SECTION 3.1:CLAYDON PIKE IRON AGE POTTERY by Grace Jones

3.1.1 INTRODUCTION
Condition of the Assemblage
Aims and Objectives of Analysis
Methodology
The Ceramic Archive

3.1.2 FABRIC ANALYSIS
Fabric Description
Fabric Quantifiacation
Source of Raw Materials
Expoitation of Resources
Regional Parallels of the Fabrics
Summary

3.1.3 FORMS
Recording and Classification
Form Groups
Regional Parallels
Summary of forms

3.1.4 CERAMIC PHASING
Sample Size
Fabric Proportions Across the Gravel Islands
Form Groups Across the Island
Illustrated catalogue of form groups across the Islands
Intra-Island Variability
Summary

3.1.5 FORM AND FUNCTION
Form and Fabric
Decoration
Surface Treatment
Use Alteration Traces
Conclusions

3.1.6 VESSEL SIZE
Thickness
Rim Diameter
Vessel Capacity
Summary

3.1.7 DISCUSSION AND CONCLUSIONS
Shifting Settlement and Social Patterns: The Evidence from Changing Fabrics
Production
Vessel Use
The Social role of Pottery
Conclusion and Recommendations

3.1.8 BIBLIOGRAPHY

3.1.9 PETROLOGICAL REPORT ON THE NON IRON AGE POTTERY AND DROITWICH SALT CONTAINERS

3.1.1 Introduction

A total of 4981 sherds of pottery, weighing 57,627g, was recovered from the settlement at Warrens Field, Claydon Pike. The form and fabric of the vessels suggested the pottery was predominantly middle Iron Age in date. The assemblage had previously been recorded by Simon Palmer in the early 1980s, and a draft report completed. The current report presents the pottery in a greater level of detail and is based upon an Masters degree dissertation for Southampton University.

Top of Page

Condition of the assemblage

Most of the pottery had been bagged according to context in Corinium Museum pottery archive boxes. Sherds that had already been drawn were boxed separately, as had the reconstructed vessels. These vessels were no longer glued but were stored in an unsatisfactory condition, divided only by layers of newspapers, causing a mixing of labels and sherds. The average sherd size of the assemblage is 11.6 g, this compares to 10 g at Thornhill Farm, 10.5 g at Gravelly Guy, 12.8 g at Guiting Power, 7.8 g at Farmoor, 6.5 g at Ashton Keynes and 6 g at Groundwell West. The condition may be described as average to poor, with some of the pottery quite abraded. Over-zealous cleaning has destroyed some of the surface treatments, leaving toothbrush impressions of the sherds. It is not certain how much this might have affected evidence of use such as burnt residues. Notwithstanding a fairly large number of sherds show evidence of sooting and food residues.

Sherd count

The coarse calcareous fabrics were particularly prone to splitting horizontally and smaller sherds often lost one or both surfaces. The initial assessment of the pottery in the 1980s recorded a sherd count of 3600. This has increased significantly, partly because sherds within a context that were found to join (both old and new breaks) were counted as a single sherd. This was particularly prominent in the case of the large reconstructable profiles. A sherd count of one clearly indicates a single vessel and in that respect is a useful indicator in an assessment of the numbers of vessels present, but not the overall ‘brokenness’ of the assemblage (Orton et al 1999). The sherd count was also complicated by the previous reconstructions as glued edges now obscured whether breaks occurred pre or post deposition. The sherd count has thus increased to 4981, although the weight has not been affected. As a result of this many of the analyses will depend on the weight.

Reconstructing the original ceramic population

The ceramic assemblage recovered from any site is usually only a sample of the original ceramic population (Orton et al 1999, 167). The archaeological record is biased towards smaller serving vessels and cooking pots that are likely to break more frequently than storage vessels, the latter are designed to withstand mechanical shock and are moved on a much more infrequent basis. These limitations have been considered with regards to the pottery analysis. A number of Roman and post-Medieval field ditches traverse the site and several features contained small abraded Romano-British sherds, however the degree of intrusion was low. The pottery is of a single period (middle Iron Age) and the problems of residuality noted on sites such as Gravelly Guy (Duncan et al forthcoming) were not found at Warrens Field.

Top of Page

Aims and Objectives of Analysis

The primary analytical aims of the pottery analysis were based upon the potential of the material, and the overall aims of the Cotswold Water Park project. These were:

  1. To discover if the three gravel islands identified with middle Iron Age occupation were in use simultaneously, or instead reflect a pattern of shifting settlement.
  2. To identify the scale of production, including the procurement of resources and the factors that may have affected the choices potters made in regards to manufacture.
  3. To ascertain the intended and actual use of the Claydon Pike ceramic vessels. Are they comparable with other sites in the region?
  4. To identify the social context of use and the role pottery might have played in creating and renewing social networks, through acts such as feasting.

Exploration of these aims were carried out through the following objectives:

  1. Identification of the fabric compositions of all sherds, and grouping into analytical units.
  2. Construction of a form series for the assemblage, identifying parallels with other sites in the region.
  3. Recording evidence of surface treatments and use-alteration.
  4. Correlation of form, fabric, surface treatments and use-alteration traces to ascertain the intended and actual use of vessels.
  5. Consideration of the sources of raw materials, mode of production and identification of non-local vessels.
  6. Assessment of the spatial distribution of fabrics and form to infer any chronological changes that may be visible across or within the three gravel islands.
  7. Recording of the sizes of vessels, assessment of the unimodal or multi-modal occurrences of vessels within a group, to identify vessels that may have been used for communal consumption of food and drink.

Top of Page

Methodology

The pottery was recorded at Oxford Archaeology using the Oxford Archaeological Roman Pottery Recording System (Booth in-house report). The recording guidelines presented in the Prehistoric Ceramics Research Group Occasional Papers 1 and 2 (PCRG 1995), have also been adhered to. Each sherd was analysed using a binocular microscope (x20 power), and a fabric series constructed. The OAU system of fabric code designation was used, this alpha numeric system assigns a letter to the two primary inclusions and a number to denote fineness. This was complimented by recording the details of each fabric on PCRG suggested sheets, including details of inclusion type, size, frequency, roundness and sorting, plus general characteristics of firing and texture. Two reference collections of the fabrics were made. Each context was recorded on a separate proforma sheet, with each related group of sherds (thought to be from a single vessel) allocated one record. Details of fabric, manufacture, firing, surface treatment, number, weight, rim type, rim diameter, percentage present, base, decoration, joins, wear and soot were recorded on the OAU sheets. A column was added for vessel wall thickness, codes detailed in section 3.1.6 (cf Morris 2000).

Every rim and base sherd, plus any decorated sherds, were allocated a featured sherd record number. Each featured sherd was drawn and recorded on an individual sheet (see section 3.1.3). Details of context, fabric, number, weight, firing, surface treatment and use evidence were recorded. A total of 240 sherds, or groups of joining sherds, were documented in this way. These sheets were used to build the form typology and formed an integral part of the analysis. A number of vessels had been illustrated previously, these were checked and a selection of further vessels were chosen for illustration.

Top of Page

The Ceramic Archive

The pottery was re-bagged by context and fabric, and placed in the original boxes. A much larger number of rims and bases have now been drawn and have been added to the original box of drawn sherds. A list of all extracted sherds can be found with the primary records. All original data sheets are currently housed at Oxford Archaeology, but will in due course be sent to Corinium Museum. One fabric reference collection will be housed at Oxford Archaeology, the other will be sent to Corinium.

Top of Page

3.1.2 Fabric analysis

A total of 33 individual fabric types were recorded from the Iron Age assemblage at Claydon Pike, although many of these were recovered in very small quantities. As a result the fabrics have been grouped according to principle inclusions for the purpose of data analysis, and to offer a clear impression of the characteristics of the assemblage (Table 1). Each fabric is detailed below.

Top of Page

Fabric Descriptions

Group 1. Fossiliferous limestone and shell

LT5 Coarse abundant fossiliferous shelly limestone

A soft, soapy fabric, with abundant (40%) fossiliferous detritus: limestone and fossil shell, 0.25-9mm, sub-rounded to sub-angular, shell fragments are often plate-like, poorly sorted. Includes rare (2%) fragments of bryozoa, most commonly seen in Jurassic deposits. The clay matrix contains rare silt sized quartz grains <0.5mm. Irregular fired, exterior surfaces are mostly oxidised, core and interior surface are often unoxidised. Breaks with hackly fracture, handmade.

LT4 Coarse abundant well rounded limestone and fossil shell

A soft, soapy fabric with abundant limestone and fossil shell (40%), limestone is <1mm and well-rounded, larger sub-angular fragments up to 2mm are also seen, shell is less frequent, 1-9mm, plate-like, poorly sorted; rare (1-2%) organic voids. Irregularly fired, oxidised exterior, unoxidied interior and core. Breaks with hackly fracture, handmade.

LT3 Abundant fine shelly limestone

Soft, soapy, slightly sandy fabric, with abundant (40%) limestone and fossil shell, 0.25-7mm, the average inclusion size is <1mm, sub-angular or plate-like, poorly sorted. Bryozoa fragments are occasionally visible. The clay matrix contains silt sized quartz grains. The exterior is oxidised, the core and interior are mostly unoxidised. Breaks with hackly fracture, handmade.

TL1 Abundant fine fossiliferous shelly limestone

A soft fabric with a slightly gritty texture, contains abundant (50%) fossil shell and limestone, well sorted, sub-angular to angular, mostly less than 0.5mm, occasional fragments up to 4mm. Fabric is mostly unoxidised. Breaks with hackly fracture, handmade.

TV4 Moderate fossil shell

A soft soapy fabric, with moderate (10%) fossil shell, poorly sorted, 0.5-4mm, angular or plate-like. Low fired and unoxidised, appears to have contained organic material which has burnt out. Breaks with a hackly fracture, handmade.

LP5 Coarse shelly limestone in a slightly sandy clay matrix

A soft soapy and sandy fabric, with very common (25-30%) limestone and fossil shell including rare bryozoa fragments, sub-angular to angular, <9mm, poorly sorted; rare to sparse (2-3%) clay pellets, moderate to well sorted, sub-rounded to sub-angular, up to 1mm, mostly <0.5mm; rare (1-2%) quartz, sub-rounded to angular, <1mm. Irregularly fired, exterior surface is oxidised, interior surface and core are unoxidised. Breaks with hackly fracture, handmade.

Group 2. Oolitic limestone and shell

LT2 Coarse oolitic limestone

A soft, soapy fabric, with very common (35%) oolitic limestone, the limestone fragments are mostly sub-angular with very well rounded oolite grains, 0.25-4mm (individual oolites are 0.25-0.5mm), poorly sorted; sparse (5-7%) shell, sub-angular, <3mm, poorly sorted; rare (1-2%) voids from organic matter. The firing is irregular, the surfaces are mostly unoxidised and the core and interior margin are oxidised. Breaks with hackly fracture, handmade.

LT1 Abundant fine oolitic limestone

A soft fabric with a slightly gritty texture, abundant (40%) oolitic limestone, well sorted, very well rounded, with a sparse to moderate amount of sub angular grains, <0.5-1mm. Sparse (3-5%) fossil shell, poorly sorted, sub-angular / plate-like, <7mm. The exterior surface and core are unoxidised, the interior margin and surface are irregularly fired. Breaks with hackly fracture, handmade.

Group 3. Non fossil shell

SL 5 Very common shell and sparse iron oxides

A soft, soapy fabric with very common (25%) white shell and occasional limestone fragments, <0.5-3mm, poorly sorted, angular; sparse (3%) iron oxides, <1 mm, sub-rounded to sub-angular; rare (1%) quartz grains, <1 mm, rounded to sub-rounded. Exterior surface and margin are mostly unoxidised, core is unoxidised, interior surface and margin are oxidised. This fabric represents a single intrusive Phase 2 (LIA - early Roman) vessel.

Group 4. Quartz, fossiliferous limestone and shell

AS5 Moderate quartz and shelly limestone

A sandy, slightly soapy fabric with a moderate amount (10%) of multi-coloured quartz grains, 0.25-2mm, most are 1mm, well-rounded to sub-angular, poorly sorted; a moderate (10%) amount of shelly limestone, 1-4mm, mostly 1mm, rounded to sub-angular, moderately sorted; sparse (5%) black grains <1mm, rounded to sub-angular, possible iron oxides. Irregular firing, mostly unoxidised. Breaks with hackly fracture, handmade.

AS4 Moderate quartz and sparse shelly limestone

A sandy, slightly soapy fabric with moderate (15%) quartz grains, 0.25-1mm, sub-angular with some more rounded grains, moderately sorted; sparse (3%) shelly limestone, <4mm, sub-rounded to sub-angular, poorly sorted; rare (1%) linear voids, possible firing cracks; rare iron oxides, <0.5mm, rounded. Irregular firing, tends to be unoxidised, hackly fracture, handmade.

LA3 Common limestone, moderate quartz and rare iron oxides

A soft, oxidised, sandy textured fabric with common (20%) limestone fragments, poorly sorted, 0.5-6mm, mostly rounded but with some larger sub-angular inclusions. Moderate (10%) quartz, 1-1.5 mm, sub-rounded to sub-angular; rare (1%) iron oxides, <1mm, sub-angular. Breaks with irregular fracture, handmade.

AV5 Common quartz and sparse limestone

A soft, soapy and slightly sandy fabric with common (20%) quartz grains, 0.5-1mm (mostly 0.5mm), sub-rounded to sub-angular, moderately sorted; sparse (3-5%) linear voids, <10mm, probably resultant from firing; sparse (3%) limestone, <6mm, sub-rounded to sub-angular; rare (1%) iron oxides, <2mm, sub-angular. Breaks with hackly fracture, handmade.

AS3 Quartz and sparse shell

A soft sandy fabric, with a moderate to common (15-20%) amount of milky and coloured quartz grains, <1.5mm, mostly <0.5mm, sub-rounded to sub-angular, larger grains may be quite rounded, moderately sorted; sparse (5%) shell, <10mm, angular, poorly sorted, rare (1%) iron oxides, <3mm, sub-angular. Breaks with hackly fracture, handmade.

Group 5. Oolitic limestone, shell and quartz

LA5 Coarse oolitic limestone with fossil shell and quartz

Soft, soapy fabric with very common (25%) oolitic limestone and fossil shell, <0.5-7mm, fragments are sub-angular with very well rounded oolites, shell is plate-like; sparse to moderate (7-10%) quartz grains, <1mm, rounded to sub-angular, well sorted. Fired in unoxidising conditions. Breaks with irregular fracture, handmade.

LA 2 Very common oolitic limestone and moderate quartz

A soft, soapy and slightly sandy fabric with very common (25%) calcareous inclusions, mostly oolitic limestone and a sparse amount of unidentified crystalline fragments, 0.5-9mm, the average grain size is 0.5-1mm, limestone fragments are sub-angular, individual oolites are very well rounded, poorly sorted; moderate (10%) quartz grains, 0.5-1mm, sub-angular to angular, moderately sorted; rare (1%) soft red clay pellets, 0.5-2mm, rounded. Exterior surface, margin and core are unoxidised, interior surface and margin are oxidised. Breaks with hackly fracture, handmade.

Group 6. Limestone and shell, quartz, possible glauconite or limonite grains

TA4 Very common fossiliferous shelly limestone, rare quartz and possible glauconite grains

A soft soapy and sandy fabric with very common (25-30%) limestone and fossil shell, poorly sorted, 0.5-7mm, sub-angular to angular or plate-like; rare (2%) quartz, sub-rounded, 0.25-1 mm, mostly 0.5mm; rare (1%) soft shiny reddish / black grains, <0.25mm, rounded to sub-rounded (possible glauconite or limonite). Irregularly fired, mostly in unoxidising conditions. Breaks with hackly fracture, handmade.

LA 4 Common shelly limestone, sparse quartz and possible glauconite grains

A soft, soapy and sandy fabric with moderate to common (15-20%) shelly limestone, 0.5-4mm, sub-angular to angular, poorly sorted; sparse (5%) quartz, sub-rounded to angular, 0.7-1mm, moderate to well sorted; sparse (5%) shiny soft black grains (possible glauconite or limonite), <0.25mm, well rounded, well sorted; rare (1-2%) soft round red inclusions, probably ferrous, sub-rounded to well-rounded, 0.5-1mm. Irregularly fired, surfaces and margins are oxidised and the core is unoxidised. Breaks with hackly fracture, handmade.

Group 7. Quartz and iron oxides / pellets

AI4 Very common quartz, sparse ferric inclusions

A soft, sandy fabric, contains very common (25%) quartz grains, 0.25-1.5mm, sub-angular with occasional more rounded grains, mostly 0.5-1mm, poorly sorted; rare to sparse (2-3%) iron oxides, sub-angular, <1-2mm, poorly sorted; rare possible iron pellets, <3mm, sub-rounded; rare linear voids. Oxdised exterior surface and exterior margin, unoxidised interior surface and interior margin, irregular core. Breaks with hackly fracture, handmade.

AI3 Very common quartz, sparse iron, possible glauconite

Sandy and slightly soapy soft fabric, common to very common (20-25%) multi-coloured quartz grains, 0.25-4mm, average size 0.5-1mm, rounded to sub-angular, poorly sorted; sparse (3-5%) iron oxides, <5mm, sub-rounded to sub-angular; sparse (3-5%) rounded glauconite grains; rare (1%) linear voids from organic matter, approximately 150mm, but variable. Oxidised exterior, unoxidised interior, the core is mostly oxidised, however the firing and resultant colours are extremely variable.

AI2 Moderate fine quartz, rare ferric inclusions

A soft, soapy and sandy fabric, contains a moderate to common amount (15-20%) of quartz, sub-rounded to sub-angular, <1mm, mostly 0.5mm, well sorted; rare iron oxides (2%), sub-angular to angular, <1.5mm; rare (1%) blackish brown inclusions, possible iron pellets, well rounded, 0.5-1mm; rare (1%) unidentified calcareous fragments, sub-angular, <1mm. Oxidised exterior, slightly oxidised interior, unoxidised core. Breaks with hackly fracture, handmade.

AI1 Very fine quartz with rare ferric inclusions

A soft, sandy fabric consisting of common (20%) very fine quartz, <0.1mm, angular, moderately sorted; rare iron oxides / iron pellets, < 2mm; linear cracks from firing. Oxidised surfaces and margins, with a light grey unoxidised core. Breaks with hackly fracture, handmade.

IA4 Moderate iron and sparse quartz

A soft, sandy and slightly soapy fabric with moderate to common (15-20%) iron, <0.5-4mm, rounded to sub-angular, poorly sorted; sparse to moderate (7-10%) quartz, 0.5-1.5mm, sub-rounded to sub-angular, moderately sorted; sparse (2-3%) unidentified grey rock fragments, 0.3-0.4mm, sub-angular to angular, poorly sorted. Irregularly fired (mostly oxidised), breaks with hackly fracture, handmade.

