drawn out of published data are mixed. At Chelmsford, black-surfaced wares (Fabrics 34 and 45), fine grey ware (Fabric 39) and North Kent fine ware decreased at this time, while sandy grey ware (Fabric 47) and BB2 increased (Going 1987, table 9). Black-surfaced ware also declined at Crescent Road while fine grey ware remained steady and sandy grey ware and BB2 increased their shares. North Kent grey ware was absent from this assemblage (Wickenden 1986, table 4). The decline of black-surfaced ware was seemingly regional, but the cause of this may be in part local, due to consumer demand as much as successful marketing; coarse grey pottery generally increased in tandem with BB2, suggesting that local producers were not in decline, but merely refocusing their markets.

With the parallel increases of North Kent grey ware and BB2 at Elms Farm, it is worth considering the possibility that North Kent supplied both fabrics, at least up until Ceramic Phase 7. Crucially, Going noted that BB2 vessels recovered from a Hadrianic/early Antonine pit at Chelmsford should perhaps be given a Kentish origin on stylistic grounds (1992a, 102), and observed that BB2 was far more common in Kent than it is in central and southern Essex at this time (1987, 112). In the absence of petrological analysis of all sherds, the weight of probability tends to support Kent for the initial supply of BB2 (although a Colchester source is assumed for the purpose of the maps).

Even if it did not supply BB2, Colchester provided a range of other wares to Heybridge. As a whole, the industry took a market share of 6%, or 12% including BB2. As at Chelmsford (Going 1987, 110), the settlement continued to receive buff ware, almost exclusively mortaria. Colchester colour-coated and samian wares reached Heybridge for the first time, although the former was unlikely to have appeared in Heybridge much before c. AD 140, some fifteen years after production began (Symonds and Wade 1999, 264-6). The ware is virtually absent from large groups dated earlier than c. AD 140, but is strongly represented in assemblages dated later, such as KPGs 22 and 23. This would seem to be in keeping with Chelmsford, which received the ware after c. AD 160 (Going 1987, table 9).

Pottery continued to be imported from Verulamium in small quantities, as in Phase 5, taking a 1% share of the market. Pottery from Hadham is again present; that fine white-slipped Hadham products occur is to be expected (Going 1999a, 297). Less so is a red colour-coated flagon (J8), from the same source, recovered from pit 4152 in the central area. The presence of this ware is unusual at this time, as it was not exported in any quantity into the region until the later 3rd century (Going 1999a, 297). Its early appearance at Heybridge suggests limited exportation, and was, perhaps, even a one-off. London fine ware and South-east English glazed ware are present still in this phase, as is early shell-tempered ware, although given that production of this fabric had ceased by c. AD 110 (Monaghan 1987, 223), its presence, which is less than 1% of the assemblage, may be residual. BB1 from Dorset, represented by a single body sherd, makes its initial appearance in pit 9029 (KPG24), which was infilled after AD 160.

Samian supply more than doubled in this ceramic phase, with Central Gaulish samian, mainly from Lezoux, showing the largest increase to form 7%. As mentioned above, Colchester samian appears for the first time during this phase, as does East Gaulish samian, initially from La Madeleine, then from Rheinzabern. Each accounts for 1% of the assemblage overall. A decline in samian imports during the early part of the phase, or, indeed, during the early 2nd century (Rodwell 1988, 98), is a phenomenon observed at Chelmsford (Going 1987, 108), but not clearly visible at Heybridge. Other continental pottery is poorly represented compared with samian. Lower Rhineland colour-coated ware and Central and East Gaulish Rhenish wares were imported during this phase, but their presence amounts to a mere handful of sherds. Gaul and southern Spain continue to supply amphoras, the latter increasing its share to form 1% of the assemblage by weight.

