: ae The Wiltshire Archaeological and Natural History Magazine Volume 103 2010 Published by The Wiltshire Archaeological and Natural History Society 41 Long Street, Devizes, Wilts. SN10 INS Telephone 01380 727369 Fax 01380 722150 Email: wanhs@wiltshireheritage.org.uk Website: http://www.wiltshireheritage.org.uk/ Founded 1853 Company No. 3885649 Registered with Charity Commission No. 1080096 VAT No. 140 2791 91 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE VOLUME 103 (2010) ISSN 0262 6608 © Wiltshire Archaeological and Natural History Society and authors 2009 Hon. Editor: Andrew Reynolds, BA, PhD, FSA, FSA Scot, FRHistS. Hon. Assistant Editor: Stuart Brookes, BA (Hons), PGDTiLL, MA, PhD. Hon. Local History Editor: James Thomas, BA, PhD, FRHistsS. Hon. Natural History Editor: Michael Darby, PhD, FRES Hon. Reviews Editor: Robert Clarke BA (Hons), Cert Ed, MIfL. Hon. Excavation and Fieldwork in Wiltshire Editor: Simon Draper, BA (Hons), MA, PhD (Dunelm.). Finds Liaison Officer: Katie Hinds, BA (Hons), MA. Editorial Assistant: Lorna Haycock, BA (Hons), PhD, FRHistS, Dip.ELH, Cert Ed. We acknowledge with thanks grants towards the cost of publishing specific papers in this volume from the following bodies: Programme Jade for ‘The Breamore jadeitite axehead and other Neolithic axeheads of Alpine rock from central southern England’ by Alison Sheridan, David Field, Yvan Pailler, Pierre Pétrequin, Michel Errera and Serge Cassen. English Heritage for ‘Early Iron Age midden site at East Chisenbury, Wiltshire’ by David Field and David McOmish. Wessex Archaeology for ‘Prehistoric, Romano-British and medieval activity at Ridge Green, Shaw, Swindon’ by Andrew B Powell; ‘Medieval remains at Pennings Road and St Andrews Road, Tidworth’ by Jonathan Milward, Andrew Manning, Lorraine Mepham and Chris J Stephens; ‘Early Bronze Age burial remains from Boscombe Down airfield, Amesbury, Wiltshire’ by Andrew Manning, Jacqueline I McKinley and Alastair Barclay; and ‘Late medieval/early post-medieval burial remains from Boscombe Down airfield, Amesbury, Wiltshire’ by Jacqueline I McKinley and Andrew Manning. The journals issued to volume 69 as parts of The Wiltshire Archaeological and Natural History Magazine (Part A Natural History; Part B Archaeology and Local History) were from volumes 70 to 75 published under separate titles as The Wiltshire Natural History Magazine and The Wiltshire Archaeological Magazine. With volume 76 the magazine reverted to its combined form and title. The cover title ‘Wiltshire Heritage Studies’ (volume 93) and ‘Wiltshire Studies’ (volume 94 onwards) should not be used in citations. The title of the journal, The Wiltshire Archaeological and Natural History Magazine, remains unchanged. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the Society and authors. Cover illustration: Silver sceatta from Marston Maisey, image courtesy of Katie Hinds Typeset in Aldine by Stuart Brookes and produced for the Society by Cromwell Press Group, Trowbridge, BA14 0XB Printed in Great Britain Contents NATURAL HISTORY, ARCHAEOLOGY and LOCAL HISTORY The petrography, affinity and provenance of lithics from the Cursus Field, Stonehenge, by Rob A. Ixer and Richard E. Bevins The Breamore jadeitite axehead and other Neolithic axeheads of Alpine rock from central southern England, by Alison Sheridan, David Field, Yvan Pailler, Pierre Pétrequin, Michel Errera and Serge Cassen The Late Bronze Age and Early Iron Age Midden Site at East Chisenbury, Wiltshire, by David McOmish, David Field and Graham Brown The Bedwyn Dyke: a revisionist view, by Ben Lennon Prehistoric, Romano-British and medieval activity at Ridge Green, Shaw, Swindon, by Andrew B. Powell Settlement in the hinterland of Sorviodunum: a review, by David Ff. James Medieval remains at Pennings Road & St Andrews Road, Tidworth, by Jonathan Milward, Andrew Manning, Lorraine Mepham and Chris Ff. Stephens The larger linear earthworks of Savernake, by Graham Bathe and Joanna Ramsay Silbury Hill: Edward Drax and the excavations of 1776, by Brian Edwards The relationship between Wansdyke and Bedwyn Dykes: a historiography, by Ben Lennon Devizes Literary and Scientific Institution: a study of change over time, by Elizabeth Gibb NOTES and SHORTER CONTRIBUTIONS Early Bronze Age Burial Remains from Boscombe Down Airfield, Amesbury, Wiltshire, by Andrew Manning, Facqueline I. McKinley and Alistair Barclay Cricklade without any Britons, by Richard Coates Celts and the Wylye, by Andrew Breeze Cunetio (1960) Hoard, by Sam Moorhead and Fohn Kentf Late medieval/early post-medieval burial remains from Boscombe Down Airfield, Amesbury, Wiltshire, by Jacqueline I. McKinley and Andrew Manning 102 130 142 181 186 257 269 289 309 312 314 BV) 320 Silbury Hill: Stukeley’s bridle and other finds, some reports from 1771 and 1850, by Brian Edwards Wiltshire dovecotes and pigeon houses: some clarifications and corrections, by RP M. Slocombe WANHS Archaeology Field Group: recent activities and future plans, by Jim Gunter REVIEWS, edited by Robert Clarke EXCAVATION and FIELDWORK in WILTSHIRE 2008, compiled by Simon Draper Highlights from the Portable Antiquities Scheme (PAS) in Wiltshire in 2008, recorded by Katie Hinds INDEX, by Philip Aslett 323 326 327) 329 336 341 349 The Wiltshire Archaeological and Natural History Society The Society was founded in 1853. Its activities include the promotion of the study of archaeology (including indusirial archaeology), history, natural history and architecture within the county; the issue of a Magazine, and other publications, and the maintenance of a Museum, Library, and Art Gallery. There is a programme of lectures and excursions to places of archaeological, historical and scientific interest, and an archaology field group for members who wish to become actively involved in the archaeological scene in the county. The Society’s Museum contains important collections relating to the history of man in Wiltshire from earliest times to the present day, as well as the geology and natural history of the county. It is particularly well known for its prehistoric collections. The Library houses a comprehensive collection of books, articles, pictures, prints, drawings and photographs relating to Wiltshire. The Society welcomes the gift of local objects, printed material, paintings and photographs to add to the collections. The Wiltshire Archaeological and Natural History Magazine is the annual journal of the Society and is issued free to its members. For information about the availability of back numbers and other publications of the Society, enquiry should be made to the Curator. Publication by the Wiltshire Archaeological and Natural History Society does not imply that the Society endorses the views expressed; the factual content and the opinions presented herein remain the responsibility of the authors. Notes for Contributors Contributions for the Magazine should be on subjects related to the archaeology, history or natural history of Wiltshire. While there is no fixed length, papers should ideally be under 7,000 words, though longer papers will be considered if of sufficient importance. Shorter, note length, contributions are also welcome. All contributions should be typed/ word processed, with text on one side of a page only, with good margins and double spacing. Language should be clear and comprehensible. Contributions of article length should be accompanied by a summary of about 100 words. Please submit two copies of the text (with computer disk if possible) and clear photocopies of any illustrations to the editors at the Museum, 41 Long Street, Devizes, Wiltshire, SN10 1NS. A further copy should be retained by the author. The editors will be pleased to advise and discuss with intending contributors at any stage during the preparation of their work. When submitting text or graphics on disk, Word or Rich Text Format files are preferred for text, jpeg or tiff format for graphics. Contributors are encouraged to seek funding from grant-making bodies towards the Society’s publication costs wherever possible. Referencing: The Harvard System of referencing (author, date and page, in parentheses within the text) is preferred: e.g. *... one sheep and one dog lay close together (Clay 1925, 69)’. References in footnotes should be avoided if at all possible. Only give references which are directly applicable, repeating as little as possible. All references cited in the paper should be listed in the bibliography using the following style, with the journal name spelled in full, and the place and publisher of books/ monographs given : For a paper: PITTS, M. W.and WHITTLE, A. 1992. The development and date of Avebury. Proceedings of the Prehistoric Society 58, 203-12 (Note that in citations Wiltshire Archaeological and Natural History Magazine is abbreviated to WANHM) For a book or monograph: SMITH, I.F, 1965, Windmill Hill and Avebury: Excavations by Alexander Keiller, 1925-39. Oxford: Clarendon Press For a paper in a book or monograph: FITZPATRICK, A., 1984, ‘The deposition of La Téne metalwork in watery contexts in Southern England’, in B. Cunliffe and D. Miles (eds), Aspects of the Iron Age in Central Southern Britain, 178-90. Oxford: University Committee for Archaeology Endnotes can be used for specific information that cannot otherwise be comfortably incorporated in the main body of the text. Illustrations need to be clear and easily reproducible, the format and proportions following that of the Magazine. If possible, all original artwork should not exceed A3 before reduction. If not supplied as computer graphic files, drawings should be produced on drafting film or high quality white paper using black ink. Detail and lettering should not be so small that it will become lost in reduction. Mechanical lettering (dry transfer or computer generated) is preferred over hand lettering. Photographs should be supplied as good quality black and white prints, and transparencies and colour prints avoided wherever possible. Original illustrations and photographs should only be sent once a contribution has been accepted. Offprints: Offprints will no longer be supplied, but authors (or the first named author where there are more than one) of full length contributions will receive an electronic pdf. file of the article as printed. WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY SOCIETY BOARD OF TRUSTEES (as at 1 January 2010) Chairman W.A. Perry, MSc Deputy Chairman D.L. Roseaman, BSc(Eng), CEng, MIMechE N.B. Harte, DSc(Econ), FRHist, FSA Hon Treasurer/Company Secretary Mrs W.P. Lansdown, FCA Other Elected Trustees Lt. Col. R.C. Chamberlain C.R. Chippindale, BA, PhD, MIFA, FSA J. Gunter, BA (Hons) Mrs V. A. Knowles, MIL, Dip. COT A. Reynolds, BA, PhD, FSA, FSA Scot., FRHistS PR. Saunders, BA, FSA, FMA, FRSA Mrs J. Triggs Nominated Trustees Cllr. C. Callow, BSc, ARCS, MBCS, Member Devizes Town Council A. X. T. Green, BA, MA, AMA, Director, Salisbury and South Wiltshire Museum Cllr J. G. Ody, Member Wiltshire Council Cllr. Mrs J. Swabey Member Wiltshire County Council In attendance: T Craig BA, MA, AMA, DipMgmt (Head of Heritage and Arts, Wiltshire Council) STAFF Director David Dawson, BA (Hons), FSA, AMA Curator Lisa Webb, BA (Hons), MA Education Officer Ali Rushent, BA (Hons), MA Sandell Librarian & Archivist Lorna Haycock, BA (Hons), PhD, DipELH, Cert. Ed, FRHistS Assistant Curator & Volunteer Co-ordinator Heather Ault, BA (Hons) Marketing & Administration Manager Karen Jones, BA (Hons) Membership Services Manager Margaret Clarke Shop Manager & Visitor Services Officer Diann Barnett, BA Wiltshire Archaeological & Natural History Magazine, vol. 103 (2010), pp. 1-15 The petrography, affinity and provenance of lithics from the Cursus Field, Stonehenge by Rob A. Ixer! and Richard E. Bevins As part of a comprehensive re-investigation into the lithology and provenance of the excavated bluestones of Stonehenge and its environs, fifteen bluestone fragments, either collected by F. FE S. Stone in 1947, or excavated from test pits within the Cursus Field in 2006 and 2008, have been described/restudied. They comprise rhyolite/rhyolitic tuffs and rhyolitic vitric/lithic/crystal tuffs together with two basaltic tuffs and are regarded as belonging to a single assemblage. Macroscopical comparisons with 6500+ bluestones excavated from within Stonehenge (in 1979/80 and April 2008) show that the Cursus Field lithics are typical Stonehenge bluestones. Microscopically the rhyolite/rhyolitic tuffs and basaltic tuffs can also be matched with Stonehenge bluestones 1n terms of their mineralogy and textures; however, many of the rhyolitic vitric/lithic/crystal tuffs have minerals/mineral assemblages not (yet) seen from Stonehenge. Although a few of the Cursus Field lithics, especially the basaltic tuffs, show some similarities to volcanic rocks from North Pembrokeshire, the majority, because of significant differences in mineralogy and textures, cannot be matched with certainty to the Lower Palaeozoic or Neoproterozoic rocks cropping out in southwest Wales, including those found on the Preseli Hills. This situation opens the possibility that, while the spotted dolerites are from the Preseli Hills, other Stonehenge orthostats together with the Altar Stone may come from a far wider and, as yet, unrecognised area, or more likely areas. Introduction The 3km long Greater Stonehenge Cursus to the north of Stonehenge has been a matter of interest since the early 18th century and has been most recently excavated as part of the Stonehenge Riverside Project (Thomas et al. 2009). Immediately south of the Cursus are open fields (the so-called Cursus Field of Stone 1947) leading down to the modern A344 and Stonehenge itself. In 1947 J. Stone reported ten bluestone fragments (used here in its original sense of a non-sarsen Stonehenge building stone, the term includes dolerites, volcanic rocks and sandstones), one within the Cursus (the notorious and mis-identified ‘Cosheston Sandstone’ discussed in Ixer and Turner (2006)), five close to the Cursus in the northwest of the field, one nearby within Fargo Plantation and three further to the south of the Cursus in the western part of the field. Stone noted that in 1934 W.E.V. Young had found ‘several bluestone chips and a small piece of micaceous sandstone’ whilst fieldwalking in the same area close to the western end of the Cursus and Fargo Plantation. Based on the anomalously high concentration (plotted in fig. 4 17.1948) and the lithological variation of the bluestones, Stone dismissed the possibility that they represented off- site dressing of the Stonehenge bluestone orthostats, tentatively suggesting instead that the lithics might be debris showing ‘the existence of a bluestone structure of late Neolithic or Grooved Ware age’ later incorporated into Stonehenge. Stone’s igneous rocks collected from the Cursus fields were thin sectioned and identified by Wallis ‘Dept. of Geology, University of Leicester, University Road, Leicester, LE] 7RH email: r.ixer@btinternet.com ? Department of Geology, National Museum of Wales, Cathays Park Cardiff CF10 3NP email richard.bevins@museumwales.ac.uk Table 1. Samples as listed in Salisbury and South Wiltshire Museum. Sandstone Recrystallised unlisted and|?Rhyolite/rhyolitic tuff ric-lithic tuff ric-lithic ash-flow tuff : unlisted and missing Bag6 1947/142.22 [312 Bag7 1947/142.21 |311 ?Rhyolite/rhyolitic tuff Rhyolitic crystal-lithic uff Rhyolite/rhyolitic tuff Section Rhyolite/rhyolitic tuff unlisted and missing Unsec- Sarsen tioned 316 ‘Lost’ as ‘four Presely rhyolites, two rhyolitic rocks perhaps from Presely and a crushed slate with microcrystalline quartz’ (Stone 1948 17). Twelve rocks were subsequently donated to the Salisbury Museum where they form the contents of a small box, number 49, with a heavily foxed label written in copper plate that reads: ‘Worked flints’ (crossed out) ‘and foreign stones from the Stonehenge Cursus 1947. Also foreign stones from the Stonehenge Car Park, from Mr Hoare’s Barrow 35, from the Cursus Field and from Wilsford, near Starveall. 142/47’. Ina later hand is written ‘+1 frag sarsen type stone’. 2 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE The contents comprise 12 lithic fragments in numbered bags (Table 1), 11 of which have been sliced (and thus probably thin sectioned). Eight have brief petrographical descriptions (Stone and Wallis 1951; Clough and Cummins 1988) and can be traced in Stone Axe Studies IT (SASII) (Table 2). These thin sections form part of the South-Western Federation of Museums and Art Galleries implement petrography collection held at Taunton Museum The additional foreign stones, from the Stonehenge Car Park, Mr Hoare’s Barrow 35 and Wilsford near Starveall cannot be matched to any entry in SASII nor are their associated thin sections in the Taunton collection; they are effectively lost. In Bag 12 the ‘sarsen type stone’ was unsectioned and is probably SASII 316 despite being identified in that volume as an ‘unsectioned chert’ and ‘lost’. This indurated sarsen is believed to be ‘a small fragment of a pink sarsen rubber’ and only one of two lithics excavated from the Cursus by Stone, the other being the so-called ‘Cosheston’ sandstone (SASII 275) now identified as a Lower Palaeozoic sandstone, not from Milford Haven but from an unknown locality (Ixer and Turner 2006). This lithic is subject to on-going studies. Current Methodology All twelve rocks have been sectioned and a polished thin section prepared from each; these were investigated in both transmitted and reflected light alongside the original seven thin sections described by Wallis (Table 2). Subsequently, an additional seven, igneous bluestone lithics collected from test pits in the same area of the Cursus Field in 2006 and 2008 (Stonehenge Riverside Project; Ixer et al. in Table 2 Cursus lithics as listed in Stone Axe Studies II. (SASIID) and identifications. SASII Listing WW} Wo] Uo Ww 14 rock Amesbury nr Stonehenge Cursus Ww Lithology 1951 275 rock Amesbury Stonehenge Cursus Cosheston sandstone |Cosheston |Lower Palaeozoic sandstone sandstone 313 rock Amesbury near Stonehenge Cursus |Calcareous chlorite Basaltic vitric tuff schist a Impure rhyolite roc 15 rock Amesbury nr Stonehenge Cursus Rhyolitic crystal-lithic tuff mh Lithology 2009 1988 sh k |Rhyolite _|Rhyolitic crystal-vitric-lithic tuff k 316 rock Amesbury nr Stonehenge Cursus | Not sectioned Indurated sarsen. from within the cursus. The 1951 lithologies are from Stone and Wallis 1951, 1988 lithologies are from SASII 1988. Note that in Stone (1947) rock 313 is described as a volcanic ash/crushed slate. THE LITHICS FROM THE CURSUS FIELD, STONEHENGE 3 press) were sectioned and polished thin sections examined. Detailed ‘total petrography’ as defined by Ixer (1994) and Ixer et al. (2004) was undertaken on all the samples. In addition the Cursus Field rocks were examined using a scanning electron microscope in backscatter mode. Both sets of lithics from the Cursus Field were compared with hand specimens described by Howard collected during Pitts’ excavations 1979/80 around the Heel Stone (Pitts 1982) and with approximately 6500 lithics, plus newly prepared polished thin sections of representative bluestones, from the Stonehenge April 2008 excavation (Darvill and Wainwright 2009). In addition, the material has been compared with major 20th century published descriptions of the non-dolerite bluestones from Stonehenge and with those for Group VIII axe-heads believed to be manufactured from South Wales rhyolites (Keiller et al. 1941; Jenkins in David and Williams 1995). Finally the Cursus Field lithics have been compared with potential source rocks from the southwest Wales region including microscopical re-examination of the original rocks (Bevins 1982; Bevins and Roach 1982; Kokelaar et al. 1985; Bevins et al. 1985, 1989, 1991, 1992, 1995; Cope and Bevins 1993; Thorpe et al. 1989). Thin section mineralogical identifications in transmitted light were made following Kerr (1959) and Deer, Howie and Zussman (1992) and in reflected light following Ixer (1990) and petrographical and lithological descriptions with reference to MacKenzie et al. (1982), Howells et al. (1991) and the British Geological Survey Rock Classification Scheme. Nomenclature It is clear when reading 20th century published descriptions of the non-dolerite bluestones from Stonehenge (Thomas 1923; Pitts 1982; Thorpe et al. 1991) and those of Group VIII axe heads (Keiller et al. 1941) that there is no uniform or consistent nomenclature and hence it is almost impossible to compare new descriptions with old. This is a problem that Howard addresses in Pitts (1982) and which is illustrated by Table B in Darvill (2006 132). Indeed, later researchers have been required to re-examine earlier material in order to try to meld existing petrographical descriptions/lithological identifications with current ones. Nomenclature of lithic fragments from archaeological sites will always be a compromise. This is because they are based at best on small rock fragments and an associated thin section/polished thin section for transmitted and reflected light microscopy. Polished thin sections are suitable for scanning electron microscope investigations, which can yield yet more important mineralogical information but can only provide limited input towards suggesting a definitive rock name largely because they are compromised by the lack of original, geological field relations for the rock fragments, an important component in assigning a rock name. Nomenclature of Stonehenge lithic debris will, therefore, remain difficult and at times controversial. This situation is particularly relevant to the fine- grained acid, igneous, non-dolerite bluestones. The proposed bluestone rock nomenclature scheme is based on that of the British Geological Survey Rock Classification Scheme. In addition, given that the lithic samples are most likely of Lower Palaeozoic age, or possibly Neoproterozoic, the terminology adopted by Howells et al. (1991) in their study of Ordovician volcanic sequences in Snowdonia, North Wales is also used. As there is a lack of consistency between these two schemes, a compromise has been determined that meets the needs of this study but which is hampered considerably by a lack of field evidence for the sub- and supra-stratal relationships of the samples being investigated. The following nomenclature, based on petrological investigations alone, comprises: Acid rocks, namely rhyolite and rhyolite tuffs It is frequently difficult to determine if a fine-grained rhyolite has an intrusive or extrusive origin both in the field and hand specimen. Sample descriptions cover features such as the presence of spherulites or flow banding. There are three main types of tuff: 1. rhyolitic lithic tuff with rock fragments, 2. rhyolitic vitric tuff with clasts of pumice/glass shards and, 3. rhyolitic crystal tuff with single crystals. The tuffs may be mixed depending on the clast content; for example a rhyolitic crystal lithic tuff carries both single mineral grains and rock clasts. Evidence of genesis is added, for example, if it seems likely that the tuff was generated by an ash flow. In such cases terms like rhyolitic vitric ash flow tuff are applied. Basic rocks namely tuffs There are two main types of tuff: 1. basaltic lithic tuff and 2. basaltic vitric tuff. Evidence for genesis, for example basaltic vitric hyaloclastite tuff, is included. + THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Such nomenclature is recommended for future publications. Results Detailed petrography of Stone’s Lithics from the Cursus Field Rhyolite/rhyolitic tuffs These rocks have fabrics that are common amongst the Stonehenge April 2008 material, the informally designated ‘rhyolite with fabric’. The loss of all primary textures due to intense recrystallisation and the lack of any geological field relations prevents a definitive lithological identification, hence the use of the term rhyolite/rhyolitic tuff. Bag 10. Rock. Cursus Field 142/1947.19 Macroscopically this lithic (the heaviest at 388g - all weights are after the removal of slices for sectioning) is a light bluish-grey, finely banded rhyolite/ rhyolitic tuff weathered to a yellowish-grey colour. Microscopically the banded, acidic rock displays a strong streaky foliation comprising alternating quartz-rich and more chlorite-rich bands. Quartz lenses and carbonate and quartz-chlorite veinlets are aligned along the foliation. The bulk of the rock is very fine-grained quartz with lenses and veinlets, comprising coarser-grained quartz mosaics; within these mosaics quartz crystals have sutured edges. Thin, simply or polysynthetically twinned, dusty, feldspar microliths lie along or at high angles to the main foliation. Locally the rock carries abundant, OP ails in Plate 1. 142/1947.21. Typical, pervasive lensoidal fabric superimposed on a planar foliation. Pale lenses comprise fine- grained quartz and streaks and lenses are chlorite-rich. Plane polarised transmitted light. Field of view ca. 7.0mm across. twinned, dusty plagioclase microliths. A single, large, simply twinned feldspar clast/phenocryst has a partial chlorite mantle. Elsewhere chlorite with blue interference colours forms veinlets or locally surrounds quartz segregations. Trace amounts of rounded zircon, TiO, mineral grains and trace amounts of chalcopyrite, pyrrhotite and sphalerite and euhedral apatite are present in the fine-grained quartz matrix. Brown fibrous/radiating titanite is associated with chlorite and limonite has replaced pyrite. Pale brown carbonate veinlets, locally associated with quartz or green chlorite, lie parallel to the main foliation. Thin, angular quartz veinlets cross-cut the main foliation and are associated with titanite. Bag 7. Rock. Cursus Field 142/1947.21 Macroscopically the lithic is a hard, medium bluish- grey, fine-grained, recrystallised, banded rhyolite or rhyolitic tuff (12.7g) with abundant pale spots up to 0.5mm in diameter. The rock has weathered to a light bluish-grey. There is a strong planar foliation with alternating dark and light bands together with a superimposed lensoidal fabric (Plate 1, 2). Microscopically this acid rock is strongly banded and intensely recrystallised with thin quartz-chlorite bands alternating with fine-grained quartz. The lensoidal fabric comprises quartz lenses, green chlorite and quartz-chlorite and rare chlorite-albite- rich lenses/bands within the main fine-grained quartz matrix. Zoned, simply or polysynthetically twinned, plagioclase microliths are rare; they have dusty cores and form laths that lie along or at high angles to the main foliation but are concentrated in Plate 2. 142/1947.21. Pale quartz-rich bands/lenses characteristically carry very small radiating TiO, aggregates (black spots). A small chlorite-TiO,, aggregate (top centre) ts present in quartz. Plane polarised transmitted light. Field of view 1s ca 2.7mm in width. THE LITHICS FROM THE CURSUS FIELD, STONEHENGE 5 Plate 3. 142/1947.22. Lensotdal fabric. A dark quartz- feldspar-opaque lens (centre) has a thin quartz-rich band above and below; the quartz carries fine-grained radiating TiO, aggregates (black spots). Thin, lath-shaped stilpnomelane (black) with characteristic bowtie shapes lies along the main foliation (bottom right). Plane polarised transmitted light. Field of view ca. 2.7mm in width. the fine-grained quartz layers. Poorly crystalline, pale-coloured titanite is associated with chlorite-rich areas. Rounded to euhedral, unzoned zircon some in clusters, lath-shaped TiO, minerals, chalcopyrite and limonite after pyrite are present but rare. The host rock is cut by thin, sinuous, quartz-chlorite veinlets. The loss of all primary textures due to an intense recrystallisation is pronounced. The presence of sinuous quartz-chlorite veinlets suggests rocks of perhaps a Neoproterozoic age rather than a Palaeozoic age. Bag 6. Rock. Cursus Field 142/1947.22 Macroscopically a light blue-grey, fine-grained, homogeneous, banded rhyolite/rhyolite tuff (22.2g) with a strong lensoidal fabric weathered to a pale orange colour. Up to 5mm thick, light olive-grey planar bands alternate with 2mm thick, light grey bands. The olive-grey bands have a uniform but pronounced lensoidal fabric with darker, 0.5—1mm long lenses. The rock is cut by stylolite-like quartz veinlets. _ Microscopically quartz-stilpnomelane + feldspar + opaque mineral lenses are present in a fine-grained matrix comprising quartz and stilpnomelane. Plagioclase (probably albite) forms rare, discrete microliths but also is present in feldspar-rich bands (Plate 3). Quartz-stilpnomelane veinlets are highly sinuous and cross cut the main foliation at high angles; individual stilpnomelane laths lie within the plane of the foliation but also at high angles Plate 4. 142/1947.22. A strongly foliated area, stilpnomelane laths lie at high angles to the foliation (top right, orientated NW-SE) and along the foliation (bottom right). The central, black lensoidal area is altered sulphide. Crossed polarised transmitted light. Field of view ca 2.7mm across. to it (Plate 4). Unzoned, rounded zircon, euhedral apatite, TiO, minerals (especially associated with plagioclase-rich areas), trace amounts of hematite replacing pyrite, chalcopyrite and possible titanite occur in minor to trace amounts. The random orientation of stilpnomelane laths is typical of its occurrence in Lower Palaeozoic volcanic rocks of the Welsh Basin, although it is rare in basic volcanic rocks generally occurring in more silicic rocks. Cross-cutting stylolite-like quartz veinlets, suggesting significant deformation, however, may be indicative of a Neoproterozoic age. Rhyolitic vitric/crystal/lithic tuffs These rocks have fabrics that are very common amongst the Stonehenge April 2008 material, the informally called ‘volcanics with sub-planar texture’ plus the sub-class ‘volcanics with sub-planar texture + voids’. The presence of vitric, crystal and rock clasts shows these rocks to be volcanic tuffs. Some of these rocks are very similar to the majority of the 2006/2008 Cursus Field test pit bluestone lithics. Bag 3. Rock. Cursus Field 142/1947.24 Macroscopically a light bluish-grey, fine-grained, banded, rhyolitic crystal-vitric-lithic tuff (29.3g) with small, linear-shaped, limonite-stained, dissolution voids weathered to a light bluish-grey to bluish-white. The greenish-grey tuff has dark yellowish-orange, limonite-rich streaks, sub-planar banding and a lensoidal fabric. Microscopically the tuff is banded/streaked 6 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Plate 5. 142/1947.24. The crystal-lithic tuff has fine-grained lithic fragments that resemble laminated mudstone (top left hand and right hand corners). Irregular single mineral grains (white) lie above dark limonite-stained streaks. Plane polarised transmitted light. Field of view ca 1.1mm across. and carries large, planar rock clasts that resemble laminated mudstone but have microphenocrysts and plagioclase megacrysts and thin, fine-grained white mica lenses in them (Plate 5). Other clasts include ~ tubular pumice (Plate 6) and angular, pale-coloured, fine-grained igneous rocks showing relict perlitic textures (Plate 7). Although there is an overall sub-planar foliation in detail, much of the fabric is chaotic and locally the foliation is draped around large, altered feldspar crystals. Coarse-grained, quartz-sodic plagioclase; quartz-opaques-?chlorite; quartz-untwinned or simply twinned feldspar + chlorite + opaque minerals; chlorite-untwinned feldspar-muscovite and polycrystalline plagioclase aggregates are present within fine-grained, banded, quartz-chlorite-white mica matrices that carry trace amounts of unzoned, euhedral zircon. Within these aggregates, cloudy plagioclase cores have thin, syntaxial, clear overgrowths and quartz occurs within mosaics. Some of these aggregates are altered single crystals but others are probably void infills especially those comprising intergrown quartz- chlorite or titanite. Pale brown, irregular, titanite aggregates are widespread and enclose small, pale- coloured TiO, minerals and are intergrown with quartz and feldspar; some intergrowths infill void spaces. Limonite is associated with clear quartz or forms thin veinlets where limonite pseudomorphs after pyrite carry small relict pyrite. Bag 9. Rock. Cursus Field 142/1947.25 Macroscopically a light bluish-grey rhyolitic crystal- lithic tuff (41.3g) has weathered to a pitted, light Plate 6. 142/1947.24. A large pumice clast lies in the centre (grey-pale with vertical streaking). A sigmoidal thin opaque- rich veinlet (black top left) partially surrounds the clast. Plane polarised transmitted light. Field of view ca. 2.7mm across. greenish-grey, slightly spherulitic surface. The tuff has thin, limonite-rich streaks associated with fine- grained bands and a chaotic fabric. Microscopically the majority of the rock comprises fine-grained quartz-chlorite intergrowths. The rock has a sub- planar foliation and some streaks comprise abundant, fine-grained feldspar microliths intergrown with chlorite. Plagioclase (probably albite) is simply or polysynthetically twinned and forms rare, large, euhedral clasts with dusty cores within clear margins; they show slight alteration to fine-grained Plate 7. 142/1947.24. A large clast (near centre) shows relict perlitic textures indicative of volcanic glass devitrification. Other clasts are smaller and some are fine- grained (dark grey, top right corner). A euhedral pyrite (black middle right hand side) ts associated with quartz (white middle right hand side). Plane polarised transmitted light. Field of view ca 2.7mm in width. THE LITHICS FROM THE CURSUS FIELD, STONEHENGE 7 year me cus Plate 8. 142/1947.25. Subhedral, dark coloured titanite (centre) encloses small lath-shaped albite (top and bottom edges of titanite). A mass of intergrown albite (patchy) lies to the left of the titanite and clear quartz and chlonite (pale grey) to its right. Plane polarised transmitted light . Field of view approximately 2.7mm in width. white mica or have chlorite rims. Irregular void spaces are common and are partially or completely infilled with medium-grained intergrowths of quartz-simply twinned albite-chlorite-titanite and very rare muscovite. The relative proportions of the four minerals vary and not all voids carry all phases. Characteristically, stubby, euhedral, end-member albite crystals enclosed within euhedral, aluminium- bearing titanite (confirmed by semi-quantitative energy dispersive analysis) occur at the edges of the voids and quartz and chlorite infill their cores (Plate 8, 9). Chlorite shows blue interference colours with energy dispersive analysis suggesting it is a mixed iron-magnesium chlorite between chamosite and clinochlore. Titanite occurs as euhedral, rhombic, brown, zoned crystals growing into chlorite and encloses rare, small, colourless TiO, laths. In some bands fine-grained diffuse titanite patches are similar to those that characterise Group VIII axes but here they are not as well developed. Sulphides are rare but include pyrrhotite, pale-coloured sphalerite, chalcopyrite and pyrite, the last two altering to limonite. Small zircon crystals are present. Very rare, graphitising carbon occurs. Natural graphitising carbon is rare in igneous rocks but has also been recognised from Stonehenge monolith 38. This tuff is very like many of the Cursus Field test pit bluestones. Bag 4. Rock. Cursus Field 142/1947.20 Macroscopically the second largest lithic (172.8g) is a light bluish-grey, homogeneous, rhyolitic crystal- Plate 9. 142/1947.25. Back scatter SEM image of euhedral albite (dark grey) surrounding titanite (white) in coarse grained secondary growth area. Matrix is dominated by quartz and albite (both dark grey) and chlorite (ght grey). Chlorite defines the foliation in the matrix. vitric ash-flow tuff with a slight planar foliation weathered to a greenish-grey colour with pale yellowish orange, limonite-rich bands. The rock carries clear or cloudy feldspar microliths within a fine-grained matrix. Limonite-stained pits have formed from weathered pyrite. Microscopically chlorite-rich clasts and large feldspar and quartz crystals with chlorite rims are present within a fine- grained, quartz-rich matrix; overall the rock shows relict pumice clasts (Plate 10) with shardic textures suggesting a non-welded, ash-flow tuff. Large, Plate 10. 142/1947.20. A large pumice clast enclosed in a fine-grained quartz-rich matrix. Random black sub- rounded to hexagonal areas are pyrrhotite/altered pyrrhotite (top edge). Plane polarised transmitted light. Field of view ca. 2.7mm in width. 8 Plate 11. 142/1947.23. Basaltic hyaloclastite with a large volcanic clast with flattened vesicles infilled with quartz. The edge of the section (bottom left) is black. A sparry calcite veinlet lies to the top of the clast. Crossed polarised transmitted light. Field of view is ca 2.7mm in width. euhedral, zoned, untwinned or simply twinned feldspar is altered/replaced with chlorite-rich rims and quartz cores; some crystals are fractured whereas most, zoned, polysynthetically twinned plagioclase crystals are pseudomorphed by coarse-grained quartz or sparry calcite. Unzoned, euhedral zircon locally forms aggregates or is intergrown with pyrrhotite and limonite. Very rare minerals include ilmenite present within titanite, magnetite, graphite and lath-shaped, pale-coloured TiO, minerals. Sulphides are widespread with abundant pyrrhotite. It shows a wide range of characteristic alterations including bird’s eye textures, fine-grained pyrite/marcasite intergrowths and finally limonite. Pyrrhotite is intergrown with galena, arsenopyrite and orange to colourless sphalerite. Altered basaltic tuffs Altered basaltic tuffs are a minor component of the bluestones excavated during the April 2008 Stonehenge excavation (included in the informal class ‘volcanics with sub-planar fabric’) and the non-dolerite bluestones collected during the 1979- 80 excavations at Stonehenge (Pitts 1982). There they totalled less than 1kg in weight mainly found as small heavily abraded fragments variable in their macroscopical properties (Pitts, 1982, 117-118). Bag 8. Rock. Cursus Field 142/1947.23 Macroscopically a greenish-grey banded, altered basaltic vitric tuff (33.2g) weathered to a pitted, greenish-grey rock with many dissolution pits. A THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE thick, light olive-grey band with limonite-stained chlorite and abundant voids due to the dissolution of calcite overlies a greenish-grey band with coarse- grained sparry calcite areas and yellow-green chlorite; the banding is sub-planar. The tuff carries rare, laminated mudstone-like clasts lying within the main foliation. Quartz-rich lenses occur within a fine-grained quartz-titanite-limonite-stained chlorite matrix. Thin tension gashes within these quartz lenses are infilled with inclusion-free calcite that does not penetrate into the surrounding silica- rich matrix but elsewhere quartz-rich areas of the matrix also are cut by thin, calcite-infilled tension gashes; all have the same orientation. Other tension gashes, all at a high angle to the main banding, are infilled with quartz-calcite-chlorite. Small, vesicular, volcanic shards are present, while larger volcanic clasts have long flattened vesicles now infilled with quartz (Plate 11). Voids are common and infilled with quartz-chlorite-calcite and, less commonly, by albite or small, euhedral, rhombic titanite. Typically, voids have a chlorite-rich + quartz rim and calcite-rich core but others comprise quartz-euhedral albite-rich zones mixed with chlorite-rich zones. Chlorite has blue interference colours and mostly forms radiating masses; minor amounts of white mica occur as thin layers. Opaque minerals include unaltered ilmenite laths associated with trace amounts of secondary hematite and TiO, minerals. Rounded titanite comprises fine-grained, poorly crystalline/spongy cores within a well-crystalline outer margin. Titanium dioxide grains replace ilmenite or form clusters/aggregates, lath-shaped pyrrhotite is very rare and pyrite is almost totally replaced by pale blue limonite with red internal reflections. Bag 5S. Rock found near Barrow 35 (Hoare No. 35) 142/1947.27 This barrow lies at SU 11466 42680, south of the Cursus and southwest of the Cursus barrow group and close to the test pits area. Macroscopically a light grey, fine-grained, altered basaltic tuff (37.4g) with a planar and superimposed sub-lensoidal fabric has weathered to a pitted light olive-grey. More mafic-rich bands are darker and resemble laminated mudstone. The olive-grey very fine-grained rock has a strong sub-planar foliation with very thin banding and much limonite along veins that lie within the main foliation. Microscopically small quartz-rich or quartz-chlorite (probably between chamosite and clinochlore)-rich lenses and very small chlorite-rich patches occur within a fine- grained, quartz-?calcite, or locally, quartz-chlorite, THE LITHICS FROM THE CURSUS FIELD, STONEHENGE 9 matrix. Thin veins and voids carry radiating chlorite partially enclosing coarse-grained, inclusion-free sparry calcite. Opaque minerals include altered iron titanium oxides that now comprise secondary ilmenite-TiO, intergrowths. Ilmenite forms small, skeletal crystals or more commonly laths associated with pale-coloured TiO, and with chlorite-rich areas and is replaced by TiO,. Small, euhedral, rhombic titanite is concentrated into dark bands and associated with TiO,. Macroscopically this lithic can be matched with Howard’s basic tuffs especially find number 212 (Pitts 1982, 118, table 2) and with the Stonehenge April 2008 lithics informally designated as sub-planar rhyolite with voids. Petrography of the Cursus Field 2006/2008 Test pit material In 2006 and 2008 the Cursus Field was test pitted towards the western end of the Cursus and close to Fargo Plantation, in order to test Stone’s suggestion that this area contained an unusually high concentration of bluestone fragments. A total of 24 fragments of sarsen and seven bluestones (Table 3) were recovered from the plough soil (Ixer et al. in press) in addition to a small ‘cache’ of hematite close to the southern boundary of the Cursus. All of the sarsens and bluestones are present towards the western end of the Cursus and lie well inside the bluestone concentration previously recorded by Stone (1947, 17, fig. 4) (Ixer et al. in press). Seven bluestones comprise four very similar, fine-grained, vitric-lithic tuffs (polished thin sections FP06 72, 74, 157 and 173). Two others (FP06 152 and FARO8 247) are rhyolite/rhyolitic tuffs, which share many, but not all, of the petrographical features of the previous four. A single, fine-grained, rhyolitic crystal tuff/ sparsely porphyritic rhyolite (polished thin sections FP06 50 and 501) differs from the other bluestones in being strongly foliated. Petrographically lithics FP06 72, 74, 157, 173 are vitric-lithic tuffs and carry clasts comprising a fine- grained, quartz-chlorite matrix with small feldspar Table 3 Test pit lithologies. Sample Number rhyolite FP06 50 l FP06 72 Vitric-lithic tuff FP06 74 Vitric-lithic tuff [FARO8 247. —__— [Rhyolite/rhyolitic tuff microliths; other clasts include chlorite-feldspar-rich clasts and yet others that have a perlitic texture. In addition, they are united in having distinctive, late- stage, albite-titanite-chlorite+ quartz segregations, where euhedral albite crystals are overgrown by euhedral to rounded titanite. A ?limonite- stained, highly siliceous, rhyolitic tuff (FP06 152) with thin feldspathic ‘shale-like’ wisps carries the distinctive late-stage, albite-titanite-quartz segregations but chlorite could not be positively identified either in the groundmass or within the segregations. A rhyolite/rhyolitic tuff with a sub- planar foliation (FARO8 247) shares many of the features of the previous five samples but, although quartz-albite-chlorite veinlets and titanite occur, the distinctive albite-titanite intergrowths are very poorly developed, if present at all. The fine-grained strongly foliated rhyolitic crystal tuff/strongly foliated, sparsely porphyritic rhyolite (FP06 50) comprises rare, small feldspar microliths and larger, possible alkali feldspar crystals/porphyroblasts, set in a very fine-grained, quartz-chlorite matrix. Later chlorite-quartz + calcite + graphite veinlets lie along the planar foliation or are sigmoidal and cut across the main fabric. Although this lithic (FP06 50) shares sufficient petrographical features to demonstrate that it is a true bluestone, it is unlike the six other test pit bluestones. Detailed petrography of Stone’s lithics away from the Cursus Field Bag 2. Rock found near Stonehenge Car Park 142/1947.26 Bluestones have not been recorded over most of the Cursus Field and there is a gap in their numbers until the vicinity of the present day Stonehenge Car Park. Macroscopically a fine-grained, moderate blue rhyolite/rhyolitic tuff (63.6g) weathered to a yellowish-grey to pale orange has a 2mm thick, quartz layer carrying limonite after pyrite. The fine-grained, banded, rhyolitic rock has a lensoidal foliation and is cross-cut by thin sinuous, chlorite- rich veinlets. Microscopically lensoidal aggregates comprising polygonal quartz mosaics within chlorite rims, discrete, simply twinned feldspar, and lath-shaped plagioclase microphenocrysts/crystal fragments with cloudy cores but clear margins, locally forming crystal clusters, are present within a fine-grained quartz-rich matrix. Chlorite forms rare, single crystals but mainly occurs as rims to quartz segregations or to highly contorted, stylolite-like, 10 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE cross-cutting, quartz veinlets. Trace amounts are present of subhedral, unzoned zircon, pale-coloured TiO, minerals, and pale-coloured, euhedral light REE-bearing epidote, as is potassium feldspar (the last two phases were confirmed by energy dispersive analysis). Sulphides are commonly associated with quartz, the most abundant being pyrrhotite, often showing its characteristic bird’s eye alteration texture. Pyrite, some altered to limonite, is present as euhedral crystals within pyrrhotite or partially replaces pyrrhotite. Orange-brown sphalerite, mixed sphalerite-chalcopyrite and chalcopyrite are rare but when present are often associated with pyrrhotite. The loss of all primary textures due to the intense recrystallisation and the lack of any geological field relations prevent a definitive lithological identification. However, cross cutting stylolite-like quartz-chlorite veinlets may suggest a Neoproterozoic age rather than a Palaeozoic age. Lithologically it is indistinguishable from the rhyolite/rhyolitic tuffs from the Cursus Field and from Stonehenge especially those designated as ‘rhyolite with fabric’. In 1935 Young recovered six pieces of rhyolite, two of volcanic ash and two of dolerite whilst excavating the Stonehenge Car Park and he suggested that they were ‘thrown out of the circles and not in a position suggestive of a trimming floor’ (Stone 1947/8 16.) This lithic records a seventh bluestone from the car park area. Bag 11. Rock found at Wilsford near Starveall. 142/1947.28 Macroscopically the light bluish-grey, fine-grained rhyolite/rhyolitic tuff (17.7g) has weathered to a yellowish-grey. The very fine-grained, homogeneous rock is very thinly banded with quartz-chlorite- rich bands alternating with fine-grained quartz. It has a planar fabric with a suggestion of streaking and a superimposed sub-lensoidal foliation. Microscopically quartz and quartz-chlorite lenses are present in a fine-grained, quartz-rich matrix with minor amounts of lath-shaped, green chlorite. Rare, discrete, plagioclase (probably albite) microliths within the fine-grained matrix have cloudy cores and clear rims. Locally, plagioclase-chlorite-TiO, mineral-rich intergrowths are present and may be rock clasts. Discrete, thin, brown, pleochroic stilpnomelane laths (confirmed by semi-quantitative energy dispersive analysis) lie both within the plane of the foliation and cut across it (Plate 12). Unzoned zircon crystals, some within aggregates and poorly crystalline TiO, grains occur. Zoned reddish-brown, Plate 12. 142/1947 .28. Stilpnomelane laths lie along the sub-horizontal main foliation and at high angles to 1t (top middle). A limonite patch after sulphide (bottom right corner) 1s present. The main matrix 1s fine-grained quartz with minor chlorite. Plane polarised transmitted light. Field of view ca 2.7mm in width. radiating sphere and light REE-bearing epidote are intergrown with coarse-grained quartz and chlorite and are concentrated within some of the bands. This is probably ‘the surface find by Newell approximately lkm south of the Normanton group of barrows’ as reported by Stone (1948 16). Lithologically it is indistinguishable from the rhyolite/rhyolitic tuffs from the Cursus Field (especially 1947/142.22) and from lithics at Stonehenge especially those designated as ‘rhyolite with fabric’. Detailed petrography of Stone’s lithics excavated from within the Cursus Bag 12. 142/1947.10 Macroscopically the sarsen (61.3g, original weight, uncut) is a greyish-red, indurated, fine-grained, unbedded, sandstone. Microscopically the rock is a vuggy, locally grain-supported quartz arenite comprising single, sub-rounded to sub-angular, detrital quartz grains with authigenic overgrowths and euhedral terminations growing into void spaces. Locally fine-grained hematite pigment or fine-grained, quartz mosaics overgrow and cement euhedral quartz. Rare, detrital, accessory minerals include green-brown tourmaline, oxidised chromite, red-orange rutile and zircon. Rock clasts are also uncommon but include rounded, chert/flint, a little metamorphic quartz, vein quartz and clasts carrying a very fine-grained hematite pigment. This is one of only two artefacts from the filling of the Cursus ditch and is a ‘small fragment ofa pink THE LITHICS FROM THE CURSUS FIELD, STONEHENGE 11 sarsen rubber’ (Stone, 1948 15). It is very similar to the indurated sarsens recovered from the 2006/8 test pits (Ixer et al. in press). Petrographical comparisons with the extensive descriptions by Howard (in Pitts 1982) show the sarsen to be typical. Discussion Stone’s 1947 bluestones and the 2006/2008 test pit bluestones (Ixer et al. in press), have been compared with the appropriate descriptions in the literature and directly, both macroscopically and microscopically, with newly thin-sectioned Stonehenge April 2008 material, Pitts 1982 material (only macroscopically as the associated thin sections cannot be traced), IPG Group VIII axe heads and with a range of in situ (outcrop) potential source rocks from South Wales. Stone’s Cursus Field and 2006/08 test pit Cursus Field lithics Stone’s Cursus Field lithics and the bluestones recovered in the 2006/2008 trial pits share sufficient features to suggest they are part of the same archaeological assemblage. Both sets are dominated by rhyolitic tuffs/fine-grained rhyolites and no spotted dolerite has been recognised from either set despite dolerite making up the majority of the standing orthostats at Stonehenge (Thorpe et al. 1991, table 1, 106) and being common (perhaps the most common) debris within the stone circle. Two of Stone’s lithics (142/1947.24 and 142/1947.25) and between four and six of the seven test pit bluestones (FP06 72, 74, 157, 173 plus perhaps152 and FP08 247) are characterised by distinctive, metamorphic, late stage albite-titanite-chlorite-quartz intergrowths. Stone’s sample (142/1947.25) especially is a close match with test pit sample FP06 173. Cursus Field lithics and Group VIII axe lithologies The Implement Petrology Committee Group VIII artefacts are described as manufactured from rhyolites with ‘an even and finely-textured micro-crystalline mosaic of quartz and possibly feldspar with scattered irregular non-homogenous accumulations of leucoxene’ (Keiller et al. 1941, 63) cropping out at Carnalw just north of the spotted dolerite exposures of Carn Menyn in the Preseli Hills. The presence of ‘spongy clusters of fine-grained titanite up to 1mm in diameter are the defining feature of this lithic’ (Jenkins in David and Williams 1995, 459) and a very typical example of the group is shown in photomicrograph A, Plate VI in Stone and Wallis (1951, 131). Comparison of the Cursus Field material with thin sections from Carnalw from outcrop material from the National Museum of Wales collections, from typical Group VIII axe-heads (Ixer unpublished), from the National Museum of Scotland collections, and descriptions in the literature (Keiller et al. 1941; Stone and Wallis 1951; Jenkins in David and Williams 1995, 459) show that none of the Cursus Field lithics would be classed as Group VIII as originally defined by Keiller et al. (1941). Indeed, examination of Carnalw sections from both museum collections shows the rocks to have a distinctive, spherulitic recrystallisation texture (as illustrated in Thomas, 1923) that is missing from classical Group VIII axe-heads suggesting that Carnalw as a source for Group VIII should be re- examined. This is in broad agreement with Jones and Williams-Thorpe (2001) who have shown that a number of Group VIII axe-heads are chemically different from rocks cropping out at Carnalw. The authors suggest that the axes were manufactured from ‘rhyolites’ from a number of localities in Wales. Cursus Field lithics and Stonehenge lithics When compared with the April 2008 Stonehenge material the Cursus Field lithics can be characterised macroscopically as typical Stonehenge material belonging to the major, informally designated classes, ‘black flinty rhyolite’ (1947/142.20), ‘rhyolite with fabric’ (1947/142.19, 21 and 22; FP06 72) and ‘volcanics with sub-planar fabric’ (1947/142.23 and 25; FP06 74, 152, 157, 173 and FP08 247) ‘+ vughs’ (1947/142.24) or to a minor class ‘phyllitic volcanics with sub-planar texture’ (FP06 50). Similarly direct macroscopical comparisons with the material from the 1979-1980 Stonehenge excavations described by Howard in Pitts (1982) show that most of the Cursus Field lithologies are similar to Howard’s ‘rhyolite Group B’, a varied but essentially tuffaceous group. One lithic (1947/142.23) belongs with her ‘basic tuffs’ as does lithic 1947/142.27 (close to Hoare’s barrow 35) and FP06 50 belongs with her ‘rhyolite Group A, especially her sample 1167 (Pitts 1982) Microscopically some of the Cursus Field lithics have equivalents with excavated lithics at Stonehenge, especially the lensoidal textures and some of the minor mineralogy seen within the rhyolite/rhyolitic tuffs can be matched very closely to thin sections of the ‘black flinty rhyolite’ and 12 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE ‘rhyolite with fabric’ excavated at Stonehenge (April 2008). In particular stilpnomelane laths present in rhyolite/rhyolitic tuffs from Stonehenge (SH08 265 and 695) have the same fabrics as those shown in Stone’s 1947/142.22 rhyolite/rhyolitic tuff. The basaltic tuffs from the Cursus Field, the Heelstone area of Stonehenge (Pitts 1982) and from the 2008 April excavation are also petrographically very similar. Although this is a minor class of lithic from Stonehenge it is of importance as the lithology is close to, and perhaps even the same as, buried Stonehenge orthostat 40c (Atkinson 1979). Cunnington (1884, 142) described this orthostat (his S28) or at least its ‘rock stump’ as ‘a soft calciferous chloritic schist ...differing entirely from any other of the stones of Stonehenge’. Thomas (1923, 262) suggested that similar calcite-bearing basic ashes cropped out north of Foel Trigarn in the Preseli Hills. However, none of the limited number of sectioned lithics from the Stonehenge April 2008 excavation (13) carries the distinctive albite- titanite-chlorite-quartz association and Howard’s petrographical descriptions using a similar number of sectioned ‘rhyolites’ (13) from the Heelstone area of Stonehenge also fail to record these titanite- bearing intergrowths. It is too soon to establish whether this petrographical difference is real or reflects different sampling strategies, namely a total sampling of the Cursus Field material but the very selective sampling for the Stonehenge lithics that concentrated on ‘rhyolite with fabric’ samples. Cursus Field lithics and Welsh field localities The Cursus Field material comprises a variety of rhyolitic and basaltic tuffs, and possibly rhyolitic lavas. The fine-grained character of the lithics combined with evidence of extensive recrystallisation, makes both identification and geological provenancing of these difficult tasks. It is possible, however, by drawing direct comparisons with previously studied sequences in southwest Wales concentrating on their detailed mineralogies and textures to infer something about the provenance of the lithics. These sequences include those published for the Ordovician rocks around Fishguard (Bevins 1982; Bevins and Roach 1982), Ramsay Island (Kokelaar et al. 1985), Mynydd Preseli (Bevins et al. 1989, 1991), and ‘Treffgarne and Sealyham (Bevins et al. 1992), the Silurian rocks of Skomer Island (Thorpe et al. 1989), and the Neoproterozoic sequences around Llangynog (Cope and Bevins 1993; Bevins et al. 1995), St Davids (Bevins et al. 1995) and to the south of Haverfordwest (Bevins et al. 1995). In addition, unpublished studies of the Roch Rhyolite Group, previously considered to be Neoproterozoic but now thought more likely to be of Ordovician age, are included in this study. Stone’s rocks (1947.142.23) and near Barrow 35 (1947/142.27) are strongly foliated altered basaltic tuffs, most probably hyaloclastites. Such tuffs are seen in the Fishguard region and also further to the west, near Abereiddi (Kokelaar et al., 1984; Bevins et al., 1992), to which they bear similarity. The lensoidal foliation, characteristic of the recrystallised rhyolite/rhyolitic tuffs (1947/142. 21 and 22: car park 1947/142.26) and test pit sample (FARO8 247) is more likely to be a metamorphic foliation developed during deformation rather than a primary flow foliation in fine-grained rhyolitic lavas or intrusives. Partial evidence for this comes from the concentration of chlorite (most probably of a metamorphic origin) in the plane of foliation, as seen in lithics 1947/142.22 and 1947/142.28 from near Starveall (in the latter case chlorite is associated with metamorphic epidote), and of stilpnomelane (a characteristic mineral of low-grade metamorphism in more silicic rocks) both in the plane of foliation and cross-cutting it, sometimes showing a sigmoidal form suggestive of a syn-tectonic origin. Although stilpnomelane is known from North Pembrokeshire (Bevins and Rowbotha m 1983; Bevins 1994) it has not been recorded in any of the rhyolites or rhyolitic tuffs in that region, in particular it is noticeably absent in the Mynydd Preseli area, as are rhyolitic rocks with a lensoidal foliation (indeed, fine-grained igneous rocks of southwest Wales tend not to show a lensoidal foliation). Correctly provenancing rocks with this strong lensoidal foliation + stilpnomelane is important as these lithics form a significant component to the Stonehenge April 2008 excavation material, as seen in thin sections Stonehenge SH08 265, 695, where they are informally classed as rhyolite with fabric. Many of the rhyolitic tuffs of the Fishguard and Mynydd Preseli region preserve original shardic and pumiceous textures, such as at Lower Fishguard Harbour, Carn Cwm and Carn Gwiber, and well illustrated in Lowman and Bloxam (1981). Such textures are only seen in sample 1947/142.24. Other silicic rocks in the Mynydd Preseli region, for example those exposed on Carn Ingli and Foel Drygarn, comprise holocrystalline microtonalites, dominated by plagioclase feldspar and with minor quartz. Rhyolitic rocks are present immediately to the south of Mynydd Preseli, comprising the Sealyham THE LITHICS FROM THE CURSUS FIELD, STONEHENGE 13 Volcanic Formation exposed between Treffgarne and Carn Afr. Original igneous textures, however, are largely preserved in these porphyritic dacites and rhyolites, which also contain the distinctive low- grade metamorphic minerals prehnite, pumpellyite and, more rarely, axinite - none of which have been recognised in any of the Cursus Field material or any Stonehenge material other than within the spotted dolerites. Also of note is the presence of coarse-grained ageregates of quartz+titanite +chlorite+albite, with titanite crystal clusters reaching up 0.5mm in Stone’s 1947/142.25 and in Cursus Field test pit samples (FP06 72, 74, 157 and 173). Despite the extensive studies of volcanic rocks in southwest Wales listed above there has been no recording of such crystal aggregates dominated by titanite, nor were any seen during the present re-examination of South Wales igneous rocks. Comparing the Cursus Field material and in situ Welsh samples show that whilst a few do share similarities to volcanic rocks from North Pembrokeshire, notably the basaltic tuffs, the majority, based on differences in mineralogy, textures and character of metamorphism, cannot be matched up with any certainty to any of the Lower Palaeozoic or Neoproterozoic rocks cropping out in southwest Wales. Conclusions The lithics collected in 1947 and in 2006/08 from the Cursus Field material are concentrated south and towards the western end of the Stonehenge Great Cursus and are consistent with Stone’s suggestion of an unusual concentration of bluestones in that area. Detailed petrography shows them to comprise a variety of recrystallised rhyolite/rhyolitic tuffs, rhyolitic vitric-crystal-lithic tuffs and minor basaltic tuffs, but that spotted dolerites - the signature rock of the Stonehenge bluestones — are absent. The 1947 and 2006/08 assemblages have sufficient lithologies in common that they can be regarded as originally being part of a single archaeological assemblage. Macroscopically the Cursus Field assemblage is identified as bluestone and can be matched closely to material identified from the Pitts 1979/80 and April 2008 excavations at Stonehenge. Microscopically, some of the lensoidal rhyolite/rhyolitic tuffs + stilpnomelane from the Cursus Field can be matched with rock excavated from Stonehenge. However, for other bluestones, especially the tuffs, there are significant mineralogical differences between the Cursus Field assemblage and macroscopically similar rocks from within Stonehenge, most notably the presence quartz+titanite+chlorite+albite aggregates in the Cursus Field rhyolitic tuffs and their apparent lack from any Stonehenge sample. It is hoped that current sampling to include a wider range of rhyolitic tuffs from Stonehenge will help to determine whether this difference is genuine or due to sampling bias. Although some of the Cursus Field lithics, especially the basaltic tuffs, show similarities to volcanic rocks from North Pembrokeshire, the majority - because of significant differences in mineralogy and textures- could not be matched up with certainty to any of the Lower Palaeozoic or Neoproterozoic rocks cropping out in southwest Wales including those found on the Preseli Hills. Postscript After this article was written, in October 2009, a recrystallised rhyolite/rhyolitic tuff with a strong lensoidal foliation and stilpnomelane, a lithology notably absent from the major rhyolitic outcrops at the higher levels in the Preseli Hills, was identified microscopically from an undistinguished outcrop at Pont Saeson in the lower ground to the north of the Preseli Hills. Although this locality is most probably not large enough to provide the origin of the abundant ‘rhyolite with fabric’ debitage found at Stonehenge, it strongly suggests that this important component of Stonehenge can be sourced to the Preseli Hills region. It is clear that an intensive search of all other outcrops in the Preseli Hills environs is required in order to determine if the still unprovenenced bluestones, especially the albite-titanite-bearing tuffs can be found im situ or if indeed they are from further afield as hinted at to date by their petrography and geochemistry. What this study strongly suggests is that none of the major rhyolitic outcrops present in the higher levels of the Preseli Hills, which have undergone decades of sampling, provides an obvious match for the Cursus Field or Stonehenge lithologies recovered thus far. The Port Saeson discovery shows the importance and robustness of detailed petrography coupled with a sampling strategy that includes unpromising outcrops in addition to major archaeologically significant ones, for example Carnalw, in helping to solve provenance issues. It is anticipated that continuing this approach will yield further 14 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE significant results and provide additional detailed provenancing. The discovery of the stilpnomelane lensoidal rhyolites is the first time a non-dolerite bluestone has been clearly provenanced, it will not be the last. Acknowledgements We are grateful to the staff of the Stonehenge Riverside Project but especially Professor Mike Parker Pearson, the South-west Implement Petrology Group, Somerset Heritage Services and Dr Vin Davis of the Implement Petrology Group but especially the Salisbury and South Wiltshire Museum either for lending rocks and thin sections and/or for permission and partial funding to make more sections. Dr Chris Blake is thanked for re-confirming the identity of some of the crucial mineral phases while Mike Lambert (National Museum of Wales) is thanked for assistance with the scanning electron microscopy. Dr Sara Lunt of English Heritage was instrumental in gaining permissions and provided sustained, overall encouragement. Dr Olwen Williams-Thorpe commented promptly on aspects of the paper. Dr Peter Webb’s comments on the whole paper were detailed, pithy and of immense value. The Constantine XI Palaeologos Research Fund is acknowledged for its usual unstinting assistance. Formal lithological descriptions are with the authors and will be deposited with the Salisbury and South Wiltshire Museum and The National Museum of Wales. References ATKINSON, R.J.C. 1979. Stonehenge. Harmondsworth: Pelican Books BEVINS, R.E. 1982. Petrology and geochemistry of the Fishguard Volcanic Complex, Wales. Geological Fournal, 17, 1-21. BEVINS, R.E. 1994. A Mineralogy of Wales. Cardiff National Museum of Wales. BEVINS, R.E., LEES, G.J. and ROACH, R.A. 1989. Ordovician intrusions of the Strumble Head-Mynydd Preseli region, Wales: lateral extensions of the Fishguard Volcanic Complex. Journal of the Geological Society of London, 146, 113-123. BEVINS, R.E., LEES, G.J. and ROACH, R.A. 1991. Ordovician bimodal volcanism in SW Wales: geochemical evidence for petrogenesis of the silicic rocks. Journal of the Geological Society of London, 148, 719-729. BEVINS, R.E., LEES, G.J. and ROACH, R.A. 1992. Petrogenesis of Ordovician igneous rocks in the southern part of the Welsh Basin. Geological Magazine, 129, 615-624. BEVINS, R.E., PHARAOH, T.C., COPE, J.C.W. and BREWER, T-.S. 1995. Geochemical character of Neoproterozoic volcanic rocks in southwest Wales. Geological Magazine, 132, 339-349. BEVINS, R.E. and ROACH, R.A. 1982. Ordovician igneous activity in south-west Dyfed. In: Bassett, M.G. (ed.) Geological excursions in Dyfed, south-west Wales. National Museum of Wales, Cardiff. 65-80. BEVINS, R.E. and ROWBOTHAM. G. 1983. Low- grade metamorphism within the Welsh sector of the paratectonic Caledonides. Geological Journal, 18 141-167. BRITISH GEOLOGICAL SURVEY. Rock Classification Scheme. http://www.bgs.ac.uk/bgsrcs/ CLOUGH, T.H.McK and CUMMINS, W.A. (eds), 1988. Stone Axe Studies, Volume 2. CBA Research Report No. 67 London: Council for British Archaeology. COPE, J.C.W. and BEVINS, R.E. 1993. The stratigraphy and setting of the Precambrian rocks of the Llangynog Inlier, Dyfed, South Wales. Geological Magazine, 130, 101-111. CUNNINGTON, W. 1884. Stonehenge notes: the fragments. WANHM 21, 141-49 DAVID, A. and WILLIAMS, G. 1995 Stone Axe-head Manufacture: New evidence from the Preseli Hills, West Wales. Proceedings of the Prehistoric Society, 61, 433-460 DARVILL, T. 2006. Stonehenge. The biography of a landscape. Stroud. Tempus Publishing. DARVILL, T. and WAINWRIGHT, G. 2009. Stonehenge excavations 2008. The Antiquaries Fournal. 89, 1 — 19. DEER, W.A., HOWIE, R.A. and ZUSSMAN, J. 1992. An Introduction to the Rock-forming Minerals, Second Edition. Harlow, Essex: Longman Scientific and Technical. HOWELLS, M.E, REEDMAN, A.J. and CAMPBELL, S.D.G. 1991. Ordovician (Caradoc) Marginal Basin Volcanism in Snowdonia. (North-West Wales). London HMSO. British Geological Survey. IXER, R. A. 1990. Atlas of Opaque and Ore Minerals in their Associations. Milton Keynes, Open University Press. IXER, R.A. 1994. ‘Does ore petrography have a practical role in the finger-printing of rocks?’ in N. Ashton and A. David (eds), Stories in Stone. Proceedings 10th Anniversary Conference. Oxford 1993. Lithic Studies Occasional Paper 4. 10-23. London: Lithic Studies Society. IXER, R.A., BEVINS, R.E. and DOVER, M. (in press) The identification and spatial distribution of lithic fragments excavated from test pits in the Cursus field near Stonehenge, Wiltshire. Lithics. IXER, R.A. and TURNER, P. 2006. A detailed re- examination of the petrography of the Altar Stone THE LITHICS FROM THE CURSUS FIELD, STONEHENGE 15 and other non-sarsen sandstones from Stonehenge as a guide to their provenance. WANHM 99, 1-9 IXER, R.A., WILLIAMS-THORPE, O., BEVINS, R.E. and CHAMBERS A.C. 2004. ‘A comparison between ‘total petrography’ and geochemistry using portable X-ray fluorescence as provenancing tools for some Midlands axeheads.’ in E.A.Walker, EWenban-Smith and FHealy (eds). Lithics in Action. Lithics Studies Society Occasional Paper 8, 105-115. Oxford: Oxbow Books JONES, M.C and WILLIAMS-THORPE, O. 2001. An illustration of the use of an atypicality index in provenancing British stone axes. Archaeometry 43, 1-18 KERR, PE 1959. Optical Mineralogy. Third Edition. New York, London Mcgraw-Hill. KEILLER, A., PIGGOTT, S and WALLIS, ES. 1941. First report of the sub-committee of the south-western group of museums and art galleries on the petrological determination of stone axes. Proceedings of the Prehistoric Society, 7, 50- 72 KOKELAAR, B.P., HOWELLS, M.F, BEVINS, R.E., ROACH, R.A. & DUNKLEY, P.N. 1984. The Ordovician Marginal Basin of Wales. In: Kokelaar, B.P & Howells, M.F (eds) Marginal Basin Geology: Volcanic and associated sedimentary and tectonic processes in modern and ancient marginal basins. Geological Society of London Special Publication, 16, 245-269. KOKELAAR, B.P, BEVINS, R.E. & ROACH, R.A. 1985. Submarine silicic volcanism and associated” sedimentary and tectonic processes: a case study on Ramsey Island, S.W. Wales. Fournal of the Geological Society of London, 142, 591-613. LOWMAN, R.D.W. & BLOXAM, T.W. 1981. The petrology of the Lower Palaeozoic Fishguard Volcanic Group and associated rocks E of Fishguard, N Pembrokeshire (Dyfed). Fournal of the Geological Society of London, 138, 47-68. MACKENZIE, W.S., DONALDSON, C.H. and GUILFORD, C. 1982. Atlas of igneous rocks and their textures. Harlow Longman. PITTS, M.W. 1982. On the road to Stonehenge: reports on investigations beside the A344 in 1968, 1979 and 1980. Proceedings of the Prehistoric Society 48, 75-132. STONE, J.ES. 1948. The Stonehenge cursus and its affinities, Archaeological Fournal, 104, 7 -19. STONE, J.ES and WALLIS, ES. 1951.Third report of the sub-committee of the south-western group of museums and art galleries on the petrological determination of stone axes, Proceedings of the Prehistoric Society, 17, 99-158. THOMAS, H.H. 1923. The source of the stones of Stonehenge, The Antiquaries Fournal, 3, 239 — 60. THOMAS, J.. MARSHALL, P, PARKER-PEARSON, M., POLLARD, J., RICHARDS, C., TILLEY, C. and WELHAM, K. 2009. The date of the Greater Stonehenge Cursus. Antiquity 83, 40 —-53. THORPE, R.S., LEAT, PT., BEVINS, R.E. & HUGHES, D.J. 1989. Late-orogenic alkaline/subalkaline Silurian volcanism of the Skomer Volcanic Group in the Caledonides of South Wales. Fournal of the Geological Society of London, 146, 125-132. THORPE, R.S., WILLIAMS-THORPE, O., JENKINS, D.G. and WATSON, J.S. with contributions by R.A. IXER and R.G. THOMAS. 1991. The Geological Sources and Transport of the Bluestones of Stonehenge, Wiltshire, UK. Proceedings of the Prehistoric Society. 57, 103-157. Wiltshire Archaeological & Natural History Magazine, vol. 103 (2010), pp. 16-34 The Breamore jadeitite axehead and other Neolithic axeheads of Alpine rock from central southern England by Alison Sheridan', David Field’, Yvan Pailler’, Pierre Pétrequin’, Michel Errera> and Serge Cassen® This article describes research recently carried out on an axehead of Alpine rock found at Marsh Farm, Breamore, and curated by the Wiltshire Heritage Museum. Investigation by one of us has clarified its findspot location, while typological and non-destructive analysis, as part of an international research project, ‘Programme JADE’, has pinpointed the source of the raw material and clarified the axehead’s place among the thousands of Alpine axeheads known from Europe. The article also discusses other Alpine axeheads — and items previously claimed to have such an origin — from central southern England, including the example from the Stonehenge area, housed 1n the Salisbury and South Wiltshire Museum, that 1s now used as a type artefact, the ‘Durrington type’. Introduction Displayed prominently in the Society’s museum is an extremely fine Neolithic jadeitite’ axehead said to have been found just 2km across the county border at Breamore in Hampshire (Figures 1-2). It was acquired when J. W. Brooke sold the Wiltshire components of his collection of artefacts to the Society in 1916. The collection spanned the years between 1880 and 1910 and, according to Brooke’s notebook, he was anxious to find a home 1 The term ‘jadeitite’ is used to refer to a type of rock composed mainly or entirely of the mineral jadeite. The terms ‘jadeite’ and ‘jade’ had previously been used by many writers to refer to this rock. for it, his ‘museum being too extensive...I offered to let the Wilts Arch Soc have all my Wiltshire specimens’. Fortunately the Society’s appeal to members for sufficient funds for the purchase was successful (Robinson 2003, 81, 84). The piece is now accessioned as number DZSWS:DM.1414. As with many early finds the exact provenance and circumstances surrounding the discovery remain obscure. According to his obituary in the Marlborough Times of 12 Feb 1954 (Anon 1954, 300), Brooke was a native of north Wiltshire, born at Burbage on 18 September 1865 and attending Sydney College, Bath, before taking employment as a surveyor and architect for the District Highways Department in both Hungerford and Marlborough. His interest in antiquities was principally as an avid ' Archaeology Department, National Museums Scotland, Chambers Street, Edinburgh EH1 1JF; a.sheridan@nms.ac.uk, * English Heritage, Swindon SW2 2YP; David.Field@english-heritage.org.uk, * 9 rue Bailly, F29200 Brest, France; yvan.pailler@club-internet.fr, ‘69 Grande Rue, F70100 Gray, France; archeo.petrequin@free.fr, > Royal Museum for Central Africa, Leuvensesteenweg, 13 B-3080 Tervuren; michel.errera@africamuseum.be, ° Laboratoire de recherches archéologiques (UMR 6566) Université de Nantes, BP 81227, F44312 Nantes cedex 3, France; serge.cassen@univ- nantes.fr (All authors except Field are Programme JADE team member) THE BREAMORE JADEITITE AXEHEAD 17 Figs. 1 (left) and 2 (right) The Breamore axehead. Drawing: Marion O’Neil. Photograph: NMS collector, particularly of coins, and he was a Fellow of the Numismatic Society. His artefacts often exceeded the space at his disposal, initially being housed at Blake House in Kingsbury Street (Evening Advertiser Friday 27 March 1936), but later moved to ‘Rosslyn’ in the London Road where he extended his house in order to cater for the growing collection. Little is known of Brooke’s early collecting, or whether he visited Breamore, but in response to a reference in the Wiltshire Archaeological Magazine (Anon 1926, 336), Brooke wrote to the Society on 10 August 1927 to confirm how he came by the axehead. ‘It was found at Marsh Farm, Breamore, by Mrs Jeans, the mother of the late Mark Feans. She used it as a paper knife but as it tore more leaves than it cut she threw it out of the window and it fell on a stone and the end was chipped. Mrs Feans told me about it, having seen one of my polished flint celts from Aldbourne. I implored her to find it, which she did, and Mark Jeans gave it to me’ (Anon 1930, 88). Unfortunately, the description provides no detail of the circumstances of the find but rather of how Brooke obtained it and, as Goddard (1921, 371) pointed out, the details are not quite as clear as could be wished. According to an obituary notice of Mark Jeans published in the Wiltshire Gazette of March 6th and 13th 1924 (Anon 1924, 509), Brooke’s daughter Elizabeth had married Mark Jeans, son of Jacob of Breamore, and an auctioneer and land surveyor. Mark was resident in North Wiltshire and Mayor of Marlborough in 1891-2, and he will have known Brooke as a result of the acquaintance with his mother. It was Mark who was responsible for disposing of the rest of Brooke’s collection on Thursday May 17th, 1917 (Wiltshire Archaeological and Natural History Society Library - Wiltshire Tracts 133). However, the Jeans family were certainly resident at Cooke’s Farm, Breamore, as Jacob W. Jeans was farming there in 1867 (Kelly 1867) through to at least 1880 (Anon 1880). Cooke’s Farm is not named on early editions of the Ordnance Survey, but it could have been an earlier name for Marsh Farm, particularly as Marsh Farm is not mentioned in the 18 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE’ directories of 1880 or 1867, but does occur in 1903 when Edward Dorrington Stanford was farming there (Anon 1903). In contrast, Marsh Farm is depicted on the Ist edition OS 25” (1843-93) map where it is shown as being located on the edge of Breamore Marsh with what appears to be a cutting at a little distance from the farm buildings and other cuttings in the marsh to the north. A series of cuttings and drains can be identified at the site today and although it cannot be demonstrated, it is conceivable that the axehead was found in one of these. The earliest record of the piece was made by John Evans (1897, 107), who recorded its presence in his seminal Ancient Stone Implements of Great Britain, writing that ‘Mr J W Brooke has a beautifully polished specimen, made of a green transparent stone, from Breamore, Salisbury. It has lost a small piece at the butt end, but is still 8 inches long. It is only 2 5/8 inches broad at the cutting end’. It was subsequently brought to attention by O. G. S. Crawford in the French journal LAnthropologie, in an article concerned with trade routes between Britain and France (1913). On that occasion Brooke provided Crawford with a drawing of the artefact, which was published in the article and from which it is clear that the chip at the tip mentioned above had already occurred. Crawford simply described the axehead as highly polished and of a ‘green stone’ that he suggested had a ‘foreign origin’. In his autobiography, Crawford, who had attended Marlborough College between 1900 and 1904, later described Brooke as ‘a rather crazy and disreputable collector’ and suggested that he ‘did not appreciate the importance of recording the exact sites of his finds: many of them were obtained from road workmen and others and the sites of many are suspect’ (Crawford 1955, 27-8). Some indeed came from road workmen, shepherds and those picking stones from fields et cetera, but Brooke’s notebooks demonstrate that, contrary to Crawford’s allegation, he generally recorded the location and circumstances surrounding the discovery of artefacts, along with the name and occupation of the finder. In the case of the Breamore axehead the family association, coupled with Mrs Jeans’ description, leave little room for doubt. The axehead was subsequently listed by Goddard (1921, 371) and published as being of ‘jade’ on a map of such pieces in Britain prepared by Lily Chitty for Cyril Fox (1933). Its significance was also recognised by Stuart Piggott who, in 1949, referred to it as comparative material for a jadeitite axehead fragment found in a chamber tomb at Cairnholy in south-west Scotland (Piggott and Powell 1949, appendix A). Having escaped attention during two previous phases of petrological work by the South West Implement Petrology Group of the South Western Federation of Museums and Art Galleries, it was finally seen and listed (as no. 300, mis-spelt ‘Braemore’) for their third report (Stone and Wallis 1951, 146) and as a result it was analysed and listed by Campbell Smith (1963, 160; no. 20) during his investigation of ‘jade’ axeheads in Britain. Supplementary reports followed which updated Campbell Smith’s initial catalogue of Alpine axeheads known in Britain and Ireland (Smith 1965; 1972; Jones et al. 1977). The Breamore axehead, along with the other Alpine axeheads from Britain and Ireland, has recently been the subject of fresh investigations and analysis as part of a major international research project, Programme JADE (Sheridan 2007a; Pétrequin et al. 2008a; in press); the results for Breamore and for other axeheads found in Hampshire and Wiltshire that are, or have been claimed to be, of Alpine rock are presented below. Breamore: the findspot location Breamore lies along the middle reaches of the River Avon, some 27km above Christchurch and 12km below Salisbury. The floodplain at this point is 1km wide and much given over to water meadows, while a first terrace some 1km wide abuts the west flank. Recent air photography has demonstrated the presence ofa series of ring ditches, which are probably levelled round barrows (M. Barber and D. Grady pers. comm.), while a high density of struck flints found on the terrace a little over 1km to the north was thought to represent a settlement site (Schofield 1986). Situated on Breamore Down, overlooking the valley, lies the Giant’s Grave, a Neolithic long barrow partially destroyed by quarrying (RCHM 1979, 10-11). While the precise location of the axehead find is unknown, Marsh Farm itself is situated at NGR SU 154177 on the edge of the floodplain and in, as the name implies, an area of former marsh land. The late Norman Quinnell, then Ordnance Survey Archaeological Field Investigator, could obtain no further information when he inspected the site for the Ordnance Survey Archaeology Division in 1959 (NMR No. SU 11 NE 15). THE BREAMORE JADEITITE AXEHEAD The axehead The axehead (Figures 1—2) is of slender form, just 14mm in maximum thickness, gently tapering from 67mm wide at the cutting edge to 36mm at the break and reaching 206mm in length. The weight is 358 gm. It is very slightly asymmetrical when viewed in profile, one side being ground flatter than the other from the midpoint towards the blade edge. Each side is straight, both in plan view and in profile, and each has been very slightly rounded to leave no sharp edges. The piece is of a pale mid-green colour (Campbell Smith described it as ‘Russian green’), darker towards the cutting edge; at its thinnest, it is translucent. Dark veins can be traced towards the cutting edge though these are paler in the centre of the face and towards the butt. A little dark brown staining occurs at the butt. The break at the butt is evidently ancient and it is not clear how it has occurred; by analogy with many other Alpine axeheads, it may have been deliberately caused. No obvious point of percussion is visible although there is some modern damage that could obscure the relevant detail. The broken surface is ‘dappled’ and part of its edge is slightly rounded, as if subject to post-breakage abrasion and stained amber/brown. A large chip, 24mm by 14mm, is missing from the blade and to judge from its unpatinated and unstained condition, this damage could have been caused when Mrs Jeans threw the axehead out of the window (see above). There is also some abrasion to part of the central face, and this too may represent relatively recent damage. Any striations visible are all post-depositional and are not part of the grinding process. Some striations at the cutting edge have dulled the gloss. The axehead has been intensively burnished leaving a very highly polished, glossy and reflective surface that has all but obliterated any evidence of grinding. A thin section has been cut into one edge for a distance of 8mm, evidently by the Implement Petrology Sub-Committee of the South Western Federation of Museums and Art Galleries (henceforth IPSC) for their third report (Stone and Wallis 1951), and the piece is labelled with the IPSC serial number, an encircled ‘300’ in ink close to the break. Given uncertainty over the precise form of the butt it is difficult to estimate the original length of the piece. However, since the axehead had probably originally been of ‘Bégude’ type (to use a Programme JADE term, explained below), it is likely that the 19 butt continued in a steady taper. When complete, the axehead could have reached 300mm, an impressive length for an example of its type. If it had been complete when in use in England, this would have constituted one of the longest known Alpine axeheads in Britain and Ireland. Previous analysis The thin section slide that had been prepared by the IPSC of the South-Western Group of Museums and Art Galleries was examined by Campbell Smith, who reported that the stone ‘seems to have a fibrous texture... The little prisms of pyroxene are very slender, 0.025 to 0.05 millimetres across, and they tend to form a felted mass, doubtless giving great tenacity to the stone’ (Smith 1963, 144). He added that ‘This, and the Canterbury axe (see no. 35) are in the best quality of any of the jade axes found in England’ (ibid., 160). Campbell Smith measured the specific gravity (3.33) and refractive index (1.667/1.680), and also undertook X-ray powder diffraction analysis, a technique described in his report, to characterise the material and to distinguish between jadeite and nephrite (calc-amphibole), the two minerals encompassed within the general heading of ‘jade’. These techniques, allied with the results of the petrological thin sectioning, allowed him to conclude that jadeite was the predominant or main mineral in the rock; in other words, to use the current term, it is of jadeitite. Close matches were found with thin sections of axeheads from Carnac (75493 H.5.Coll. Bertic), France, and another jadeitite axehead said to have been found at Spiennes, Belgium. Source of the jadeitite: traditional ideas and new fieldwork It had long been thought, by some, that the axeheads found in Brittany had a local source. Campbell Smith (1963, 154) discussed the claims in detail but rejected them. Similarly, the discovery of boulders of an impure translucent green rock (not a jadeitite) found on the shore of Lake Neuchatel were considered and rejected as a source of the Brittany axeheads. However, a source in Piedmont, of which a pale 20 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE green version was provenanced to Mont Viso, in the North Italian Alps south west of Turin, appeared to be more promising (Damour 1881) and Bishop et al. (1977, 1) considered this the likeliest source as workshops producing axeheads of Alpine rock had been identified in the Piedmont, but they were nevertheless cautious as other sources of jadeitite were known in the north and north-eastern parts of the Alps. Recent work by Anne-Marie and Pierre Pétrequin, latterly as part of Programme JADE, has led to the discovery of sites in the Mont Viso massif where jadeitite, eclogite and omphacitite were extracted and worked — in accordance with Damour’s prediction — at high altitude, between 1800 and 2400 metres above sea level. Spectroradiometric analysis by Michel Errera of raw material samples and debitage from these sites, and also from other sites around Mont Beigua near Genoa, has demonstrated that most of the Alpine axeheads found across Europe came from these areas. Analysis and precise sourcing of the Breamore axehead by Programme FADE The Breamore axehead has recently been re- analysed by Michel Errera of Programme FADE, using a wholly non-destructive technique known as scattered reflectance spectroradiometry. This technique, borrowed from the field of remote sensing (e.g. as used in the planet Mars probe), measures the light reflected by a target — in this case, the axehead — and compares it with the light reflected by a pure white reference material. The ratio, or reflectance, is then compared — using standard spectroradiometric curve-fitting software — with a database of comparative material which, in this case, comprises several thousand analyses of Alpine raw material samples, working debris and over 1700 Alpine axeheads from across Europe. (For details of the technique and its application to Alpine axeheads, see Sheridan 2007a, 22—27: Pétrequin et al. 2006a (in English); Pétrequin and Pétrequin 2004: Pétrequin et al. 2003; 2004; 2005a; 2005b; 2006a; 2006b; 2006c; 2008a; 2008b; in press; Errera 2003; 1995; Errera et al. 2006). The specific gravity of the axehead was also re-measured and the result (3.33) matched that obtained by Campbell Smith. The result of the spectroradiometric analysis (analysis nos. SCTL_025 and SCTL_026) confirmed that the material is indeed Alpine jadeitite and indicated that the most likely source location is a free-standing block near Pontinvrea/Giovo Ligure, Liguria, in the Mont Beigua massif near Genoa (Pétrequin et al. 2005a, 2005b). This block was spotted during the 2004 systematic prospection of the area by Pierre and Anne-Marie Pétrequin. In the valley that leads down from Bric Tamburo (at 773m above sea level) towards Pontinvrea, such blocks can be found along the bank of the stream called Ca’ de Parei, near La Fornace, between 600m and 520m in altitude. The largest of these blocks is 2 x 1 x 1.5m (Figure 3) and is probably in its original position, surrounded (as many such blocks are) by a sludge of decayed serpentinite. This enormous fragment of what geologists term a ‘boudin’ of jadeitite (so named because of its resemblance to a blocd pudding) shows lithological variation from the surface to the interior of the block, as follows: e On the surface, it is a fine-grained, bluish rock. In thin section (Programme JADE reference: LM 138), the rock is very fine-grained with blue amphibole and traces of sphene. Whitish zones may correspond to a colourless amphibole or to phyllosilicates. No interstitial quartz was spotted. Analysis using X-ray diffraction (XRD, undertaken at the GeaDue laboratory, Bologna) showed that glaucophane, chlorite, feldspar, pumpellyite and lawsonite are well represented even if jadeite was not noted. e Beneath the surface the rock appears macroscopically to be a pale green, medium- grained jadeitite, with small white and mid-green grains. In thin section (LM 139) it is a coarse- grained quartz-jadeitite. The pyroxenes tend to cluster or to occur in elongated prisms. Quartz comprises 40-50% of the rock, the rest consisting of clinopyroxenes. Some pyroxene crystals intrude into the quartz. The clinopyroxenes show a localised and very limited retromorphism. Associated mineral: sphene. Analysis using XRD reveals quartz, jadeite and feldspar; Italian geologist Claudio D’Amico interprets this as a quartzo-albitic jadeitite. e A third sample taken from the heart of the block where the jadeitite is of the highest quality (in terms of its potential use for making axeheads) shows some characteristics that are identical with the aforementioned sub-surface material. Macroscopically it is a lightly saccharoidal jadeitite, with marked oblique bedding, of pale green colour with several dark green speckles. The thin section (LM 140) shows a rock identical 21 Fig. 4. The La Fornace jadeitite has narrow natural fissures owing to the inclusion of diaclases. Fire-setting against the rock allows the extraction of long blocks which can then be shaped by hammering. Photo: PR Pétrequin THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE bo bo Données : JADE (base juin 2009, P. Pétrequin. dir.) Haches de type Bégude repoli ke Carrigres Fond : Esri WBM. SRTM ou a poli camacéen : Nb/commune CAO : J. Desmeulles et E. Gauthier — Université de Franche-Comté, UMR CNRS 6249 pol oa Aoit 2009 e | @ 5 @ Fig. 5 Distribution of repolished axeheads of Bégude type and of axeheads with Carnac-type polish. Based on Programme JADE results as of fune 2009, collated by P Pétrequin; cartography by F. Desmeulles and E. Gauthier, Université de Franche-Comté. to LM 139, with a little secondary colourless marked grain which is lightly ‘stretched’ obliquely amphibole which appears on the pyroxenes. with respect to its fracture surfaces (Figure 4). It A slight retromorphism can be observed, a is capable of being exploited by fire-setting, to little more marked than in LM 139. The same detach elongated blocks which come away along minerals (quartz, jadeite and feldspar) were these natural fracture planes. This block provides identified through XRD. such a close match for the Breamore axehead that it may well have been the parent material. Even more This block is thus a distinctive jadeitite, with a well remarkably, the same block may well have been used THE BREAMORE JADEITITE AXEHEAD to make the exquisite axehead from Canterbury to which Campbell Smith referred. This axehead is of slightly different shape from the Breamore example and may have been made at a different time, which suggests that people — most probably specialist producers of special axeheads — returned to the same block at different times over several centuries (or else successively exploited different fragments taken from the same block). Even more remarkably, a near- perfect match also exists with a third axehead, from Battignies (Binche) in Belgium, currently in the Musée de Mariemont. The phenomenon of matching axeheads to parent blocks has also been possible in the case of four jadeitite axeheads, one from Fife in eastern Scotland and three from northern Germany (Pétrequin et al. 2008b), and further examples are expected as the processing of the spectroradiometric results for the whole of Europe progresses. It should, however, be noted that not a single working flake or roughout has been found in or below the Fornace valley, even though the jadeitite blocks from this area constitute the best candidates for the Breamore and Canterbury axeheads. This could be because the mountain-torrent erosion in this Mediterranean area, on the steep slopes, is not favourable to the preservation of traces of Neolithic exploitation (unlike at some other areas around Mont Viso, where plentiful traces of extraction and initial roughing-out have been found: Pétrequin et al. 2005b). One is most likely to recover traces of working further downslope, on the ancient terraces and in recent alluvial deposits: by analogy with the known sites elsewhere, these should produce evidence for extraction by fire-setting and initial roughing-out by hammering, the two techniques best adapted to the working of this more or less grainy type of jadeitite (which is particularly hard to flake by direct hard hammer percussion, as borne out by personal experience). Nevertheless, in the collections of the Museum of Ligurian Archaeology at Genoa, there are several roughouts for jadeitite axeheads of Bégude type — the type from which the Breamore axehead had been created — and these probably come from the Pontinvrea source. At least three of these come from Sassello, and another from Giusvala, just a few kilometres from the Pontinvrea/Giovo Ligure jadeitite blocks. These roughouts offer direct proof of the local exploitation of the jadeitite, even if the source areas themselves have not yet produced working debris. Finally, it should be noted that these jadeitite blocks are situated far upstream in the Erro valley, 23 in the heart of the Mont Beigua massif. In this mountainous terrain with its extremes of climate, traces of Neolithic occupation are few and scanty, except in the Sassello basin and the Bormida valley to the west. It seems most likely that most of the people who exploited the jadeitite (and other rocks) of the Mont Beigua massif lived in villages lying to the south, facing the Mediterranean coast (which would have been a day’s march away from the source area) or further to the north, among the rich agricultural lands of the east bank of the Tanaro river, close to Alba (two days’ march away: Traverso 1898; 1901; 1909; Pétrequin et al. 2005b). In these conditions, the jadeitite of Pontinvrea/Giovo Ligure represented, for Neolithic people, a rare raw material which could only be exploited in difficult conditions. Indeed, it is possible that the value of jadeitite axeheads — of which certain examples, including Breamore, have been found over a thousand kilometres from the Alps — was largely based on the rarity of the primary and secondary sources of this material in the mountains of the Beigua massif; mountains which, on a good day, may be seen from afar but whose summits may disappear during storms or after snowfall. Typology The Breamore axehead has been variously described over the last 50 years. In 1963, Campbell Smith devised a tripartite classification for axeheads of Alpine rock, with Types I (thin and triangular), II (plump) and III (short and squat) each being subdivided further. Class I axeheads were triangular in plan with a thin lenticular cross-section and more or less pointed butt and it was within this class, sub-class Ia (with acute-angled side edges) that Campbell Smith placed the Breamore axehead. He assumed that the butt was chamfered and that the axehead formerly measured about 260mm in length. He compared it to an exceptionally long and narrow axehead that had been found embedded in clay almost a metre below the surface at Lyme Park (Lyme Handley) in Cheshire (Anon 1932, 167) that measured 270mm from butt to blade edge and 71mm in maximum width. He also suggested that there was a superficial resemblance to an axehead from St Jean, Chateaudun, Eure-et-Loire, in the Ashmolean Museum (Reg no. ASH 1927.5804). Work by Clive Bishop et al. in the 1970s refined this typology and distinguished between short thick functional and long thin ceremonial 24 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE axeheads (Woolley et al. 1979). Like Campbell Smith they assessed the shape in terms of length, width and thickness, but illustrated this graphically by comparing the ratio of length to width (L/W) against the ratio to thickness (W/T) and visually contrasting the outline of the plan shape with the cross-section. The thin axeheads were generally triangular in outline and those with higher W/T ratios noted as being composed of a ‘distinctive coarse pale green jadeite’. Long, thin ‘torpedo’ types known from Brittany were then not thought to be represented in Britain although it should be noted that the Breamore example approaches this form (see below). With a width more than three times the thickness, the Breamore example is of the long and thin, ceremonial type. More recently, Programme JADE has developed a materio-typo-chronology pertaining to all Alpine axeheads in Europe (the main features of which are presented, for example, in Pétrequin et al 2008a; in press; and Sheridan 2007a, 26). Within this scheme, axehead types are named after the first axehead that was used to characterise the type. The Breamore axehead has been classified as being of Bernon type —that is, along Alpine axehead, originally of Bégude type, that has been thinned down and repolished, almost certainly in the Morbihan area of Brittany. This type of axehead, which had been transformed from a thick-sectioned to a thin-sectioned object and sometimes polished to a mirror-like sheen, is well known in the Gulf of Morbihan, where magnificent examples have been discovered in association with stone bangles, themselves made from Alpine rocks (Cassen et al. 1999). This transformation of axeheads in the Morbihan around and just before 4500 BC — which represents a significant investment of labour and skill — is believed to have been a strategy used by the local leaders, in order to create objects that were distinctively Morbihannais and hard to imitate, thereby confirming and enhancing these individuals’ power (Pétrequin et al 2008a; in press). Only a handful of ‘Carnac style’ Alpine axeheads are known in Britain, with or without the mirror- like polish (Figure 5); significantly, the Canterbury specimen is one of these (along with a fragment of a Tumiac-type axehead, with aborted perforation, from High Peak enclosure, Devon; the aforementioned example from Lyme Park, Cheshire; and the famous axehead found beside the Sweet Track in Somerset). The chronology of their appearance in Britain will be discussed in the next section. Date \ Since so little is known about the findspot of the | Breamore axehead, nothing can be gleaned about its | date of deposition. Discoveries of Alpine axeheads | from dated (or datable) contexts are very rare, not | only in Britain but across Europe; that said, the | | best-dated of all Alpine axeheads in the whole of | Europe is the slender example, of jadeitite, found | beside the Sweet Track in the Somerset Levels (Coles et al. 1974). The trackway’s construction is _ dated dendrochronologically to 3807/3806 BC and it | seems to have been used over a short period, perhaps a dozen years (Hillam et al. 1990, 218). The axehead || (which, as noted above, is of a type that had been | reshaped in the Morbihan) was found along with | an axehead of mined flint, probably from a Sussex _ mine, in mint condition, and also with several carinated bowls. To judge from what is known of the | typochronology of Alpine axeheads at a European scale, the Sweet Track alpine axehead must have been | several centuries old when it was deposited. The only other British axehead whose deposition date can be estimated with a degree of confidence | is the fragment found in a simple chamber tomb | at Cairnholy (I) in the Stewartry of Kirkcudbright, | south-west Scotland (Piggott and Powell 1949). Given the style of this monument and the presence of a bowl of the Carinated Bowl tradition found in the forecourt, it has been argued elsewhere (Sheridan | 2007b) that this axehead fragment was probably deposited around the same time as the Sweet Track | specimen, that isc. 3800 BC. Three Alpine axeheads | have been found in causewayed enclosures: the | aforementioned fragment from High Peak, Devon | (Pollard 1966, 51; Smith 1963, 158, no. 12) and two | from Hambledon Hill, Dorset (Smith 2008, 630, fig. 10.1 $3, S11; note that the jade-nephrite example | S3, in private hands, cannot now be traced and its Alpine origin double checked). While the enclosures themselves have been dated, with the initial enclosure construction at Hambledon Hill having taken place during the 37th century BC (Bayliss et al. 2008, table 4.2), for example, unfortunately the axeheads were all stray finds and may not have been deposited during the use of the enclosures. Even though we cannot date the deposition of the Breamore axehead, Programme FJADE’s review of all the available sources of chronological information regarding Alpine axeheads in Europe has enabled its likely date of initial manufacture to be estimated. Cassen and Pétrequin’s research into THE BREAMORE JADEITITE AXEHEAD Breton examples (1999) has shown that the Breamore axehead is identical to those that were already in use in the Morbihan region when the giant Carnac tumuli, such as the tumulus Saint-Michel at Carnac, were built around 4500 BC (Pétrequin et al. 2002). The association of these axeheads with stone ring- bangles, as at Saint-Pierre-Quiberon, is furthermore characteristic of material dating to the first half of the fifth millennium (Herbaut et al. 2000). As with the stone bangles, it seems most unlikely that axeheads of Bégude type and their ‘reincarnations’ as thinned and super-polished Carnac-style axeheads continued to be produced much after 4500 BC. This raises the question of whether the Breamore axehead could have crossed the Channel during the second half of the 5th millennium or whether, like the Sweet Track example, it was centuries old when (imported and) deposited. There is no evidence for contact between the late Mesolithic communities of southern England and their cross-Channel neighbours (most of whom were already farming during the second half of the 5th millennium) and it seems more likely, at least to some of the authors, that the Breamore axehead was brought over from Normandy or northern Brittany, as part of the process of the Neolithisation of south-west Fig. 6 Axeheads and other objects of Alpine rock in southern central England. Key: 1-2 Devon: 1. Brownstone Farm, Kingswear (chisel; a second, possibly from Devon, 1s not marked); 2. High Peak, Sidmouth; 3—5 Somerset: 3. Quantoxhead; 4.Sweet Track, Somerset Levels; S. St. Cuthbert Out, Ebbor Gorge; 6-14 Dorset: 6. Wootton Fitzpaine (axehead pendant); 7. Netherbury; 8. Nether Cerne; 9. Dorchester; 10. Preston (axehead pendant); 11. Hambledon Hill (2 axeheads, one not available for verification by Programme JADE); 12. Newton Peverill; 13. Parkstone; 14, Handley Common; 15-18 Wiltshire: 15. Beckhampton; 16. A barrow near Stonehenge’; 17. Winterslow; 18. Teffont Magna; 19-22 Hampshire: 19. Breamore; 20. ‘Barton’; 21. near Beaulieu; 22. Titchfield (plus possible example from Hillhead). Cartography: Alison Sheridan, using Programme JADE results as of October 2009. ZS England, by an immigrant farming community just after the beginning of the 4th millennium (Sheridan et al. 2008. See also French et al. 2007, 282 and fig. A4.4.3, and Whittle 2007, fig. 4, on the radiocarbon dates from the Fir Tree Field shaft on Cranborne Chase that indicate that the Mesolithic—Neolithic transition at this site occurred some time between 4120 and 3870 cal BC, with a pointed-butt ground flint axehead being found among the Early Neolithic material.) It is known that Carnac-style axeheads, along with certain distinctly Morbihannais beliefs and practices, were circulating out of the Morbihan in various directions during the second half of the fifth millennium, as the distribution map (Figure 5) demonstrates (Pétrequin et al. 2006a; in press; Cassen 2009); furthermore, copies of Carnac-style axeheads that date to this period are known from north-west Iberia (Pétrequin et al. in press). It is therefore possible that examples had been circulating around north-west France for several centuries by the time the Breamore axehead (and the other Carnac-style axeheads in Britain) was carried across the Channel as a precious heirloom. Other Alpine axeheads in central southern England The Breamore axehead forms part of a sizeable cluster of Alpine axeheads (and other artefacts) from central southern England that extends westwards along and near the coast, from the mouth of the Wessex river systems that lead into the Solent as far as Kingswear in Devon, and inland to northern Wiltshire (Figure 6), with a further small cluster in northern Somerset. A range of types is present but, as Bishop et al. (1977, 5) pointed out, a larger number of hachettes and plump axeheads occur in the south than elsewhere. This section will briefly describe these objects and give their latest material identification; it will also deal with specimens that had, at some time in the past, been published as, or considered to be, Alpine axeheads but which can be dismissed. It is not intended to present a detailed catalogue of all specimens from south central England, since a comprehensive catalogue covering the whole of Britain and Ireland is planned for publication elsewhere. Attention will instead focus on examples from Wiltshire and Hampshire, whose findspots and find circumstances have been studied in detail by one of us. 26 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Wiltshire Four definite Alpine axeheads are known from Wiltshire (Figure 7) and at least two claimed examples can be discounted. The first of the definite examples (Figure 7.1) was found near Lopcombe Corner, Winterslow, some 10km to the east of Salisbury, and is in the Salisbury and South Wiltshire Museum (Reg. no. SSWM 1958.100; Smith 1963, 164, no. 42). This is half the size of the Breamore axehead, measuring 123mm in length and 61mm in maximum width. It is triangular in plan with pointed butt and slightly curved sides, but Campbell Smith describes it as type IIa with the side edges slightly flattened. It is pale grey-green in colour with green flecks. It was found in ploughsoil by the donor in 1948, around 183 metres south of the Andover Road and about 830 metres from Lopcombe Corner at approximately NGR SU 257358. Campbell Smith suggested that there is a superficial resemblance to another from Eu, Seine Inférieure, in the Ashmolean Museum (Reg. no. ASH 1927,5801). According to the Programme JADE classification this is a Chelles type axehead; spectroradiometric analysis confirmed Campbell Smith’s suspicion that the rock is a jadeitite, probably from Mont Viso (analysis nos. BMus _ 161, BMus 162). The second (Figure 7.2; Smith 1963, no. 41), said to have derived from ‘a barrow near Stonehenge’ was, according to Thurnam (1871), one of the ‘most beautiful and elaborately polished stone celts known’. It certainly is striking, being composed of highly burnished Alpine rock (see below) and in an almond or teardrop shape with a sharply pointed butt. Thurnam noted ‘delicate golden veins’. It measures 177mm by 77mm wide and 28mm thick. A footnote in Thurnam’s paper adds that ‘It was formerly in the Leverian Museum, and is now in the possession of Rev W H Winwood of Bath’. (The Leverian Museum, in London’s Leicester Square, was established in 1773 and mostly displayed ethnographic material including much of what Captain James Cook had brought back to England during his travels. However, the collection was dispersed by auction in 1806 and the museum closed the following year.) The axehead was among those described by John Evans in his Ancient Stone Implements (1897) and of more recent finds is remarkably similar to one recovered from Monzievaird (Sheridan 1992) and is almost identical in form, although not in size, to that found in Kirkby Lonsdale, Cumbria (Jones et al. 1977, 289). Campbell Smith (1963, 164) indicated that the Winwood Collection was given to the Manchester Museum in 1921 and the axehead was then placed on permanent loan to the Salisbury and South Wiltshire Museum where it is accessioned as SSWM 28/59 (02919). It was initially ascribed to Durrington parish by Campbell Smith (Smith 1963, 164), although he corrected the error in a supplement to his first paper (Smith 1965, 30) and subsequently placed it in the parish of Amesbury. Despite the authoritative sounding attribution, with its tantalising potential of a sealed context perhaps comparable with the Cairnholy axehead, the origins of the piece unfortunately remain uncertain. However, there is no record of any other axehead being recovered from a barrow around Stonehenge and, despite their extensive barrow excavations in the area, Cunnington and Hoare (Hoare 1812) do not refer to such an artefact. Given the extreme rarity of finds of ground stone axeheads in long barrows here or indeed elsewhere in the country (Kinnes 1992), it is conceivable that this is the ‘celt’ found in the long barrow almost directly north of Stonehenge that was in the custody of Sir Andrew Fountaine — an antiquary (1676-1753) who succeeded Isaac Newton as Warden of the Mint in 1727 —at the time of Stukeley’s work in the Stonehenge area (Burl and Mortimer 2005, 29). Stukeley depicts this long barrow as lying about 1 mile north of the cursus (ibid. pl 1) and says of it that it is ‘remarkably large’. There are two long barrows north of the cursus. One, ~ the Durrington 24 long barrow (SU14SW23), lies just over a kilometre away, beside the road in front of Larkhill Camp. This is a very shallow mound reaching no more than Im in height and in long grass can be easily missed by observers. At just over 2km north of the cursus, however, the Knighton long barrow (Figheldean 27) (SU 14NW14) provides dramatic contrast. It lies in approximately the correct position and, at 3m in height, is at least twice as high as any other example in the Stonehenge environs. Stukeley’s ‘remarkably large’ is certainly an appropriate description for it. This axehead has been used by Programme JADE to define the so-called ‘Durrington type’ of Alpine axehead, and even though (in the light of the comments above) it would seem that ‘Figheldean’ is a more likely parish name for the findspot, the term ‘Durrington’ has stuck. Numerous axeheads in France, Italy, Britain and along the Rhine are of this type. Spectroradiometric analysis of the ‘eponymous’ axehead indicates that it is of eclogite, from the Alps (analysis nos. BMus_ 165, BMus_166). (Campbell Smith had described the stone as ‘chloromelanite’: Smith 1963, 164). Upon examination by two of the authors, it was found that this axehead shows signs THE BREAMORE JADEITITE AXEHEAD 3 Fig. 7 Axeheads of Alpine rock from Wiltshire: 1. near Lopcombe Corner, Winterslow; 2. probably Knighton long barrow, Figheldean (also known as ‘Durrington’); 3. Beckhampton; 4. Teffont Magna Down (axehead reused as metalworker’s tool). Drawings: Yvan Pailler. ; si 28 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE of having been made from a thermal flake and has a narrow band of peck-marks around two-thirds along the body, in the area where a haft would be placed (had the axehead been hafted). The third Wiltshire example (Figure 7.3), which has hitherto not featured in any publication concerning Alpine axeheads (and was described as being of ‘pyroxene granulite’ in Stone and Wallis 1951, 142 and in Clough and Cummins 1988, 156), is a fragment from a large flat triangular axehead of ‘Altenstadt’ type, found at Beckhampton and curated by the Alexander Keiller Museum, Avebury. Spectroradiometric analysis has confirmed that the material is jadeitite (analysis nos. BMus_ 146, BMus_ 147). The fourth item was found on Teffont Magna Down around 1900, and is housed in the Salisbury and South Wiltshire Museum (SSWM 71/1950); it has not featured in previous listings of Alpine axeheads. The findspot area was investigated by M. Fletcher for the Ordnance Survey Archaeological Division in April 1974, but the donor (details in Salisbury Museum) could not give a more accurate siting as he was given the object. Short and squat (58mm long by 46mm wide), and initially described by the Museum as an ‘unperforated Neolithic polished greenstone macehead’, the item is in fact a Puy-type Alpine axehead (to use the Programme JADE terminology) that has been subsequently reused as a metalworker’s tool; this explains its blunt, squared-off appearance. Spectroradiomentric analysis has confirmed that the material is dark green jadeitite, possibly from Voltri in the Mont Beigua massif (analysis nos. BMus_163, BMus 164). There are two axeheads that are recent ethnographic manuports rather than Neolithic Alpine axeheads. One of amphibolite, which had been part of Sir Richard Colt Hoare’s collection at Stourhead and had been believed to have been dug up at the entrance to Scratchbury Camp, Norton Bavant, by Cunnington (Stone and Wallis 1951, 289), was dismissed as probably Australian by Smith (1963, 171); Smith pointed out that the Stourhead collection was not confined to British antiquities. This axehead is in Wiltshire Heritage Museum (Reg. no. DZSWS:STHEAD.11). Programme JADE concur that it is an ethnographic object. The second, in the Natural History Museum (reference no. E 5887), was found at Broad Town (Smith 1963, 164, no. 40; Jones et al. 1977, 293, no. 5) and its material was identified as an unusually black nephrite by Campell Smith. Spectroradiometric analysis by Programme JADE (analysis nos. BMus_ 252, BMus_253 has confirmed that the material is amphibolite, and typologically the axehead is typical of New Caledonian lenticular- sectioned axeheads, so this would appear to be another relatively recent ethnographic manuport. Hampshire Three other Alpine axeheads (from ‘Barton’, near Beaulieu, and Titchfield), and a possible fourth (from Hillhead), are known from Hampshire in addition to the Breamore example (Figure 8.1—4). All are small and all are in the care of the Winchester City Museums Service (Reg. nos. WINCM:ARCH 2212-2215); three of them had previously been in the collection of W. Dale, and over the years there has been some confusion regarding findspots and about the number of axeheads concerned. One of the Dale collection axeheads (Figure 8.3; WINCM:ARCH 2212, Dale 79) was published by Dale as having come from ‘near the sea at Barton’ (Dale 1907, 265), but in a previous publication he had cited its findspot as ‘near Milford’ (Dale 1905, 185). Milford is about 4km from Barton and the assumption is that the find spot was somewhere between the two. The axehead is figured (Dale 1905, fig 4) as a small triangular hachette, evidently in good condition and a little over 2 1/2” (70mm) long. The Dale Collection was purchased by public subscription and deposited in Winchester City Museum from where the artefacts were loaned to Hampshire Museum Service; the ‘Barton’ axehead is currently on display at the Red House Museum, Christchurch. The axehead — one of two mentioned by O. G. S. Crawford in his LAnthropologie article (1913) — was accidentally listed twice by Campbell Smith in his 1963 inventory: once as no. 18, “‘Beauliew’ (and as having formerly belonged to Lord Montagu) and again as no. 23, ‘Lymington, Barton-on-Sea, Barton, near Milton’; he subsequently corrected the error (Smith 1965, 30). He described it as being of Type IIIb shape. This axehead was examined by Programme JADE (analysis no. BMus_153); classified as a Chelles type axehead, it was found to be of omphacitite, an Alpine rock. (Smith had previously described the rock as a garnet-pyroxenite. ) The other small axehead (Figure 8.4) — actually an adze-head — that Smith had described as coming from ‘Beauliew (Smith 1963, no. 19; WINCM: ARCH 2213, Dale 81) would indeed appear to have come from near Beaulieu, or at least there is no evidence to contradict this. Crawford (1913) reported that it belonged to Lord Montagu (Anon 1926, 336) although by the time of its presentation to Winchester City Museum it formed part of the 29 cm Fig. 8 Axeheads (and an adze-head), definitely and possibly of Alpine rock, from Hampshire (other than Breamore). 1. Hillhead; 2. Titchfield; 3. ‘Barton’; 4. near Beaulieu. Drawings: Yvan Pailler. 30 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Dale Collection. Like the ‘Barton’ axehead, it is currently displayed at Christchurch. At only 29mm long, this is almost a miniature object and could conceivably have been used as an amulet. Described by Smith as being of the same material as the ‘Bartor’ axehead (1963, 160), its reanalysis by Programme JADE (analysis no. BMus_ 154) has confirmed that it is of eclogite, another Alpine rock. Typologically, Smith described it as a Type IIIb axehead, but the Programme JADE classification is as an adze. The third Alpine axehead was found near the River Meon at Titchfield (Figure 8.2; Smith 1963, no. 24; WINCM:ARCH 2214, Dale 80). Broken at the butt end and with a length of 80mm, Smith classified this as a “Type Ila’ axehead and described the rock as a ‘composite with pyroxene as the main constituent. Hardly likely to be of sufficient purity to rank as jadeite properly speaking’ (Smith 1963, 160-1). In Programme JADE’s typology this is a small ‘Durrington type’ axehead, made of retromorphosed omphacitite, and with an Alpine source (analysis nos. BMus_150 and BMus_151). The possible fourth example, from Hillhead (Figure 8.1; WINCM:ARCH 2215), was found some time before 1945 a few kilometres downstream from the Titchfield axehead at the mouth of the River Meon. This, too, is small, at only 62mm long; Campbell Smith classified it as a Type IIc example and reported that ‘the X-ray powder pattern is indistinguishable from that of jadeite’ (Smith 1963, 16, no. 210). According to the Programme FADE classification this, too, is a small Durrington type axehead; while macroscopically the material resembles jadeitite or eclogite, the spectroradiometric result shows it to be of chlorite schist (analysis nos. BMus_ 148 and BMus_ 149). Further research may be required to determine whether the rock is definitely Alpine or not. Several other axeheads found in Hampshire had previously been published as, or thought to be, of Alpine rock but are now known not to be. In his 1963 inventory, Campbell Smith mentioned one in the Russell-Cotes Museum that had been found at Tuckton, near Bournemouth, and confirmed that it is most probably a recent ethnographic manuport from New Guinea (Smith 1963, 171. Victoria Pirie, formerly Senior Arts and Museums officer at the Russell-Cotes Museum, is in accord with that view: pers. comm.). In the same publication Smith mentioned another axehead, found at the base of the cliff at Hordle (Hoardwell), roughly equidistant between Barton and Milford. This had been obtained by T. W. Shore about 1893 and placed in the Hartley Museum at Southampton (Shore 1905, 21). It could not be found at Tudor House Museum by Campbell Smith, but was included as No. 22 in — his catalogue. Encountered subsequently while on | loan to Christchurch Museum, it was thin sectioned and found to be of altered igneous rock, possibly a Cornish greenstone (Smith 1963, 160; 1972, 408-9). Doubt about an Alpine origin also surrounds a small axehead that had been found in the garden of a house (incidentally called ‘Beaulieu’) in St Vincents Road, Ryde, on the Isle of Wight (Reg. no. IWCAC.1893.1). This was analysed by Programme JADE (analysis nos. BMus_ 374 and BMus 375) | and found to be of amphibolite or glaucophanite, — not necessarily of Alpine origin; research continues into the sourcing of this rock. A small black axehead found at Shawford Down, notified to Programme JADE after the analysis programme had ended, does not look to be a likely candidate; and the current whereabouts of an alleged example from Bushy Down Farm, Soberton, could not be ascertained, so this specimen could not be checked. Finally, an axehead found at Droxford and published as being an Alpine import in 1987 (Schofield 1987) was found, when examined by the Programme JADE team, to be a recent ethnographic manuport: a kula axehead, of amphibolite, from Woodlark Island to the south- east of the Trobriand Islands, Papua New Guinea (analysis nos. BMus_ 348, BMus_ 349). Discussion and conclusions This review of the Breamore axehead has served to underline its importance, and its long and complex ‘biography’, and has also shown how much new evidence has been obtained through a review of the existing documentary evidence and through Programme JADE. In terms of the broader cluster of Alpine objects in southern central England, while it seems likely that the majority were brought over during the early centuries of the fourth millennium during one or more processes of Neolithisation (see Sheridan et al. 2008 for a discussion), it is clear that a few had probably arrived much later. Two small perforated ‘axehead pendants’, from Wootton Fitzpaine and Preston (both Dorset: Jones et al. 1977, 289, no. 88 and Smith 1963, 159, no. 15 respectively) are a type of artefact known from Late Neolithic contexts in northern France, associated THE BREAMORE JADEITITE AXEHEAD - with pottery of the Seine-Oise-Marne tradition and gallery graves, and dating to the late 4th to early 3rd millennium BC. Two examples are also known from Jersey. It is conceivable that the Wootton Fitzpaine and Preston examples were acquired from _cross-Channel contacts around this time, as part of a continuing (if intermittent) pattern of contacts with north-west France — and possibly the Channel Islands — from this part of Britain. Earlier evidence for such contacts, post-dating the initial appearance of Neolithic artefacts and practices in southern and south-west England, includes the adoption of certain elements of Chasséen ceramic design (notably trumpet-shaped lugs on ‘Hembury ware’ bowls) from the late 38th/37th century BC; it is possible that the small number of Breton fibrolite axeheads found in south-west England, and of Type A dolerite axeheads from the Plussulien quarries in central Armorica found in southern England, were imported around this time (Smith 1963, 171, ‘C’; Clough and Cummins 1988, 146, no. 266; Le Roux 1979, fig 3). Possible evidence for later cross-Channel contacts comes in the form of two small double-ended chisels of Alpine rock in Torquay Museum (Jones et al. 1977, no. 81, plate 23 and Chandler pers comm.). One of these, from Brownstone Farm, Kingswear, Devon, comes from an Early Bronze Age round barrow and was associated with cremated human remains that have recently been radiocarbon dated (for Programme FADE) to 3420+30 BP (SUERC-21367, 1880-1630 _cal BC at 20). Whether this, and the other very similar (but unprovenanced) chisel in Torquay Museum arrived during the Early Bronze Age is one of many questions about Alpine objects that remain to be answered through future research. Acknowledgements Paul Robinson, former curator of the Wiltshire Heritage Museum, is thanked (along with the Trustees of the Wiltshire Archaeological and Natural History Society) for allowing the loan of the Breamore axehead to Programme JADE and for encouraging us to report on the results of our research. Many museum curators are thanked for their kind co-operation in facilitating study and _analysis of objects mentioned in this article to _ Programme JADE: Jane Ellis-Sch6n at Salisbury and - South Wiltshire Museum; David Allen of Hampshire County Museums Service; Helen Rees of Winchester - Museums Service; Rosamund Cleal of the Alexander Bh Keiller Museum, Avebury; Corina Westwood of the Isle of Wight Heritage Service; Rob Kruszynski and colleagues at the Natural History Museum and Barry Chandler at Torquay Museum. Mr and Mrs Horn, owners of the Droxford axehead, are also thanked. Gill Varndell of the British Museum is warmly thanked for hosting Programme JADE during their 2008 campaign of analysis. Michel Rossy (Université de Franche-Comté, Besancon) is thanked for his examination of thin sections; Massimo Ghedini (GeaDue laboratory, Bologna), for undertaking XRD analyses; and Claudio D’Amico (University of Bologna), for his comments on the XRD results. Finally, Andrew Reynolds is thanked for his patience as editor. References Please note: due to the large number of cited publications authored by Pierre Pétrequin and colleagues, the listing has been simplified so that entries up to 2004 have been listed in the alphabetical order of authors, while the multiple publications for 2005, 2006 and 2008 are prefaced by ‘PETREQUIN et al. 2005a =’ et cetera. ANNABLE, F. and SIMPSON, D. D. A., 1964, Guide Catalogue of the Neolithic and Bronze Age Collections in Devizes Museum. Devizes: WANHS ANON., 1880, Kelly’s Directory, Hampshire. London: Kelly & Co ANON., 1903, Kelly’s Directory, Hampshire. London: Kelly & Co ANON. 1924. Wiltshire Obituaries: Mark Jeans. VANHM 42, 501-511 ANON. 1926. Evidences of prehistoric trade between Wiltshire and France. WANHM 43, 336 ANON. 1930. Highly polished greenstone celt of Breton type. WANHM 45, 88 ANON. 1932. Jadeite celt from Cheshire. Antiquaries Fournal 12, 167 ANON. 1954. Wiltshire Obituaries: Joshua Watts Brooke. WANHM 55, 300-302 ANON. 2007. Dorset County Museum. CBA Wessex News, September 2007, 22-23 BAYLISS, A., HEALY, F, BRONK RAMSEY, C., McCORMAC, G. and MERCER, R., 2008, ‘Interpreting chronology’, in R. Mercer and F. Healy (eds), Hambledon Hill, Dorset, England. Excavation and Survey of a Neolithic Monument Complex and its Surrounding Landscape, 378-411. Swindon: English Heritage BISHOP C., WOOLLEY, A., KINNES, I. and HARRISON, R. 1977. Jadeite axes in Europe and the British Isles: an interim study. Archaeologia Atlantica 2, 1-8 BURL, A. and MORTIMER, N. (eds), 2005, Stukeley’s ‘Stonehenge’: An Unpublished Manuscript 1721-1724. 