IA3 Well sorted common quartz, sparse iron oxides

A soft, sandy fabric, with common (20%) quartz grains, mostly well sorted, <0.5mm, some larger grains 1-1.5mm, sub angular. Sparse (3%) iron oxides, 1-2.5mm, sub-angular. Mostly unoxidised, breaks with hackly fracture, handmade.

A5 Low fired sandy fabric with rare iron oxides

A fairly hard fabric, with sandy texture containing common (20%) quartz grains, rounded to angular, 0.5-4mm, poorly sorted, includes milky opaque grains and also multi-coloured grains; rare iron oxides (<1%), angular, <1mm. Irregularly fired, slightly oxidised exterior, unoxidised interior, unoxidised core. Breaks with hackly fracture, handmade.

AR2 Fine common quartz and sparse sandstone

A fully oxidised soft, soapy fabric. Contains very common (25-30%) quartz grains, <0.5mm, sub-angular to angular, very well sorted; sparse (3%) sandstone fragments, 1-5mm, sub-rounded to sub-angular; sparse (3%) dull red fragments, 0.5-1mm, sub-rounded, possibly clay pellets; rare iron oxides, <1mm, sub-rounded. Breaks with hackly fracture, handmade.

AB5 Common quartz, moderate iron, sparse shell

A coarse sandy fabric with common (20%) quartz grains, 0.5-1mm, sub-angular to angular, a few more rounded grains can also be seen, well sorted; moderate (15%) iron oxides and soft reddish grains (possible iron ore), rounded to sub-rounded, 0.2-1mm, moderately sorted; sparse (3-5%) fossil shell, 1-4mm, sub-angular, poorly sorted. Fired in oxidising conditions, breaks with hackly fracture, handmade.

Group 8. Quartz with possible glauconite, sparse calcareous inclusions

AB3 Very common quartz, sparse fossil limestone and possible glauconite grains

Soft, very sandy fabric with very common (25-30%) quartz grains, <0.5-1.5mm, mostly <1mm, sub-angular, poorly sorted; sparse (5%) fossiliferous limestone, including bryozoa fragments, 1-3mm, sub-angular to angular, poorly sorted; sparse (7%) soft blackish red grains, possibly glauconite, 0.5-1mm, rounded, very well sorted. Irregularly fired, the surface is mostly unoxidised, breaks with hackly fracture, handmade.

AB2 Common fine quartz, moderate possible glauconite grains

A soft, sandy fabric with a hackly fracture. Contains a common amount (20%) of sub-rounded to sub-angular quartz, 0.5-1.5mm, the average size grain is 0.5mm, poorly sorted; moderate amount (10%) of possible glauconite grains, well rounded, <0.25mm, very well sorted; rare (2%) iron oxides, sub angular, <1mm. May also have up to 1% shell, <2.5mm. Surfaces are oxidised, core unoxidised, handmade.

AC3 Common quartz, rare clay pellets

Soft sandy fabric with common (20%) multi-coloured quartz grains, <0.5-1.5mm, mostly sub-rounded although smaller grains can be more angular, poorly sorted; rare dull soft red round patches, possible clay pellets, 2mm; rare soft rounded shiny black grains, possible glauconite, <0.25-0.5mm. Breaks with hackly fracture, handmade.

Group 9. Quartz and grog

AG 3 Moderate quartz and grog

A soft, soapy and sandy fabric with a moderate (15%) amount of quartz, 0.5-2mm (mostly 0.5-1mm), sub-rounded to sub-angular, moderately sorted; moderate (10%) black grog, 0.5-1mm, angular, poorly sorted; sparse (5%) black iron oxides, 1-1.5mm, sub-angular to angular, poorly sorted. Quite a vesicular fabric, breaks with hackly fracture, handmade.

Group 10. Malvernian Fabrics

MLA: Peacock (1968, 415) fabric Group A. Pottery in this fabric group consists of a sandy clay matrix with deliberately added temper including ‘angular rock and mineral fragments’ from both igneous and metamorphic rocks. The postulated source of this fabric is ‘the Pre-Cambrian outcrop forming the narrow ridge of the Malvern Hills’ (ibid. 419).

B1/SUP>: Peacock (1968, 421) fabric group B1 contains abundant angular fragments of Palaeozoic limestone, added as temper to the clay. Further work by Morris (1983, 120) has suggested that this pottery may originate from the Woolhope area of east Hertfordshire.

Both fabrics are thought to represent the work of specialist potters involved in organised trade (Peacock 1968, 414).

Top of Page

Fabric Quantification

The quantification of each fabric, by count and weight, is presented in Table 2. The percentage for both measures are also shown, results are given to two decimal places.

Source of Raw Materials

The Warrens Field site lies on first terrace river deposits, near the confluence of the rivers Coln and Thames. ‘The gravel is predominantly oolitic limestone pebbles with some sand’ (Robinson, section 3.4). The geology of the area is dominated by drift deposits and the solid Jurassic strata. The Oxford Clay outcrops within 1-2 km of the site. Within 3-5 km deposits of the middle Jurassic Great Oolite Group can be found, principally the Forest Marble and Cornbrash Formations. At 6 km minor deposits of the Kellaways Clay can be located. Within 5-8 km are pockets of other Great Oolite deposits such as Athelstan Oolite. To the south of the site Coral Rag outcrops at a distance of around 9 km. Cretaceous deposits of Ferruginous sands and Lower Greensand are seen at 11 km and 14 km respectively. Gault Clay and Upper Greensand are located further still, at 18 km and 23 km (information from Ordnance Survey Geology Maps, sheets 235, 236, 252, 253).

The Oxford Clay is highly fossiliferous, containing abundant ammonites and bivalves (Sumbler et al 2000, 71). The Cornbrash ‘consists of a complex succession of limestones, interbedded with marls’ (ibid. 68). The deposit contains shell detritus and peloids, fossils are also commonly seen at some levels, ‘brachiopods and bivalves are the most common’ (ibid. 69). The Forest Marble is dominated by mudstone, although limestones occur throughout. ‘The most common rock-type is a coarse-to medium-grained, shell-fragmental and ooidal grainstone, commonly containing scattered large shells of pectinaceans and oysters’ (ibid. 68), towards the top of the formation the shell material is dominated by broken oyster shells. Bryozoa debris is a minor but characteristic feature of the deposit. The Kellaways Clay consists of ‘sandy or silty mudstone with thin muddy sandstone beds…Molluscan faunas, particularly bivalves and ammonites, are relatively abundant’ (Sumbler 1996, 46). The Coral Rag deposits are oolitic and typically contain ‘a large proportion of clay, in which are numerous fragments of shells’ (Richardson et al 1963, 87).

Top of Page

Exploitation of Resources

Just over 91% of the assemblage was made from clays with abundant fossiliferous limestone inclusions (fabric group 1), which may have been obtained from the immediately local Oxford Clay, or the deposits of Cornbrash or Forest Marble located approximately 3-5 km to the north. Descriptions of the local geology suggest that calcareous inclusions occurred naturally in the clay, and would have required little or no further added ‘temper’ (here defined as purposely added inclusions). The Oxford Clay may also have been the source of the sandy group 7 fabrics; Kellways Clay deposits, and also the Ferruginous Sands (located at 11 km distant) offered alternatives. Pockets of sandy clays were also located in the gravels. In addition, the Kellaways clay may have provided the raw materials for group 4. Oolitic limestone recorded in fabric groups 2 and 5 probably originated from the gravel, deposits such as the Athelstan Oolite may also have been suitable. Coral Rag was an alternative, outcropping 9 km to the south. The non-fossil shell in group 3 may have originated from the gravel terrace deposits. The dominant fabric in group 7 (AI3) contained glauconite grains, indicating a Greensand origin. The presence of glauconite is also suspected in fabric groups 6 and 8. Lower Greensand is located 14 km to the south-east of the site, with Gault Clays a further 4 km to the south-east (18 km from the site) and Upper Greensand a total of 23 km distant.

Morris (1997) has advocated the use of a model of resource procurement, such as Arnold (1985), in order to define ‘local’ wares and thereby identify the presence of non-local wares. Arnold’s study of ethnographic data indicated that 33% of communities used a clay source located within 1 km of the production area, this was termed the ‘preferred threshold’. In total 84% of communities were prepared to travel up to 7 km (geodesic distance) to a clay source, and this represented ‘the upper limit of the maximum range of exploitation’ (Arnold 1985, 50). If additional tempering material was required, potters would be willing to travel up to 10 km. A radii of 7 km around a site thus represented its resource area (Arnold 2000).

This model suggests that the Claydon Pike fabrics indicate local resource procurement, and therefore local production. The exception is the presence of glauconite grains in sandy fabrics AI3, AC3, AB2 and AB3, which together account for 2.9% of the total assemblage. This indicates a Greensand source, and these febrics may therefore be seen to represent non-local procurement. The Malvernian fabrics, group 10, are non-local and part of a regional distribution network. They are often found on sites where Droitwich Briquetage is present, here they represent only 0.1% of the total weight of the assemblage. The distance to the source is approximately 65 km (Morris, this section).

Morris (1997) has questioned why non-local resources were utilised when alternatives were available locally. She presents the case of Danebury where ‘local sandy fabric pottery and non-local glauconitic sandy pottery were found in the same phases of occupation’. Was the glauconitic sand fabric in some way superior to that of the local sandy fabric, or does this represent ‘the maintenance of exchange networks’ (Morris 1997, 38) within regional communities?

Overall, the inclusions in the fabrics may occur naturally in the clays, and represent minimal processing, with the exception of the single grog-tempered sherd in group 9, and the Malvernian fabrics. However, fabric LT3 may also indicate the tempering of clay as the inclusions appear to have been crushed. The presence of larger quartz grains in a matrix of silt-sized grains may also be indicative of tempering, iron oxide inclusions could also have been added to the clay.

Top of Page

Regional Parallels of the Fabrics

Parallels for the middle Iron Age fabrics present at Claydon Pike were sought from 12 other sites in the Upper Thames and Southern Cotswolds. Bibliographical references are listed after the site name for each heading, in the text thereafter sites are referred to by name only. Claydon Pike may be abbreviated to FCP.

Thornhill Farm (Timby 2004)

Thornhill Farm is located less than 1 km from Claydon Pike and as such shared the same resource area (see above). It showed a smaller range of fabrics (13 types), perhaps unsurprisingly as the assemblage is much smaller than that from Claydon Pike, with 699 sherds and 7009 g attributed to the middle Iron Age (Period A). The Thornhill Farm assemblage was dominated by the presence of fabric C24, a ‘reddish-orange, brown or grey ware with a moderate to common frequency of inclusions comprising various fossiliferous fragments – shell, bryozoa, limestone and discreet ooliths’. It accounted for 70% of the total weight from this period and is comparable with the group 1 fabrics from Claydon Pike, predominately LT5 and LT3. Also seen at Thornhill is fabric C29, accounting for just over 1% of the weight of the period, containing a common amount of oolitic limestone, this may equate to group 2 fabrics at Claydon Pike. A small amount of Palaeozoic limestone tempered fabric B1 was seen at both sites, at Thornhill it becomes the dominant fabric in the late Iron Age phase. Sandy wares were not frequently seen at Thornhill, fabric R00 is described as very sandy but with no further visible inclusions. It represented only 0.7% of the total number and weight for the period, and may be similar to the group 7 sandy fabrics at Claydon Pike. One surprising aspect of the Thornhill assemblage was the presence of 166 sherds from a coarse Malvernian rock-tempered vessel from a single pit. The vessel may be earlier in date, possibly a late Bronze Age urn.

Ashville, Abingdon (DeRoche 1978)

At Ashville, Abingdon, a two fold fabric division was used to classify the assemblage: Fabric 1, containing abundant calcareous fragments, usually shell, and Fabric 2, sherds without significant calcareous inclusions, usually sandy. Fabric 1 is probably fairly similar to group 1 at Claydon Pike, with Fabric 2 perhaps encompassing some of the more sandy fabrics such as group 7. At the Ashville site Fabric 1 was characteristic of Period 1 (c 550-300 BC) but was also common in Period 2 (c 300-50 BC). The sandier Fabric 2 was rare in period 1 but dominates in Period 2.

Farmoor (Lambrick 1979)

A similar situation can be seen at Farmoor, where shelly fabrics dominate in the earlier period, Phase I, c 550-300 BC, and are then replaced by more sandy fabrics in Phase II, lasting into the second and possibly the first century BC. Coarse shell fabric B3 is most likely paralleled by FCP LT5, medium shell B2 by LT3, however there does not appear to be a parallel for the dark fine shell fabric B1 seen at Farmoor. An ochreous shelly fabric (DB) may equate to FCP LP5, at Farmoor this fabric is characteristic of the early period. The medium sandy fabric A3 is probably rather similar to FCP AI3, with parallels for the finer A2 and A1 found in FCP AI2 and AI1. A number of mixed calcareous fabrics were also recorded at Farmoor. The fine AB1 (possibly similar to AS4 at FCP) was regarded as fairly characteristic of Phase II, whereas the coarser AB2 was seen as a more ubiquitous fabric that lasted throughout the Iron Age, parallels may be seen in AS5 and LA5 at Claydon Pike.

Watkins Farm (Allen 1990)

Ten dominant fabric groups were identified at Watkins Farm, the largest was group 2 (sandy), which equates to group 7 at Claydon Pike, fabric 2a having much in common with FCP AI3. The next most commonly occuring inclusions were calcareous, represented by group 6. This included fabric 3a, a coarse to fine shelly limestone which may be similar to FCP LT5 and LT3. Fabric 3 contained coarse limestone fragments, fabric 3b oolitic limestone (FCP group 2) and 3c fossil shell. Group 5 (mixed sand and calcareous inclusions) and group 3 (coarse sand and ferric and calcareous inclusions) also formed significant proportions of the assemblage, the only other notable fabric was group 4 (ferric inclusions). The sandy fabrics appeared to increase over time. Whilst the range of fabrics at Watkins Farm is fairly similar to Claydon Pike, the proportions are not, due to a greater presence of the sandy fabrics. Watkins Farm is dated to the later middle Iron Age, c 250 – 50BC.

Mingies Ditch (Wilson 1993)

At Mingies Ditch the dominant fabrics consisted of a sandy ware and a shelly limestone fabric. The calcareous fabric was believed to be local in origin, the sandy fabric may have a Greensand source as 67 sherds exhibited clear black glassy grains suggesting a possible glauconitic sand for at least part of the group. Upper Greensand deposits are located 16 km to the south of Mingies Ditch, Lower Greensand 9-10 km to the east. An analysis of the proportions of the calcareous and sandy fabrics did not suggest chronological patterns as had been seen on other sites in the region. The site is dated c 380 – 110 BC.

Ashton Keynes (Edgeley-Long unpublished)

The Iron Age fabrics from Ashton Keynes were dominated by a coarse fossil shell and limestone fabrics, S1, paralleled at FCP by LT5. It accounted for 70% of the total weight of middle to later Iron Age pottery. A finer version, S3, was also present, similar to FCP LT3. Palaeozoic limestone fabric Bı represented 8.4% of the assemblage, however its presence on this site is probably late Iron Age in date. Oolitic limestone fabrics represent 8% of the assemblage, handmade sandy fabrics are poorly represented.

Gravelly Guy (Duncan, Lambrick and Barclay in prep)

The Iron Age pottery at Gravelly Guy was divided into eight main fabric groups on the basis of the predominant inclusion: shelly (SH), sandy (SA), calcareous gravel (CG), limestone (LM), alluvial (AL), flint (FL), organic (OG) and no added temper (NO). The predominant sandy fabrics include SA6 with abundant clear quartz grains, AS7 with abundant multi-coloured quartz and SA20 with possible glauconite grains; two of the SA fabrics also contained shell. This group is paralleled by groups 7, 6 and 4 at Claydon Pike. The calcareous gravel fabrics included CG1 with large amount of rounded limestone and fossiliferous detritus plus occasional quartz, CG22 contained high amounts of quartz. This group is paralleled by FCP group 1. Two limestone fabrics were recorded, LM37 and LM24, both oolitic, similar to FCP group 2. The alluvial fabric AL3 contained non-fossil mollusc shell, with rare quartz and calcite. The early Iron Age ceramic phases at Gravelly Guy are dominated by shelly and to a lesser extent sandy fabrics. In the middle Iron Age IV and V phases sandy and calcareous gravel fabrics dominate (predominately calcareous in MIA IV, sandy equalling in MIA V), although shelly fabrics are still present, possibly residually. By LMIA VI very little shelly material is seen. The Claydon Pike material is most similar to ceramic dating group MIA IV, the transitional earlier middle Iron Age.

Groundwell Farm (Gingell 1983)

Shell-tempered fabrics made from clays weathered from the Corallian Beds accounted for much of the assemblage. This class includes fabric A, a coarse fossil shell and limestone, accounting for 17.6% of the assemblage, and was commonly used for jar forms. It parallels FCP group 1. Fabric B consisted of an oolitic limestone fabric, seen in 34.2% of the pottery, this was mostly utilised for strongly constructed jars and also cooking vessels, this fabric is probably very similar to FCP group 2. The most popular fabric (40%) was C, a coarse sand with occasional flint or limestone fragments, used for both jars and bowls (FCP group 7). Less common fabrics include D (3.5%), a fine sand with rare flint; fabric E (0.1%), detritus including sandstone fragments; fabric F (2.5%) which was similar to fabric B but with a high proportion of shell and limestone; and miscellaneous fabric G, 1.6% of the assemblage.

Groundwell West (Timby 2001)

Sixteen fabrics were identified at Groundwell West, characterised by eight principle inclusion types: fossil shell; limestone and fossiliferous material; sandy with limestone; sandy with organic and limestone inclusions; sandy with iron and limestone; grog-tempered and flint-tempered. The dominant fabric was shelly SH1 representing 32% of the total weight, followed by fossiliferous limestone LI2, accounting for 27% of the weight of the assemblage. The remaining calcareous fabrics total 14% by weight with sandy fabrics totalling 7%. The LI2 fabric is probably most closely paralleled amongst the FCP fabrics by group 1, whilst Groundwell West fabric LI4 is similar to the oolitic fabrics from Claydon Pike, group 2. The Groundwell sandy fabrics are relatively fine and may equate to FCP AI1, AI2 or AI4. The sand and limestone, plus sand, iron and limestone fabrics are also paralleled, however the more organic, flint and grog fabrics are not.

Deer Park Road, Witney (Timby 1995)

The assemblage from Deer Park Road was smaller than many of the other sites in the region (only 366 sherds), and therefore a smaller number of fabric types were seen. Calcareous fabrics include a coarse fossil shell (H1), two fabrics with varying proportions of oolitic limestone (H2 and H3) and one of limestone and fossil shell (H4). The latter two are represented at FCP by groups 2 and 1 respectively. Sandy fabrics included two of very fine sand (S1 and S2) and one with glauconite sand inclusions (S3), possibly similar to FCP groups 7 and 8. A small number of sherds were also attributed to a fabric with ferric inclusions. The calcareous inclusions ‘account for 64% by weight compared to 35% sandy wares’ (Timby 1995, 80).

