South Gas Pipeline Project Site 512 Land North of Steynton Archaeological Excavation

for Rhead Group on behalf of National Grid

CA Project: 9150 CA Report: 13355 Event DAT108845 April 2014

South Wales Gas Pipeline Project Site 512

Archaeological Excavation

CA Project: 9150 CA Report: 13255 Event DAT102846

Alistair Barber, Senior Project Officer, prepared by Christopher Leonard, Project Officer and Jonathan Hart, Senior Publications Officer

date 9 April 2014

checked by Karen E Walker, Post-Excavation Manager

date 22 April 2015

approved by Martin Watts, Project Director, Head of Publications

signed

date

issue 01

This report is confidential to the client. Cotswold Archaeology accepts no responsibility or liability to any third party to whom this report, or any part of it, is made known. Any such party relies upon this report entirely at their own risk. No part of this report may be reproduced by any means without permission.

© Cotswold Archaeology

Cirencester Milton Keynes Andover Building 11 Unit 4 Office 49 Kemble Enterprise Park Cromwell Business Centre Basepoint Business Centre Kemble, Cirencester Howard Way, Newport Pagnell Caxton Close, Andover Gloucestershire, GL7 6BQ MK16 9QS Hampshire, SP10 3FG t. 01285 771022 t. 01908 218320 t. 01264 326549 f. 01285 771033 e. [email protected] © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

CONTENTS

SUMMARY ...... 2

1. INTRODUCTION ...... 3

2. RESULTS (FIGS 2–7) ...... 6

3. PROJECT TEAM ...... 10

4. REFERENCES ...... 11

APPENDIX A: CONTEXT DESCRIPTIONS ...... 13

APPENDIX B: THE FINDS ...... 15

APPENDIX C: PALAEOENVIRONMENTAL EVIDENCE BY JAMES RACKHAM ...... 16

APPENDIX D: THE RADIOCARBON DATING BY SEREN GRIFFITHS ...... 32

LIST OF ILLUSTRATIONS Fig. 1 Site location plan (1:25,000) Fig. 2 Plan of archaeological features (1:250) Fig. 3 Sections (1:20) Fig. 4 Photograph: trough 264075, looking east Fig. 5 Photograph: pit 264003, looking north Fig. 6 Photograph: trough 264039, looking west Fig. 7 Photograph: burnt mound 264030, looking north-west Fig. 8 A Bayesian model for the results from Site 512

GLOSSARY CA – Cotswold Archaeology CAP – Cambrian Archaeological Projects DAT – Archaeological Trust HER – Historic Environment Record MHA – Milford Haven to Aberdulais gas pipeline NAL – Network Archaeology Ltd NLMJV – Nacap Land & Marine Joint Venture UPD – Updated Project Design

1 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

SUMMARY

Project Name: South Wales Gas Pipeline Project Location: Site 512, Land North of Steynton, Milford Haven, Pembrokeshire NGR: SM 9147 0833 Type: Excavation Date: 20 July–1 August 2006 Location of Archive: To be deposited with RCAHMW (original paper archive) and Scolton Manor Museum (material archive and digital copy of paper archive; accession number 2008.1) Site Code: MHA 06

An archaeological excavation was undertaken by Cotswold Archaeology during groundworks associated with construction of gas pipelines (part of the South Wales high pressure gas pipeline scheme) between Milford Haven and Aberdulais, and Felindre and Brecon, which were conducted between 2005 and 2007.

Two palaeochannels were found either side of a slight topographical rise close to an existing stream. Two troughs were cut into the higher ground between the channels and were partially overlaid by a crescent-shaped burnt mound. The mound was composed of up to seven layers and was probably the result of prolonged but intermittent activity. Radiocarbon dates from the mound layers and one of the troughs indicate a Middle Bronze Age date, which was confirmed by the recovery of pottery sherds of the same date from the mound deposits. An older but residual piece of charcoal may indicate that the mound had earlier origins and could therefore have been contemporary with the final use of a ring ditch and associated urned cremations found 300m to the west at pipeline Site 513.

A pit containing charred cereal remains was cut through the mound and was radiocarbon dated to the early medieval period. This pit contained abundant cleaned cereals and it is possible that the abandoned mound provided a useful working platform for early medieval cereal processing or food production activities.

2 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

1. INTRODUCTION

1.1 NACAP Land and Marine Joint Venture (NLMJV), on behalf of National Grid, commissioned RSK Environment (part of the RSK Group) to manage the archaeological works (non-invasive surveys, desk based assessment, evaluation, watching brief, and open area excavation) on a 216km-long section of pipeline from Milford Haven (Pembrokeshire) to Brecon (in Powys). The high pressure gas pipeline (part of the 316km-long pipeline route from Milford Haven to Tirley in Gloucestershire) was required to reinforce the gas transmission network. The archaeological work performed in advance of this pipeline was undertaken in a number of sections by a number of archaeological companies. The westernmost section of 122km, from Milford Haven to Aberdulais, was investigated by CA (then Cotswold Archaeological Trust) during 2005–2007 with some additional excavation work carried out by CAP. The section of 89km, from Felindre to Brecon was investigated by CA during 2006–2007 and CAP during 2007. Assessment reports on the works were completed in January 2012 (NLM 2012a, 2012b) and the current reporting stage was commissioned in February 2013.

1.2 In July and August 2006 CA carried out an archaeological excavation at Site 512, Land North of Steynton, Milford Haven, Pembrokeshire (centred on NGR: SM 9147 0833; Fig. 1). The objective of the excavation was to record all archaeological remains exposed within the site during the pipeline construction.

1.3 The excavation was carried out in accordance with professional codes, standards and guidance documents (EH 1991; IfA 1999a, 1999b, 2001a, 2001b, 2001c, 2001d and IfA Wales 2008). The methodologies were laid out in an Archaeological Management Plan (RSK 2006) and associated Written Statements of Investigation (WSIs) and Method Statements.

The site 1.4 The site lies within a pasture field on ground that slopes gently down to the south towards a stream which flows along the southern field boundary (Fig. 1). The site lies at 40m OD and the surrounding landscape is characterised by gently rolling hills and stream valleys north of Milford Haven.

3 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

1.5 The underlying solid geology comprises the Milford Haven Group Argillaceous Rocks and Interbedded Sandstone of the Devonian and Silurian periods; no superficial deposits are mapped (BGS 2013).

Archaeological background 1.6 No archaeological remains were identified within the site during the preliminary Archaeology and Heritage Survey (CA 2005) and none are recorded in the immediate vicinity on the Dyfed Archaeological Trust Historic Environment Record (DAT HER). However, Prehistoric sites are recorded in the wider vicinity both by the DAT HER and within the Archaeology and Heritage Survey report. These include a standing stone (NPRN 305351), Long Stone Burial Chamber chambered tomb (ibid., NPRN 276027) and an eroded round barrow (CA 2005, ref. ID 35). A further cropmark possibly representing a barrow has been identified c. 400m to the south of the Site (CA 2005, ref. ID 28). During the pipeline construction, a ring ditch with urned Bronze Age cremations was found 300m east of the site at Site 513, adjacent to a small group of Neolithic pits (CA 2014a) whilst at pipeline Site 505, 800m south- east of Site 512, a barrow ditch with Early Bronze Age pottery was uncovered (CA 2013). A burnt mound is located 3km north-west of the site (NPRN 276065) and a second possible burnt mound is located at Barretts Hill, 1km south-east of the site (HER ref. 3183). Further mounds were recorded along the length of the pipeline (Hart et al. forthcoming).

1.7 Features were first identified within the site during a geophysical survey undertaken in advance of the pipeline construction (Bartlett 2005). A subsequent evaluation by CA revealed an undated pit containing burnt stone together with further burnt stone deposits within adjacent hollows (CA 2009, evaluation Site 34, trenches 115–116). These features were subsequently recorded during the excavation.

Archaeological objectives 1.8 The objectives of the archaeological works were:- • to monitor groundworks, and to identify, investigate and record all significant buried archaeological deposits revealed on the site during the course of the development groundworks; and • at the conclusion of the project, to produce an integrated archive for the project work and a report setting out the results of the project and the archaeological conclusions that can be drawn from the recorded data.

4 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

Methodology 1.9 Following consultation between RSK ENSR Environment, Dyfed Archaeology (formerly Cambria Archaeology (Heritage Management)) and CA, an initial watching brief during topsoil stripping (coded as Site 264) was followed by an archaeologically controlled machine strip undertaken to define the archaeological remains identified during the evaluation. The limits of a burnt mound were identified. The mound was divided into quadrants, three of which were hand excavated. The fieldwork followed the methodology set out within the WSI (NLM 2006). Site 512 was excavated and recorded using a continuation of the numbering system used initially for watching brief Site 264, with 264 prefixes throughout.

1.10 Where archaeological deposits were encountered written, graphic and photographic records were compiled in accordance with CA Technical Manual 1: Fieldwork Recording Manual.

1.11 The post-excavation work was undertaken following the production of the UPD (GA 2012) and included re-examination of the original site records. Finds and environmental evidence was taken from the assessment reports (NLM 2012a) except where the UPD recommended further work, in which case the updated reports were used. The archaeological background to the site was assessed using the following resources:- • the Archaeology and Heritage Survey which was undertaken in advance of the pipeline construction and which examined a 1km-wide corridor centred on the pipeline centre line, including the then existing HER record (CA 2005b); • Dyfed Archaeological Trust HER data, and • other online resources, such as Google Earth and Ordnance Survey maps available at http://www.old-maps.co.uk/index.html. All monuments thus identified that were relevant to the site were taken into account when considering the results of the fieldwork.

