Archaeological Investigations of a Later Prehistoric and a Romano-British Landscape at Tremough, Penryn, Cornwall 9781407301242, 9781407320991

Citation preview

BAR 443 2007  GOSSIP & JONES  ARCHAEOLOGICAL INVESTIGATIONS

Archaeological Investigations of a Later Prehistoric and a Romano-British Landscape at Tremough, Penryn, Cornwall James Gossip Andy M. Jones

BAR British Series 443 9 781407 301242

B A R

2007

Archaeological Investigations of a Later Prehistoric and a Romano-British Landscape at Tremough, Penryn, Cornwall James Gossip Andy M. Jones with contributions by

Paul Bidwell, Wendy Carruthers, Rowena Gale, Anna Lawson-Jones, Joanna Mattingly, Henrietta Quinnell, Roger Taylor, Carl Thorpe and Rachel Tyson

BAR British Series 443 2007

ISBN 9781407301242 paperback ISBN 9781407320991 e-format DOI https://doi.org/10.30861/9781407301242 A catalogue record for this book is available from the British Library

BAR

PUBLISHING

CONTENTS List of Figures Acknowledgements Summary

iii v vii

Chapter 1

Project background Introduction Aims Location and setting The investigations Historical background

1 1 1 1 2 2

Chapter 2

The results Introduction The Neolithic (c. 4000-2500 cal BC) Early Neolithic Later Neolithic Environmental evidence Discussion of the Neolithic landscape The Bronze Age (c. 2200-1000 cal BC) Early Bronze Age Middle Bronze Age Environmental evidence Discussion of the Bronze Age landscape Later Iron Age and Romano-British period (c. 350 cal BC-AD 400) The Later Iron Age The Romano-British period Other Romano-British features Discussion of the Later Iron Age and Romano-British landscape The early medieval and later periods (AD 410-1540)

6 6 6 6 8 8 8 13 13 14 22 22 22 22 23 25 26 27

Chapter 3

Interpretation Neolithic Bronze Age The post-rings Structured deposition Order and architecture Iron Age and Romano-British Post-Roman and early medieval

28 28 31 31 34 37 40 50

Chapter 4

The artefacts Prehistoric, Roman and early medieval pottery Possible Early Neolithic (2000) Grooved Ware (2002) Second millennium cal BC: Trevisker and Trevisker-related (2002) Trevisker ceramics (2000) Baked clay objects of Bronze Age date Colourant Metalworking mould fragment or oven brick from a pottery clamp? Loomweights Later Iron Age and Roman period ceramics Post-Roman imported wares (2000) Early Medieval Spindle Whorl

51 51 51 51 57 65 67 67 67 67 69 78 78 78

i

Petrography Ironwork Stonework Neolithic (2000) Grooved Ware pits Structures of the 2nd millennium cal BC Roman period contexts Flint Medieval glass The medieval coin Post-prehistoric artefacts

78 81 81 81 81 83 85 88 96 96 97

Chapter 5

The ecofacts Calcined bone Plant remains Charcoal

100 100 100 107

Chapter 6

Radiocarbon dating Introduction Strategy Results

112 112 112 112

Appendices Appendix 1: Tables of structural details Appendix 2: Charred plant remains

119 125

Bibliography

135

ii

LIST OF FIGURES Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 Figure 15 Figure 16 Figure 17 Figure 18 Figure 19 Figure 20 Figure 21 Figure 22 Figure 23 Figure 24 Figure 25 Figure 26 Figure 27 Figure 28 Figure 29 Figure 30 Figure 31 Figure 32 Figure 33 Figure 34 Figure 35 Figure 36 Figure 37 Figure 38 Figure 39 Figure 40 Figure 41 Figure 42 Figure 43 Figure 44

Areas of archaeological recording 2000-2003. Archaeological anomalies identified by the geophysical survey. Early Neolithic features. Later Neolithic features in Area A (see Figure 5 for the wider context of the pits). Early and Middle Bronze Age landscape. Early Bronze Age features in Area A. Early Bronze Age Structure 66. Photograph of Structure 66 from the south-west. Middle Bronze Age Structure 102. Middle Bronze Age Structure 335. Photograph of Structure 335 from the east. Middle Bronze Age Structure 712. Middle Bronze Age Structure 392. Photograph of Structure 392 from the south-east. Distribution of finds within Bronze Age Structures 66, 102, 392 and 335. Reconstruction of the post-rings (conjectured posts are solid). The Iron Age and Romano-British landscape (west). The Iron Age and Romano-British landscape (east). Distribution of finds within the Romano-British enclosure. Romano-British Structure 338. Distribution of finds within Romano-British Structure 338 (quantities of pottery sherds numerated). Photograph of Romano-British enclosure, with adjacent field system visible (top left). Grooved Ware PP1-2 pit [193], PP3-6 pit [293]. Grooved Ware PP7 pit [300], PP8-9 pit [329], PP10-10A pit [331], PP11 pit [333], PP12-14 pit [494]. Grooved Ware PP15-17 spread (496). Trevisker/Trevisker-related ceramics PP18-27 Structure 392. Trevisker/Trevisker-related ceramics PP28-30 Structure 102, PP31 Structure 335, PP32, Structure 232, PP33A-34, ditch [596]. Trevisker/Trevisker-related ceramics PP35-36 old land surface (17), PP37 pit [25]. Clay mould/fire brick SF 1022, Structure 66. Local Roman period ceramics PR1-8 Structure 338 Phases 1-3. Local Roman period ceramics PR9-13 spread over Structure 338, PR14 enclosure ditch [306]. Local Roman period ceramics PR15-22 Enclosure Ditch [565]. Later Iron Age and Local Roman period ceramics PR 23 pit [729], PR24-27 pit [424], PR28 pit [377], PR29 ditch [30], PR30- 1 u/s Field 4. Neolithic axe S1. Grooved Ware pits. Stonework S2 pit [296], S3 pit [300], S4 pit [494], S5 spread (496). Middle Bronze Age stonework S6 Structure 392, S7 ditch [19] Field 4. Roman period stonework S8-11 from Structure 338. Roman period stonework S12-14 from Structure 338, S16 from pit [377]. Roman period rotary quern S15 from Structure 338. Worked flint artefacts from Field 2 (L1 and L2), Field 3 (L 3 to L6) and Field 4 (L7 to L14). Worked flint artefacts from Field 7 (L15 to L22). Worked flint artefacts from Area A (L23 to L31) and Area B (L32 to L34). Date ranges from Tremough Phase 1 and 2 excavations. Date ranges from Bronze Age Timber structures in Cornwall.

iii

3 4 7 9 10 12 13 14 15 16 17 18 19 20 35 36 41 42 43 44 47 48 55 56 58 59 63 66 68 71 73 74 76 80 82 84 86 87 88 92 93 94 114 116

ACKNOWLEDGEMENTS James Gossip and Andy Jones would like to thank the following for their help and co-operation: Combined Universities in Cornwall and Falmouth College of Arts for funding the excavation, archiving and final publication. Within HES, the authors would like to thank Carl Thorpe for his artefact illustrations, Sean Taylor, Dick Cole, Pete Dudley, Anna LawsonJones and Imogen Wood for their work on-site, Graeme Kirkham, Henrietta Quinnell and Peter Rose for their comments on the text and Richard Bradley and Tim Darvill for information on Bronze Age timber-circles. We would also like to thank the many people who worked on the excavations in 2000 and 2002. Henrietta Quinnell is grateful to Paul Hands of the School of Earth Sciences, University of Birmingham, for the preparation of thin-sections. The Ordnance Survey mapping included within this publication is provided by Cornwall County Council under licence from the Ordnance Survey in order to fulfil its public function to disseminate information to a wider audience. Persons viewing this statement should contact Ordnance Survey where they wish to licence Ordnance survey mapping for their own use.

v

SUMMARY Following an archaeological assessment and three geophysical surveys, a succession of archaeological excavations, test pitting and monitoring was carried out by Historic Environment Service, Cornwall County Council at the site of the Combined Universities in Cornwall’s campus at Tremough, Penryn between 2000 and 2004. Excavations were focused upon the investigation of a Neolithic flint scatter and pit groups, including some containing Later Neolithic Grooved Ware, together with Early to Middle Bronze Age timber structures and a small enclosed Romano-British settlement, situated upon a plateau overlooking the Fal estuary. Subsequent smaller-scale monitoring and test pitting was carried out at the east end of the plateau and on its slopes. This work confirmed the results from geophysical survey, which had indicated that there was a marked decline in significant archaeological features beyond the plateau. Twenty-two radiocarbon determinations ranging between 3950 BC and AD 1160 were obtained. These dates fell into four major groups; the Early Neolithic, the Later Neolithic, the Early to Middle Bronze Age and the Later Iron Age to RomanoBritish period. A single earlier medieval date was also obtained from a field ditch. From the beginning of the medieval period the site was part of the manor of Tremough and it continued to be an estate with parkland into the earlier part of the twentieth century. Although the archaeological record for this period is comparatively sparse, historical research undertaken for the assessment revealed a remarkable history which continues into the present with the foundation of Cornwall’s first university.

vii

CHAPTER 1

PROJECT BACKGROUND

Introduction

uncovered, as well as five circular post-ring structures and other features dating to the Early to Middle Bronze Age. In the light of these discoveries, the research objectives became focused upon issues relating to understanding the complexity of the phasing of the site, identifying long-term change across the project area, and learning more about the function of the post-rings.

The Historic Environment Service (Cornwall County Council) was commissioned by Falmouth College of Arts and later the Combined Universities in Cornwall (CUC) to undertake a programme of archaeological recording at Tremough, Penryn (SW 726 347), in advance of the construction of the CUC campus (Fig. 1). During the course of the project approximately eight hectares were subjected to geophysical survey and four hectares on the plateau were topsoil stripped. In addition, some 35 test pits were dug, and archaeological monitoring during site works took place at numerous locations across the site.

Location and setting The Tremough site is situated approximately 100m to 120m OD on an elongated, roughly level, 600m long spur looking south east over the Fal estuary and to the sea beyond. It lies immediately north west of Penryn (Fig. 1) and prior to the development comprised an area of farmland and former parkland on the edge of the urban area. The underlying geology varied across the site as it lies on the junction between the clays and slates associated with the Devonian sandstone and slates of the Mylor Series and the igneous Carnmenellis granite (Geological Survey of Great Britain 1974). The majority of the exposed bedrock is characterised as metamorphic with killas and vein-quartz. The natural clay subsoil ranged from a bright pale yellow to a deep rusty orange across the site. The overlying soiltype is classified as a Stagnogley soils and Rankers.

The project was significant because it provided the first opportunity in Cornwall to undertake a staged programme of assessment before investigating a large area in a controlled manner. This enabled the study of long-term changes in the organisation and development of a selfcontained block of land. Aims Initial objectives for the excavations were based on an understanding of the site which derived from previous assessment, field-walking, evaluation and geophysical survey.

Prior to the archaeological recording there was little to suggest that there were significant archaeological remains within the project area. There were no upstanding or documented sites recorded in the county Historic Environment Record and the types of crops in the cultivated parts of the site were not of a type which produced crop-marks that could be identified from the air. However, it was known that the Tremough place-name was of medieval origin and the farmland was characterised as ‘Anciently Enclosed Land’ (Cornwall County Council 1996). ‘Anciently Enclosed Land’ is made up of farming settlements documented before the seventeenth century AD and field patterns of medieval origin. Investigations across Cornwall have demonstrated that it has a high potential to contain buried archaeology dating to the prehistoric and Romano-British periods (Cole and Jones 2002-3; Jones and Taylor 2004; Taylor 2005). This potential was confirmed by the geophysical surveys which revealed a high concentration of archaeological features across the top of the plateau. This had been extensively altered in the later eighteenth and nineteenth centuries following the construction of the present house. There is evidence that during this period the field layout was reorganised to

The objectives were to: Investigate the character of the buried archaeology within a topographically well-defined block of lowland Cornwall. Increase understanding of prehistoric and RomanoBritish settlement activity, which had been indicated by finds recovered from the walkover and evaluation trenching. Clarify the archaeological potential of the anomalies identified by the geophysical survey to become better informed about the impact of proposed development on the archaeological resource. However, the topsoil stripping of the area revealed an even greater complexity than had been indicated by the assessment and geophysical surveys. In addition to the expected linear field systems and Romano-British settlement activity, Early and Later Neolithic pits were 1

ARCHAEOLOGICAL INVESTIGATIONS OF A LATER PREHISTORIC AND ROMANO-BRITISH LANDSCAPE AT TREMOUGH make larger rectilinear fields and an ornamental landscape which included Italianate and terraced gardens were laid out around the house.

excavated by a machine supplied by the client under the supervision of an archaeologist. Very few finds and no archaeological features were recovered in any of the test pits, which highlights the potential inapplicability of this technique on rural sites.

The investigations Between 2000 and 2004 HES carried out several stages of archaeological recording which helped to identify the archaeological potential of the site. They include, desktop assessment, geophysical survey, evaluation trenching, test pitting, field-walking, watching briefs and excavation. The methodologies for these stages of recording are briefly outlined below.

Field-walking Field-walking was carried out in advance of the Phase 1 excavations. Fields 2, 3 and 7 (Fig. 1) had been cultivated which meant that it was possible to systematically walkover them to recover artefacts and identify areas of earlier activity (Lawson-Jones 2002). In particular, Field 4 produced quantities of prehistoric lithics, RomanoBritish pot sherds and a Neolithic greenstone axe. Field 4 became the focus of investigation during the Phase 1 and 2 excavations.

Assessment In 2000 a desk-top survey and walkover was carried out for all four hectares covered by the proposed campus (Jones and Lawson-Jones 2001). This included two large geophysical surveys, extensive historical research and landscape characterisation.

Watching briefs Between 2000 and 2004 watching briefs were undertaken across the campus area (Lawson-Jones 2002; Dudley 2003; Gossip 2004). The three largest were situated in the car-park area to the east of the ‘Fort’; ahead of soak-away construction in the grounds to the east of the house; and in advance of the construction of the Digital Media Centre. In all three cases the ground was stripped down to the natural subsoil by a swing-shovel under the supervision of an archaeologist.

Geophysical survey Assessment of the project area resulted in two magnetometer surveys by Geophysical Surveys of Bradford (GSB 2000; 2001) and a third by Stratascan (Mercer 2001) at recommended locations across the site. In total, approximately eight hectares were surveyed. The surveys were located across the main plateau within predominately ‘Anciently Enclosed Land’ and at several locations on the eastern side of the complex (Figs 2 and 18). Numerous anomalies were identified, including a rectangular enclosure at the eastern end of the plateau, which became known as the ‘Fort’, a smaller ‘D’ shaped enclosure at the western end and a possible Bronze Age roundhouse in Field 3 (Fig. 2).

Excavations The Tremough excavations were carried out in 2000 and 2002 (Lawson-Jones 2002; Gossip 2003a) (Fig. 1). The HES team was aided by volunteers, including university students from Bristol, Exeter, and Truro College and members of Cornwall Archaeological Society. The excavation area was stripped by swing-shovel, through the topsoil and then cleaned and planned. In 2000 the excavation area was surveyed by plane table, whereas in 2002 the investigated area was planned using an EDM (Electronic Distance Measurement). Subsequent plans and sections were produced using the EDM or drawn to scale by hand. With the exception of the unstratified artefacts, the majority of the finds recovered were recorded three dimensionally so that their distribution was plotted. Soil samples were taken from those features and layers considered to have the greatest potential for palaeoenvironmental analysis.

Evaluation works Three phases of evaluation recording have taken place (Lawson-Jones 2001; Gossip 2003b; Gossip 2006). The first was used to confirm the identity of anomalies revealed by the geophysical survey. The second was carried out at the eastern end of the plateau near the ‘Fort’ to assess the level of erosion caused by construction traffic. The last consisted of two 20m long trenches being excavated into the interior of the ‘Fort’ prior to soil being deposited over the site. In each case a machine fitted with a flat bucket was provided by the contractor and trenches excavated to the top of the subsoil were supervised by an HES archaeologist. Sections of the trenches were drawn and archaeological features planned and recorded.

Historical background Joanna Mattingly No estate records survive for Tremough. What exists are mostly stray deeds and claims to title, material from other gentry archives, universal materials like taxation and muster lists, the tithe map and apportionment, censuses, newspapers and trade directories, together with miscellaneous materials like turnpike and auction records, diaries and garden notes (Mattingly in Jones and Lawson-

Test pitting Thirty-five test pits have been excavated across the area of the campus (Lawson-Jones 2002; Cole 2002). All were 2



Figure 1: Areas of archaeological recording 2000-2003.

Project Background

Figure 2: Archaeological anomalies identified by the geophysical survey.

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough



Project Background Jones 2001). This dearth of estate records reflects the frequent changes of ownership and status at Tremough by contrast with nearby estates like Enys and Trefusis, or even Carnsew, with greater continuity of ownership. However, enough survives for Tremough to reconstruct in partial form its probable historical development from the thirteenth century.

streams. To the west was the larger estate of Carnsew, named after a prominent rock on its land and later the site of famous granite quarries. After 1200, both Tremough and Carnsew were effectively free tenements, both apparently practising open field agriculture (ibid, liv, for socage; tithe apportionment books for ‘gwel’ names at Carnsew; P Herring pers. comm.).

The 800 year spread of documentary references confirms the impression derived from the archaeological record that Tremough was a prime settlement site. Although no remains have been excavated, medieval pottery has been found in the Italian garden flowerbeds to the west of the house (B Mehen pers. comm.) and in unstratified contexts elsewhere (Thorpe below). Tremough’s proximity to the medieval town and port of Penryn also has implications, not least in offering a potential livelihood for younger sons. In view of these two elements, a defensible medieval precinct at Tremough (P Herring pers. comm.) with a centrally-placed house like those formerly at Godolphin, and at Hall near Bodinnick, or Cotehele, is not improbable (Herring 1998, 200-2). The medieval context of the site is therefore worthy of some further comment.

Nicholas de Tremough was one of the four leading parishioners in 1309 and it is possible that the Tremoughs created a medieval precinct around their house in the thirteenth or early fourteenth centuries. However, a more likely scenario is that if this precinct existed, it was the work of the Carnsews. Walter de Carnsew was the spokesman on behalf of all the parishioners in 1309 and already held a third, if not two thirds, of the holdings at Tremough. The Tremough family are not certainly noted at Tremough after 1538 and subsequent references to them in Carnsew rentals appear to relate to property in Penryn. The medieval house was replaced by the present highstatus Queen Anne house, although there may have been one or more house on the site in the interim. During the second half of the nineteenth century the Shilson family rebuilt part of the house and remodelled the interior.

Tremough was part of the bishop of Exeter’s manor of Treliever before 1066 and may be one of the 30 villein tenements mentioned in the Domesday Book in 1086 (Thorn and Thorn 1979, 2). The name Tremough appears to mean ‘the estate of swine’ (Padel 1985, 167, 223). In 1208-13, Tremoh comprised two fiscal holdings of an acre and a half-acre, Cornish (Fox and Padel 2000, l, liv-lv). The estate of Tremough presumably centred on the present estate, which has boundaries formed largely by natural



Tremough’s history of educational use extends from Benjamin Barwis’s 1799-1802 Tremough Academy and use by a convent school from 1943 to 1998 to the Falmouth College of Arts, which made its campus here in 1998, and its current role as the focus of the Combined Universities in Cornwall.

Chapter 2

The results Introduction

dated to the Early Neolithic period (c. 4000-3500 cal BC) and was associated with flint scatters, pits and a possible enclosure ditch. The second dated to the Later Neolithic period (c. 2900-2300 cal BC) and was associated with shallow pits and small-scale structured deposition of Grooved Ware and flint. The Early Neolithic activity was located in a different part of the site from the subsequent Later Neolithic pit group.

The results from the desk-top assessment, evaluation trenching, field-walking and especially the geophysical surveys meant that further archaeological recording would be necessary in mitigation of the campus development. In particular, the geophysical survey identified a rectangular enclosure the ‘Fort’, with a ‘D’ shaped enclosure on its southern side at the eastern end of the plateau and a smaller curvilinear enclosure at the western end (Fig. 18). Evaluation trenching of features at the western end of the plateau confirmed that there were archaeological features of Iron Age or Romano-British date (Lawson-Jones 2001). The ‘Fort’ was not eventually excavated as it was taken out of the development envelope.

Early Neolithic The Early Neolithic activity was revealed during the fieldwalking and subsequent topsoil strips in Fields 4 and 7 (Lawson-Jones 2002). No evidence of Early Neolithic activity was found during the Phase 2 excavations in Areas A or the western end of Area B (Field 4) (Gossip 2003a).

Although geophysical survey detected a large number of anomalies, the responses from some archaeological features were relatively weak and some significant features, including five Bronze Age post-rings subsequently discovered by excavation, were not detected.

Field-walking in Field 4 recovered 130 flints and another 10 were recorded during the excavations (Lawson-Jones below). A Group I greenstone axe, probably from the Mounts Bay area was also found during field-walking but this could date to either the Early or Later Neolithic. The overwhelming majority of the flint, 123 pieces, in this area was pebble flint which had probably been derived from local beaches. Although Field 7 could not be systematically field-walked because it was scrub-covered, topsoil stripping in this area led to the recovery of an even greater quantity of lithics. A total of 205 flints were found within the ‘L’ shaped trench and the access road corridor. Remarkably, 115 of them were derived from a nodular source outside Cornwall and were associated with a knapping floor (107) (Lawson-Jones below). A second group of larger and more abraded nodular flints was also found in Field 7, to the north of the flint knapping floor. This assemblage was totally different in character from the tiny pieces of knapping debris found in (107).

The results of the two phases of archaeological excavation are presented below by period, together with evidence from the evaluations, watching briefs, environmental analyses and artefactual studies. The probability distributions for the radiocarbon determinations (Tables 18 and 19 and Fig. 43) have been calculated using OxCal (v3.10). The 95% level of probability is used throughout this report. The Neolithic (c. 4000-2500 cal BC) The excavations at Tremough led to the identification of two significant phases of Neolithic activity. The first Table 1: Descriptions of Early Neolithic pits

Near-circular pit, 0.5m-0.6m in diameter and up to 0.1m deep with concave sides and a round base. Filled by (12), a charcoal-rich, red-brown clay loam. Circular pit, 0.6m in diameter and up to 0.18m deep, with concave sides and a round base. Filled by (14), a charcoalrich black silty loam, which contained burnt stone. Circular, steep-sided pit, 0.8m in diameter and up to 0.19m deep and an uneven base. Sealed by a large, burnt stone, laid horizontally and filled by (22); a very dark grey-brown silty loam containing charcoal and burnt stone. Oval pit, 1m by 1.2m and up to 0.2m deep, with sloping sides and a round base. Filled by (24), a mixed orangebrown fill, very disturbed by animal burrowing. Near circular pit, steep-sided and flat-bottomed. It measured 0.65m by 0.7m and was up 0.19m deep. Filled by (29), a mixed silty loam, which contained charcoal and burnt stone. A flat stone was located at the base of the cut.