Uley Bury (Saville and Morris 1983)

The fabrics from the Iron Age assemblage at Uley Bury are dominated by local calcareous fabrics, predominately one of oolitic limestone and fossil shell (FCP group 2) and one of only fossil shell, together accounting for 68% of the assemblage. A calcitic fabric accounts for a further 21% and Paleaozoic limestone fabric Bı 8.7%. Sandy and micaceous fabrics occur only in small quantities.

Guiting Power, Gloucestershire (Saville 1979)

The fabrics from this Cotswold site were predominately limestone and / or shell, including Palaeozoic limestone. Non calcareous sherds including quartz were present in lesser quantities.

Top of Page

Summary

The pottery from Claydon Pike mostly represents local procurement of resources. Local clays may have contained sufficient natural inclusions for use, requiring little added temper. The presence of glauconite grains in several of the sandy fabrics indicates an exchange or trade of vessels constructed from a Greensand fabric. Longer distance regional trade is represented by the presence of Malvernian fabrics, albeit in small quantities. The coarse Jurassic fossiliferous limestone fabric LT5 that dominates the assemblage is fairly ubiquitous and characteristic of the locally produced pottery in the region during the early and middle Iron Age. On the whole the range of fabrics seen at Claydon Pike readily finds parallels at most of the other sites. The most striking difference is the proportions in which they appear, a factor that appears to represent chronological differences, with a shift from calcareous to sandy fabrics during the middle Iron Age. At Claydon Pike calcareous fabric group 1 accounts for 91.2% of the total weight of the assemblage. The only other fabrics with a noticeable presence are the sandy fabrics totalling 3.4% and mixed calcareous and quartz fabrics with 3.8%. These proportions tend to be seen on other sites throughout the region that are dated earlier within the middle Iron Age period.

Top of Page

3.1.3 Forms

Recording and Classification

The middle Iron Age vessels recorded from Claydon Pike were handmade and as such are unique. Every rim, base and decorated sherd was recorded onto a proforma feature record sheets (see ‘Methodology’, section 3.1.1). Each record was allocated a featured sherd (FS) number and also a rim or base number to record details of form. Where possible very similar rims were allocated the same rim number, however most showed subtle differences leading to a large number of rim types. These individual rim forms therefore needed to be grouped to form a classification. The purpose of constructing a classification was to make the data more manageable, allow details of fabric, surface treatment and use to be combined with vessel shape to indicate intended and actual use, to quantify the presence of forms, and to search for parallels in the region. ‘The principal is that the similarity of entities within groups does not occur by chance but reflects something inherently significant in their nature’ (Rice 1987, 274). Descriptions of the form groups is shown in Table 3, with the quantification of vessels present and figure number.

The classification has been based upon rim form and predicted profile. The assemblage was characterised by slack forms, and in many ways there is a forms continuum, with the gap between two form groups gradually becoming filled with intermediate forms (Orton et al 1999, 78). This was particularly noticeable in groups such as four and five, divided by barrel or globular profiles. Here a very basic version of the ‘Envelope System’, devised by Clive Orton, was employed to assess the validity of the divide between groups. The profile of each sherd in the group was traced onto a single sheet, each starting from the same point. Although the two groups showed separate clusters of profiles, the least globular profile of group 4 was found to be fairly close to the more rounded example of group 5, therefore indicating a continuum between the forms. As the two groups did not overlap the division has been left in place. This classification represents a ‘splitting’ rather than ‘lumping’ approach (Orton et al. 1999). Group 15 is unique in that it is primarily classified by decoration. This is a result of the chronological distinctiveness of fingertip decoration.

Parallels for each form group were sought from 12 key early and middle Iron Age sites in the region. The later Iron Age site at Salmonsbury was also consulted. Detailed searches amongst illustrated forms are presented below.

Definition of forms

Bowls have been classified as vessels whose height is less than their maximum width. Jars are defined as vessels whose height is greater than their maximum width. If the diameter of the mouth is equal to or greater than the maximum diameter of the body the vessel may be described as having an unrestricted orifice (Rice 1987, 212). Unfortunately few profiles were reconstructable and divisions into jar / bowl forms have thus been problematic. To avoid incorrect interpretation, the term ‘vessel’ has been used in most cases.

Top of Page

Form Groups

A total of 128 vessels were identified to form using rim and profile characteristics. As noted above a description of each form group is provided in Table 3.

Miscellaneous

A large number of the drawn rims were too small to be allocated to a specific form group. These were sub-divided into four categories based on the characteristics of the sherds. For the purposes of the database they have been numbered 16-19.

16. Squared or flattened rim, slightly in-turned or upstanding, unknown vessel profile.

17. Rounded rim, in-turned or upstanding, unknown vessel profile.

18. Rounded rim, slightly shaped, unknown vessel profile.

19. Flattened rims, slightly shaped, unknown vessel profile.

Bases

Bases were mostly plain with flat bottoms and obliquely splayed walls. Five forms were identified.

B1: Base with obliquely splayed wall, usually 70-80º from horizontal. A clear indication of finger-pinching was seen at the lowest point of the wall, a number of the bases showed finger imprints in this area. The pinching causes a slightly protruding foot. Twenty one vessels exhibited this base form.

B2: Base with slightly splayed or expanded foot. Wall angle approximately 60-70º from horizontal. Only five vessels were associated with this base form.

B3: A plain simple base, clean angle between wall and base. Angle of wall is 60-80º from horizontal. Four vessels represented.

B4: A plain simple base, the wall rounds to the base without any form of pinching. Examples are more splayed than B3 and may indicate a bowl form. Represented by six vessels.

B5: Footring base with pronounced foot. Seen on only two vessels.

Lids

One possible lid was recovered from context 264/F/1 (FS 070). This had a convex domed profile, and a diameter in the range of 16-18 cm. The potter appears to have run their finger around the outer exterior edge, possibly to finish it. A slight recess runs around the edge of the interior, probably to allow it to sit on top of a vessel. Coil built, and heavily sooted on the interior.

Top of Page

Regional Parallels

Group

Group 1 equates to Ashville group B3, barrel-shaped jars, seen predominately in Period 2 in both sandy and shelly fabrics. Examples include fig. 32, 20 and 23; fig. 42, 123 and 138; fig. 43, 150. At Farmoor most of the barrel shaped vessels had shaped necks and therefore are seen as part of group 3, however one neckless example is fig. 23, 106, in a sandy fabric, period II. At Mingies Ditch there are parallels in coarse calcareous fabrics, fig. 34,2 and fig. 35, 32; a sand and shell fabric was also used for this form, fig. 34,12. Several examples were noted at Watkins Farm, fig. 21,5; fig. 22,9; fig. 23,22; fig. 24,18; fig. 25,2. This form also has parallels at Gravelly Guy in forms 57 and 62, fig. 7.2,22; fig. 7.5,79 and fig. 7.11,178 mostly in a coarse calcareous fabric, but also a sandy form. At Groundwell Farm the form can be seen in a calcareous fabric, fig.14,2 and also a more sandy fabric, fig. 15,1. Two examples have been illustrated from Deer Park Road, fig. 6 numbers 6 and 7, in a limestone and fossil shell fabric. At Salmonsbury there is a first century BC form seen in fig. 14.5, although this example has a slightly rounder body, approaching a group 2 form. At Guiting Power a clear parallel is shown in fig. 9,1 in a fossil shell and oolitic fabric.

Group 2

There are no close parallels for this group at Ashville, however the form may have been encompassed by group B3. At Ashton Keynes it is equal to group 1, ovoid jars, fig 5: R1 and R6. Similar forms can be seen at Watkins Farm, fig. 23, 5 and 17; fig. 25, 1. At Farmoor it is seen in a shelly fabric in phase I, fig. 21, 19. No parallels were noted at Mingies Ditch or Groundwell Farm. At Gravelly Guy similar examples are seen in a coarse calcareous fabric, fig. 7.2, 28; fig. 7.8, 138; and fig. 7.9, 141, forms 56 and 60. At Guiting Power it is paralleled in a fossil shell and oolitic fabric, fig. 9, 3.

Group 3

Group 3 is also paralleled by Ashville group B3, barrel jars, examples include fig. 42, 121, a shelly fabric vessel seen in Period 2. This form was popular at Farmoor, with one coarse shelly example in Phase I, fig. 21, 16; all other examples were Phase II, mostly in a medium or fine shelly fabric, illustrated vessels include fig. 22 numbers 38, 45.46 and 51. At Mingies Ditch the form is seen in both calcareous and sandy fabrics, fig.34 numbers 11,13 and 15. At Watkins Farm it is also popular, examples include fig. 20, 1; fig. 21 numbers 5, 6; fig. 23, 14 and 24. At Ashton Keynes it was the most frequently seem form and is paralleled in group 7, fig. 7, R13, in a coarse shelly limestone fabric. At Gravelly Guy it does not appear as a distinct form, and may be described as a hybrid of straight-sided barrel jar form 57 and globular forms 55 and 58, in both calcareous and sandy fabrics. Examples include fig.7.1,15; fig. 7.2, 25; fig. 7.6, 96, 98, 103; fig. 7.7, 16. At Groundwell Farm the form can be seen in a sandy fabric, fig.16,1. One possible parallel occurs at Groundwell West, fig. 15.22, in a shelly fabric. At Deer Park Road it resembles fig.6 numbers 3 and 12, both seen in a sandy fabric. At Salmonsbury two examples can be found, fig.14, numbers 1 and 2. At Uley Bury the form may be paralleled by fig. 10.31. Similar vessels have been recovered from Guiting Power, fig. 9, numbers 2, 4, 13 and 14, they have been constructed from a palaeozoic limestone fabric.

Group 4

This form is similar to Ashville form D0, globular vessels with smoothed surfaces. Most examples are Period 2 and Fabric 2 (sandy). Illustrated vessels include fig. 42, 139; fig. 44, 181; fig. 46, 211; the globular jar B2 also has several similarities including fig. 32, 18 and fig. 49, 264. At Watkins Farm parallels include fig. 20, numbers 14 and 16; fig. 22, 5; fig. 24, 22. At Mingies Ditch comparable vessels are seen in a coarse calcareous fabric, fig. 34, 6 and fig. 35, 26. The form was not seen at Farmoor or Groundwell Farm. It resembles Gravelly Guy forms 55 and 56, fig. 7.1, 18; fig. 7.7, 123; fig. 7.9, 154 and fig. 7.10, 165, seen in calcareous, oolitic limestone and sandy fabrics.

Group 5

This group was more difficult to parallel due to the similarities with group 4. Comparable examples at Ashville were mostly globular bowls such as fig. 34, 32; fig. 52, 336; both period 3, and fig. 47, 236; and fig. 48, 255, both Period 2, all Fabric 2. At Watkins Farm examples include fig. 20, 15; fig. 21, 10; fig. 22, 8 and 10. At Farmoor parallels are seen in phase II, fig. 42, numbers 49, 52 and 70, in a sand and shell fabric and also a medium shell. No examples were found at Mingies Ditch or Groundwell Farm. It was also unparalleled at Gravelly Guy due to the lack of barrel-shaped vessels with distinctive rims, the closest examples are fig. 7.4, 67 and fig. 7.8, 128. The only possible parallel at Deer Park Road is jar form fig. 6.1. Two similar vessels can be seen from Guiting Power, fig. 9 numbers 5 and 12, both in an oolitic limestone fabric.

Group 6

Group 6 at Claydon Pike is equal to group A2 at Ashville. Examples include fig. 31, 5 in Fabric 1, and fig. 50, 280 in Fabric 2, both period 2. At Farmoor it is seen in phase II, fig.22, 50 and fig.23 numbers 92 and 102, all in a coarse shell or shell and sand mixed fabric. At Mingies Ditch fig.3 4, 5 is comparable, in a fine sandy fabric. Externally expanded rims are seen on middle Iron Age forms 55, 56 and 62 at Gravelly Guy, fig. 7.2, 21; fig. 7.6, 102; fig. 7.7, 107; fig. 7.7, 120 and 124, fabrics include shelly, sandy and oolitic limestone. Examples in a calcareous fabric were seen at Groundwell Farm, fig. 13, numbers 14 and 15; fig. 14, 1. At Deer Park Road it is paralleled by fig. 6.2, seen in a shelly fabric. At Uley Bury two examples can be seen, fig. 10 numbers 39 and 40.

Group 7

Group 7 equates to group A1 at Ashville, such as fig. 31, 3 and fig. 42, 137, both in a shelly fabric, the second example was shown to be Period 2. No parallels were noted at Watkins Farm, Mingies Ditch or Groundwell Farm, and only one example was found at Farmoor, fig. 44, 101, a fine sandy fabric in Phase II. Two examples have been illustrated at Gravelly Guy, fig. 7.4, 66 and fig. 7.5, 80 on forms 55 and 60, both in a shelly fabric. At Deer Park Road this form is seen in an oolitic limestone fabric, fig. 6.8.

Group 8

This form group can be paralleled with Ashville group A3, fig. 31, 10, in a shelly fabric. At Watkins Farm, Farmoor and Groundwell Farm no parallels were noted. Two examples were seen at Mingies Ditch, fig. 34 numbers 1 and 3, both in a coarse calcareous fabric. At Gravelly Guy the rim form is seen predominately on form 55 in a shelly or calcareous gravel fabric, fig. 7.4, 7, 29, 77 and 78; fig. 7.5, 82.

Group 9

The Ashville parallels are mostly globular vessels, again group D0. They are seen in a sandy fabric, mostly in Period 2. Illustrated examples include fig.46,210; fig.43,156; fig.48, 254 and 256; fig.34, 33 and fig.50, 289. At Mingies Ditch fig.34, 16 shows a possible parallel in a coarse shelly fabric with some sand. The Watkins Farm examples include fig.20 numbers 6 and 16; fig.21, 26 and fig.25, 23. Two parallels were noted at Farmoor, fig.42 numbers 53 and 73, both Period II. At Gravelly Guy bowl varieties of forms 54, 56 and 63 resemble this group, such as fig.7.1,11; fig.7.4,67; and fig.7.6,105. Here they are seen in shelly, sandy and alluvial fabrics. At Groundwell Farm examples are seen in both a sandy and calcareous fabric, fig.16, numbers 2 and 9; fig.13, 17.

Group 10

Parallels at Ashville include fig. 44, 185, a sandy fabric; and fig. 47, 222 in a shelly fabric, both Period 2. At Farmoor possible parallels include fig.21 numbers 3 and 10, however their forms are uncertain. At Watkins Farms examples include fig.23, 3; fig. 25, 1 and fig.25 numbers 15, 17, 19. Two parallels were seen at Mingies Ditch, fig.34, 2 and fig.35, 32, the fabric of the latter was noted as coarse calcareous. At Gravelly Guy similar forms can be seen in fig. 7.5.81 and fig. 7.8,130 in oolitic limestone and alluvial fabrics. No parallels for this group were noted at Groundwell Farm, however it is seen at the slightly earlier Groundwell West, fig.14 numbers 17 and 18, both in a shelly fabric. At Uley Bury one parallel was noted, fig.10, 46.

Group 11

No real parallels for this form group were seen from Farmoor, Mingies Ditch, Groundwell Farm or Groundwell West. At Ashville it is seen in fig. 42, 133 and fig. 45, 201, both in sandy Fabric 2 and of Period 2. At Watkins Farm examples include fig. 23, 7 and fig. 21, 3. At Gravelly Guy it is possibly paralleled by form 53, such as fig. 7.1.9 and fig. 7.11, 181. One possible parallel was seen at Deer Park Road, fig. 6.11, in an oolitic limestone tempered fabric. Similar vessels have been seen at Uley Bury, fig. 9.12; fig. 12 numbers 41,50, 51 and 56.

Group 12

Group 12 at Claydon Pike finds parallels in the coarse jars of Ashville, particularly fig. 32, 21; fig. 42, 121 and fig. 48, 259, one example was in a coarse calcareous fabric, the others a sandy fabric, all were Period 2. At Farmoor this form was again seen in both sandy and calcareous fabrics in Phase II, fig. 22, 45; fig. 23 numbers 75 and 109. At Watkins Farm it is paralleled in the following examples: fig. 21 numbers 7 and 25; fig. 22 numbers 7 and 10. One example was located at Mingies Ditch, fig.35, 30, in a medium sandy fabric. This form was quite popular at Gravelly Guy, paralleled by forms 58 and 59. Examples include fig. 7.2, 24, 25, 26 and 27; fig. 7.5, 95; fig.7.10,159 and fig. 7.12, 185. The forms are seen in calcareous gravel, ooilitc limestone and sandy fabrics. Examples from Groundwell Farm include fig.14, 29 and fig.15, 2 in calcareous and sandy fabrics. At Groundwell West it is seen in a fossil limestone and a shelly fabric, fig. 15, 19 and 20. A possible parallel at the Cotwolds site of Uley Bury is seen on fig. 9.12 and fig. 10.49.

Group 13

A parallel for the complete jar profile was seen at Watkins Farm, fig. 20, 17, however no parallels could be found at Ashville, Farmoor, Mingies Ditch or Groundwell Farm. Only one possible parallel was found at Gravelly Guy, fig. 7.1, 19, in an oolitic limestone fabric.

Group 14

Group 14 represents a curious form that appears to be out of character with the rest of the assemblage. No Iron Age parallels could be found at Watkins Farm, Farmoor, Mingies Ditch, Groundwell Farm or Gravelly Guy. The only possible parallels are wheel-thrown examples such as Ashville fig. 50, 296, a Period 3 necked bowl, or fig. 52, 355.

Group 15

Finger impressed decoration is seen predominately in period 1 at Ashville, mostly in Fabric 1, calcareous. Examples are seen on a number of different forms including expanded rim vessels (A), coarse-ware jars (B1) and angular vessels (C1). Illustrated vessels include fig. 31, numbers 6, 8 and 9; fig.32 numbers 14 and 16; fig. 33, 28; fig. 38, 78-80 and fig. 40, 110. Only one possible example was noted at Watkins Farm, and one was illustrated at Mingies Ditch, fig. 35, 34. At Farmoor, finger impressions were seen on rims 1 and 17 and shoulders 18, 24 and 38. Most were in a coarse shelly fabric and seen in Phase I, the exception was the final example in a sand and shell fabric, Phase II. This decorative technique is also seen on a number of forms at Gravelly Guy, predominately in the early Iron Age, and mostly on shelly fabrics. Illustrated examples of the technique include fig. 7.1, 4,7 and 8; fig. 7.4, 61, 62, 65 and 69; and also the later fig. 7.9, 144. No examples were found at Gravelly Guy. At Groundwell West fingertip impressions can be seen on illustrated vessels fig. 14, numbers 1,3,6 and 8, however it was not recorded at the later Groundwell Farm.