1.12 The archive and artefacts from the excavation are currently held by CA at their offices in Kemble. Subject to the agreement of the legal landowner the artefacts will be deposited with Scolton Manor Museum under accession number 2008.1, along with a digital copy of the paper archive. The original paper archive will be deposited with the RCAHMW.

5 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

2. RESULTS (FIGS 2–7)

2.1 This section provides an overview of the excavation results; detailed summaries of the recorded contexts, finds, environmental samples (palaeoenvironmental evidence) and radiocarbon dates are to be found in Appendices A, B, C and D. Full, original versions of the specialist reports are contained within the archive.

2.2 The natural geological substrate 264002, comprising pink/red sandstone and yellow- white sandy clay with stones, was cut by archaeological features and overlain by burnt mound deposits. Two palaeochannels (264031 and 264033) ran through the site, one on a north/south alignment, the other more north-west/south-east. These flowed either side of a slight natural rise in ground level and were perhaps braids or successive channels of the same water course, although this may have related to seasonal run off (see Appendix C for further discussion); a later channel, 264017, is discussed below. They contained silt-clay fills indicative of gradual silting.

Neolithic 2.3 A single unstratified Neolithic flint scraper fragment was recovered.

Bronze Age 2.4 Two cut features (264075 and 264039) were found on the higher ground between the palaeochannels and given the presence of a burnt mound (see below), these are interpreted here as troughs. Trough 264075 was sub-rectangular in plan, 2.7m long, 1.3m wide and 0.4m deep (Fig. 3, section AA and Fig. 4). The smaller trough, 264039, was 1.7m long, 0.87m wide and 0.4m deep and was more oval in plan (Fig. 3, section BB and Fig. 6). Both troughs had generally vertical sides and flat to slightly rounded bases.

2.5 Trough 264039 contained an initial silting fill (264040/264054) which included crumbs of undecorated prehistoric pottery. This was overlain by a backfill of charcoal and burnt stone (264041/264055) and the final fill (264056) comprised natural infilling but included one fragment of fresh flint microdebitage, not closely dateable. All three deposits contained frequent fuelwood charcoal, derived mainly from alder and hazel with some oak also present. A charcoal fragment from fill 264056 was radiocarbon dated to 1450–1300 cal. BC (SUERC-55487).

6 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

2.6 Trough 264075 also contained an initial silting fill (264076). This was overlain by stony sandy silt fill 264077, perhaps a deliberate backfill derived from upcast material (possibly, for example, from the excavation of a previous trough, although this cannot be proven). Both fills contained fuelwood charcoal and a charred hazel nutshell was recovered from fill 264077.

2.7 The troughs were partially sealed (and, in the case of trough 264075, partially filled by) by a crescent-shaped burnt mound, 264030 (Fig. 2, Fig. 3, section CC and Fig. 7). This mound directly overlaid the natural substrate, largely on the higher ground between the palaeochannels. As well as partially sealing the troughs, it also overlaid the width of the western palaeochannel and very slightly overlaid the eastern palaeochannel. There was no evidence for a former land surface beneath the mound. The mound itself covered an area 12m to 18m in extent and was typically 3.5m to 5m wide and up to 0.5m deep. It had accumulated within and over natural undulations in the ground surface and was composed of up to seven layers of charcoal with varying proportions of burnt mudstone, sandstone and gravel.

2.8 Finds from the burnt mound comprised two Bronze Age pottery sherds from upper layer 264016 and a single piece of fresh flint microdebitage from layer 264063. Close dating of the pottery is difficult, although it is possible that they belong to the Middle Bronze Age bucket and barrel urn traditions dateable to c. 1400-1000 cal. BC (Appendix B), a period from which pottery is rarely discovered in Wales (Ken Murhpy pers. comm. to Jonathan Hart). Bulk and column samples from the burnt mound recovered charcoal which was dominated by the remains of fuelwood, mainly from oak and alder supplemented by other readily available species. A small number of charred barley and wheat cereal grains and a few charred hazel nutshell fragments were also present.

2.9 Radiocarbon dating of charcoal from the mound produced Middle Bronze Age dates of 1510–1390, 1510–1400 and 1500–1320 cal. BC (SUERC-55494, -55495 and - 55488), ranges compatible with the tentative identification of the pottery and with the radiocarbon date range from trough 264039.

Early medieval 2.10 Pit 264003 (Fig. 3, section DD and Fig. 5) was cut through the burnt mound. It was 3m long, 0.7m wide and 0.5m deep and included a 1.15m-long gully (264047) extending eastwards from its eastern edge. This gully was far shallower than the pit

7 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

(0.15m deep) and was perhaps used to drain excess water from the pit towards the palaeochannel on the eastern side, probably indicating that the higher ground on the western side of the pit was a working area. Several in situ pitched slabs found against the southern edge of this pit were possibly part of a stone lining, or perhaps chocks for a wooden or wicker lining.

2.11 The lowest fill within the pit was clay silt 264004. Samples from this yielded charred cereal grains, mainly hulled barley and free-threshing wheat as well as oats, along with charred weed seeds and charcoal. It was overlain by what was probably a deliberate backfill of charcoal and unburnt pebbles (264005) deposited following the disuse of the pit. Samples from this fill included large quantities of charred cereal grains including six-row hulled barley, free-threshing wheat and oats. These were cleaned grains that had been processed and the charcoal derived from fast-burning heath species and it is likely that these remains point to crop processing on the site. Radiocarbon dating of charcoal from fill 264005 returned a single date of 1740–1560 cal. BC (SUERC-55486) and two statistically consistent dates between cal. AD 660– 880 (SUERC-55492 and -93). It is likely that the earlier date is from residual material and that the early medieval dates are more indicative of the age of the feature. The final fill (264006) was a natural infill within what must have survived as a remnant hollow.

Undated 2.12 A small shallow pit (264023) and a pit/posthole (264051) were found immediately north-west of the burnt mound (Fig. 3, sections EE and FF respectively). Both were undated but may have been related to the mound: in particular, feature 264023 may in fact have been a natural hollow into which mound material collected.

2.13 A third palaeochannel, 264017, truncated the western edge of the burnt mound. It contained a silt fill devoid of finds.

Discussion 2.14 The remains provide a well-preserved example of a burnt mound and associated troughs located close to a water course. No hearths were present, although it is possible that these lay beneath the unexcavated part of the mound, or had been entirely truncated, and it is possible that small fragments of fired clay recovered from the troughs (see Appendix C) may have derived from hearths. The charcoal from the mound and trough was dominated by fuelwood from readily available species. The

8 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

presence of cereals is of note but these were present in very small quantities and do not suggest that the mound was used for cereal processing. The mound lies within 50m of an extant stream channel, and this was the likely source of water for the mound; the palaeochannels running through the site may have been seasonally available but not, perhaps, a reliable source of water. The source of the stone is uncertain, although it may have been transported from the stream channel, or even from the palaeochannels or any nearby exposures such as tree-throw pits.

2.15 Analysis of two column samples taken through the mound suggests that the mound formed as the result of a sequence of deposition events which included a period of hiatus represented by a weathering deposit. Bayesian analysis of the radiocarbon dates from the mound and trough suggest that this activity began during 1510–1420 cal. BC (95% probable) and ended in 1470–1380 cal. BC (96% probable), although this of course only reflects the dated material. The older charcoal fragment from early medieval pit 264003 may not derive from the mound but, if it did, would push the start date for the mound back to 1740–1560 cal. BC. Although the start and end dates for the mound activity overlap, the fact that the mound comprised several layers and included at least two troughs suggests that it developed as a result of episodic use. Bayesian analysis of radiocarbon dates from sequences within two burnt mounds at Town Farm Quarry, Burlescombe, Devon indicated that one mound was in use over a period of between 10 and 170 years and had gone out of use before the second mound formed, itself in use for up to 60 years (Gent 2007, 37). It may well be that the mound on the current site similarly accumulated over decades, as the result of periodic events.

2.16 It is worth noting that the known dating for this mound suggests that it was not contemporaneous with the Early Bronze Age ring ditch and associated cremations found at Site 513, 300m to the west (CA 2014a). however, Bayesian analysis of the radiocarbon dates from the ring ditch cremations suggest that the cremation site was in use between 2020–1890 and 1870–1700 cal. BC (95% probable; ibid.), and if the earliest radiocarbon date from the mound site is taken into account (i.e., the residual charcoal from early medieval pit 264003; 1740–1560 cal. BC), then it is just possible that the mound site was in use towards the end of the use of the cemetery.

2.17 During the excavation it was assumed that pit 264003 was a further trough for the mound; however the radiocarbon dates from this pit, along with the fact that it was cut through the mound, showed it to have been an early medieval feature.

9 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

Excavations at a burnt mound near Yate, South Gloucestershire revealed a number of early medieval pits close to a Bronze Age burnt mound (CA 2014b). This hints at a possible model of re-use of some burnt mound sites during the early medieval period, suggesting that they remained visible within the landscape (presumably differentiated by the overlying vegetation) and were perhaps chosen as they provided useful working platforms for activities which in this case may relate to cereal processing or food production. The fuelwood charcoal and processed cereals from the pit indicate that it was related to cereal processing and it is possible that it was the remains of a field oven used for drying crops.

2.18 It is not known whether the western channels still occasionally flowed when the burnt mound activities commenced. Certainly one had silted up during the lifetime of the mound, and had been sealed by a large part of the mound. The easternmost channel was only slightly overlaid by mound deposits and could still have flowed or remained as a damp area.

3. PROJECT TEAM

Fieldwork was undertaken by Stuart Joyce, assisted by Greg Jackson, Vicky Rees, Sian Reynish, Sean Rice and Carina Summerfield. This report was written by Alistair Barber, Jonathan Hart and Christopher Leonard with illustrations prepared by Daniel Bashford. The archive has been compiled by Jonathan Hart and prepared for deposition by Hazel O’Neill. The fieldwork was managed for CA by Clifford Bateman and the post-excavation was managed for CA by Karen Walker.