Pit [11] Filled by (12) Pit [13] Filled by (14) Pit [21] Filled by (22) Pit [23] Filled by (24) Pit [28] Filled by (29)

The Results

Figure 3: Early Neolithic features. Differences between Fields 4 and 7 also extended to the below-ground evidence. In Field 4, three pairs of pits were identified (Fig. 3). The largest, pit [25], contained Bronze Age pottery (Quinnell below) and produced a Middle Bronze Age radiocarbon determination. The five remaining pits were not associated with pottery, were smaller, more circular in plan and all but one, pit [23] contained evidence for burning in the form of charcoal and burnt stones (Table

1). Pit [21] produced an Early Neolithic determination of 4850±55BP, 3770-3510 cal BC (AA-44601). By contrast, Field 7 to the north east was devoid of pits and the only certain Neolithic feature was ditch [76]. This ran on a broadly north–south alignment for approximately 8m. It was up to 0.9m wide and 0.5m deep, slightly curvilinear in plan and terminated 2m from the eastern side of the access 

ARCHAEOLOGICAL INVESTIGATIONS OF A LATER PREHISTORIC AND ROMANO-BRITISH LANDSCAPE AT TREMOUGH of the identifiable Later Neolithic activity (2900 to 2300 cal BC) was confined to the north west end of Area A (Figs 4 and 5). Twelve small pits were identified in this area, although two of these, [296] and [298], may be of Romano-British date, and two were of uncertain date [497] and [499]. However, eight produced Grooved Ware. All were shallow, 0.15m-0.25m deep and were between 0.7m and 1m in diameter (Table 2).

road. The terminal showed no sign of either a posthole or stone-packing associated with the southern side of the potential entrance that it formed. Unfortunately an opposing ditch was not revealed in the Access Road excavation. A notable characteristic of the ditch was the very pale, compact nature of the majority of its two fills (76) and (77), and the very stony character of upper fill (77). The latter appeared to have been deliberately backfilled into the ditch cut. It is unlikely that this stony material represents up-cast formed when the ditch was cut since the quantity of stone appeared disproportionably great, and the blocks were fairly large and unlike those seen in the ditch sides. A Neolithic radiocarbon determination 4995±50 BP, 39503660 cal BC (AA-44604) was obtained from charcoal within (77), together with a possible Early Neolithic sherd of pottery (Quinnell below). Although the date is a little earlier than the determination from pit [21], there is sufficient overlap to suggest that they could have been contemporary.

To the north of pit [494] was spread (496), an orange, yellow sandy silt which covered an area measuring approximately 1m by 1.5m and 0.1m thick. It contained 33 sherds of Grooved Ware, including vessels PP15, PP16 and PP17, two worked stone objects SF1055, S5 polisher or grinder and SF1056 white vein-quartz cobble and four flints (Quinnell and Lawson-Jones below). This deposit may represent a spread of occupation-related activity. Environmental evidence The environmental analyses shed little further light on Neolithic activity at Tremough (Carruthers and Gale below). No evidence was recovered for cereal cultivation, and the only identified food remains were fragments of hazelnut shells in three of the Later Neolithic pits, [193], [293] and [331], and undated pit [298]. The Early Neolithic pits provided no evidence for the surrounding environment, but the charred material from some of the Later Neolithic pits included small fruits and seeds from common weeds of grasslands and disturbed soil. As a whole, the grassland taxa originate from a damp, fairly open grassland habitat, so the area may have been disturbed or grazed for some time (Carruthers below). Oak and hazel charcoal was recovered from both Early and Later Neolithic pits, suggesting that timber was locally available (Gale below).

None of the other features revealed in Field 7 can be demonstrated to be of a Neolithic date. Eighteen pits and postholes and 25 stakeholes were uncovered (LawsonJones 2001, 46-7). Only two produced artefacts, however: some medieval glass from pit [58] and a post-medieval potsherd from pit [62]. In general the pits in Field 7 had fills which were darker, and less compact than the fill of ditch [76], and none were pits of the kind found in Field 4. The 25 stakeholes formed a 4.5m long north-west–southeast alignment [95], and were filled by compact clays, which resembled the fills of ditch [76]. It is possible that they belong to the earlier Neolithic and that they formed some kind of fence. Later Neolithic

Discussion of the Neolithic landscape One of the most remarkable aspects of the investigations at Tremough was the recovery of 165 sherds of Grooved Ware, representing some 21 vessels (Quinnell below). All

The Neolithic evidence at Tremough was derived from non-monumental activity. The evidence did appear to

Table 2: Descriptions of Later Neolithic pits Pit [193] Filled by (194) Pit [293] Filled by (294) Pit [331] Filled by (332) Pit [333] Filled by (334) Pit [300] Filled by (299) Pit [329] Filled by (330) Pit [492] Filled by (493) Pit [494] Filled by (495)

Sub-oval pit, 0.82m-1.15m across and up to 0.3m deep, filled by (194), a dark orange-brown silty clay. Sherds of Grooved Ware PP1 and PP2 and a flint burin were found. Oval pit measuring 0.7m-0.83m and up to 0.25m deep, filled by mid-brown silt (294). Sherds of Grooved Ware PP3, PP4, PP5 and PP6 were found. Sub-oval pit, 0.95m-1.3m, across and up to 0.18m deep, filled by (332), a dark brown sandy silt. Sherds of Grooved Ware PP10, and PP10a, a flint beach cobble SF1029 and three flints were found. Oval pit, 0.65m-0.83m across, and up to 0.16m deep, filled by (334), a mid orange-brown sandy silt. A sherd of Grooved Ware PP11 and large flint flake were found. Sub-circular pit, 0.8m-0.91m across and up to 0.18m deep, filled by (299), a pale grey-brown sandy silt. Sherds of Grooved Ware PP7 and a beach cobble grinding tool S3 were found. Oval pit, 0.61m-0.71m across and up to 0.18m deep, filled by (330), a dark brown clay silt. Sherds of Grooved Ware PP8, PP9 and a nodular flint flake found. Pit, measuring 0.6m in diameter and up to 0.3m deep, filled by (493), a light orange-brown clayish silt. Sherds of Grooved Ware and a discoidal chert knife were found. Oval pit, measuring 0.95m-1.1m across and up to 0.2m deep, filled by (495), a mid-brown sandy silt. Sherds of Grooved Ware PP12, PP13, PP14 and were found. A number of flints, a small saddle quern S4, an aplite stream cobble SF1051 and a fragment of a beach cobble SF1045 were also found.

8

The Results



Figure 4: Later Neolithic features in Area A (see Figure 5 for the wider context of the pits).

Figure 5: Early and Middle Bronze Age landscape.

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough

10

The Results indicate differing activities across the site, particularly in the Early Neolithic period between c. 3900 cal BC and 3500 cal BC.

have originally been deeper. Descriptions of the post-rings can be found below in appendix 1. Early Bronze Age

Although a much smaller area was investigated in Field 7 – 0.6 hectares against 2.4 hectares in Field 3 – there were sufficient differences between the two areas to indicate a zonation in the types of activity which were occurring across the site. In particular, activity in Field 3 was marked by pits containing burnt material and a scatter of locally available flints. By contrast, activity in Field 7 was delineated by a shallow ditch which cut off the east end of the plateau from Field 3. Charcoal-rich pits were absent and the scattering of pits and postholes could not be convincingly assigned to this phase. A line of stakeholes [95] may represent some kind of partition. However, non-local nodular flint was worked in this area and it is possible that ditch [77] acted as division between the working of different lithic resources and possibly in the character of the occupation. The collection of two types of nodular flint from two separate areas which had been worked differently within Field 7 (Lawson-Jones below), could also hint that there were further spatial subdivisions in the working of flint.

Structure 66 This structure comprised an arc of six postholes (Figs 6, 7 and 8), assumed to form a ring approximately 6m in diameter, although the eastern extent of the postring was not excavated. The structure was devoid of internal features or deposits. The postholes were with the exception of [67], which was 0.73m by 0.43m-0.45m, of a reasonably consistent size but shallow, with an average depth of only 0.24m. All the post-holes were filled with silty clay deposits and packing stones were evident in most (appendix 1, Table 21). The elongated and slightly irregular posthole [67] appears to have been disturbed, perhaps by the collapse or removal of the post. Posthole [73] produced part of a fired clay block (SF1002), possibly a fragment of an object connected with either pottery production or metalworking. The unweathered nature of the object suggests that it had been deliberately deposited within the posthole (Quinnell below). Since there is no evidence that metalworking or pottery production took place close to Structure 66, the block must have been brought to the structure from elsewhere and deliberately buried within the posthole. One sherd of Trevisker pottery was recovered from posthole [67] and another tiny sherd was found in [71]. No cereal grains were recovered from the structure. Posthole [75] produced a radiocarbon determination of 3704 ± 38 BP, 2210-1970 cal BC (Wk14995). A second posthole [69] gave a radiocarbon determination of 3380 ± 38 BP, 1770-1530 cal BC (Wk14994), perhaps suggesting the renewal of the post-ring several centuries after its original construction.

The Later Neolithic saw a shift in activity towards the western and of the plateau, beyond any of the Early Neolithic remains. Here too, pits were dug but there was little evidence of burning and instead Grooved Ware, flint and stones were selectively incorporated into the pit fills. Radiocarbon dating indicates that this took place between c. 2900 and 2300 cal BC. A small spread of pottery and flint located to the north of the pit group suggests that some occupation-related activity took place in the area adjacent to the pits. The evidence from the environmental analyses was extremely limited but could be taken to suggest that by at least the Later Neolithic the plateau had been cleared; given the number of Early Neolithic features, this may have happened earlier in the fourth millennium cal BC. Significantly none of the securely dated environmental samples revealed any evidence for pre-Bronze Age cereal cultivation.

Pit group This pit group, comprising an arrangement of pits in an incomplete circle with a diameter of 2m, was situated 18m to the north of Structure 66 (Fig. 6). The features were generally shallow pits between 0.12m and 0.19m in depth 0.46m and 0.8m in diameter (Table 3). Most were filled by silty clays and all contained charred plant remains, and an abundance of charcoal. The principal deposit in all pits was fragments of hazelnut shell with smaller quantities of charred cereal grains also present (Carruthers below). The greatest numbers of these were produced by pit [57], fill (58). Narrow roundwood species including hazel, holly and oak were also identified (Gale below). Heat-affected stones were also present in the pit fills, but as no evidence for in situ burning was found it is unlikely that the features functioned as hearth-pits. Two very similar radiocarbon determinations were obtained from the pit group. Pit [59] produced a date of 3668 ± 45 BP, 2200-1920 cal BC (Wk14992) and pit [64] a determination of 3677 ± 45 BP, 22001930 cal BC (Wk-14993).

The Bronze Age (c. 2200-1000 cal BC) The Bronze Age activity on the site spanned almost the entirety of the second millennium cal BC, starting with Early Bronze Age dates from Structure 66, and from pits [59] and [64], and continuing with Middle Bronze Age Structures, 102, 335, 392 and 712. This period saw a continuation of pit-related activity, but also the appearance of five circular post-built structures arranged on a roughly north west to south east alignment, a linear arrangement of posts and the construction of a probable sub-rectilinear building (Fig. 5). Other activity dating to the Middle Bronze Age included pit [25] (Field 4) and an old land surface [17] (Field 4). All of these features have suffered some truncation from later agricultural activity and would 11

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough Table 3: Early Bronze Age Pit Group Details Context [57]

Diameter

Depth 0.22m

[59]

0.8m

0.14m

[61]

0.8m

0.18m

[64]

Width 0.76m

0.58m

Length 0.9m

0.73m

0.19m

[79]

0.57m

0.12m

[81]

0.46m

0.12m

Comments Steep-sided pit filled by (58), silty clay, large quantities of charcoal and burnt stone. Steep-sided pit filled by (60), dark silty clay, some burnt stone and charcoal. Bowl-shaped pit filled by (62), friable silty clay, some root disturbance. Steep-sided pit filled by (65), silty clay, frequent charcoal, granite stones and quartz. Shallow pit filled by (80), sandy clay, frequent charcoal, some stone. Shallow pit filled by (82), silty clay, large quantities of charcoal.

Figure 6: Early Bronze Age features in Area A. 12

The Results

Figure 7: Early Bronze Age Structure 66.

13

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough

Figure 8: Photograph of Structure 66 from the south-west. Middle Bronze Age

within posthole fills included rose seed, elderberry, and mallow interpreted as wild foods, and oak and hazel charcoal fuel. The hearth-pit [157] produced cereal grains and disturbed ground species, including waste from the final stages of cereal processing suggesting that this may have taken place next to the hearth-pit immediately prior to use (Carruthers and Gale below).

Structure 102 This structure consisted of a ring of twelve postholes evenly spaced 1.9m and 2m apart, forming a ring approximately 7m in diameter. The post-ring encircled a probable hearthpit [157] and four post or stakeholes [133] [149], [153] and [155] (Fig. 9). Two additional postholes and five shallow pits ([151], [159], [112], [145], [139], [141], and [143]) were located on the exterior of its south eastern side, and may have demarcated the entrance. Partly overlying these features was a stony spread [125] containing sherds of Trevisker-related pottery (PP30). Two further sherds of this vessel were found within pit [139] (Quinnell below).

Three radiocarbon determinations were obtained from Structure 102. Posthole [109] produced a date of 3180±39BP, 1530-1380 cal BC (WK-14996), posthole [112] a date of 3111±37BP, 1460-1290 cal BC (WK14997) and hearth-pit [157] a date of 3009±40 BP, 13901120 cal BC (WK-14998). Structure 335

The postholes ranged between 0.17m and 0.32m in depth, had steep or vertical profiles and generally flat bases, and were between 0.17m and 0.45m in diameter. Packing stones were recorded in five of the postholes; in one of these [126] the post-pipe (128) had been preserved. Postholes [112] and [145] were close together and may indicate replacement. Most of the posts were filled by silty clay fills (appendix 1, Table 22). A total of 29 sherds of Trevisker-related pottery were recovered from the structure, with 11 sherds of PP30 from pit [139]. Pottery was also found in postholes [112] (PP29), [123], [133], [147] (PP28) and spread (125), (PP30) (Quinnell below). Other finds included beach veinquartz pebbles from [107] and a possible pebble hammer and beach cobble from (125). Charred plant macrofossils

Structure 335 (Figs 10 and 11) was located in the middle of the timber post-ring alignment and comprised a ring of ten postholes forming a structure with a diameter of 6.3m. The postholes were spaced 1.9m and 2.1m apart and their diameters ranged between 0.10m and 0.3m. They were between 0.10m and 0.20m deep with the exception of [369] and [349] which were very shallow; all may have been truncated by agricultural activity. Postholes [371], [265] and [261] had steep, near-vertical profiles. Others had more concave sides and presumably had been affected by erosion or collapse as the result of removal or decay of the posts. Posthole [345] was cut by posthole [261]; this was perhaps a replacement post consolidating the 14

The Results

Figure 9: Middle Bronze Age Structure 102.

15

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough

Figure 10: Middle Bronze Age Structure 335.

16

The Results

Figure 11: Photograph of Structure 335 from the east. 0.41m deep. Two internal features were recorded, [704] and [710], themselves close to the western arc of postholes. Feature [704] was a possible posthole or post-base, its fill packed with small angular stones. Posthole [710] possessed an apparently intact post-pipe. Any other internal features would probably have been obliterated by Romano-British ditch/gully [596], which contained two sherds of PP33 and one sherd of PP34 Trevisker pottery, presumably derived from disturbance to the structure. Postholes [606] and [608] were located on the south east side of the Structure. Both reached a depth of 0.35m, which was much deeper than all but one of the other posts in the ring. This might imply that taller posts were located in a position comparable with the clearly defined entrances into Structures 102 and 392. A shallow pit [367] containing burnt stone was found on the east side of the Structure.

structure. The postholes all had friable silty clay fills and lacked stone packing (appendix 1, Table 24). Internal features included four shallow postholes or post ‘impressions’ ([341], [349], [365] and [263]) close to the south western perimeter of the structure and [361], a posthole just to the south west of the centre of the structure. Four features were recorded close to but outside the structure which may have been associated: postholes [363], [375], [388], and pit [604] (cut by ditch [383]). Postholes [363] and [375] were situated on the south east side of the post-ring and could have helped to demarcate an entrance into the structure. Two possible stakeholes (not numbered) could be seen on the north eastern edge of the structure, one adjacent to posthole [371] and the other adjacent to posthole [357]. A single sherd of Trevisker pottery was recovered from the fill of internal posthole [361]. Thirteen sherds of PP31 pottery were recovered from external pit [604] (Quinnell below). As with the Early Bronze Age Structure 66, no cereal grains were recovered and there was no hearth-pit. No material available for scientific dating was available.

An elongated, shallow pit [600] cut the northern arc of the structure. The eastern end of this pit had been cut through by a later posthole [729], the fill of which (730) produced one sherd of Trevisker pottery in addition to RomanoBritish pottery. The Bronze Age material in this feature was almost certainly incorporated as a result of disturbance. No cereal grains were recovered from the structure and there was no surviving hearth-pit. Due to the uncertainty over disturbance to the site no material was submitted for radiocarbon dating.

Structure 712 Structure 712 comprised a partial ring of ten postholes (appendix 1, Table 26) disturbed and cut through by features associated with the Romano-British settlement, including the main enclosure ditch [565] (Fig. 12). The post-ring was incomplete in its north eastern quadrant, but it is likely that the posts originally formed a ring measuring approximately 7-8m in diameter. The surviving postholes were filled by stony silty clay fills and had diameters between 0.2m and 0.45m, and were between 0.12m and

Structure 392 Situated at the north western end of the alignment Structure 392 was the most complex of the Middle Bronze Age postring structures (Figs 13 and 14). The structure comprised two approximately concentric rings of postholes, with 17

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough

Figure 12: Middle Bronze Age Structure 712.

often contained large stones as post-packing and were filled by silty clay deposits. Structural postholes in the post-ring ranged between 0.15m and 0.44m in depth and 0.15m to 0.42m in diameter. The western edge of the structure had been cut through by the later Romano-British enclosure ditch [565], truncating outer ring postholes [783], [785], [787] and [789]. Posthole [325] contained fragments of a light reddish-brown piece of baked clay with wear suggesting

additional features between the post-rings and outside. A group of four closely-spaced postholes and pits was located just to the south of the centre of the structure (appendix 1, Table 25). The outer ring of the structure was approximately 8m in diameter, the inner-ring about 5m. Spacing between the outer-ring postholes ranged between 1.9m and 2.2m, while those of the inner-ring varied from 1.5m to 2.2m. The majority of postholes had vertical sides and flat bases; they 18

The Results

Figure 13: Middle Bronze Age Structure 392. that the material was being ground for use as colourant. Posthole [419] contained a wolframite stream pebble.

post-ring may have been associated with an entrance to the structure, ([511], [439], [321], [422], [314], and [316]). Postholes [439], [478] and [316] (PP19) contained sherds of Trevisker-related pottery. These features comprise pits and postholes and stratigraphic relationships suggest more than one phase of building or renewal of posts. Outside the outer post-ring were elongated shallow postholes [483] to the south and [485] to the north. Entrance posthole [485]

Wider spacing between the postholes of both rings on the south eastern edge of the structure, and the presence of external features, suggests an entrance from this side. A cluster of features on the south eastern side of the structure, between the inner and outer post-rings and outside the outer 19

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough

Figure 14: Photograph of Structure 392 from the southeast. contained one sherd of Trevisker-related pottery and a large broken muller S6. This object is likely to have been deliberately buried within the posthole (Quinnell below). Cut through [485], were postholes [501] and [563], and cut through [483], was posthole [480]. Just to the west of posthole [483] was posthole [487]. Together these features form an almost symmetrical arrangement on this side of the structure. Between these and the structure’s central features were [314], a small stakehole; [511] a posthole set between the inner and outer post-rings; [316] a shallow pit cut into posthole [504]; posthole [510] and deposit (508). Pit [401], between the inner and outer post-rings on the west side, contained four sherds of Trevisker-related pottery, a quartzite beach pebble and a stream cobble rich in cassiterite.

The central features were cut into by shallow pit [320], a probable hearth-pit. The primary fill (319) produced 44 sherds of Trevisker-related pottery including PP21-27, fragments of burnt bone from a sheep or goat (Higbee below), four quartzite beach pebbles were also found within this hearth-pit (Quinnell below) and a ripple-flaked flint knife. The latter is diagnostically an Early Bronze Age piece, and therefore possibly residual or curated and selected for deliberate deposition (Lawson-Jones below). The hearth-pit fills were charcoal-rich with, oak, hazel, hawthorn and willow/poplar used as firewood and also contained cereal grain (mainly emmer wheat) (Carruthers and Gale below). However, it is possible that some of this material could have been the result of contamination: the stratigraphy was very shallow and the structure lay within the Romano-British settlement, which produced a large cereal assemblage. A small amount of RomanoBritish pottery was recovered from Structure 392, which highlights this problem.

Steep-sided posthole-type features [602], [613] and [623], pit [618] and probable hearth-pit [320], which cut them, lay together near the centre of Structure 392. Posthole [623] contained in situ and disturbed post-packing stones, and charcoal from (622) within this feature produced a radiocarbon determination of 2964±40 BP, 1370-1040 cal BC (WK-15003). Adjacent to these was [618], a concave, bowl-shaped pit. Deposits in all features, particularly [602], [613] and [618], were noticeably rich in charcoal within the silty clay matrix. The top of posthole [602], fill (603), comprised a spread of stone. Features [618] and [613] produced 16 and 17 sherds of Trevisker-related pottery respectively. Posthole [623] contained charcoal identified as oak and gorse.

Adjacent to the hearth [320] was (318), a shallow spread of stony, charcoal-rich silty clay 0.05m deep. This incorporated five sherds of Trevisker-related pottery (including sherd PP20). A radiocarbon determination of 2947±43 BP, 13001010 cal BC (WK-15002) was obtained from the primary fill of the hearth-pit. In total 174 sherds of pottery were recovered from Structure 392. All are distinctive Trevisker-related ceramics likely to date to between the 13th and 11th centuries cal BC and 20

The Results Structure 232

some may have been deposited with a degree of formality (Quinnell below).

Close to the eastern extent of Area B a probable structure (Fig. 5) formed by shallow postholes was revealed but had been disturbed by roots and agricultural practices (appendix 1, Table 23). The overall length and width of the probable irregular, sub-rectangular structure was approximately 13m by 3m. It was formed by 10 possible postholes ranging from 0.3m to 0.8m in diameter and 0.1m to 0.27m deep, mostly with disturbed concave profiles. Fills were generally light brown clay silts with very few inclusions, with the exception of charcoal flecks and small fragments of naturally derived stone, but in places with clear evidence of in situ post-pipes and post-packing stones ([208], [244]). A single piece of pottery was recovered from (209), the fill of posthole [208], comprising a rim sherd of probable Trevisker pottery PP32. Although the postholes had been disturbed and showed little formal arrangement it is likely that they form a structure relating to the Middle Bronze Age phase of occupation.

Posthole alignment 37, 38 and 85 This comprises three groups of postholes on an approximate north – south alignment to the east of Structure 335 (Fig. 5 and Table 4). They may have formed a fence-line with an overall length of approximately 17m. The southernmost Group 37 comprised five postholes 0.3m to 0.5m in diameter and 0.17m to 0.26m deep, spaced 1.2m to 1.8m apart. Two postholes, [29] and [31], each produced a single sherd of Trevisker-related pottery, and posthole [29] also provided a radiocarbon determination consistent with the pottery of 3060±51 BP, 1440-1130 cal BC (WK14991). Charcoal and charred cereal remains recovered from the postholes in very small quantities included hulled and possibly naked barley and emmer/spelt wheat grains and weeds and tubers typical of grasslands or disturbed/ cultivated land.