Assemblage composition

A number of significant trends are evident in this phase. Platters decrease to form 2% of the assemblage measured by EVE. Bowls fall from 7% in Ceramic Phase 5 to 3% in Phase 6. Meanwhile, the proportion of dishes increases to 18%. Platters are now largely confined to samian f18/31, itself transitional between a platter and a dish, and the A4 bead-rimmed type from North Kent. There were three principal dish forms at this time: - the plain-rimmed B1, the shallow bead-rimmed B2 and the deeper, bowl-like, bead-rimmed B4, all available mainly in BB2 and black-surfaced ware. The average diameter of B2/B4 dishes in Phase 6 and that of platters and bowls in Phase 5 is almost identical. Either eating habits had radically altered during the mid-2nd century or, as is more likely given vessel sizes, dishes, mainly in BB2, were replacing and serving the same functions as black-surfaced ware platters and bowls. This trend is reflected at Chelmsford, where dishes very soon became the dominant open form after AD 125 (Going 1987, 110; 1992a, 103). The eventual disappearance at Chelmsford of all platters except samian was complete by c. AD 160, and by around AD 140 at Heybridge. Samian bowls f37 and f38 were imported from Central and East Gaul respectively, while the dish f31, also from Central Gaul, was present after AD 140.

The proportion of mortaria remains largely unaltered. Vessels were imported from Colchester, though a smaller number came from local or East Anglian sources and Verulamium. A D1 mortarium in a white-slipped sandy grey ware is present, though unusual in this fabric. Bowl-jars are also rare in this phase; the ledge-rimmed E2 type replaced the E6 of Ceramic Phase 5.

As expected, jars remain the most prolific vessel class, although their share is slightly reduced at 57%. Use of the G44 storage jar continued. There are fewer high-shouldered jars (G16-20) than in Phases 4 and 5, as represented by EVE. These were replaced by a narrower suite of vessels, namely the G22 with stabbed shoulder, the increasingly prevalent G23 jar, and the carinated G29 jar. Other jar forms were present, including the oval-bodied G24, which became more popular later in the Roman period. This pattern is repeated at Chelmsford (Going 1987, 110).

Beakers retain their share of the assemblage in Ceramic Phase 6. Poppy-headed H6 beakers, present in fine grey and North Kent wares, replaced H1 and H7 beakers as the commonest drinking form, a little later than at Chelmsford. Roughcast and colour-coated H20 beakers were imported from Colchester and the Lower Rhineland. Large, folded, grey ware beakers (H34/H35), typical of the late 2nd century onwards, appear towards the end of the phase. Cups increase their share to 4%. Typically, cups remain restricted to samian wares; f27, the commonest cup of the previous two phases, is superseded in Phase 6 by f33. Flagons, at 2%, continue to form a small proportion of the assemblage. The ring-necked J3 flagon is largely absent from this phase. Lids remain steady at 1%, and the K3 and K6 types present in Phase 5 are also present in Phase 6.

Ceramic Phase 7: c. AD 170-210

Figure 278
Figure 278: Ceramic Phase 7 pottery supply map (c. AD 170-210)

Locally made pottery regains its very dominant position within the assemblage, accounting for 84% measured by EVE. For the first time, sandy grey ware takes a larger market share than black-surfaced ware. The former provides 39% of the principal local wares (black-surfaced ware, fine and sandy grey ware, and storage jar fabrics), while the latter accounts for 26%. Fine grey ware takes a reduced 4% share. The decrease of the proportion of storage jar fabric is considerable - from almost 50% by weight in Ceramic Phase 6 to 13% in Phase 7. Storage requirements were presumably being fulfilled using alternative means, such as barrels.

The Hadham products reaching Heybridge increased and include oxidised ware, and grey and black-surfaced wares. These reduced fabrics were presumably made in competition to BB2. However, the industry was still a minor supplier at this time, taking a 1% share of the market by weight; only the black-surfaced fabric is sufficiently abundant to be represented by EVE (1%). This ceramic phase is devoid of Hadham white-slipped fine ware, although unsourced white-slipped fabrics are present.

The Colchester industry consolidated its market position, increasing its supply of buff and colour-coated wares to 7% by EVE. North Kent grey ware is virtually absent. Forming 3% of the assemblage, supply of BB2 from Essex or Kent continued, though perhaps at a slightly reduced volume from Ceramic Phase 6. London supplied fine reduced ware and possibly mica-dusted ware; both are represented by body sherds. Mica-dusted ware was present in assemblages at Colchester until c. AD 225 (Symonds and Wade 1999, 245), although normally considered to be residual by this time in central Essex (Going 1987, 5). Heybridge received negligible quantities of Nene Valley colour-coated ware at the end of the phase. This industry did not yet threaten Colchester as the main British supplier of fine ware. Mancetter-Hartshill mortaria and BB1 vessels are also present during this phase.