10 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

4. REFERENCES

Bartlett, A. D. H. 2005 Milford Haven to Aberdulais Proposed Gas Pipeline. Report on Archaeogeophysical Survey of Proposed Gas Pipeline 2005 BGS (British Geological Survey) 2013 Geology of Britain Viewer. Online resource at http://mapapps.bgs.ac.uk/geologyofbritain/home.html accessed 4 November 2013 Britnell, W. J., Silvester, R. J., Gibson, A. M., Caseldine, A. E., Hunter, K. L., Johnson, S., Hamilton-Dyer, S. and Vince, A., 1997 ‘A Middle Bronze Age Round House at Glanfeinion, near Llandinam, Powys’, Proc. Prehist. Soc. 63, 179–97 CA (Cotswold Archaeology) 2005 Milford Haven to Aberdulais Gas Pipeline: Archaeology and Heritage Survey. CA typescript report 04147 CA (Cotswold Archaeology) 2009 Milford Haven to Aberdulais Natural Gas Pipeline: Archaeological Evaluation. CA typescript report 09078 CA (Cotswold Archaeology) 2013 South Wales Gas Pipeline Project. Site 505, Steynton Ring Ditch, Steynton, Pembrokeshire: Archaeological Excavation. CA typescript report 13248 CA (Cotswold Archaeology) 2014a South Wales Gas Pipeline Project. Site 513, Land North- West of Steynton, Milford Haven, Carmarthenshire: Archaeological Excavation. CA typescript report 13261 CA (Cotswold Archaeology) 2014b Autumn Brook, Yate, South Gloucestershire: Post- excavation Assessment and Updated Project Design. CA report 14149 EH (English Heritage) 1991 The Management of Archaeological Projects 2 GA (Groundwork Archaeology) 2012 Milford Haven to Aberdulais and Felindre to Brecon High Pressure Gas Pipelines: Updated Project Design Gent, T. 2007 ‘Bronze Age Burnt Mounds and Early Medieval Wells at Town Farm Quarry, Burlescombe,’ Proc. Devon Archaeol. Soc. 65, 35–46 Gibson, A. 2013 Milford Haven to Aberdulais and Felindre to Brecon Gas Pipeline: The Neolithic and Bronze Age Pottery from the Archaeological Investigations, Gibson typescript report no. 121 Greig, J, 1991 ‘The British Isles’, in W. van Zeist, K. Wasylikowa and K-E. Behre (eds) Progress in Old World Palaeoethnobotany. Rotterdam, 229-334 Hall, R. 2008 ‘Assessment Report for Fired Clay and Daub’, in NLM 2012a Hart, J., Rackham, J., Griffiths, S., and Challinor, D. forthcoming ‘Burnt mounds along the Milford Haven to Brecon gas pipeline, 2006-7’, forthcoming in Archaeologia Cambrensis IfA (Institute for Archaeologists) 1999a Guidelines for Finds Work. IfA, Birmingham

11 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

IfA (Institute for Archaeologists) 1999b Standard and Guidance for Finds and Ecofact Studies and Curation. IfA, Reading IfA (Institute for Archaeologists) 2001a Standard and Guidance for the Collection, Documentation, Conservation and Research of Archaeological Materials IfA (Institute for Archaeologists) 2001b Standard and Guidance for Field Evaluation IfA (Institute for Archaeologists) 2001c Standard and Guidance for an Archaeological Watching Brief IfA (Institute for Archaeologists) 2001d Standard and Guidance for Archaeological Excavation IfA Wales (Institute for Archaeologists of Wales/Cymru) 2008 Introducing a Research Framework for the Archaeology of Wales, online resource at http://www.archaeoleg.org.uk/intro.html accessed December 2008 NLM (Nacap Land and Marine) 2006 Milford Haven to Aberdulais Natural Gas Pipeline: Scheme of investigation for a programme of archaeological works NLM (Nacap Land and Marine) 2012a Milford Haven to Aberdulais High Pressure Gas Pipeline: Archaeology Assessment of Potential for Analysis NLM (Nacap Land and Marine) 2012b Felindre to Brecon High Pressure Gas Pipeline: Archaeology Assessment of Potential for Analysis Pannett, A. 2009 ‘Assessment Report for Lithics’, in NLM 2012a RSK (RSKENSR) 2006 Milford Haven to Aberdulais Natural Gas Pipeline: Archaeological Management Plan. Nacap Land and Marine Final, RSKENSR Environmental Ltd Savory, H. N. 1980 Guide Catalogue of the Bronze Age Collections. Cardiff, National Museum of Wales Stanford, S. C. 1982 ‘Bromfield, Shropshire-Neolithic, Beaker and Bronze Age Sites, 1966- 79’, Proc. Prehist. Soc. 48, 279–320 Ward, G. K. and, Wilson, S. R. 1978 ‘Procedures for Comparing and Combining Radiocarbon Age Determinations: a critique’, in Archaeometry 20, 19–31

12 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

APPENDIX A: CONTEXT DESCRIPTIONS

Context Fill of Interpretation Description L W Depth Spot (m) (m) (m) date 264000 Topsoil Brown-grey clay silt 0.15 264001 Subsoil Yellow-brown sand clay 0.25 264002 Natural Red sandstone and yellow-white stony sandy clay 264003 Pit Sub-circular, steep sides, flat base 3.1 0.7 0.5 264004 264003 Pit fill Mid grey-brown clay silt 3.1 0.6 0.15 264005 264003 Pit fill Dark grey-brown pebbly clay silt 3.1 0.6 0.35 Cal AD 660– 880 264006 264003 Pit fill Mid grey-brown clay silt 3.1 0.55 0.25 264007 Natural hollow Irregular with uneven, gently- 2.5 0.45 sloping sides and base 264008 264007 Burnt mound Yellow-blue silt clay 0.5 0.1 264009 264007 Burnt mound Light red-brown silt clay 1 0.1 264010 264007 Burnt mound Charcoal and burnt stone 1.6 0.2 264011 264007 Burnt mound Yellow-red clay silt and burnt stones 1.05 0.2 264012 Natural Light pink-brown silty clay 264013 Natural hollow Sloping sides, concave base 3.15 0.3 264014 264013 Burnt mound Mid pink-brown stony silty clay 0.8 0.1 264015 264013 Burnt mound Mid grey-brown clay silt with 0.9 0.3 abundant burnt stone 264016 264013 Burnt mound Dark grey-black silt and burnt stone 2.95 0.45 ?MBA 1510– 1390 cal BC 264017 Palaeochannel U-shaped profile 1.35 0.35 264018 264017 Palaeochannel Mid brown stony silty sand 1.05 0.15 fill 264019 264017 Palaeochannel Light pink-brown silty sand 1.3 0.2 264020 264017 fill Mid brown silty sand 0.8 0.1 264021 264007 Burnt mound Mid orange-yellow silty clay 0.95 0.15 264022 264007 Burnt mound Mid orange-red silty clay 1.55 0.25 264023 Pit Sub-circular with gently sloping 0.65 0.5 0.05 sides and concave base 264024 264023 Pit fill Dark grey brown sandy silt with 0.65 0.5 0.05 burnt stones 264025 264007 Burnt mound Grey-pink-brown silty clay 2.5 0.15 264026 Palaeochannel = 264033 264027 264026 Palaeochannel Light brown-grey silty clay 4.7 0.3 fill 264028 Natural hollow Steeply sloping sides, concave base 1 0.25 264029 264028 Burnt mound Dark orange-brown silty clay 1 0.25 264030 Burnt mound Group number given to all mound 18.0 5.0 0.5 deposits 264031 Palaeochannel Irregular profile 8 0.6 264032 264031 Palaeochannel Light brown-grey silty clay 8 0.6 fill 264033 Palaeochannel Irregular profile 1.5 0.15 264034 264033 Palaeochannel Light yellow-grey silty clay 1.5 0.15 fill 264035 Natural hollow Circular with moderately sloping 1.55 0.3 sides and uneven base 264036 264035 Burnt mound Mid grey-black silty clay with burnt 1.55 0.2 stones