Other features

Four metres to the north was Posthole Group 38 comprising a row of three postholes 0.28m to 0.38m in diameter and 0.19m to 0.33m in depth. To the west was feature [55], an amorphous cut, which was probably a result of modern disturbance or natural process. Postholes [51] and [39] were 1.4m apart, while postholes [39] and [40] were 2.1m apart.

A large pit and remnants of an old land surface were found to the north east of Structure 232 in Field 4 (Fig. 5). Pit [25] was oval in plan, measuring 1.15m long, 0.88m wide and 0.33m deep, and was filled with burnt stones, and charcoal. However, none of the burning had taken place in situ. It also contained large fragments from a Trevisker vessel (PP37), and a sherd from another (Quinnell below). A radiocarbon determination from pit [25] produced a Middle Bronze Age date of 3080±55 BP, 1460-1190 cal BC (AA-44602).

At the northern end of the alignment was Posthole Group 85. This comprised a tight arc of four postholes and a stakehole with a diameter of 0.3m. All the postholes were filled by similar dark orangebrown silty clay deposits which might indicate a degree of contemporaneity. Only occasional charcoal flecks were recorded from the fills (Table 4).

The old land surface (17) was found at the eastern end of the 2000 excavation area. Two substantial patches remained (Fig. 5), and were found to contain 29 sherds of Trevisker

Table 4: Posthole alignment 37, 38 and 85 details Context Width Length POSTHOLE GROUP 37 [23] [25] [27] 0.39m 0.48m [29] [31] POSTHOLE GROUP 38 [39] 0.3m 0.38m [40] 0.3m 0.38m [51] [55] 0.32m 0.82m POSTHOLE GROUP 85 [33] [35] 0.28m 0.4m [43] 0.18m 0.22m [47] 0.09m 0.18m [53] [83]

Diameter

Depth

Comments

0.36m

0.26m 0.22 0.6m 0.26m 0.17m

Posthole filled by (24), dark orange-brown silty clay. Posthole filled by (26), dark orange-brown silty clay. Natural feature or shallow depression filled by (28). Posthole filled by (30), dark orange-brown silty clay. Posthole filled by (32), dark orange-brown silty clay.

0.33m 0.33m 0.19m 0.3m

Posthole filled by (41), dark orange-brown silty clay. Posthole filled by (42), dark orange-brown silty clay. Posthole filled by (52), dark orange-brown silty clay. Filled by (56), natural or modern disturbance.

0.31m 0.25m 0.17m 0.37m 0.16m 0.25m

Posthole filled by (34), dark orange-brown silty clay. Posthole filled by (36), dark orange-brown silty clay. Posthole filled by (44), dark orange-brown silty clay. Posthole stakehole filled by (48), dark orange-brown clay silt. Posthole filled by (54), dark orange-brown silty clay. Posthole filled by (84), dark orange-brown silty clay.

0.4m 0.5m

0.28m

0.14m

0.18m 0.25m

21

ARCHAEOLOGICAL INVESTIGATIONS OF A LATER PREHISTORIC AND ROMANO-BRITISH LANDSCAPE AT TREMOUGH pottery (including PP35 and PP36) and 29 baked clay fragments from cylindrical loomweights (Quinnell below). A radiocarbon determination 3055±55 BP, 1440-1120 cal BC (AA-44603) was obtained from ditch [30] close to old land surface (17). The determination cannot be regarded as providing a date for the ditch which contained much later pottery. However, as it was sampled at a point where it crossed through old land surface (17) the determination is likely to relate to this material.

Area A and the western end of Area B were associated with circular timber post-ring structures, a pit group and a timber alignment (Fig. 5). The earliest phase of activity in this area was associated with the Early Bronze Age pit group and Structure 66 at the southern end of Area A dated to between 2200 cal BC and 1900 cal BC. This appears to have been followed by the construction of the more complex Structure 102 to the north east between 1530-1380 cal BC and concluded by the use of Structure 392 in the period 13701140 cal BC. If this was the case, it would appear that the complex developed from south to north and that there was growing complexity in individual elements of the site over time. It is also the case that far greater numbers of artefacts were deposited within Structure 392 than in Structure 66. However, caution is required as Structures 335 and 712 could not be radiocarbon dated, and the determinations from Structure 392 were from the central hearth-pit, which may not have been a primary feature. Furthermore, it appears that posts may have been replaced, which means that an earlier phase at Structure 102 may have been missed and activity in Structure 66 might have continued quite late. Structure 392 was most complex in terms of its architecture and the artefacts deposited within it, and this was located at the north western end of the post-ring complex.

Finally, a circular anomaly approximately 8m diameter was located by geophysical survey about 80m north west of the old land surface in Field 3 (Fig. 2). This feature was not excavated but is similar to anomalies which have been produced by sunken-floored roundhouses in other parts of Cornwall (for example, Jones 1998-9). It may therefore represent a roundhouse site. Environmental evidence Environmental analyses shed little further light on Bronze Age activity at Tremough (Carruthers and Gale below). Some evidence was provided for consumption of wild resources, including hazelnuts, during the Early Bronze Age, with greater amounts of cereals being found in later Middle Bronze Age contexts. Notably, these occurred in the greatest quantities in the hearth-pits within the more complex post-ring Structures 102 and 392, with the largest amount in the latter. Some caution is required as this Structure was located within the Romano-British enclosure; which also contained an assemblage of emmer grains. Although the excavated features did not provide much evidence for the surrounding environment, charred material from some them in the form of tubers and other macrofossils indicates a continuation of an open environment (Carruthers below). However, oak and hazel charcoal again suggests that timber was obtainable locally (Gale below) and suitable timbers for the post-rings were obviously available.

To the east, the remains identified were more ephemeral: a probable sub-rectangular structure, a remnant old land surface and an isolated pit. The structure is undated but may well have been contemporary with some of the post-ring structures to the west. If so, the difference in shape of the structure may well be significant and this will be discussed below. The remnant old land surface and the pit are both dated to the same period as the post-ring structures by radiocarbon dating and artefactual association. Again, these features are different from those to the west. The finds in the old land surface could represent midden material associated with prehistoric arable plots or fields – or may indicate that there is more intensive activity, such as settlement in the close vicinity. An unexcavated circular anomaly identified by geophysical survey may possibly represent the site of a sunken-floored Middle Bronze Age roundhouse, but without excavation the date of this feature is uncertain.

Discussion of the Bronze Age landscape The Bronze Age period saw the first monumental activity at Tremough in the form of timber post-ring structures and post-settings, the construction and use of which may have spanned the best part of a millennium between c. 2000 and 1100 cal BC. Although comparable post-rings have been identified in Cornwall at Early Bronze Age round barrows, few other free-standing post-rings of this period are currently known in Cornwall. Middle Bronze Age activity in Cornwall more typically takes the form of settlements comprised of sunken-floored roundhouses in the lowlands and stone-walled roundhouses in the uplands. The activity at Tremough does not therefore immediately fit comfortably with what is currently known of Early Bronze Age ceremonial sites or Middle Bronze Age settlements. These issues will be more fully explored below.

The evidence from the environmental analyses was limited but may suggest that during the Bronze Age the plateau remained cleared, and that some cereal cultivation took place in the vicinity. Comparable evidence for cultivation has come from other archaeological investigations of Middle Bronze Age settlement sites across Cornwall and southern Britain (Nowakowski 1991; Campbell and Straker 2003). Later Iron Age and Romano-British period (c. 350 cal BC-AD 400) The Later Iron Age

However, the evidence does seem to show differing activities across the site and spatial development over time.

During the 2000 excavations elements of a rectilinear field system were identified, comprising ditches [15] 22

The Results and [30], aligned north–south, and ditch [19] which was aligned east–west (Fig. 17). The ditches measured up to 0.9m wide by 0.5m deep. Their upper fills contained pottery consistent with a date during the currency of South Western Decorated Ware in the Later Iron Age, the fourth to first centuries cal BC (Quinnell below). Ditch [19] also produced a saddle quern fragment S7 in its fill (20). To the north east, excavations in Field 7 (access road) revealed a short section of east–west orientated ditch [7], possibly part of the same field system, which provided a radiocarbon determination of 2070±55 BP, 350 cal BC–AD 60 (AA44600). No associated Iron Age settlement focus was found, although it may have been located in the area of the ‘Fort’ at the eastern end of the plateau.

a small rim sherd of probable Oxfordshire ware. This material is likely to have been derived from the RomanoBritish settlement as a result of manuring and scattering of rubbish from the occupied site, perhaps reusing the earlier field system.

Ditch [5] was located within the access road (Fig. 17). A length measuring approximately 17m was exposed which was aligned north-east–south-west and was 0.92m wide by up to 0.17m deep. No artefacts were recovered but it may be part of the Later Iron Age field system located in Area B to the south, as the dark clay loam fill (6) was similar to those in the Iron Age ditches.

Structure 338

The Romano-British period Romano-British activity at the site comprised a ‘C’ shaped enclosure which was probably of two phases and a complex, circular post-built structure (Figs 19, 20 and 22). These appear to have been established at the western end of the Iron Age field system. A small number of pits and other ditches are likely to be associated with activity at this time.

Structure 338 was initially identified following topsoil stripping as a roughly circular area of dark soil (Figs 20 and 21). This overlying series of deposits (312), (416), (513), (515), (516) and (503) was removed by hand in a series of layers to reveal a complex of features including postholes, pits, gullies and hearth-pits. A table of the features within Structure 338 can be found in appendix 1.

Six sherds of unstratified Romano-British pottery were recovered from Fields 3 and 4 within the area enclosed by the Later Iron Age field system (Quinnell below), perhaps suggesting continuity in the use of the fields.

Ultimately, it became apparent that these features formed an oval structure measuring 8m by 6.7m on a north-east– south-west axis, its perimeter defined by an arrangement of postholes with interconnecting gullies on the north western and south eastern sides. Spacing between postholes was irregular as a result of the re-cutting and consolidation of original features. A space of 2m between two large postholes [754] and [725] with diameters of 0.61m and 0.65m respectively on the southern side of the structure probably formed the entrance. The remaining postholes were of variable depth with those forming the outer ring tending to be shallow, only [754] and [725] being deeper than 0.2m. At least two gullies, [793] and [713] appear to have been wall-trenches, since they link postholes and have stakeholes driven into their bases.

To the west, and investigated during the 2002 excavations, were the remnants of a brick-shaped field system comprising five sub-rectangular plots divided by ditches. These broadly correspond with linear features recorded in Field 3 (Lawson-Jones 2002). The ditches were shallow and narrow, no greater that 0.3m deep and 1m wide, and were orientated east–west and north–south (Fig. 17). Further west, the north-west–south-east orientated ditch [698], along with ditches [590] and [716] had an uncertain stratigraphic relationship with Romano-British enclosure ditch [565] (also shown as [306]), but probably formed the western extent of the field system. It is likely that agricultural disturbance during later periods truncated many of the ditches, although it is equally probable that they were never very substantial. The southern boundary of the field system formed by ditches [169], [173], [179] and [181] may have been incorporated into the later medieval field layout, since these fall on the line of medieval ditch [381]/[18] to the west and runs towards a ditch recorded during investigations in 2001 (LawsonJones 2002).

Internal elements comprised an irregular ring of steep-sided and flat-bottomed postholes ([440], [731], [748], [643] and [458]), 0.2m to 0.46m deep. Features [763] and [645] were similar, measuring 0.6m in diameter and up to 0.46m deep. Both were steep-sided and flat-bottomed and may have supported an internal structure. Two linear scatters of stakeholes were found between the two postholes and to the north of [645]; these appear to be truncated by a pit [611], and post-pad [627]. In the centre and south western quadrant of the structure was a complex of inter-cutting hearth-pits (cuts [680], [773], [628], [626], [433] and [429]) from which extended an elongated gully [466] on the south western side, perhaps a flue [466] connecting hearth-pits [773] and [743]. In the southern corner of hearth-pit [743] was a well-defined stone packed posthole [754]. These pits had depths between 0.25 and 0.45m and varied in shape from oval cut [680], which measured 0.75m long and 0.54m wide, to the somewhat amorphous

The brick-shaped field system recorded during the 2002 excavations (Figs 17 and 21) is probably the western extent of the Later Iron Age field system suggesting that the Romano-British Structure 338 and associated enclosure were added at a later date to the west (see below). No dating evidence was recovered from the brick-shaped fields in this area. All Roman material in the area of the field system came from unstratified contexts and included a Type 20 bowl, body sherds of BB1 and 23

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough porphyritic elvan (Quinnell below). Also recovered from this posthole was S9, a slab with a depression, perhaps used as a spindle base, and S10, a whetstone (Quinnell below). Other stones appeared to ‘seal’ this feature, reinforcing the conclusion that the objects had been deliberately buried. The posthole’s position described above as forming one side of the entrance may have influenced this ‘structured’ deposit.

[674], measuring 2.6m long and 1.3m wide. The edges of the hearth-pits had been scorched by intense burning and fills contained significant quantities of charcoal. Shallower pits with concave profiles ([750], [343], [690] and [727]) could be seen in the southern half of the structure to either side of hearth-pit [773] and may have been associated with the hearth-pits. Deposits in postholes and gullies generally comprised mid reddish-brown silty clays with varying proportions of stone and charcoal. The internal hearth-pits and other pits contained larger quantities of charcoal.

Phase 2 appears to belong to the early third century AD, and comprises a rather mixed assemblage but including diagnostic forms such as a Type 23 plain-rimmed bowl PR6 from pit [611] (late second century to fourth century) and a storage jar sherd from posthole [669] which has incised decoration typical of storage jars probably introduced sometime in the third century. Posthole [339] produced pottery PR5 consistent with this phase 2 activity in the third century AD and a beach cobble utilised as a whetstone.

The recutting of existing postholes – [768] by [669], [766] by [572], [725] by [770] and [659] by [661] – suggests several phases of activity. Posthole [611] was backfilled and cut by pit [645], and hearths [773] and [680] were stratigraphically earlier than hearth-pit [626]. Posthole [731] was backfilled and then cut by pit [690] which contained a number of fragments of burnt granite which may have derived from one of the hearths.

Continued use (phase 3) to the late third century is indicated by a Type 22 rim (gully/hearth-pit [743]) from a Cornish flanged bowl almost certainly first introduced in the late third century AD. Vessels PR7 and PR8 from pit [343] are consistent with this date, the Type 12 form of PR7 not being produced after c. AD 300.

Posthole [431], which perhaps consolidated the roof structure, the stone post-pad in the top of backfilled pit [611] and hearth-pit [674] represented the latest phase of activity. During this phase the hearth area was extended and a stone-setting placed within the hearth-pit to contain the fire.

Pottery recovered from a deposit overlying the structure and including Type 22 Cornish flanged bowls PR12 conforms to a date range from the late third into fourth century AD (phase 4). This deposit (mixed with (416) and (503)) overlaid all features belonging to Structure 338 and had the appearance of a disturbed cultivation soil, although it may represent a deliberate ‘closure’ of the structure subsequently disturbed by agriculture. It suggests that the structure went out of use in the early fourth century AD. Deposits overlying the structure contained pottery PR9, PR10, PR11, PR12 and PR13, and S13, a fragment of stone, a mortar perhaps subsequently used as a smelt pot (Quinnell below), and above this S14 and S15, a whetstone and a fragment of greisen rotary quern respectively. The rim form of S13 suggests manufacture in the third or fourth centuries (Quinnell 1993, 33).

Two radiocarbon determinations were obtained. Hearthpit [628] provides the earliest date of 1844 ±38 BP, cal AD 70-260 (Wk-15004), while later hearth-pit cut [674] gave a date of 1770±38 BP, cal AD 130-390 (Wk15005). At least three phases of activity were suggested by stratigraphic relationships between features recorded during excavation of the structure, as tentatively outlined by the Phase 2 archive report (Gossip 2003a). Closer examination of the stratigraphic record, the analysis of the ceramics (Quinnell below) and the radiocarbon dating uphold these relationships but it is sensible to regard activity within the structure over a period of around 150 years as being a more or less continuous process of consolidation and maintenance rather than as distinct phases of construction or remodelling. Reuse of the structure, perhaps for a variety of functions, has compounded the mixed nature of the artefact assemblage. The abandonment process may also be complex, and it is possible that the structure was deliberately sealed with soil and rubbish rather than simply left to decay.

Plant macrofossil assemblages from the structure contained emmer, spelt and bread-type wheat, hulled barley and oats, mostly from hearth-pit deposits. Chaff present in the hearth-pit deposits suggests cereal processing waste was also being used as fuel or tinder for the fire. Other species varied little from the Bronze Age deposits, with a similar range of grassland/disturbed ground, weed and wild food taxa represented. The post-abandonment layer produced very few cereals, no chaff and just a few grassland weed seeds and tubers.

The pottery assemblage suggests that there is an initial phase of activity in the late second century, the most distinctive forms being Type 10 bowl sherds from posthole [336] (PR1 and PR2), posthole [645] (PR4) and gullies [466] and [519] (PR3). A posthole thought to belong to this initial phase of activity [754] (755) contained a worn but intact mensuration weight S8 made of fine-grained

Samples from hearth-pit fills also indicated the consistent use of narrow roundwood, predominantly from oak and gorse and/or broom, with some birch, hazel, willow and 24

The Results bowl PR22 from deposit (752), an upper fill of the ditch, was curated or had been redeposited from an earlier fill into which the ditch was cut, then all dates could be placed within the third century AD. Stone artefacts recovered from the ditch include two possible whetstones, part of a saddle quern, and iron nail fragments. The radiocarbon determination 1910±42 BP, cal AD 0-220 (WK-14990) was obtained for the ditch, but since the charcoal from which this date was derived may be redeposited it can only be taken as providing a terminus ante quem.

poplar. Hearth-pit fuel deposits confirmed that oak was still the major source of firewood in the Romano-British period but there was a marked increase in gorse/broom, perhaps suggesting the preferential selection of this type of fuel or changes in the vegetation composition. This might indicate that rough ground was present nearby. Enclosure ditches Two curvilinear enclosure ditches were recorded towards the northern extents of the Phase 2 excavations (Fig. 19). Ditch [306], believed to be the earlier, formed a ‘C’ shaped enclosure, open on its southern side and measuring 48m by 28m, with an area of 0.08 hectares. To the west the ditch had been superseded by ditch [565], although the base of the original ditch [306] was recorded in its base as [679] – recognizable as an earlier cut for almost its entire length. A short section of ditch, [22], may represent a remnant of the ditch along its southern edge, the enclosure perhaps forming more of a complete circuit prior to truncation by later features and modern agricultural activity. The width of the ditch ranged between 0.7 and 1m at the top of the cut, narrowing to a concave base 0.3m wide. The maximum depth was 0.4m and it was filled by a homogeneous brown silty clay suggesting gradual silting and erosion during its use. Sherds of Type 4 jars PR14 recovered from the ditch fill indicate a terminus post quem in the late second century AD. A Cordoned sherd and a necked rim sherd unlikely to date after AD 150 may have been curated prior to deposition. A single hobnail was also recovered from the ditch fill (307).

The phase 2 enclosure ditch [565] would have produced a significant amount of upcast, the stone and clay of which was presumably incorporated into a bank. Although none of this bank material survived, the formation of deposits within the enclosure ditch suggested that a bank was constructed on the inside of the ditch. If this was the case it can be assumed that Structure 338 would have almost abutted the bank, assuming that the two were contemporaneous. The small, shallow gully [519], attached to the end of [565] to the north of Structure 338 is unlikely to have provided material for a bank, and may have functioned as a drainage channel running into the main ditch. Parallel gullies [596] and [310] projected southwards from the south western terminus of ditch [565], cutting through Middle Bronze Age Structure 712, obliterating elements of it. This almost certainly explains the presence of Middle Bronze Age finds in gully [596], including two sherds of PP33, and one sherd of PP34 Trevisker-related pottery and three flint flakes, as well as Romano-British gabbroic sherds. The gullies may represent a structure, perhaps a bank, defining the entrance to the enclosure. Pit/posthole [729] just to the north of the gullies is probably associated. This too contained both Trevisker-related and RomanoBritish pottery (Type 20 bowl); it cut pit [600], possibly part of Structure 712.

The phase 2 ditch [565] re-cut the western half of [306], forming a smaller ‘C’ shaped enclosure measuring 25m by 28m, open on its southeastern edge and enclosing an area of 0.03 hectares. This ditch was wider, being 1.58m at the top, and up to 0.7m deep, with a very steep profile to a flat base between 0.3 and 0.4m wide. The ditch became deeper as it ran anti-clockwise. Fills were a sequence of mid or dark grey-brown clay silts with large quantities of angular and rounded granite in places, and indicate the gradual silting and erosion of the ditch and bank from its upcast, with perhaps a more rapid final and deliberate infill event. The three shallow linear ditches [698], [590] and [716] radiating northwards from the enclosure ditch did not have firm stratigraphic relationships with the enclosure ditch and their phasing is uncertain. It is possible that they represent a remnant of the western extents of the earlier brick-shaped field system.

Other Romano-British features A number of isolated features were found in addition to those already described associated with the Later Iron Age field system and Romano-British settlement in Areas A and B. These included some deep pits of uncertain function in Areas A and B and fragments of a ditched field system in Fields 3 and 7. These latter features were exposed during the construction of the access road in 2002, a car-park in 2000 and by evaluative trenching in 2003 (Lawson-Jones 2002; Gossip 2003b; 2004).

Notable finds included sherds of a very large Type 13 jar PR15 from ditch fill (566) which was possibly manufactured in the later second century AD. In total 193 sherds of pottery (PR 15-22) were recovered from the ditch fills of the later cut [565], many of which have fresh breaks and appear to have been covered very soon after deposition. The date range for this pottery is late second century to late third century or very early fourth century AD. However, if the abraded, late second century Type 19

Areas A and B Four pits containing Romano-British artefacts were located outside enclosure [306] (Fig. 17). Pit [424]/[525] was situated 15m to the south of the enclosure. It was 3m in diameter and approximately 2m deep and filled with numerous dumped deposits. Twenty two sherds of Romano-British pottery were recovered from these 25

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough Age or Romano-British date. The final ditch in Field 7, [1] was located to the southeast of the ‘Fort’. It was up to 0.86 m wide by 0.22m deep. A single sherd of RomanoBritish pottery was recovered from the silty clay fill (2). The evidence from ditches in Field 7 suggests that they were in use during the Later Iron Age or Romano-British periods.

deposits, including a sherd of Samian Ware that had been re-used as a spindle whorl. Type 3 and 4 jar rims PR 24 – PR27 and a well-made Cordoned sherd indicate a later second century date for the infilling of this feature. Pit [377]/[470] was a small shallow pit a short distance to the west of pit [424], measuring 0.8m by 1.1m and 0.2m deep. Within fill (379) were 70 sherds of pottery sealed by layer (378), a greyish white gritty clay silt with small eroded granite inclusions. The sherds had been laid with their outer surfaces uppermost, and represent 80-90% of a Type 4 jar (PR28) (Quinnell below). These were apparently deliberately smashed, perhaps by the beach cobble hammerstone S16 recovered from the same deposit. The cobble appears to have been used for a repetitive percussive process, probably breaking-up or crushing stone (Quinnell below). A radiocarbon determination was obtained from the deposit 1926 ±46 BP, 40 cal BC–AD 220 (Wk-15000).