Bases

B1: This base form in seen in Periods 1 and 2 at Ashville, in both fabrics 1 and 2. Examples include fig. 37, 73; fig. 44, 192, 193, 195. At Farmoor it is again seen in both periods, and shelly and sandy fabrics, fig. 21.36; fig. 22, 67; fig. 23, 81 and 94. It is quite commonly seen at Mingies Ditch, fig. 34, numbers 10, 14, 18, in sand and calcareous fabrics. Watkins Farm examples include fig. 20, 7, 8 and 16. At Gravelly Guy this base is seen on some straight-sided vessels, but also more rounded forms. Groundwell Farm examples include fig. 13.7 and 8. At Guiting Power it resembles fig. 9,10.

B2: At Ashville this base is seen in both fabrics, and paralleled in fig. 44, 196; fig. 48, 262; fig. 49, 279. At Mingies Ditch parallels include fig.34, 15 and 29, in sandy and calcareous fabrics. Parallels at Watkins Farm are shown in fig. 25, 18 and fig. 23, 15 and 26. It can be found at Groundwell Farm in fig. 13.9; fig.14 numbers 22, 23, 26 and 27; fig. 15, 4.

B3: This form is seen in Period 2 at Ashville, in fig. 42, 136; fig. 44, 189, 190; fig. 47, 248 and fig. 49, 278, all are fabric 2. At Farmoor examples are found in Phase I and II, fig. 21, 27; fig. 23, 80, in sandy and shell/sand fabrics. Watkins Farm parallels include fig. 21, 24 and 25; fig. 23, 16; fig. 24, 20. At Groundwell Farm examples are seen in fig. 14, 18-21 and 30; at Groundwell West fig. 14, 13.

B4: Three parallels were found at Ashville, fig. 44, 191 and 194; fig. 49, 277, Period 2, fabric 2. At Farmoor it can be seen in fig.22, numbers 42, 58 and 74, in sandy and calcareous grit fabrics. At Mingies Ditch it is associated with a calcareous bowl form, fig.35, 33. The Watkins Farm parallels are fig.16, 9; fig. 23, 10 and 12; fig. 24, 17. At Gravelly Guy this base is associated with globular vessels. One example is found at Groundwell Farm, fig. 15.33. In the Cotswolds it is seen at Guiting Power, fig. 9 numbers 8 and 11; and at Uley Bury fig. 10, 43.

B5: One parallel for this form was seen at Ashville, fig. 50, 315, period 3. At Groundwell Farm two examples can be found, fig. 14.35 (oolitic) and fig.15.9 (sandy). It is also seen at the Cotwolds Uley Bury site, fig.10, 42.

Top of Page

Summary of Forms

The assemblage at Claydon Pike is dominated by barrel-shaped vessels, probably jars, with plain or slightly shaped rims, which account for 30% of the total number of recognised vessels. Forms with expanded rims (groups 6,7 and8) represent 20% of the vessels. 15.6% have shaped necks and internal bevels, and are mostly globular in profile. 11% were recorded as bowl forms (groups 9 and 10), 8.6% as ovoid jars (group 2).. A further 4.7% are globular vessels with rolled or beaded rims (group 11), and 3.1% were straight-sided vessels (group 12). A single high shouldered jar (group 13) and a vessel with a strongly everted rim (group 14) were also recorded, the latter is late Iron Age/early Roman in date. Vessels demonstrating finger-tip or nail decoration (group 15) account for 5.5& of the assemblge.

The earlier periods at Ashville and Farmoor (550-300 BC), and the early Iron Age phases at Gravelly Guy (EIA I & II) were dominated by angular forms, shouldered vessels, expanded rim forms and vessels decorated with fingertip / nail impressions. Possible angular jars are also seen at Groundwell West. These forms are either absent or present in only small quantities at Mingies Ditch, Watkins Farm, Ashton Keynes, Groundwell Farm, Deer Park Road, Uley Bury and Guiting Power. Where these earlier forms are seen, they often continue into subsequent phases, but become far less common. The presence of expanded rims and fingertip decoration would suggest an early focus for the Claydon Pike assemblage, although the complete lack of angular vessels may rule out a date earlier than the middle Iron Age. This early focus is also reflected in the fabrics which are dominated by calcareous inclusions (see section 3.1.2). Barrel-shaped and globular vessels are characteristic of the middle Iron Age in the region and are seen throughout the period, with well-developed globular bowls and beaded rims becoming increasingly popular towards the end of the period.

The Claydon Pike forms may therefore be described as fairly typical of the middle Iron Age assemblages from the Upper Thames Valley. Typologically, they appear to span the middle Iron Age, with a focus on the earlier part of the period, and fewer forms present from the latter part. No evidence of wheel throwing or other late Iron Age characteristics are present, the exception to this is that single Phase 2 vessel recovered from context 218.

Top of Page

3.1.4 Ceramic phasing

One of the key questions it was hoped the ceramic assemblage might answer was whether the three gravel ‘islands’ were occupied simultaneously, or if the site represented a pattern of shifting settlement. Throughout the region, fabric proportions and forms are seen to be chronologically significant. ‘In the Upper Thames Basin it has been observed that early-middle Iron Age sites tend to show a steady decrease over time in the use of calcareous fabrics (shell/limestone), with a concomitant increase in sandy wares’ (Timby 2001, 21-2). Timby presents the example of Groundwell West where 86% of the assemblage was composed of shelly wares, with only 7-8% sandy fabrics. At nearby Groundwell Farm a shift is seen to 40% sandy and 52% shelly wares. ‘Such a marked difference in character would strongly suggest a chronological difference between the two sites…..the focus of occupation appears to have been at Groundwell West in the earlier Iron Age, perhaps shifting or expanding to include Groundwell Farm in the middle Iron Age’ (Timby 2001, 23-4). This pattern is repeatedly seen at sites such as Ashville, Farmoor, and Gravelly Guy. Analysis of the proportions of fabric groups on the three gravel islands at Claydon Pike has thus been carried out to see if such patterns are present, accepting that the differing fabric proportions represent chronological factors (ibid.).

Top of Page

Sample Size

Hodder (1974, 72) has suggested that 30 sherds is a minimum sample size for the analysis of spatial trends, less than this may distort the results. The PCRG also suggest a minimum of 25-30 sherds (PCRG 1995). In terms of weight this might be expected to equate to approximately 300 g, based on a mean sherd weight of 10g. Contexts with less than 25-30 sherds, or 250-300 g will therefore not be used to phase features. At Claydon Pike 146 contexts produced a total of 4981 sherds, 57.6 kg. Of these, 88 contexts produced less than 10 sherds. Seventeen contexts produced 10-19 sherds, six contexts produced 20-29 sherds. A total of 35 contexts contained a ‘reliable’ number of sherds, 13 produced 30-<50 sherds, 10 produced 50-<100 sherds. Six contexts contained 100-<200 sherds, two contexts contained 200-<300 sherds, two contexts contained 300-<400 sherds, one context contained 400-<500 sherds and one contained 500-<600 sherds.

Top of Page

Fabric Proportions Across the Gravel Islands

A total of 57502 g was recorded from the three islands, 10657g from Island 1, 25076g from Island 2 and 21769g from Island 3. Analysis has been conducted using percentages of total weight, the number of sherds has not been included here due to the reasons presented in section 3.1.1. The proportions of the different fabric groups across the gravel islands are presented in Table 4, percentages are given to one decimal place.

Using a model of increasing sand to decreasing calcareous inclusions over time, it can be shown that island 3 is the earliest in date and island 1 the latest. 99% of the island 3 assemblage is composed of a coarse fossiliferous limestone and shell fabric (group 1), with oolitic limestone (group 2), quartz and shell (group 4), and quartz and iron (group 7) represented by less than 1% each. Not only is there a clear focus on the group 1 fabrics, but the other three fabrics are present in minute quantities.

Island 2 contained the largest assemblage and it is perhaps not surprising that all fabric groups are represented here, although it must be noted that it is only marginally larger than the island 3 total. The calcareous group 1 accounts for 87% of the total weight, the only other significant fabrics are group 5, oolitic limestone and quartz and the increasing sandy fabrics of group 7, now accounting for 3.5%.

Island 1 shows a shift again with a slight decrease in calcareous group 1 to 85%, and an increase in sandy group 7 to 9%. Other fabrics with more than 1% include the possible glauconitic sand group 8, and quartz and calcareous group 4.

Non-local pottery and Droitwich Briquetage

The Malvernian Wares and Droitwich Briquetage from Claydon Pike have been reported on by Morris (this section), and are summarised here. The Malvernian Wares are seen only in minute quantities, and are present on Islands 2 and 1. Both Bı and Group A wareswere part of a regional distribution network during the middle Iron Age period. Examples of both briquetage fabrics FT1 and FT2 were found at Claydon Pike, including a variety of the former, FT1a, that is often seen in earlier Iron Age collections. This fabric was found only on Island 3, along with sherds of FT1 and FT2. FT1 dominated the Island 3 assemblage, accounting for 61% of the salt container weight. On islands 1 and 2 FT2 sherds are more common (57% and 87% respectively). The difference in the proportions ‘is a pattern of technological change’, also seen at the production source where FT1 is slightly more common in the earlier phase ‘but completely overshadowed by FT2 in the later phase’ (Morris this section). The evidence from the salt container material also suggests that Islands 2 and 1 ‘were occupied slightly later in the Iron Age than Island 3’ (ibid.).

Top of Page

Form Groups Across the Islands

A total of 128 vessels were identified by form across the three islands. Thirty two vessels were identified on Island 1, 58 on Island 2 and 38 on Island 3. These are tabulated by group in Table 5.

Island 3 was dominated by barrel-shaped vessels, groups 1 and 3 account for 47.4% of the total number of vessels on the island. Expanded rim vessels (groups 6, 7 and 8) account for 15.8%, and those decorated with fingertip decoration (group 15) represent 10.5%. Vessels with shaped necks and internal bevels are the only other fairly commonly seen form, with groups 4 and 5 totaling 7.9%. On Island 2 vessels with expanded rims become the most common, representing 29.3% of the total number of vessels identified on the island. This is followed by barrel-shaped vessels, 20.7%; vessels with shaped necks and internal bevels, 15.5%; bowl forms, 12.1% and ovoid jars, 8.6%. A shift is again seen on Island 1 where barrel-shaped vessels and internally beveled vessels are equally dominant, each accounting for 25%. They are followed by bowls, 15.6%; ovoid jars, 12.5%; expanded rim vessels, 9.4% and globular vessels with rolled /beaded rims, 9.4%.

If the chronological variations suggested by the fabrics are accepted, then it may be argued that barrel-shaped vessels, expanded rim vessels and those decorated with fingertip impressions decrease in number with time. Ovoid jars, internally beveled vessels, bowls, globular vessels with rolled / beaded rims and straight-sided vessels increase with time. This pattern is borne out at other sites in the region.

Top of Page

Illustrated catalogue of form groups across the Islands

Fig. 3.1.1: Pottery from Island 3, S 20

  1. FS 125. Form group 13. Flattened rim with rounded edges, slightly thickened on exterior and interior. Constricted neck, sharp high shoulder, walls slope inwards, wide mouthed jar. A band of sooting is present around the shoulder. 425/G.
  2. FS 128. B1. Plain base, obliquely splayed. Smoothed exterior, pitted interior. 425/G.
  3. FS 210. Form group 3. Flat-topped rim, slightly shaped on the exterior and expanded on the interior. Shoulder area is marginally wider than the rest of the vessel. Barrel-shaped jar. Smoothed exterior with some sooting, pitted interior. 371/A/1.
  4. FS 047. Form group 4. Flat-topped slightly shaped rim, slightly rounded on the exterior forming a minor bead effect. Constricted neck, barrel-shaped profile. Minor external sooting. 371/A/2.
  5. FS 059. B1. Large base with little curvature. Wall to base join exhibits finer-pinching. 371/C/1.
  6. FS 037. Form group 7. Flat-topped rim, internally expanded, almost straight-sided profile. Slightly smoothed exterior and interior surface. 371/A/2.
  7. FS 122. Form group 12. Slightly flattened rim, exterior is rolled to form a very irregular bead, minor internal bevel. Vessel walls are quite straight sided, jar form. Very well smoothed exterior, sooting present on exterior. 425/F/3.
  8. FS 120. Form group 3. Flat-topped rim, very lightly shaped which causes a slight pulled effect on the interior. Barrel-shaped profile, probably a squat jar. Finger smoothed and wiped exterior with sooting. 425/G/3.
  9. FS 20. B1. Obliquely splayed base, slightly pinched in at wall / base join. Smoothed exterior. 371/F/2.
  10. FS 044. Form group 3. Flat topped rim, clay was irregularly rolled over the top creating a slight exterior lip. Barrel-shaped vessel. 371/A/2.
  11. FS 130. Form group 1. Un-differentiated rim from barrel-shaped vessel. 425/F.
  12. FS 039. Form group 1. Flattened un-differentiated rim with rounded edges, barrel-shaped vessel. 371/J/4.
  13. FS 041. Form group 3. Rounded rim with slightly constricted neck, barrel-shaped profile. Smoothed exterior. 371/A/2.
  14. FS 061. Form group 3. Fairly plain rim, vessel wall slopes out to probable slack-shouldered vessel. Sooted exterior. 371/J/4.

Fig. 3.1.2: Pottery from Island 2

Enclosure 6

  1. OPDW 34. FS 085. Form group 9. Slightly everted / shaped rim, globular body profile: rounded bowl. Smoothed exterior. 264/A/2.
  2. OPDW 37. FS 071. Form group 7. Flat-topped rim, thickened and squared on the interior. Neckless vessel, probably barrel-shaped. 264/A/2.
  3. OPDW 36. FS 081. Form group 6. Rim is pulled on the exterior and also bevelled on top. Fairly straight-sided vessel, neutral form. 264/A/2.
  4. OPDW 38. FS 074. B1. Obliquely flared base, slight pinching between wall and base. Internal residue and abrasion on the interior. 264/A/2.
  5. OPDW 41. FS 078. Misc. D. Riml appears to have been damaged during the firing process. Unknown vessel form. 264/A/3.
  6. New print publication drawing 1: FS 082. Form group 1. Flattened un-differentiated rim, barrel-shaped vessel, probable jar. Upper exterior sooting. 264/D/1.
  7. OPDW 39: FS 084. Form group 6. Flat topped rim, thickened on the exterior and very slightly thickened on the interior. Fairly straight sloping sides, probable jar. 264/B/1.
  8. OPDW 33. FS 060. Form group 3. Flat-topped rim, slightly shaped and upstanding. Slack-sided vessel profile, very squat jar, height is probably very similar to diameter. Small amount of internal burnt residue. 264/A/2.
  9. FS 090. B3 Plain simple base, wall rounds gently to the base. 264/D/2.
  10. FS 104. Form group 3. Rounded rim, clay has lapped over the top creating a mini bead. Probably a barrel-shaped jar. Smoothed, slurry finish on exterior. 267/A.
  11. FS 083. From group 10. Flat topped plain rim, squared edges, globular shaped profile, probable bowl. Smoothed exterior. 264/A/2.
  12. FS 076. Form group 6. Flat-topped rim, thickened on the exterior, internal bevel. Barrel-shaped profile. 264/A/3.
  13. FS 072. Form group 1. Flat-topped un-differentiated rim, barrel-shaped profile. Minor external sooting. 264/A/2.
  14. Structure 9

  15. OPDW: 58. Form group 9. Flattened slightly in-turned rim with a slight bead. Rounded profile, neutral form. Burnished exterior. 228/H/1.
  16. FS 106. Form group 5. Flat-topped rim, slightly everted and rolled on the exterior to create a bead effect, internally bevelled. Barrel-shaped profile, probable jar. Internal burnt residue. 228/H/1.
  17. FS 113. Form group 5. Flat-topped rim, with internal bevel and slightly shaped exterior. Slack / barrel sided vessel. External sooting. 228/H/1.
  18. FS 092. Form group 1. Flat topped square un-differentiated rim. Plain form, barrel-shaped vessel. 228/H/1.
  19. FS 091. Form group 12. Flat topped rim, thickened on the exterior to form an irregular roll of clay / bead. Slack profile, possible jar form. 228/H/1.
  20. FS 107. Form group 4. Upstanding rim, vessel profile is probably rounded. 228/H/1.

Fig. 3.1.3: Pottery from Island 1, E 2/S1

Enclosure 2

  1. FS 176. Form group 12. Flat-topped rim, slightly shaped on exterior causing bead effect and constriction in neck area. Slack-sided profile. 8/AE/2.

  2. FS 179. Form group 11. Flat-topped rim with rounded edges forming an irregular external bead. Vessel profile unknown, but probably a closed form. Smoothed exterior. 8/AH/1.

  3. FS 182. Form group 10. Un-differentiated rim on open form vessel, probable bowl. Exterior is smoothed. 8/B/1.

  4. FS 185. Form group 7. Flat-topped squared rim, internally expanded, unknown vessel profile. Smoothed exterior and interior. 8/C/4.

  5. FS 193. Form group 9. Fully everted rim with bead and slightly constricted neck. Appears to be an open bowl form. 8/O/1.

  6. FS 195. Form group 9. Slightly everted rim causes minor neck constriction, vessel is probably a bowl. Lower exterior is burnished. 8/P/1.

  7. FS 198. Body sherd decorated with two diagonal pairs of comb impressed lines, crossing at approximately 70°. Smoothed exterior. 8/Q/1.

  8. FS 196. Form group 4. Fairly flat-topped upright rim, sloping outwards slightly. Necked vessel with rounded profile. 8/Q/1.

  9. FS 200. Form group 5. Rounded rim on vessel with slightly constricted neck and internal bevel. Barrel-shaped profile. Burnished exterior. 8/R/2.

  10. Structure 1

  11. FS 166. Form group 4. Flat topped rim, slightly everted and upstanding. Vessel is necked with internal bevel, barrel to rounded profile. Smoothed exterior with minor sooting. 21/AM/2.
  12. FS 169. Form group 10. Un-differentiated rim, top is irregular. Rounded profile, open form. Smoothed interior. 21/B/2.
  13. FS 171. Form group 4. Upstanding rim, slightly shaped on upper exterior. Short neck, internally bevelled. Well sprung rounded walls. External sooting, internal pitting and abrasion. 21/G/1.
  14. FS 206. Form group 3. Flat-topped rim, slightly expanded on the interior. Barrel-shaped jar with wide mouth. B2 base. Smoothed exterior. 21/C/1.
  15. FS 156. B4. Base with pitted interior. 59/C/1.
  16. FS 174. Form group 1. Un-differentiated rim on barrel shaped vessel. Slightly smoothed exterior with a small amount of external sooting. 8/A/5.
  17. FS 177. Flat-topped rim with slightly constricted neck and very narrow bead. Barrel-shaped vessel profile, slightly smoothed exterior, minor exterior soot and internal burnt residue. 8/AF/2.
  18. FS 180. Form group 7. Flat-topped rim, expanded on the interior. Orientation uncertain. Smoothed surface. 8/AH/2.
  19. FS 197. Form group 9. Rounded rim, slightly shaped, vessel profile is rounded, probable bowl form. Very smoothed exterior, almost burnished. 8/Q/1.
  20. FS 194. Form group 1. Flat topped un-differentiated rim on barrel shaped vessel. External sooting. 8/P/1.
  21. FS 183. Form group 10. Un-differentiated rim on open form, probable bowl. Smoothed and wiped external surface. 8/B/1.
  22. FS 190. B1. Base with smoothed exterior walls. 8/J/1.
  23. FS 199. Form group 4. Upstanding rim on necked vessel, globular shaped profile, probable jar. 8/Q/1.
  24. FS 164. Form group 2. Un-differentiated rim on rounded profile vessel. Smoothed exterior with minor sooting. 21/A/2.
  25. FS 165. Form group 3. Slightly shaped rim from a barrel-shaped vessel. Exterior has been smoothed and wiped. 21/A/2.
  26. FS 168. Form group 3. Flat topped rim from barrel-shaped vessel. Smoothed exterior and interior. 21/B/1.
  27. FS 203. Form group 2. Squared off rim creating an internal bevel. Rounded profile, closed form. 8/V/1.