13 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

264037 264035 Burnt mound Mid grey-black silty clay with burnt 1.55 0.15 stones 264038 264035 Burnt mound Mid grey-black silty clay with burnt 1.55 0.2 stones 264039 Trough Oval with steeply sloping sides and 1.7 0.85 0.4 concave base 264040 264039 Trough fill Mid to dark pink grey-brown 1.7 0.65 0.1 silty sand 264041 264039 Trough fill Black-grey clay silt with burnt stone 1.7 0.65 0.15 264042 264039 Trough fill Light to mid pink-yellow-grey silty 1.7 0.35 0.15 sand with burnt stone 264043 264039 Trough fill Pink-grey silty sand with burnt stone 1.7 0.4 0.25 264044 264039 Trough fill Yellow sandy silt with burnt stone 1.7 0.2 0.1 264045 Pit = 264003 264046 264045 Pit fill Mid grey-brown silty clay 3.1 0.7 >0.15 264047 Gully, part of E/W aligned, moderately-sloping 1.15 0.15 0.15 pit 264003 sides, concave base 264048 264047 Gully fill Mid grey-brown silty clay 1.15 0.15 0.15 264049 context not used 264050 context not used 264051 Pit/posthole Oval with moderately sloping sides 0.45 0.2 0.1 and flat base 264052 264051 Pit /posthole Mid grey-brown clay silt 0.45 0.2 0.1 fill 264053 Trough = 264039 1.7 0.9 0.4 264054 264053 Trough fill Mid to dark grey-brown silty sand 1.7 0.9 0.1 264055 264053 Trough fill Black-grey clay silt with burnt stones 1.7 0.9 0.15 264056 264053 Trough fill Light to mid yellow sandy silt 1.7 0.9 0.2 1450– 1300 cal BC 264057 Natural hollow Gently sloping sides; base not 0.6 >0.2 reached 264058 264057 Burnt mound Dark grey-brown silty sand 0.6 0.2 264059 264057 Burnt mound Dark grey-brown silty sand 0.6 0.2 1500– 1320 cal BC 264060 context not used 264061 context not used 264062 Natural hollow Circular with gently sloping sides 2.5 0.15 and irregular base 264063 264062 Burnt mound Mid grey-brown silty clay 2.5 0.15 264064 Natural hollow Moderately sloping sides and 6.1 0.8 concave base 264065 264064 Burnt mound Dark grey-brown silty sand 6.1 0.25 264066 264064 Burnt mound Dark grey-brown silty sand 6.1 0.8 264067 264064 Burnt mound Dark grey-brown silty sand 6.1 0.65 264068 264064 Burnt mound Dark grey-brown silty sand 6.1 0.2 264069 264064 Burnt mound Mid pink-brown silty clay 6.1 0.3 264070 Land drain 264071 264070 Land drain fill 264072 Natural Light yellow-brown sandy clay 264073 Natural hollow Gently sloping sides and concave 1.2 0.45 base 264074 264073 Burnt mound Mid grey-brown sandy silt 1.2 0.45 264075 Trough Sub-rectangular in plan, vertical 2.5 1.2 0.25 sides, flat base 264076 264075 Trough fill Mid grey-black sandy clay 2.5 1.2 0.4 264077 264075 Trough fill Brown-black sandy silt 2.5 1.2 0.35

14 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

APPENDIX B: THE FINDS

Prehistoric pottery (Gibson 2013) Weight (g) No of Contexts Periods Represented 115 2 Bronze Age

Bronze Age 264016: Rim sherd and body sherd (113g) in a hard, well-fired and heavy fabric. The sherd has not been cleaned but would appear to be black throughout, averaging 11mm thick and containing abundant stone inclusions up to 7mm across. Many inclusions break the outer surface whilst the inner surface has a thick carbon deposit. The rim is pointed with an internal bevel. There are occasional fingernail impressions across the rim. The rim diameter has been in the region of 300mm. Bronze Age Urn.

Discussion The lack of decoration on the outside of this vessel make it difficult to assign it to a particular tradition. The fingernail impressions on the rim bevel can be found on Food Vessels, such as the globular vessel from Welsh St Donats, Glamorgan (Savory 1980, Cat. No. 393.1), on Collared urns such as that from Fishguard, Dyfed (ibid Cat. No.410) and on Cordoned Urns such as that from Penmaenmawr, Gwynedd (ibid cat. No,403.1). The coarseness of the fabric, however, and the plainness of the vessel combined with the rounded profile of the rim may suggest that the sherd belongs to the bucket and barrel urn traditions of the Middle Bronze Age. Sparse fingernail impressions, though admittedly not on rim bevels, and a similarly coarse and fabric have been found on a Middle Bronze Age assemblage from Glanfeinion, Powys and radiocarbon dated to 1400-1170 cal BC (Britnell et al. 1997) whilst a similar assemblage from Bromfield, Shropshire, extends the date range into the beginning of the 1st millennium BC (Stanford 1982).

Miscellaneous Ceramics 264040: undiagnostic crumbs

Post-Roman pottery (Courtney and McSloy 2013) Two sherds of post-medieval pottery (76g) were recorded as unstratified finds. A bowl and a jug were identifiable in fabric NDGT (North Devon Gravel-Tempered Ware). This fabric was the most commonly recognised type from the pipeline excavations and is dateable to the 16th to 18th centuries.

Flint (Pannett 2009) Two struck flints were recovered, comprising a single piece of microdebitage and a scraper fragment. The former (recovered from trough fill 264056) is not closely dateable, the latter is probably Neolithic in date but was recovered as an unstratified find.

Fired/burnt clay (Hall 2008) Heavily-abraded featureless fragments of fired/burnt clay were recovered from 14 contexts. Due to the small size of the fragments no significant forms were identified.

15 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

APPENDIX C: PALAEOENVIRONMENTAL EVIDENCE BY JAMES RACKHAM

Human Bone A collection of cremated human bone was recovered from site RDX182, but no context is recorded. The assemblage comprises fifty three bone fragments (48g) including several long bone fragments, rib, vertebra and skull. It is possible that this group was recovered from site 182.4 not 182.2 and relates to the cremations found at Site 513. Owing to the unstratified nature of the finds and their insecure context they were not subjected to post- excavation analysis.

Environmental samples Two samples were taken from the evaluation (Table 1) and as part of the main investigation a series of eighteen environmental samples were taken from the troughs and burnt mound deposits (Table 1) and two series of column samples, a total of twenty one samples, were taken through the burnt mound (Table 3).

Table 1. Bulk environmental samples taken from Site 512 sample context feature description processed wt kg processed vol (l) 11501 11504 11503 Burnt mound deposit nd 6 11502 11506 11503 Burnt mound deposit nd 6 264000 264024 264023 Pit fill 5 4 264001 264040 264039 Trough fill* 20 18 264002 264041 264039 Trough fill* 10 8 264003 264043 264039 Trough fill* 12 8 264004 264016 264013 Burnt mound deposit 50 36 264005 264016 264013 Burnt mound deposit 54 39 264006 264005 264003 pit fill 31 27 264007 264004 264003 pit fill 10 8 264008 264059 264057 Burnt mound deposit 38 28 264009 264058 264057 Burnt mound deposit 10 6 264010 264056 264053 Trough fill* 23 18 264011 264043 264039 Trough fill* 33 20 264012 264055 264053 Trough fill* 17+ 12+ 264013 264054 264053 Trough fill* 20 18 264021 264072 264003 pit fill 10+ 8+ 264036 264074 264073 Burnt mound deposit 38 28 264037 264077 264075 Trough fill 28 12 264038 264076 264075 Trough fill 30 20 * 264039 and 264053 are the same trough; + extra sample processed but weight and volume not recorded; the evaluation trench samples are not located on plan.

The samples were processed in the manner described in the assessment reports (Giorgi and Martin 2009; Martin and Rackham 2009) with the additional refloating of the dried <2mm sample residues from the bulk samples (Table 1) that had been retained whose flot volume is indicated in Table 2 as ‘2nd flot’. This second flot was then sorted for charred macrofossils and the residue re-dried and checked with a magnet to recover any further magnetic material. The column samples were processed by the EAC and were all subjected to second floating when originally assessed (Martin and Rackham 2009).

16 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

Table 2. Data for the environmental bulk samples from Site 512

st 1 nd pro- 2 burnt sample context flot residue burnt cessed flot pottery stone coal magnetic comments no no vol wt g clay vol l vol g. ml 11501 11504 6 110 85 nd D 2270 B HNSx1; Triticum sp grain x1; Hordeum sp grain x1; indet grain x6 11502 11506 6 56 80 nd 11540 C Indet grain x3; # HNSx1 264000 264024 4 140 3 nd D 764 A 264001 264040 18 43 9 nd E A+0.1g Large stones- 2366g 264002 264041 8 64 15 nd E 1562 D 264003 264043 8 35 13 nd D D Stone>10mm 2294g 264004 264016 36 400 400 nd B 3712 A Large stone- from 10kg 12kg 264005 264016 39 498 800 nd D+12600 D A Stone-6056g 264006 264005 27 520 140 nd C 264007 264004 8 46 25 nd C B Stone-667g 264008 264059 28 114 1 nd 11100 264009 264058 6 90 2 nd D D D 264010 264056 18 54 3 nd D 1290 D Burn flint-E 264011 264043 20 69 5 nd C B Stone-1564g 264012 264055 12+ 236 13 nd C B Stone-1092+g 264013 264054 18 84 2 nd D B Stone-528g 264021 264072 8+ 7 0.5 nd 374+g D 264036 264074 28 86 41 nd D E C HNSx2 264037 264077 12 122 50 nd D B+0.1g Stone- 7586g;CBM-D; HNSx1 264038 264076 20 85 17 nd E 1972 B * abundance rating – E= 1-10 items; D=11-50, C=51=100, B=101-200, A=>200; + - present in flot; nd – no data; HNS hazel nutshell fragment; # this weight in the Cotswold processing record appears to be an anomaly since the sample was only 6 litres and 11.5kg of stone would be more than 6 litres.

17 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

18 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

19 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

The bulk samples from the troughs and pits produced a range of debris typical of burnt mounds, with burnt stone, fired clay and a generally high magnetic fraction. The processing records record pottery from context 264040 but they were undiagnostic crumbs. Burnt stone included mudstone and sandstone. The processing records for some of these samples are not complete. There is no consistent record of the presence and quantity of burnt stone, several samples having the stone residue weighed but no indication of whether it was burnt, and three samples, 05,12, 21, have only a partial record of the quantity of material processed or recovered so these samples cannot contribute to discussions of the density of material in the mound which therefore has been restricted to the column samples.

Column samples The two series of column samples were taken in the manner proposed for all burnt mounds. They were taken from a 50x50cm square in 5cm thick spits against two of the sections through the mound (Fig. 1) from the top to the base (Fig. 2). The objective was to establish whether it was possible to identify stratigraphy within the mound deposits that might reflect different episodes of use.