32 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE New Haven and London: Yale University Press CASSEN S., 2009, Exercice de stéle. Une archéologie des pierres dressées. Réflexion autour des menhirs de Carnac. Paris: Editions Errance CASSEN, S. and PETREQUIN, P. 1999. La chronologie des haches polies dites de prestige dans la moitié ouest de la France. European Fournal of Archaeology 2 (1): 7-33 CLOUGH, T. H. McK. and CUMMINS, W.A., 1988, Stone Axe Studies, Volume 2. London: Council for British Archaeology (Research Report 67) COLES, J., ORME, B., BISHOP, A. C. and WOOLLEY, A. R. 1974. A jade axe from the Somerset Levels. Antiquity 48, 216-220 CRAWFORD, O. G. S. 1913. Prehistoric trade between England and France. L-Anthropologie 24, 641-9 CRAWFORD, O. G. S., 1955, Said and Done. The Autobiography of an Archaeologist. London: Weidenfeld and Nicolson DALE, W. 1905. Neolithic implements from the neighbourhood of Southampton. Proceedings of the Hampshire Field Club 4 (1898-1903), 183-185 DALE, W. 1907. [Report on] Neolithic implements from Hampshire. Proceedings of the Society of Antiquaries 21 (1906-7), 263-266 DAMOUR, A. 1881. Nouvelles analyses sur la jadéite et sur quelques roches sodiféres. Bulletin de la Société Francaise de Minéralogie 4, 157-164. ERRERA, M. 1995. Application de la spectroradiométrie a Pidentification des pierres précieuses. Musée royale de l’Afrique centrale, Tervuren (Belgique), Département de Géologie et de Minéralogie, Rapport annuel 1993 & 1994, 109-128 ERRERA, M. 2003. Application de la spectroradiométrie a l’étude des lames polies: exemples auvergnats, in Les matiéres premieres lithiques en Préhistoire. Préhistoire du Sud-ouest supplément 5, 161-167 ERRERA, M., HAUZEUR, A., PETREQUIN, P and TSONEV, T., 2006, ‘Etude spectroradiométrique dune hache trouvée a Svoboda, district de Chirpan (Bulgarie), in P. Valev (ed), Interdisciplinary Studies XIX, 5-22. Sofia: Archaeological Institute and Museum BAS EVANS, J., 1897, The Ancient Stone Implements Weapons and Ornaments of Great Britain London: Longman, Green, Reader, and Dyer FOX, C. 1933. The distribution of man in East Anglia, c2300 BC-AD 50: a contribution to the prehistory of the region. Proceedings of the Prehistoric Society of East Anglia 7, p |V fig 6b FRENCH, C., LEWIS, H., ALLEN, M. J., SCAIFE, R. and GARDINER, J. 2007. Prehistoric landscape development and human impact in the upper Allen valley, Cranborne Chase, Dorset. Cambridge: McDonald Institute Monographs GODDARD, REV. E. H. 1921. Stone implements of uncommon type found in Wiltshire. WANHM 41, 365-377 HERBAUT, EF and PAILLER, Y., 2000, ‘Les anneaux en pierre dans le massif armoricain’, in Eléments d’architecture. Exploration d’un terte funéraire a Lannec (Erdeven, Morbihan). constructions et reconstructions dans le Néolithique morbthannats. Propositions pour une lecture symbolique, 353-385. Chauvigny, Association des publications Chauvinoises (Mémoire XIX) HILLAM, J.,GROVES, C. M., BROWN, D. M., BAILLIE, M. G. L., COLES, J. M. and COLES, B. J. 1990. Dendrochronology of the English Neolithic. Antiquity 64, 210-220 ; HOARE, R. C., 1812, The Ancient History of South Wiltshire. London: Richard Miller JONES, V., BISHOB A. C. and WOOLLEY, A. R. 1977. Third supplement of the catalogue of jade axes from sites in the British Isles. Proceedings of the Prehistoric Society 43, 287-93 KELLY, E. R., 1867, The Post Office Directory of Hampshire including the Isle of Wight. London: Kelly & Co KINNES, I. A., 1992, Non-megalithic Long Barrows and Allied Structures in the British Neolithic. London: British Museum Press (Occasional Paper 52) LE ROUX, C.-T, 1979, ‘Stone axes of Brittany and the Marches’, in T-H.McK Clough and W.A. Cummins (eds), Stone Axe Studies, 47-56. London: Council for British Archaeology (Research Report 23) PETREQUIN, P, CASSEN, S., CROUTSCH, C. and ERRERA, M., 2002, ‘La valorisation sociale des longues haches de l’Europe néolithique’, in J. Guilaine (ed), Matériaux, productions, circulations du Néolithique a l’Age du Bronze, 67-98. Paris, Editions Errance PETREQUIN, P, CROUTSCH, C. and CASSEN, S. 1998. A propos du dépot de La Bégude : haches alpines et haches carnacéennes pendant le Ve millénaire. Bulletin de la Société Préhistorique Francaise 95 (2), 239-254 PETREQUIN, P.,, ERRERA, M., CASSEN, S. and CROUTSCH, C. 2003. De la pétrographie aux approches sociales: la circulation des grandes haches alpines en Europe occidentale pendant le Neolithique, in Les matiéres premieres lithiques en préhistoire. Préhistoire du Sud-Ouest supplément 5, 253-275 PETREQUIN, P and PETREQUIN, A-M., 2004, Rapport de prospection au Monviso (2003-2004) Commune: ONCINO (Cuneo, Piémont). Besancon: Centre National de la Recherche Scientifique Université de Franche- Comté PETREQUIN et al. 2005a = PETREQUIN, P, ERRERA, M., CASSEN, S., BILLAND, G., COLAS, C., MARECHAL, D., PRODEO, F and VANGELE, F 2005. Des Alpes Italiennes 4 l’Atlantique au Ve millénaire. Les quatre grandes haches polies de Vendeuil et Maizy (Aisne), Brenouille (Oise). Revue Archéologique de Picardie n° special 22, 75-104 PETREQUIN et al. 2005b = PETREQUIN, P, PETREQUIN, A. M., ERRERA, M., CASSEN, S., CROUTSCH, C., KLASSEN, L., ROSSY, M., GARIBALDI, P., ISETTI, E., ROSSI, G. and DELCARO, D. 2005. Beigua, Monviso e Valais. AlPorigine delle grandi asce levigate di origine alpina THE BREAMORE JADEITITE AXEHEAD in Europa occidentale durante 11 V millenio. Rivista di Scienze Preistoriche LV, 265-322 PETREQUIN etal. 2006a = PETREQUIN, P, CASSEN, S. and CROUTSCH, C., 2006, ‘Imitation ou convergence: Les haches néolithiques a talon perforé au nord-ouest des Alpes’, in L. Baray (ed), Artisanats, sociétiés et civilisations: hommage a fean-Paul Thevenot, 163-177. | 24° supplément a la Revue Archéologique de l’Est ~ PETREQUIN et al. 2006b = PETREQUIN, P, ERRERA, M., PETREQUIN, A-M. and ALLARD, P 2006. The Neolithic quarries of Mont Viso, Piedmont, Italy: initial radiocarbon dates. European Fournal Archaeology ) 9(1), 7-30 ~ PETREQUINetal. 2006c = PETREQUIN, P PETREQUIN, A-M., ERRERA, M., CASSEN, S. and CROUTSCH, C., 2006, ‘Complexité technique et valorisation sociale: haches polies de Nouvelle-Guinée et du Néolithique alpin’, in L. Astruc, F. Bon, P-Y. Milcent and S Philibert (eds), Normes techniques et pratiques sociales. De la simplicite des outillages pré- et protohistoriques, 419- 433. Antibes: Editions APDCA (XXVI Rencontres internationales d’archéologie et d’histoire d’Antibes) PETREQUIN etal. 2008a = PETREQUIN, P, SHERIDAN, J. A.. CASSEN, S., ERRERA, M., GAUTHIER, E., KLASSEN, L., LE MAUX, N. and PAILLER, Y., 2008, ‘Neolithic Alpine axeheads, from the Continent to Great Britain, the Isle of Man and Ireland’, in H. Fokkens, B. J. Coles, A. L. van Gijn, J. P Kleijne, H. H. Ponjee and C. J. Slappendel, Between Foraging and Farming: an Extended Broad Spectrum of Papers Presented to Leendert Louwe Kooymans, 261-79. Leiden: Leiden | University (Analecta Praehistorica Leidensia 40) | PETREQUIN er al. 2008b = PETREQUIN, P, : PETREQUIN, A. M., ERRERA, M. and KLASSEN, L. 2008. Aus den italienischen Alpen nach Thiiringen. Zur naturwissenschaftlichen Untersuchung eines herausragenden neolithischen Fundes des Berliner Museums fiir Vor- und Frihgeschichte. Acta | Praehistorica et Archaeologica 40, 241-57 - PETREQUIN, P, SHERIDAN, J. A., CASSEN, S., ERRERA, M., GAUTHIER, E., KLASSEN, L., LE MAUX, N., PAILLER, Y. and PETREQUIN, A. M., in press, ‘Eclogite or jadeitite: the two colours involved in the transfer of Alpine axeheads in western Europe’, in M. R. E. Edmonds and R. V. Davis (eds) Stone Axe Studies IIT. York: Council for British Archaeology (Research Report) | PIGGOTT, S. and POWELL, T.G. E. 1949. The excavation | of three Neolithic Chambered Tombs in Galloway, 1949. Proceedings of the Society of Antiquaries of Scotland 83 (1948-9), 103-161 _ POLLARD, S. H. M. 1966. Neolithic and Dark Age Settlements on High Peak, Sidmouth, Devon. Devon Archaeological Exploration Society 23, 35-59 | R.C.H. M., 1979, Long Barrows in Hampshire and the Isle of Wight. London: HMSO _ ROBINSON, P, 2003, ‘The Museum Collections’ in J. H. Thomas (ed) Wiltshire Archaeological and Natural History 33 Society: the first 150 years, 79-89. Devizes: WANHS SCHOFIELD, J. 1986. Archaeology and historic buildings in Hampshire. Hampshire County Council Annual Report 1986, 32 SCHOFIELD, J. 1987. A jadeite axe from Droxford, Hampshire and its regional context. Proceedings of the Prehistoric Society 53, 483-485 SHERIDAN, J.A., 1992, ‘Scottish stone axes: some new work and recent discoveries’, in N.M. Sharples and J.A. Sheridan (eds.), Vessels for the Ancestors: Essays on the Neolithic of Britain and Ireland in Honour of Audrey Henshall, 194-212. Edinburgh: Edinburgh University Press SHERIDAN, J. A. 2007a. Green treasures from the magic mountains. British Archaeology 96, 22-27 SHERIDAN, J.A., 2007b, ‘From Picardie to Pickering and Pencraig Hill? New information on the ‘Carinated Bowl Neolithic’ in northern Britain’, in A.W.R. Whittle and V. Cummings (eds), Going Over: The Mesolithic- Neolithic Transition in North-West Europe, 441-492. Oxford: Oxford University Press (Proceedings of the British Academy 144) SHERIDAN, J.A., SCHULTING, R., QUINNELL, H. and TAYLOR, R. 2008. Revisiting a small passage tomb at Broadsands, Devon. Proceedings of the Devon Archaeological Society 66, 1-26 SHORE, T. W. 1905. The origin of Southampton Water. Proceedings of the Hampshire Field Club 5, 1-25 SMITH, I., 2008, ‘Stone axes’, in R. Mercer and E Healy (eds), Hambledon Hill, Dorset, England. Excavation and Survey of a Neolithic Monument Complex and 1ts Surrounding Landscape, 630-2. Swindon: English Heritage SMITH, W. CAMPBELL. 1963. Jade axes from sites in the British Isles. Proceedings of the Prehistoric Society 29, 133-172 SMITH, W. CAMPBELL. 1965. The distribution of jade axes from sites in Europe with a supplement to the catalogue of those from the British Isles. Proceedings of the Prehistoric Society 31, 25-33 SMITH, W. CAMPBELL. 1972. Second supplement to the catalogue of jade axes from sites in the British Isles. Proceedings of the Prehistoric Society 38, 408-411 STONE, J. E S. and WALLIS, E S. 1951. Third report of the sub-committee of the south-western group of museums and art galleries on the petrological determination of stone axes. Proceedings of the Prehistoric Society 17, 99-158 THURNAM, J. 1871. On Ancient British Barrows, especially those of Wiltshire and adjoining Counties (Part II Round Barrows). Archaeologia 43(2), 285-544 TRAVERSO, G. B., 1898, 1901, 1909. Stazione neolitica di Alba. Alba: Libreria Sansoldi WHITTLE, A. W. R. 2007. ‘The temporality of transformation: dating the early development of the southern British Neolithic’, in A. W. R. Whittle and V. Cummings (eds.), Going Over. The Mesolithic—Neolithic Transition in North-West Europe, 377-98. London: British 34 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Academy (Proceedings of the British Academy 144). typological survey’, in T. H. McK Clough and W. A. WOOLLEY, A. R., BISHOP, A. C., HARRISON, R. Cummins (eds.), Stone Axe Studies, 90-6. London: J. and KINNES, I. A., 1979. ‘European Neolithic Council for British Archaeology (Research Report jade implements: a preliminary mineralogical and 23) Wiltshire Archaeological & Natural History Magazine, vol. 103 (2010), pp. 35-101 | The Late Bronze Age and Early Iron Age Midden Site at East Chisenbury, Wiltshire by David McOmish, David Field and Graham Brown with contributions by Shahbaz Ashraf, Joanna Bagust, Martyn Barber, Wendy Carruthers, Deborah Cunliffe, Roy Entwistle, Sarah Inskip, Caroline Fenkins, Richard Macphail, Elaine Morris, Frances Raymond, and Dale Serjeantson Chance discovery of a mound near East Chisenbury, Wiltshire that preserved an intricate series of deposits akin to those - excavated at Potterne (Lawson 2000) has transformed our understanding of the Late Bronze Age — Early Iron Age _ transition in southern Britain. The landscape setting 1s remarkable with the remains of a substantial enclosure adjacent to the north-east, as well as a number of convergent prehistoric linear features and a substantial circular hollow. Ephemeral and fragmented components of a later prehistoric field system and the remains of more recent ridge-and-furrow cultivation overlie the major components of this complex. This report subsumes preliminary accounts (McOmish 1996: Brown et al. 1994) and sets out the scale of the fieldwork at East Chisenbury. Following the introduction, sections outline the local landscape history, archaeological background, earthwork survey and other field interventions. Short reports are included relating to various materials recovered during the fieldwork and the article concludes with discussion of the material culture and a general review of the site, its significance and context. Introduction Archaeological investigation on the ridgetop at East Chisenbury took place initially in May 1992 and gathered impetus when it became clear that the military authorities were planning a new road across what was then presumed to be a natural hillside (Figure 1). Survey and excavation of two test pits demonstrated that this was not the case and that the hill itself was a constructed monument, set amidst a complex of other sites. The site is located towards the northern end of the Avon valley where the river leaves Fig. 1 View of the mound at East Chisenbury looking west. The central ranges of the Salisbury Plain Training Area can be seen and the massive Compton English Heritage, 24 Brooklands Avenue, Barrow is visible slightly right of centre in the this image, just above the curve of Cambridge, CB2 8BU the midden mound. © David Norcott 36 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE SOO mates X ~~ yad metres < ~ *S0 metres a yen y: ic a ; UPAVONICR. § Upaven Airfield Fig. 2 The site lies on the eastern flanks of the Avon Valley, immediately to the north-east of the village of East Chisenbury, itself close to the north-eastern periphery of the Salisbury Plain Training Area. © Crown Copyright. Unauthorised reproduction infringes Copyright and may lead to prosecution of criminal proceedings. English Heritage 100019088. 2009 THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 37 the Vale of Pewsey and begins to cut through the chalk en route to the coast at Hengistbury Head (Figure 2). It sits close to the southern limit of an ill-defined spur of Middle Chalk that projects into the valley of the River Avon, 0.6km from the village of East Chisenbury in the parish of Enford (NGR SU 145 532) at a height of approximately 147m above Ordnance Datum (OD). The spur is generally aligned on a north-east to south-west axis and flanked on its northern side by a sharply defined and now dry re-entrant coombe that extends back from the main river valley for a distance of nearly 1km. This valley, part of a network of palaeo-tributaries feeding into the River Avon from the east, provided shallow gradient access to the higher chalk plateaux. To the south, the topography shelves gently, descending to the 100m contour mark some 600m to the south of the site. Likewise, this edge is demarcated by a now dry valley, less sharply profiled than that to the north, again aligned on a north-east to south-west axis, but slightly more complex in that it bifurcates with a pronounced arm extending northwards 400m to the east of the site. These dry valleys thus define the spur and demarcate a downland ridge that rises to a height of 165m above OD in the area of Upavon Airfield, 1km to the north-east of the site. To the west, the natural ridge drops gently to the valley floor and it appears that the archaeological _ deposits have accumulated at that point where the natural declination becomes accentuated. False- crested locations are more frequently seen in the positioning of earlier burial monuments and it may well be that the same sorts of topographical preferences were being exercised on this occasion. The complex thus occupies a striking landscape setting with wide and extensive views to the east, the : far distance punctuated by a ridge of Upper Chalk extending northwards from Chisenbury Field Barn rising to a height of 166m above OD in the area of Upavon Down. The most pronounced section of this ridge, further to the south-east, is crowned by the multivallate hillfort of Sidbury Hill, at 224m above OD, the highest point on this part of Salisbury Plain, and a range of other, less monumental and, largely later, Iron Age sites have taken advantage of these natural high points. To the south, a dissected plateau would have been apparent with the level Middle Chalk ranges cut into by a number of dry valleys sapping back from the Avon Valley. Indeed, an extensive view of the river valley would have been afforded from the site itself. The northern horizon is equally remarkable and the daunting south-facing escarpment edge of the Marlborough Downs forms a spectacular backdrop. This block of high chalk downland is punctuated by a range of significant sites, some earlier, such as the Neolithic causewayed enclosures at Knap Hill and the juxtaposed megalithic tomb at Adam’s Grave, the enclosure and linear earthwork complex at Rybury/ Clifford’s Hill (Oswald et al. 2001; Brown et al. 2005), as well as other contemporary sites, including the Early Iron Age enclosures at Martinsell Hill and Giant’s Grave. The southern fringes of the settlement complex at All Cannings Cross (Cunnington 1923) would have been evident, so too the newly discovered Early Iron Age deposits in the vicinity of Stanton St Bernard. The visual impact of this dramatic location is diminished by a modern shelter belt of trees planted earlier in the 20th century, which has curtailed views to the west from the spur top. Paradoxically, this narrow plantation forms a prominent ‘eye- catcher’ from the valley floor but in doing so masks the location of the site. The views west, given clear sightlines, would have been spectacular, with a vista that encompassed a significant stretch of the river course as well as the central portion of the Salisbury Plain; an area that hosted a number of significant contemporary sites such as the enclosures at Widdington Farm and Casterley Camp hillfort. One of the most significant viewlines extends south-westwards across the Avon valley towards the re-entrant channel at Compton. This view iS prominently interrupted by the massive and presumed Neolithic round mound on the southern slope above Water Dean Bottom (McOmish et al. 2002, 39). Landscape history and previous research The earliest documented investigations at the site are those of William Cunnington and Sir Richard Colt Hoare who visited the area in the late 18th century in preparation for the Ancient Wiltshire series (Hoare 1812). Hoare, having already investigated a circular enclosure nearby at Chisenbury Trendle, some 400m to the north-east and now largely obliterated by the runway at RAF Upavon, observed the eastern arc of a bank and ditch enclosure (with external ditch) of considerable magnitude and depicted it as cutting off the chalk spur in the manner of a cross-ridge boundary on Philip Crocker’s map of Everley Station (Hoare 1812, facing page 178). No mention of a io 4) > >) mound was made at this stage with Hoare noting that, ‘On the same ridge of hill, and at a short distance to the west, are the remains of another work forming the segment of a circle, and presenting a bolder rampart than the former. In the neighbourhood of this latter work we picked up some rude British pottery, and I was informed by the farmer that the soil 1n the neighbourhood of these works was remarkably fertile.’ (abid., 192). Thereafter, the site has attracted little attention. Neither the enclosure nor pottery scatter feature on early Ordnance Survey maps and visits in the early 20th century by a range of people including Maud Cunnington, excavator of All Cannings Cross, (who in 1913 commented that no trace could be seen of the earthwork reported by Hoare), failed to document either the enclosure or the dense spread of pottery. The enclosure was visible, though, on a Crawford air photograph taken in 1921 which appeared to show a ‘segment of a circle’ shaped bank apparently overlying a more angularly shaped cropmark enclosure, the latter most likely the remains of ancient fields in the area. Interestingly, the curvature of the ‘segment’ is opposite to that shown by Hoare. The site was under the plough at the time of Crawford’s reconnaissance and may have remained so over the subsequent decades. Grinsell (1957) was aware of an important prehistoric site in the area, having reported bone implements, spindle whorls, and decorated fineware ceramics, but he noted no earthwork remains. A later Ordnance Survey description, made in 1973, comments that the feature noted by Hoare consisted of a ‘single bank with flanking ditches’ following a course across the valley, ignoring local topography ‘more likely to be a land boundary than a trackway.’ In addition, the then Ordnance Survey Archaeology Field Investigator confirmed ‘Iron Age or Medieval pottery in small quantities but not sufficient to prove the existence of a settlement.’ (OS 495 — SU 15 SW 22). The site remained in cultivation at that time but it is most surprising that no mention whatsoever was made of an earlier and more intensive excavation during the Second World War. Happenstance is often an important factor in archaeological research and so it has proved in this case. Quite by chance, on a visit to the site during an open day in July 1992, Jon Cotton of the Museum of London remarked that he had been assessing archaeological finds left by the recently deceased Squadron Leader Tom Walls of Epsom and that a large number of these were discovered during digging in the vicinity of Upavon airfield. Further news emerged from Dr Gill Varndell and THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Dr Stuart Needham of the British Museum that another individual, Squadron Leader N. N. E. Bray had deposited finds from Upavon at the museum in 1945. Both Bray and Walls had been stationed at RAF Upavon and had evidently joined forces to work on the site of a pottery scatter. Both museums supplied tantalising archive material including written accounts of artefacts, areas of excavation and personal correspondence between the excavators and established archaeologists of the time. It is clear from surviving documentation that Walls discovered the site shortly before 1945. Evidently, he had noted it from the air but it is unclear if he observed spreads of dark soil, cropmarks or earthwork remains (letter from John Musty to Tom Walls 1:v:60 BM). On a ground visit, Walls collected pottery similar to what he had found at Chisenbury Trendle, a site almost levelled as the earthworks posed a hazard to landing aircraft (Cunnington 1932). Walls proceeded quickly to dig at East Chisenbury, initially on his own and then with Bray, giving the position of the dig as about ‘J mile from site to Watch Office’ marked with an ‘x’ on a sketch plan; a position that makes it clear that he investigated the site described here — his trench was probably located on the northern edge of the mound at its intersection with a ‘Celtic’ field lynchet (see below). Interestingly, on his sketch two locales are indicated — ‘site’ and ‘B. SITE’ further to the north- east. No detailed report survives of this exploration other than a hand written note accompanying drawings of artefacts recovered. ‘Here are very rough & confused notes made during excavation:- There appears to be a floor at 1’. Leaden clayey-earth, potsherds, lumps of chalk and flints & at this level too the bone implements have been found. The potsherds and flints are usually coated in patches with some dirty yellowish thick deposit which suggests a dirty living floor. This deposit is firm and resists working & scrubbing. Most of the animal bones have been cracked open either by teeth — there are examples with teeth marks around the broken ends — and some have been hacked or sawn in half presumably to extract marrow. The potsherds are found at all levels down to 2" and are fragmentary & confused. Red & yellow ochre (sic) in small lumps & rough lumps of unbaked clay of a light reddish yellow colour.’ By late August 1945 it is clear that Bray was working on his own and in a letter to Walls dated Wednesday THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 39 . 22 August 1945, he commented, ‘Tam now going to advance pick-a-pte into the hill and things become more & more extraordinary...’ _ Bray then proceeded to outline his main discoveries including basic plans and sections of the trench. _ Three rough plans are shown. The first of these labelled ‘(a) Plan’ shows the area of excavation in - relation to Walls’ earlier work and indicates that Bray _ dug an area at least 10 feet by 8 feet in extent and that the trench was placed on the north-facing slope of a ‘Celtic’ field lynchet that apparently truncates the ‘hill’ on the north. Walls’ smaller trench lay a short distance west, again at the foot of the lynchet. Two elevations are shown, ‘(b) Elevation’ and ‘Front elevation’; the latter drawing is of most interest as it _ is supplemented by short descriptions of structures and layers encountered. At the time of writing, Bray had dug to a depth of 6 feet but had not yet _ reached the base layer: complex stratigraphy had | been revealed — ‘Where I have marked A’, is a built up pillar — at ‘B’ and ‘C” there are windows, as it were, full of bones pottery etc right beyond the apex of the bank!!....The pillar is 2’ 6” wide & about 2’ high...’ The so-called Front elevation includes layers _ described as ‘rough cement’ or being ‘full of bones pottery etc’. Ata depth of close to 6 feet a more diagnostic , clay full of bones & floored with flints.’ Beneath this was (a lens of (unexcavated) flinty gravel. Given the quality and the quantity of finds from | their excavations, Walls and Bray were convinced that they had found a settlement of some importance and communicated this to the archaeological authorities of the day. Samples of pottery and animal _ bones were sent to Alexander Keiller for comment _ and he presented information about the site to the | ' Council of the Wiltshire Archaeological Society. | From this correspondence it is evident that Walls _ also wrote to C. E C. Hawkes, and W. E. V. Young, | the latter then Curator at the Museum in Avebury, who, drawn into on-going discussion about the site, took a swipe at Keiller (and, amusingly, Maud Cunnington) in the process — ‘In Mr Keiller’s letter to you of Nov. Ist, I note he states that “Mr Young 1s of the opinion that this site, in so far as one can accurately identify it from the description given to me at the time of Squ.Ldr. Walls’ excavations, 1s identical with one which Mrs Cunnington excavated in some small measure some years ago.” This statement 1s rather misleading....I know of no site excavated by Mrs Cunnington — “in some small measure some years ago”. What I did say to Mr Keiller when he brought the finds into the Museum to show me was, that the pottery and the bone objects appeared to be definitely of the ALL CANNINGS CROSS type, and this opinion I still adher (sic) to’. T hope you will accept this correction, as I certainly did not express at any time the opinion Mr Keiller saw fit to attribute to me. At the same time I am not bothering to point this error out to Mr Keiller as it would only be a waste of time, and would probably lead to a lot of un- necessary correspondence of the kind he revels in.’ Animal bone from East Chisenbury was sent to Judith E. King of the British Museum (Natural History), but she replied that ‘nothing much can be said about the bones as they are very fragmentary and no measurements could be taken for comparisons with specimens in the Museum.’ Nevertheless, she was able to identify a number of species, mainly ox and sheep or goat, with smaller representations of pig, horse and a kind of mouse. With the significance of the site appreciated by contemporary scholars it is surprising that no follow- up work took place. Apart from the brief mention of pottery, bone implements and spindle whorls in the Victoria County History of Wiltshire Vol II (Grinsell 1957), derived from material presented to the British Museum by Bray (B 1945, 10-5, 1-35 and 1946, 10-3, 1-21), no other discussion has taken place. This is remarkable given the intensity of fieldwork on the Wessex chalklands throughout the 20th century, although such neglect is perhaps better attributed to the site’s location on a restricted military training area, albeit tenanted farmland. Artefacts from Walls’ and Bray’s excavation continued to circulate in archaeological circles and some fifteen years after their discovery, the late John Musty, then Curator of the Salisbury, South Wiltshire & Blackmore Museum, wrote to the British Museum to ask if the pottery sherds might be given to Salisbury. In a letter dated March 20, 1960, he queried — ‘Somewhile ago I received from Mr Brian Hope-Taylor, via Prof. Richard Atkinson, some extremely interesting pottery which I gather had been found by you at East Chisenbury. One day I might possibly make an investigation of the area, but in the meanwhile I write to ask if you would be prepared to present these sherds to the Museum, as we are not well represented in material of this type.’ Despite Musty’s aspirations, no further research 40 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE was undertaken at East Chisenbury and the site was forgotten. The re-discovery of the East Chisenbury midden The existence of a significant archaeological site at East Chisenbury was first brought to the attention of the authors by Nell Duffie, a member of the Salisbury Plain Training Area (East) Conservation Group. Mrs Duffie had been informed of the discovery of large numbers of sherds of prehistoric pottery being brought to the surface by rabbit and badger digging and had gathered a small collection of the most impressive fragments. From this assemblage it was immediately apparent, both from the form of the ceramics as well as the large size and excellent condition of the collection, that it had come from a site of some importance dated to the Early Iron Age. The large sherd size and the fresh and clean nature of the fracture lines suggested that the material had come from a largely undisturbed and protected context that had avoided sustained ploughing. A reconnaissance visit soon after as part of a programme of investigation of Salisbury Plain Training Area archaeological sites by the Royal Commission on the Historical Monuments of England resulted in the surface collection of several hundred sherds of pottery, all, seemingly, of one period, and amassed from a relatively small area of surface disturbance within the modern strip plantation on the hilltop. As well as the ceramics there were substantial quantities of ancient animal bone, fragments of sarsen, flint and chalk. Detailed inspection of the topography revealed that the ‘hill’ appeared to be artificial and what were presumed to be cultural deposits apparently survived as a prominent mound on the ridge top. In addition, there is a large elliptical enclosure immediately to the north-east, remains of multi-period field system on the north-west and south-east, as well as an adjacent massive hollow. This remarkable setting, even for a Wessex prehistoric site, was subsequently embellished further by at least six linear features that converge on the hilltop. The earthwork remains The mound The mound on first observation appeared to be polygonal in outline (Figure 3), with a maximum diameter ofc. 150m and a basal area of approximately 2ha. However, the mound is severely truncated to the north and, to a certain degree, to the north-west and by projecting its line it is feasible to suggest an earlier oval outline in excess of 200m on the north- west to south-east axis and 180m on the north-east to south-west line. The highest point of the mound is at the centre and this is at least 3m above the current ground level. It is estimated, therefore, that the mound might contain, as a minimum, 40,000 cubic metres of material. Plough truncation is evident all around the edge of the mound and it must be assumed that cultivation, as well as erosion from a more recent trackway, will have effectively stripped a large part of the surface area. Ancient fields The fragmentary remains of ancient fields can be seen to the north and south-east of the mound and enclosure. As noted previously, those on the north-west cut through the midden deposit and are, therefore, later in date, and it is feasible that they belong to a later prehistoric or Romano-British episode of cultivation. Aerial photographs of the vicinity of the mound suggest at least two phases of field system; a common observation on the Salisbury Plain Training Area. Indeed, their sharp profile, particularly that of the lynchet closest to the mound, indicates that they may have been re-used in more recent times. The earthworks to the north are particularly difficult to disentangle and they are likely to incorporate elements of cultivation, the mound, as well as a section of the nearby enclosure boundary. The tripartite linear feature that truncates the mound to the north-west has evidently been used as a double-lynchet trackway at some stage and marks the boundary between at least two ancient fields — one to the north-west, the other to the south-east. The largest and best preserved of the agricultural features is that contiguous with the mound on the north-west and consists of a plough step 0.6m in height. This is a positive lynchet that originated as the result of soil build-up at the northern edge ofa field gently sloping to the north- west. The scarps flanking this on the north-west are shallower to a height of c. 0.4m and 0.5m and are negative lynchets created by the plough biting into the steepening slopes here. These linear scarps are truncated by a trackway immediately north of the mound and beyond an even more confused pattern is apparent in the field remains. A field system can also be seen to the south- THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 41 East Chisenbury uae we angers Teg et TEE qm RAG S \e x \ Y SLUT TTT > \ Man i S METRES becca ———— seer —— 0 FEET Fig. 3 RCHME earthwork survey plan. The earthwork survey was undertaken by staff from the RCHME as part of a rapid response to potential threat to the site. It was carried out at the height of summer in lush vegetation and as a result some detail was obscured. Nonetheless, the main components, including the mound, enclosure and quarry pit’ were identified. A new survey, in ideal ground conditions, would certainly lead to the discovery of new details. = = = een 42 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE: east of the deposit, parallel to, but some 200m distant from their north-western counterparts. Field boundaries are represented by two parallel shallow lynchets aligned north-east to south-west, each south-east facing and reflecting the underlying natural topography. These lynchets are less well- preserved than those to the north and must have been over-ploughed on numerous occasions. They appear to form a double-lynchet trackway following the line of the enclosure and it may well be that the northernmost lynchet of this pairing resulted from cutting into a counterscarp bank. Only one intervening field boundary was noted during the survey and this consists of a curvilinear plough- eroded bank 0.1m high, which converges with the enclosure on the north-east. The enclosure The enclosure to the north-east of the mound (Figure 3) appears partly bivallate with a single bank 15m wide at the base rising to a height of 0.1m above internal ground level. Externally, the bank descends into a shallow ditch 0.3m deep and up to 10m wide and is flanked, particularly noticeably on the north, by a lower spread bank with a slight external ditch used in recent times as a trackway. Ploughing and erosion have altered the profile of the enclosure which is indicated by the slight plough step evident on the external face of the internal bank. At the extreme northern apex of the enclosure the bank profile seems more prominent and an interruption in its line may mark the presence of a narrow gap entrance 2m wide. Viewed from a distance this area appears to be naturally raised, the swelling extending over an oval area at least 30m in diameter, and it may be that the enclosure boundary overlies an earlier feature here or, less likely, a natural eminence. To the south-east only 110m in length was identified at the time of survey and consisted of a bank with an external ditch; no external additions were evident. Recent clearance of vegetation and grass-cutting has revealed additional detail, although our original survey plan reproduced here remains unamended. The enclosure boundary largely follows the course outlined by the 1992 survey, but it is now clearer that it extends on a fairly straight route on a north-east to south-west alignment, skirting the southern edge of the mound before becoming lost in the plantation at its southern limit. There is, therefore, no surface Stratigraphic relationship between the enclosure and mound on the south-east. The boundary, heavily spread by cultivation, when eventually fully exposed in the vegetation on the hilltop consisted of a bank 10-12m wide and up to 0.1m in height above internal ground level. The ditch is similarly massive, up to 15m wide in places, with a maximum depth of 0.2m. The line of the enclosure on the north-west is much more problematic as is its relationship to the mound and this is compounded in that a linear block has been removed by military activity close to the midpoint. This ‘side’ is markedly straight, presumably as a result of cultivation and other later activity, but the angled change in alignment at the north-western apex seems too sharp for the anticipated course of the enclosure boundary. The remains here consist of a series of linear hollows and low banks, of which the outermost feature extends as an embanked hollow away from the enclosure to the north-east for a distance of at least 0.4km. The line of the enclosure in this area, therefore, cannot be identified with complete assurance. It is likely, however, given the convergence of major plough lynchets and later trackways, that the enclosure boundary lies within these linear earthworks. Although at least 50% of the enclosure boundary is now lost, it is probable that a minimum area of c. 4ha was enclosed internally. Apart from the curving lynchet noted above, few internal features were observed due to the intensity and presumed prolonged period during which ploughing took place on the hilltop. The remains of ridge-and-furrow cultivation scar the interior of the enclosure and overlie the mound. On a recent visit to the site, a long low linear ridge (not on plan) extending from the mound to the north- east and apparently underlying the enclosure, was observed. This is associated with the swelling on the enclosure boundary noted earlier and could be a natural ridge in the chalk. Other, slighter, scarps and banks may result from later activity and damage. One of these, the small bank that bifurcates from the innermost linear feature close to the northern apex of the mound, may well be related to an earlier line of enclosure, pre-dating the late stages of mound development. The hollow The massive crater immediately north-west of the enclosure and perched on a terrace above the coombe, appears on first inspection to be a quarry hollow (Figure 3). It is oval in plan with two squared off corners, reaching a maximum diameter in excess of 100m, and has been terraced into the north-facing slope so that its rear face is more elevated than any other aspect. To the north, the hollow flattens out and THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 43 opens, in an entrance-like fashion, on to the steep slopes above the re-entrant valley that leads from Avon valley. At least two concentric ‘benches’ or tiers are evident within the hollow and best observed where the hollow reaches its greatest dimensions. Even though it has been overploughed in the historic period, the hollow retains a distinctive profile and there is a strong suspicion that it was deliberately constructed rather than excavated as part of surface extraction of chalk. Although undated the hollow is of a scale that dwarfs other medieval and post- medieval quarrying on the chalk downland and its morphology finds no analogy with more recent remains of quarrying. Two slight scarps emanating perpendicularly from a lynchet and apparently cut by it to the south of the hollow may well result from more recent military activity. Alternatively, they may be the remnants of one of the linear features, noted by Entwistle (pers. comm.), that approach the site from the north. Another plausible explanation is that these ephemeral scarps mark the site of Walls’ and Bray’s excavations. The excavations (May 1992 — August 1993) Following initial re-discovery of the site in 1992 a substantial part of the mound was threatened by the development of a new routeway across this section of the Salisbury Plain Training Area. The construction of a road would have involved the digging of a preparatory trench 15m wide to varying depths but normally down to bedrock. Early plans suggested that an ideal route for the track lay close to the northern flank of the plantation that masked part of the mound. The site was thus at risk but at a time prior to any knowledge of the work by Walls and Bray. Further work was thus required on the ‘hill’ in order to characterise the field monuments and to test the survey information. In the face of serious doubt by military authorities of the archaeological nature of the site, an auger survey and two small test pits were opened in May 1992 (Figure 3) with the aim of investigating what appeared to be a natural ‘hill’. The auger survey undertaken by staff from Reading University, with the position of boreholes surveyed by the RCHME and shown here (Figure 4) plotted on to a depiction of the contours, confirmed the magnitude of the mound. Measurements were undertaken in linear Fig. 4 Location of trenches and auger samples. A number of auger points (black dots) were established and these helped to define the extent of the archaeological deposits and influence the location of the test trenches. These are marked as ‘A’ and ‘B’ on plan and retain this nomenclature throughout the report text. strips radiating out from the centre point every 45° and showed largely that where earthwork survey suggested the cultural deposits terminated, very little (<0.4m) survived below ground. Recent animal damage to the south of the mound revealed an ashy, material rich soil, to a depth of at least 0.4m and similar to that found during excavation, indicating that the main deposit was not confined to the mounded area. Indeed, it would appear likely that the entire ground surface, both of the mound and a wide buffer surrounding it for a considerable distance, consists of archaeological material, either in-situ deposit or plough-derived material unrecognised during the auger survey. On several occasions the auger probe, 2m in length, failed to find the base of the deposit indicating that the natural ground level might be relatively undulating and that debris and soil has filled these pockets to a much greater depth than that suggested by the above-ground remains. Excavation, initially by the report authors as well as Frances Raymond and Roy Entwistle, then of the University of Reading, was continued into a second season by the former. This long period reflected both the complexity of the deposits uncovered and the slow pace of progress because the work could only be carried out in addition to the normal RCHME survey programme. +4 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 1: aici of deposits Trench A vontext Description Pe [Tsoi A friable sandy loam, covering the whole trench, with 25% small to medium-sized sub-angular eee nodules. Good, clear but irregular clarity of junction oe A friable loamy material dark brown in colour, covering the whole trench, with 10% medium sub-rounded chalk nodules. The deposit is quite loose and sandy, fibrous, gluey and unctuous, oa 4 Midden _ {A soft clayey loam, covering the whole trench, with 10% very small sub-rounded chalk fragments and 1% small sub-angular flint, and with frequent flecks of charcoal and potsherds. Clarity of junction with underlying layer 5 clear, less so with layer 6. 1OYR3/3. fragments. In places the charcoal was quite dense. Clarity of junction with the underlying layer is clear. 10OYR3/2. 10 Clay Loam |Friable clay loam with 5% very small sub-rounded chalk fragments and occasional small flint and the upper part of the deposit probably mixed by cultivation. Clarity of junction with the 5 Chalky A chalky marl, fine chalk matrix, incorporating medium sub-angular nodules of chalk c. 0.03m in marl size, covering north-west corner of the trench only, 0.5m by 0.6m in extent. Has the appearance of being deliberately rammed to make a floor. Clarity of junction with associated layers good. Some charcoal, bones and potsherds within feature flakes. Clarity of junction with underlying layer was good. 10YR3/3. Eee Ploughsoil or OLS gluey, and containing lenses of clay, chalk and ashy material that occur regularly but which extend for only a few cms before fading, with 5% sub-rounded chalk and flint nodules, and with many small flecks of charcoal. An organic stain, c. 0.05m by 0.03m occurred in the south-east junction with the underlying layer is irregular. 1OYR2/1. [Saal Ce sate A soft silty clay loam distinguished from layer 7 by less frequent charcoal constituent. Clarity of junction with the underlying layer was clear. 1|OYR3/2. corner of the trench at 0.8m deep. 10YR3/2. Much ash and charcoal Midden {A soft, silty loam with 5% very sub-rounded chalk fragments and with 25% small charcoal underlying layer clear and smooth. 10YR3/2. Pee Midden material similar to layer 4. A soft clayey loam across the whole of the trench again with frequent flecks of charcoal. 10YR3/3. Dark grey clay of even consistency, with good clarity of junction with underlying layers. Included A soft, loamy clay, covering the whole trench, variable in consistency, both fine, crumbly and A soft, silty clay loam with many flecks of charcoal. Although clear in places, the clarity of blade of a socketed axe and shale bracelet. Fire- Large nodules of burnt flint within a dark grey loamy clay matrix. Clarity of junction with cracked underlying layer poor. flint 13 [Clay Loam | i Clay Loam |An orange clay, friable and crumbly, being the surface of clay-with- flints. Small natural flints occur in profusion 0.1m further within the deposit. A small stakehole, 0.1m in diameter by 0.19m deep, containing darker softer loamy material occurred at the centre of the trench close to the west section. Occasional pieces of bone and sparse potsherds in the upper part of the layer. Table 2: pata of deposits Trench B Mira Description flint was abundant and some fragments of bone were also present. One small piece of sheet bronze was also recovered. ial A friable loamy material dark brown in colour with 10% medium sub-rounded chalk nodules. This deposit covered the whole trench and was quite loose and sandy, fibrous, gluey and unctuous, and the upper part of the deposit probably mixed by cultivation. Clarity of junction Cleared by spade to a depth of 0.2m, but like all other deposits was sieved. Pottery and burnt with the underlying layer clear and smooth. Pottery, flint and animal bone was abundant and what appeared to be at least one well-formed coprolitic stool. Chalk nodules up to 0.15m across, with smaller pieces of chalk rubble filling the interstices. Uneven thickness across the trench; chalk layer appeared immediately under topsoil in western half of the trench. Appeared discontinuous and disappeared in the north-west corner. In the south, the edge of the deposit is concave as though a pit had been cut through it, but there was no obvious change in the consistency of the midden material. In the east the chalk layer ended in a straight kerb aligned NNE-SSW. Chalk dives away to south and east, slightly ‘rubblier’ in places but solid elsewhere. Upper surface very level and with the appearance of having been trampled or smoothed. Clarity of junction good 4 Midden Midden material of variable thickness across the whole trench. Thin lens 0.05m thick at north-east corner widening to the south-west Upper surface again fairly level but with a small depression on the north side of the trench and on the south-east appears to spread across the outer edge of chalk spread. Loose, friable, sandy material, containing much decomposed bone from which the layer may have derived its orange colouring. Lenses of debased bone are spread across the surface of this layer. The possible coprolitic material also occurred in some quantity. Orange sandy layer contains fragments of decayed bone/coprolite-like material stained orange as per soil. Much pottery came from the junction of this layer with context 6 along with the perforated spindlewhorl (Fig 11). THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 45 Table 2: Continued Chalky Marl |South east corner of trench. Hard compact chalk layer. Angular chalk fragments to a maximum 0.15m across with smaller rubble infill. Again, smoothed upper surface but mixed with small amounts of midden material. Felt to be trampled or exposed to elements and only apparent over a small area in the south-east corner of the trench Grey-brown sandy material of even consistency. Contained lenses of very soft decomposed bone. Deposit of variable width and almost absent from the south-western quadrant of the trench Flecks of charcoal appear in places. In fact, the grey colouring is probably due to the charcoal content. Clarity of junction is clear. Human skull fragments and other human fragments in this context. The skull occurred in the middle of the trench immediately adjacent to the southern section and appeared to be deliberately placed with the exterior surface uppermost. This context produced large amounts of pottery and bone as well as an antler tine. Isolated semi-circular lens of midden material to a depth of 0.07m in the southern part of gee trench — did not extend to edge of excavated area. Contained many nodules which were thought to be coprolites. ed Marl | Loose chalk, angular fragments up to 0.15m across. Nodules set amongst loose powder chalky rubble in the south-west corner of the trench, but the chalk appeared more compact towards the west. Approximately 2m from the north-west corner it dipped away leaving a scoop or gulley oat [ae c.lm long extending in a north-south alignment into the section. This was filled with loose chalk rubble but also large flint nodules stained yellow-green with occasional signs of being affected by fire. Slight traces of a level of chalk blocks laid crazy-paving style above the flint. |Clarity of junction clear. A few fragments of bone recovered from this material. |D-shaped deposit of soft and friable midden material within chalk layer 8 was initially considered to be a posthole but later considered to be a localised high point of the underlying Context 4. In the northern part of the trench at midpoint and close to the section. It was semi- circular in outline and appeared to cut throu h Context 8. noted primarily at the south-east end of the geal Sandiviched between two chalk layers, and prised a loose and friable soil with a high panic content. Slatted irregular edges. Chalky Marl |Hard compact chalk layer c. 0.03m thick. Noted only in the south-east corner of the trench and consisting of chalk rubble rammed and tightly packed but sealed and cemented by puddle chalk. Midden Isolated deposit of ?midden material below layer 14. Located in the south-east corner of the trench and to a depth of 0.02m. It had a fibrous texture of a brown/grey colour and contained much charcoal. aa Puddled Puddled chalk layer 0.07m thick across predominantly the south-east area of the trench. Chalk Context 15 sits within a shallow depression on the surface of this layer. ee | Noted principally in the south-east corner of the trench, it had a mixed colour and consistency eee (8 05m thick. Containing much charcoal and ashes. Probabl hearth residue. il Present across whole trench. No consistent thickness varying between 0.01m and 0.24m. Orange/brown in colour but lighter than that in Context 6 from which it is indistinguishable in Decomposed Bone the southern half of the trench. Tip lines of ash and almost decomposed bone visible but not possible to pursue these for more than a few centimetres at a time. At the base of this context there was an extensive layer of decomposed bone. Orange in colour and spread over most of the trench in patches, it was very thin and of one thickness, 0.02cm, almost as if it had been carefully and deliberately laid. Artefactually rich context with much pottery (some decorated), spindle whorls, as well as animal bone and an antler tine, but lack of the material thought to be coprolites. The lower part of the layer appeared to contain small pot fragments than the upper but also a greater quantity of bone. Two sarsen boulders were embedded in this layer, their upper surfaces exposed in the layer above. One at the centre of the trench the second close by. Both were associated with decorated potsherds and the easternmost example had an antler tine associated with it. Clarity of junction at base is clear. Extremely thin spread of material, just lcm thick within layer 18, that appeared to be debased or decomposed bone. Extending extensively across the trench, orange in colour and extremely friable and very difficult to trace and fragmented when touched. It was of one thickness and gave the impression of having been laid down in one, deliberate, episode. Midden material containing more charcoal and ash than Context 18. Of variable thickness and deepest at 0.33m on the south-east with greater clay and stone content in the west. Present across entire trench apart from in the far south-west. In the upper levels of the context, on the north side of the trench, an irregularly-shaped lens of orange clay is clear with some good concentrations of charcoal. Artefactually rich with pottery, animal bone and spindlewhorl, but the potsherds generally rather smaller. A sarsen boulder occurred embedded within the N section at this level and a second smaller boulder just 62 long in the SW corner of the trench. Close to the latter was a large antler beam lying west-east. 46 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE © Table 2: Continued Thin loose chalk layer. Small extent of chalk rubble extending into the trench on the south side close to the south east corner. Overlaid by layer 20 (Fig 6 plan 6). Grey-brown earthy clay-like consistency. Fairly standard width across entire trench at 0.1m thick and a depth of 1.45m below the surface at the southeast corner. Contains potsherds and bone but in smaller fragments that midden material. Human bones present. 25 Posthole Possible posthole or hollow within chalk layer (Context 21); 0.1m wide and to a depth of 0.2m. Filled with grey/black soil derived from Context 20. ce Lens of midden material at western end of the trench. Appears to be the same as Context 20 but underlies Context 22. '27.___|Chalky Marl |Circular patch of rammed chalk. Close to south east corner of the trench. (Fig 6 plan 7). Hearth Burnt layer. Sealed by and covered entirely by Context 27 and consisting eee of blackened Material earth — hearth/oven material? (Fig 6 plan 7). Chalk Lens 24 Ploughsoil/ OLS Posthole (Fig 8) Closes to western section. cn wide at mouth narrowing to a base 15cm wide. It is filled with crumbly dark brown/grey soil and is overlain by Context 24 but cuts into the clay-with-flints at the base of the deposit (Context 36). Small number of potsherds within the fill. (Fig 6 plan 7). Hollow Irregular hollow in south west corner of the trench containing large flint and chalk ‘nodules in —conjoined /fill. Resolved into four intersecting postholes a-d of which d contained a number of large flints postholes that could have been used as packing. All sealed by Context 24. (Fig 6 plan 8). Posthole with no internal stratigraphy filled with dry, crumbly dark grey soil. Small, circular, posthole, 25cm wide at its mouth, See AU See and 28cm deep close to the north eastern Posthole Flat-bottomed, 20cm wide, straight-sided and 10cm aces Posthole with no internal stratigraphy filled with dry, crumbly dark brown-grey soil, cut into the clay-with-flints (Context 36). It is overlain by layer 24. (Fig 6 plan 7). Sandy, light and crumbly deposits containing some charcoal flecks and a little more midden- like situated immediately above the Clay-with-flints. Considered to be a possible lower ploughsoil on the old ground surface. Friable consistency and midden-like in colour and texture with some charcoal staining. Human bones present as well as a bronze tanged chisel. Clay-with- |Clay-with-flints. Ata minimum depth of 1.6m below current ground level. : flints Puddled Puddled chalk feature occurring against and within the west section. (Fig 6 plan 7). Chalk Burnt area below chalk feature 37 (Fig 6 plan 7). Potsherds, stone and animal bone present Hearth Rubble io _Posthole _Posthole with fla base, Containing burnt sarsen Irugments pottery and charcoal, ——_—_ 41 —([Posthole _| 42 Possible een Thin layer of crumbly soil beneath chalk features 27 and 28. Posthole with no internal stratigraphy filled with dry, crumbly dark grey soil. (Fig 6 plan 8). Potter stone animal bone abundant charcoal also present. oe 6 plan 8). Posthole Sea with no internal stratigraphy filled with dry, crumbly dark grey soil and which cut through the hearth of Context 30 and its capping Context 29. Contained two large flints with een ES Hearth, reddened chalk Benen features 27 and 28 another two lying beside it that could be the remnants of packing. (Fig 6 plan 8). Eleven fragments of animal bone were also recovered. The first test pit (Trench A), measuring 2m x cautious method of recording and control. This 1m, was placed in the lee of the plantation a little to allowed, ultimately, the recognition of a number the west of, but close to, the centre of the ‘hill’. In of key layers and contexts outlined below (Table 1 view of the unexpected and dramatically undulating and Figure 5). In order to accommodate access, at a deposits, excavation proceeded in 10cm spits as a depth of 0.8m, it was decided to concentrate efforts THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 47 Trial Trench A Trial Trench B West section North section East section South section Fig. 5 Section drawings. One section in Trench A and four in Trench B were recorded during the excavation. The key letters refer to context/layer numbers detailed in the main text. 48 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE SHALLOW DEPRESSION Plan 2 Plan 3 Plan 5 2) Plan 7 Plan & Fig. 6 Selection of plans recorded within Trench B. Numbers 2-6 are sequential within the deposits, 1.e. plan 2 1s at a higher location with the matrix of deposits. Plans 7 and 8 lie on the old land surface at the base of the mound. Key letters relate to context/ layer classifications within text. a Ig cen me THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 49 on a lm square only, leaving a stepped arrangement with the base of the deposits reached at no less than -a depth of 2.0m. A second test pit (Trench B) was initiated simultaneously within the plantation on the south- western periphery of the mound and in an area _ heavily affected by rabbits and badgers (Figure 4). This trench covered an area of 2m by 0.5m and was only partly excavated. It was covered over and work resumed the following year when it was enlarged to Facing N Facing E 3m by 2m and excavated to a maximum depth of 1.67m. Stratigraphy in this larger area was complex with layers and features displaying little consistency in either extent or depth. As a result Trench B was excavated using layers and features (Table 2) here numbered as sequential contexts (Figures 5, 6 and 7): Discussion of the deposits Although set several metres apart, there was a general 0 20 centimetres A TTA Fig. 7 Profiles of post-holes and other cut features at the base of the deposit. These probably relate to pre-midden structures. Again, labelled numbers relate to context/layer classifications within text. 50 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE match between the deposits encountered in the two Test Pits. In Trench A, significant quantities of material considered to be ‘midden’ debris were recognised to a depth of 1.5m. In places, e.g. Layer 4 in Trench B, these were more bone rich and orange in colour; elsewhere there was a greater charcoal component and the deposits took on a greyish hue. Artefactual debris, including worked bone and a wide variety of pottery forms, was present in all contexts within Trench A, but with noticeably higher quantities in Contexts 2-4 and 6-9. These were separated by a number of chalky lenses, some of them compacted and composed of fine chalk soil and fragments of small chalk lumps. It is noteworthy that the metal artefacts, the broken blade end of a socketed axe as well as a tanged chisel, were recovered from the basal deposit, which may well have been a pre-midden land surface. While Trench B displayed a similar stratigraphy, the slightly larger area allowed more detailed investigation of the deposits. Consequently, it could be determined that the layers of ‘midden’ material often dipped away and were interspersed with patchy lenses that could not always be excavated or recorded satisfactorily on the ground. There is a suspicion that deposits, of what are apparently homogenous material in terms of colour and texture, mask a sequence of depositional events difficult to recognise during excavation. For example, the thin deposit here referred to as debased bone or calcareous matter (19), was present in a number of places and appeared as a dried, slurry-like material of the kind often encountered around farmyards. It was extremely difficult to follow as it dipped and dived, crumbling when trowelled, and appears to represent the detritus of an im situ process having flowed into hollows and dips. The deposits uncovered contrasted with those from Trench A in one respect; while superficial deposits of chalk were recorded in the former, those in Trench B took the form of substantial rammed chalk horizons. In both trenches these separated major episodes of deposition. The difference here may be merely a product of the limited area of Trench A, it perhaps being positioned on the edge of similar horizons recorded elsewhere in the auger survey, but there was some structural variation in the chalk evident in Trench B. Some deposits of chalk, for example Context 3 in Trench B, were relatively loose and friable and covered reasonably large areas, while others, particularly those that had been rammed hard and appeared as cement-like, were restricted to areas of a little more than a metre across. One place, uncovered in the north-east corner of Trench B and depicted in its east section, displayed a sequence of chalk deposits, both loose and solid, that appear to have been disturbed. It is not absolutely clear how this had occurred or what it represents, but it may have involved removal of a feature such as a post. The imprint of a potential post, not recognised in plan, is nevertheless visible in the section within context 14. These small platform-like features appear to be too small to have been living floors, but they nevertheless seem to mark an event requiring a small level base that took place during the longer term build up of ‘midden’ material. In certain cases, contexts (e.g. 27, 28 and 29) appear to have provided capping for burning events. The further significance of the chalk deposits will be considered below. In summary, the mound was composed of complex and diverse archaeological deposits. The layers uncovered varied greatly in terms of colour, texture, and thickness and contained a substantial artefactual assemblage. The most productive contexts in terms of artefacts were 4, 6 and 20 in Trench B, each interleaved with spreads of compact chalk of varying extent and depth. At the base of the mound, a series of post- and stake-holes (Figure 7) cut into the surface of the Clay-with-flints were separated from the mound layers by a well-developed horizon, probably a ploughsoil, that appeared to mark an old land surface. Here, deposits in the respective Test Pits differed slightly in that there was a greater build up of material at this level in Trench A than B. The ploughsoil was of similar thickness in A and B and at a not dissimilar depth below the surface, 1.63 deep in A and 1.45m in B. Sealed by it were what might be described as occupation or settlement features, that is post holes and hearth like features. In Trench A, a pile of burnt flint nodules overlay what appeared to be an earlier ploughsoil some 0.25m in thickness that, in turn, sealed a small stakehole cut into the sandy, friable, loamy surface of the Clay-with-flints and associated with animal bone, potsherds and bronze implements. In Trench B, a similar sandy, friable loam, Context 35, formed the surface of the Clay-with-flints, containing flecks of charcoal and fragments of human bone and here also a second deposit, layer 24, may represent a similar second event. Cut into the Clay-with flints were post holes, each some 0.2m in diameter, filled with dark, loose and friable soil, some containing small amounts of pottery and animal bone (Figures 6, 7 and 8). No sequence could be detected and no structure was identifiable either, although it is expected that 1 } ; i | eR ag A oT IRATE a a eS ae ee THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 51 Fig. 8 A number of cut features including shallow pits and post holes, as well as hearths, were revealed beneath the midden deposit. These almost certainly represent the remains of a pre-midden settlement characterised by post-built structures and associated with decorated post-Deverel Rimbury pottery and the Llyn Fawr phase metalwork. these features continued well into the sections of the trench. Alongside, and in one case cut into by a posthole were small chalk platforms, their surfaces reddened presumably by fire. These latter are termed hearths though they also resembled collapsed ovens. A lack of excessive amounts of charcoal indicates that these features were swept clean and they differed from those further up the midden profile only in having a reddened or burnt component. Excavations on the route of the tank track Having demonstrated the archaeological nature of the deposits at East Chisenbury and provided a context for the discovery of large numbers of pottery sherds in the nearby shelter-belt plantation, the _ military authorities accepted the need to re-route the proposed tank track. The new route approached from the north and then looped in a wide arc at least 150m to the west of the mound around the lip of the spur and thus on three sides. Clearance of the ground was _ supervised by Roy Entwistle and Frances Raymond. Their observations (Entwistle and Raymond, pers. comm.) resulted in the recognition of six cut linear features approaching the mound on the north and south. Those on the north consisted ofa linear ditch and pit alignment: a rare combination of features. Due to time and resource constraints no excavation was undertaken of these features and the only dating evidence from the top of the features were sherds in the 800 to 400 BC date range, although a close association with the midden and enclosure complex is beyond doubt. Contents of the deposits The amount of material recovered from such a small excavation is impressive and is listed here in three tables. Almost 2000 pieces of pottery were found in rabbit scrapes or other unstratified locations (Table 3). While weight rather than artefact counts have greater significance, numbers are listed here as an indication of the quantities of material from respective layers. The brief report below is a signpost to content and there remains much scope for further analytical work. In the case of charcoal, daub and possible coprolite samples, numbers refer to the number of plastic bags containing the material rather than to weight or number of pieces. All soil from both trenches was sieved and thus there is 54 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE. Table 6: Soil samples and Notes on Soil Micromorphology West Section Sample 1. Depth 112-120cm. 5YR3/1-2.5/1 Massive to partially fragmented sediment of type d and discontinuous 1-2cm thick layer of type e fabric — dark, reddish- brown, finely pellety (PPL), low to non-birefringent (XPL), with abundant fine calcite, including <20um-size faecal spherulites, quartz silt/fine sand; inclusions of dark yellowish burned bone, coprolite, shell, rare vesicular silica (ashed phytoliths — ‘fused ash’) and many wood charcoal; type c is very dominantly amorphous organic matter and brown plant tissues, with rare diatoms and occasional phytoliths (humic soil masking?). As with sample 2, suspected use of humic soil/peat as fuel, animal bedding material. Sample 2. Depth 98-106cm. 5YR4/1 Generally massive without obvious layering of mixed type d material — a homogenised version of microfabrics types a and b?; inclusions of yellowish organic stained sediment, ash, phosphatic impregnated sediment (Blue Light), phosphate stained flint; fine bone present. Weathered and biologically homogenised animal and domestic bedding material. Sample 3. Depth 72-80cm. SYR3/1 Dominantly poorly-formed medium prisms of well-preserved stratigraphy, with crumbs and fine blocky; finely banded material similar to type a (sample 4) with again very low humic content, except for some fine bands rich in phytoliths containing occasional to many fine O.M. fragments. Another layer of material (type c) is dominant in the lower half of the slide; it is more brown because it contains abundant charred/aged organic matter, fewer faecal spherulites, but higher amounts of fine calcite (micritic ash?), pale yellowish to clear inclusions of phytolith-rich (many articulated) material with few faecal spherulites, but abundant fine calcitic probable grass ash (probable burned herbivore sheep/goat?) in a sediment containing very abundant (up to 60%) diatoms. Charcoal, chalk and fine bone, shell also occur. Of particular interest are coarse (3mm-1cm) sections through phosphate-replaced plant stems. Some layers as 4 but with more evidence of probable dumping of ashed herbivore coprolites (dung), and/or possible in situ stabling; also includes mineral-replaced plant stems — from cess pit? Sample 4. Depth 52-60cm. 7.5YR3/2 Crumb to fine subangular blocky with relict fine (200-600mm) banding in places; a) very dominant speckled grey (PPL), highly birefringent (XPL) and grey (OIL) caicitic (multiple-sized calcite spherulites; faecal spherulites, calcium oxalate and calcite crystals present) sediment containing very abundant phytoliths and rare diatoms with many amorphous organic matter; b) very few c. 200mm thick bands of very pale speckled yellowish grey (PPL), non-birefringent (XPL), grey (OIL) phytolith-rich (80%) material (occasional amorphous organic matter); coarse mineral inclusions are frequent silt to fine-sand-size quartz, medium sand-size quartz and flint, with few chalk fossils, few biogenic calcite, very few gravel-sized chalk; anthropogenic and organic inclusions are many fine charred organic matter fragments and occasional coarse wood charcoal (e.g. probable oak); occasional bone; many type a) coprolites up to 3-4mm, pale yellowish brown with very abundant plant materials — articulated sheets of phytoliths (up to 2mm), other tissue remains, probable pollen, spores; with possible nematode eggs present; other inclusions within (human type) coprolites are occasional to abundant calcite crystals, calcium oxalates and faecalspherulites and quartz silt; rare to occasional small round pale yellow with no apparent inclusions (possible bird type/guano?); all coprolitic material is BL autofluorescent and, therefore, phosphatic. The soil also contains in situ aged roots, occasional enchytraeid-like excrements with old channels and abundant areas of worm-worked (sometimes mammilated) soil. Heterogeneous midden deposits of possible grass and grass/herb dumps/ occupations and/or ashed sheep/goat dung residues with possibly in situ deposition of (human?) faecal waste. (Sediment made up of domestic bedding material and/or sheep/goat dung residues?) North Section Sample 5. Depth 133-142cm. Junction of SYR4/4 and underlying Clay-with-Flints soil. Loose crumb to subangular blocky mixture of calcitic midden type d and silt rich Clay-with-Flints soil, which is speckled and dotted pale brown (PPL), has low interference colours (XPL), pale yellow brown (OIL), with many fine amorphous organic matter and charcoal; soil contains occasional thin dusty clay void coatings and many fine ferruginous impregnations; soil has a fine channel and vughy porosity and includes some charcoal and bone. Silty soil and midden material is often juxtaposed in large earthworm excrements. Earthworm-worked junction between Clay-with-Flints topsoil (Ah and Eb); occupation effects (coatings) before burial by the midden? Sample 6. Depth 76-85cm. Junction of 5YR4/1 midden and underlying chalk floor. Finely fragmented type a) material with frequent gravel to small stone-size chalk, few yellowish coprolites containing articulated plant tissues, faecal spherulites; phosphate replacing plant tissues also present. Relict layered material (in fragments) at base of slide, both type a and c material; areas of yellowish (phosphate?; BL) staining. As with sample 2, with weathering of deposits giving rise to movement of phosphate or inputs of phosphatic waste? _ THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 3) ae 23 |e f Fig. 9. a: Photomicrograph of typical layered ‘midden’ deposits, mainly composed of mineralogenic plant and dung(?) residues | including weakly humic stained layered articulated phytoliths (LP); also present are charcoal (Ch), rare bone (B), chalk (Ck) and quartz sand (OS). Plane polarised light (PPL), frame width is ~5.5 mm. b: As Fig 1, crossed polarised light (XPL), showing | layered calcitic residues (crystallitic bfabric); layered phytoliths (plant opal) across the centre of the image are non-birefringent. c: Detail of Fig 2, illustrating calcitic ‘spherulites’ contributing to crystallitic bfabric; both calctum oxalate (CO) residues of plant leaves and more numerous calcite faecal spherulites (FS) relict of dung of ruminants such as sheep and goats, are present. Frame width is ~0.33mm. d: Photomicrograph of probable human coprolite/cess, probably composed of calcium phosphate, with typical vughs (voids) and layered articulated phytoliths (AP). PPL, frame width is ~5.5 mm. e: Detail of Fig 4, showing long lengths of layered articulated phytoliths (AP), likely residues of ingested cereal material such as bran; large arrow locates area of Fig 6. Frame width is ~1.1mm. f: Detail of Fig 5 (large arrow), blue light fluorescence microscopy; void (resin) is not autofluorescent, but coprolite/cess with phosphate embedded layered phytoliths is highly autofluorescent implying the likely presence of hydroxyapatite. 56 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE | Results The midden components are exceptionally well preserved and therefore a great many microfabric types and materials can be recognised that are worthy of interpretation. Of particular note are the large number of faecal spherulites (Brochier et al. 1992; Canti 1997, 1999) present in this apparently mainly ashed deposit, that also contain calcite ash and calcium oxalate crystals (Figure 9a-c). This type of material is also exceptionally rare, with apart from Mediterranean analogues, the only example being from a farm mound-like deposit of Middle Saxon date at West Heslerton, North Yorkshire, that appears to contain ashed sheep/goat coprolites (Macphail et al. forthcoming). Much of the Chisenbury material is highly calcareous, much more so than the deposit at Potterne (Macphail in Lawson 2000). The Potterne deposit also seems to contain a generally far higher proportion of quartz sand soil (Greensand) compared to East Chisenbury. Both these differences could relate to the contrasting parent materials of the two sites, East Chisenbury being on the Chalk. Even so, whilst Clay-with-flints material is present at the base of the mound, little clay and other soil mineral matter are apparently mixed into the fine matrix of the deposit compared with Potterne. On the other hand, possible trampling features were noted at the Clay-with-flints/‘midden’ junction (sample 5) and fragments of a peaty soil occur as inclusions in sample 1. The differences between the two deposits could relate to both weathering and to real variations in the composition of the deposits. Firstly, it has to be noted that calcareous residues are more likely to be preserved in a base-rich chalk environment. Secondly, East Chisenbury seems to have a different make-up compared with the deposit at Potterne, and, for example, includes no ‘fused soil’ (commonly phosphate cemented soil) and few ‘fused ash’ (e.g. sample 1) and contrasts also in the compacted chalk surfaces, which are absent from Potterne. On the other hand, both sites contain ubiquitous coprolitic residues (Fig 9 d-f) and material termed ‘pale nodules’ that appeared to be human coprolites. The last were thoroughly studied at Potterne and may have a cess-pit origin. Similar amounts of phosphate were found at Chisenbury, compared to Potterne (see below) and clearly inputs of cess/human coprolites were contributory (cf. phosphate at Norman period, Monkton cess pit). In comparison, the remains of a Late Bronze Age/Early Iron Age settlement soil at Salford, Bedfordshire (Dawson 2005), although disturbed and accretionary, show little chemical and micromorphological evidence of being formed in the same way as the deposit at East Chisenbury. Another main difference between the two deposits relates to the composition of the fine fabric. Whereas the main components of the fine fabric at Potterne are fine sand-size quartz and glauconite, fine charred and amorphous organic matter, phytoliths (and some diatoms) with a little fine calcitic ash, at East Chisenbury the most dominant component in places are calcitic ‘spherulites’. Spherulites, sensu lato are crystals with an obvious optical cross under XPL (Courty et al. 1989, plate 3f). At East Chisenbury there are both calcium oxalates that form naturally in plants (such as druses) and more commonly, <20 uwm-size spherulites, sometimes termed ‘faecal spherulites’ of calcite that are commonly residual of sheep/goat dung (Brochier 1983 and Brochier et al. 1992; Canti 1997, 1999)(Fig. 9c). In addition, the numbers of phytoliths are also higher or concentrated into bands (Fig 9a-b). Fine calcitic grass ash is also present, as well as some layers being dominated by diatoms. The origins of such fine layering needs to be investigated, for example, surfaces and how they weather under exposed and ‘protected’ circumstances have been studied from tell sites as well as in caves (Matthews and Postgate 1994: Matthews 1995). Two major calcareous sediment types were observed: 1. Type a, is composed of phytoliths, calcium oxalates (from leaves and roots) and faecal spherulites, with thin interbedded layers of type b, which is composed purely of phytoliths. Such material is exactly comparable to the sediments from the Neolithic cave of Arene Candide, Liguria, Italy, which have been interpreted as resulting from domestic occupation — formed from the slow oxidation of layers of grass (type b) and herb/grass human bedding. Such material could have been dumped, or relate to in situ occupation, the presence of faecal spherulites simply testifying to their ubiquity in calcitic environments, where sheep/goat dung is present, for example. 2. ‘Typec, is composed ofa similar calcium oxalate/ faecal spherulite mixture, but also contains high amounts of fine grass ash and inclusions of what appear to be ashed herbivore coprolites, possibly those of sheep/goat. To some authors just the presence of ’spherulites’ is enough to confirm the presence of sheep/goat dung (Brochier _ THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 57 1983, 1992), but here, at West Heslerton and in Mediterranean caves (e.g. Arene Candide) there are also fragments of sheep/goat coprolites as defined by Courty et al. (1989; 1992). This type of material also contains some layers with extremely high numbers of diatoms. Diatoms are often associated with herbivore stabling, because it contains trampled soil from muddy puddles, watering holes, and is ingested during drinking (cf. Deansway, Macphail 1994; see arguments for herbivores at Potterne where [soil and freshwater] diatoms were similarly identified). At East Chisenbury, there is also the presence of partially burned peaty material (sample 6); possibly the diatom-rich material could come from a similar ‘wet’ source, but diatom identifications would be necessary for such an interpretation. At Potterne, anthropogenic components and pedofeatures were semi- quantitatively counted so that the impact of earthworm working (burrows and biogenic calcite) and animal trampling (dark, dusty clay coatings etc.) could be assessed (Macphail in Lawson 2000). The brief assessment of the East Chisenbury thin sections showed that fine stratigraphy was preserved in places although earthworms had produced totally reworked soils in others — a situation found commonly as at Arene Candide. There appears to be a lack of textural features at East Chisenbury, except for those found in the basal ‘occupation soil’ (Clay-with-flints interface; sample 5). The apparent paucity of textural features in the deposit generally, however, suggests that there was possibly much less trampling by animals which produce large amounts of liquid waste (e.g. cattle). The other features of (cattle?) concentrated stocking/stabling at Potterne, namely, phosphate-cemented ‘fused soil’ (see below) is also apparently absent. | Comparison of East Chisenbury with Potterne Although there are a large number of similarities at the field archaeological scale, at the microstratigraphic level, the microfabrics deposits are quite different at the sampled locations, even whilst they may contain similar types of domestic waste inclusions. Potterne is more organic and less calcareous. It also seems to contain far less well-preserved stratigraphy at the finest level. Whereas it has been argued that cattle dung (and possibly also pig dung) and related stabling material are major components at Potterne (hence large quantities of included soil and the presence of phosphatic solutions), at East Chisenbury one interpretation of the abundant phytoliths present may suggest that sheep/goat dung is the prime domestic animal waste material being added to the deposit here. No sheep/goat dung was identified at Potterne, although animal bone evidence indicates that it became more dominant upwards in the sequence. However, if the phytolith- calcium oxalate- faecal spherulite mixed deposits do relate to human bedding, this is the first time it has been identified in the British Isles (Macphail et al. 1997). Again, the lack of textural features (dusty clay coatings) at East Chisenbury indicates that (animal) trampling was not a significant factor in the site formation process and is a possibility that the compact chalk surfaces here could have been laid in order to avoid sheep hooves stirring up the deposit (see Macphail and Goldberg 1995 and Macphail et al. 2004 for Butser Ancient Farm stabling comparisons). Oxidation of once-organic remains of dung-rich deposits would lead to the preferential preservation of the mineral residues, hence the importance of phytoliths, calcium oxalates and faecal spherulites. The possibility that a higher proportion of the midden deposit at Chisenbury is unburned, compared to Potterne, may be reflected in the relatively high %LOI measured (8.8% LOI; see below). Conclusions This assessment (and it must be emphasised that very much more work needs to be done), poses as many questions as answers. Nevertheless, it is quite clear that the well-preserved microfabrics and components at East Chisenbury contain a large amount of potential information, but it would be necessary to carry out greater detailed study such as was undertaken at Potterne to confirm the tentative conclusions in this assessment. Still, this has shown that the deposit at East Chisenbury is very well preserved indeed, equalling microstratified deposits that are normally only preserved in ‘protected’ contexts, and more common to Mediterranean caves and tells of the Middle and Near East (Goldberg and Macphail, 2006). The Plant Remains, by Wendy 7. Carruthers © During a visit to the site by the author in April 58 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 1993, five Sltr soil samples were taken from the east section of Trench B. Both ashy and red-brown soil layers were sampled in a vertical sequence down through the east section. Well-preserved charred and mineralised (see Green 1979: Carruthers 2000) plant remains were frequent in all of the samples, as can be seen in Table 7. Considering that only a few, small, assessment samples were examined, Table 7: charred and mineralised plant remains from East Chisenbury assessment samples Triticum cf. aestivum-type (free-threshing bread-type wheat grain) l Triticum spelta L. (spelt glume base) Primulaceae (scarlet pimpernel etc. seed) nN aed nm Es Sie —_ ene Ba Pale Fate Biba ex Eat Ee Sil ie ele a as al SS SE HHS Sfp NI} NO — SS Cena Sherardia arvensis L. (field madder nutlet) AD Nutrient-rich soil indicators Erica dat otnangneieasheneyCDR I arophviaeee Crave apetmoneearsea Da | i HAyoscyamus niger L. (henbane seed) Dn 1 | THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 59 Grasslands, disturbed ground etc. _KEY : Habitat Preferences _* = plant with economic use (Context er A en eee ela as olan Taxa (contd.) sample no. 2 Ea MINERALISED PLANT REMAINS (contd.) z) iS) S = SS io = = ~ = Ss & = Q is} ise) =. 2 = = = Ss wn = w~ (ay ion Sc Come (oa (@) isn) QO = “o) pals) Q i=" (@) 3 (@) wa” ie) >) — S Sa ~~ ~ OI as = ° = y 3 a gS . : au. Q (2) 9) —= (oon a gq ie) 3 ie) Cc n _ 3 jo) Coa ¢ = (a) (a) Qa. pet) Q a io) =) (a) YS’ Qe o 7) See as Fumaria sp. (fumitory) Lamium — type (dead-nettle nutlet) CDY Carex sp. (sedge) GMPw fo) & = & = S i= : a g =. C Q & | 3 3 oO = QO mH gq e) = : (Se io) Lom) ue) — .@) fel) n : e len} (a) n Q (a) jon Ww a s) W| Wl “i CO} bw — | N ioe) n = nutrient-rich soils D = disturbed fe es ee —_— — — — BSS N SN i—) “J — i=) Pele erled| eo eeat | Bae Ee a NM Nn RO] SO] Ww = MN =) M = marsh, bog P = ponds, ditches etc. | c = calcareous soils cS sandy soils S = scrub | A = arable G = grassland W = woodland _d = dry soils w = damp/wet soils Y = waysides | C = cultivated a surprisingly wide range of plant remains was i , recorded. The results are discussed below by making comparisons with the plant remains from Potterne, Wiltshire (Carruthers 2000: Straker 2000). There _ are interesting similarities and differences between _ the assemblages of the two sites that may provide an | insight into their origins. i | H = hedgerow The Charred Plant Remains Charred plant remains were present in all five samples, with fewer seeds being present in the visibly ashy sample (sample 4, context 15) than the red-brown soil samples. Cereal grains, chaff fragments (emmer/spelt glume bases and straw culm nodes) and a few common weeds of cultivated and 60 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE disturbed soils were recorded (e.g. cleavers (Galium aparine), black bindweed (Fallopia convolvulus)). The cereal crops represented were free-threshing bread- type wheat (Jiiticum cf. aestivum-type), spelt wheat (T. spelta, confirmed by the identification of glume bases) and barley (Hordeum sp., probably hulled six-row barley, but too poorly preserved to confirm the identification). Emmer wheat was also probably present but the majority of glume bases could not be identified beyond emmer/spelt (7: dicoccum/spelta). Traces of possible rye, oat awn and a large legume fragment (most likely Celtic bean) hint at other taxa that could have been cultivated or growing as weeds. Although it is difficult to characterise such small assemblages (67 charred remains in total), the frequent charcoal fragments, fine ash and charred plant remains are indicative of mixed domestic waste, perhaps the sweepings from hearths and bonfires. The hulled wheat chaff fragments may have come from small-scale, piecemeal final processing of spikelets prior to cooking. Because hulled wheats were probably stored and transported in spikelet form (Hillman 1981), the presence of chaff on the site does not help to determine whether the grain was being grown locally, or whether it had been brought to the site for feasting purposes. The mineralised remains (see below) demonstrated that straw had been deposited in the midden, and soil micromorphology (Macphail this report) has indicated that much of the material was derived from sheep/goat ‘stable’ waste, in this case, waste from sheep pens or byres. The charred cereal remains, therefore, could also have come from this type of waste, perhaps representing burnt fodder, animal bedding or dung. Apart from the large legume fragment, all of the cereals at East Chisenbury were also recovered from the midden-like deposit at Potterne. Although it is difficult to make comparisons between the large number of samples from Potterne and these five assessment samples, it is interesting to note that three of the eight identifiable grains were of the free-threshing bread wheat type, a crop that is not common on sites of this period and that was present only in small numbers throughout the deposit at Potterne (Straker 2000). There is some indication from other sites that crops such as bread wheat are more likely to occur in ritual deposits than as general waste. This could be because free-threshing wheats are under-represented in general waste deposits, being less likely to become charred during processing (since they do not require parching to free the grain from the chaff), or it may be because free- threshing wheat was more highly valued and so was more likely to be used for special occasions. Clearly, larger scale sampling is needed at East Chisenbury before statistically valid comparisons can be made with the Potterne data. In addition to the charred plant remains, more fragments of silicified ash were present in the East Chisenbury samples than at Potterne. The apparent increased occurrence of this delicate material could be due to less trampling of the deposit at East Chisenbury (this report above). Identifiable material was not recovered from the small assessment samples. Where large concretions of silicified plant material are present, however, they can be a useful source of information concerning cereal processing, as they often comprise burnt chaff (glumes, awns etc.) (Robinson and Straker 1991). The Mineralised Plant Remains As at Potterne, mineralised plant remains were present in all of the samples and the quantities varied rather erratically, indicating that preservation had probably occurred in small pockets of dumped organic waste (Carruthers 2000). The wide variations in seed concentrations from 14 to 260 remains per Sltr makes comparisons with Potterne difficult, but the average figure was c. 200 remains per 10ltr compared to c. 50 fragments at Potterne. The occurrence of mineralised rootlets at Potterne and an encrusted mineralised layer confirmed that mineralised preservation had occurred in situ. At East Chisenbury the site of mineralisation was more difficult to determine, as no rootlets or visible mineralised layer were present. Slight differences in the conditions of preservation could have prevented rootlets from becoming preserved at East Chisenbury. The fact that only dicotyledonous rootlets (broad-leaved plants) appear to have become mineralised at Potterne, even though monocotyledonous plants (grasses etc.) were obviously present (see explanations in McCobb et al., 2003) demonstrates that this type of preservation can be dependent on very specific conditions being met. Differences in the rate of deposition of waste material on the midden or the porosity of the soil could have prevented rootlets from becoming preserved at East Chisenbury. More samples need to be examined to clarify this difference. At both sites the assemblages were dominated by the same range of nutrient-rich soil indicators, such as stinging nettles (Urtica dioica), Chenopodiaceae and henbane (Hyoscyanus niger), i.e. the type of vegetation that would have grown on a midden-like deposit, thus THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 61 -making it likely that 7m sztw mineralisation had also taken place at East Chisenbury. For mineralisation to have taken place, high concentrations of decaying organic waste must have been present, producing _very moist, nutrient-rich conditions at least in small pockets throughout the deposit (Carruthers 2000). The presence of frequent mineralised fly puparia, fly eggs, worm cocoons and millipede segments in _the samples provided further evidence for the highly organic and probably smelly conditions. However, as anyone with a compost heap knows, the odours from composting vegetation can be dampened by adding bonfire ash, and smells soon die away once composting begins. The very visible stratigraphy in East Chisenbury Trench B, with soil layers -alternating with ashy layers and chalk, may indicate the deliberate dampening of odours, soaking-up of | liquids and the creation of hard-standing areas. | Although the high occurrence of nitrophilous taxa showed that East Chisenbury was similar to Potterne in some ways, other mineralised remains demonstrated that there were important differences between the sites. Of course, it should be noted that only minute fractions of each site have been examined, so the data from both sites may not be ‘representative of the whole picture, particularly East Chisenbury. The initial impressions from this assessment suggest, however, that more remains that are likely to have originated from human faecal material were present in the East Chisenbury samples. At Potterne, the only remains from edible | taxa (e.g. crab apple pips, Malus sylvestris) that ‘may have come from human faeces was present in the pre-deposit features but not the deposit | itself. Although over 800ltr of soil were processed for mineralised plant remains, only 3 apple/pear _pips were recovered from the site as a whole. No mineralised concretions containing cereal bran (a good indicator of faecal material) were recovered |from Potterne. In contrast, in the 25ltr of soil processed from East Chisenbury, 3 apple/pear seed embryos (Malus sylvestris/Pyrus communis) were | recorded and 3 concretions containing cereal bran were found. Seeds of poppy (Papaver sp.), either from the field or opium poppy are numerous. These | and the mustard/charlock (Brassica/Sinapis sp.) may | also have been eaten as flavourings, but this is more | difficult to prove because they also occur as arable ‘weeds. Both species of poppy were introduced to | Britain, originally as contaminants of cereal crops. Mineralised byre and penning waste concretions containing matted straw fragments were fairly frequent in sample 2, context 11, and these were not recovered from Potterne. Whether these differences relate to better preservation conditions at East Chisenbury or the deposition of different types of waste could be addressed in future by a carefully designed sampling programme. Conclusions The assessment has shown that even small samples from midden-type deposits can provide a wide range of charred and mineralised plant material, including important direct evidence of diet from a period that does not usually produce this type of data. Nodular accretions and coprolites Nodules of various forms and sizes were encountered from a number of levels during the excavations (see Tables 4 and 5 also ‘pale nodules’ in Macphail above). Some of these were rather amorphous and in some cases it was difficult to distinguish whether the material concerned was daub residue, debased bone or indeed a natural deposit. Other pieces appeared to be more recognisable as human or dog faeces. Well preserved pieces, i.e. those taking on good recognisable form as stools were considered to be coprolites and were sent to the Department of Environmental Science at Bradford University for analysis by James Ottaway, who unfortunately, died before the assessment was completed and the samples have now been lost. Ottaway also took samples from an augered hole 5m from the east end of Trench B. Other more amorphous pieces from Context 6 in Trench B were initially submitted to Dr John Penn of Kingston University, Faculty of Science, who determined that there were two distinct materials. The first was bone, in which case it was badly decomposed, the second a fine grained granular material which tests with trisodium phospahate indicated was not coprolitic. Calcareous nodular accretions, by Shahbaz Ashraf Further less recognisable samples were provided for analysis to Shahbaz Ashraf, based at the University of East London. A total of sixteen samples were analysed using a variety of techniques. Initially, basic soil tests were applied to derive each sample’s carbonated, phosphate, humic acid, and pH content. 62 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE After this, samples were analysed for their elemental composition by X-ray Fluorescence (XRF). Thin- layer chromatography (TLC) was then used to analyse the sample’s organic components, which were extracted from them by solvent extraction using hexane. Flame Photometry was then used to measure the calcium and potassium concentrations of the samples and, finally, the scanning electron microscope (SEM) was used for a closer physical analysis of the soils. All samples were carefully analysed for the presence of anything which may derive from coprolites, e.g. seeds, roots, leaf particles, bone fragments, feathers, etc. No real evidence for these indicators was found. Instead, a small number of samples were found to contain other organic fragments. The main test for coprolites is the presence of coprostanol. Coprostanol appeared in low levels in two samples, Trench A soil sample 7 and Trench B soil sample 3. This may be enough to suggest that these particular samples are coprolitic. The TLC results indicated that at least six of the samples showed the presence of cholesterol, but as this is now known to occur widely in many non-animal components as well as coprolites, its discovery in the East Chisenbury samples is not conclusive. A single hair was found in Trench B sample 1. Examples of various known animal and human hair were examined in order to determine the origin of this hair. It had medium-sized spacing between the edges of its overlapping scales, and was not very fine. Therefore, cat, cow, and horse hair were immediately dismissed as their hair tends to very thin with small spacing between the scales. The unknown hair also had a very clear indentation running along it, and so, animal hair known to be completely rounded (such as mohair and wool) was dismissed too. The indentation also ruled out a dog or a human as the source of the hair. The closest match was that of rabbit. It may well be that this is a modern contaminant, though the possibility that it was hare, a species present in prehistoric Britain, was at the time not considered. Results obtained from the XRF analysis showed high levels of calcium in all the samples. Furthermore, flame photometry analysis indicated amounts of sodium within almost all samples. This, therefore, suggests that a number of the samples analysed may well be wood ash concretions, but the rest remains enigmatic. General observations from this study fail to provide evidence to suggest that any of the analysed samples were of a coprolitic origin. The particle size of the samples, although revealing clay-like material matching other known coprolites, also did not indicate anything to suggest a coprolitic origin, such as vegetable matter. Furthermore, none of the samples showed any yellow precipitate to indicate the presence of phosphate. XRF analysis also supports this conclusion as, although most of the samples showed phosphorus, not one sample produced the high peaks as would be expected from coprolite or bone residues. Animal Bone, by Dale Serjeantson, Joanne Bagust and Caroline Jenkins The quantity of animal bone recovered from the two trenches was surprisingly large in view of the fact that only a very small percentage (0.02 per cent) of the mound was excavated. This account is based on an unpublished report by Jo Bagust, a study of the age at death of the cattle and sheep by Caroline Jenkins and a consideration of the animal husbandry at East Chisenbury in the context of contemporary settlements in the south of England (Bagust nd: Jenkins nd: Serjeantson 2007). It will focus on the meat yield from the mound, the management of the domestic animals, and the possibility that some of the deposition of bone may derive not from rubbish disposal but from ritual practices. All the excavated deposits were sieved using 4 mm mesh, so the bones examined include all significant fragments. The only bones which might have been missed are of small birds and small fish, neither of which would be expected on Downland sites of the 1st millennium BC in southern Britain. The remains are very well preserved, though some, like those from Potterne, had concretions or encrustations on the surface. They are very fragmented, mostly from butchery and dog gnawing, something which was perceptively observed by Walls on the animal remains he excavated in 1945. Just below 30 per cent were identified to species, a slightly higher percentage than on some prehistoric sites. A few were found in articulation. Fifty-four tools and other objects of bone and antler were also found, a proportion which is unusually large in relation to other material excavated, as Morris (below) points out. Animals and quantity of meat The number of identified bones (NISP) shows that THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 63 _ Table 8: Species Representation Trench A SPECIES ) [Sheep/goat|609_[54% [55 [66.3% _ 2) sheep were the most important domestic animal -at East Chisenbury and that the other domestic animals, cattle, pig, horse, and goat, were also eaten. Bones as well as antler of red and roe deer were found in both trenches, but dog remains were only found in Trench B (Table 8). The percentage of pigs is lower than at Potterne, Wallingford and Runnymede, all | sites with unusually high numbers of pigs, but it is higher than at most Late Bronze Age and Iron _Age sites (Serjeantson 2007). A single bird bone | was found, probably of mallard; as birds were not eaten regularly at this time, the best explanation for the duck is that its feathers were used for personal | decoration. | The minimum number of animals (MNI) | Table 9: Species Representation Extrapolated meat Yields for the Whole Site and Per | Year, (assuming that 0.02% of the mound was excavated and that the site was used for 100 years-see below. Sheep _ divided into adults and lambs, since one-third were below six months). Even if these figures greatly exaggerate the : quantity of meat available, it must nevertheless have been enormous. was calculated for each trench based on the most frequent element. These figures summed were the basis for an estimation of the number of animals in the whole mound and an extrapolation was made to show the quantity of available meat. The age at death of the sheep, discussed below, shows that one third of the sheep were under six months of age. As these would have given less meat than adult sheep, they are shown separately (Table 9). Dressed carcass weights are based on data in Dahl and Hjort (1976) and Newton (1983). One characteristic of MNI is that it declines in relation to the number of identified fragments, so these calculations must overestimate the numbers. Further, the assemblage comes from a very small sample of the mound, which, it is worth pointing out, may not be typical of the whole. However, even after the problems with MNI have been taken into account, the figures do suggest that a very large number of animals were consumed at East Chisenbury. The meat was not, however, the only food, which the animals provided; the sheep and cattle also provided milk and cheese. Sheep husbandry The eruption and wear on the mandibles was recorded in order to reconstruct the husbandry practices of sheep and cattle (Jenkins nd). None of the juvenile mandibles of sheep and goat were in fact from goat, as none had the accessory pillar. Not all mandibles were complete so those mandibles and teeth that would fit into several stages were assigned to an age stage in proportion to the more complete ones (Payne 1973, 293-296). There are clear age biases in the sheep: more than one third are young lambs (age stages A and B) and the remainder are young and mature animals between two and four years; no jaws were found from animals over six years of age (Table 10). Payne (1973) suggested that age stage B lasted up to the age of six months, but recent work (Jones 2005) shows that it probably covered a shorter period of two to four months. As the lambs would have been born in April, the deaths would have occurred in June and July, i.e. summer. One contemporary site, Runnymede, has a similarly high percentage of lambs culled or dead at the same very young age (Serjeantson 2007), but at most contemporary and later sites such as Ashville (Wilson 1978), Danebury (Grant 1984, Grant et al. 1991), Grimes Graves (Legge 1981; 1992), survival after four months is over 80 per cent. There are several possible explanations for the high percentage of lambs at East Chisenbury and Runnymede. One is that an unusually large 62 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE After this, samples were analysed for their elemental composition by X-ray Fluorescence (XRF). Thin- layer chromatography (TLC) was then used to analyse the sample’s organic components, which were extracted from them by solvent extraction using hexane. Flame Photometry was then used to measure the calcium and potassium concentrations of the samples and, finally, the scanning electron microscope (SEM) was used for a closer physical analysis of the soils. All samples were carefully analysed for the presence of anything which may derive from coprolites, e.g. seeds, roots, leaf particles, bone fragments, feathers, etc. No real evidence for these indicators was found. Instead, a small number of samples were found to contain other organic fragments. The main test for coprolites is the presence of coprostanol. Coprostanol appeared in low levels in two samples, Trench A soil sample 7 and Trench B soil sample 3. This may be enough to suggest that these particular samples are coprolitic. The TLC results indicated that at least six of the samples showed the presence of cholesterol, but as this is now known to occur widely in many non-animal components as well as coprolites, its discovery in the East Chisenbury samples is not conclusive. A single hair was found in Trench B sample 1. Examples of various known animal and human hair were examined in order to determine the origin of this hair. It had medium-sized spacing between the edges of its overlapping scales, and was not very fine. Therefore, cat, cow, and horse hair were immediately dismissed as their hair tends to very thin with small spacing between the scales. The unknown hair also had a very clear indentation running along it, and so, animal hair known to be completely rounded (such as mohair and wool) was dismissed too. The indentation also ruled out a dog or a human as the source of the hair. The closest match was that of rabbit. It may well be that this is a modern contaminant, though the possibility that it was hare, a species present in prehistoric Britain, was at the time not considered. Results obtained from the XRF analysis showed high levels of calcium in all the samples. Furthermore, flame photometry analysis indicated amounts of sodium within almost all samples. This, therefore, suggests that a number of the samples analysed may well be wood ash concretions, but the rest remains enigmatic. General observations from this study fail to provide evidence to suggest that any of the analysed samples were of a coprolitic origin. The particle size of the samples, although revealing clay-like material matching other known coprolites, also did not indicate anything to suggest a coprolitic origin, such as vegetable matter. Furthermore, none of the samples showed any yellow precipitate to indicate the presence of phosphate. XRF analysis also supports this conclusion as, although most of the samples showed phosphorus, not one sample produced the high peaks as would be expected from coprolite or bone residues. Animal Bone, by Dale Serjeantson, Joanne Bagust and Caroline Fenkins The quantity of animal bone recovered from the two trenches was surprisingly large in view of the fact that only a very small percentage (0.02 per cent) of the mound was excavated. This account is based on an unpublished report by Jo Bagust, a study of the age at death of the cattle and sheep by Caroline Jenkins and a consideration of the animal husbandry at East Chisenbury in the context of contemporary settlements in the south of England (Bagust nd: Jenkins nd: Serjeantson 2007). It will focus on the meat yield from the mound, the management of the domestic animals, and the possibility that some of the deposition of bone may derive not from rubbish disposal but from ritual practices. All the excavated deposits were sieved using 4 mm mesh, so the bones examined include all significant fragments. The only bones which might have been missed are of small birds and small fish, neither of which would be expected on Downland sites of the 1st millennium BC in southern Britain. The remains are very well preserved, though some, like those from Potterne, had concretions or encrustations on the surface. They are very fragmented, mostly from butchery and dog gnawing, something which was perceptively observed by Walls on the animal remains he excavated in 1945. Just below 30 per cent were identified to species, a slightly higher percentage than on some prehistoric sites. A few were found in articulation. Fifty-four tools and other objects of bone and antler were also found, a proportion which is unusually large in relation to other material excavated, as Morris (below) points out. Animals and quantity of meat The number of identified bones (NISP) shows that Se eee speprnrerensssapeenscceeneepeeniersee | Year, (assuming that 0.02% of the mound was excavated | MNI 1 Zoey eee OR | Total weight (kg) _ [3825000 THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 63 Table 8: Species Representation Trench A 3.2% Cow om |23:5%e)/5 Se T61% Mie enn | 440 | 13169604 12-99% Trench B MNI WN [Sheep/goat|609 [5 Goat ——([5_———([0.44% ; eae nm Bl W|N]W]oo SIAR RS SNES Rl || OO |b NI 1p [00 19° SS |D |W Oo | She ISSISSISSISSIISS [Duck _—_{1_——*([0.08% [1 1.2% sheep were the most important domestic animal | at East Chisenbury and that the other domestic animals, cattle, pig, horse, and goat, were also eaten. Bones as well as antler of red and roe deer were found in both trenches, but dog remains were only found in Trench B (Table 8). The percentage of pigs is lower than at Potterne, Wallingford and Runnymede, all sites with unusually high numbers of pigs, but it is higher than at most Late Bronze Age and Iron Age sites (Serjeantson 2007). A single bird bone was found, probably of mallard; as birds were not | eaten regularly at this time, the best explanation for the duck is that its feathers were used for personal decoration. The minimum number of animals (MNI) Table 9: Species Representation i i Extrapolated meat Yields for the Whole Site and Per and that the site was used for 100 years-see below. Sheep divided into adults and lambs, since one-third were below Six months). Even if these figures greatly exaggerate the quantity of meat available, it must nevertheless have been enormous. was calculated for each trench based on the most frequent element. These figures summed were the basis for an estimation of the number of animals in the whole mound and an extrapolation was made to show the quantity of available meat. The age at death of the sheep, discussed below, shows that one third of the sheep were under six months of age. As these would have given less meat than adult sheep, they are shown separately (Table 9). Dressed carcass weights are based on data in Dahl and Hjort (1976) and Newton (1983). One characteristic of MNI is that it declines in relation to the number of identified fragments, so these calculations must overestimate the numbers. Further, the assemblage comes from a very small sample of the mound, which, it is worth pointing out, may not be typical of the whole. However, even after the problems with MNI have been taken into account, the figures do suggest that a very large number of animals were consumed at East Chisenbury. The meat was not, however, the only food, which the animals provided; the sheep and cattle also provided milk and cheese. Sheep husbandry The eruption and wear on the mandibles was recorded in order to reconstruct the husbandry practices of sheep and cattle (Jenkins nd). None of the juvenile mandibles of sheep and goat were in fact from goat, as none had the accessory pillar. Not all mandibles were complete so those mandibles and teeth that would fit into several stages were assigned to an age stage in proportion to the more complete ones (Payne 1973, 293-296). There are clear age biases in the sheep: more than one third are young lambs (age stages A and B) and the remainder are young and mature animals between two and four years; no jaws were found from animals over six years of age (Table 10). Payne (1973) suggested that age stage B lasted up to the age of six months, but recent work (Jones 2005) shows that it probably covered a shorter period of two to four months. As the lambs would have been born in April, the deaths would have occurred in June and July, i.e. summer. One contemporary site, Runnymede, has a similarly high percentage of lambs culled or dead at the same very young age (Serjeantson 2007), but at most contemporary and later sites such as Ashville (Wilson 1978), Danebury (Grant 1984, Grant et al. 1991), Grimes Graves (Legge 1981; 1992), survival after four months is over 80 per cent. There are several possible explanations for the high percentage of lambs at East Chisenbury and Runnymede. One is that an unusually large 64 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 10: Age of Sheep at East Chisenbury A B to 4 C 2 2 D 4 6 E 5 5 F 6 7 G 4 4 H 0 0 I 0 0 4 5 > i number died natural deaths. Shepherds expect a natural mortality of about 10 per cent in the first few weeks of life, and this probably accounts for the small percentage dead in the first two months (age stage A), but environmental or climatic conditions are unlikely to have been particularly unfavourable or we should expect the same pattern at other sites and with cattle and pigs. It could, however, be a result of penning a large number of lambs together at the site. The analysis of the plant remains and the soils make it clear that animals, mainly sheep, were indeed penned at the site, which would make them vulnerable to disease. Grant considered that lambing took place within the hillfort at Danebury, and that the remains were of lambs which had died there. However, penning also took place at Potterne, and there most of the remains are from rather older lambs. The large numbers at East Chisenbury which were dead at four months would suggest a very low standard of husbandry among prehistoric shepherds, and it is unlikely that prehistoric domesticated sheep would have been neglected in a way likely to cause unwanted casualties. An alternative explanation is that the slaughter of lambs was deliberate, and this seems more likely. In flocks of sheep milked intensively, the peak age at which male lambs not required for the flock are slaughtered is 2-4 months (Payne 1973; Ryder 1993; Serjeantson 2007). After that age the ewes continue to give milk for a further 2-3 months. This would provide an economic explanation for the early cull. It does however suggest that ewes were milked more intensively at East Chisenbury (and Runnymede) than elsewhere in southern Britain. We are familiar with the milking of cattle in prehistoric Britain, but the fact that sheep too must have been milked is less familiar, though it is accepted for other parts of Europe. The intensive management for milk proposed here, in which lambs are slaughtered so that the ewes can be milked, is a strategy that involves risk, but this is lessened if the site is the centre of a network of settlements. The scale of the PAYNE |APPROX. AGE RAW COUNT CORR COUNT! FINAL CORR COUNT midden mound and its associated earthworks makes it very likely that East Chisenbury was the focus of a wide network of settlements, as argued elsewhere in this report. | It also raises the question of why the male lambs were not retained as wethers to provide wool. During the Bronze Age the wool of sheep underwent the change from ‘hairy’ to ‘woolly’ type, making sheep and their coats much more desirable. That wool and weaving were very important at East Chisenbury is emphasised by the spindle whorls found, which made up 16 per cent of all small finds. The greatly increased importance of wool from the Bronze Age onwards is reflected in the larger numbers of sheep found on settlements from this period onwards. The young age at which sheep were slaughtered at East Chisenbury indicates that older animals providing wool were killed elsewhere. Cattle Cattle remains from East Chisenbury were sparse when compared to those of sheep; there are fewer than one third as many as sheep and goats together (Table 8). Most of the remains are from the body and they include some articulated cattle joints. The few jaws and teeth, only nine of cattle compared to 40 of sheep, suggest that, unlike the sheep, most cattle were butchered off site. The jaws were assigned to tooth wear stage (TWS) (Grant 1982, 98-100). Only one is from a mature animal; the remainder are from young animals (MWS < 15) and the loose teeth are also from young calves, with the majority being the deciduous premolar (Dpm4) in the very early stages of wear. This is similar to some contemporary sites, including Grimes Graves, Bishop’s Cannings Down and Dean Bottom. This age distribution raises the same possible explanations as with the sheep, but it is generally accepted that it results from specialised cattle milking (Legge 1981, 1992). The present analysis allows comparison with other contemporary sites in southern England (Serjeantson 2007), summarised in Table 11. THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 65 Table 11: Comparison of East Chisenbury with contemporary Late Bronze Age Assemblages showing types of animal husbandry based on relative numbers and age at death. Runnymede Grime’s Graves DoE Few [DeanBottom |) | D Ritual Activity? The presence at East Chisenbury of so many young lambs — and also calves - of approximately similar age may not reflect an activity that was solely economic. The lambs may have been brought to the site from afar for the celebration of annual ceremonial rituals and sacrifices. One of these took place in summer, an event recalled by the many Lammas Fairs still celebrated in the British Isles in July and August. The deposition of the bones examined up to now do not suggest that ritual was involved. On settlements where a larger area has been excavated, deposits of animal remains have been encountered in articulated groups which seem to have a ritual origin, but ' when the finds from individual contexts at East Chisenbury were examined no discernible pattern was seen. A few elements were found in articulation, but none represented significant or substantial parts of animals. J. D. Hill has noted that ‘dog, horse and wild species were treated in different ways’ at many Iron Age sites (Hill 1995, 103), but that was not evident in the trenches excavated here. Nevertheless horses and dogs were eaten at East Chisenbury, as chopping and cut marks were seen on the bones. The fact that ritual deposition of animal bones cannot be detected in this assemblage does not mean | that such activities were not taking place. Grant | (1991, 113) has argued that we cannot separate economic and ritual behaviour nor assume that human behaviour was motivated purely by either. The deposition of midden material itself, including food waste, may be a ritual activity, as argued later in this report. If the inhabitants of the site consumed on average 250g of meat per day, the population would have been just over 1200 people; at 100g per day, it would have been more than 3000 (Tabie 9). These figures seem high for a prehistoric settlement — (the population of Upavon in 1941 i.e. before military expansion was just 430. Today at 1300, it is N PIGS SHEEP CATTLE HUSBANDRY HUSBANDRY Unintensive Unintensive ntensive [Few v. few Unintensive v. man ntensive Dean Bottom Unintensive ntensive Down Farm Unintensive Burderop Down Unintensive v. few nintensive East Chisenbur ntensive ?unintensive tensive still dwarfed by the midden figures). An alternative, and more likely, explanation would be that meat consumption took the form of feasts to which people were invited from other settlements. The scale of the site and its central role, together with the quantity of food consumed, makes it likely that feasting, no doubt of a ceremonial nature, took place. The Human Bones, by Sarah Inskip The human remains from East Chisenbury consist of seven bones, all well preserved (Table 12). They include both adult