Top of Page

Intra-Island Variability

An attempt was made to see if this pattern might be projected onto feature groups within the islands. The total number and weight of sherds present in a single structure or enclosure feature was calculated and those with under 30 sherds were excluded from analysis. The largest feature assemblage came from the enclosure ditches associated with Structure 20 (S 20) on Island 3, amounting to 14979g, over 10 kg greater than any other feature at Claydon Pike. For each feature the number of forms present was recorded, however the results were inconclusive with most features showing only one or two examples of the different vessel groups, the range of forms present usually directly related to the sample size. The percentages of the total weight of fabric groups present in each feature are presented in Table 6 (Island 3), Table 7 (Island 2) and Table 8 (Island 1).

Island 3

The group 1 calcareous fabrics accounted for 99% of the total weight of island 3 vessels, and a breakdown into features has not added to this picture. All features contain this fabric in similar proportions, the only difference is linear boundary (LB) 372 where 7.1% of sandy fabrics were present. Stratigraphically LB 372 is truncated by LB 363 and in this particular case the fabric variation does not appear to represent chronological variation.

Island 2

A greater number of features were present on Island 2, and a greater variation in fabric proportions is shown than the Island 3 assemblage. The highest proportions of the group 1 fabrics were seen in Structures 5 and 6, plus Enclosures 6 and 7, these features also had relatively low ratios of sandy fabrics (group 7). Structures 7, 9 and 10, plus LB 322 contained the lowest proportions of the group 1 fabrics and the highest proportions of sandy group 7. Enclosures 5, 6 and 7, and linear boundary 322 are stratigraphically related. Of the enclosures, E5 contained the lowest proportions of group 1 fabrics and highest proportion of group 7, thus suggesting a later date, this is confirmed by the stratigraphic relationships. LB 322 truncates all three enclosures and this has a greater ratio of sand to calcareous than the three enclosures, again supporting the model. Structures 5 and 6 overlap in plan and their stratigraphic relationship is uncertain. Unfortunately most of the pottery from S 6 comes from a single vessel (hence 100% of fabric group 1) and the data is therefore unreliable for phasing of these two structures. However, the rim of the vessel is expanded on the exterior and interior (form group 7), such forms are often seen in the earlier part of the period, tentatively suggesting S 6 may be earlier than S 5. Structures 7, 9 and 10 are stratigraphically isolated, however all show lower proportions of group 1 fabrics than structures located to the immediate east and as such may be slightly later in date, S 9 shows the greatest variability. To the south E 4 / S 11 also contains relatively high group 7 to group 1 proportions. Structure 12 is stratigraphically isolated but contains only calcareous fabrics and as such may be slightly earlier.

Island 1

Four feature groups were excavated on Island 1. Structure 1 was located within E 2, the latter is truncated by S 2. Structure 3 is located to the north and does not have any stratigraphic relationships with the other features. Structure 2 contained the greatest ratio of sand to calcareous that has been seen on the site and as such may be the latest feature. Structure 3 also contains relatively high proportions of sandy fabrics (groups 7 and 8) to calcareous fabrics, and therefore may also be relatively late. The ratios seen in E 2 and S 1 are slightly curious, with the structure apparently containing earlier pottery than the enclosure that surrounds it. The mean sherd weight from Enclosure 2 was 6.4 g, much lower than the average assemblage weight of 11.6 g. A large proportion of the pottery came from the upper fill, and may represent some tertiary infilling from nearby later Structure 2.

Top of Page

Summary

The model of decreasing calcareous fabrics to increasing sandy fabrics has been applied throughout the region in discussions of phasing. Its application to the Claydon Pike assemblage has produced interesting results, suggesting a settlement shifting eastwards throughout the middle Iron Age, although a tighter chronology could not be defined. This is also reflected in the Malvernian fabrics and Droitwich Briquetage. Analysis of the fabric proportions within feature groups on each island indicates differences in the feature groups. Where stratigraphic relationships are recorded these concur with the fabrics, the exception is the relationship between linear boundaries 363 and 372 on Island 3. The fabrics can therefore be used to elucidate relationships and confirm stratigraphic observations, although absolute dates are not possible.

Top of Page

3.1.5 Form and function

Form and Fabric

to ascertain if specific fabrics were being selected for certain vessel types. The numbers of vessels present in each fabric are shown in Table 9.
Little variation is shown in the fabrics of each form group, perhaps not surprising in an assemblage dominated by coarse calcareous fabrics. The barrel-shaped vessels (form groups 1 and 3) are mostly constructed from a paste with abundant coarse calcareous inclusions, LT5 and LP5. Three vessels occur in a finer version of this fabric, LT3 and one in the very fine LT1. There are two examples of mixed calcareous and sand fabrics, LA5 and AS4. The ovoid jars are also dominated by the coarse calcareous fabrics, with one vessel of LT3 and one mixed fabric. This pattern in repeated in the expanded rim vessels (groups 6, 7 and 8), dominated by group 1 with the exception of one vessel in a mixed fabric (LA5); and also in the bead / rolled rim vessels of group 11, and the straight sided vessels of group 12.

Seventy five percent of barrel to globular shaped vessels with shaped necks and internal bevels (groups 4 and 5), were constructed from the coarse group 1 fabrics, however the remaining 25% were made from sandy fabric. Of the 14 bowls in the assemblage (groups 9 and 10) nine vessels were made from a coarse calcareous paste, and three in finer versions LT3 and LT1. Two vessels were found in a sandy fabric. Group 13 was represented by a single vessel constructed in the LT5 fabric. Group 15 is mostly found in coarse calcareous fabrics, but also includes a low-fired fossil shell ware (TV 4) and a sandy fabric (A5). The single Phase 2 vessel, form group 14, was unique in fabric (SL5)

Other sites in the region have recorded some correlation between fabric and form, particularly fine and sandy fabrics with globular forms. At Ashville 95% of globular forms were seen in fabric 2. The Watkins Farm sandy fabrics ‘were clearly deliberately chosen when making bowls’ (Allen 1990, 39).

Top of Page

Decoration

Four examples of decoration were recorded, accounting for only 0.1% of the middle Iron Age assemblage (by number and weight). Two of these were almost identical, consisting small impressed circles (possibly from a comb) creating two pairs of diagonal lines, which cross at approximately 70º. One example was recovered from context 22/Z/1 (FS 149), and one from 8/Q/1 (FS 198). The fabrics varied only slightly, the former is AI3, the latter IA3, both fabric group 7. Both were carefully smoothed on the exterior, FS 198 was also smoothed on parts of the interior. No parallels were identified at the sites studied, only Watkins Farm had an example of similar decoration, fig. 25.5 (Allen 1991).

The surface of bowl form FS 235 was broken by a tooled curvilinear decoration seen just under the rim. Again no parallels were located from the sites examined. Incised lines were recorded on a very small burnished sherd in B1 fabric (context 155/F/2). A much deeper (1mm) line had been incised horizontally across a thick smoothed sherd from context 386/B/1, fabric LT5. The vessels of form group 15 are all decorated with finger-tip or finger-nail impressions. In one case the decoration is seen along the top of the rim, with the impressions running diagonally (FS 010). In two examples the decoration spans the top of the rim and the upper exterior edge of the rim: FS 115 takes the form of slightly oval impressions, and FS 233 has finger-nail impressions. FS 089 and FS 100 have a band of fingertip decoration running around the shoulder area of the vessel. The remaining two examples, FS 068 and FS027 are very small and their detail uncertain, but both appear to have impressed circular decoration.

Finger impressed decoration was fairly commonly seen throughout the region, particularly in the earlier Iron Age. Of the 12 other sites examined in the region, middle Iron Age pottery decoration appeared limited to incised and tooled lines. Gravelly Guy contained the greatest number of examples with 462 sherds decorated, however this accounts for only 1.1% of the total assemblage. At Groundwell West four sherds were decorated (0.1%); 5 sherds at Uley Bury (0.9%); 8 sherds at Guiting Power (1.2%); 7 sherds (0.2%) at Mingies Ditch, and 2.2% of the Watkins Farm assemblage was decorated. Both Farmoor and Ashville had quite large proportions of finger-tipped decoration in their earlier phases, however this information is not quantified. Only one middle Iron Age sherd from Ashton Keynes was decorated.

Top of Page

Surface Treatment

The surface treatments applied to the Claydon Pike vessels include smoothing, wiping and burnishing. Haematite coating and tool trimming, seen at other sites in the region, were not recognised. The most popular treatment was external smoothing (abbreviated to SM), seen on 26.2% of the assemblage by weight, although this translates to only 12.8% of the total number of sherds. It was also noted occasionally on the upper interior of vessels. At Watkins Farm Allen (1991, 38) noted that ‘smoothing is not always a deliberate surface treatment, sometimes it results purely from pinching up the pots’. External wiping (WP) was found on 0.3% of the total number and weight of sherds, and just under 2% were burnished (BU). Evidence of surface treatment appeared to have been affected by post-excavation cleaning of the material. Surface treatments recorded within each fabric group are presented by percentage of total weight in Table 10.

The interior of a number of vessels appeared to have an almost waxy, dark greyish brown coating. It is not certain if this represents some kind of reaction between the contents of the vessel and the calcareous inclusions in the paste, or perhaps some kind of sealant. Work by Schiffer (1990, cited in Skibo 1992, 156) has shown that vessels ‘without an impermeable surface treatment have a much lower heating effectiveness and may be unable to boil water’. Skibo’s own ethnoarchaeological studies amongst the Kalinga in the Philippines found the effectiveness of the Kalinga vessels greatly reduced when the internal resin wore away. Once this occurred, pots were no longer used for their original purpose, but might be re-used for other processes such as roasting coffee beans (Skibo 1992, 162). This study indicates that internal wall porosity might reduce the effectiveness of cooking vessels, but the application of a sealant could negate this effect. The use or possible use of sealants does not appear to have been explored within the Upper Thames region. At the middle Neolithic site of Runnymede beeswax was noted on one of the 22 sherds sampled by Needham and Evans (1987), however the additional presence of glucose led the authors to suggest a honey constituent. As it was not recorded from the body of the vessel its use as a sealant has been ruled out (Needham and Evans 1987, 25).

Surface treatment and fabric

As noted above, the presence of surface treatments has been tabulated by fabric type and weight in Table 10. Any assessment of the percentages of fabrics showing surface treatments is slightly biased by the small representation of many of the fabric groups within the assemblage. However, some overall trends may be presented. Smoothing is seen in most fabric groups, however the highest percentages were seen on the mixed quartz and calcareous fabrics. Within the calcareous groups higher proportions of smoothing were noted on the finer fabrics. This smoothing usually took the form of a slurry finish, probably a simple case of the potter smoothing over the exterior of the vessel with wet hands. Wiping was exclusively seen in group 1, each example was the coarse LT5 fabric. Wiping may have been preferred to smoothing in some cases as this calcareous fabric was so coarse, containing sharp inclusions that might be more safely smoothed with organic matter or textiles rather than the potter’s hands. Burnishing is seen almost exclusively in the sandy fabrics and is a characteristic treatment of the Malvernian fabrics. It is rarely seen on calcareous fabrics, no doubt because of the difficulty to burnish these vessels.

Surface treatment and form

The correlation between surface treatment and form is shown in Table 11. The highest recorded instances of surface treatments are seen on the bowl forms (groups 9 and 10), here 12 out of 14 vessels (86%) are smoothed or burnished. Approximately one third of the barrel / globular vessels with shaped necks and internal bevel (groups 4 and 5) have smoothed or burnished surfaces. Similarly, 32% of barrel shaped vessels (groups 1 and 3) are also smoothed or wiped. Within the expanded rim group, 27% of vessels show smoothing or wiping, however the internally expanded group 7 has particularly low rates of surface treatment.

Top of Page

Surface treatment recorded from other sites in the region

Within the region a general correlation between burnishing and sandy fabrics can be seen, usually associated with globular vessels. At Ashville wiping or smoothing of vessels, particularly exteriors and rim areas, is carried out in all periods. Manufacturing marks such as finger depressions are often still preserved on vessels without such treatments. As sandy fabrics increase in use, so do the occurrences of burnishing, particularly on globular vessels (DeRoche 1978, 54-70). Coarse and fine sandy fabrics from Farmoor are more commonly burnished or smoothed than other vessels. A similar picture is presented from Watkins Farm where the finer sandy and alluvial fabrics are more commonly burnished than other fabrics, and a correlation between burnishing and globular vessels can be seen. Burnishing was recorded for 21% of the assemblage, mirrored by a high proportion of sandy fabrics (Allen 1990. The predominant technique at Gravelly Guy was also burnishing, although here accounting for 5.5%. Smoothing was rare, and seen on globular vessels where present (Duncan et al. forthcoming). At Mingies Ditch smoothing and wiping was recorded, but burnishing was not (Wilson 1993). At Guiting Power surface treatment was not found on coarse vessels, where recorded it was mostly on smaller vessels (Saville 1979). No evidence of an internal surface treatment designed to reduce permeability has been recognised at the other sites, with the exception of internal burnishing.

Top of Page

Use Alteration Traces

The term ‘use alteration’ was advocated in preference to ‘use-wear’ by Hally (1983) and Skibo (1992) and will be used here as it encompasses not only wear, but the addition of carbonised remains. The burnt residue (RS) seen on the interior of vessels ‘is caused, simply, by charring of food. Food residues that adhere to or are absorbed into a vessel surface become carbonised when the pot is heated to a sufficient temperature’ (Skibo 1992, 148). External sooting (SO) is caused by the accumulation of the by-products of fuel-combustion (Hally 1983, 7). Three predominant substances are found in the soot, ‘distilled resins, oxidised resins, and free carbon' (ibid.). Experiments by Hally found that two types of soot were produced when a vessel was exposed to fire. One consisted of solid carbon that could be removed by rubbing, the other, located further from the flames, consisted of distilled and oxidised resins and solid carbon that hardened on cooling and could not be removed (ibid.). Experiments by Skibo (1992) recorded similar results. On the vessels in Hally’s study, soot was most commonly seen on the upper portion of vessels, and at the point of maximum diameter.

Abrasion of the vessel surface will be influenced by a number of factors including the strength of the vessel, the nature of the surface and the presence of existing cracks and voids, the size and frequency of inclusions, surface treatments and overall shape (Skibo 1992, 108). Repeated stirring and scraping weakens the surface of a vessel and will eventually cause damage to the interior surface (AB). This will be acute in areas already weakened through poor firing (Hally 1983, 20). Vessels composed of calcareous fabrics can show pitting (PT) if they held acidic contents as this would have led to the dissolution of the inclusions. Pitting and abrasion were recorded only where they could be distinguished from post-depositional damage.

The relationship between use alteration traces and the different morphological form groups is shown in Table 12.

‘The presence of soot on a vessel is direct evidence that the vessel was positioned over an open fire during use. The absence of such deposits is direct evidence that a vessel was not so positioned during use’ (Hally 1983, 10). There are of course exceptions to the rule, such as the case of accidental sooting on a vessel placed close to a fire. The location of the soot may indicate the position of the vessel in relation to the fire. Sooting on the base shows that a vessel was suspended over the fire, the absence of soot from the lower wall and base indicates a vessel was stood in the fire ‘as resinous vapors will not condense on lower vessel walls where temperatures are high’ (ibid.).

Of the 38 vessel bases recovered from Claydon Pike there were no deposits of external soot, and only two cases of internal burnt residue. The incidences of use alteration traces on the bases are shown in Table 13. As discussed above, the lack of sooted deposits on the lower exterior of vessels may indicate that vessels were placed in the fire as opposed to being suspended over it. The rims of many of the vessels are plain and in-curving with little neck definition, and therefore might not be suited to suspension over a fire. As might be expected, evidence of abrasion caused by stirring and scraping, and pitting caused by attrition or acidic contents, was concentrated in the lower portions of vessels. Pitting would have occurred only on those areas of the vessel that were in contact with the contents, this is unlikely to be the uppermost portion of the vessel. In the case of abrasion, even in modern day cooking it is the lower portion of vessels that may suffer the worst damage from scrapping and stirring, and the base is the area where food is most likely to stick. The neck area may also exhibit abrasion. The pattern of pitting on a complete profile from form group 3 (FS 210) was particularly clear, as pitting was seen on the interior base and lower walls, whilst the upper interior was free from pitting. Of the remaining identified forms, pitting and abrasion was seen on one vessel in form group 4. Pitting with sooting was recognised on one vessel in group 4 and one in group 15.

Sooting and burnt residues were recorded on 214 of the 1851 records on the database, a total of 11.6% of all records, 13.5% of the total number of sherds and 20.1% of the weight. Count of number and weight appear to be biased by the occurrences of complete vessels with sooted or burnt residues which inflate the figures, and the number of records is the most reliable measure in this case.

Actual Vessel Use

A number of observations of the actual use of vessels can thus be made, although form groups containing fewer than five vessels may be unreliable for such conclusions. Group 5 vessels, forms with shaped necks, internally beveled rims and barrel-shaped profiles, all showed evidence of use in cooking or heating. Their more globular-shaped counterparts, group 4, indicate this use on a third of cases. The plain barrel shaped vessels (group 1) are often used for cooking, with evidence on 53% of these vessels, those with more shaped rims (group 3) had only slightly less evidence, with sooting or residue adhering to 44% of the vessels. The ovoid jars (group 2) were also sometimes used for this purpose. The expanded rim vessels were seldom used for cooking: groups 6 (externally expanded) and 8 (internally and externally expanded) were never used for this purpose, 20% of internally expanded vessels (group 7) did exhibit evidence of cooking. The bowls were rarely used for cooking, in group 9 the only example (accounting for 11%) is the bowl with curvilinear decoration, 20% of group 10 are sooted. Within the remaining classes of rolled or beaded rims (group 11) and straight sided vessels (group 12) this form of evidence is seldom seen. No conclusions may be drawn about the single vessels in groups 13 and 14. The criteria for inclusion in group 15 is fingertip decoration, however the example with evidence of sooting would appear to be a closed barrel-shaped form.