Table 3. The two column sample series through the burnt mound deposits at Site 512. Spit 1 was taken from the east-facing section of 264064 (Figs 1 and 2) and spit 2 from the north-facing section of 264013 (Fig. 1). sample no. context no. feature depth in column sample vol. sample wt. (L) (kg) Spit 1/column 1 264014 264066 264064 Burnt mound 0-5cm 10 14 264015 264066 264064 5-10cm 8 9 264016 264067 264064 10-15cm 10 12 264019 264067 264064 15-20cm 10 12 264020 264067 264064 20-25cm 9 12.5 264022 264067 264064 25-30cm 10 12 264024 264067 264064 30-35cm 10 12 264025 264067 264064 35-40cm 10 13 264028 264067 264064 40-45cm 10 13 264029 264067 264064 45-50cm 12 12 264032 264069 264064 50-55cm 7 11 264033 264069 264064 55-60cm 6 9 Spit 2/column 2 264017 264001 264013 Burnt mound 0-5cm 10 15 264018 264001 264013 5-10cm 10 14 264023 264016 264013 10-15cm 10 15 264026 264016 264013 15-20cm 16 23 264027 264016 264013 20-25cm 10 15 264030 264016 264013 25-30cm 10 14 264031 264016 264013 30-35cm * * 264034 264016 264013 35-40cm 10 15 264035 264016 264013 40-45cm 10 15 * - sample missing, not delivered to EAC

One sample from spit 2 (sample 264031) is lost and two flots are unaccounted for (the 2nd flot of sample 264018 and the 1st flot of sample 264027. A flot that was incompletely labelled (reading only ‘spit 2’) may be either the missing flot of sample 264018 or 264027 and has been recorded separately. The residues are characterised by varying proportions of sub-angular, sub-rounded and often fire-cracked, small-medium stones including sandstone, mud/siltstone and occasional white quartz, and also burnt gravel. Much of the mud/siltstone is fired and often has the appearance of degraded/worn fired clay. The archaeological finds recovered from the two column spits are restricted to fired-cracked stone, small quantities of magnetic material which consists entirely of magnetised stone and sediment crumb and also a trace of coal in the upper samples of column 2 (Table 4).

20 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

Table 4. Site 512. Archaeological finds from processed column samples 512014-264020 and 264022-264035, taken from Sections 512016 and 512041. sample context pro- resi- 1st 2nd FCP fired mag- coal Comment no. cessed due flot flot > earth/ netic vol (l) vol vol vol 7mm daub wt.g. ml ml ml wt. g wt.g Column 1 264014 264066 10 5000 65 66 P 3 P Charc charcoal mineralised. 264015 264066 8 3500 2 120 P 2 P Charcoal mineralised; 2x flint, <1g. 264016 264067 10 4500 150 150 P 1 charcoal mineralised. 264019 264067 10 4000 100 150 P 1 charcoal not so miner-alised in appearance. 264020 264067 9 5500 100 325 P 2 Charcoal not mineralised in appearance. 100% of 1st flot & 25% of 2nd flot assessed. 264022 264067 10 5000 500 45 6260 5 50% of 1st flot & 100% of 2nd flot assessed. 264024 264067 10 6000 300 200 P 2 50% of 1st & 2nd flot assessed. 264025 264067 10 6500 200 250 P 2 100% of 1st flot & 50% of 2nd flot assessed. 264028 264067 10 6000 200 200 8308 <1 264029 264067 12 4500 70 130 P <1 264032 264069 7 4000 5 60 P <1 1x flint, 1g. 264033 264069 6 3500 <1 5 3930 1 Column 2 264017 264001 10 6000 15 2 P <1 Charcoal mineralised. 264018 264001 10 7000 8 ? P 2 2nd flot missing? charcoal mineralised. 264023 264016 10 7000 16 10 P 2 charcoal mineralised. 264026 264016 16 8500 95 110 P 3 charcoal mineralised. 264027 264016 10 5500 ? 135 P 2.3 1st flot missing. 264030 264016 10 5000 100 200 P 2 charcoal mineralised. 264031 264016 ? ? - - - - Missing/not delivered? 264034 264016 10 4500 20 160 P 1 charcoal mineralised. 264035 264016 10 6000 7 63 P <1 charcoal mineralised. ? ? ? ? 200 charcoal mineralised. One unlabelled flot from spit 2; likely to be either the second flot of sample 264018 or the first flot of 264027. £/#- sherd no/weight in g; P – present

21 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

Table 5. Site 512. Environmental finds from processed column samples 512014-264020 and 264022-264035, taken from Sections 512016 and 512041. sample context pro- resi- 1st 2nd char- charred charred comment no. cessed due flot flot coal grain* seed* vol (l) vol vol vol */* (ml) (ml) (ml) Column 1 264014 264066 10 5000 65 66 5/5 1 Indet. grain (5) 264015 264066 8 3500 2 120 5/5 cf. Triticum (1); indet. grain (2) 264016 264067 10 4500 150 150 5/5 1 cf. Hordeum (2); indet. grain (1) 264019 264067 10 4000 100 150 5/5 1 Hordeum vulgare (indet) (1), indet. grain (3) 264020 264067 9 5500 100 325 5/5 No charred plant remains 264022 264067 10 5000 500 45 5/5 est. 2 1 Hordeum vulgare (5) (1 hulled, 1 naked, 3 indet); cf. Hordeum (3), indet. grain (4); Corylus avellana shell fragments (2); very occ. charred culm node & small round wood/twig. 264024 264067 10 6000 300 200 5/5 est. 2 est. 1 Hordeum vulgare (indet.) (3), cf. Hordeum (2); indet. grain (3); Corylus avellana.shell fragments (3) 264025 264067 10 6500 200 250 5/5 1 Hordeum vulgare (4) (3 hulled, 1 indet); indet. grain (4); Corylus avellana shell fragment (1); very occ. small round wood/herbaceous stem fragments 264028 264067 10 6000 200 200 5/5 1 Hordeum vulgare (5) (3 hulled, 2 indet); Triticum (1); indet grain (2); very occ. culm nodes 264029 264067 12 4500 70 130 5/5 1 1 Hordeum vulgare (indet.) (1); Triticum dicoccum glume base (1); indet. grain (4), Galium aparine (1); cf. G. aparine (1); indet. seed frag (1). 264032 264069 7 4000 5 60 5/5 No charred plant remains 264033 264069 6 3500 <1 5 2/3 No charred plant remains Column 2

22 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

264017 264001 10 6000 15 2 3/5 No charred plant remains 264018 264001 10 7000 8 ? 2/3 No charred plant remains 264023 264016 10 7000 16 10 5/5 1 No charred plant remains 264026 264016 16 8500 95 110 5/5 1 cf. Triticum (1), indet. grain (4) 264027 264016 10 5500 ? 135 5/5 No charred plant remains 264030 264016 10 5000 100 200 5/5 1 1 Indet. grain (2); Corylus avellana shell fragment (1) 264031 264016 ? ? - - - Missing/not delivered? 264034 264016 10 4500 20 160 5/5 1 Hordeum vulgare (1) (hulled/twisted); cf. Hordeum (1); Corylus avellana shell fragment (1) 264035 264016 10 6000 7 63 5/5 1 Corylus avellana shell fragments (5) (<0.1g) ? ? ? ? 200 5/5 1 Hordeum (2), cf. Hordeum (1), indet. grain (6) * frequency – 1=1-10; 2=11-50; 3 =51-150; 4 =151-250; 5 =>250 items; */* frequency of charcoal >2mm/<2mm; £/#- sherd no/weight in g; est. – estimated abundance score; P – present.

The flots range in volume between 17 and 300ml in column 1 and 8 and 545ml in column 2. The flots primarily consist of comminuted charcoal, which is frequently mineralised, with only very occasional fragments of charred herbaceous stems in three samples, some of which may be from cereals/wild grasses, and small round wood. Small quantities of charred cereal grain were present in twelve samples although preservation was poor and just over half the grains could not be identified. Hordeum vulgare (barley) was the best represented cereal, with occasional and small numbers of grains in eight samples including evidence for hulled barley in four samples and a twisted hulled grain in sample 264034 showing the presence of six-row hulled barley. There was also a naked barley grain in sample 264022. Single grains of Triticum (wheat), including tentative identifications, were recovered from three samples while another sample 264029 contained a glume base belonging to the hulled wheat Triticum dicoccum (emmer). The same sample also produced evidence for Galium aparine (cleaver) a potential arable weed of loam and sometimes clay soils. Several fragments of Corylus avellana (hazel) nutshell were identified in six column samples

Discussion Despite the relatively close proximity of the two column spits, there are tangible differences in the sequences of the samples. Pit 264003 contained a large concentration of fairly clean cereal grain. The column of samples nearest the grain-rich feature (spit 1/column 1) yielded the most amount of cereal grain, primarily from deposit 264067, whilst the column of samples to the east of pit 264003 produced fewer numbers of grain, from deposit 264016. But, the cereals in the mound are different from those in the pit. Furthermore, the relatively grain-rich spit/column 1 appears to contain evidence for more than one depositional event (Fig. 3), whilst spit 2 shows no such pattern (Fig. 4) although it could be interpreted as two periods of deposition (see below). In the first series of samples (column 1) the basal three samples are characterised by a high fine sediment proportion, but with 40% of the samples composed of heated mudstone and sandstone greater than 6.7mm, of which stones over 5cm comprise about 30%. This might represent a primary dump of