and juvenile remains suggesting a minimum of two and a maximum of six individuals. No information was available for sex or stature, but bone fusion provided a small amount of information on age. The proximal and distal joint surfaces of the two phalanges (Nos. 3 and 4) are fused suggesting a minimum age of above fourteen years if female and sixteen if male. The sternal end of the clavicle (No. 6) is unfused, suggesting an age under twenty six (Scheuer and Black 2000); however the size of the bone suggests a younger age. Measurement of the Table 12: Human bones from East Chisenbury. CB6_ |Left parietal Same individual as s reagent number 2. Button lesion cag) ECB 6 Same individual as number 1 1. |E 2 3 |ECB6 {2nd hand phalanx |Age = 14 years + 4 |ECB 24 |2nd hand phalanx |Age = 14 years + 5 |ECB 24 |Right ulna shaft Juvenile? eel ECB 35 |Right clavicle Juvenile, age 8 — 11 ears 66 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE clavicle length indicates an age between eight and eleven years (Black and Scheuer 1996). The juvenile remains could in theory be from the same individual. The only pathology is a small button lesion (5.2mm x 4.5mm) on the anterior of the left parietal. The exact aetiology of button lesions in unknown, but it is suggested to be a result of evolutionary background (Eshed et al. 2002). Other theories suggest traumatic and developmental causes. Pottery, by Frances Raymond The pottery (Figure 10) has not been the subject of as much detailed work as the material warrants and the following information is based on a preliminary assessment and partial analysis. The two trenches produced a total of 8083 sherds (1514 from Trench A and 8073 from Trench B), which have not been quantified by weight. The fabrics are characterised by a range of inclusion types which point to several different production centres. These include combinations of flint, sand, iron ore, mica, chalk, organic filler, glauconite altered to limonite, shell including fossil shell, limestone and ooliths. As at Potterne, wares containing material from distant sources are in the minority (cf. Morris 2000). The relative proportion of fabrics alters through the deposit in a manner which points to a chronological progression. Wares containing a mixture of flint and sand are dominant in the earliest horizons (82% in Trench A, Contexts 18 to 20; and 83% in Trench B Contexts 28 to 32 and 36 to 46). These occur alongside fabrics filled mainly with sand; micaceous sand; and sand and shell. In Context 35 in Trench B, which produced the largest ceramic group from the lower part of the midden, flint and sand tempered fabrics comprise 72% of the assemblage. Sand and shell tempered wares are fairly prominent in this group (16%), while micaceous sandy wares make up 9% of the pottery from the horizon. Rare fabrics include fossil shell and ooliths or limestone; limestone and micaceous sand; and an organic filler. Above the lowest horizons the proportions begin to change, with flint and sand tempered wares declining (55% in Trench A, Context 17), micaceous sandy fabrics rising (36% in Trench A, Context 17) and fabrics containing shell making up most of the rest (7% in Trench A, Context 17). This trend continues with the percentage of flint tempered wares falling further in the central part of the midden (27% in Trench A, Contexts 9 to 15). The micaceous sandy wares become dominant in these horizons (65% in Trench A, Contexts 9 to 15) and pre-eminent in the upper deposits (94% in Trench A, Contexts | to 7). One of the notable traits of the assemblage, displayed by sherds throughout the midden profile, is the hard character of many of the fabrics. This is typical of the ceramics of the Late Bronze Age/ Earliest Iron Age in this part of Wessex. Although there is no clear evidence for re-firing at East Chisenbury (cf. Brown 2000), much of the pottery must have been fired at a relatively high temperature or for extended periods of time to achieve the degree of hardness. As with other aspects of the technology, this represents a significant change in pottery production alongside the introduction of the new stylistic repertoire. Part of this process of style change led to a greater emphasis on ceramic decoration. At East Chisenbury this is reflected by a relatively high proportion of decorated fine and coarse wares. Fingertip or fingernail impressed rows are amongst the most common designs and are predominant on jars, where they place emphasis on rims and shoulders. Horizontal and vertical applied cordons, generally decorated with fingertip rows, also occur infrequently on the necks and shoulders of jars. This apparent use of decoration to stress the importance of the upper parts of vessels is echoed amongst the furrowed bowls, which are also a prominent element within the assemblage. More complex geometric angular and curvilinear motifs occur on some of the finer bowls and are less common. They include parallel lines, diagonal lines, infilled triangles, bordered zig-zags, split herringbone, swag, and ring and dot designs. Most of this more complex decoration occurs above the shoulder, although occasional vessels with motifs on the lower body are represented. Without exception all of the decoration was applied before firing and is largely deeply tooled or impressed. In rare instances white inlay survives in the geometric motifs or on furrowed bowls. Vessels with wiped surfaces are predominant, again presumably reflecting the relatively high frequency of jars within the assemblage. However, a notable percentage of sherds are burnished (approximately 40% from Trench A). This surface treatment is occasionally confined to the exterior, but mostly occurs on both surfaces, suggesting that it was applied to open vessels with visible interiors (i.e. bowls). In Trench A approximately 10% of the burnished fragments have burnished surface coatings (59 | | | THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 67 1 2 Fig. 10 The pottery forms at the site were dominated by long- and short-necked furrowed bowls as well as a wide range of decorated coarsewares. sherds). These occur in various shades of red (33 sherds, 56%), cream (19 sherds, 32%), pale brown (5 | sherds, 8%), pale yellow (1 sherd, 2%) and pale grey (1 sherd, 2%). A similar range of hues are represented _ amongst the assemblage from Trench B, but have not been quantified. On the few occasions where there is evidence for vessel style, burnished surface coatings _ appear to have been applied to furrowed bowls and to vessels with carinated shoulders. The red and cream surface coatings are particularly striking in contrast with the colour range in the rest of the assemblage which is predominantly black and shades of grey and brown. The repertoire of styles is extremely varied; the vessels range in size from miniature examples with rim diameters of 4cm to those measuring up to 30cm. 68 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE The furrowed bowls virtually all have tripartite profiles and flaring simple rims. Those with long necks similar to Bowl Type 2 at Potterne which have a currency extending from the late 8th to the early 6th century cal BC (Morris 2000) were recovered from the upper deposits in Trench B (Contexts 2, 3, 4 and 6). A cup of comparable style and date came from one of these same horizons (Morris 2000, Cup Type 82), while a plain carinated bowl sharing this long-necked flaring profile was found in the central midden deposits in Trench A (Context 11). The shorter necked tripartite furrowed bowls comparable to Bowl Type 3.3 at Potterne, which was used over a similar time period from the 7th to early 6th century calBC (Morris 2000), came from Contexts 4, 6, 18 and 24 in Trench B. These include re-fitting sherds from one vessel found in both Contexts 6 and 18 and joining fragments from a second bowl derived from Contexts 18 and 24. Furrowed bowls with bipartite profiles are rare in the excavated assemblage from East Chisenbury and include an example from the lower midden deposits in Trench A (Context 14). The profile recalls Bowl Type 3.1 at Potterne, which again has a date range between the 8th and 7th centuries cal BC (Morris 2000). A plain carinated bowl belonging to this type group came from the same context, with further plain and decorated examples from Contexts 18 and 22 in Trench B. Other biconical vessels, all decorated carinated bowls, came from the lower horizons in Trench B (Contexts 18, 20 and 22). These are comparable to Bowl Type 1 at Potterne, that emerged rather earlier in the 10th or 9th centuries cal BC and was produced until the early 6th century cal BC (Morris 2000). Biconical cups with a more restricted currency spanning the 8th and 7th centuries BC (Morris 2000, Cup Type 81) are also represented in these deposits (Contexts 18 and 22). Occasional fragments from open shallow bowls with a currency spanning the 9th and 8th centuries cal BC (Morris 2000, Bowl Type 11) occur in contexts close to the top and bottom of Trench B (Contexts 3 and 35). Shouldered or situlate jars are by far the most common of the jar forms represented at East Chisenbury. These are not chronologically sensitive since they were produced over a long time period between the 10th and 6th centuries cal BC (Morris 2000, Jar Type 51). Carinated jars with everted rims are also present, including one decorated example from Trench B (Context 18), but again these were used over a similarly extended time span (Morris 2000, Jar Type 31 or 32). Vessels with a more restricted currency include jars with globular profiles typical of the 8th to 6th centuries cal BC (Morris 2000, Jar Type 20) from the upper horizons in Trench B (Contexts 2 and 6) and from the central deposits in Trench A (Context 12). The example from Trench A has vertically mounted lugs, again a feature that recurs at Potterne. Occasional fragments from jars with horizontal applied cordons also occur (Trench B, Context 18). These vessels were in use between the 9th and 7th centuries cal BC (Morris 2000, Jar Type 56). A rare form combining applied horizontal and vertical cordons in a grid pattern, decorated with fingertip rows, was recovered from the upper horizons in Trench B (Contexts 2 and 3). Apart from these types the East Chisenbury assemblage additionally incorporates a few lid fragments and miniature vessels, including an open thumb pot. The ceramics from East Chisenbury are thus all of Late Bronze Age/Earliest Iron Age date. As at Potterne (Lawson 2000) sherds with Plain Ware affinities occur in deposits at the base of the mound (Contexts 19 and 20 in Trench A; and Contexts 28- 45 in Trench B). The flint and sand tempered wares that predominate in the lower part of the midden at East Chisenbury are similar to many local Plain Ware fabrics (Raymond 1994). However, these East Chisenbury assemblages display qualities which most probably emerged at the end of the Plain Ware horizon. Some of the surface finishes, for example, and the evidence for firing at high temperatures or for extended periods of time, are not at all typical of this earlier pottery. Unfortunately, there are few featured fragments from the lower midden horizons, but sherds from furrowed bowls and vessels decorated with All Cannings Cross geometric motifs are represented in Context 35 in Trench B, suggesting that the deposit formed no earlier than around 800 cal BC. The open bowl from this layer is of a type that went out of use at the end of the 8th century cal BC (Morris 2000, Bowl Type 11), so that in combination the evidence points to a formation date between approximately 800 and 700 cal BC for context 35. The pottery from the layers above this suggests that the deposits accumulated relatively rapidly between the beginning of the 7th and early 6th centuries cal BC. Although it is difficult to identify stylistic trends using the preliminary data available, notable changes in the preferred fabric range point to a chronological progression. Later material belonging to the All Cannings Cross/Meon Hill stylistic series is entirely absent from both trenches and from all unstratified THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 69 - groups. This would suggest that the activity at - East Chisenbury is most likely to have ceased by - the end of the 6th century BC before the new types appeared. The pottery from Trench B has provided the most detailed insight into the way in which the - mound may have formed. The deposits seem to have _ been placed in this area during a small number of _ punctuated episodes (see interpretation of deposits above). In fact much (not all) of the stratigraphy could be attributed to a single event. Joining sherds from a number of vessels unite Contexts 4 to 24. The re-fitting sherds are large and in fresh condition, all of the joins are crisp and some are between pieces which both retain food residues. In addition, multiple joining fragments from single vessels occur in different contexts. This makes it very unlikely that the re-fitting sherds can be explained in terms of later disturbance. Far from being homogenous, the composition of assemblages from separate layers shows considerable variation, as does the relative quantity of pottery from the different horizons. The three largest groups occur in association with chalk floors or deposits of chalk and include: 1772 sherds stratified immediately below a chalk floor (Context 6); 904 sherds from a compacted deposit of chalk (Context 18) and 1012 sherds from a similar layer (Context 24). Few of the assemblages from other horizons are composed of more than 250 fragments of pottery. The intermediate groups include 538 sherds from the upper midden (Context 2) overlying a chalk floor; 340 sherds from the deposit immediately below this same chalk floor (Context 4); and 358 sherds from Context 35. The vessels do not appear to have been shattered in situ, for in spite of extensive reconstruction none is totally complete. This may, of course, partly be a reflection of the small size of the trench which only sampled the various deposits, increasing the chances that some of the missing pottery lay just outside the excavated area. However, it is evident from the pottery alone that the formation of the midden was a complex process. The fresh sherds, for example, occur alongside fragments exhibiting | signs of moderate to heavy abrasion and for the most | part the sherd size is relatively small. As suggested at | Potterne (Gingell and Morris 2000) it seems probable that many of the vessels were broken during practices _ beginning elsewhere and culminating in the area of Trench B. Whatever the circumstances, the fresh condition | of many of the sherds and the large proportion of fragments retaining food residues (not investigated further) indicate that at least part of the assemblage is unlikely to have been exposed to weathering. This pottery was either deposited soon after breakage or was protected deliberately until it was finally placed in the midden. The precise significance of this material is irrecoverable, but the variation in vessel style and the differences in the range of ceramics placed together in separate layers, suggests the expression of a complex set of relationships. Whether these were confined to a single community or involved more than one group of people is at present open to debate. Other materials, by Elaine L. Morris A total of 59 worked objects was examined including an amber bead, a shale armlet fragment, a chalk bead, eight ceramic and three chalk spindle whorls, 25 bone objects and 19 worked antler or antler fragments (Figure 11: Table 13). A single piece of glass was also recovered. It was sent through the post but was lost in the process. The objects are presented as a catalogue by material (Table 14) with short discussions throughout the text for significant types. This is followed by a concluding section that briefly explores functional groups and also the frequency of types by comparison to similar assemblages in southern England. Discussion by functional group and comparison to other site assemblages The range of 59 objects included personal ornaments, small tools for spinning and leatherworking, larger tools, and also partly worked pieces and offcuts from bone and antler. Five objects used to adorn the body were identified — single examples of amber, chalk and glass beads, a shale armlet and the perforated mandible of a badger; three perforated bone plaques may also have performed this function. Amber and glass beads and shale armlets are not uncommon artefacts on Late Bronze Age sites in southern England such as at Potterne and Runnymede Bridge, All Cannings Cross has a chalk bead, and Potterne has perforated bone objects including a badger jaw and a variety of flat plaque pendants similar to All Cannings Cross. The presence of only single examples of amber, glass and shale at East Chisenbury is similar in relative frequency to Potterne and Runnymede based on the relative quantity of pottery recovered from each of these sites (Lawson 2000, table 45). Evidence for textile working at East Chisenbury THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 70 10 cm Fig. 11 Illustration of a small number of special finds from East Chisenbury. These include worked antler tines, decorated spindlewhorls, worked and utilised bone. THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY Table 13: Small finds | CONTEXT WORKED BONE ~ = 5 oO & 24) 2 3 Oe! vo| x 8) ol sd Sra/ g| Ss] = CH UE a) pas ad Herel) ov alee: cw. gqolqgo Vag «@~D gOsysy Cal ul ae ‘soles! SI Se se 65 Sal Hl wo] S a Q, Oo; Sa-n YY Ol Sm o o| og] oO as| as! Oo] 2 oe] ag] Ao Ba ae] al aul 6 TRENCH A oe ao ee PE one eo ee ee : Pee ee pie (Se see fees eee ie shuttle tip oe — other = iyo) — — — — antler S a nie IS [= |S |e eae © |S [mn |B IS - S = z Q x w& 1 Je |! Feb ee NS. 4 I 2 I {l Table 14: Catalogue of small finds Glass 1. Annular bead, opaque, 95% present; diameter c. 5-10mm, thickness, c. 2mm; context 22, trench B. 7 __ This bead was lost when sent to the specialist; the envelope arrived damaged and the bead was not inside. The measurements | and details of the bead are approximate based on the excavators’ memories. Their description is not dissimilar to Guido’s | Class 8 small opaque yellow annular Iron Age beads of British design and origin which are fine, thin and small in size (1978, | 73-6, plate II, 13). Glass bead production at Meare Lake Village allows for a date as early as 5th century BC (Henderson 1987) which is supported by the recovery of a similar bead at Conderton Camp, Worcs. (Thomas et al 2005, fig. 56, G2). | Amber 2. Annular bead, 50% present; diameter 20mm, thickness 12mm, weight 2g; perforation diameter 6mm; very poor condition, | cracked and shattered; context 35, trench B. Within the Beck and Shennan (1991) classification scheme, this is a medium-large, annular bead of group type 2 because it has an unusually large central hole and the object’s thickness is more than half its diameter. However, with a ratio of 6/10, it may well be classified as a type 3, flattened globular bead. It has a C-cross section. The annular bead is a recognised 72 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 14: Continued Late Bronze Age type, while the flattened globular form is more often recovered from Early Bronze Age deposits (Beck and Shennan 1991, 53-7, tables 4.6-4.7). Amber beads of all types have been found at three other Late Bronze Age/Early Iron Age settlement sites in southern England including Potterne (Beck and Shennan 1991, 103, 166-7, fig. 11.24; Healy 2000, fig. 86, 1-5), Runnymede Bridge (Longley 1980, 31, 33, fig. 18, 69-70) and Queen Mary’s Hospital, Carshalton (Adkins and Needham 1985, 41, fig. 15, 417). An extremely similar bead, which had been broken and re-drilled presumably for repair, was recovered from an Early Iron Age (cp. 3) pit at Danebury (Cunliffe 1984, 396, fig. 7.42: 5.1). Shale 3. Armlet, 12% present; diameter 60mm, width 6mm, thickness 3mm, weight 1g; very good but fragmented condition; special find no. 6, context 17, trench A. Fragments of broken shale armlets are found on Late Bronze Age/Early Iron Age settlements in southern England at some considerable distance from the sources of shale located along the coast of the Isle of Purbeck. These inland occupation sites include All Cannings Cross (Cunnington 1923, plate 26, 2-4), Brean Down (Foster 1990, fig. 112, 37-39), Potterne (Wyles 2000, figs. 80-1), Runnymede Bridge (Longley 1980, fig. 18, 65-66; Needham 1991, fig. 65, SH1), and now East Chisenbury. At both Brean Down (Foster, op cit., fig. 112, 40-41) and Potterne (Wyles, op cit., fig. 81, 17-20), shale roughouts or ring blanks were also recovered; at Potterne these were found throughout the deposit. This transportation of the unfinished product away from the source area and primary or roughout production sites, such as Eldon’s Seat (Cunliffe and Phillipson 1968) and Rope Lake Hole (Cox and Woodward 1987) in Dorset during the first millennium BC, to occupation sites for final artefact working was a form of regional trading which occurred from the Late Bronze Age through to the Iron Age, as at Danebury (Cunliffe 1984, fig. 7.41, 4.4-4.5; Laws 1991a, fig. 7.40, 4.15). Ceramic Spheroid whorl 4, Spindle whorl, spheroid form, 50% present; fabric: a fine silty clay matrix with rare (<1% concentration) shreds of mica; diameter 41.5-42.0mm, height > 25mm, extant weight 39g, perforation diameter 6mm; unoxidised but irregularly oxidised exterior surface; generally good condition and with a concretion on exterior only; context 9, trench A. Similar examples of the slightly flattened, spheroid form were recovered at All Cannings Cross (Cunnington 1923, plate 25, 2-3), Potterne (Hall and Seager Smith 2000, fig. 64, 3-4), and Danebury (Poole 1984, fig. 7.46: 7.32 and 7.34). Biconical whorls 5. Spindle whorl, biconical form, 50% present; fabric: moderately sandy fabric containing moderate (10%), subrounded quartz sand, <0.5mm across, and with rare mica shreds, glauconite pellets and larger grains of quartz up to 1mm; diameter 34mm, height 22.5mm, extant weight 14g, perforation diameter 6mm; roughly faceted around the girth by irregular, finger-tip impressions; unoxidised firing; context 1, trench B. 6. Spindle whorl, biconical form, c. 35% present; fabric: sandy fabric containing common (20-25%), subrounded quartz sand, <0.5mm, with rare calcined flint fragments and possible glauconite; diameter > 34mm, height 25mm, extant weight 12g., perforation diameter 5mm; burnished; mainly unoxidised firing with zone of oxidised effect in perforation; context 24, trench B. Biconical spindle whorls are a relatively common type of Late Bronze Age object (Longley 1980, fig. 17, 52-53; Adkins and Needham 1985, fig. 13, 406-407; Bradley, et al. 1980, fig. 20, 6-7), and similar, roughly executed as well as more carefully decorated, examples have been found at Runnymede Bridge (Needham 1991, fig. 66, C7-C9). The biconical form is a type which was also utilised during the Iron Age, as at Budbury hillfort (Wainwright 1970, fig. 19, 145) and Danebury hillfort (Poole 1984, fig. 7.46: 7.39, 7.45). Burnishing is an unusual surface treatment for later Bronze Age whorls although this site has two examples; both recovered from context 24 (see Cat. No.11 below). Cylindrical whorls 7. Spindle whorl, slightly convex cylindrical form, complete; fabric: fine, sandy clay matrix with sparse to moderate (5-10%), soft, rounded chalk fragments, up to 1 mm across; diameter 36mm, height 33mm, total weight 54g, perforation diameter 6mm; decorated irregularly around convex surface with numerous 6 mm deep and 3-4 mm diameter indentations; unoxidised; context 4, trench B. 8. Spindle whorl, slightly waisted cylindrical form, complete; fabric: moderate to common (15-20%), subrounded quartz sand, <0.S5mm across, and rare (1%), larger quartz sand grains up to 1.0mm; diameter 36-40mm; height 28.5-31.0mm; total weight 55g; perforation diameter 7-9mm; unoxidised; cessy concretion on surface; context 18, trench B. 9. Spindle whorl, cylindrical form, 50% present; fabric: as for Cat. No. 7; diameter 36.5-39.5mm, height 34.5mm, extant weight 32g; perforation diameter 6mm; unoxidised core, irregularly fired surfaces; context 18, trench B. THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 73 Table 14: Continued 10. Spindle whorl, convex cylindrical form with slightly expanded upper and lower sides, 50% present; fabric: common (20%), subrounded quartz grains, < 0.5mm; diameter 28mm, height 25mm, extant weight 12g, perforation diameter 5mm; unoxidised; context 6, trench B. 11. Spindle whorl, convex cylindrical form, possibly 45% present; fabric: very common to abundant (30-40%), subrounded quartz grains, up to 1.5mm across with the majority between 0.3-0.6mm; diameter 29mm, height > 19mm, extant weight 16g; perforation diameter 4mm; burnished; unoxidised; context 24, trench B. Fine examples of the cylindrical type of spindle whorl have been found at Potterne (Hall and Seager Smith 2000, fig. 64, 11), All Cannings Cross (Cunnington 1923, plate 25, 5), Budbury (Wainwright 1970, fig. 19, 143) and Danebury (Poole 1984, fig. 7.46: 7.36-7.38), but the all-over style of decoration of Cat. No. 7 may be unique to the East Chisenbury collection. Cylindrical whorls, such as Cat. No. 8, which have a waist-effect due to slight squeezing of the whorl girth while still leather-hard are present in the Early Iron Age collections from Danebury (Poole 1984, fig. 7.46: 7.35) and All Cannings Cross (Cunnington 1923, plate 25, 6 and 10), but not amongst the published whorls from Aldermaston (Bradley, et al. 1980), Potterne or Runnymede (Longley 1980; Needham 1991). As for Cat. No. 6, Cat. No. 11 whorl is unusual due to the burnished surface treatment. Chalk Disc whorls 12. Spindle whorl, disc form, c. 20% present; diameter 80mm, thickness/height 13-16mm, extant weight 26g; multiple, irregular scored marks on one surface along face; context 3, trench A. 13. Spindle whorl, disc form, c. 35% present; diameter c. 60mm, thickness/height 18mm, extant weight 38g; hourglass profile perforation indicating manufacture from each surface for minimum 8mm hole; scratched on one face; context 20, trench B. 14. Spindle whorl, disc form, c. 25% present; diameter c. 46mm, thickness/height 8-10mm, extant weight 6g; perforation diameter c. 10 mm; unstratified, rabbit scraping. Disc-shaped chalk whorls are common artefacts on later Bronze Age and Iron Age sites located on or near chalk deposits. Similar whorls were found at Potterne (Seager Smith 2000a, 216, fig. 85, 20-21), All Cannings Cross (Cunnington 1923, plate 25, 15-16), Danebury (Brown 1984, fig. 7.61; Laws, Brown and Roe 1991, figs. 7.64-7.66), Winnall Down (Fasham 1985, fig. 64, 1-7) and Maiden Castle (Wheeler 1943, 294, plate 33, except no. 6; Laws 1991b, fig. 169, 1-9, 12) to name a few collections. The estimated weights of the larger examples from East Chisenbury suggest that they may not have been whorls but simply perforated chalk discs of unknown function (cf. Laws, Brown and Roe 1991, 404). Sub-spheroid bead 15. Bead, sub-spheroid form, complete; diameter 30-32.5mm, thickness/height 16-17mm, total weight 17g, perforation diameter 6mm, displays lopsided wear; context 6, trench B. Cat. no. 15 is very small in weight and the lopsided wear suggests that the object was suspended. Therefore, this chalk object is most likely to have been used as a bead. Animal Bone | The major tool groups are assigned according to Smith (1983) and Seager Smith (2000); nearly all groups are represented. The bone identification to species and anatomical part was conducted by Clive Gamble. Group 1: Pointed tools (see also Cat. No. 46) | Large pointed tools (1.1) 16. Ovicaprid, tibia, proximal end; damaged at general working end but not at tip itself; length 102mm; deposits of cessy concretion adhering to surface; polished from use along entire shaft; unstratified, trench A. 17. Horse, metapodial, proximal end; probably broken in antiquity but continued to be used; polished from use along entire length of tool; length 92mm; context 4, trench B. 18. Ovicaprid, tibia, proximal end; broken at tip, highly polished along tip end half of tool; extant length 110mm; context 6, trench B. 19. Unidentifiable to species or anatomical part, possibly rib or scapula; broken working tip end; well-smoothed from use; extant length 75mm; context 6, trench B. 20. Cow, ulna; broken at handle end; polished from use on two handle sides and all over pointed end; excellent shape for holding in the right hand; length 90mm; context 24, trench B. 74 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 14: Continued 21. Ovicaprid, metatarsal, distal end; complete with very sharp point; highly polished all over from use; length 99mm; context 35, trench B. Small pointed tools (1.2) 22. Ovicaprid, metacarpal, proximal end; complete; length 77.5mm; traces of polish along entire shaft; special find 2, context 4, trench A. 23. Ovicaprid, metatarsal, distal end; highly polished nearly all over tool; cessy deposit covering much of surface area; length 75mm; context 19, trench A. 24. Unidentifiable to species or anatomical part; broken at both tip and handle ends; incised line apparently demarcating tip from handle; smoothed around tip; extant length 48mm; context 1, trench B. 25. Horse, metapodial, proximal end; slightly damaged at tip; polished all over from use and holding; extant length 65mm; polished all over from use; context 6, trench B. 26. Ovicaprid, metatarsal, distal end; broken at working point end; unwashed; extant length 63mm; context 35, trench B. Large and small points are usually found to have been made from the shaft and proximal end of ovicaprid long bones, as is the case for the majority of examples from East Chisenbury, but horse and cow bones are also known to have been selected (Sellwood 1984, 382). Various types of points, or awls, are extremely common bone tools found on many Late Bronze Age/Early Iron Age sites such as at All Cannings Cross (Cunnington 1923, plate 10), Brean Down (Foster 1990, fig. 113, 45-55), Eldon’s Seat (Cunliffe and Phillipson 1968, plate Sb, 33-50), Potterne (Seager Smith 2000b, fig. 89, 1-9) and Runnymede Bridge (Longley 1980, fig. 14, 48-51; Foxon 1991, 148, fig. 64, B1-B3). A great variety of uses was possible with these pointed tools but it is indeed interesting to note that there are two general size lengths for bone points which suggest either different functions or different hand sizes of the users of these tools. Shuttle tip or gouge (1.3) 27. Ovicaprid, metatarsal, proximal end; broken at middle of tool, handle end remains; extant length 40mm; well-smoothed from use; context 24, trench B. Gouge-like pointed tools are found in Late Bronze/Early Iron Age sites in southern England, such as at All Cannings Cross (Cunnington 1923, 82-8, plates 8, 1-16, 9, 1-16 and 12, 8), Brean Down (Foster 1990, fig. 114, 56-60), Eldon’s Seat (Cunliffe and Philipson 1968, plate 5b, plate Sb, 23-32) and Potterne (Seager Smith 2000b, fig. 90, 10-15), and they are extremely common on Iron Age sites such as at Danebury where they were found throughout the occupation of the site (Sellwood 1984, 382-387, figs. 7.33-7.34; Cunliffe and Poole 1991, 359, fig. 7.32). They are often referred to as possible shuttle tips or leather-working tools. Needle (1.6) 28. Unidentifiable to species or anatomical part; broken at eye end; well-worn from use at tip end; extant length 68mm; context 24, trench B. Miscellaneous bone points (1.9) 29. Unidentifiable to species or anatomical part, splinter; broken opposite tip end; usewear polish on tip end; extant length 33mm; context 24, trench B. 30. Unidentifiable to species or anatomical part, spongy core of bone exposed; probably broken; polished along length; extant length 48.5mm; context 6, trench B. 31. Unidentifiable to species or anatomical part; broken, tip end present; polished from use; extant length 31mm; context 24, trench B. Needle/pin shank (1.10) 32. Unidentifiable to species or anatomical part; broken, tip end present; sub-circular cross section; highly polished from use; extant length 32 mm, width 3 mm, thickness 2 mm; special find no. 4, context 11, trench A. 33. Unidentifiable to species or anatomical part, broken at both ends; sub-rectangular cross section; well-polished from general use; may have been a needle which broke and was reused as a pin due to old break at one end repolished over the break; extant length 42mm, width 3-5mm, thickness 3mm; context 15, trench A. Needles and pins are also commonly found amongst large assemblages of bone tools from later prehistoric sites, as at All Cannings Cross (Cunnington 1923, plate 6, 8-23), Budbury (Wainwright 1970, fig. 18, 133), Potterne (Seager Smith 2000b, fig. 90, 21-26), Runnymede Bridge (Longley 1980, fig. 43) and Danebury (Sellwood 1984, fig. 7.32; Cunliffe and Poole 1991, fig. 7.31). THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 1D Table 14: Continued Group 2: Bladed tools Worked ribs (2.5) 34. Large ungulate; cut and pared at one end, broken at one end; numerous scratches or repeated knife cuts along length edge and across width at cut/pared end on one face only; extant length 102mm, rib width 25-27mm; unstratified, from rabbit scrapings, trench B. 35. Large ungulate; cut and pared at both ends, complete; numerous scratches or repeated knife cuts across both ends and part of length on one face, only across only one end on other face; very highly polished on unscratched end face and length; length 165mm, width 30-34mm; context 6, trench B. 36. Large ungulate; cut and pared at one end, broken at one end; repeated knife cuts or scratches at cut end; extant length 93mm, width 43mm; context 6, trench B. _ 37. Large ungulate; complete, cut and pared at both ends; heavily used due to repeated knife cuts, scratches or as a result _ of binding across both ends and longitudinally on both sides - very worn appearance across the ends; length 155mm, _ width 32mm; context 6, trench B. _ 38. Large ungulate; cut and pared at one end, broken at opposite end; well-polished/smoothed at the rounded end, also repeated knife or scratch marks across rounded end, diagonally at opposite end, and along side lengths of both faces; extant length 152mm, width 33mm; context 6, trench B. 39. Large ungulate; broken at both ends; small area of repeated knife cuts at one end and a knife-cut notch along one side; _ extant length 60mm, width 22mm; context 6, trench B. 40. Large ungulate; cut and pared at one end, broken at opposite end; repeated knife cuts or scratches across rounded end on both faces and along one face with polished zone on opposite face; extant length 127mm, width 30mm; context 18, trench B. _ 41. Large ungulate; broken at both ends; repeated knife cuts or scratches along both faces and across one end; extant length 130mm, width 27mm; context 18, trench B. _ This is a significant collection of these unusually treated worked rib fragments or blades. Similarly worked and used - artefacts, also described as ‘rib bones worn and striated by use as if in a machine or contrivance of some kind’ (Cunnington 1923, 23), offcuts and polishers, have been recovered from All Cannings Cross (Cunnington, op cit, plate 15, 8, 10, 15, 17), Budbury (Wainwright 1970, fig. 18, 130), Eldon’s Seat (Cunliffe and Philipson 1968, 225, plate 5b, 15-22), Linton (Cambs.) | (Fell 1952, fig. 2, 4), Potterne (Seager Smith 2000b, fig. 91, 31-36) and Runnymede Bridge (Foxon 1991, 148, fig. 64, B6-B9). This class of artefact is very distinctive and deserves further research to determine its function. Group 5: Dress, decorative or gaming artefacts Decorations pierced for suspension (5.5) | 42. Possibly tooth enamel, possibly pig canine; polygonal, flat plaque, broken; perforated at broken end, worn surface on _ broken end side of perforation indicates suspended use; extant length 40mm, width 10-15mm, thickness 1mm; unstratified, _ from rabbit scrapings, trench B. | 43. Undiagnostic to species or anatomical part; bone fragment split, exposing spongy interior and subsequently smoothed _ from wear; broken at both ends; narrow, incised groove cut on the diagonal around one end to separate suspended end from spongy end of object; perforation size similar to that of needle (Cat. no. 29); extant length 60mm, width 12mm, _ thickness 8mm; context 21, trench B. _ 44. Badger, left mandible, incomplete, missing incisors end of mandible and end attached to skull; perforated at skull end; extant length 67mm; perforation 3-4mm diameter; perforation worn on one side of mandible and at skull end indicating | suspension as a pendant; context 24, trench B. Several similar perforated sub-rectangular and ovate flat pendants were found at All Cannings Cross (Cunnington 1923, plate 6, 1-7), Brean Down (Foster 1990, fig. 115, 66) and Potterne (Seager Smith 2000b, fig. 92, 37-43). Perforated animal teeth and mandibles are not infrequent objects used as pendants at later prehistoric sites such as at All Cannings Cross | (Cunnington 1923, plate 6, 36-37) and Potterne (Seager Smith 2000b, fig. 93, 56-65). It would not be inappropriate to | suggest that these items were used as indicators of clan association or used as amulets. It is possibly of some interest that ' Suspended bone plaques and animal tooth or jaw pendants are not found within the abundant worked bone assemblage from the Early to Late Iron Age site at Danebury. Group 6: Miscellaneous | Other pieces of worked bone (6.6) | 45. Possible split rib of cow or horse; broken at both ends; single, narrow knife cut mark diagonally across one face and 76 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 14: Continued repeated scratch marks diagonally across other face; extant length 45mm, width 25mm, thickness 2mm; context 6, trench B. 46. Ovicaprid metatarsal; complete length with epiphysis missing; pared by cut along shaft in the process of creating either a large (1.1) or small (1.2) pointed tool; length 125mm; context 6, trench B. Evidence which demonstrates actual bone-working on later prehistoric sites in Wiltshire is common (Smith 1983; Seager Smith 2000b), and an identical example of a partially complete tool has been recovered from All Cannings Cross (Cunnington 1923, plate 9, 17). 47. Horse scapula, small adult; reduced, rounded blade end due to use as shovel; polish visible along part of spine; extant length 285mm; context 18, trench B. 48. Two joining fragments from a slightly tapering rectangular flat strip, incomplete; one end and cross-section complete, both long sides are bevelled in profile; unidentifiable to specie or anatomical part; one surface highly polished but also bearing multiple linear scratches which may be post-depositional in origin, other surface spongy structure of bone; possibly similar to Cat. No. 12 pendant/plaque; not likely to be a gaming piece since it is not square in shape; length 24mm, width 20-23mm, thickness 1.5mm; context 9, trench A. 49. Tapering, convexo-rectangular, flat plaque, broken; unidentifiable to specie or anatomical part; highly polished on both flat faces; possibly etched with single line across short end and also along each convex, tapering side of the same flat surface face; length 45 mm, width 6.5-14.5 mm, thickness 1.5 mm; special find 3, context 10, trench A. 50. Split rib fragment, unidentifiable to specie; broken at both ends; highly polished on both surfaces; extant length 24mm, width 15mm; unstratified, trench A. Group 7: Antler Antler tines (7.1) 51. Red deer antler tine, broken; very smoothed/polished along area from repeated cut marks across one side in same location; a zone of repeated cut marks on reverse; extant length 183mm; unstratified, from rabbit scraping, trench B. 52. Red deer antler tine, chopped off antler with clear cut marks and snapped; damaged by breakage at chopped end; polished at point of tine; extant length 114mm; context 1, trench B. Antler picks or pointed tools (7.2) 53. Roe deer antler, shed; nearly complete, longer tine slightly damaged at pointed end; both tines polished with polish extending over damage in antiquity; naturally roughened zone (with the appearance of ‘candle wax’) suitable for improving grip; extant length 210mm; context 18, trench B. 54. Roe deer antler, burr/pedicle removed by cut and snapped; facet along lower shaft zone; longer prong slightly damaged at pointed end; both tines very well polished; extant length 210mm; context 18, trench B. 55. Roe deer antler, shed; longer tine missing due to breakage; shorter tine polished; extant length 155mm; context 10, trench B. Miscellaneous worked antler pieces (7.3) 56. Red deer antler and skull fragment; damaged at working tip end of antler; originally cut across beam now abraded; extant length 95mm; well-polished from use; context 6 or 10, trench A. 57. Red deer antler fragment; cut and polished at one end, broken along the length and one end; abraded; extant length 33mm; context 9, trench A. Worked antler (7.5) 58. Red deer, antler, shed; on double prong, one tine removed by cut with other tine broken; remaining tine, well polished and worn; possible two-handed tool; extant length 247mm; unstratified, from rabbit scrapings, trench B. 59. Red deer antler, shed; double prong, both tines broken; pathology — antler did not stand upright on the head but reached along the deer’s back and single prong not grown (J. Bagust, pers. comm.); zones of cut marks on both sides; extant length 430mm; context 20, trench B. The use of red deer antlers as raw material for tools is a common later prehistoric activity. Numerous examples of red deer antler points being removed from the main antler branch have been identified at All Cannings Cross (Cunnington 1923, plate 13, 15), Potterne (Seager Smith 2000b, fig. 95, 82-90 and 98, 102-103), Runnymede Bridge (Longley 1980, fig. 14, 40-42; Foxon 1991, fig. 65, B12) and Danebury (Cunliffe and Poole 1991, fig. 7.39, 3.371-3.377). THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY Table 15: Spindle whorls No. Present] Weight | Weight Comments 39g. 14g. ores ee cee eae 260 ee ee eee 8 ES ee ee eee Ea Ruebed bumied 78g. 28g. BSP esac 0% 2s [ote ediom Bp [is talk fe 35% te [0m Ylree Sorted by Weight Serco ay IS ger [S[k oe Pe oma a SN ae ee oes |ccamic] 50% 2g. 5 | Ae small | Pee Bee (5 fea 5% [le [ate —_fomall [ora EC MEDIUM: 54 - 78g. RP [ane SOP Fee ne CC Ee ie a8) a) |ceramic) 100% | 55a) | ee] medium | ium eine = ee ceramic | 50% a2) ee [medium | ee ESC (CR ECC BoP a3 tele Fe 77 Fig. 12 Decorated and perforated spindlewhorls 78 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE is presently limited to an array of fired clay and chalk spindle whorls (Table 15), but no clay or. chalk loom weights were identified. This is not surprising considering the relative quantity of pottery recovered or area excavated compared to several other sites of similar periods (Lawson 2000, table 46). The array of whorls is extremely varied in terms of fabric, shape, decoration and surface treatment which suggests that Late Bronze Age/ Early Iron Age whorls are very likely to be personal objects distinctive to individuals, and therefore not unlike bodily adornments. East Chisenbury had relatively similar ratios of fired clay to chalk whorls (8:1) as All Cannings Cross (6:1), but far fewer than the clay and stone whorls at Potterne (13:1) and there are no examples of potsherd whorls, 16 of which were found at All Cannings Cross and 19 at Potterne (Lawson 2000, table 45). The weights of the whorls, if adjusted to reflect their probable complete sizes, fall into three distinct weight groups (Table 15), and it is worth considering in future research whether these groups are found amongst other assemblages and if they are indicative of different types of yarns being used to make textiles at East Chisenbury (c.f. Parsons 1972). Midden sites in the Wiltshire area commonly have worked bone objects in their artefact assemblages, especially large and small pointed tools and gouges which are most likely to have been used as part of leather-working equipment and other everyday tasks, while needles and pins belong to textile sewing. The points of polished antlers may also have been used in leather-working, or as possible pitchforks (Duncan and Mackreth 2005, 135). The shoulder blade shovel is not unique; a Bos version with similar wear pattern was recovered at All Cannings Cross (Cunnington 1923, plate 14, 3). The actual manufacture of bone and antler tools, as evidenced by the discarded antler and bone-working waste at the site, is also a common activity on midden sites in Wiltshire of this date (Smith 1983; Seager Smith 2000b). Despite the small areas excavated in the two trenches, the range of non-metalwork objects is remarkably similar to assemblages of artefacts from midden sites in Wiltshire in particular and also in general to Late Bronze Age/Early Iron Age sites in southern England. Metalwork, by Martyn Barber Copper Alloy I. The socketed axehead (Figure 13 upper). Trench A Context 11 Old Land Surface. Max. extant length: 55mm. Max. blade width: 47mm. Blade thickness: 3mm. Width at break: 38mm. Breadth at break: 18mm. Size of socket at break: 28mm x 11.5mm. Weight: 154g. The fragment comprises the lower portion of the body of a socketed axehead, including the blade end. There is some minor corrosion, and some soil adhering to the face and within the socket. Colour- wise, the fragment presents a range of light greens and browns giving an overall khaki appearance. The sides appear relatively straight, diverging as they approach the blade and giving the fragment a wedge-shaped appearance. Although the axehead has clearly been subject to post-casting working, this has stopped short of bringing it into a usable state. The faces and sides are generally smooth and flat, with no obvious imperfections, but the casting seam remains clearly visible along both sides and across the blade end, standing up to 1mm proud of the body of the axehead. Although the casting seam has clearly been subject to a certain amount of working to bring it to this state, it seems that there was no intention of completing the job of removing it entirely. Consequently, the blade itself has never been sharpened, with what should be the cutting edge remaining up to 3mm in breadth. The break is irregular and its cause unclear — it cannot have happened in use — although there is a vague indentation on one face just below the break, plus another smaller dent lower down on one side. These suggest the possibility that one or more blows to the body of the axehead may have contributed to or caused the breakage. The break shows an uneven thickness of metal around the body of the axe, ranging from a little over 3mm to a little under 2mm. Discussion The fragment almost certainly comprises the lower part of a Sompting-type axehead. These have been described by Burgess (1969), Schmidt and Burgess (1981) and most recently by Coombs et al. (2003). They tend to be both larger and heavier than other socketed axehead forms from the British Isles, and possess a distinctive body form. They are usually wedge-shaped, typically with straight sides diverging from the socket mouth towards a cutting edge which may show little or no sign of blade expansion through use or working. In cross-section, the upper body is usually broader across the sides than it is wider across the face, producing a socket whose shape is usually — and incorrectly — described as THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY Fig. 13 A small amount of metalwork was recovered from the excavations including a bronze tanged chisel (Trench B) and the broken blade end of a socketed axe (Trench A). 79 80 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE ‘back-to-front’. It is, of course, turned through 90° from the norm and not 180°. This particular feature is not unique to Sompting axeheads — it is known to occur sporadically on some slightly earlier axehead forms — but it is most commonly associated with this type. Extending the lines of both faces and sides on the East Chisenbury fragment suggests that the upper body would indeed have conformed to the Sompting style. Sompting axeheads are also distinctive in terms of body decoration, commonly featuring ribs, pellets and roundels in a variety of permutations and combinations. Unfortunately, the East Chisenbury example has been broken below the point at which such decoration would normally be visible. Another key feature of the type was noted by Coombs et al. (2003, 205) — ‘a large proportion of the axes are as straight from the mould and have never been sharpened or show little signs of wear’. This is certainly true of the East Chisenbury example. Indeed, arguably, as with many Sompting axes, there was no intention to bring the axe to a finished, usable state, which raises further questions about its breakage and deposition. Sompting axeheads are normally associated with the Llyn Fawr metalworking phase, belonging to the latest Bronze Age or earliest Iron Age. The few radiocarbon dates associated with metalwork of this phase suggest that the transition to Llyn Fawr from the preceding Ewart Park phase occurred around 800 BC, although the duration of the phase beyond 800 is unclear (Needham et al. 1997, 98). Sompting axeheads are found sporadically across the British Isles, with noteworthy concentrations in the Thames Valley, East Anglia, North Lincolnshire and Yorkshire, and the ‘Wessex’ region. 2. Tanged chisel/knife. (Figure 13 lower) Trench B Context 35 Old Land Surface. Max. length: 62mm. Max. blade width: 24mm. Max tang thickness: 5mm. Weight: 13g. A tanged blade, comprising an asymmetric blade with a tang of broadly rectangular cross section. In face view, the sides of the tang begin at a point, diverging sharply for a few millimetres before turning to follow a more parallel but still gently diverging course, eventually moving more noticeably apart towards the blade. In side view, the tang is at its broadest at around the object’s mid point, the sides then converging gently in both directions. A considerable amount of soil still adheres to the object, which is otherwise a fairly dark green colour. As far as it is possible to tell, the object is ‘finished’ —1.e. itis in a perfectly usable condition, although it is debatable whether it has actually been used. The asymmetry visible on the blade may have been an intentional as-cast feature. Simple tanged blades, described variously as chisels or leatherworking knives (Eogan 1964; Burgess et al. 1972; Roth 1974; O’Connor 1980), occur from the Middle Bronze Age onwards, with a clear break between blade and tang, evident in the form of break, shoulder or some form of projecting stop being common from Wilburton phase metalworking onwards. However, Ewart Park and Llyn Fawr associations have produced a range of forms, including both simple and complex items. Consequently, there is little else that can be said about this particular example. Other pieces Copper alloy fragment from topsoil above Trench B. Length: 29mm. Max width: 10mm. Max thickness: 3mm. Weight: 3g. Small, irregularly-shaped fragment of copper alloy. One side is generally flat, the other slightly convex. May be a fragment of a larger item, or possibly metalworking spill/waste. Not necessarily Bronze Age or even prehistoric. Possible strap end/fitting from the topsoil above Trench B. Length: 31mm. Width: 14.5mm. Max thickness: <0.S5mm. Weight: 1g. Small, flat item of copper alloy. Central portion near square in shape, but the ends are triangular, one larger than the other. There is an off-centre perforation c3mm long within the central section, through which the object would presumably have been attached or suspended. Probably not prehistoric. Possible ornament/fitting unstratified from lynchet. Length: 37mm. Breadth: 14mm. Thickness: 3mm. Weight: 3g. Near-oval shaped item comprised of two strands of bronze, both of sub-circular section, twisted around each other. There is no evidence that this is anything other than its original shape. In any case, it would have been too small to fit around the wrist or ankle of anyone other than a very young child, and too large to fit around a finger. It may, therefore, be a strap of belt fitting of some kind. Unlikely to be prehistoric. Coin unstratified. Almost certainly Roman in date. THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 81 - Tron Tanged blade from Trench B layer 17 (close to rabbit hole). Max extant length: 80mm. Max extant width: 13mm. Max thickness: 5.5mm. Weight: 13g. Tanged blade, the blade having a straight top edge which merges into a similarly straight, flat- - topped tang, though the latter is noticeably broader _ than the top of the blade. The blade narrows towards the point, the edge diverging gradually from the top until the blade reaches its widest point close to the tang. Discussion Knife blades such as this are far from uncommon _ on Iron Age sites, though they have been subject _ to little discussion. Four classes were recognised among the Danebury assemblage (Selwood 1984, _ 349-51), the East Chisenbury example falling into _ Class 2 — knife blades terminating with a tang for insertion into a handle. The blade shape further places it within form c of Class 2. There is at present no evidence of any chronological sensitivity among the identified Classes. Tron fragment from Trench B layer 17. | 21mm xl5mm x 13mm. Weight: 7g. One small _ fragment of iron was recovered during the excavations. | Itcomprises a small, amorphous lump, with rust and _ corrosion preventing any further identification. As Collard et al. 2006 have made clear, the presence _ Table 16: Stone from Trench A * Several pieces of flint in All have concretions attached. deg ee See Pea ey Peg | eee isa Sees fe eee [Burnt & pot _ [140 — some very large Ee es frags (Chalk [I frae oe ae Igudidledichalke lo | of objects or fragments of iron at a site like East Chisenbury should not occasion any surprise. Stone, by David Field Struck flint was initially recovered from rabbit scrapes on the surface, but several hundred pieces were also recovered from excavated contexts. As can be seen from Tables 16 and 17, occasional flakes occurred in various deposits, but the greatest quantity was recovered from Trench B layers 24 and 35. Only one piece appeared to be residual, a fine long tertiary flake from Layer 24, completely out of character with the rest of the assemblage and probably Mesolithic or Early Neolithic in origin. Most of the pieces are crude flakes of the form usually attributed to post-Neolithic knapping. Three flint nodules and a tabular piece were present which is hardly surprising given that raw material is never very far away on the Chalk. Formal cores are rare. Of two from B 24, one was multi-platform and two hammerstones from the same deposit were broken cores. Similarly, cores from A 9, B 20 and from B 35 were also used as hammerstones. Some of the pieces, particularly from B 35 appear to have been shattered rather than flaked, presumably placed on an anvil and smashed with a hard object. The high proportion of primary and secondary pieces indicates that knapping took place on or near the site and that most pieces appear to have eae ees ae Eben Ae Eo be iS eae elie DET Ss OT Se ee ee a Os Oe a Eee ee a eee Co Pa ae eee ee ae ae ee ee ei eS Ce eS ee ae ea a Bee ie ere a ea ee a ae ee eae ee oe eee a ee eae ee Peon! lg ae Se el a eee es ee Pe ere ae LA ee ee eS ee a eee ee ee ee ee 2s as Be ie WB 6 a ies Bees ies Ea 4lumps|4small pieces| | |i frag |_| Ee ee ee eee ee ee eee eae eee se Ee 82 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE been struck with the aim of producing immediately usable products. A lack of formal tools is evident, but worked flakes are present in layer B 3 and B 20, while 38 of 295 flakes from layer B 24 were utilised in one form or another, pieces having abraded, spalled, or in a few cases retouched edges. One utilised piece may conceivably have been a scraper and a further scraper-like utilised piece was recovered from B 8. Some of the material from A 11 has concretions attached and similar material forming a conglomerate attached to a natural stone and a ?dog jaw from B 36 (note: not included in bone report), has a group of minute flint flakes attached to it. It is difficult to account for these. They may be sweepings of the minute detritus from a flint knapping episode but in view of the shattered material from B 35 it could be that these were deliberately crushed to form filler for pottery. Some layers, particularly the upper layers, but also B 24 and 35, contained considerable numbers of burnt or calcined flints used as potboilers. Some of those from A 2 and B 35 are very large and utilised natural nodules but flaked pieces were also used. Sarsen, varying from small flakes to significant boulders, forms an important component of the assemblage. Two large pieces may have been split from the same boulder which had been used in some Table 17: Stone from Trench B by layer alate [Primary flake 5° }5) [| Primary pieces [2 [10| cau Mertiary flake 2 |S: real ay ee ee peel sche spherical el craper es el [Burnt & pot _[82|70]3 Pebble wey ae Re ee od is [aa] ea we a Real ea 3 ] Es Ee) ie Lane) al ied aia] grinding process, possibly in a circular fashion, as the surface displaying grinding is level and not dished. The two pieces which measure 36cms by 30cms by 22cms thick and 40cms by 37cms by 22cms are at the very limits of what an individual might comfortably carry. They do not join but appear to be from the same boulder, which must have been at least 1m across if not larger. The outer edge displays the curvature of the natural boulder but the inner edge is flaked, one scar presumably reflecting the size of the hammer measuring 11cm wide. Each piece has one face that is ground flat and one that has been split on two faces, presumably using the method of heating and pouring water across the stone since all split surfaces are relatively straight. Smaller fragments of sarsen come from B 22; two from B 24, one of which was pecked into a rounded shape possibly to be held in the hand, with one surface flattened; two small flakes from B 26; a piece 9cms by 8cms by 7cms ground on one face from B 28; 15 pieces, some possibly from the same boulder and two with a ground face. A slab from B 18 18cms by 17cms by 7cms ground flat on one surface and along one edge. A one-piece saddle quern 29cms by l6cms by 5cms, which has brown and amber stains on its grinding surface, comes from B 20. Chalk nodules occur in B 2, 2 A, 3.3, 6, 20, 24 and — ON — bo Ww N ms cSReS eee eSees — SESE e SSeS aRe aa a a P| a | P| ea ee EON pia a il aise Ea ae ieee ia at aaa ae ba THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 83 Table 17 continued: Trench B Stone by layer ES es ee se (peimany Sake; Ai oso |S Nim _ 35. One of the pieces from B 6 was fist sized. Chalk _ may have been residue from the chalk platforms, _ alternatively it was a component in pottery making _ or perhaps used as paint. Three lumps of clay occurred in B 35. Each lump has a flat surface and one has two as though from a _ structure. The clay is not reddened but it appears to _ havea high chalk content and there are striations and | hollows indicative of leached vegetable matter. One | piece has small fragments of flint embedded in it. One fragment of limestone was recovered from | layer B 24 and a flake from B 18. The nearest source of limestone is in the Calne area of Wiltshire, some | 30km distant. A sandstone grindstone or quern, 25cms by 15cms by 4cms, with a slightly laminar structure and only very slightly dished, comes from B 18. From | the same layer is a siliceous stone that may be flint, but is water rounded and then pecked into an oval shape. It has been squared on two sides. Chronology _ During the early 1990s, the middle centuries of the Ist millennium BC were seen as a carbon 14 black hole with problems in terms of dating. The iii Anas eel hearth Be leo eee etn ie ee eee WRrimanypiecese: (cn52 (ie I3l (8 [| ak alee (23) th | Secondary make ARE ht ale aes) Sevondary pieces 789) A | 16) Wenuanyatiakemerr menace cba eae Seg lee es Tertiary pieces COE ey eee Rae Pe a Ce See eee ea eer Oe ialaked@piccem ame MIG awl be Bn igiakedoduicm me Or oe te We ee Care: 2 a cs eee ee ee ee Pema ee ea spherical Linkin = ee ah Sa aa See eS hee ee erapecmeammtec ares rr aie Perel. ORO el Gli ree ey [Pima Ce a ee i a ee eee eee Rite TT See Sane eee eee eee eee Neel ee ee ee ree ee Somer eS ee eee ee ee a eee ei Cua oe abel i ee ee Cey, See ee eae aaa aes Pudde Gaik Te Sen ae eee eae ae WHammerstone 2.01 |2* i )c |e). Seeks eae ae fimesroncmmene Mines Meili viaie (lt oy Cie “oes aa ee ee eee eee ae eee als ales! likely errors were considered to be greater than the chronology provided by the artefactual data, so on advice no attempt was made to obtain C14 dates. Instead magnetic dating was attempted. Fifteen samples from a rammed chalk layer with evidence of burning (Context 29 and 30) were taken by Tony Clark but with uncertain results. Sadly, Tony died before a report could be prepared. The presence of a Sompting socketed axe from the friable loamy clay (Trench A context 11) at the base of the sequence suggests that the initial activity on site, represented by post and stake holes and possibly hearths, occurred no earlier than 800BC. The pottery supports this, and while there are a number of forms with wide date ranges in the earlier half of the 1st millennium BC those with more restricted dates bracket the period between c. 800 and c. 600 BC. Several types represented in the basal layers occur no earlier than 800BC at Potterne. All Cannings Cross geometric motifs on pottery from Context 35, indicate that the settlement at the base was certainly active no earlier than the 9" century BC, while pottery from the midden deposits indicates that they accumulated subsequently at sometime between c. 750 and a little after 600. Crucially, absolutely no Scratch-Cordoned Ware was recorded on site, a pottery type that is ubiquitous on nearby Early-Middle Iron Age sites and which 84 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE is introduced by the middle of the 6th century cal BC at the very latest, suggesting that the process of deposition had ceased by this date. Discussion of the Mound and Enclosure Complex at East Chisenbury The results of the trial trenches confirmed the interpretation suggested by surface survey, namely, that the mound did, indeed, form a cultural construction comprising an enormous quantity of placed material deposits. In the first trench, these were recorded to a depth of 2m, while in the second, sited closer to the edge of the deposits, they existed to a depth of 1.67m. Auger borings in the vicinity of the centre of the mound failed to bottom at 2.5m. Variously, these deposits comprised ashed grass and dung along with non-burnt materials, some bone rich, others with a greater component of charcoal and particular strata contained large quantities of butchered animal bone, broken pottery and other artefacts (Figure 14). In places, chalk deposits, some hard packed and cement-like, intervened and separated certain layers. Although sparse, the description of the deposits encountered by Walls and Bray in 1945 is similar, with their reference to cement-like chalk floors and a ‘pillar’ with ‘windows’ full of pottery and bones. The artefactual material, bone points, furrowed bowls, etc., is also apparently identical. The scale of the deposit immediately presents difficulties in interpretation. Three basic possibilities emerge. First, that the deposits represented domestic settlement refuse and that, as Cunnington had speculated at All Cannings Cross, the chalk platforms were floors on which buildings stood. If this were so it was evident that the inhabitants were living amongst their rubbish which built up around them forming what would appear to be the first recognised tell site in Britain. The second is that the deposits represent a sequence of sporadic but massive depositional events whereby huge quantities of meat (sheep and cattle) were brought to the site and consumed and their bones >... Fig. 14 A very wide range of pottery forms were recovered during the excavation. Perhaps most diagnostic are the long- and short-necked furrowed bowls (middle centre and left) as well as the incised and decorated wares (bottom centre and middle right). Decorated coarsewares such as that shown top centre, were a significant component of the assemblage. _ THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 85 _ discarded alongside a vast array of pottery vessels, some of which were involved in food preparation, storage and consumption. These components are accompanied by very large amounts of bedding, both animal and human, and substantial quantities of it burnt. Comparisons with Biblical feasts such _ as the Passover might be entertained, where the account says that ‘tens of thousands’ of lambs were _ slaughtered for a single feast, or in a modern context with rock festivals, where enormous quantities of detritus are left behind after a weekend of revelling. The final possibility is that the deposits encountered had accumulated elsewhere in discrete clusters and had been moved to the site at a later and final stage. Or perhaps there is a combination of these. In any event it is clear that a substantial component of the site comprised what might be considered to be disposed domestic refuse and consequently the site _ has been termed a ‘midden’. Since completion of the investigations a number _ of developments have taken place that directly impact upon our interpretations, analysis and understanding of the deposit at East Chisenbury, and indeed, other superficially similar deposits further afield. Needham and Spence have now published their detailed analysis of the refuse disposal pattern at Area 16 East, Runnymede Bridge (Needham and Spence 1996), while the Potterne excavation report has been published (Lawson 2000 formerly available as interim reports Gingell and Lawson 1984, 1985). More recently a major landscape assessment of the Vale of Pewsey in the Late Bronze Age-Early Iron Age has re-investigated the site of All Cannings Cross and revealed a number of contemporary sites _ nearby (Barrett 2003), while a project examining the Iron Age and Romano-British settlement and landscapes of Salisbury Plain has also come to fruition (Fulford et al. 2006). All of this work affords us now a much fuller means by which we can assess the deposits at East Chisenbury and place them in a wider context. It should be stated, however, that the analyses presented here are neither comprehensive nor detailed but rather seek to offer a commentary on what is evidently a highly complex sequence of archaeological remains. Lawson commented (on Potterne) that, ‘The unprecedented nature of the site and quantity of recovered data led, perhaps inevitably, _ toa rather introspective view’ (2000, 3). Work at East Chisenbury had a similar effect and has led to the same sorts of contemplation — the scale of the site, the complexity of its landscape setting, and the vastness of its associated material culture are remarkable. The Mound The unusual nature of the site, especially the mound, the contrast with anything that preceded it, was contemporary with it, or indeed, post-dated it, certainly within a prehistoric context, immediately sets East Chisenbury apart from the vast majority of sites/settlements on the chalk downlands in later prehistory. The deposits and cultural material encountered have been defined as a midden, and it is useful here to outline a number of points which have been made concerning middens and middening activities in general: a midden being:- ... an occupation deposit relatively rich in refuse, including archaeologically less tangible elements — notably decayed organics and comminuted structural clay — and with evidence for the deliberate and sequential accumulation of refuse at one location (Needham and Spence 1997, 80). Furthermore, Needham and Spence (ibid., 84-5) have outlined four broad categories in terms of function, noting that these cannot be exclusive of each other:- i. Middens as ‘economic’ function. ii. Middens from production and processing. iii. Middens from site management/social practice. iv. Middens as symbolic structures. This builds on analysis and discussion of a number of earlier sources, especially Schiffer (1976; 1987), whose quadruple division of refuse categories 1s still widely quoted. In this he refers to primary sources of discard, material or rubbish left at source of use; secondary, where waste has been deposited away from area of use; de facto, material discarded at point of use but still had a real value; and provisional refuse, clusters of debris accumulated before re-collection and final deposition. The life history of material incorporated within these various processes is sometimes referred to as a waste stream, but better termed refuse cycle (Needham and S¢rensen 1988). The accumulation of refuse within a domestic setting is clearly determined by a number of guiding principles, both planned and haphazard, and was likely to have been influenced by the scale of the contributing communities, the intensity and frequency of disposal and the relative importance of other socio-economic factors, such as agricultural practice. For example, the collection of midden material may have been encouraged as a component of a risk management strategy to ensure continued 86 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE fertility of the soil as part of the manuring process, although, as Needham and Spence point out, *...the practice of accumulating refuse does not automatically point to the practice of manuring’ (1997, 84). Midden mounds are not necessarily linked to a refuse curation strategy, however, and Schiffer notes that more frequently waste collects in a thin veneer across the surfaces of most sites (1987, 45). If, indeed, East Chisenbury is a mound comprising a series of midden accumulations, which were not built up deliberately, its scale and extent make it even more remarkable. Our 21st century perception of refuse is largely ruled by the dustbin and dustcart, a relatively modern innovation originally introduced to clear towns of the heaps of ash (dust) from domestic hearths. There was until recently no ‘domestic refuse’, all detritus being sorted and passed through a line of totters, etc., and effectively recycled (Eede 1971: see also Chapman 2000). Similar processes no doubt operated in prehistory, yet at East Chisenbury we see ‘refuse’ carefully curated and used in a completely different way. There are glimpses of this kind of treatment from relatively early in the Neolithic period, deposits of ‘domestic debris’ being deposited in some long barrows, at for example, Skendleby 1, Lincolnshire (Philips 1936) or in causewayed enclosure ditches, for example at Whitehawk (Curwen 1934; 1936). Ashbee (1976) found similar deposits of ‘occupation earth’ in chambered cairns on the Isles of Scilly and suggested that there may have been a received association between settlement debris and plant development and growth. One of the best examples is probably in the filling of the passage and chambers at West Kennett long barrow, where deposits of chalk rubble were interleaved with layers of earth containing charcoal, beads, pottery and previously burnt animal bone. The potsherds from at least 150 vessels comprised several distinct types, Ebbsfleet, Mortlake, Grooved Ware and Beaker representing over 500 years and as much as 1000 years in chronology, but crucially all deposited at one time. Conjoining sherds from the same vessel were found scattered both horizontally and vertically throughout the deposit, including in different chambers (Piggott 1962, 27, 33, 68-71). The excavator made the point that the material could only have been collected from elsewhere and deposited as one during the Beaker period, i.e. the latest of the pot types. Similar processes may have taken place at Tynings Farm, Somerset, where round barrows excavated by Taylor (Taylor 1931-5, 67-102) were found to contain deposits of animal bones, potsherds, flint, black ash and charcoal. Taylor went to great pains to emphasise that the material had not accumulated in situ, but had been collected and brought in from elsewhere. What seems apparent from the East Chisenbury deposit is that there is broad uniformity within each of the stratigraphic units observed in both Trench A and B. This creates, perhaps, an illusory impression that the deposits were laid down in only one or two major episodes. The processes of pre-deposition and accumulation at East Chisenbury appear more complex than that first suggested by both the observed stratigraphic units and, in particular, the pottery sequence. Within each broad layer, some of which were up to 0.4m in thickness, although there appeared to be a general homogeneity, there was, in fact, a great deal of variation. Small lenses of material were apparent; patches, smears and spreads of charcoal, small clusters of bone or pottery could be seen. This admixture may well represent separate components in isolated waste streams and reflect the different treatment of varied waste types at different times and in different places. It is plausible that there were a number of discard and accumulation strategies at East Chisenbury that saw particular sorts of debris residing in separate locales (provisional refuse) before final collection and deposition in what became the midden mound — in this case, the ‘locus of ultimate incorporation’ (Needham and Spence 1996, 24). It is entirely feasible that different sorts of materials were treated in different ways. At East Chisenbury, it might be reasonable to interpret a thick layer of midden such as Layer 3, Trench A or Context 4, Trench B as having apparent uniformity but really comprising only one major constituent, ashed sheep dung, mixed at the ultimate stage of disposal with materials drawn from other geographically distinct depositional contexts, each with its own waste accumulation trajectory. Looking in detail at the deposits in Trench B, for example, there were at least nine broad bands of midden deposition. These varied in thickness across the excavated area and were occasionally separated by layers of compact chalk, though these, too, were frequently interwoven with layers of midden material. The composition of each layer, on initial inspection indicative of one ultimate depositional event, has been characterised as largely deriving from human bedding and/or sheep/goat stabling waste, ashed material, dung, straw and grasses but within each there were also varying quantities of animal bone, ranging from burned and heavily fragmented through to a few articulated joints, as THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 87 well as fire-cracked flint, charcoal and unburnt flint and pottery, etc. Much of our initial assessment of the mound deposits was predicated by the assumption that the deposits arose from in situ deposition as part _ of continuous or, at least, sequential phases of _ occupation. This now seems unlikely. Macphail’s _ analysis suggests that much of the parent material within the midden build up was unweathered. If this was in situ it implies the provision of a roof over parts of the site; alternatively, and perhaps more likely, the material must have collected elsewhere from a location unexposed to the elements. The good condition of bone surfaces and the small number of articulated bones found support this view. Macphail notes that signs of animal trampling were limited to _ the base of the deposit above the Clay-with-flints. Given the depth of the deposits overlying here, this implies that enormous quantities of bedding and animal dung were gathered and laid down, no doubt originating from the very large number of animals which must have been kept - or at least killed — at the site. One possibility is that much of this material derived from roofed or covered stables at another location for sheep (and to a much lesser extent, cattle and pigs) and that the waste they generated was moved and stored elsewhere. Some of this occurred in an ashed form, though whether it was burnt in situ _ or brought in as ash and deposited in that form is _ not clear. The lenses and layers of midden material are soft and friable and would not have survived _ (as they so self-evidently have) the trampling that | would have accompanied occupation, or indeed, in | situ penning of cattle or sheep. The accumulation process outlined above, characterised as comprising one major constituent | gathered from areas of initial accretion and then ' embellished with a number of additives derived from different locales at different times and rates, makes best sense of the deposits uncovered. This, then, could explain the paradox identified by Raymond (this report) in which the lowest levels of the deposit include a diverse range of vessel forms, both early and late together. It is worth speculating, therefore, that these distinctions cannot be explained at East Chisenbury by chronological change but are, rather, determined by some form of ‘choice’. _ The mixture of different groups of ceramic vessels _ surely points to the translocation of previously deposited material to its final resting place in the _ mound. There is a possibility that some re-working or re-ordering of the mound deposits may have taken place, particularly when chalk floors were laid or posts inserted or withdrawn, but the chalk surfaces themselves effectively seal, to a degree, the depositional events. This implies that either in the mound or at another location nearby, there were other dumps of refuse, perhaps collections of specific categories of waste such as bone or pottery, which accumulated over a long period of time and were then moved to the site of final incorporation. In fact, smaller spreads of refuse were noted in the plantation to the south-east of the mound and these could, indeed, be the remnants of other smaller accretions of debris — lesser dumps of material that were moved and ended up in the area of the larger midden mound. Future research on the site should focus on a detailed investigation of the compact chalk horizons exposed in both trenches. Similar sorts of rectangular ‘structures’ were uncovered by Cunnington at All Cannings Cross (1923) though termed ‘pavements’ by her, and these varied in size greatly from 1m? through to areas close on to 10m x 6m. These were accompanied by deep soil sometimes fire reddened and other midden debris and were interpreted as the bases for rectangular huts. It is not possible to assign a definitive purpose for the spreads of chalk encountered at East Chisenbury but they clearly played several important roles in fhe mound composition. In no instance were they shown to be associated with hearths or burnt areas; in fact, their ‘clean’ appearance was noted frequently by the excavation team. On occasion it was also observed that their upper surfaces appeared worn or heavily compacted with a thin veneer of fine chalk soil wash. This may represent the effects of trampling by humans or animals but it might also conceivably derive from exposure to the elements — heavy rain can have the same effect on chalk land surfaces. In short, though, it does not seem that the chalk floors are occupation surfaces and there are a number of alternative suggestions. The uppermost chalk horizon is, however, particularly noteworthy. It occurs at the same depth and with a similar thickness in both trenches (Layer/Context 5) and auger survey confirmed its presence across the entire mound. It is reasonable to suggest that at a late stage in the process of accumulation, the entire deposit was sealed within a thick layer of chalk. This would have presented a striking sight; a gleaming white mound perched in a prominent location and it would undoubtedly have underscored the visual impact of the mound. It did not, however, mark the end of depositional events as there are truncated midden deposits on top of this layer. The chalk layers may well be some form of 88 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE capping device placed on top of the refuse in order to seal it or to provide stability for the deposit or neutralise smells or acidity. The excavation was not extensive enough to determine the overall extent of the chalk layers within the main body of the mound, and it may well be that they are spatially localised events marking the end point in sequences of deposition. Interestingly, these lesser chalk spreads occur at similar levels in both trenches suggesting, possibly, a chronological link and possibly a contextual one too. A midden deposit consisting of both animal and vegetable waste products would have been a noxious mass of detritus and a potential source of disease and illness in general, although the inhabitants would not have been aware of the epidemiological connection between the two. With this in mind it is noteworthy that one potential by- product of the chemical interface between the chalk floor and the underlying midden deposits would have been the production of calcium polysulphide, lime sulphur, a well known insecticide or fungicide. Again, the significance of this might become clearer with further, more extensive, investigation. The floors may have been more intimately connected with the depositional process, either as punctuations in the sequence of dumping, or alternatively, as platforms from which the deposition was undertaken. Detail on the sequence and process of deposition, however, eludes us at present. One other possibility is that the spreads of chalk are the re-deposited remains of floors or walls from demolished structures. It is worth noting that the chalk surfaces may have been accompanied by timber structures — these could have been used to revet areas of substantial deposition (Dave Norcott pers. comm.). At least two posthole cuts were noted in the sections, though not in any clear structural relationship with the chalk levels and, furthermore, in the north-western corner of Trench B it is obvious that a subsequently chalk-filled feature, marked by arcing bands of flint and chalk, cuts through one of the chalk contexts. Although, the meaning of the chalk floors is unclear, the origin of their parent material is more tangible. The large circular hollow on the north-western flank of the enclosure is the best candidate for the source of the chalk incorporated in the mound. Its close co-location with the mound and enclosure complex and a lack of similarity to other more recent extraction pits strongly hint at a prehistoric date for the hollow. It was, however, not simply a source of raw material for incorporation in the mound and, like the mound, its ‘construction’ must have been a substantial undertaking matched by its significance for the local community. Its form and setting is amphitheatre-like with an entrance approach from the dry valley to the north and an interior that is double-tiered. Further work at the site should certainly seek to clarify the form and chronology of the hollow but it may well be that this significant location, adjacent to the midden, was a place of public congregation and social ritual. At the base of the deposit and separated from it by an earthworm-turned ploughsoil, were a series of features that formed part of an earlier settlement. Two circular hearths were noted along with a number of shallow post-holes, presumably part of timber Late Bronze Age structures. These features were associated with sherds that displayed Plain Ware affinities and it is unclear whether these belonged to an open site or one that lay within the enclosure. Apart from being the earliest features on site their precise chronological status could not be determined. The settlement The nature of the ‘settlement’ features beneath the midden is not understood and remains as enigmatic as the midden itself. That one posthole should cut through one of the small chalk platforms implies that they were not used contemporaneously and while the pottery from this level is of All Cannings Cross type it is quite possible that some or all of these features are independent of each other. The platforms are relatively close set to one another and are therefore unlikely to represent hearths from separate domestic dwellings. They could, of course, represent some industrial function. The burning events on the chalk platforms, however, appear to be small-scale and the lack of associated charcoal, or oven walling, suggests that they may have been short-lived. It is possible to entertain a ceremonial or even sacrificial function for these and it is worth considering that were it not for the reddening of the platforms there would be little to distinguish them from similar platforms higher up the midden profile. Macphail mentions possible trampling features at the Clay-with-flints junction and in marked contrast to the midden material, notes the presence of earthworm-worked loose crumb and sub-angular blocky material at the Clay-with-flints junction. Whether this occurred during everyday use of the site or after its use had come to an end, is unclear. At some point use of the site appears to have changed, firstly, part of the site, and subsequently, the whole area coming into cultivation and this is indicated --—- -- — THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 89 Fig. 1S A view of the archaeological complex taken from the north side of the dry valley to the north-east of the site. Recent tracks have gouged hollow ways that ascend the slope towards the mound, and these mirror the approach axis of the linear boundaries noted by Entwistle and Raymond. The outline of the enclosure can just be discerned as slightly darker band of grass immediately downslope from the mound, which is marked by white posts. by the potential ploughsoils close to the base of the deposits and pre-dating the accumulation of midden material. This implies a clear chronological distinction between settlement and cultivation and between this cultivation episode and the midden. The (?pre-midden) settlement was evidently returned to agriculture and the subsequent midden formed on a green field site. The Enclosure The precise nature of the enclosure and its date remains uncertain as is its relationship to both the ‘settlement’ and the midden. Curvature of the enclosure as depicted on plan (Figure 3) suggested that the midden partially overlies it. Recent _ observation in more favourable ground conditions, however, identified slight traces of an earthwork that followed a different course set away from the midden on its southern arc. It is possible that this reflects an episode of cultivation influenced by curvature of the mound, but alternatively it could represent an enlargement of the enclosure belling out again similar to the hillfort at Quarley Hill (Hawkes 1938); an enclosed site associated with linear boundaries and an early midden deposit too. A plot of the features depicted on air photographs (NMR SPTA SU15SW) depicts the curvature at the north-east with an arm swinging out towards the south-east. If this extension is indeed part of the enclosure then it is one of the largest on Salisbury Plain, and lies in a relatively prominent landscape position (Figure 15. This contrasts with the locational preferences of a number of other, later, Iron Age hillforts and enclosures in the local area. Sidbury Hill, for example, occupies a prominent ridge-top situation with commanding views in all directions. Both sites, however, are the focus of a number of linear earthwork boundaries. At Sidbury, five lines of linear earthwork radiate out from the hilltop and the northernmost segment has a construction date 90 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE sometime between 1000-750 cal BC (Bradley et al. 1994). These features are probably related to an early phase of enclosure here, unconnected to the developed hillfort, but instead associated with the outer line of enclosure, clearly the earliest in the sequence on the site. Similarly, at least six linear features have been noted emanating from the site at East Chisenbury. At one stage it was presumed that the concentration of these features was connected to, and reflected the contemporary significance of, the midden deposit but this now seems unlikely. The most likely scenario is that the linear boundaries are associated with the earliest documented phase of activity here, represented by post-holes and hearths sealed beneath a well developed ploughsoil at the base of the mound, and associated with Plain Ware post-Deverel- Rimbury pottery and Llyn Fawr phase metalwork. This may or may not have been set within an enclosure; if the Plain Ware settlement was within an enclosure it is entirely anomalous on this part of Salisbury Plain where, without exception all of the known settlements of this date are open. In this way, the East Chisenbury sequence replicates a pattern witnessed at a wide range of local and regional sites — Sidbury Hill, Lidbury, Casterley Camp, to name but a few. In all of these other cases, though, it is apparent that the linear boundaries precede the first phase of enclosure, often by a very short time period, and that these linears cut through and therefore decommission existing field patterns. If this is correct then it seems best to surmise that the linear boundaries were the earliest identifiable monumental component of the hilltop at East Chisenbury, perhaps focussing on a currently ill- defined pre-existing settlement or even some other structure such as a long or round barrow; and here it is worth bearing in mind the suspicion that a pre- existing mound was slighted by the construction of the enclosure boundary on the north-east. Generally thought of as land unit boundaries, their purpose in relation to the site in question remains uncertain and a major subject for further investigation. It is plausible that any associated occupation, the construction of the enclosure and the earliest phases of midden development, are broadly contemporary. Why So Big a Mound? The investigation at East Chisenbury immediately raises a number of fundamental questions. Two of the most basic are: 1. Is East Chisenbury unique or was it typical of all settlements at the time? ii. What were the social-political-economic conditions that caused the remains of vast quantities of ashed bedding, dung, huge numbers of animals, pottery, and an unusually large number of artefacts, to end up in the mound? Despite attempts otherwise, knowledge of settlements of the Late Bronze Age, especially on Salisbury Plain and in the area around the site, remains insubstantial. By way of contrast, settlements immediately pre- dating the mound and enclosure, such as the earliest phase at East Chisenbury, appear to consist of small-scale spreads of unenclosed activity, often in inconspicuous locations on valley sides off the crests of the ridges or hilltops and frequently in watershed locations or close to springlines and ponds. It is likely that many of these open settlements had been abandoned before All Cannings Cross type ceramics emerged, prior to 800 BC. Intriguingly, sites that appear to have been abandoned often had restricted views to the linear boundaries that may well have defined individual ‘territories’ (Fulford et al. 2006). Frances Raymond (pers. comm.) has speculated that there was a move towards locations with wider views that may have been an expression of the increasing importance/prominence of relationships between, rather than within, communities and that this was also emphasised by the related material culture changes. Based on artefactual evidence and landscape associations it is apparent that these were agricultural communities in which both arable and pastoral concerns were of significance. A number of sites on Salisbury Plain and, indeed, the Marlborough Downs, such as Dean Bottom (Gingell 1992) have revealed evidence for shallow spreads of midden deposit within an enclosure boundary and spreading across upper ditch fills. This mirrors evidence from earlier Deverel-Rimbury sites such as South Lodge (Barrett et al., 1991, 144-183) and indicates that refuse was generated as part of contemporary social and economic practices and that disposal strategies developed in order to structure its curation and deposition. Assuredly contemporary activity in the area surrounding East Chisenbury is rare indeed. This may simply reflect the nature of settlement at this time — largely unenclosed and seemingly without a similar intensity of occupation. Simultaneous developments took place at Lidbury 2km to the east THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY of the midden where an open settlement probably of the 8th — 7th centuries BC was associated with a - number of linear boundaries (Cunnington 1917). - Isolated finds of Early All Cannings Cross pottery have been made further to the east close to a cluster of enclosures at Coombe Down (Fulford et al. 2007); but as with Lidbury it seems best to view these enclosures as later developments. On the northern side of Coombe Down there are a group of three enclosures, one of which is concentric in _ arrangement. The small inner enclosure dates to the _ Early Iron Age, but excavation revealed a storage pit inside the enclosure in a position that would have placed it below the bank. This produced an ambiguous pottery assemblage, which on balance was probably contemporary with the midden material. There was also residual Plain Ware from all three enclosure excavations — suggesting that _ like East Chisenbury, each site had early origins. The large enclosure complex on the southern side of Coombe Down produced two residual 8" to early _ 6th century BC sherds, although the earliest of the _ excavated features were Early Iron Age in date. The Everleigh enclosure has similar residual Late Bronze Age Plain Ware and produced other pottery, which could only be bracketed between 800 and 400 BC. A residual sherd from a furrowed bowl came from | the enclosure ditch at Warren Hill on the other | | side of Tidworth. To the west of the River Avon, | finds from features such as pits or ditches point to | an activity of an uncertain extent in the vicinity of the later enclosure at Widdington Farm and, further | west, at Casterley Camp, finds of Late Bronze Age metalwork would suggest that this latter site played an important focal role for the local community. Further afield a number of contemporary sites lie on the Lower Chalk/Upper Greensand interface _in the Vale of Pewsey. Very few of these have been | Investigated but it is clear that a number in the | area of All Cannings Cross, Erlestoke, Westbury, _ and Steeple Ashton have produced the same sorts _of morphological and artefactual cues (Barrett and _ McOmish 2009). Slightly later activity has been recorded at _Chisenbury Trendle, where ceramics, including I | ' scratch-cordoned bowls, were found on the old | ground surface underneath the bank of the enclosure, _ indicating that this is considerably later than that at East Chisenbury. Although the evidence is scant, a _ basic sequence is apparent in which open settlements | dating to the 8th century BC was subsequently | abandoned or enclosed by the late 7th/early 6th | century BC. This dichotomy, of course, is rather 91 simplistic and it is more likely that a range of site types was evident from open, to partly and fully enclosed, with a number of locales oscillating between the two (McOmish 1989), although nothing approaching the scale, complexity and context of East Chisenbury has yet been found. The nearest analogies, self evidently, are with sites such as Potterne (Lawson 2000) and All Cannings Cross (Cunnington 1923) but as has been stated, both of these lack the monumental status of East Chisenbury. This may well be as a result of post-abandonment events. Although both Potterne and All Cannings Cross occupy markedly different topographical niches, it seems probable that the midden deposits at each would have been substantial enough to create a very prominent mound. The deposit at Potterne lies in an elevated position on a west-facing slope of greensand above the Rangebourne valley, and is probably the remnant of a previously more extensive spread of debris whose core lay on higher ground to the east. Later cultivation as well as housing and road construction, and not forgetting activities associated with the later cemetery, have all had a deleterious impact on the deposit here. Similarly, recent fieldwork at All Cannings Cross has shown there to be a substantial mounded earthwork spread over 1.5ha, largely across a ridge of Lower Chalk and coated in deposits at least 0.6m in thickness and probably as deep as 2.5m. Other spreads of debris close by indicate that before the impact of later cultivation, these deposits were once much more extensive and probably survived in a monumental form. Undeniably, the recently discovered deposit at Stanton St Bernard, 600m to the east of All Cannings Cross, and which survives to a depth of 1.4m over an area of 2.5ha, still retains a domed appearance even after centuries of ploughing (Barrett and McOmish 2009). Given a similarly aggressive history of destructive ploughing, East Chisenbury might have completely succumbed to destruction. Instead, the midden deposits here endure in a spectacular fashion, despite the impact of cultivation since the middle of the Ist millennium BC. If in excess of 2m of deposits currently survive in places on the site, then it must be presumed that at an earlier stage the residues were even thicker and highly visible from a great distance. Indeed, it may well be the case that cultivation has removed all but the basal layers of a once even more substantial series of deposits. The intensity or longevity of ploughing on the hilltop in the post-Roman period is unknown but despite the presence of ridge-and-furrow remains across 92 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE the mound, cultivation may well only have started to bite deeply into the deposits in the period after the mid-18th century reaching a nadir in the middle part of the 20th century. Romano-British and later pottery was recovered from both the Potterne and All Cannings Cross excavations and although no similarly dated material was found during the excavation at East Chisenbury there have been isolated finds of pottery made nearby. Interestingly, a 4th century coin hoard was unearthed on the north-western perimeter of the plantation at East Chisenbury apparently buried in the slope of the ancient field boundary that cuts across the midden (Ordnance Survey 495 record card). It is tempting to view this as a deliberate act of deposition at what was still a recognisably important and propitious location. It is clear, however, from the evidence at East Chisenbury, as well as at Potterne and more recently at All Cannings Cross and Stanton St Bernard, that these debris accumulations belong largely to the 8th and 7th centuries BC. Earlier pulses of deposition are apparent at Potterne though, and further afield at Runnymede Bridge, but the massive accelerations in deposition belong to a phase when a series of significant transformations were being worked in the Wessex landscape. This has been widely acknowledged (see Barrett 1989; 1994 for fuller discussion) and is simply characterised as a reorganisation in the settlement pattern alongside radical developments in related material culture and, as Serjeantson (2007) has argued, animal husbandry. The decline of bronze as a prestige material is evident and the distinctive Llyn Fawr metalwork phase, marking the death throes of an extra-regional, possibly international, flow of goods exchange, belongs to this period. Conversely, this was matched by a corresponding surge in local craft production and specialisation, and, at All Cannings Cross, evidence of early iron-working. The most pronounced transformation though is witnessed in the appearance of what is a new repertoire of ceramic forms — early All Cannings Cross pottery with its characteristic vessel shape and wide range of decorative motifs, and a distinction between coarse and fine wares. It is these ceramics, in the absence of radiocarbon dating, that have provided a dating scheme for the deposit at East Chisenbury. The most significant components of the early All Cannings Cross sequence are evident — short- and long-necked furrowed bowls and related ceramic forms — and at the base of the mound there are fragments of post- Deverel-Rimbury Plain Ware. This would give a date range of between the early 8th and late 7th centuries BC for the accumulation and underlying settlement. This accords well with the finding of the Llyn Fawr socketed axe and tanged chisel close to the basal deposits and the close co-location of a number of linear earthworks elsewhere on the Salisbury Plain Training Area, whose primary phases are dated to the 10th/9th centuries BC. No later ceramics, such as scratch-cordoned bowls, were noted so it is clear that the process of accumulation had ceased by the end of the 6th century BC. The exact chronology of deposition remains unknown but it may well have taken place over a very short period of time, perhaps as little as a century. This is a startlingly short history of collection, but if our hypothesis about the refuse accumulation strategies is correct, it probably reflects the movement of material already deposited at one or more primary sources to its final resting place at East Chisenbury. What marks sites like East Chisenbury apart is the scale of the levels of deposition. Assuming some degree of uniformity across the mound - and the auger survey along with the evidence from Walls and Brays work do not indicate otherwise - the animal head count, based on numbers of mandibles only, suggests that, of an assumed one hundred year period of occupation, up to 1400 sheep/goat, were killed and consumed each year, alongside 600 cows and 450 pigs. Doubling the period of occupation does little to assuage the enormity of these numbers. These are staggering figures and cannot be representative of a ‘normal’ farming community. It has already been suggested that many of these animals may have been culled in the spring, summer or autumn, and if so, this concentrates the ‘use’ of the sites in even more intensive periods of activity on or, more likely, close to the site. It is difficult to escape a conclusion that, at least in part, points to the vast quantities of meat produced and consumed as being the product of a series of complex social rituals including feasting. These figures suggest that if several thousands of animals were dying or being killed at the site each year, the herd and flock sizes must have been enormous. Pryor independently reached the conclusion that sheep herds were very large in the Late Bronze Age, based on territorial organisation in the Fens (Pryor 1996). This would have brought with it anumber of requirements for stock management such as grazing rights and the availability and supply of fodder and water. It is easy to appreciate the strains this may well have put on the landscape but it does make sense of the apparent landscape reorganisation that saw large tracts of pasture THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 93 develop in the 10th-7th centuries BC (McOmish et al. 2002). The problems involved in controlling herds of animals that may have matched the size of post- medieval herds, potentially numbering hundreds of thousands, suggests that East Chisenbury drew upon a very wide but as yet, ill-defined, hinterland. _ Macphail’s identification of grass/herb bedding is also relevant here, for it must surely include hay, and it indicates the extent to which the pastures would have been managed in order to support such large _ numbers of animals. While the presence of cattle might be taken to imply dairying, this aspect of economy is given additional emphasis by Serjeantson’s observation of the large number of lambs killed within four _ months and thus leaving lactating ewes which | also provided milk. Whenever and wherever processing took place, this would have created _ its own demands, in particular, those relating to _ butchery. Only one possible knife fragment was found at East Chisenbury but other suitable items of equipment such as saws were absent, yet large numbers of bones exhibited cut marks. Of the other local sites, only All Cannings Cross has produced evidence of a suitable tool kit (Cunnington 1923, plate 20, opp. p125). There are other indications of | activities taking place elsewhere. Large numbers of _ spindle whorls were present at East Chisenbury, and | there must be enormous numbers contained in the rest of the mound, but there were, surprisingly, no _ loom weights or perforated slabs — items elsewhere representative of domestic activity of this period. Similarly, Serjeantson has pointed out that older wool animals are absent from the assemblage and must have been killed elsewhere. The available evidence, from a very small sample size, also strongly suggests that arable cultivation was an important factor in the vicinity of East Chisenbury. Wheat, oats, and barley arrived at the _ Site in a processed state and arable weeds occurred | ina variety of contexts alongside a number of fruits, _ including a surprisingly large quantity of poppy seeds. Given the extent of pasture on the chalk downland this cultivation can only have taken place on the valley floor and lower slopes, perhaps immediately around presently unknown settlement _ locations. It is unlikely that the ‘construction’ of the massive midden mound at East Chisenbury conforms to some initial design. Instead, it represents the persistence of deposition in a particular locus over _ along period of time, perhaps as much as 150 years. It may be envisaged that it emerged out of a more general spread of waste alongside smaller dumps of detritus, and, as it grew, further additions of material referenced and enhanced its ‘role’ as a repository of waste. In some ways this recalls the complex deposits assessed by Barber at a number of locations on the Western Isles (Barber 2003). Here, a range of sites were investigated and a number incorporated features and structures closely akin to those observed at East Chisenbury. The composition of sites such as Baleshare and Balelone included natural and accreted sediments as well as substantial quantities of midden detritus. Significantly, at these sites, and in marked contrast to East Chisenbury, the physical properties of the deposits had been altered to create plaggen soils, fertile environments for arable cultivation. The importance of middens as symbolic features or status indicators is widespread both in antiquity e.g. the Zerramara of the central Italian Bronze Age (see, for example, Pearce 1998); the Neolithic ashmounds in southern India of 3000- 1200BC (Johansen 2004); arguably, the farm mounds of the Orkneys and Scandinavia (Bertesen 1979) and in more recent times when an individual has been prosecuted for stealing cow dung from a cattle enclosure during an annual sacred fertility ceremony in Swaziland (Canadian National Post 16 March 2000). This is succinctly summarised by Needham and Spence (1997, 85) who saw ‘ that refuse heaps could have developed a particular symbolism, in certain societies...Refuse has links with fertility where the value of green midden as fertiliser was recognised, and more generally to the cycle of death and renewal’. The constitution of the midden mound at East Chisenbury with its complex stratigraphy and enormous collections of debris, identify it as something that was unusual within its contemporary environment. What seems clear is that a large proportion of the detritus recovered, including large quantities of animal bone and pottery, may well represent consumption of the product of the agricultural cycle, through feasting. Spielmann (2002), for example, has suggested that the need for social ceremony drives the intensification of food production and that the increase in economic effort is a product of ritual performance. Regardless of the symbolic properties of these collections of waste, sites such as East Chisenbury, Potterne and All Cannings Cross all point to a series of significant transformations in the regional landscape during the early 1st millennium BC and, furthermore, that this was directly related to, and focussed intensely on, the fertility of the land and the success of the agricultural cycle, in particular, 94 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE its pastoral component. These were communities undergoing rapid change technologically, culturally and ideologically, but also environmentally. The massive accumulations of debris mark a shift from upland chalk landscapes dominated by small- scale dispersed settlement groups to one where enclosed elements, in particular hillforts, played an increasingly important role. Thus, in highly general terms, we may be looking at smaller groups of individuals, families, or communities, coalescing into larger, though still small-scale, polities. Socially conditioned and highly structured activities such as feasting then played a pivotal role in choreographing the scale and form of social reproduction evident on these sites. As Barrett notes (1989, 313), the ideology underpinning the structuring of age, gender and other social relations, including the establishment of new forms of authority, may well have been communicated through a series of symbolic events revolving around the preparation, display and consumption of food. This, then, is a ritualised process and much of the debris from this was ultimately incorporated with other settlement and agricultural waste in the mound. It is hard to disagree with Needham and Spence’s assertion that some middens may have assumed “‘...connotations of affluence and social success, indeed perhaps even become a mark of territorial dominance.’ (1997, 85). Ata developed stage in its life history, the waste collection and deposition strategies at the East Chisenbury midden would have taken an overtly monumental form (before compaction and cultivation of the mound it must have been of considerably greater dimensions and dwarfed all but the largest barrows); the mound of debris rising up from the surrounding area making a bold statement about relative status and the significance of waste accumulation, perhaps as a repository of agricultural wealth in the form of manure. That it was not excessively plundered is apparent in its monumentality and makes the point eloquently that it may well have been the accretive process, the adding of settlement and agricultural waste, that underpinned its value and ‘role’ within the contemporary social milieu. A wide range of artefacts encountered within the deposit includes craft items such as spindle whorls, bone pins and awls, as well as a number of personal ornaments, beads and pendants. These can be regarded, of course, as everyday objects, but the same cannot be said for the pieces of human skull and other skeletal fragments found within the deposit. Although burial practices in the Early Iron Age are not well understood, Briick (1995, 255) comments that: Attitudes towards the dead as part of the general category of ‘refuse’ must also have been very different from today. Human remains were used, displayed, and finally placed in special deposits in locations very different from today’s cemeteries....ncorporated into specific contexts in the settlement.’ Nonetheless, the evidence, such as it is, characterises a diverse and, to us, unfamiliar mortuary practice — one that witnessed human bone apparently becoming intermixed with a range of other detritus. All of the bones from East Chisenbury are single and disarticulated and include juveniles and adults. To modern minds there is, perhaps, an assumption that in some way the human remains incorporated in the deposit had somehow or other become regarded as rubbish, as part of a waste stream that ultimately terminated in the final deposition. This apparent disregard for a properly observed mortuary ritual is surprising and stands in stark contrast to practices that preceded it and, were current elsewhere in the British Isles and indeed, developed thereafter. At the moment of their final inclusion the human bones were decayed/abraded to a stage where they may well have been inseparable from other animal remains and, at East Chisenbury Trench B, occur throughout the entire deposit. This argues against, for example, a single act of human burial or deposition or the possibility that an earlier burial site had been disturbed and that the remains then became incorporated in the debris flow. It may well be more appropriate to view the human remains within the deposits, and those, too, from sites such as All Cannings Cross, Stanton St Bernard, and Potterne, as having gone through a long and complex series of mortuary rituals leading to their final inclusion alongside a range of other ‘domestic’ material. This alludes to detailed and complex mortuary practices better documented in Neolithic and Early Bronze Age contexts and indicates very forcefully, that there are a whole range of mortuary activities in the Early Iron Age that have not been anticipated and are, as yet, poorly perceived. The stages from death to incorporation of body parts are not at all clear. The notion that these individuals are indicative of communities that eschewed a burial process or were unaffected by the loss of life is untenable but it is a paradox when faced by human remains discovered in an ostensibly non-mortuary location. Cunnington too, recognised this in her excavations at All Cannings Cross where up to 12 individuals are represented from her trenches: ‘They [human remains] were all quite detached from one another, and were mixed promiscuously with the fragments THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 95 of other bones, pottery, etc that had accumulated as rubbish on site....1t 1s difficult to account for these pieces of human skulls apart from any other human bones....it hardly seems probable that the people of the village would have treated the remains of their own people with such scant decency...’ (Cunnington 1923, 40). Cunnington noted only the remains of skull _ fragments and indeed, posed the question ‘where are the other bones?’. More recent excavations at the site have produced other human remains, including an adult humerus (Barrett and McOmish 2009) and so it would seem that Cunnington either did not recognise or failed to collect other human skeletal fragments. The predominance of skull pieces led Cunnington to suggest that the remains were related to a custom of head-hunting, or at the very least, a fetish for collecting heads — a belief that has subsequently gained widespread acceptance as being a feature of Iron Age communities in north-western Europe. The evidence from Early Iron Age deposits such as East Chisenbury, All Cannings Cross, and Potterne points to mortuary practices altogether more varied and complex. Although caution must be exercised due to the small sample size, it is plausible that taphonomic | processes are at work and that, as is expected, larger skeletal elements survive disproportionately. Long bones and skulls are over-represented generally. _ Within the massive accumulations of detritus, | all body parts are represented but with a marked _ absence of the smallest bones — this is as true of human material as it is of cattle, sheep and pigs. Thus Cunnington’s (and others) interpretation of the | preponderance of skulls as witnessed at All Cannings Cross and, initially at East Chisenbury, is illusory. _ We know nothing of the status of the individuals concerned — some may have been killed, fatalities as a result of internecine conflict, even sacrificial victims or, more likely, all may have died of natural _ causes. Regardless, a significant transformation _ from living individual to fragmentary, disarticulated _ bone is apparent. Thereafter, the mortuary ritual remains unclear. Once again, it must be presumed that at an ultimate stage, the human bone, already in a fragmentary condition (and probably not | immediately identifiable as human), was moved from a previous source to become a constituent of _ the mound. Finally, we must therefore acknowledge that the deposits, like those at Potterne and All Cannings _ Cross, host the remains of a large number of corpses. At Potterne a minimum of 15 individuals were found from a total excavated area that amounts to only 0.75% of the known site. The implications of this are clear — there may well be the remains of thousands of individuals contained within the deposit; in the case of Potterne close on to 2000 people. Even if only two bodies are represented from the excavated East Chisenbury assemblage, this would give a total maximum number of 10000 individuals based on a site sample of 0.02%. McKinley (2000, 96) has argued that it cannot be assumed that there was an even distribution of human bone across the Potterne deposit, and so this scaling up to arrive at a total number of individuals is not relevant — despite the fact that human bone was widely distributed across the site. There is perhaps understandable caution in accepting the standard methods of calculating total numbers of individuals when it is applied to human bone rather than animal bone, particularly when it derives from the same context, for the inescapable conclusion must be that these sorts of deposits are repositories for large numbers of human bodies (or fragments of). The problem lies, of course, in the small sample size, rather than the techniques of analysis applied. Regardless, sites such as East Chisenbury, offer huge potential in the study of later prehistoric mortuary rites. Accumulation of debris at the midden mound appears to have terminated abruptly in the early 6th century BC. The reasons for this are unknown. It may be that at this time the enclosure surrounding the midden reached its fullest extent but in general the large-scale accrual and deposition at East Chisenbury either ceased or shifted to a different place. This was accompanied again by another fundamental shift in the local landscape — the development of high profile, monumentally prominent, enclosures. A number such as Lidbury, Chisenbury Trendle, and close to Coombe Down, emerged at this time and together with developed hillforts such as Sidbury Hill, suggest that these sites played an important role as arenas for the sort of social reproduction and ideological communication witnessed earlier at East Chisenbury and similar locations. Acknowledgements An anecdotal description of the discovery of the site was published in 3rd Stone (McOmish 1999) but prior to the introduction of PPG 16, and to the employment by the MoD of a dedicated archaeologist. The chief purpose of the investigation was to demonstrate to the Ministry of Defence that the mound was not a ‘natural’ eminence and 96 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE to confirm that it had an important archaeological value. This was achieved and the results of our work were accepted by the military who kindly amended plans for the proposed tank road accordingly. We thank members of the SPTA Conservation Groups, the Range Liaison Officers in particular John Lock who provided every assistance along with a remarkable amount of enthusiasm, several SPTA Commandants and in particular, Lt Col. Rooke, who very kindly agreed to re-route his proposed tank track. Thanks are also due to Nell Duffie for raising initial interest in the site, the former County Archaeological Officer Roy Canham MBE, for his enthusiastic support, and Roy Entwistle and Frances Raymond for their guidance, assistance and hard work during the excavation. A great many other people spent time on the site, gave up personal leave to work on it (always unpaid), and provided excellent advice and guidance among them Robin Burgess, Deborah Cunliffe, Nigel Fradgley, and Cathy Tuck from the RCHME, additionally J. Denford, G. White, Phil Gage and Audrey Summers. Some of these, and others, including Eileen Rollo, Nora Morris, Diana Fautley, James Rankin, Julie Bushall, Polly Williams, Morag Rankin, Margery Williams, and Nell Duffie spent a considerable amount of time carrying out the post-excavation washing, sorting, and cataloguing. Deborah Cunliffe and Mark Corney provided illustrations and drew the finds to her usual high standard. In addition, we have benefitted from the discussions and advice of a wide range of colleagues most notably Dr David Norcott, Dr Andrew Lawson, Dr Andrew Fitzpatrick, Dr Stuart Needham, Mark Corney, the late Tony Clark and by no means least, the late Desmond Bonney, Head of Archaeology in the Royal Commission on the Historical Monuments of England whose subtle but important influences in the matters of archaeological fieldwork will stay with us. Subsequently, the mound has been cordoned off to vehicular traffic by wooden posts and part of the plantation that lay on the edge of the mound has been cleared and other conservation measures taken. Dr Ian Barnes, Richard Morton and most recently, Dr Richard Osgood were instrumental in this. References ADKINS, L. and NEEDHAM, S. 1985. ‘New Research on a Late Bronze Age Enclosure at Queen Mary’s Hospital, Carshalton’, Surrey Archaeological. Collections 76, 11-50 BAGUST, J., (ND), The animal bones from East Chisenbury, Wiltshire: Ritual rubbish or simply a midden deposit? Unpublished University of Leicester MA dissertation, 1996 BARBER, J., 2003, Bronze Age farms and Iron Age farm mounds of the Outer Hebrides. Scottish Archaeological Internet Reports 3. Edinburgh: Society of Antiquaries of Scotland BARRETT, J. C., 1989, ‘Food, gender and metal: questions of social reproduction’, in M. L. S. Sorenson and R. Thomas (eds.), The Bronze Age-Iron Age Transition in Europe: 304-320. Oxford: British Archaeological Reports International Series 483 BARRETT, J. C., 1994, Fragments from Antiquity: an archaeology of social life 2900-1200 BC. Oxford: Blackwell BARRETT J. C.2003. Unpublished project outline: www. shef.ac.uk/archaeology/prospectiveug/fieldwork/ allcannings.html BARRETT, J. C., BRADLEY, R. and HALL, M., 1991, Papers on the Prehistoric Archaeology of Cranborne Chase Oxford: Oxbow Books BARRETT, J. C. and MCOMISH, D., 2009, ‘The early Iron Age in southern Britain: recent work at All Cannings Cross, Stanton St Bernard and East Chisenbury, Wiltshire’, in M-J Rouliére-Lambert (ed.), De l’age du Bronze a Page du Fer en France et en Europe occidentale (X° — VIF’ siécle av. F.