The rectangular enclosure (known as the ‘Fort’) identified by geophysical survey was not affected by the development at Tremough, but two evaluative trenches have recently been excavated across the enclosure ditch and into its interior (Fig. 18). These confirmed that the ditch was a substantial feature, nearly 4m across and over 2m deep. The trenches produced evidence for a stone-walled structure and substantial postholes, possibly associated with timber structures were also uncovered. The minimal pottery assemblage from the internal features may include an Early Iron Age component, as well as Cordoned Ware dating from the first century AD to second century BC, suggesting that the enclosure may have been a focus for occupation during both the Iron Age and the RomanoBritish periods (Gossip 2006; Gossip and Jones in prep).

Pit [185] was located 62m to the east of the RomanoBritish enclosure [306], and pit [217] in the corner of Area A, some 45m to the south west of pit [424]. Each of these pits measured approximately 3m in diameter, and they were 1.25m and 1.6m deep respectively. They had been backfilled with a succession of dumped clay silt deposits, some of which were extremely rich in charcoal. The pits had been cut through the natural subsoil and bedrock and their function is uncertain.

Discussion of the Later Iron Age and Romano-British landscape Activity during the Iron Age was only identifiable by the brick-shaped system of fields in Fields 3 and 4 and Area B. It is likely that these fields were established and worked during the Later Iron Age since a ditch in Field 7 provided a radiocarbon determination of 2070±55 BP, 350 cal BC-AD 60 (AA-44600). Whatever the exact date of their formation, it is likely that they were open for some time and it is probable that although the ditches may have silted up by the end of the Iron Age, they were still in use and defined by hedges or banks. There is no definite settlement contemporary with the Later Iron Age fields, but it is possible that the field system was associated with the ‘Fort’ in the eastern area of the site. This interpretation is tentative, since most of the enclosure remains unexcavated and the character of the occupation within it is uncertain.

On the basis of similar morphology the two may be contemporary with pit [424]. It is possible that they had been dug in order to quarry bedrock, and subsequently used as rubbish pits for organic waste. Field 7 Four ditches were located in Field 7 during small-scale interventions (Fig. 18). Located at the southern end of the ‘L’ shaped trench was a single ditch [7]. The cut measured 1.1m wide and was up to 1m deep. It contained two fills, which were securely sealed beneath plough-soil layer [107]. The upper fill (8) was a firm, compacted clay silt, almost certainly naturally produced during the ditch’s final phase of silting. Basal context (9) was a sticky, compact silt, which had formed while the ditch was functioning as an open drainage ditch. Charcoal contained within context (9) produced a radiocarbon determination of 2070±55 BP, 350 cal BC-AD 60 (AA-44600), suggesting a Later Iron Age or possibly a very early Romano-British date for this ditch. A probable southern extension of this ditch, numbered [102] was recorded during monitoring work in 2003.

Later Iron Age activity is not represented at the western end of the site where the ‘C’ shaped enclosure and Structure 338 provide a firm Romano-British date. Structure 338 appears to have been constructed at some point in the late second century AD, and this may be contemporaneous with the initial, larger phase of enclosure ditch [306]. Evidence from the pottery assemblage suggests that this ditch was fairly quickly replaced by a smaller curvilinear ditch [565], probably at the very end of the second century or into the third century AD. Structure 338 seems to have continued in use from the late second until the late third or possibly into the fourth centuries AD. It is possible the Iron Age field system saw continued use into the late second century; although all Romano-British finds from the field system were unstratified, this is likely to represent

In addition, a section of ditch [104] on the western side of the ‘D’ shaped enclosure attached to the ‘Fort’ was recorded. This feature proved to be 1.3m wide by 0.8m deep and was filled by (105), a brown clay silt. No finds were recovered but it seems likely that it is of Later Iron 26

THE RESULTS from the fill (4). Several sherds of early medieval grassmarked ware (Quinnell below), broadly contemporary with the radiocarbon date, were mostly recovered from unstratified contexts in Field 4.

manuring of these fields with refuse from the settlement and enclosure. Given the small dimensions of Structure 338, an internal area of no more than 35m2, it is perhaps surprising that it was so long-lived, perhaps as long as 100-150 years. Rather than successive large-scale phases of construction the evidence suggests the structure saw frequent consolidation or patching. Internally the structure saw much activity, with the burning of successive hearth-pits and the excavation and backfilling of pits.

Excavation work did not reveal evidence for the medieval settlement at Tremough. This may be due to the effects of fairly extensive landscaping during the post-medieval and later periods. However, if the present Tremough house is on much the same site as its medieval predecessor, which could have been established as early as the sixth or seventh centuries, this would explain why no evidence for the early settlement was found. It is interesting to note that the few recorded ditches which were associated with medieval fields, such as ditch [19] and the recorded hedge boundaries (Lawson-Jones 2002) appeared to show little continuity with the later prehistoric and RomanoBritish field system, implying that there was a major reorganisation of the agrarian landscape between these periods.

The increasing importance of arable cultivation is borne out by the cereal grain assemblage, and indirectly by the wood species. The Romano-British period saw the introduction of spelt alongside emmer wheat, while to a lesser extent bread-type wheat, hulled barley and oats were also being grown. The presence of cleavers has been shown to be an indicator of autumn sown crops, indicating perhaps a greater reliance throughout the year on cereal production. During this time it is possible that there was a more open landscape with more extensive scrub, suggested by a charcoal assemblage showing an increased occurrence of gorse, broom and oak and hazel roundwood; although this could also indicate management of hazel coppicing or its use in hedgerows. Hazel, gorse and broom typically colonise disturbed or more open land, indicating a scrubbier and conspicuously less wooded landscape in the Romano-British period than in the Bronze Age, a process that began with the establishment of fields in the Later Iron Age (Carruthers and Gale below).

Creation of the ornamental parkland landscape in the late eighteenth and nineteenth centuries would have reduced areas available for cultivation and this may have helped to preserve the earlier ploughsoil, which was found across much of the site. Traces of medieval strip fields also seem to have survived in Field 3, presumably from a time when Tremough was covered by sub-divided open arable fields. The possible strips take the form of changing levels in the subsoil, revealed by topsoil stripping along the southern edge of the field (Lawson-Jones 2002).

The early medieval and later periods (AD 410-1540)

A significant quantity of medieval artefacts (AD 1066 to 1540) was found, especially as unstratified finds within Field 4. Their presence illustrates a continuation of domestic settlement and activity within the immediate area. The collection of medieval artefacts typical of kitchen midden disposal (Thorpe below) indicates that there was nearby settlement.

Few excavated features were found to be of early medieval date and archaeological evidence for the period consists of a small number of sherds of pottery and a radiocarbon date. Ditch [3] was located in the haul road between Fields 3 and 7 and was aligned north west to south east (Fig. 17). The cut measured 0.45m wide by 0.1m deep, and was exposed for a length of approximately 20m. There was a fork at the southern end of the cut. The eastern fork was numbered [106] but no stratigraphical relationship could be established, as both cuts were filled with identical pale-brown clay loams. A radiocarbon determination of 1015±45 BP, cal AD 890-1160 (AA-44599) was obtained

The pottery from Tremough is predominantly standard local domestic wares, but a small fragment of glass, thought to come from a medieval goblet or beaker (Tyson below), is more unusual and may hint at the high status settlement which is indicated by the historical sources (Mattingly above).

27

CHAPTER 3

INTERPRETATION

In the beginning

The investigations at Tremough were important as they provided an opportunity to investigate a self-contained block of land and record changes in the way it was used and organised from the onset of the Neolithic into the historical period.

In common with much of southern England, there was a dearth of evidence relating to settled Neolithic occupation in terms of cultivation or house-type structures (see papers in Cotton and Field 2004; Thomas 1999, 9-10). Across Cornwall there are a growing number of Early Neolithic flint scatters and pit sites with associated radiocarbon determinations which do not appear to be associated with permanent settlements (see below and Cole and Jones 20023; Jones and Reed forthcoming). None of the excavated areas appeared to provide any evidence for permanent settlement. It could be suggested that the lack of evidence for long-term permanent settlement activity could exist outside the excavated areas. However, a substantial part of the plateau has been investigated and there is absence of environmental evidence for sedentary occupation; for example, no evidence for cereal production was found and wild food species continued to be used.

The discovery of features dating to the Neolithic, Bronze Age, Iron Age and Romano-British periods at Tremough enables the study of one of the longest archaeological sequences in lowland Cornwall. The emerging picture is of a topographically well-defined block of landscape containing evidence of both continuity and change, seeing successive groups of people redefining space which during some instances refers back to the past but at others clearly related to contemporary practices in the wider landscape. One of the dominant themes to emerge from all periods in this study is evidence for structured deposition – the way people chose to discard their artefacts and other materials. While it is probable that for much of this time these practices are likely to have been influenced by underlying beliefs associated with religiosity (Insoll 2004), it is argued here that most were carried out routinely as part of an accepted, ritualized way of behaving. This model of domestic and ritual life as superimposed layers of prehistoric life has been explored by several writers (Bradley 2005; Hill 1995), and this has guided the way that we have interpreted the deposition of discarded items at Tremough from the Early Neolithic through to the Romano-British period.

At Tremough, Pit [21] produced a radiocarbon determination of 4850±55BP, 3770-3510 cal BC (AA44601) broadly consistent with other Early Neolithic pit sites in the county. The nearest of these is at Poldowrian on the Lizard. There Early Neolithic pottery and flints were found in association with a pit group and two Neolithic determinations 5180+150 BP, 4350-3650 cal BC (HAR4323) and 4870+130 BP, 4000-3350 cal BC (HAR-4052) were obtained on bulked samples which contained oak (Smith and Harris 1982, 49). A Neolithic date of 33302910 cal BC was obtained from one of a group of charcoalrich pits at Trenowah, on the St Austell Distributor Road, and excavations at Tregarrick, Roche revealed ten pits arranged into three clusters with associated dates ranging from 3790 cal BC to 3370 cal BC (Cole and Jones 20023). Most recently, four fourth-millennium radiocarbon determinations have been obtained from pits containing charcoal, hazelnut shell, sherds of pottery and some cereal grains at Portscatho: 4952±45 BP, 3910-3640 cal BC (Wk-13258), 4818±48 BP, 3710-3380 cal BC (Wk13256), 4805±51 BP, 3700-3380 cal BC (Wk-13257), 4713±51 BP, 3640-3370 cal BC (Wk-13259) (Jones and Reed forthcoming).

The following discussion is therefore arranged in chronological order but within each period there are a number of recurring themes which include the character of settlement and zonation of activities, the role of placed deposits, the influence of the past and cosmology, and the role of social memory. Neolithic Two distinct phases of Neolithic activity were identified, separated by perhaps more than half a millennium. The first occurred during the Early Neolithic from c. 3900 cal BC to 3500 cal BC and the second during the Later Neolithic c. 2900 cal BC and 2300 cal BC. Both were associated with pits although the character of the deposits differs between the two periods.

The radiocarbon dating of Cornish pit sites suggests a widespread tradition of pit digging between c. 3700 cal BC and 3300 cal BC; these were often associated with the burial of pottery, lithics, hazelnut shells and charcoal. Although filled with apparently domestically originated 28

Interpretation material, the nature of many of the deposits found at Tregarrick and Portscatho (Cole and Jones 2002-3; Jones and Reed forthcoming) illustrate how the deposition of domestic refuse may have been given become more than simple rubbish disposal. It is possible that this routine, even subconscious method of ‘clearing up’ was part of an accepted way of acting following short-term occupation of a site. The concept of domestic and ritual activities as being inextricably entwined practice has recently been explored by several writers (Bradley 2005; Richards 2005).

have linked with group status or membership of the community. The contrasting lithic scatter areas may have been created by groups of people who seasonally occupied the site on a temporary basis. It is possible that the plateau was initially cleared for grazing by transitory pastoralists, or that its distinctive topographical setting and views over the Fal estuary made it a natural focal point for groups to meet and exchange flint and perhaps other perishable items, which have not survived, such as animal hides or harvested plants.

However, what is striking about the five Early Neolithic pits at Tremough is the lack of formality in the deposition of artefacts, which is widely found elsewhere across southern England (Thomas 1999, 64-74; Garrow et al 2005; Cole and Jones 2002-3). Indeed, they appear to have been shallow pits, devoid of the usual range of artefacts and perhaps have most in common with the charcoal-rich pits at Trenowah. Instead, there was a greater spatial dimension to the way in which different kinds of archaeological features and flint resources were spread across the site as a whole, which suggests that activities may have been ‘zoned’.

And at the end The next visible phase of phase of activity occurs some 500 years later, from c. 2900 to 2300 cal BC. This gap is significant, as to date no Middle Neolithic Sites are known in Cornwall (c. 3300-3000 cal BC) and the Peterborough ceramics which were current elsewhere during this period are more or less absent from the county. However, it is likely that this apparent gap is largely the consequence of a lack of immediately diagnostic artefacts and that it will be narrowed or closed by further programmes of radiocarbon dating.

It has been suggested by a number of writers that activities within sites (Healy 2004; Pryor 1998) and even landscapes (Hey 1997; Parker Pearson 2000) may have been structured so that different places became associated with different practices and experiences. At Tremough this argument could be supported by the physical separation of different types of archaeological feature and perhaps by contrasting spreads of lithic working debris.

The Later Neolithic in southern England is traditionally characterised as a period during which large-scale monuments such as Avebury and Durrington Walls were constructed and of social feasting associated with the consumption of large quantities of pig meat and the use of Grooved Ware (Wainwright 1971; Parker Pearson 1993, 72; Barrett 1994, 145). Until recently, evidence of Later Neolithic activity in Cornwall was largely confined to a handful of monuments dated to this period on the basis of their form (for example, small Class I henges at Castilly, Castlewich and the Stripple Stones) and a small number of sites which had produced Grooved Ware, often from uncertain or residual contexts (Jones 2005, ch 2). More recent archaeological investigations have began to redress this balance and have produced a growing body of Grooved Ware sites and these are beginning to be better understood in terms of both date range and context (Nowakowski forthcoming; Exeter Archaeology 1997). A picture is starting to emerge which is perhaps more comparable with other regions where monuments such as henges are scarce (Bradley et al 1993; Peterson 2004) and Grooved Ware was instead deposited into non-monumental contexts, such as small pit groups.

Evidence for the zonation of activities is perhaps suggested by ditch [76], which has an Early Neolithic radiocarbon determination of 4995±50 BP, 3950-3660 cal BC (AA44604). Although only exposed for a short length, it may have cut off the narrow eastern end of the plateau from the rest of the site. It appears to have demarcated a division in lithic resources and in the distribution of features, as illustrated by the finds from Field 3 and Field 7. To the east of ditch [76] the primary evidence for Early Neolithic activity consists of an area of delicate in situ nodular flint knapping and a second, probably concurrent area of ‘heavy duty’ nodular core reduction (Lawson-Jones below). By contrast, to the west a local beach flint assemblage was found which was probably contemporary with the scatter of pits. The subdivision of the area suggests a complex pattern of zones appropriate for different knapping activities which may have been determined by the accepted traditional practices of different groups who used the site. The ditch was not deep but it could have served to mark a distinction between those who had access to and were able to work ‘imported’ nodular flint from elsewhere, and those who had to rely upon the availability of local flint pebbles for their tools. Hearth-pits could be interpreted to represent the fires of visitors to the site who perhaps had less access to the eastern end of the plateau. Such divisions could

The Grooved Ware assemblage from Tremough fits this general pattern and is significant because it forms the largest assemblage yet discovered in Cornwall, constituting sherds of about 21 of the approximately 40 vessels so far recorded. Two radiocarbon determinations were obtained from the Grooved Ware pits. That from pit [331] dated to between 2870-2480 cal BC with pit [494] yielding a date of 2570-2290 cal BC. Few other radiocarbon dates associated with Grooved Ware are available from Cornwall. Two 29

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough and/or freshly broken, and five contained items of worked stone (S3 and S4, SF1029, SF1051, SF1045, SF1055 and SF 1056). Layer (496) was a shallow spread of material that contained a round glossy pebble S5 with extensively worked edges and attractive white and black marbling in conjunction with large decorated Grooved Ware sherds PP15, PP16 and PP17, which appear to have been fragmented prior to deposition. The burial of conjoining broken sherds occurs in several of the pits, namely [193], [293], [300], [329], [331], [494]. Several of the sherds showed signs of abrasion suggesting curation for a long time prior to deposition (see Quinnell below). Two of the Grooved Ware pits [193] and [331], and pits [296], [298] of possible Later Neolithic date also included hazelnut shells and pignut tubers which were probably food residues.

pits at Trevorva Cott, near Probus gave determinations between 2880-2208 cal BC (J Nowakowski, pers. comm.) and another from a pit at Trevone calibrates to 2900-2610 BC (Quinnell below). The primary phase of Davidstow site 22 produced two sherds of Grooved Ware and gave a determination of 4130±70, 2890-2550 cal BC (HAR-6643) (Christie 1988). In addition to these dates, a radiocarbon determination falling between 2950-2050 cal BC was obtained from a pit beneath a mound at Poldowrian which contained a possible undecorated Grooved Ware vessel (Harris 1979). These determinations fall within the nationally accepted date range for Grooved Ware use, which is currently dated from approximately 2900 to 2300 cal BC (Garwood 1999). A small spread of material was recorded which might represent the last vestiges of an occupation-related deposit. It contained 33 sherds of Grooved Ware, including vessels PP15, PP16 and PP17, a polisher/grinder, a vein quartz cobble and four flints. However, there was little to suggest permanent settlement activity and even traces of simple structures such as those found in Wales were absent (Britnell 1982; Gibson 1999, 36-42). The recovered charred plant materials were devoid of cereals but did contain hazelnut shells. If the radiocarbon determinations are taken at face value low-key activity may have occurred over several centuries between c. 2800 and 2300 cal BC. In this context, it seems likely that these pits were created by groups of people who only temporarily or sporadically occupied the site. The limited environmental evidence indicates that a grassland habitat predominated at the time. It is likely that some grazing took place and the pits were probably created by occasional visitors to the site, revisiting this part of the plateau as part of the seasonal round.

The nature of the deposition of artefacts within the pits does not preclude them from having had a primarily ‘domestic’ function. Indeed, it is highly probable that many of the vessels had been used and broken before they were deposited, while others such as the hazelnut shells had probably been generated by the daily consumption of food. Again, it is likely that distinctions between everyday activities and ritual practices, or the sacred and the profane, were far less clearly drawn in prehistoric societies than they are in the modern, post-industrial world (Hill 1995; Bradley 2005). Everyday objects of prehistoric life could perhaps have achieved some kind of value as a result of their own biographies and memories, as they became material fragments reminiscent of people, places and activities (Edmonds 1995; Woodward 2002b). Certainly two of the Grooved Ware vessels PP6, and PP11, may well have been broken long before burial and curated, whereas fresh nodular flint may have been associated with distant places.

The possible zoning of activities apparent in the Early Neolithic is not reflected by the Late Neolithic pits, although it is interesting to note that the area of the site with the greatest concentration of Early Neolithic features and artefacts was apparently devoid of Later Neolithic activity. The eight to ten pits in this spatially tight grouping at the western end of the plateau shared the same general characteristics, being shallow, concave features containing a range of structured depositions of Grooved Ware pottery, nodular and freshly broken flint-work and worked stone within single episodes of backfilling. The deliberate selection and deposition of these objects is comparable with other Cornish Later Neolithic pit assemblages (Nowakowski forthcoming) and supports the argument that ritualized depositions continued to be made in Cornwall throughout the Neolithic period (Cole and Jones 2002-3, 134, Jones 2005, 137-8).

In addition to having functional uses, decorated Grooved Ware pottery, flint and stone tools may well have carried symbolic associations – as indeed could ecofacts such as hazelnut shells and charcoal. It should not be assumed that a discarded or buried object needed to hold ‘value’ in the modern sense of the word to enable it to be deployed in ritualized patterns of discard (Moore 1986, 111-15). Anthropological study of non-industrialised societies has shown that objects can be categorised and given meanings (Douglas 1966) and that items can inherit symbolic significance regardless of their ubiquity or the energy that would be needed to replace them (Moore 1982). However, whereas objects such as the pottery and distinctive pieces of flint and stonework could have been placed into the pits with a degree of formality, it is perhaps more likely that charcoal from fires and food residues such as hazelnut shells would have entered the pits more casually, being dumped into the open pit as part of more routine activity associated with the structured clearance of deposits which had been generated by the process of inhabiting the site. It should also be noted that at Tremough not all of the occupation material was incorporated within pit fills – a

It can also be argued that many of these artefacts have a distinctiveness of colour, shape, material or feel that may have been influential in their selection for deposition (see papers in Jones and MacGregor 2002). Eight of the pits contained decorated pottery PP1-17, seven included one or more flint objects many of which were finely worked 30

Interpretation The post-rings

small spread of material (496) that had not been cleared away survived near to the pit group. This might imply that it was not imperative that all residues associated with domestic-related activity were cleared away. The spread itself may represent the last remnant of a midden or marker mound, where occupation-related material was stored until the time was right for burial.

Neolithic and Early Bronze Age timber monuments have usually been discussed in terms of the ceremonial functions they may have performed or the ritual activities which may have taken place within them (Gibson 1998; Pitts 2000, 256-9). By contrast, post-rings dating to the Middle Bronze Age are usually taken to be roundhouses – the reason being that there is good evidence for a large expansion in settlement activity at this time across southern Britain and they leave exactly the kind of signature in the ground that a roundhouse would (Moore and Jennings 1992; Barrett et al 1991; Johnson and Rose 1994). At the same time, the rise of clearly visible settlement activity often provides a watershed for the way in which the Bronze Age is talked about. In short, the second half of the Bronze Age has frequently been discussed in terms of agricultural expansion and settlement organisation (Ellison 1981; Fowler 1983; Bowen and Fowler 1978). However, Joanna Brück has pointed out that Middle Bronze Age settlements are far removed from being simple unproblematic functional agricultural units and that activities such as the deposition of human bone, which would traditionally have been associated with ‘ritual’ activity also took place within them (Brück 1995; 1999). More recently, Richard Bradley (2005, 73-5) has questioned the wisdom of separating prehistoric ritual and domestic spheres at all, and has pointed to the links that existed between different kinds of later prehistoric structure. With this in mind, the first question to be raised is where the post-rings fit within our knowledge of Cornish Bronze Age structures.

The important point is that the Later Neolithic pits at Tremough are among a growing body of Cornish Neolithic sites which provide evidence for the controlled discard of materials generated by communities, a practice which is likely to have been considered the accepted way of ‘doing things’. Summary Currently, there is a division between those who view the Neolithic as associated with the establishment of sedentary societies (for example, Darvill 2004), and others who argue that the change was more ideological and that communities continued to be mobile in nature (for example, Thomas 1999). In actuality these models need not be mutually exclusive, as they are to a large extent based on evidence drawn from different areas (Bradley 2003) and it is likely that there was a high level of regional variation across Britain. At Tremough, there is no evidence from the excavations for permanent or long-term settlement during the either the Early or Late Neolithic periods. The Early Neolithic pits, ditch and flint scatters do not appear to have been associated with houses, fields or cultivated areas and the picture that emerges is one of transitory groups occupying the site for short periods of time. During the Later Neolithic activity shifted westwards but if anything occupation-related activity appears to have been on a smaller-scale. However, by this time activity is found in the form of structured deposits being placed into pits. The 500 year or so gap between the two phases means that it is unlikely that there was a straightforward sense of continuity between the two phases beyond – if the area remained cleared – a persistent memory of place and a more general regional tradition of ritualized deposition of artefacts within pits is more likely. This mixture of change and tradition is one which continues into the Bronze Age with the construction of the first built structures and the deposition of artefacts into an even wider range of contexts.