23 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

fired stone which may have undergone some weathering and sediment infilling the voids in the dump. The flot volume increases upwards from a charred flot size of only 8ml per litre to 222ml in sample 264029, and charred grains, including one of barley, an emmer glume base and several weed seeds, appear in this sample. In the next three samples, 28, 25 and 24 the larger fired stone component increases at the expense of fine silts, and the flot volume rises to over 400ml per litre. Barley is consistently present along with a single wheat grain. This deposit (the top of 264069 and the base of 264067) would suggest a new dump of fired stone on a previous heap, indicating a return to the same site. This episode of dumping probably continued to sample 264020, and the slight drop in coarse stone (>6.7mm) and the rise in the finer sediment fractions suggests some weathering of the mound and the incorporation of inwash and blown silts. A relative drop in flot size and an increase in the silt component of the deposits above 264020 suggests a period of weathering and slower build up of material. The top three samples may be the weathered and silted remains of a third episode of dumping or the weathered down top of the second episode. On site the lowest of these three was assigned to context 264067, which might imply their association with the deposits below (ie the second dump episode), but the change in archaeological layer recording might merely reflect the weathering of the upper part of this layer. A marked drop in the flot volume in these upper samples might reflect charcoal being washed out of the dump during weathering. No organic remains have survived in the deposits. In the second spit a slightly different picture is suggested. From the base of the sequence the contribution to the residue by stones over 5cm tends to drop, but is never very high. The basal two samples, 264034 and 264035, show a high silt content but almost all the stone is >6.7mm with very little weathered burnt stone. Perhaps the initial dump of burnt mound material was flooded and infilled with fine sediments from the flood waters. The upper samples in the series show a reduced and slowly reducing silt component, a suggestion of weathering of the exposed fired stones leading to an gradual increase in the <2mm and 2-6.7mm residue fractions, but a fairly consistent and large 6.7-50mm fraction, and a slight reduction of the >50mm stones. This picture could possibly be explained by the weathering of a single dump of fired stones on top of the earlier dump. A little barley, possibly wheat and hazelnut shell occurred sporadically through the sediments, while the largest flots are associated with two of the bottom three samples in the suggested second dumping episode.

24 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

Fig. 3. Bar diagram of the residue components of column samples from Spit 1 (column 1), from Section 264013 through the burnt mound and overlying deposits, with a brief summary of the finds by sample interval.

% unwashed sample weight Context Mag. Flot vol/ 264 Wt/10l 10l 0% 20% 40% 60% 80% 100% 066 coal 3 131 Indet. cereal 14 066 Coal, flint x2 2.1 152 Wheat 15 067 2 300 Barley 16 067 1.5 250 Barley 19 067 2.2 472 20 067 5 545 Barley, hazelnut 22

24 067 2 500 Barley, hazelnut

sample no. 25 067 2 450 Barley, hazelnut

28 067 <1 400 Barley, wheat

29 067 <1 222 Barley, emmer wheat

32 069 Flint x1 <1 93

33 069 Fired earth- - 8 1g res.<2mm res. 2-7mm res. >7mm res. >5cm <1mm sediment

25 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

Fig. 4. Bar diagram of the residue components of column samples from Spit 2 (column 2) from Section 264064 through the burnt mound and overlying deposits, with a brief summary of the finds by sample interval.

%unwashed sample weight Context Finds Mag. Flot vol/ 264 Wt/10l 10l 0% 20% 40% 60% 80% 100% 001 <1 17

17

001 2 ? 18

016 2 26 23

016 3 205 Wheat 26

016 2.3 200? ?unlabelled flot 27

sample no 016 2 300 Indet cereal, hazelnut 30

31 016 Sample never delivered

34 016 1 180 Barley, hazelnut

35 016 <1 70 Hazelnut

res. <2mm res. 2-7mm res. >7mm res.>5cm <1mm sediment

26 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

Conclusion The two sequences of column samples from the burnt mound deposits were taken in the hope of answering specific questions relating to the nature and intensity of activity at the site, the longevity of the site and whether evidence for seasonality could be detected. Determining whether the activities were seasonal, and if so the time of year they potentially took place is not really answerable on the basis of the material recovered from the column samples. Detecting whether the sequences were formed from single or multiple events is possible and spit 1 seems to contain evidence for two or three events, whilst spit 2 appears to be derived from two events. The longevity of the activity at the site is more difficult to establish; radiocarbon dating remains such as nutshell, grain or roundwood from different points in the two sections is a possibility, but the margin of error may be too close to differentiate the episodes of activity satisfactorily. The silt element in the lower samples could have been introduced during just one season of flooding. Both silt/mudstone and sandstone appear to have been selected for heating in the fires. It might be useful to consider the quantities of mudstone and sandstone in the column samples from all the burnt mound sites and compare this with the immediate geology of each site to establish whether there is any selection or not, but unfortunately the burnt stone has not been retained from all the samples.

Charred plant remains (John Giorgi) Eight of the 20 bulk soil samples produced charred plant remains, five from burnt mound deposits 11504, 11506, 264016, 264074, and two from fills of trough 264075 and pit 264003. Most of the samples, however, only contained traces or small amounts of identifiable remains (Table 2) although the two fills of pit 264003 produced very large quantities of cleaned but very poorly preserved and fragmentary cereal grain (Table 6). The charred botanical remains in the five samples from the burnt mound were broadly similar to the results from the column samples, consisting largely of charcoal together with small numbers of poorly preserved cereal grains in four of the samples including evidence for barley (including one hulled grain) in three, and wheat in two sampled fills. There were also traces of hazelnut shell in three deposits and a seed of Raphanus raphanistrum (wild radish) in sample 264004, a weed usually found in acid sandy soils. The two samples from pit 264003, on the other hand, contained very rich charred cereal grain assemblages together with a large amount of charcoal, the two sampled fills 264005 and 264004 producing densities of 114 and 82 grains per litre of processed soil respectively. As noted above, however, preservation was very poor and 88% of the counted grains could not be identified while cereal fragments smaller than 1mm were not quantified. The relatively small number of identifiable grains showed barley, including evidence for hulled barley, to be the best represented cereal, followed by smaller but almost equal numbers of wheat and oat grains. The well- preserved wheat grains were identified as Triticum aestivum/turgidum (free-threshing wheat) although it was not possible to establish whether these were from hexaploid and/or tetraploid species. Similarly the poor condition of the oat grains and absence of diagnostic floret bases made it impossible to establish whether the oats represent cultivated and/or wild species; the size (>6mm in length) and morphology (widest towards the embryo end) of a few of the oat grains, however, may tentatively point to the presence of Avena sativa (common oat). The few other identifiable charred plant remains in the feature included a small amount of large but poorly preserved legumes in fill 264005, classified as Vicia/Lathyrus/Pisum (vetch/tare/vetchling/pea). These seeds ranged in size from 4mm to 5mm displaying morphologies similar to Vicia faba (horse bean) but with no diagnostic surface features allowing a definite or even tentative identification as such. Both fills also produced very small numbers of weed seeds including Raphanus raphanistrum and Fallopia convolvulus (black bindweed), the latter often found in sandy soils and usually associated with spring-sown cereals. A hazelnut shell fragment was recovered from sampled fill 264005 and also in the fill of trough 264075.

27 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

Table 6. Site 512. The charred plant remains from the bulk samples Feature type BURNT MOUND PIT Feature no. 264015 264003 Context no. 264016 264005 264004 Sample no. 264004 264005 264006 264007 Sample volume (l) 32 39 27 8 Flot 1st 2nd 1st 2nd 1st 2nd 1st 2nd Vol. flot (ml) 400 400 498 800 520 140 46 25 Cereal grains Triticum. aestivum/turgidum free-threshing wheat 4 4 T. cf. aestivum/turgidum ?free-threshing wheat 11 7 Triticum sp(p) wheat 1 1 13 2 cf. Triticum sp(p) ?wheat 2 1 10 14 Hordeum vulgare L. barley hulled indet. 1 20 12 H. vulgare L. barley indet 5 125 3 69 cf. H. vulgare ?barley 1 1 19 26 Avena spp. oat 2 1 8 cf. Avena spp. ?oat 14 7 40 2 Avena/Poaceae oat/large wild grasses 12 10 Cerealia indet. Indet. grains (estimate) 8 4 4 7 2679 135 377 77 Indet. grain frags Cerealia indet. +++ (<1mm) Other plants hazel nut shell Corylus avellana L. 1 fragments Atriplex/Chenopodium sp. orache/goosefoots etc. 1 Fallopia convuluvulus (L.) A 3 1 1 Love black bindweed Polygonum sp. knotgrass 1 vetch/tare/vetchling/pea 13 Vicia/Lathyrus/Pisum spp. (large legumes >4mm) Raphanus raphanistrum L. wld radish 1 2 6 Charcoal +++++ +++++ +++++ +++++ Total nos. of items 11 4 13 9 2927 148 576 81 Item density (per litre of processed soil) 0.5 0.6 113.9 82.1

The two charred plant assemblages from pit 264003 consist of fully processed and cleaned cereal grain (free- threshing wheat, hulled barley and oats), which may have been accidentally burnt while being dried before milling and/or storage or judging by the very poor preservation and fragmentation, as a result of a small conflagration, the remains subsequently being deposited into the pit. The few weed seeds may tentatively suggest the use of sandy soils for cultivation and possibly the spring sowing of cereals. There is also the possibility that pulses may have been grown from time to time. The cereal assemblage of free-threshing wheat and oats (and the complete absence of any evidence for hulled wheats) in the pit is consistent with a medieval date.

Charcoal (Dana Challinor) Two samples from burnt mound deposit 264013 and one each from the fills of pit 264003 and trough 264053 were studied, following standard procedures. The condition of the charcoal, in common with other, assessed flots, was very poor; with high levels of sediment infusion and orange staining. The charcoal was also extremely soft and crumbled easily, inhibiting fracturing. For this reason, most of the charcoal examined was selected from the larger fractions (>8mm and >4mm) as it was considered that the 2mm sieve would yield largely unidentifiable material.