Central Gaulish samian forms the largest single group of continental fine wares in this phase. Rather than an overall decrease in the supply of new samian, reflecting the decline in output after AD 170 (Marsh 1981, 185), the Elms Farm figures are mixed, with Central Gaulish samian decreasing in terms of EVE, but increasing by weight. This must reflect the increased residual element to the samian assemblage - a reasonable expectation if samian vessels were valued items and thus had longer lives than other pottery. New vessels did not perhaps replace older ones, but rather added to them, enlarging the pool of vessels current and in use at a given time. East Gaulish samian, Central and East Gaulish Rhenish wares were also imported, but with no apparent change in quantity. By this time, Lower Rhineland colour-coated ware had all but ceased to be imported, its presence reduced to a few sherds. Dressel 20 amphoras were imported from southern Spain in greater numbers, forming 3% of the assemblage by weight.

Assemblage composition

Generally, the range of forms present in this ceramic phase is more restricted than in Phase 6. Notably, platters, as a class, have disappeared altogether, fitting a pattern identified at, for example, Great Dunmow (Going and Ford 1988, fig. 52) and Chelmsford (Going 1987, table 10). Bowls, comprising samian f37 and the mica-dusted Cam 41, are reduced in number, forming just 1% of the assemblage measured by EVE. Dishes, then, were virtually the sole type of ceramic eating vessel, produced in a range of reduced fabrics, from locally made black-surfaced and sandy grey wares to Hadham ware and BB2. The bead-rimmed B2 and B4 types are by far the most common of dishes, followed by samian f31 and then the plain-rimmed B1 and B3. Mortaria take a 5% share of the assemblage. This increase and the fact that local potters were inspired to produce mortaria at this time may well testify to an increasing popularity of specific cooking styles (that is, recipes requiring the mixing, grinding or pounding of ingredients). Cups take a reduced 1% share, never regaining former levels. The class is, again, restricted to samian; f33 had by this time totally replaced f27.

Jars again make up the largest class, forming 56% of the assemblage. Jar types typical in Phase 6 (G22-4) remain so in Phase 7. There is little evidence that the carinated G29 jar survived into the late 2nd century. By far the most common jar form, however, was the ledge-rimmed jar (G5). Local production of this form, along with the oval-bodied G24 and G25 jars, is attested. The G5 jar was undoubtedly commoner in kiln assemblages than the G24 or G25 types. There is no suggestion in these assemblages that ceramic lids were made to fit the G5 jar. Indeed, typological development of lids was stagnant; no new lid form had been introduced since Ceramic Phase 4, and just one type (K6) is represented. Beakers take 11% of the assemblage. The appearance of the poppy-headed beaker at Heybridge was relatively fleeting and had disappeared by the late 2nd century to be partly replaced by an increasing range of Colchester products, including the roughcast H20 beaker and barbotine-decorated H24 beaker. Folded grey ware beakers, introduced during Phase 6, became more common in Phase 7. The range of flagons is somewhat meagre, restricted to bead-rimmed or tall, narrow-necked types.

Ceramic Phase 8: c. AD 210-260

Figure 279
Figure 279: Ceramic Phase 8 pottery supply map (c. AD 210-260)

In contrast to earlier phases, the quantities of pottery, along with form diversity, are much reduced. This reflects the general economic trend during the 3rd century, which continues into the next phase. The similarities between the pattern of pottery supply to Heybridge (Wickenden 1986 and this report) and Chelmsford (Going 1987, 113) are more marked than any differences. Thus, in both settlements, locally produced pottery is, again, predominant, accounting for 85% of the overall Ceramic Phase 8 assemblage at Elms Farm measured by EVE. The proportions of black-surfaced and sandy grey ware remain relatively stable. Fine grey wares substantially increased their share to 17%, a surge that may have been in response to successful marketing of Hadham fine fabrics or a general increase in the use of dishes. After its decline in Phase 7, storage jar fabric in this phase returned to something resembling former levels.