Roundhouses or… Currently, there is little evidence for Early Bronze Age settlement activity, and with the possible exception of structure 1624 at Gwithian site GMXV (Nowakowski et al forthcoming) no structures, which have been identified as dwellings of this period (Jones 2005, 44). By contrast, over the last couple of decades, a pattern has emerged for the nature of Middle Bronze Age settlement in Cornwall which suggests that the landscape was occupied by permanent or semi-permanent settlements across a variety of topographical zones. However, to date post-ring roundhouses of the type known from southern Britain as far west as Honiton in east Devon (see papers in Barrett and Bradley 1980; Barrett et al 1991; Moore and Jennings 1992; Fitzpatrick et al 1999) do not appear to be a feature of the Cornish Middle Bronze Age. Instead, two types of roundhouse are found which vary with topography.

Bronze Age The majority of the excavated features belong to the Bronze Age, with a peak of activity between 1500 and 1100 cal BC. This was mainly associated with the construction and use of five post-rings and with more ephemeral activity in the form of pits, fence-lines and old land surfaces which may have been associated with occupation. The evidence raises questions about the function and appearance of the post-ring structures, and the relationship between these structures and other contemporary activities taking place on the site will be explored.

In the lowlands, sunken-floored roundhouses ranging in date from the middle of the second to the first millennia cal BC have been found at numerous locations including Trevisker, Trethellan Farm, Trevilson and Callestick (ApSimon and Greenfield 1972; Nowakowski 1991; Jones and Taylor 2004; Jones 1998-9). Excavated lowland roundhouses have been found to comprise single or double post-rings from 6m to 10m in diameter set within circular 31

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough in Fields 4 and 7, the survival of deposits outside Structure 102 and in situ walling associated with a Romano-British structure within the ‘Fort’ at the eastern end of the plateau (Gossip 2006) indicate that this process is unlikely to have entirely removed internal features or evidence of walling. Furthermore, the depths of the postholes comprising the post-rings are, with average depths ranging between 0.17m-0.4m (excluding Structure 712), comparable with well-preserved roundhouse and ceremonial sites such as those at Trethellan and Stannon site 6 (Nowakowski 1991; Jones forthcoming a), again suggesting that truncation may not have been severe.

hollows dug into the ground; these often contain central hearths. It is apparent that these were circular, roofed structures with a single entrance, usually on the southern side. Evidence of inhabitation is often found in these buildings in the form of occupation deposits, charred plant macrofossils and refurbishment. There is good evidence that many activities associated with them were ritualized, especially during their abandonment when structured demolition and infilling of the house-hollows took place (Nowakowski 1991; Jones and Taylor 2004). In the Cornish uplands such as Bodmin Moor, Penwith and the Lizard, substantial stone-walled roundhouses are found, which when excavated are often found to contain internal post-rings and features such as paved areas or clay floors (Dudley 1956; Mercer 1970; Smith and Harris 1982; Jones forthcoming a). In common with their lowland counterparts entrances typically face towards the southern part of the horizon. Artefactual and environmental evidence from many of these structures suggests that they too were occupied as domestic living spaces. This division between upland and lowland zones may be too crude as some recently investigated roundhouses appear to embody structural elements from both architectural traditions (Taylor 2005).

The Tremough post-rings therefore differ greatly from what is known of later Bronze Age roundhouses in Cornwall. An interpretation of these structures, based on their morphology, patterns of artefact deposition (see below) and dearth of environmental evidence would suggest that they were never used as conventional domestic structures. That they were carefully designed constructions with a degree of time and effort invested in their construction cannot be doubted, and radiocarbon determinations suggest a long period of use possibly spanning several centuries. ...Shrines?

The five post-rings at Tremough are rather different from those described for the Middle Bronze Age and the Iron Age. They comprise simple post-rings on a level surface with no indication of stone walling or any kind of wallline. Internal features, such as paving or stake-settings are scarce with the exception of Structure 392, which includes not only an internal ring of postholes but central postholes. The regular, circular plans of the Tremough post-rings also differ from those of many excavated lowland roundhouses, which can have irregular spacing between postholes, particularly where post renewals have taken place. In addition, there are only two instances where hearthpits occur and only one of these, in Structure 392, was a substantive central feature. The most complete Structures, 102, 335 and 392, are associated with external features, such as entrance postholes at Structures 102 and 392, and a possible fence-line to the east of Structure 335, suggesting that the approach to these structures was formalised and from the south-east. This orientation was reflected in the overall layout of the site which also aligned from south east to north west. Finally, there was a dearth of evidence for occupation activity, the spaces enclosed within the post-ring were not worn and spreads of occupationrelated material such as charcoal were absent. Substantive quantities of grain were limited to Structure 392 and artefacts were largely confined to deposits situated within post-holes.

As the post-ring Structures do not sit comfortably within what is known about domestic structures it is possible that they were in some way related to ceremonial monuments. Structure 66 dates to the Early Bronze Age and could be argued to be allied with a wider body of timber-circles, which are found during this period. The presence of free-standing timber-circles and post-rings associated with barrows has been well-documented throughout the European prehistoric record (Gibson 1998 ch 4; Glasbergen 1954). In Britain, post-ring structures were first identified with the discovery of Woodhenge between 1926 and 1928 (Gibson 1998, 100). Subsequently, numerous timber-circles have been identified. Alex Gibson has compiled a gazetteer of some 100 timber-circle sites throughout Britain. These range from complicated Later Neolithic structures with multiple rings of postholes with large diameters – for example the Sanctuary, or Stanton Drew (Pollard 1992; David et al 2004) – to more modest, simpler constructions such as at Charnham Lane, Berkshire (Ford 1991, 179), which is of a similar date to the Structures at Tremough: 3360±40 BP, 1750 cal BC to 1530 cal BC (BM-2737). Chronological issues are also highlighted by recent work at Croft Moraig stone circle, Perthshire, where a timbercircle was probably constructed within the stone circle during the Late Bronze Age (Bradley and Sheridan 2005). Thus, the dating of timber post-rings suggests a currency for this form of monument from around 2800 cal BC to 1000 cal BC (Gibson 1994; 1998). As such, it is unlikely that they were all constructed for the same reasons or were generally associated with any particular over-arching ideology.

Despite being covered by topsoil with an average depth of 0.3m-0.4m, damage to the sites through later ploughing will have occurred. This could be argued to have had a major impact upon the survival of cut features and occupation deposits. However, the preservation of old land surfaces 32

Interpretation Comparatively few timber-circles have been identified in south-west Britain. The six published examples in Dorset vary greatly in size, association, date and form. At Conygar Hill two Neolithic pit-rings were excavated, one of which possessed an inner ring of postholes 8m in diameter which produced sherds of Neolithic pottery (Smith et al 1997, 4852). The structure at Mount Pleasant was 38m in diameter and was situated within a Later Neolithic henge monument which was probably used between 2900 cal BC and 2200 cal BC (Wainwright 1979; Woodward 1991, 136). The 9m diameter post-ring structure set within a gully at Litton Cheney was also situated within a henge, although this site was found to contain cremation deposits and sherds from Early Bronze Age Collared Urns, so its actual date is open to question (Catherall 1976, 88). Two final examples from Dorset are both much later in date. One of a pair of excavated barrows at Poole was found to cover a post-ring with a diameter of approximately 8m (Case 1952, 152), which has subsequently been radiocarbon dated to between 1610 cal BC and 1400 cal BC (CBA Radiocarbon Index for Great Britain website). On Ogden Down, two concentric post-rings with an outer diameter of 17m dating from 1220 cal BC to 910 cal BC encircled an earlier round barrow (Green 2000, 115-6). These latter sites are contemporary with most of the Tremough post-rings.

date to those at Tremough and may have been associated with a late phase of activity within the Stannon cairn complex prior to its incorporation within a roundhouse settlement, whilst the timber-setting in site 9 is likely to be contemporary with the settlement. Lastly, stake-hole rings have been identified as part of multi-phase barrows, for example at Tregulland, Trelowthas and Davidstow site I (Ashbee 1958; Jones and Nowakowski 1997; Christie 1988). This last group of features are much less substantial than those at Tremough but could be argued to relate to a wider tradition of circular monument construction (Bradley 1998; Case 2004). However, ceremonial monuments dating to the Middle to Late Bronze Age are uncommon and until recently there were few comparable free-standing timber post-rings of these periods. On Dartmoor, Structure 804, a free-standing post-ring with a diameter of 3.3m, occupied the centre of the Shaugh Moor enclosure (Wainwright and Smith 1980, 89). This structure has been radiocarbon dated from 1400 cal BC to 920 cal BC and is again likely to have been contemporary with the latest phase of the Tremough Structures. Within Cornwall a possible large diameter timber-circle or hengiform monument has recently been discovered at Goss Moor, although this monument is currently undated (Stuart Foreman pers. comm.). The closest Cornish parallel for the Tremough structures lies with a timber monument at Bosmaugan, where a structure of approximately 8m in diameter enclosed a central pit which contained Trevisker pottery (Cole 1999). This has been radiocarbon dated to between 1400 cal BC and 1000 cal BC. However, this site was found in isolation within a pipeline corridor and its wider context is uncertain.

Devon has few post-rings that can be associated with ceremonial monuments. An elliptical setting of pits was identified within the henge at Bow, but they are likely to date to the Later Neolithic: they are unexcavated and it is possible that they never held upright timbers (Griffith 1985, 311). At Farway, a Bronze Age cairn covered an oval central platform, and five pits containing charcoal and a ring of postholes with a diameter of approximately 3.5m (Pollard 1967, 32).

Perhaps the most interesting comparisons are found in Wales at Moel y Gaer and on the Isle of Man at Billown. At Moel y Gaer a post-ring structure with a diameter of 7.4m enclosing a stone-filled central pit was found during the excavation of an Iron Age hillfort (Guilbert 1976, 311-12). This structure was largely respected by the later roundhouses and was interpreted as a possible open air ‘sanctuary’ (ibid, 312). With the exception of a single quernstone the structure was devoid of finds. A radiocarbon determination of 3750±100 BP, 2200-1650 cal BC (HAR-1195) was obtained from one of the postholes. The determination has a wide standard deviation and the material submitted for dating is not specified which means that the site could be younger if mature oak was included in the sample (CBA Radiocarbon Index for Great Britain website). Nonetheless, the Moel y Gaer post-ring does highlight the possibility of similar structures occurring within later prehistoric settlements. This point is reinforced by the site at Billown where two timber post-rings were sited on the northern side of a Middle Bronze Age settlement (Darvill 2003). The first was a simple penannular ring approximately 11m in diameter. The second, referred to as the ‘shrine’, was a 7m diameter structure consisting of concentric post-rings, approached by a path from the south east. This structure was radiocarbon dated from 1309 cal BC to 1138 cal BC

A small number of post-rings have been recognised in Cornwall, which are mostly known from barrows and cairns (Table 20 and Fig. 44). They are predominantly simple, single rings which vary in diameter from approximately 5.5m to 20m in diameter (Miles and Miles 1971; Jones forthcoming a). The earliest of these was found beneath Davidstow site 22 (Christie 1988; Jones 2005, 100-2) and consisted of a post-ring some 10m in diameter which was associated with pits and Grooved Ware pottery. A later Neolithic determination of 2890 cal BC to 2550 cal BC was obtained from one of the pits. The remaining Cornish monuments with post-rings are all of Early Bronze Age date. At Cocksbarrow a cairn-ring contained a double ring of posts with an entrance on the south eastern side and two phases of posts were recorded in the bank at Caerloggas I (Miles and Miles 1971; Miles 1975, 58; Jones 2005, 89-92). A post-ring identified during excavations at Stannon site 6 was found within an earlier ring-cairn which was radiocarbon dated from 1600 cal BC to 1400 cal BC and a second possible timber-setting within site 9 dated from 1500 cal BC to 1100 cal BC (Jones forthcoming a). The post-ring in site 6 is the nearest in 33

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough and has been interpreted as an open structure (ibid, 1315). It is worth noting that where contextual information is available, Middle Bronze Age post-rings, whether freestanding as at Billown and Tremough or set within earlier monuments as at Stannon, seem to be located inside or near to areas of settlement activity.

remain speculative. However, given the fact that the sites appear to have been maintained over several centuries and the alignment respected, it is perhaps likely that they had substantial timbers, which, if connected by lintels may not have needed such deep postholes to remain upright. There is insufficient evidence to show whether the Tremough structures were roofed, although if comparisons are drawn with ceremonial monuments – particularly where timber has survived (Brennand and Taylor 2003) – it is perhaps likely that the simpler ones such as Structures 66, 712 and 335 were not. However, the possibility must remain that the latest, Structure 392 with its double postring and entrance posts may well have been roofed and that some or all of the others could have been.

An open or shut case? Another important question which arises in any consideration of these timber monuments is their appearance; put simply, were these structures roofed or open post-rings? It can be reasonably assumed that the postholes supported vertical posts, probably of oak given the amount of mature oak charcoal. The post-rings were all roughly circular in form, with suggestions of entrances or approaches evident at Structures 102, 335 and 392 on their south eastern sides (Fig. 16). Only one Structure, 392, had more than one ring of posts – implying either a smaller second phase circle, or a contemporaneous inner circle; although it is likely that the inner-ring was contemporary and the site was more complex than the others.

Structured deposition Given that the post-ring complex could have been used for up to a millennium, when compared with lowland roundhouse settlements (Nowakowski 1991) and many excavated barrows (Jones 2005, ch 3), the amount of material deposited within the structures was relatively small and does not suggest intensive use. However, several deposits within the post-rings show that they were the focus for structured deposition, but that the types of artefact selected for deposition and the context were different from those of the Later Neolithic, and this change became more marked over time. This can be demonstrated by a consideration of the evidence from the earliest postring Structure 66, and the latest Structure 392. Structure 66 is dated from 2210 cal BC to 1970 cal BC and was contemporary with the group of Early Bronze Age pits. These contained some deposits which were similar to the Later Neolithic pit group, namely charcoal and hazelnut shells, but were devoid of pottery and flint, and instead contained a few cereal grains and burnt stones. So, as with the Later Neolithic pits, it is possible to argue that they represent the ritualized burial of remains which were associated with the consumption of food. However, the construction of Structure 66 marks a departure from this picture by providing the first architecture on the site and gives perhaps the clearest evidence of deliberate deposition within a structure in the form of fragments of a mould or an oven brick from a pottery clamp SF1022 from posthole [73]. These were fresh and unabraded when deposited, but do not appear to have been used within the structure (Quinnell below) and may well have been associated with transformative processes of raw materials into durable artefacts – either of clay to pottery or ore to metal – rather than with the immediate consumption of food. It may well be significant that these were the first objects to find their way into the first durable site.

The dry-land nature of the site means that the preservation of the post-rings at Tremough was limited to the postholes. This presents problems in reconstruction and a range of alternative possible reconstructions can be made for both large and more modest post-rings. This is demonstrated by the open and roofed versions of the Knowth timbercircle, the Sanctuary and Woodhenge, Sarn-y-Bryn-Caled (Eogan and Roche 1994; Pollard 1992; Musson 1971; Gibson 1994). It has also been suggested elsewhere that vertical posts may have been joined with a complete ring of lintels, thus emphasising the circularity of the structure as illustrated by suggested reconstructions of Sarn-y-BrynCaled (Gibson 1994; 1998, 109) and Itford Hill (Russell 1996, 12). This cannot be proven, but would certainly have helped define post-rings at Tremough as distinct circles. The reconstruction of post heights is also problematic. Gibson (1998, 107) has suggested a ratio of 1:4 for the depth of the posthole below ground to height above, but suggests that this may be over cautious since posts are unlikely to have been carefully trimmed and were possibly wider in diameter at the base, adding greater stability. In this case, taller posts would have not required such deep postholes. The general regularity of postholes at Tremough suggests the careful selection of relatively straight timbers, or at least the trimming or shaping of the bases. Average posthole depth for the Tremough post-rings ranges between 0.23m to 0.24m, with an average of 0.28m for the inner post-ring of Structure 392. Using Gibson’s model these suggest a post height of at least 0.96m, with a possible height of 1.12m for the inner-ring of Structure 392 (Gibson 1998, 106-8). The exception to this is Structure 335 which has an average posthole depth of 0.14m; it is likely that this reduced depth is a result of truncation in this area by later activity. Because of possible truncation, the actual heights of the timbers within the post-rings must

From the middle of the second millennium cal BC, there were further changes as four further post-ring structures were built but fewer pits were excavated. Indeed only one, pit [25] is dated to the Middle Bronze Age and this is found far to the east of the post-rings. Instead, ceramics 34

Interpretation of an unusual form for Cornwall, with both Trevisker and possible Deverel-Rimbury affinities – referred to here as Trevisker-related pottery (Quinnell below) – and other artefacts, including worked stone, were deposited in postholes and features which were associated with the post-rings.

below) from the central hearth-pit [320] (Fig. 15). The inclusion of the last object within a hearth-pit radiocarbon dated to the latter part of the second millennium cal BC suggests that it had been curated for several centuries. Posthole [326] produced fragments of baked clay (SF 1034) in a fine fabric which showed evidence of being ground, probably for use as a colourant (Quinnell below). With the exception of a few grains from Structure 102, Structure 392 contained the only significant quantity of cereal grain to be found in a prehistoric context at Tremough. Although problems exist with possible contamination from later Romano-British disturbance, it is possible that food was processed and consumed in this Structure. It is interesting to note that the complexity of the finds assemblage is

The greatest concentrations of artefacts were found in two of the post-rings, Structure 102 and especially Structure 392. Items which were selected for disposal in the latter Structure include Trevisker-related ceramics belonging to a minimum of 12 vessels, quartz, cassiterite and wolframite pebbles, a large, broken muller S6 and a ripple-flaked flint knife of probable Early Bronze Age date (Lawson-Jones

Figure 15: Distribution of finds within Bronze Age Structures 66, 102, 392 adn 335. 35

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough front (south east) to the rear (north west) of the structures and the stonework follows a similar pattern, but is found to the south west of the pottery. This is strongly suggestive of a pattern of intentional deposition, which may well have been influenced by underlying ideas of cosmology and alignment based on the south east – north west orientation which may have structured much of the building activity which occurred on the site during this phase.

matched by its architecture and its key position at the north west end of the post-ring alignment. A second significant pattern during the second millennium cal BC is the actual locations which were chosen as appropriate places for deposition within the Structures (Fig. 15). This took the form of an axis which seems to have followed a line that had been established by the overall alignment of the post-rings. Put simply, there appears to be a general tendency for artefacts to be deposited along a south east to north west axis. This is particularly marked in Structures 102 and 392 and perhaps to a lesser extent within Structure 66 (although this post-ring was incomplete). In these, pottery was found in a band from the

However, there are subtle differences between the two structures (Fig. 15). Within Structure 392 the greatest numbers of finds were deposited within the central hearth-pit and may have been intended to close what might have been the focal point of the building. The

Figure 16: Reconstruction of the post-rings viewed from the south (conjectured posts are solid). 36

Interpretation majority of artefacts associated with Structure 102, however, came from a spread located in the entrance. This deposit may also have represented a closing down of the site. Spreads of materials (including ceramics, charcoal clay and quartz pebbles) used in blocking entrances are known both at Early Bronze Age barrows – for example Davidstow 8 (Christie 1988; Jones 2005, 107) – and at the later Bronze Age roundhouse at Callestick (Jones 1998-9), again providing a link between roundhouses, round barrows and the Tremough post-rings. However, the point to make here is that the actions at Tremough were probably carried out against the back-drop of a settled landscape and should be seen as being part of a wider spectrum of ritualized practices which occurred on and near to settlements from the middle of the second millennium cal BC onwards (Brück 1995; Masefield et al 2003; Bradley 2005).

the temporal order. A similar argument has been made by Parker Pearson (1996) to explain the relationship between cosmology and roundhouse architecture. He argues that activities within later prehistoric roundhouses and the cyclical nature of life were linked by the passage of the sun; with the commonly found eastern orientation of entrances in roundhouses creating an axial symmetry about a line between the door and the rear. Such a division may have structured activities within the roundhouse, with some parts of the house being associated with particular tasks or considered auspicious or inauspicious (Parker Pearson and Richards 1994, 47-9; Yates 1991). Similar considerations could also have been at work at Tremough, with underlying cosmological ideas underpinning the spatial settings of the post-rings, the orientation of their entrances and the pattern of deposition within.

As above so below

Another consideration at Tremough is the role that the landscape might have played in the siting of the postrings, as the orientation of the Structures aligns towards the Fal estuary and the coast. The importance of landscape and the proximity of distinctive natural features, such as rocky outcrops or the sea to the siting of Early Bronze Age ceremonial monuments is well-documented in Britain (Tilley 1994; 1995; Bradley 1991; 2000; Watson 1991; Jones 2005; Jones and Quinnell forthcoming) and a large number of Cornish sites, including entrance graves, barrows and cairns, are located in the coastal zone (Christie 1985; Reynolds 2005; Jones 2005). The siting of roundhouse settlements has been less well-studied but there are indications that some Middle Bronze Age settlements may have been located near to distinctive natural features, for example roundhouses at Leskernick and Kynance Gate were constructed near to rocky outcrops or had large rocks incorporated within them (Bender et al 1997; Jones forthcoming b). At Tremough, the view towards the water and the sea beyond could have conveyed a variety of meanings and might have been linked to celestial cosmologies associated with the passage of time.

As discussed above, the Structures were laid out on a south east to north west axis. Where entrances were identifiable (Structures 102 and 392, and possibly 712), they shared the same orientation and in Structures 102 and 392 the deposits of artefacts followed the same axis. One possible inspiration for this orientation could have been a desire to maintain a celestial alignment, a factor which has been recognised as an influence upon both complex and simple ceremonial monument construction (Ruggles 1999; Owoc 2001a; 2001b; Bradley and Sheridan 2005), as well as on the orientation of domestic buildings (Richards 1990; Parker Pearson and Richards 1994; Oswald 1997). Indeed, Richard Bradley (2000, 76) has noted that on Dartmoor both cists and roundhouse doorways are frequently aligned to the south east. Owoc (2001a, 195-6) has argued that some Cornish barrows, such as Cocksbarrow, had entrances aligned towards the mid-winter sunrise, on the shortest day of the year, for the purposes of manipulating time and space and the linking of the deceased with a wider perception of

The material used for the construction of the post-rings is also worthy of comment. Cornwall possesses a large number of upstanding monuments which survive because they were built of stone. It is therefore easy to forget that timber had an equally important role to play in roundhouse construction and was used in many ceremonial sites (Miles 1975; Jones 2005; forthcoming; Reynolds 2005). Timber-rings with ceremonial connotations would have been visible in the landscape at the time that the first post-rings at Tremough were erected. In Wessex, it has been proposed that timber may have been associated with the living and stone with the ancestors. It has been argued that this explains why some timber sites were made permanent by the replacement of wood by stone (Parker Pearson and Ramilisonina 1998; Pitts 2000, 2578). However, in Cornwall the reverse scenario seems to be the case: where relationships are found, timbersettings seem to be late additions to many sites (Jones forthcoming a). An implication from this might be that,

Order and architecture The radiocarbon based chronology for the post-rings (radiocarbon dating below) could be taken to suggest that they more-or-less developed from south east to north west across the site over time, beginning with Structure 66 and the adjacent pit group, probably at the beginning of the second millennium cal BC, and culminating with Structure 392 closer to the end of the second millennium cal BC. However, the picture is likely to be more complex, as a second determination from Structure 66 suggests the renewal of the structure between 1770 cal BC and 1530 cal BC long after its initial construction, possibly at a time when the other post-rings were being erected. The radiocarbon determinations do emphasize two significant points: first that the area was used over several centuries for a limited range of practices and, secondly, that certain structuring principles behind the layout of the architecture were maintained over a similar period.