28 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

Six taxa were positively identified (Table 7); Quercus sp. (oak), Alnus glutinosa (alder), Corylus avellana (hazel), Populus/Salix (poplar/willow), Maloideae (hawthorn group) and Cytisus/Ulex (broom/gorse). Low to medium levels of vitrification were observed in the oak of <264005> and the broom/gorse in <264006>, which also displayed frequent radial cracks (a common feature in charred specimens of this taxon). Roundwood characteristics were frequent, especially notable in sample <264006>, which comprised many small twigs, some with pith and bark edge. The poor state of charcoal in this sample (strong orange staining, vitrified and very crumbly) inhibited further analysis. The poorest sample in terms of preservation was <264053> which contained mostly small fragments and a relatively high level of indeterminate fragments. It is likely that more taxa were present in this particular sample than could be positively confirmed.

Table 7. Charcoal from burnt mound and trough fills at Site 512

LBA LBA EMed. Feature type burnt mound trough pit Feature number 264013 264053 264003 Context number 264016 264055 264005 Sample number 264004 264012 264005 264006 Quercus sp. oak 6 (r) 4 14 (h) 1r Alnus glutinosa alder 3 (r) 9 5 Gaertn. Corylus avellana L. hazel 4r 1 1 Alnus/Corylus alder/hazel 4 (r) 7 2 1r Populus/Salix poplar/willow 2 3 Maloideae hawthorn group 7 (r) 2 (r) cf. Maloideae hawthorn group 1 Cytisus/Ulex broom/gorse 26r Indeterminate diffuse porous 4 1 2r Indeterminate 4 (b) 4 2(b) Total 30 30 30 30 h=heartwood; r=roundwood; brackets denotes presence in some frags only

The picture of the fuel use at Site 512 is, to some degree, compromised by the condition of the charcoal. The two samples from context 264016 of the burnt mound produced comparable assemblages derived from the same taxa, but with some differing quantities, which is difficult to evaluate, given the limitations of the dataset. Sample <264005> contained considerably more charcoal, with larger fragments and more oak, than <264004>. Perhaps this relates to discrete episodes of burning (sample 264004 was immediately above sample 2643005), in which case it is clear that the surrounding woody resources of oak, alder, hazel, hawthorn group and willow and/or poplar were generally exploited. Alder and willow/poplar prefer wet ground habitats and would have thrived by the adjacent stream/channel. The assemblage from trough 264053 is similar to those from the burnt mound samples, and the fuel would have derived from the same wood supply. However, the assemblage of predominantly broom/gorse twigs from pit 264003, one of the grain rich deposits, is notably different to all the other samples and indicates a heath type habitat. Site 512 is located not far from the Pembrokeshire coast, where heathland grows today, and it is likely to have been available as a resource in the early medieval period. Gorse (and, to an extent, broom) burns with a high but fast heat, making it suitable for kindling or specific activities, such as warming bread ovens (Gale and Cutler 2000, 260). It would not, however, provide the longer-lasting heat from oak, hazel or even alder which seem generally to have been preferred fuel for burnt mound activities, and in this context is almost certainly functionally related to the cereal assemblage in the deposit.

29 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

Discussion This burnt mound site is unusual, and different from the others along the pipeline route in producing deposits that are rich in charred cereals, and one charcoal assemblage in which broom or gorse dominates. The radiocarbon dates obtained from charcoal from the cereal rich feature, 264003, however resolved this peculiarity in identifying the ‘trough’ as an early medieval pit cut into the burnt mound. The cereal assemblage in pit 264003 (free threshing wheat, oats and barley) fits much better in the post-Roman period than the Bronze Age, and with the cereals and gorse/broom charcoal present we can confidently assign the fills to the period reflected by the radiocarbon dates, the 7-9th centuries AD. This raises a question concerning the date of the other troughs, 264039 and 264075, but one of these is radiocarbon dated to the Bronze Age and the other contains a typical burnt mound assemblage of burnt stone and charcoal with no evidence for the cereals present in 264003. The cereal assemblages from 264003 are clean and lack chaff and weed seeds indicating they are ready for storage, milling or consumption. Gorse/broom is used as a fast burning fuel in ovens and it is possible that pit 264003 was receiving material accidentally burnt during drying, perhaps for grinding. The third sample from this feature, <264021>, was taken from the northern end from beneath the pit cut and produced a little burnt clay and no other finds. The cereal assemblage from pit 264003 indicates a cleaned cereal crop, while the abundance of gorse/broom charcoal might reflect ‘oven’ fuel if the charcoal is gorse (Lucas 1960). The implication would seem to be that this debris derives from local early medieval occupation in the immediate vicinity. The apparently secure burnt mound deposits also include a little charred cereal grain and chaff and because this material includes evidence for glume wheat and barley but no oats, a typical Bronze Age assemblage, these finds are taken as contemporary rather than possible contaminants from the later activity on the site. It is also evident from the distribution of the charred cereal remains in the burnt mound deposits, with cereal grains occurring throughout the column of samples and present in some of the lower spits that these are contemporary with the mound. This makes this mound unusual, with more cereals present in the two columns of samples than almost all of the other burnt mounds on the project put together. This need not indicate a different activity at this mound, but it might suggest proximity to nearby settlement or another type of activity area which the other very limited finds - 2 pottery sherds and two flint flakes – might also lend support to. The burnt mound deposits extend to 129.6 square metres and with an average thickness across section DD (Fig. 1) of 0.31m the total volume of the mound is approximately 40 cubic metres. With an average density of stone in the samples in column 1 of 730g per litre a ballpark figure for the total weight of stone in the mound is 29.2 tonnes. A little of the mound has been removed by the later ‘palaeochannel’ so a figure of 30 tonnes is a guesstimate for the whole mound. This represents a large burnt mound in terms of the sites excavated along the pipeline. The site lies approximately 50m from the stream on the valley floor. The description of the linear features on the site as ‘palaeochannels’ is a misnomer. The site lies on sloping ground, a fall of about 1 in 16 to the south west, and these features could only be palaeochannels if a spring rises on the hillside above and flows down to the brook in the valley. There is no evidence for this in the topography, and no spring is indicated on the 1st edition OS map and a ‘channel’ 8m wide seems too broad for a spring fed channel. It seems more probable that these features represent scour and run-off channels during periods of high rainfall. The fact that one underlies the burnt mound might even suggest that the landscape had been cleared prior to the build up of the mound. Such features are more pronounced in arable than pastoral landscapes. It is difficult to conceive of these palaeochannels as being the water source for the site in the Bronze Age which might imply that water had to be transported to the site from the stream in the floor of the valley. Stone is also problematic given the quantity in the mound. A tree throw pit might have been responsible for the original selection of this location but most of the 30 tonnes, and this just represents the burnt stone in the surviving mound, must have had to be carried to the site. The bed and banks of the stream afford another potential source just 50m away but one might have expected the

30 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

mound to move closer to the stone and water resources once any local stone source was used up. The fuel resources are likely to have been nearby, and certainly alder and poplar/willow trees would have lined the stream and oak and hazel woodland if not actually on the site were probably nearby with trees of the hawthorn group fringing the woodland or in hedgerows. With alder, poplar/willow and Maloideae a significant element of the fuel, and the scour or run-off channels beneath, this burnt mound may have been in a partially cleared landscape on the hillside. The charcoal in pit 264003 is quite clearly different from the other assemblages (Table 7) with a dominance of gorse or broom suggesting rough grass or heathland edge habitats in the early medieval period, but if this assemblage is functional, as it appears to be, it may have been brought from further afield. This site is unusual in that there does not appear to be a water source or ready supply of stone within a few metres, but these resources may have had to be carried some 50 metres or more. There is some evidence for occupation activity, as reflected in the charred cereals, pottery and flint flakes. It is a relatively large burnt mound with perhaps as much as 30 tonnes of burnt stone in it. This implies some longevity and suggests perhaps some reason, other than resource availability, for the continued use of the mound even after the resources had to be brought from the floor of the valley. Fifty metres may not seem any great distance, but most burnt mound sites on the pipeline route seem to lie within as little as ten metres of the water source and probably not much more from stone and fuel.

REFERENCES Carruthers, W. 2008, Felindre to Brecon Natural Gas Pipeline. Assessment report from charred plant remains. For Cambrian Archaeological Projects Greig, J, 1991, The British Isles, Progress in Old World Palaeoethnobotany (eds W van Zeist, K Wasylikowa and K-E. Behre), Rotterdam, 229-334 Lucas, A.T. 1960 Furze. A survey and history of its use in Ireland. The Stationery Office for the National Museum of Ireland. Martin, G and Giorgi J, 2008 Milford Haven to Aberdulais High Pressure Gas Pipeline. Assessment Report for the Archaeobotanical Remains from Llwyn, Llangain (Site 501). Unpublished Report for NACAP Land and Marine for National Grid Robinson, M, 2000, Further considerations of Neolithic charred cereals, fruit and nuts, Plants in Neolithic Britain and beyond (ed A S Farbairn), 85-90

31 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

APPENDIX D: THE RADIOCARBON DATING BY SEREN GRIFFITHS

For the analysis, radiocarbon measurements were produced on short-life, single entity charred plant remains. Samples with the ‘SUERC-‘ laboratory code were pretreated using an acid-base-acid process. Samples were combusted and graphitized and then dated by Accelerator Mass Spectrometry (AMS). The results are conventional radiocarbon ages, quoted according to the international standard set at the Trondheim Convention. The results have been calibrated using IntCal13, and OxCal v4.2. The date ranges have been calculated using the maximum intercept method, and have the endpoints rounded outward to 10 years.

Seven radiocarbon dates were produced on shortlife samples from deposits associated with the burnt mound and a trough (Fig. 8). The three radiocarbon dates from the mound are statistically consistent (SUERC-55488, - 55494, -55495; T’=0.2; T’5%=6.0; df=2; Ward and Wilson 1978), and could be of the same actual age. A weighted mean taken prior to calibration suggests that if the activity represented by these results represented a single ‘archaeological event’, these burnt mounds deposits could have been associated with activity in the 15th century cal BC. The result from the fill of trough 264053 (SUERC-55487) reflects the evidence from the burnt mound deposits and suggests activity in the 15th or 14th centuries cal BC.