-C.): 565-75. Actes du XXX* colloque international l AFEAK, co-organisé avec l APRAB BECK, C. and SHENNAN, S., 1991, Amber in Prehistoric Britain. Oxford: Oxbow Monograph 8 BERTELSEN, R. 1979. Farm mounds in North Norway, A Review of Recent Research. Norwegian Archaeological Review 12, 48-56 BLACK, S, and SCHEUER L. 1996. Age Changes in the Clavicle From The Early Neonatal Period to Skeletal Maturity. International Journal of Osteoarchaeology 7, 425-434 BRADLEY, R., LOBB, S., RICHARDS, J. and ROBINSON, M. 1980. Two Late Bronze Age Settlements on the Kennet Gravels: Excavations at Aldermaston Wharf and Knight’s Farm, Burghfield, Berkshire. Proceedings Prehistoric Society 46, 217-295 BRADLEY, R., ENTWISTLE, R., and RAYMOND, EF, 1994, Prehistoric Land Divisions on Salisbury Plain: the work of the Wessex Linear Ditches Project. London: English Heritage Monograph 2 BROCHIER, J. E. 1983. Bergeries et feux néolithiques dans le Midi de la France, caractérisation et incidence sur le raisonment sédimentologique. QUATAR Band 33/34, 181-193 BROCHIER, J. E., VILLA, P and GIACOMARRA, M. 1992. Shepherds and Sediments: geo-ethnoarchaeology of pastoral sites. Journal of Anthropological Archaeology 11, 47-102 BROWN, L. 1984. ‘Objects of stone’, in B. Cunliffe. Danebury, An Iron Age Hillfort in Hampshire, Vol. 2, The excavations 1969-1978: the finds: 407-426. London: THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 97 Council for British Archaeology Res. Rep. 52. BROWN, L., 2000, ‘The Pottery’, in B. Cunliffe and C. Poole (eds.), The Danebury Environs Programme. The Prehistory of a Wessex Landscape. Volume 2, 65-113. Oxford, Oxford University Committee for Archaeology Monograph 49 BROWN, G., FIELD, D. and MCOMISH, D., 1994, ‘East Chisenbury Midden Complex, Wiltshire’, in A. P. Fitzpatrick and E. L. Morris (eds.), The Iron Age in Wessex: recent research (Association Francaise D’Etude de LAge du Fer), 46-49. Salisbury: Trust for Wessex Archaeology BROWN, G., FIELD, D. and MCOMISH, D., 2005, The Avebury Landscape: aspects of the field archaeology of the Marlborough Downs. Oxford: Oxbow Books BRUCK, J. 1995. A place for the dead: the role of human remains in Late Bronze Age Britain. Proceedings of the Prehistoric Society 61, 245-277 BULLOCK, P, FEDOROFF, N., JONGERIUS, A., STOOPS, G., and TURSINA, T., 1985, Handbook for soil thin section description Wolverhampton: Waine Research Publications BURGESS, C. 1969. Some decorated socketed axes in Canon Greenwell’s collection. Yorkshire Archaeological Fournal 42, 267-72 BURGESS, C., COOMBS, D. and DAVIES, D. G., 1972, “The Broadward Complex and Barbed Spearheads’, in FE Lynch and C. Burgess (eds.), Prehistoric Man in Wales and the West. Essays 1n honour of Lily F Chitty, 211-284. Bath: Adams & Dart. CANTI, M. 1997. An investigation into microscopic calcareous spherulites from herbivore dung. Fournal of Archaeological Science 24, 435-444 CANTTI, M. 1999. The production and preservation of faecal spherulites: animals, environment and taphonomy. Fournal of Archaeological Science 26, 251-258 CARR, G. and KNUSEL, C., 1997, ‘The ritual framework of excarnation by exposure as the mortuary practice of the early and middle Iron Ages of central southern Britain’, in A. Gwilt, and C. Haselgrove (eds.), Reconstructing Iron Age Societies: New approaches to the - British Iron Age, 167-173 Oxford: Oxbow Books CARRUTHERS, WENDY. J., 2000, ‘The mineralised plant remains’, in A. J. Lawson, Potterne 1982-5: animal husbandry in later prehistoric Wiltshire, 72-84, 91-95. Salisbury: Wessex Archaeology Report 17 CHAPMAN, J., 2000, “‘Rubbish dumps’ or ‘Places of deposition’? Neolithic and Copper Age settlements in central and western Europe”, in A. Ritchie (ed.), Neolithic Orkney in its European context, 347-362. Cambridge: McDonald Institute COLLARD, M., DARVILL, T. and WATTS, M. 2006. Ironworking in the Bronze Age? Evidence from a 10" century BC Settlement at Hartshill Copse, Upper Bucklebury, West Berkshire. Proceedings of the Prehistoric Society 72, 367-422 COOMBS, D., NORTHOVER, P. and MASKALL, J., 2003, Tower Hill Axe Hoard’, in D. Miles, S. Palmer, G. Lock, C. Gosden, and A. M. Cromarty, Uffington White Horse and its Landscape: Investigations at White Horse Hill, Uffington, 1989-95, and Tower Hill, Ashbury, 1993-4, 203-224. Oxford Archaeology: Thames Valley Landscapes Monograph 18 COURTY, M. A., GOLDBERG, P. and MACPHAIL, R. I., 1989, Soils, micromorphology and archaeology. Cambridge: Cambridge University Press COURTY, M. A., MACPHAIL, R. I. and WATTEZ, J., 1992, ‘Soil micromorphological indicators of pastoralism; with special reference to Arene candied, Finale Ligure, Italy’, in R. Maggi, R. Nisbet and G. Barker (eds.), Archaelogia della pastorizia nell’Europa meridionale II, 127-150. Revista di Studi Ligura A. LVII, Bordighera COX, P. and WOODWARD, P, 1987, ‘The Kimmeridge Shale’, in N. Sunter and P J. Woodward, Romano- British Industries in Purbeck, 164-72. Dorchester: Dorset Natural History and Archaeology Society Monograph 6 CUNLIFFE, B., 1984, Danebury, An Iron Age Hillfort in Hampshire, Vol. 2, The excavations 1969-1978: the finds London: Council for British Archaeology Research Report 52 CUNLIFFE, B., 1993, Danebury. London: Batsford/English Heritage CUNLIFFE, B. AND PHILLIPSON, D. W. 1968. Excavations at Eldon’s Seat, Encombe, Dorset. Proceedings of the Prehistoric Society 34, 191-237 CUNLIFFE, B. AND POOLE, C., 1991, Danebury, An Iron Age Hillfort in Hampshire, Vol. 5, The excavations 1979- 88: the finds. London: Council for British Archaeology Research Report 73 CUNNINGTON, M. E. 1917. Lidbury Camp. WANHM 40, 37-74 CUNNINGTON, M. E., 1923, The Early Iron Age Inhabited Site at All Cannings Cross Farm, Wiltshire. Devises: George Simpson and Co. CUNNINGTON, M. E. 1932. The demolition of Chisenbury Trendle. WANHM 46, 1-3 CURWEN, E. C. 1934. Excavations at Whitehawk Neolithic Camp, Brighton, 1932-33, Antiquaries fournal 14, 99-133 CURWEN, E. C. 1936. Excavations in Whitehawk Camp, Brighton. Third season, 1935, Sussex Archaeological Collections 77, 60-92 DAHL, G. and HJORT, A., 1976, Having Herds: Pastoral Herd Growth and Household Economy. Stockholm: Department of Social Anthropology DAWSON, M., 2005, An Iron Age Settlement at Salford, Bedfordshire. Bedford: Bedfordshire Archaeology Monograph 6 DUNCAN, H. B. and MACKRETH, D., 2005, ‘Registered and Non-Ceramic Bulk Artefacts’, in M. Dawson, An Iron Age Settlement at Salford, Bedfordshire, 125-137. Bedford: Bedfordshire County Council EEDLE, M DE G. 1971. Street Cleansing and Refuse Collection from the Sixteenth Century to the 98 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Nineteenth Century (with special reference to Surrey). Surrey Archaeological Collections 68, 161-181 EOGAN, G. 1964. The later Bronze Age in Ireland in the light of recent research. Proceedings of the Prehistoric Society 30, 268-351 ESHED, V., LATIMAR, B., GREENWALD, C. M., JELLEMA, L. M., ROTHSCHILD, B. M., WISH- BARATZ, S. and HERSHKOVITZ, I. 2002. Button Osteoma: Its Etiology and Pathophysiology. American Fournal of Physical Anthropology 118, 217-230 FASHAM, P J., 1985, The Prehistoric Settlement at Winnall Down, Winchester. Gloucester: Hampshire Field Club Archaeological Society Monograph 2 FELL, C. I. 1953. An Early Iron Age settlement at Linton, Cambridgeshire. Proceedings of the Cambridge Antiquarian Society 46 (1952), 31-42 FOSTER, J., 1990, ‘Other Bronze Age artefacts’, in M. Bell. Brean Down Excavations 1983-1987, 158-175. London: English Heritage Archaeological Report 15 FOXON, A., 1991, ‘Worked skeletal remains’, in S. Needham. Excavation and Salvage at Runnymede Bridge 1978. The Late Bronze Age Waterfront Site, 148-9. London: British Museum Press and English Heritage FULFORD, M. G., POWELL, A. B., ENTWISTLE, R. and RAYMOND, F, 2006, Iron Age and Romano-British settlements and landscapes of Salisbury Plain. Salisbury and Reading: Wessex Archaeology and University of Reading GINGELL, C., 1992, The Marlborough Downs: a later Bronze Age landscape and 1ts origins. Devizes: Wiltshire Archaeological and Natural History Society Monograph 1 GINGELL, C. and LAWSON, A. J. 1984. The Potterne Project: Excavations and Research at a Major settlement of the Later Bronze Age. WANHM 78, 31-4 GINGELL, C. and LAWSON, A. J. 1985. Excavations at Potterne, 1984. VWANHM 79, 101-8 GINGELL, C. and MORRIS, E., 2000, ‘Form series’, in A. Lawson, Potterne 1982-5: Animal husbandry in Later Prehistoric Wiltshire, 149-52. Salisbury: Wessex Archaeology Archaeological Report 17 GOLDBERG, P. and MACPHAIL, R. I., 2006, Practical and Theoretical Geoarchaeology. Oxford: Blackwell Publishing GRANT, A., 1982, ‘The use of tooth wear as a guide to the age of domestic ungulates’, in B. Wilson, C. Grigson and S. Payne (eds.), Ageing and Sexing Animal Bones from Archaeological Sites, 91-108. Oxford: British Archaeological Reports British series 109 GRANT, A., 1984, ‘Animal husbandry’, in B. Cunliffe, Danebury, An Iron Age Hillfort in Hampshire, Vol. 2, The excavations 1969-1978: the finds, 496-548. London: Council for British Archaeology Research Report 52 GRANT, A., RUSH. C. and SERJEANTSON, D. D., 1991, ‘Animal husbandry’, in B. Cunliffe and C. Poole (eds.), Danebury: An Iron Age Hillfort in Hampshire Vol. 5, The excavations 1979-1988: the finds, 447-487. London: Council for British Archaeology Research Report 73 GREEN, EF, 1979, Phosphate Mineralisation of Seeds from Archaeological Sites. fournal of Archaeological Science 6, 279-284 GRINSELL, L. V., 1957, ‘Archaeological gazetteer’, in R. B. Pugh and E. Crittal (eds.), The Victoria History of the Counties of England. A History of Wiltshire: Volume 1, Part 1, 21-272. London: Oxford University Press GUIDO, M., 1978, The Glass Beads of the Prehistoric and Roman Periods in Britain and Ireland. London: Society of Antiquaries Research Report 35 HALL, K. and SEAGER SMITH, R., 2000, ‘Fired Clay Objects’, in A.J. Lawson, Potterne 1982-5; Animal Husbandry in Later Prehistoric Wiltshire, 179-85. Salisbury: Wessex Archaeology HENDERSON, J., 1987, “The archaeology and technology of glass’, in J. M. Coles, Meare Village East, The excavations of A Bulleid and H St George Grey 1932-1956, 170-82. Hertford: Stephen Austin and Sons HEALY, EF, 2000, ‘Amber’, in A. J. Lawson, Potterne 1982-5; Animal Husbandry in Later Prehistoric Wiltshire, 219-20. Salisbury: Wessex Archaeology HILL, J. D., 1995, Ritual to Rubbish in the Iron Age of Wessex: A study on the formation of a specific archaeological record. Oxford: British Archaeological Reports 242 HILLMAN, G., 1981, ‘Reconstructing Crop Husbandry Practices from Charred Remains of Crops’, in R. Mercer (ed.), Farming Practice in British Prehistory. Edinburgh: Edinburgh University Press JENKINS, C., 1996, Do the sheep remains from East Chisenbury support the hypothesis for sheep milking in Southern Britain 1n the first millennium BC? Unpublished University of Southampton BA Dissertation JOHANSEN, P. G. 2004. Landscape, monumental architecture, and ritual: a reconsideration of the south Indian ash mounds. Journal of Anthropological Archaeology 23, 309-330 JONES, G. G., 2005, “Tooth eruption and wear observed in live sheep from Butser Hill, the Cotswold Farm Park and five farms in the Pentland Hills, UK’, in D. Ruscillo (ed.), Recent advances in ageing and sexing animal bones. Proceedings of the ICAZ Conference, Durham, 2002. Oxford: Oxbow HOARE, R. C., 1812, The ancient history of Wiltshire, Vol. 1. London: Miller KNIGHT, V. K. 2004. Characterising elite midden deposits at Moundville. American Antiquity 69, 304-321 LAWS, K., 1991la, ‘Objects of Kimmeridge shale’, in B. Cunliffe and C. Poole (eds.), Danebury, An Iron Age Hillfort in Hampshire, Vol. 5, The excavations 1979-88: the finds: 368. London: Council for British Archaeology Research Report 73 LAWS, K., 1991b, ‘The chalk’, in N. M. Sharples, Maiden Castle, Excavations and field survey 1985-6, 210-14. London: English Heritage Archaeological Report 19 LAWS, K., BROWN, L. and ROE, EF, 1991, ‘Objects of stone’, in B. Cunliffe and C. Poole (eds.), Danebury, An THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 99 Tron Age Hillfort in Hampshire, Vol. 5, The excavations 1979-88: the finds, 382-404. London: Council for British Archaeology Research Report 73 LAWSON, A. J., 2000, Potterne 1982-5; Animal Husbandry in Later Prehistoric Wiltshire. Salisbury: Wessex Archaeology Archaeological Report 17 LEGGE, A. J., 1981, ‘The Agricultural Economy’, in R. Mercer (ed.), Grimes Graves, Norfolk: Excavations 1971- 72: 79-103. London: HMSO LEGGE, A. J., 1992, Animals, Environment and the Bronze Age Economy. London: British Museum Press LONGLEY, D., 1980, Runnymede Bridge 1976: Excavations on the Site of a Late Bronze Age Settlement. Guildford: Surrey Archaeological Society LYMAN, R. L., 2008, Quantitative Paleozoology. Cambridge: Cambridge University Press MACPHAIL, R. I., 1987, Soil report on the Late Bronze Age Midden at Potterne, near Devizes, Wiltshire. London: Ancient Monuments Laboratory Report 144/87 MACPHAIL, R. I., 1994, ‘The re-working of urban stratigraphy by human and natural processes’, in A. R.Hall and H. K.Kenward (eds.), Urban-rural connections: perspectives from environmental archaeology, 13-43. Oxford: Oxbow Monograph 47 MACPHAIL, R. I., 1995a, Monkton, Kent (MMP94): assessment of soils. Canterbury: Canterbury Archaeological Trust unpublished report MACPHAIL, R.I., 2000, ‘Soils and microstratigraphy: a soil micromorphological and micro-chemical approach’, in A. J. Lawson, Potterne 1982-5: Animal Husbandry in Later Prehistoric Wiltshire, 47-70. Salisbury: Wessex Archaeology Archaeological Report 17 MACPHAIL, R. I. and GOLDBERG, P,, 1995, ‘Recent advances in micromorphological interpretation of soils and sediments from archaeological sites’, in A. J. Barham and R. I. Macphail (eds.), Archaeological sediments and soils: analysis, interpretation and management, 1-24. London: Institute of Archaeology MACPHAIL, R. I., COURTY, M. A., HATHER, J. and WATTEZ, J., 1997, ‘The soil micromorphological evidence of domestic occupation and stabling activities’, in R. Maggi (ed.), Arene Candide: a - Functional and Environmental Assessment of the Holocene Sequence (Excavations Bernabo Brea-Cardini 1940- 50), 53-88. Rome: Memorie dell’Istituto Italiano di Paleontologia Umana MACPHAIL, R. I., CRUISE, G. M., ALLEN, M. J., LINDERHOLM, J. and REYNOLDS, P. 2004. Archaeological soil and pollen analysis of experimental floor deposits; with special reference to Butser Ancient Farm, Hampshire, UK. fournal of Archaeological Science 31, 175-191 MACPHAIL, R. I., CRUISE, G. M., ENGELMARK, R. and LINDERHOLM, J., 2000, ‘Integrating soil micromorphology and rapid chemical survey methods: new developments in reconstructing past rural settlement and landscape organization’, in S. Roskams (ed.), Interpreting Stratigraphy 9, 71-80. York: University of York MACPHAIL, R. I., CRUISE, G. M., GEBHARDT, A. and LINDERHOLM, J., Forthcoming, ‘West Heslerton: soil micromorphology and chemistry of the Roman and Saxon deposits’, in J. Tipper (ed.), West Heslerton Anglo-Saxon Settlement. Yedingham: Landscape Research Centre MATTHEWS, W., 1995, ‘Micromorphological characterisation and interpretation of occupation deposits and microstratigraphic sequences at Abu Salabikh, southern Iraq’, in A. J. Barham and R. I. Macphail (eds.), Archaeological sediments and soils: analysis, interpretation and management, 41-74. London: Institute of Archaeology MATTHEWS, W. and POSTGATE, J. N., 1994, ‘The imprint of living in an early Mesopotamian city’, in R. Luff and R. Rowley-Conwy (eds.), Whither environmental archaeology?, 171-212. Oxford: Oxbow Monograph 38 MCCOBB, L. M.E., BRIGGS, D. E. G., CARRUTHERS, W. J. and EVERSHED, R. P. 2003. Phosphatisation of seeds and roots in a Late Bronze Age deposit at Pot- terne, Wiltshire, UK. Fournal of Archaeological Science 30, 1269-1281 MCKINLEY, J., 2000, ‘Human bone’, in A.J. Lawson, Potterne1982-5: Animal husbandry in Later Prehistoric Wiltshire: 95-100. Salisbury: Wessex Archaeology Ar- chaeological Report 17 MCOMISH, D., 1989, ‘Non-hillfort settlement and its implications’ in M. Bowden, D. Mackay and P. Topping (eds.), From Cornwall to Caithness: some aspects of British Field Archaeology, 99-110. Oxford: British Archaeologi- cal Reports 209 MCOMISH, D. 1996. East Chisenbury: ritual and rubbish at British Bronze Age — Iron Age transition. Antiquity 70, 68-76 MCOMISH, D. 1999. East Chisenbury Midden and Later Prehistory on Salisbury Plain. 3 Stone 36, 12-16 MCOMISH, D., FIELD, D., and BROWN, G., 2002, The Field Archaeology of the Salisbury Plain Training Area. London: English Heritage MORRIS, E., 2000, ‘Fabrics, form series, dating and sum- mary’, in A. J. Lawson, Potterne 1982-5: Animal hus- bandry in Later Prehistoric Wiltshire: 140-176. Salisbury: Wessex Archaeology Archaeological Report 17 MURPHY, C. P, 1986, Thin section preparation of soils and sediments. Berkhampstead: A B Academic Publishers NEEDHAM, S. and SPENCE, T., 1996, Refuse and Disposal at Area 16 East Runnymede: Runnymede Bridge Research Excavations, Volume 2. London: British Museum NEEDHAM, S., BRONK RAMSAY, C., COOMBS, D., CARTWRIGHT, C. and PETTIT, P. 1997. An inde- pendent chronology for British Bronze Age metalwork: the results of the Oxford radiocarbon accelerator pro- gramme. Archaeological Journal 154, 55-107 NEEDHAM, S. and SORENSEN, M. L. S., 1988, ‘Run- nymede refuse tip: a consideration of midden deposits and their formation’, in J. C. Barrett and I. A. Kinnes 100 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE (eds.), The archaeology of context in the Neolithic and Bronze Age: recent trends. Sheffield: Department of Archaeology NEEDHAM, S. and SPENCE, T. 1997. Refuse and the formation of middens. Antiquity 71, 77-90 NEEDHAM, S., 1991, Excavation and Salvage at Runnymede Bridge 1978. The Late Bronze Age Waterfront Site. Lon- don: British Museum Press and English Heritage NEWTON, J. and FRANCE, J., 1983, Farm Animals. Lon- don: Harper Collins O’CONNOR, B., 1980, Cross-Channel Relations in the Later Bronze Age. Oxford: British Archaeological Reports International Series 91 OSWALD, A., DYER, C. and BARBER, M., 2001, The Creation of Monuments: Neolithic enclosures in the British Isles. Swindon: English Heritage PARSONS, M. H., 1972, ‘Spindle whorls from the Teoti- huacan Valley, Mexico’, in M. W. Spence, J. R. Parsons and M. H. Parsons (eds.), Miscellaneous Studies in Mexican Prehistory, 45-81. Ann Arbor: University of Michigan Museum of Anthropology Papers 45 PAYNE, S. 1973. Kill-off patterns in sheep and goats: the mandibles from Asvan Kale. Anatolian Studies 23, 281-303 PEARCE, M. 1998. New research on the terramare of northern Italy. Antiquity 72, 743-746 PIGGOTT, S., 1962, The West Kennett Long Barrow. Lon- don: HMSO PHILLIPS, C. W. 1936. The Excavation of the Giants Hills Long barrow, Skendleby, Lincolnshire. Archaeologia 85, 37-106 POOLE, C., 1984, ‘Objects of baked clay’, in B. Cunliffe Danebury, An Iron Age Hillfort in Hampshire, Vol. 2, The excavations 1969-1978: the finds: 398-407. London: Council for British Archaeology Research Report 52 PRYOR, F 1996. Sheep, stockyards and field systems: Bronze Age livestock populations in the Fenlands of eastern England. Antiquity 70, 313-24 PUGH, R.B.and CRITTALL, E., 1957, The Victoria History of the Counties of England. A History of Wiltshire: Volume I, Part 1. London: Oxford University Press RAYMOND, F, 1994, ‘The Pottery’, in R. Bradley, R. Entwistle and F Raymond, Prehistoric Land Divisions of Salisbury Plain: the work of the Wessex Linear Ditch Project, 69-90. London: English Heritage ROBINSON, M. and STRAKER, V., 1991, ‘Silica skel- etons and Macroscopic Plant Remains from Ash’, in J. Renfrew (ed.), New Light on Early Farming. Edinburgh University Press ROTH, H. 1974. Ein Ledermesser der Atlantischen Bronzezeit aus Mittelfranken. Archaeol. Korresponden- zblatt 4, 37-47 ROULIERE-LAMBERT, M-J. (ed.), 2009, De l’age du Bronze a l’age du Fer en France et en Europe occidentale (X° — VIF’ siécle av. F.-C.). Actes du XXX°* colloque international l AFEAK, co-organisé avec l APRAB RYDER, M. L., 1983, Sheep and Man. London: Duck- worth SCHEUER, L. and BLACK, S., 2000, Developmental Fuvenile Osteology. London: Academic Press SCHIFFER, M. B. 1976. Behavioral Archaeology. New York: Academic Press SCHIFFER, M. B. 1987. Formation processes of the archaeological record. Albuquerque (NM): University of New Mexico Press SEAGER SMITH, R., 2000a, ‘Worked Stone’, in A.J. Lawson. Potterne 1982-5; Animal Husbandry in Later Prehistoric Wiltshire, 213-219. Salisbury: Wessex Ar- chaeology Archaeological Report 17 SEAGER SMITH, R., 2000b, ‘Worked Bone and Antler’, in A. J. Lawson, Potterne 1982-5; Animal Husbandry in Later Prehistoric Wiltshire: 222-240. Salisbury: Wessex Archaeology Archaeological Report 17 SELLWOOD, L., 1984a, ‘Objects of Iron’, in B. Cunliffe Danebury: An Iron Age hillfort in Hampshire Vol. 2. The excavations 1969-1978: the finds, 346-371. York: Council for British Archaeology Research Report 52b SELLWOOD, L., 1984b, ‘Objects of bone and antler’, in B. Cunliffe Danebury: An Iron Age hillfort in Hampshire Vol. 2. The excavations 1969-1978: the finds: 371-396. York: Council for British Archaeology Research Report 52b SERJEANTSON, D., 2007, ‘Intensification of Animal Husbandry in the Late Bronze Age? The Contribution of Sheep and Pigs’, in C. Haselgrove and R. Pope (eds.), The Earlier Iron Age in Britain and the Near Continent, 80-93. Oxford: Oxbow Books SCHIFFER, M. B., 1976, Behavioral archaeology. New York: Academic Press SCHIFFER, M. B., 1987, Formation processes of the ar- chaeological record. Albuquerque: University of New Mexico Press SCHMIDT, P and BURGESS, C., 1981, The Axes of Scotland and Northern England. Prahistorische Bronzefunde 9/7 SMITH, M. J. C., 1982, The worked bone from Iron Age sites in Wiltshire. Unpublished University of Southampton dissertation SPENCE, M. W., PARSONS, J. R. and PARSONS, M. H., (eds), 1972, Miscellaneous Studies in Mexican Prehis- tory. Ann Arbor: University of Michigan Museum of Anthropology Papers 45 SPIELMANN, K. A. 2002. Feasting, Craft Specialization, and the Ritual Mode of Production in Small-Scale Societies, American Anthropologist 104 (1), 195-207 STRAKER, V., 2000, ‘Charred cereals and weed seeds’, in A. J. Lawson, Potterne 1982-5: animal husbandry 1n later prehistoric Wiltshire: 84-95. Salisbury: Wessex Archaeol- ogy Archaeological Report 17 TAYLOR, H. 1931-35. The Tynings Barrow Group-Second Report. Proceedings of the Bristol Speleological Society 4, 67-119 THOMAS, N., 2005, Conderton Camp, Worcestershire; a small middle Iron Age hillfort on Bredon Hill. York: Council for British Archaeology THOMAS, N., GUIDO, M. and HENDERSON, J., 2005, THE LATE BRONZE AGE AND EARLY IRON AGE MIDDEN SITE AT EAST CHISENBURY 101 ‘Glass beads’, in N. Thomas, Conderton Camp, Worces- tershire; a small middle Iron Age hillfort on Bredon Hill, 162-3. York: Council for British Archaeology WAINWRIGHT, G. J. 1970. An Iron Age Promontory Fort at Budbury, Bradford-on-Avon, Wiltshire. WANHM 65, 108-166 WATTEZ, J.. COURTY, M. A. and MACPHAIL, R. I., 1990, ‘Burnt organo-mineral deposits related to animal and human activities in prehistoric caves’, in L. A. Douglas (ed.), Soil Micromorphology: a basic and applied science, 431-440. Amsterdam: Elsevier WHEELER, M., 1943, Maiden Castle, Dorset. London: Society of Antiquaries WILSON, B., 1978, “The animal bones’, in M. Parrington (ed.), The Excavation of an Iron Age Settlement, Bronze Age Ring-ditches and Roman Features at Ashville Trad- ing Estate, Abingdon (Oxfordshire) 1974-76, 110-138. London: Council for British Archaeology Research Report 28 WYLES, S., 2000, ‘Shale and Jet’, in A. J. Lawson, Potterne 1982-5; Animal Husbandry in Later Prehistoric Wiltshire, 208-12. Salisbury: Wessex Archaeology Archaeological Report 17 Wiltshire Archaeological & Natural History Magazine, vol. 103 (2010), pp. 102-29 The Bedwyn Dyke: a revisionist view by Ben Lennon The Bedwyn Dyke is posited as one of a series of enclosing cross valley dykes that express a local polity and define a territory of valuable agricultural land possibly associated with Chisbury hillfort. Cognate features can be found across the local landscape closely correlated to soil and topography. Other features displaying the same general characteristics may be less obvious due to continued or subsequent use as routes or boundary features. The Bedwyn Dyke does not appear to have been conceived as a continuous feature but as a series of demarcating boundaries across open chalk valleys. The gaps between the sections lie on clay soils, which may have been occupied by light woodland or wood pasture. This would have provided a sense of complete visual enclosure within the valley. The area enclosed by the dyke system 1s almost entirely visible from the Iron Age hillfort of Chisbury Camp and this may be significant, although the link 1s not conclusive. It 1s conjectured that the dyke was created as an expression of territoriality or as a statement of prestige rather than any meaningful form of defence. In the absence of any recovered stratified material a period of construction between the late Bronze Age and the middle Iron Age 1s posited as most likely. A similar relationship between the late Iron Age oppidum at Forest Hill, Marlborough, 1s also considered 1n relation to a similar pattern of dykes in the surrounding landscape. other similar features to the east have been rather neglected. Whilst they remain substantial features, with sections that rival parts of the Wansdyke in terms of scale, the dissociation with the Wansdyke has left the Bedwyn Dyke and its associates without an archaeological context. The purpose of this paper is to gain a better understanding of the Bedwyn Dyke by examining its topographical and historical setting, form and structure. Introduction The features collectively known as the Bedwyn Dyke comprise a bank and ditch complex of variable form and size in the area around Bedwyn and Chisbury in north-eastern Wiltshire. Historically, the dyke has been linked with the Wansdyke and it has often been supposed to have formed an eastern extension of that feature beyond the limits of its apparent terminus at New Buildings to the west of Savernake Forest. It has also had strong associations with Chisbury hill fort, to which it appears to be attached. Over the last two hundred years the nature and context of the Wansdyke has been re-assessed on a number of occasions, and it is now commonly held that the Bedwyn Dyke does not form a part of the Wansdyke frontier. Since the development of this hypothesis by Fox and Fox in 1960 the majority of the archaeological work has focussed on the nature of the Wansdyke, whilst the Bedwyn Dyke and General Description of the area Bedwypn lies to the east of Savernake Forest in north- east Wiltshire close to the border with Berkshire. The area lies on the cusp of four major geophysical divisions of the landscape of southern England: the Thames basin, the Marlborough Downs, the The Old Sweet Shop, 79 Primrose Hill, Lydney, Glos. GL15 5SW. benlennon@tiscali.co.uk THE BEDWYN DYKE: A REVISIONIST VIEW 103 ™~ pe / a sS ~ ban’ “S } aN Eastern section : of Wansdyke. Valley gravel & alluvium Chalk Ne Clay with flints Bagshot and Reading Beds Greensand Fig. 2 Soils of the Bedwyn area 104 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE | | : LOW ese | ELS NX Cunetio (y ~AM_588 | : \\ AM 850 gabe eee : Forest Hill > ——s “= pitfort “EAM S89 AM 628. Wansdyke va We ZS 238 = oi 1295 ro i AS © AY Martinsell” = Sag 12246~2 AM 865 _ gee ilPertS 5(. M TIDSECN A MOSS), O13 boa RR contacts Scheduled Ancient Monument x Ee Se cae Vee o~ N 1 se sero6) 2s VY onisbury Woodland Le) i ae 7 ; 5 Say, i Fig. 3 Scheduled Ancient Monuments in the area. Table 1 Scheduled Ancient Monuments in the area around Bedwyn M 26709|Chisbury hillfort SU 278 659 AM 162. |Fosbury hillfort SU 319 564 AM 850. |Forest Hill hillfort SU 208686 AM 864_ |Pottery kilns at Column Ride AM 869_ {Pottery kilns, Tottenham Park M 12218|Undated bowl barrow near Saddle Ride SM 12251 SM12252 SM 12246 THE BEDWYN DYKE: A REVISIONIST VIEW Vale of Pewsey and the Salisbury Plain (Keefer 1997, 1)(Figure 1). The village of Great Bedwyn sits at the bottom of a valley spanning the Bedwyn Brook which rises in the south west and flows to the north east to meet the River Kennet, a tributary of the Thames. To the west lies the clay and flints _ plateau largely occupied by Savernake Forest. To _ the east lie the Berkshire Downs (eastern part of the Marlborough Downs) and to the south the land falls away towards the Vale of Pewsey and the Salisbury Plain. The valley in which Great Bedwyn sits comprises upper chalk and valley gravels, but on the higher ground to the east and west the soils quickly change to heavier clay (Figure 2). Description of the Dyke The area has a long history of human habitation and is rich in archaeology with well known deposits of Paleolithic flints found in the vicinity. Bronze Age and Iron features are also widespread with a number of hillforts across the plateau. The Bedwyn Dyke itself is a Scheduled Ancient Monument (SAM) and there are a number of other nationally important sites from various periods within the area (Table 1 and Figure 3). In addition to this the Wiltshire Historic Environment Record (HER) holds details of a wide range of unscheduled archaeological sites. The Bedwyn Dyke is described in the Wiltshire HER as an undated ditch. The extent and visibility of the scheduled feature varies enormously. At some points it forms a large bank and ditch some 3m from the bottom of the ditch to the top of the bank. In other places it exists merely as a crop mark. Its scheduled extent begins at Chisbury hill fort and runs in a south-easterly direction for 2km across the valley of the Bedwyn Brook before turning sharply to the south west for a further 0.5km. In order to gain a clearer understanding of the structure of the accepted sections of dyke it will be useful to describe and characterise its different components (Figures 4 and 5). Section! The first section stretches for some 600m from the south eastern corner of Chisbury hillfort to the pumping station at Brown’s Lane. Chisbury hill fort sits atop a cap of London Clay. As the dyke leaves the hillfort it passes briefly through a narrow band of the Reading Beds before entering into the chalk valley. The dyke appears to be physically joined to 105 \) Chisbury hillfort \ Section 4 X XS \ \ Section 2 Section 3 _ Section4 © / Section 5 Woodland 200 0 200 400 600 Meters oe 5 Section 6 Fig. 4 Characterisation of the Scheduled extent of the Bedwyn Dyke. the hillfort and the tarmac road shares its route. The bank is represented on the eastern side of the road but has been highly modified by the road and agricultural activity. The ditch, which lies to the east of the bank, has been much filled in through agricultural activity. The bank is variable but can be as high as 3m (Figure 5a). Section 2 The second section is approximately 200m long and runs from the junction of the pumping station, where it leaves the tarmac road, until it is interrupted by the railway line. This short section is one of the most intact parts of the whole dyke although impinged on to a degree by surrounding agricultural activities. The bank rises to a steady 2m above ground level and is largely covered in woody vegetation including some large poplars (Figure 5b). The ditch is clear and intact (although much filled in with organic debris). The overall width of the structure is approximately 10m. For a length of 150m between sections 2 and 3 the assumed course of the Dyke has been heavily disturbed by the construction of the Kennet and Avon Canal and the railway, as well as the natural 106 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE ; a) Bedwyn Dyke north of the Pumping Station in Great Bedwyn (SU 283 653). west Road iY _— YYZ Ys b) Bedwyn Dyke south of the Pumping Station in Great Bedwyn (SU 283 652) yest c) Bedwyn Dyke at Jockey Copse (SU 290 644) WEST — Track 2200tllddddéds Past east Wa d) Bedwyn Dyke at the southern end of Round Copse (SU 289 637). north sauth Fig. 5 Profiles of the Scheduled sections of the Bedwyn Dyke. course of the Bedwyn Brook. This section lies entirely upon the chalk soils. The missing section falls into the valley gravel associated with the floodplain of the Bedwyn Brook. Section 3 The line of the dyke now rises up the southern side of the chalk valley From the railway track in the valley bottom the line of the dyke is now co-existent with the parish boundary between Great Bedwyn and ~ Little Bedwyn until the junction of sections 5 and 6. This is a relatively recent alignment of the parish boundary and until the late 20th century it ran on a roughly parallel course some 300m to the south west along Galley Lane. The whole 400m of section 3 between the sewage works on the Kennet and Avon THE BEDWYN DYKE: A REVISIONIST VIEW 107 Fig. 6 Aerial photograph from 1946 showing section 3 (Burntmill Lock to Fockey Copse). canal and the corner of Jockey Copse has been lost through ploughing. Curiously the sinuous course of _ this section has been fossilised as a ditch and bank _ feature on the Ordnance 1:10 000 series and this exact course can still be seen on aerial photography as a crop mark (Figure 6). Curiously the modern OS 1:10 000 scale map correctly identifies this snaking course whereas the 1880s Ist Ed. OS 1:2500 map merely indicates (incorrectly) the course of the dyke. _ This suggests that the feature was already ploughed out by the late 19th Century.! Section 4 This section runs for approximately 625m just inside the remains of Jockey Copse and Foxbury Wood. This was a continuous tract of woodland until the early part of the 20th century since when it has become fragmented and lost to agriculture. The dyke is shown to have been virtually continuous on the 1880s lst Ed. OS, but the central section has suffered from woodland loss and subsequent farm track construction. Despite the loss of the surrounding woodland the easternmost 150m of this section has remained intact and is covered in woody vegetation Similar to section 2 (Figure 7). The dyke passes across the chalk valley but comes to an abrupt halt as it enters the Reading Beds. Where the dyke is still intact the bank stands around 2m high. The ditch is mostly evident but much in-filled with organic material and the overall structure is approximately 16m wide (Figure 5c). Section 5 Section 5 has completely disappeared and is evidenced only through crop marks. The section continues for 375m across an open field beginning in the narrow band of the Reading Beds before clipping an outcrop of the London clay. For most of this section the crop mark appears as a double line with a maximum width of 13m (Figure 8). The southernmost 100m of this section is shown as still extant on uncultivated ground on the 1880s OS 1:2500 1st Ed. This was still in existence in 1953 when Crawford described it thus - ‘just before it meets the parish boundary it is still unploughed and under grass; it no longer consists of a single bank and ditch but is a broad ditch, twelve feet wide between two banks, with an overall width of nearly seventy feet’ (Crawford 1953,121). The feature is shown on a Savernake estate plan of 1820 (WSA x6/4) on which it can clearly be seen as a hedged field boundary for its whole length between Jockey Copse and Round Copse. 108 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 7 Section 4: The Bedwyn Dyke at Foxbury Wood. Overall width approximately 16m. Section 6 The junction between sections 5 and 6 is a complex one where several physical and political changes occur. From the agricultural land the scheduled line of the dyke now enters the woodland of Round Copse. This is the boundary between the London Clay and the Reading Beds. It is also the area where the parishes of Great Bedwyn and Little Bedwyn join with that of Shalbourne, and a series of banks and ditches veer off in different directions (Figure 9). The line of the scheduled dyke now takes a rather abrupt turn to the south west, through Round Copse and back across the chalk soil. It meanders down through Round Copse for 500 m terminating at the road running from Great Bedwyn to Shalbourne. Throughout this section the earthwork resumes its more typical form of a large bank and ditch. Due to the abrupt turn in the course the earthwork now ‘faces’ south east with the bank now on the north west side. At the northern end of Round Copse the earthwork consists of a shallow bank and ditch approximately 16m wide in totality but no more than 1m deep at any point. As it moves away from this junction it retains its width but the vertical dimensions become increasingly exaggerated until the distance between the bottom of the ditch and the top of the bank is around 2.5m (Figure 5d and 10). At the northern terminus of this section the earthwork is overlain by the parish boundary bank which physically separates the Round Copse section from the ploughed out remains of section 5. This is the line of the old parish boundary between Little and Great Bedwyn and here comprises a broad bank and ditch to rival that of the Bedwyn Dyke. It is approximately 13m wide evident as a broad bank and ditch facing northwards (Figure 11b). At the corner of the wood where this earthwork intersects the Bedwyn Dyke the bank is topped by a large old oak pollard. This is probaby an old parish boundary marker of some antiquity. At this corner the old parish boundary takes a sharp westerly turn where the bank and ditch disappear and the boundary is henceforth marked by a spring which was previously the northern extent of Round Copse.’ Forty metres i THE BEDWYN DYKE: A REVISIONIST VIEW Fig. 8 Section 5: Aerial photograph showing crop mark across London Clay. Bedwyn Dyke ~ Section 4 2 16m wide Section 5 Old Parish boundary marked by spring _ Great Bedwyn Bedwyn Dyke section 6 16m wide Parish boundary Soils 0 Chalk o ( London clay Reading Beds Pd Soils y YA Ploughed out croy A @ Round Copse Bank and 13m wide 50 100 150 200 250 Meters | (eee aw Fig. 9 Map of the junction of the three parishes at Round Copse. 109 110 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE | Fig. 10 The Bedwyn Dyke at the southern end of Round Copse. Overall width approximately 16m. to the south of this corner the parishes of Little and Great Bedwyn join with the parish of Shalbourne. The large earthwork continues south between Great Bedwyn and Shalbourne parishes for another hundred metres or so until it completely disappears into an area of old mineral working. The boundary between Shalbourne and Little Bedwyn strikes out in an easterly direction across the London Clay with a small bank and ditch no more than 7m wide and 1m in height. When considered from any direction the line of the parish boundary here appears to have been aligned in order to go around the northern edge of this section of the Bedwyn Dyke. The natural and obvious line would have cut through the dyke, buta conscious effort appears to have been made to round its northern end. The eroded state of the Bedwyn Dyke at the northern end of Round Copse could be explained by the removal of much material in order to create the later parish boundary earthwork Curiously there is little evidence to suggest the continuation of the dyke to the south west of the Shalbourne- Great Bedwyn road. Despite the chalk soil there are no known recent crop marks relating to the extension of the dyke in this direction. Crawford made a claim for a chalk crop mark across this field in 1921 but omitted it from his 1953 map (Hostetter and Howe 1997, 366). Fox and Fox identified a similar mark (Fox and Fox 1960, fig. 16, pl. VI B) but this is difficult to see on modern aerial photography.’ In conclusion, the scheduled area of the Bedwyn Dyke comprises a number of disjointed sections of earthwork varying in form and character. A few observations on this character can be drawn at this point. e All of the sections of dyke that conform to the typical pattern of a large bank and ditch run on chalk soils and peter out on the clay soils. west and the ditch to the east. e Where a section runs over clay soils and has been ploughed out, the structure can be deduced from old records or crop marks, the construction of the feature does not conform to the above pattern. A number of other bank and ditch systems On these sections the bank lies generally to the _ THE BEDWYN DYKE: A REVISIONIST VIEW Tt a) Boundary between Great Bedwyn and Little Bedwyn parishes (SU 293 629) variable in form and construction. | the Bedwyn Dyke Prior to the pioneering work of Fox and Fox in the | late 1950s the Bedwyn Dyke was widely considered | part of the Wansdyke system. This view seems to _ have been formed at least as early as the 17th century. The link between the Wansdyke and the Bedwyn Dyke may have been provoked to some degree by the | writings of John Aubrey, the 17th century antiquary and travel writer. Aubrey travelled on the Wansdyke and made certain vague observations regarding its | Structure and formation adding that ‘they say it runnes into Savernake Forest? (unpub. cited in Hoare 1819, 17). Although Aubrey did not appear to have spent | much time on the subject it was clear that ‘they’ connect to the scheduled earthwork. These are Historical perceptions of q b) Boundary between Shalbourne and Little Bedwyn parishes (SU 2923 639) c) Bedwyn Dyke at the southern end of Round Copse (SU 289 637) Fig. 11 Profiles of earthworks in Round Copse. already had a perception about the continuation of the Wansdyke beyond its apparent terminus at New Buildings in Savernake Great Park. An early cartographic reference can be found on an early plan of the Manor of Chisbury by Thomas Atkinson dating from 1719 (Figure 12). The plan shows the area around Chisbury village, which lies to the southern extreme of the manorial bounds. The section of Dyke which runs south from Chisbury hill fort is clearly marked ‘Wodens Ditch’ in a number of places. This conception of the continuation of the Wansdyke through Savernake Forest was perpetuated by a string of early antiquaries including Sir Richard Colt Hoare and William Stukeley, as well as the later pioneers in archaeology such as General Pitt-Rivers and O. G. S. Crawford. In 1960, in their thorough examination of the Wansdyke they concluded that the Wansdyke did, in fact, end at New Buildings in Savernake Great Park (Fox and Fox 1960). It was Fox and Fox who coined the term Bedwyn Dyke to A \S \ UOOKKgn . \\ \\ \N A K A wt" QKC y N > aly | WN WG a ¥ ag 5 Pat Aye «nls Senne qos, oe Fig. 12 Thomas Atkinson’s Plan of the Manor of Chisbury, 1719 showing a feature referred to as Wodens Ditch (WSA 1300/372). apply to all sections of dyke that lay to the east of Savernake Forest and had previously been thought of as part of the Wansdyke. In the 1980s Hostetter and Noble Howe reconsidered the status of the sections of what had now become known as the Bedwyn Dyke. At the level of the wider landscape they drew attention to the strategic importance of this area as an east-west trending upland between the Thames basin and the Vale of Pewsey, and potentially a major junction of several communication routes. The several sections are noted to generally block presumed lines of communication through the Vale of Ham and the Bedwyn Valley. They concluded that the whole system did not constitute a convincing military structure and would at best only hinder small scale raiding (Hostetter and Howe 1997, 364). In terms of dating the origin of the system a range of possibilities were entertained from the Iron Age to the post-Roman period with a conclusion that the latter seems most likely, possibly late 4th or 5th Centuries (Hostetter and Howe 1997, 365). In 2006 Andrew Reynolds and Alex Langlands rejuvenated the notion of an extended Wansdyke frontier running from the Bristol channel as far as Inkpen Beacon. Drawing parallels with the apparently deliberate discontinuity of other cognate earthworks such as Offa’s Dyke they implied that the feature represented a political construct conceived on a grand scale. They went on to argue that such a large undertaking could not have been executed during the politically disparate period following the withdrawal of the Roman garrisons, but was much more likely to have been associated with the formation of Wessex (Reynolds and Langlands 2006). Other cognate features in the local landscape It has been argued above that few of the sections historically claimed for the Wansdyke or Bedwyn Dyke actually form part of a cogent system. However, it will be posited below that the accepted sections do form the basis of a system, but that its character and extent has been overlooked in favour of the desire to _THE BEDWYN DYKE: A REVISIONIST VIEW identify an extensive landscape frontier. In order to put the scheduled earthworks into context a number _of other cognate features in the local landscape have _been examined for comparison. Postern Hill, north Savernake _ Forest Hill at the northern end of Savernake Forest is _ thought to have been the site of an Iron Age oppidum prior to the Roman invasion and establishment of ' the Roman town of Cunetio in the early lst century -AD (Corney 1997). There are a number of linear earthworks in the immediate vicinity possibly related to the oppidum site. In the wider landscape, beyond the immediate bounds of the Forest Hill | site lie other linear earthworks that seem to bear 'some similarity to those of the Bedwyn Dyke. Of ' particular note is an earthwork in the northern end of Savernake Forest at Postern Hill (SU 200683). This is a north-south linear earthwork, around 700m in length. From Postern Hill, where it begins on the Clay-with-flints soil it heads north towards to the chalk valley. As it leaves the confines of the woodland and the Clay-with-flints it seems to peter out as shown on the modern OS map. However, it is actually only temporarily interrupted by a lateral trackway of more recent origin. After 50m it re- emerges on the chalk (in perfect alignment) as a _much larger entity running down into the valley to the edge of the modern A4. On this chalk section it has remained concealed by a very wide hedge or narrow woodland. The ditch appears to have been _ eroded through the passage of much traffic and there _ are several striations through the bank where tracks have been cut through for access purposes. These form into a series of braided tracks striking off to _ the east onto the main oppidum area (Figure 13). The relatively undisturbed section on the clay at Postern Hill is approximately 16m across and faces westward from the Forest Hill site with the bank to the east _ (Figure 16a). On the chalk where it has formed into _a hollow way it is as wide as 30m. hill Barn | A second feature in the area to the eastern side | of the Forest Hill site is to be found at Hill Barn, | approximately 500m to the east of the site of the | Roman town of Cunetio (Figures 14 and 16b). This earthwork comprises a large ditch and bank with a north-south orientation currently under rough pasture. The whole feature averages 17m wide with _a bank on the west, thus facing away from Forest | Hill. Again the feature begins on the Clay-with-flints _ and heads directly north down into the Kennet iis) Fig. 13. Hill-shaded LiDAR image of the earthwork feature at Postern Hill, Savernake. The well-known section of dyke runs through Savernake Forest off the Clay-with-flints. North of the woodland edge the line becomes confused with other landscape features. On the chalk the earthwork re-emerges until it 1s interrupted by the expansion of Marlborough and the A4. Peter Crow, Forestry Commission. valley moving quickly onto chalk. After 150m it meets Chopping Knife Lane and appears to stop. However, it can be detected to the north of this point for another 100m towards the Kennet. Here a farm track or old drove road has overlain the ditch and eaten into a part of the bank. The spoil of the bank may have been used to level out the ditch and create a running surface (Figure 16c). The remains of the ditch are clearly discernible to the east and the feature aligns perfectly with the section above Chopping Knife Lane. Noke Wood A similar section of earthwork is also to be found on the edge of Noke Wood where it forms the western edge of the woodland and lies 100m to the south of Golden Arrow Bungalow on the modern A4 (SU 266 678). The clearly visible extent of the feature is approximately 300m running north to south across valley gravel and chalk. It faces east (away from 114 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 14 Aerial photograph of earthwork feature at Hill Barn, Stitchcombe. Overall width approximately 17m. Forest Hill) with the high bank on the west and an overall maximum width of around 15m. At the southern end the feature merges with the London Ride, a long linear trackway laid out during the 18th century in order to connect Tottenham House with the London Road (now the A4) (Figure 15). A number of large beech from the late 18th century now partially sit on top of the earthwork. Heading in a northerly direction the earthwork quickly veers off to the west forming the edge of the woodland and cutting directly across the valley. Once out of the woodland it quickly disappears into an old chalk pit and is not visible in the arable land beyond. To the north of the A4 (at Golden Arrow Bungalow) the feature is not clearly discernible but may extend to the other side of the valley in the form of a field and parish boundary which aligns with the feature. Here this is the boundary between Little Bedwyn and Froxfield parishes. These three features are good examples (but not the only ones) of fragments of earthwork that seem to demonstrate many of the characteristics of the Bedwyn Dyke in terms of morphology, size and orientation. Their exact purpose remains uncertain but it would appear to involve the delineation of rich valley soils. Two further examples of this arrangement provide further possibilities. 115 Fig. 15 Earthwork feature at Noke Wood looking north with the bank on the right. Overall width approximately 13m. Great Lodge Bottom and Cheval Bottom, | ‘Two features in Savernake Forest have been identified _as cross valley dykes at Great Lodge Bottom (SU 211 670) and Cheval Bottom (SU 229 660). The earthwork _at Great Lodge Bottom has long been known and is -a Scheduled Ancient Monument. The feature at Cheval Bottom has only recently been identified _as a distinct earthwork and appears to have been | Incorporated into an extensive woodbank created by the Earl of Hertford in the last decade of the 16th _ century (Ramsay and Bathe 2008, 171). Both features | are morphologically distinct from those previously _ Mentioned in that they comprise a prominent double bank with an intervening ditch and a total width of | approximately 17m (Figure 17). With the exception _ of the double bank arrangement they are comparable in their occupation of the chalk valleys with the _ features previously considered. However, whereas the others appear to be interrupted by agricultural _ activity beyond the confines of the woodland or pasture, these features appear to be complete, and there is no suggestion from the LiDAR images that they were ever intended to extend on to the clay soil (Ramsay and Bathe 2008, 173). A relationship between the cross valley dykes and nearby small enclosures disclosed by LiDAR survey has also been observed (Lennon and Crow 2009). These typically enclose around 0.5 Ha and sit on the nearest available flat site on the clay above the chalk valleys, seemingly overlooking the area delineated by the cross valley dykes (Figure 18). No work has been carried out to attempt to date these enclosures, but they may be regarded as being typical of those that were created from the middle Bronze Age and throughout the Iron Age, and seem to conform to a general model of farmstead; small enclosed settlements of single or extended family size, 30-50m across (Cunliffe 1984,34). LiDAR survey has also disclosed the traces of negative lynchets forming on the side of these valleys suggesting early agricultural activity along the chalk valleys but not extending onto the surrounding clay (Lennon and Crow 2009). Work on prehistoric territories on the North 116 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE b) Earthwork feature at Hill Barn above Chopping Kini 8 < = c) Earthwork feature at Hil Bam below Chopping Kefle Lane (SU 222 637) OGhalk im arable = se i A West qd) Earthwork in Noke Wood (SU 260 677). Valley gravel and chalk in woodland. Fig. 16 Profiles of linear earthworks to the north of Savernake Forest THE BEDWYN DYKE: A REVISIONIST VIEW Fig. 17 Hill-shaded LiDAR images of cross valley dykes at Great Lodge Bottom and Cheval Bottom. Peter Crow, Forestry Commission. York Moors has indicated that from the Bronze Age, societies may have been organised at several levels including that of the household and the community into which is was incorporated (Spratt 1991, 477). This seems to have some resonance with the features of the Savernake landscape. Assuming that these features are contemporaneous, the relationship between the small enclosures and the cross valley dykes is one that is both curious and attractive. The agricultural soils of the valleys appear to be apportioned or divided through the use of the cross valley dykes. The enclosures occupy a prominent position above the valley and are not generally sited on the chalk itself, but on the nearest level site on the surrounding Clay-with-flints. The enclosures overlook the delineated land in a protective way that emulates that of many Iron Age hillfort sites. Given the size of the earthwork features dividing the land here it might be conjectured that they are not merely functional, but are an expression of symbolic territoriality, or prestige. The question that must now be asked is whether this is a model that can be applied to the relationship that can be applied to the Bedwyn Dyke at the community level. Bridging the gap Spratt (1991) has cautioned against inferring the existence of prehistoric systems where the data on settlements is sparse. In particular the following principles should apply. 1. The observation that artificial features are related to the natural features of the landscape does not necessarily imply that they formed a territorial system. 2. The observation that features enclose an area of various ecological zones may not always imply a territorial system. 3. The observation of different sizes of earthworks cannot always be attributed to social factors. 4. The observation that features are related to later boundary systems does not always mean that the former were parts of earlier boundary systems. 118 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE valley gravel chalk 1m clay Fig. 18 The relationship between the cross valley dyke and a small enclosure at Cheval Bottom, Savernake. Spratt goes on to suggest that the solution to the problem of logic is to look for combinations of these attributes. When a number of these occur in combination then the likelihood of a territorial system increases. With this in mind I shall consider the possibilities of a revised Bedwyn system. The relationship between the Bedwyn Dyke and the Bedwyn Valley can be said to mirror very closely that of the dry chalk valleys and cross valley dykes found in the nearby Savernake Forest, but on a much larger scale. It might also be surmised that there is a strong relationship between Chisbury hillfort, the chalk soils and the enclosing dyke. Again, this pattern is mirrored in microcosm with the small enclosures and cross valley dykes found within Savernake Forest. Chisbury itself sits upon an outcrop of London clay in close proximity to the chalk. However it must be observed that the scheduled area of the dyke only crosses two sections of valley but does not appear to enclose the valley on the southern side. This appears to make sense if the Bedwyn Dyke was constructed to deter a perceived threat to the valley from the north. However, this does not account for the abrupt right angle turn through Round Copse. The enveloping nature of the known sections of dyke would suggest that a southern equivalent section may also have existed to complete the enclosure of the valley chalks. One feature which may offer a plausible bridge in the gap is an old hollow way called Hatchet Lane. Hatchet Lane currently takes the form of a sunken lane that extends from Bloxham Copse in the north-west and runs in a south-easterly direction down into the Bedwyn valley (Figure 19). The lane is distinctively sunken with a high bank on the northern side (facing away from Chisbury hillfort). It is notable that along much of its length the bank and ditch lie on the counter-slope with the bank lying down hill of the ditch. It runs through agricultural land and has been heavily ploughed on both sides leading to the generally erosion of © the bank and a narrowing of the lane in many places (Figure 20). However, at the north western end where it enters into woodland it is very wide THE BEDWYN DYKE: A REVISIONIST VIEW and comparable with the other cross valley dykes | previously mentioned. Here a smaller post medieval | bank and ditch appears to have been incorporated | into the bank in order to create a continuous linear | boundary feature. This is evident as the end of the _ original bank can still be seen (Figure 21). Hatchet Lane is limited by the extent of the chalk soils to the west. To the east it extends as far as the Mill House at the southern extent of Great Bedwyn where it is interrupted by the railway and the Kennet and Avon canal. Beyond the canal its progress may have been incorporated into the road to Brail Farm or may have _been ploughed out. Beyond Brail Farm there is no | discernible feature visible either on the ground or | on the LiDAR imagery. Holloways such as this are common features in _ the English landscape. Very often they are formed along Anglo-Saxon estate boundaries (Hoskins | 1977; Rackham 1993). However, in this case the | boundaries of the Anglo-Saxon estates that relate to | | | | | Bedwyn do not pass along this line. Nor is it a lane created purely through the passage of traffic. High banks on the eastern side, although much eroded by ploughing, suggest a deliberate construction. The earliest references to this lane date from the mid 14th century when it is referred to as Hachezate and later la Hachgate probably referring to a ‘hatch gate’ _ into the forest (Gover et al., 1939, 333).* Although a route to the east of the canal cannot be positively identified it is a seductive prospect that this be a fragment of a cross valley dyke that appears to bear 119 Fig. 19 Hill-shaded LiDAR image of Hatchet Lane. Total width at this point 14.5m. Peter Crow, Forestry Commission. the necessary attributes. When placed alongside the known sections of the Bedwyn Dyke the whole system now envisaged represents a quasi-enclosed landscape defined across the chalk by dykes and along the intervening ridges by clay soils and lower fertility sands. There is little evidence to suggest that the clay soils were cultivated to any degree prior to Roman colonisation, although this is not to say that the clays and sands would have been occupied by dense woodland even though the modern pattern is such. As a timber and fuel resource, areas such as these may well have been heavily exploited and formed suitable land for grazing animals, possibly among light woodland cover. With this model of the landscape in mind the Bedwyn ‘system’ now begins to take shape. If the upper margins of the valleys were lightly wooded (utilised wood pasture) and the open land on the chalks (possibly agricultural) were defined by cross valley dykes, this would provide a complete sense of visual enclosure to the valley amounting to an area of 300 hectares. However, there are still two short gaps in system. Both run from the clay soils of the Bedwyn Brails to join up with Round Copse in one direction and Brail Farm in the other (Figure 22). There may be a number of reasons why these sections are not evident on the ground (see below). From Chisbury hillfort virtually the whole of the enclosed valley is visible with the exception of some dips and hollows (Figure 23). To the opening of the valley lying to the 120 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Profiles of Hatchet Lane a) SU 267 640 south north arable holloway arable b) SU 268 639 south north arable holloway arable c) SU 271 638 south north arable holloway arable d) SU 274 637 south north arab le holloway arable . : “ oF Rar c oy 3 ie POISE SS # og & Fs F < L : é Kr 2 s Pe ef Fs ‘ ves : y $ ‘é 3 8, PF PM “es YG Cae < &S sé sf Fis ss é é é PS Ay 3 Fee mS s oa are . Py is “Oo Ce Ke ere PS FE SS b Bi & 2 Bact ss. < . . 5 es Se SF SS Sa s Fig. 20 Profiles of Hatchet Lane. north east much of the valley is also visible and the Similar relationships between hillforts and enclosed area appears to be sited ‘behind’ the hillfort large external enclosures are known from the in this respect. Beyond Hatchet Lane to the south archaeological record, notably at Quarley and — west the visibility quickly deteriorates as it does over Sidbury on the border between Wiltshire and Harandene beyond the Round Copse section. Hampshire. In these cases the linear features ae i 4 i i ‘THE BEDWYN DYKE: A REVISIONIST VIEW defining large areas of the landscape were built prior to the establishment of the nearby hillforts (Cunliffe 1995, 33). In both these cases the enclosures are almost completely continuous but may have been highly visible in the open chalk downland. At Bedwyn the dykes are fragmentary. The reason for ) this may relate to purpose and practicality. _ It is proposed that the scheduled length of the ; Bedwyn Dyke forms part of a system of cross valley dykes and that this system encloses the valley : centred on what is now the village of Great Bedwyn. | The dykes appear to bear a relationship to the hillfort at Chisbury but there may be other relationships | which require further examination. | The system proposed is characterised by the following attributes: ¢ All of the sections of dyke run over open chalk | land and are constructed to be visible features in the landscape ° The features are typically between 15 and 20 m __ wide in totality and consist of a bank and ditch | Fig. 21 Hatchet Lane at the northern terminus showing the large older bank (with bicycle in the foreground) and the smaller bank behind. e The dykes face ‘outward’ from the enclosed valley with banks on the inside and ditches on the outside e The enclosed valley and the dykes are mostly visible from Chisbury hillfort and conversely the hillfort is a visible and conspicuous feature from most points in the valley e On the heavier clays and infertile soil light woodland would have been the dominant feature of early landscapes and these would have provided a sense of visual enclosure from within the valley e The enclosed area is centred on what is now the village of Great Bedwyn The purpose of the Bedwyn Dyke system The physical layout proposed for the dykes at Bedwyn 122 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Cross valley dykes bank e@eoece ditch Se presumed section valley gravel & alluvium chalk clay sandy beds Visible Visibility ae Chisbury hillfort Cc / Not visible Cross valley dykes bank e@eeceece ditch ed eet presumed section Soils 7/7, veavy and infertile soils Soils , ilometers Fig. 23 Landscape visibility from Chisbury hillfort. THE BEDWYN DYKE: A REVISIONIST VIEW _ represents a cohesive and convincing alternative to the extended frontier that has previously been claimed. However, if the proposed system was conceived as a complete and unified entity then several questions remain. Firstly, what was the purpose of such a system; and secondly what may _ have been the period of its initial construction? With reference to purpose several features stand out. The dykes do not appear to have been conceived as a continuous feature. They run across what is presumed to have been an open, possibly agricultural, landscape on good quality soils attractive to early farming communities. The absence of dykes on the heavy clay soils suggests that they were not defensive in nature, a point already _ observed by Hostetter. Nor do they appear to define _alarge and extensive frontier as does the Wansdyke. Their absence on the heavy clay soils, presumably _ some form of light woodland cover, suggests that _ their extension through these areas was not crucial to their primary purpose. It is the visibility of the dykes that appears to form the crux of their purpose. They are visible from the hillfort and the hillfort is _ visible from the majority of the valley. They would _ also have been highly conspicuous features to visitors _ from outside the valley. It has already been noted _ that this valley lies on an important communications _ network between the Thames watershed and the Vale _ of Pewsey. It therefore seems likely that the dykes represent an expression of small scale territoriality as a visual expression of domination in a contested landscape. Such an expression may only have been required where there was a likelihood of incursion from beyond the established community. | Possible period of construction | The period over which such a system of dykes could have been established is very wide. Large linear features were common from the middle Bronze Age through to the early medieval. The terminus ante quem for the dykes is provided by Cynewulf’s charter of Little Bedwyn of AD 778. By this time the dyke is referred to as a feature but does not appear to form any significant boundary to the estate. Reynolds and Langlands (2006) suggest the _ Mercian occupation of Berkshire as a context for the | extended Wansdyke frontier. This occupation was _ Initiated a year after Cynewulf granted this charter to earl Bica. The fact that the Mercian incursion 123 took place in Berkshire rather than Wiltshire lends credence to the effective nature of the Wansdyke across Wiltshire, comparable to an early Maginot line which was merely circumnavigated by the invading forces. Fox and Fox also suggested that the Bedwyn Dyke was constructed as a defensive structure to protect the ‘metropolis’ of Cissa in the 7th century (Fox and Fox 1960, 20). At the time of the Cynewulf charter of AD 778 the boundary of the estate granted to Earl Bica runs virtually through the middle of the modern day Great Bedwyn. This indicates that at that time Little Bedwyn was perhaps the nucleated centre of the estate and that Great Bedwyn did not exist, but grew up later based on the river crossing. It seems implausible that the Bedwyn Dyke would run through the middle of an estate that had just been granted to a favoured noble if it was a contemporary delineating feature. In the Cynewulf charter the dyke is recognised as a boundary feature but no special status is conferred upon it. Furthermore, no settlement site of the Anglo-Saxon pagan period has yet been identified in the Bedwyn area (Eagles 1997, 379). The earliest cemeteries with any number of graves dating from the 5th Century are to be found as far away as Collingbourne Ducis (11km to the south), and Black Patch, Pewsey (13km to the south west). Noble Howe and Hostetter also considered a late Roman or early post Roman date for construction. The circumstantial evidence for this is based on the Roman villa estate centred on Castle Copse, perhaps linking with the hillfort at Chisbury. The Castle Copse villa participated in the market economy of the late 4th century and persisted in a less opulent form into the 5th century following the withdrawal of the Roman administration. Within this context they suggest a possible construction date of the late 4th or 5th century drawing parallels with the last phase of the remodelled Bokerley Dyke and the defensive walls at Cunetio (Hostetter and Howe, 1997, 365). Noble Howe and Hostetter also hint at the possibility of long term continuity in the area linking the tribal settlement of Chisbury with the Romano- British villa at Castle Copse and the later estate of Bedwyn. In terms of circumstantial evidence for continuity in the area there is certainly a case to be made. However, if the extent and purpose of the Bedwyn Dyke conjectured above is accepted then this would create a number of problems for the late Roman/post-Roman date. The villa site at Castle Copse occupies an elevated position perched on the London clay and Reading beds of what is now 124 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE | the woodland of the Bedwyn Brails (SU 283 6290). In the model presented above this would lie on the projected line of the dyke, although it is suggested that the area would be wooded rather than open. It may be assumed that when the villa was at the height of its economic wealth in the 3rd and 4th centuries it would have lain at the heart of the surrounding estate rather than on the fringes. From the villa site the land falls away to the south onto the chalk of Harandene and the Shalbourne Valley, whilst to the north lie the fertile chalk and valley gravels of the Bedwyn Valley. Recovered material from the villa site suggests a rather generalised mixed farming output from the estate (Hostetter and Howe 1997). This is perhaps not surprising given the mixed nature of the surrounding soils and the location of the villa itself on soils that would have been difficult to cultivate by earlier farmers. The case for an Late Bronze Age or Iron Age context Eagles, Hostetter and Howe all considered an Iron Age date as a distinct possibility for the construction of the dyke but on balance appeared to favour a post- Roman date. Given the model proposed above, a late Bronze Age or Iron Age date is more appealing. Although the dating of the hillfort at Chisbury is largely undetermined, the recovery of some material suggests a degree of occupation during the middle Iron Age and possibly earlier (Robinson 1997). As Hostetter and Howe point out, Chisbury ceased to evolve in a way that other hillfort sites may have. Ifit ever truly was a locus of territorial power this seems to have been completely eclipsed by the Forest Hill oppidum by the late Iron Age. The natural resources prevalent across the area are an important limiting factor to changes in land use throughout the prehistoric period. The technology available throughout the Bronze and Iron Age precluded any meaningful agricultural expansion on to the heavy clays for the purposes of cultivation. However, Bronze age axe heads are relatively common spot finds along the edges of the clay plateau indicating that some clearance or woodland exploitation may have been undertaken on the plateau from the Bronze Age onwards. There is little evidence from any of the other sources that there was any extensive level of cultivation during this time. The archaeological record has so far indicated that some areas of difficult soils or harsh environments were sparsely settled. Extensive settlements would have naturally occurred on the lighter soils; the chalks, limestones and river valleys. Throughout the period there were certainly variations in temperature and rainfall across the British Isles which may have affected demographic movement and land practices, and we should not discount the prospect of demographic shifts from the marginal lands of the extreme north and west having a gradual but extensive impact on demographic change. This in turn may have led to localised social tension and increased territoriality. It is generally assumed that as the Iron Age progressed the population increased in line with growing commerce and wealth. It is likely then that pressure began to be exerted on potential agricultural soils that had not currently been exploited on a large scale. This would have possibly forced farmers along the dry chalk valleys of areas like Savernake which, although long and narrow, may have afforded a combination of cultivatable soils and surrounding woodland for materials and pasture. However, the lack of water features in this area would have been a limiting factor and ponds would have been an important resource. Some archaeologists view the Iron Age as one which was characterised by the most intensive impact by prehistoric humans (Simmons 1981,290). On the fertile soils of Wessex there were probably expansions in population with increasing emphasis on mixed agriculture, increasingly nucleated settlements and societies developing into hierarchical tribal structures. This may have had an effect on major earthworks already in existence at this time (Spratt 1991, 476). The principal limitation on the expansion of agriculture was farming technology in relation to the major soil types. The plough used in Iron Age Britain was a simple ‘bow-ard’, the wooden share of which was tipped with a short iron sheath. This tool was ideal for the lighter chalk soils but had immediate limitations. It lacked a mouldboard and couter and was thus was unable to turn a furrow. The later Belgic farmer improved this with longer and wider iron shares enabling them to exploit heavier soils (Frere 1994, 269). Cows, sheep and pigs supplemented the agricultural economy. The crops sown were barley, rye, beans and early forms of wheat including emmer. An important development during this period was the use of spelt that enabled winter sowing to supplement the spring grown crops (Frere 1994, 7). During this period it is unlikely that the simple THE BEDWYN DYKE: A REVISIONIST VIEW - bow-ard would have been able to cope with the heavy Clay-with-flints that cover the majority of _ the Savernake plateau. Nor would the crop types of emmer and spelt have been suitable for these heavy - soils. We can assume then, on technological grounds alone, that little agricultural incursion would have taken place on the clay plateau with the possible exception of the dry chalk valleys such as Great Lodge Bottom and Cheval Bottom in Savernake. The landscape around Danebury and Quarley hillforts in Hampshire is particularly rich in linear features. These appear to have been laid out at various times commencing with extensive ‘spinal linears’, up to 10km in length, and associated subsidiary features which divided the landscape into large territories from the late Bronze Age onwards. These are complemented by ‘local linears’; distinct from the large scale systems but which often overlie earlier field systems and are associated with Iron Age enclosures of various sizes (Spratt 1991, 453). Similar large-scale territorial boundaries from around the same period have been identified in the Yorkshire Vales, Chilterns and Cambridgeshire, as well as the Berkshire Downs (Spratt 1991). The development of small, well defended hill forts in the Middle Iron Age seems to indicate social, economic and perhaps religious centres of substantial | individual territories of approximately 90-120km7. (Robinson 1997, 24). Some of these were new, as was the case with Chisbury. Others were developed from earlier hill top settlements known from the Late Bronze Age onwards. These were generally larger and better defended enclosing an average area of | 6-12ha (Robinson 1997, 24). Chisbury is typical of early hillforts created on new sites and may have been sited because of its position which overlooked an established agricultural hinterland. Cunliffe (1984, 19) describes these as ‘usually constructed as contour works, averaging 5ha in extent. They were defended by a rampart, faced inside and out with | timber, stone, or a combination of the two, fronted _ byaberm and ditch, and were usually provided with _ two entrances facing each other on either side of the fort’. Chisbury encloses approximately 6ha which is in the average range as defined by Cunliffe. It has a bi-vallate form with an additional rampart to the south west. There have been no recorded pottery finds here after the Late Iron Age (possibly around 100BC) and its abandonment is consistent with other Iron Age hillforts in Wiltshire.’ Later Iron Age finds to the west suggests that settlement moved to this area following the abandonment of the hillfort (Robinson 1997, 28). 125 After 800BC Wessex saw dramatic changes in social structure. New hillforts were constructed across the landscape with high levels of defensive characteristics with structures indicating the centralisation of power and communal functions. The creation of these new organisational structures has been seen by some as being a symbolic act of reordering social structures governed by a newly emerged elite (Cunliffe 1984, 31), although the identification of definitive boundaries of tribal groups throughout the Iron Age period is fraught with difficulty. Despite this, progress has been made by plotting distributions of distinctive stylistic motifs in pottery and, for the later Iron Age at least, distributions of tribal coinage issues. While the outer boundaries are invariably difficult to discern, hubs of production and distribution are well defined. The political landscape of the southern Marlborough Downs in the Late Iron Age was evidently complex. From the sources available it would appear to constitute an interstitial territory nestling between some of the main tribal groups known from historical accounts and numismatic evidence. The bulk of knowledge is derived from coins struck in gold, although silver, copper and bronze were first introduced to Britain during this period (Sellwood 1984). In Wiltshire the main tribal coinages associated with this period are attributed to the Durotriges, Dobunni and the Atrebates. Another series of coins found in this area of eastern Wiltshire centred around Pewsey and Marlborough area suggests an independent administration. This was possibly based on the oppidum at Forest Hill. The coinage is distinct but the tribal origin is not recorded (Robinson 1997). Numismatic evidence suggests that the Dobunni were dominant in NW Wiltshire with the Durotriges occupying the majority of southern Wiltshire as far north as Salisbury Plain. Durotrigian evidence is further supported by the widespread use of ceramic style as far as the River Wylye and the southern edge of Salisbury Plain (Corney 2001, 6-8). Evidence of the Atrebates is common on the eastern edge of Wiltshire but rarely found to the west of the River Avon. It has been suggested that the River Kennet may have been the southernmost boundary of a Sub-Dobunni tribal unit (Figure 24). This would make the area around Forest Hill strategically important in the later Iron Age as it overlooks tribal. territory across the Kennet to the north. What is clear is that by the end of the Iron Age at least the Savernake area existed as a confluence 126 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE ~ Nothem ot SS Debunni_— x \ Lf. — | Re ereny e Cost ~ Sab 7 a ~~ ts arte < ~ ~ | Ji Southems | oe, _ Dobunni \\ é CS cS J EZ - Atrebates \~\ ee aa «i ~\ AC i Dutotriges ‘ Durotriges NO = . > {YAY | SS _ a a ee, z NX = > el xe Fig. 24 Main tribal divisions of Central southern England in the late Iron Age (after Sellwood 1984). between several tribal groups. This may have resulted in increased trade and commerce, but also occasionally increased levels of friction. Being at the edge of so many tribal territories it is possible that at some point a small but independent tribal entity emerged for a short period. This may have given rise to the location of the oppidum at Forest Hill. However long it lasted it does not have appeared to have survived the impact of the Roman invasion. Throughout the Iron Age the Thames basin, represented here by the River Kennet, formed a distinct tribal boundary that would re-emerge as the defining northern limit to the kingdom of Wessex in the centuries following the collapse of the Roman administration. It is within this context that we must view the area of the southern Marlbororough Downs. The area around Savernake may have even had its own distinctive cultural identity that did not owe allegiance to either the Atrebates, the Durotriges, the Dobunni or the Belgae. As with all frontiers tensions may have occasionally arisen and it may have been intermittently necessary to assert the limitations of the tribal influence. This may have been increasingly important as populations grew and land pressure created a high demand for light fertile soils. The physical context of the perceived system of dykes is also suggestive of an Iron Age date. Visibility of the enclosed area is excellent with a little less than 90% of the enclosed area visible from the ramparts of Chisbury hillfort (Figure 23). Although the chalk soils extend along the valley in both directions to the north east (the Kennet) and south west (the Vale of Pewsey), the valley sides quickly open up to become very wide. It is essentially the narrowest neck of the valley that is enclosed. Visibility to the south reduces dramatically beyond the dyke. The majority of the dyke itself is also visible from the hillfort. The Hatchett Lane section lies in a slight blind spot which may be in part due to its considerably eroded state. Extensive linear ditch systems were created from the late Bronze Age through to the Iron Age indicating the organisation of the landscape on a large scale. The building of hillforts created foci by which large areas of surrounding landscape were divided as at Woolbury (Robinson 1984, 3). A Pre-Roman date of construction might also account for the predicted missing sections between Brail Farm and Ivy Copse, and between the western end of the Round Copse and Bedwyn Brail. The Roman-British villa at Castle Copse was clearly the centre of a large and wealthy estate. It is likely that the lands immediately around the villa itself would have been intensively cultivated (by contemporary standards). It seems likely, therefore, that any pre- existing earthworks that may have interrupted the highly organised field of the Roman farming system would have been removed. This may have also have had symbolic connotations as well as practical. However, only 350m of predicted dyke are missing from the system as it is conceived above. This amounts to around 8% of the total presumed dyke (see below). In conclusion, a phase of construction from the late Bronze Age to the Middle Iron Age appears to be most likely given the sum of the evidence. However, it is not inconceivable that the enclosing dyke pre-dated the hillfort itself which may have been positioned to overlook the enclosed area and to guard from the north. On the areas of London Clay and Reading sands it is unlikely that much agriculture would have taken place but that these areas would have been wooded to some degree. It should not be assumed that this woodland was dense or impenetrable. These heavy clay soils support a variety of ground flora suitable for a range of grazing ungulates, especially cattle and pigs. The lopping of branches may have offered a supplementary food source. However, light woodland of this kind would still have provided a sense of enclosure to the valley as a whole and it would not have been necessary for the dyke to extend into the woodland areas. The return on labour expended on creating a dyke through these areas is likely to have been low given the heavy nature of the soil and the likely occurrence of features such as tree roots. Furthermore, any approaching aggressors or opportunistic settlers would be more likely to approach along the valley _ THE BEDWYN DYKE: A REVISIONIST VIEW que = Known secions of dyke eeeeee § Dyke suspected but anaccounted for — Provisional territorial boundary Soils valley gravel & alluvium chalk clay sandy beds greensand Fig. 25 A provisional model for the system of dykes around Forest Hill oppidum and the Bedwyn dyke. bottoms rather than over the clay ridges. A general pattern of animal husbandry during the Iron Age is thought to include the herding of sheep on the chalk downlands of Central and Southern England. Cattle also appear to form a significant , though less important, component of domesticated stock on the downs, and these are likely to have been more important on the heavier soils (Grant 1984, 116). Conclusions It has been posited above that the Bedwyn Dyke represents a small scale tribal expression of identity. It is also important to consider why there should still be sections missing from the model posited above. The missing sections lie between the corner of Round Copse and the Shalbourne — Bedwyn road running across to the Bedwyn Brail and from the other side of the Brail down to the Canal Bridge and thus onto Hatchet Lane. It is not anticipated that any dyke would have run over the crest of the hill that is now occupied by Bedwyn Brail, but it would fit the theory for a dyke to run over the chalk from each side and terminate at the change in geology perhaps on the Reading Beds. The first and most obvious option is that the theory posited above is incorrect. It remains provisional and theoretical although it accords closely with field observations and current data. The second option is that the dyke system was never completed. This may have occurred for any number of reasons, a reduction in perceived political hostilities, conquest by a rival polity, or perhaps the intervening gap was completely shrouded in woodland. The third option is that the dyke was completed but has been partially destroyed. If it is accepted that the road from the Hatchet Lane Bridge to Brail Farm occupies the predicted line of the dyke then around 350m of dyke are missing from each side of Bedwyn Brail. Out of a predicted 4500m this amounts to around 8% of the total presumed missing or destroyed. It is noteworthy that the villa site at Castle Copse is situated directly on the predicted line of the dyke. The villa appears 128 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE to have functioned for several hundred years and reached a high degree of sophistication and wealth. It clearly sat at the middle of a large and relatively important estate possibly related in some way to the walled town of Cunetio. Given these facts and in light of the discussion above it is suggested that any dyke structure that penetrated this area would have been obliterated over the centuries of management associated with the villa at Castle Copse. This may have occurred incidentally through a re-ordering of the agricultural field system, or possibly as a deliberate and symbolic removal of a pre-existing socio-political entity. Reflecting this idea back to the later oppidum at Forest Hill we might consider a similar set of relationships between the central site and the cross-valley dykes in the surrounding landscape. The features identified above all share the same characteristics of facing away from the hillfort (with the two short exceptions within Savernake Forest which have two banks of equal height) and are generally confined to the chalk, running out quickly on the clay ridges. If these features are joined in a similar way to those of the Bedwyn Dyke it can be seen that they enclose a large area of land in much the same way (Figure 25). A number of sections may have been lost but the general pattern remains the same be it on a much larger scale. In this case the River Kennet would form the northern boundary with the cross-valley dykes defining territory on the chalk to the south, east and west. It must be assumed that some form of light woodland covered much of the clay soils and the tilled chalk. Those fragments of cross valley dykes that lie further out along the same valley systems (such as that at Noke Wood) may have represented a gradual expansion of territory reflecting growing wealth, power and population. There are a number of cross valley holloways which could have conceivably incorporated large dykes in much the same way as has presumably happened below Chisbury, Hatchet Lane, and below the Hill Barn Dyke. These outstanding issues are beyond the scope of this paper and require more detailed study. Acknowledgements I would like to acknowledge the help and assistance of Graham Bathe as a constant source of stimulation and encouragement in the preparation of this paper. Endnotes 1 This addition during the early 20th Century was almost certainly the work of O. G. S. Crawford who spent many years surveying for the Ordnance Survey and worked on the revisions that were produced in the 1920s. 2 On older maps this spring marks the boundary of Round Copse but in recent decades the area of uncultivated land to the north has scrubbed up to become woodland. 3 One lynchet type feature can be seen occasionally on aerial photographs and on LiDAR survey from the road junction slightly to the south east. However, this does not join appear to join up with the Round Copse section and is shown as a farm track on the 1880s OS Ist Ed. 1:10560. 4 There were several such gates into the forest. Another was probably at Leigh Hill on the southern end of the modern forest on the A346. This is marked by Hatt Gate Cottage (SU 214 643). 5 It has been considered elsewhere that Chisbury may have belonged to a later group of double ditched enclosures that occurred after the abandonment of the larger hillforts and may date to the Roman conquest period or even later (Cunliffe 1984, 34). References CORNEY, M., 1997. The Origins & Development of the “Small Town” of Cunetio, Mildenhall, Wiltshire. | Britannia 28, 337-350 CRAWFORD, O. G. S., 1953. The East End of the Wansdyke. WANHM 55, 119-125 CUNLIFFE, B. and MILES, D. (eds), 1984, Aspects of the Iron Age in Central Southern Britain. Oxford: OUCA Monograph 2 CUNLIFFE, B., 1984, ‘Iron Age Wessex: Continuity and Change’, in B. Cunliffe and D. Miles (eds), Aspects of the Iron Age in Central Southern Britain, pp. 12-45. Oxford: OUCA Monograph 2 CUNLIFFE, B., 1995, Iron Age Britain. London: B.T Batsford Ltd/ English Heritage EAGLES, B., 1997, ‘The Area around Bedwyn in the Anglo-Saxon Period’, in E. Hostetter and T: N. Howe (eds), The Romano-British Villa at Castle Copse, Great Bedwyn. pp. 378-97. Bloomington: Indiana University Press FOX, A. AND FOX, C.,1960. Wansdyke Reconsidered. Archaeological Fournal 115, pp. 1-48 FRERE, S.,1994, Britannia: A History of Roman Britain. London: Pimlico GRANT, A., 1984, ‘Animal Husbandry in Wessex and the Thames Valley’, in B. Cunliffe and D. Miles (eds), Aspects of the Iron Age in Central Southern Britain, pp. 102-119. Oxford: OUCA Monograph 2 THE BEDWYN DYKE: A REVISIONIST VIEW _ GOVER, J. E. B., MAWER, A. AND STENTON, E M., 1939, The Place-names of Wiltshire. English Place-name Society. Cambridge: Cambridge University Press - HOARE, Sir R. C., 1819, The Ancient History of Wiltshire. London: Lackington, Hughes, Harding, Mavor & Jones HOSKINS, W. G., 1977, The Making of the English Landscape. London: Book Club Associates - HOSTETTER, E. AND HOWE, TN. (eds.), 1997, The Romano-British Villa at Castle Copse, Great Bedwyn. Bloomington: Indiana University Press KEEFER, D. K., 1997, ‘The Regional Setting: Environmental Geology and Geography’, in E. Hostetter and T. N. Howe (eds), The Romano- British Villa at Castle Copse, Great Bedwyn, pp. 1-19. Bloomington: Indiana University Press LENNON, B., and CROW, P, 2009. LiDAR and 1ts role in understanding the historic landscape of Savernake Forest. WANHM 102, 245-261 RACKHAM, O., 1993, The History of the Countryside. London: J.M. Dent | RAMSAY, J. and BATHE, G., 2008. The Great Inclosure of Savernake with a note on Cross Valley Dykes. WANHM 101, 158-75 REYNOLDS, A. and LANGLANDS, A., 2006, ‘Social Identities on the Macro Scale: A Maximum View of Wansdyke’, in W. Davies, G. Halsall and A. Reynolds (eds), People and Space in the Middle Ages, 300-1300, pp. 13-44. Turnhout: Brepols 129 ROBINSON, M., 1984, ‘Landscape and environment of the Central southern Britain in the Iron Age’, in B. Cunliffe and D. Miles (eds), Aspects of the Iron Age in Central Southern Britain, 1-11. Oxford: OUCA Monograph 2 ROBINSON, P. 1997, ‘Chisbury Camp and North East Wiltshire in the Iron Age’, in’ E. Hostetter and T. N. Howe (eds), The Romano-British Villa at Castle Copse, Great Bedwyn, pp. 1-11. Bloomington: Indiana University Press SIMMONS, I. G., 1981, ‘Culture and Environment’, in I. G. Simmons and M. J. Tooley (eds), The Environment in British Prehistory, 282-291, London: Blackwell SELLWOOD, L., 1984, ‘Some tribal boundaries viewed from the perspective of numismatic evidence’, in B. W. Cunliffe and D. Miles (eds), Aspects of the Iron Age in Central Southern Britain, pp. 191-204. Oxford University Committee for Archaeology Monograph 2 SPRATT, D. A., 1991. Recent British Research on Prehistoric Territorial Boundaries. Fournal of World Prehistory 5:4, pp. 439-79 Primary Sources WSA x6/4 Map of Savernake Forest Estate. 1820 WSA 1300/372 A catalogue of maps by Charles Price for Lord Bruce Forestry Commission. LiDAR survey of Savernake Forest, 2006. Wiltshire Archaeological & Natural History Magazine, vol. 103 (2010), pp. 130-41 Prehistoric, Romano-British and medieval activity at Ridge Green, Shaw, Swindon by Andrew B. Powell Archaeological works 1n advance of the construction of a new building at Shaw Ridge Primary School, Shaw, Swindon, revealed part of an open Iron Age settlement, a complex of ditches representing up to four phases of a Romano-Bnitish field system, and a range of features of 11th/13th-century date possibly associated with the medieval settlement of Shaghe. Introduction Between August 2006 and July 2008 a programme of archaeological works, comprising evaluation, excavation (Area 1) and watching brief (Areas 2-5), was undertaken at Shaw Ridge Primary School, Shaw, Swindon. The work, commissioned by Swindon Borough Council in support of a planning application for the construction of a new school building, revealed evidence predominantly of Iron Age, Romano-British and medieval dates (Figure 1). The site, measuring c. 0.6ha centred on NGR 411765 184950, occupied the high part ofa generally flat east-facing spur of land, at c. 114m above Ordnance Datum, that slopes gently down to the south and north where tributary streams drain into the River Ray, which in turn flows north into the River Thames. All archaeological features were cut into the underlying Kimmeridge Clay (BGS, Map Sheet 252) and sealed by compact orangey-grey silty-clay subsoil. Although there is a series of Iron Age hillforts, such as Liddington Castle and Barbury Castle, sited along the Ridgeway to the south and east of Swindon, the evidence for settlement in the area of the site, west of Swindon, consists generally of isolated features and a thin scatter of finds of Iron Age pottery. Substantial evidence for Romano- British pottery industry in the area has been revealed during extensive developments west of Swindon since the 1970s. The closest kiln sites lay approximately 140m north-east of the site near Upper Shaw Farm, where there was also evidence of a Romano-British building, and 380m to the south- east near Eastleaze Farm (Youngs et al. 1983; Frere 1984, 323). The considerable spread of the industry is also indicated by the discovery of twelve, mostly double-flue, kilns at Whitehill Farm, 800m to the south-east, which, typological evidence suggests, were in use from the 2nd to the 4th centuries AD, and which have parallels with kiln sites at Savernake (Anderson 1979). , Medieval settlement in the vicinity of the site is represented by a pattern of small ditched enclosures and associated earthworks visible in aerial photographs immediately to the north and extending into Area 2 (Canham 1982). The earthworks were subject to survey and excavation in 1982-3 and produced evidence of late medieval settlement, including house platforms, field boundaries, ridge and furrow and a hollow way (Youngs er al. 1985, 205-6), probably associated with the medieval settlement of Shaghe (Gover et al. 1939, 36). Wessex Archaeology, Portway House, Old Sarum Park, Salisbury SP4 6EB PREHISTORIC, ROMANO-BRITISH AND MEDIEVAL ACTIVITY AT RIDGE GREEN, SHAW 131 Digital Map Dale © (2004) AYS Olctal Map Company 200000 The Site }~« °”"4" erste! ‘ 4 @ Chippenham Site boundary < | Devizes Trowbridge Salisbury® PRN pele Fi Primary School | pests 185000 pon Limits of excavation \ a PA 8 —~eE | oy : ee oo SA 184900 waa Archaeological feature \ a ea \X Ly Disturbance e ie Rey } NK VY Reproduced from ordnance Survey data © Crown Copyright 2008. All rights reserved. Licence Number 100024296 ae Fig. 1 Site location and plan showing all features 132 Results The majority of archaeological features recorded on the site were ditches (Figure 1). These proved to be predominantly Romano-British, although some, particularly in Area 1, were of medieval and post-medieval/modern date. Their phasing, however, both within and between these periods, is problematic. First, the site was subject to flooding following prolonged wet weather during both the excavation and watching brief stages, significantly limiting archaeological visibility and hampering the determination of a number of stratigraphic relationships. Secondly, many ditches, particularly in Area 1, contained varying quantities of both Romano-British and medieval pottery, and while some of the later material may be intrusive, it is also likely that some of the Romano-British pottery was residual in later, but otherwise undated, features. In addition, the ditches were laid out on a variety of orientations, and while some appear to be associated and hence broadly contemporary, many comprise only short lengths, with the result that few combine to form readily identifiable features such as fields, enclosures or trackways. It is possible that some were dug, not as field or other boundaries, but on an ad hoc basis for localised drainage purposes, the ground conditions encountered during the fieldwork suggesting that drainage may have been a recurrent concern. It has not been possible, therefore, to establish a definitive phasing for the site, and the chronology suggested below, which describes possible broad phases of activity rather than exactly contemporary episodes, is only one of a number of possible interpretations of the evidence. Middle-Late Bronze Age Evidence for low-level activity on the site during the Middle—Late Bronze Age is provided by 18 sherds (63g) of flint-tempered pottery which, although undiagnostic, are characteristic of the Deverel- Rimbury and post-Deverel-Rimbury ceramic traditions of the region (Table 1). The distribution of the sherds showed no obvious focus of activity, the only possible feature of this date being a small subcircular pit (1461) (Figure 2) whose upper of two fills produced six of the sherds along with a small amount of oak charcoal; a further five (residual) sherds were found in a later ditch c. 10m south-west of the pit. It is possible that some of the worked flint assemblage from the site (Table 2), which consists Table 1: Pottery totals by ware type Date range Ware type Middle- Flint-tempered wares Late Bronze Age Iron Age Calcareous wares ‘Detrital’ wares Sandy wares sub-total Iron Age Romano- Late Iron Age/early Romano- British British sandy wares Grog-tempered (Savernake- Saxon type) wares Coarse oxidised wares Coarse greywares Black Burnished ware Oxon colour coated ware Samian sub-total Romano-British Organic-tempered ware Medieval Minety-type wares Laverstock-type coarseware Modern Kennet Valley wares West Wiltshire type ware Other medieval flint-tempered wares Miscellaneous medieval sandy wares sub-total medieval Yellow ware Total Table 2: Finds totals by material type Material type Pottery Middle-Late Bronze Age Tron Age Romano-British Saxon Medieval Modern Ceramic building material Fired clay Flint Burnt flint Stone Burnt stone Glass Copper alloy Iron Slag Animal bone Clay pipe No. THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE No. Weight sherds (g) 18 63 19 70 41 241 117 584 177 —- 895 12 78 186 3229 39° 1150 533. 2545 3 24 2 3 1 ] 776 6030 5 9 131 1493 2 seh 440 3636 18 60 5 146 18 234 614 5613 1 3 1591 12613 Weight (g) 12613 63 895 6030 9 $613 5 763 246 67 66 3132 1956 62 6 195 149 666 2 PREHISTORIC, ROMANO-BRITISH AND MEDIEVAL ACTIVITY AT RIDGE GREEN, SHAW 133 Structure Key: game (\Vliddie-Late Bronze age lron Age ~ Other period feature Disturbance 185000 Roundhouse Fig. 2 Prehistoric features - entirely of chronologically undiagnostic waste flakes _ (one of them burnt), dates to this phase, although they could equally be of Iron Age date; the same applies to the small quantity of burnt flint. _ Iron Age A moderate assemblage of Iron Age pottery (177 Sherds weighing 895g), much of it badly abraded and residual in later features, was recovered from across the site (Table 1). It appears to have derived from an Iron Age settlement, part of which — comprising a roundhouse, another possible structure and a number of pits and postholes — was revealed in Areas 5 and 2 (Figure 2). The pottery is in three broad fabric groups — calcareous (mainly shelly), wares with sparse quartz in a fine silty matrix (sandy ware), and ‘detrital’ fabrics which contain randomly sorted inclusions of coarse quartz, patinated gravel, and coarse iron oxides. The wares show some similarities in the clay matrix, but with varying amounts and varieties of inclusion types. Although the only diagnostic form is a rim sherd from a small, gently convex vessel, probably of Middle Iron Age date, the three fabric 134 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE groups have been broadly dated as Iron Age. The roundhouse, on the south-eastern edge of the site, comprised a circular gully (1801), approximately 9.5m in internal diameter, with a 2m wide entrance gap on the east side (Figure 2). A break on the opposite side was at least 1m wide, but its southern terminal was obscured by a later ditch. Part of an earlier cut (1800) on the north side suggests that the building may have been repaired. The gully, which averaged 0.4m wide and was up to 0.2m deep, produced 37 sherds of Iron Age pottery. Four postholes were recorded inside the roundhouse (1674, 1729, 1745 and 1786), concentrated towards the north-west. Although only one (1674) contained any dating evidence (a single Iron Age sherd) and they formed no obvious arrangement, it is possible that they all had some structural function associated with the building. A number of pits were also recorded inside the roundhouse. Two (1436 and 1465), both undated, were in closely corresponding positions on the north and south sides, although their small size (0.7m and 0.8m in diameter, and 0.2m and 0.1m deep, respectively) and absence of finds makes them hard to interpret. Two further pits (1418 and 1438) both contained single sherds of Romano-British pottery; this is quite possibly intrusive, deriving from a Romano-British ditch, recut on more than one occasion, which crossed the eastern side of the roundhouse. Environmental samples from gully and pit contexts within the roundhouse contained very few charred plant remains; these include a glume base of spelt wheat (Znticum spelta) and a few grains of barley (Hordeum vulgare); a barley grain from pit 1436 was twisted and so probably of 6-row barley; weed seeds comprised occasional seeds of dock (Rumex sp.) and vetch (Vicia sp.). Some 40m north-west of the roundhouse was a second cluster of small features, associated with a length of curved gully (1518) containing 15 sherds (92g) of Iron Age pottery, ten of them of Late Iron Age date (and one residual Middle Bronze Age sherd) (Figure 2). The gully, which comprised two segments, c. 0.4m wide and 0.1m deep, separated by a narrow gap, lies on the north-eastern arc ofa circle with a projected diameter of c. 15m. It is probably, therefore, not a roundhouse gully, although the proximity of a number of postholes may indicate some other form of structure. Of the small features close to the gully that were excavated, only four (1337, 1390, 1500 and 1616) contained any dating evidence (in each case Iron Age pottery), although feature 1618 is clearly contemporary with features 1616. Posthole 1390 was one of five small pits or postholes immediately inside the gully, the rest of which were undated. There was also a small hearth (1333) located just inside the gap between the two gully segments, on the base of which were nine pieces of burnt stone (1436g) within a charcoal-rich fill. In contrast, posthole 1337 lay just outside the gully, and appears to be the southernmost of a line of postholes immediately inside, and parallel to, the ditch of a later, Romano-British enclosure (1514, below). While it seems unlikely that the eight sherds (46g) of Iron Age pottery from posthole 1337 were residual, this remains a possibility given the proximity of the Iron Age gully; the posthole also contained fragments of animal bone and a single struck flint. Pit 1500, the base and sides of which showed evidence of im situ burning, may also have been a small hearth. A sample from it produced a few hazelnut (Corylus avellana) shell fragments, a possible glume base of emmer wheat and a seed of spikerush (Eleocharis palustris). Aligned on it, to its north-west, was a pair of small elongated features (1616 and 1618), possibly postholes, both of which also displayed in situ burning, and had burnt fills containing charcoal and fragments of burnt flint. It is possible, therefore, that these three features (which also lie inside the circle projected from the line of gully 1518) may have formed part of a structure that burnt down. A quantity of Iron Age pottery was recovered from features further north-west, in Area 2. Four sherds (66g) came from a shallow irregular feature (1236), possibly a natural tree hollow, and six residual sherds (30g) from adjacent sections of a probable medieval ditch (1804, below) that cut across it (Figure 4). Eight sherds came from a small truncated pit (1213) and another one from a length of otherwise undated ditch (1224). Possibly also of this period was an oval pit (1611), 15m north-west of the roundhouse. It measured 0.9m by 1.2m and was 0.3m deep, and appeared to have had a post set within it, surrounded by flint packing. Within the packing were four pieces of fired clay, one of which appears to be part of a loomweight, probably of later prehistoric date, although its form is uncertain. Romano-British Activity in the Romano-British period is represented primarily by a complex of ditches covering the whole site. While a number of pits were also recorded, —~_ = PREHISTORIC, ROMANO-BRITISH AND MEDIEVAL ACTIVITY AT RIDGE GREEN, SHAW 135 some close to the ditches, these were generally thinly dispersed and suggest no clear focus of settlement activity, although the recovery of 776 sherds (6030g) of Romano-British pottery from the site indicates that there was probably a settlement in the vicinity. Also recovered was a small quantity of ceramic building material, all of it, including one imbrex roof tile, of Romano-British date. The pottery assemblage consists almost entirely of coarsewares, most of which are likely to have been locally produced — grog-tempered wares (‘Savernake’ types) in north Wiltshire, and the greywares and oxidised wares in north Wiltshire or Oxfordshire (Table 1). Three sherds of Black Burnished ware (BB1) from south Dorset were also recognised. The generally abraded nature of the assemblage means that any finer, colour-coated wares present remained largely unrecognised. Some of the sandy wares may be local products from kilns in the Swindon area, while others, particularly the oxidised wares, may originate from the Oxfordshire production centre. Diagnostic sherds, however, were scarce, although the presence of bead rim and everted rim jars point to an early Romano-British date — some sandy sherds and some of the coarser grog-tempered wares could even be pre-conquest, although this cannot be definitively proved. The low level of chronologically diagnostic pottery (and the site conditions previously described) has made it hard to phase the Romano-British features, although it is evident from the varied orientations and intersecting nature of the ditches that the site saw a number of phases of activity, with some large-scale changes being made to their layout over time (Figure 3). While the recorded stratigraphical relationships between ditches provide some supporting evidence for the suggested phasing that follows, it is only one of a number of possible interpretations of the site. Phase I The earliest phase of activity appears to comprise a largely irregular array of ditches forming small enclosures and other features possibly constructed _ for controlling the movement of livestock. At the west, in Area 5, a curving ditch (1515) forms the north side of a possible enclosure, bounded to the south-west and south-east by ditches 1519 and 1520, _ respectively. To its north was a possible west—east aligned droveway defined on its northern side by ditch 1513, the latter turning northwards to its west and creating a funnel shape. A number of other short lengths of ditch in Area 5 (1371, 1488, 1799 and 1806) and Area 3a (1258) may represent a continuation of this enclosure and droveway system towards the east. To the north-east, in Area 1, two roughly parallel ditches (1147 and 1152), both aligned approximately north-south, splayed outwards from each other at their northern ends, this arrangement again suggesting a livestock control function, although this is far from certain. The ditches were irregular in form, possibly through repeated recutting; ditch 1147 narrowed from 2.5m wide and 0.3m deep at the south to 0.6m wide and 0.1m deep at the north, while ditch 1152 widened from 0.7m wide at the south to over 2m at the north. Although both contained significant amounts of medieval pottery — 28 sherds (311g) in ditch 1147, and 65 sherds (453g) in ditch 1152 -—in both cases this is considered to be intrusive. For example, all the medieval pottery in ditch 1152 came from the single fill in one excavated section towards its northern end, but the two adjacent sections to the south produced between them 93 sherds (1282g) of Romano-British pottery (from a grog-tempered bead rim storage jar, and a necked jar and a cordoned jar both in coarse greyware), and no medieval pottery. Phase 2 The northern ends of ditches 1147 and 1152 were subsequently cut across by a substantial east—west aligned ditch (1148), up to 3m wide and 0.9m deep, whose slightly curved line ended at the west at a rounded terminal 4m from a north-south aligned ditch (1143/1531). The gap between these two ditches may represent an access point between adjacent fields. East-west ditch (1511) at the north end of Area 5, appears to be aligned on ditch 1148. Ditch 1143/1531 was cut at the north by a medieval ditch, and although not visible further north may have been obscured by later features. To the south, it ran south for over 70m through Areas 3 and 5, cutting across the Iron Age roundhouse and the line of the possible Phase 1 droveway. There was evidence that it had been recut in places and it had a variable profile, being 1.1-2m wide and up to 0.6m deep. It is quite possible that some of the Phase 1 features in Area 5 remained in use during this phase. Phase 3 This phase is characterised by a much more regular system of rectilinear field boundaries with axes aligned north-east to south-west and north-west to south-east. The field system cuts across and slights 136 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Key: @=a——8 Romano-British phase | ss Romano-British phase || Other period feature Disturbance 411700 Key: gum ROomano-British phase III * Romano-British phase IV Other period feature Disturbance i 185000 a He 184900 ° S foe) - é St 185000 SQ asses SQuaqaq ns HOOK emt oe WV g-1528 50m Fig. 3 Romano-British features by phase PREHISTORIC, ROMANO-BRITISH AND MEDIEVAL ACTIVITY AT RIDGE GREEN, SHAW 137 - many of the earlier features, indicating a wholesale reorganisation of the agricultural landscape In the centre of the site, three ditches (1514, 1524 - and 1523) form the western, southern and eastern sides of a rectangular field measuring approximately - 25m by 25m, whose northern side is formed by a long curving ditch recorded in Areas 2 (1807), 5 (1510) - and 3 (1810). This curving ditch cuts across Phase 2 | ditch 1143/1531 and possible Phase 2 ditch 1511. The line of ditch 1524 was subsequently extended to the south-east by ditch 1532, where _ it defined the southern side of an adjacent field. A _ number of lengths of ditch in Areas 1 (1604) and _ 3b (1242 and 1244) may represent a continuation _ of this rectilinear field system to the north-east, _ while in the south of Area 5, ditches 1528, 1533 and 1534 represent elements of additional fields or enclosures, with gaps at their corners providing access between them. A charcoal-filled pit (1298), c. 0.5m in diameter and 0.2m deep, located just outside the gap between ditches 1528, 1533 and 1534, may _ be associated with them, although this can not be proven and its function is unclear. To the west of ditch 1514, an L-shaped ditch (1516), with arms 1.3m and 7m long, appears to form a small ‘annexe’ to the central enclosure. Although it produced a single sherd (2g) of medieval pottery, _ its position and alignment in relation to the Phase | 3 ditches and the absence of any other medieval features in the area suggest that a Romano-British date is more likely. As mentioned above, this feature overlies the curved Iron Age gully (1518), and it bounds same group postholes and small pits that lay within the arc of the gully (Figure 2). The position of this features may be related to the 8m long line of six possible postholes that run along the inside of ditch 1514 (the southernmost of which, 1337, may be Iron Age, above). Phase 4 A small group of apparently associated ditches in Area 5, with north-south and east-west orientations, are not easily associated with any of the above phases. Although they appear to cut across the Phase 3 field system, they have very limited extent and therefore cannot be viewed as a reorganisation of the site on the scale of that seen in Phase 3. Ditches 1522/1526 _and 1527 are arranged as a right-angled corner, | with a short ditch (1763) running west from ditch 1522/1526. To the north, the line of ditch 1522/1526 was continued, after ac. 5m wide gap, by ditch 1521. | Ashallow pit (1446) on the south side of the gap may | be associated, although this cannot be proven. To the south, two slightly diverging ditches (1529 and 1535) ran south from the corner. Other features Two possibly associated oval pits were recorded within Area 3a on the north side of Phase 3 ditch 1810, one either side of Phase 2 ditch 1143. To the east, pit 1268 measured c. 1m by 1.6m and 0.4m deep. It contained four fills, the sterile primary fill, which showed evidence of waterlogging, overlain by two dumped layers and a silting layer. Of the dumped layers, the upper was richer in charcoal but both contained pottery (111 sherds, 688g) probably deriving from the same vessels. To the west, oval 1266 measured 0.7m by 1m and 0.3m deep and had a single fill containing four sherds. Saxon and medieval A low level of Early/Middle Saxon activity is indicated by five undiagnostic body sherds in coarsely organic-tempered fabrics, all of which were intrusive in Romano-British contexts in Area 5. Evidence for medieval activity was concentrated in Area 1, the features from this area producing 96%, by weight (565 sherds, 5400g), of all the medieval pottery from the site (Table 1). Despite the aerial photographic evidence for medieval earthworks to the north of the site and extending into Area 2, no medieval features were recorded in Area 2. As in the Romano-British period, most of the medieval features were ditches, although a number of other feature types, some of uncertain function, were also recorded (Figure 4). In this period, again, the determination of a secure phasing was hampered by poor ground conditions and the chronologically unspecific nature of much of the medieval pottery assemblage; consequently only a single medieval phase is defined. The pottery sherds were all small and heavily abraded, hampering identification. They consisted almost entirely of coarsewares, in flint-tempered, calcareous and flint-tempered/calcareous fabrics, many of which fall within the ceramic tradition of ‘Kennet Valley wares’ which extended across west Berkshire and northeast Wiltshire from at least the 11th to 13th centuries; diagnostic sherds derive exclusively from jars, with thickened and/or clubbed rims. Some of the calcareous wares may originate from sources within north Wiltshire, such as Lyneham, and there are certainly oolitic wares of Minety type, including some glazed wares. The latter include jars with everted, possibly wheel-finished rims, and tripod pitchers, suggesting a date range 138 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE iil 185000 12241 RR \S SEAAMBUEETS COUNTESS — er Key: ma (Viedieval seams’ Post-medieval/modern Other period feature Disturbance Fig. 4 Medieval and post-medieval/modern features from 12th to 13th centuries for these wares. There were 18 sherds of micaceous West Wiltshire type ware (possibly deriving from the Warminster area). Other wares are sparsely represented — one sherd of Laverstock-type coarseware from the Salisbury area, and a handful of miscellaneous flint-tempered and sandy wares from other (unknown) sources. Overall the date range is 11th to 13th centuries, although the presence of the Minety-type jar forms may compress this into the 12th/13th century. One of the main features of probable medieval date was a ditch running approximately ENE-WSW through Areas 1 (1155), 2 (1804) and 3. The only excavated sections were in Area 2 and although the only pottery from it was six Iron Age sherds, these are almost certainly residual, possibly deriving from a tree hollow (1236, above), cut by the ditch at this point, that also contained Iron Age pottery. The ditch, which was up to 2.3m wide in Area 1 was considerably smaller in Area 2 (up to 0.9m wide and 0.2m deep), where it ended at a rounded terminal. Despite the absence of medieval pottery from ditch 1155/1804, a medieval date is strongly suggested by its relationship to other proven medieval features, in particular ditches 1146 and 1149 which run southwards from it at right angles. The more regular of these, at the east, was ditch 1149, which measured 2.4m wide and 0.3m deep. Ditch 1146, to the west, had a more sinuous line; towards the south it was over 2m wide and 0.55m deep, but at the north, where it appeared to splay out into a number of narrow branches, it was less than 0.5m wide and 0.2m deep. Neither of these medieval ditches continued north of ditch 1155/1804 which appears to represent a significant boundary, and one that continued, slightly altered, into the post- medieval and modern periods (below). The line of ditch 1149 was crossed by that of a narrow sinuous ditch (1150) with which its stratigraphic relationship was not established. The — latter’s dimensions — up to 0.6m wide and 0.1m deep —and its irregular line indicate that it had a different function to the larger ditches. A large quantity of medieval pottery was recovered from its western end where it intersected with feature 1153 (below). Further west, ditch 1144 may have continued the - PREHISTORIC, ROMANO-BRITISH AND MEDIEVAL ACTIVITY AT RIDGE GREEN, SHAW 139 - line of ditch 1150, although at 1.3m wide and 0.6m deep, it was significantly different in form. A number of other less easily identifiable _ features, as well as a single pit, were recorded among the medieval ditches. Pit 1070, which was c. 1.7m - in diameter and 0.6m deep with moderately steep _ sides sloping to a narrow base, produced sherds of - prehistoric, Romano-British and medieval pottery from its lower fill, but no other finds apart from a fragment of fired clay. A possibly circular feature - (1016), cut by ditch 1144, was 1.5m wide but only 0.1m deep. A short length of ditch (1123) aligned north-south was truncated c. 2m south of its terminal by modern ditch 1154. Another linear feature (1130), 3.5m long, up to 0.5m wide and 0.35m deep, ran west _ from close to pit 1070, its single fill containing 66 sherds (654g) of medieval pottery, as well as small quantities of animal bone, slag and two iron nails. To its south-east and parallel to it, linear feature 1151 petered out at the east but was 1.2m wide and 0.2m deep at the west where it abutted a north-south aligned feature (1153). Feature 1153 was 10m long, up to 2.3m wide and _ 0.7m deep at the centre, but it narrowed at either _ end where it was only 0.3m deep. Its single fill _ contained a large quantity of pottery and charcoal. _ Anenvironmental sample was dominated by grains of free-threshing wheat (Triticum aestivum sl), with only a few of barley. Weed seeds included those of stinking mayweed (Anthemis cotula), a species associated with the cultivation of heavy, clay soils. Such finds are generally common from medieval sites in the Thames Valley from the later Saxon period onwards (Robinson and Wilson 1987). The _ sample also contained a probable seed of broad bean (Vicia faba) and a possible seed of lentil (Lens culinaris), the former a common crop and the latter a possible crop in the region during this period (cf. Greig 1991). Several fruits and probable thorns of | hawthorn (Crataegus monogyna) may indicate the use | and burning of hedging or scrub material, perhaps | relating to clearance activities. The general location of these features appears | to represent a focus of activity, since features 1130 and 1153 and the western terminal of ditch 1150 produced approximately 42% (by weight) of all the medieval pottery from the site. A number of other features, some undated, are of probable medieval date based on their apparent stratigraphical relationship to dated features; others can not be dated. | Post-medieval and modern The Ist edition Ordnance Survey map of 1889 (and later editions) shows two tree-lined field boundaries crossing the site. One, running approximately east-west, corresponds to a ditch recorded in Areas 1 (1154), 3 and 2 (1221) (Figure 4). As surveyed in Area 1, ditch 1154 was at least 4.5m wide, but in Area 2 was no more than 1.8m wide and 0.3m deep. However, a parallel ditch (1205) to the south in Area 2 may be part of the same boundary, giving it a similar width to that in Area 1. This boundary may have been a replacement for the medieval boundary represented by ditch 1155/1804 (above). The other mapped field boundary, which runs south-south-east from ditch 1154, corresponds to ditch 1145. This was still visible at the start of the excavation as a linear hollow with a hedge along its edge and appeared to be more recent in origin than ditch 1154, possibly a product of parliamentary inclosure of the late 18th/early 19th century. Running perpendicular from it to the west-south-west, and possibly associated with it although not shown of the OS map, was an unexcavated ditch (1156), both these ditches cutting medieval ditch 1155. Discussion The initial excavation, of Area 1, revealed evidence only of Romano-British and medieval activity, so the uncovering during the subsequent watching brief in Areas 2—5 of part of an Iron Age settlement and (albeit limited) ceramic evidence of Middle— Late Bronze Age activity, has added a significant prehistoric component to the archaeology of the site and the wider area. The Iron Age settlement appears to have been open, as no ditches could be firmly dated to this period, and others, such as ditch 1224 in Area 2 contained only single, possibly residual, Iron Age sherds. The roundhouse was the only clearly recognisable structure, but the low level of finds and environmental evidence from this and other Iron Age features means that little can be said about specific activities on the site during this period. Most of the pottery does not allow more precise dating of the settlement than a general Iron Age date; one sherd recovered is possibly characteristic of the Middle Iron Age. In contrast, ten Late Iron Age sherds from the possible curved gully structure may indicate a relatively late date. The suggestion that some of the coarser grog-tempered wares from Romano-British features could potentially be of pre-conquest date raises the possibility of a degree of continuity of occupation from the Iron Age into the Romano-British period, although this was far from 140 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE certain. Nonetheless, the presence of bead rim and everted rim jars within the Romano-British pottery assemblage point toa Ist to 2nd century AD date for the start of the Romano-British activity on the site. The repeated remodelling of the Romano-British field and/or enclosures, evident in their varied orientations and frequent intersections, suggests that the site was used over an extended, although not necessarily continuous, period, even if the dating evidence cannot specify its duration. The Romano-British pottery from the site was generally abraded and it is possible that most of the finds derived from domestic waste material incorporated within manure and spread over fields, with the settlement possibly at some distance from the site. Apart from scattered pits of uncertain function, there were no features which could be shown to be directly associated with settlement activity. The only possible candidate was the small rectangular ‘annexe’ (1516) on the west side of the central, Phase 3, enclosure. If this was the site of a structure, and the presence of a number of undated postholes within it allows for that possibility, it may be significant that it directly overlay the Iron Age curved gully structure (to which the postholes might equally have belonged). There is little to indicate a specific function for the Romano-British field system, although the outward splaying ends of some of the parallel ditches in Phase 1 may hint at funnels leading into droveways and therefore a farming regime incorporating animal husbandry. This evidence, however, is ambiguous, and the features assigned to Phase 1 do not form any coherent pattern. Certainly the more regular layout of the Phase 3 field system points to a more organised, and possibly more centralised agricultural regime. Unfortunately, as with the Iron Age, the environmental samples from Romano-British features contained few cereal remains, these comprising hulled wheats of emmer and/or spelt (Ziticum dicoccum/spelta) and barley, and a few weed seeds, including spikerush, remains of which are common on Thames Valley sites and can be attributed to the cultivation of wetter, perhaps flooded, marginal soils (Jones 1988). The range of plant remains from both the Iron Age and Romano-British features are generally consistent with those from other sites within the general area, such as Groundwell West (Stevens and Wilkinson 2001), Salisbury Plain (Stevens 2006) and Swindon Gateway (Wessex Archaeology 2007). Emmer and spelt were the predominant cereals in the Iron Age and Romano-British period in north Wiltshire (Stevens 2006) and the Upper Thames Valley (Robinson and Wilson 1987). The low numbers of plant remains may reflect the poor preservation conditions, with relatively shallow features penetrated by roots from overlying thick shrub. Other contemporary sites, such as Thornhill Farm and Roughgrounds Farm (Robinson 2004; Allen et al. 1993), also produced few plant remains. The uneven distribution of medieval pottery from the site, with almost all of it coming from Area 1, may reflect the proximity of the earthworks associated with the medieval settlement of Shaghe to the immediate north of the site. No medieval features, however, were recorded north of ditch 1804 in Area 2, where two north-south features are visible in aerial photographs. No clear settlement structures were identified in Area 1, although a number of non-ditch features contained substantial quantities of medieval finds, and could indicate a settlement focus. In contrast with the earlier periods, some of the environmental samples from this period were relatively rich in plant remains, being dominated by free-threshing wheat but containing also barley, broad bean and possibly lentil, with weed seeds such as stinking mayweed pointing to the cultivation of heavy, clay soils. Acknowledgements Wessex Archaeology is grateful to Swindon Borough Council for commissioning the excavation, and in particular to Mike Blissett, Project Planning and Design Officer, of Swindon Borough Council (Technical Services). The fieldwork was monitored by Roy Canham, and Melanie Pomeroy-Kellinger, County Archaeologists to Wiltshire County Council who are also thanked for their help and advice. The project was managed for Wessex Archaeology by Brigitte Buss and Caroline Budd, and the fieldwork directed by Phil Harding, Mike Dinwiddy, Kevin Ritchie and Vasilis Tsamis, who also undertook the post-excavation analysis with contributions by Caroline Budd and Andrew Powell. The pottery analysis was undertaken by Lorraine Mepham, and the environmental analysis by Chris Stevens. The illustrations are by Rob Goller. The archive and finds, currently stored at the offices of Wessex Archaeology, Old Sarum, Salisbury under the project codes 63150-63153, will be deposited in due course with Swindon Museum and Art Gallery under accession number B2007.11 PREHISTORIC, ROMANO-BRITISH AND MEDIEVAL ACTIVITY AT RIDGE GREEN, SHAW 141 References ALLEN, T.G., DARVILL, T.C., GREEN, L.S. and JONES, M.U., 1993, Excavations at Roughgrounds Farm, Lechlade, Gloucestershire: A Prehistoric and Romano-British Landscape, Oxford: Oxford Archaeological Unit/ Oxford University Committee for Archaeology - ANDERSON A.S., 1979, The Roman Pottery Industry in North Wiltshire, Swindon: Swindon Archaeological Society Report 2 CANHAM, R. 1982. Aerial photography in Wiltshire 1975-81. WANHM 76, 3-19 FRERE, S.S. 1984. Roman Britain in 1983. Britannia 15, 266-332 GOVER, J.E.B., MAWER, A. and STENTON, FM., 1939, The Place-names of Wiltshire, Cambridge: Cambridge University Press GREIG J., 1991, ‘The British Isles’, in W. van Zeist and W.A. Casparie (eds), Plants and Ancient Man: studies in the palaeoethnobotany, 229-334, Proceedings of the 6th Symposium of the International Work Group for Palaeobotanists, Rotterdam: A.A. Balkema JONES, M.K., 1988, ‘The phytosociology of early arable weed communities with special reference to southern England’, in H. Kuster (ed.), Der Prahistoriche Mensch und Seine Umwelt, 43-51, Stuggart: Forschungen und Berichtezur vor-und Friihgeschichte in Baden- Wiurttemberg, Band 31 ROBINSON, M., 2004, ‘The plant and invertebrate remains’, in D. Jennings, J. Muir, S. Palmer and A. Smith (eds), Thornhill Farm, Fairford, Gloucestershire: An Iron Age and Romano-British Pastoral Site in the Upper Thames Valley, 133-45, Oxford: Thames Valley Landscape Monograph 23 ROBINSON, M.A. and WILSON, R., 1987, ‘A survey of environmental archaeology in the South Midlands’, in H.C.M. Keeley (ed), Environmental Archaeology: a regional review 2, 16-100, London: HBMC Occasional Paper 1 STEVENS, C.J., 2006, ‘Charred plant remains’, in M.G. Fulford, A.B. Powell, R. Entwistle and F. Raymond, Iron Age and Romano-British Settlements and Landscapes of Salisbury Plain, 152-8, Salisbury: Wessex Archaeology Report 20 STEVENS, C.J. and WILKINSON, K.N., 2001, ‘Economy and environment’, in G. Walker, B. Langton and N. Oakey (eds), An Iron Age Site at Groundwell West, Wiltshire: Excavations in 1996, 33—42, Cirencester: Cotswold Archaeological Trust WESSEX ARCHAEOLOGY 2007, Swindon Gateway, Swindon, Wiltshire: Integrated Archaeological Evaluation Report (Phases I and 2), unpublished client report ref 63890.4 YOUNGS, S.M., CLARK, J. and BARRY, T.B. 1983. ‘Medieval Britain in 1982’, Medieval Archaeology 27, 161-229 YOUNGS, S.M., CLARK, J. and BARRY, T.B. 1985. ‘Medieval Britain in 1984’, Medieval Archaeology 29, 158-230 Wiltshire Archaeological & Natural History Magazine, vol. 103 (2010), pp. 142-80 Settlement in the hinterland of Sorviodunum: a review by David Ff. James with a contribution by Bill Moffat This paper examines Romano-Bnitish settlements in the hinterland of Sorviodunum out to a radius of 19.2 km (12 miles) using County Sites and Monuments records, ‘grey literature’ and unreported museum information as well as previously published material. The pattern and density of sites 1s compared with other regions of Roman Britain with regard to market economy, relative wealth and trade and industry. The location of sites relative to the road and river network and to Sorviodunum 1s investigated as 1s the fate of settlements following the withdrawal of the Roman army at the beginning of the Sth century. The results of small-scale archaeological investigations inside Sorviodunum since 2001 are presented 1n the Appendix. Introduction Previous research by the present author has established that the Roman settlement of Sorviodunum was located in Stratford-sub-Castle covering an area of up to 45ha (James 2002). The precise role and influence of the settlement, however, remained largely unknown, although earlier research had concluded that as it was a medium sized ‘small town’ at the convergence of at least five major Roman roads, Sorviodunum would almost certainly have been an important market/trading centre and possibly a town where Imperial taxes were collected. It is clear that a study of the hinterland of Sorviodunum might provide indications of the marketing population (through the density and type of settlements in the surrounding countryside); an indication of the types of trade and industry; and reveal the nature of the economy and whether or not the town was a local administrative centre. Evidence relating to the fate of Sorviodunum and settlements in its hinterland at the end of the Roman period is examined in the light of recent historical and archaeological research. Since the publication of the previous paper eight years ago a number of further small, yet significant, finds have been made in the Sorviodunum area mainly due to archaeological watching briefs at housing construction or extension projects: a description of these is incorporated as an Appendix and their impact reviewed. Research Aims & Methods The principal aim of the research presented in this paper is to examine the currently known archaeological evidence for Romano-British occupation in the countryside around Sorviodunum by examining the type, frequency and location of settlements, the likely impact on the Roman town of farming and industrial activities and how the evidence compares with regions elsewhere in Wessex. The research draws on information from County Watermead, Mill Lane, Stratford-sub-Castle, Salisbury, Wiltshire, SP1 3LJ SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 143 I DOBUNNI Oo Bakes ee —o met: Meare ‘ewe, SAN Pie | ie ee 2 \ agers Noe sono ~ ~AT, Archaeological Sites and Monuments databases, archived ‘grey literature’ reports,! and unreported _ material held in Museum archives together with _ previously published information. | Landscape and geography: _a description of the hinterland of Sorviodunum | For the purposes of this study the hinterland _ of Sorviodunum is defined as extending to the _ approximate halfway point between it and the | nearest Romano-British ‘towns’. The six nearest _ towns are: Cunetio 36.3km to the north, Levcomagus (East Anton) 28.1km to the east, Venta Belgarum (Winchester) 34.4km south east, Bitterne 36.9km _ to the south, Vindocladia (Badbury) 34.4km to the - south west and Verlucio (Sandy Lane) 38.4km to the north east. Apart from Levcomagus, all of these towns lay between 34.4 and 38.4km from Sorviodunum. | To accommodate all six halfway distances, the hinterland of Sorviodunum is taken as a circle of / Be ~ 0 ® { 5 pore Si oe Bath +t. so iicenhall So Silchester @ TY. “SS ATREBATES Ch FBR. a Fig. 1 Late Iron Age Tribal Boundaries (after Cunliffe 1993) Showing the Location of Sorviodunum and the Hinterland 19.2km (12m) radius. For people settled within this hinterland Sorviodunum would have been the nearest substantial marketing town and trading centre. Interestingly, the only direction apparently without a town of any significance is to the west. The Romano-British ‘temple’ site at Cold Kitchen Hill, however, is located just over 37km away beside the Roman road from Sorviodunum to the lead mines in the Mendips.’ About three-quarters of the hinterland is in modern day Wiltshire, with the remainder in Hampshire (22.5%) and Dorset (3.3%). The geology of the hinterland The geology of the hinterland must be treated in two sections, to the west and east of the River Avon. To the west the chalklands continue as ridges running approximately east to west to Cranborne Chase and Dorset with the rivers Wylye, Nadder, and Ebble in the valley floors. To the extreme west the river Nadder passes through the Vale of Wardour with its complex geology of Gault, Greensand and Wealden deposits which have long been a valuable source of building material. To the east of the river Avon, the chalk is covered by deposits of tertiary sands and clays 144 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE at heights below c.76m (250ft) and the landscape changes from one characterized by open grassland and arable farming to include more woodland but still nevertheless supporting a productive agricultural regime. Settlement Background At the time of the Roman conquest of Britain in AD43, small clans of indigenous Iron Age people existed within a framework of large centralized tribes. From at least the middle of the 1st century BC intertribal rivalries became increasingly intense with rapid changes occurring to tribal territories (Millet 2005, 26). Pottery evidence and coin distributions provide a good indication of the areas occupied by each of the major tribes in the period just before the Roman invasion. Locally the principal tribes were the Durotriges, who appear to have occupied the area to the west of the river Avon and south of the river Wylye down into Dorset, and the Atrebates, or a sub-division within them known as the Belgae, to the east and north of these rivers (Cunliffe 1993, 205-13) (Figure 1). Settlement evidence from southern Britain leading up to AD 43 shows a basic pattern of farmsteads and hamlets with a few larger nucleations set in a landscape of so called ‘celtic’ fields. These settlements were fairly densely distributed on good soils and somewhat less-so on poorer soils. Houses were usually circular in plan often with a large floor area of up to 15m diameter and a cone shaped roof of thatch is inferred. Farmsteads were often enclosed by a bank and a ditch more than 1m deep, normally with a palisade or earthen bank, which probably served to exclude or hold stock and may well have served a tenurial function suggesting that land was effectively privately controlled at this time (Milllett 2005, 27). There were some larger Iron Age settlements. Over most of southern Britain, hill-top enclosures, or hillforts, formed ‘central places’ serving a variety of functions to judge by the results of excavations. Common in the early and middle Iron Age, such sites in certain areas had fallen out of use, for example in the territory of the Atrebates, but not in others such as that of the Durotriges (bid, 27). In this latter area hillforts remained in use and some contained large numbers of circular houses within substantial defensive enclosures. Around the beginning of the Ist century AD diverse focal settlements came into existence in southern Britain (Millett 2005, 27-9; Cunliffe 1993, 222); these were generally located in river valleys and referred to as Oppidum. Certain of these oppida became tribal capitals while others were focal settlements for smaller tribal groupings, the Pagi in the Roman period. Archaeological evidence indicates that initially the rural settlement pattern changed very little following the Conquest (ibid, 248). Romanization of the Romano-British native rural population was a very patchy process with some areas such as eastern and southern Britain adopting Roman styles and material goods quite readily, while other regions such as western and northern Britain appear to have been more resistant. Relatively few Iron Age rural sites appear to have been completely abandoned, while many show continuity of occupation through the period of conquest; although only a few have been excavated in the hinterland of Sorviodunum they are likely to have been common. Classification of settlement types has long posed problems. In order to simplify categorizing the many hundreds of sites covered by this paper, a tripartite division into ‘villages’, ‘villa type buildings’ and ‘farmsteads’ is used (King 2004, 349). The definition of a villa is a source of debate (Hingley 1989, 21, 1991, 75-6; Scott 1993, 1-8; Smith 1997, 10-11), but it is generally agreed that substantial domestic buildings in the countryside that indicate wealth or status can be classed as villas (Hingley 1989, 31). Villas are largely considered to be farmhouses or country houses of the native British aristocracy and the majority of them are relatively easy to define because they have common characteristics, such as mortared rectilinear masonry, ceramic or stone roof tiles, clusters of rooms, porticoes, and bath buildings. Embellishments include decorated wall plaster, window glass, mosaics, under-floor heating and ornamentation in the classical style (Blagg 2001). It is likely that many villa type buildings were half- timbered with stone/flint footings. Hingley (1989, 30) notes four basic plans for villa-type buildings, starting with a simple rectangular building with three or four rooms, sometimes called ‘cottage houses’ (Richmond 1969); aisled houses or halls with rooms either side of a passageway (in the case of the halls often with rooms located at one end of the structure); corridor or winged corridor houses with a corridor or veranda along the front of the building; and finally L-shaped and courtyard buildings which were usually of the grandest scale. Recent work has emphasized the diversity of architecture, specific purpose and time- SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW frame of villa-type buildings in the Wiltshire/East Hampshire area, although with common broadly agricultural/industrial use (Walters 2001, 144; ~ Cunliffe and Poole 2008). The distinction between villas and farms is ill- defined. Farmsteads were much more variable and simple in their nature than villa type buildings. _Earthfast timber, clay and daub and dry stone _ walling techniques are used, usually for circular, ovoid or simple rectangular buildings, presumably with thatched roofs. In plan, farmstead buildings usually only describe a single room. Hamlets were apparently quite common, consisting of up to six small farmsteads, and these may well have been dwellings associated with a single extended family _ (Hingley 1989). Villages are the most difficult to define because they grade into villas and farms/ hamlets on the one hand but can also be large enough to be regarded as small towns (ibid, 76-80). Perhaps the most useful definition of a village is a grouping of more than six buildings (the number can be very large, see below) of approximately equal size and apparent status. Such an approach separates villages from large villas (which can easily have attained village size in terms of population). Villages that straddle Roman roads usually take the alternative _ definition of roadside settlements or, if large enough, ‘small towns’ (King 2004, 350). Villages located alongside roads undoubtedly had their origins in being part of the network of settlements supporting _ travellers and merchants using the Roman road | System. The status of hamlets and villages probably varied. Those located close to large villas may have been settlements of either villa slaves or freeborn tenants. Those further away from villas were perhaps the homes of freeborn tenants or freeholders. Villages could be either linear, that is strung out _ along one main road, or nucleated where dwellings clustered around a central focal point and there were a number of side roads. Examples of small compact and linear villages on Salisbury Plain just outside the Sorviodunum hinterland can be found at - Knook East and Chisenbury Warren, where each had approximately 20 houses (McOmish et al. 2002, figs 4.11 and 4.14). Larger villages on Salisbury Plain, , such as Charlton Down, may have consisted of up to _ 200 households, covering 25ha and were as large as _ certain principal towns in Roman Britain. Charlton Down had a number of streets which converged on | a pond at the centre and a shrine, while part of the | valley just outside Charlton Down was dammed to retain water (ibid, figs 4.6 and 4.7). 145 Two final categories of settlement widely seen in Roman Britain are industrial sites, such as pottery kilns and stone quarries, and religious sites, with shrines and/or temples. The hinterland of Sorviodunum could be expected to have examples of both these types of sites. Settlement in the Hinterland Against the general background of Roman settlement types presented above, what is the character of the known evidence in the hinterland of Sorviodunum? Table 1 lists the numbers of findspots recorded on the Wiltshire, Hampshire, and Dorset Sites and Monuments Records and shows the breakdown between different types of settlement, including industrial/miscellaneous sites.* Figure 2 shows a plot of the principal settlement locations, in relation to roads and rivers, while these are listed in Table 2 by type and county. When a given site is first mentioned in the text it is followed by its entry number (in Table 2) in square brackets. The sites are broken down into 4 types, village settlements (VG), villa buildings (V), farmsteads and other settlement sites where the evidence is insufficient to be able to place the site in a particular category (S) and, lastly, miscellaneous sites (M) which are either industrial or religious in nature. Table 1: Summary of Romano-British Sites in Sorviodunum Hinterland ‘Type of Site _[ Wiltshire |Hampshire|Dorset | Total_ PSS Ca Villages Buildings unidentified sites Religious Roads Irackways|13 0 Nee 2 2 33 194 Village Type Settlements (38 sites listed in Table 2) Village settlements are listed as either VG or VG?; the latter being only possible sites. Brief details of the sites, including distance from Sorviodunum, known main roads and navigable rivers, are given in Table 2. The sites are distributed across the hinterland with 146 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE a>e : 3 | Map Produced by Wikshire Archaeology Service, Wikshire & Swindon Heritage 0 ia FP lns, Centra, Chippenham SN15 3QN Fig. 2 Plot of the Principal Settlement Sites in the Hinterland (Listed in Table 2) many located close to or beside roads or rivers. Only 12 sites have been subjected to modern excavation, while a further site at Woodyates (Dorset) was investigated in the late 19th century by Pitt Rivers (1892) and again in 1959 by Rahtz (1990). Many of the possible sites have been discovered through field walking, revealing heavy pottery/ building material scatters, sometimes supplemented by aerial photographic evidence or test pitting. The apparent by size of these sites ranges from around 70ha, as at Stockton Earthworks [21], to more than 40ha as at Teffont Evias [22] and at least 28ha at Downton New Court Farm [13], to small sites, probably little more than hamlets, at around 3ha, as at Britford [8]. Certain sites are of such a scale and/or located alongside major transport arteries that they were very likely local market centres (Hingley 1989, 25-9), for example Stockton Earthworks [21] and Woodyates [148]. The large number of 4th century coins found at Woodyates suggests a shrine. The most recent excavations on village sites are those at Butterfield Down, Amesbury [2], Netheravon Down, on the Netheravon/Figheldean parish boundary [18], Suddern Farm, Over Wallop [112] and Crystal Hollow, Godshill [110] both in Hampshire. At Butterfield Down [2] rescue excavations in 1990s revealed a late Romano-British village based on a mixed farming economy covering at least 6ha, which included a corn drier, timber buildings and a possible rural shrine (Rawlings and Fitzpatrick 1996). On Netheravon Down [18] excavations in 1991/2, supplemented with evidence from aerial photographs, geophysical surveying and surface collection revealed a nucleated 1st-4th century AD rural settlement with a corridor villa contained within an Iron Age enclosure. Aerial photographs also revealed a small polygonal enclosure, 0.Sha in area, of unknown date and function but conceivably a shrine or temple. Pottery covered the whole of the Roman period, while the early part of the assemblage is similar to that recovered from the 1959 excavations at Bower Chalke [6, 7](Rahtz 1990). Keyhole archaeology revealed seven graves, parts of five structures, and two corn-driers. Environmental evidence indicated the cultivation of spelt wheat and six-row barley and a predominance of sheep/goat in SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 147 Table 2: Principal Settlement Sites in the Sorviodunum Hinterland 8ha site of former IA polyganol enclosure containing a nucleated settlement and ‘villa type’ building Site Name |NGR Type|km |km |km_ |Comments Sites and References from |to to Monuments Sorv.|road |river No 1 {Allington |Boscombe |SU189394 |VG |9.5 |2.1 |3.4 |Timber structures, pits, burials, SU13NE301 | WANHM 54, Down and possible enclosure covering an 124-68 (Areas RB R West area > 30ha and S 2 {Amesbury |Butterfield |SU166414 |VG Single room dwellings, storage SU14SE324 | WANHM 85,156; pits, corn drying ovens covering |SU14SE331 |86, 158; 89, 1-43; >6ha, plus 5 nearby cemeteries 94, 244 (SU14SE353/4) containing around 300 burials 3 |Barford St}Hamshill |SU058332 |VG |7.7 1.7 |Evidence of Banjo encs, circular SU03SE301 |WANHM 62, 118- Martin and rectangular huts, also field sys- 21; 69, 185; 72-3, tem, multiple ditched enc of 16ha 205(80); Britannia 4,178; AntiqJ 74, 96-7 14 |Berwick |Berwick |SU056402 |VG? |12.2 3.1 |Settlement site discovered by SU04SE689 | WANHM 45, 174, StJames {Down (centre) Hoare who found pottery and (site 15); Hoare coins 1812, 95 5 |Berwick |Yarnbury |SU036403 |VG {13.0 |5.6 |2.6 |Enlarged IA hillfort of 11 ha which|SU04SW300 |RCHME 1999, 116- from Hoare onwards producing 8, Fig 62; WANHM considerable quantities of pottery, 39, 401; 45, 195; brooches and other artifacts. 46, 198-213; Hoare RCHME survey in 1990’s 1812, 90; Grinsell 1957, 38 S$U003212 |VG {16.8 |3.2 |3.5 |Excavations in 1959 revealed pot- | SU02SW305|WANHM 57, 395; tery, daub, quern fragments and intensive occupation behind early IA Ist C AD ditches 1.4km west of the site no 7 _|7_ |Bower East Chase |SU017212 |VG {15.6 |2.0 |3.6 |Excavations in 1959 revealed simi- | SU02SW304| WANHM 57, 395; Chalke Farm lar intensive occupation behind 83, 1-49 SM 39/91 | the early IA 1st C AD Great Ditch Banks to site no 6, 1.4km E ae Woodbur 1996R.700 Bulford |Milston S$U211457 |VG? |16.1 |1.6 {4.5 |Evidence of settlement found by P WANHM 45, 179 Downs Farrer from pottery etc in areaof |SU24NW302|(Site 42a); Grinsell field systems and ditches 1957, 52 10 |Bulford |Beacon SU195428 |VG? }12.7 {1.1 {3.2 |Roman pottery and other evidence | SU14SE300 | VWANHM 45, 180 | Hill of occupation found by P Farrer (Site 42b); Grinsell 1957552 11 |Chilmark |Whithy ST947319 |VG? |18.6 |3.9 |4.3 |Exact size of site unknown, village |ST93SW303 | WANHM 34, 414- Beds or villa. Nearly 5000 coarse and 8; SM 96R/703 fineware sherds inc samian and NE iron objects and rotary querns 12 |Clarendon|Home SU187305 |VG? {5.6 |2.7 |2.7 |Extensive scatter of R material Not Yet James and Gerrard |Park Copse suggests linear settlement Recorded 2007, 32, Figs 14 and 16 13 {Downton |New Court }|SU168220 |VG |10.2 {2.0 {0.7 |Fieldwalking and trenching over |SU12SE309 |WANHM 84, 143; Farm 28ha indicates evidence for exten- WCAS 1990.021 sive settlement with pits and clay lined oven | |14 |Downton |Standlynch|SU207243 |VG? |10.5 |2.J |2.3 |2 enc covering 8ha, surrounded by |SU22SW301 | WANHM 59, 113 $U204238 celtic fields, surface? pottery and |SU22SW302 large area of ground disturbance at $U20282390 found by Crawford $U147435 |VG {11.9 }3.3 {0.5 |Timber framed buildings some SU14SW301 |WANHM 40, 101; roofed with tiles, evidence of 66, 76-128; 87, pottery, dressed stone, pits, corn 125-7 drying ovens, 2 infant burials. Site covering area >6ha {16 |Enford Fifield SU139498 |VG {18.2 |6.4 Settlement covering~4ha, from SU14NW305 | WCAS 1994.03; Folly pottery, tile and other artifacts McOmish et al found 2002 {17 |Gt Wish- $U060355 |VG |8.2 {1.5 {1.1 |IA ramparts levelled. Area ~40ha |SU03NE300 | WANHM< 35, | ford or Grovely two late coin hoards found 115-31; 45, 194 (site | Earth- 105), 500; 65, 199 | works | | SU147471 |VG |15.4 4.7 |0.2 SU14NW302|WANHM 86, 8-57 SU14NW303 |(Site A); 92, 7-32; McOmish et al 2002 Fig 3.31, 82 148 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 2: Continued Parish Site Name |NGR Type|km |km |km_ |Comments Sites and References from |to to Monuments Sorv.|road |river No 19 |Orcheston|Church $U073481 |VG_ {17.4 7.2 |Planned linear village with up to |SU04NE300 | WANHM 45, 199 Pits 60 buildings covering ~ 9 ha and (Site 131); 80, 242 surrounded by 400ha of extensive (89); McOmish et field systems al 2002, 18-20 and Figs 1.17& 1.18; Britannia 23, (Fig 20), 298 Stapleford nearb work card 1974 21 |Stockton |Stockton |ST973362 |VG |16.6 2.6 |In 70ha of earthworks, masonry ST93NE300 | WANHM 43, 389- found, evidence of streets seen on 94; 56, 109; 67, 174; AP’s, possible LC 70-1, 136; 72-3, 206; 77, 57-8 22 |Teffont Teffont ST992308 |VG |14.1 |4.4 |0.1 |Finds of pottery, brooches, iron ST93SE301 |WANHM 35, 503; Evais Evais (centre) knife, hypocaust(?) made over area 36, 142; 38, 329; 45, Quarry > 40ha include large nos burials 485-6 23 |Winter- |Parsonage |SU065409 |VG {11.7 |3.4 {7.0 |Settlement revealed during evalua- |SU04SE310 |WCAS 2002.054 bourne Down tion excavation in 2002 Stoke 24 E of Roll- |SU086432 |VG? |12.4 {1.1 {5.4 |Settlement site discovered by SU04SE302 |WANHM 45, 209, estone Hoare (site 176); Grinsell 1957, 127; Hoare 1812, 116-7 25 Winter- SU089424| VG? |11.6 5.0 |Heavy pottery scatter recorded by |SU04SE303 | WANHM 45, 209, bourne Stoke Down Hoare (site 175); Grinsell 1957, 127; Hoare 1812, 116-7 26 |Winter- |All Saints |SU228325 |VG 0.7. |7.0 |Occupation and finds of fineand |SU23SW326 |WANHM 53, 258- slow Church coarseware pottery, and metalware 9; 72-3, 207; WCAS over an area of ~6ha 1999.004 27 |Wylye Bilbury SU010362 |VG {13.1 {1.1 [1.4 {IA hillfort of 6ha originally dis- SU03NW300|WANHM 58, 32, Rings covered by Hoare and producing 243-4, 468-9; 60, much pottery, coins, brooches, 135; 62, 120; Hoare | bracelets and querns 1812, 108; Grinsell 1957, 129 S 28 {Allington |S of Upper |SU207383 |V 0.4 {0.3 |Corridor type built in 3 phases, U23NW300 | Scott 1993, WZ 8; 15m x 7m, walls 1m thick, subse- Algar 1970; 1971 quently length doubled, Chilmark stone tiles 29 |Amesbury|NE of $U157423 |V 11.2 {1.1 |0.1 |Building at least 9.67m wide by SU14SE357 | WCAS 2003.152 Countess 12m long, walls 0.8m thick of com- Services pacted chalk and flint facing 30 |Amesbury|Earls Farm |SU182420 |V 11.4 |0.3 {2.2 |Pottery and foundations found by |SU14SE308 |Scott 1993, WZ 13; Down P Farrer WANHM 45 173; Grinsell 1957, 30 31 |Amesbury|Boscombe |SU178401 |V? {9.4 2.2 |Sherds and box flue tile, found SU14SE315_ |Scott 1993, WZ 14; Airfield during barrow excavation in 1930, WANHM 432-58; suggests possible building nearb Grinsell 1957, 207 32 |Amesbury| Earls Farm |SU184419 |Ve {11.4 2.6 |Pottery, oyster shell and tile frag- |SU14SE316 |Scott 1993, WZ Down ments found during excavation 15; WANHM 59, of barrow indicating a possible 39-40; Grinsell 2nd-4th C building nearby (and |WsyGa ts associated with find 14SE315? 33 |Barford St|Crouch’s |SU043325 |V 1.8 {1.1 |Pottery inc samian, segments of SU03SW302 |WCAS 2003.129 Martin Down limestone floor, roof tiles and pieces of stone building rubble 34 |Barford St}|Field S of |SU060331 |V 7.4 Pottery inc samian and BB ware, |SU03SE316 |WCAS 2003.129 Martin Hamshill mortarium rims (lst-2nd C), frag- Ditches ments of plaster/cement 1.6 35 |Berwick |Yarnbury |SU036403 |V W300) 5:85 2:6 SU04SW300 | Scott 1993, WZ 156; VCH 1.1, 38; RCHME 1999, 116-8; WANHM 39, 401; 45, 195; 46, 198-213; Hoare 1812, 90; SM 1962.64; DM 1985.117; 1995.64 RCHME geophysical survey revaled suspected rectangular masonry structure Langford Steeple walls and 4thC coins found in WANHM 57 212-6 church restoration in 1873 suggests presence of building close by >? 0 (5. St James; | Hillfort 36 |Britford |StPeter’s |SU163284 Roman bricks re-used in Church |SU12NE303 |Scott 1993, WZ 34; Church SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 149 - Table 2: Continued ra No) [e) “I 44 45 6 0 Zz 4 Mr Site Name |NGR Type|km |km Comments References from |to Monuments Sorv.|road |river No Clarendon Park 38 |Clarendon|Fussell’s 39 |Clarendon|Gt Gilberts |SU188298 140 |Downton |E of Moot |SU181211 |V 11.6 |1.4 {0.3 {Corridor type with at least 7 SU12SE301 Lane rooms, bath house, mosaic floor and corn drying surface built late 3rd C {41 |Durnford |N of Coffee|SU143370 |V 5.4 {1.0 {1.2 |Bldg discovered during excavation |SU13NW300| Scott 1993, WZ 72; | Farm of oval cropmark. Flint wall foot- WANHM 56 245; ings, pottery, coin 57 173-5 (43 |Fighel- dean Fonthill Odstock Pitton and Farle 3 |South Newton Stapleford km Sites and to Nr Best’s |SU210291 |}V? {8.1 {4.1 |3.8 |Significant quantities of late RB SU22NW303|James and Gerrard Farm sherds, tile and brick fragments 2007, 34-5, Fig 14; over area >4ha Arch Con 1996, 40; Gazetter 37, No 319 SU191318 |V? {5.8 {1.4 {3.9 |Substantial lynchets, much early |SU13SE318 |James and Gerrard RB pottery, c 20m stretch of flint 2007 34, Fig 14; and mortar walling, possible Arch Con 1996, 39, Temple site? Gazetter 26, No 192 NF and BB pottery, roofing slate, James and Gerrard Dairy Cot- tage box flue tiles, pilae bricks and 2007, 31-5, Figs 14, tesserae, density of metal residues 15 and 16; Arch suggests possible semi-industrial Con 1996, 39-40; site Gazetter 20, No 136 Scott 1993, WZ 71; WANHM 55 176-8; 58 303-41 RB pottery, tile fragments, coin and other artifacts found over an area of ~ 4ha indicating possible SU139498 |V? |18.1 |6.4 SU14NW305 | WCAS 1994.03; DM 1993, 12 building E of Alton |SU153467 |V 15.1 |4.1 {0.2 |Roman pottery and building SU14NE301 |Scott 1993, WZ 81; Parva foundations WANHM 45 188, (site 84a); Grinsell 1957, 70 SW of ST945353 |V? {19.2 4.3 |Centre of area of concentrated scat-|ST93NW300 |SM 1969.9; SM High Park 6” map, WCAS, fieldwork Sir J Wil- loughb ters of RB pottery inc samian and 1st C local wares East Grim- |SU234275 |V 10.9 {5.3 {5.9 |Excavations by Heywood Sumner |SU22NW301 |Scott 1993, WZ stead in 1914-24 revealed a corridor villa 90; Sumner 1924; with 3 bath houses Grinsell 1957, 75 Pottery and tile fragments found during excavations in 1963-8 Bath house found in 1907, tessel- |SU14NW301 |Scott 1993, WZ lated floor in 1936, other finds in- SU182356 |V? ed Nethera- |SU148482 16.6 15.1 [0.2 von House SU147471 |V a 0.2 SU149310 |V 3 1 128; WANHM 45 490-1; 47 538 SU14NW303 |McOmish et al 2002, Fig 3.31 A late RB settlement, partially SU13SW312 |Scott 1993, WZ excavated 1941-8, finds include 146; VANHM hearths, glass, pottery, roof tile, 52 394-9; 53 258; coins, iron slag and a sandal Grinsell 1957, 102 von Road c NW of SU153267 |V? 4.7 |1.4 |Large amount of pottery and brick }|SUI2NE311 |WCAS, L Earley Longford found in area indicating possible pers comm Farm building V 3.5 {2.4 |Various finds around farm include |SU22NW300|] Scott 1993, WZ stone tiles, box tile, horseshoe, 134; Grinsell 1957, clude roof tiles, ridge tiles, coarse blackware pottery and 6 coins Corridor villa found in 1993 during geophysical survey of area inside 8ha IA polygonal enclosure, and 198 coins found by metal detectorist in 1987 Nethera- von Down Nethera- Farley $U231297 Farm nails, very large quantities of pot- 96 tery and a coin of Antonius Pius N of Biggs |SU215304 |V? {8.3 {2.8 {1.9 |Reputed earthwork which could be |SU22SW313 |Scott 1993, WZ 135 Copse a possible villa or a camp Camp Hill |SU111338 : Settlement revealed in 1933 dur- —|SU13SW300 |Scott 1993, WZ Reservoir ing reservoir construction. Finds 148; WANHM 46 include sherds, painted plaster, 389; Grinsell 1957, a rotary quern, and remains of a 106 SU03NE305 |WCAS, OS field- work card 1974 Floor and foundations partly ST93SE302 |Scott 1993, WZ underlie a mound. 225 coins found 170; WANHM 38 on floor in c. 1920, with sherds. 329; Grinsell 1957, Site may be a temple 13 small oven with a flue Much pottery, building 6 material+slag suggests iron work- ing and building Ilml NW {SU057384 |V? {10.1 |4.2 {1.2 of Stapl- eford Upper $1T983317 |V 1520535 Holt Copse LUN) 150 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE a ee ee Boscombe Rd/New Covert Amesbury 7 Excavated by R R Clay in 1925, RB sherds found in upper levels Table 2: Continued Parish Site Name |NGR Type|km |km {km |Comments Sites and References from |to to Monuments Sorv.|road |river No 56 |Tidworth |Above SU227474 |V 18.4 {2.8 |0.5 |Site of building indicated by BB |SU24NW307|WCAS 1994.031 Flood ware and fine greyware, fragments Plain of limestone roofing tile and con- Tidworth siderable amounts of animal bone 57 |West Hollyflow- |SU258271 |V 13.3 {5.4 |6.3 |2 Bldgs one aisled one courtyard |SU22NE300 {Scott 1993, WZ Dean er Field type. Pt of 3rd also discovered. 179, HA 100; Painted wall plaster, pottery, glass, WANHM 13 33-5; tiles, roofing slates, marble frag- 22 243-50; 45 185; ments, piece of mosaic and coins Grinsell 1957, 119, AD 205-353 VCH Hants 1, 311-2 58 |Whitepar- | Abbot- SU243236 |V? |13.7 |9.1 |6.1 |Box flue tile found suggesting SU22SW303 |Scott 1993, WZ ish stone Farm possible building nearby, Taylor 184; WANHM 62, records accounts of much RB 79-102; SMAR material found in this area 1933-4, 1 59 |Winter- |Broomhill, |SU242338 |V 11.1 |0.7 {5.9 |Rectangular flint foundations exca-}|SU23SW306 |SM 1954.49; SMAR slow East Win- vated in 1954 and finds of pottery, 1955, 11 terslow roofing tiles, ironware and a coin Winter- |EofOld |SU236324|V?_ {10.3 6.3 |Possible villa site recorded by L SU23SW318 |Scott 1993, WZ 187 slow Manor V Grinsell in 1949 but not listed Farm in VCH 61 |Winter- |W of West |SU228325 |V 5.6 |Building to W of Church. Finds |SU23SW317 {Scott 1993, WZ slow Winter- include coins, pottery, key, frag- 188; WANHM slow Ch ments of bronze and lead 72-3, 207 62 |Alderbury|Whaddon |SU199260 |S: 8.7. |7.1 {1.7 |RB sherds, oyster and other shells, |SU12NE302 |SM 1932.120; Brick- quern stone (pudding?) 1934.71; Grinsell works 1957, 23 63 |Allington | West SU194389 |S 9.55 {1.8 |4.1 |RBsherds anda London coinof |SU13NE315 |WCAS, OS field- Boscombe Carausius found where other coins walking/surveying apparently found previously record card 019; SM 6” Shortt Map 64 {Allington | Little SU198383 |S 9.3. |1.0 {4.8 |RB sherds, and bronze coin of SU13NE303 |Grinsell 1957, Boscombe Constantius II (AD 341-6) found 26; WANHM 45, Down by P Farrer 177(32a); Archaeo- logical Review, CBA Grps 12 and 13 Vol 5 (1970), 2 65 |Allington;|NW of $U190391 |S 9.4 : 3.6 |RB sherds were found during SU13NE314 |WANHM 54, 135 Idmiston |Boscombe excavation by Richardson in 1949 (Area ‘p’); Grinsell in upper fill of inner ditch and low 1957, 25-6 down in outer ditch SU164410 |S Two RB middens containing NF |SU14SE303 |SM 1951.84; SM pottery and spindle whorl, 2 more 2.5” map pits 50m from middens and a N-S road 67 |Barford St}Crouch’s |$U043325 |S 1.7 _|Finds during fieldwork in 1999- SU03SW302 |WCAS 2003.129 Martin Down 2000 inc bracelet, 1st C samian, lo- cal pottery, segments of limestone floor, roof tiles and pieces of stone building rubble Barford St|Field S of |SU060331 |S 7.4 1.7. |Collection of RB pottery inc BB SU03SE316 |WCAS 2003.129 Martin Hamshill ware, mortarium rims (lst-2nd Ditches C) small fragments of samian and pinkish plaster/cement . Broad SE of S 11 2 SU02SE302 |WANHM 45, 212; Chalke Knighton Arch J 104, 33-4 Hill Farm of ditch around a kite-shaped (Fig 2) enclosure Without |Barn fragments 263 > ? ? 2 ? 9; 91, 11-33; WCAS bish; slight earthworks RCHME 1994.005 and 018 71 |Chilmark |W of East |ST971324 |S? 16.2 |3.0 |2.6 |RB coarseware sherds and Kim- ST93SE306 |WCAS, R Lever Farm meridge shale spindle weight from pers com Eyewell site below? 72 |Chilmark |Eyewell $T971322 |S 16.2 |3.2 |2.4 |Ovenorcorn dryer, and pitsand |ST93SE311 Fm ditches containing domestic rub- {and 312 surveyed. Cists aligned E-W 73 |Chitterne |Gdn of the |ST992441 |S? {18.8 |8.8 |5.0 |NFand imitation samian sherds |ST94SE301 |Archaeological Grange plus bronze coin of 3rd C Review, CBA Grps 12 and 13, Vol 4 WANHM 85, 158- | ~ SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 151 from |to to Monuments Sorv.|road |river N oO SU171313 |S 3.8 |1.6 |2.8 |Evidence of postholes, ditches, SU13SE302 |James and Gerrard pottery, and inhumation graves in 2007, 30-5, Fig 14; 18ha enc indicates presence of at WANHM 84, 116- 9; 91, 162; 92, 33-8; Arch Con 1996, 40 Gazetter 22, Nol51 eee ee SUI2NE308 |James and Gerrard 2007, 30-6, Fig 14; - Table 2: Continued Lvat vu I | fo SU166295 Dense scatter of RB sherds and crop mark enc least one farmstead Arch Con 1996, 40, Gazetter 30, No 234 SU12NE309 |James and Gerrard 2007, 30-6, Fig 14; Arch Con 1996, 39-40; Gazetter 16, No 102 ST94SE305 |Grinsell 1957, 58; WANHM 38, 228; 45, 183-4 site 59; Hoare 1812, 80 Many RB sherds, coin of AD ST93NE306 |Grinsell 1957, 166; 364-7, and other metalwork found WANHM 58, 422 during barrow excavation by (site C); 59, 206; Vatcher in 1958 DM 1963.23 Codford |SE of RB corn drying oven or hypocaust, |ST94SE306 |Grinsell 1957, Settlement sherds, brooch 58; ALM 53 site (ST94SE305 at 17.3km at Clay Pit Clump $U091211 7.8 |RB pottery fragments and 8 coins |SU02SE304 |WCAS, R Canham Bissett Toyd Farm found in 2002 pers comm : /|81 | Dur- Dur- $U124438 12225» 2 Large quantities of samian and SU14SW300 |Grinsell 1957, 66; WANHM 45, 186 (site 72a); WCAS, OS fieldwork card 1969; RCHME 1979, 24 S S rington |rington other RB sherds 2 |Ebbes- Fifield S$U005257 |S? 9 14.1 |RB Sherds, brooch, coins and SU02NW301 | Grinsell 1957, 69; hobnails WANHM 42, 464- ? 5) S? ? $U182297 Stray finds and shovel pit testing revealed Ist and 2nd C pottery “I ~—] I “I Ke) [o-) ST995421 RHMC plotted C-shaped enclosure from AP’s, RB sherds odford |S of Clay Pit Clump C Codford |Lamb ST990391 Down vu S 2) ~w z . ST998419 of Barrow G38 vie et Down E 14. : bourne Bavant Wake Down 5; 45, 188 183 |Ebbes- Monks ST985219 |S 17.8 |5.1 |7.3 |Many RB sherds noted by RR ST92SE300 |SM 6” map bourne |Hole Clay Wake Bae 0.2 |Lead coffin discovered in N face of |SUI15SW315 |WCAS, R Drew Quarry in 1984 pers comm es 2.1 |6. | Enford 5 |Fovant Chiselbury |SU018281 1 6.5 |2.3. |RBcoarseware sherd, 2 coins (one |SU02NW300|Grinsell 1957, 71; Camp of Constantine), and lead spindle WANHM 45, 189; whorl DM 1975.78 SU02NW305 | 186 |Fovant |Fir Hill | [SU006295 187 | Laver- W of Ende |SU158341 |S ‘ 3.2 |Site of enclosed settlement, RB SU13SE307 | WANHM 47, 409; stock Burgh pottery inc rim of 3rd C cooking Grinsell 1957, 81 7 pot 8 2 co Laver- Cockey SU169316 |S 3.6 |0.4 |3.2 |Quantity of RB sherds, worked SU13SE324 |WCAS 1999.015; stock; Down and burnt flint found during water WANHM 94, 250-1 Clarendon pipeline watching brief in 1999 Park New Bishop- SU144324 |S 1.4 {0.2 {1.6 |RB sherds and tile fragments SU13SW311 |WCAS 1991.006 ; | Sarum down with dense scatter of burnt flint 1992.021 . partially excavated in 1992 ee eee Sarum down found during evaluation in 1991 1 |New Bishop- $U151319 |S 1.8 Sarum down Fm 9 1.6 {Scatter of pottery, tile and burnt SU13SE315 |WCAS 1991.006; flint and 26 pits found during 2 1992.021 evaluations in 1991-2 92 |New Highfield |SU133308 |S SU13SW315 | WANHM 46, 579- | Sarum 612; 48, 292-300; Britannia 4, 178; anus (AD 260-73 Grinsell 1957, 102 93 |Odstock |In SU135244 |S? |7.3 14.9 Pheasantry Copse 3.6 |RB sherds, and many pot boil- SU12SW300 |WCAS, OS field- ers/burnt stones found during OS work card 1954 fieldwork 0.5 |0.5 |Finds from 19thC excavation in- clude metalware, Samian and NF pottery, and hoard of 25 Antonini- 152 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 2: Continued from |to to Monuments Sorv.|road |river No 94 i SU204323 |S Wea 5.2. |RB sherds and brooches found SU23SW310 | WANHM 52, 395; over long period at ‘tumulus’ (site 58, 243; Grinsell excav in 1961). Also coins early to 1957, 96; SM 2.5” late, iron key, oyster shell Map 95 |Pitton Winter- $U202324 |S 7.1 10.7 |5.2 |Probable RB horseshoe enclosure |SU23SW307 | WANHM 58, 470; bourne containing a late 4th C cemetry of JRS 53, 149; SM Down 37 cremations and 14 inhumations 1963.112 Firsdown excavated in 1962 (see entry above Shrewton |NE of SU050446 |S 15.4 {12.6 |7.1 |Small circular posthole struc- SU04SE306 |WANHM 839, 44-72 Madding- ture with hearths and pits; 2 ton Fm burials,querns, nails, bone pins etc 97 |Tisbury |W of Ash- |ST945308 |S 18.9 |3.9 |1.4 |Irregular quadrilateral enclosure |ST93SW304 | WANHM 34, 414- ley Wood excavated in 1903-4 finding ditch 20; Grinsell 1957, Down 3-4 ft deep with RB coarseware and 114 and 270 (218) samian in primary silt Tisbury |Little $T945320 |S? {18.9 |3.8 |7.4 |Skeleton in stone lined grave ST93SW302 |WANHM 39, 103; Ridge found in 1914 45, 206; Grinsell 1957, 114 (bldg at $T941314 19.3km The Bushes ee oS ae Hospital excavation in 1997 {WCAS 1997.002 100}Winter- |High $U073422 |S 12.1 |8.4 : Pottery, coins etc found by Hoare |SU04SE687 |WANHM 45, 209 bourne Down (site 174); Grinsell Stoke 1957, 126; Hoare 1812, 95 101}Winter- |NW of $U228326 |S 0.5 |7.1 |RB sherds found during excava- U23SW309 |Grinsell 1957, 127; slow All Saints tion of earlier date pits WANHM 53, 258- Church 9; WCAS 1999.004 S 102|}Winter- |All Saints |SU228324 |S? |9.5 {0.7 {7.0 |8 coins of 3rd and 4th Cinclof |SU23SW326 |WCAS 1999.004 slow Church Allectus, 3 frgs of bronze and 35 lumps of lead scattered over 50 sq yds found in 1936? RB sherds and possible piece of imbrex found during evaluation excavation to Church in 1999 103 | Barford St} NW of Fri- |SU065330 |M {7.0 1.8 |Roughly rectangular mound with |SU03SE305 |WCAS, OS Martin ars Peak RB sherds and bldg rubble, pos- fieldwalking 1974; | slble temple WCAS 2003, 129 104|Tisbury | Lady ST963312 |M 17.2 |4.3 |1.4 |Industrial quarrying site for ST93SE308 | WANHM 34,421-2; Down reputed remains of RB ‘beehive pit 58, 471-2 type’ extraction of Chilmark stone 105 | Winter- SU225329 |M 9.3 {0.2 |7.1 |Excavated 1959, constructed in 3 |SU23SW308 |WANHM 57, slow Winter- superimposed stages, pottery late 396-7; Antig J 43, slow 3rd- early 4thC possible amphithe- 197-213 atre or theatre? ib) Hamipsbite ® ic | "oh PL) tulle ainsi, Mees tout elit Vs se 0S 106 |Bossing- |Bossington |SU325310 |VG {19.2 |0.2 |2.5 |155 sherds mainly Alice Holt and [55722 Old Sarum to NK tile and floor tile, briquetage, Braishfield Gas nails, animal bones and flat bot- Pipeline, M D tomed ditch, AP’s suggest village Smoothy with field system 107 |Breamore | North SU161184 |VG {13.7 {4.0 |0.4 |Large RB site, pits, features and 29628 . |Light at al 1995, Street heavy pottery scatter over at least 76-84, Fig 28 15ha 108 |Buckholt |Horseshoe |SU279334 |VG? |14.6 |1.2 {12.6 |Occupation debris; Stukeley 41760 VCH Hants Vol 1 Woods reported Roman antiquities often 312, 322-3; Hoare found, site is possibly Brige 1821, 58-60 109 |Fording- |Bicton SU150129 |VG? {18.9 {8.1 |0.1 |Heavy pottery scatter on terrace 29723, 29868 | Light et al 1995, bridge edge overlooking river covering > 76-84, Fig 25 7ha suggesting possible settlement 110 | Godshill S$U165142 |VG_ |17.8 Settlement with street plan, house |29817-21 Archaeology in platforms, timber buildings major Hampshire HCC ditches IA to 4thC Annual Reports 1989-92 111 ]Godshill |Island SU215151 |VG {18.5 {2.1 |4.9 |Earthworks covering at least 0.7ha |27775, Berks Arch Thorns (centre) comprising terraces, platforms, Services, 2000; sites of numerous pottery kilns, 42282, 42338 | PHFCAS New For- and lynchettes est Section Reports 14-19 (1975-81 112 |Over Wal-|Suddern |SU280377 |VG {15.8 |4.2 |7.6 |Enclosure ~4ha in size with triple |30233-6 Cunliffe and Poole, lop Farm ditches, ovens, pottery and roof 2000a slab, shallow pits, grave io” em) © = QO an ey Lend oO — o fo} SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 153 Table 2: Continued Parish Site Name |NGR Type|km |km |km_ |Comments Sites and References from |to to Monuments Sorv.|road |river No a3 Shipton SU244459 |VG Village and associated fields cover- |41487 Defence Estates Bellinger | Bellinger ing over 20ha plotted from APs TAEM, 21, 1997 114 | Whits- Castle $U127197 |VG? |12.0 |4.7. 14.4 |Sherds Ist - 4th C found in grav- |29567 PHFCAS, 43, 63-81 bury Ditches of tiles, bricks, mortar and animal elly-silt overlying IA occupation 15|Brough- |W of $U294331 |V 16.1 |1.3 |4.8 22374 Scott 1993, HA 22; ton Church VCH Hants 1, 312 bones possible RB settlement 116|Buckholt |Buckholt 1 |SU293319 |V? |16.3 {0.2 |4.5 |Roman pottery, piece of tile keyed |30222-4 Scott 1993, HA 23; for plaster, 2 small lumps of iron, Hoare 1821, 58-60 and field system About 1 mile W of Church, 7 exploratory pits dug by A Potrell in 1972 revealed short length of collapsed walling, a little samian, coarsewares, an imbrex, fragments errs {117 [Buckholt_[Buckholt 2 |SU285319 |V? /}118 | East Dean |SW of SU281270 |V Aisled farmhouse, much window {22156 Scott 1993, HA 33; | glass, portland roof slates, animal VCH Hants 1, 312 bones, oyster and snail shells, pot- | tery, metal objects and nearly 200 | coins mostly 4th C a — \oO tT fet) n Cond al therley 3: Stone, tile, wall plaster, pottery inc |22153 Scott 1993, HA much fineware, 298 coins, harness 35; Master 1872a; fittings, spindle whorls and iron 1872b implements 0.4 |Wall plaster and area of flint nod- |55948 AVAS Unpublished ules evidence for a building only Field Survey 0.4km from site no 141 9: lst-4th C site covering 15ha with |22912, 22925 |Scott 1993, HA ERB timber bldgs and LRB 41; PHFCAS 6 (pt masonry Aisled Hall, double cor- 4) 341-2; Current ridor villa and two crop processing Archaeology 188, buildings (both with corn-driers) (2003), 344-51; Cunliffe and Poole 25430-1 30223-4 Cunliffe and Poole, i] |120|Godshill [Folds SU164157 |V? Farm | Holbury |SU286274 |V Wood $U275411 2008 121|Grateley |Grateley South |]122 [Over Wal-|Over Wal- |SU306399 19.2 |3.6 lop lop Over Wal-|Suddern [|SU280376|V? |15.8 |4.2 d Farm 2000a also a rubbish pit, containing , _ Rock- West Park |SU120170 |V 14.7 |8.4 |3.6 27779 Scott 1993, HA bourne 793 JRS 33, 75; 35, 88; 53, 150, 164; 555,217, 228; 56, | 214, 219-20, 225; Britannia 11, 394; RCHME 1983 [ ae ‘ ip /|126 | Whits- Whitsbury |SU129192 |V 1257-3), 3135 21592 Scott 1993, HA 102; bury Church PHFCAS 43, 63-81; samian and a quern fragment, and SM Manuscript Rpt | a hypocaust with flint walls. No and plan coins found ||127|Braemore |Braemore |SU144194 |S 12.3 |4.5 |2.2 |Heavy pottery scatteron Wedge |29596 Light et al 1995, | Wood of wood, full extent unknown, pos- 76-84, Fig 26 sible small hamlet 128|Braemore |Outwick |SU145176 |S? [14.1 {5.7 [1.7 |Small pottery scatter over 0.25ha, Light et al 1995, possible occupation site 76-84, Fig 27 \|129|Braemore |S Charford |SU162194 |S? |12.7 {2.7 |1.0 |Heavy pottery scatter on two adja- |29674 Light et al 1995, | cent sites covering ~3ha, possible 76-84, Fig 23 hamlet? '|130|Bramshaw|Eyeworth |SU224150 |S LOOM Le7 At least 3 structures, quantities of |31254 Wise 1863; PHF- i Wood pottery but no sign of kiln CAS New Forest Section Ann Rpts | Nos 14-18 (1975-80 3 inhumation burials, grain store, Purbeck stone roofing slab, or tile brick fragments, oyster shell, pot boiler and pottery, pit and 2 coins Excavations in 1991 and 1996 by Cunliffe found evidence of hexago- nal stone roofing slabs and possible peripheral villa activities Rich, large courtyard type villa, Purbeck stone roof over the cor- Scott 1993, HA 71; Crawford 1929, 18; PHFCAS 17, 60-3 12 I o Ss 22 3 24 Destroyed walling, tile, brick, pot- tery, and alien stone discovered in 1922. 3km SW of second Thruxton villa at SU 298462 lying outside hinterland. RB pottery found near Church, 154 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 2: Continued Parish Site Name |NGR Type|km |km |km_ |Comments Sites and References from |to to Sorv.|road |river 131|Damer- {Soldiers $U082176 |S? {15.2 {5.2 |7.3 |11.5ha polygonal enclosure, 16160 Sumner 1914 ham Ring probable RB cattle enclosure (and settlement?) overlying existing celtic field system bridge heath 76-84, Fig 23 133|Fording- |Upper SU154165 |S? 15.5 0.4 |Possible settlement site indicated {29672 Light et al 1995, bridge Burgate by several pottery scatters over 76-84, Fig 23 area of 18ha, IA to 4thC [o> _ S n SU054187 15.3 |2.5 [10.3 |Heavy pottery scatter suggests 29471 AVAS Unpub- Down occupation site lished Field Surve RB finds by Pitt Rivers ditch of 2ha LBA enc in 1894/6. Pit con- — —" co Pitt Rivers 1898, 185-215 taining RB artifacts found in 1953 200m NW. So}; NM Nn NN}; WN Ww Settlement 120m x 90m 16174 RCHME Unpub- lished field surve Considerable quantity of sherds on yo be 22187 PHFCAS 14 (Pt 2), 136-94; Unpublished OS fieldwalking NW slope of hill towards Quarley Manor Archaeology in Hampshire, HCC Annual Report for 48 29589 N 3 8 Zz S — a) No} Nn 5 i) fo) = 3 o) =) = wn 135 | Martin Martin SU043200 |S? {14.7 10. Down Down 137|Quarley |Quarley $U265427 |S?} 17.2 Hill 139 | Rock- Spring SU108215 |S? {10.5 14.3 c 39ha polygonal enclosure, bank = |21713 Sumner 1914 bourne Pond with ditch on both sides, occupa- Enclosure tion inside and outside 140 |Bramshaw] Church $U220150 |M__—-}18.8 {1.9 {5.5 |Occupation site, a series of banks PHFCAS New Green and ditches, kiln Forest Section Ann (1975-80); Swan 1984 141/Godshill |Armsley |SU165161 |M_ {15.9 {5.1 {0.1 |Possible industrial settlement, 27808 Edwardes 1927, 56- floors (2), dipping ponds or wells 62; Musty 1973 2 143 |Godshill |Crock $U215143 |M_—s—-} 19.2 {2.8 |5.1 |Earthworks associated with 6 kilns |42384 Berkshire Arch Hill Main Services 2000 Group 144/Godshill |Island $U213153 |M_ {18.3 {2.1 {4.6 |Kiln and hut excavated by Sumner |27772 Sumner 1927; | Thorns in 1925 Berks Arch Ser- vices 2000 145 |Godshill |Pitts Wood |SU197146 |M? |18.3 {3.4 |3.3 |Masses of pottery recorded during |54606 Pasmore 1967 drainage ditch work in 1965 now dense conifer plantation, kiln site? c) Dorset 147 | Cran- Toby’s SU047155 |VGP {18.3 {3.2 |10.2 |AP’s suggest unenclosed settle- 3005 034 RCHME Dorset bourne Bottom ment of at least 4.5ha overlying v, 15 Celtic fields 148 |Pentridge | Woodyates |SU033198 | VG 5 |0.1 {12.6 |Settlement excavated by Pitt Riv- |3017015 {Arch J 104, 62-78; ers in 1888-90, finds date from Ist 118, 65-99; Pitt C but some 1200 coins found are Rivers 1892, 3-239; from the period 275-400 AD sug- Rahtz 1963a; gesting a market and/or shrine RCHME Dorset Vv, 55 V plaster, 2-4th C pottery and 5 4thC 150 | Cran- NE of SU053141 |S 19.2 |4.3 |9.3 |RB pottery found in ditch contain- |3005 038 PDNHAS 86, 119; bourne Manor ing oyster shells, tiles and a coin of RCHME Dorset Farm Constantine I v, 15 151 }Pentridge | Woodyates |SU016195 |S 16.8 {1.2 |14.1 |RB occupation debris of pottery, {3017 014 PDNHAS 49, 77; Manor coins, oyster shells and animal RCHME Dorset bones v, 54 18.5 13.8 |Pottery from 3-4th C found in ear- |3026 036 PDNHAS 71, Down 138 | Rock- Rockstead |SU127172 |S 14.5 |7.4 |2.9 |Heavy pottery scatter of Ist-4th C bourne Farm over area 50m x 70m terraced into hillside Rpts Nos 14-18 Godshill |Ashley $U200147 3.0 |3.5 |Three kilns excavated by Sumner |28786 Sumner 1927; Swan Rails in 1926 1984 SU132160 Possible kiln site dating to last 54618 PHFCAS 39, 69-75; bourne Farm quarter of 3rd C Swan 1984, 138 feYMDOnSet sinie ce TU ee ai ag a ee TY 15. 149 |Pentridge |SW of $U013192 17.3 1.3 |14.3 |Building indicated by flint nod- 3017 047 Green 1990 ules, Purbeck stone tile fragments, coins lier IA 4 sided enclosure defined 69-70; 72, 92-3; by a bank and external V shaped 73, 104; RCHME ditch enclosing 3.6ha Dorset v, 99 SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 155 _ Table 2: Continued - WCAS Key VG Village type settlement V Villa type structue S Farmstead or other settlement type M Misc eg Industrial, Religious etc AVAS Avon Valley Arc Soc DM Devizes Museum Accession No PHFCAS Proc Hants Field Club Arch Soc JRS Journal of Roman Studies PDNHAS Proc Dorset Nat Hist and Arch Soc RCHME Royal Comm Hist Mon England SM Salisbury Museum Accession No SMAR Salisbury Museum Ann Rpt WANHM Wilts Arch and Nat Hist Mag Wilts County Arch Services Distances shown in italics are from a conjectured road | the economy followed by cattle. At Suddern Farm [112] a small part of a roughly circular (approximately 4ha in area) multiple- ditched enclosure was investigated in 1991 and | 1996. Occupation ran from the Iron Age through _ the Roman period and after the mid-1st century AD the enclosure ditches were levelled and the area of occupation expanded (Cunliffe and Poole 2000a). There was a possible short break of occupation | between the mid-3rd and mid-4th centuries. The _ economy in the early Roman period showed an increase in the number of sheep to cattle while, by _ the later Roman period, cattle provided some 80% of - the total meat yield. Charred plant remains indicated | that spelt wheat and hulled six-row barley were the _ main crops, while several groups of small bread _ ovens were also discovered. Non-local goods were _ available to the inhabitants with sherds of amphorae and copies of Gallo-Belgic beakers, with 25% of the pottery coming from the Poole region. The site at Crystal Hollow, near Fordingbridge | (Hants)[110] is barely 300m from the River Avon and appears to have been a small settlement with _ two streets. Occupation began in the early Iron Age and the latest pottery is late 3rd or early 4th century _ AD (Light et al. 1995, 84). Field survey and test pitting has been undertaken over the last 10 years on Salisbury Plain (McOmish et al. 2002; Fulford et al. 2006), at Clarendon Park _ (James and Gerrard 2007), Downton, New Court Farm (WCAS 1990.021), Grovely Woods (WCAS 2003.129) and Yarnbury hillfort (RCHME, 1999, _ 116-8 and fig. 62). The Salisbury Plain study lay _ mostly to the immediate north of the hinterland _ and established the location of 11 villages with one, at Church Pits Orcheston [19], falling inside the hinterland study area. Here a planned linear village was identified with up to 60 buildings covering over 9ha (McOmish et al. 2002). The archaeological survey of Clarendon Park identified a number of sites with one, at Home Copse [12], classed as a possible linear village-type settlement following the discovery of an extensive scatter of Roman material. The 267 pottery sherds recovered from shovel pits included Alice Holt, Savernake, samian, Oxfordshire colour-coated and white ware, New Forest colour- coated ware and Dorset Black burnished ware (James and Gerrard 2007, 32, 212). In 1989 the archaeological evaluation of 28ha of land at New Court Farm, Downton [13] involved machine trenching and field walking on an area of rectilinear cropmarks together with part of a complex of circular and penannular features visible on aerial photographs. Results suggest a settlement area with two distinct phases of activity, one in the earlier 1st century AD, the other in the 2nd and 3rd centuries. Amongst the features investigated was a clay lined oven (Anon 1991, 143). Fieldwalking in Grovely Woods over a period of 7 years has resulted in the identification of a number of settlements, two of which are almost certainly villa type buildings [33, 34] (WCAS 2003. 129). Extensive fieldwalking on either side of the river Avon at Bicton, Fordingbridge [109] and at North Street, Breamore [107] revealed dense pottery scatters and other features covering areas of >7 and 15ha respectively suggesting two possible village- type settlements (Light et al. 1995). At Yarnbury [5] RCHME surveying and field walking together with aerial photographic evidence suggests continuous occupation from before and throughout the Roman period, including a suspected rectangular masonry structure (RCHME, 1999, 116). Villa Type Buildings (47 sites - listed as V in Table 2) Nearly two thirds of the villa type buildings are located within 2km of main roads and rivers, particularly the Avon, and on the more fertile soils in the Avon river valley to the east and northeast of Sorviodunum. Their elevation above sea level ranges from 35m, alongside the river Avon, to the highest at 167m, at the Yarnbury Hillfort site [35]. Only 19 (40%) sites have received any form of intrusive investigation and the majority of these were excavated in the late 19th and early 20th centuries, and then only partially and very poorly by modern standards. The remainder of the sites is identified variously by artifacts found during field walking or trenching for services, aerial photography and geophysical investigations. The sites fall into two groups. In the first are 29 sites where intrusive 156 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 3. Numbers and Distances of Settlements from Major Roads and Rivers Villages (38), [8° 7 [20 | 16 51422773387 A | 42 25 i 66a BS OAS SEC oe Villas (47)=_]13 7128) |229 8/472) 298 Mle2NN/38 812s | 1/2300 149 15/3200 680) 43 ea ae 44 oe Cunetio or survey investigation has revealed significant evidence of wall foundations and/or building debris on the surface, while the second group of 18 sites only suggest possible ‘villa type’ buildings, having produced large quantities of pottery, the occasional coin, and only one type of building material, for example tile, or brick: they are marked V? in Table 2: Of the first group, seven villas have been excavated in sufficient detail to provide a plan of individual buildings (often more than one). These are West Dean [57], East Grimstead [45], Downton [40], Allington [28], Rockbourne [124], Holbury Farm, East Dean [118], and Grateley South [121] in Hampshire. At Netheravon Down [48], the building plan has been obtained through geophysical survey. The biggest villa excavation in the hinterland is that at West Park, Rockbourne [124], which was discovered in 1942 and dug from 1956-74 by A.T. Morley Hewitt with further limited work in 1978 and 1982 by I. P Horsey. The excavations showed that the villa developed steadily in size over 300 years and ultimately contained over 40 rooms in five buildings around three sides of a courtyard (RCHME 1983). Most of the building materials (Table 4 below) used were of relatively local origin. Two corn driers were excavated, and part of a briquetage tray suggesting the import of salt. The excavation evidence indicated decay and eventual collapse of the villa after c AD400 but not violent destruction. There is no evidence for any disturbance at the time of the barbarian incursion in AD367 and the subsequent disorder seen elsewhere in Britain. A hoard of over 7000 coins found beside the north range of buildings has a latest date of AD305 (ibid, 131). After Rockbourne the largest villa complex is that at West Dean [57] which is between the two other villa sites at East Grimstead [45] 2.5km to the west and Holbury Farm [118] 2.2km to the east. They lie in the valley of the River Dun, a tributary of the River Test which the Dun joins 4.9km east of Holbury. They are all located 5-6km south of the Roman road from Sorviodunum to Venta Belgarum. West Dean was first discovered in 1741 when a tessellated pavement was uncovered and removed but it was not excavated until 1846 and then again in 1871 resulting in the discovery of three buildings; two corridor villa buildings and an aisled hall of the 3rd and 4th century. East Grimstead was discovered in 1914 and excavated over the following 10 years by Heywood Sumner who uncovered an L-shaped corridor type villa with 14 rooms and two bath- houses surrounded on at least on 3 sides by a wall. A third bath-house (with only a cold room) dating to the mid 3rd to late 4th centuries was also discovered 40m to the east of the walled enclosure (Sumner 1924). In 1869, in Holbury Wood [119] 3km east of West Dean villa, the Rev. G. S. Master recovered a ‘cartload of pottery’ representing more than a thousand separate pots, some 30 coins (AD 268-340) and a small bronze? object (Master 1872a, 1872b). This was followed in 1870 by the excavation, 400m to the south, of a structure 20.3m long by 16.4m wide known as Holbury villa [118], which was subsequently classified as an aisled hall (Scott 1993, 83). While there may well have been two separate villas at Holbury, it is also possible that the aisled hall was part of a complex of buildings as at neighbouring West Dean. Downton villa [40], barely 300m from the river Avon, was discovered and initially excavated by Morley-Hewitt in 1953. This investigation was followed up in 1955-6 by Philip Rahtz (1963b) on behalf of the Ancient Monuments Inspectorate. A j | ! | | | | | | ] 4 SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 157 corridor villa with at least 7 rooms occupying an area of more than 5ha was found, which had been built in the late 3rd century and probably abandoned before the mid 4th century. A bath house was attached to the southeast corner of the main range of rooms. Decorated mosaic pavements were found in the >» principal room of the villa and in other rooms, while the corridor and two other rooms were paved with _red/off white tesserae. The rooms had painted wall , plaster and the fragments recovered showed that 16 or more colours had been used and that there had ‘been at least 2 re-coatings of some pieces. The villa had a large corn drying oven and outbuildings. Two villas northeast of Sorviodunum, both near the Roman road to Silchester, are to be found at Allington [28] and at Grateley South [121]. The | corridor villa at Allington was constructed in three phases from the late 3rd to the 4th century. The final building was approximately 30m long by 7m wide and divided into five rooms with a corridor on the east side. The villa at Grateley South, first recognized by J. P. Williams-Freeman in 1910, and further defined by subsequent investigations including aerial photography, was eventually | excavated between 1998 and 1999 by the Oxford University Institute of Archaeology as part of its | Danebury Roman Environs Project. A complex of four to five structures included an aisled hall, a strip house villa with flanking corridors and two crop _ processing buildings set in an area of some 15ha. The site originated as a Late Iron Age banjo enclosure followed early in the Roman period by timber buildings. From the end of the 3rd century and into the 4th century AD the settlement was upgraded with a series of masonry buildings arranged around three sides of an open area that had once been the space occupied by the banjo enclosure. First was the aisled hall, built on a virgin site, followed by the villa and crop processing buildings which replaced _ earlier timber structures. The settlement was largely dependent on grain production: a large double corn- drying oven in existence in the mid- 3rd century was _ replaced by two double corn-drying ovens, one in _ each crop processing building (Cunliffe and Poole 1998, 1999, 2008, 177; Cunliffe 2003, 346). The excavation provided useful information on crop processing in an area of rich soil ideal for ' growing corn (Cunliffe 2003, 347). Excavations at a number of other villa sites in the Danebury Environs _ Project area, but located just outside the Sorviodunum ' hinterland, also revealed aisled halls and other vernacular buildings typical of Hampshire Roman dwellings. Each site had developed to meet the needs of the economy and social life of the farming estate. Work at Grateley South has revealed continuity of occupation from the Iron Age to the end of the 4th or early 5th century AD. While continuous occupation of the site is difficult to prove, that the same pattern of occupation is seen over a wide area implies that systems of landholding and agrarian exploitation established by the Middle Iron Age continued little changed into, and possibly throughout, the Roman period (Cunliffe 1993, 219; Cunliffe and Poole 2008). A further 14 sites have been investigated in a more limited way. Sites close to the river Avon include, Figheldean [43]; two near Netheravon [47, 48], (just over a kilometre apart), Countess Services at Amesbury [29], Coffee Farm, Durnford [41]; two sites a little further south at Camp Hill [53] and Netheravon Road, Salisbury [49]. The remainder are to the south of Sorviodunum at Clarendon Park [39], Whitsbury [126] in Hampshire, and Sixpenny Handley [149] in Dorset, and to the east at Winterslow [59], and in Hampshire at Over Wallop [122] and Broughton [115] and to the northeast towards Andover 3.9 km north of Grateley South at Thruxton [125]. The latter villa is one of two in the general area of Thruxton the second being a further 3.4km to the northeast and outside the hinterland boundary. The information gained from these sites is sufficient to rate them as villas by the finding of walls and, in a number of cases, limestone roofing slabs (Table 4 below). In two cases some idea of the building plan has been obtained. At Amesbury [29] excavations in 2003 revealed part of a substantial masonry building at least 12m long and 9.7m wide, located only 140m from the river Avon. Limited pottery finds date to site the 3rd/4th century (WCAS 2003.152). On Netheravon Down [48] geophysical investigations in 1993 revealed the plan of a corridor villa approximately 40m long by 20m wide inside an 8ha polygonal enclosure of late Iron Age date (McOmish et al. 2002, fig 3.31). Some 198 coins were found over an area of 150m by 100m by a metal detectorist in 1987,* giving a date range of the early 2nd to late 4th century (see coin section below). At the remaining seven sites in the first group [30, 33, 34, 51, 55, 56, 61], no actual foundations have been found, although surface finds suggest that each one is almost certainly a villa by the quantity and/or quality of artifacts recovered. Of the second group of 18 sites, villa type buildings are suggested by large quantities of sherds and the occasional coin, although only one type of building material, e.g. tile 158 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE or brick. Two sites of the second group are reported as possible villa sites by well known authorities but lack supporting evidence, one by Leslie Grinsell in 1949 [60] and the other by Eleanor Scott in 1993 [52]. A number of sites show continuity from the Iron Age. The earliest appearance of masonry buildings was at West Park Rockbourne [124] in the late Ist or early 2nd century and by the late 3rd century all the villas had them. One villa, Downton [40], ceased to function around the mid 4th century, while the rest were occupied until the latter part of the 4th century or in some cases possibly even into the 5th century. Farmsteads and Unidentified Settlements (57 sites listed as S in Table 2) Very few farmstead sites have been excavated in the Sorviodunum hinterland and usually only in part due to the fact that they lie in the path of proposed construction works. In 1993 the construction of an oil pipeline booster station near Maddington Farm, Shrewton [96] revealed a farmstead and associated burials. The site proved to be part of a small farmstead of probable 3rd- 4th century date, comprising a small circular posthole structure, for agricultural rather than domestic purposes, with associated hearths and pits, adjacent to field boundary ditches which formed the focus for a small inhumation cemetery. Environmental evidence suggested a typical low- level mixed farming economy with both arable and pastoral elements. Sheep bones were the most numerous with cattle second (McKinley and Heaton 1996; McKinley 1999). At Eyewell Farm, Chilmark [72], several burials were disturbed during the construction of a new farmhouse in 1990. Subsequent investigations between 1991 and 1994 revealed occupation throughout the Roman period, including a grain drier, stone structures, ditches and pits and part of a linear later Romano-British inhumation cemetery which included several cist burials. The 16 coins recovered ranged from Hadrian to the House of Valentinian. A T-shaped grain drier used to process spelt wheat is of a well known and relatively long-lived type popular in the later Roman period. A similar drier, probably of 4th century date, was found nearby at Durrington Walls [15] (Wainwright 1971). Sheep, cattle, pig, horse and dog bones were present suggesting a mixed agricultural regime (Fitzpatrick and Crockett 1998). At Cockey Down, Laverstock/Clarendon Park [74], water pipeline excavations in 1989 and 1996 identified 3 phases of activity (Lovell 1999) for farming purposes from the Iron Age until the late Romano-British period when a corn drier was constructed and two ditches dug. Whether occupation was continuous could not be confirmed. During the late Iron Age/Early Romano-British period an enclosure ditch was in-filled and a round house and small sub-rectangular enclosure were constructed; at least 12 burials were made, possibly with others outside the area of excavation. Fieldwalking and other surveys at Grovely Woods (WCAS 2003.129) and Clarendon Park (James and Gerrard 2007) and at sites in Hampshire, close to the river Avon near Breamore and Fordingbridge (Light et al. 1995) indicate numerous farms or small hamlet-type settlements. At the present time it is not possible to a the type of dwelling at 30 of the sites listed in Table 2 Miscellaneous Sites - Industrial and Religious (10 sites listed as M in Table 2) Of the eight industrial sites six are New Forest (Hampshire) pottery kilns [140,142-6], two others are industrial but of differing type. At Armsley, Godshill [141], also in Hampshire on the edge of the river Avon, a small settlement operating over most of the Roman period was evidently concerned with iron smelting and other metallurgical activities and might have served as a base for itinerant craftsmen (Musty 1973). At Tisbury [104] the reported remains of a Romano-British beehive pit, created by the extraction of Chilmark stone, is the only record | to date of what must have been a thriving stone quarrying industry in the hinterland in the Roman period because it is found on so many Roman sites. Finally there are two sites of a religious nature; a possible temple site northwest of Friars Peak at Barford St Martin [103] and a possible amphitheatre or theatre site at Winterslow [105] excavated in 1959 (Vatcher 1963). Discussion Density and Distribution of Settlements (Table 3) SMR records from the three counties shown in Table 1 reveal an average settlement density of 0.53 sites/km* The figure for Wiltshire is somewhat lower than for Hampshire with 0.46 sites/km* compared with 0.75 sites/km? respectively. The difference is possibly due, in part, to the fact that a large area of the hinterland in Wiltshire lies on chalk compared to the more fertile areas in Hampshire. Salisbury i Sa =. — = ES SS eseaeagusSeSSouspesespemssseeersing sees. _ SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 159 | Plain has low settlement density of 0.26 sites/km? (McOmish et al. 2002, fig 4.1), while in Hampshire _ a figure of 0.6 was obtained in the Basingstoke area and 0.8 around Chalton (Millett 1990, Table 8.3). These differences, of course, also reflect the intensity of archaeological exploration across the region. Extensive field survey in Clarendon Park, for example, revealed 13 potential sites in 16km/ giving a | density of 0.8 (James and Gerrard 2007, fig 14), while _ Millett found that settlement density data (available in 1990) derived from literature searches or SMR records seriously underestimates likely real values. Field surveys undertaken in southern Britain gave a reasonably consistent set of figures with an average of 0.8 sites/km? (Millett 1990, 183-4, Table 8.3). This suggests that the hinterland figure of 0.53 sites/km7? is _ likely to be an underestimate of the true settlement density. Even so, taking 0.53 sites/km’ and using population figures per site suggested by Millett? gives a potential population of between 12,000 and 30,000 people in the hinterland of Sorviodunum quite apart from the inhabitants of the town itself (James 2002, n. 10). The proportion of the hinterland population | that would have travelled to market at Sorviodunum, or taken advantage of it indirectly is open to | question. Much debate has taken place regarding the | importance of villas, villages and farmsteads, their relationship with neighbouring settlements and their possible role as re-distributors of goods and services (Condron 1995). A link has been drawn between the prosperity of a given town and the number of villas in its hinterland (Branigan 1987). The present study found that previous researchers have consistently underestimated the number of villas in the Salisbury Plain area (eg McComish et al 2002, 104-6; Cunliffe 1993, 254, and Fig 7.1), especially in the Avon valley to the north of Sorviodunum. The level of ‘wealth’ in _ Settlements was variable (see below). By the first half | ofthe 4th century, Romanization in large parts of the hinterland resulted in many settlements exhibiting modest levels of wealth. Having money and/or goods to trade/barter would have created a thriving economy and attracted people to the markets of _ Sorviodunum and also to local centres that may have existed at sites such as Stockton Earthworks [21] and Woodyates [148]. Distances of villages and villas from known major roads listed in Table 2, reveal that 21% of the villages and 28% of the villas lay within 1km of the major roads (see Table 3). These figures rise to 42% and 49% respectively when navigable rivers are taken into account. Nearly half of villages and villas were within 2 km of a road, rising to around two-thirds with navigable rivers taken into account. These figures are not surprising given the high concentration of major roads and rivers in the hinterland. Roads and rivers centred on Sorviodunum would have provided an effective communication network between the town and its hinterland. The 38 definite and possible villages lay within 5km of a major road or navigable river, while for villas the figure is 94%. Comparison of villa data from the Sorviodunum hinterland with that for other areas in Roman Britain is shown in Table 3. Studies from the home counties around London (Sheldon et al. 1993; Green et al 1997); around Cunetio (Hostetter and Flusche 1997) and from across the Wessex region generally (Walters 2001), are in broad agreement with between 25-33% of the villas lying within lkm of a major road and 38-66% within 2km. The Sorviodunum data agrees well with that from the Cunetio hinterland at 1, 2 and 5km. Comparison of the figures for villages with those from villas in the Sorviodunum hinterland also shows broad agreement from 1-Skm. The number of villa type buildings in the Sorviodunum hinterland is impressive with 47 possible sites with an average distance of 12km. This compares with 24 possible villa sites around Cunetio whose average distance is 10.1km (Hostetter and Flusche 1997, 64-5). In the Cunetio hinterland 63% of villas are between 5km and 10km from the town, with a marked fall-off beyond 10km. Villas around Sorviodunum show a different fall-off pattern to those around Cunetio with 28% lying between 5km and 10km, a further 28% between 10km and 15km and 38% between 15km and 19.2km. Villas so far discovered in the region out to 10km are almost identical:16 for Sorviodunum and 17 for Cunetio. A plot of villa density against distance shows the density around Cunetio peaking at between 8km and 10km while for Sorviodunum the peak occurs further out at between 12km and 14km. Higher villa density further out from Sorviodunum is indicative of a substantial town acting as a market centre carrying out administrative functions (Hodder and Millett 1980; Millett 1990, 189-95). Gregson (1988, 24) suggests that a more gradual fall-off away from towns may indicate a preference for placing villas away from urban influence. Another significant recent discovery is the number of villa type buildings identified in the Avon valley north of Amesbury. Over 8km stretching up into Salisbury Plain eight possible sites have been located. Sites beside the Avon to the south of Amesbury are few but this — Table 4: Materials Used in Building Construction Site eee Walle et] RSE | TQuoins|Floor =|" “|Roof” |) | |Hypocaust "(2 a General Plaster | Glass Ceramic Pilae |Channel ae oe! Britannia 13, 211 eR 190-1; Scott 1993, WZ 8 [Amesbury [29]|K. 9 <|oe sXe a aS ee IRE eS 2003005 2 a Lee rene artin (Beach’syBarniy|! ciel ogi |e tie ae OES [ina aT lf ak ORR A pal lec 2007 ei [38] Arch Con 1996; James and Gerrard, 2007 [Downton\ [40] |X RSS TS! |X Xe IS a REX CS ee Xe Box || Beane WANES 8 oe [Durnford [4174] X 2 SS, CS a ECS Ae Box ae (AINE stead [45 FS, GS South [121] Box? 188, 344-51; Cunliffe and Poole 2008 (Holbury: [119] XW See XOX SMe Re Sree: = RC amas ees Pe a |e IN stereo anisiae ae I A 2000b House [47] PeS 294; 45, 198; 47, 538; 91,161; 94,148-53 Re | |) eee Road [49] 258; Grinsell 1957, 102; Scott 1993, WZ 146 ee ee) ae [122] PHFCAS 17, 60-3; Scott 1993, HA 71 a 14 a Farley [51 124 CS Archaeologia 22, 49; Royal Arch Inst 1951, 241-5 lca ED Pee De 53 X XP ax X lees x F and C? | Hoare 1821; WANHM 13, 33-5; 22, 243-50 eee boned ne i elon I a 57 | 59 ames 2002 Key would likely change with a programme of field walking either side of the river AC Ashlar Chalk towards Sorviodunum. a pe ores A plot of the density of Sorviodunum villages with regard to distance shows CSceGhilmnarlaStone a peak at between 14km and 16km, slightly further out than the villas, in other F Flint words about halfway between Sorviodunum and its neighbouring towns. FS Ferringous Sandstone GS Green Sandstone Pa gape aeigna Gere Settlement Development PH Portland Hexagonal Settlements grew steadily during the Roman period often on former Iron Age PL Purbeck Limestone sites and reached a peak in both numbers and size in the late 3rd to early 4th PS Portland Stone century (Table 2). Figure 2 shows a greater density of ‘villages’ lying to the west RC Rammed Chalk of the Avon than to the east. This compliments recent CBA research findings RS_ Red Sandstone . wan: : : SSuStonie: SlahsrarSlates on rural settlement in Roman Britain which found that in the Salisbury TS Tisbury Stone Plain/Danebury environs area while the majority of sites developed into a OL Oolite linear form of small enclosures and paddocks, there were local differences. To S_ Stone the west of the Avon settlements tended to be larger (>6ha) than those on the M_Forei : : 5 . orergu Marble more fertile soils to the east which developed into large farmsteads or small ‘hamlets’ (1-3ha in size) and into villas (Taylor 2007, 81-6). 60 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE —————— _ SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 161 | Construction Materials (Table 4) - Table 4 summarises the various construction materials used in the more substantial buildings - in the hinterland (and just outside) to illustrate the _ distances that these items would have travelled. The _ information in Table 4 was obtained from excavation and fieldwalking reports and the description of - materials, including stone identifications, is that _ of individual authors. Despite a lack of data from certain sites, it is clear that large quantities of building material were imported, sometimes over considerable distances, for example Purbeck and Portland stone from Dorset and Pennant sandstone from the Mendip Hills. It has been calculated that the weight and number of stone tiles required to roof Building 4 of the Grateley South villa, which had dimensions of 9m x 23m, was of the order of 40 tons and over 7000 slabs (Cunliffe and Poole 1999). Whether such loads were transported by mule, cart _ or by water has given rise to much speculation, | although river transport was the cheapest (Fulford 1975; Allen and Fulford 1996). Ceramic tiles for roofing, hypocausts and other constructions would _ also have created sizeable cargoes. Relatively low cost, but bulky, LHS stamped ceramic tiles manufactured _ at kilns in Minety on the Wiltshire/Gloucestershire _ border 60km north of Sorviodumun are attested in the ' hinterland (James 2002; Millett 1990, 176-8; Darvill 1979). Allen and Fulford (1996, 268) have described the movement of large consignments of stone from the Purbeck area for military purposes, but civilian _ consumption is also evident. Pottery (Table 5) _ Pottery is a key source of information regarding | the economy, relative wealth, the nature of regional | communications and the function of settlements in the hinterland. Unfortunately, within the study _ area only a small number of sites have produced large enough assemblages for detailed analysis. As a consequence comparisons are reliant on research conducted in areas just outside the hinterland, to the north on Salisbury Plain and to the north east in the environs of Danebury. Recent studies in these regions provide a picture of economic life probably much as it was in the northern part of the hinterland, especially on the chalk lands (McOmish et al. 2002; Fulford et al. 2006; Cunliffe and Poole 2008). The major local pottery industry was located in Savernake Forest c. 35km north of Sorviodunum and 4km south of Cunetio. Savernake Ware (SW) first appeared in the second quarter of the Ist century and continued possibly until about the end of the 2nd century (Hodder 1974; Timby 2001). Oxfordshire Ware including mortaria (OXW) was manufactured nearly 80km north east of Sorviodunum (Young 1977). About 40km southeast, near Poole, South East Dorset Black Burnished Ware (BB1) was produced from the early 2nd century until the late 4th century (Allen and Fulford 1996). Just south of Farnham, Surrey, some 70km from Sorviodunum, Alice Holt-Farnham potteries (AFW) produced only grey coarsewares throughout the Roman period. Finally, and partly within the hinterland itself, the New Forest pottery industry (NFW) covered an area of at least 60 sq km east of the River Avon (Fulford 1973; 1975; Swan1984) and lay 15.9 km south east of Sorviodunum. The New Forest industry manufactured mortaria and slipped wares from AD c.260 until the beginning of the last quarter of the 4th century (Tyers 1996). More localised pottery kilns producing mainly coarsewares were located near Swindon, Westbury and Warminster, while other settlements in the hinterland produced pottery for limited local use as at Allen’s Farm, Rockbourne and Spring Pond, Martin (Sumner 1914). Imported wares were restricted mainly to the early Roman period. The most widely distributed ware in the hinterland was terra sigillata or samian ware from south, central, and east Gaul, which was imported to Britain between c. AD 40-225 (Tyers 1996). Finewares from other continental potteries were imported in much smaller quantities. Detailed research into the marketing areas of BB1, AFW, NFW, and OXW pottery has been summarized by Fulford (1989, fig. 5). To provide a broader picture of the distribution of pottery in the hinterland from the major southern British industries and also continental imports, Table 5 summarizes the information from sites with reliable data. For the purpose of comparison, results from Fulford’s study of Salisbury Plain are also provided (Fulford et al. 2006). Several significant points emerge from Table 5. With the exception of two sites, BB1, OXW and NFW are present at all sites; samian is also found at all sites, albeit in relatively small quantities, especially those sites in existence in the early Romano-British period; SW pottery is found on the majority of sites in the northern half of the hinterland and also on Salisbury Plain, while AFH pottery is found in very small quantities at only three sites.° Percentages of OXW and NFW wares are equal at Butterfield Down and Figheldean, while OXW predominates further to the north on Salisbury Plain. To the south, mainly New Forest 162 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 5: Evidence of Pottery Industries, Artifacts and Agriculture Spelt, Sheep/goat 55%, barley cattle 36%, pig 4%, horse 3% <3 2 x es ~~ 8 D 2 op Vv Ln v : Jeseisiaigis |€ 8 = | E : 2 = SSElSIEIEIS [= |e r g g : g elgieigigisizi2 je |g E é 2 E a A\S|AlAal<|olzlz |A |= Ss S < 4 Bower X} |X/X X|/<5 |Ist- |3 nails, cleat, brooch no]}1 Rotary,1 not not recorded WANHM Chalke 2nd Beehive recorded 83, 1-49 6,7 Greensand ce X|X|X|/X} |X|X]6.7 |3rd- |Numerous Cu alloy, iron |X early |and over 900 coins inc Sth {small hoard of 8 gold; 28 Millstone frags of metal working (Greensand) Downton |X xX X}un Numerous Cu alloy and Villa [40] iron objects xX X| |X|X]7.4 |3rd- |Many nails, 10 iron and 9 bronze objects inc 8 coins, a spoon, 2 bracelets, and 2 knife blades Eyewell |X X| |X|X}]<5 |{Ist- |Very small nos iron X |3 Rotary Spelt, em- Farm [72] 4th |objects (Greensand) |mer wheat, barle X| |X|X] |X|/X]<5 |1st- |Large nos iron objects, 4th |mainly nails, 1 Cu Alloy 1 millstone X|X|X}X}/X|X|X}<5 |1st- |Objects include 50 coins, 4th |bronze encased casket, iron tripod candle holder, brooches, pins, rings and bracelets knives, shears, trowel, stylus and a spoon bit sandstones X|X} |X] |X|X {un |1st- |715 coins and hoard of X |Numerous Sheep 42% cattle RCHME, early |7715, very large nos of pieces of 33% 1983 Sth |domestic, agricultural and Saddle and carpentry tools, 7 styli, flat quern and many luxury items i stones . . Xx X X no Sheep/goat 59%, cattle 34%, pig 6% cattle in LRB; pig, horse, and dog 3 Rotary (2 Nearly twice as Greensand many sheep as cattle throughout whole period; pig, sandstone), horse and dog 1 Saddle and present Shrewton |X X| |X|X1]7.0 |3rd- |Large number of nails; 7 Rotary, WANHM [96] 4th |cleats, awl, knife, 2 Cu 3 made of 1/2 |89, 44-72 Alloy Greensand cattle; horse and pig present ».4 X Xx Ist- |Large nos of iron and not Sheep 46%, cattle 4th |Cu alloy objects inc over recorded |51%, pig 2.8% 1200 coins, 3 styli, spoons, Venus figurine, part of a balance arm X| |X/X/X/X/X17.0 |1st- | 25 coins, a dozen Cu Al- early |loy, and 28 iron plus nails Sth Hinterland Edge Beach’s X| |X|X/X/X}X Ist- {5 coins, nos of small Cu Barn 4th _|alloy trinkets, cleats, and nails X|X|X|X} |X}]X]3.0 | Ist- [15 coins, < 5 Cu alloy and 4th {5 iron objects (apart from nails and cleats) Coombe |X|X/X|/X] |X|]X]13.0{1st- |13 coins, <5 Cu alloy, and Down early |5 iron objects (apart from South Sth | nails and cleats Notes: Wealth Category 3 Rotary (of which 2 are Greensand) Cunliffee 1/2 }and Poole 2000a Large nos Spelt, 3 times as many rotaryand |hulled6 |sheep as cattle saddle stones |row barley mainly sreensand Many pieces | Spelt, of quartz barley conglomer- ate querns® Many pieces | Spelt, sheep 35%, cattle |2 |Fulford et of Greensand|emmer, 40%, pig 7%, horse al 2006 querns> barle 18% 1. Very poor (barely above subsistence) a. From Forest of Dean 2. Fairly poor b. From Vale of Pewsey 3. Modestly wealthy 4. Very wealthy ae sheep 22%, cattle Fulford et 57%, pig 12%. 2/3 |al 2006; Horse 9% Harding sheep 64%, cattle Fulford et 19%, pig 6%, horse 18% SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 163 fine and coarseware pottery was consumed, although - the Avon Valley field walking survey found only small quantities of New Forest finewares with up to 5% at just one site and much lower levels elsewhere (Light et al. 1995, 78). Table 5 lists the ratio of fine to coarsewares where calculation is possible.’ For early Romano-British - rural sites less than 5% of assemblages are finewares. _ Attwo later Romano-British sites the ratio is 7%. At - Beach’s Barn villa on Salisbury Plain at the northern _ edge of the hinterland the ratio is 9%. At the Grateley South villa, in the later Romano-British period the ratio is under 5% with the finewares comprising almost exclusively NFW and OXW (Cunliffe and Poole 2008, 114, table 5). These figures can be compared with sites in Oxfordshire and Warwickshire (Henig and Booth 2000, fig 6.11; Evans 2001, 32). In those counties, the ratio for early Romano-British rural sites ranges between 1% and 5%, while in the later part of the period it rises to between 13% and 18%. For early Romano-British towns and villas the values are between 7% and 20% while in the later period they range between 13% and 30%. Overall, these figures show a higher fine to coarseware ratio than that observed in the Sorviodunum hinterland, which could be due to several factors including wealth, status and communications. At the Downton villa, with the exception of samian ware, the finewares appear to be almost solely NFW, while around 16% of the pottery excavated was BB1. A high proportion of the pottery from Crystal Hollow appears to be NFW and proto NFW (T. Steptoe pers. comm.), although the Avon Valley field survey indicates that despite close proximity to the NFW kilns certain rural settlements were too poor to purchase fineware pottery in any quantity (Light et al. 1995, 78). In the south-west of the hinterland BB1 dominates with relatively small quantities of NFW and OXW. For example, at Bower Chalke nearly 73% of the assemblage is BB1 (Allen and Fulford 1996, 244), while in the north-east, at Grateley South, BB1 in the Late Iron Age represents over 45% of the wares but reduces markedly to 12.1% in the Roman period (Cunliffe and Poole 2008, 113). Study of BB1 finds shows a distribution corridor from the production centre along the Roman road known as The Portway to Sorviodunum and on to London (Allen and Fulford 1996, 239-44, fig 8). This distribution was maintained despite competition more locally from NFW (:bid., 266). While the widespread distribution of BB1 was mutually exclusive to that of NFW, paradoxically the BB] industry appears to have been unable to supply certain sites which lay only a few kilometers from the factories. This aspect finds a direct parallel with the observations on NFW pottery finds in the Avon Valley described above. The overall impression obtained from the frequency of wares on individual sites is that trade was carried out over considerable distances very likely in wagon-trains with other commodities, such as foodstuffs and salt (ibid, 268). Pottery may have been marketed from local centres such as Sorviodunum, or sold by itinerant traders travelling by road or river. Recent survey of areas of the New Forest around known kiln sites has produced evidence for settlement and agricultural activity (Berkshire Archaeological Services 2000) indicating that potters lived alongside their kilns. The sheer volume of pottery manufactured, calculated as being in excess of that produced in the medieval period, would have necessitated full time employment (Dark and Dark 1997, 132-3) and the recent discovery of settlement evidence alongside the kiln sites suggests that continuous manufacture is a distinct possibility. Evidence drawn from across the hinterland and from just outside reveals a network of long distance trading of pottery after AD 350 up to at least AD 400, although the New Forest industry appears to have gradually reduced production during the second half of the 4th century. A possible reason for this decline may have been the inability of poor people to buy wares in the face of increasing taxation in the late Roman period (Fulford 1975, 134-6). A further postulation is that a loss of customers could have resulted from increased ‘tribal’ tensions between civitates (Laycock 2008, 158-62). Coins (Table 6) Coin evidence can be used to gain general insights into the wealth, economy and longevity of settlements in the hinterland. Table 6 lists coin numbers for sites where at least 30 identified coins have been found according to the 21 periods commonly used for analyzing Romano-British coin finds. Earlier work on Sorviodunum coin finds concluded that the coin loss profile for Sorviodunum was comparable to that for other small towns in western Britain but not with rural settlements (James 2002, 16).° Butterfield Down and Sorviodunum both have significant percentages of Period 21 (AD 338-402) coins, 9% and 7% respectively, a strong indicator that life in these two settlements continued to the end of and probably beyond the Roman withdrawal from Britain. Sites with smaller percentages of Period 21 164 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 6: Coin Data Down [18, 48 eta: Sete Down [2 Ditches [3 Stockton Earth- works [21 coins are the villages of Netheravon Down, Stockton Earthworks and Woodyates and the villa sites at Holbury and Rockbourne. By contrast the village at Hamshill Ditches and the villas at Grateley South and East Grimstead have little or no evidence of either Periods 20 and 21, although the number of identifiable coins from these sites is relatively few at 36, 38, and 44 respectively (Table 6). Just north of the hinterland, 37 coins from three sites on Salisbury Plain with Period 20 and 21 coins show a similar percentage to coin-lists from Sorviodunum and Butterfield Down (Fulford et al. 2006, tables 5.5 and 5.9). At Suddern Farm 12% of the 25 coins are from period 21 with 68% of the coins falling in the years 348-402 (Cunliffe and Poole 2000a, 115). Elsewhere in the hinterland small numbers of coins from Downton villa and Durrington Walls/Stonehenge (Moorhead 2001, table 1) appear to show the early demise of the former site and late activity at the latter settlement, which is close to the village at Butterfield Down. A number of Late Roman coin hoards (post- dating AD 350) have been found in Wiltshire (Draper 2006, 32-3, table 2) and 30% of these lie within or on the edge of the hinterland. Whatever the circumstances of deposition, it is clear that material wealth was still available to some people, most probably landowning farmers up to and perhaps beyond the cessation of widespread coin use in the early 5th century. Some hoards may have been deposited for ritual reasons hinting at the persistence of Romano-British religious practice into the early Sth century (zbid., 33). oS) tw iS) - [vs — Wo Reference aR fain fleck a Rawlings and Fitzpatrick 1996 166 |SM1961.11, 1970.66, 1973.12 256 |Algar, DJ (pers aig 364-378 _— NR | diel Poole 2008 re Br — i) lon) Other metal objects (Table 5) Metal finds can be a useful indicator of settlement wealth according to the sophistication and range of items found. The wealthiest sites reveal luxury and exotic objects, yet assigning a degree of wealth is subjective because so few sites have been extensively excavated to modern standards.’ A brief summary of the range of metal artifacts is given in Table 5, from which it can be seen that the majority of assemblages are rural in character containing tools for agricultural and domestic purposes but also jewellery and trinkets. At Bower Chalke, Eyewell Farm and Shrewton the number of finds is small while at Durrington Walls a slightly larger number and range of items suggest more prosperous occupation. The finds from Suddern Farm are similar to those from Durrington Walls, while Butterfield Down was certainly wealthier, yielding 8 gold solidi and copper alloy finds such as armlets, finger rings, brooches, a probe or pin, spoons, key ring and fragments of figurines and a sceptre-head (Rawlings and Fitzpatrick 1996). The Downton villa produced very few finds but the site was short lived. At Woodyates a large and varied iron and copper-alloy assemblage indicates a prosperous roadside settlement (Pitt Rivers 1892). A similar picture emerges from the Stockton Earthworks settlement beside the Roman road to the Mendips, where a significant metalware assemblage, mostly excavated in 1923 by Nan Kivall (1926), includes nearly 300 coins, two styli, and many brooches, bangles, rings, and knives suggesting a level of prosperity not too dissimilar to Woodyates. | SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 165 ‘Both roadside settlements appear to have been more - prosperous than Butterfield Down. The Rockbourne villa was evidently very wealthy with many iron and copper alloy items including seven styli, a seal box and silver objects. The villa also produced -evidence of smithing of iron and copper-alloy. The villa at Grateley South did not produced anything like the range and sophistication of the objects from -Rockbourne, although other finds indicate a wealthy settlement (Cunliffe and Poole 2008). | Comparing the results of the present study with those from recent excavations on Salisbury Plain | just outside the hinterland at Chisenbury Warren, | Coombe Down and Beach’s Barn, enables a hierarchy -of wealth to emerge (Table 5). Fulford et al. (2006, 211) found parallels in the range and abundance of both ferrous and non-ferrous assemblages from ‘Suddern Farm, Coombe Down and Boscombe Down “West (Richardson 1951). Overall, Table 5 shows that the degree of prosperity ranges from very poor (level 1) at Chisenbury Warren, and Bower Chalke; through slightly better off (level 2) at Butterfield -Down, Durrington Walls and Coombe Down; to fairly prosperous (level 3) at Grateley South and to very wealthy (level 4) at Rockbourne. Where evidence is insufficient, the wealth level of a few settlements in Table 5 has been rated as either one of two levels. The Rockbourne villa not only exhibits the highest level of prosperity but also marked Romanization from an early date (RCHME 1983, 141) in contrast to most other sites where this process occurs later in the Roman period. Agriculture (Table 5) The Roman occupation took place in the context of much longer term change in agrarian practice. Agricultural intensification in many places changed the balance of crop species and introduced new methods of cultivation, horticulture and hay making. During the Iron Age a predominance of sheep was the norm. In the early Roman period, however, lowland Britain experienced a halt or even a decline in arable production, while the technology of storage and distribution was enhanced. In the later Roman period the impact was quite different, resulting in some places with the culmination of a process of agrarian change of much longer duration than the Roman presence itself (Jones 1989, 134). Are these wider processes observable in the hinterland? A summary of the available evidence is given in Table 5. Spelt wheat and barley with emmer wheat appear at two sites with bread wheat at one - the villa at Downton beside the river Avon in an area with mixed riverine soils suited to the growing of this crop (Rahtz 1963b, 303; Dark and Dark 1997, 110). At Grateley South wealth was largely dependent on grain production. In the mid-3rd century one double corn drier was increased to two when two corn processing buildings were constructed each with its own double drier. Careful analysis of the plant remains showed that in each pair of ovens the left-hand one was used to heat sprouted spelt wheat to prepare it for the malting process and beer making, while the right-hand oven was raised to a greater heat probably to parch the spikelets of spelt to make the glumes brittle so that they could more easily be separated from the grain. Cleaned and fully threshed wheat was perhaps stored in sacks in the attic of one of the buildings in the villa complex; the processing was sophisticated and well ordered (Cunliffe and Poole 2008, 177). The Salisbury Plain evidence shows that the range of cereal crops conforms to other Iron Age and Romano-British sites in Wessex with a predominance of spelt and barley and the almost total absence of emmer. The significant number of corn driers found in the hinterland, and just outside, suggests that at these sites the processing of grain was greater than the amount needed for daily domestic consumption (Fulford et al 2006 , 217) suggesting the production of a surplus, perhaps for the payment of taxes, and rent, or for marketing or conversion into other commodities. Three main animals provided food in the Roman period namely cattle, sheep/goats and pigs although at most sites pig bones are much less abundant. Throughout the Romano-British period there was a trend towards increasing numbers of cattle, at the expense of sheep, which was most marked at forts, towns and villa sites, while native settlements appear to have continued the Iron Age pattern of a predominance of sheep (King 1989, 53-4). Variations can occur due to local soil/pasture conditions, for example sheep not being suited to wet low lying sites because of the likelihood of infection by liver fluke. Sites with pasture close to rivers tend to have greater numbers of cattle bones and sheep are best suited to well-drained downland sites (Dark and Dark 1997, 112). Table 5 shows that, where figures are available, cattle and sheep/goat were the major species present at all sites but with varying percentages. Limited amounts of pig, horse and dog bones were present at most sites. The later Romano-British sites at Butterfield Down, Durrington and Shrewton, together with the much longer-lived sites at Grateley South and Suddern Farm all had more sheep than 166 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE. cattle. At Butterfield Down sheep appear to have been kept to maturity for their wool and milk while the age of cattle suggests they were kept for meat (Rawlings and Fitzpatrick 1996; Fulford et al. 2006, 212). The longer-lived village plus villa settlement site at Netheravon Down, close to the river Avon, had more sheep in the early Roman period which changed to a regime with more cattle in the later Roman occupation (Graham and Newman 1993, 10). At Rockbourne villa, of 600 or so animal bones identified, sheep and cattle accounted for 42% and 33% respectively. At Grateley South the former Iron Age ‘banjo’ site had nearly twice as many sheep as cattle throughout the whole Romano-British period. The high proportion of sheep at this site suggest an economy based on the Iron Age tradition (Cunliffe and Poole 2008, 164). At Woodyates roadside settlement there were slightly more cattle than sheep over the whole Roman period (Pitt Rivers 1892). Results from just outside the hinterland on Salisbury Plain ranged from a high percentage of sheep at Chisenbury Warren to slightly more cattle than sheep at the higher status site of Coombe Down, only 2km away (Fulford et al. 2006) and a relatively increased frequency of cattle bones over sheep/goats at the ‘villa’ site of Beach’s Barn barely 3km from Chisenbury Warren (Harding 2007, 88). At Rotherley, just outside the southwest edge of the hinterland, excavations of a 2ha farming settlement in 1880 by Pitt Rivers found a 12m diameter circular house within a circular enclosure or stockade. Nearby were 2 rectangular corn stores and probably other farm buildings (Pitt Rivers 1888; Hawkes 1947). Animal bone evidence showed 43% sheep and 32% cattle, a similar result to Rockbourne villa despite the large disparity in ‘wealth’ between the two settlements. Animal bone evidence from sites excavated so far paints a complex picture suggesting a number of possible factors determining animal husbandry. Amongst these are status or level of wealth; agricultural opportunities resulting from soil conditions; tenanting or ownership of land; and continuity, or otherwise, of occupation from the Iron Age. On the basis of evidence accumulated so far it would be difficult to predict what the animal husbandry is on any, as yet, unexcavated site either in or just outside the hinterland. Evidence from Chisenbury Warren and Rotherley indicates poor agricultural communities with little sign of any significant wealth. This contrasts with the evidence from Stockton Earthworks, 16km to the east, and Butterfield Down, 9km to the north of Sorviodunum, which were both more prosperous villages. Fulford et al. (2006, 210-11) compared evidence from Suddern Farm and Coombe Down with that found at Boscombe Down West (Richardson 1951). All are located on chalk at sites of former Iron Age bivallate enclosures. Despite the lack of published faunal remains and reliable quantitative data from Boscombe Down West there are parallels in the range and abundance of metalwork and in the ratio of fineware to coarseware pottery. Burial and religion The preferred burial rite in late Iron Age Britain was inhumation (Whimster 1981), which continued, with some modifications, throughout the Roman period when cremation was also practised (Philpott 1991). Roman burial evidence in the hinterland was, until recently, confined to individual graves and poorly recorded cemeteries excavated some years ago. The recent find of five cemeteries with a total of c.300 burials in Amesbury, probably associated with the village found at Butterfield Down [2], has significantly increased the overall numbers. These burials, all from the late Roman period, are mostly inhumations with two cremations and are one of the most significant cemetery finds associated with — a village type settlement discovered so farin Roman _ Britain. Many burials had grave goods, one was contained in a limestone (probably from Chilmark) coffin and consisted of a richly adorned female with a child in her arms and evidence of expensive, imported, footware slippers. Complete analysis of this significant and nationally important group of burials is expected to take some time (A. Fitzpatrick pers. comm.). The cemeteries lie to the south west of Butterfield Down village, beyond a Bronze Age linear ditch, and between the village and the river Avon. Until the recent Amesbury discoveries the largest cemetery excavated was at Winterbourne Down [95] with 37 cremations and 14 inhumations with smaller cemeteries at Maddington Farm [96] (9 burials), Eyewell Farm [72] (7 burials) and at Boscombe Down West [1] (which contained a cemetery of at least 15 inhumations) and unknown numbers at Yarnbury [5] and Teffont Evias Quarry [22]'° (Foster 2001, 165-6). Apart from Winterbourne Down these cemeteries © comprised mostly inhumations. Some burials were coffinless, others were placed in stone-lined cists, or coffins of wood, stone or lead. About 30% of single burials had grave goods but this figure dropped to 17% amongst those burials from cemeteries (ibid., SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 167 - 169), however publication of the Butterfield Down cemeteries data may well change these figures. -Hobnails from footware are common and occur in both male and female burials. The introduction of -hobnailed footware is Roman; however its inclusion ‘in graves appears to be a largely rural practice in southern England (Philpott 1991, 71). Most coffin burials are 3rd and 4th century but uncoffined ones occurred throughout the Roman period, as did cremations. The site at Winterbourne Down has the largest recorded number of cremations and the majority of them were associated with beakers or jars, many burnt, and many had grave goods (Foster 2001, 173). By the end of the 4th century unfurnished, extended inhumations were almost universal and practised by both Christians and Pagans (Philpott 1991, 226). There are, as yet, no securely dated 5th century burials from the area and the question of burial practice, continuity, and siting of cemeteries is still very much open (Eagles 2001). Evidence for religion normally comes from finding sites of temples or shrines, votive offerings to the gods and from certain apparel. In the hinterland such evidence is very sparse with a miniature stone altar excavated from a well at Stockton Earthworks [21] in 1923 (Nan Kivell 1926 pl. v, fig 11); and the discovery at Winterslow of a crescent shaped earthwork tentatively identified as the remains of a religious theatre (Robinson 2001, 158). Christianity became the official Roman religion in AD 313; however tangible evidence for Christianity in the hinterland is very limited. At the Butterfield Down cemeteries [2] there is some hint of Christian burials particularly as two of them contain coins with the Chi-Rho symbol. In other parts of Britain medieval churches have been found on or near the sites of former Roman villas and in the hinterland there is close association of Romano-British material with the 8" century Anglo-Saxon church at Britford, the 11" century church at Winterslow and recently a number of Romano-British fieldwalking finds were made by the author near the 12th-century church in Stratford sub Castle, less than 0.5km from the suburbs of Sorviodunum. Land Ownership Resolving land ownership and tenurial relations among rural settlements is very difficult from archaeological evidence alone. However recent work on Salisbury Plain suggests different types of ownership by the different character of finds at each site with a possible link between the nucleated village of Chisenbury Warren and a villa in the Avon valley and independent status for another settlement close by at Coombe Down (Fulford et al. 2006, 214). At Coombe Down there is evidence of settlement continuity from the Iron Age through to the Anglo- Saxon period while Chisenbury Warren appears to have folded at the end of the 4th century AD. The origin and development of Chisenbury Warren village may have been linked to the land owner’s desire to exploit particular agricultural produce markets for the Roman State administration. When the Roman requirements ceased the village died. The Salisbury Plain study provides a strong argument for a variety of tenurial arrangements (ibid., 217-8).!! Clearer evidence of the exact relationship between village and villa may well emerge over time but current evidence supports the view that nucleated settlements on Salisbury Plain, such as Chisenbury Warren, formed part of larger villa estates (ibid., 209). There are a number of sites in the hinterland where a villa and village are in close proximity. Location by geophysical survey of a villa type building inside the 8ha Romano-British village settlement at Netheravon Down [18, 48], located beside the river Avon, suggests a strong villa- village link (McOmish et al., 2002, fig 3.31), while a similar situation is suggested by the presence of a possible villa inside Yarnbury Hillfort [5, 35] village settlement (RCHME 1999, 116-8). In addition, the excavators of the Suddern Farm settlement [112, 123] found archaeological evidence to suggest a possible ‘villa type’ building nearby (Cunliffe and Poole 2000a, 202). A similar close relationship was discovered during the detailed field survey in Clarendon Park between Great Gilberts Copse villa and a linear village at Home Copse [12, 39] located only 0.6km apart (James and Gerrard 2007, fig 14). Other possible tenurial links might have existed between Hamshill Ditches village and a villa 200m to the south [3, 34] and also the possible village in Castle Ditches at Whitsbury (Hants) and a villa site 0.5km to the south [114, 126]. Finally it has been suggested that a number of the settlement sites in the lower Avon Valley around Fordingbridge may have been on lands forming part of the Rockbourne villa estate [124](Light et al., 1995, 78). Certainly, archaeological evidence analysed so far from the excavations at the Crystal Hollow village settlement [110] 5km south east of the villa would suggest a similar poor level of wealth to that at Chisenbury Warren. 168 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Sorviodunum Since the comprehensive archaeological review carried out 8 years ago (James 2002), a number of further discoveries have taken place in the Roman ‘small town’ and are detailed in the Appendix and summarized in Table Al. The two most significant discoveries are part of a large defensive ditch and evidence for three or four more buildings including a lst century AD structure possibly associated with industrial activities. The former discovery suggests a possible association with a fort or a mansio. There is, at present, architectural evidence for at least five definite buildings (Figure Al) and anomalies seen on the resistivity plot from the Albertine Cottage site could represent the remains of a further structure. In addition, aerial photographs suggest further structures (Moffat 2005). The picture gradually emerging from limited excavations and geophysical work is that Sorviodunum was a medium size ‘small town’ in existence throughout the whole Roman period. Evidence from the hinterland settlements suggests it would have acted as an important trading and communications hub, well served by a network of roads and the river Avon, and having a population density large enough to require the presence of government officials to administer tax collection. Clearly much more archaeological work needs to be done to establish further details of the town’s history, its public and private buildings and whether there was a fort, during the early Roman occupation, and a harbour for boats transporting goods and people up and down the river Avon both from outside the hinterland and from the many sites inside located close to the river (Figure 2). The 5th and 6th Centuries Esmonde Cleary (2004, 425) has pointed out that between AD350 and AD450 a huge range of highly visible archaeological sites and materials cease to exist and that so many of these go into decline at about the same time in the second half of the 4th century particularly towns, villas, mosaics and pottery. Late Roman provincial culture of the elite appears to be progressively ceasing. It is suggested here that it is that the archaeology, for a variety of reasons, declined both quantitatively and qualitatively while at the same time there is no evidence that there was a decline in the ability of the State to extract surplus. It would appear that the elite were unable or perhaps unwilling to maintain the traditional means of status display. It has also been suggested recently that there was a re-emergence of pre-Roman tribal enmities towards the end of the Roman military occupation that resulted in disruption of goods production and markets. This was exacerbated when the Roman army withdrew at the beginning of the 5th century (Laycock 2008). Much debate, sometimes heated, occurs over the exact events following the Roman military withdrawal at the beginning of the 5th century. This ranges from the belief that life continued, at least for quite some time, virtually unaffected by the departure of the Roman presence (Dark 2000), to a more or less dramatic collapse of society in the face of the Germanic invasions (Esmonde Cleary 1989; Faulkner and Reece 2002; Wickham 2005). Recently a more ‘half way’ approach has been taken which suggests that of the four Roman Provinces in Britain, created at the end of the 3rd century, only Britannia Prima survived intact as an entity for possibly up to 200 years after AD 400 (White 2007). Faulkner and Reece (2002, 75) argue that evidence from archaeological excavations between 1969 and 1996 when examined using properly defined terms, analyzed statistically, and interpreted in context, show the Roman Britain way of life ended inc.AD _ 400 having long been in decline, which began in the later 3rd century and fast gathered pace after the mid 4th century. White (2007, 195-207) on the other hand draws on a broader range of evidence including the results of recent, more comprehensive open area excavations in larger towns to show that life in them continued, but with a gradually changing style of living and some sort of ‘official’ management or administration. As a result of this research the boundary between the provinces of Britannia Prima and Maxima Caesariensis (ibid., 98, fig 38) is shown to pass near to Sorviodunum (Figure 3). It is worth noting that other researchers have positioned the Britannia Prima eastern boundary so that it encompasses the area thought to have been occupied by the Belgae (zbid., 39, fig. 11, 1, 2 and 3). Whichever is the case a large part of the hinterland could have become ‘Saxonised’ early on while part, to the west, remained ‘British’ in Britannia Prima. So what is the archaeological evidence? Very few comprehensive area excavations have been ~ undertaken in the hinterland. Consequently the evidence is of a very fragmentary nature, partly from limited excavated finds, and partly from historical documents and place-name studies (White 2007; Eagles 2001; 2004). Eagles (2001, 215) states that SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 169 : Dorchester & yr @ principal towns ..--°” Civitas boundary Britannia Prima boundary Fig. 3 Britannia Prima Boundary (White 2007) in Relation to Sorviodunum and its Hinterland there is no evidence to suggest that any part of Wiltshire or Dorset was in Saxon hands by the middle of the 5th century and that a grave (VI) probably of the third quarter of the 5th century in a cemetery at Winterbourne Gunner may have belonged to an initial group of foederati. The appearance and spread of new material culture in the late 5th and in the 6th centuries has until recently been restricted to sites at Winterbourne Gunner, Petersfinger, Harnham Meadows, Charlton, and Alderbury (zbid., 206) and from metal detecting and field walking in the Avon valley at Breamore and Fordingbridge (Light et al., 1995, 79). More recently, archaeological evaluations by Wessex Archaeology at Countess East, Amesbury, 170 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE | (WCAS 2003, 152) revealed four ‘Anglo-Saxon’ Sunken Featured Buildings (SFB) and artifacts suggesting a 5th- to 8th-century settlement close to the river Avon; at Harnham Meadows evidence of an early Anglo-Saxon settlement only 300m north of the 19th-century find of 64 graves by Akerman in 1853 (Jackson 1854) and in 2007/8 the discovery of an SFB in Wilton and another at the site of the new Park and Ride facility at Petersfinger (WCAS pers comm). At Coombe Down Enford, just outside the hinterland boundary on the Salisbury Plain, hand made organic and stamped pottery was found (Fulford et al., 2006, 121). Despite these more recent finds, which suggest that evidence for occupation in the Anglo-Saxon period may have been more widespread than previously believed, there is very little detailed archaeological evidence to be able to draw firm conclusions. There are hints of boundaries, probably of the civitates, at certain places in the hinterland. In the Roman period part of the Durotriges eastern boundary with the Belgae is likely to have followed the line of the east-facing Bokerley Dyke (Eagles 2001, 213) and in the post Roman period the New Forest appears to have been part ofa Jutish kingdom in south Hampshire whose western limit is unknown until the late 7th century (Yorke 1989, 92). Another part of the Durotriges eastern boundary is suggested by the Old English place name of Teffont (‘Spring on the boundary’), which is supported by the contrasting dates, 6th and 7th centuries, and nature of the Anglo-Saxon burials found to the east and west of the settlement. Possibly the western area may have been part of a different territory from the east and remained in British hands until the 7th century when it finally became subject to Saxon conquest (Eagles 2001, 213). The Anglo-Saxon Chronicle recounts that the conquest by Cerdic and Cynric of the Avon Valley and the Salisbury area was ultimately achieved following a victory on Charford in the early 6th century’’ and it also states that Cynric defeated the Britons at Salisbury later on in the 6th century (AD 552). Charford (‘Cerdic’s Ford’) is about 12 km south of Sorviodunum and on the Hampshire side of the present county boundary with Wiltshire; it may once of lain in the northern limit of the territory of the Jutes in south Hampshire where it met the land of the Wilsaete — the dwellers of the Wylye recorded in the Chronicle in AD 800 and 878 (ibid., 204). Dark (2000, 148) argues that there is an apparent cultural division at the Wansdyke, which coincides closely with the limits of Iron Age Dobunnic/Durotrigan coin distributions. This may suggest that in the late 5th century, the Durotriges controlled their pre-Roman lands and the Dobunni ruled in what is now Dorset and Somerset implying that the late Roman civitas of the Belgae had collapsed, but it is uncertain whether this extended so far west. The archaeological evidence available suggests that the area to the west and south of the Wylye valley was possibly part of Britannia Prima while that to the north and east was outside and as a consequence fell more rapidly under ‘Saxor’ influence, particularly if the Belgae civitas had collapsed. It leaves the status of Sorviodunum and especially the part played by the Old Sarum hillfort as crucial. It may be that the hillfort resumed its former function with what remained of the population of Sorviodunum possibly withdrawing into it.!3 Whether it became part of Saxon territory by the later 5th century is open to debate. However the presence of Anglo-Saxon cremation burials in southeast Wiltshire predates Cynric’s supposed victory over the Britons at Old Sarum in 552 by at least 50 years (Draper 2006, 37). This suggests a blending of British and Anglo- Saxon culture in eastern Wiltshire with Sorviodunum and part of its hinterland becoming a buffer zone between the Atrebates and the Durotriges civitates with the Anglo-Saxons fighting on behalf of the Atrebates state or an Anglo-Saxon successor to the Atrebates state (:bid., 38). Eagles (2004) suggests Saxon acquisition of civitas Durotrigum was piecemeal and drawn out and that from a 5th century nucleus around Old Sarum Anglo-Saxon culture slowly spread westwards. Conclusions The hinterland study has examined an area of nearly 1200 km? around Sorviodunum, and established a clearer picture of the different types of settlement and their range of prosperity/status. It has also confirmed the important role that Sorviodunum would have played both as a market trading and goods re-distribution centre for the people in the surrounding rural countryside and also as an administrative centre for collecting taxes. The study enabled a number of specific conclusions to be drawn:- 1. Analysis of SMR data has revealed an overall settlement density of 0.53 sites per km? comparable with the average density found elsewhere in Roman Britain. However the | SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 171 information is drawn from SMR databases, which have been found to consistently underestimate the true density of around 0.8 sites per km? found in intensely surveyed areas. In one such intensely surveyed area in the hinterland, Clarendon Park, it reaches a figure of 0.8 sites per km’. A calculation of the likely population in the hinterland using the lower figure of 0.53 suggests that there could have been, at the height of its prosperity, between 12,000 and 30,000 people, though this does not include those living in Sorviodunum itself. The hinterland has a large number of definite and possible villa-type sites, with 47 discovered so far, which is nearly twice that recorded around the neighbouring ‘small town’ of Cunetio. The number of villas around Sorviodunum has a different fall-off pattern from that around Cunetio. Around Cunetio they peak at between 8km and 10km while for Sorviodunum the peak occurs further out at between 12km and 14km. The pattern of higher villa density occurring further out in the case of Sorviodunum is, according to some, indicative of a substantial town acting as a market centre and carrying out administrative functions for the Roman authorities as well (Hodder and Millett 1980; Millett 1990, 189-95). The recent discovery of a high concentration of villa type buildings in the Avon valley to the north of Amesbury (one per km over a distance of 8km) suggests that a programme of field walking beside the river Avon to the south, in the direction of Sorviodunum, could find evidence of more. The hinterland had at least 38 possible village type settlements, many quite substantial in size and some, such as Stockton Earthworks and Woodyates, located alongside major roads, possibly functioning as local trading/market centres. A plot of village numbers with distance shows a peak between 14km and 16km, slightly further out than the villa sites. This is perhaps not unexpected since it occurs at around the ‘halfway distance’ between Sorviodunum and its neighbouring towns. The distances of the villages and villas from known major roads were examined revealing that nearly half lay within 2km of the major roads. These figures rise to around two thirds when navigable rivers are also taken into Us account. It would appear that the Romano- British settlements were located strategically to make good use of the available main routes of communication for access both to their neighbours and to the markets. There was a wide diversity in wealth and possibly in tenurial status of the numerous settlement sites. This ranged from very wealthy villas to some quite poor villages and farmsteads but with the majority of the sites having a prosperity lying somewhere in between. Evidence from sites on Salisbury Plain, just outside the hinterland, suggests there was no single overarching landlord, such as the State, but a number of different landlords. The present study shows potentially similar examples of tenurial relationships across the hinterland and a much larger range of prosperity than on Salisbury Plain. Overall the hinterland had a thriving mixed agricultural and industrial economy which appears to have reached its highest level of ‘wealth creation’ in the later Roman period. Limited agricultural evidence suggests a broadly similar mixed farming economy was in operation in the hinterland to that elsewhere in Wessex. Spelt wheat and barley were the main crops and sheep rearing predominated particularly at those sites on chalk; however there was a wide variation in animal husbandry from site to site. The high incidence of corn driers suggests a surplus of crops was available for paying taxes or rent to a landlord and would have enabled some degree of marketing or bartering. The nationally important discovery of a large number of burials outside the village at Butterfield Down will, once analyzed fully, provide a valuable insight into the health and dietary regime of one of the hinterland’s more prosperous settlements. Sorviodunum and its hinterland were at their most prosperous in the late 3rd and early 4th centuries. There is evidence in some areas of the hinterland for a decline towards the mid 4th century with the apparent fall into disuse of the villa at Downton and the decline of the pottery industry in the New Forest in the second half of the 4th century. However elsewhere life appears to have continued until the end of the 4th century and at a number of locations, where chaff tempered ware has been found, possibly well into the 5th century AD. 172 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 8. Small-scale excavations and field surveys in Sorviodunum carried out since the earlier study, reported in 2002, have revealed more buildings, including one of Ist-century date; a large defensive ditch; and the possible site of a harbour beside the river Avon. The fact that Sorviodunum was at the centre of a comprehensive road and river communications network is undoubtedly one reason why it would have developed into an important market/trading town during the Roman period. It would have been a strategic marketing and redistribution centre for the various types of basic and luxury goods sought by the population in the hinterland and beyond. From the significant number of people living in both Sorviodunum and the hinterland it would also have been a centre for collecting taxes. 