The Colchester industry continued to supply colour-coated vessels. However, the supply of Colchester buff ware declined, despite the industry flourishing at this time (Bidwell 1999, 496). Heybridge continued to receive BB2. Nene Valley colour-coated ware is barely present, but in any case the industry may have preferred to concentrate on the export of white ware mortaria, which appeared at Heybridge at the end of this phase. BB1 sustained its tenuous grip on the market, its presence largely limited to body sherds.

The supply of samian to Heybridge reflects established regional patterns. East Gaul became the sole samian provider, though its products amount to less than 1% of the assemblage; most samian in this phase is residual Central Gaulish. Central and East Gaulish factories continued to supply Rhenish ware, however, but not beyond the early 3rd century. Amphoras from Gaul and southern Spain continued to be imported during this phase, but in reduced quantities.

Assemblage composition

The proportion of dishes has increased to 27%. The majority of dishes were made locally, with a smaller amount produced in BB2 fabric or imported from Hadham. The usual dish forms are represented, including the plain-rimmed B1 and B3 (the latter are more frequent), and the ubiquitous bead-rimmed B2 and B4 types. Two new forms were introduced by the mid-3rd century; the incipient bead-and-flanged B5 and the stubby bead-and-flanged B6.1. The East Gaulish samian f31 dish is also present.

The rise of dishes is offset against a decline in jars, with the range of different jar forms becoming restricted. The ledge-rimmed G5 remains one of the more common forms, along with the oval-bodied G24-5. The everted-rimmed G9 is also more common than in Ceramic Phase 7. Just one new jar form is introduced; the small G42 storage jar. By this time, non-samian bowls are entirely absent from the assemblage, although bowl functions were perhaps served by the E2 and E5 bowl-jars, which account for 2%. Mortaria, taking a reduced 2% of the assemblage, are restricted to the hammerhead-rimmed D11 type, and the reeded Nene Valley varieties. The range of beaker types present is broadly similar to that in Ceramic Phase 7. Colchester colour-coated roughcast beakers (H20) and grey ware folded beakers (H34/35) are the most common. Flagons are very poorly represented and few diagnostic forms could be identified. However, one example present is in Hadham white-slipped ware. A single lid type, the black-surfaced K2, occurs in this phase.

The range of forms present is more restricted than in Ceramic Phase 7. This narrowing of the range available for use seems to be a sustained phenomenon, but trends identified here are broadly reflected at other sites, too. The increasing dominance of dishes has already been discussed in respect of Group 464/820 at Great Dunmow (Going and Ford 1988, fig. 52). At Chelmsford, dishes increased in number, while jars declined during the 3rd century. Bowls and mortaria also decreased (Going 1987, table 10). B1, B2/B4 and B5 dish forms are present in the mid-3rd century kiln assemblages recovered from Orsett (Cheer 1998, 99). In none of these cases are dishes commoner than jars, although the growing importance of dishes is apparent. This is also the case at Heybridge.

Ceramic Phase 9: c. AD 260-310

Figure 280
Figure 280: Ceramic Phase 9 pottery supply map (c. AD 260-310)

Heybridge was receiving pottery from a wider range of regional sources in Ceramic Phase 9 than in Phase 8. As ever, locally produced pottery was predominant, taking an 81% share of the assemblage measured by EVE. The proportions of the major local wares (black-surfaced ware, fine and sandy grey ware) are largely unchanged from Phase 8; the proportion of storage jar fabric has fallen once again, and now forms 21%. Ceramic Phase 9 marks the initial appearance of Rettendon ware, the distinctive late Roman flint-tempered fabric. This accounts for 2% of the assemblage by EVE.

The supply of Colchester products declined during this phase and the colour-coated products are largely residual. The supply of BB2 from Essex or Kent also decreased. A reasonable cause for this decline was the increased importation of BB1, which forms 4% of the assemblage by EVE. A range of Hadham wares is represented. Reduced wares, particularly the grey fabric (HAR), remained the most common of these. The Nene Valley industry increased its supply of colour-coated ware, filling the gap left by the Colchester industry, and reflecting the regional expansion of the industry at this time (Going 1987, 114; Bidwell 1999, 496). Importation of Nene Valley white ware mortaria continued, although the Oxfordshire industry, which also exported white ware mortaria, provided competition. Continental imports are present in this phase, but all are residual.