A single result (SUERC-55486) from pit 264003 is much older than the other scientific dates from the site, and indicates activity in 1740–1530 cal BC (95% probable), most probably in the 17th century cal BC in 1690–1610 cal BC (68% probable; SUERC-55486; Fig. 1). The two other results from this context (SUERC-55492 and - 55493) are statistically consistent (T’=0.2; T’5%=3.8; df=1; Ward and Wilson 1978), and could be of the same actual age. These results are produced on different species of charcoal, providing independent estimates for the burning activity associated with the formation of fills within pit 264003. If these results represented a short-lived ‘archaeological event’, a weighted mean taken prior to calibration indicates that this could have occurred in the later 7th or 8th centuries cal AD. Interpretation of the phasing of the features at Site 512 suggested that pit 264003 might represent the latest activity at the site. The radiocarbon chronology from this feature might suggest that SUERC-55486 was produced on a piece of charcoal residual within this context, while the two other measurements from this feature represent much later migration period activity.

The periodicity of Bronze Age activity at the site could include multiphase activity associated with the use of burnt mounds (as represented by SUERC-55486), which is poorly represented within the radiocarbon sample, or that the radiocarbon dates from the site reflect activity of much shorter duration focusing on the 15th century cal BC with much earlier and later ephemeral and poorly understood human presence in the area. If the statistically consistent results from the burnt mound fills represent the main focus of activity at the site this might be estimated to have begun in 1510–1420 cal BC (95% probable; or 1500–1440 cal BC 68% probable; FirstFill; Fig. 8), and ended in 1470–1380 cal BC (96% probable, or 1350–1310 cal BC 9% probable; or 1450–1400 cal BC 68% probable; LastFill; Fig. 8).

32 © Cotswold Archaeology South Wales Gas Pipeline Project Site 512: Archaeological Excavation

Context Feature Sampled Laboratory ref Measured δ13C Calibrated date (95%) material age 264005 Pit 264003 Corylus sp. SUERC-55486 3359 +/-30 -27.6 1740–1560 cal BC charcoal (GU35181) 264005 Pit 264003 Cytisus/Ulex SUERC-55492 1268 +/-30 -25 Cal AD 660–780 charcoal (GU35184) 264005 Pit 264003 Alnus/Corylus SUERC-55493 1247 +/-30 -25.3 Cal AD 670–880 charcoal (GU35185) 264016 Burnt mound Corylus sp. SUERC-55494 3167 +/-30 -25.2 1510–1390 cal BC 264030 charcoal (GU35186) 264016 Burnt mound Alnus sp. SUERC-55495 3175 +/-30 -26.8 1510–1400 cal BC 264030 charcoal (GU35187) 264056 Trough Alnus/Corylus SUERC-55487 3126 +/-30 -26.4 1450–1300 cal BC 264053 charcoal (GU35182) 264059 Burnt mound Alnus/Corylus SUERC-55488 3155 +/-30 -26.3 1500–1320 cal BC 264030 charcoal (GU35183) Dating undertaken by Beta Analytic, Miami and Scottish Universities Environmental Research Centre

Fig. 8 A Bayesian model for the results from site 512. Each distribution represents the relative probability that an event occurred at a particular time. For each distribution two ranges have been plotted. The range in outline represents the calibrated radiocarbon result, the solid distribution represents the posterior density estimate- or Bayesian statistical model output. The large square brackets and CQL2 OxCal keywords define the model exactly.

33 A NNyferyfer

AAC Cothi

PEMBROKESHIREPEMBROKP MBROROKESESHIRSSHIREH RE n Lly AfonA LlynfiLly dau Brecon w Tywi

A CyninCyni on frfron Ty stern Cleddau AfronAAf Tywi ddau WessternWes Eastern Cleddau Cleddau CywynCy A Cywyn Cyw FawrFawFaawr endraet Milford Gwendraeth Fach draeth GwendraethGwendr Fach Haven g ugu

Site 512 R Loughor

Aberdulais Rhy

AAR Rhondda Fach ondda F Felindre A Rhondda Fawr

awr 0 25km SWANSEASSWSWANSWWAANANSEEAA

Site 512

Site 513 Fig 2 Inset

Site 505

Site 286

N Cirencester 01285 771022 Cotswold Milton Keynes 01908 218320 Milford Haven to Aberdulais pipeline Andover 01264 326549 Archaeology w www.cotswoldarchaeology.co.uk Felindre to Brecon pipeline e [email protected] PROJECT TITLE 0-75m contour South Wales Pipeline. Site 512, Land North of 75m contour Steynton, Milford Haven, Pembrokeshire 200m contour FIGURE TITLE Site location plan 400m contour 600m contour 0 1km FIGURE NO. Reproduced from the 2005 Ordnance Survey Explorer map with PROJECT NO. 9150 DATE 19-02-2014 the permission of Ordnance Survey on behalf of The Controller DRAWN BY DJB REVISION 00 of Her Majesty's Stationery Office c Crown copyright Cotswold Archaeology Ltd 100002109 APPROVED BY LM SCALE@A4 1:25,000 1 N

SM 91470 08338

264033

pit land E 264023 drain

E posthole 264031 264051 F Evaluation Trench 116

F

trough 264075

A

264016 N A

D Evaluation D Trench 115 gully Site 512 264047 C

pit 830 264003 1740-1560 cal. BC 660-880 cal. AD burnt mount 264030 B 264017 1510-1390 cal. BC 1510-1400 cal. BC 1500-1320 cal. BC B SM 915 0 1:5000 100m trough 264039 1450-1300 cal. BC pipeline centreline burnt mound

C early medieval palaeochannel 2005-6 evaluation 264033 trench

Reproduced from the digital Ordnance Survey Explorer map with the permission of Ordnance Survey on behalf of The Controller of Her Majesty's Stationery Office c Crown copyright Cotswold Archaeology Ltd 100002109

Cirencester 01285 771022 Milton Keynes 01908 218320 Cotswold Andover 01264 326549 Archaeology w www.cotswoldarchaeology.co.uk e [email protected] 0 500m PROJECT TITLE South Wales Pipeline. Site 512, Land North of Steynton, Milford Haven, Pembrokeshire

FIGURE TITLE Plan of archaeological features

PROJECT NO. 9150 DATE 19-02-2014 FIGURE NO. DRAWN BY DJB REVISION 00 APPROVED BY LM SCALE@A3 1:250 2 Section AA Section BB

S NW E SW NE 40.5m 39.75m AOD AOD

264042 264044 264043 264041 trough 264074 264040 264039

264077

trough 264076 264075

Section CC

S N 40.5m AOD

264066

264065 264067

264069 264068 natural 264069 hollow N natural hollow

Site 512

830

Section DD Section EE Section FF

SM 915 0 1:5000 100m NE SW SN NE SW 264052 40.25m 40.25m 264024 40.25m AOD AOD AOD pipeline centreline 264023 264051 264006 Reproduced from the digital Ordnance Survey Explorer map with the permission of Ordnance Survey on behalf of The Controller of Her Majesty's Stationery Office c Crown copyright Cotswold Archaeology Ltd 100002109

264005 Cirencester 01285 771022 pit Milton Keynes 01908 218320 264003 Cotswold Andover 01264 326549 264004 Archaeology 01m w www.cotswoldarchaeology.co.uk e [email protected]

PROJECT TITLE South Wales Pipeline. Site 512, Land North of Steynton, Milford Haven, Pembrokeshire

FIGURE TITLE Sections

PROJECT NO. 9150 DATE 19-02-2014 FIGURE NO. DRAWN BY DJB REVISION 00 APPROVED BY LM SCALE@A3 1:20 3 4

Trough 264075, looking east (scale 1m)

5

Pit 264003, looking north (scale 2m)

N Cirencester 01285 771022 Cotswold Milton Keynes 01908 218320 Archaeology w www.cotswoldarchaeology.co.uk e [email protected]

pipeline centreline PROJECT TITLE Fig. 4 South Wales Pipeline. Site 512, Land North of Steynton, Milford Haven, Pembrokeshire 830 Fig. 5 photograph locator Fig. 5 FIGURE TITLE Site 512 Photographs

FIGURE NO. Reproduced from the digital Ordnance Survey Explorer map with PROJECT NO. 9150 DATE 26-02-2014 the permission of Ordnance Survey on behalf of The Controller DRAWN BY DJB REVISION 00 of Her Majesty's Stationery Office c Crown copyright 01:2500 50m SM 915 Cotswold Archaeology Ltd 100002109 APPROVED BY LM SCALE@A4 N/A 4 & 5 6

Trough 264039, looking west (scale 0.3m)

7

Burnt mound 264030, looking north-west

N Cirencester 01285 771022 Cotswold Milton Keynes 01908 218320 Archaeology w www.cotswoldarchaeology.co.uk e [email protected]

pipeline centreline PROJECT TITLE South Wales Pipeline. Site 512, Land North of Fig. 6 Steynton, Milton HavenPembrokeshire 830 Fig. 5 photograph locator Fig. 7 FIGURE TITLE Site 512 Photographs

FIGURE NO. Reproduced from the digital Ordnance Survey Explorer map with PROJECT NO. 9150 DATE 26-02-2014 the permission of Ordnance Survey on behalf of The Controller DRAWN BY DJB REVISION 00 of Her Majesty's Stationery Office c Crown copyright 01:2500 50m SM 9155 Cotswold Archaeology Ltd 100002109 APPROVED BY LM SCALE@A4 N/A 6 & 7