37

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough In Area B/Field 4 a sub-rectangular building, Structure 232, was revealed near to the remnants of old land surface [17] and pit [25] which were dated to the Middle Bronze Age. Further to the north in Field 3 a circular anomaly was identified by the geophysical survey which may represent a sunken-floored roundhouse, but this was not excavated. Structure 232 was not radiocarbon dated and its dating relies on a single sherd of Trevisker pottery recovered from within the site. However, although rectangular structures dating to the Middle to Late Bronze Age are less common than circular buildings they have been recorded elsewhere, for example at Down Farm in Dorset and at Easton Lane in Hampshire (Barrett and Bradley 1980; Barrett et al 1991, 201; Fasham et al 1996, 50). Only two other examples have been recorded in Cornwall. At Penhale Moor, a small hollow-set structure measuring 5.5m long by 2.2m wide was constructed beside a contemporary sunken-floored Middle Bronze Age roundhouse (Jones et al 1994). A second example was found within the Middle Bronze settlement at Trevisker, structure B, and was again situated beside a sunken-floored roundhouse. It measured approximately 5.5m long by 2.5m wide (ApSimon and Greenfield 1972, 312). Both rectangular structures were interpreted as ancillary buildings. If Structure 232 was contemporary with the post-rings then it provides an interesting contrast in the use of architecture between the eastern end of the site, with its old land surface which contained occupationrelated material in the form of Trevisker pottery and broken loomweights, and the carefully laid out circular post-ring alignment in the west associated with placed deposits of Trevisker-related pottery. It may be no coincidence that a fence-line appears to have separated the two areas (posthole groups 37, 38, and 85). Zonation of space has been argued for in other Bronze Age settlements in southern Britain (Parker Pearson and Richards 1994, 51; Brossler 2001), for example with middens being created in some areas and activities involving feasting taking place in others.

if wood were metaphorically associated with the living, it was used to keep sites alive for the communities who used them. Another possible metaphor associated with the post-rings may have been with dwellings (Bradley 1998, 150-8). Few Early Bronze Age residential structures are known in Britain, and those are often insubstantial and circular (for example, Hey 1997, 108). This raises the possibility that the first structures at Tremough could also have been referencing the house, albeit on a larger-scale. However, it is interesting that contemporary activity around the earliest post-ring, Structure 66, did not take place within it but was located several metres to the north; this was associated with the circular arrangement of pits, whose contents may have been generated by occupation-related activity. Over time it seems likely that with the increase of permanent settlements in the landscape and the demise of ceremonial monument construction, referencing of the roundhouse would have increased – and this is perhaps reflected in the layout of Structure 392, which possessed a house-like plan, and may have been roofed. Taken together these elements may indicate a degree of continuity or development from cosmological principles which were first evident in the Early Bronze Age and later found at Middle Bronze Age roundhouses. It is possible that a fusion of older traditions associated with celestial orientation and topographical features influenced the layout and use of the post-rings. As Colin Richards (2005, 249-58) has argued in Orkney, shared cosmological principles could exist across both domestic and ceremonial monuments. They allowed a wider comprehension and categorisation of the world as it was experienced in ‘daily life’ to be expressed. However, cosmological ideas are not static constructs and would have been subject to changing circumstances in the outside world, challenges to tradition, and to the foibles of social memory, and thereby transformed over time.

Although it is evident that ritual and domestic practices are frequently closely entwined, informing and influencing one another (Hubert 1994; Bradley 2005), it is also true that the subdivision of space between places associated with more formal ritualized practice and with occupation activities is documented in the anthropological and archaeological record (Kent 1993; Richards 2005). For example, in her discussion of the Lio settlements, Howell (1995, 155) notes the presence of structurally different ceremonial houses, which are located near to temples and away from the houses which are found in the fields. Victor (2001, 193) has argued that Bronze Age cult houses in Sweden were not located near to settlements but were often found close to graves. Similar relationships can be posited for areas of settlement and burial in southern England. At Itford Hill a Middle Bronze Age barrow surrounded by a 5m diameter post-ring was found to the north of a nearby contemporary roundhouse settlement (Holden 1972). Radiocarbon dating and analyses of the pottery supported evidence for links between the two sites, with parts of one vessel being found in one of the houses and in the barrow (Bradley 1998, 150-

Unity and division Some evidence is present for zonation of activities across the plateau. The post-rings were located at its western end and appear to have been used for intermittent ritualized activities. They do not appear to have been intensively used and other activities, such as roundhouse construction, do not seem to have taken place in the immediate vicinity. It would be tempting to describe this setting as a ‘ritual landscape’; however, to do so would be to ignore the fact that many of the deposits within the post-rings directly referred to the domestic world and thereby create a false dichotomy between two interrelated spheres of life (Barrett 1994; Richards 2005, 205). Furthermore, there is some evidence for occupation activity to the east of the post-rings in Area 4/Field 4 and Field 3, which indicates that the post-rings did not stand alone in an empty landscape, or one reserved solely for ceremonial purposes. 38

Interpretation 2). It has also been suggested that the site was a ceremonial structure rather than a simple barrow (Drewett et al 1988, 114). As previously discussed, the Middle Bronze Age ‘shrine’ at Billown was also located on the edge of the settlement area and approached by a formal path (Darvill 2003). The formal route of approach and the association with settlement-related activity has obvious resonances with Tremough.

had ‘life-cycles’ resulting in a sequence in which only one was used at a time with others in various stages of decay and the sites of the oldest being actively avoided. Under this scenario, a visitor to the site around c. 1200 cal BC might have seen a freshly constructed Structure 392 amidst an open area in which other sites were rotting or survived only as scrubby patches. Alternatively, it is possible that older sites were maintained and that if the site had been approached from the south west in the later second millennium cal BC an onlooker would have been faced by a series of structures which although made from similar materials would have had contrasting uses and subtly different appearances. The first structure to be encountered would have been Structure 102 with its large entrance posts and off-centre hearth-pit. To the west would have been the ‘ancient’ Structure 66, which may have had associations with the past and an ancestral time. To the north west were Structures 335 and 712, which were sited close together and could have formed a pair of open-rings. Beyond this was the house-like Structure 392. Its architecture appears to have referred both to the earlier sites and to roundhouses. This referencing of both past and present is also reflected in the deposits within it, which included a curated Early Bronze Age flint knife and cereals. If this structure were roofed, the visitor would have entered what looked like roundhouse, but would immediately have been met by an inner-ring of posts and internal posts which may have carried resonances both from earlier traditions and contemporary roundhouses.

In Cornwall there is also evidence from other excavated Middle Bronze Age sites that certain structures were set slightly apart from the main settlement area. At Stannon, site 9, a ring cairn located on the southern edge of a settlement was probably reused by people who occupied the adjacent roundhouses (Jones forthcoming a). However, in the lowlands there is evidence that new structures were built, such as Trethellan structure 2192 and structure 26 at Callestick, which were set apart from the main settlement areas, may never have had residential functions and could have been used for ritualized activities (Nowakowski 1991; Jones 1998-9). Seen in this light, the post-rings at Tremough could be argued to have been situated on the edge of a settled landscape. Walking the line A final point of consideration is to do with how the postrings themselves may have been experienced. The earliest, Structure 66, appears to have been renewed several centuries after its original construction and the regular spacing between the other structures implies that there was an awareness of where things had gone before. In other words, rather than considering the alignment as a simple progression it seems probable that there was a process of referencing earlier sites both in terms of the renewal of older post-rings and the continued borrowing of older architectural traditions. This would also mean that rather than being a ‘finished product’, the entire complex was an ongoing project (Barrett 1994, 24-9; Richards 2005, 252), periodically being reworked – with timbers being intermittently replaced and newer structures constructed. The main point, however, is that this does not imply that meanings associated with the structures were either the same as one another or unchanging, but that transformations were made which reflected changing social realities. Indeed, this is indicated by differences between the sites and the increasing complexity of the architecture from the simple southern post-ring Structure 66 to the house-like northern Structure 392. Again, similar observations have been made about both Early Bronze Age monument complexes and some Middle Bronze Age settlements in Cornwall (Jones 2005; forthcoming; Taylor 2005) where differences between sites may well have been linked to the way that they were used and between the groups who used them.

Summary In this section, we have attempted to identify what the post-ring structures might have looked like and how they relate to what is currently known about Cornwall in the second millennium cal BC, a period first associated with ceremonial monuments and later by settlements. These phases have often been treated by archaeologists in very different ways. To illustrate this latter point it might be useful to consider the pictures made by John White in the sixteenth century (Alexander 1976). White made illustrations of Native American settlements in New England, which included settlements, daily activities ceremonies and timber-rings. An illustration of one of the latter sites is frequently reproduced in the archaeological literature in relation to ceremonial monuments and stone circles (Burl 1979, 205; 2000, 69; Webster 1993, 69; Gibson 1998, 99), yet it is interesting to see that another drawing which shows the context of the post-ring, on the margins of the settlement and set amongst the fields of maize is not (Alexander 1976, 83). This is not to argue that the Middle Bronze in Cornwall was comparable with New England in the sixteenth century, but merely to make the point that we wished to avoid being too restricted in our focus. For this reason parallels with both roundhouses and barrows were examined. By pursuing this avenue, it was found that parallels existed with postsettings associated with earlier barrows and a handful of

The available evidence suggests two possible scenarios for the development of the complex. In the first, the sites 39

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough Field systems have been identified throughout Cornwall as ancillary elements to Iron Age and Romano-British enclosures or rounds (Herring 1994; 1998). Knowledge of many of these results from close scrutiny of aerial photographs and the identification of crop-mark sites by the National Mapping Programme, for example aerial photographs of land inland of the Rumps cliff castle are strongly suggestive of associated field systems. Unfortunately, the dating of most of these is uncertain since they remain unexcavated. However, some excavated field systems have been shown to have direct relationships with rounds, such as those at Boden (St Anthony-inMeneague), Richard Lander School (Truro) and Grambla (Wendron) (Gossip forthcoming; 2005; Saunders 1972). Geophysical survey undertaken at Boden appears to show field systems radiating outwards from the round, and these may be of similar Late Early Iron Age date (Linford 1998, Fig. 2; Gossip forthcoming). The Richard Lander School excavations revealed a system of rectangular fields extending south east from a sub-rectangular enclosure. The enclosure is unexcavated but the field ditch fills produced late Iron Age pottery consistent with first century cal BC occupation of an adjacent unenclosed settlement (Gossip 2005).

other free-standing Middle Bronze Age post-ring sites. We also argued that comparisons could be made with roundhouses and that these are likely to have increased with time. It is therefore possible that the Tremough postrings may have echoed the post-rings associated with barrows, but these links were increasingly supplanted by ideas that were associated with the settled landscape. In short, it has been suggested that the changing nature of buildings found in the wider landscape may have become reflected in the post-rings at Tremough. Iron Age and Romano-British The Iron Age field system The best evidence for Iron Age activity is provided by the brick-shaped field system which extended across Fields 7 to 3. This consisted of a series of square fields each of around 1000 square metres defined by ditches alongside which were probably earthen banks, possibly hedge-topped. Although the excavated remains were fragmentary, they extend over an area of around 3 to 3.5 hectares to the west of the ‘Fort’ enclosure. Geophysical survey (Figs 2, 17 and 18) suggests that the fields may have continued for some distance east of the ‘Fort’ and that they were generally aligned with this enclosure. Ditch [7] produced a Later Iron Age radiocarbon determination of 2070±55BP, 350 cal BC-AD 60 (AA-44600), from its lower fill (9) and ditch [30] a Later Iron rim PR29. PR29 and a small quantity of other sherds from upper ditch fills are all of Later Iron Age date. The only Roman sherd from ditch fills comes from ditch [1].

The preservation of ecofactual material from Tremough was poor and there is no data from Iron Age field ditch contexts. Unfortunately, there was very limited material from contexts associated with Iron Age fields at Trenowah (J Jones in Johns forthcoming) and few of the limited cereal finds at Stencoose could be specifically linked to the Iron Age (Jones 2000-1, 78, table 2). The rather better evidence for the Romano-British period will be discussed below.

The field system appears to stop broadly in the area of the Bronze Age post-rings. It is possible that some memory of the special character of the area was involved in the choice of position for the field system or it was an area of pasture or rough ground in woodland.

Occupation in the Iron Age and the character of rounds No Iron Age structures were identified by the excavations. Perhaps the best candidate and most likely location for a later Iron Age settlement was the rectangular enclosure of approximately 0.17 hectares (50m by 38m) known as the ‘Fort’, which may be equivalent to the class of enclosure known as ‘rounds’ found across Cornwall. Evaluation trenching of the ‘Fort’ confirmed that it was enclosed by a substantial ditch and suggested that it may have been occupied during the Iron Age and Romano-British periods. Archaeological recording to the south of the ‘Fort’ also investigated ditch [104] of the adjacent ‘D’ shaped enclosure (Fig. 18). Its western side was defined by a relatively substantial ditch, likely to have been associated with a stock-proof boundary.

Only two other Cornish Iron Age field systems have been investigated and radiocarbon dated. The earlier example was found at Trenowah on the St Austell North East Distributor Road. Radiocarbon determinations from a stony bank and a field ditch suggest initial setting out during the Early Iron Age between 770 cal BC and 390 cal BC (Johns forthcoming) with subsequent additions through the Later Iron Age and early Romano-British period. At Stencoose, a field system was recorded which indicated use from the later Iron Age into the Roman period. A radiocarbon determination falling between 162 cal BC and AD 218 and finds from the ditches imply that the area was enclosed before the Romano-British period (Jones 2000-1, 82). Both the Trenowah and the Stencoose field systems appear to have developed over time, and therefore are not closely comparable to the Tremough brick-shaped system. The Tremough system is the first brick-shaped system in Cornwall to be tied to the Later Iron Age both by artefacts and by a radiocarbon determination.

Rounds are small, enclosing an area of typically less than one hectare. They are usually univallate enclosures and often located on hill-slopes. Their prevalence as upstanding monuments and crop-mark sites in the Cornish landscape (see for example, Jones and Taylor 2004, Fig. 34) means that they have dominated discussions of settlement activity in later prehistoric and Romano-British Cornwall (Johnson 40

Figure 17: The Iron Age and Romano-British Landscape (west).

Interpretation

41

Figure 18: The Iron Age and Romano-British Landscape (east).

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough

42

Figure 19: Distribution of finds within the Romano-British Enclosure.

Interpretation

43

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough

Figure 20: Romano-British Structure 338. and Rose 1982, 151-207; Quinnell 1986, 115; ibid 2004, 211). Air photograph transcription by the National Mapping Programme indicates that the number of potential rounds in Cornwall may exceed 2000, with a density in places of two per square kilometre (Quinnell 2004, 211; Andrew Young pers. comm.). Other enclosures have been identified within 1.5km of the ‘Fort’ at Bellevue, Mabe Burnthouse, Treluswell, Burnthouse and Round Ring.

These enclosures, if rounds, are also likely to have been occupied during the Later Iron Age and/or Romano-British periods and give an indication of the density of lowland settlement at the time. Although only a small number of rounds have been excavated, it has been established that some such as that at Threemilestone have origins in the Middle Iron Age 44

Interpretation Iron Age to Romano-British occupation – continuity and change?

(Schwieso 1976, 65-6: Quinnell 2004, 212, table 12.1) and others remained in use until the fifth or sixth centuries AD. At one time rectangular examples were thought to have resulted from Roman military contact (Dudley 1954), but excavations at Tregilders (Trudgian 1977) and Trevinnick (Fox and Ravenhill 1969) suggested dates in the first century AD with no indication of Roman influence. The recently excavated univallate rectangular enclosure, a probable round, at Boden, St Anthony-in-Meneage, has chronological modelling of radiocarbon determinations suggesting that activity commenced around 425-390 cal BC (at 68% probability). This dating is supported by the presence of Late Early Iron Age ceramics and an absence of Middle Iron Age South Western Decorated Ware, demonstrating the presence of a substantial subrectangular enclosure in the Late Early Iron Age (Gossip forthcoming). The rectangular form of the ‘Fort’ cannot, therefore, be taken as necessarily indicating RomanoBritish origin, although its continued occupation into the Roman period is likely.

Given the chronological range of the artefactual assemblage and the radiocarbon dates, it is likely that there was a degree of continuity on the site between the Iron Age and Romano-British periods. The construction of Structure 338 and its surrounding enclosure could be indicative of agricultural expansion during the second century AD, with the development of a smaller farming settlement at the western end of the site away from the main centre at the ‘Fort’ to the east. If correct, this would perhaps suggest intensification in the arable economy with the greater use of cereal types such as spelt wheat and an indication of ‘year round’ cereal production. Most sites in Cornwall show general evidence of continuity. At Stencoose, the Later Iron Age field system continued in use into the Romano-British period and beyond (Jones 2000/2001). At Carn Euny there was continuous occupation from the fifth century cal BC until the fourth century AD, although structure morphology changed from timber-built roundhouses to stone-walled courtyard houses in this time (Christie 1978, 373-74). At Chysauster, the Romano-British courtyard settlement was set within an extensive field system: Iron Age pottery suggested activity during this period, although no precursors to the courtyard houses were found (Christie 1987). Similarly, the round at Reawla produced Iron Age pottery but contemporary structures were absent (Appleton-Fox 1992). An Iron Age field system prior to establishment of the round is evident at Trethurgy, but no evidence of contemporary settlement was found (Quinnell 2004, 215). The construction of the rounds at Reawla and Trethurgy in the second century AD together with others such as Grambla provide the principal evidence for agricultural and population expansion at this time (Quinnell 2004, table 12.1).

Rounds probably developed from changes within the structure of local Iron Age societies, with later expansion in their numbers caused by population growth and/or social reorganisation resulting from contact with the Roman world. A significant increase in the number of rounds occurs in the second century AD (Quinnell 2004, 212), around the time when enclosure [306] and Structure 338 at Tremough were constructed, and when also Trethurgy, the only totally excavated round (Quinnell 2004), was established. Since Johnson and Rose (1982, 155) identified the diversity of this settlement-type, archaeological investigation has raised questions of their function, status, inter-site relationships and chronology (Appleton-Fox 1992, 116-19; Quinnell 2004, 213-14). It has been proposed that rounds were places of permanent settlement whose inhabitants controlled their own resources and that their proliferation during the Later Iron Age and Romano-British periods was a result of a stable social structure (Herring 1994; Quinnell 2004). An interim interpretation of a round has been proposed: ‘a permanent settlement with substantially built houses whose inhabitants merited the distinction of a formal bound or enclosure, which may have held significance for their status beyond its provision of protection or defence’ (Quinnell 2004, 213). This interpretation implies that not all enclosures of appropriate size may prove to be rounds. The reference to status also assumes the presence of nonenclosed settlement, always difficult to identify within the Cornish landscape. Open or lightly enclosed Iron Age settlements have been identified at Bodrifty and Kynance Gate, (Dudley 1956; Thomas 1960) and a probable Iron Age transhumance structure has been found at Stannon Down on Bodmin Moor (Jones forthcoming a). A sofar unique unenclosed lowland settlement of ten circular houses was recently excavated at Richard Lander School, Threemilestone (Gossip 2005). As the Richard Lander School site was located near to at least one, and possibly two, contemporaneous rounds, the existence of open Iron Age settlement at Tremough cannot be ruled out.

Romano-British orientation

settlement

–

function,

form

and

The small enclosure [306] (phase 1) and [565] (phase 2) which surrounded Structure 338 has few parallels, none excavated, in Cornwall. At Small Hill Farm, an enclosure has been identified close to the round (c. 200m away) but other comparable sites with associated small enclosures are rare (Johnson and Rose 1982, Fig. 5). However, work carried out as part of the National Mapping Programme has identified comparable sites from crop-marks visible on aerial photographs, for instance in the Camel valley (Andrew Young pers. comm.). After some decades in the early third century, enclosure [306] was reduced in size, from 40m by 26m (0.08 hectares) to 25m by 19m (0.03 hectares). The reason for this contraction is uncertain, but the deeper ditch of [565] would have provided a more substantial bank. In both phases the ditch circuit was incomplete with no indication of a formal entrance, although the east side of the entrance is presumably marked by the definite terminal of [565]. 45

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough post-ring provided the main roof support. This Iron Age form is clearly demonstrated in the settlement at the Richard Lander School where there were nine circular structures with internal post-rings (Gossip 2005).

No parallels for these incomplete ditch circuits or lack of entrances are known in Cornwall, and in these aspects, as well as its small size and the presence of only a single structure, the enclosure differed markedly from rounds. The use of the ditch also differed (see below). The siting of Structure 338 in the centre of the first enclosure, with a fairly even spread of space all round between it and the bank is again different to the arrangement in rounds. The presence of a single structure implies control by its inhabitants over the working space in the enclosure as opposed to the shared use of space in rounds (Quinnell 2004, 224).

Some of the smaller buildings at Trethurgy are comparable to Structure 338 in size: house A2 was 8.5m by 6.5m and house A3 6.5m by 5.5m (Quinnell 2004, 183). These of course were stone walled and only house A2 had internal hearth features comparable to those in Structure 338. House A2 had a life-span of possibly 175 years within the fourth and fifth centuries AD and was replaced by A3 with a life-span of some 50 years (ibid) within the sixth century. There was also a single small oval timber building, House D1, at Trethurgy which measured approximately 5m by 6m and dated to the late second century AD (ibid 2004, 48-9). It is not surprising that there are so few parallels for Structure 338. Almost all Roman period structures in Cornwall known so far come from rounds which have survived because of their solid stone buildings. Buildings outside surviving enclosures have been difficult to locate but a few scattered examples demonstrate that oval as opposed to circular building in various forms also occurred in open settlements. A small Romano-British structure has been excavated at Carngoon Bank, Lizard, and comprised pits and stakeholes surrounded by an outer gully and measuring 7m long and 6m wide (McAvoy et al 1980), while at the Romano-British salt-making site of Trebarveth, St Keverne, a symmetrical oval structure measuring 5.5m by 4m was set into the ground and retained by stone walling (Peacock 1969). Porth Godrevy near Gwithian was an irregular oval stone-walled structure 8m by 5m with some internal postholes (Fowler 1962). The recent excavation of a Roman period settlement at Atlantic Road, near Newquay also revealed parts of oval structures with stone walls (Reynolds forthcoming). These examples are all in coastal locations and appear to have involved activities other than those connected with the usual farming round to various degrees. Structure 338 may be presumed to be unique in discoveries so far in that it represents a single house within a farming unit in a lowland setting.