Assemblage composition

Dishes account for 23% of the assemblage. As at Chelmsford (Going 1987, 114), straight-sided B1 and B3 dishes are now the most prolific types, vastly outnumbering bead-rimmed B2 and B4 types (not represented at Chelmsford at this time). Bead-rimmed dishes, however, remain more common than flanged B5 and B6 dishes, the latter both introduced towards the end of the 3rd century, and it is likely that bead-rimmed dishes were produced locally well into the second half of the 3rd century. Notably, the Hadham and BB1 factories were almost exclusively supplying straight-sided and flanged dishes at this time, suggesting that the continued production of bead-rimmed dishes was a phenomenon restricted to producers local to Heybridge. East Gaulish samian dishes (f31) are present, but residual in this phase.

Local or East Anglian mortaria production appears to have increased, perhaps in response to the importation of Nene Valley and Oxfordshire mortaria. Hammerhead-rimmed D11 and wall-sided D13 mortaria were the preferred shapes. Ledge-rimmed E2 bowl-jars and the bead-rimmed E5 types are again represented.

Jars account for 52% of the assemblage. The ledge-rimmed G5 jar is ubiquitous. Interestingly, the high-shouldered G23 jar, last seen in Ceramic Phase 7, makes a dramatic reappearance, becoming the second commonest jar form, while the oval-bodied G24 lies third. This should not overly concern us, as there is little typological difference between these two forms; both types were present at Chelmsford at this time (Going 1987, 114). The everted-rimmed G9 jar gained in popularity, appearing in BB1 as well as in local wares. New jar forms include the frill-rimmed G26 and large narrow-necked G38, both made locally. Beakers, forming 14% of the assemblage, were the sole class of drinking vessel, and restricted to locally made folded beakers (H33-H35) and barbotine-decorated beakers from the Nene Valley (H28, H32, and H41). Contemporaneous Colchester colour-coated beakers are entirely absent. Flagons are rare, while lids are non-existent.

Ceramic Phase 10: c. AD 310-360

Figure 281
Figure 281: Ceramic Phase 10 pottery supply map (c. AD 310-360)

Locally produced pottery is, as always, predominant, forming 88% of the assemblage measured by EVE. Both sandy grey ware and black-surfaced ware have increased their proportions by weight. The proportion of storage jar fabric has further declined to 7% by weight. The level of production of the fabric was minimal, if it continued at all. While the range of sources that supplied Heybridge increased from Ceramic Phase 9, the regional industries made very little impact upon the market. No industry increased its market share; at best, the more successful manufacturers merely retained their markets. As at Colchester (Going 1999a, 297), Hadham increased its supply of oxidised ware, though the overall share of Hadham fabrics remained the same as Phase 9 at 3%. Colchester buff mortaria (D1 and D11) and Essex/Kent BB2 bead-rimmed dishes (B2/B4) are present in Ceramic Phase 10, but both wares are likely to be residual by this time. Supply from the Nene Valley industry continued, colour-coated products being the chief imports, but painted ware and white ware mortaria were also reaching Heybridge.

The Rettendon ware producers failed to capitalise on their initial market successes. Rettendon ware took 2% of the market share by EVE at Heybridge, compared to 17% at Chelmsford (Going 1987, 115). This difference reveals an interesting picture of local trade patterns. Between 40% and 50% of the pottery at Chelmsford during the first half of the 4th century was from production sites within the town itself (Moulsham Street) and at Inworth (Going 1987, 89). These and other known Rettendon ware production sites - at Sandon and Rettendon itself - were located away from principal roads that lead to Heybridge. In terms of the distribution of this fabric, Heybridge, like Great Dunmow (Going and Ford 1988, 70), fell outside its catchment area. It is reasonable to suggest, then, that the low proportion taken by Rettendon ware should be expected given these factors; the distribution of Rettendon ware was restricted and essentially local. A mid-4th century group recovered from Chigborough Farm (Horsley and Wallace 1998, table 7) supports this suggestion. Rettendon ware makes up just 1.5% of this assemblage. As local production of a flint-tempered fabric during the early 4th century is attested, it is possible that some of the pottery labelled as Rettendon ware at Elms Farm is in fact of local origin.