9. Inthe later Roman period Sorviodunum appears to have been located on or near the border between the Romano-British provinces of Britannia Prima and Maxima Caesariensis. The archaeological and historical evidence suggests that by the end of the 4th-century Sorviodunum and a large part of the hinterland lay in a buffer zone between the Atrebates and the Durotriges civitates. This would have had some considerable influence upon the prosperity of the area and on its longer-term future in the Anglo-Saxon period. Much more archaeological evidence is needed, particularly from the late Roman and early Anglo-Saxon periods, before a clearer picture is obtained. In Summary The density of settlement with its large number of villages and villa sites suggests the hinterland countryside was an intensely worked agricultural landscape by the late 3rd, early 4th century AD. It was almost certainly producing an excess production of food for the Roman State either through direct sale, probably from villa estates, and/or from the population at large via taxes. The rapid decay of the economy at the end of the 4th century is most likely attributable to the loss of the Roman military market. The very comprehensive network of roads and rivers was actively exploited by the settlements many of which were founded in the Iron Age, some developing into villas, others remaining as relatively poor farmsteads. The larger settlements developed into thriving villages for various reasons such as roadside facilities for travellers, as homes for tenants working on villa estates and as centres for working on industrial processes and natural resources. The end of the Roman way of life may have been fairly rapid for those settlements whose livelihood relied upon producing food for the Roman State. For others it may have been much more slowly with their life style changing gradually as they adopted that of the Saxon incomers over a period of decades or much longer. Further Research Much more archaeological work needs to be done before the full extent of the role of Sorviodunum, and its relationship with the hinterland settlements, can be established. The main priority for the future should be extensive geophysical survey of the land on either side of the Roman road RR4a, where it passes through Sorviodunum, with selective trenching to determine the extent and role of other buildings. In addition, a number of sites in the hinterland would benefit from more detailed investigation to enable their precise function and degree of prosperity to be determined. These include Stockton Earthworks, Hamshill Ditches, and the village-villa complexes at Netheravon Down and Clarendon Park. Appendix, by Bill Moffat Excavations and Watching Briefs in Sorviodunum -2001 to 2008 (Table Al) Since the last review of the archaeological evidence was published (James 2002) there have been a number of housing developments within the area of Sorviodunum on either side of the Roman Road running through the town (Margary 1955, road RR4a) which have allowed small scale, but controlled recording. There have also been two small exploratory investigations using geophysics and testpitting. At 6 Castle Keep the footings trenches of a house extension revealed the chalk floor and one flint wall footing of a building at least 6m wide perpendicular to the Roman Road. The room or space also contained a small keyhole oven with a Purbeck limestone base. The structure was c. 25m away from the road and its full extent was not exposed. Adjacent to the building on its west side was a small ditch, -SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 73 ) Table Al: Summary of Excavations and Watching Briefs in Sorviodunum 2001-8 1 |2001/2)Sorvio- 13453185 | dunum Cottage (formerly Silverdale) {2 |2002 |9Castle {13443183 Keep industries ms 2004 |3 Castle 13513177 Keep 2004 |6 Castle Keep stead i “SI wai ‘113 12008 lOurne 113473196 Cottage on the same alignment and beyond that, a large pit. Stratification and material collected from deposits within the building and features suggest that they were of the same phase, being 3rd to 4th century AD in date. Beneath the chalk floor and on an east- ‘west alignment was a somewhat larger ditch, at least 2m in width and over 2m deep. Material from this feature is earlier, perhaps 2nd century. A full analysis of the finds and stratigraphy has not yet been completed as it is part of an ongoing project (Moffat forthcoming). | These results are similar to, though less complex than, those obtained at Sorviodunum Cottage (formerly ‘Silverdale’), a property in the Stratford Road on the northern corner of Castle Keep, and in almost exactly the same position relative to the Roman Road. An evaluation and watching brief | GR SU_|Nature of Finds Two small evaluation trenches in 2001 revealed evidence of a RB building overlying earlier ditches. Subsequent excavation of the new house footings in 2002 revealed 4 phases of RB activity, including an early soil sealed ditch, at least 3m wide on top of which were two distinct early RB phases of building construction. These included floors, 2 ovens, a substantial stone built well, 2.5m in diameter, and large quantities of Purbeck, Chilmark and varieties of other non local stone. Pottery was early and mainly coarseware with limited Two small test pits sited to investigate deposits disturbed by a new conservatory re- vealed 2 soil layers with a building demolition layer of rubble and stone sandwiched between. The uppermost soil horizon dated to the 3rd-4th centuries and the lowest to the Ilst-2nd centuries. Finds included fragments of box flue tiles and tegula, and mainly coarseware pottery from the N Oxfordshire, Savernake, NE, BB and local | Watching brief on an extension to the house revealed easternmost corner of an enclo- | Close sure ditch ca 5.5m wide which had filled up in the 2nd centur ee House sibly a building or enclosure 2003 |24 Shake- |13665186 |No archaeological features found, only one undateable RB sherd of a coarse, grog speare tempered fabric Close 1 2003 |Spindles |13383196 |Surface finds of oyster shells and RB sherds in rear garden resulted in a 1x1m testpit being dug which revealed 3rd-4th centuries NF and BB pottery, ceramic and lime- stone tile fragments, and robbed out remains of a flint and chalk wall ~0.8m wide 2003 |Rose- 13413204 |Small 1x1m testpit in rear garden revealed RB boundary ditch (at least 1m wide and bower 1m deep and running approx NE-SW) with small quantities of 3rd-4th centuries pottery and domestic rubbish Watching brief on a house extension. No Romano-British material was recovered. Excavations did not proceed below the rammed chalk surfaces of a medieval farm- 13463176 |House extension footings revealed chalk floor and one flint wall footing at least 6m wide, a small oven and 3rd-4th centeries finds. Underneath the floor was an earlier, perhaps 2nd century, ditch at least 2m wide and over 2m deep Watching brief on a house extension revealed pottery from two phases of RB activit | 2005 13483190 | Watching brief on a house extension revealed no material of Romano-British date. ) wood 12 |2005 {Albertine |13653202 |Fieldwork in the paddock behind Cottage revealed the line of RR4a and structures | Cottage adjacent to it. Excavation within one structure dated it to the Ist Century AD with (formerly Portway Cottage evidence for iron working in the vicinity Watching brief on extension and garage at the former Stratford Inn. No significant Romano-British material was recovered, although a single, shallow and undated NW- SE linear was recorded at the eastern corner of the new garage. quantities of finewares was carried out between 2001 and 2003 prior to the construction of a new house on the site of a demolished bungalow; just the initial phase of this work was reported earlier (James 2002). Excavation of the footings for the new house exposed a column of stratification 21 strata deep, the physical thickness of which averaged 1m. Because of the form of construction works, the watching brief and the stratigraphy were considerably simplified with several features recorded only as groups. The deposits relate to five main phases of activity, of which the lower four (and the bulk of the strata) were Romano-British. These comprised early cut features followed by a soil accumulation phase, then two phases of building. The earliest phase consisted of four large pits and a group of intercutting pits lying to the SW 174 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE. of a large ditch or channel, which was at least 3m broad and of unknown depth. Its orientation was difficult to determine but lay between the present line of the adjacent Stratford Road and north- south. Soils sealed these features. The deposits varied in thickness from 200 to 600mm. Their greater depth and compaction over the large ditch suggested that they constituted a deliberately dumped consolidation layer, within which some soil formation had taken place. The phases of building are considerably more complex. At their most simplified they relate to two distinct phases of construction, indicated by two sets of superimposed chalk floors. Associated with these were two ovens, several beamslots and pits and a substantial stone built well 2.5m in diameter. At the south-eastern extremity of the site the edges of more substantial floor makeup layers were discovered, suggesting that these phases represented the remains of service rooms and ancillary buildings belonging to a town house fronting onto the Roman Road. The pottery assemblage comprised a small proportion of imported and local finewares but predominately local coarsewares. On the basis of the fabric types, and vessel forms present, the pottery assemblage may be dated to the early Romano-British period. No characteristically later fabrics such as finewares from the Oxfordshire or New Forest industries or characteristically later vessel forms were recorded. It is possible that there was also a slightly earlier influence due to the presence of very dark grey coarse sandy fabrics which may represent the continued use of late Iron Age sandy fabrics in the area. As well as pottery the site produced a single fragment of blue-green glass, a fragment of worked bone and an entire quern from Chaterhouse on Mendip, which had been reused as a post pad. During the excavation it was particularly noticeable that the upper deposits contained large quantities of Purbeck and Chilmark stone, varieties of greensand and other non-local stones. Other assemblages have recently been collected from Beech Tree House in 1999 (and reported previously in James, 2002), Castle Close (2002), and The Glen (2004). While that from The Glen simply shows that two phases of Romano-British activity took place, and can provide no more precise dating, those from Beech Tree House and Castle Close result from larger scale construction works which have provided more data. The assemblage at Beech Tree House comprised two phases of activity, one in the lst-2nd centuries and one in the 2nd-3rd, starting no earlier than AD70 and continuing perhaps as late as AD270. Structural evidence was restricted to the first phase and consisted of two postholes and two pits, sealed in the later phase by a group of rough yard surfaces. Both phases are consistent with deposits resultant from activity in the backyard of a property. The significance of the assemblage lies in the comparatively rare presence of New Forest Ware, particularly in the near vicinity, and in the relative absence, as at Sorviodunum Cottage of mortaria, ceramic roof tiles and nails. In April 2002 a watching brief was carried out at Castle Close, a detached house on the northeastern side of Stratford Road. It exposed the easternmost corner of an enclosure ditch c. 5.5m wide. As far as can be determined this enclosure appears to have been rectilinear and on the same orientation as the Roman town. Deposits revealed inside the angle of this ditch suggest that it was the outer of a double ditch system, and pottery from the site suggested that they had silted up, or been filled in during the 2nd century. This evidence was interpreted as being indicative of a lst-century fort. In order to test this hypothesis, a geophysical survey was carried out within the grounds of Orchard House, over an area within which it was reasonable to expect that the enclosure ditch, or ditches, might appear. — The results of this survey, carried out in 2003, were negative in so much that they showed that the enclosure did not extend so far to the northwest. They did however reveal a further enigma, in the form of a large rectilinear anomaly apparently a large building or enclosure with an indication of an entrance on its short southern end. It is quite likely that this is part of the post-medieval farm of which Orchard House is the remnant; however it is on a very different alignment, inconsistent with the regularity shown in all other medieval and post-medieval properties in Stratford sub Castle. Investigations to its north and west produced no evidence of distinctive archaeological activity; however a small test pit dug immediately south of the structure in the rear garden of ‘Spindles’ bungalow produced significant quantities of Romano-British building material and domestic refuse, consistent with the upper layers exposed in recent works alongside RR4a (James unpublished data). ) In 2005 an exploratory investigation was carried out, under professional supervision, by the Spire Group of the University of the Third Age (U3A) to assess the suitability of the site and its environs as a base for a Community Archaeological SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 175 Projet (Moffat 2005). Shovelpitting, geophysical ' surveying and testpitting were conducted between the OS designated line of Roman road RR4a -and the modern Portway on the land behind Albertine Cottage. The shovelpitting technique turned out to be inappropriate because the level | Albertine Cottage | Fig. Al Archaeological Setting of Recent Finds in Sorviodunum of agricultural disturbance of the site was minimal and the archaeological remains are buried by soil accumulations commonly deeper than the usual depth of shovelpits. The geophysical surveying was successful in that for the first time it provided an accurate plot of a length of the roadway showing 176 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE ° that it lay ca 5m south of its previously plotted position and on a slightly differing alignment. It also showed a number of features including possible structures alongside the road. A single 5m x 2.5m testpit was excavated at the edge of the Roman road (RR4a). It exposed the details of road construction which was a layer of large-medium sized flint cobbles bedded in a coarse, hard, lime mortar and laid directly on the natural clay subsoil that separates it from the underlying natural Coombe Rock (a pulverised chalk deposit laid down during the last Ice Age). The coarse layer of cobbles and lime mortar was covered by a layer of finer flint gravels and softer, friable lime mortar. This accords with the road designs recorded by Stratton during his excavations in the 1960s and 70s (James 2002). The testpit also revealed evidence of a roadside structure separated from the road by what appeared to be a robber trench. The form of wall which was originally contained there is unclear although it is likely to have been stone and with a width of only 350mm it would have not been particularly substantial. The wall is likely to have supported a timber framed building whose precise dimensions are unclear; however the high resistivity around that area indicates a complex structure. Within the part of the structure excavated was a rudimentary cobble floor consisting largely of small to medium flint cobbles with a matrix of yellowish brown lime mortar. Pottery dates the building to the Ist century AD, and the small amount of finewares together with the presence of frequent lumps of iron slag within the earliest fills suggests that it was not a grand building. The comparatively early date of the pottery assemblage is of particular interest as the nearest building excavated previously was that conducted by Stratton in 1977 (James 2002), only 12m to the south on the opposite side of the road, which was of 3rd or early 4th-century date and exceptionally grand by comparison. In most urban environments one would normally expect an element of rubbish and debris from that building to have found its way into the deposits exposed in the testpit. Though inconclusive, the absence of late material may suggest that Strattons building was part of a mansio, or other public building within an enclosure that was by that time empty of other activity, and kept clean of rubbish. Thus the combination of fieldwalking, evaluation, excavation and watching briefs has continued to build up a picture of a large urban settlement alongside the Roman Road below the Old Sarum hillfort. While some preliminary work using air photographs has already been carried out (Corney 2001), to compliment it an earthwork survey, particularly over the Scheduled Area south of Castle Keep, would be of inestimable value. Restricted observations made during the winter of 2003/4 showed that AP parchmarks seen running perpendicular to the Roman Road adjacent to the bank of the river Avon could be easily traced on the ground and appeared as long, low mounds on which the vegetation was extremely thin. These blended into the riverbank, and were pierced by a deep cutting through which the stream adjacent to Castle Keep runs. To the southeast of the stream and within the line of the banks described above was a rectilinear depression, an occasional pond in which marsh grass grows all year around. It is possible that this area is where the Roman town had its harbour for boats using the river Avon as a water highway. There is, at present, some architectural evidence for at least five definite buildings (Figure Al) and anomalies seen on the Albertine Cottage resistivity plot could represent the remains of a further structure. In addition aerial photography pictures suggest there may well be quite a number more. Geophysical survey over a wider area could provide valuable data in determining the size density and planning of the town. Also over the last 40 years a large amount of artefactual material has been collected and published in a very piecemeal — fashion. A thorough collation and examination of all the material is essential. Acknowledgements The author is extremely grateful to the large number of people who provided assistance during the course of this study. These include the staff of the three County Archaeology Departments in particular Helena Cave-Penny (Wilts CC) who also produced Figure 2, Rachael Salter (Hants CC) and Claire Pinder (Dorset CC). Particularly useful were the discussions and encouragement received from numerous archaeological experts and the invaluable support provided by the Directors, Peter Saunders — and Adrian Green and their staff at the Salisbury and South Wilts Museum. SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW 177 Notes 1. Held by the Wiltshire, Hampshire and Dorset County Archaeological Departments. 2. Cold Kitchen Hill is classified as a ‘Temple’ site; however its coin loss pattern suggests a more complex story. 3. The numbers were those current at July 2008. A list of SMR sites is available from the three County Archaeological Departments. -4. A coin list is available from the Salisbury and South Wilts Museum where the coins are accessioned under ) entry SM 1987.83. 5. Taking the rural population density figures for the optimum period of the first half of the 4th century AD as between 20 and 50 persons per site (Millett 1990, | 183-6). | 6. Because of the difficulty experienced by some | excavators in positively identifying Alice Holt- Farnham pottery it could well have been present at many older excavations. 7. The ratios of fine to coarseware pottery have been calculated from the numbers of individual pots wherever possible. 8. See earlier paper on Sorviodunum (James 2002) and also | Reece 1991; 1993; and 1995. | 9. Where area excavations are large enough the density | of finds per square metre is used as a comparator with other sites and also provides an indicator of the relative wealth. 10. It is estimated that in excess of 30 burials have been found over a period of many years in the area of Teffont | Quarry. 11. Fulford et al. (2006, 217) point out that there is no direct evidence to distinguish between the types of owner, whether private or imperial, of the various estates, individually or collectively, but the potential relationship of villa and dependent village settlement is perhaps no different on Salisbury Plain than in another part of Lowland Britain where it is generally assumed | to be private. 12. Given as AD519, though this is considered by many | scholars to be unreliable 13. The earliest evidence of Anglo-Saxon occupation found so far is in the form of two possible 5th-century burials just outside Old Sarum (SMR 13SW400) anda 7th-century buckle found inside the hillfort (SM OS C36). References | | ALGAR, D. J. 1970. Rescue and Research Work in the Salisbury Area, 1969. VWANHM 65, 209 ALGAR, D. J. 1971. Rescue and Research Work in the | Salisbury Area, 1970. WANHM 66, 190-1 ‘ALLEN, J. R. L. and FULFORD, M. G. 1996. The Distribution of South-East Dorset Black Burnished Category 1 Pottery in South-West Britain. Britannia 27, 223-81 ANON. 1991. Excavation and Field Walking in Wiltshire 1989. VWANHM 84, 143 ARCH CON, 1996, Clarendon Park Salisbury, Archaeology, History and Ecology, Vol 1 (Text) and Vol 2 Illustrations and Gazetteer. Archaeological Consultancy, King Alfred’s College Winchester. Unpublished Report for English Heritage BERKSHIRE ARCHAEOLOGICAL SERVICES, 2000, Report on an Archaeological Earthwork Recording Survey in the Island Thorns Inclosure, Fritham, The New Forest, Hampshire. Unpublished Client Report BERSU, G. 1940. Excavations at Little Woodbury, Wilts. Proceedings of the Prehistoric Society 6, 107-10 BLAGG, T. E C. 2001, Roman Architectural Ornamentation in Britain. Oxford: Britsh Archaeological Reports 329 BRANIGAN, K. 1987, The Catuvellauni. Gloucester: Sutton CONDRON, E 1995, ‘When is a town not a town? ‘Small towns’ on the Nene and Welland in their context’, in A. E. Brown (ed.), Roman Small Towns in Eastern England and Beyond, 103-18. Oxford: Oxbow Monograph 52 CORNEY, M. 2001, ‘The Romano-British Nucleated Settlements of Wiltshire’, in P. Ellis (ed.), Roman Wiltshire and After, 5-38. Devizes: Wiltshire Archaeological and Natural History Society CRAWFORD, O. G. S. 1929, Air Photography for Archaeologists. Southampton: Ordnance Survey Professional Papers 12 CUNLIFFE, B. 1993, Wessex to AD 1000. London: Longman CUNLIFFE, B. 2003. Roman Danebury. Current Archaeology 188, 344-351 CUNLIFFE, B. and POOLE, C. 1998, The Danebury Environs Project: 2. Grateley South Excavation 1998, Interim Report. Oxford: Institute of Archaeology CUNLIFFE, B. and POOLE, C. 1999, The Danebury Environs Project: 3. Grateley South Excavation 1999. Interim Report. Oxford: Institute of Archaeology CUNLIFFE, B and POOLE, C. 2000a, The Danebury Environs Project. Vol 2 Pt 3. Suddern Farm, Middle Wallop Hants. Oxford: Institute of Archaeology Monograph 49 CUNLIFFE, B and POOLE, C. 2000b, The Danebury Environs Project. Vol 2 Pt 6. Houghton Down Stockbridge, Hants. Oxford: Institute of Archaeology Monograph 49 CUNLIFFE, B. and POOLE, C. 2008, Danebury Environs Programme. Vol 2 Pt 2. Grateley South, Grateley, Hants, 1998 and 1999. Oxford: English Heritage and Oxford School of Archaeology Monograph 71 DARK, K. 2000, Britain and the End of the Roman Empire. Stroud: Tempus DARK, K and DARK, P. 1997, The Landscape of Roman Britain. Stroud: Sutton DARVILL, T. 1979, ‘A Petrological Study of LHS and 178 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE TPF Stamped Tiles from the Cotswold Region’, in A. McWhirr (ed.), Roman Brick and Tile, 309-50. Oxford: British Archaeological Reports (International Series) 68, DRAPER, S. 2006, Landscape, Settlement and Society in Roman and Early Medieval Wiltshire. Oxford: British Archaeological Reports 419 EAGLES, B. 2001, ‘Anglo-Saxon Presence and Culture in Wiltshire AD c. 450 — c. 675’, in P. Ellis (ed.), Roman Wiltshire and After, 199-234. Devizes: Wiltshire Archaeological and Natural History Society EAGLES, B. 2004. Britons and Saxons on the Eastern Boundary of the Civitas Durotrigum. Britannia 35, 234-40 EDWARDES, H. S. W. 1927. Excavations at Armsley. Proceedings of the Hampshire Field Club and Archaeological Society 10,, 56-62 ESMONDE CLEARY, A. S. 1989, The Ending of Roman Britain. London: Routledge ESMONDE CLEARY, A. S. 2004, ‘Britain in the Fourth Century’, in M. Todd (ed.), A Companion to Roman Britain, 409-27. Oxford: Blackwell EVANS, J. 2001, ‘Material Approaches to the Identification of Different Romano-British Site Types’, in S. James and M. Millett (eds), Britons and Romans: Advancing an Archaeological Agenda, 26-35. York: Council for British Archaeology Research Report 125 FAULKNER, N and REECE, R. 2002. The Debate About the End: A Review of the Evidence and Methods. Archaeological fournal 159, 59-76 FITZPATRICK, A. P and CROCKETT, A. D, 1998. A Romano-British Settlement and Inhumation Cemetery at Eyewell Farm, Chilmark. WANHM 91, 11-33 FOSTER, A. 2001, ‘Romano-British Burials in Wiltshire’, in P. Ellis (ed.), Roman Wiltshire and After, 165-78. Devizes: Wiltshire Archaeological and Natural History Society FULFORD, M. G. 1973. New Forest Pottery. Britannia 4, 160-78 FULFORD, M. G. 1975, New Forest Pottery. Oxford: British Archaeological Reports 17 (2000 Edition) FULFORD, M. 1989, “The Economy of Roman Britain’, in M. Todd (ed.), Research on Roman Britain 1960-89, 175-201. London: Britannia Monograph 11 FULFORD, M. G, POWELL, R, ENTWISTLE, R and RAYMOND, F. 2006, Iron Age and Romano-British Settlements and Landscapes of Salisbury Plain. Salisbury: Wessex Archaeology Report 20 GRAHAM, A and NEWMAN, C. 1993. Recent Excavations of Iron Age and Romano-British Enclosures in the Avon Valley, Wiltshire. WANHM 86, 8-57 GREEN, D. SHELDON, H. HACKER, M. WOON, C. and ROWLINSON, H. 1997. The Distribution of Villas in Some South-eastern Counties: Some Preliminary Findings from a Survey. London Archaeologist 8:7, 187-95 GREEN, M. 1990. Ancient Roman Building and Medieval Earthworks Near Woodyates Manor, Proceedings of the Dorset Natural History and Archaeological Society VA) GREGSON, M. 1988, ‘The Villa as a Private Property’. in K. Branigan and D. Miles (eds), Villas Economies. Economic Aspects of Romano-British Villas, 211-33. Sheffield: University of Sheffield GRINSELL, L. V.1957. ‘Archaelogical Gazetteer’, in R. G. Pugh and E. Crittall (eds), Victoria County History of Wiltshire Vol. 1, part 1, 21-272, London: Oxford University Press HARDING, P. 2007. Two Possible Iron Age ‘banjo’ enclosures and a Romano-British villa and settlement at Beach’s Barn, Fittleton, Salisbury Plain. WANHM 100, 83-90 HAWKES, C. E C. 1947. Britons, Romans and Saxons Round Salisbury and in Cranborne Chase. Archaeological Fournal 104, 27-81 HENIG, M. and BOOTH, P. 2000, Roman Oxfordshire. Stroud: Sutton : HINGLEY, R. 1989, Rural Settlement in Roman Britain. London: Seaby HINGLEY, R. 1991, ‘The Romano-British Countryside: the significance of rural settlement forms’, in R. F J. Jones (ed.), Roman Britain: Recent Trends, 75-80. Sheffield: Sheffield University Press HOARE, Sir R. C. 1812, The Ancient History of Wiltshire. London: William Miller HOARE, Sir R.C. 1821. Auncient Wiltescire - The Roman Era. London: Lockington HODDER, I. 1974. The Distribution of Savernake Ware, WANHM 69, 67-84 HODDER, I and MILLETT, M. 1980. Romano-British Villas and Towns: A Systematic Analysis’. World Archaeology 12:1, 69-76 HOSTETTER, E. and FLUSCHE, L. 1997, ‘Villa Siting’, in E. Hostetter and T. N. Howe (eds), The Romano- British Villa at Castle Copse, Great Bedwyn, 61-68. Bloomington: Indiana University Press JACKSON, Rev J. E. 1854. Anglo-Saxon Cemetery at Harnham. WANH™M 1, 196-208 JAMES, D. J. 2002. Sorviodunum — A Review of the Archaeological Evidence. VWANHM 95, 1-26 JAMES, T. B. and GERRARD, C. 2007, Clarendon: Landscape of Kings. Logaston: Windgather Press JONES, M. 1989, ‘Agriculture in Roman Britain: The Dynamics of Change’, in M. Todd (ed.), Research on Roman Britain 1960-89, 127-34. London: Britannia Monograph 11 KING, A. C. 1989, ‘Villas and Animal Bones’ in K. Branigan and D. Miles (eds), The Economy of Romano-Bnitish Villas, 51-9. Sheffield: Department of Archaeolog and Prehistory KING, A. C. 2004, ‘Rural Settlement in Southern Britain’, in M. Todd (ed.),A Companion to Roman Britain, 349-70. Oxford: Blackwell LAYCOCK, S. 2008, Britannia the Failed State. Stroud: Tempus LIGHT, A, SCHOFIELD, A. J and SHENNAN, S. J. 1995. SETTLEMENT IN THE HINTERLAND OF SORVIODUNUM: A REVIEW The Middle Avon Valley Survey: A Study in Settlement History. Proceedings of the Hampshire Field Club and Archaeological Society 50, 43-102 LOVELL, J. 1999. Further Investigations of an Iron Age and Romano-British Farmstead on Cockey Down, near Salisbury. WANHM 92, 33-8 MARGARY, I. D. 1955, Roman Roads in Britain. London: Phoenix _ MASTER, Rev G. S. 1872a. Roman Remains Found at Holbury, near Dean. WANHM 13, 33-41 _ MASTER, Rev G. S. 1872b. On a Roman Building at Holbury near Dean. WANHM 13, 276-9 MCKLINLEY, J. I. 1999. Further Excavations of an Iron Age and Romano-British Enclosed Settlement at Figheldean near Netheravon. WANHM 92, 7-32 | MCKINLEY, J. I. and HEATON, M. 1996. A Romano- British Farmstead and Associated Burials at Maddington Farm, Shrewton. WANHM 89, 44-72 | MCOMISH, D., FIELD, D. and BROWN, G., 2002, The ) j | | / Field Archaeology of the Salisbury Plain Training Area. Swindon: English Heritage MILLETT, M. 1990, The Romanization of Britain. Cambridge: Cambridge University Press MILLETT, M. 2005. (Revised Edition). Roman Britain. London: English Heritage MOFFAT, W. 2005. Archaeological Investigations at Albertine Cottage, Stratford-sub-Castle, Wiltshire. Unpublished Client Report, Pathfinders Archaeological Reconnaissance MOORHEAD, T. S. N. 2001. ‘Roman Coin Finds from Wiltshire’, in P Ellis (ed.), Roman Wiltshire and After, 85-106. Devizes: Wiltshire Archaeological and Natural History Society _ MUSTY, J. W. G. 1973. The Roman Pond and Dipping | PITT RIVERS, A. 1892. Vol III Excavations in Bokerly Dyke Well at Armsley, Godshill, Hampshire. Proceedings of the Hampshire Field Club and Archaeological Society 30, 35-8 NAN KIVELL, R. de C. 1926. Objects Found During Excavations on the RB Site at Stockton Earthworks 1923, WANHM 43, 389-94 | PASSMORE, A. 1967, New Forest Pottery Kilns and Earthworks. Southampton: Cadnam | PHILPOTT, R. 1991, Burial Practices in Roman Britain. Oxford: British Archaeological Reports 219 PITT RIVERS, A. 1888. Vol II Excavations in Cranborne Chase near Rushmore, 1880-1888. Privately printed and Wansdyke, 1888-1891. Privately printed PITT RIVERS, A. 1898. Vol IV Excavations in Cranborne Chase near Rushmore, 1893-1896. Privately printed RAHTZ, P A. 1963a. An Excavation on Bokerly Dyke 1958. | Archaeological Journal 118, 65-99 /RAHTZ, P. A. 1963b. A Roman Villa at Downton. _ WANHM 58, 303-41 -RAHTZ, P A. 1990. Bower Chalke 1959: Excavations at Great Ditch Banks and Middle Chase Ditch. VANHM | 83, 1-49 /RAWLINGS, M. and FITZPATRICK, A. P. 1996. | | 179 Prehistoric and Romano-British Settlement at Butterfield Down, Amesbury. WANHM 839, 1-43 RCHME, 1979, Stonehenge and its Environs. Edinburgh: Edinburgh University Press RCHME, 1983. West Park Roman Villa Rockbourne, Hampshire. Archaeological Fournal 140, 129-50 RCHME, 1999, Bowden, M. (ed.), Unravelling the Landscape. Stroud: Tempus REECE, R. M. 1991, Roman Coins from 140 Sites in Britain. Cirencester: Cotswold Studies REECE, R. M. 1993. British Sites and their Roman Coins. Antiquity 67, 863-9 REECE, R. M. 1995. Site Finds in Roman Britain. Britannia 26, 179-206 RICHARDSON, K. M. 1951. The Excavation of Iron Age Villages on Boscombe Down West. WANHM 54, 123-68 RICHMOND, I. R. 1969, ‘The Plans of Roman Villas in Britain’, in A. L. E Rivet (ed.), The Roman Villa in Bnitain, 49-70. London: Keegan Paul ROBINSON, P. 2001, ‘Religion in Roman Wiltshire’, in P. Ellis (ed.), Roman Wiltshire and After, 147-64. Devizes: Wiltshire Archaeological and Natural History Society SCOTT, E. 1993, A Gazetteer of Roman Villas in Britain. Leicester: Leicester University Archaeological Research Centre Monographs | SHELDON, H., CORTI, G., GREEN, D. and TYERS, P. 1993. The Distribution of Villas in Kent, Surrey, and Sussex: Some Preliminary Findings from a Survey. London Archaeologist 7:2 40-6 SMITH, J. T. 1997, Roman Villas: A Study in Social Structure. London: Routledge SUMNER, H. 1914, Excavations on Rockbourne Down Hampshire. London: Chiswick Press SUMNER, H. 1924, Excavations at East Grimstead Wiltshire. London: Chiswick Press SUMNER, H. 1927, Excavations on New Forest Roman Pottery Sites. London: Chiswick Press SWAN, V. G. 1984, The Pottery Kilns of Roman Britain. London: Royal Commission on Historical Monuments Supplementary Series 5 TAYLOR, J. 2007, An atlas of Roman settlement in England. York: Council for British Archaeology Research Report 151 TIMBY, J. 2001, ‘A Reappraisal of Savernake Ware’, in P. Ellis (ed.), Roman Wiltshire and After, 73-84. Devizes: Wiltshire Archaeological and Natural History Society TYERS, P. 1996, Roman Pottery in Britain. London: Routledge (1999 Reprint) VATCHER, F de M. 1963. The Excavation of the Roman Earthworks at Winterslow, Wilts. Antiquaries Journal 43, 197-213 WAINWRIGHT, G. J 1971. The Excavation of Prehistoric and Romano-British Settlements Near Durrington Walls, Wiltshire, 1970. WANHM 66, 76-128 WALTERS, B. 2001, ‘A Perspective on the Social Order 180 of Roman Villas in Wiltshire’, in P. Ellis (ed.), Roman Wiltshire and After, 127-46. Devizes: Wiltshire Archaeological and Natural History Society WHITE, R. 2007, Britannia Prima: Britain’s Last Province. Stroud: Tempus WHIMSTER, R. 1981, Burial Practices in Iron Age Britain. Oxford: British Archaeological Reports 90 WICKHAM, C. 2005, Framing the Early Middle Ages. Oxford: Oxford University Press THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE WISE, J. R. 1863, The New Forest: Its History and Scenery. London: Gibbings YORKE, B. 1989, ‘The Jutes of Hampshire and Wight -and the Origins of Wessex’, in S. Bassett (ed.), The Origins of Anglo-Saxon Kingdoms, 84-96. Leicester: Leicester University Press YOUNG, C. J. 1977, Oxfordshire Roman Pottery. Oxford: British Archaeological Reports 43 Wiltshire Archaeological & Natural History Magazine, vol. 103 (2010), pp. 181-85 Medieval remains at Pennings Road & St Andrews Road, Tidworth by Fonathan Milward, Andrew Manning, Lorraine Mepham and Chris F. Stevens | _ Archaeological fieldwork within a small redevelopment area adjacent to Holy Trinity Parish Church in North Tidworth _ uncovered evidence from the early prehistoric to post-medieval period. Two distinct phases of early (1 1th—12th century) and | later (12th —14th century) medieval features and structures were identified divided by a property boundary maintained _ for a considerable time. Features either side of the boundary showed clear differences in the nature of activities. { | | Introduction | Thesite lies immediately east of Holy Trinity Church within the historic core of North Tidworth (NGR | 42350 14900)(Figure 1) and is bounded to the east _by Pennings Road (A338), the historically important | Salisbury to Oxford Road and by St Andrew’s Road. The underlying geology is weathered ‘coombe’ chalk bedrock, overlain along the eastern edge of the site | by calcareous fluvial gravel deposits (Geological : Survey of Great Britain 1:50000 Drift Sheets 282 Andover and Sheet 283). A shallow but pronounced north-south terrace ran through the excavation area, marking the floodplain boundary of the River _ Bourne, a winterbourne, east of Pennings Road. __ The redevelopment area had been previously occupied by a substantial 19th century rectory which was demolished and replaced by retail premises in ‘the 1970s. An evaluation of the site in 1995 (Wessex /Archaeology 1995) demonstrated that medieval | and post-medieval features survived within the ‘south-western corner of the site (an area ofc. 0.1 ha); these were excavated in 2006. A watching brief was carried out within the remainder of the site although |no further surviving archaeological features were MW i identified. ] Results The excavation revealed c. 70 features, mainly medieval boundary ditches, pits, quarries and post- holes indicative of pens/light structures, covering two broad phases of activity within the 11th—12th and 12th—14th centuries. Earlier artefacts recovered from later medieval contexts include 21 undiagnostic prehistoric worked flints, a sherd of Romano-British pottery and two sherds of early-middle Anglo-Saxon pottery. 11th — 12th century The earliest medieval activity was mainly at the eastern edge of the site. Post-holes ranging between 0.27m and 0.4m in diameter were associated with three short gully sections (104, 106 and 140). Two isolated features lay to the south-west, including an elongated pit (196) and a pit or possible ditch terminal (427), which extended beyond the southern boundary of the excavation. Only four features contained datable pottery, all of flint- tempered Kennet Valley type (Table 1) (post-hole 114, gully 140, pits 196 and 427) and all of these features contained very similar fills which could be distinguished from those of later date. | Wessex Archaeology, Portway House, Old Sarum Park, Salisbury SP4 6EB = ee TO 182 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Trowhrida HPOWDTMAGS Tidworth @ 11-12th century ~ O8VECr | 12-14th century Post-medieval Modern services 148980. Hedgeline 463 263- Structure 664 < 0 10m 47 | fea oe | Foundations of 19th century rectory - 00SEzr | Fig. 1 Site location and excavation area 12th-14th century The dominant feature of this period was an east—west boundary that extended beyond the limits of the excavation and ran perpendicular to the line of Pennings Road. This boundary separated two plots with evidence for markedly different activities. The boundary was defined by an intercutting sequence of shallow narrow ditches (456-8 and 460) typically 1.0-1.5m in width and 0.5m in depth, containing single sandy silt fills with moderate amounts of animal bone, pottery and ceramic building material. Pottery included both flint-tempered and chalk-/flint-tempered Kennet Valley wares, West Wiltshire type wares, Laverstock-type coarsewares and finewares, and sandy wares of Newbury C type (see Table 1). The Kennet Valley wares, part of a widespread regional ceramic tradition, could have been supplied from a very local source at or close to Ludgershall, less than 5km from Tidworth (Gerrard and King 2000, 195-6). Northern plot Features to the north of the boundary were predominantly chalk quarries and rubbish pits, separated from possible roadside properties by a north-south orientated boundary ditch 455 (0.75m wide, 0.50m deep) at the eastern edge of the MEDIEVAL REMAINS AT PENNINGS ROAD & ST ANDREWS ROAD, TIDWORTH 183 Table 1: Late Saxon and medieval pottery totals by fabric type Description No. sherds Weight Late Saxon flint-tempered ware (Powell et al. 2009, fabric F400) 2 18 Michelmersh-type ware (Mepham and Brown 2007, fabric E403) ] 30 Laverstock-type fineware (Musty et al. 2001, 139-40) 28 437 Laverstock-type coarseware (Musty et al. 2001, 140-1) 10 72 West Wiltshire type coarseware (Smith 1997, 21-2) 13 75 Kennet Valley ware A (flint-tempered) (Mepham 2000, fabric E441) 218 3143 Kennet valley ware B (chalk-/flint-tempered) (Mepham 2000, fabric E442) 140 1854 Newbury C ware (sandy) (Vince 1997) 6 34 TOTAL 418 5663 excavation area. At least 11 intercutting quarry pits were identified in the central area of the plot. These were 2.0-4.6m in diameter, capped by a 0.34m deep silty clay alluvial deposit (321). The largest, partially excavated feature was pit 313, with a diameter of at least 4m and a depth of at least 2.5m ascertained using an auger. Two further pits (311 and 312) c. 4.6m _ in diameter were comparatively shallow at 0.8m and 1.0m respectively. All the quarrying pits exhibited _ the same pattern of backfilling with a distinct ) sequence of fluvial deposition interspersed with turf formation and deposits of clean chalk and flint. A small group of intercutting pits was found in the north-west corner of the excavation area (168, 176, 230, 256, 258, 282, 285 and 462). The pits were 1.8-2.8m in diameter and 0.36—0.6m in depth and _ each contained a single fill with small quantities of animal bone, shell and pottery. Pit 282 produced the | largest group of datable pottery (50 sherds), including | Kennet Valley, West Wiltshire and Laverstock types. The group is clearly chronologically mixed, as it included both a probable tripod pitcher sherd of 12th century date and Laverstock-type fineware of 13th or _ early 14th century date. The fill of pit 168 was rich | in grains of free-threshing wheat and barley. | Southern plot Activity in the southern plot was distinctly pastoral in nature. A low lynchet, c. 6m wide, was evident i in the southern section of the excavation area, | directly sealing an early medieval pit or possible | ditch terminus (427). The crest of the lynchet (0.6m in height) lay on the top of the natural step | corresponding to the edge of the flood terrace and tailed off towards the west. The edge of the lynchet ‘partly corresponded with irregular and diffuse | patches of root disturbance in the natural geology | (group 463), indicating that it had developed against |a pre-existing hedgeline forming the corner of a paddock or a field. A very short section of ditch (group 668) at the western extent of the lynchet contained pottery and a fragment of lava quern and may have represented part of an enclosure. In the south-west corner, immediately to the south of the former hedgeline, was part of a small sill beam structure (664), at least 3.0m by 5.0m, which extended beyond the limits of excavation. The structure was defined by two sill beam slots (217 and 263) associated with small stake- and post-holes (207, 215 and 228). The structure contained a slight depression and internal stake-hole, which had silted up. A single sherd of 12th—13th century pottery was recovered from the depression. The line of the hedge continued to the east of the terrace step in the form of a shallow ditch 413, 1.6m wide and 0.54m in depth, terminating at its western end in a group of eight irregular intercutting pits (278, 288, 307, 322, 367, 369, 371 and 452) each 0.2-0.5m deep. Environmental samples from ten medieval pits demonstrate a low diversity of charred plant remains. Pit 168 was relatively rich in, and dominated by, grains of free-threshing wheat (Triticum aestivum/ turgidum sl.) followed by barley (Hordeum vulgare sl.) and lesser amounts of rye (Secale cereale) and possibly oats (Avena sp.; Table 2). The grain had been relatively well cleaned with no chaff such as rachis fragments present. Only a small number of seeds of larger seeded species, typical of arable fields, are present and other crop remains include pea (Pisum sativum) and bean (Vicia faba). Medieval records from Wiltshire indicate that free-threshing wheat and barley (including bere) were the most important local crops along with oats and some legumes (Harrison 1995). The results reflect this trend and the few grains of rye might indicate its occurrence as a weed or only as a minor crop. The relative paucity of weed seeds and rachises implies that crops arrived at the site and were stored as almost clean grain after threshing, winnowing and sieving had taken place. 184 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 2: Quantified charred plant remains from pit 168 Pit 168 Context 169 Sample 8 Volume (litres) 16 Species Name Common Name Hordeum vulgare (grains _ barley 35 indet) H, vulgare (grains hulled) 8 Triticum aestivum/turgidum free-threshing wheat 120 (grains) Secale cereale (grains) rye Sacielk Cereal indet. (grains) cereal 5 Cereal indet. (culm cereal 2 nodes) Corylus avellana hazelnut frag. 3 Agrostemma githago corncockle l Silene sp. campion ] Rumex sp. dock ] Malva sp. mallow 1 Vicia/Pisum sp. Vetches/pea 5 Vicia/Lathyrus sp. vetches/tare 15 Centaurea sp. knapweed, l cornflower Anthemis cotula stinking mayweed 3 Avena sp. oats 22 Bromus/Avena sp. brome grass/oats 2 The weed seeds provide only limited information on crop practices in the field. The presence of stinking mayweed (Anthemis cotula) indicates cultivation of heavier clay soils, while both wild mustard and runch are often more common on sandier soils. Generally free-threshing wheat tends to perform better on the heavier soils, while barley, along with legumes, performs better on the poorer soils. The samples can then be seen as reflective of the mixing of charred waste and grain from crops grown on varying soil types. Post-medieval and modern activity Following a substantial hiatus, evidence of early post-medieval activity was limited to a small number of isolated pits and two wells. One large elongated pit (134), 6.5m by 1.8m and 0.5m in depth, lay towards the north-east corner of the excavation. It contained two dark and humic silt fills, with a small quantity of 16th century pottery, animal bone and shell within the basal fill. One well (298), 1.5m in diameter and at least 1m deep, had a chalk block lined shaft with a later brick wellhead. The second, larger, brick-lined well (336) was 2.4m in diameter and at least 1.6m deep. It lay in the north-west corner of the excavation within the footprint of the later 19th rectory, with the remains of the southern wall footings evident along the northern edge of the excavation area. Discussion Relatively little archaeological investigation had previously taken place immediately around Holy Trinity Church. The excavation reported on here has thus made a useful contribution to our understanding of the medieval development of the town. While no pre-medieval features were found, a small number of Romano-British and early—middle Anglo-Saxon pottery sherds indicate earlier activity. Although Romano-British settlement in Tidworth is suspected, the only evidence so far is restricted to small pottery assemblages. Documentary evidence suggests an Anglo-Saxon date for the foundation of Tidworth (Ekwall 1991, 462) and recent excavation at the Matthew Housing Estate, less than 160m to the east of the site, has identified early—middle Anglo-Saxon features relating to settlement activity (Godden et al. 2002). Although the extent of the excavation was limited, a well-dated sequence of earlier medieval features (11th-13th century) with carefully zoned activity was revealed. This activity continued into the 14th century, before being apparently curtailed until the early post-medieval period. No firm evidence of domestic buildings was uncovered, although the undeveloped area adjacent to Pennings Road may contain roadside structures. The small number of 11th—12th century post- holes and gullies which form the earliest phase of medieval activity may reflect light structures or temporary stock pens likely to have been associated with nearby roadside structures. However, most surviving features date to the 12th—14th centuries, a period in which the site, and presumably a wider area along Pennings Road, appears to have been formally defined and developed. The pastoral nature of the activity in the southern half of the excavation area suggests that the east-west boundary may have marked the southern edge of the settlement of North Tidworth, separating a tenement property from pastoral land to the south, rather than defining two separate plots. To the north of the boundary, extensive quarrying indicates exploitation of the underlying chalk, presumably for building and possibly related to minor renovation of the adjacent church, though the major rebuilding of the present structure is believed to have taken place in the 15th century (Currie 1995). There is little surviving evidence of further activity until the post-medieval period. This hiatus broadly correlates with the arrival of bubonic plague MEDIEVAL REMAINS AT PENNINGS ROAD & ST ANDREWS ROAD, TIDWORTH 185 in England in the mid-14th century. Tidworth appears, from documentary evidence, to have been particularly badly affected, with the possible loss of all the recorded tenants (Croman 1991, 4). The long-term effects of this event may explain the limited evidence for activity until well into the post- medieval period. Acknowledgements The project was funded by J. S. Bloor Homes (Newbury) Limited and Wessex Archaeology wish to acknowledge the assistance of Robert Clark and Tom Cole. The project was monitored by Sue Farr and Helena Cave-Penney of the Wiltshire County Archaeology Service (WCAS) and their advice is gratefully acknowledged. The project was managed by Andrew Manning The excavation was directed by Jonathan Milward and assisted in the field by Ruth Panes, Hayley Clark, Neil Fitzpatrick, Paul Cooke, Oliver Good, Naomi Hall and Lee Newton. The subsequent watching brief was undertaken by Jamie Wright, Steve George, Steve Thompson and Phil Harding. Illustrations are by Will Foster and Rob Goller and the text was edited by Julie Gardiner. The full project archive is currently held at the offices of Wessex Archaeology at Old Sarum, Salisbury, Wiltshire under the project codes 60900-2. In due course the archive will be deposited with the Salisbury and South Wiltshire Museum. References CROMAN, D. J., 1991, A History of Tidworth and Tedworth House. London: Phillimore CURRIE, C.R.J., (ed.) 1995, Victoria County History of Wiltshire 15: Amesbury Hundred, Branch and Dole Hundred, 153-63. Oxford: University Press EKWALL, E., 1991, The Concise Oxford Dictionary of English Place-names (4th edition). Oxford: Clarendon Press GERRARD, C. and KING, R., 2000, The pottery, in Ellis, P (ed.), Ludgershall Castle: excavations by Peter Addyman 1964-1972, 181-200. Devizes: WANHS Monograph 2 GODDEN, D., HAMILTON-DYER S., LAIDLAW, M. and MEPHAM, L., 2002. Excavation of Saxon pits at Tidworth 1999. WANHM 95, 240-8 HARRISON, B., 1995. Field systems and demesne farming on the Wiltshire Estates of Saint Swithun’s Priory, Winchester, 1248-1340. Agricultural History Review 43:1, 1-18 MEPHAM, L., 2000, Enborne Street and Wheatlands Lane: medieval pottery, in Birbeck, V., Archaeological Investigations on the A34 Newbury Bypass, Berkshire/ Hampshire, 1991-7 (Technical Reports), 52-66. Salisbury: Wessex Archaeology MEPHAM, L. and BROWN, L., 2007. The Broughton to Timsbury Pipeline, part 1: A Late Saxon pottery kiln and the production centre at Michelmersh, Hampshire. Hampshire Studies 62, 35-68 MUSTY, J., with ALGAR, D., GERRARD, C. and HADLEY, J., 2001, Pottery, tile and brick, in Saunders, P (ed.), Salisbury and South Wiltshire Museum Medieval Catalogue Part 3, 132-212. Salisbury: Salisbury & South Wiltshire Museum POWELL, A.B., CHANDLER, J., GODDEN, D., MEPHAM, L., STEVENS, C and KNIGHT, S., 2009. Evidence for Late Saxon and medieval occupation near Salisbury Street, Amesbury. WANH™M. 102, 188-201 SMITH, R.W., 1997, Excavations at Emwell Street, Warminster: the early economy and environment of a Wiltshire market town. Salisbury: Wessex Archaeology VINCE, A.G., 1997, Excavations at Nos 143-5 Bartholomew Street, 1979, in Vince, A.G., Lobb, S.J., Richards, J.C. and Mepham, L., Excavations in Newbury 1979-1990, 7-85. Salisbury: Wessex Archaeology Report 13 WESSEX ARCHAEOLOGY, 1995, The Old Rectory Site, Pennings Road, North Tidworth, Wiltshire, Archaeological Evaluation. Salisbury: unpublished client report 38801 Wiltshire Archaeological & Natural History Magazine, vol. 103 (2010), pp. 186-256 The larger linear earthworks of Savernake by Graham Bathe! and foanna Ramsay?’ Field survey of 35 linear earthworks in the Savernake area revealed features dating from Bronze Age to modern times. In many cases field remains can be related to archival sources, which throw light on their original or adopted functions. Ancient alignments have frequently been re-employed for a variety of purposes over centuries or even millennia. A pre- occupation with the notion that the Wansdyke once extended beyond its current terminus has resulted in a profusion of fanciful claims as researchers have sought to join the dots, distracting attention from more prominent features which have been ignored because they are out-of-line or have their ditches facing the ‘wrong’ way. Current data suggests that any former extension of the Wansdyke could not have been lost through ploughing after 1130, and circumstantial evidence suggests that it never continued through the study area. Other features, currently assumed to form coherent structures, are shown to have sections with widely disparate morphologies, where their attribution is tenuous. Suggestions for resolving these issues, and avenues for further research are identified. cross-valley dyke, a braided track, an inclosure bank, another cross-valley dyke, and part of an old road. Fixation with the Wansdyke and its course at the regional scale has diverted attention from sound observation at the local level, an essential pre- requisite if linkages amongst supposed fragments are to be claimed. Pre-occupation with the Wansdyke, and to a certain extent also with dykes mentioned in Anglo-Saxon charters, has tended to detract from major features on differing alignments. This paper seeks to complement existing research by presenting a catalogue of the larger linear earthworks in Savernake placed within a landscape context. ‘Tt is very difficult to be sure of anything’ (Crawford 1953) Introduction Maps depicting archaeological features in Savernake resemble star charts, luring observers into conjecturing relationships amongst juxtaposed point sources. Despite the apparent proximity of their components, such ‘constellations’ are, of course, purely human constructs, based on mythical beasts. By the early 18th century, and at least 100 years before Colt Hoare published Ancient Wiltshire, it was conjectured that the Wansdyke had once extended beyond its known terminus in Savernake Great Park. Since then, a feverish scramble to identify relics has resulted in a rash of contenders drawn across the Methodology face of Wiltshire. Sometimes four competing claims have been staked within a kilometre of each other. In one case a putative section of the Wansdyke some 3.2km long within Savernake itself incorporates features which can now be identified as a small The study area shown in Figure 1 is centred on the surviving parts of Savernake Forest and extends eastwards to incorporate Chisbury and Bedwyn and was defined so as to correspond with concurrent archival research on the history of Savernake, with | Byeley in Densome, Woodgreen, Fordingbridge, SP6 2QU, gmbathe@aol.com 2 Hilltop, 39 Lode Hill, Downton, Salisbury, SPS 3PW, jokramsay@yahoo.co.uk 187 THE LARGER LINEAR EARTHWORKS OF SAVERNAKE ‘palaquinu 1Xa} U1 Sans YNM Daly KPMG ayDUsaavy | ‘Sh Seats OO OO jesumen ’ “IH ¢ io 9) 18 some adjustments to take advantage of data emerging from aerial imaging. Within the study area, linear features or complexes were chosen for field visits selected because they met up to five criteria: a) their potential importance had been identified from archival sources, b) their location suggested possible links with features documented in Anglo-Saxon charters, Forest perambulations, or other territorial boundaries, c) they appear as linear features on maps or aerial imagery, d) they had been noted during concurrent fieldwork, and, e) they had been subject to previous research and debate. New sites were added during the course of the survey, while others, especially those identified solely from aerial images, were rejected when field inspection revealed that apparently coherent features were adjoining but unrelated. The final list comprises 35 linear features or complexes. Whilst this is not a comprehensive catalogue, the intention has been to study all earthworks claimed at one time or another to have been part of the Wansdyke together with a selection of further notable features. Field visits took place during winter when ground features were least obscured by vegetation. Basic information was recorded regarding location, topography, habitat, land use, current function of earthworks, structural details (number of banks, ditches, counterscarps), consistency of course (straight, curving, sinuous, meandering), gaps, and correspondence with parish, township or other boundaries. Correlations were made with features visible from maps or aerial imaging. In each case, cross sectional profiles were recorded at representative sections. Drop-distances to ground level were measured every metre from a levelled, horizontal cord suspended between poles. The cord was aligned at right angles to the feature, using the highest bank as a benchmark. Satellite-generated grid references were recorded for each location (accurate to 8m) and compass bearings taken. Alongside fieldwork, existing information relevant to the study area was collated. The major sources inspected were: 1. A detailed LiDAR (Light Detection and Ranging) aerial survey of Savernake Forest and peripheral areas, flown for the Forestry Commission in 2006. This technique uses high density laser pulses, capable of penetrating canopy woodland. Digital Terrain Models, that represent the ground surface after all vegetation has been removed, were processed by the THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Forestry Commission to generate image files revealing ground features at high resolution. The technique and its role in understanding the historic landscape of Savernake are described in Lennon and Crow (2009). 2. Aerial photographs, particularly those held by the National Monuments Record at Swindon, and RAF images taken after heavy wartime tree felling exposed underlying features. 3. Data emerging from English Heritage’s National Mapping Programme for Savernake which provides a synthesis of archaeological information from multiple photographic and LiDAR sources. 4. Cartographic sources for Savernake. Of particular importance are maps of manors and farms surveyed by Charles Price and Thomas Atkinson for Lord Bruce around 1719 (WSA 1300-372), together with detailed estate maps of 1786 (WSA 1300-360), 1820 (WSA1300-375L) and 1846 (WSA 3354MS). 5. Archival information relating to Savernake. This is voluminous. Major sources include Royal Forest perambulations, Forest and manorial court records, inclosure awards, surveys, deeds and conveyances, employment records, legal disputes (especially those relating to tithes), and details of landscaping, including advice given by Lancelot (Capability) Brown. These sources are held at the Wiltshire and Swindon Archives, National Archives and at Longleat House. Site accounts 01 Pantawick Site 1, Figure 1. Points A to G on aerial photograph, Figure 2. Location West of Brown’s Farm near the A338, 1km south of Marlborough. Description This feature is partly ploughed and is known from 19th century maps (eg WSA1300-375L; WSA 3354; WSA2027 - Figure 3) and aerial photographs, which show four fields whose combined perimeter forms a striking ovoid structure, centred on SU19256765, 850m wide, with a perimeter ofc. 2.2km. The location occupies some of the highest land in Savernake at THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 189 Fig. 2 Aerial Photograph, Site 1 Pantawick. NUR SU1967/11. © English Heritage, NMR, Crawford Collection. 190m OD, and lies on a natural domed ridge covered with Clay-with-flints. The northwestern quarter adjoins a small field known as Pantawick (marked P on Figure 2), whose boundary may represent an unploughed section of the ovoid enclosure. This comprises a small bank with two steps dropping 2.6m to lower ground within Pantawick itself (Figure 65). Research, documentation and correspondence with boundaries A Roman villa is known from aerial photographs and can be broadly located (labelled RV on the aerial photograph in Figure 2). Fourth century Roman pottery has been recovered from the site (Hostetter and Howe, 1997) and it has been conjectured that Pantawick could be a Romano-British or earlier site (Crittall 1959, 420). A settlement known as Ponter Wyke was present in 1490 (Gover, Mawer and Stenton 1939), which became the site of a Forest keeper’s cottage. From 1575 Pantawick was at the northern edge of the newly constructed Savernake Great Park, and the keeper controlled a beat, which extended outside the Park and into the northern (unpaled) Forest. The Receiver’s account for 1618 reveals that the keeper of Pantawick Lodge, at the northern end of the Park, was dismissed, and the pales removed (WSA 9-22-206). In the same year, the Earl granted to his youngest grandson, Sir Francis Seymour, 100 acres of pasture in the area that became Brown’s Farm, ‘lately taken out of the said Earl’s park, and divided from the residue of the Park by pales on the south part’. The re-alignment of the pale indicates a clear intention to continue managing the rest of the Park for deer. The revised park boundary is marked K-L on Figure 2. The 190 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE wy RS ~_ N \ ~ a on WS \S\ Y \ ‘ ptm Fig. 3 Extract from map of Savernake by A M May (enhanced and relabelled), 1845 showing Pantawick. (WSA 2027). ovoid structure might relate to Woodhouse or Widow’s Coppice, one of a series of coppices along the northern edge of Savernake, partly forged out of woodlands mentioned in perambulations from about 1300, including Hurlmore, Granham and Manton Coppices. Records reveal the systematic grubbing of the coppices, with the sole remnant of this massive woodland, once extending to 277 acres (112ha), being a small parcel of Manton Coppice. Woodhouse Coppice was conveyed with Brown’s Farm and Pantawick from the Earl of Hertford to Francis Seymour in 1618. The conveyed land was also bordered ‘by the coppice and woody ground called Widow’s alias Woodhouse Coppice and Granham Ragge on the north’, and by Pantawick Lodge ‘lately granted to Sir Gilbert Prynne and William Mountjoy’ in the north-west (WSA 9-22-14; 9-22-206). The coppice was described as 60 acres, ‘formerly parcel of the Great Park, but now severed from it’. Men had been employed ditching, quick-setting and hedging the new bounds of Woodhouse Copse, and carrying row (deadwood) from Woodhouse Rag for it in 1616 (WSA 192-44). There was no mention of Widow’s Coppice by the mid 17th century and it may have already been grubbed (WSA 9-22-225). Whilst nearly all the coppices were eliminated, their rounded outline formed of banks and hedges persisted in a landscape otherwise covered in rectilinear fields. From points D-K, the Pantawick earthwork is the parish boundary between Savernake, in the Hundred of Kinwardstone, and Marlborough (previously Preshute) within the Hundred of Selkley. The parish and hundred boundaries also extend across the site south from point F to a corner at H and then eastwards to J, defining a rectangle, just visible and arrowed on Figure 2. This area might equate to rectangular woodland shown on Andrews’ and Dury’s map of 1773. It is likely that the same lines mark the bounds of the King’s demesne. Perambulations of 1330, of which several versions exist, state that the villages of Manton, Preshute and Elcot together with their pasture and lands, were to remain in the Forest of Savernake, even though they lay outside the demesne (WSA 1300-65; PRO: E 32-220; 12-12-22; E146 bundle 2 no. 34). Observations Aerial photographs and maps suggest that the boundary of the feature may not have been strictly curving, but a series of short straight sections. Boundary trees also marked the course, one of which survives. The soil mark visible on Figure 2 appears narrowly defined, rather than as a broad bank. Conclusion This feature is clearly of interest in a local context. Excavation and further research might elucidate whether it is connected with the Roman Villa, whether it defines Woodhouse Copse, or is related to Pantawick Lodge. 02 Furze Hill Site 2, Figure 1. Points A to F on LiDAR image, Figure 4. Location This earthwork is situated in the northern part of modern, wooded Savernake, south east of Marlborough. Description The earthwork comprises a minor bank and ditch, extending in a regularly curving loop partly in woodland and partly around the grounds ofa private residence known as Furze Coppice. Variable in scale, but not more than 75cm high and always with the ditch on the outside (Figure 65), in many places the ditch is in-filled, although to the north it has sometimes been augmented with pits excavated along its base. In part it forms the boundary between the residential grounds and the woodland each side. The bank has some fine specimen oaks, including the Cathedral Oak (Figure 5). THE LARGER LINEAR EARTHWORKS OF SAVERNAKE Fig. 5 Members of Bedwyn History Society measuring the Cathedral Oak situated on top of (and presumably post-dating) the bank on Furze Hill (Site 2). Studies of oaks of known age in Windsor, give a predicted age of over 1000 years for a tree of 995cm girth in that locality. Research, documentation and correspondence with boundaries The monument broadly equates with the feature shown as Isbury Coppice on a coarsely drawn map prepared in association with a tithe dispute in 1688 (WSA 1300-46). It also accords with an area named Furze Coppice on a 1786 estate map of Savernake, which depicts a sweeping landscaped avenue penetrating the site (Figure 6) (WSA 1300-360). A 6” to the mile Ordnance Survey map of 1900 shows the same bounds, with the name Furze Coppice applied to the residence at the eastern end. A lodge here retained its name into the 21st century (demolished 2008). Isbury Coppice is mentioned in Court Rolls from the Elizabethan period, when hazel, ‘roddes and rowe’ were frequently stolen (WSA 192-47B; 1300-87). The 192 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE ee RG * oes Fig. 6 Extract from a Plan of the Forest of Savernake and Tottenham Park (showing Furze Coppice - Site 2) prepared for the Earl of Ailesbury 1786, re-orientated north. (WSA 1300-360). change of name is explained in a lawsuit concerning the extent of St Margaret’s Common in 1762 when it was reported that ‘Isbury Coppice was grubbed and 1s now overgrown with furze and 1s called Furze Coppice’ (WSA 9-2-388). Observations The Cathedral Oak (Figure 5), which sits on the southern bank at SU20616798, has a girth of 995cm at a height of 1.5m. Studies of oaks of known age in Windsor have shown that, in the conditions prevailing there, a tree of such dimension might average a little over 1000 years old (White 1998). As a slow growing pollard, the Cathedral Oak might be even older. Whilst the validity of applying growth and age profiles measured from one location to another has not been proven, the oak (the third largest in Savernake) is certainly of considerable antiquity. Conclusion All of the available evidence indicates that this feature marks the bounds of a former, large coppice. Of interest however is the Cathedral Oak, whose size indicates that the coppice bank is of considerable antiquity. Even allowing for exceptional growth, the feature would appear to be early for woodbanks in Savernake. The possibility that the bank was constructed around a pre-existing tree, or that the earthwork represents an early enclosure, later adopted as an coppice within the Royal Forest, cannot be excluded. 03 Postern Great Dyke Site 3, Figure 1. Points G-H-A-K on LiDAR image Figure 4. Location On Postern Hill, 400m south of Marlborough, south- west of Forest Hill. Description This feature comprises a single bank and ditch extending for 520m on level ground of the Savernake plateau, overlying Clay-with-flints. It traverses young broad-leaved, semi-natural oak and birch woodland, with considerable scrub to the north. The feature is substantial, but without steep sides, being THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 193 some c.17m wide, and up to 1.41m from top of bank to bottom of ditch, with the ditch to the west (Figure 65). It is sinuous in its course, and variable in height, showing signs of erosion in places. Its northern end at point K, SU 20096858, is close to another large bank along the crest of Postern Hill at an angle of 060°. From here to point A at SU 20086849 the feature has a significant cover of hawthorn scrub and bramble, and is ill-defined. From point A southwards the feature is prominent, and falls within the bounds of Furze Hill (Site 2). At point H, SU 20086841, the feature changes direction, sharply but smoothly, from 190° to 138° and continues for a further 350m petering out at point G, SU 20246821. A major track traverses the feature near the southern end, and a smaller one in the centre. Research, documentation and correspondence with boundaries The feature does not accord with any parish, compartment or coppice boundary. Aerial photographs show the site largely open until 1930 (eg NMR SU202686). The feature is not shown on 18th or 19th century estate maps (WSA 1300-360; WSA1300-375L), but appears to be present on Cary’s engraving of earthworks in Colt Hoare’s Ancient Wiltshire (1819). Observations A further feature F-M, comprising a sinuous ditch, is situated 200m south west of the dyke. There are complex holloways leading to the edge of Postern Hill to the northeast, at point L, and a potential field system on the slopes of the hill. Conclusion This feature remains undated and unidentified. It is of larger scale than most other linear earthworks in the study area, with the exception of the Bedwyn System (Site 35). It appears akin to dykes found nearby at Rainscombe and Martinsell. A prehistoric origin seems likely, and it may be associated with the Iron Age oppidum and Romano British settlement at Forest Hill Farm. There is no current evidence that it was linked to other dykes within a large scale territorial or defensive arrangement, although that possibility cannot be excluded. Further work is needed to determine the relationship between the various earthworks in the locality and geological changes of slope on this northern crest of Savernake. 04 Forest Hill Crescent Site 4, Figure 1. Points A to C on LiDAR image, Figure 7. Location This site is 1km east of Marlborough between Forest Hill Farm and the A4. Description This minor feature is just 240m long, and appears almost as the arc of a true circle projecting southwards, either side of the lane leading to Forest Hill Farm at SU 20956840. It traverses an area of scrub and large hawthorn trees on the Clay-with- flints, with either end lost in the arable to the north. Within the encircled section the land is higher, as if forming a raised platform. The bank and ditch are poorly defined, up to 87cm in height from top of bank to bottom of ditch and less than 7m wide (Figure 65). Research, documentation and correspondence with boundaries The 1786 Savernake estate map (WSA1300-360) appears to show an avenue of trees alongside the western arc of the feature. Both this source and a map of 1820 (WSA 1300-375L) suggest that the bounds of the Forest may have followed the western arm A-B, but then continued directly east to point D on Figure 7 (Site 5). The remains of a late Iron Age and Romano-British complex lie adjacent to Forest Hill Farm, thought to have been a regional oppidum prior to the Roman invasion (Corney 2001). Observations The crescent resembles the feature found in Shepherds Wood (Site 6, Figure 8), 800m to the east. Conclusion This feature is undated, and its purpose is not identified. It could be modern, associated with landscaping works of the Forest, or part of a sweeping entrance to Forest Hill Farm. Alternatively, it may be linked to local Iron-Age occupation. Further work on settlement and land-use patterns in this area may throw light on its origins. 194 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 7 LiDAR aerial image of Forest Hill (Sites 4 and 5), 2006. © Forestry Commission. 05 Forest Hill Banks Site 5, Figure 1. Points B-D-E-F on LiDAR image, Figure 7. Location This feature is parallel with and just north of the A4, lkm east of Marlborough. Description This feature is distinctive amidst a plethora of banks and ditches in this location (Figure 7). It extends for 580m in open broad-leaved woodland on the Clay- with-flints from point B at SU 20956838, to point F at SU 21606817, maintaining a gently meandering course. At point B it emerges as a single bank and ditch branching off the crescent-shaped feature of Site 4. Within 100m, at SU 21056837, point D, it appears as a stepped structure. Here it comprises a bank and ditch, south facing, with a brief rise again before plunging into a steep trench, 3.57m deep. The whole structure is 30m wide (Figure 65). The large trench merges with others in the vicinity. Further east the bank and ditch dissociate from the trench. At point E, SU 21436825, the bank is only 40cm high, which defines a step in ground level of 70cm between the two sides (lower to the south), as if the feature is partly formed by a lynchet. Further east at point F it cannot be traced. A small bank runs north east from here to point L, crossing a pit. There is a small bank F-G, which straddles the dip in ground level, joining up with a differently conformed straight bank which runs from H to J, at Cock-a-troop Lane, where it meets Site 6. Research, documentation and correspondence with boundaries The course of two Roman Roads conjoins at point I (described, Site 6). A woodland known as Levy (Levi/Leafy) Copse is recorded north of this site, and was cleared after 1846 (WSA 3354 Box 1&2). The small bank F-L may mark its eastern extent. Point F is in the vicinity of Red Vein Corner, shown on an 18th century estate map (WSA 1300-360). From here the main London road descended south-east to Red Vein Bottom. The same map, and one from 1820 (WSA1300-375L), show the feature marking the edge of the Savernake Estate. Part of the feature accords closely with a Forest perambulation undertaken following a Royal decree of 1299 (Edward I). This perambulation traced the King’s highway as far as Lachenhardescrofte. Leonard’s Field, west of Cock-a- troop Lane, probably marks the general area of the croft. Section H-J departs from this perambulation ) VET OT RL eT PL eee pees ore en oe per nant THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 195 Fig. 8 LiDAR aerial image of Shepherds Wood (Site 6), 2006. © Forestry Commission. which continued by the King’s Highway through the bottom of the valley (Red Vein Bottom) (WSA 212B/7189). Observations The trench at point D appears to be a major holloway. Deep structures of this kind, constraining traffic in narrow passageways, must have been a major security risk to travellers. On one night in 1782 no less than five stagecoaches were attacked by highwaymen in Savernake, including two near Red Vein (WSA 1300-2511). LiDAR imagery shows further extensive braided tracks to the south, skirting both sides of Bowles Oak, a major woodland block in this vicinity since at least the 16th century (WSA 1300-87). The single bank and ditch appears to have been aligned with the trench, and inserted part way down it (i.e. post-dating it). Conclusion This feature appears to be an ancient boundary of the Forest used in 13th century perambulations and presumably therefore also marking the King’s demesne. In part it aligns with a substantial holloway representing the King’s Highway of the same period. — 196 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 06 Shepherds Wood Site 6, Figure 1. Points E to G, H tol and J to K on LiDAR images, Figure 8. Location Shepherds Wood lies north of the A4, 2km east of Marlborough. Description This feature, distinct amidst a profusion of banks and ditches, forms a wavy line (sometimes with sharp crests) across the south-facing slopes of Shepherds Wood and is situated mainly on the Clay-with-flints, with some chalk exposed on the lower slopes and valleys. Much of the woodland is plantation beech, with semi-natural oak woodland near the A4, and mature hawthorn and birch, resembling scrubbed- over pasture woodland. A parcel at Hern Corner is conifer plantation. The suspected length of the feature is 1.65km, although its course is sometimes indistinct, especially each side of the Stitchcombe Road. It commences by Cock-a-troop Lane, at SU 21836825, point E on Figure 8, where it forms the lower of two ditches running eastwards which diverge briefly. It curves slightly south, obscured partly by pits, before rejoining at point KF From here, apparently the same feature turns acutely south in a great arc to point G. From here to point H it probably continues, although is difficult to map with certainty amidst dense undergrowth. From H it undulates further to point I, just south of an earthwork enclosure, where its course is less clear. For the most part the feature represents a simple change in ground level, at its deepest 78cm, with lower ground to the south. A shallow, poorly defined bank up to 39cm in height is present in places. The whole feature may be just 5m wide, with indistinct edges (Figure 65). Field survey provides no evidence that the feature continues to Hern Corner, although documentary sources (described below) suggest that it may have. At Hern Corner, north of a straight bank J-L cutting across it, a single bank and ditch extends northeast for 20m before turning sharply in a regular southerly arc to point K, where it forms another point at SU 23956825. This appears to conform to the morphology of the remaining structure, with a bank of up to 32cm above ground level and ditch no more than 65cm below the bank top, south facing. A further bank and ditch runs close to the northern edge of the woodland, from A to L, and Fig. 9 Puthall Gate, looking east. Early 20th century postcard. presumably at one time bounded it. Between points S and F the woodland extends beyond this line, but there is no northern bank where it abuts the field. Hern Corner itself, as defined by J-T-L, is a separate woodland compartment and did not form part of the Savernake Estate when the woodlands were leased to the Forestry Commission in 1939. There is a bank and ditch between L and M, which, for most of its length, defines the edge of East Croft Coppice. At point M, Puthall Gatehouse stood (Figure 9), opposite point I of the Crabtree-Amity feature (Site 13). Cock-a-troop Lane (leading to Mildenhall) is deeply banked on each side. The course of two Roman Roads from Cunetio (Mildenhall) diverge and extend across the site just west of the lane at point B. Section B-C is the road to Sorviodunum (Old Sarum) whilst B-D marks the road to Venta Belgarum (Winchester). There is an undated earthwork enclosure (marked on Figure 8) at SU 22806832 - a Scheduled Ancient Monument (AM589). LiDAR suggests that there may have been a wide entrance towards it (marked P on Figure 8). There is a possible second enclosure labelled R. LiDAR also reveals a series of parallel, south-east facing lynchets, marked N. Research, documentation and correspondence with boundaries The feature does not accord with parish boundaries and cannot be related to surviving Anglo-Saxon charter bounds nor with Royal Forest boundaries. Although a 1299 perambulation went via the cross at Cukeresthorpesende [Cockatroop] it then descended to the bottom of the valley [Red Vein Bottom] well south of Shepherds Wood (WSA 212B/7189). Cartographic evidence suggests that this apparently inconsequential earthwork has been employed as a boundary feature. In particular, a coarse sketch by — i aN - ne ; > Puthall | - » Gate os, & Mile Stone é House Fig. 10 Extract of sketch map by Fohn Ward, 1795, showing wavy bounds of forest, currently within Shepherds Wood (Site 6). (WSA 1300-363A&B). Savernake steward John Ward, attached to a letter addressed to Lord Bruce in 1795, appears to show it in its entirety, as if a familiar feature and representing the boundary of the Forest (Figure 10) (WSA 1300- 363A&B). A 1786 map of Savernake (extract, Figure 11) suggests that Shepherds Woods [plural] formed a discrete block lying beyond the more open ground connected with the Forest, bounded on the south by the wavy earthwork. This map shows a series of tree-lined avenues, one of which curves around the western edge of the feature (WSA 1300-360). Large oaks mark this boundary today. An 1820 estate map is similar, showing Shepherds Wood as a densely wooded parcel, bounded by the wavy earthwork (WSA1300-375L). Observations The feature at its eastern end near Hern Corner appears to be broadly parallel to the large earthwork enclosure. Although having the stepped appearance of a lynchet, it does not define a rectilinear field pattern and cuts across other lynchets parallel to each other at point N. Conclusion This feature cannot be dated, although its wave-like alignment is unusual in Savernake and quite unlike coppice boundaries. The line was recognised as a boundary feature in the late 18th century and its edges respected in the layout of avenues. Whilst it might represent bounds constructed around the edge of an irregular dense stand of trees, there is little evidence of banking and ditching, and it mainly consists of a simple change in ground level. Although perhaps only of local relevance, its unusual conformity over a long distance merits further investigation. Fig. 11 Extract from a Plan of the Forest of Savernake and Tottenham Park (showing Shepherds Wood — Site 6) prepared for the Earl of Ailesbury 1786, redrawn. (WSA 1300-360). 07 Puthall Park Site 7, Figure 1. Points A to E on LiDAR image, Figure 12, and aerial photograph, Figure 13. Location This feature is 3.5km east of Marlborough, just north of the A4, adjacent to Puthall Farm. Description Puthall Park is on the northern (south-facing) valley of the A4, mainly on Clay-with-flints, with Upper Chalk in the lower reaches. A pasture adjacent to the park on the eastern side marks the deserted medieval village of Henset. Most of the site is bounded by a bank, points A to D. The western section (A-B) is of interest because of its apparent double or stepped structure, with the ditch to the west (outside the park). It maintains a smoothly meandering course aligned at 338°. The feature is 1.53m high from floor to top of bank and 14m wide (Figure 65). There are semi-mature oaks on the bank and along both sides of the lane, with arable to the west; the interior of the park is mainly beech plantation. The bank between points B and C is just 80cm from ditch floor to bank- top. Section C-E along the eastern side is straighter and appears morphologically distinct, with a wider ditch up to 5m. This section is connected to another ditch, running east-west, D-E just north of Henset. There is a major forestry track crossing the site. The south eastern edge of the park appears to have been cut off by a track that extends to Puthall Farm. Research, documentation and correspondence with boundaries An Anglo-Saxon charter boundary recorded in a 198 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 12 LiDAR aerial image of Puthall Park (Site 7) and Hens wood (Site 8), 2006. © Forestry Commission. grant of landin AD 778 (Sawyer 1968, Cat. No. 264), ran along a continuous dyke in Puttan ealh (Grundy 1919; Crawford, 1921). Options for the course of this are given under Hens Wood (Site 8). There is a pedestrian right-of-way, Axford Lane, immediately adjacent to the feature, which forms the boundary between Mildenhall and Little Bedwyn parishes. The northern end of the Park divides Ramsbury from Bedwyn. The location appears on Price’s 1717 map of Puthall Farm (WSA 1300-372), under the name of Puttall Park coppice (Figure 14), where the site of Henset village is labelled Moam’s. Puthall had been part of the King’s estate of Bedwyn in the 12th century, when it was transferred to the keeper of Marlborough Castle (Crowley 1999, 59). The boundary of Savernake abutted the land of William of Puthall in the 13th century (Brentnall,1942). For much of its early history Puthall was held by Fig. 13 Aerial Photograph of Puthall Park (Site 7) and Hens wood (Site 8), 1946, after war-time felling. (RAF5110, CPE/UK, 1821). THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 3) ££ Lis] sonal = SA Pitt Close Puttall Gate Crbm Crhrcouvale Sroroy of {PUTTEL FARM on the Bt Soon CHARLES Ld Bruce by C Pric 1717 Galver’s Close 199 \ 4444PE White Furlong — 0% Seale of 60 Boles G BO 60 Fig. 14 Survey of Puthall Farm by Charles Price, 1717 (relabelled). (WSA 1300-372). ecclesiastical institutions. In 1368, Henry Esturmy conveyed Puthall, including a carucate (an area ploughable by one ox-team), 10 acres of pasture and 20 acres of wood, to Holy Trinity Priory, Eston (WSA 1300-20). At the dissolution of the monasteries the property was acquired, along with the site of the Priory itself, by Edward Seymour, later Duke of Somerset. Subsequent rentals of Seymour lands in the 16th century mention woodland at East Croft Coppice, but not Puthall Park (Longleat House Archives, Seymour Papers, Vol 12). Observations _A cross section through the feature suggests that the ‘step’ in its bank is probably an artefact of the holloway along Axford Lane causing a secondary change in ground level (Figure 65). The bank and ditched section alongside this is modest and comparable to other banks of the park. There is a meandering shallow ditch traceable WSW — ENE across the park, possibly linked to Henset. LiDAR suggests the presence of many lynchets in the area. Although the site could be an ancient deer park, there are no references prior to 1717. A profusion of wild daffodils in this wooded area makes it distinct from nearly all other parts of Savernake. It is possible that the land was imparked after its acquisition by the Duke of Somerset. The presence of an external ditch, ill-suited to keeping animals in, suggests that it may not have been first built as a park, but adopted for this purpose afterwards. A long, wide avenue of trees extends through the park and joins Axford Lane. This avenue and the tracks through nearby Hens wood have mounds alongside them (visible on LiDAR image Figure 12) as if part of a single planned system. The park might relate to landscaping of Tottenham and its hinterland, rather than a deer enclosure. Conclusion Whilst the Park and Henset are both of interest their boundary features are of local significance only 200 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE and there is no reason to consider them part of any large-scale structure, nor the dyke at Puthall in the 778 bounds (considered under Site 8). 08 Hens Wood Banks Site 8, Figure 1. Points C-G-H and N to P on LiDAR image, Figure 12, and aerial photograph, Figure 13. Location Hens Wood is 4km east of Marlborough, on the northern side of the valley of the A4. Description Almost the whole of Hens Wood overlies Clay-with- flints. The wood has a bank and ditch system near its southern and north-western edge. A characteristic of Hens Wood is that parts of it are divided into strips by roughly parallel, meandering banks and ditches with at least six in the south-western quarter of the wood, aligned north-south, and six more outside the study area, in the north-western quarter, running east-west. These are not readily distinguishable from footpaths on aerial imagery, although two can just be seen either side of point P on Figure 13, heading at an angle of 170° towards the southern boundary. In some cases small gaps are visible on the perimeter banks of the wood, as if permitting access to each strip through independent gates. Southern Section: This extends from point C, at SU 23816835, where it abuts Puthall Park, to the junction with Knowle Hens Wood, point H, at SU 25116840. It does not form the boundary of the woodland, but lies just within it, on largely level ground. There is hazel coppice and oak south of the bank with plantation woodland to the north. Nineteenth century Ordnance Survey maps show a boundary stone at SU 24516832 and a post at SU 23826835. The feature consists of a single bank and ditch, with the ditch southwards, characterised by being erratically sinuous, especially along its course of 1.03km from C to G. It is slightly more regular along a further 360m from G to point H at the edge of Knowle Hens Wood. The feature is about 10m wide, and up to 1m high (Figure 65), although both bank and ditch are variable. There are no gaps except for a track to Voronzoff Gate and a levelled area with a public path across it. Knowle Hens Wood has almost no discernible bank around it from H to J. Dense vegetation makes it difficult to assess whether the bank and ditch continues as the division between Hens Wood and Knowle Hens Wood. North-western Section: At point C there is a wide eroded gap between the banks of Hens Wood and Puthall Park, resembling a former passageway for animals. The Hens Wood bank and ditch extend from point C north, maintaining the ditch on the outside of the wood. It then turns east from point N (SU 23816860) alongside a right of way leading to the Kennet Valley. Here the feature is fairly regular, up to 0.7m high. At SU 24006865 where the bank deviates slightly to the south, it joins with one of the strip division banks, referred to above. At point O (SU 24216870), where the track turns north, the feature continues east and into the woodland to point P (SU 24416870). Here it appears as a broad ditch between two banks, 8m wide and up to 1.1m deep, before dividing erratically and merging with ground level (Figure 65). Research, documentation and correspondence with boundaries The southern section C-H forms the boundary between the parishes of Little Bedwyn and Ramsbury. The parish division does not extend around the edge of Knowle Hens Wood between points H and J, but goes between the two woods, with a major bizarre stagger in its course, K-L-M, before dropping back to the southern edge at point J. Section C-G also forms the edge of Puthall, as shown on Price’s map of 1717 (Figure 14). The southern section is a possible 13th century perambulation boundary of Savernake Forest. Through a decree signed at Clarendon in 1299, Edward I formally accepted (initially at least) perambulations undertaken to eliminate all lands afforested by Henry II, Richard and John (WSA 212B-7189). A section of the perambulation went along the eastern part of Red Vein Bottom, as far as the Croft of the Priory of Saint Margaret at La Putte [Puthall], thence ascending by the aforesaid croft as far as the croft which 1s called Hobbesare, and thence always by the hedge as far as the wood of Puttehale, to the house of Richard de Tunbrigge. Brentnall (1942) interpreted this line as along the edge of Hens Wood. It is clear from various sources that the name Hobbs (as referred to in Hobbesare) related to land both north and south of the London Road (eg WSA1300-47; 9-2-388). The name features on Price’s 1717 map of Puthall (Figure 14). An alternative interpretation for the perambulation would be an alignment along the modern boundary of Savernake Forest, well south of Hens Wood. Whichever boundary is correct, it is likely to define the extent of the King’s own holding. The Anglo-Saxon charter of 778 noted THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 201 above, recording a grant by Cynewulf, King of the Saxons, to Earl Bica of land at Bedewinde, mentions Puthall (K133, quoted by Grundy, 1919) and may have aligned with the edge of Hens Wood. The boundary went through Standen to a certain dyke, and by the same continuous dyke in Puttan ealh, and thus by the long agger to Bulcan Pit, and thenceforth by the long dyke across the street to a certain stony slope. Standen is unidentified. There are many pits in the area; Pit Close is shown on Price’s map (Figure 14) and a large pit is visible on the LiDAR image north east of Henset (Figure 12). The ‘street’ appears to be a reference to the Roman Road from Cunetio (Mildenhall) towards Spinae (Speen), only minor sections of which are known. A possible fragment of this, revealed by the mapping undertaken by English Heritage in association with the National Mapping Programme, is shown as R-S on Figure 12. This also aligns with a suspected fragment of Roman Road at Hill Barn, south of Chopping Knife Lane, south east of Mildenhall. Hens Wood was described as rank coppice by the Forestry Commission in 1948, when, after war-time clear felling, the area was planted with Japanese Larch and Norway Spruce. The Chairman of the Commission was dismayed to find that conifers had been planted here, contrary to the policy of broadleaves in Savernake and its outliers (PRO F43/150). These conifers were themselves clear-felled in 2008. Observations There is presumably a link between the names of the deserted medieval village of Henset, next to Puthall, and Hens Wood. Neither of these names carry a possessive apostrophe, although the etymology is unclear. It seems possible that the southern and northern boundaries of Hens Wood are not contemporary. The erratic course of the southern section is quite distinct. At point P on the northern section, the double feature seems to be a former track leading into the core of the wood. This can be seen on the aerial photograph (Figure 12) taken in 1946 immediately after the felling referred to above. The whole of the wood, as revealed on this photograph, is riddled with animal paths, suggesting that the site was used extensively as pasture, probably for protracted periods to cause such erosion. Conclusion The wood itself is clearly of interest, as a massive enclosure (135ha) with an abundance of animal tracks and subdivision into parallel strips. In addition, the southern boundary, which forms the division between Ramsbury and Bedwyn, and the bounds of Puthall Farm, could be part of a long distance territorial boundary. It may be the dyke referred to in the 8th century charter (or is otherwise close to it) and which defined the King’s demesne in the 14th century, thereafter being employed as a perambulation boundary for the Forest of Savernake. 09 Bushell’s Linear Site 9, Figure 1. Points A-B on LiDAR image, Figure 15 and (partly) (A)-(B) on aerial photograph, Figure 16. Location This feature is in the western part of the modern (wooded) Savernake crossing Great Lodge Bottom. Description The earthwork extends from point A at SU 20786722, east of High Trees Farm, to point B at SU 21466675 in Great Lodge Drive near Thornhill Pond. It ventures within 50m of the dyke in Great Lodge Bottom (Site 10) and comprises a number of irregular ditches and apparent banks. From point A, beside a pit and two enclosure banks (marked G and H on Figure 15), the feature extends on a fairly straight course of 120° for 930m, obliquely crossing the valley of Great Lodge Bottom (where its course is interrupted by gravel pits) and the straight avenue known as Great Lodge Drive (point F). At point B near New Road Bottom it is also obscured by pits, but may continue to Thornhill Pond (point N on Figure 15). Certain scars cut into the slope here may continue its line (point M) or may be unrelated workings of mineral veins. Whilst maintaining a regular course, the feature is inconsistent in structure. North-west of Great Lodge Bottom up to four ditches converge and diverge. On the southeastern flank of the valley two deep ditches merge into one (Figure 65). Mounds between nearby ditches give the impression of banks, but may reflect general ground level. At its widest this feature is 27m across, but only 80cm from ditch floor to adjoining ground level. Research, documentation and correspondence with boundaries Part of the course of this earthwork appears to have been interpreted as a section of Wansdyke by Major Cs SSW ~ DRNDU0X\RF ~ Fig. 15 LiDAR aerial image of Bushell’s Linear (Site 9) and Great Lodge Bottom (Site 10), 2006. © Forestry Commission. and Burrow (1926), drawing on fieldwork by H. C. Brentnall, and linking it with Great Lodge Bottom Dyke (C-D, Site 10). The feature does not appear on maps, nor does it correspond with any boundaries. Where it commences, at point A on Figure 15, there are two nearby banked areas, one of which, marked H, is similar to other small enclosures in Savernake, possibly of Iron Age date (Lennon and Crow 2009). The other part, enclosing a pit, may represent a former brickmaker’s property. A map of 1762 showing the bounds of St Margaret’s sheep common states at this point ‘Here Bushell’s house stood’ (WSA 1300-1842 A & B). Observations This feature has no banks and is most substantial on sloping ground. The ditches merge and peter out on hilltops, observations consistent with it being a braided track. Conclusion There is no evidence that this is a deliberately constructed feature, part of an administrative boundary or defensive feature, nor connected to the dyke in Great Lodge Bottom. Instead, it appears to be an unmade track, possibly linking Thornhill with a brickmaker’s premises, which became redundant when the house was abandoned before 1762. oe \ Fig. 16 Aerial photograph 1934 (when conditions were more open) showing part of Bushell’s Linear (Site 9) and Great Lodge Bottom (Site 10). (NMR SU 2165/1, CCC9120/9340). © English Heritage, NMR, Crawford Collection. THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 203 10 Great Lodge Bottom Dyke Site 10, Figure 1. Points C-D on LiDAR image, Figure 15, and 1934 aerial photograph, Figure 16. Location This feature straddles a small part of Great Lodge Bottom centred on SU 21156705. Description This is a prominent earthwork extending for 120m at right angles across the shallow wooded valley of Great Lodge Bottom, 500m south-west of the former keeper’s lodge at Braydon Hook (L on Figure 15). It crosses a small chalk valley with a gravel floor, but ceases at the Clay-with-flints on each flank. The earthwork consists of a single bank and ditch, orientated at 150°, with the bank to the south west (Figure 17). The dyke is generally straight, with a very slight S-shape in its course to the south-east. Bank height varies, although it does not appear eroded and may have been constructed inconsistently. At its greatest it is 1.68m from the ditch floor to bank top, and 19m wide (Figure 65). The dyke is crossed by a track along Great Lodge Bottom, where several pits have been dug along the valley. A causeway across one pit permits access along the ditch of the earthwork. The south-eastern arm of the earthwork has a gap along its western edge and ends abruptly (cf Shovel Bottom Dyke, Site 12). The north western arm is wider, with a more rounded ditch and a lip where its depth changes abruptly; a later pit has been dug across its northern end. Research, documentation and correspondence with boundaries Major and Burrow (1926), drawing on initial fieldwork by H. C. Brentnall, considered this feature part of the Wansdyke and linked it with another linear system (Site 9) to extend the line. The feature is not shown on estate or other maps prior to the Ordnance Survey, nor does it accord with parish, bailiwick, or management boundaries. No trace of it can be located in the voluminous Forest documentation. It forms part of a Scheduled Ancient Monument (SAM 866). The correspondence between Savernake cross valley dykes and geology was reported by Ramsay and Bathe (2008) who observed that Shovel Bottom and Great Lodge Bottom dykes are confined to exposures of chalk in the valleys. They conjectured that the dykes were constructed to demarcate valuable arable land, Fig. 17 The southern arm of Grerat Lodge Bottom Dyke (Site 10), looking north (downhill). whilst semi-natural habitat on the Clay-with-flints adjoining was a common pool resource for grazing, fuel and hunting. Observations The remnants of small fields appear to abut the dyke (labelled P on Figure 15) and there are prominent enclosures nearby (labelled H and J on Figure 15). The juxtaposition of certain enclosures and cross- valley dykes in Savernake was noted by Lennon and Crow (2009). The feature is close to a series of pits (K) on both sides of the valley, which resemble boundary markers due to their equidistant spacing, but which equally might be simple mineral extractions. Conclusion This earthwork is a discrete feature that may define local boundaries on scarce chalk soils, possibly related to adjacent small fields and nearby enclosures. The cross-valley dykes of Savernake offer an interesting field of research, but there is no reason to suppose they form part of any long distance territorial or defensive boundary. 11 Savernake Great Inclosure Site 11, Figure 1. Shown as circuit on map, Figure 18. Location — This large feature encircles the area of Eight Walks and Luton’s Lye in Savernake Forest. 204 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE \\ Mildenhall ” {\ Parish .-” m wa § EEG Litde 4 Bedwyn K Parish Ae SARS Ss. . ee S< Savernake Bank visible on ground Course from maps/ LIDAR |} - » Civil parish boundaries cs) 500 4 i Shovel Bottom Fig. 18 Map of Savernake Great Inclosure (Site 11). Description This ovoid earthwork comprises a bank and external ditch enclosing a level plateau of 217ha and is 2.2km across at its widest. The perimeter is crossed by the dry valleys of Shovel Bottom, Bagden Lawn and a tributary valley of Bagden carrying the ride called Drury Lane. The circuit is traceable on the ground for 85% of its course, but not with certainty where it sweeps around Savernake Lodge (G-H). LiDAR images suggest that it probably formed a single continuous structure, although the most substantial section in Shovel Bottom, A-B, may not be contemporary with the remainder and is treated here as a separate feature (Site 12). Elsewhere the earthwork is variable in scale but generally less than 1m high (Figure 65). Along its western edge, between points D and E, the bank is constructed on a natural ridge which exaggerates its vertical dimension. The circuit conforms with natural topography for nearly its whole length, but for a section of 400m from points H and A, between Drury Lane and Shovel Bottom, where it is up to 80cm high and takes a straight course. Research, documentation and correspondence with boundaries The feature is shown on all detailed maps of Savernake, the first of which was prepared in 1786 (WSA 1300-360). Crawford (1920) labelled the feature as an ‘ancient enclosure (park or wood)’ on a map illustrating the Anglo-Saxon bounds of Bedwyn and Burbage. Ordnance Survey Explorer maps label part of the bank as ‘Park Pale’. Major and Burrow (1926) interpreted a section of the bank, from Drury Lane, over Shovel Bottom and as far as Pigsty Track, (H-A-B-C) as part of the Wansdyke. Ramsay and Bathe (2008) compared measurements along the earthwork and demonstrated that, where largest, 60-120cm high, it forms the boundary of known coppices. Elsewhere it is a faint bank, 20cm high and 4m wide. Ramsay and Bathe (2008) relate the whole structure to a controversial and probably abortive inclosure ordered by the Earl of Hertford in 1594 to exclude commoners’ animals and permit the re-establishment of woodland (WSA1300-1484). Sections of the bank form the division between the parishes of Savernake, Little Bedwyn and Mildenhall. These three parishes meet at the Amity Oak, point F. Part of the bank was formerly also the division between North and South Savernake. Observations The bounds of the inclosure were adopted to define an area of open park-like ground, during landscaping in the 18th century, partly in association with advice from Lancelot (‘Capability’) Brown (WSA 1300-1917 & 1918). Conclusion This feature represents an attempt to inclose a large area of common land in Savernake. With the exception of Shovel Bottom, where an existing structure may have been re-utilised, it is a discrete feature dateable to the late Tudor period. 12 Shovel Bottom Dyke Site 12, Figure 1. Points A-B on map Figure 18, and LiDAR image and drawing, Figure 19. Location This feature is in Savernake Forest, 800m SSE from Eight Walks. Description This earthwork extends for 155m between points A and B at the southwestern end of Woolslade at SU 229660 and comprises a bank, ditch and counterscarp, up to 1.56m high (ditch floor to bank top) and 16m wide (Figure.65). It straddles the valley of Shovel (Cheval, Shoul) Bottom, cutting into the chalk underlying the Clay-with-flints. Shovel Bottom track runs through it at D-E, but a sunken area here, equating with the ditch of the feature, suggests that the dyke might have formerly crossed the whole valley. The eastern side, A-D, maintains an THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 205 100 metres Fig. 19 Shovel Bottom Dyke (Site 12) shown on LiDAR aerial image © Forestry Commission, and detailed site plan. angle of 264°, some 10° out of line with the western arm. Approximately 80m west of the track the dyke re-aligns from 254° to 294°. At this junction of these otherwise straight sections, there is a gap in the southern bank at point K with a slope from the ditch floor (as if this formed the route along which spoil was dragged out). At the western end, where the feature meets the Clay-with-flints, it ends in an abrupt head wall (point B) whilst to the east it peters out (point A). At each end it connects with minor banks, B-C and A-H, described under Site 11. Research, documentation and correspondence with boundaries This earthwork appears on maps showing the Great Inclosure (Site 11) from 1786 (WSA 1300-360), although these do not depict its current morphology. An 1820 estate map appears to show it as a simple line (WSA1300-375L). Modern Ordnance Survey Explorer maps label this section ‘Park Pale’. Major and Burrow (1926), referencing H. C. Brentnall, incorporated the cross-valley dyke as a section of the Wansdyke through Savernake. Ramsay and Bathe (2008) posited that it was an ancient, perhaps pre- historic, structure, incorporated within an inclosure constructed to exclude commoners’ animals in the 16th century (see Site 11). Like the cross-valley dyke in Great Lodge Bottom (Site 10), the feature is confined to the chalk, and may have separated territories of rare and valuable agricultural soil in the valley, whilst the Clay-with-flints remained undivided, perhaps used as communal pasture and woodland. Lennon and Crow (2009) note earthwork enclosures in proximity to both dykes. Observations There is a tumulus in the same valley 400m NNE of the Shovel Bottom dyke and potentially analogous dykes further along Shovel Bottom at SU 227657 and in New Pond Bottom at SU 216666, another chalk valley with tumuli nearby. Conclusion This dyke is undated but resembles the dyke in Great Lodge Bottom and others in Savernake. A defining characteristic of these dykes is that they cease at the point of leaving the chalk and are thus unlikely to have been part of large-scale territorial or defensive systems. 13 Crabtree-Amity Linear Site 13, Figure 1. Points A to I on LiDAR image, Figure 20. Location This feature is situated on the southern side of the valley of the A4, abutting the edge of Savernake Forest, 4km east of Marlborough. 206 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Description This structure (ifa single feature) extends for 1800m on a gentle north-east facing slope overlying Clay- with-flints on higher land, with chalk and alluvial gravels exposed in the valleys of Crabtree and the A4. It commences at SU 244666, point A, along the edge of the property known as Owl’s Castle, where it appears to have been modified. Between points B and C, within commercial larch plantation, a single bank and ditch 7m wide and 90cm high, with the ditch to the south-west, erratically changes course and varies in height and depth (Figure 65). In the valley of Crabtree Cottages it has been ploughed or eroded, but can be followed as a faint mark across a lawn near the buildings. Between points E and F it is smaller and less erratic in character. From points F to G it forms the edge of Horseleaze Wood, which intrudes into its apparent course at the north western end (point G), as if the feature was Fig. 20 LiDAR aerial image of Crabtree-Amity Linear (Site 13) and Little Frith (Site 14), 2006. © Forestry Commission. THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 207 constructed around the woodland. From points G to I it maintains a regular and straight course, possibly eroded in places, crossing a small valley at point H. At point I (SU 232679) it meets the A4. Research, documentation and correspondence with boundaries From points F to I the feature accords with the boundary of Puthall Farm on Price’s map of 1717 (Figure 14). An estate map of 1786 shows a single hedged bank along the whole of the north east edge of Savernake, from Holt Pound to Puthall, prior to the construction of Owl’s Castle, possibly linking the Crabtree-Amity section and the Bishop’s Rag feature (Site 18) (WSA1300-360). It shows Horseleaze Wood intruding into its course. The parish boundary between Savernake and Bedwyn lies broadly parallel to the south-west, before branching out from the Amity Oak and crossing the earthwork along the valley at point H. An interpretation of the 1299 perambulation of Savernake Forest, drawn to define the King’s own holding, aligns closely with this feature. The boundary went from Red Vein Bottom to the croft of Puthall, thence ascending by the aforesaid croft as far as the croft which 1s called Hobbesare, and thence always by the hedge as far as the wood of Puttehale, to the house of Richard de Tunbrigge (WSA 212B-7189). This identifies Puthall Wood as Horseleaze Wood (sometimes called Puthall Horseleaze), with Richard of Tunbrigge’s house equating with Timbridge. An alternative reading places the perambulation north of the A4, along the bounds of Hens Wood (Site 8). The boundary of the 778 King Cynewulf charter passed along or close to this feature (K133, quoted by Grundy, 1919; Sawyer 1968, Cat. No. 264). Observations At Owl’s Castle this feature runs along the south- western edge of the property, while the earthwork of Bishop’s Rag (Site 18) runs north-east behind _ the same property. There is no evidence that they were connected, although they are broadly similar with south facing ditches. The erratic section B-C resembles the southern edge of Hens Wood (Site 8). At point I, Puthall Gatehouse formerly stood, regulating traffic and preventing animals straying along the main road (Figure 9). Conclusion The presence of a forest-facing ditch (to keep animals in) coupled with the gatehouse, suggests that this feature formed the Forest boundary as part of the pale containing deer and commoners’ animals. Its small scale might suggest a late origin; early pales often involved more massive earthworks (Watts, 1996). It is possible that it was developed from a pre-existing structure defining the extent of the King’s demesne in the Forest perambulation of 1299, or even the charter boundary of AD778. However, the parish boundary to the south-west provides an alternative alignment, as this formerly defined the edge of the extra-parochial land, equating with the King’s demesne. 14 Little Frith Linear Site 14, Figure 1. Points J to K on LiDAR image, Figure 20. Location This feature runs east-west on a high, level ridge on the Upper Chalk and gravels, immediately south of the valley of the modern A4. Description The feature extends for c. 370m from point J at SU 237675 at Horseleaze Wood, almost due east to point K at SU 243675. Largely running through semi- mature beech plantation, with semi-natural oak, hazel, ash and hawthorn nearer its eastern end, it serves no modern function. At point K a deep and prominent circular ditch (feature L on Figure 20) is partly eroded and obscured by pits to the north. This earthwork resembles a number of curvilinear and potentially pre-historic enclosures within Savernake (Lennon and Crow 2009). The linear feature consists of a minor bank and ditch, erratic in its course, with the ditch to the south. A sharp detour is evident at SU23856745 as if to avoid or encompass a pre- existing feature. At only 60cm high and 6m wide (Figure 65) the feature is frequently eroded and barely perceptible in places. An adjacent enclosure (L) has a substantial ditch with a low, narrow external bank and broader one within (Figure 65). This regular structure curves gently, with sharper curves at the SW and SE ends. The linear feature adjoins this enclosure on its western side but does not break through the outer bank; there is a lip across the ditch at this point. North of this point the ditch is shallower and the banks significantly diminished, possibly by livestock. Further east, Little Frith has a substantial coppice bank (M) measuring 96cm from ditch floor to bank-top (Figure 65). 208 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 21 Extract of Map of Little Bedwyn and Froxfield, 1846, showing Little Frith (Site 14), apparently amended to show boundary changes. Re-orientated to north. (WSA 3354MS). Research, documentation and correspondence with boundaries This feature is shown as the boundary of Puthall Farm on Charles Price’s map of 1717 (Figure 14)(WSA 1300-372) and the division between Puthall and the settlement of Littleworth near Timbridge (which presumably also included Little Frith). At the eastern end, the boundary extends south around the enclosure and is partly delineated on an 1820 estate map (WSA1300-375L). A map of 1846, upon which subsequent amendments have been coarsely drawn, shows the line of this feature crossed out, and another (approximating to the modern one) added, with a new track (Figure 21). This map also shows a further fenceline extending north of the enclosure, and probably along its western side. Little Frith coppice itself pre-dates 1549 when a court presentment accused Sir Edward Darrell of killing a buck there (WSA 1300-86). Observations Although Puthall features in the 778 Saxon charter (K133), this earthwork cannot be positively identified with that recorded in the bounds. The croft of Puthall was a regular feature in Forest perambulations. The land was granted to the Priory of Easton by Sir William Esturmy in the 14th century (WSA 9-15-243; 1300-6574). Littleworth, which seems to be bounded by this feature, might be related to Lytleferme, which was associated with 421 acres let to John de Havering in 1302, on condition that he enclose the land ‘with a great dyke 6 feet deep and 7 feet broad, with a hedge, and make the crest of the dyke so that the king’s deer cannot get out of the close but can enter without hindrance’ (Patent Rolls, 30 Edward I 1302). Little Frith coppice, presumably part of the Littleworth estate, has a rounded outline suggesting that it may have been enclosed within an undifferentiated landscape, although LiDAR shows that it encompasses a presumably earlier, unrelated rectilinear field pattern (Figure 20). Conclusion This inconspicuous linear feature is linked with a potentially early enclosure at its eastern end, although fieldwork suggests that it post-dates the enclosure and is essentially unrelated. The bank predates 1717 and might be connected with the establishment of the medieval croft at Puthall or the Littleworth enclosure. 15 Noke Wood Ditch Site 15, Figure 1. Points A to D on 1846 map, Figure 23. THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 209 Fig. 22 Noke Wood ditch (Site 15) looking north. Location This feature lies within Noke Wood, just south of the A4, 3km west of Froxfield. Description This small bank and ditch extends for c. 300m in a roughly north-south direction in the moderately sloping valley of the A4 and on a mixed geology of chalk, Clay-with-flints and alluvium (Figure 22). The ditch lies on the eastern side and forms the western edge of the woodland at its northern end, but otherwise serves no modern function. The wood consists of conifer plantation within former coppice, with relict hazel, young oak, birch and hawthorn. At its southernmost (uphill) point, where it is least substantial, it commences at the edge of the woodland, with no trace in the adjoining field. The feature is partially infilled c. 100m north of its starting point, where there is an S-shaped deviation in its alignment, while the eastern side is breached by a ditch 90m from the northern edge of the wood. Between this latter point and its terminus the feature is substantial, extending up to 14m across ~and 2m between top of bank and bottom of ditch (Figure 65). Research, documentation and correspondence with boundaries 3 Noke Wood is first recorded in 1547, when a Forest Court heard accusations that servants of Sir William Harbard had waylaid and killed a pricket with greyhounds there (WSA 1300-86). Lancelot (Capability?) Brown advised on various changes to the junction of the adjoining London Ride with the Turnpike (A4) including a sweeping entrance in 1767-68 (WSA 1300-1918; 1300-1920). The feature Fig. 23 Extract of Map of Little Bedwyn and Froxfield, 1846, showing Noke Wood ditch (Site 15). (WSA 3354MS). is Shown on an 1846 map of Froxfield and Bedwyn (WSA 3354 MS)(Figure 23, points A to B), which also shows a track extending below the wood to the main road (points B to C), corresponding with the boundary between Froxfield and Little Bedwyn parishes (points C to D). The feature is also apparent on the 1817 one-inch-to-one-mile Ordnance Survey map. Lennon (2010) contends that the Noke Wood Ditch resembles other cross-valley dykes in the wider Savernake area. Observations The substantial ditch with only a minor bank suggest that much of the material has been removed by natural or human agency. Conclusion This feature resembles an abandoned holloway, perhaps pre-dating the London Ride which may have been aligned with it. 16 East Wansdyke (New Buildings) Site 16, Figure 1. 