Use alteration traces in the region

Recording of sooting and burnt residues at other sites in the region is variable, with many of the earlier reports making no mention of traces of use. At Groundwell Farm (Gingell 1982), burnt residues were most commonly seen on vessels made from the coarser calcareous and sandy fabrics, little evidence of such residues were seen at Groundwell West (Timby 2001). Of the 366 sherds from Deer Park Road (Timby 1995), only two had evidence of use in cooking / heating. At Thornhill Farm (Timby 2004) external sooting or internal burnt residue was commonly seen on vessels in the coarse calcareous fabric, the same was recognised at Gravelly Guy (Duncan et al. forthcoming) where residues occurred most commonly on globular vessels, and less so on those with barrel-shaped or straight-sided profiles. The Ashton Keynes (Edgeley-Long unpublished) coarse calcareous fabric was most often used for cooking vessels, with traces of use seen most commonly on the barrel-shaped vessels. Pitting is recorded on 532 sherds in the calcareous fabrics at Gravelly Guy.

Top of Page

Conclusions

The calcareous group 1 fabrics dominate each form group at Claydon Pike and allow little discussion of correlation between fabric and form. However, it can be shown that the barrel / globular vessels with shaped necks and internal bevel (groups 4 and 5), and to a lesser extent the bowl forms, are also associated with sandy, finer fabrics. These fabrics tend to appear slightly later than the coarse calcareous fabrics, and are seen on sites throughout the Upper Thames region during the middle Iron Age. This change in fabric is usually associated with the adoption of more globular vessels, often burnished. Smoothing is widespread throughout the assemblage and is commonly seen in middle Iron Age assemblages in the region. It would appear to represent a simple ‘finishing’ of the vessel, however it may also have performed a more technical function. At Claydon Pike burnishing is almost exclusively seen on sandy fabrics, where found on identified form types these are mostly bowls. Burnishing is quite labour intensive and is rather decorative. The association between burnishing and bowls may be indicative of serving vessels which are meant to be seen, and possibly used in the sharing of food at social occasions.

The barrel-shaped vessels (groups 1 and 3), ovoids (group 2), and vessels with shaped necks and internal bevels (groups 4 and 5) all appear to be fairly frequently used in cooking and heating activities. The rolled / beaded rimmed and straight-sided vessels are occasionally used for cooking. The vessels with expanded rims appear to have been used for purposes other than heating. The bases present in the assemblage were never associated with sooting, suggesting that vessels may have been placed directly in the fire during cooking, rather than suspended over it when soot could have adhered to the base. Pitting and abrasion was rarely found on the upper portions of vessels, but became more common on the lower internal wall and base of vessels.

Top of Page

3.1.6 Vessel size

During recording it was observed that many of the sherds appeared to be quite thick, in contrast to a previously recorded middle to late Iron Age assemblage from Ashton Keynes, c 20 km to the south west. The range of rim diameters was also in stark contrast, the largest seen at Ashton Keynes was 22 cm, but at Claydon Pike diameters up to 38 cm were recorded. Woodward stated that ‘ethnographic studies (e.g. Miller 1985) have demonstrated that vessel types primarily designed for specific functions are characterised by closely defined size parameters’ (Woodward 1997, 27).

Thickness

The thickness of body sherds was recorded using a numerical code for 2mm ranges. These codes follow Elaine Morris’s system of recording, where 1 = <5mm; 2 = 5-<7mm; 3 = 7-<9mm; 4 = 9-<11mm; 5 = 11-<13mm; 6 = 13-<15mm; 7 = 15-<17mm; 8 = 17-<19mm; 9 = >19mm. Sherds where thickness was ambiguous, or one or both surfaces were missing, were recorded using code 0. The total numbers and weights for each thickness code are presented in Table 14.

The mean sherd weights for the thickness categories indicate that thicker vessels break into heavier, and probably larger sherds than thinner vessels. Overall the most commonly seen thickness is code 4 (9-<11 mm), accounting for 24.6% of the total number of sherds and 30.7% of the weight. This is followed by code 3 (7-<9 mm) and code 5 (11-<13mm). The heavier sherds in the code 5 category led to variation in the number and weight, with code 3 accounting for 24% of the number and 15.3% of the weight, and code 5 representing 12.9% of the total number and 21% of the total weight. The total number and weight per thickness code for each fabric group were recorded, however no relationships were revealed. All fabrics were most commonly seen in thickness code 3 or 4, usually followed by code 5. It would therefore appear that a certain fabric was not used to create a designated wall thickness, this instead might be related to overall vessel size.

Top of Page

Rim Diameter

Rim diameters ranged from 10 cm to 38 cm. The most commonly seen sizes are vessels with diameters ranging from 14-18 cm. However, there is another peak at 26 cm and a small number of very large vessels, 32-38 cm diameter. The varying sizes are thought to relate to varying categories of use. A number of ethnoarchaeological researchers have recorded vessels of a single class made in different sizes for different purposes, such as Longacre’s study of Kalinga pottery in the Philippines (Longacre 1991). Hally’s (1983, 55) work on aboriginal (AD 1550-1770) pottery from north-west Georgia, USA, revealed several distinct classes of vessel, and variations in size within these categories. He recorded two vessel types ‘identical in all physical characteristics except size’ the smaller ones were used in cooking / heating, the larger one was not (Hally 1983, 22).

The group 1 vessels range steadily from 10 cm diameter to 24 cm, with a peak at 18 cm. For this discussion small is described as <14 cm, medium 16-20 cm, large 22-26 cm and very large as >28 cm. Group 1 vessels may therefore be described as small to large. The group 2 ovoids are mostly small, ranging from 10-14 cm, with a further peak at 24 cm. Within group 3 most vessels could be described as small to medium (10-16 cm), however there are small numbers of vessels present in much larger sizes, at 22 cm, 28 cm and 36 cm diameter. Group 4 vessels show a fairly steady range up to 20 cm, peaking at 14 cm, however larger vessels are again occasionally seen, at 24 cm and 28 cm. Group 5 appears to divide into small and medium/large vessels. The expanded rim vessels show a slightly different picture to the rims so far recorded. Those expanded on the exterior (group 6) have no really small vessels, although they do occur in 14 cm diameter examples. A range is seen from 14-22 cm, with a peak at 26 cm. The internally expanded vessels (group 7) are only found in large sizes, 26-28 cm, and those expanded internally and externally (group 8) again have no small vessels, diameters range from 18-32 cm. The bowl forms (groups 9 and 10) are almost exclusively small to medium, 14-18 cm diameter, with a larger size of 26 cm also seen in group 9. Group 11 diameters are mostly small, 10-14 cm, with a larger example seen at 26 cm; group 12 rims vary from 14-20 cm. No conclusions can be drawn about groups 13 and 14 as they are represented by a single vessel.

Within each group the presence of use-alteration has been recorded against the diameters to ascertain if the patterns noticed by Hally (above) and other authors are present at Claydon Pike. In the majority of groups displaying sooting or burnt residues these traces were usually seen on vessels with diameters of 20 cm or less. In six groups a clear distinction was thus seen with all cases of cooking / heating evidence occurring only on vessels with diameters of 20 cm or less. Larger vessels in these groups did not show evidence of this form of use. Sooting was seen on larger vessels in group 7 (26 cm) and group 13 (38 cm), however in each of these groups smaller vessels were not present.

A comparison of thickness codes for each of the diameter classes within the groups was made. On the whole smaller diameters correlated with thinner walls, and most vessels 14 cm and less had a wall thickness of 7-<11 mm. Medium sized vessels with a diameter of 16-20 cm varied from 7-<13 mm in thickness. Larger vessels, 22-26 cm diameter, showed a good deal of variation, from code 4 to 7, (9-<17 mm), the most common thickness in this class was code 5 (11-<13 mm). Very large vessels, 28-38 cm in diameter were also very variable in thickness, from a code 3 to 7, the thinner walled examples were from the expanded rim group. Overall there is a good deal of variability and no clear trends between thickness, vessel diameter and form group were noted, other than the very basic generalisation presented above.

A comparison of vessel size with other sites in the region was not carried out. This is because diameter and thickness are not quantified in most of the published reports (other than the material selected for illustration), and it was beyond the scope of this report to go back to the original vessels and paper archives to gather data. However, it was possible to plot the diameters of vessels from Ashton Keynes (author’s data), Thornhill Farm (information from J. Timby), and Watkins Farm (histogram published in Allen 1990). This indicated Claydon Pike did have a wide range of vessel sizes, in stark contrast to the small to medium vessels that dominate the Thornhill Farm and Ashton Keynes assemblages.

Top of Page

Vessel Capacity

Only five complete profiles could be reconstructed. The shapes of these vessels are very simple, and in all but one the mouth is of similar dimensions to the height, and access to the vessel is unrestricted. Their capacities have been calculated as if they were cylindrical in shape, using the formula V=p r2h. This is will not give a precise measurement as the vessels are not completely cylindrical, but will offer a good estimate. A more accurate method is described by Rice (1987, 221-222), whereby the vessel is divided into horizontal slices, each one a cylinder. The volume of each cylinder / slice is calculated using the formula above and the results summed to give the total number of cubic centimeters. In the capacity calculations the internal height and radius have been used so as not to include the wall thickness, the results are shown in Table 15.

Woodward’s analysis of vessel size at Cadbury Castle recorded the existence of small, medium and large size vessels in each phase. Height variations indicated three classes: low vessels (40-130mm), intermediate vessels (140-190mm) and tall vessels (200mm and more). The vessels in Table 15 fit into the ‘tall’ category, with the exception of FS 246, which is ‘low’ (Woodward 1997, 29).

The largest vessels found at Claydon Pike are FS 060 (36cm diameter) and FS 125 (38cm diameter). Both showed evidence of use in cooking or heating processes, FS 060 contained a small amount of burnt residue, and FS 125 had a ring of soot around the widest part of the vessel. Such large vessels, the latter with a capacity of 30 litres, suggest the preparation of food for a communal meal (see section 3.1.7).

Top of Page

Summary

This analysis has revealed that a wide range of vessel sizes were used at Claydon Pike, from quite small pots, 10 cm diameter, to the very large vessels, maximum diameter 38 cm. Vessel wall thickness was variable, most commonly ranging from 7-11 mm, although thicker walled examples are also seen. The wall thickness appeared to be associated with the overall vessel size, with thicker walls seen on larger vessels, partly because of the wall strength required to support large vessels during the drying process. A certain amount of control over this problem is shown by three of the expanded rim vessels with diameters 26-28 cm, and a wall thickness of 7-9 mm.

Vessels in several of the form groups appeared to be made in a number of size ranges. The barrel-shaped vessels (groups 1 and 3), ovoids (group 2) and those with shaped necks and internal bevels (groups 4 and 5) were mostly manufactured in small to medium sizes, however larger vessels are also seen in these groups. Observations of use alteration have shown that the small to medium vessels (<20 cm diameter) are quite often associated with sooting and burnt residues, whilst larger vessels within these groups are not. This suggests that a single vessel form might be used for more than one purpose, and this is influenced by the size of the vessel. The expanded rim vessels were rarely made in the smaller sizes, and a peak in diameter is seen at 26 cm. They were rarely used for cooking, although their function is unknown, and the expanded rim may have served a specific technical purpose. The bowl forms (groups 9 and 10) are mostly 14-18 cm diameter and would therefore suggest individual serving vessels.

Most of the vessels that could be related to cooking or serving represent individual use, however the two very large vessels with evidence of sooting or burnt residues suggest food preparation and therefore consumption on a much larger communal scale. This is clearly shown in the capacities of the vessels, with a small low vessel (13 cm diameter) able to hold one litre, a medium tall vessel (16 cm diameter) able to hold nearly four litres, but the very large 38 cm diameter vessel able to hold approximately 30 litres when full.

Top of Page

 

3.1.7 Discussion and conclusions

The pottery from the middle Iron Age occupation phases at Claydon Pike has been analysed to try and elucidate the chronology of the settlement and identify if the three gravel islands were occupied simultaneously, or if the settlement shifted over time. This aim was assessed using a spatial consideration of the fabrics and forms in use. The second aim was to discover the scale of the production of the ceramic vessels. Thirdly, it was hoped that the intended and actual use of the vessels might be revealed through an assessment of the morphological variables and traces of use alteration. Finally, an attempt was made to try and reveal the social role that pottery might have played at Claydon Pike, by assessing the evidence for non-local pottery and feasting.

 

Shifting Settlement and Social Patterns: The Evidence from Changing Fabrics

The use of calcareous fabrics during the early Iron Age and earlier part of the middle Iron Age is well attested throughout the Upper Thames Valley and is seen as a chronological marker. Throughout the middle Iron Age the use of sandy fabrics increases, and this is often associated with new globular forms (at Ashville, Abingdon for example). This shift from calcareous to sandy fabrics is evidenced at Claydon Pike and has been used to assess the first aim of this project: were the three gravel islands occupied simultaneously or is a pattern of shifting settlement evidenced? Although the differing proportions of calcareous to sandy fabrics is not as pronounced as is seen on some of the other sites such as Ashville, it is still present. Comparison of the percentages of the two fabric groups revealed that the most westerly island (Island 3) was probably the first to be occupied, with the settlement then shifting eastwards, first to Island 2 and then Island 1. This pattern was also recognised in the Malvernian wares and Droitwich briquetage (Morris this section). The different forms present also appeared to follow the pattern, with the chronologically earlier finger-tip decoration seen only on Islands 3 and 2, and the low numbers of expanded rim vessels recovered from Island 1. Within each gravel island a number of relationships could be established, these were for the most part confirmed by the stratigraphic relationships. Whilst answering one question this trend towards an increasing use of sandy and mixed (calcareous and sand) fabrics created another: why were potters changing fabrics?

There may be technical reasons for such a change, sandy fabrics may have been found to be slightly easier to work than the calcareous fabrics, or easier to fire. However, quartz inclusions are technically ‘detrimental in low-fired cooking vessels’ because silica ‘undergoes a crystalline inversion at 573ºC’ (Rye 1981, 34), although the ubiquity of the use of quartz in ceramics throughout British history suggests the problem was not a major one. The change might be related to style, as these sandy vessels are often burnished, a finish that is difficult, although not impossible to achieve on the coarser calcareous fabrics. This in turn may have social connotations with a desire to own more carefully finished and possibly aesthetically pleasing vessels, particularly as many of the new vessels throughout the region are globular bowls and therefore may have been used as serving vessels in socially charged occasions. At Claydon Pike the shift is most clearly seen in the globular vessels with shaped necks (group 4) and the bowl forms. These forms quite often exhibit surface treatments, the bowls are rarely used for cooking activities, however sooting or burnt residues were discovered on one third of the globular vessels.

The simultaneous use of local sandy and non local glauconitic fabrics for the same vessel forms may also offer clues to the social role of pottery at this time. As stated in section 3.1.2, Morris (1997) questioned why glauconitic fabrics were imported at Danbury when local sandy fabrics were also in use. It may be that the glauconitic pots were in some way technically superior, or that the pots represent the maintenance of exchange networks (Morris 1997, 38).

Top of Page

Production

The Raw Materials

The pottery in use during the middle Iron Age phase of occupation at Claydon Pike was mostly made from a fabric paste containing abundant inclusions of coarse limestone and fossil shell. This group of fabrics accounted for 91.2% of the total weight of the assemblage, 85% of this was represented by a single fabric, LT5. This fabric was distinctive, containing small fragments of bryozoa that pointed to a local origin from the Jurassic strata. The ingredients for such a paste were available locally in the Oxford Clay, Cornbrash and Forest Marble deposits.

Modes of production

The ceramic vessels recovered from the middle Iron Age occupation at Claydon Pike were all handmade. Coiling was clearly evidenced on 73 out of 1851 records, in these cases the coil join often provided a weak point on the vessel prone to breakage. No investment in equipment was identified at the site, or specific areas associated with potting. It is therefore proposed that the manufacturing of vessels was carried out on a part-time domestic basis, at the level of household production (after Peacock 1982). A lack of drying facilities and kilns suggest that production was very much affected by seasonality, and would have to be scheduled to avoid conflicting with subsistence activities (Arnold 1985). The weather would have affected even the gathering of raw materials, for example the clay collection would not have taken place in winter as the ground would have been too cold and hard. Early spring and late autumn are optimal times as the deposits are not obscured by herbaceous plants (DeRoche 1997, 21).

Ethnographic studies indicate that at a household level pottery is made and organised by women (Skibo 1995, 83; Hill 2002, 83). Gender issues are almost invisible in archaeology, yet this is one area where they may be studied. A cross-cultural study of 185 non-industrial societies by Murdock and Provost in 1973 (in Skibo 1995, 86) looked at female participation in non-industrial societies. They found that the only activities with higher levels of female participation than the manufacture of pottery include cooking, vegetable preparation and fetching water, activities that also involve ceramics. In this respect pottery is made by women for women to use in subsistence strategies, but at a social level may be used as a tool by both sexes. Skibo triumphs the technological achievements of cooking vessels, labeling it ‘women’s technology’.

Intended vessel use

‘The production of every pot requires the potter to make a series of ‘choices’ selecting from a range of possible raw materials, tools, energy sources, and techniques. In this sense every pot is the unique result of a series of choices between alternative techniques’ (Sillar and Tite 2000, 3).

As has been shown, the vessels from Claydon Pike were mostly manufactured from local materials. The choices made in the selection of raw materials offer an insight into the third aim of the project, what was the intended use of the vessels?

The uses of inclusions

The inclusions in a clay are useful for several reasons, particularly as clay is not a good thermal conductor, although it is a good insulator (Sinopoli 1991, 15). During the forming of the vessel they can help reduce excessive plasticity and improve workability. The presence of inclusions helps prevent shrinkage during drying and thus allows the fabric to dry without warping or cracking. During firing the inclusions can open up the structure of the clay and allow trapped water to escape as steam, rather than spalling. They can also help to halt cracks that have already started, and stop them from spreading (Braun 1983, 122-3; Sinopoli 1991, 15).

Thermal and mechanical shock resistance

A vessel’s ability to withstand mechanical shock through use such as pounding and grinding, as well as being knocked or moved around, is increased with hardness and strength, properties that can be manipulated by the composition of the paste, and the firing temperatures (Rice 1987, 228).

If the vessel is used for cooking or heating activities it can be exposed to thermal stress. This is usually caused by one part of the vessel (eg the exterior) heating up faster than another part (such as the interior). One of the most effective ways to counteract the problems caused by thermal stress is to use a fabric ‘recipe’ that contains inclusions with a similar thermal expansion rate to the clay. Calcium carbonate (CaCO3), which occurs in shell and limestone, is ideal in this respect. However, during firing the temperature must be kept below 750º as above this the calcium carbonate becomes unstable and will decompose (Rye 1981, 33). This can be averted by the addition of sodium chloride, however Woods (1986, 186) has noted that there is very little evidence for the addition of salt to calcareous ceramics in Britain.