Ceramic Phase 10 marks the probable initial appearance in the mid-4th century of late shell-tempered ware from Harrold in Bedfordshire or Lakenheath in Suffolk, though accounting for less than 1% by EVE, and Oxfordshire red colour-coated ware, though no rims survive. Heybridge was receiving no new pottery from continental sources. The appearance of East Gaulish and Central Gaulish samian in this phase is entirely residual.

Assemblage composition

Dishes account for 32%, measured by EVE. The straight-sided B1 dish is the most prolific type of its class, followed by the bead-and-flanged B6. This remains the case until the end of Ceramic Phase 11. These dish types were available in a range of fabrics, including local reduced wares, BB1 and Hadham wares. The proportion of grooved straight-sided B3 dishes declined dramatically, suggesting that production of the form had ceased. Notably, the type is absent from production waste assemblages, which date to the late 3rd or first half of the 4th century. Of the dish population, bead-rimmed B2/B4 dishes continue to form a substantial proportion. Given that these types are almost certainly residual, as they are also absent from production waste assemblages, this reflects their overwhelming presence in previous ceramic phases, and the length of time that some examples must have been in use. The flanged B5 dish is absent, having fully developed into the bead-and-flanged B6.

Bowls reappear in this phase after their temporary absence in Ceramic Phase 9. Forms include the hemispherical flanged C8 bowl, which originated from three sources, Hadham, Oxford and the Nene Valley. The presence of a 'Castor box' lid (K7), also from the Nene Valley, provides indirect evidence of the C18 bowl, which, together, form a set. Mortaria form a smaller proportion than Ceramic Phase 9, and are present in buff ware fabrics and Oxfordshire white ware (D5). Nene Valley white ware mortaria are largely absent from this phase. A broader range of bowl-jars is present. The common E2 and E5 vessels are represented (the latter produced at Heybridge), as is the grey ware E3 bowl-jar and the E1 and E6 bowl-jars from the Hadham region.

Most jar forms present in Ceramic Phase 9 are also present in Phase 10. The oval-bodied G24 is the commonest form during this phase. The ledge-rimmed G5 jar accounted for 10% of jars, but, like bead-rimmed dishes, its presence may well be residual. Significantly, the form was absent from the production waste assemblages. New jar types include the narrow-necked G37. The proportion of beakers is down from Phase 9. The class is restricted to Nene Valley products, namely barbotine-decorated H32 and folded H39 types. The absence of locally produced beakers, and the decline of the class overall, suggests that, by the 4th century, the inhabitants of Heybridge were drinking out of non-local vessels and, to a larger extent, vessels made from materials other than clay. Flagons are not present in any of the Key Pottery Groups.

Ceramic Phase 11: c. AD 360/70 - 400+

Figure 282
Figure 282: Ceramic Phase 11 pottery supply map (c. AD 360-400+)

The final decades of the 4th century began to see a fall off in the amount of locally produced pottery and a further increase in the amounts of Romano-British regional traded wares. Locally made wares decline to 65%, while the traded wares rise to 35%. Within the locally made category, fine and sandy grey wares stand at 41% in contrast to the 62% found at Crescent Road (Wickenden 1986, table 11). Meanwhile black-surfaced wares stand at 19%, while Rettendon wares increase slightly to 3%. This is nothing like the 12% for Rettendon ware at Crescent Road.

Romano-British regional traded wares originated from a variety of sources and are strongly represented in all well-dated groups of this period. Fine wares, too, increased in volume and were derived from three sources, the Nene Valley, the Oxfordshire potteries and Hadham. Nene Valley colour-coated ware is the most important fine ware at 11% by EVE, followed by Oxfordshire red colour-coated ware at 3%. Hadham oxidised ware is the least important at just 1%. The range of Oxfordshire products is wider than in any previous period and now includes colour-coated, parchment, white and white-slipped wares. This seems to have been due, not to increased production, but to an expanded market area for the industry at this time. Oxfordshire products account for 12% by weight of the pottery in KPGs 37 and 38.

The only coarse regional traded ware in any quantity is late shell-tempered ware, which accounts for 7% by EVE. Other traded wares at Heybridge comprise Alice Holt grey ware and Portchester D ware, although the quantities are negligible and account for less than 1% by weight in any assemblage. At Crescent Road, Alice Holt grey wares also accounted for less than 1% of the assemblage by weight. The quantities of BB1 are minimal, and, where present, this fabric is clearly residual. There is no evidence for the export into Essex of the latest cooking pot types (cf. Holbrook and Bidwell 1991, fig. 29, type 21). This is in keeping with our overall understanding of BB1, which was not widely distributed subsequent to the loss of the northern markets in the third quarter of the 4th century (Gillam 1976, 59). BB1 is not present at Great Holts Farm, Boreham (Martin 2003).