210 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Location The reputed eastern terminus of Wansdyke is 2km south of Marlborough, and 1km west of Cadley. Brief Description This part of the Wansdyke is the eastern section of a structure traceable from Morgan’s Hill near Devizes, 19.5km to the west. This section lies within a wooded strip of oak and sweet chestnut, crossing the former Savernake Great Park west to east at a broad angle of 104° with gentle meanders (Figure 24). Approximately 20m wide and up to 3.75m between bank-top and ditch floor (Figure 65), there are differences in ground level either side of the earthwork. The ditch lies to the north and the feature is inconsistent in height and depth, diminishing in scale to the east where, at its terminus 200m south- west of New Buildings (SU 19596643), there is no certain ditch and the bank peters out. Research, documentation and correspondence with boundaries Wansdyke has been subject to extensive research and debate as summarised by Reynolds and Langlands (2006) and Lennon (2010b). Fox and Fox (1960) concluded that the Eastern Wansdyke was a separate earthwork to both the West Wansdyke in Somerset, and from the Bedwyn Complex (Site 35). Extensive archives of the Royal Forest of Savernake fail to mention Wansdyke extending beyond its current terminus, until antiquarian claims emerge in the 18" century. Conversely there is circumstantial place-name evidence that Wansdyke ended near its current location in the 14th century. A Roll of Henry Sturmy, Warden of Savernake Forest, of attachments and inquisitions of vert and venison up to the Festival of St Margaret the Virgin, in the 3 year of Edward ITI, [1329] made accusations of unauthorised felling of oaks between La Houkslade and Dychesende (WSA1300-59C). Whilst the evidence is not conclusive, La Houkslade appears to be the valley of Braydon Hook, before Great Lodge was built (thereafter known as Great Lodge Bottom), and Dychesende would therefore equate with the Wansdyke at New Buildings. Wansdyke carried a paling fence in the 17th century, presumably a subdivision of Savernake Great Park. Workmen were paid 2s4d for four days work ‘taking up the pale upon the Wednesditch’ in 1618 (WSA192-44). Crawford (1953) claimed that excavations of the Wansdyke in the 1920s had ‘proved that it continued beyond its present end’ near New Buildings. He reflected that ‘one would have expected Fig. 24 The meandering course of the Wansdyke (Site 16) across the former Savernake Great Park, looking west. to find it in the virgin tracts of Savernake Forest... but with one possible exception, one does not’. However, Crawford overstated the findings on which this conclusion was drawn. Whilst Brentnall (1924) asserted that ‘it seems clear enough that Wansdyke did not always end where it ends to-day’, excavation had merely demonstrated that it continued for a further 18 feet! Whilst he conjectured ‘If 18 feet, why not 36, 54, or 72 feet?’, the excavations themselves had failed to reveal more. Green’s (1971) excavations of the East Wansdyke near its terminus revealed buried soils yielding pollen from trees, grasses, pasture herbs, ferns and ruderal species (weeds). Trees were predominantly oak and hazel, with birch, alder, and ash, and traces of lime, holly and beech. There was a very little cereal and heather pollen. Green concluded that the location had been open country prior to the construction of the Wansdyke, with abundant woodland nearby. Sherds of Savernake Ware dating from the lst century AD were reported. Reynolds and Langlands (2006) provided a revisionist view of the Wansdyke, re-opening the prospect of it as a frontier defining a major polity, extending from Maes Knoll in Somerset, via the East Wansdyke and the Bedwyn System (Site 35) and on to Inkpen. Observations The ditch of Wansdyke contains many hundreds of sarsen stones, some of several tonnes, cast in from adjoining ground when troublesome to the plough. Such activity and ongoing animal burrowing have changed the morphology of the dyke. Interpretation of Green’s environmental samples requires re- evaluation as he deduced that the presence of tree pollen was evidence of woodland, whereas it is no more than evidence of trees. Recent ecological ee _ THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 211 and historical analysis has thrown doubt on the received wisdom that Britain was once cloaked in closed canopy woodland (Vera 2000). Certainly for Savernake there is abundant evidence of wood pasture, but dense woodland seems largely to have been confined to medieval encoppicement and modern planting. Whilst Wansdyke remains undated, with theories for its construction ranging from Roman to mid Saxon times, there is no evidence that this part of Savernake changed significantly until the Great Park was converted to arable in the 17th century. An Elizabethan survey describes the Great Park as ‘newly paled out of the Forest, all heath, green and ferny ground’, comprising 1650 acres of pasture and 300 acres of heath and waste (WSA 1300-90). The pasture included trees, with the survey, listing 1400 acres for the Great Park, Brimslade, and ‘the Forest Unpaled’ (modern Savernake) combined, suggesting little distinction between them. The former presence of trees in the Great Park is confirmed by a legal case of 1806 in which a witness, aged 76, said his great grandfather had remembered when it was possible to walk from Levett’s (a keepers lodge) to Great Lodge under their shade (WSA 1300-1853). Green (1971) notes that the Romano-British pottery kilns in Savernake would have needed a plentiful supply of timber, although such fuel need not have involved the cleaving of large trees. Analysis of charcoal from kilns at Alice Holt, Hampshire, for example, has shown that small diameter roundwood of almost any species was used, together with grasses and non- woody plants (Birkbeck et al. 2008). An understanding of the landscape prior to construction of the Wansdyke, may reveal why it did not continue beyond its putative terminus. The notion that it had entered dense woodland impenetrable to a military force has no ecological credence. An alternative interpretation is that the Wansdyke ceased on reaching wild and uncultivated terrain, possibly used as a common resource for pasture and fuel. Savernake is known to have been common land from at least the 12 century. The massive efforts necessary to define and defend territory in non-arable land might be unjustified. A similar situation may have prevailed at the micro- scale with the cross-valley dykes of Great Lodge and Shovel Bottoms (Sites 10 and 12). If Wansdyke had been ploughed beyond its current terminus it is difficult to envisage when this might have occurred. Its location fell well within the Royal Forest from the time that it first appears in the earliest surviving Pipe Roll of 1130, while it was part of the King’s demesne until, by Letters Patent, Edward VI conveyed Savernake to the Duke of Somerset in July 1547 (WSA 1300-166). After the Duke’s execution, and during the minority and later imprisonment of his son the Earl of Hertford, the land remained a parcel of the private Forest, until from c. 1575 it fell within land imparked as Savernake Great Park. The park was broken up gradually from 1618, and by 1651 was all agricultural land (WSA 9-22-14/22/24/25/28/33/206). The first detailed map by Charles Price in 1717 (WSA1300-372) bizarrely fails to show the Wansdyke at all. Its course fell across two holdings, whose ‘homestalls’ were shown in the locations now occupied by Wernham’s and High Trees Farms. The unremittingly regular field pattern, all recorded as arable, did not align with the Wansdyke for much of its course, and it awkwardly bisected three fields, with the easternmost one (where it currently terminates) called Wansditch Field. This field pattern is little altered today. Written sources for Savernake do not suggest that the Wansdyke itself was ploughed in modern times. By this reasoning, the most likely opportunity for ploughing is prior to 1130, although heavy clay soils, difficult terrain, and extra-manorial and extra-parochial status, militate heavily against the prospect. The unlikelihood that the Wansdyke ever continued through Savernake does not preclude the possibility that it may have been part of an ongoing frontier, as envisaged by Reynolds and Langlands (2006). Offa’s Dyke is also now seen as a discontinuous earthwork, incorporating gaps across woodland and other topographic features. Pitt Rivers (1892) similarly observed that the Wansdyke ‘diminishes 1n size, or 1s wanting, in places where forests may have existed, and that it increases in places where forests are unlikely to have grown’. He contemplated that within wooded country ‘the earthen mound and ditch may have been replaced by an abatis of felled trees, no trace of which remains’. He acknowledged that this idea ‘must be taken for what it is worth, 1n the present state of our evidence’. Current ecological thinking suggests that it is unlikely that trees could have then grown so densely (and been so expendable) in Savernake to permit construction of such a barrier, or to maintain it thereafter. Some other form of fencing or dead hedge could have replaced the earthwork, although it may be noted that court records for Savernake report endless theft of deadwood from such structures. Conclusion — Available evidence suggests that Eastern Wansdyke did not extend beyond its current terminus, near 212 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE _ New Buildings, where it probably ceased within open wood pasture. The theory that it extended beyond this point could be better entertained by envisaging major caesura in its course from this point, rather than its construction and subsequent loss. 17 Bagden Complex A series of related earthworks mapped as Site 17, Figure 1. Points A to H on LiDAR image, Figure 25, together with a temporary diversion from this line, points E to G, and a nearby pair of features points J to K. Location The line A to H runs within the heart of modern (wooded) Savernake, almost parallel with the Grand Avenue (points N to O), aligned north-west to south- east, and straddling Bagden Lawn at SU 232664. The earthworks incorporate part of the Roman Road running from Cunetio (Mildenhall) to Venta Belgarum (Winchester) together with features which gradually diverge from this alignment. Description The valleys of Bagden Lawn and Drury Lane have chalk and alluvial deposits, whilst the generally wooded higher ground alongside overlies the Clay- with-flints. Section I, Woolslade and Bagden Coppice: points A to B (Figure 25). This section extends on gently sloping, mainly wooded land at right angles to the valley in which the Lawn is situated. It comprises a wide series of minor and inconsistent banks and ditches along a straight course (Figure 65) running on the same alignment as section II (see below), but not certainly connected with it. Section II, Bagden Lawn: points B to C (Figure 25). This earthwork comprises a consistent, wide and deep ditch which runs across Bagden Valley, with two almost straight but not exactly aligned sections. The ditch is flat bottomed in part. Across the Lawn itself, where it divides fields, there is a modest bank along the NE edge, with a substantial ditch 7m wide and 1.2m deep. To the south-west the ground dips away sharply by 1.6m (Figure 65). A line of mature oaks along the NE edge (Figure 26), includes a veteran oak 6.6m girth. The ditch extends beyond Fig. 25 LiDAR aerial image of Bagden Complex (Site 17 sections I-VII), 2006. © Forestry Commission. THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 213 Fig. 26 Bagden Lawn (Site 17, section II) looking north west. Veteran oak alongside single ditch seem to represent the southernmost of a double line of trees present in 1786. the pasture and into woodlands either side, where there are major head-walls blocking movement along the ditch floor. Section III, south-east of Bagden Lawn: points C to D (Figure 25). In the woodland south-east of the Lawn, the ground level changes across an area 30m wide (Figure 65). Alongside the ditch which extends from the lawn, the ground level has been reduced by nearly 3m, in a wide funnel-shaped depression visible north east of point C (LiDAR Figure 25). A considerable quantity of spoil appears to have been removed from this area. Section IV, Woodland: points D to E (Figure 25). This feature continues the alignment of A to D through gently sloping woodland, but without the massive reduction in ground level characteristic of Section III. The total width exceeds 30m, and comprises (from SW to NE) a ditch, bank, further ditch, a small secondary bank, an apparently levelled area and a step to a lower platform (Figure 65). Whilst the alignment remains constant, the size of banks and ditches is variable. Section V, Curving Bank besides Drury Lane: points E to G (Figure 25). This deviates from the strict alignment of previous sections, and comprises a single curving bank and ditch, 1m high from ditch- floor to bank-top (Figure 65), either side of the drive to Savernake Lodge (Drury Lane), which projects as an arc towards the south west. At point G it joins the bank of Birch Copse. Section VI, Linear Hollows through Birch Copse: points J to K (Figure 25). These extend across gently sloping conifer plantations east of the Savernake Lodge drive. At least two ditches, up to 1m deep, with no obvious banks (Figure 65) Puthall Gate 4 Mile Stone Stock # Common Sadiers Stock Corner <" Chisbury Lane Fig. 27 Sketch map by fohn Ward 1795 (redrawn) showing the Old Marlborough Road through Savernake before closure (black arrows, left) and a new proposed route to the London Road (grey arrows, right). (WSA 1300-363A&B). are regular in direction and depth, but of variable width. Section VII, Roman Road, south east of Drury Lane: points G to H (Figure 25). This section continues through plantation woodland and follows the broad alignment of A to G. It comprises a large ditch, 1m deep and 10m wide, with no bank on the SW side (Figure 65); an apparent bank to the east may result from erosion associated with a nearby path. 214 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Research, documentation and correspondence with boundaries Three 18th century maps reveal a convergence of features at Bagden. The old Bedwyn to Marlborough Road passed along the course from at least points A to G, and possibly along the hollows of J to K. Its route is shown on a rough map sketched by estate steward John Ward (Figure 27)(WSA 1300-363). The map was prepared when planning the closure of the road in association with the construction of Savernake Lodge in 1795 (WSA 9-35-261). This was a grand mansion to a design by Sir John Soane, encasing and enlarging a former keeper’s lodge. The mansion burned down inc. 1861 (Anonc. 1900). An estate map of 1786, prior to the building of the new house, shows the road crossing Bagden Lawn, with trees on each side (WSA 1300-360), forming the boundary between Woolslade and Bagden. Alongside the construction of the Lodge, the entrance drive may have been made more imposing with a curved bank, as shown on a coarse, undated map (Figure 28)(WSA 9-1-509). The Roman road to Winchester runs through the complex, although on a slightly different alignment (142°, cf 139° for the ditch across Bagden Lawn). Observations and Conclusion This complex represents a convergence of imperfectly aligned and otherwise un-connected features — a Roman Road, the old Bedwyn-Marlborough Road, coppice banks and landscaping for Savernake Lodge. Section VII represents the Roman Road, which is unaffected by these later works. The arc-shaped bank (Section V) is associated with the Lodge. Section VI may represent scouring along the old Marlborough Road, or possibly a mineral excavation. Sections I and IV have earthworks relating to both the old Marlborough Road and the Roman Road, as well as bearing the marks of erosion. Section I also has coppice banks and ditches, Section II has been subject to more substantial change. The 1786 map shows a double avenue of trees where the old Marlborough Road crossed Bagden Lawn on a causeway, perhaps constraining traffic. The south- western ditch survives, flanked by oaks, of which some would have been present in 1786 (the largest oak there has a girth commensurate with a tree 400 years old). If there was a corresponding north- eastern ditch this has been infilled, and its trees removed. Surviving oaks may have been retained in the grounds of the Lodge and it is probable that the ditch served as a ha-ha, of which several were constructed at nearby Tottenham Park. An 1820 oo x IG : CEB eee Sonne SS Fig. 28 Coarse sketch showing landscaping and changes to the entrance to Savernake Lodge, probably late 18th century. (WSA 9-1-509). estate map records ‘sunk fence’ across Bagden Lawn at this point (WSA1300-375L). It appears that there was substantial raising and levelling of ground to the north east across Bagden Lawn (Section II), perhaps in association with landscaping for the mansion, turning rough pasture into a park lawn. The head walls at each end of the ditch across Bagden Lawn are similar to the cross-valley dykes of Shovel Bottom and Great Lodge Bottom; hence the Bagden Lawn ditch might have augmented an earlier feature. None of this complex is likely to be associated with defensive or large-scale territorial definition. 18 Bishop’s Rag Dyke Site 18, Figure 1. Points A to E, LiDAR image Figure 29. Location This feature is situated in Birch Copse, near the north eastern edge of modern Savernake. THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 215 Fig. 29 LiDAR aerial image of Bishop’s Rag (Site 18) and Holt Pound, 2006. Arrows show course of parish boundary ditch and Forest boundary known as the ‘Mereway’ through Bedwyn Common. © Forestry Commission. Description The earthwork is on level or gently north-sloping land, close to the boundary between the Clay-with- flints of the forest and the chalk of the agricultural land to the north east. A single bank and ditch with the ditch to the south west (facing the forest) runs for 500m in dense conifer plantation from point B (SU 24456660) to point D (SU 24656620)(Figure 29). It maintains a broadly direct course of 145°, with localised gentle meandering. The bank height is irregular, as if tree roots have locally prevented erosion of friable soils. At its greatest the feature is 7m wide and 84cm from ditch floor to bank-top (Figure 65). The line of the ditch appears to continue indistinctly on a more northern course some 170m from point B to A along the edge of a residential property (but see below) and as a surface feature in woodland 200m south east (D to E) where it becomes lost on leaving the woodland. At D the feature is joined by a bank extending to the east. A significant opening where the earthwork appears to make way for a ride (point C), this continues east as a holloway. Other gaps appear to be breaches caused by modern forestry vehicles. 216 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Ste BROS . o Timbridge ~ Common Ground py j Uf, MOVIIV A ty WY fre py Y : Fig. 30 Extract from a Survey of Knowle Farm surveyed by Charles Price, 1717 (relabelled) showing Bishop’s Rag (Site 18) and Belmore (Site 19). (WSA 1300-372). Research, documentation and correspondence with boundaries Two Saxon charters involving adjacent land, one of 778 concerning Little Bedwyn (Sawyer 1968, Cat. No. 264), the other a document of 968 covering land in Great Bedwyn, Grafton and Burbage (K1266), might relate to this feature as both use identical perambulation points west from Horsel Geat via Rames Dene Geate, Holhrygc Gete, Hadfeld Geate, to Baggan Geate (Grundy, 1919). Whilst intermediate stages cannot be identified, Horsel appears to equate with Horsehall, near SU 266668, and Baggan with Bagden, the valley just north west of Bishop’s Rag. Bishops Rag was mentioned in 1549 when a court book summoned John Waren for slaying a ‘doe in milk’ there by bow and arrow (WSA 1300-86). The bank is the boundary between the Forest and Knowle on a 1717 map by Charles Price (Figure 30), when it marked the edge of 20 acres of Bishop’s Rag woodland (WSA 1300-372). The ride here is labelled on the 1717 map ‘Road from Pewsey to London’. A Forest estate map of 1786 shows this ride labelled “To Ramsbury’ and with Bishop’s Rag sub-divided into strips (Figure 31)(WSA 1300-360); this is presumably the tree nursery recorded at Birch Coppice in 1785 (WSA 1300-1953). The line of the feature also shows on an estate map of 1820 (Figure 34)(WSA1300-375L) and maps of Little Bedwyn and Froxfield of 1846 (Figure 32)(WSA 3354MS). Observations The gap in the dyke for the Pewsey to London Road could be that mentioned in 1547 when a servant of Edward Seymour, Duke of Somerset, killed a buck with his greyhounds in ‘Le Sowthfyld near Bishops Gate’ (WSA 1300-86). The 1716 map shows a gate at bg: MOE EPP d tg, ey i Af 5 « pag & pln 2B i” py, ae, pt SN aS OO Fig. 31 Extract from a Plan of the Forest of Savernake and Tottenham Park (showing Bishop’s Rag — Site 18) prepared for the Earl of Ailesbury 1786. (WSA 1300-360). this point (Figure 30). The 1846 map may show the line of the earthwork formerly curving northwards and continuing to Timbridge Farm. Amendments to this map indicate subsequent removal of boundaries (Figure 32). The same map shows a line continuing south to Holt Pound (point F on Figure 32) and then via Holt Coppice te the north-east corner of Bedwyn Common. If this continuous cartographic line represents a single feature, then this ditch was the division (successively from west to east) between Timbridge and Littleworth, between the Forest and Knowle, and the Forest and Chisbury (WSA 1300-372). A section of it appears in a 1300 perambulation of Savernake (the Mereway or Merryway)(WSA 1300-63), the division between Chisbury and Bedwyn Commons (WSA 9-7-42), and was used when the single Bedwyn parish was split into Little and Great Bedwyn in c. 1547 (Crowley 1999). Significantly, the western part did not bound the edge of extraparochial land in the 19th century, which is likely to have defined the King’s holding. At the far northern corner of Bedwyn Common at SU2 63660, the parish boundary meets the same ditch which runs through Chisbury Wood (Park Copse). The ditch could therefore be the one referred to in a legal case reciting the bounds of a sheep common in 1681, which went ‘along a ditch to Holt Pound and so along the same ditch to the Lords Way leading to Knowle Farm’ (a point half way along the boundary of Bedwyn Common) (WSA 9-3-388). The dyKe at Bishop’s Rag, like that in Chisbury Wood Park Copse THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 2 Fig. 32 Extract of Map of Little Bedwyn and Froxfield, 1846 (relabelled), giving later amendments, and showing Bishop’s Rag (Site 18). (WSA 3345MS). (Site 26) has a south facing ditch. At the northern end it is unclear whether the feature around the residential property known as Owl’s Castle (A to B on Figure 29) is a modern deviation. This could be the edge of Timbridge Coppice shown on the 1786 map (Figure 31). The same map, which shows no property here, appears to show a line continuing along the same course to Crabtree, suggesting a much longer feature. Conclusion The Bishop’s Rag earthwork may be up to 2.4 km long. If related to the Chisbury Wood (Park Copse) feature, it would measure a further 1.6km, longer again if continuing through Crabtree. It is likely that much of this length represents a Forest pale with its ditch on the Forest side to contain deer. It might be an adaptation of an earlier structure, demarcating a significant landholding, perhaps related to the 778 and 968 charter bounds, and land held by the King. 19 Belmore Site 19, Figure 1. Options for its course are shown on LiDAR image and 1946 aerial photograph, Figure 33. Location This site lies immediately east of Savernake, 800m south of Timbridge Farm. Description One part of this earthwork extends 0.35 km straight across agricultural land in a shallow chalk valley from Birch Coppice, at SU 247663 (point A on Figures 33 and 34) to Belmore Copse at SU 250664 (B). Here there is a simple step in ground level between two arable fields, with the higher land to the south. Midway along this course it 1s joined by another feature (E to F), discernible as a soil mark. From point B it penetrates into the wood to point C, and maintains a meandering course some 0.25km near the edge of Belmore Copse to Sicily Cottages THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 18 2 Lower ission. ry Comm ~ © Forest . 2006 isbury Common (Site 20), isbury Common (S ite 19) and Ch rial image of Belmore (S iDAR ae 1 photograph 1946 33 Top L 2. 1 , CPE/UK, 1821) F2112 . (RA ite 20) h 19) and C. ute ing Belmore (S show Aeria THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 219 (Holt Coppice) © a Fig. 34 Extract from Savernake Estate map c1820, showing bank marked Wansdyke (here marked with black arrow and re-labelled) equating with section E-F on LiDAR image and aerial photograph of Fig. 33 (WSA 1300-375L). It also shows the Bishop’s Rag dyke (Site 18), here marked with white arrows. SU 253663 (D). Within the copse, which is situated on drift deposits, the feature has a mounded bank and ditch (Figure 35). Its alignment is variable and inconsistent even over short distances, incorporating a sharp, almost 90° curve part way through its course. At its greatest, the height from ditch to bank-top is only 84cm, and the feature just 8m wide (Figure Fig. 35 Bank and ditch within Belmore Copse, but not forming the boundary of the wood, and with the ditch on the wrong side to keep animals out of the coppice. Looking east. 65). The earthwork here has no function, and lies close to, but never forms, the coppice boundary, and indeed has the ditch facing north (on the wrong side to exclude animals). It meanders within or just outside Belmore Copse, where mature oaks, both sides of the modern fence, suggest there was once another wooded area to the south. Research, documentation and correspondence with boundaries The area is shown on a map of Chisbury prepared by Thomas Atkinson in 1719 (WSA 1300-372) (Figure 36). Part of the line may accord with the northern part of an area marked Holts. The feature E-F-B-C-D is shown labelled Wansdyke on a Savernake estate map of c.1820 (Figure 34). This may derive from Colt Hoare’s observations. Following a suggestion by the Earl of Ailesbury, he had inspected the bank in this location, concluding that it bore ‘the usual character of the Wansdyke’. His map in Ancient Wiltshire is too small to give a precise delineation but appears to accord with the same route. In his volume ‘Excavations on Bokerly Dyke and Wansdyke’, Pitt-Rivers (1892) mapped the Belmore section as ‘Wansdyke’, but did not envisage that this had ever been connected physically to the Wansdyke at New Buildings. His drawing of the profile (the only one known) is labelled ‘section of Wansdyke 1n a ploughed field to the east side of Savernake Forest and to the west of the road from Bedwin Common to Knowl’. It shows little more that a modest change in ground level of 4ft Sin Chisbury Common _” Fields inclosed 1703 , =~"Ghisbury Common r - Fields inclosed 1703 0 “NS Fig. 36 Extract from a Map of the Manor of Chisbury surveyed by Thomas Atkinson, 1719 (relabelled) showing edge of Belmore (Site 19 — see also Fig 30) and Chisbury Common (Site 20). (WSA 1300-372) 220 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE (1.3m) across a 50ft (15.2m) stretch, with barely any discernible bank or ditch. Pitt-Rivers suggested that the earthwork continued to Chisbury hillfort, but he proffered no course. Crawford (1953), reflecting that traces of the Wansdyke ought to be visible in Savernake, considered that the bank (E to F), which he thought continued into Belmore Copse (B to D), was the only discernible contender. He marked it with a question mark on a map, but concluded that it was unlikely to be the Wansdyke. Major and Burrow (1926) accepted that the Wansdyke crossed this valley. They considered that the diagonal feature E to F continued beyond (and was unrelated to) the bank running across the valley A to C, but that either of these might be the Wansdyke. Fox and Fox (1960) reported that the bank E to F had formerly been impressive at a distance but was since ploughed out. They considered that it might be the accidental product of levelling after quarrying chalk. Howe and Hostetter (1997), whilst noting the doubt expressed by Crawford, considered that its validity should not be completely discounted. Reynolds and Langlands (2006) noted earlier claims that this site was part of an ancient dyke system, and the doubts expressed by both Crawford and by Fox and Fox. Observations The straight feature (A to B) across agricultural land in the valley, appears to be a simple lynchet. However, it is joined by the diagonal mark E to K which in turn is probably related to the bank and ditch in Belmore. The latter is certainly a deliberately constructed feature. Its size is unremarkable and it resembles many local coppice banks. Here however, the ditch is on the wrong side to keep animals out of Belmore. It may have been a boundary to prevent access of animals into the land known as Holts (Figure 36). This was formerly open to the Forest and considered part of it, and deer at one stage wandered from the main block, through Holt, and on to Bedwyn and Stock Commons (also part of the Forest). However, when Chisbury Common was inclosed in 1703, tenants of the new arable land complained to Lord Ailesbury about deer entering via Bedwyn Common. As communication with the rest of the Forest had already been constrained to a narrow neck of land near Holt Pound, Lord Ailesbury decided to drive the deer entirely from Chisbury, Stock and Bedwyn Commons, and prevent their re-entry by a new gate erected near the pound (WSA 9-3-388). Holt was a regular perambulation feature of the Forest from the time of Henry III. The earliest record of Holt is in a document that mentions Michael de Columbarius (d.1235). The bounds of the Forest are recorded as passing along a hedge between the arable and the pasture to the wood of William de Holt (WSA 2667-24-5-1). Holt Coppice was recorded on Atkinson’s map of 1719 and was grubbed up sometime after 1846 (WSA3354MS). Land at Timbridge to the north, probably included the pasture called ‘7Zymerruggedoune’ which formed part of the holdings of the hereditary wardens of the Forest referred to in an Inquisition Post Mortem in the time of Edward I (Fry, E.A., 1908). Conclusion This feature may be part of an enclosure, possibly to keep animals out of a licensed purpresture. It might incorporate existing lynchets. There is no evidence that it formed any larger territorial or defensive boundary. 20 Chisbury Common Site 20, Figure 1. Options for its course are shown on LiDAR image and 1946 aerial photograph, Figure 33. Location The earthwork is 400m north of Bedwyn Common. Description This feature extends from points H to L for 600m across the former Chisbury Common from Sicily Cottages (SU 253663) to Lower Horsehall Cottage (SU 258665), within the parish of Little Bedwyn. It occupies an open arable landscape, which is level or gently sloping to the northeast. The farmland here is situated on drift deposits of alluvium and gravels, with an abundance of flints. The earthwork comprises a simple but large change in ground level marking the division between two fields with the lower land to the south-east. Straight, although variable in height, there is no bank or ditch. Near Lower Horsehall the change in ground level is at its greatest at 1.59m (Figures 37 and 65). Research, documentation and correspondence © with boundaries This earthwork carries a pedestrian right-of-way and is close to a feature conjectured by Major and Burrow (1926) to bridge the gap between Belmore Copse, where a small bank is visible, and Lower Horsehall THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 221 Fig. 37 Former Chisbury Common, near Lower Horsehall, looking south west. Hill (D to L), where a meandering track was considered to mark the ongoing course. Fieldwork undertaken by Brentnall and Gidney in the early 20th century, and accepted by Major, suggested that a long distance feature extending between Savernake and Chisbury divided near Sicily Cottages, with one route extending south through Sicily Clump. Observations The north-eastern end of this site connects with Horsehall, whose name is echoed in Horsheal geat, a perambulation point in the 778 and 968 charter bounds considered above (Grundy, 1919). The change of ground level here, and the speculative line of Major and Burrow, are situated within the former Chisbury Common, an area of 144 acres north of Bedwyn and Stock Commons. Although frequented by deer and probably largely indistinguishable from the adjacent (extra-manorial) Forest, Chisbury had not been part of the demesne land of the King, and was not encompassed within Forest perambulations after the massive disafforestation of 1330. In the mid 16th century rights of pasture were held by Monk’s, Holt’s and Horse Earl of Hertford’s Hill Farms (WSA 9-1-95). Sir Edward Seymour, son of Protector Fig. 38 Fohn Rawlins map of 1703, for the inclosure of Chisbury Common so that the holders would have ‘their just proportion of ground’. (WSA 9-8-26 to 28). Somerset, purchased the Manor of Chisbury in 1602, and brought it within the Savernake Estate (WSA 9-1-287). At the end of 1701, articles of agreement were drawn up between Thomas, Earl of Ailesbury, as Lord of the Manor, and five customary tenants, appointing commissioners to lay out paths and highways, and to divide the common so that all the holders would have ‘their just proportion of ground according to their lawful interest of common’. John Rawlins, a surveyor from Boxford was appointed, and presented his results in 1703, Figure 38 (WSA 9-8-26 to 28). The division between the fields accords with a highway established following Rawlins’s recommendation. The former Chisbury Common is shown on Atkinson’s map of Chisbury of 1719, annotated with details of its inclosure (Figure 36). Conclusion There is no evidence to suggest this is a deliberately constructed feature and it more likely to be a lynchet that developed after the inclosure of 1703. The area has abundant lynchets, visible on aerial photographs and recorded by the English Heritage National Mapping Programme. 21 Warren Ditch Site 21, Figure 1. Site B to C on LiDAR image, Figure 39. Location | This stretch of ditch is situated alongside the north-western edge of the Durley to Bedwyn Road, 222 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE a ee immediately north-east of Warren Farm, at SU 248655 (point A). Description This feature is on chalk soils with the ground level sloping gently north east to Holt Pound. A broad, deep ditch, orientated at 040° supports a narrow woodland strip of mature ash, elder, hazel and birch between the road and a farm track. The ditch has probably been considerably altered by road construction. There is no obvious bank (which would have lain partly under the modern road), and the feature is largely visible as a change in ground level, with the lower land to the north- west. At its deepest it is 1.56m (Figure 65). The feature may continue south-west of Warren Farm, where there is a significant ditch on the opposite side of the road, partly on Clay-with-flints (L to M on Figure 39). The original ground level here is difficult to determine due to a major explosion of stored wartime ammunition on 7 July 1945 (WSA 3345 Boxes 3-5). Research, documentation and correspondence with boundaries This ditch does not feature on maps of Savernake, nor does it correspond with parish or other boundaries. A 1786 estate map shows a short track along this course, but which does not then continue as a road to Durley (WSA 1300-360). The Warren had been a part of Havering’s Heath, an area of the Forest which included Bedwyn and Stock commons. A rabbit warren was established there from at least the 16th century, at a place called Blackfault (WSA 9-22-42). There was only one building, occupied by the warrener, who was also the keeper for Havering’s Heath (WSA 9-7-42). In 1582 this was William Applegate (WSA 1300-87) and the Lodge was sometimes known by his name. In the 18th century The Warren was at the heart of legal disputes, as conflicts arose concerning whether this land was within the Forest or neighbouring manors, whether it fell within a parish, whether it was common land of Durley, and whether it could continue to claim Forest exemption from tithes once ploughed and converted to agriculture. One of the legal cases recited the bounds of asheep common of Stock Farm made in 1681, which went from Straight Corner (H on Figure 39), to Stepstile (J), along a green way to Applegate Lodge, to the three rabbit burrows, and ‘from thence along the Ditch to Holt Pound [D] and so along the same ditch to the Lords Way leading to Knowle Farm [EJ]. It then went along the Cross Ways (the division between Stock and Bedwyn Commons) to Straight Corner (WSA 9-3-388). The Warren bounds extended beyond the sheep common at Knowl Gate ‘via the Merryway [the ditch dividing Chisbury Common and Great Bedwyn Common] to Bewley Gate’ (point F on Figure 39, the north eastern point of Bedwyn Common) (WSA 9-7-42). A similar route was taken in reverse by residents from Burbage when they sought to incorporate The Warren within their parish by a process called ‘possessioning’, as reported in 1708. Their march went from Holt Pound to Applegate Lodge and then on to Stepstile (WSA 9-7-42). Observations If Applegate Lodge was situated on the same site as the current Warren Lodge and Warren Farm, this provides prima facie evidence that the Warren Ditch was ‘the same ditch’ which continued to Chisbury Lane and beyond. However, as the site of Applegate Lodge and the three rabbit burrows is unknown, the ditch referred to in the 1681 Sheepslade perambulation might have reached Holt Pound from the north. It is also significant that the ditch at Warren Farm is on the north-eastern side of the feature, whilst that at Chisbury Wood is to the south west. Conclusion The origin of this feature is unclear. It might be deliberately constructed or derived from a lynchet, or even an artefact of modern road construction. Documentary sources indicate that either Warren Ditch, or another ditch entering Holt Pound from the north, are part of a longer distance earthwork, passing on to Chisbury Wood, possibly continuing to Great Bedwyn. 22 Bedwyn Common Mound Site 22, Figure 1. Shown with line as point G on » LiDAR image, Figure 39. Location The site is located at the far north-eastern corner of Bedwyn Common. Description The earthwork consists of a simple mound 31m long, 18m wide and 130cm high, orientated at 070° with no certain ditch or source for the material. It is on almost level ground derived from alluvial gravels, at $SU261659. The feature is marked on the 2004 SS ee = —— THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 223 Fig. 39 LiDAR aerial image of The Warren, showing Warren Ditch (Site 21), and location of Bedwyn Common Mound (Site 22, marked G), 2006. Arrows show course of parish boundary ditch and Forest boundary known as the ‘Merewavy’ through Bedwyn Common. © Forestry Commission. edition Ordnance Survey 1:25,000 map as a ‘mound’, in contradistinction to tumuli marked 200m south east. There are gravel pits just below the site. An overgrown track, parallel with the nearby fence, crosses part of it. Rabbit burrows reveal mixed gravel and clay. Scattered bottles, brick and pottery, suggest it may have served as a rubbish heap. Research, documentation and correspondence with boundaries Crawford (1921) interpreted this feature as a small but perfect fragment of a hitherto unrecorded section of Wansdyke. Having forgotten about the feature, he stumbled across it again thirty years later and entirely rejected his earlier claim, although he felt sure it was an ancient feature (Crawford 1953). He concluded that a long barrow was the most plausible explanation whilst expressing concern about the absence of a ring ditch. Howe and Hostetter (1997) stated that if its “existence cannot today be ascertained [its] validity [as part of a possible dyke] should perhaps not be completely discounted’. Observations The feature is close to an ancient perambulation of Savernake which extended ‘along the Mereway’ (the division between Bedwyn and Chisbury Commons) (WSA 212B/7189; 9-3-388). It is close to Belling Gate or Bewley Gate, at the north-east corner of the Common (which was part of Savernake). Here a lane from Marlborough to Hungerford formerly crossed from west to east, and continued through Chisbury Wood (WSA 1300-372). The earthwork is also at a junction where land was exchanged between the parishes of Great and Little Bedwyn sometime before 1753 (WSA 9-3-398). The possibility of the feature being an ancient boundary marker was recorded in 1760 during a legal dispute about whether Bedwyn Common was part of Savernake Forest, or within the adjoining manor called Stokke, Ford and West 224 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE | Bedwyn. Lawyers defending a suit of trespass against Lord Bruce claimed ‘we shall be able to show a high mound at the extremity of Bedwin Common against Lord Verney’s lands which appears to be the natural bound and extremity of the Forest’ (WSA 9-3- 388). A mound called the Holtball featured in many medieval Forest perambulations (e.g. WSA 1300- 65). Related perambulation points tend to suggest it was 1km west, near Holt Coppice (SU252658), although the Holtball has never been located. While it is possible that this feature represents it, there may have been a number of mounds or balls used as perambulation markers. The nearby gravel pits lie downhill, and are unlikely to have provided spoil for the mound, although there may have been further pits in the field to the west, now infilled and ploughed level. Accordingly, a prosaic explanation remains possible. Conclusion This enigmatic site could bea tumulus, perambulation marker, spoil heap or other feature, but is unrelated to any long distance territorial or defensive boundary. 23 Chisbury Wood — Park Copse Site 23, Figure 1. Points E to H on LiDAR image Figure 40, and 1946 aerial photograph, Figure 41. Location This feature is 500m south-west of Chisbury Hillfort. Description This earthwork occupies high, gently south-easterly sloping ground, and crosses the whole of Chisbury Wood. It extends some 900m, commencing along the edge of a small pasture, point E on Figures 40 and 41 (SU 269660), and continues through the wood from point F along a straight course, south east, to point H (SU 277654). A significant single bank and ditch, with the ditch to the south-west, varies in size either side of a major forestry track. From Point F to G, situated on Tertiary deposits, it is conspicuously large, extending 10m across and 1.56m high from ditch floor to bank top (Figures 42 and 65), although it is somewhat inconsistent in structure, sometimes eroded on the bank and varying in depth. Although its overall course is straight, at a local level it meanders, with a significant S-shaped bend towards the south-east. From Point G to Point H, where the feature often overlies chalk, it is smaller and less sinuous, up to 8m wide and 1m high. The feature ceases upon reaching the field below the wood where it is imperfectly aligned with a hedgeline (H to I) along a lynchet continuing south east, although here, in contrast to the earthwork, the lower side is to the north (by 1.5m). Within the wood the feature is crossed by modern forestry tracks and an avenue of semi-mature beech. Research, documentation and correspondence with boundaries Despite its scale, little attention has been given to this feature. The earthwork has variously defined the divisions between Chisbury Manor and Bedwyn Prebendal Manor (Figure 43)(WSA 1300-372), Park Copse and the copses of Bewley and Bar Field Copse and formed the boundary between Great and Little Bedwyn parishes, when Bedwyn parish was divided some time before 1554 (Crowley 1999). Observations There is evidence that this earthwork formerly extended significantly to the west. A hedgerow marked its course in 1946 (points D to E, Figure 41). This in turn is connected with a ditch which ran from Holt through Bedwyn Common, also a 16th century parish division, and a perambulation mark of Savernake known as the Mereway (= boundary way). It is possible that the earthwork in Chisbury Wood is part of a line extending at for at least 2.8km. The feature might also partly coincide with the bounds of the 778 and 968 charters considered above. Going west from the Bedwyn Brook, the former bounds ran by the long thicket (spznetz) at Horselget (= Horsehall Gate) to Ram’s Dean (unidentified). The later charter travels in the opposite direction, using the same perambulation features. Thus the earthwork in Chisbury Wood (which may indeed be the thicket of 778) is on an alignment corresponding © with these bounds. Park Copse, recorded as 40 acres in 1719 (WSA 1300-372), may represent the 40 acres of woodland recorded in Domesday for Chisbury. It seems likely that the Copse is on the site of the former Chisbury Park. In 1261 the King licensed Mathew de Columbarius, Lord of the Manor of Chisbury, to take four bucks and four does from the Forest of Savernake for stocking his park (Close Roll, 1259-61). The park had ceased to be managed for rearing or hunting deer by 1547, when a survey described it as a pasture of an estimated 40 acres, THE LARGER LINEAR EARTHWORKS OF SAVERNAKE ee Bn ty Sp tii ti YY ee SO Fig. 40 LiDAR aerial image of Chisbury Wood, showing Park Copse (Site 23), The Lawn (Site 24), The Scar (Site 25) and Chisbury Wood North (Site 26), 2006. © Forestry Commission. let for its ‘herbage, pannage, pasture, grassing and feeding’ for 40 shillings. The tenants also purchased the underwood and scrub in the park, but the Lord of the Manor retained great oaks and timber trees (WSA 9-1-5). There is no mention of a separate Park Coppice in 1547, although a map of 1612 (Figure 44) records Park Coppices (plural) measuring 60 acres, of which 50 were wood (WSA 9-8-153H). In 1719 Park Coppice was again listed as 40 acres, whilst Lawnfield comprised 10 acres of arable. Despite the clear association between the park and the coppice, the main earthwork passing through Chisbury Wood is not ideal for enclosing deer as the ditch is external. It is possible that the park was established in existing woodland, using available banks and dykes, rather than being constructed anew. Conclusion The earthwork describes part of the boundary of Chisbury Park, which is itself perhaps based on an earlier woodland, but it extends beyond. It can be related to an earthwork recorded in the 13th century, while it denoted manor and parish divisions. 22 6 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE . Fig. 41 Aerial photograph Chisbury Wood, 1946, after wartime clear-felling, showing Park Copse (Site 23), The Lawn (Site 24), The Scar (Site 25) and Chisbury Wood Site 24, Figure 1. Points K to M on LiDAR image 24 Chisbury Wood - The Lawn Figure 40, and 1946 aerial photograph, Figure 41. | Chisbury Wood, 200m south west of the hillfort. Description Fig. 42 Chisbury Wood, Park Copse bank looking east, ditch The earthwork runs for 380m across gently south- Location | This feature is in the north-eastern corner of | to south. sloping Upper Chalk, close to the junction with THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 227 Tertiary sands and gravels. It commences at point K (Figures 40 and 41), where it connects with a small straight bank (J to K) extending south-west to a much longer earthwork (Site 23). The feature curves ESE to point L where it curves sharply through almost 90° heading NNE and ceasing at the edge of the wood. At its greatest it is c. 8m wide and 1m from ditch floor to bank top (Figure 65). The ditch is to the south east. Research, documentation and correspondence with boundaries This inconsequential feature might be disregarded as acommonplace local bank but for its inclusion on Atkinson’s 1719 map of Chisbury under the name ‘Woden’s Ditch’ (Figure 43). Atkinson shows open land here and continues the line NNE along the edge of a small field below the hillfort, which turns east through a right angle parallel with the hillfort; it joins another dyke which extends due south from it. All three sections are labelled Woden’s Ditch. The feature is also shown on a map by George Wyther, of Standen, Chisbury and Froxfield dated 1612, where it separates part of Park Copse from an unwooded area called ‘The Lawn’ (Figure 44) (WSA 9-8-153H) which had been let as pasture in 1547 (WSA 9-1-5). This open area is also shown on an estate map of Savernake of 1820 (WSA1300-375L)(Figure 64), and appears as ‘Lawn Wood’ on a map of 1846 (WSA 3345MS). It serves no function today, occurring in the middle of a mixed beech and pine plantation. Observations Points J to K appear to be the former edge of Park Copse as shown on Atkinson’s map of 1719 (Figure 43). With at least two dog-leg turns, its small-scale ‘ Woden's Ditch Fig. 43 Extract from a Map of the Manor of Chisbury surveyed by Thomas Atkinson, 1719 (relabelled) showing Chisbury Wood Park Copse (Site 23), The Lawn (Site 24) and part of the Bedwyn System (Site 35g). (WSA 1300-372). ‘The Park Coppices — Land 60.41.10 | Wood 50.1.25 Fig. 44 Extract from map, A New Plotte of Lands belonging to North Standen, The Demesne of Chesburie and the Farme of Froxfylde, 1612, by George Wyther, showing Chisbury Wood Park Copse (Site 23), The Lawn (Site 24), Chisbury Wood North (Site 26) and part of the Bedwyn System south of the hillfort (Site 35g). Re-orientated north, and relabelled. (WSA 9-8-153H). construction 1s consistent with a barrier to prevent animals wandering from The Lawn into Park Coppice. The feature might also have bounded part of Chisbury Park (see Site 23). The purpose of the earthwork between the wood and hillfort, as shown in 1719 and since ploughed-out, remains speculative, but it is probably a late- or post medieval construction. Conclusion This feature is of interest because it is named Woden’s Ditch on a map of 1719, the earliest cartographic representation for the Wansdyke continuing to the east of Savernake. The source for this claim remains unknown. The earthwork appears to define an area of pasture identifiable from the mid 16th century. 25 Chisbury Wood Scar Site 25, Figure 1. Points B to C on LiDAR image Figure 40 and aerial photograph Figure 41. Location This site is in Chisbury Wood west of the hillfort and parallel with Chisbury Lane. Description This earthwork extends for 350m on level or 228 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE gently south-facing ground on an almost east-west alignment over soils derived from Tertiary clays and gravels within mature spruce plantation. It peters out within woodland to the east, and within a small pasture to the west. Undetectable on the ground or from aerial imagery beyond these points, it shows well on a 1946 aerial photograph following wartime clear-felling (Figure 41). With up to four ditches running parallel, a variable sequence of merging and diverging banks and ditches is inconsistent in all dimensions. Use of the area by modern forestry vehicles may had a significant effect on the morphology of the field remains. At widest it is 30m, although the height from ditch to adjoining ground level does not exceed 80cm (Figure 65). Research, Documentation and Correspondence with Boundaries This feature is mapped as a perfectly straight double bank bearing the label ‘Wansdyke’ on the 1924 Ordnance Survey 25”/mile sheet XX X13. Crawford (1953) considered it as a fragment of Wansdyke but, finding it densely planted with young conifers noted that it was very difficult ‘to be sure of anything’ and he ultimately regarded it as doubtful. Fox and Fox (1960) concluded that ‘as far as it can be studied in a dense plantation, it is negligible in scale, character and extent’. Hostetter and Howe (1997) keep open the prospect of this feature as a dyke, having reported a double bank and double ditch, which they considered was in the most logical position for an earthwork running west from the hillfort. They also suggest that, because it is ignored by the division between Little and Great Bedwyn parishes, it might post-date that division (mid 16th century). Observations The feature does not appear deliberately constructed, and is too inconsistent to have formed an effective barrier. Apparent banks are probably original land levels standing proud after scouring along ditches either side. The feature resembles a heavily braided track, made by animals or wagons negotiating clayey soils and aligned with a major gateway into Chisbury hillfort (point A on Figures 40 and 41). No parish, manorial or other boundaries accord with it. Considerable traffic must have contributed to the scale of this feature, although it does not obviously link with any drove road. The fact that it is not shown on either Wyther’s map of 1612 (Figure 44) or Atkinson’s map of 1719 (Figure 43) suggests that its use had earlier ceased. At its western end, the bank of Chisbury Wood appears to overlie it, although a modern vehicle track appears to have breached the bank in one place. Conclusion Field evidence indicates an ancient, track, potentially related to Chisbury hillfort, but unrelated to any territorial or defensive dyke. 26 Chisbury Wood North Site 26, Figure 1. Points F to N on LiDAR image, Figure 40, and 1946 aerial photograph, Figure 41. Location This earthwork marks the present boundary of Chisbury Wood, west of the hillfort. Description This feature crosses Tertiary sands and gravels, and consists of a wooded bank dividing a small pasture adjoining Chisbury Lane and a modern conifer plantation. It extends for 370m from the main bank through Chisbury Wood (point F) to a sharp corner (N), curving increasingly eastwards towards Chisbury hillfort. Its small bank and ditch is variable in height with a maximum of 98cm from bank top to ditch floor. There is a substantial change in ground levels facing north (Figure 65). The feature post- dates Chisbury Wood Scar (Site 25). Research, Documentation and Correspondence with Boundaries Major and Burrow (1926) interpreted this feature as the Wansdyke, positing that an ongoing change of ground level occurring to the east marked its former course. On Atkinson’s map of 1719 (Figure 43), this line is shown as the division between the southern field Broad How and a small field beside Church Mead. Wyther’s map of 1612 also shows this line ~ continuing as far as the hillfort (Figure 44). Observations This bank forms the northern boundary of Park Copse and appears later than the main ditch extending east-west through Chisbury Wood (Site 23) and Chisbury Wood Scar (Site 25). Conclusion There is no evidence that the earthwork formed any long distance territorial or defensive function. Its association with Park Copse suggests it was also THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 229 the boundary of the 13th century Chisbury Park, which may have centred on pre-existing woodland mentioned in the Domesday Survey (see Site 23). 27 Mud Lane Site 27, Figure 1. Location Mud Lane runs from Hat Gate at SU211643 to Martinsell Hill, at SU184646. Description Mud Lane is a wide track 1.84km long with variable banks and ditches. At its eastern end its course is interrupted by the Midland and South Western Junction Railway. Wooded along most of its length, with mature oak including occasional pollards, the feature meanders only slightly in its course and runs consistently along the southern edge of the Savernake plateau, broadly east-west, with the Pewsey Vale sloping away to the south. Apart from a small section at its eastern end, the appropriately named Mud Lane is on Clay-with-flints, close to its southern junction with the Upper Chalk. There are banks and ditches to the north and south and the track itself is also frequently ditched. To the north there is a bank up to 70cm high (Figure 65), # g cy coh 4 fore burbig? onde ,, aoe ia. < \ : oe ; Gi arnt e se coo OORT peo z 4, D sideman @ fos ww ay 3 dwhilnrd t horbdg occasional deeper ground may represent a former holloway or mineral winning associated with the kilns at Brick Kiln Copse. Research, documentation and correspondence with boundaries This feature is included because of its known historic associations. Mud Lane connects with Martinsell Hillfort to the west, and probably relates to a feature known as the Wodeditch mentioned in Forest records. A perambulation undertaken following a Royal decree of 1299, goes from Apsehullmede (the name survives at Apshill Copse), by the covert of Fwode, and so always by La Wodediche, as far as the eastern corner of the croft which is called Boneclyne (Martinsell Hill) (WSA 212B/7189). A coarsely drawn attempt to reconstruct a 14th century perambulation in association with a tithe dispute of 1668, shows the Wodeditch (arrowed in Figure 45). A Court Roll of 1334 reported Robert de Bilkemore for felling one oak worth 18d on the Wodedych between Wydenrych (Witheridge, towards Hat Gate, a former common of Wootton Rivers (WSA 9-21-44)) and la Bers (towards Clench Common) (WSA 1300-59). He was also reported for taking another worth two shillings, also on the Wodedych at a place called Goddermanscroft, and taking it to his house at Wyke (WSA 1300-59C). These incidents, coupled with the perambulation section always by the Wodeditch, suggests an extensive feature, possibly , wen Guy ie i had Qucts beer ae | A ns Poorer cath Sere them ark Fig. 45 Coarse map reconstructing early 14 century perambulation of Savernake, probably prepared in association with a tithe dispute of 1668, re-orientated to north (WSA1300-47). The map shows three apparent dykes, marked by arrows. The Magnum Fossatum (Site 28) is shown to the east, an area near Ram Alley (Site 29) is also marked fossat, and the Wodeditch, described under Mud Lane (Site 27) 1s to the west. 2 230 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE equating with much of Mud Lane itself. The route is now a bridleway which links with the drovers’ road known as Tan Hill Way, near Martinsell. Mud Lane was also the boundary of Savernake Great Park, an area extending to cl0km2 with a perimeter fence of cl4km, constructed for the Earl of Hertford in the 1570s (WSA 9-22-204). The record states that 104 lugs (approx 523m) of park pales had accidentally been erected on the neighbouring land of Sir Edward Baynton of Bromham. Mud Lane is the boundary between the parishes of Savernake and Wootton Rivers. Observations A map of Savernake Great Park by Charles Price in 1718, shows Mud Lane (un-named) going only from The Hat to Brick Hill Copse at SU 194646 (WSA 1300-372). This might be a depiction of ownership, rather than an indication that the lane ceased at this point. Conclusion Documentary evidence indicates a long distance feature named Wodeditch connected with Martinsell hillfort and equates with at least part of Mud Lane. It carried a park pale in Tudor times. The link with Tan Hill Way suggests it could have formed a long distance trackway. Despite good archival sources, the modern feature is variable in morphology and insignificant in scale and it remains doubtful if it is a single coherent structure. It is possible that the Wodeditch was never a dyke (c.f. Wodensditch), but a lesser ditch bounding the wood or Forest of Savernake. Further work might unravel the relationships between surviving structures. 28 Magnum Fossatum Site 28, Figure 1. Points B to C on combined LiDAR image, and aerial photograph, Figure 46. Location This earthwork is in the northern part of Tottenham Park, running east-west approximately midway between Tottenham House and St Katharine’s Church. Description Invisible on the ground, it is not possible to identify the scale or structure of this feature, although faint traces can be discerned by remote sensing. It evidently extended from Straight Corner (SU 257646, marked B on Figure 46) to the former location of Bentley Gate (SU 245643, point C). Knowledge of its existence derives largely from archival records. Research, documentation and correspondence with boundaries The Magnum Fossatum or great dyke was a regular feature in medieval perambulations of Savernake. In 1300 the bounds of the Forest extended from the Pasture of Stock (the eastern edge of Bedwyn Common, point A), ‘and so always by the great dyke as far as Bentlewell’ (sometimes recorded as Bentley Gate) (WSA 212B-7189). The gate features on a map of about 1700 (WSA 9-6-769) and its location can be confirmed by reference to a legal case of 1708, when Burbage parishioners were reported for illegally beating the bounds and going from ‘Stepstile on the east part, along the pale of Tottenham Park, to Bentley Well or Bentley Gate’ (WSA 9-7-42). The feature is close to a massive chalk pit (between C and D, Figure 46), while the well itself was probably associated with Bentley Close, a half-acre plot let to the warrener in 1623 (WSA 9-22-19), which appears to feature on LiDAR images (point D). The Magnum Fossatum is one of three dykes shown on a coarsely drawn early map of Savernake (WSA1300-47) (Figure 45). This map shows perambulation stages ofc. 1300 alongside post-medieval features. H. C. Brentnall, who worked extensively on Forest perambulations, thought that the map was an attempt of about 1550 to reconstruct the Forest bounds. He considered that certain mistakes of the cartographer were understandable, but proclaimed that the ‘course drawn for the Magnum Fossatum between Durley Heath and Tottenham Park 1s less excusable. There 1s no trace whatever of a great ditch on that line ... unless you suppose that the Ha-Ha was constructed on part of 1t. You cannot obliterate a great ditch’ (WSA 1300-47A). From Bentley Gate the Forest bounds at this time continued along a ‘vallum’ — possibly a reference to the adjacent Roman Road. Observations Palaeographical and contextual evidence, coupled with idiosyncratic spelling (based on misreading of medieval texts), enable the map in Figure 45 to be linked to a dispute of 1668, when the Forest bounds of Edward I were cited in an attempt to evade tithes (WSA 1300-46A; 1300-100). Despite the assertions of Brentnall, who had been unable to trace the identity of Bentley Gate, archival evidence and remote sensing are consistent with the THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 231 Fig. 46 Upper LiDAR aerial image of part of Tottenham Park, 2006. © Forestry Commission. Lower Aerial photograph of part of Tottenham Park, 1946. (RAF 4112, CPE/UK, 1821). Images show supposed course of the Magnum Fossatum (Site 28), from Straight Corner (B), to Bentley Well and Bentley Gate (C/D map of 1668, confirming that the dyke formed the northern boundary of Tottenham Park at that time. The grounds were extended between 1718 (WSA 1300-372) and c. 1730 (WSA 1300-358) and the dyke would then have become a redundant feature across open parkland. It is not clear when it was levelled. However, Lancelot Brown, who was commissioned to advise Lord Bruce later in the century, spoke of the need to ‘fill up holes, and take off the abrupt hills and hiccups’ to produce a smoothly graded scene in 1770 (WSA 1300-1926). Landscaping and ploughing have therefore been able to destroy virtually all traces. Conclusion This feature appears to have been an early historic or prehistoric earthwork and the term ‘magnum’, in contradistinction to other ‘fossae’ nearby, suggests a substantial structure. Given its destruction above 232 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE ground, it is not possible to say whether it could have been part of a defensive system. Its use as a Forest perambulation feature during the reign of Edward I suggests that it also defined the King’s land and hence a former extensive landholding or transfer. 29 Ram Alley Complex Site 29, Figure 1. Points A to B, C to Fand G to H on 1930 aerial photograph, Figure 47. Location These sites occur immediately north, north-west and south-west of Ram Alley, near Burbage. Description This complex of earthworks lies on Upper Greensand, with alluvial deposits along the valley floor. Site A to B isc. 300m long and runs within a woodland strip of mature oak, ash, and hazel, with elm, elder, hawthorn and sallow south-west of Ram Alley in a gentle S- shaped curve alongside the road which slopes from Easton to Burbage Common. In section there is a (Qe Fig. 47 Aerial photograph of Ram Alley, 1930, showing alignment of extant or apparent former features (Site 29). (NMR 9967, SU2863 1-2). © English Heritage, NMR, Crawford Collection. THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 233 2.85m high bank with a steep drop of 1.66m to road level (Figure 65). There is no ditch, and south east of the road the ground returns to the same height of the bank opposite, maintaining this level across arable fields towards Bowden. Points C to D are ploughed-out and lost, although Ordnance Survey maps suggest it extended for 200m towards the valley floor immediately north of Ram Alley hamlet. It is marked as an earthwork on OS Explorer Maps (2004) and the approximate course is a hedge-line (since grubbed up) on a 1930 aerial photograph (Figure 47). Points E to F are on the same alignment as C to D and consist of an insignificant ride between Leigh Hill Copse and Square Copse. There is no obvious earthwork other than vehicular tracks and coppice banks. Site G to H is in the north-east quarter of Ram Alley Copse and consists of a single massive hollow 320m long, aligned almost due north, through park-like sycamore woodland, with dense bracken in open areas. It begins as a barely perceptible depression at point G and gets progressively deeper towards the north, up to 4m deep, where its course meanders briefly before reaching the valley floor. There is no bank on either side (Figure 65). Research, documentation and correspondence with boundaries It is clear that parts of this complex are associated with the former Brimslade (Deer) Park and also with features recorded in Anglo-Saxon Charters and 14th century Savernake perambulations. Points A to B form the modern boundary between Easton and Wootton Rivers civil parishes which continues north along C to E Both A to B and C to F form part of the bounds of Brimslade Farm in 1842 (Figure 48), where labelling of neighbouring parishes suggest this land was still extra-parochial, a relict of its former Forest status. This map also suggests that there may have been deliberate attempts to make boundaries symmetrical, a situation later emphasised by triangular, circular and square coppice designs. Crawford (1921) described the section C to D as a pack-trail, with a curved course, its concave side facing eastwards, cut through by the Brimslade Park boundary and therefore pre-dating it. A right-of- way persists along C to D through the middle of a ploughed field, and along E to K although these two sections are severed by the Kennet and Avon Canal and main line railway. Anglo-Saxon Charters of 968, equating with land in Bedwyn (discussed above), and of 961 equating with land in Burbage (cited by Grundy, 1919), go PLAN OF BRIMSLADE FARM WILTS 1S42 froxx Wootton =< res Sey Tr WHO EDF SP ESS Bel eR ee OS ie a 4 NY f oS ge Apsehill North ie Bea » ¥ E +0 Coppice Round Gees Quarters 42 5°00 4 . pea Sh Coppice ege 08) He ea le hayes ud ee Ps 4 Oe a os, Ga KG = e or O 256 6. Oo “eet: South ve Round we, : AS PARISH Fig. 48 Plan of Brimslade Farm, 1842, showing geometrical coppices within contrived, symmetrical, shield-shaped outline, derived from the Tudor Brimslade Park. Features described under Site 29, Ram Alley Complex, relate to some of the eastern and south-eastern edges. Redrawn from WSA MS3354 Brimslade. from Byrbaec towards Ig-feld and then to Bydan Hammes. Byrbaec equates with Burbage (although not necessarily the heart of the current village), and the name of Bydan Hammes survives at Bitham Pond in Savernake at SU 221652. The area of Savernake converted to Brimslade Park was known as Iwode or Ywood, and appears to relate to [g-feld. Crawford (1921), endorsed by Gover, Mawer and Stenton (1992), considered that the component ‘Alley’ in the name Ram Alley, was likely to be derived from Ig-lea (or Y-lea, like Ywood). Crawford considered that the charter bounds extended along the line C to F as shown on Figure 47. A Forest perambulation of 1300 follows a similar route (WSA 212B-7189) coming south ‘by the Swynepath, as far as the wood of the Prior of Motesfonte, and thence descending by the dyke as far as the head of Iwodesmede’ . A parish boundary runs along his dyke between Leigh Hill Copse and Square Copse. It is shown with the label ‘fossat’ on a 17th century reconstructed perambulation (Figure 45)(WSA 1300- 47). The mead and Prior of Motesfonte’s wood are also both shown on this map and there are many other documents referring to these together as if they were adjacent features, including the sale of the woodland to the Earl of Hertford in 1599 (WSA 9-6-6). The perambulation continued ‘always by the extremity of the wood as far as Erthbyr’. The location of this last earthwork is unclear but appears to be 234 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE near Ram Alley. The 17th century reconstructed perambulation (WSA 1300-47) places it south west of the hamlet (Figure 45). Points A-B of Figure 47 equate with the label Park Pale on modern OS maps and the extent of the park probably equated with the area of Brimslade Farm in 1842 (Figure 48). Brimslade itself was apparently imparked after 1585 and the park was conveyed by the Earl of Hertford to his wife Lady Frances, as part of her Jointure, in 1602 (WSA 1300- 90), and was probably forged from an area already managed for deer. In 1536, accounts from Wolfhall refer to work at Ywode, including setting ‘quyke frythe on the gret dyche in Padsdeyn ... mending the red deer park hedge... [and] hedging on the gret hedge and in Ywodes Meade’ (Longleat House, Seymour Papers Vol IX, ff 55-56). The reference to a ‘gret dyche’ suggests that this feature already existed, rather than something newly created for the park. Crawford (1921) interpreted points G to H of Figure 47 as a series of deep pack-trails (or holloways). Observations The absence of a bank for sections A to Band Gto H does not suggests a deliberately constructed feature and the marked changes in ground level probably result from heavy traffic. Conclusion It appears that sections A to B, C to D and G to H represent holloways or scars from early roads across the valley of the Pewsey Avon. These were effectively closed by the construction of Brimslade Park in the 16th century, if not earlier, which pushed the crossing eastwards to nearer its current alignment (point J on Figure 47). Both 14th and 16th century references to a dyke or great ditch appear to be represented by the line C to K which has been used successively as an Anglo-Saxon, Forest, park and parish boundary. Hence there is potential that it is part of a long distance territorial feature. 30 Durley — Tottenham Linear Site 30, Figure 1. Points B to C on LiDAR image, Figure 49, and aerial photograph Figure 50. Location This feature is situated 1km south west of Tottenham House. Description The surviving earthwork extends for 270m across the Upper Chalk, in a gentle west-east curve between Durley and the bounds of Tottenham Park, on the Fig. 49 LiDAR aerial image of area between Durley and Tottenham Park (Site 30), 2006. © Forestry Commission. THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 235 Fig. 50 Aerial photograph of the area between Durley and Tottenham Park (Site 30), 2000. (NMR 18717/18 SU 2363/4). © English Heritage, NMR. south facing slopes above the valley of the Kennet and Avon Canal. It commences at point B on Figures 49 and 50 at a corner projecting into a large field to point C where it meets a ha-ha (C to D) near the Octagon summerhouse of Tottenham Park (E). The feature marks a dramatic south sloping change in ground level of over 3.5m. There is no bank or ditch, but there is an inconsistently defined step part way down the slope, and possibly a second (Figure 65). It supports semi-natural broadleaved woodland of ash, hazel and hawthorn, and marks the division between arable fields at different levels. Research, documentation and correspondence with boundaries An aerial photograph (Figure 50) appears to show the same feature continuing west to Savernake Road. The surviving part forms the modern boundary between Great Bedwyn and Burbage parishes, although a 6”/mile Ordnance Survey map of 1900 shows the parish boundary running east of this. The same map shows trees marking the line of section A to B. At its eastern end the feature appears obscured by landscaping associated with Tottenham Park. The Octagon summer house was constructed in 1743 (WSA 1300-362) and in 1769 Lancelot Brown advised that the land around should ‘be made level and decent’ (WSA 1300-1920). Observations The absence of a bank or ditch suggests that this is not a deliberately constructed feature. Although just a simple steep slope, soil marks where it has been ploughed (section A to B) show a double line (Figure 50). Similar soil marks suggest that cognate features extended elsewhere across this land. Conclusion Field evidence suggests a lynchet. 31 Bedwyn Brail - Chalkpit Bank Site 31, Figure 1. Site A to B, Figure 51. Location This earthwork is situated at the easternmost projection of Bedwyn Brail. Description A substantial ditch and bank 85m long and aligned at 105° is situated in semi-natural woodland of old ash coppice, with hazel, maple, birch and some oak on Tertiary sands and gravels, close to the edge of the Upper Chalk. The feature is c. 15m wide, with a bank 1m high to the north, which defines the edge of the woodland (Figure 65). The ditch is broad and wide and has been used as a modern vehicular track. The feature curves slightly south at each end, while the eastern end appears to have been ploughed in. An open area of level grass and bracken to the south has exposures of sand. There is a substantial chalkpit at the western end of the feature at SU 286632. Research, documentation and correspondence with boundaries The feature does not appear on early maps of the area, such as an Inclosure Award map of 1792 (Figure 52)(WSA EA68), Benjamin Haynes map of Wilton of the same date (WSA 9-27-59), nor an 1820 estate map (Figure 61)(WSA1300-375L). The north eastern edge of Bedwyn Brail has been subject to major change since the 18th century. The Brails were planted in 1791-2 upon an earlier sheep and cow common, following Parliamentary inclosure (Sandell, 1971; WSA 1300-2539 and 2540). Prior to this, two small parcels together called ‘Brail Woods’, were situated near the north-eastern edge of the area. Engravings by J. Cary, reproduced by Colt Hoare (1819) and Smith (1884), and a map of Wiltshire by Andrews and Dury (1773), all show the former 236 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE ff Chisbury Little Bedwyn —~ | { sani V1 WN J \. Zb Chisbury Wood _} Na Pumping £ f Za\'< y = Sin hod Misti Foxbury edwyn U y Merrell ‘ \ & Down Wilton Brail Tidcombe © ‘y Gate Kv si Hillbarn Fig. 51 Map of Bedwyn System and Brails complex, showing Sites 31-35. Brail Woods. Cary’s engravings appear to show small sections of dyke within the wood (Figure 53). It is not possible to pinpoint the location of the depicted dykes on modern maps, nor the location of Brail Woods within which they were shown. Observations Whilst the earthwork could be a holloway associated with chalk workings, the estate map (Figure 61) suggests a northern extraction route, and the bank Broile Gates N Great Bedwin Lands / Broie. Bann Spe < ¥ See NTO Broile ¥ ry Private Real < = linen RS a | Poors S Horn Stile” [ Allotnjent ¢ ow 9 S as é \ < \ ay § Ire \ Hil & The Cow, Common ye" & 8 a*, Closes ; & : Layings Stile me ; : \ att, woot Ms f I ip { , } cod Ne < s r x The Sheep Common i ¥ S HY ba ; a 4 a § Tideombe Gate S--~ da Buegeot i f 3 z if Sesame HEATHELL CLOSES } } OS oy sa | i Pn | f Fig. 52 Map of the Wilton Inclosure Award 1792, showing area now occupied by Bedwyn and Wilton Brauls, including the division between the Cow Down and Sheep Down (Site 33). Redrawn from WSA EA68B. itself is proud of ground level both sides suggesting deliberate mounding. Conclusion This short section of constructed earthwork may be either one of the dykes depicted in Colt Hoare’s work or located nearby and akin to them. No purpose or age can be deduced, nor whether this earthwork formed part of a larger feature. : % py > : -* = ws REISS Gea S Seo PRESS : RSS Great Boetw tr 3 é Seay ; AN DES A es > . VERS ve x: Broil & \ floc rete $ a Fig. 53 Extract of ¥ Cary’s engraving from Colt Hoare’s Ancient Wiltshire, 1819, showing purported dykes, including short sections in two small parcels labelled Bedwyn Woods near the north eastern corner of what is now Bedwyn Brail (see Site 31). THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 237), 32 Bedwyn Brail Conduit Ditch Site 32, Figure 1. Points C to K Figure 51. Location This feature is in the southern part of Bedwyn Brail woodland, north of Hillbarn Farm. Description This substantial feature can be traced for 520m as a continuous single trench, banked on both sides. It extends from SU 284625 just south of the Broadwalk (a landscaped avenue extending across the Brail, aligned with Tottenham House and the Ailesbury Column), to the edge of the wood at point E, SU 281621. The ditch is positioned along the western edge of the plateau of Bedwyn Brail, with a significant slope to the west. Virtually the entire length is situated on a thin deposit of London Clay overlying Reading Beds, with complex sandy, gravely and clayey soils causing springs to erupt, even near the tops of hills. The feature comprises three main parts. Section C to D is aligned at 200°. At its northern end it emerges indistinctly from surrounding land, gradually increasing in dimension further southwards. At its greatest, this section is 15m wide (8m between bank-tops) and 1.4m deep. At point D a small valley approaches the feature from an angle of 125°. A pond iS Opposite on the downhill (western) side. Section D to E is aligned at 209° and comprises a wider and deeper structure with steeper sides (Figure 54). At its greatest, this section is 20m wide (12m between bank-tops) and 2.7m deep (Figure 65). From point Fig. 54 The Conduit in Bedwyn Brail (Site 32) southern section, looking north. E, at the edge of the wood, across rough pasture to point F (a former entrance into the common known as Tidcombe Gate), most of the original ditch has been filled in, partly with 20th century material. Its course as a curving feature can be traced clearly on mid-century aerial photographs (e.g. NMR, sortie RAF797, 1942) and less distinctly from residual ground features. Research, documentation and correspondence with boundaries The feature is not shown on the Inclosure Map for Grafton and Wilton Tythings (WSA Great Bedwyn EA68 Plan B) following the Award of 1792, when it would have extended across the part of the common pasture known as the Cow Common (Figure 52). Neither is it shown on Benjamin Haynes’ map of Wilton of 1792 (WSA 9-27-59) which shows the new plantations of Wilton and Bedwyn Brail, nor on the 1820 Savernake Estate Map (WSA1300-375L). The feature does not accord with any parish, township, or land management boundaries. It has been labelled ‘conduit’ on OS maps since the late 19th century. This derives from its deduced purpose as an open ditch supplying water to the Duke of Somerset’s planned mansion at The Brail, which commenced in 1548 and had been abandoned by 1551 (Jackson 1875; Bathe 2006). The trench was located amongst dense scrub by the Rev George Stallard and Maurice Adams in the spring of 1874 (Wiltshire Archaeological Notes 1932; Adams 1905). Their search had been stimulated by the discovery amongst the Longleat Archives of evidence that a conduit had been built at The Brails (Longleat House: Seymour Papers vol IX, Thynne Papers vol II; Jackson, 1875). Stallard wrote to Jackson in May 1874 stating ‘I find that the old people do not call this Conduit the ‘Wansdyke’ but ‘tthe moat’ or ‘the ditch” (Wiltshire Archaeological Notes 1932). Fieldwork from 2007-2009 by the WANHS Archaeological Field Group under the direction of Jim Gunter revealed the presence of a brick conduit (water flowing south) up to 2m deep below the ditch floor along its length from point C to near to point E, with a side channel entering from the south-east at point D. Excavation also revealed pre-historic pottery, including a Deverel-Rimbury cremation urn apparently in situ in one of the banks. Observations Preliminary examination of excavated trenches and pre-Iudor ceramic material from section D to E suggests that the conduit has encroached on an 238 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE existing pre-historic structure. References in the Wolfhall bailiff’s accounts of payments to workmen for ‘dikinge the gret diche’ at The Brails in 1548 would appear to make more sense if the great ditch was a pre-existing feature (Longleat House, Seymour Papers Vol 9, ff 92-145). Conclusion Part of these earthworks derive from mid 16th century construction of a brick-lined conduit, although the southern part of its course, on a different alignment, was perhaps constructed upon, or at least intruded upon, a pre-historic feature. There is no evidence for a long-distance territorial or defensive structure, although its possible association with major dykes forming part of the Bedwyn System (Site 35) deserves investigation. Ongoing research by the Archaeological Field Group should help elucidate the sequence of events associated with this feature. 33 Bedwyn Brail - Common Bank Site 33, Figure 1. Points H to J, Figure 51, with adjoining features to north (G to H) and south (J to K). Location This feature is within the mature broadleaved woodland of Bedwyn Brail. Description Poorly fertile Eocene deposits of the Reading Beds and London Clay overlie the Upper Chalk and give variable sands, clays and gravels, with scattered springs and swallow holes. The overlying earthwork comprises two sections. From point H toI this 380m long feature is regular and gently meanders aligned at 138°. It comprises a bank and ditch up to 1.43m high (Figure 65). The ditch is to the north, even when, in the western section, on the uphill side, causing water to pond against it (Figure 55). At the western end, point H, at SU 282629, the dyke tapers to an abrupt end. There are two main cuttings through the feature, for the main track along the ridge of the Brail, and another smaller route. At the eastern end of this section, point I, there is a sharp corner as the feature sweeps south. This second section, I-J, has a different character and maintains a virtually perfect straight course of 203° for 680m to Hillbarn Farm where it marks the eastern edge of the Brails, near BEY ARE Fig. 55 The Common Bank in Bedwyn Brail (Site 33). Much of this feature was the boundary between the Cow and Sheep Downs (see Fig. 52). The ditch 1s invariably to the north and east, even when changes in topography causes this to be on the uphill side, resulting significant ponding. the junction between the London Clay and Reading Beds. Here, the plateau has an open, grassy, park-like habitat and appears to have been levelled. Outside the area is denser woodland with hazel coppice. The ditch of this section is east and downhill of a more modest bank, up to 87cm high, whose height appears exaggerated by the local topography. The feature passes through progressively wetter soils to the south, where it is frequently eroded and barely discernible. Near the northern end of this section is a gap of 15m, apparently caused by the passage of animals. Research, documentation and correspondence with boundaries Castle Copse Roman Villa is within 100m of the northern section of the bank. The dyke was not mentioned by Colt Hoare (1819), despite his searches in the Brails. The straight section of this earthwork, I to J, is labelled ‘Wansdyke’ on an 1820 Savernake Estate map (WSA1300-375L). This length, and at least part of the northern section H to I, formed the division between the Cow Common and Sheep Common, as shown on the Wilton Inclosure Award of 1792 (WSA EA68B), Figure 52. The earthwork does not correspond with parish or township boundaries. Nineteenth century OS maps show a feature connected with the western edge of the northern section, between point Gand H. There is no feature there today, nor any indication that the existing bank has been cut or curtailed. A shallow bank nearby, on an alignment of 018°, might have linked with the THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 239 earthwork, but is not obviously related and may be a lynchet. The southernmost point of the straight section, at point J, appears to link with a curving line across Hillbarn Farm to point K, apparent on 19th century OS maps, passing through Heathhill Closes as shown on both the Inclosure Award map, Figure 52, and Benjamin Haynes’ map of Wilton of 1792 (WSA 9-27-59). This alignment extends west along the road passing Wilton Windmill to point L. It is possible that the surviving section forms part of a longer conceived course. Observations Either section of this feature could be associated with landscaping of the Duke of Somerset’s unfinished mansion and park from 1548 (Jackson, 1875; Bathe, 2006). During preliminary ground work for this over 6000 man days were ‘spent in making and planting a new dych ... at doddsdowne bussh’ (Longleat House/Seymour Papers/Volume IX, ff. 92-145). The alignment could also equate with a perambulation of the park recorded in a letter by John Berwick in November 1548 (Longleat House/Thynne Papers/ Volume II, f.30). These bounds extended ‘over the felde to the close corner at Wiltons towns end viz. to the nether ende of the lane/ whiche desendith from the broile’ suggesting that it adjoined the road running past the modern windmill to point L (Figure 51). Conclusion Despite few certainties, this feature is of significant potential interest; the two sections may not be contemporary. Presently undated, the northern section H to I could relate to other pre-historic or Roman features in Bedwyn Brail. As a pre-existing structure, it might have been adapted to form part of the Duke of Somerset’s park bounds in the mid 16th century. It appears that the eastern straight section, I to J, which must have been planned and surveyed, was added later and might also represent the bounds of Somerset’s mansion. This line was the division between the Sheep and Cow Commons in the 18th century and might have been constructed for that purpose. Archaeological investigation might resolve the sequence of these features. There is no evidence for a territorial or defensive structure. 34 Bedwyn Brail — Hillbarn to Harding Field Mark Site 34, Figure 1. Points M to N, Figure 51. Location This feature extended across arable fields east of Bedwyn Brail. Description Aerial photographs suggest a feature c. 900m long and c. 33m wide between the minor road east of Hillbarn Farm, at SU 284621 and a road west of Harding Farm at SU 291628. The arable fields across which it can be traced are level or sloping gently to the north east and overlie Upper Chalk. There are no surviving features and the aerial images are indistinct, but show a pair of parallel linear marks, curving slightly along a general course of 040°. Research, documentation and correspondence with boundaries The feature is visible on at least four low level aerial photographs taken in 1975 (NMR 2862/2/217-220). It does not accord with any boundary, although it links with the Grafton-Great Bedwyn parish boundary along the road from Wilton to the cross roads south-east of Harding. Its course cuts diagonally across current and former field boundaries, which do not correspond with its course. It would have run across the former Sheep Common, as shown on the 1792 Inclosure Award map (Figure 52), but is not indicated on that map, and cannot be identified on Andrews’ and Dury’s Map of Wiltshire of 1773. The feature is north of the extensive ‘Celtic’ field system on Wilton Down. The road south of Hillbarn was established through the Inclosure Award, when it was labelled The Broadway and described as 40 feet wide. The road west of Harding Farm appears older and is presumably the Bedwyn to Andover Road which features in a 14th century Forest perambulation (WSA 1300-65). Observations A shortcut between the two roads, thus avoiding a detour to the east, seems a plausible explanation for this feature as it appears to be too broad and regular for a road. The feature describes a course that broadly accords with records of the ‘Ditch of Lynley’. This ditch or dyke is mentioned in at least four surviving perambulations of The Brails, which were outliers of the Royal Forest of Savernake, during the reigns of Edward I and Edward III (c1300-1330) (WSA 2667/24/5/1; 212B/7189; 1300-65; PRO E 22-220). The perambulations extend from the Bedwyn- Andover Road, via various stages to ‘the eastern corner of La Lynelyghe (otherwise spelled Lynley, Lyllegh or Lynlegh)’. All four perambulations then continue 40 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE = al =a ee metres (approx) Fig. 56 Aerial photograph of the Bedwyn Complex and Brails, 1946, showing sections of the Bedwyn System (Site 35). (NMR 501, RAF/CPE/UK/1821, 4118 — 4120). © English Heritage, NMR, RAF Photography. THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 241 ‘by the ditch/dyke’ (in two cases: ‘always’ by this) ... ‘between the pasture of Wilton and the covert of the Lord King? ... ‘to the Muleway’. The Muleway equates with the bridleway skirting Bedwyn Brail and leading towards Great Bedwyn - possibly a corruption of Millway. The name Lynley has not survived, although a wood with this name was removed from Forest Law by a 1330 perambulation and presumably relates to the same area. Conclusion The relevance and identification of this superficial field mark is far from certain and a wholly prosaic explanation is possible. It is similar to a section attributed to the Bedwyn System (Site 35). Failure of the earthwork to accord with modern field boundaries and 14th century references to the Ditch of Lynley broadly at this location make this feature of potential interest. (approx) Fig. 57 Aerial photograph of the northern part of the Bedwyn System (Site 35) near Chisbury, 1946. (RAF Flight, CPE/UK, TS21): 242 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 35 The Bedwyn System Introduction Seven geographically connected sections are grouped under this heading because of their putative (but not unequivocally proven) relationship. They are collectively mapped as Site 35 on Figure 1, shown as sections B-P and Q to Zb on map, Figure 51, and aerial photographs Figures 56 and 57. 35a Bedwyn System: Harding Valley Site 35, Figure 1. Points B-PR Figure 51 and aerial photograph Figure 56. Location This ploughed-out feature is near Great Bedwyn, 700m south-east of Jockey Green. Description This feature is known from written accounts, maps and aerial photographs up to c. 1964. It extends for 450m from point B on the edge of Bedwyn Brail, at SU 288631, to a road fork at point B SU 290635 over almost level arable draining gently to the north-east. Underlying soils are derived from Upper Chalk. There are no visible traces. Research, Documentation and Correspondence with Boundaries Aerial photographs taken between at least 1930 and 1964 (e.g. NMR 9967; 501; 11135) show a slightly meandering white chalk mark (Figure 56). Near point P a branch extends westwards and merges with the right-of-way leading to Castle Copse (point Pa), whilst the main route continues to where the Bedwyn-Shalbourne road forks. A feature broadly along this line is shown on 19th century OS maps, sometimes with the word ‘bank’ or showing a straight track. Colt Hoare (1819) considered it a branch of the Wansdyke, which crossed the Bedwyn-Shalbourne Road, looped south-east over what is now the A338, went across Shalborne Downs and then swept south-west around Botley Hill. He claimed he had traced its whole distance himself, although his map of the earthwork, engraved by Cary, shows only an intermittent feature. It depicts a dyke between Round Copse and the two small woods then located Cc Ohissuay Castft Fig. 58 Undated sketch map, thought to be preparatory to Colt Hoare’s Ancient Wiltshire, 1819, showing the purported course of the Wansdyke, and apparent dykes between Round Copse and Castle Copse (now part of Bedwyn Brail). Re- orientated north. (WSA 383-907). near the north-eastern corner of the modern Brail (Figure 53). A coarse sketch map, perhaps prepared whilst researching his Ancient Wiltshire, shows rectangular ‘earthenworks’ between Round Copse and Castle Copse (part of Bedwyn Brail) edged by a bank and ditch, with a double line suggesting a dyke crossing the valley here (Figure 58). A Savernake estate map, prepared about 1820 (Figure 61), concurrent with Colt Hoare’s observations, shows the feature as a field boundary, but with both branches of the fork, near point PB as parts of a staggered track connected to a chalk pit within the Brail (WSA1300-375L). In his analysis of the Anglo-Saxon bounds of Bedwyn and Burbage, Crawford (1921) marked this feature as part of the Wansdyke (which he interpreted along the line of P-Q-R). He considered that the gate or gap described as ‘dices gaete aet harandene’, in the 968 charter (discussed previously)(K1266) was located . at the end of this feature, between points P and Q, the latter at the south-western tip of Round Copse. Major and Burrow (1926) accepted the same route for a branch of the Wansdyke as a bank along the eastern edge of Bedwyn Brail, although they were unable to relocate the route which Hoare claimed to have ‘very accurately laid down’. By 1953, Crawford reflected that he ‘did not feel at all satisfied’ with identification of dykes in the vicinity of Round Copse, noting that ‘there are many old banks and ditches hereabouts’. He dropped any mention of the section between the Copse and the Brail in his article on the East End of THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 243 Wansdyke in 1953 and omitted it from a map of the dyke in a book published the same year (1953b). Fox and Fox (1960), equating the Harding valley with the Harandene of the 968 charter, considered that reference to a gate made it ‘appear’ that the earthwork continued across the valley, although they saw no valid reason for calling it Wansdyke. Burne (1953) re-located the chalk mark, speculating that it defined part of a preliminary, but subsequently abandoned, branch of the Wansdyke that ran as far as Wilton Windmill. Howe and Hostetter (1997) considered the feature approximating to points B-P on Figures 51 and 56, was a ‘fairly well-attested ... section of dyke mentioned 1n the Saxon Charter’, and that it must have dog-legged twice here. Observations Grounds for an ancient linear feature are based on the assumption that the @ices gaete’ in the 968 charter is a gap in the earthwork and that ‘aet harandene’ was in the valley floor. Even if these assumptions are reasonably founded, the gap can be explained only by accepting that there were right angles in the dyke, since ploughed-out. Grundy (1919 and 1920) and Lennon (2010) place the gate north of Round Copse, while other options are plausible. Field evidence is also equivocal. Whilst the former presence of a chalk mark is indisputable, its division near point P is directly related to tracks in this area (Figure 61). There were massive chalk pits at the eastern side of Bedwyn Brail, some of which are ancient - a ‘grete pytte’ was mentioned in a park perambulation of 1548 (Longleat House/Thynne Papers, Vol II, £30). These may have needed a surfaced extraction route away from the wet, clayey soils of the hill. Conclusion Documentary evidence for a dyke in this locality is weak, while cartographic sources identify the feature as a track for most of its course. Archaeological excavation might resolve the issue. 35b Bedwyn System: Round Copse Site 35, Figure 1. Points Q to S, Figure 51, and aerial photograph, Figure 56. Location This feature runs south-west to north-east through Round Copse, c. 1km south-east of Great Bedwyn. Description The south-western part of the wood overlies the Upper Chalk, with the north-eastern part situated on Reading Beds which produce a mosaic of sands, gravels and clays. The feature extends for 520m from the Bedwyn-Shalbourne road at SU 28926370 and snakes north-east through the woodland, rising to almost 30m at SU 29316399 on the edge of Merrelll Down. Highly variable in form, its appearance is often emphasised by changes in topography. At the south-western edge of the wood an overgrown track develops into a sunken way within 30m. At 70m from the road, at SU 29006370, there is a substantial holloway, 1.3m below ground level, with a bank 40cm high and 2.5m wide on the north-western edge. The whole structure is 12m wide. The bank here is too small to account for the material from the sunken area, which has been eroded or removed. At SU 28996376, the lower western coppice bank meets the feature and overlies it. Between this latter point and SU 29056379, where the lower eastern edge of the coppice branches off, there is a considerable bank 2m high and 14m wide with a domed top 2m wide, but lacking ditches (Figure 65). Immediately Fig. 59 Oak Pollard in Round Copse (Site 35b), near the meeting point of the three parishes of Shalbourne, Great Bedwyn and Little Bedwyn. 244 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE north of this point, the bank is emphasised by an apparent cutting on its south eastern side. Between here and point R (SU 29156395) a small bank or ridge is accentuated by a sunken path alongside. At point R, the path joins another one running east- west. Between point R and the north-east corner of the wood, the feature again comprises a small ridge, with sunken path alongside. Near the corner, at SU 293640, the bank is almost indiscernible, with a steep slope leading to a deep channel to the north exaggerating the prominence of this natural ridge. Between its course and the eastern side of the wood an old oak pollard with a girth of 450cm (Figure 59) is close to the meeting point of three parishes. The eastern side of the woodland, between points S and Sa, has a single bank and ditch, 11m wide and up to 1.25m deep, with the ditch to the east (Figure 60), which forms the boundary between Great Bedwyn and Shalbourne parishes. A further bank with its ditch to the north runs east from point S to T and is the division between Little Bedwyn and Shalbourne parishes (Figure 65). Research, documentation and correspondence with boundaries This feature is part of the Bedwyn Dyke Scheduled Ancient Monument (AM593). A detailed 1820 estate map (WSA1300-375L) indicates that the track and holloway in the lower part of the Harding Valley had not been established by this date (Figure 61). The map shows nothing related to the putative dyke from Q to S. Colt Hoare (1819) mapped an earthwork here, which he interpreted as a branch of the Wansdyke, but ‘not a part of that great rampart’. He Fig. 60 The large bank forming the boundary between Shalbourne and the Bedwyn parishes (described under Site 35b). Fig. 61 Extract from Savernake Estate map, c1820, showing land interpreted as part of the Bedwyn System, Site 35. Section B-P relates to 35a Harding Valley; Q-S to 35b, Round Copse; and S-U to 35c, Merrell Down. The map also shows the approximate location of Site 31, Bedwyn Brail Chalkpit Bank. linked it with the section to Bedwyn Brail, south of the Bedwyn-Shalbourne Road (Section 35a). Major and Burrow (1926) considered the feature to be ‘a south-eastern branch of Wansdyke’. By contrast Burne (1953) considered that the section looping from Chisbury through Round Copse almost to Wilton Windmill, represented an earlier course of the Wansdyke, later abandoned in favour of a branch eastwards from Merrell Down. In his analysis of the Anglo-Saxon charters of Wiltshire, Grundy (1919, 1920) considered that the ‘dices gaete aet haradene’ (dyke’s gate at Harding) was at the northern part of the Copse. Working from the same evidence, Crawford (1921) accepted that the feature through Round Copse was part of the Wansdyke. In contrast to Grundy, he deduced that the charter bounds ran along the Bedwyn-Shalbourne Road from Jockey Green, meeting the dyke at the far south-western corner of the wood, where he determined dices | THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 245 gaete to have been. By 1953 Crawford considered the Round Copse earthwork to be a branch of the Wansdyke, claiming uncertainty about the whole section. Fox and Fox measured the feature near the road as 45ft (13.7m) wide, whilst at the north- eastern corner it was a small flat topped bank and ditch of just 25 ft (7.6m). They viewed this structure as linked with a double bank system across Merrell Down to the north (35c), but ‘saw no valid reason ... for calling any of this earthwork Wansdyke’. Hostetter and Howe (1997) report a bank and ditch up to 18m wide running through Round Copse dyke which they argue is linked to the Bedwyn ‘system’ and mentioned in the various charters. Observations The scale of this earthwork varies in accordance with the underlying geology. Only when overlying Upper Chalk does it attain a significant size. The remainder, on the Reading Beds, is broad and shallow. Sizeable chalk pits are found along the western edge of the wood with sand and chalk pits to the south east. A massive pit c. 170m long straddles the parish boundary to the east. Major brick-works at Merrell Down north-west of the copse, in existence since at least the 18th century, provided material for the Kennet and Avon Canal (WSA 9-10-261). Brick- making was undertaken at Folly Farm, south-east of Round Copse, and in the nearby Brails since the 16th century. The scale of the holloway in the south- western part of Round Copse suggests substantial traffic, perhaps partly in association with mineral working, apparently after 1820 (Figure 61). Conclusion Despite being long-recognised and with a purported link to the Bedwyn Dyke and pre-Conquest charter evidence, doubt remains that this feature is a unified construction. Its scale is exaggerated by natural and man-made feature, and while the southernmost part has a modest bank it is otherwise a large post-1820 holloway. To the north, a massive bank lacking a ditch (Figure 65) is perhaps best explained by mineral extraction. The majority of the feature is no more than a small mounded ridge, given slight prominence by the sunken track alongside. The larger scale of the parish boundary bank at the corner (Figure 60) might provide an alternative course for the dyke mentioned in pre-Conquest charters. Archaeological excavation would help resolve outstanding questions. 35c Bedwyn System: Merrell Down Site 35, Figure 1. Points S to U, Figure 51, and aerial photograph Figure 56. Location This feature is 1km east of Great Bedwyn, north-east of Jockey Green. Description This section, on the London Clay and Reading Beds, crosses the level ground of Merrell Down. It was ploughed sometime after 1953, and is known only from historical accounts and aerial photographs. It sweeps in a broad arc across the former downland, from Foxbury Wood (point U) at $U29136425 to the north east corner of Round Copse (point S) at SU29316399. Here it meets the section 35b, and also the large parish boundary bank (S to Sa) between Great Bedwyn and Shalbourne. It features on aerial photographs, both prior to ploughing and as prominent soil and crop marks afterwards, even into the 21st century. These, and OS 25”/mile maps of 1900 and 1924 (sheet XX XVIII), all show a double system for most of the course, some 20-27m wide, merging to a single structure 100m east of point U at Foxbury Wood. Research, documentation and correspondence with boundaries The Merrell Down feature is shown as a wide curving strip of woodland on a Savernake Estate map of 1820 (WSA1300-375L), with rough ground to the south-west and arable to the north-east (Figure 61). Colt Hoare (1819) interpreted the line across Merrell Down as Wansdyke, but without providing measurements or description. He initially thought that the route turned an ‘unaccountable angle’, descending through Round Copse, but subsequently concluded that this was ‘an erroneous direction’, deducing that the dyke must have gone east towards Bagshot Mill, between Bird’s Heath Farm and Newtown Common Farm (Figure 53), although he did not locate any junction with Merrell Down. Major and Burrow (1926) concluded that there was an earlier arm of the dyke extending eastwards, a junction, almost at right angles, on Merrell Down, just north of the bridleway (probably near SU 29256413), although its course became indiscernible almost immediately. There is a contrast in scale between this and adjoining sections. With tree cover removed it was described as a ditch 12 feet (3.7m) wide between two 246 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE banks, with an overall width of nearly 70 feet (21.3m). Crawford recorded the site as ‘still unploughed’ in 1953 and found this contrast in morphology ‘rather disconcerting’ but remained convinced that the twin structure across Merrell Down could be followed, without a break, as part of the Bedwyn Dyke. Hostetter and Howe (1997) accepted the course across Merrell Down as part of the same system. Lennon (2010) postulates that the crop mark across Merrell Down is modern, contending that there was a gap in the dyke here where hill woodland stood and that this represents the ‘dices gaet’ recorded in the 968 charter (K1266). Across Merrell Down the feature forms the boundary between Great and Little Bedwyn parishes, although this alignment was not adopted until the latter 20th century. The feature is a component of the Bedwyn Dyke Schedule Ancient Monument (AM593). Observations Arable land on the 1820 estate map shows that Merrell Down was not all ancient and unploughed land (Figure 61). A short straight section of earthwork visible on an aerial photograph (Figure 56) could be a continuation of the dyke in Foxbury Wood (section 35d) to which it is connected. The second parallel line may thus represent a feature, pre-dating inclosure and ploughing, or a secondary addition. Such a wide double bank suggests a track. Twenty acres of ‘Merrell Down and drove’ were included in the letting of Harding Farm in 1575 (WSA 9-14-223). The 1820 map shows no track within the woodland, but land to the north east divided into three arable fields. The former division between two of these might be the right-angle junction described by Major and Burrow (1926), which they interpreted as the original line of the dyke. This ‘branch’ cannot be traced on aerial photographs although the course of the curving feature is clear. Conclusion This crop-mark forms part of a continuous line from Chisbury hillfort and therefore appears to be a part of the Bedwyn System. The twin banked structures described in the mid 20th century, and persisting as crop marks, appear to relate to a broad woodland strip. It remains unclear whether the feature is linked with section 35b through Round Copse, or with the prominent parish boundary ditch along its eastern side. Archaeological excavation could resolve these issues. There is no tenable evidence for supposing that it branched part way across the Down. 35d Bedwyn System: Foxbury Site 35, Figure 1. Points U to V, Figure 51, and aerial photograph, Figure 56. Location This feature is 1km east of Great Bedwyn. Description This earthwork runs for 160m in a straight course orientated at 310° across a level plateau. It extends from point U at SU 29136425, where its eastern course (if the same feature) is traceable only as crop marks across Merrell Down (section 35c), to point V at SU 29046440, an area heavily modified through mineral workings with pits related to brick-making since at least the 18th century. It forms a wooded bank and ditch, 14m broad, and 1.65m deep, ditch to the north-east (Figure 65), with fine mature oaks, beside a track, and is the boundary between arable fields. This formerly wooded landscape was extensively grubbed up in the late 20th century. The bank is eroded along part of its course (highest where protected by large oak root-plates), but is always a substantial structure (Figure 62). The Fig. 62 Large bank and ditch forming part of the Bedwyn System at Foxbury (Site 35d). THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 247 feature overlies Tertiary deposits (Reading Beds and London Clay). Research, documentation and correspondence with boundaries This section is labelled ‘Wansdyke’ on the Ist edition OS 1”/mile map of 1817. Colt Hoare (1819) and Major and Burrow (1926) included it as part of the Wansdyke. Fox and Fox (1960) described it as ‘crest- sited and well marked ... along the edge of Fockey Copse and Foxbury Wood, measuring 40ft (12.1m) overall and 20ft (6.05m) on the scarp’. The feature is part of Bedwyn Dyke Scheduled Ancient Monument (AM593). In the latter 20th century it was adopted as the revised boundary between Great and Little Bedwyn parishes. Observations | Similarities of scale and orientation between a remaining fragment in the Bedwyn Valley (section 35f) and the Bedwyn-Shalbourne parish boundary dyke along the edge of Round Copse (section 35b), suggest that these earthworks are related. Conclusion Field observations, historical accounts and aerial imagery indicate that this feature is one of the least modified sections of an earthwork traceable to Chisbury. 35e Bedwyn System: Jockey Copse to Burnt Mill Site 35, Figure 1. Points V to X, Figure 51, and aerial _ photograph, Figure 56. | Location _ This feature is on the eastern side of the River Dun _ valley, north-east of Great Bedwyn. Description This earthwork extends for 970m from point V, a surviving fragment of Jockey Copse, at SU 29046440, through grubbed up woodland and another parcel of Jockey Copse (point W), to the Kennet and Avon Canal north-east of Burnt Mill Lock, at point X, | SU 28326495. With a gently meandering course on | ageneral alignment of 310°, no part of this earthwork retains its original morphology. At point V, on the top of the hill, the land has been considerably altered by mineral excavation of Tertiary clays, and there is almost no surviving structure, although its course is marked by a modern track. Further west the feature overlies Upper Chalk and can be traced as an eroded, shallow bank and ditch approximately 10m wide and 0.5m high, often little more than a change in ground level. It can be followed through Jockey Copse to point W, where it emerges into an arable field. Here, on the north-west facing side of the valley, the dyke is indistinct at close quarters, but shows as a soil mark on aerial photographs (e.g. NMRS501 of 1946, Figure 56, and NMR SU2864/3 of 1972). Its course is discernible from the air and from the far side of the valley in certain ground conditions or shallow snow. Research, documentation and correspondence with boundaries This feature is part of the Bedwyn Dyke Scheduled Ancient Monument (AM593). It is labelled ‘Wansdyke’ on the first edition OS map of 1817, and its course (but not morphology) was described by Colt Hoare in 1819. Burrow and Major (1926) published a drawing showing the dyke at Burnt Lock (since ploughed) resembling a banked green lane. Whilst rejecting its identification as the Wansdyke, Fox and Fox (1960) described it as well marked along the edges of Jockey Copse and Foxbury Wood, implying that there might then have been little distinction in morphology, whereas now only the latter (35d) has the structure of a dyke. The feature is the modern boundary between Great and Little Bedwyn parishes. Observations It is likely that a track ran alongside this feature, linking it with the road to Chisbury. Conclusion Although heavily modified, available evidence suggests is part of a continuous feature linking (at least) Chisbury and Merrell Down. 35f Bedwyn System: Spaines Site 35, Figure 1. Points X to Z, Figure 51, and aerial photograph, Figure 57. Location This feature is north-eastern of Great Bedwyn. Description This section extends across the Bedwyn valley from 248 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE — Burnt Mill, point X at SU 28436495 to Z, at SU 28236525, on the road between the Bedwyns. Across marshy ground by the River Dun, from the Kennet and Avon Canal to the mainline railway, no ground feature is discernible, although its presumed course is marked by a line of poplars. Beyond the railway, at point Y, a banked hedge with a ditch on the north- east side, 14m wide and 2.8m deep (Figure 65) is perhaps exaggerated by a 1.8m change in ground level between the fields at this point. The feature diminishes in size to the north-west, merging with ground level nearer the road, where its course is marked by a hedge. Research, Documentation and Correspondence with Boundaries Colt Hoare (1819) did not describe this section, but considered it part of the Wansdyke. Major and Burrow (1926) identify it as a field boundary marking the dyke’s course. From the canal to the railway the feature defines the modern boundary between Great and Little Bedwyn parishes, after which the parish boundary leaves the course of the presumed dyke, dog-legs south-west along the railway and then resumes its mid-l6th century alignment, running between Tyle Field and Barr Field towards Chisbury Wood. This section lies within the Bedwyn Dyke Scheduled Ancient Monument (AM593). Observations Construction of both the canal and railway are likely to have impacted upon the topography of the valley. It is not clear whether the only sizeable part of this earthwork, adjacent to the railway at point Y, is unmodified and represents a dyke. Conclusion Whilst a continuous feature, it cannot be demonstrated from surviving field evidence that this was a true dyke. 35g Bedwyn System: Pumping Station to Hillfort Site 35, Figure 1. Points Z to Zb, Figure 1951, and aerial photograph Figure 57. Location This feature is 200m north-east of Great Bedwyn. Description This section commences at the Thames Water pumping station between Great and Little Bedwyn at point Z, SU 28236525 and may extend as far as Chisbury hillfort at Zb, SU 28026575. The line meanders on the hillside, with an overall angle of 337° degrees to the assumed terminus and is situated on the south-east facing flank of the River Dun or Bedwyn Brook; apart from a small section on Reading Beds, the feature overlies Upper Chalk. For 100m alongside the road from the pumping station to the hillfort, Z to Za the earthwork is massive and comprises a steep, single east-facing bank and ditch, 19m wide and 3.2m high (Profile, Figure 65; photo Figure 63). North of this point the earthwork is heavily modified, comprising a variable roadside bank with no evidence ofa ditch. It seems likely that any former dyke here has been modified by holloway erosion, lynchet development or other agricultural use. Research, documentation and correspondence with boundaries The major section Z-Za is absent from George Wyther’s detailed 1612 map of the demesne of Chisbury (WSA 9-8-153H), which includes ‘Bryer Fig. 63 The massive earthwork of the Bedwyn System south of Chisbury Hillfort, amidst dense scrub near the Pumping Station (Site 35g). The bank here is larger than any other earthwork in Savernake except the Wansdyke. THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 249 Fig. 64 Extract from map of the Savernake Estate c1820, Chisbury hillfort at top. The map gives no hint of the massive excavation, hitherto claimed to be part of the Bedwyn System (Site 35), between points Z and Za. The field arrowed is The Lawn, described under Site 24. Pea Fylde’ to the north-east, but not the road to the south-west. Thomas Atkinson’s 1719 map of Chisbury Manor (WSA 1300-372) shows a significant feature (a presumed bank) east of the Chisbury road from the Bedwyn Road as far as the hillfort (Figure 43). No distinction is made for the major excavation where the pumping station now stands. Minor parcels of land are shown alongside the road, including a wood (named New Leaze Coppice in | l _ the accompanying ledger), and a broader meander of presumed roadside waste, visible on aerial _ photographs. Perhaps significantly, alongside the road, but to the west, a third flecked line is labelled Woden’s Ditch on the 1719 map. Significantly, Atkinson does not equate the Wansdyke with the present earthwork, also absent from the 1820 estate map (Figure 64)(WSA1300-375L). Colt Hoare (1819) stated that the feature (part of the Wansdyke) _ “descends [from the hillfort] to the road which leads to i Great Bedwyn’, without mentioning the earthwork at this junction. He considered that the dyke departed from the road near the ramparts of the hillfort and then skirted a declivity in the hill before joining the dyke through Park Copse (Sites 23 and 24). Pitt-Rivers (1892) examined the large feature in this locality, which he accepted as part of the Wansdyke. His measurements of a feature 61ft5in (18.72m) wide and 9ft8in (2.95m) deep and his cross section are virtually identical to those made during this study (Figure 65). He considered that the Wansdyke continued south from Chisbury Hillfort to the area of the Brails, although his small-scale map is inadequate to discern the exact course. Major and Burrow (1926) argued that the earthwork was part of the Wansdyke and continued all the way to the hillfort. They concluded that it changed to the south-west side of the road 10 rods (50m) from the outermost vallum and went through the centre of the fort to Chisbury village before joining up with a supposed dyke to the north of Park Wood (Site 26). Fox and Fox (1960) considered that the dyke began ‘well-below the hillfort? and that road construction obscured whether it ever continued thus far. A curve in the road south-east of the hillfort ensures that the hill is climbed obliquely and less steeply, arguably a deviation more appropriate to a trackway than a dyke. Fox and Fox measured the large section Z to Za as 55ft (16.8m) overall and 20ft (6.1m) on the scarp, comparable with the East Wansdyke in West Woods. Hostetter and Howe (1997) note that apart from the section at the pumping station the course of the dyke is marked by a 3m high lynchet. The earthwork is part of the Bedwyn Dyke Scheduled Ancient Monument (AM593). Observations Section Z to Za is out of character with other parts of the Bedwyn System, which raises questions about its origin and attribution. This concern is exacerbated by the fact that it does not feature on Atkinson’s map, whose ledger frequently describes parcels of only one rood (0.1ha), nor on the 1820 estate map or in Colt Hoare’s description. It is plausible that this excavation is a linear quarry. While the earthwork profile (Figure 65) suggests a close correlation between excavated and deposited material, the area is heavily modified by lynchets and road construction and further work is needed to identify whether material has been removed from site. Hostetter and Howe (1997) felt that agricultural activity could have filled in ‘the presumed ditch ...up to the top of the bank’ beyond this stretch. This view represents a substantial loss of topsoil and it is questionable whether the topography favours such a drift of material in the upper reaches. 250 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE | Even assuming the identification of this feature is correct, explanation is needed as to why the Bedwyn road separates the most dramatic part of the Bedwyn System from a slightly misaligned and inconsequential field boundary opposite (35f). Conclusion Like other elements of the Bedwyn System, this earthwork generates many questions. Detailed work on profiles and archaeological excavation along the course of the presumed infilled ditch might elucidate whether the feature has modern origins and is a single integrated structure. General conclusions concerning the Bedwyn System There is strong evidence for a geographically continuous system, of some kind, for most (if not all) of this feature. However, surviving portions display disparate morphological characteristics and are not unequivocally related. The largest earthwork south-east of Chisbury hillfort (point Z, Figure 51) is unequalled in scale in Savernake except for Wansdyke itself, but does not feature on early maps and adjoins a section where barely any earthworks are discernible. The least modified dyke appears to be the section U to V at Foxbury, which is broadly analogous to two other parts, one northwest of Burnt Lock (point Y), where the adjoining railway construction could have modified local topography, and the other in Round Copse forming the parish boundary between Great Bedwyn and Shalbourne (S to Sa). The latter is not generally regarded as part of the Bedwyn Dyke system, but maintains a coherent line with the route across Merrell Down and sweeps down to the road. Lennon (2010) postulates that these dykes were part of an arrangement fanning out from Chisbury, from which they were largely visible, enclosing a territory centred on Great Bedwyn. Sections included within the system grouped and described here generally have a bank with a ditch facing north-east and are highly conspicuous, meandering around a broadly steady orientation, often with an adjacent track. A thoroughfare is traceable alongside 90% of the 1900m Bedwyn System from Chisbury to Merrell Down (including the section drawn as a green lane near Burnt Mill by Major and Burrow, 1926). The exception is Jockey Copse, where a broad ditch might represent the same track, connecting other known parts of the route. A continuous track seems probable, but does not mean that a dyke was present along its entire course. The ends of the system present the greatest uncertainty, with little evidence that the Harding Valley feature (B-P) is either a dyke or linked to the remainder of the feature. The generally accepted course through Round Copse is also problematic, with stark changes of direction and a combination of structures, including a mounded ridge with no adjoining ditch. The double feature of Merrell Down is unlike any other section. At the opposite end it is unclear whether a dyke continues to Chisbury hillfort, or whether the feature at the pumping station is modern or part of a formerly more extensive system. Localised archaeological excavation could address many of these difficulties. A useful exercise would be to investigate parallels with pre-historic land divisions on Salisbury Plain, where dykes appear to spread out from a central high-point, where a hillfort was later inserted (Bradley et al., 1994). Conclusions This paper has sought to provide a catalogue of surveyed sites, together with summaries of existing documentary evidence and research. Tentative conclusions offered in each case open up the prospect of further research, rather than definitive identifications or attributions. The survey has confirmed that Savernake is riddled with many ancient and interesting features. Features which can be dated span the Bronze Age to modern times and many alignments have been used successively for different purposes. Banks attributed to modern estate boundaries appear to follow 14th century Forest perambulations and are likely to relate to the King’s land, and potentially, also to boundary marks . in Anglo-Saxon charters. LiDAR aerial imaging has advanced our knowledge in a spectacular way. It has enabled the detection of features which cannot be seen on the ground, either because they are cloaked in vegetation or because they are too subtle to discern at close quarters. Areas once regarded as pristine and virgin woodland have at various times been ploughed and settled. Savernake, like much of southern England, has undergone spasmodic periods of intensive use followed by abandonment. A better model is needed to understand those woodlands currently classified THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 251 17 01 ae Panterwick | 02 : Furze Hill - : : i 17 OE: Postern Gt Dyke Bagden Ill 04 | Forest Hill Crescent 05 Forest Hill Bank 17 Bagden IV 06 47 Shepherds Wood Bagden V 07 47 Puthall Park Bagden VI 08 Hens Wood south 17 Bagden Vil 08 Hens Wood central 18 ‘ 09 Bishop’s Rag Bushell’s Linear 19 410 pet tb Great Lodge Bottom : 14 Chisbury Common Savernake Gt Inclosure 24 Warren Ditch 12 Shovel Bottom 23 Chisbury Wd 13 Park Copse Crabtree-Amity ; 44 24 Chisbury Wd Little Frith Linear Eig itan 25 Chisbury Wd 14 Scar Little Frith Enclosure 26 Chisbury Wd 14 | North Little Frith Coppice 27 Mud Lane 15 Noke Wood 29 Ram Alley Brirnslade Pk 16 Wansdyke 29 Ram Alley Copse 30 Durley- | |Tottenham Linear Yi, 31 Bedwyn Brail Chaikpit Bank 32 Bedwyn Brail Conduit north 32 Bedwyn Brail a Conduit south 7 L 33 Bedwyn Brail Common Bank N 33 Bedwyn Brail Common Bank E 35b Bedwyn System Round Copse 35b Bedwyn System Round Cop parish bo. 35d Bedwyn System Foxbury 35f Bedwyn System Spaines 35g Bedwyn System Pumping Stn 0 metres Fig. 65 Cross sections of earthworks within the study area. Vertical scale 1s exaggerated to emphasise site characteristics. Standard 1.83m (6ft) human figure inserted to provide visual impression. Site data is provided in Fig. 66. as ‘ancient’. Existing research provides a salutary reminder of the need for thorough field recording. Published records from 1819 to 1997 are often grossly inadequate for evaluating claims, due to a failure to map, give exact locations, or even label cross-sections. In contrast, Major and Burrow (1926) provided records of every location visited, with detailed maps, and over a hundred paintings executed in the field. This laid them open to castigation in the archaeological world, because their mis-attributions were readily exposed. However they enabled other workers to assess their claims, and in this respect, if no other, their work was exemplary. Inadequate descriptions of morphology and scale are especially frustrating when it is suspected that sites might have been modified since first recorded. The naming of sites needs to be ruthlessly avoided unless identification is certain as it carries connotations of origin, function, and relationship. In Savernake there are dozens of sites which have, at various times, been called ‘Wansdyke’ (but only in the last 300 years). None of these can be demonstrated to deserve that epithet and until substantive evidence is available such naming is misleading, if not pretentious. Where the functional relationship between geographically linked sections remains unproven, it may be better to refer to such areas as systems or complexes. Even where continuous earthworks have been claimed, such as the Bedwyn System (site 35), dramatic inconsistencies in scale and morphology raise serious questions of attribution to a single coherent system. Local questions require resolution before uniting this complex with large- 252 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE No. |Site Name IGridref SU: |L_[R__| 3 4 [Forest Hill Crescent _ [21025 68404180 [0 | 16_|Shepherds Wood _—[22838 68187|150 [330 | [8 [Hens Wood south _—164632_68343|206 [26 | [8 [Hens Wood central __ [24227 68708|8 [188 | [9 |Bushell’s Linear [21138 66953(211 [31 | 114 [Little Frith Enclosure [24111 67455 |240 [60 | 26613 67816 18802 66543 @) SS D lon c 5 @ fe) 5 =| fo) Ss Nn N N NI ON ON R \O 5 vs) 5 > — — fo?) (®) ie) oO Nn oO i) — —“S WM S [oy Ww N Let (=) 2 34 34 36 36 2 30 26 25 Chisbury Wd Park Copse|27166 65749 |224 |4 14 6 8 2 88 84 6 8 8 10 DD 0 Durley-Tottenham 24287 6329010 Linear [Bed Brail Conduit north |28243 62369|296 [116 | i eal lal St i ned ll Bank N a ee Bank E a eet ra Op a ee parish bo. 35d |[Bedwyn System Foxbury |29151 64322 [34 [214 | 35£ [Bedwyn System Spaines [28340 65083 [56 [236 _| 54 54 56 56 22 204 50 180 146 305 4 294 336 348 200 122 108 116 118 130 142 300 214 236 242 Ww Mr it je} pes} ct 58. <= . 