The calcareous fabrics in use at Claydon Pike would therefore appear to be quite well suited to a use as cooking pots. The presence of sooting and burnt residues on many of these vessels certainly supports their use as cooking pots, although not exclusively so. However, these fabrics were also widely available locally and opinion is divided as to whether potters made a conscious decision to use pastes with thermal properties, or simply chose clays from locally available sources. Woods argues that there is ‘little evidence that potters in ancient Britain were aware of the problems of thermal shock and the potential properties of their raw materials in relation to it’ (Woods 1996, 168). Whilst accepting that the presence of inclusions in a vessel is important to prevent spalling, Woods (1996, 169) states that if a vessel was able to survive firing then it would probably be able to withstand the heat used for cooking. However, she adds that if a particular paste performed badly then it is unlikely that a potter would continue to use it. Braun also states that when pottery is a domestic craft, those making the pottery are also those who are using it and will therefore be ‘exposed to evidence of the success or failure of their vessels in use’ (Braun 1983, 112). Woods concludes that the important factor is not the choice of inclusion, but the coarseness of the pot and therefore the quantity of inclusions (Woods 1996, 170). Rice argues that because a fabric has certain thermal properties does not necessarily mean this was intended, ‘correlation does not prove cause’ (Rice 1996, 140).

The coarse calcareous fabrics at Claydon Pike would therefore have been advantageous in the construction of vessels that would be subjected to thermal and mechanical shock. However their use also in bowl forms which require neither of these properties, and instead show a preference for smooth surfaces, might suggest that the primary reason for using these fabrics is access to resources. The increasing preference for sandy fabrics in the later part of the settlement, and the apparent importation of glauconitic vessels, suggest that proximity was not the predominant influence in the choice of raw materials, at least in the later occupation. Authors such as Rice (1996) and Sillar and Tite (2000) have warned against trying to follow singular lines of enquiry, and the technical choices made by potters were done so ‘in a rich context of tradition, values, alternatives, and compromises’ (Rice 1996, 140). A wide range of factors will affect the choice of raw materials including the local environment, and the potter’s physical ability to extract the clay. The choice is further affected by the potter’s perception of the clay source as suitable, and may be affected by political concerns such as the control and ownership of resources (Sillar and Tite 2000, 7).

Form: Pots as containers

Most pots are implements: they are made to be used as containers’ (Braun 1983, 107).

A ceramic vessel may be designed for a wide range of uses, these may be divided into three main categories: transfer, processing and storage (Rice 1987). The potter may therefore consider whether the vessel will be required to hold liquid or dry goods, be used over heat, how robust it will need to be, the volume it will need to contain, and the weight once full. This is not to say that a family would have a pot for every occasion, there are of course vessels that would have multiple uses, as well as occasions of re-use. The vessel properties that may be considered by the potter include permeability (of importance for water vessels, but detrimental for cooking vessels); toughness and strength (for processing foods, transport or storage) and thermal shock resistance for cooking vessels (Sillar and Tite 2000, 8). ‘Each category of vessel use requires a different combination of attributes of form and composition to achieve a product that meets its special needs’ (Rice 1987, 209). Attributes of form include the size of opening and whether this is restricted or unrestricted, this in turn relates to accessibility of contents, stability and centre of gravity, and the weight and volume once full (Sinopoli 1991, 84).

The majority of vessel forms identified at Claydon Pike were quite open in terms of access to the contents, and were mostly quite squat. Evidence of neck constriction was recorded on forms 3, 4, 5, 6, 8, 9, 11, 12, 13 and 14, but in most cases it was quite subtle. Rice notes that vessels designed for boiling ‘are likely to have a relatively open orifice for adding or removing food, but a slight constriction or a low neck helps prevent boiling over and reduces evaporation’ (Rice 1987, 239-40). She adds that a survey of the ethnographic data (by Henrickson and McDonald 1983) found that they are often quite short and squat (ibid.), traits seen in the Claydon Pike vessels. In a number of cases the neck constriction may have been pronounced enough to tie a cord around a vessel intended to be suspended, for example over a fire. Vessels with internal bevels, such as groups 4 and 5, may have been designed to be used with a lid or another method of sealing the vessel. The larger, thicker walled vessels in the assemblage may have been intended as storage vessels.

The open and fairly shallow forms of groups 9 and 10 appear well suited to a use as serving vessels, their size mostly suggests individual serving vessels. Although mostly seen in the coarser calcareous fabric, 21% are in finer calcareous fabrics, and 15% in the sandy fabrics. The highest recorded instances of surface treatment (86%) are seen on these bowl forms, this includes the greatest proportions of burnishing from the assemblage. Surface treatment may serve at least two purposes here, to reduce permeability and to signify the social aspect of food, bowls that were meant to be seen (Rice 1987, 240). One bowl of 26 cm may have been used as a communal serving bowl. As is the case with the selection of raw materials, choices of morphological variables are not just made on technical merit but are a response to social concerns and expectations.

Thickness

The thickness of a vessel affects its ability to conduct heat, as well as to resist thermal and mechanical shock. As wall thickness increases, so too does the ability of the vessel to resist mechanical shock; as wall thickness decreases the ability to resist thermal shock and conduct heat increases. Therefore, thinner walled vessels make better cooking pots, and thicker walled vessels are more suited to storage vessels, or those used for processing of food stuffs (such as pounding) as they are more likely to withstand mechanical shock (Braun 1983, 118-19). However, the picture is complicated by the size of the vessel, as large vessels will require strong and therefore relatively thick walls to support the vessel during the drying process. Compromises therefore need to be made, particularly when making large cooking vessels. This can sometimes be executed through the manipulation of the inclusions. The thickness of vessels at Claydon Pike was variable but was most commonly within a range of 7-11 mm. No overall trends between thickness and fabric, or thickness and form were found, other than a basic relationship between increasing vessel size to increasing thickness. However the potters at Claydon Pike appear to have been able to combat some of the forming problems that result from making thinner walled large vessels, as can be seen in form group 7 where a 28 cm diameter vessel had walls 7-9 mm thick. Larger, thicker walled vessels were still used for cooking, as can be seen from the evidence of the single group 13 vessel, 38 cm diameter and 11-15 mm thick, sooted around the shoulder area.

Top of Page

Actual vessel use

Recording of the presence of external soot, internal burnt residues, pitting and abrasion, has added to the picture and indicated the actual use of vessels.

Vessels use to cook or heat

The presence of external soot or internal burnt residues indicates that a vessel was probably used for cooking or heating, at least at some point in its life. At Claydon Pike evidence of cooking was commonly seen on the barrel-shaped vessels, including form 1, with plain, in-curving rim; form 3 with a slightly constricted neck area; and form 5 with shaped neck and internal bevel. The more globular profiled version of form 5 (form 4) and the ovoid jars (group 2) were also used in this capacity. These vessels were mostly constructed from a fabric containing abundant calcareous inclusions, although sandy fabrics were also used for just under a third of the form group 4 vessels.

The form of the vessels used for cooking is variable. Forms 2 and 4 had quite rounded profiles, which may have increased the vessel’s ability to conduct heat, whilst the barrel-shaped profiles of forms 1, 3 and 5 are much more slack. This may be a factor of chronology, as forms 2 and 4 appear to become more popular with time, possibly in response to an increased concern with thermal performance. The vessels used for cooking had mostly quite open mouths and only slightly restricted orifices. Approximately one third were smoothed on the exterior. Analysis of rim diameters showed that traces of use for cooking were almost exclusively restricted to vessels of 20 cm diameter and less. Larger vessels in each of these form groups did not appear to have been used for cooking. The complete lack of sooting on the base or lower portions of vessels suggests that vessels were placed directly in the fire for cooking (after Hally 1983).

Serving Vessels

The form, size and surface treatments that characterise the bowls of groups 9 and 10 suggested they were intended to be used as serving vessels. This form is rarely found with traces of sooting, and no evidence of pitting or abrasion was recorded, the archaeological record would therefore appear to support that this was their actual use.

Top of Page

Vessels of undetermined function

The expanded rim vessels occur almost exclusively in coarse calcareous fabrics. Most showed slightly restricted orifices. Surface treatments, namely smoothing and wiping, were recorded from 27% of these vessels, with the externally expanded group 6 showing the lowest occurrences of surface treatment. No evidence of soot or residues was found on the group 6 or 8 vessels, however two out of the ten group 7 vessels did appear to have been used for cooking, and a further vessel had a pitted interior, possibly indicating abrasion or acidic contents. Like the bowls, the expanded rim vessels did not include any particular small vessels. Group 6 showed a steady range from 14-22 cm, with a peak at 26 cm. Group 7 was quite unique in the assemblage, exclusively found in larger sizes, 26-28 cm diameter. Group 8 ranged from 18-32 cm in diameter. The use of these vessels is unknown. The distinctive expanded rim is seen as a chronological indicator in the region, usually (although not exclusively) occurring during the earlier part of the period. The rim may have a specific range of functions that it is useful for, however no clues have been found as to what this might be. As a group they are rarely used for cooking and tend to occur in larger sizes. Possible uses might include communal serving vessels (although access is slightly restricted by the rim form), storage or transportation.

A similar problem was encountered in understanding the function of vessels in form groups 11 (straight-sided) and group 12 (rolled rims). This is particularly evident because of the low numbers of examples in each group, six for group 11 and four for group 12. They are mostly found in the coarse calcareous fabrics and show relatively little evidence of surface treatment (although two of the group 12 vessels were smoothed on the exterior), and each group contains a single sooted example. The group 11 vessels were mostly small, with one larger (26 cm) example; group 12 ranged from 14-20 cm in diameter. There are therefore few clear indications as to their actual use.

Top of Page

The Social Role of Pottery

The final aim of the analysis was to discover the social role of pottery. The manufacture of pottery will be affected both directly and indirectly by the social and economic context. This is seen in the control of resources, level of technology, scheduling of the stages of manufacture, and also the social situations in which vessels are used. Social concerns may have been one of the factors that prompted the widescale adoption of sandy fabrics throughout the region. These fabrics are associated with much more rounded vessel forms that may have been found preferable due to their greater ability to conduct heat, however they are also used for the well-finished bowls. The advantages of being able to burnish such fabrics to a greater finish than the calcareous fabrics may have had techno-function (after Skibo 1992) advantages such as the reduction of permeability, but may also have performed a socio-function due to the increased decorative effect. A change in fabrics, and a greater use of bowl and globular forms, may therefore indicate a greater emphasis on serving vessels and elaboration of subsistence traditions. The small amounts of Malvernian pottery recovered and the presence of glauconitic sandy fabrics, indicate that vessels may have been brought in from up to 14-23 km away, and approximately 65 km for the Malvernian vessels. The use of non-local fabrics when similar local versions are available has led Morris (1997) to suggest that this may signify ‘the maintenance of exchange networks’ (ibid. 38).

Feasting: Food for thought

Large vessels of 22 cm diameter and greater were seen in most form groups. As noted above these were rarely associated with evidence of cooking. Very large vessels (< 28 cm diameter) were recorded from form groups 3, 4, 7, 8 and 13. Slight smoothing was seen on the exterior of one vessel in group 3, and the exterior and interior of a vessel from group 8 had been given a slurry finish. Minor internal pitting was recorded on the group 7 vessel. Vessels of this size without evidence of use may have been used for storage or some other function that is unlikely to leave a visible trace. However, a small amount of internal burnt residue on the 36 cm diameter vessel from group 3 (FS 125), and sooting observed around the widest part of the 38 cm diameter vessel in group 13 (FS 060), suggest that the latter two vessels may have been used for cooking. If this is the case, their large size would suggest the preparation of a communal meal or feast, with FS 125 able to hold around 30 litres of liquid when full.

In the earlier prehistoric period in Britain ceramic ‘feasting sets’ have been found on a number of occasions. Woodward (after Cleal 1990) presents the Neolithic site at Coneybury, near Stonehenge, where a pit contained a very large carinated vessel, ‘several open or neutral uncarinated bowls, moderate in size…small shallow bowls or cups…one small neutral bowl with neck…and one small closed bowl or jar’. The pit also contained a deposit of animal bone and was interpreted by Cleal as ‘the debris from a single major feast, or a period of feasting’ (Woodward 1999, 4). A further ‘feasting set’ deposit has been identified by Woodward in a late Bronze Age pit at Alcester, and evidence has also been recorded for Southern England (Woodward 1999, 6). ‘Thus in the Neolithic and later Bronze Age alike, ceramic feasting sets abound’ (ibid). In the middle Iron Age pottery becomes far more common, ‘and for the first time there is a wider range of vessels possibly used in cooking’ (Woodward 1999, 8).

The analysis of the fabric, form and use-alteration of ceramic vessels is an attempt to understand how they were used in the daily life of prehistoric peoples. It can help reconstruct past behaviour and offer clues as to the social role of pottery. These insights are often triggered by ethnographical and ethnoarchaeological studies that illustrate archaeologically invisible or unrecognised processes. One such case is that of feasting, a subject which is receiving more and more attention in the literature, but until recently had been very much neglected. Dietler and Hayden (2001, 2) argue that ‘feasts are an extremely significant aspect of social life on a worldwide scale, and that understanding them is crucial for apprehending and comprehending many social and cultural processes’.

Various definitions may be put forward for the term ‘feast’, however most authors seem to agree that it involves communal eating and / or drinking, typically an unusual meal to mark an unusual occasion (Dietler and Hayden 2001, 3-4). Clarke (2001, 145) describes feasts as ‘ritualized’ and states that ‘the meal is not eaten solely for subsistence’. Having defined a feast, why is it important? Hayden believes that feasts, along with gift-giving, kinship and language are a major component of what he terms ‘social technology’. This is defined as ‘the creation and maintenance of social relationships that are predicated on securing access to resources, labor, or security’ (Hayden 2001, 26).

Field study by Michael Clarke amongst the Akha of Northland Thailand offers a case in point. Akha society is mostly egaliterian, however there are various positions of administration, religious leadership, plus a group of elders. Persons in such positions not only give feasts to confirm their roles, ‘very often based on proof of spiritual potency which is demonstrated through economic success’, but are also invited to feasts to win or renew favour (Clarke 2001, 145-7). A lack of long-term economic security has created a need for ‘a structured and dependable life-crisis support network’. This works not only for times of crisis when crops might have failed, but also to maintain relationships so that co-operatives might be formed at certain key points in the agricultural cycle (ibid. 148-9). Clarke identified two levels of feasting. The first is seen within the clan and lineage and involves ‘alliance formation and mutual assistance’. The feast is non competitive and may focus on activities such as ancestor veneration, name-giving or sickness curing (ibid. 149). Competition for labour, spouses, political control and so forth are dealt with between clans and lineages and take the form of larger feasts, such as weddings and funerals, and are usually more impressive displays of wealth (ibid. 150).

The value of Clarke’s study is that it not only characterises the relationships between feasts and the societies that create them, but it specifically records the ways these feasts might be visible in the archaeological record. The range of vessels required for a feast is considerably larger than those required for daily life. Families are unlikely to own the number of serving vessels required for a large feast and will normally have to borrow, but families involved in giving feasts on a regular basis are more likely to own more than the average number of serving vessels. Clarke suggests that the occurrence of feasting can be seen in the presence of ‘multimodal size distribution of cooking vessels in large assemblages’ (ibid. 159). Households not hosting feasts might be expected to have a more unimodal distribution of sizes (Hayden 2001, 49). Skibo has further recorded that amongst the Kalinga, in the village of Guina-ang, pottery is divided according to everyday and ceremonial use, and the vessels for each are stored in separate locations (Skibo 1992, 66-7).

The pottery from Claydon Pike has suggested a shifting settlement pattern, and the stratigraphy suggests a maximum of four house site in use at any one time, with only one in use during some periods. It would therefore appear that the assemblage represents the ceramic debris of a single extended family group for each generation. Within this assemblage there does appear to be a multimodal distribution of vessels within the form groups, although for the most part these do not show any evidence of use. However the unusual size of two of the vessels associated with sooting and burnt residues may be evidence of a communal meal, which may have had a greater significance in the maintenance of social networks, in much the same role that the importation of non-local pottery may have played.

The identification of feasting is only the first step in the interpretative process. It needs to be placed in their socio-economic context by combining other forms of archaeological evidence (Dietler and Hayden 2001, 9-18).

‘without an adequately theorized and contextualized analysis of feasting, the mere documentation of the existence of such practices will not yield the kind of understanding of prehistoric societies and their social, political, and economic dynamics for which this domain of activity holds such heuristic promise’(ibid. 18).

Top of Page

Conclusions and Recommendations

‘pottery was part of the subsistence strategy for obtaining and distributing food and a primary contribution to the process of staying alive in later prehistory’ Morris 2002, 54).

Hill (2002, 79) argues that the initial use of pottery in the earlier Neolithic period may have been restricted for more ceremonial occasions than daily activities. He adds that across Britain ‘pottery appears as part of the new cultural, rather than purely subsistence orientated package that marks the Neolithic’ (ibid. 79). The initial adoption of pottery created a container medium that was malleable in the un-fired form, yet once fired could be used to hold liquid and dry goods for storage, processing or serving, and could be used directly over or on an open fire. This revolutionised food consumption and allowed food to be boiled. Arnold (1985, 232-3) notes that agricultural crops that were fairly resistant to disease contain a toxin that is also poisonous to humans. Ceramic vessels permitted a greater range of detoxifying processes than baskets and wooden vessels as they could be heated. They also prevent animals from getting to grain. Furthermore, the use of ceramics meant the customs associated with eating and drinking, both at a domestic and a communal level, ‘could be elaborated to a high degree’ (Woodward and Hill 2002, 2).

At Claydon Pike pottery manufacture was carried out at the household level, probably by women and children, utilising local resources. Technical choices of fabric, form, wall thickness, and the treatment of surfaces were made to produce vessels suited to cooking, serving and other uses such as storage and transport, within specific contexts such as domestic or feasting (Sillar and Tite 2000, 4). Two very large pots used for cooking activities point to communal sharing of food, and may have been used in feasting celebrations ‘for displaying wealth and sharing to reinforce or re-negotiate relationships’ (Morris 2002, 55). The settlement shares a regional concern for more sandy globular vessels, and these increase in use towards the end of the settlement’s life. The importation of vessels made from a glauconitic fabric and the Malvernian wares may be further evidence of the importance of maintaining social networks. Sources of sandy clays were available locally, so to bring in similar vessels from some distance suggests social, political and cultural reasons (Morris 1997, 38; Hill 2002, 77).

The role of pottery in elucidating the archaeological record is expanding. The recognition that potters work amongst a complex web of economic, social, political organisation; technical and environmental constraints, and their own beliefs and ideologies (Sillar and Tite 2000, 4), will allow a much fuller understanding of the people that made and used these vessels, and why they chose to do so.

Top of Page

Future work

Large scale excavation in the Upper Thames Valley has produced vast amounts of data concerning not only concerning the ceramics, but also the structure of settlements, the economy and the environment. It seems that now is the time to construct a synthesis of the manufacture and use of pottery, not only using information provided by form and fabric, but placing a greater emphasis on vessel use and size. In the case of the earlier reports, it will be necessary to go back to the original archive to obtain details of rim diameters and evidence of use. This can then be integrated with other forms of evidence to create an overall picture of the producers and users of pottery in the region.