The amount of imported pottery reaching Heybridge is negligible. Two sources, Aquitania (céramique à l'éponge) and the Rhineland (Mayen ware), are represented, neither of which appear in the selected key pottery groups. Although these fabrics seem to occur in slightly greater quantities compared to Chelmsford, the amounts involved are perhaps not commensurate with trade in ceramics per se. It is likely that the small numbers of vessels represented reached Heybridge as either chattels or as make-weights alongside other commodities. Moreover, neither of these fabrics is recorded at Crescent Road (Wickenden 1986, table 11). As in Ceramic Phase 10, supplies of amphora seem to have all but ceased and, where present, mainly comprise residual Dressel 20 or earlier types. Although not in a quantified group, a single Kapitän 2 'Hollow foot' amphora was recovered from Well 5806, in a fill (5768) dated to the second half of the 4th century.

During the early 5th century, as may be evidenced by KPG39, there was a major decline in the availability of locally made pottery and an upsurge in the amount of Romano-British traded wares. Imports are not evident. The amount of grey wares has fallen to 18% by EVE, while black-surfaced wares declined to just 2%. Rettendon wares have entirely vanished. It is notable that late shell-tempered ware now stands at 34%. This figure is greater than the 15% at Chelmsford (Going 1987, 116) but compares well with the 21% in the 'late shrine' group at Great Dunmow (Going and Ford 1988, table 2). Oxfordshire products account for 31%, the bulk of which comprises red colour-coated ware at 29% by EVE. This figure is far in excess of the 8% seen in the late shrine group at Great Dunmow (Going and Ford 1988, table 2) and the 3% at Chelmsford (Going 1987, table 9). The amount of Nene Valley colour-coated ware remains stable at 9%, while Hadham wares have risen to 6%. Presumably other fabrics like Alice Holt grey ware and Portchester D ware also continued to reach the settlement in small quantities, but it is impossible to say whether supply continued from across the Channel.

Assemblage composition

In the late 4th century, open and closed forms are present in roughly equal proportions. Open classes account for 48% of the assemblage with a distinct bias in favour of dishes (32%). Next in importance are bowls (8%) followed by mortaria (6%). Lids are exclusively Nene Valley colour-coated 'Castor-box' types (K7) and account for just 1% of the assemblage. Of the dish types, the plain-rimmed B1 predominates with eighteen examples, followed by the bead-and-flanged variety (B6) with sixteen. These forms occur in a variety of local grey and black-surfaced fabrics. The B1.2-type dish in Nene Valley colour-coated ware is the least common variant. A number of fragmentary bead-rimmed B2 and B4 types and the incipient-beaded flange-rimmed B5.1 form were also identified in grey and black-surfaced fabrics, but these are clearly residual. The bulk of the mortaria are derived from the Oxfordshire manufactory with the white ware form, Young's M22 (1977), occurring in reeded and unreeded versions. The other mortarium, a D14-type from the Nene Valley, is residual. Most of the bowl forms occur in Oxfordshire red colour-coated ware and include the form C8, imitating samian f38, and a number of vessels that can be referenced in Young (1977). The C8 also occurs in Hadham oxidised ware alongside several bowl-jars.

Closed forms account for 52% of the assemblage with jars comprising the largest class at 36%, a figure broadly comparable with the 38% seen in Ditch 21 at Great Sampford (Martin 1998, table 3). The forms recorded at Elms Farm comprise the late shell-tempered ware necked jar (G27) and the common late necked jar (G24), which is present in sandy grey ware and Rettendon ware. The G27-type jar is the main jar form in the group at 7%. Of the storage jar types only the high-shouldered G44 is evident and this is likely to be residual. There is thus little sign of the recovery in the incidence of storage jars as seen at Chelmsford (Going 1987, table 10). Indeed, the figures are very much in keeping with those from Crescent Road (Wickenden 1986, table 11). Beakers comprise 9% of the assemblage and consist of folded types in both fine and coarse fabrics. Flagons account for 7% of the assemblage, a figure much lower than the 12% seen at Chelmsford (Going 1987, table 10). The only flagon type present is in Oxfordshire white ware and corresponds to Young's W25 (1977), which is a rare form in Essex.