In both phases the enclosure may have contained livestock close to Structure 338 when needed, perhaps tethered to impede interference with other activities. The enclosure may also have contained small horticultural plots, as suggested adjacent to houses within Trethurgy (Quinnell 2004, frontispiece) and could also have been a place where grain was processed and stored. For such purposes a complete circuit would have been necessary, perhaps a bank topped by hedging to make a more substantial barrier, or some temporary fencing, such as hurdles, to complete the circuit of bank as necessary. Such a barrier would have served to keep livestock away from activities around Structure 338 as well as controlling any brought within the enclosure. Structure 338 consisted of a complex pattern of intercutting postholes, pits and scoops defined by a ring of postholes forming a building slightly oval in plan and measuring 8m by 6.7m. This was occupied for a long period, probably from around AD 170 until after AD 300. The Structure was built entirely of timber, with stones used only for the packing and support of posts. The plan suggests an oval ring of outer posts, linked by timbers forming a ring beam, which supported the principal rafters. Walling would have been placed between these posts: there is no evidence for an outer wall line (Fig. 20). Internal postholes may have provided additional supports for the roof structure, although in a structure of this small size such additions are more likely to be alterations than original features. An entrance between two wider-spaced and exceptionally large postholes was situated on the southern side.

The doorway in Structure 338 was oriented to the south, towards the entrance into the enclosure. Although orientation may have been influenced by the direction of the sun to ensure the best available natural light and avoid prevailing winds, there may have been some symbolic stimulus dictating the construction of entrances. Parker Pearson (1996, 119-20) has argued that cosmology may have had an important role in the layout of Iron Age houses, especially with regard to the orientation of roundhouse entrances on the eastern side, thus facing sunrise and relating life in the roundhouse to the daily cycle of light and darkness around the house. Given the evidence for continuity between the later prehistoric and Roman periods in other aspects of daily life, we could expect earlier traditions to continue into the Romano-British period (see also Bradley 2005). This may demonstrate axial symmetry

Structure 338 represents the first clear example in timber of a Romano-Cornish oval house. Oval structures are recognised as the dominant house form for the Roman period in Cornwall, as exemplified by the stone-walled structures within Trethurgy Round (Quinnell 2004, 10.12). The houses at Trethurgy, and those from elsewhere in Cornwall discussed in its report, are generally larger than Structure 338, with the biggest buildings 12m to 17m long and 7m to 9m wide. They also tend to be more distinctly oval in shape with a greater difference between length and width and the unifying form of construction was the use of the outer wall as the main roof support (S Blaylock in Quinnell 2004, 10.4). This represented a change from the main Iron Age house form in which an internal circular 46

Interpretation of roundhouses between door and rear which in turn may be reflected in the structured deposition of artefacts; in fact the largest groups of finds were recovered from features along this axis, northwards from the entrance (Fig. 21). The north–south axis is repeated in the general layout of the settlement; for occupants of the structure the viewpoint along this axis towards the south would take in the wide entrance to the enclosure and pit [377] containing the deliberately broken vessel PR28 and hammerstone S16.

period development from a broad Iron Age background, suggesting the need for the regular measurement of commodities. Weights of this kind have come from rounds at Trethurgy, Porthmeor and Reawla and another from the courtyard houses at Chysauster – two have come from an enclosure at Killigrew in which metalworking appears to have been the principal activity (Cole and Nowakowski forthcoming). The presence of S8 at Tremough increases the range of sites at which these weights occur.

It is possible that the association of a stone weight S8 with the Structure 338 entrance was deliberate. At least sixteen carefully made mensuration weights have now been found (Quinnell 2004, 140) and are thought to be a Cornish Roman

Within Structure 338 was a series of intercutting hearthpits showing repeated and successive use. At Trethurgy and other rounds hearth-pits seem to be connected with domestic activity (Quinnell 2004, 185) although no other

Figure 21: Distribution of finds within Romano-British Structure 338 (quantities of pottery sherds numerated). 47

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough

Figure 22: Photograph of Romano-British enclosure, with adjacent field system visible

(top left).

brick-shaped fields and may represent episodes of activity on the edge of the field system before the construction of the enclosure. The presence of a scatter of unstratified Roman sherds across the field system probably indicates continued use throughout the Roman period.

site has a building with such a proportion of its interior taken up with these. The lack of metallurgical debris at Tremough suggests that here too the hearth-pits were broadly domestic. Activities such as the processing of milk for cheese or the smoking of meat come to mind and these would accord with the fairly low levels of cereal processing, at least around the Structure.

Increased reliance on cereal crops and evidence for a more open landscape is apparent from Romano-British contexts at Tremough although it is likely that this pattern was an expansion and continuation of the Iron Age agricultural regime. Animals must also have been kept, but acidic soil conditions meant that only a few crumbs of burnt animal bone were recovered from the site. Other Cornish sites with better levels of preservation have produced animal bones. In particular, Atlantic Road contained cattle and other domesticated species, with sheep/goat forming the dominant part of the assemblage (Reynolds, forthcoming). Here, the assemblage was seen to change from a predominantly young population indicating animals butchered for meat in the earliest Romano-British phase, to an older population indicating a greater use of wool and milk and animals kept partly for manuring in the later phases.

The stone weight S8, various whetstones, a possible smelting crucible re-worked from a broken stone mortar (S13) and the heavily-used hammerstone S16 from outside the enclosure could all have been associated with metalworking. This metalworking is likely to have been fairly small-scale, at least in the Structure area, primarily related to the needs of the occupants. It is very similar to the low-scale metallurgy identified at Trethurgy and other rounds (Quinnell 2004, 233). Other enclosure sites are now providing evidence for more concentrated metalworking as at Killigrew (Cole and Nowakowski forthcoming) and at Little Quoit Farm (Lawson-Jones 2003). These may have been liminal places set apart from centres of settlement and from such sites products would have gone into local exchange networks. The position and function of Structure 338 and its enclosure

Structure 338 was small but much time and energy was invested in its construction and maintenance, indicating that it was intended to be permanent. The relative absence of non-local fabrics might imply that it was a low status structure (see Quinnell below). The enclosure within which it was sited was also solidly constructed, with its ditch used for special depositions (see below). Structure 338 and its enclosure represent part of a larger but poorly understood pattern of open or lightly enclosed Romano-

Structure 338 within its successive enclosures [306] and [565] was situated at the west end of the Iron Age brickshaped field system. The enclosure ditches probably cut through a series of earlier ditches, [698] [590] and [716] although the relationships were not clear cut. These earlier ditches appear irregular when compared to those of the 48

Interpretation British settlements with less substantial dwellings. At Tremough the Structure and its enclosure were built on the edge of a field system which may have had as its principal settlement the higher status enclosed ‘round’ (the ‘Fort’) to the east. The most probable reason for its addition to the local agricultural landscape was to provide additional space and accommodation at a time of expansion, with much of the activity in the Structure probably focused on the processing of farm produce.

ditch [306] contained sherds which had been fairly evenly spread along its length, the majority of which were freshly broken. In the Phase 2 ditch [565], the majority of pottery was deposited in the northern arc of the ditch, increasing towards its north eastern terminus, and the closest point to Structure 338: a second concentration occurred close to the south western ditch terminus. Almost all the sherds, except Type 19 PR22, which shows evidence of curation, were freshly broken, and had been quickly covered after deposition. They included a group of conjoining sherds PR15 representing one third of a vessel. The rapid covering suggests a certain degree of ritualization of this act. The structured deposition of ‘rubbish’ within the enclosure ditch shows a similar emphasis on boundaries as was found at Stencoose, perhaps marking ownership of the enclosure in a symbolic way. Away from Cornwall, on sites where conditions for preservation have been more favourable, significant deposits of animal bones have been recorded in similar contexts (Merrifield 1987, 37-40). Unfortunately, no evidence for these survived at Tremough, but they may have been present alongside the deposits of pottery, perhaps deliberately incorporated into the ditches as part of a feasting ceremony (Fitzpatrick 1997, 78-9).

Structured deposition in the Romano-British period Structured deposition in Romano-British settlements has been identified across Britain and taken to indicate the survival of local practices throughout the Roman period (Clarke 1997, 80; Richardson 1997, 88). The deliberate selection and placement of artefacts is now accepted as commonplace in Iron Age contexts (Hill 1995), and the continuance of long established ritualized practice is likely. It is probable that many aspects of life during this period in Cornwall showed a ‘distinctive mix of indigenous Iron Age traditions with adapted Roman ideas’ (Quinnell 2004, 236). That there was notably less direct Roman influence within the region than elsewhere makes the nature of the deposits within Romano-British pits and ditches of particular interest (Quinnell 1986).

Within Structure 338, the selection of objects for deposition is most clearly demonstrated by the mensuration weight S8 from posthole [754] on the west side of its entrance (Fig. 20). The weight, although worn, was still complete and was found in conjunction with possible spindle bearing S9 and whetstone S10. These finds appear to have been sealed in place by covering with stone. The outer postholes also contained significant pottery deposits including PR5 and PR6 in posthole [339], which also included whetstone S11, and PR1 and PR2 from posthole [336], the condition of which indicated deposition and protection soon after being broken (Quinnell below). Another deposit of note was the Type 16 storage jar in posthole [669].

Recent writers have begun to address the issue of deliberately discarded objects in Cornish Roman period sites. Structured deposits identified at Trethurgy include unusual ceramics, part of a Trethurgy bowl and a rotary quern buried in the construction of a possible shrine structure G, and a pit [78] with a saddle quern close to a probable granary (Quinnell 2004, 237). Hearth-pit [294] contained deliberately selected and buried material including a stone mensuration weight which was probably a closing deposit for house A1b (ibid). At Stencoose the terminal of Romano-British field ditch ([6]) was the focus for deliberate deposition of ceramics, perhaps marking the boundary to a field system. A pit ([21]), thought to mark the same alignment as the field system ditches, had also been filled with a large number of pottery sherds, iron and a broken stone object, suggesting special, ritualized emphasis of these alignments and boundaries (Jones 20001).

The inclusion of fresh, unabraded sherds (including PR9PR13) and of stone mortar or smelt pot S13, whetstone S14, and fragmentary rotary quern S15 in levels sealing the disused structure show a similar pattern, supporting the notion of a deposit deliberately covering the abandoned house and including carefully selected artefacts. This suggests a ritualized abandonment process. These layers were subsequently disturbed by agriculture and bioturbation.

At Tremough there is evidence for the structured deposition of artefacts in Structure 338 and its enclosure ditches (Figs 19 and 21). The condition of artefacts, their placement within the enclosure ditches and posthole-packings and their likely dates suggest that this took place soon after construction of the enclosure ditches, and during construction, use and abandonment of Structure 338.

Outside the enclosure, the broken vessel PR28 and used hammerstone S16 from pit [377] also provides evidence of structured deposition. The pot had been deliberately smashed, perhaps using the hammerstone, and its 70 sherds (80-90% of the vessel) were placed with their outer faces uppermost in the pit before being backfilled. The hammerstone showed abrasive wear consistent with repeated use for breaking up stone, perhaps as part of a metalworking process. The pit is some distance to the south of the enclosure’s western entrance terminus, but

During both phases of the enclosure represented by ditches [306] and [565], fresh sherds of pottery were deposited in distinct groups (fig. 19). These present a very different patterning to the dumps of midden material sometimes found in round ditches (Quinnell below). The Phase 1 49

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough it related to continued use of the fields during the second century AD. Enclosure [306] was replaced by smaller enclosure [565] in the third century AD with Structure 338 continuing in use. The enclosures and Structure 338 relate to agricultural expansion, with the complex of intercutting multi-phase hearth-pits possibly connected with the processing of farm produce. A number of structured deposits, particularly associated with the southern entrance to the structure, suggest that ritualized discard of artefacts, whether domestic rubbish or status object, was as much a part of life for Romano-British populations as it was for their ancestors. The building went out of use in the fourth century AD and was purposefully sealed by a deposit of earth containing a number of specially selected objects. The site is particularly important as providing evidence for a timber version of a Cornish Roman-period oval house and for clear details of a small, non-round enclosure with a single structure, something never before located for this period in Cornwall.

its special nature may have been derived from ideas associated with marking of boundaries. In Iron Age Wessex, deliberately broken, decorated pottery has been recognised as being selected for deposition at spatial and conceptual boundaries, such as enclosure entrances (Hill 1995, 131). In Cornwall, a pit containing pottery and burnt bone was found close to the enclosure ditch at Pollamounter, a Romano-British site (Jones and Taylor 2004, 39) and reference has already been made to pits located near to the upper edge of the Stencoose RomanoBritish field system (Jones 2000-1). The range of items selected for deposition reflects the main activities within the enclosure. The majority comprise fragments of pottery, sometimes used, probably for cooking, and buried soon after breakage. Stone items are also well represented, including whetstones and the only quern and weight, all of which show use prior to breakage and discard; these are items perhaps linked to specific individuals or functions, with related memories and biographies. The need to reinforce a sense of belonging and continuity may have been at the root of the manner of their deposition. It is possible that as agriculture intensified there was an increasing need to justify ownership. This was expressed in the ritualized deposition of everyday artefacts with metaphorical links to agriculture and the production of food and possibly metalwork. Individuals are likely to have been guided by a set of functional and symbolic requirements, ‘providing a mnemonic for day to day action within ordered space’ (Moore 1982, 79). At times this may have had an informal and deeply embedded tradition, such as the dumping of certain domestic rubbish in a specific area of ditch. At others, the deposition of artefacts and other materials may have been part of a more complex ritual or overtly ceremonial nature, such as during the ‘closure’ of Structure 338. The selection of domestic artefacts could be taken to imply a mundane world view, but this should not be assumed to be so. As was the case in prehistory, the world of the Romano-British farmer was probably one where even everyday actions could be imbued with a symbolism, where ritual need not be distinct from the profane in terms of either the content or location of deposits (Hill 1995, chapter 10).

Post-Roman and early medieval periods Structure 338 was abandoned by the end of the Roman period and there is no evidence from the limited evaluation of the ‘Fort’ for post-Roman occupation (Gossip and Jones in prep), but the radiocarbon date 1015±45 BP, cal AD 890-1160 (AA-44599) from ditch [3] in Field 3 and unstratified early medieval finds (Quinnell below) from Fields 3 and 4 indicate some form of activity across the site, which included access to post-Roman amphorae from the Mediterranean (Bidwell below). There were also sherds of grass-marked pottery dating from the seventh century onward. Unfortunately, all but one sherd of this was unstratified (Quinnell below). Continued use or re-use of late prehistoric and Roman sites is frequent in Cornwall, for example at Boden where the round was constructed around 400 cal BC, had occupation in the Late Iron Age and early Roman periods and saw renewed activity in the sixth or seventh centuries AD (Gossip forthcoming). The archaeological evidence for this period at Tremough is extremely limited and the settlement foci unknown.

Summary

Documentary evidence confirms that the site was settled and farmed throughout the medieval period and later but again our information is limited to mostly unstratified finds, including ceramics and glass (see Thorpe and Tyson below), and ditches associated with removed field boundaries. The majority of the artefacts found are typical of assemblages from fields close to farming settlements – deriving from domestic midden material utilised for the manuring and improvement of fields. It is therefore possible that postRoman and medieval activity was located in the area of the site occupied by the current house.

During the Later Iron Age, brick-shaped ditched fields were laid out across the plateau. This field system extended westwards from the rectangular enclosure the ‘Fort’, probably the focus of settlement and possibly a round, and may have been part of a larger agricultural landscape. To the south of the ‘Fort’ was a smaller ‘D’ shaped enclosure, a possible stock-enclosure associated with the field system and the ‘Fort’. The addition of enclosure [306] at the western end of the field system with Structure 338 inside

50

Chapter 4

The artefacts Prehistoric, Roman and early medieval pottery Henrietta Quinnell with petrographic comment by Roger Taylor

less compact than the other gabbroic material from the site. The fabric has many parallels among the ‘medium ware’ gabbroic group at Carn Brea as defined by Smith (1981, 162) (observation by author of material in the Royal Cornwall Museum), although the ascription of a small featureless sherd on fabric alone can not be certain. Radiocarbon determination AA-44604 which calibrates to 3950-3660 cal BC at two sigma comes from this context and suggests a possible Neolithic date. The sherd may have been redeposited in the deliberate infill of ditch [76] together with the charcoal which provided the determination. Given the quantity of Neolithic lithics and the axe S1, the presence of Neolithic pottery might be expected. However, a fabric this soft would not survive unless protected within a stratified deposit and any comparable material in surface levels would have disintegrated (P1 in archive report).

This report presents the assemblages from the 2002 excavations together with those from watching briefs and other works in 2000; for the latter, the detailed account included in the archive report (Lawson-Jones 2002) has been revised for inclusion and referenced as Quinnell and Taylor archive report. The material includes a possible Early Neolithic sherd, Grooved Ware, Trevisker and Treviskerrelated material from the second millennium BC, a little Later Iron Age material, a sizeable Roman period assemblage and a small early medieval group including imported amphora sherds, grass-marked and bar lug wares. Fabric descriptions are included under the sequential chronological headings below. Supporting petrographic data from microscopic examination, described as PS No 1 etc, and thin-section study is included at the end of this section on pottery.

Grooved Ware (2002) (Figs 23-25) The assemblage consists of 175 sherds weighing 1933g and representing at least 21 vessels. It was found in eight pits and one spread in a group in the north western corner of the site. Of the 12 pits in this group only one contained pottery other than Grooved Ware, [296] with one small Romano-British sherd almost certainly intrusive. From ceramic data therefore all pits within the group may well be of Later Neolithic date.

Recording of sherd abrasion is based on the system devised by Sorenson (1996) for Bronze Age midden material at Runnymede with some modifications. The following descriptors and numeration are used in the report and on the database. Very fresh

1; recent breakage, hardly ever applicable

Fresh

1/2; colour of core slightly patinated but unaltered surfaces with sharp corners and edges

The fabrics are soft with open matrices containing numerous tiny air pockets. They are therefore likely to have been subject to considerable post-depositional alteration by ground water percolation and bioturbation. Sherd surfaces have generally been compacted by tooling and are less likely to have been affected than edges. However, in some cases there is a marked variation in weathering on the two sherd surfaces reflecting the way the piece lay in the ground. A number of sherds retain residue on their interiors although the exterior and edges are abraded. It is therefore uncertain how far the recorded degree of abrasion reflects the condition of sherds at the time of deposition and likely that most sherds were deposited in a fresher condition that the degree of abrasion indicates.

Moderate abrasion  2; core colour patinated, some definition in the sharpness of corners lost Abraded

2/3; core colour patinated, slight rounding of corners and very slight erosion of surfaces

High abrasion 3; core colour patinated, rounding of corners and of sherd outline, surfaces somewhat eroded

Generally, the material appears to have been freshly broken when deposited and differential weathering suggests that decorated surfaces were normally placed uppermost. Some sherds however, such as PP6, PP11, PP16 and PP17 are much more abraded than the remainder from their context,

Possible Early Neolithic (2000) Fill (77) in ditch [76], Field 7 (Not illus). A featureless scrap (3g, abrasion 2/3) of gabbroic fabric, softer and 51

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough On PP5 a vertical cordon has fingernail decoration and is bordered with incised lines. Paired fingernail decoration occurs on PP3. Close-set grooves on PP2 form patterns on both the inside and the outside of the vessel. PP4 has closeset grooved herringbone which appear to have covered the whole lower part of the vessel. PP8 and PP10 have evenly spaced horizontal cord impressed lines; these vessels are so similar even to their distinctive reduced interiors that they may in fact come from the same vessel. Horizontal impressed cord occurs alone on PP16. Probable combstamping occurs on PP10A.

and are also small; such pieces may well have been broken long before burial and curated. Fabrics Ga.1 Gabbroic 76% by sherds, 70% by weight, 61% by vessel numbers. The principal components all indicate a source on the gabbro some 15km to the south (Thinsections 1, 4, 5). Thin-section 1 contained small fragments of micaceous slate. Amount of inclusions vary but particles 0.5mm-1.5mm, occasionally up to 5mm, may form 20% of the body. The fabric may be oxidised or reduced, often with considerable variation in the same vessel: a distinctive feature is the darkness of reduction, sometimes 5YR 3/1 very dark grey, and the brightness of oxidisation, sometimes 2.5YR 4/8 red, with these two extremes sometimes occurring on the same sherd.

Comparanda and dating The general shape of the vessels with a number of slightly curved walls, the simple rims and the overall range of decorative techniques and designs, especially the use of impressed cord, indicates that the assemblage belongs within the Durrington Walls sub-style (Longworth 1971). The close-set grooved herringbone of PP4 may however, be more appropriate to the Clacton sub-style. The three southern British sub-styles defined by Longworth in 1971, Durrington Walls, Clacton and Woodlands, were seen as contemporary variations but more recent studies and the consideration of the radiocarbon dates now available indicates that Clacton and Woodlands may be sequential rather than contemporary (Garwood 1999, 157). The possible Clacton vessel PP4 is from pit [293] which has no date but which lies in the south of the Tremough group – as does pit [331] with the earlier of two radiocarbon determinations, 2870-2480 cal BC (Wk-14999). Pit [331] contains PP10 and PP10A. The later date 2570-2290 cal BC (Wk-15001) is from pit [494] in the north of the group, with undecorated pieces PP12-14. As the dates do not overlap, they should indicate deposition of Grooved Wares over a period of several centuries. The Durrington Walls sub-style was used throughout the depositional sequence but pit [293] with Clacton possible sub-style vessel and pit [494] with the earlier date lie in the southern part of the group. The later date comes from pit [494] to the north. This data may indicate that pits were dug at Tremough from south to north over several centuries.