Top of Page

 

3.1.8 BIBLIOGRAPHY

Allen, T G, 1990 The Iron Age Pottery, in T G Allen, An Iron Age and Romano-British Enclosed Settlement at Watkins Farm, Northmoor, Oxon, Thames Valley Landscapes: The Windrush Valley, Volume 1. Oxford, 32-46

Allen, T, 2000 The Iron Age Background. In M Henig and P Booth, Roman Oxfordshire, Gloucestershire

Arnold, D E, 1985 Ceramic Theory and Cultural Process, Cambridge

Arnold, D E, 2000 Some Concepts Linking INAA and Society from the INAA of Ethnographic Pottery. Paper presented at The Theoretical Archaeology Group, Oxford UK, 18-20 December 2002

Booth, P Oxford Archaeological Unit Roman Pottery Recording System: An Introduction. OAU in-house report

Booth, P, 2001 Appendix 1: Iron Age and Romano-British Pottery. In A Smith and I Scott

Braun, D P, 1983 Pots as Tools. In J A Moore and S A Keene (eds.) Archaeological Hammers and Theories, New York, 107-134

Clarke, M J, 2001 Akha Feasting. An Ethnoarchaeological Perspective in M Dietler, and B Hayden (eds.), Feasts. Archaeological and Ethnographic Perspectives on Food, Politics and Power, Washington, 144-167

DeRoche, C D, 1978 The Iron Age Pottery, in M Parrington 1978, 40-74

DeRoche, C D, 1997 Studying Iron Age Production. In A Gwilt and C Haselgrove (eds.) 19-25

Dietler, M and Hayden, B, 2001 Digesting the Feast – Good to Eat, Good to Drink, Good to Think: An Introduction, in M Dietler and B Hayden, B (eds), Feasts. Archaeological and Ethnographic Perspectives on Food, Politics and Power, Washington, 1-20

Duncan, D, Lambrick,G, Barclay, A, forthcoming Final Bronze Age to Middle Iron Age Pottery, in G Lambrick and T Allen, forthcoming

Dunning, G C, 1976 Salmonsbury, Bourton-on-the-Water, Gloucestershire, in Hillforts: later prehistoric earthworks in Britain and Ireland (ed. D W Harding), 76-118, London

Edgeley-Long, G P, unpub. Cleveland Farm, Ashton Keynes: The Iron Age Pottery. Unpublished report for Southampton University Course AY636, 2001

Gingell, C, 1982 Excavation of an Iron Age enclosure at Groundwell Farm, Blunsdon St Andrew, 1976-7, Wiltshire Archaeol Mag 76, 34-75

Gwilt, A and Haselgrove, C 1997 Reconstructing Iron Age societies, Oxford

Hally, D J, 1983 Use Alteration of Pottery Vessel Surfaces, in North American Archaeologist, vol. 4 (1), 3-26

Hayden, B, 2001 Fabulous feasts: a prolegomenon to the importance of feasting, n M Dietler and B Hayden (eds.), Feasts. Archaeological and ethnographic perspectives on food, politics and power, Washington, 23-64

Hill, J D, 2002 Pottery and the Expression of Society, Economy and Culture, in A Woodward and J D Hill (eds.), 75-84

Hodder, I, 1974 The Distribution of Savernake Ware, Wiltshire Archaeological Magazine, 69, 67-84

Jennings, D, Muir, J Palmer, S and Smith, A, 2004 Thornhill Farm, Fairford, Gloucestershire. An Iron Age and Roman pastoral site in the Upper Thames Valley, Thames Valley Landscapes Monograph 23, Oxford Archaeology

Lambrick, G, 1979 Finds: The Iron Age Pottery, in G Lambrick and M Robinson 1979, 35-46

Longacre, W A, 1991 Sources of Ceramic Variability Among the Kalinga of Northern Luzon, in W A Longacre (ed.) Ceramic Ethnoarchaeology, Tuscon, 99-111

Miles, D, 1997 Conflict and Complexity: The Later Prehistory of the Oxford Region. Oxoniensia LXII, 1-20

Millett, M, 1987 A Question of Time? Aspects of the Future of Pottery Studies, in Bulletin of the University of London Institute of Archaeology, 24, 99-108

Morris, E L, 1983 Salt and Ceramic Exchange in Western Britain During the First Millennium BC. Unpublished PhD thesis, University of Southampton

Morris, E L, 1997 Where is Danebury Ware? In A Gwilt and C Haselgrove (eds.) 1997, 36-39

Morris, E L, 2000, Pottery Assessment, in E L Morris, A Crosby, R Leech and T Machling, The Nevis Heritage Project Interim Report 2000, University of Southampton, 21-33

Morris, E L, 2002 Staying Alive: The Function and Use of Prehistoric Ceramics. In Woodward, A and Hill, J D (eds.) 2002, 54-61

Needham, S and Evans, J, 1987 Honey and Dripping: Neolithic Food Residues from Runnymede Bridge, Oxford Journal of Archaeology VC, 21-28

Orton, C, Tyers, P and Vince, A, 1999 Pottery in Archaeology, Cambridge

P.C.R.G. 1995 The Study of Later Prehistoric Pottery: General Policies and Guidelines for Analysis and Publication. Occasional Papers 1 and 2

Peacock, D P S, 1968 A petrological study of certain Iron Age pottery from western England, Proc Prehist Soc 13, 414-427

Peacock, D P S, 1982 Pottery in the Roman World: an Ethnoarchaeological Approach, London:

Rice, P M, 1987 Pottery Analysis. A Sourcebook, Chicago

Richardson, L, Arkell, M A, and Dines, H G, 1963 Geology of the country around Witney, Memoirs of the geological survey of Great Britain, Explanation of sheet 236, London

Saville, A, 1979 Excavations at Guiting Power Iron Age site, Gloucestershire, Committee for Rescue Archaeology in Avon, Gloucestershire and Somerset, occasional papers no. 7, Bristol

Saville, A and Morris, E, 1983 Iron Age Pottery, Fired Clay and Briquetage, in A Saville, Uley Bury and Norbury Hillforts. Rescue Excavations at Two Gloucestershire Iron Age Sites, 14-21, Western Archaeological Trust, Excavation Monograph No. 5

Sillar, B and Tite, M S, 2000 The Challenge of ‘Technical Choices’ for Materials Science Approaches in Archaeology, in Archaeometry, 42 (1) 2-20

Sinopoli, C M, 1991 Approaches to Archaeological Ceramics, New York

Skibo, J M, 1992 Pottery Function: A Use-Alteration Perspective. New York

Skibo, J M, 1995 The Clay Cooking Pot. An Exploration of Women’s Technology, in J Skibo, W Walker and A Nielsen, Expanding Archaeology, Salt Lake City, 80-91

Smith, A and Scott, I, 2001 Cotswold Water Park Project Gloucestershire, Revised Archive Assessment and Post-Excavation Research Design. Client report, unpublished, Oxford Archaeology

Sumbler, M G (ed), 1996 British Regional Geology: London and the Thames Valley, London

Sumbler, M G, Barron, A J M and Morigi, A N, 2000, Geology of the Cirencester District, Memoir for Sheet 235, British Geological Survey, London

Timby, J, 1995 The Pottery, in G Walker, A Middle Iron Age Settlement at Deer Park Road, Witney: Excavations in 1992, Oxoniensia LX, 78-82

Timby, J, 2001 The Pottery, in G Walker, B Langton and N Oakey, An Iron Age Site at Groundwell West, Blunsdon St. Andrew, Wiltshire. Excavations in 1996, Cirencester: Cotswold Archaeological Trust, 19-26

Timby, J, 2004 In Jennings et al. 2004

Wilson, D, 1993 Iron Age Pottery, in T G Allen and M A Robinson, 1993, 70-75

Woods, A J, 1986 Form, Fabric, and Function: Some Observations on the Cooking Pot in Antiquity, in W Kingery (ed.), Technology and Style, Ohio, 157-172

Woodward, A, 1997 Size and Style: An Alternative Study of Some Iron Age Pottery in Southern England, in A Gwilt and C Haselgrove (eds.), 26-35

Woodward, A, 1999 When Did Pots Become Domestic? Special Pots and Everyday Pots in British Prehistory, in Medieval Ceramics 22-23 (1998-1999) 3-10

Woodward, A and Hill, J D (eds.), 2002 Prehistoric Britain, The Ceramic Basis. Prehistoric Ceramics Research Group. Occasional Publication 3, Oxford

 

Top of Page

3.1.9 Petrological report on the non-local Iron Age pottery and Droitwich salt containers

by Elaine Morris

The pottery and fired clay from Claydon Pike were examined to determine if certain types of very distinctive Iron Age pottery from western England (Peacock 1968, Morris 1983) and Droitwich salt containers (Morris 1981, 1983, 1985) could be found within the collection. Only 22 g. of this non-local pottery (18 g. of Group A, with Malverninan rock inclusion, and 4 g of Group Bı, with Paleozoic limestone inclusions) and 351 g. of Droitwich salt container material were identified. This very low quantity of these artefacts suggests that Fairford may be at the southeastern edge of their respective distributions. Fairford is 60 and 65 km from the two pottery sources and 70 km from Droitwich.

In the Upper Thames Valley, these pottery types and Droitwich salt containers were not found at Mount Farm (G. Lambrick, pers. comm.), Ashville (Parrington 1978) and Farmoor (Lambrick and Robinson 1979) but the latter material was recovered from Mingies Ditch (Allen and Robinson 1993), Watkins Farm (Allen 1990) both of which are 75 km from Droitwich. Group A (Malvernian) pottery was also found at Groundwell Farm, Swindon (Gingell 1982), and Droitwich salt container sherds have subsequently been identified in that collection (Morris 1985). This site is located 68 km from the Malvern Hills and 79 km from Droitwich.

Six samples of ceramic material, four pieces of fired clay/daub and two pieces of salt container, were removed for thin sectioning and petrological analysis to show how different the locally derived clays used at the Claydon Pike sites are from the Droitwich clays used in the production of the salt containers. Table 16 lists the contexts which contained the salt container sherds from the Claydon Pike assemblage, while Table 17 lists the contexts with Groups A and Bı pottery. provides the context recording for the thin sectioned samples retained in the School of Humanities (Archaeology) at the University of Southampton.

Groups A and Bı Iron Age pottery

Detailed form and fabric descriptions of this material have already been presented elsewhere (Peacock 1968). The source for the inclusions in the Group A fabric pottery is located in the vicinity of the Malvern Hills in Worcestershire west of the river Severn. Early work on the inclusions in the Group Paleozoic limestone fabric could not determine which of several was the likely source for this limestone (Peacock 1968, 421-2). Subsequent quantitative work on the distribution of the Group pottery has favoured the Woolhope Hills in Herefordshire as the most appropriate source for these inclusions (Morris 1983, 116-122). At least one vessel of each fabric type was identified in the Claydon Pike collection from Island 2. This area is believed to date slightly later than Island 3 where, interestingly, Groups A and Bı pottery were not found. This information supports the interpretation that Groups A and pottery were first produced during the 5th-4th century BC for a localized distribution, or core area, and that a wider distribution developed from the 3rd-1st centuries BC through an exchange network which eventually incorporated the Upper Thames Valley (Morris 1983, 112-6).

Droitwich salt containers

This material has been described in detail and illustrated elsewhere (Morris 1985). Two general fabric types, a sandy type (FT1) and an organic-tempered type (FT2) have been defined. The former includes a specific sub-variety (FT1a) which contains clay pellets and is often found in collections from earlier iron age sites such as Crickley Hill, Shenberrow and Chastleton. All of the material was made from Keuper Marl clays found in the immediate Droitwich area and used to produce oxidized, vase-shaped porous containers. These in turn were used to dry and transport salt from the brine salt springs at Droitwich to hillfort and non-hillfort sites in the region. The FT1a sherds were found on Island 3 where FT1 sherds predominate in the salt container collection (61% by weight). On Islands 1 and 2, however, FT2 sherds are much more common (57% and 87% respectively). The observed difference in proportion between FT1 and FT2 on the islands is a pattern of technological change also found at the production source where in the earlier phase, FT1 was slightly more common but completely overshadowed by FT2 in the later phase.

The fired clays

The fabric descriptions presented below utilise the comparative charts recommended by the Prehistoric Ceramics Research Group (PCRG 1997).

Fired Clay Fabric 1- Oolitic and Calcite fabric

Macroscopic Definition

Numerous pieces of oolitic limestone characterize this poorly-wedged, sandy, fired clay fabric. When oxidized, the fabric ranges in colour from yellowish to reddish orange (Munsell 5YR 5/6-5/8; 2.5YR 4/6-4/8). The limestone fragments tend to be rounded and can measure up to 6.0mm across but are generally smaller in size. They occur in a common to abundant frequency.

Microscopic Definition

In a single thin section, this fabric contains a 25-35% concentration of sub-rounded pieces of oolitic limestone and a 1-5% concentration of sub-angular to sub-rounded calcite. Both types of limestone measure from 0.05-2.0 mm across. There is also a 1% concentration of quartz grains (0.2mm or more across), 5-10% concentration of medium-sized quartz (0.1-0.2mm) and 5-10% concentration of fine quartz. All of these grains are sub-angular to sub-rounded in shape. The clay matrix of this fabric is slightly micaceous and rich with iron oxides. Since the majority of inclusions are sub-rounded in shape, this is probably a naturally gritted clay.

Fired Clay Fabric 2 - Slightly Sandy Fabric

Macroscopic Definition

This is a black, reduced fabric which is rich with iron oxides and contains a sparse amount of rather rounded quartz grains. There is also a very rare concentration of minute, possibly shelly limestone fragments.

Microscopic Definition

In thin section, this iron-rich clay matrix contains a 1% concentration of sub-rounded quartz grains (0.5-0.2mm across), 5-10% sub-angular to sub-rounded grains (0.2-0.1mm) and 5-10% finer grains. Single pieces of sub-rounded micaceous siltstone (0.6mm), micaceous mudstone (0.9mm) and clay pellet were identified, but no calcareous matter was observed. This is probably a naturally gritted clay.

Fired Clay Fabric 3-Very Sandy Fabric

Macroscopic Definition

This is a very sandy fired clay but the quartz grains are much more angular and smaller than for the previous fabric types. This results in a very ‘micaceous-looking’ fabric with a very sandpaper-like feeling to the surface. It is irregularly fired from black to pink and pale orange (Munsell 1OR 5/8; 2.5YR 6/8; 5YR 6/6-7/6).

Microscopic Definition

In thin section, there is a 50% concentration of subangular to angular quartz grains in a micaceous, and iron-rich clay matrix. The quartz ranges as follows: 1% or less of large grains (0.5-0.2mm), 25% medium-sized grains and 25% finer grains.

Fired Clay Fabric 4-Very Fine or Ungritted Fabric

Macroscopic Definition

This is a very dense, iron-rich clay which contains very rare amounts of fine quartz. It ranges in colour from grey to pale orange/buff (Munsell 7.5YR 7/6).

Microscopic Definition

The clay structure of this fabric in thin section is very clumped together with pellet-like ferruginous fragments, occasionally containing one or two pieces of quartz. Otherwise, there is only a 1% concentration of very fine, sub-angular quartz (less than 0.1mm across). This is a naturally ungritted clay.

Sources for the Fired Clay Fabrics

The geology of the area up to 15 km around Claydon Pike consists of Oxford Clay, Cornbrash, Inferior Oolite and Upper Lias strata. Fired Clay Fabric 1, rich with oolites, probably came from the Inferior Oolite deposits which are generally full of oolitic limestones and often relatively free from shell fragments, particularly in the Lower Freestone Series (Kellaway and Welch 1948, 65). The three sandy, limestone-free fired clay fabrics could come from the Oxford clays or possibly from the clay facies of the Upper Lias (Kellaway and Welch 1948, 55-60 and 71-9). The petrological analysis of the four fired clay samples demonstrate how different the utilised clays in the Claydon Pike area are from those in the Droitwich area. It is unusual to discover that the closest possible source for Fired Clay Fabric 1 is located more than 10 km from Claydon Pike. Fired clay materials are generally expected to represent immediately local clay sources, if suitable types of clay exist.

Discussion

More than 50 pieces of Droitwich salt container material have been identified in the collections from all three Claydon Pike islands. A few sherds of the type generally associated with earlier Iron Age occupation (FT1a) were found at Island3, but FT2 material was also recovered from the area. Mainly FT2 and a few sherds of Groups A and Bı pottery were found at Islands 2 and 1 supporting the interpretation that these two islands were occupied slightly later in the Iron Age than Island 3.

The presence of salt container material at Claydon Pike, Mingies Ditch and Watkins Farm, but not amongst the Iron Age collections from Farmoor, Ashville and Mount Farm, suggests that either we are witnessing the maximum distribution of this material in an exchange system or that the salt in this area was used primarily on sites associated with a strong emphasis on pastorialism. The similarly limited distribution of the very distinctive Groups A and Bı pottery in this area favours the interpretation that the salt and pottery represent commodities in a restricted exchange network.

Bibliography

Allen, T G, 1990 An Iron Age and Romano-British Enclosed Settlement at Watkins Farm, Northmoor, Oxon, Thames Valley Landscapes: The Windrush Valley, Volume 1. Oxford

Allen, T G and Robinson, M A, 1993 The Prehistoric Landscape and Iron Age Enclosed Settlement at Mingies Ditch, Hardwick-with-Yelford, Oxon, Thames Valley Landscpaes: The Windrush Valley, Volume 2. Oxford

Gingell, C, 1982 The Pottery, in C Gingell, Excavation of an Iron Age Enclosure at Groundwell Farm, Blundson St. Andrew 1976-7. Wiltshire Archaeological Magazine 76, 54-63

Lambrick, G and Robinson, M, 1979 Iron Age and Roman Riverside Settlements At Farmoor, Oxfordshire, Oxford Archaeological Unit and the Council for British Archaeology, CBA Research Report 32

Morris, E L, 1983 Salt and Ceramic Exchange in Western Britain During the First Millennium BC, Unpublished PhD thesis, University of Southampton

Morris, E L, 1985 Prehistoric Salt Distributions: Two Case Studies from Western Britain. Bulletin of the Board of Celtic Studies 32, 336-379

Parrington, M, 1978 The Excavation of an Iron Age Settlement, Bronze Age Ring-Ditches and Roman Features at Ashville Trading Estate, Abingdon (Oxfordshire) 1974-76, Oxford Archaeological Unit and the Council for British Archaeology, CBA Research Report 28

Peacock, D P S, 1968 A Petrological Study of Certain Iron Age Pottery from Western England, Proceedings of the Prehistoric Society 34, 414-426

P.C.R.G. 1995 The Study of Later Prehistoric Pottery: General Policies and Guidelines for Analysis and Publication. Occasional Papers 1 & 2

Top of page