By the early 5th century, open forms account for 45% while closed forms stand at 55% of the assemblage with jars accounting for 44%. However, leaving aside the dominant late shell-tempered G27 jar, most of the rims are too fragmentary to identify type with any certainty. The only other type certainly identified is a bifid-rimmed vessel similar to the G28. A possible flagon is present in Oxfordshire white-slipped ware, and a sandy grey ware beaker rim of indeterminate form is the only other closed form represented in the group. Next in importance to the jar is the dish category (24%). The main type is the plain-rimmed B1 in Hadham black-surfaced ware (B1.3) and Nene Valley colour-coated ware (B1.2). Bead-and-flanged dishes were only recorded in fine grey ware. Bowls accounted for 18% of all vessels but are exclusively in the fine Nene Valley colour-coated, Oxfordshire red colour-coated and Oxfordshire parchment wares. In Nene Valley colour-coated ware, the range was restricted to the 'Castor-box' (C18) and the E4 bowl-jar. By comparison, the range of Oxfordshire red colour-coated forms recorded is less restricted. These include the ubiquitous C8 and C25 bowls and E4 bowl-jar, as well as Young's forms C69 and C79 (1977). The only identifiable form in Oxfordshire parchment ware was a P17-type bowl (1977). Mortaria account for just 2% of the assemblage and were supplied solely by the Oxfordshire potteries. Although there was considerable fabric diversity, the range of identifiable forms was confined to the wall-sided D12 in red colour-coated ware and the flanged white-slipped ware WC7 (1977, 122).

Post-Roman

There is sufficient evidence from the Crescent Road and Elms Farm excavations to indicate settlement activity in the early Saxon period. Continuity of occupation has been argued (Drury and Wickenden 1982, 31-4) and, in order to test this hypothesis against the Elms Farm data, a search was made to identify substantial deposits of latest Roman pottery that also contained large quantities of Saxon sherds. Only one group was identified, the fills of pit 14529 (KPG40). This group was therefore studied to see whether any insights could be provided into the relationship between Romano-British and Anglo-Saxon material cultures.

Although probably residual, the Roman pottery is here discussed entirely from the standpoint of latest Roman supply/production. The aim was to identify any unusual fabrics and forms that were not represented in the three key pottery groups. Several observations are possible. Firstly, there are no unusual fabrics present. This confirms the picture provided by KPG39 in that there is no evidence for contemporary local Romano-British pottery production. Generally, the range of fabrics present is typical of Ceramic Phase 11 contexts and there is nothing to suggest that any of the pottery had been deliberately selected as at West Stow, Suffolk, where there seems to have been a particular liking for red pots (Plouviez 1985, 84).

Perhaps, the only notable feature of the group is the amount of Alice Holt grey ware, at 3% by EVE, and the presence of B6 bead-and-flanged dishes (Lyne and Jefferies 1979, Type 6C) in this fabric. These are exceptionally rare in Essex and are only otherwise known on sites close to the Thames as at Aveley (Martin 2002) and London as at Leyton (Greenwood 1979, fig. 13.76). There is very little dating evidence for these vessels, but what there is suggests that they could be among the latest Romano-British ceramics in Essex.

Even though rare late forms are present in the group it is highly likely that all of the Roman pottery is residual. That this is the case is suggested by the high incidence of grey wares. These account for 30% by EVE of the assemblage, while late fabrics are poorly represented. For instance, late shell-tempered ware stands at just 3%, a figure more typical of late Ceramic Phase 10 horizons. The research archive contains a study that further investigates the types of pottery deposited, and the condition of the sherds, in these and other selected groups. The results are interesting; one conclusion appears to confirm that low amounts of late Roman fabrics are, indeed, more likely to indicate a 5th-century assemblage.


 PREVIOUS   MORE DETAILS   NEXT   CONTENTS   ISSUE   HOME 

only search Internet Archaeology

Internet Archaeology is an open access journal. 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.

University of York legal statements