Ga/Gr Gabbroic/granitic 17% by sherds, 25% by weight, 29% by vessel numbers. Microscopic scanning and Thinsection 3 indicated a mix of gabbroic clay with a ‘granitic’ component, biotite and muscovite, probably deriving from granitic or gneissic rocks situated in, or adjacent to, the gabbro. Inclusion size and quantity and colour are as Ga.1. Vq Vein-Quartz 7% by sherds, 5% by weight, 9.5% by vessel numbers. Crushed angular vein-quartz has been added to a clay which includes siltstone, hornfels, biotite and a granite fragment (Thin-sections 2, 6) which are appropriate to a clay source local to the site. Vein-quartz inclusions (0.5-5mm) form 25% of the fabric and the variation between reduction and oxidisation, and the deep colours these produce, is similar to those on gabbroic Ga.1. There are no variations in decorative motifs on vessels in the three fabrics which need be significant in such a small assemblage; all vessels (PP1, PP5, PP7, PP15) which have panels defined by vertical or oblique cordons are Ga.1. Ga.1 and Ga/Gr are distributed fairly evenly across the pit group but Vq fabric is only found in pit [329]. Form and decoration

The Tremough assemblage provides the seventh definite occurrence of Grooved Ware in Cornwall; details of four are given in the gazetteer compiled by Longworth and Cleal (1999) nos 42-45. Two of these, in addition to Tremough, are from pits, both in the process of publication. That from Trevorva Cott, Probus (Longworth and Cleal gazetteer no 45) consisted of a minimum of seven vessels from two pits, mostly in the Durrington Walls sub-style but with two vessels of Clacton type (Longworth forthcoming a). All the vessels from Trevorva Cott are of gabbroic fabric and dates from the two pits are respectively 28802460 cal BC (AA-29731) and 2590-2208 cal BC (AA29732) (Nowakowski forthcoming). Decoration on the Trevorva Cott vessels is generally more complex than at Tremough, with most illustrated pieces having grooved or incised designs covering the exteriors; fingernail is

Rim and base forms have been described according to Longworth’s (1971) system. Rims are simple rounded or slightly pointed with one flattened example (PP12) and one with a simple inturn (PP7) and come from vessels with straight or, more commonly, slightly incurving sides. Bases are generally simple with one protruding example (PP14). Decorative techniques include applied cordons, impressed cord, grooved lines, incised lines, fingernail and possible stab-and-drag. All sherds with decoration are described and illustrated. Insufficient of each vessel survives for complete decorative patterns to be reconstructed. Cordons combine horizontal with oblique (PP1), horizontal and vertical in a panel arrangement (PP7), horizontal (PP9) and vertical and oblique (PP15). 52

The Artefacts well-represented but there are no cord impressions. The single pit at Trevone, Padstow (SW 8895 775, not listed in Longworth and Cleal 1999), contained sherds from five vessels, four of Durrington Walls and one of Clacton sub-type. Petrography is not available but a radiocarbon determination 4175+50 BP (AA-26412) (from Pomoideae charcoal) calibrates to 2900-2610 BC. The Trevone pit vessels again have incised or grooved decoration covering their exteriors but in more complex arrangements than on those at Trevorva Cott; neither cord or finger impressions are present, but one sherd has stabbed decoration (Longworth forthcoming b). This latter is assigned to the Clacton substyle but the remainder of group is Durrington Walls. The mix of a possible Clacton sub-style vessel with a mainly Durrington Walls group at Tremough therefore reflects the rather more definite mix of the two sub-styles in the two other Cornish pit finds. The Tremough group however, has by far the simplest range of decorative motifs, with only one vessel PP4 having the whole exterior surface covered. Only Tremough has impressed cord, represented by three vessels.

incised decoration on rim and body. Zennor I (ibid) with complex internal and external cord decoration was not present for examination but may also be Grooved Ware. Tremough is thus the most simply decorated assemblage of Grooved Ware recognised in Cornwall and the only one with sherds of probable undecorated vessels, apart from Zennor. It is possible that simpler Grooved Ware vessels have not been recognised. Of these the most likely is the small plain straight-sided and open topped vessel with a simple lug from a pit at Poldowrian, St Keverne (Harris 1979, Fig. 9); possible Grooved Ware allocation is supported by the radiocarbon determination 2950-2050 cal BC (HAR 3108). A plain vessel similar to that from the Poldowrian pit except for the lug came from a cist at Trevedra Common, St Just (Thomas 1961, Fig. 18). Some of the sherds collected from the surface of Field 24 at Polcoverack, St Keverne, with incised or fingernail decorations may also belong within the Grooved Ware category (Smith 1987, figs 20-21); all this material has been identified as of gabbroic fabric.

One of the other Cornish Grooved Ware occurrences comes from a cave and two from old land surfaces beneath barrows. The former is an early twentieth century find at Carrick Crane Crags, St Keverne (Longworth and Cleal gazetteer no 44). Sherds from four vessels in the Durrington Walls style come from the floor of the Carrick Crane Crags cave; the illustrations (Patchett 1950, Fig. 1) show three vessels with overall exterior incised decoration and a fourth with possible fingernail; the published fabric descriptions suggest the use of gabbroic clay which would be appropriate for the location. One of the two barrow old land surface finds came from Trevone, Padstow (Longworth and Cleal Gazetteer no 43), a short distance from the pit group described above. This consisted of sherds of a single vessel in Durrington Walls style, with the exterior completely covered with incised decoration, from the old land surface beneath one cairn and a plain sherd in a similar fabric from a similar context beneath an adjacent cairn (Longworth 1972). The fabric can not at present be assigned to a source and there are no dates. The final occurrence comes from the barrow at Davidstow Moor Site XXVI (Longworth and Cleal gazetteer no 42). Three sherds from the same vessel came probably from the old land surface beneath the barrow (Healy in Christie 1988, Fig. 71, 121); they are decorated with grooved lines and triangular impressions, ascribed to the Durrington Walls sub-style but perhaps more at home in Clacton. The fabric is gabbroic (Williams in Christie 1988, 161), the first identified Grooved Ware gabbroic fabric from Cornwall and the more remarkable as Davidstow Moor is on the north side of Bodmin Moor, the other side of the county from the source area on the Lizard. The final occurrence comes from Zennor Quoit (SW 469380 not in Longworth and Cleal). Recent examination by the author confirms that the vessels published by Patchett (1944, Fig. 2) as Zennor II and Zennor III are definitely Grooved Ware and made of a fine granitic fabric: one is plain and the other has simple

The scarcity of Grooved Ware in Devon is even greater than in Cornwall. There are only three clearly acceptable occurrences, although occasional possible sherds have been found such as that from Widegate Nursery, Elburton (Watts and Quinnell 2001, 14). That from Three Holes Cave, Torbryan (Longworth and Cleal gazetteer no 59) includes, from unstratified material, a sherd decorated with incised lines and stab marks (Rosenfeld 1964, pl 1b), perhaps of the Clacton sub-style. The second occurrence includes sherds from a vessel with complex stab-and-drag decoration from Ash Hole Cave in Torbay. This has recently been identified by the author and is not in the Longworth and Cleal gazetteer; it is probable that the c. 2000 other sherds from the cave include further Grooved Ware vessels. The third occurrence from Bow (not in Longworth and Cleal gazetteer) consists of 57 sherds from three shallow pits: there are three vessels with grog temper and one of granitic fabric (Brown 2003). Decoration includes incised and grooved lines and also fingernail, the only site in Devon and Cornwall apart from Tremough to display this characteristic. The Bow material is probably to be assigned to the Durrington Walls sub-style. Details of the assemblage Pit [193] (194) (Fig. 23) PP1 was probably deposited as one large sherd and several smaller ones, which have disintegrated into small fragments. The largest, illustrated, sherd of PP1 is very abraded on its decorated surface and fresh on the inside. PP2 was probably deposited as a group of small sherds. No other vessels appear to be represented. PP1 Ga.1. Sherds from girth of large vessel, diameter c. 260mm, simple horizontal cordon with oblique cordon 53

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough Table 5: Details of Grooved Ware in pits and spread (496) Fabric

[193] (194) pit Total [193] [293] (294) pit

Total [293] [300] (299) pit [329] (330) pit Total [329] [331] (332) pit Total [331] Pit [333] (334) Pit [492] (493) [494] (495) pit

PP1 PP2 PP3 PP4 PP5 PP6

PP7 PP8 PP9 PP10 PP10A PP11 P12 PP13 PP14

Ga.1 20 0 20 0 0 8 1 1 10 3 0

0 0 2 29 31 1

Total (496) Totals

19 4 5 2 30 17 2 1 18 38 133

Overall totals

175

Total [494] (496) spread

PP15 PP16 PP17

Weight Ga/Gr

Vq

8 8 1 4

2 7 9

9 2

6 6 12

2 4

30

12

Abrasion

Minimum vessels

195 65 260 111 78 41 11 8 29 278 227 78 44 53 175 88 11 174 273 4 39 105 13 85 3 206 305 17 17 132 471 1348 488 97 1933

2 1/2 2 2 1/2 2/3

1/2 1/2 1/2 1/2 2 2/3 2 2/3 2 2 2 2 2 2 2/3 2/3 2/3

1 1 2 1 1 1 1

4+ 1 1 1 1 3 1 1 1 3 1 1 1 1 1 3 1 1 1 3+ 13+ 6+ 2 21+

Mean sherd weight 9.8g 8.1g 9.9g 111g 19.5g 36g 11g 4.2g 16.4g 75.7g 8.7g 7.3g 8.8g 8.3g 44g 5.5g 6g 8.3g 4g 9.8g 5.5g 3.3g 17g 1.5g 7.3g 17.9g 8,5g 17g 7.3g 13.6g

below and a possible second; part of simple base angle Longworth A not included in illustration.

part is missing. A base angle sherd indicates part of a further vessel in Ga/Gr.

PP2 Ga/Gr. Sherds from rim and girth of large vessel, diameter c. 300mm, decorated with groups of grooved lines on both exterior and interior, exterior lines interspersed with sparse fingernail, interior lines becoming shallow and more appropriately described as tooling. Simple rounded closed rim, Longworth Form 5.

PP3 Ga/Gr. Body sherd with paired fingernail decoration, with single oblique incised line. Maximum diameter c. 250mm. The angle as illustrated must be regarded as tentative. PP4 Ga/Gr. Basal sherds with close-set incised herringbone design which are irregular in places. Internal base diameter 110mm.

Pit [293] (294) (Fig. 23) PP3 was deposited as a single sherd; the interior is fresher than the decorated surface. PP4 was deposited as one large and one small sherd; on the former the interior is fresher than the decorated surface. PP5 was probably deposited as one sherd, with residue on the fresh interior set downwards; a number of small sherds have broken off. PP6 was abraded when buried, but fresh breaks indicate

PP5 Ga.1. Girth sherd, vertical cordon with finger tip/nail decoration, oblique incised lines to one side, surface flaked away on the other. PP6 Ga/Gr. Body sherd with horizontal and oblique incised lines; single circular ‘impression’ may have been caused by loss of an inclusion. 54

The Artefacts

Figure 23: Grooved Ware PP1-2 pit [193], PP3-6 pit [293].

55

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough

Figure 24: Grooved Ware PP7 pit [300], PP8-9 pit [329], PP10-10A pit [331], PP11 pit [333], PP12-14 pit [494]. 56

The Artefacts PP13 Ga.1. Plain rim of Form 4, pointed closed, with marked inturn, internal diameter 200mm.

Pit [300] (299) (Fig. 24) PP7 appeared to have been deposited as three sherds or one sherd which has subsequently broken.

PP14 GW. Gabbroic Fabric, very hard. Complete base with Form C slight protruding foot.

PP7 Ga.1. Vessel with slightly curved/biconical shape, rim sharply inturned Form 12. Internal rim diameter 210mm. Horizontal cordon with vertical cordons below, all plain; a scar in the gap on the horizontal cordon indicates that it may originally have been continuous.

Spread (496) (Fig. 25) PP15 had been deposited as large fresh sherds which had fragmented, while PP16 and PP17 were abraded when deposited. It is not certain whether the other featureless sherds belonged to PP15 or to another vessel.

Pit [329] (330) (Fig. 24) PP8 and PP9 were deposited as groups of fresh small sherds. PP9 and body sherd SF1025 from a separate vessel are the only vessels in Vq.

PP15 Ga.1. Non-joining sherds representing rim, girth and base of slightly open vessel, maximum internal diameter 260mm. Plain rim of Form 1 pointed upright. Base angle simple Form A with three close spaced plain cordons, centre vertical and the outer ones slightly oblique. This has been reconstructed in the illustration with the three other sherds, a girth sherd with plain vertical cordon, another with plain vertical and oblique cordons and a third with a plain vertical cordon.

PP8 Ga/Gr Girth sherds with three evenly spaced lines of horizontal impressed cord. Internal diameter c. 210mm. PP9 Vq Rim, Form 2 rounded upright, non-joining girth sherd with plain horizontal cordon.

PP16 Ga.1. Rim of Form 1 pointed upright with two horizontal twisted cord lines beneath. Second sherd, probably from PP16, suggests spacing of cord may not have been regular.

Pit [331] (332) (Fig. 24) PP10 was deposited as two sherds, probably with decoration uppermost. PP10A was worn before breakage. The other sherds, in Ga.1, were buried as a group and probably represent one vessel including a base angle.

PP17 Ga.1. Plain rim of Form 3 flattened upright with possible incised line beneath.

PP10 Ga/Gr. Rim of open vessel, Form 1 pointed upright, with horizontal lines of single cord impressions beneath. Internal rim diameter 160mm.

Second millennium cal BC: Trevisker and Treviskerrelated (2002) (Figs 26-28)

PP10A Ga.1. Body sherd with four impressions of a similar stamp which has at least three rectangular depressions along it and probably shows the use of a coarse combstamp.

The assemblage from 2002 consists of 230 sherds weighing 4822g and represents a minimum of 25 vessels.

Pit [333] (334) (Fig. 24)

Macroscopic examination indicated that most of the assemblage was gabbroic admixture with some sherds apparently of gabbroic fabric without additions. Microscopic and thin-section study showed that some admixture fabrics contained inclusions not found in the gabbro area. However, detailed studies, especially of the Trethellan assemblage (Williams 1991), have shown that it is not possible to group sherds reliably without thinsections and so overall percentages of fabric groups are not be presented.

Fabrics

PP11 was buried as a small and abraded sherd. PP11 Ga.1. Thin pointed rim of Form 4 type from a closed vessel. Pit [492] (493) Ga/Gr. Body sherds represent the minimum of one vessel. Pit [494] (495) (Fig. 24)

Gabbroic Ga.2 Well-worked matrix, good surface finish, few inclusions larger than 2mm. PS Nos 18, 21, 23 confirm identification. Very similar to Grooved Ware Ga.1; generally oxidised, mostly 5YR 6/6 reddish yellow, occasionally 5 YR 5/1 grey, the highly fired distinctive colours of the Grooved Ware fabric being absent. A minimum of four vessels.

PP12 was buried as two sherds, PP13 as several sherds which have subsequently fragmented. The base PP14 was buried in a complete state and upside down. PP12 Ga.1. Plain rim of Form 3 flattened upright, from a slightly open-topped vessel, internal rim diameter 160mm.

Gabbroic admixture 1 GaAd.1 Well-worked matrix, good surface finish, sparse to common inclusions which vary 57

Figure 25: Grooved Ware PP15-17 spread (496).

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough

58

The Artefacts from 3mm to 12mm, colour as Ga.2. Thin-section 7 shows the added inclusions to be micro-granitic, Thin-sections 9 and 10 to be dolerite and Thin-section 11 serpentinite; PS17 confirmed the presence of gabbro rock. All these added materials occur in the immediate area of gabbroic clays. Thin-section 7 showed a loessic component in the gabbroic clay. A minimum of 13 vessels.

Structure 392 (Fig. 26) There were a minimum of 2 vessels in Ga.2, 8 vessels in GaAd.1 and 2 vessels in GaAd.2, a minimum of 12 vessels overall. The Table shows sherds to be generally abraded, certainly more so than those in the Grooved Ware pits, but this may be due to the action of groundwater and other bioturbation factors on large soft sherds. Average sherd weight was 20.4g; the only contexts which varied much from this were pit [401] with an average of 54g and the inner post-ring with an average of 9g. However, the average sherd size of the illustrated vessels PP18-27 is 43g.

Gabbroic admixture 2 GaAd.2 As GaAd.1 but Thinsections 8, 12 and 13 showed the presence of tourmaline, hornfels and other material of which the source is likely to be the metamorphic aureole surrounding a granite, almost certainly very local to the site. A minimum of eight vessels.

Form, affinities and chronology

Gabbro admixture fabrics have been extensively discussed, especially by Parker Pearson (1990), by Williams (1991) and Quinnell (1998-9a). The studies on which these discussions are based showed that admixtures all could be sourced to the immediate gabbro clay area, although earlier comments, notably that on the eponymous assemblage from Trevisker (ApSimon and Greenfield 1972, 355), suggested the mixing of gabbroic materials with clays local to the site. The fabrics from Tremough clearly show that some gabbroic clay was been transported to the area of the site and there mixed with crushed minerals, and possibly some clay, of local origin. This has already been demonstrated in a limited way with the assemblage from Stannon (Quinnell forthcoming c) and the transport of gabbroic clay in an unfired state identified from Bronze Age levels at Gwithian (Nowakowski 2004). These clear results indicate that some gabbroic clays were being transported in the second millennium cal BC to be manufactured near the sites on which vessels were to be used, with the addition of local temper. The transport of clays may have been frequent. Now that some admixture inclusions have been identified as sourced away from the gabbroic, it is reasonable to reconsider those admixtures that could be sourced to the gabbro as possibly sourcing near to the sites on which the ceramics were used. Recent studies have emphasized the symbolic function of inclusions, and that these should not be thought of in terms of practicality but rather in terms of the addition of materials which added meaning to the vessels (Woodward 2002a). The tradition of using gabbroic admixture fabrics in the Cornish Bronze Age, fabrics in which inclusions were large and noticeable, suggests that it was important for the symbolic connections of vessels to be visible.

Contexts within Structure 392 produced two radiocarbon determinations, Wk-15002 calibrating to 1300-1010 BC at two sigma from hearth-pit [320] and Wk-15003 calibrating to 1370-1040 BC from posthole [623], (622). These two similar dates cover the 13th to the 11th centuries BC, the Middle Bronze Age during which Trevisker pottery appears to have been the only ceramic current in Cornwall. The assemblage from Trethellan (Woodward and Cane 1991) with radiocarbon determinations covering the 15th to 13th centuries BC forms a good reference for current understanding of Trevisker material during the later second millennium cal BC. All distinctive pieces from Structure 392 have been illustrated as PP18-27 and these together present characteristics which may be interpreted as falling within the Trevisker style combined with some unusual features. The flat-topped expanded rim of PP18 is a standard Trevisker rim form but the fingernail decoration along it is uncommon and the girth cordon with finger tip impressions unknown. If a vessel of this size, shape and decoration had been found in Dorset or the Wessex area it would be described as a typical bucket urn with DeverelRimbury affinities, see for example the range from the Simons Ground cemetery (White 1982). The finger tip decoration on rim and cordon however, also forms part of a wider tradition found for example at the Lesser Garth Cave in Glamorgan (Savory 1980, Fig. 72 no 505:4). PP19 and PP20 are both body sherds with incised herringbone, on PP19 combined with a horizontal line. PP20 appears to come from a vessel with an incised decoration around its girth bordered by horizontal lines and would be quite at home within a Trevisker assemblage (Woodward and Cane 1991, Fig. 46, no 36); PP23 is broadly similar to PP20. PP19 and PP25 have coarser, unbordered, herringbone incisions, again found at Trethellan and elsewhere (Woodward and Cane 1991, Fig. 48, no 43). The slightly everted and expanded rim with an incised line beneath on PP22 comes from a vessel with Trevisker rim form with the very top of a bordered incised design beneath. The plain cordon on PP26 has Trethellan parallels (ibid, Fig. 46, no 37) but the heavy cordon with fingernail PP21 has no recognised parallels. Fingernail impressions form 5% of the decorative motifs at Trethellan and were therefore a technique in general use within the Trevisker tradition. PP24 with fingernail impressions on a girth sherd could

GaAd.2 vessels were found in Structures 392, 102, Posthole Group 37 and ditch [596] and in all cases were accompanied by GaAd.1 vessels. This indicates that its use was current from the 15th to the 11th centuries BC. Its absence from Structures or pits dated before this period can not be assessed due to their general scarcity of ceramics. The practice of transporting gabbroic clay to the Tremough area and mixing it with inclusions and possibly other material of local origin covers the whole of the second half of the second millennium cal BC. 59

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough

Figure 26: Trevisker/Trevisker-related ceramics PP18-27 structure 392.

be matched by No 4, Fig. 40 at Trethellan. Finally PP27 with its four rows of cord impressions has no recorded parallel. The vessels are too incomplete to be assigned to the classification devised by Parker Pearson (1990, 10).

of contact with the ceramics of Wessex, or to the specific choice of vessels with particular decorative motifs for use in a Structure of the type of 392. All Middle Bronze Age Trevisker assemblages so far studied come from domestic assemblages within houses with either sunken floors or with stone walls and inner post-rings; on such undoubted domestic structures the type of vessel appropriate for

The presence of unusual features may be due to the comparative late date of the assemblage, to some form 60

The Artefacts Table 6: Pottery from Structure 392. Features grouped from the outside inwards. Intrusive Roman gabbroic sherds from [323] [414] [401] (318) 2 Context PP Entrance features [483] (484) posthole [480] (479) posthole [485] (486) posthole [501] (502) posthole Outer post-ring clockwise from entrance [314] (315) [325] (326) [398] (397) PP18 [327] (328) Between post-rings [401] (402) pit [401] (435) pit Inner post-ring [511] (512)

Sherds/weight

Abrasion

Comment

1/16 5/89 1/1 3/80

2 1/2 1/2 2

Ga.2 Ga.2 Ga.2 Ga.2

1/36 8/88 1/146 2/15

2 2 2 2

Ga.2 GaAd.1 GaAd.2 GaAd.1

4/350 3/28

2 2

GaAd.1 PS22 TS 9 dolerite GaAd.1

10/93

2

GaAd.1

[455] (454) [439] (438) [449] (448) [478] (477) On or within inner post-ring [316] (317) pit

3/12 4/38 1/4 6/68

1/2 2 3 2-3

GaAd.1 GaAd.1 GaAd.1 Ga.2 PS23

1/16 6/70 1/11 4/26 5/51 5/17 5/232 1/20 1/22 1/24 1/14 1/71 1/58 39/786 23/500

2 2-3 2 2-3 2-3 3 2 2 2 2 1/2 2 2 2-3 2-3

Ga.2 GaAd.1 GaAd.1 with gabbro GaAd.1 GaAd.1 GaAd.1 GaAd.1 PS21 TS10 with dolerite GaAd.2 GaAd.2 GaAd.1 with gabbro GaAd.1 with dolerite GaAd.1 with serpentinite GaAd.2 GaAd.1, GaAd.1, GaAd.2 PS28 TS13 PS27

22/293

2-3

GaAd.1, GaAd.2

1/8 1/12 1/46 1/13 174/3554

2 2 2 3

GaAd.1 GaAd.1 GaAd.1 GaAd.1 PS25 TS11 with serpentinite

(318) spread [567] (568) posthole [320] (577) hearth-pit upper fill [320] (319) hearth-pit main fill

[613] (614) posthole cut in hearth-pit (320) [618] (619) posthole cut in hearth-pit (320) [623] (622) [504] (505) pit [665] (666) posthole [565] (566) RB ditch Total

PP19 + PP20 +

PP21 PP22 PP23 PP24 PP25 PP26 PP27 +

the functions carried out within the buildings may well have been different to those carried out within post-ring structures. Our perceptions of Trevisker material may be influenced by that found on domestic settlements away from the Lizard area. It may be noted that much of the Bronze Age pottery published from Polcoverack on the Lizard (Smith 1987), which comes from surface collection and for which therefore the context of use is unknown is difficult to characterise and includes a range of pieces with fingernail decoration not assigned to any particular style. The same appears to be true for the assemblage from

Kynance Gate only covered by interim reports (Thomas 1960). The nature of deposition in this Structure is generally uncertain. The unusual PP18 was found in [398], the posthole which lies at its rear directly opposite the centre of the entrance and for this deliberate deposition seems likely. PP19 occurred in [316] which lies on a direct line through the entrance to [398] and the remaining pieces are either in central spread (338) (PP20) or the central hearthpit [320] (PP21-7). The occurrence of all these pieces 61

Archaeological investigations of a later prehistoric and Romano-British landscape at Tremough PP25 [320] (319) hearth-pit GaAd.1 with dolerite. Girth sherd with slashes forming herring bone design. Fabric 30% inclusions