5 WN ce DN ct oO 5 rg Cc =) SS bo co we Ne) Nr [o> wn US) — ix NO Fig. 66 Table of site data relating to cross sections drawn in Fig. 65, showing number, name, grid reference, and compass orientation in degrees to left and right as drawn. scale polities. Some of the false leads seized upon in earlier research might have been avoided through a more vigorous inter-disciplinary approach, involving documented history and archaeology coupled with modern concepts of landscape ecology. The present work has revealed how light can be thrown on many of the features in Savernake through an examination of surviving archives. Consulting key documents, including detailed cartographic sources, might have removed certain fanciful claims which have remained speculative for decades, or at least offered alternative plausible explanations. Structures with a prosaic origin, such as the former Bedwyn-Marlborough Road (closed 1795), have been interpreted as ancient earthworks. Even where documentary sources are inconclusive, current research has thrown considerable doubt upon identifications previously made for several features, including important parts of the so-called Bedwyn Dyke. Where archival sources are incomplete, surviving documents frequently complement one another. In broadest terms, many aspects of the landscape flow from the extent of the King’s demesne in Anglo-Saxon and Norman times. The extent of Royal holdings appears to dictate the location of extra-manorial and extra-parochial land, areas encompassed by Forest perambulations, and those areas free from obligation to pay tithes, poor rate and other local taxes. Legal disputes concerning tithes, whilst tortrous to unravel, are an under-used resource in understanding the landscape, since tithe-free status can reveal the King’s demesne before the Lateran Council of 1180, with plausible links to pre-Norman Royal holdings (Cunningham, 1777). A better appreciation is needed concerning how the historic and prehistoric landscape may have appeared, and also the former extent of Savernake. Claims have sometimes been made that the Wansdyke ceased west of Savernake, although its current terminus would have been in the heart of Savernake until the 16th century. Rather, it is Savernake that has retreated from its former extent. Similarly, it has been suggested that the Wansdyke © ended upon reaching dense woodland. There is no current evidence that Savernake had continuous tree cover before coppices were established in medieval times or major plantations enclosed from the 18th century. Instead, records suggest that it was an open park-like landscape with groves and large trees, but not dense closed-canopy woodland in the modern sense. Hence the notion that Savernake ever had woodland impenetrable to a military force has little credence. A different model is needed to explain the observation that long distance earthworks may cease at various stages along their course especially ee THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 253 in clay landscapes. One interpretation is that these areas were unenclosed and uncultivated common land, used collectively for pasture, hunting and fuel, where the vast efforts necessary to define territory could not be justified. Over 290 years, after it was first mooted, there is not a single piece of field evidence to support the theory that Wansdyke continued beyond its current terminus (and strong circumstantial evidence that it could not have done so). Yet until now the idea itself has shown no sign of abating. Wansdyke provides a prime example of the ‘subtle and inevitable hold that theory exerts upon data and observation’ (Gould, 1989). On occasion archaeology seems to have turned from an evidence-based to a faith-based discipline. At a local level claims can persist tenaciously generating a momentum of their own. Indeed, in one case, where a putative section of Wansdyke was later entirely rejected by its original discoverer as ‘wishful thinking’, other researchers sought to resurrect it, stating that it should ‘not be completely discounted’. Current evidence suggests that if the Wansdyke was ever ploughed out beyond its current terminus, this must have occurred before 1130, which is an unlikely prospect. Despite this, there is no reason why caesura could not occur at any scale. It has already been postulated that cross-valley dykes defined territory at the micro-scale and Offa’s dyke is now recognised as a continuous earthwork with gaps in certain terrain. While there is no evidence that Wansdyke continued through Savernake, this research does not deny the possibility that it resumed at some point beyond the study area. The survey has identified further work, which could be fruitful in expanding our understanding of Savernake, including comparisons between the Bedwyn System and the pre-historic land divisions on Salisbury Plain. Detailed evaluation of the banks through Shepherds Wood, an investigation of Hens Wood as a possible extensive former wood pasture, a review of cross-valley dykes, more detailed work on the Magnum Fossatum, an inspection of the ovoid bounds of Pantawick, and further mapping of the complex of features around Postern Great Dyke are much needed. It would also be useful to examine whether a single line, possibly equating with Anglo- Saxon charters, can be traced from Chisbury Wood, through Bedwyn Common, and along Bishop’s Rag and the Crabtree-Amity Linear. There is also scope using LiDAR and excavation to confirm (or refute) the course of the elusive Roman Road from Cunetio to Spinae. The fact that none of these are physically linked to the iconic Wansdyke should not detract from their interest or scope for more detailed investigation. Acknowledgements We would like to thank Ben Lennon and Peter Crow of the Forestry Commission for access to LiDAR images and for vigorous debate; Forester Chris Sorenson concerning access timing; Dick Greenaway for advice on survey techniques; Joan Davies for contributions to the history of Savernake; Richard Charles for consent to visit private land at The Brails, and much else; Jim Gunter and the WANHS Archaeological Field Group for survey data; Stella Maddock for information on Colt Hoare’s work near Bedwyn; Simon Crutchley and Fiona Small for evaluating tentative interpretations, and for emerging data from the English Heritage National Mapping Programme; Jill Butler and Ted Green for advice on predicting tree age based on girth, and Matthew Steadman Jones for assistance with fieldwork. References Original Sources Wiltshire and Swindon Archives (WSA) WSA 9-1-5 Schedule of Deeds 12th century -1725. WSA 9-1-95, Survey of the manors of Pewsey, Shalbourne Westcourt alias Dormer. Collingbourne Kingston. Chisbury and Bedwyn Prebend 1552. WSA 9-1-287, Account Book 1769-1770, with wider notes. WSA 9-1-509, Coarse map showing designs for entrance banks, Savernake Lodge, late 18C. WSA 9-3-388 Legal brief and associated papers relating to law suit, concerning trespass on common land in Great Bedwyn, 1760-1761. WSA 9-3-398, Enclosure Act for Great Bedwyn, Little Bedwyn and Preshute, 1790. WSA 9-5-12, Lease and release in fee of farmhouse and land near Burwood’s Heaths, between John Hunt of London, and Thomas, Lord Bruce, October 1766. WSA 9-6-6, Letters of attorney relating to the seisin of Mottesfounte Coppice May 1599. WSA 9-6-769, Undated sketch map of part of Durley Common adjoining Tottenham Park, c1700. WSA9-7-42, Legal Briefs relating to land disputes, 1589-1786. WSA 9-8-26 to 28, Papers relating to Chisbury Common inclosure, 1702-1722. 254 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE WSA 9-8-153H, Map labelled A New Plotte of Lands belonging to North Standen, The Demesne of Chesburie and the Farme of Froxfylde, 1612, George Wyther. WSA 9-10-261, Accounts of work at the brick yard, Merrelll Down, apparently for the Kennet and Avon Canal Company, 1795-96. WSA 9-14-223, Lease to John Chocke of Harding Farm in Great Bedwyn, 1575. WSA 9-21-44, Survey of Wootton Rivers 1574. WSA 9-22-14, Transfer between Sir Edward Seymour, Earl of Hertford and Sir Francis Seymour of 100 acres lately taken out of Savernake Great Park, 1617. WSA 9-22-42, Lease Durley Lodge and Warren, 1673. WSA 9-22-204 Sale of strip of land for The New Park 1575: WSA 9-22-14/22/24/25/28/33, Conveyances of land in former Savernake Great Park, 1618-1651. WSA 9-22-206, Sale Brown’s Farm Widow’s Coppice, and Pantawick Lodge from Sir Edward Seymour to Sir Francis Seymour, 1618. WSA 9-22-255, Legal Cases relating to Granham Farm and coppice and Jennings’ Farm, 1759-1777. WSA 9-27-59, Benjamin Haynes’s Plan of Wilton in Great Bedwyn, 1792. WSA 9-35-261, Letters from John Soane concerning designs for Savernake Lodge, 1795. WSA 383-907, Undated sketch map of Round Copse and Castle Copse (Bedwyn Brail), possibly preparatory to Colt Hoare’s Ancient Wiltshire (1819). WSA 1300-20, Grant by Henry Sturmy of land at Puthall, part of Wexcombe, to the Holy Trinity of Eston, 1368. WSA 1300-46, Map 1668 labelled A Survey of that Part of Savernacke Forrest which 1s pretended to bee in the Parish of Preshut. WSA 1300-46A, Abstracts of evidence in tythe dispute, Davis and others v Hitchcock 1668. WSA1300-47 Coarse map reconstructing early 14C perambulation of Savernake (probably associated with tithe dispute of 1668). WSA 1300-47A, Notes from H C Brentnall relating to WSA 1300-47. WSA1300-59C, Roll of Henry Sturmy, Warden of Savernake Forest, of attachments and inquisitions of vert and venison, 1329. WSA 1300-65, Copy of Bounds of Savernake forest, divided into sections for towns, woods and groves outside forest, 14th century [Edward III]. WSA 1300-87, Curiae Liber Savernake 1577-1609. WSA 1300-90, Survey of properties of Edward Earl of Hertford late 16C. WSA 1300-100, Summary of Evidence in tythe dispute, Davis and others v Hitchcock 1668. WSA 1300-166, copies of Letters Patent Edward VI, 1547, granting land to Edward Seymour, Duke of Somerset. WSA 1300-358, Map of Tottenham Park, early 18C. WSA 1300-360, A Plan of the Forest of Savernake and Tottenham Park prepared for the Earl of Ailesbury 1786. WSA 1300-362, Letters relating to works at Tottenham, 1743. WSA1300-363, Letter, John Ward to Lord Bruce, 1795, concerning closure of the Old Marlborough Road across Savernake. WSA 1300-372, Volume of maps by Charles Price and Thomas Atkinson, cl717, and Survey of Lands held by Lord Bruce. WSA 1300-375L (and WSA X6-4) Savernake Estate Map, c1820. WSA 1300-1484, Volume, compiled by William Longueville, in 1718, containing extracts of legal cases 16-18th centuries. WSA 1300-1917, Copy of letter from Charles Bill, probably to Lancelot Brown, describing proposed landscaping at Lewdon’s Lye, Savernake, 13 March 1767. WSA 1300-1912 to1927, Minutes of Mr. [Lancelot] Brown’s advice concerning Tottenham 1765-1770. WSA 1300-1928, Sketch Map in hand of Charles Bill relating to proposed landscaping, undated but apparently referred to in WSA 1300-1917 of 1767. WSA 1300-1953, Nursery Book, list of young trees in Savernake nurseries 1785-1787. WSA 1300-2511, Letter from Brounton, Tottenham Park, to Lord Bruce, Dec 15 1782. WSA 1300-2539 & 2540, Instructions and memoranda concerning work at Tottenham, 1784-5 WSA 192-44, Accounts of workmen, Savernake, 1616. WSA 212B-7189, Copy of Perambulation of Savernake [Edward I] WSA 2027, Map of Savernake Forest, surveyed by A M May, 1845. WSA 2667-24-5-1, Copy of Perambulation of Savernake [Edward I]. WSA 3354 MS, Map of Little Bedwyn and Froxfield, 1846. WSA 3354 MS, Map of Brimslade Farm, 1842. WSA 3354, Unsorted material relating to the Savernake Estate. WSA EA68, Inclosure Award map of 1792. WSA EA68 Plan B, Great Bedwyn Inclosure Award, Wilton Tything, 1792. The National Archive Perambulations of Savernake: PRO E 22-220, Perambulation of Edward III. PRO Chancery Bundle 12 12-22, and PRO E 146 bundle 2 no. 34. PRO F43/150, Notes from Forestry Commission Inspections. Longleat House Archives Thynne Papers, Vol II, Letters John Berwick to John Thynne, 1548-9, ff 30, 45, 62, 88, 98, 129. Thynne Papers, Vol II, Letters Bryan Teshe to John THE LARGER LINEAR EARTHWORKS OF SAVERNAKE 255 Thynne, 1549, ff 41, 60, 68, 72, 83, 87, 100. Seymour Papers, Volume IX, Money disbursed by Bailiff of Wolfhall, 1536, ff 55-56 and March 1548 — February 1549, ff. 92-145. Seymour Papers, Volume XII, Rent Roll of Edward Seymour, Viscount Beauchamp, 1540 (ff. 3-354). WANHS Library, Devizes Wiltshire Archaeological Notes, 1932, Manuscript Letters Rev G. Stallard to Canon J.E. Jackson, May-December 1874. Aerial Images and Data Photographs held by National Monuments Record, Swindon: NMR 6114, RAF797, H23/40, 1942. NMR 2862/2/217-220, 812217-812220, 1975. NMR SU2863 1-2, 1930, CCC8997/9973, 9967. NMR 501, RAF/CPE/UK/1821, 4118 — 4120, 1946. NMR 11135, OS/64160, 1964. NMRI10326, OS/72295, 336, 1972. NMR SU 2165/1, 10-5-1934, CCC9120/9340. NMR SU202686, 20-10-1930. NMR SU1967-11. NMR 18717/18, SU2363/4 February 2000. NMR SU2864/3 April 1974, 094. Photographs held by Forestry Commission: ~ RAF Flight, CPE/UK, 1821, November 1946, 58 Squadron. Photographs 2112, 2117, 2118, 4106, 4112, 5110. LiDAR (Light Detection and Ranging) Data Images derived from Digital Terrain Models, following aerial survey of Savernake Forest and peripheral areas, flown for the Forestry Commission in 2006. English Heritage National Mapping _ Programme for Savernake Preliminary (unpublished) maps collating information on _ archaeological features detectable from a range of aerial surveys, coupled with detailed ground survey. Printed and Published Sources _ ADAMS, W. M., 1905, Savernake in the Vale, Chapter X Wilton, Marlborough Times. ANDREWS, J. and DURY, A., 1773, A Map of Wiltshire. London: Andrews and Dury ANON., c. 1900, Savernake Forest [Leaflet]. Westbury: W. Michael BATHE, G. M., 2006. The Duke of Somerset’s Grand Mansion at The Brails, Great Bedwyn: A Review of the Evidence. WANHM 99, 236-245 BIRKBECK, V., JONES, G., POWELL, A. B., and SEAGER SMITH, R. H., 2008. A Roman Pottery Kiln, Kiln Furniture and New Vessel Forms from Alice Holt, Hampshire. Proceedings of the Hampshire Field Club and Archaeological Society 63, 110-128 BRADLEY, R., ENTWISTLE, R., and RAYMOND, F, 1994, Prehistoric Land Divisions on Salisbury Plain. London: English Heritage BRENTNALL , H. C., 1923. Excavations on the Line of the Wansdyke. Marlborough College Natural History Society 72, 90-97 BRENTNALL , H. C., 1924. Excavations on the Line of the Wansdyke. Marlborough College Natural History Society 73, 71-74 BRENTNALL , H.C., 1942. The Metes and Bounds of Savernake Forest. WANHM 49, 391-434 BURNE, A. H., 1953. Wansdyke West and South. WANHM 55, 128 CORNEY, M., 2001, ‘The Romano-British Nucleated Settlements of Wiltshire’, in P. Ellis (ed.), Roman Wiltshire and After, Papers in Honour of Ken Annable, 5-38. Devizes: WANHS CRAWFORD, O. G. S., 1921. The Anglo Saxon Bounds of Bedwyn and Burbage. WANHM 41, 281-301 CRAWFORD, O. G. S., 1953a. The East End of the Wansdyke. WANHM 55, 119-125 CRAWFORD, O. G. S., 1953b, Archaeology in the Field. London: Phoenix House CUNNINGHAM, T., 1777, A New Treatise on the Laws concerning Tithes. London: J. Johnson, G. Robinson, R. Baldwin and J. Bew CRITTALL, E., 1959, A History of the County of Wiltshire: Volume IV. London: Oxford University Press CROWLEY, D. A., 1999,A History of the County of Wiltshire: Volume XVI, Oxford: Oxford University Press FOX, A., and FOX, C. 1960. Wansdyke Reconsidered. Archaeological Journal 115, 1-48 FRY, E. A., 1908, Abstracts of Wiltshire Inquisitions Post Mortem, Henry III to Edward II. London: British Records Society GOULD, S. J., 1989, Wonderful Life: the Burgher’s Shale and the Nature of History. New York: W.W. Norton & Co. GOVER, J. E. B., MAWER, A. and STENTON, E M., 1939, The Place-names of Wiltshire. Cambridge: Cambridge University Press. GREEN, H. S. 1971. Wansdyke: Excavations 1966 to 1970. WANHM 67, 129-146. GRUNDY, G. B. 1919. The Saxon Land Charters of Wiltshire. Archaeological Fournal 26, 143-301. GRUNDY, G. B. 1920. The Saxon Land Charters of Wiltshire. Archaeological Fournal 27, 8-126. HOARE, R. C.,1819, The Ancient History of Wiltshire. London: W. Miller HOSTETTER, E., and HOWE, T. N., 1997, The Romano-British Villa at Castle Copse, Great Bedwyn. Bloomington: Indiana University Press. HOWE, T. N., and HOSTETTER, E., 1997, ‘Observations Db isn op) on the Bedwyn Dyke’, in E. HOSTETTER and T.N. HOWE, The Romano-Bnitish Villa at Castle Copse, Great Bedwyn, 359-369. Bloomington: Indiana University Press. JACKSON, J. E. 1875. Wulfhall and the Seymours. WANHM 15, 140-207. LENNON, B., and CROW, P. 2009. LiDAR and its role in understanding the historic landscape of Savernake Forest. VANHM 102, 245-261. LENNON, B. 2010a. The Bedwyn Dyke: A revisionist view. VANHM 103. LENNON, B. 2010b. A Historiography of the Wansdyke. WANHM 103. MAJOR, A. EF, and BURROW, E. J., 1926, The Mystery of the Wansdyke. Cheltenham: E.J. Burrow & Co. RAMSAY, J. K., and BATHE, G. M. 2008. The Great Inclosure of Savernake with a note on Cross Valley Dykes. WANHM 101, 158-75. REYNOLDS, A., and LANGLANDS, A., 2006, ‘Social Identities on the Macro Scale: A Maximum View of the THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE - Wansdyke’, in W. Davies, G. Halsall and A. Reynolds (eds), People and Space in the Middle Ages, 300-1300, [PAGE NUMBERS?]. Turnhout: Brepols. SANDELL, R. E., 1971, Abstracts of Wiltshire Inclosure Awards and Agreements. Devizes: Wiltshire Records Society. SAWYER, P. H. 1968, Anglo-Saxon Charters: An Annotated List and Bibliography. London: Royal Historical Society. SMITH, A. C., 1884, Guide to British and Roman Antiquities of the North Wiltshire Downs, Marlborough: WANHS VERA, FE W. M. 2000. Grazing Ecology and Forest History. Wallingford: CABI. WATTS, K. 1996. Wiltshire Deer Parks. WANHM 89, 88-98. WHITE, J., 1998, Estimating the Age of Large and Veteran Trees in Britain. Edinburgh: Forestry Commission (augmented by unpublished tables showing relationships between age classes and size of oaks). Wiltshire Archaeological & Natural History Magazine, vol. 103 (2010), pp. 257-68 Silbury Hill: Edward Drax and the excavations of 1776 by Brian Edwards This paper provides an introduction to two letters by Edward Drax that shed light on the excavation of Silbury Hill in 1776. They describe progress during the excavation and details of features and finds. Introduction There was an air of mystery about ‘Colonel Drax’ when the BBC announced details of the proposed investigation of Silbury Hill in 1967. Drax was associated with the vertical shaft excavated in 1776, the earliest excavation to have focused on the centre of the mound. The only known post-excavation source, James Douglas’s Nenia Britannica (1793), revealed that having excavated Silbury Hill under the ‘supposition of it being a place of sepulture’, the late Colonel Drax of Dorset had recovered what he thought to be whalebone from the centre of the mound, but had been persuaded by Douglas that the artefact was actually a slip of wood. Drax had suggested the mound had been ‘raised over a Druid oak; and he thought the remains of it were discovered in the excavation’. The late Duke of Northumberland, Sir Hugh Smithson Percy (c.1714-1786), had also featured in Douglas’s report, as co-director alongside Drax. Whilst there was no further indication of the Duke’s role, Drax it seemed had possession of the ‘whalebone’, in addition to ‘several specimens of relics found in barrows which he opened on the [Dorset] Downs’. Drax had also featured in the index to Douglas’s tome, whereas the Duke had not.! ‘On the assumption that the Duke provided the patronage and the money, and that the work was done by Colonel Drax’, the lesser known figure posed a line of enquiry for intrigued historian Professor A. C. F Beales (1905-1974),’ of King’s College, University of London. His interest had been welcomed by Richard Atkinson (1920-1994), the archaeological leader of the 1968-70 BBC excavation, who suggested that Beales ‘may well be right in thinking’ that the leader of the Archaeological Institute’s 1849 Silbury tunnel, John Merewether (1797-1850), Dean of Hereford, ‘made no very serious hunt for references’ to the 18th century excavation.’ On making enquiries Beales was informed by the War Office librarian, D. W. King, that whilst there was no Drax in the regular army in that period, in 1759 there was a Colonel Edward Drax with the Dorset Militia, during the Seven Years’ War, who was not listed when Militia Lists were next drawn up for the war with America.‘ Beales had stated that his ‘thoughts were of course to locate’ Drax and ‘see what may have happened to his own papers,’ but what may have been gleaned is not known at this time.° Atkinson, confessing to a fifteen-hour working day, gratefully replied that he would follow up possible sources as soon as time allowed.° From accessible texts it could have been discovered that Edward Drax’ of Charborough Park was born around 1726, and that he was educated at Eton and Hertford College, Oxford.’ On the death of his father Henry in 1755, Edward succeeded him The Wiltshire Archaeological and Natural History Society, 41 Long Street, Devizes, Wiltshire, SN10 1NS io) ‘nn (o/6) THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 1 Edward Drax as a young man in the 1740s. From a private collection. as MP for Wareham, remaining an MP until 1761. The following year Edward married Mary Churchill (1743-1820) at Sturminster Marshall.’ Their only child, Sarah Frances Drax, was born in 1769. Sarah married in 1788, not long before the death of her uncle Thomas Erle Drax (1721-1788). Having died without issue, Thomas’s property seemingly passed to Edward, who died in 1791.!° From such sources as Richard Gough’s Camden’s Britannia (1806) which, as well as mentioning the 1776 excavation, indicated Thomas Erle Drax’s ownership of a seat at Maddington, an impression might perhaps have been gained of the extensive property and wealth inherited by the brothers following the death of their mother Elizabeth in 1759." George Pitt (1720-1803) has not hitherto been suggested as having any connection with Silbury Hill, but in addition to links with the Dorset Militia and Parliament he and Edward Drax evidently encountered one another through society.!? In 1776 George Pitt became the first Lord Rivers, and is revealed as Edward’s confidante through two letters sent to him in November that year. After some teasing preliminaries regarding a letter that went astray, an upbeat Edward Drax reports on the progress at Silbury Hill. Two letters from Edward Drax to Lord Rivers regarding his excavations of a barrow and Silbury Hill dated November 1776 (A working transcription with grammatical and other insertions in parentheses)” Bath November 4th 1776 My Dear L[or]d On my Arrival here'* yesterday from Silbury Hill I found your kind favour. That my Letter Mist its way I can impute only to a gross neglect in the post office in having no one Clerk that can Decypher The Druidical Character in which mine was wrote[,] but I hope your L[or]dship[’]s interest with L[or]d [le] Despencer and Mr Thynne will get that rectified, but for the future I shall direct to Hertford Street, as Hertford bridge seeming more difficult of Access than I hope Kings-Bridge!® will be found. Autour de [About] Silbury Hill it is my Duty to give some Account as I am at present left commandant in the absence of my Syeniors. Your liberal Hint I give your L[or]dship Credit and Thanks for;!’ we have had 4 miners from Mendip at work there ever since Thursday sennight[.]!®> When we had got about ten feet down, I found it was absolutely necessary to have it Timbered or the men’s lives would be endangered. The first six feet I found had been moved when Trees were planted there in 1723, Some few animal Bones bits of Deers Horns &C were thrown in with the earth and a great many small snail shells[,] but they certainly came from the Surface & when the trees were planted were thrown in. As we got deeper down I found the Hill still composed of Large chalk stones as big as a man’s Head thrown in Loosely and great interstices still remained between them[.] Still as we go down we continue finding pieces of Deers Horns of a very Large size[,] in all appearance either of large Stags or else the mo[o]se Deer. At 23 foot a Large piece & at 28 feet a Large piece at 30 Feet another and nothing else as yet[.] When I left them Saturday night we were 31 feet down and as we have by my measurement 125 feet to go before I expect to find the Deposit[,] what ever it may be[,] it will be a fortnight or perhaps 3 weeks before we get to our desired object. I have agreed with these miners at 19s p[e]r fathom for 20 fathom[,] we finding the wood[.] In digging a path to go up the Hill[,] just under the surface I found a key of a very odd and Antique make perhaps 4 or 500 years old but quite modern compared to the object of our Search. I had prepared a Barrow[,] one that I thought of the oldest make,”” and as it was on the Hapken [Hackpen] or Snakes Head and pointed directly to it I thought was most certainly belonging to and coeval with the Temple[,] but as I found you did not come and the men stood still for want of Timber[,] I got Through it Saturday and found it to be of the greatest antiquity made of black mold brought from the neighbouring meadow. Where the urn is usually deposited I found only a Cavity[.] The urn or vesicle SILBURY HILL: EDWARD DRAX AND THE EXCAVATIONS OF 1776 259 in which the bones had been, being entirely consumed by the prodigious length of Time nothing remained in it but a vast heap of burnt bones[,] some Human and Some animals, many Horses Teeth entire[;] but in searching with my Hands amongst Them I found an odd im- plement of what use I cannot conceive as yet — it was ‘Two Stones cut with exactness to fitt of an oblong square 2 Inches & a half Long[.] Another implement of Stone likewise sharpend at the End Half a foot long and Shaped with a Tool[.] I fancy (perhaps it is only an antiquarians Reverie) but I fancy I can make out some Characters upon it very like Hebrew or my own Handwriting;”’ when you happen to be down I should like to open another Barrow which I think may produce some urn or other antiquity as it Lays immediately near the Sanctuary and is more perfect than the other[,] but L{or]d Pembrokes leave must be had as he is L[or]d of the mannour There.”! The Country people wish to have it search[e]d as they say (but that is not my reason) that great lights have been seen on it in the night. Saturday I measured Silbury Hill as accurately as I could and am amazed at the ground it stands upon{,] not much less than most of the Egyptian Pyramids except the great one which I think stands on near 11 Acres. My measure agrees very well with Dr Stukelys except in the perpendicular Height Which he makes 170 feet[,] Mine but 125[,] but I believe he reckons what it was at first and allows for the Grounds settling which is visible from the west side[,] but nous verrons [we shall see] when we get down, on the whole it is vast[.] Dr Stukely says it contains 13,558, 809 cubic feet and would cost £20000 to raise it now but I say according to the rate of moving Earth here which they do as cheap as any where at 34 per Cubic Yard it would cost £56495 — 0s — 9d. So much my Dear Lord for Silbury Hill till next week. I cannot take the Liberty of troubling his Grace of Northumberland with a scrawl of This kind but if your L[or]dship should happen to be sitting by him in the House [of] L[or]ds or elsewhere[,] it would be perhaps Some amusement to his Grace if you will be so kind as to Shew him the enclosed measurement and acquaint him how far we have proceeded. I believe by this time your L[or]dship will find by experience that when a man gets upon his hobby Horse it is a hard matter to get him down 260 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE but I Should not shew the sincere regard I have for everything Belonging to L[or]d Rivers if I did not spare his Eyes and his Patience by concluding this unscrutinable [inscrutable] scrawl. My wife and little Sly Boats” beg their kindest remembrance to Miss [Mistress] Pitt and yourself [my] L[or]d and Mr Geo[rge] Pitt Iam my Dear L[or]d Your obliged and ever Sincere Humble serv[an]t Edward Drax The second letter was written some two weeks later: Bath Monday” Mr Dear L[or]d On getting out of my Chaise from Silbury This morning, I found your kind favour, and as The Uncourtly Gout gives dreadful sounds of preparation I will give a short Account of my last expedition, if I can make it short; for when one’s writing to those one likes writing its so like conversation it is not easy to leave of[f]. To Answer then as short as possible I have hitherto till this last expedition gone by myself and like an old Druid slept a[t] Beckhampton within the precincts of the Temple” but This Time I have taken my faithfull Sarah with me/to do likewise/and glad I am That I did, as we spent yesterday one of the finest days eve that ever the sun shone on very Luminously in the Patriarchal way. We rode over and contemplated all the wonderfull remains of the old Patriarchal Religion with which the environs of Abury Abound more perhaps to those who have a Turn for enjoying and discovering those kind of works than any part of the known Globe. In our ride in a sequestered vale we met With a work which I prevailed on my fellow labourer To get of[f] and Draw It is a stupendous stone Sat up, According to the Account as your Lordship remembers in Scripture; /for I confess I don’t/ and Abraham sat up a stone and call’d on the Name of the Lord.” I will prevail on her to send it you. These sort of works in my opinion have more simple Grandeur, and st[rJike with more religious admiration than the finest works of Greece and Rome; in my opinion there is the same comparison between them as between the finest Dress[e]d works of Mr [Capability] Brown and the Great works of nature. The Gout puts me in mind to stop and tell you that when I left Silbury Saturday night we had got down 56 feet: still made-ground, no more Horns this time. (my faithful Sarah was with me) I’m a saucy fellow to joke on such a subject, perhaps tis owing to my great security that I dare —- But when we came to the depth of 95 feet we struck upon a Thing which I am sanguine enough to hope will lead to a great Discovery. It was a perpendicular Cavity That as yet appears Bottomless it is just 6 inches over; we have followed it already about 20 feet[;] we can plumb it about Eleven Feet more” but as a great deal of loose chalk has unfortunately fallen in, at that Depth, is a stoppage but as at present a Strong wind comes up the Hole enough almost to blow out a Candle it must have some communication with the Air or some great Cavity somewhere, at first what the miners call a Damp or foul Air come out of it into the Shaft so that they could Hardly breath[e] nor would a Candle Burn, but that is over and now a strong wind comes up|[;] as it is in the very centre of this Great Hill and goes perpendicularly down it is matter of Astonishment The country people when we were at work at the Barrow said that on the Hill opposite on the other side the River Kennet, on a great Long Barrow set round with stones,’ the people at work there to get the stone had discover[e]d a Great Hole which they always had an opinion had a communication under Silbury Hill tho[ugh] on the other side of the river and a mile off. The Temple of Abury is no further. I wait © with impatience the event of 2 or 3 days more as by that Time we shall have come to where the stoppage is, and then I hope shall make a further Discovery. Something that is now perished must have remained in this hole to have kept this Cavity open as the ground is loose chalk Stones and visibly nothing to Support it, my Dear L[or]d I must once more give you leave to put on your night Cap which I believe this long scrawl will hasten. My wife who will not think her time ill bestowed in drawing this Druid Altar for you (called in our language a Kistvaen) begs to join in kind wishes with little Sly to all yours at Stratfield Say[e] I am my Dear Lord With great Truth sincerely Your very humble servant Edward Drax What emerges from these extraordinary letters is that Drax not only recognised Silbury Hill as lying within an ancient designed landscape, that he compared with the orchestrations of his highly acclaimed contemporary Lancelot Brown, but also SILBURY HILL: EDWARD DRAX AND THE EXCAVATIONS OF 1776 261 had an understanding of what is now regarded as archaeological stratigraphy and made notes and measurements in recording his discoveries. He furthermore reflected upon and entered into deductions related to his discoveries in the light of his knowledge of earlier records and his contemporary findings. Some foreground Having stated his intention to excavate one or two barrows in addition to Silbury Hill, in a letter to Sir John Pringle (1707-1782) some six months prior to the dig,”® Drax set the miners to work on a barrow that he evidently had identified earlier. In focusing on a barrow ‘of the oldest make’, and searching Silbury Hill for a central burial, his intentions are clearly signalled as relating to the very beginnings of this designed landscape. Drax furthermore described the selected barrow ‘on the Hakpen or Snakes Head’ as ‘most certainly belonging to and coeval with the Temple’ and he associated Silbury Hill as belonging to and part of the ‘Antient Druidical Temple of Abury’. In describing his plan in these terms an integrated vision and approach is illustrated, that has never hitherto been credited of Silbury’s earliest excavation. 7 Whilst the influence of Stukeley is evident, there is the additional possibility that Drax may have accessed the researches of John Aubrey (1626- 1697). Aubrey had left his manuscript of Monumenta Britannica with the publisher Awnsham Churchill (1658-1728), with whom it remained after the antiquary’s death. After Awnsham Churchill’s demise the manuscript had passed to Awnsham Churchill the younger, nephew of his namesake,* from whom the Revd. John Hutchins (1698-1773), sought permission to copy extracts.*! This seemingly took place shortly before Edward Drax took his father’s place as MP for Wareham, where Hutchins was the rector. The extracts made by Hutchins were subject to ongoing attention, and an enduring interest in Aubrey seems something that would surely register with the owners of the manuscript.” There was also a particular atmosphere of consciousness flowing from an intense interest in primary sources from the mid 1760s. This developed after Thomas Percy, chaplain to the Duke of Northumberland, published Reliques of Ancient English Poetry in 1765, which ignited interest in folklore and induced obsessive documentary searches.*? In 1770 a reminder of Aubrey’s unique position in relation to primary knowledge appeared in the very first volume of Archaeologia, while the antiquary’s contribution was further highlighted in a new edition of the enormously popular Camden’s Britannia that appeared in 1772.* The following year Awnsham Churchill died,* so questions perhaps arose surrounding the future of Aubrey’s manuscript, as the apparent inaccessibility of Monumenta Britannica after Aubrey’s death had a tendency to intensify the focus on the manuscript. This has traditionally been viewed from the perspective of antiquaries rather than the Churchill family, but one way or another it seems that Aubrey’s manuscript and its whereabouts may have come to the attention of the architect of the first major excavation of Silbury Hill, for on the 16" April 1762 Awnsham Churchiil was a witness at the marriage of his daughter Mary to Edward Drax.*° The Drax proposal The plan to ‘search the contents of Silbury Hill’ was unveiled in the first half of 1776, when Drax sought an introduction to Sir John Pringle, President of the Royal Society.*’ As a Fellow of the Royal Society and knowing Pringle well, the Duke of Northumberland could have facilitated this introduction if only by letter, but the introduction was sought through a Mr Castle, which suggests that the Duke was distanced from if not unconnected with the plan at that time.*® From his letters it appears Drax had opened barrows prior to 1776, possibly in his native Dorset. Furthermore he had likely encountered Silbury Hill on many occasions when travelling between his homes in Queens Parade, Bath, and Portman Square, London.” In the outline of the plan communicated to Pringle, Drax stated that he had contemplated excavating Silbury Hill for ‘some time’ as it ‘has never been thoroughly examined into’. He proposed ‘to sink a shaft of some 10 feet diameter in the Centre of the Plain in which it stands’, which notably differs from the centre of the summit, and estimated the cost of the excavation at ‘less than a £100’. Albeit Drax appears likely to have had access to the means to readily cover such a sum, a subscription had been suggested by people ‘of eminence both for their rank in life and taste and curiosity in works of this kind’.*° In the following century, the Devizes painter and historian James Waylen (1810-1894) suggested that this subscription had been originated by ‘several 262 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE SSS: Fig. 2 Mary, Sarah, and Edward Drax c.1774. From a private collection. gentlemen of Bath and Bristol’,*! which is evidenced as feasible by Drax’s Bath residence, family and circle.” It was perhaps one of the subscribers who suggested Drax approach the Royal Society, rather than the Society of Antiquaries. The President of the Royal Society, however, was recorded as viewing Drax’s proposal as ‘relating to matters of Antiquity solely’, that he ‘deemed more properly to the Antiquaries’.”” Although Pringle had seemingly followed protocol, the President of the Royal Society, who was also a Fellow of the Society of Antiquaries, appeared in person to put Drax’s case to a meeting of the Society of Antiquaries on 27 June 1776 chaired by Sir Joseph Ayloffe (c.1709-1781). In terms of archaeological excavation in this country the proposal was extraordinary, for nothing on the scale of Silbury Hill had been attempted in Britain before, but the Antiquaries ‘declined being any way concerned in the Adventure’.** The Antiquaries’ rejection has been interpreted as a matter of finance,* more besides is perhaps suggested by the recorded minutes stressing that the committee declined being in ‘any way’ involved. In April 1784 Edward Drax was elected a Fellow of the Society of Antiquaries.* This is perhaps what brought Drax into contact with James Douglas, who at that time was very active and a regular at meetings.’ In a letter dated 12 January 1787, sent to the antiquary Major Hayman Rooke, James Douglas mentioned the ‘research made a few years back into Silbury-Mount by Mr Drax and the late Duke of Northumberland: who employing miners from Cornwall for the purpose of exploring it, sunk into the very heart of it to its base without making any sepulchral discovery’.** Douglas did not mention in this letter what Drax had found, but it is interesting to note that three years on Rooke would publish his famous survey of the Duke of Portland’s ancient oaks.” Recalling the Drax post If not quite what he expected, Edward Drax’s ‘great discovery’ was simply that the great mound was raised over a ‘Druid oak’,° and the association of SILBURY HILL: EDWARD DRAX AND THE EXCAVATIONS OF 1776 263 ‘oak’ and ‘druid’ with Silbury’s vertical void was a conclusion steeped in national identity that was topically forthright. Memorable not least as the hiding place adopted by Charles II after the Battle of Worcester in 1651, the oak held a particular place in English identity and attracted meaning from navigation and naval superiority to independence and liberty,°! the theme centre stage in 1776. From the mid 1760s American communities adopted Liberty Trees and Liberty Poles that, as a feature of revolutionary identity during the War of Independence, were deployed in militia flags and highlighted in satire as being in competition with the motherland’s oak.** Americans tended to cast their Liberty Trees in the role of being old and anciently established, but this was an area of the propaganda war in which Britain could draw on precedence. In 1755 the Wiltshire rector William Cooke had proffered that druids had particular reasons for preferring mistletoe cut from oak and, in 1784, that other Wiltshire rector and future Archdeacon, William Coxe, was reporting a ‘withered trunk, the remains of a tree’ he had witnessed whilst travelling in Denmark and that those disposed to the ‘hypothesis that circles of stones as druidical relics, might, with a small degree of enthusiasm, have considered this very tree as the central oak...’.»? Between these times the oak was a focus of William Mason’s highly influential dramatic poem Caractacus (1756), which was based on the Roman assault on Anglesey and featured a ‘central oak’. Such was the symbolic potency, as Ronald Hutton has indicated, that upon Mason’s death Thomas Gisborne called for the author of Caractacus to be buried beneath a central oak.** The central oak of Caractacus had even made its debut centre stage at Covent Garden shortly after Drax completed his Silbury Hill excavations.» From the first Drax appears to have deduced that a pole or tree was evidenced at the centre of Silbury Hill, and his conclusion that it was a druid oak was plainly a received credibility for some time. Editing Camden’s Britannia in 1806, Richard Gough reported that ‘Major Drax digging perpendicularly through this hill [in] 1777 found only a rotten post and rusty knife’..° The finding of a more substantial pole at the centre of Silbury Hill was promoted in 1823, when Henry Browne published a booklet featuring Richard Hickley’s memories of 1776. Hickley was employed as bailiff to manage the Avebury manor estate by Arthur Jones, who gave permission to Drax to excavate Silbury Hill and barrows on his land. Having continued managing the estate for absentee hereditary owners after the death of Jones, any antiquary with a particular interest in barrows on manor land seems likely have come into contact with Hickley, who was regarded as ‘well acquainted with the property and every circumstances of the country’.°’ That eminent list would include John Merewether, the future figurehead of the 1849 Silbury Hill tunnel, whose father living at nearby Blacklands Hickley knew well enough to converse with.*® What Richard Hickley said to Browne, and quite possibly anyone that showed interest in barrows on the estate, was that he was present when the excavators of 1776 excavated a barrow and in Silbury Hill: found a piece of timber continued down the whole way, evidently for a centre from whence to take up the measurement of the hill in working it upwards.*? The suggestion that ‘a piece of timber continued down the whole way’ may indicate that on reflection Drax may have thought further traces had been encountered prior to the void being recognised. It seems more likely, however, that this was translated as a single pole or tree left 1m situ and not therefore permanently extended or supplemented above a certain height. This at least appears to be the conclusion of the Revd. Edward Duke, who was renowned not only for his researches and writings, but also for excavations famously recorded by Sir Richard Colt Hoare. Duke had a considerable presence in archaeological thought at this time, and of Silbury Hill he authoritatively stated in 1846: A slip of oak is produced, which, I have no doubt, was the ultimate remains of an upright log, placed as a centre, around which this aspiring mound was raised for I will not call it a barrow. If an objection be taken, that wood must have been utterly perished in an artificial mound, that was probably coeval with the Pyramids of Egypt, I will meet it by saying that I have seen the remains of wood in barrows, and that heart of oak, immured in chalk, is almost imperishable. Yet here, I believe it to have been the last remains of one entire log, and thus far a visible and substantial evidence of the vast antiquity of Silbury Hill, which, as it is not a barrow, so neither is a planetary temple, or a place of worship,...° It is such a leap from a slip of wood to an ‘entire log’ that Duke seems certain of this knowledge, to the extent he had concluded that there was no further question to address concerning the interior of the mound. 264 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Timber at the very centre of Silbury Hill proved a feature of the report made by Flinders Petrie, after he excavated the mound in 1922. This excavation proceeded with A. D. Passmore of WANHS present, in an unspecified capacity, alongside Petrie’s son and some labourers. Passmore added a short note to the famous Egyptologist’s published report, but the amateur’s knowledge perhaps deserves greater credit. In publishing a report on Silbury Hill two years earlier, Passmore mentions a central timber, providing a likely source for the suggestion in Petrie’s report that a cord stretched from a tree in the centre of Silbury was used to form level layers of construction.” Very much in the shadow of Silbury’s Hill’s more famous myths and legends, Drax’s central post perhaps had a surviving legend of its own. Responding to a request in 1967 for memories of the 1849 tunnel, which had intermittently re-opened through collapse between 1910 and 1923,° a local man, then aged 65, reported that his mother used to refer to Silbury as the Wooden Hill. Understandably this was at first thought to have been a dialect corruption of the neighbouring Waden Hill, but apparently not, as the interviewee said his mother had related that it was because Silbury Hill had been ‘erected over a wooden structure’. He further related that his mother was told a story when she was around eight years old (c.1880), that a couple went to bathe in the ‘swamp’ which was in the middle of the hill, that they let themselves down the mine shaft from the top by a ladder and were never seen again. Apparently a story to keep children off the hill, Drax’s excavation and his Druid oak were perhaps remembered in local myth.™ A clear path The devastating collapse discovered in the summit of Silbury Hill 29 May 2000 has raised questions associated with the state of backfilling within the vertical shaft excavated in 1776. It is then interesting that one of Drax’s immediate tasks was to create a path that would enable the transportation of equipment up the hill. As a path had existed for over a century, the new path was evidently different, probably on account of having to be less steep. The early recognition that shuttering would be necessary, to prevent the walls of the shaft from collapsing, demanded regular transportation of a large amount of timber to the top of the mound throughout the dig. The size and weight of this timber suggests the need for a more gradual path than that existing, and for the new route to have taken an entire circuit to reach the top seems not at all unlikely. We should note there would have been a need to keep this path clear and unblocked for the duration of the works. Not only would hoists and other mining equipment have to be returned to ground level, but the timber would have been valuable enough to be recovered when the excavation was complete. As spoil could not be cast over the lip where it would block the path, it suggests that the spoil from the shaft was stored on the summit. Mary and Edward Drax Whilst the opportunity was taken to place these informative letters in the public domain together with some preliminary observations as soon as possible, the letters are so abundantly rich that any fuller discussion of them is beyond the space afforded here. Future research might well add more. In a letter written on 5th February 1777, extending a ‘thousand’ thanks to the artist Coplestone Warre Bampfylde (1720-1791) for ‘one of the most beautiful landscapes’ she ‘ever saw’, Mary Drax expiained that her husband Edward had not written himself as his ‘Old Enemy’ the ‘Gout, has tied up his Hands’.® We are at once aware that Mary may have written letters that featured Silbury Hill as well as Edward, and we of course know that she produced drawings as did their daughter Sarah. There is then the possibility that more sources may have survived that could radically reshape future accounts. Acknowledgements Thanks are due to Guy de la Bédoyere, June Bennett, Christine Buckley, Joanna Corden, Jason Bowerman, Adrian James, Fachtna McAvoy, Mike Pitts, Zoe Stansell, the Drax family, the Avebury Archaeological and Historical Research Group, the British Library, the Library of the National Monuments and Record Centre, the Library and Archives of the Royal Society, the Library of the Society of Antiquaries, and the WANHS Library. I am grateful also to Dave Field, Kate Fielden, and Jim Leary for work on the transcription and their input on the draft of this paper. Particular thanks to Chris Drakes, without whose enthusiasm the letters would not have been found at this time. SILBURY HILL: EDWARD DRAX AND THE EXCAVATIONS OF 1776 265 Notes ] James Douglas, Nenia Britannica (1793) pp. 158-161. Dictionary of National Biography (1921-22 edition in all instances unless otherwise stated). See also Cecil Roberts, And so to Bath (London, Hodder and Stoughton, 1940) pp. 94-5. For references to the Duke’s scientific connections see Dave Field, Graham Brown, Bernard Thomason, The Investigation and Analytical Survey of Silbury Hill (English Heritage, 2002) p.15. British Fournal of Educational Studies 23 (1975) pp.4-6. Alexander Keiller Museum, Avebury: Atkinson Silbury Hill Archive - Richard Atkinson to Beales, 11 December 1967. John Merewether, “The Examination of Silbury Hill’ in Memoirs Illustrative of the History and Antiquities of Wiltshire and the City of Salisbury Communicated to the Annual Meeting of the Archaeological Institute of Great Britain and Ireland Held at Salisbury July 1849 (London, George Bell, 1851) pp.73-81. Alexander Keiller Museum, Avebury: Atkinson Silbury Hill Archive. The papers are not catalogued, but see: Beales to Atkinson 6 December 1967; Atkinson to Beales, 11 December 1967; Beales to Atkinson 20 April 1968; and Atkinson to Beales 25 April 1968. Alexander Keiller Museum, Avebury: Atkinson Silbury Hill Archive. Beales to Atkinson 20 April 1968. Atkinson to Beales 25 April 1968. Alexander Keiller Museum, Avebury: Atkinson Silbury Hill Archive. Atkinson to Beales 25 April 1968. Atkinson was also in receipt of a letter from Joan Evans making him aware of the reference to the 1776 excavation in her History of the Society of Antiquaries (1956) p.152, and as a consequence stated an intention to consult the relevant minutes. The author would like to thank and acknowledge Christine Buckley (Pers. Comm. Postcard 3 April 2002), June Bennett (Pers. Comm. Letter 11 April 2002), and Guy de la Bédoyére (Pers. Comm. Letter 7 April 2002) who all made contributions during the pursuit of information about Edward Drax. Particular thanks are due to Chris Drakes (Pers. Comm. telephone calls and e-mails 2002-2009), for his enthusiasm and support in addition to supplying Drax related references. Chris Drakes also runs a website (www. drakesfamily.org). John Burke, Burke’s Genealogical and Heraldic History of the Landed Gentry of Great Britain and Ireland (London 1847) p. 350. Dave Field, Graham Brown, Bernard Thomason, The Investigation and Analytical Survey of Silbury Hill (English Heritage, 2002) pp.15-16. Eton College Register 1698-1752 (Eton, 1927), pp.109- 110. Alumm Oxontenses 1715-1886 (Oxford, 1891) p. 387. Christine Buckley (Pers. Comm. Postcard 3 April 2002). Guy de la Bédoyére (Pers. Comm. Letter 7 April 2002). John Burke, Burke’s Genealogical and Heraldic History of the Landed Gentry of Great Britain and Ireland (London 1847) p. 350. John Nichols, Literary Anecdotes 10 11 12 13 of the Eighteenth Century (6 Vols), Vol. 1. London, 1812) pp.149-151. Edith Hobday, The Register of Sturminster Marshall, Dorset, 1563-1812 (1901) p.266. Dates for Mary Churchill taken from John Hutchins, History and Antiquities of the County of Dorset (4 vols) Vol. III (Westminster, John Bowyer Nichols, 1868) p.503. John Burke, Burke’s Genealogical and Heraldic History of the Landed Gentry of Great Britain and Ireland (London 1847) p. 350. Sir Lewis Namier, John Brooke, The History of Parliament: The House of Commons 1754- 1790 (London, History of Parliament Trust, 1964) p.340. John Nichols, Literary Anecdotes of the Eighteenth Century (6 vols), Vol. 1. (London, 1812) pp.149-151. Richard Gough (ed) Camden’s Britannia (4 vols), Vol. 1 (1806) p. 159. John Britton, Edward Brayley, The Beauties of England and Wales or Delineations Topographical, Historical and Descriptive of Each County Vol. IV (London, 1803) pp. 474-5. John Hutchins, History and Antiquities of the County of Dorset (4 vols) Vol. III (Westminster, John Bowyer Nichols, 1868) pp.497-508. John Bernard Burke, A Visitation of the Seats and Arms of The Noblemen and Gentlemen of Great Britain and Ireland Vol. 11 (London, Hurst & Blackett, 1853) p.166. Wiltshire and Swindon Archives 148/1: ‘Survey of the estates of Richard Erle Drax Grosvenor esq. [m. 1788, Sarah Frances Drax] by H. C. Wright; estates in the manor of Coate including Coate Manor Farm, Bishops Cannings; estates in the manor of Etchilhampton including Etchilhampton Farm, All Cannings; Bushyvowle Farm, Wootton Bassett; Barn Hill Farm in Clyffe Pypard and Broad Hinton; estates in the manor of Northcombe, Urchfont including Eastcott and Wedhampton Farms; Charlton Farm; estates in Tilshead; estates in Orcheston St George and Orcheston St Mary including Orcheston Farm; estates in Maddington including Maddington farm; 2 estates in Westbury including Bratton Farm’. See also Elizabeth Crittall, ‘Westbury’ in Elizabeth Crittall (ed), Victoria County History, A History of Wiltshire: Volume 8 (1965) p.161. W.R. Powell, PM. Tillott, ‘Bishop’s Cannings’, in R. B. Pugh and Elizabeth Crittall (eds), A History of the County of Wiltshire: Volume 7 (1953), pp. 187-197. Not available to Beales, but see also such as D.A. Crowley ‘Etchilhampton’, in Elizabeth Crittall (ed), Victoria County History, A History of Wiltshire: Volume 10 (1975), pp. 72-73. Jane Freeman, ‘Maddington’, ‘Orcheston St Mary’, ‘Tilshead’, in C.R.J. Currie (ed), A History of the County of Wiltshire: Volume 15 (1995), pp. 206-7, 231, 268. Rodney Legg, Dorset Families (Tiverton, Dorset Books, 2002) pp.28- 33: An Account of the Hospital for the Maintenance and Education of Exposed and Deserted Children (London, 1759) p.5. Sir Lewis Namier, John Brooke, The History of Parliament: The House of Commons 1754-1790 (London, History of Parliament Trust, 1964) pp.282-3, 340. British Library: Drax (Edward). Letters to Lord Rivers 266 14 15 16 17 18 19 20 21 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 1776. Add. 34813 ff. 110, 112. Working transcription by Brian Edwards, Dave Field, Kate Fielden, Jim Leary. Another example of Edward Drax’s writing (Drax (Edward). Letters to the Duke of Newcastle 1756. Add. 32868 ff. 350, 384.), catalogued within the British Library from an entirely different cache, was used by comparison to verify the same hand. The transcription was then arrived at as a working model for the purposes of ongoing research. Particular thanks and acknowledgements are due to Chris Drakes for his enthusiasm and promptings with Drax related references that led to the finding of these sources, and Zoe Stansell for kindly assisting with remote referencing. Queen’s Parade, Bath. Joint Postmaster Generals: Lord le Despencer [Sir Francis Dashwood (1708-1781)] Joint Postmaster General 1766-1781, and Mr Thynne [Henry Frederick Thynne, First Baron Carteret of Hawnes (1735-1826)] Joint Postmaster General 1770-1789. Vicary Gibbs (ed) The Complete Peerage of England Scotland Ireland and the United Kingdom (London, St Catherine Press) Vol. ITI (1913) pp.68-9, Vol. IV (1916) pp. 284-5. This was seemingly located in the area of Hyde Park and Hertford Street. The ‘liberal hint’ was perhaps Pitt suggesting that Drax should be left to get on with the work without interference. A possible alternative is that the hint could relate to the deployment of miners. It could of course be neither of the above. This suggests a start date of 24 October 1776. As Drax names long barrows elsewhere this is perhaps a reference to either two barrows together or a single barrow aligned by the ridge, possibly Barrow G30 or Barrow G31. See L.V. Grinsell, ‘Archaeological Gazetteer’ in R. B. Pugh and Elizabeth Crittall (eds), A History of the County of Wiltshire: Volume 1, Part I. (1957), pp.21-279, especially 208. Note also that James Douglas revealed in a letter to Colt Hoare that he held ‘three of the round rayed blue verditer beads of the highest period, which were once in the possession of Dr. Stukeley and mentioned in his Abury; from a barrow near the Hackpen hill, the head of the great Python,..’. See transcript of the letter from Douglas to Colt Hoare 25 June 1812 (Wiltshire and Swindon Archives: Colt Hoare Papers 383.907) in Ronald Jessup, Man of Many Talents: An Informal Biography of James Douglas 1758- 1819 (London and Chichester, Phillimore, 1975) pp. 271-4. The initial suggestion that there is something that looks like Hebrew on an artefact is perhaps a serious reflection by Drax, for it can be seen from his references to Patriarchs he believes Britain to have been inhabited by their direct descendants. The reference to that etched being something akin to his own writing, is presumably a joke. This barrow is evidently one of those adjacent to the Sanctuary and east of the Ridgeway. 22 23 24 25 26 Dy] 28 29 30 32 33 34 35 36 Presumably Sarah Frances Drax (1769-1822). The date is not known, but possibly 18 November 1776. This could be a reference to Stukeley, but equally relating to the burial being searched for within Silbury Hill. Given Drax’s humour it was perhaps both. This location appears to be the Devil’s Den. Drax is playfully writing to Pitt of course, and full well remembers his Scripture. This suggests an upright of some 150 mm diameter at the uppermost point with a height of 10.5 - 11m above the original ground surface with perhaps 1.5m below. This is very much a working estimate. West Kennett Long Barrow. The Minutes of the Society of Antiquaries June 27th 1776, Vol. XV pp. 2-4. The Minutes of the Society of Antiquaries June 27th 1776, Vol. XV pp. 2-4. John Nichols, Literary Anecdotes of the Eighteenth Century (6 vols) Vol. 1 (London, 1812) pp.149-151. John Britton, Memoir of John Aubrey (London 1845), pp. 87-9. William Long, Abury Illustrated, reprinted from WANHM IV (Devizes 1858). Michael Hunter, Fohn Aubrey and the Realm of Learning (1975) pp. 89-90. John Fowles, ‘John Aubrey and the genesis of Monumenta Britannica’, in John Fowles (ed) and Rodney Legg (an.), John Aubrey’s Monumenta Britannica, Vol. II (Sherborne, Dorset, 1982) pp. 613-4. Anthony Powell, Fohn Aubrey and his friends (1963) p. 288. Powell suggests that the younger Awnsham Churchill owned Aubrey’s manuscript in 1755 and 1780. John Britton, Memoir of John Aubrey (London, 1845) p. 88. Michael Hunter, John Aubrey and the Realm of Learning (New York, 1975) pp. 20-7. John Nichols, Literary Anecdotes of the Eighteenth Century (6 vols), Vol.1. (London, 1812) p. 150. Michael Hunter, John Aubrey and the Realm of Learning (New York 1975) pp. 205-6. Rosemary Sweet, Antiquaries: The Discovery of the Past in Eighteenth- Century Britain (Hambledon and London, 2004) p. 257. Howard Colvin, Essays in English Architectural History (New Haven and London, Yale, 1999) pp.214-6. Thomas Percy, Reliques of Ancient English Poetry (1765). Nick Groom, The Making of Percy’s Reliques (Oxford 1999) pp. 145-192. Michael Hunter, ohn Aubrey and the Realm of Learning (New York 1975) pp. 205-6. Date of Awnsham Churchill’s death cited in John Caswell, Lewis Arnold Dralle (eds) The Political Fournal of George Bubb Dodington (Oxford Clarendon Press, 1965) p. 449. Christine Buckley (Pers. Comm. postcard 3 April 2002). Guy de la Bédoyére (Pers. Comm. Letter 7 April 2002). John Burke, Burke’s Genealogical and Heraldic History of the Landed Gentry of Great Britain and Ireland (London 1847) p. 350. John Nichols, Literary Anecdotes of the Eighteenth century comprising Biographical Memoirs of William Bowyer and many of his learned friends (6 SILBURY HILL: EDWARD DRAX AND THE EXCAVATIONS OF 1776 a7 38 By) 40 41 42 vols), Vol. 1 (London 1812) pp. 150-1. Edith Hobday, The Register of Sturminster Marshall, Dorset, 1563-1812 (1901) p.266. The Minutes of the Society of Antiquaries June 27th 1776, Vol. XV pp. 2-4. It seems a possibility that Edward Drax knew the Duke of Northumberland. Edward’s eldest sister Elizabeth Drax (1720-1791), Lady Berkeley, who later married Robert Nugent, Viscount Clare, was a close friend of the Duchess of Northumberland. See James Greig (ed), The Diaries of a Duchess: Extracts from the Diaries of the First Duchess of Northumberland, Elizabeth Seymour Percy, 1716-1776 (London, Hodder and Stoughton, 1926). Another close friend and also a relative of the Duchess, Elizabeth Macie, was in owning a property in Queen’s Square, Bath, a near seasonal neighbour of Drax’s Queen’s Parade property. Elizabeth Macie was the Duke’s mistress and mother of the Duke’s son James Smithson (1765-1829). The Duke was also, with some probability apparently, suggested as being the natural father of Elizabeth Macie’s other child, Henry Louis Dickenson, who was born in August 1771. The Duke and Duchess were regular visitors to Bath with prolonged seasonal stays, evidently with the Duke making additional unaccompanied visits. Heather Ewing, The Lost World of Fames Smithson: Science, Revolution and the Birth of the Smithsonian (London, Bloomsbury, 2007) pp. 30-45. Somerset Record Office DD\GC/103 : Letters, Mary Drax of Bath to Coplestone Warre Bampfylde 5 February 1777. James Douglas, Nenia Britannica (1793) pp. 158-61. We might consider the possible influence of Bath architecture and John Wood on Drax’s proposal. See John Wood, A Description of Bath (London, W.Bathoe, T: Lownds, 1765). Ronald Hutton, Blood and Mistletoe: The History of the Druids in Britain (New Haven and London, Yale, 2009), pp.104-7. David Gadd, Georgian Summer: Bath in the Eighteenth Century (Adams and Dart, 1971). R.S. Neale, Bath A Social History 1680-1850: or A Valley of Pleasure, Yet a Sink of Inquity (London Routledge, 1981) p.165. The Minutes of the Society of Antiquaries June 27th 1776, Vol. XV pp. 2-4. James Waylen, A History Military and Municipal of the Town of Marlborough (Marlborough 1854), pp. 405-6. L. B. Namier ‘Charles Garth and His Connexions’, The English Historical Review, Vol. 54, No. 215 (July 1939), pp. 443-470; L. B. Namier ‘Charles Garth, Agent for South Carolina’: Part II (Continued). The English Historical Review, Vol. 54, No. 216 (October 1939), pp. 632-652. Heather Ewing, The Lost World of James Smithson: Science, Revolution and the Birth of the Smithsonian (London, Bloomsbury, 2007) pp. 30-45. Report of the inaugural meeting of the Bath Agricultural Society in Western Flying Post, or Sherborne and Yeovil Mercury, Monday 15 September 1777. The author extends thanks to Dr Norman Beale for his mutual collaboration in confirming this source. 43 44 45 46 47 48 49 50 il 34 53 267 The Minutes of the Society of Antiquaries June 27th 1776, Vol. XV pp. 2-4. The Minutes of the Society of Antiquaries June 27th 1776, Vol. XV pp. 2-4. Joan Evans, The History of the Society of Antiquaries (1956) p. 152. Edward Drax was elected a Fellow of the Society of Antiquaries on 1 April 1784. His name appears in membership lists in 1786 and 1787, but he subsequently resigned from the Society at some unknown date. The author is obliged to Adrian James, Assistant Librarian at the Society of Antiquaries, for the information regarding Drax’s election date and subsequent resignation. Ronald Jessup, Man of Many Talents: An Informal Biography of Fames Douglas 1758-1819 (London and Chichester, Phillimore, 1975) pp. 27-9. Society of Antiquaries Correspondence Files and Rooke Papers, Soc. Ants. MSS, transcript in Ronald Jessup, Man of Many Talents: An Informal Biography of ames Douglas 1758-1819 (London and Chichester, Phillimore, 1975) pp. 239-241. Hayman Rooke, Descriptions and Sketches of Some Remarkable Oaks in the Park at Welbeck in the County of Nottingham, a Seat of the Duke of Portland, to which are added observations on the age and durability of that tree and remarks on the annual growth of the acorn (London, J. Nichols, 1790). James Douglas, Nenia Britannica (1793) p.161. William Tighe, The Plants: A Poem (London, James Carpenter, 1808) pp. 69-140. David Hackett Fischer, Liberty and Freedom: A Visual History of America’s Founding Ideas (Oxford University Press, 2005) pp. 19-49. Amelia Rauser, ‘Death or Liberty: British Political Prints and the Struggle for Symbols in the American Revolution’, Oxford Art Fournal 21.2 (1998) pp. 151-171. Simon Peter Newman, Parades and Politics of the Street: Festive Culture in the Early American Public (University of Pennsylvania, 2000) pp. 25-6. Frank Paine, Diary of the American Revolution from Newspapers and Original Documents Vol.1. (New York and London, 1860) pp. 831-2. Benson J. Lossing, The Pictorial Field Book of the Revolution, Or, Illustrations by Pen and Pencil of the History, Biography, Scenery, Relics, and Traditions of the War of Independence (New York, Harper, 1860) pp. 233, 254, 466-7, 526. Richard Frothington, The Life and Times of Foseph Warren (Boston, Little Brown, 1865) pp. 37, 54-61, 120. William Cooke, An Enquiry into the Patriarchal and Druidical Religion, Temples, &c, being the substance of some letters to Sir Hildebrand Jacob Bart., Wherein the Primeval Institution and Universality of the Christian scheme is Manifested (London, Lockyer Davis, 1755) p. 50. William Coxe, Travels into Poland, Russia, Sweden and Denmark (3 vols), Vol. III. (London 1784) p. 424. See also George Richards, The Aboriginal Britons: A Poem (1791) p.20. 268 54 56 57 58 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE William Mason, Caractacus, A Dramatic Poem (London, J. Kanpton and R. and J. Dodsley, 1759) p. 64. Thomas Gisborne, Elegy to the Memory of the Revd. William Mason (London, 1797) p.9, cited in Ronald Hutton, Blood and Mistletoe: The History of the Druids in Britain (New Haven and London, Yale, 2009) p.117-8. George Watson (ed), The New Cambridge Bibliography of English Literature 1660-1800 (Cambridge, 1971) pp. 669-670. Richard Gough (ed) Camden’s Britannia (4 vols), Vol. 1 (1806) p. 159. Wiltshire and Swindon History Centre 271/3: Book of Memorandums proper to be preserved for whoever may possess this property. See also Wiltshire and Swindon Record Office 184/6 Letters 1791-1798 of Richard Hickley bailiff of the farm at Avebury to Mrs Adam Williamson in Jamaica being monthly descriptions of farm work and local news. See also Wiltshire and Swindon Record Office 184/5&7. Sir Thomas Bernard, ‘An Account of the Introduction of Straw Platt at Avebury’, in Sir Thomas Bernard (ed), Reports of the Society for Bettering the Conditions and Increasing the Comforts of the Poor (1805) pp. 90-111. John Merewether, ‘The Examination of Silbury Hill’ and ‘Diary of the examination of barrows and other earthworks in the neighbourhood of Silbury Hill and Avebury, Wilts, in July and August 1849’ in Memoirs Illustrative of the History and Antiquities of Wiltshire and the City of Salisbury Communicated to the Annual Meeting of the Archaeological Institute of Great Britain and Ireland Held at Salisbury fuly 1849 (London, George Bell, 1851) pp.73-81, 82-107. Wiltshire and Swindon Archives 727/3/35: Letter, Richard Hickley to the Revd. Mayo, Wimborne, 26 December 1817. Gentleman’s Magazine 59 60 61 62 63 64 65 XXXII (June 1850) p.562. Henry Browne, An Illustration of Stonehenge and Abury (Salisbury, Brodie and Dowding, 1823) pp. 38-9. See also, Brian Edwards, ‘A Missing Drawing and an Overlooked Text: Silbury Hill Archive Finds’, WANHM 95 (2002) pp. 88-92. Sir Richard Colt Hoare, Ancient Wiltshire Vol. 1 (1812) pp. 212, 221. Gentleman’s Magazine XCIV (April 1824) pp. 311-315. The Christian Remembrancer XII (July-December 1846) pp. 467-477. The Living Age XI (October-December 1846) pp. 159-60. fournal of the British Archaeological Association TX (1854) pp. 104-5. Edward Duke, The Druidical Temples of the County of Wilts (Salisbury, Brodie, 1846) pp. 40-2. A.D. Passmore, ‘Silbury Hill’, WANHM XLI (1920) pp. 185-6. W.M. Flinders Petrie, ‘Report of the Diggings in Silbury Hill, August 1922’, WANHM XLII (1924) pp. 215-8. Atkinson Archive, Alexander Keiller Museum, Avebury. Subsequent to adopting a date of 1915 for the horizontal tunnel becoming accessible Atkinson received a number of letters suggesting a much earlier date. Of these there seems no reason to doubt an account given by Bill Deacon, who had lived opposite the mound at Silbury Hill Cottages since at least the turn of the century and was ‘emphatic’ that he had discovered the tunnel during one of his regular rabbit hunts aged 16 in 1910. Atkinson Archive, Alexander Keiller Museum, Avebury: A.PA. Tucker, Wiltshire County Council, letter to Hester Atkinson 24 October 1967. Somerset Record Office DD\GC/103 : Letters, Mary Drax of Bath to Coplestone Warre Bampfylde 5 February 1777. Wiltshire Archaeological & Natural History Magazine, vol. 103 (2010), pp. 269-88 The relationship between Wansdyke and Bedwyn Dykes: a historiography by Ben Lennon For at least three hundred years the linear earthwork now known as the Bedwyn Dyke was held to be a part of the Wansdyke system. It was not until 1960 that the Wansdyke was finally established by Fox and Fox as terminating at New Buildings to the west of Savernake Forest. From this point forward the nature of the Bedwyn Dyke was divorced from studies relating the Wansdyke. Since this time the Bedwyn Dyke has been mentioned only in relation to the study of other features in the area and no critical reassessment has been carried out. This paper charts the historiography of the east end of Wansdyke to examine how the concept of an extended Wansdyke frontier became established and was perpetuated for so long. Introduction The Bedwyn Dyke is the commonly accepted name for the linear earthwork stretching from Chisbury hillfort, south east across the Bedwyn Valley to Round Copse, a distance of approximately 2.8km (Figure 1). The whole of this section is a Scheduled Ancient Monument described as an undated linear earthwork. The earthwork crosses from one side of the chalk valley to the other in fragmented sections _ with variable morphology. For the purposes of this _ paper this section of earthwork is referred to as | the Bedwyn Dyke. The earthwork is variable in _ dimensions but where most obviously intact has an approximate width of 17m with a bank of 2-3m in height lying generally to the south west (Figure 2). Anglo-Saxon Charters | The first historical record that appears to relate } specifically to the Bedwyn Dyke is an Anglo-Saxon | charter granted by King Cynewulf to Earl Bica in AD 778 (Sawyer 264; Sawyer 1968; also Kemble 133 and Birch 225).! The bounds of this charter have been deduced by Crawford and found to be roughly consistent with the old bounds of Little Bedwyn and Froxfield parishes (Crawford 1922). The bounds begin their perambulation to the north east of Littke Bedwyn working in a clockwise direction and progressing around the east side of Burridge Heath (Figure 3). The fifth boundary mark on the perambulation runs thus; Inde in longum praedicti septi in quoddam Vallu[m] in Harandene (trans: ‘thence by the aforementioned long hedge to a certain dyke at Harandene’). Harandene is the valley along which the Great Bedwyn-Shalbourne road now runs. The name is preserved in several place-names such as Harding Farm and Harding Copse at the head of the valley. An early 18th-century field name of Harding between Brook Street and Galley Lane lies at the bottom of the valley (WSA15/11/07).’ It can also be found on Christopher Saxton’s late 16th century map of Wiltshire as a settlement name. The vallum or dyke is generally accepted to be the Bedwyn Dyke (Crawford 1922, Grundy 1919), which is approached here from the east along the parish The Old Sweet Shop, 79 Primrose Hill, Lydney, Glos. GL15 5SW. benlennon@tiscali.co.uk 270 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 0 i 1 fe 25 a | a pili = [ 5 Kilometers Fig. 1 The area surrounding Bedwyn boundary between Shalbourne and Little Bedwyn (which is still a long hedge) and encounters the dyke at the aforementioned junction of the three parishes at Round Copse. The next part of the perambulation runs thus; Sicque per hoc Vallum pertingit ad ilum agellum qui dicitur Tatan Edisc, (trans: ‘and so by this dyke reach that field called Tata’s enclosure’). From this point the modern parish boundary between the Bedwyns runs north and north east along a ploughed out section of the dyke, onto the extent section at Foxbury Wood heading in the direction of Chisbury. The old parish boundary doglegs around the north end of the extant section of the Bedwyn Dyke along the aforementioned 13m wide bank and ditch until it meets the head of the spring.‘ From here it follows the spring along the northern edge of Round Copse until, at the edge of the woodland, it takes a sharp turn north around the edge of a field until it turns again to meet the Bedwyn Shalbourne road at Jockey Green (Figure 4). It is worthy of note that the vallum is given no name in this charter. Several other dykes are mentioned in this charter and all are equally anonymous. The next section of the perambulation runs thus; Et sic per occidentalem plagam ejusdem agelli jacet in ittos tumulos . . . torum deinde in Bedewindan, (trans: ‘and thus by the western edge of the same field where the mound (barrow?) lies. Then to Bedwyn’). The barrow is now lost but the inference is that it lay somewhere between Jockey Farm and Great Bedwyn, possibly in the areas between Galley Lane and Brook Street. The reference to Bedewindan is generally regarded as referring to the Bedwyn Brook rather than the settlement of Great Bedwyn which later arose at this point (Gover, Mawer and Stenton 1939, 332 Crawford 1922, Grundy 1919).° The second reference to the dyke appears in a charter of King Edgar to the Monastery of Abingdon, dated AD 968 (S756, K 1266, B 1213). This charter, in Old English rather than Latin, largely describes the bounds of Great Bedwyn parish as opposed to that of Little Bedwyn described in the previous charter. Where the two are contiguous they are remarkably consistent, indicating long-term stability in these estate boundaries (Figure 5). The first clause runs thus; 4rest fram Bedeuuindan to haran grafan norpeuueardan up et pere dices geate et haran dene. (trans: ‘First from the Bedwyn Brook to the north side of Haran Grafan at Dyke’s Gate at Harandene’). Crawford posits that dices geate refers to the point along the Great Bedwyn-Shalbourne road where the Bedwyn Dyke emerges from Round Copse and now visually terminates (Crawford 1922, 76). This eens WANSDYKE AND BEDWYN DYKES: A HISTORIOGRAPHY 271 a) Bedwyn Dyke north of the Pumping Station in Great Bedwyn (SU 2783 653). west ) Bedwyn Dyke south of the Pumping Station in Great Bedwyn (SU 283 652) west east c) Bedwyn Dyke at Jockey Copse (3U 290 644) west east d) Bedwyn Dyke at the southern end of Round Copse {5U 289 637). nor south Fig. 2 Profiles of the scheduled sections of the Bedwyn Dyke between Chisbury Hillfort and Round Copse. line has the virtue of being consistent with the more straightforward of the lines from Cynewulf’s charter. However, on closer inspection this is far from clear. Assuming the crossing place on the _ Bedwyn Brook is roughly consistent with that of the - previous charter then the line along the valley to the _ point where the Bedwyn Dyke emerges from Round _ Copse is topographically so obvious as to render the qualifying landmarks superfluous. This suggests that this is not the point of dices geate, but that the other landmarks provide the key to its location. That all of the landmarks lie within the Harandene valley is at least clear. Following the old parish boundary between the Bedwyns (in reverse this time) the reference to going to the north of Haran grafan could easily refer to the point at Jockey Green where the old parish boundary turns away from the road to edge along the northern boundary of Round Copse (Haran 272 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Horselget 3 , N S : (Chisbury) i 7 (Little Bedwyn) es De ee Coase ; COU, a \\ en ree, : fe ® : Abs : te Vd . “a (Great Bedwyn) ne) se ts Ao” <0" SS © ie LS Na eo Modern woodland boundary 3 1 Kilometers — Fig. 3 Map with and interpretation of Cynewulf’s charter of AD 778 (S 264). Chisbury hillfort. Extant sections of the Bedwyn Dyke” , oo » * ; » Little Bedwyn Parish Fatish boundayy . ra a / _ SS Burridge wa Heath Pos _ ockey. . : Copse ~ ‘ : Foxbury — Wood oe payeh poungary Ploughed out dotible bank 4 attributed to the’ Bedwyn Dyke Great Bedwyn Parish Shalbourne Parish Bedwyn Brail 0 500 1000 1500 Meters naa A as re se TNE os mS Fig. 4 Parish boundaries and features of the Bedwyn Dyke or WANSDYKE AND BEDWYN DYKES: A HISTORIOGRAPHY 273 .. re) "Me : te ee, Ung, an 4 (Great Bedwyn) oe ° S Oy S & Modern woodland boundary 1.5 ‘Kilometers Fig. 5 Map with an interpretation of Edgar’s charter of AD 968 (S 756). _ grafan perhaps translating as the Hare’s grove). This would bring us to the point on the north-east corner of Round Copse where there is a gap in the large ditch and bank sections of the Bedwyn Dyke. Dices _ geate offers no clues as to the nature of the geate. This | | could refer to a gap large or small. The ploughed out section of double bank described by Crawford (1953a, 121) is morphologically very different from the rest of the extant Bedwyn Dyke sections and could well have been added later to bridge a gap here. | This would then present a credible location for dices _ geate. Grundy also suggested this point to bea likely location for Dices geate (Grundy 1919, 274). The next clause offers a further clue; ford pon >’ | be uurtwalan per se haga ut cymé, be pam wyrtwalan to pedes pape (trans: ‘thence forth by the uurtwalan _ where the haga emerges and by the wyrtwalan to Paedes Path’). Crawford and Grundy both take an interpretation of uurtwalan as meaning ‘foot of the slope’ but wyrtwalan as ‘woodbank’ (Crawford 1922, _ 76; Grundy 1919, 284). It seems more probable that these are different renderings of the same word and _ finds a point in time in which wu is in simultaneous use as w. Rackham offers an alternative definition based on a literal interpretation of wyrt + walan as ‘plant-wall’, suggesting that these were linear features usually associated with woodland edges or possibly even the woodland edge itself (Rackham 1993, 82). If Rackham’s interpretation is correct then this would suggest a direction to follow the woodland edge (or its associated linear boundary feature) to where the haga emerges. A haga is generally interpreted as a linear earthwork created to form an enclosing boundary (Hooke 1998, 21). In this case it is clearly distinguished as a feature from the uurtwalan/wyrtwalan, suggesting perhaps that the latter was not bank and ditch but the woodland edge itself. This presents the issue of the nature of the haga which could either have been the scheduled line of the Bedwyn Dyke as it emerges from Round Copse or it could be the large bank on the parish boundary (if such a feature existed at this time). Either way the parish boundary is now continued along the woodland edge back down to the junction of the Shalbourne-Bedwyn road (south of Folly Farm) before heading south along what was the old Salisbury road in the direction of Botley Down. The deviation around Round Copse is a peculiar one when a very straightforward line from Jockey Green to the junction to the south of Folly Farm 274 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE presents a perfectly straight line along the valley bottom. Nevertheless the explicit directions in the charter suggests this more convoluted route. There are no other known charters from the Anglo-Saxon period that explicitly mention the dykes in this area. However, Wansdyke is specifically mentioned on a number of occasions within Wiltshire charters. In these documents it is invariably referred to as Wodnes dic and references are often accompanied by related names such as Wodnes beorg or Wodnes dene. References to Wodnes dic are common throughout Anglo-Saxon charters of Somerset and Wiltshire and are all of the 10th century. Three instances deal with the same section of the Wansdyke around Tan Hill (S368, AD 903; $647, AD 957; S685, AD 960). Further to the east the section of the Wansdyke that runs through West Woods is mentioned in two charters (S424, AD 933 and $449, AD 939). The distance between Tan Hill and West Woods is a little under 6km and the Wansdyke is a more or less a continuous feature between Morgan’s Hill and the eastern side of West Woods. Critically, despite the huge number of dykes and similar features that are described in charters to the east of Marlborough (including Berkshire and beyond) Wodnes dic is nowhere used to describe any other feature. In the 10th Century at least we must assume that the concept of the Wansdyke as an entity appears to be confined to the sections lying west of its termination at New Buildings. Later medieval references There are few, if any, explicit references to the section of the Bedwyn Dyke from the later medieval period. One of the clearest references to the east end of Wansdyke that comes from the medieval period is a perambulation of Savernake Forest thought to date from the 14th century (WSA 1300/65). The perambulation describes a forest that had reached its zenith and was subsequently experiencing a period of reduction and fragmentation. One section which mentions the Wansdyke explicity runs thus: Beginning at the western corner of Boneclyne and thus crossing the way which lies between the aforesaid Boneclyne and the wood of the Abbot of Hyde by a certain pit, recently made in the same place, and thus go down by the same way as far as Drayston. And from there as far as Crokerway and by the same way as far as Stumer to the new boundary pits and thus as far as Stokesgrene and thence as far as Redshard and thence go down by the Wodnesdich as far as the Oare Road N + [Savernake Forest} Wansaye| ver {New Buildings] [Savernake Great Park] \* — Wodedich Face NG [Mud a Wydmroche i Oar Route of we Perambulation ar Ss “~Boneclyne [Martinsell hill] 0 1 2 Kilometers Fig. 6 Map with and interpretation ef part of the perambulation of Savernake (14th Century) showing the relationship between Wodnesdich and Wodedich. and thus by the aforesaid road as far as the wood of Nicholas of Hampton which wood is (certainly?) in the King’s hands due to ancient waste. (WSA 1300/65). The section of Wansdyke mentioned here lies between West Woods and the Oare road (NGR SU 169 664). However, there is another feature that is mentioned in the charter and is referred to as Wodedich. Which [?] bounds begin at the head of the Wodedich at the eastern corner of the purpresture aforesaid of the Wydmroche and thus always descending by the Wodedich as far as __ western [eastern] corner of Boneclyne. (WSA 1300/65). Despite the close similarity in the names this appears not to be the Wansdyke but another prominent dyke feature that ran between Martinsell (Boneclyne) and the area around Hat Gate or Leigh Hill (Wydmroche)( Figure 6). Wodedich possibly derives its name from a woodland name Jwode mentioned in the same perambulation in this area. In its simplest form it may translate as ‘wood bank’. As with the Anglo- Saxon charters the Wansdyke does not appear to be mentioned by name again anywhere deeper within or to the east of Savernake Forest. One other medieval reference may hold a clue to the perceived end of the Wansdyke at this time. It comes from the roll of Henry Esturmy, sometime warden of Savernake and dated around 1333. . certain persons from Berwick and Sampson, Bilkemores carter, with dhur, Dagsone and Waghleg of Eastwick, felled 2 oaks in February of that year between La Houkslade and Dychesende, worth 6/- OO WANSDYKE AND BEDWYN DYKES: A HISTORIOGRAPHY 275 By, Wy yp WOOO) eee | fiyj ll eo UMMM), VA Fig. 7 Thomas Atkinson’s Plan of the Manor of Chisbury, 1719 showing a feature referred to as Wodens Ditch. and removed all the timber therefrom wither they would with 3 cartloads of the lop of the same, which went to Robert’s house at Wyke (WSA 1300/59C). La Houkslade is probably the valley now called Great Lodge Bottom, one end of which is now known as Braydon Hook. This shallow valley runs west out of the modern Savernake Forest petering out to the east of New Buildings; the end of the main run of the eastern Wansdyke. If this is Dychesende then it suggests that the Wansdyke was thought to end at _ this point in the medieval period. It is also unlikely that this area was cultivated prior to this time, being _ within the bounds of the forest, and the Wansdyke _ consciously terminated at this point. 18th- and 19th-century perceptions It was during the post-medieval period that the _Wansdyke became firmly linked with the Bedwyn Dyke. Whether association became entrenched by a verbal tradition, misplaced whimsy, or a desperate search for meaning in the linear earthworks that litter the Wessex landscape, remains open to debate. From the beginning of the 18th century the Bedwyn Dyke and those earthworks lying to the east were to be conceptually bound to the Wansdyke for the next 250 years. This link was reinforced in two ways; first through the recurrence on maps, and secondly by entrenchment of theories of antiquarian observers and commentators. The earliest evidence of the cartographic link comes from an estate map produced by Thomas Atkinson® in 1719 for Charles Lord Bruce, of Tottenham House. The plan for the manor of Chisbury, drawn in 1719, shows the area around Chisbury village, which lies to the southern extreme of the manorial bounds. The section of the Bedwyn Dyke which runs south from Chisbury hill fort is labelled ‘Wodens Ditch’ in a number of places (Figure 7). Clearly, at this time the Bedwyn Dyke was already regarded as an extension of the Wansdyke proper. Curiously, the plan does not show the earthwork connecting directly with Chisbury 276 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE WN \ Fig. 8 Hill-shaded LiDAR image of Chisbury hillfort and Chisbury Woods showing the feature described on Thomas Atkinson’s Plan dated 1719. Peter Crow, Forestry Commission. hillfort, but turning away westwards 110m south of the southernmost rampart and running parallel to it for a distance of about 275m before turning at right angles to the southwest and curving around to the west before disappearing into the dense woodland of Park Coppice. Atkinson was a fine surveyor. Many of the features that he recorded on his maps are still evident on the ground today. It is highly unlikely that a feature of this description did not actually exist if he recorded it. Despite this, no map or aerial photograph has since recorded such a deviation or any such feature across these fields. However, LiDAR survey carried out by the Forestry Commission in 2006 of Savernake Forest and surrounding woodlands shows that such a feature did, and still does, exist within the woodland area (Figure 8). The feature shown has now been ploughed out in the fields to the west of Chisbury Lane, but re-emerges as soon as it enters the woodland. It is considerably smaller than the scheduled sections of the Bedwyn Dyke with an overall width never exceeding 7m. The feature lacks the graceful curve suggested by Atkinson and instead displays a rather severe elbow turning the earthwork in a westerly direction. This feature is imperfectly suggested in the 1820 plan of Savernake Forest Estate (SWRO x6/4) in that it represents the woodland boundary at that time (Figure 9). Ironically, it is also shown as the woodland boundary on George Wyther’s map of North Standen dated 1612 (WSA 9/8/135/h). The ploughed out section across the field to Chisbury Lane is also partially shown as a field boundary (Figure 10). On neither the 1612 nor the 1820 maps is there any reference to these features having a name. It would appear that the woodland boundary shifted back and forth up to this feature before the woodland was extended eastwards to its present boundary around the middle of the 19th century. A second early mention of the Wansdyke appearing beyond Savernake can be found on a Common Awards Map in Berkshire and dated 1733 (Victoria County History of Berkshire Vol. IV, 1924, cited in Crawford 1953 (b), 257). The name used is Wans Dyke and it occurs some miles to the east of Bedwyn near Old Dyke Lane, 600m west of Inkpen church WANSDYKE AND BEDWYN DYKES: A HISTORIOGRAPHY 277 Ee. Fig. 9 Part of the Savernake Forest estate plan dated 1820 (WSA x6/4) showing Chisbury and the boundary of Park Coppice. (SU 351 637). Crawford later found the section of dyke referred to and named it Wansdyke so as to appear on the 1920s edition of the Ordnance Survey. The name has subsequently become fossilised on the modern 1:10 000 Ordnance Survey map. This link between the Wansdyke and the Bedwyn Dyke may have been provoked to some degree by the writings of John Aubrey, the 17th century antiquary and travel writer. Aubrey travelled on the Wansdyke and made certain vague observations regarding its structure and formation adding that ‘they say it runnes into Savernake Forest’ (unpub. cited in Hoare 1819, 17). Although Aubrey did not appear to have spent much time on the subject it was clear that ‘they’ already had a perception about the continuation of the Wansdyke beyond its apparent terminus at New Buildings. No doubt ‘they’ were the local inhabitants, innkeepers and local gentry that Aubrey would have encountered on his travels throughout Wiltshire. In other words it was already established as a perception of the local populace that the Wansdyke did continue into Savernake and beyond by the time of Aubrey’s sojourn in 1663. | 5 Dr William Stukeley was the next antiquary to speculate on the nature and extent of the Wansdyke and claims it for the Belgae, being the last and northernmost of their advancing frontiers (Stukeley 1776). However, Stukeley does not appear to make many claims as to its extent beyond Savernake. In 1791 the Rev. J. Collinson, historian of Somerset, Fig. 10 Part of George Wyther’s Map of North Standen dated 1612 (WSA 9/8/135/h) showing Chisbury and the boundary of Park Coppice. was much more explicit in his understanding of the extent of the Wansdyke both east and west: It commences at Andover, in Hampshire, and then passes nearly 1n a straight direction to Great Bedwin, in Wiltshire; it then crosses the Forest of Savernake, and the wild Marlborough Downs, where it appears in its pristine state, exceedingly deep, and flanked by a very lofty mound, after the manner of the elevated rampire of a castle. (Collinson 1791, 22, cited in Hoare 1819,19). Despite the growing perception that the Wansdyke did indeed extend beyond Savernake Forest this does not appear to have become widely accepted by map- makers until the early 19th century. Andrews and Dury’s map of 1773 shows the path of the Wansdyke from Morgan’s Hill crossing the Marlborough Downs and entering into West Woods. However, the gap then is considerable. No sections of the Bedwyn Dyke are claimed to be the Wansdyke and the next section that is marked thus is at Scots Poor on Chute Heath (SU 285 561), 10km directly south of Chisbury. This was in accordance with Collinson’s widely supported view regarding the course of the Wansdyke. Richard Colt Hoare was perhaps the first to challenge and critically examine earlier claims for the course of the Wansdyke. Hoare saw that the 278 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE claims for its eastern course were becoming more and more fanciful and sought to introduce a degree of objectivity. Commenting on Collinson’s observations he curtly notes: As the historian of Wiltshire, the whole of which county is traversed from east to west by WANSDYKE, it becomes me to correct the errors of former writers on this subject, and to enter more minutely than they have done into the detail of this interesting, and hitherto unexplained relict of antiquity. Many an arduous journey has been undertaken, and many successive days have been spent in the deliberate investigation of its course: the chase has been pursued with ardour, and with some degree of success, though, according to the sportsman’s natural wish, it has not terminated in death (Hoare 1819, 20). Having set forth his stal! Hoare went on to suggest that the scatter of disjointed banks and ditches that can be tentatively followed from Andover to Round Copse is a ‘branch of the Wansdyke: not a part of that grand rampart’ (Hoare 1819, 20). From Andover Hoare describes a northerly route running over Chute Downs, through Collingbourne Woods, ever north and; From thence it directed its course to Shalbourn Down, crossing the valley, and great road leading from Salisbury to Hungerford, penetrated Brail Wood; from whence issuing, it crossed the road leading from Great Bedwin to Shalbourn, and ascending the high ground at the corner of a wood [Round Copse] , united itself with the real WANSDYKE on Merril Heath.(Hoare 1819, 20) Hoare was very clear regarding his line of the Wansdyke. Working from the west he followed its line to Ivy’s Farm (to the west of New Buildings), and then on the recommendation of the Earl of Ailesbury (Aylesbury) he located an earthwork on the far side of Savernake emanating from Birch Copse and crossing the valley to Belmore Copse. He noted that this earthwork faced north as did the Wansdyke and he drew the logical conclusion. It is only a fragment, but it didn’t take Hoare long to push further and discover the earthwork shown on Atkinson’s map a century earlier and which he was still able to describe very clearly: ...the scent was recovered by the information of a very respectable tenant of Cheesebury farm, who from a long residence on the spot, had become perfectly acquainted with the dyke, and had frequently noticed its course. He told me that it formed the boundary between the parishes of Great and Little Bedwin, and that it directed its course through a wood to the east of Cheesebury Camp, named Park copse, from the lower corner of which it issued, and seems to have followed the line of the modern hedge, and to have skirted the declivity of the hill to the south of the camp; then crossing a deep lane, to have entered another small copse on the east side of it. Here it made a sharp angle to the south where its ridge 1s very visible and bold, and descended to the road which leads to Great Bedwin. (Hoare 1819, 30) This earthwork had brought Hoare to Chisbury hillfort. Following his inspection of the ramparts of the fort he concluded that the relationship between the hillfort and the earthwork was different to the relationships he had found with the other three hillforts that occur along the Wansdyke (Maes Knoll, Stantonbury and Bathampton). He saw these three as being raised at the same time as the Wansdyke and constructed as part of it.’ This separation clearly vexed Hoare who eventually concluded that Chisbury hillfort represented a far earlier construction of British origin and that the Wansdyke was formed around it (Hoare 1819, 31). From Chisbury Camp Colt Hoare went on to describe the route of the earthwork along Chisbury Lane, across the Bedwyn Valley and up to the top of Round Copse where he observed the sudden turn to the south west through Round Copse. He makes no mention of the now ploughed out section prior to entering Round Copse. After some deliberation Hoare then concluded that the section diverting through Round Copse was not the Wansdyke, but the branch alluded to earlier that he traced from Andover. Following more deliberation, agonizing, and a few false starts, Hoare eventually concluded that, although very fragmentary, the Wansdyke headed east from the corner of Round Copse across Merril Down, towards Bagshot Mill and arriving ‘very loftv and bold’ near Prosperous Farm. After a brief resurgence near Sadler’s Farm Hoare finally suspended his search and concluded that the Wansdyke probably headed for, and terminated at, Walbury hillfort some 2.5km south of Inkpen (Figure 11). Whatever the strengths and frailties of Hoare’s conclusions, his survey of the antiquities of Wiltshire, and in particular the Wansdyke, represents a huge achievement in what must have been very difficult and challenging circumstances (even for a well- heeled aristocrat of the early 19th century). Hoare saw the Wansdyke as an east-west boundary and would have no truck with the prospect of it heading south into Hampshire. He even abandoned the large section through Round Copse and relegated it to the —E——E—E =< Based on his own observations, particularly at the _and fifty years. His was the benchmark by which ' the Ordnance Survey One-Inch (‘Old Series’) for | Wiltshire, produced in 1817, incorporated Hoare’s | projection for the Wansdyke taking the line of the _Bedwyn Dyke (Figure 12). Once entered onto the Ordnance Survey record the name of Wansdyke became repeated on all subsequent editions to form a meaningful opinion on the east end of ' Wansdyke was General Augustus Pitt-Rivers as part -in 1892. During his investigations Pitt-Rivers lig EN J nt a tell Mill ‘made a number of observations and measured a TR 6 : large number of profiles of the various parts of east | Wansdyke. These included a section that Hoare had WANSDYKE AND BEDWYN DYKES: A HISTORIOGRAPHY Jif) J fs i; } ‘BERKSHIRE * : > %3, Lereenham Fark Ca <> i Bead 3 = VLE Fig. 11 Sir Richard Colt Hoare’s map showing the course of the Wansdyke from Savernake to Prosperous Farm. rank of a branch line in order to make his preferred Bedwyn Dyke below Chisbury hillfort. Pitt-Rivers route fit a continued eastward path. In the end even offered a new insight regarding missing sections of he had to conclude that he could make no case for the dyke such as those passing through Savernake it continuing much into Berkshire. However, he Forest: was the first to challenge Stukeley’s widely accepted notion that the Wansdyke was of Belgic construction. dilave eiseaiiione suepesica, tar ih pidces Were ihe Dyke passed through a forest, the earthen mound and ditch may have been replaced by an abattis of felled : a ; Bie junction with the Roman road on Morgan’s Hill he trees, no trace of which, of course, remains at the present steadfastly attributed the Wansdyke to the Saxons (Hoare 1819, 27). Hoare’s account of the Wansdyke set the accepted course and extent for the next hundred all others would be measured. The first edition of ¥ throughout the 19th century.® ¢ | ros A 3 The last major antiquary of the 19th century » telle : ae in of his Excavations in Bokerly and Wansdyke published \ Se investigated to the east of Savernake Forest (between Fig. 12 Part of the First Edition Ordnance Survey ‘Old Birch Copse and Belmore Copse) and a section of the Series’ of 1817 showing Chisbury and Wansdyke. 280 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE . @ Marlborough Buildings @ Burbage Chisbury 4 Kilometers Walbury hillfort @ Fig. 13 Some sections of earthwork claimed for the Wansdyke during the 18th and 19th Centuries. time. If the Roman Road from Marlborough to Bath were made at an earlier date than the Dyke, as now appears probable, nothing would be more likely than that, in places where it passed through a forest, an abattis should have been laid in front of the bank of the road, as a defence. But in places where the line of defence left the road, as on the tops of the hills, where no trees grew, a deeper ditch and bank would be necessary...It is, however, to be observed that the rampart diminishes 1n size, or 1s wanting, in places where forests may have existed, and that it increases in places where forests are unlikely to have grown. (Pitt- Rivers and Garson 1892, 246) After some deliberation and consideration of the evidence Pitt-Rivers concluded that it was likely that the Wansdyke was created during the Roman or post-Roman period (Pitt-Rivers and Garson 1892, 29-30). It is also clear from his accompanying maps that he considered that the Wansdyke did indeed extend beyond Savernake Forest but opted for the southern course from Chisbury, through Round Copse. He also showed the short sections that make up the course to Walbury but fell short of labelling them as the Wansdyke (Figure 13). The 20th century If the 18th and 19th centuries are to be regarded as the period in which the conceptual extent of the Wansdyke expanded, then the 20th century, in contrast, can be said to be that in which it paused for breath and gradually contracted. The Reverend C. S. Taylor, vicar of Banwell (Somerset) was the first scholar of the 20th century to critically examine the date of Wansdyke. Basing his work largely on the observations of preceding antiquaries, and using historical records such as Anglo-Saxon charters and the Anglo-Saxon Chronicle he sought to age the Wansdyke more accurately. By and large Taylor concurred with the previous assertion that the Bedwyn Dykes were part of the same system as Wansdyke and made many similar observations in the field. He found the ploughed out section between Birch Copse and Belmore Copse, but failed to find the place where the Chisbury Lane Section joined with Chisbury hillfort (Taylor 1904,136-7). Based largely on the historical record and observations regarding the establishment of parochial boundaries Taylor suggested a construction date for Wansdyke of the mid-7th century and linked it firmly to political conflict between the West Saxon and Mercians kings (Taylor 1904, 154-5). In 1923 HC Brentnall excavated a trench to the eastern terminus of the continuous section of the Wansdyke at New Buildings and claimed to have found evidence of a ditch, thereby ‘proving’ that the Wansdyke did indeed continue on towards Savernake Forest (Fox and Fox 1960, 16; Crawford 1953) (a) e119): In 1926 Major and Burrow published The Mystery of the Wansdyke. Although based on extensive fieldwork, Major and Burrow mistook a number of relatively modern features for sections of the Wansdyke and were heavily criticized for it by later writers. Although a popular book, well presented, and beautifully illustrated by Ed Burrow, it appeared eee WANSDYKE AND BEDWYN DYKES: A HISTORIOGRAPHY 281 Path of Wansdyke (Major and Burrow) Linear features Bank eeeeee Ditch Cd) Old tracks Lynchet Woodland to do little to advance the field of study. However, it does offer occasional insights regarding the nature of the Bedwyn Dyke and serves to highlight some of the common problems of trying to track long linear earthworks across the landscape. With this in mind it is worth pausing for a moment to consider some of Major and Burrow’s observations. Major had worked closely with the aforementioned H.C. Brentnall during the compilation of the book _and. they shared many of the same conclusions regarding the extension of the Wansdyke beyond _Savernake. Unfortunately, in charting the course 'of the Wansdyke, Major and Burrow paid scant regard to the size or form of many of the features that they encountered. The line that they concocted through Savernake and on to Chisbury consisted of a miscellany of earthwork features of varying form and origin. In order to demonstrate this point | Figure 14 shows a section of the line of Major and | Burrow’s Wansdyke as it passes through a part of | Savernake. The illustration also shows the type | of features that have been identified initially from 500 _ Fig. 14 Sections of earthwork in Savernake Forest claimed for the Wansdyke by Major and Burrow in 1926. Features shown here were initially identified using LiDAR followed by ground survey. LiDAR survey. Within this short section Major and Burrow incorporate a part of the Earl of Hertford’s inclosure (from the late 16th century), a portion of the old Bedwyn to Marlborough road, coppice banks running through Birch Copse and Belmore Copse, and a large bank of unknown origin running across the intervening field. Much of the rest of Major and Burrow’s Wansdyke between New buildings and Chisbury is similarly dubious. To the east of Chisbury Major and Burrow made the pertinent observation that the Wansdyke appears to veer rather abruptly to the south. Between Foxbury Wood and Round Copse they identified a division in the course of the dyke with the easternmost branch cut though by that of the southern branch (which leads to Round Copse). On this basis they deduced that the southern branch through Round Copse was the later of the two and presumed an easterly course toward Baverstock Copse, Prosperous House and on to Inkpen (Major and Burrow 1926, 114). The point at which this junction occurs is precisely that at which Crawford 282 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE. Fig. 15 Hill-shaded LiDAR image of a section of earthwork in the north of Chisbury Wood tenuously considered to be part of the Wansdyke by OGS Crawford. Peter Crow, Forestry Commission. described a double ditch 12 feet wide.’ Given its very different structure this is possibly a distinct feature, which would render Major and Burrow’s observations completely redundant. Despite the spurious quality of some of their own deliberations Major and Burrow were sceptical regarding Colt Hoare’s southern branch of the Wansdyke running south from Round Copse, although they did dutifully follow its line. Given the shortcomings of Major and Burrow’s work they did eventually make the critical conclusion that the entire system as they conceived it was a composite work, possibly formed over a number of different periods (Major and Burrow 1926,135). Perhaps more pertinent was the work of O.G.S. Crawford, one time archaeological adviser with the Ordnance Survey, progenitor of Antiquity, and pioneer of the use of aerial photography in landscape archaeology. Fox and Fox summarized the archaeological consensus with a cutting opening line from their 1960 critique of the Wansdyke by saying: A four page appendix by the late O.G.S. Crawford in his Archaeology in the Field in 1953 contributed more to its understanding than the 140 pages of Albany Major’s obscurities. (Fox and Fox 1960,1) Crawford made numerous field observations over a long period of time and was responsible for assigning the appellation of Wansdyke to a number of earthworks on the 1920s OS maps that had previously not been named. He was later repentant of his haste in the light of growing knowledge and experience (Crawford 1953 (b), 257). Crawford’s observations largely concurred with previous works with some modifications. He tenuously attributed a new section in the northern part of Chisbury Wood (Park Coppice) but overlooked the previous section in Park Coppice (from Atkinson’s plan of 1719), a mere 500m to the south east (Figure 15). He also identified a morphological distinction in the section between Jockey Copse and Round Copse, and considered the possibility of a medieval adaption for the purposes of enclosure (Crawford 1953 (a), 121). He, too, puzzled over the Round Se WANSDYKE AND BEDWYN DYKES: A HISTORIOGRAPHY 283 0 1 2 3 4 Kilometers @ Marlborough | E + ane “8 Chisbury ee ; ttre New econ "gr tf @ Westcott Po Bedwyn Buildings Park — Copse fj ies * Prosperous Coppice o i House L. Great ®@ “8 Inkpen Bedwyn H e 7 Daniel's A _¢ Bedwyn Copse rae Old Dyke _ oe Br ails Lane ake a Walbury urbage hillfort @ Fig. 16 Sections of earthwork considered a part of the Wansdyke by OGS Crawford. Copse section and felt unsatisfied regarding the apparent jumble of banks and ditches in this area. Crawford identified short sections at Westcott Copse (SU 318 646), Daniels’s Lane (SU 322 644), and Old Dyke Lane (SU 351 636) (Crawford 1953 (a), 122- 4), which he assumed formed part of the Wansdyke (Figure 16). He also had his suspicions regarding a section of woodbank at the head of Cheval Bottom in Savernake Forest which appears as a large double _ bank and single ditch (SU 229 660).'° However, he let this pass and did not attribute it to Wansdyke (Crawford 1953 (b) 257). In the same year (and in the same journal) Burne _ considered the previous evidence but did little to _ advance the study of the east end of Wansdyke. _ He even went as far as to suggest three potential endings for the Wansdyke; at the Windmill, Wilton (SU 275 615), Great Botley Copse (SU 295 600), and Inkpen Beacon (SU 356 625). By his own admission his reasoning was fanciful, suggesting a succession of surveyors in a politically changing climate (Burne 1953, 133-4). If nothing else Burne’s paper inadvertently points out a critical fact in the study of large cross-country linear earthworks; it 1s a more attractive prospect to join the dots to create a single frontier than it is to consider the fragments as individual entities. In 1960 Fox and Fox published the most in depth and critical analysis so far produced. With a wealth of knowledge gained through the study _ of linear earthworks in East Anglia and Wales, Fox and Fox spent a large part of the late 1950s embroiled in fieldwork on the Wansdyke. The work largely succeeded in undermining many of the myths that had grown up around the Wansdyke over the intervening centuries. Among the litany of revelations was the relationship between East and West Wansdyke and the intervening Roman road. Brentnall’s claim for the excavated ditch beyond New Buildings was repudiated (Fox and Fox 1960, 16). More critically for the purposes of this study Fox and Fox are clear in their conviction that the Wansdyke, as a conceived earthwork, ends at New Buildings. This is particularly evident in the way that they go on to give a separate, but no less detailed, account of the earthworks between Savernake and Round Copse, at the same time coining the name Bedwyn Dyke. The section that is now scheduled is given the most comprehensive consideration and its description is largely recognizable as the features we see today. However, Fox and Fox were clear that, as a stand alone defensive earthwork, the Bedwyn Dyke did include the section through Round Copse but was unlikely to have extended eastward beyond this point as had been previously supposed (Fox and Fox 1960,18-20). They also recognised one of the most important defining characteristics of the Bedwyn Dyke as being a cross valley dyke, stretching across the valley chalk and possibly petering out on heavier clay soil. They drew attention to comparisons with other cognate linear earthworks such as Rowe Ditch, Hereford (Fox and Fox 1960, 20). They concurred with Crawford on the interpretation of Cynewulf’s charter of AD 778 that Dices geat lay on the Shalbourn-Great Bedwyn road and drew attention to a diagonal chalk mark, 55ft wide, and 284 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE |. extending across the road to the south.'! They also suggested that the distinctive south west curve at Round Copse indicated a local defence work curving around Great Bedwyn and its hinterland (Fox and Fox 1960, 19). Attention was drawn to the links between the later use of Chisbury as a likely stronghold of Cissa, a Wessex sub-king of the area in the late 7th or early 8th centuries. Cissa was identified from the Chronicle of Abingdon Abbey by Stenton and his name resides in Chisbury (Fox and Fox 1960, 20).!? On this basis Fox and Fox concluded that the Bedwyn Dyke system probably had an Anglo-Saxon origin but was certainly distinct from, and independent of, the Wansdyke or any other local features. With regard to the Bedwyn Dyke they concluded that: ‘Tt therefore seems probable to us that the Bedwyn dyke was constructed to define and defend, if need be, the north-eastern part of the original Bedwyn settlement or estate. The Little Bedwyn charter shows that this was in being by the end of the 8th century.’ (Fox and Fox 1960, 20) From this we can infer that not only is the Bedwyn Dyke system distinct from the Wansdyke, but that Fox and Fox regarded the fragments to the east of Round Copse as separate from those close to Chisbury, which they suggested relate specifically to a historical Bedwyn estate. It is from this point forward that we can begin to view the Bedwyn Dyke as a distinct landscape feature in its own right, re-emerging from its eclipse by the Wansdyke. Yet, despite the insight of A. and C. Fox, some scholars continued to consider all of the previous earthworks lying between Savernake Forest and Inkpen as a single system. In the 1980s a series of excavations at the Roman villa site in Castle Copse (within Bedwyn Brail) were conducted by Indiana University. The resultant publication by Hostetter and Howe was a wide-ranging account of the Bedwyn area from the Iron Age through to the Saxon period including a section devoted to the Bedwyn Dyke system. Hostetter and Howe defaulted to a consideration of all sections east of Savernake Forest. Despite the book’s comprehensive section on the geology of the area Hostetter and Howe made an initial observation that ‘virtually all sections run on clayey soils which have eroded’ (Hostetter and Howe 1997, 359). This was a major error as the section between Chisbury and Round Copse runs almost entirely on chalk soils. All the sections observed by Crawford and Colt Hoare were re-examined but other sections within Bedwyn Brail were also considered as potential candidates for the system. Despite similarities the conclusion was that none of the features within Bedwyn Brail could be attributed to the Bedwyn Dyke system and are more likely to be associated with Celtic or Roman field terraces, post-medieval field boundaries or game park enclosures. They also noted that no southern extensions to the system had been observed (Hostetter and Howe 1997, 361). Crawford’s observation of the double bank between Jockey Copse and Round Copse was accepted uncritically without considering morphological discrepancies or the possibility of later adaptations. Among the more astute observations was that the dyke system always strives to keep the ‘outside’ downhill and avoids running along low features. To achieve this view the ‘inside’ must be regarded as that area enclosing Chisbury hillfort, the villa at Castle Copse or the village of Great Bedwyn (Hostetter and Howe 1997, 363). At the level of the wider landscape attention is drawn to the strategic importance of this area as an east-west trending upland between the Thames basin and the Vale of Pewsey, and potentially a major junction of several communication routes. The several sections are noted to generally block presumed lines of communication through the Vale of Ham and the Bedwyn Valleys. However, they conclude that the whole system does not constitute a convincing military structure and would at best only hinder small scale raiding (Hostetter and Howe 1997, 364). In terms of dating the origin of the system a range of possibilities were entertained from the Iron Age to the post-Roman period with a conclusion that the latter seems most likely, possibly late 4th or 5th centuries (Hostetter and Howe 1997, 365). 21st-Century Research In 2006 Andrew Reynolds and Alex Langlands rejuvenated the notion of an extended Wansdyke frontier running from the Bristol channel as far as Inkpen Beacon. Drawing parallels with the apparently deliberate discontinuity of other cognate earthworks such as Offa’s Dyke they implied that as an entity the Wansdyke frontier represented a political construct conceived on a grand scale. They went on to argue that such a large undertaking could not have been executed during the politically disparate period following the withdrawal of the Roman garrisons, but is much more likely to have been associated with the formation of Wessex WANSDYKE AND BEDWYN DYKES: A HISTORIOGRAPHY 285 Baggan Gaon (Timbridge Farm) . gete Hadfeld~ os "(Belmore Copse) geate _ (Birch Copse) Section of dyke on the parish boundary observed by Crawford and. later dismissed. ff : / (Horsehall Hill) i (Chisbury / Lane Farm) / / Horselget rames dene ~~~. g -_ ee i (Chisbury) ‘Section of lynchet observed by as geate of de = Hoare, Crawford and others. *. . - : . (Park Coppice) (Bedwyn — y _ Section of the parish boundary bounded by a bank and ditch. a) 1 =e a é 2. Kilometers [ Fig. 17 Map with an interpretation of the line of gates (geat) in Anglo-Saxon charters along the eastern edge of Savernake Forest (S264 and S756). (Reynolds and Langlands 2006). A novel observation in this work is the regular occurrence of geat features in the Anglo-Saxon charters, particularly those previously mentioned Sawyer Nos. 756, 688 and 264).!° Taking a cue from Peter Fowler’s work on the East Wansdyke in West Woods (Fowler 2001) Reynolds and Langlands observed nine references to gates lying close to the projected course of the Wansdyke through the Savernake area: “Building on Crawford’s 1942 map of the features in these boundary clauses there are nine references to gates, all in close proximity to the dyke as plotted by Colt Hoare and Major and Burrow. Crawford suggests that these ‘gates’ represent breaks in hedgerows, yet 1t 1s likely that they are related to more substantial boundaries. Straet Gate, located in the south-east corner of the Great Bedwyn estate, for example, is the point where the Roman road from Cunetio (Mildenhall) to Venta Belgarum (Winchester) passes through a possible stretch of dyke, identified by Major and Burrow.” (Reynolds and Langlands 2006, 21) This is a keen observation that requires a greater degree of scrutiny. The geats referred to are interspersed among the three charters aforementioned. Four features appear in quick succession along the eastern side of Savernake Forest roughly co-existent with the modern parish boundary of Little Bedwyn. In the Cynewulf charter of AD 778 (S264), which we followed earlier as far as the Bedwyn Brook, the perambulation passes over the brook and continues thus: ...et sic in longum illius spineti in horselget . et continuo... . rames dene geate . et extenso tramite ejusdem septi . to holhrygc gete . et eodem septo to hadfeld geate . et eodem septo to baggan gete . Et sic in tlum septum to Bradan Leage... [trans: ...and thus by the long thicket at Horsel Gate. And continue..... to Rames Dene Gate. And continue [along the] track [by the] same hedge to Holhrygc Gate. And by the same hedge to Hadfield Gate. And by the same hedge to Bagga’s Gate. And so by that enclosure to Bradan leage...] In the charter of Edgar (S756) some two hundred years later this line is approached from the opposite direction but still yields the same line as follows: ... bonne ut to bradan leage per cudhardes peo ut ligo, ponne on baggan geat, ponne on headfeld geat, bonne on pone hledredan beam, ponne on hrammes dene geat, bonne on hors heal get ut on beocces heal, panon to Bedewindan. [trans:... Then out to the Bradan leage where Cutharde’s path emerges. Then on to Bagga’s Gate. Then on to Headfeld Gate. Then to hledredan tree. Then to Hrammes Dene Gate. Then to Horse heal Gate. Out to the Beocces heal. Then to the Bedwyn Brook. Despite a couple of interjected points on the later charter the line is identical. Although little firm evidence of these points now remains, place name and topographical evidence suggests their location. One valley is mentioned explicitly (rames dene/ hrammes dene). Baggan geat can probably be identified with the valley now occupied by Timbridge Farm (Figure 17). At the head of this valley lay the medieval lodge called Bagden Lodge (now Savernake Lodge SU 233 337). Called Baggeden in 1290 (Gover et al., 1939, 353), the place-name suggests a valley based on this area. Horselgeat is aname that persists in the modern form of Horsehall Hill (Gover et al., 1939, 335). The horselget was probably on the parish boundary near to Chisbury Lane Farm (SU 266 661). This would suggest that rames dene/hrammes dene would be the shallow valley between this point and Sicily Clump, and holhrygc gete (replaced later by hledredan beam) would be on the broad ridge (as the name rygc suggests) where the parish boundary crosses the road on Bedwyn Common (Sicily Clump SU 255 658). This would place hadfeld geate/ headfeld geat somewhere in the region of Owl’s Castle in Birch Copse (SU 243 667)."4 This line of gate features aligns very well with two of the earthworks attributed to the Wansdyke between Savernake Forest and Chisbury hillfort. The first is the lynchet lying between Birch Copse and Belmore Copse. The second is the earthwork found by Crawford, and later dismissed, lying 250m to the southwest of Chisbury Lane Farm on the parish boundary. Although very faint it is vaguely visible on the LiDAR survey carried out by the Forestry Commission in 2006. So, potentially there are a series of earthwork features that align with a projected route of the Wansdyke. However, the language of the charters does not support this as being the major earthwork as suggested by Reynolds and Langlands. It is not helpful that the Cynewulf 286 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE | charter of Little Bedwyn (S264) is in Latin and the later charter of Eadgar (S756) is in Old English. This makes a comparison of the features identified difficult. Despite this, some germane observations can be made. The term used in the earlier Latin charter for the long feature on which the gates are situated is described as a septum. As was mentioned in relation to the section at Round Copse this term appears to be used as a word for a hedge. This is to be distinguished from vallum that is used to describe a larger feature more akin to a dyke. The septum appears to run continuously for a distance of 6km from Bedwyn to the northern part of Savernake where a vallum is encountered. In the later of the charters the hedge is not mentioned at all, perhaps suggesting that the route was so well-defined no feature was required. : It is in this same charter that Dices geat is encountered (see above). Returning briefly to this section it will be noted from the passage below that there are several linear features mentioned: 4Erest fram Bedeuuindan to haran grafan norpeuueardan up et pere dices geate et haran dene, ford pon’ be uurtwalan per se haga ut cymd, be pam wyrtwalan to pedes pape, ponone wid helmes treowes, bonne on embrihtes get, ponne wid stet gcetes, bonne on hundan dene neopewerde’ [trans: First from the Bedwyn Brook to the north side of Haran Grafan at Dyke’s Gate at Harandene. thence forth by the uwurtwalan where the haga emerges and by the wyrtwalan to Paedes Path. Then towards Helmes Treowes. Then to Aembrihte’s Gate. Then towards Staet Gate [probably a misspelling of straet gaete]. From this passage it does not seem that a large linear feature is in any way singularly connected with any of these gates other than dices geate. Moreover, it suggests that dices geate is singled out as being characterized by its relationship to a dyke, whereas the others are not indicated in the same way. Fowler’s work on the charters of Oare (S242, AD 933) and East Overton (S449, AD 939) were based on specific references to gates along Wansdyke ( Wodnes dic)(Fowler 2001). The references in the charters relating to Savernake do not seem to portray a large linear feature in the same way. Although the high numbers of references to gates in the Savernake area does seem curious there appears to be little other evidence currently available to support the theory that this is a continuation of the Wansdyke. WANSDYKE AND BEDWYN DYKES: A HISTORIOGRAPHY Conclusion The historiography of the Bedwyn Dyke in relation to the Wansdyke is complex and is characterised by a desire to create monumental frontiers by linking disjointed fragments of earthwork together. There is little evidence to suggest that there was any conception of the Wansdyke extending beyond Savernake before the post-medieval period. However, once the idea of monumental earthworks of antiquity took hold, it continued to fire the imagination of antiquaries for hundreds of years. Following on from the work of Fox and Fox in conceptually severing the Bedwyn Dyke from the Wansdyke, scholars such as Hostetter have still tended to consider the remaining scattered fragments of earthwork features between Savernake and Inkpen as being part of a singly conceived “system” rather than a series of independent features. Despite the conclusions of Fox and Fox in 1960 the idea of the extended Wansdyke frontier continues to be revisited in different forms. Endnotes 1 All Anglo-Saxon charters will be referred to be the reference given in SAWYER, PH., 1968, Anglo-Saxon Charters: an annotated list and bibliography. London: Royal Historical Society. For cross-referencing purposes numbers allocated by Birch and Kemble are also provided. Henceforth these numbers will be prefixed by S, B or K respectively. See References for full citation. 2 Harandene is though to be derived from OE har + denu meaning ‘boundary valley’ (Gover, Mawer and Stenton 1939, 333). However, it is not clear whether the boundary in question was that of the dyke or that of the charter (or both). 3 Crawford and Grundy both refer to all of the Bedwyn - Dyke system as part of the Wansdyke. 4 The parish boundary appears to have been changed during the 1980s. 5 Ekwall claims a back-formation with the stream name coming from that of the village or even a woodland name (Ekwall 1928, 30). 6 The Atkinson maps are held in the Wiltshire and Swindon Archive where they are identified as being the work of Charles Price. However, a number are signed by Thomas Atkinson. The relationship between Atkinson and Price is not clear although there are many stylistic similarities in the work. 7 Inhis work looking at the hillforts of Stantonbury and 10 11 12 13 14 287 Maes Knoll Ian Burrow concluded that these Iron Age hill forts had been incorporated into the Wansdyke as an expression of territoriality, but were probably never manned or the Wansdyke patrolled. The evidence for this was based on a lack of a constructed walkway between the hillfort and the Wansdyke itself (Burrow 1981). Interestingly, Greenwood’s map of 1820 still only shows its extent as far as West Woods and no further. This was still in existence in 1953 when Crawford described it as ‘just before it meets the parish boundary it is still unploughed and under grass; it no longer consists of a single bank and ditch but is a broad ditch, twelve feet wide between two banks, with an overall width of nearly seventy feet’? (Crawford 1953,121). The feature is shown on a Savernake estate plan of 1820 (WSA x6/4) on which it can clearly be seen as a hedged field boundary for its whole length between Jockey Copse and Round Copse. Crawford erroneously described this woodbank as part of the enclosing bank of the Bailiwck of Bagden. This has more recently been identified as a woodbank established by the Earl of Hertford from 1594 (Ramsay and Bathe 2008). However, the section observed by Crawford remains in doubt (see below). It should be noted that this crop mark does not join on to the end of the Round Copse section but seems to flow from the road junction two hundred metres to the south east. It is certainly shown as a farm track on the 1:10560 OS 1880s Ist Ed. Gover et al. have drawn attention to the fact that in the medieval history of Abingdon Abbey there is a story of a West Saxon noble named Cyssa who was regulus in a district which included Wiltshire and the greater part of Berkshire. His dominium included the Episcopal see of Malmesbury but the metropolis of his kingdom was Beduinde, and in the southern part of the that urbs he built a stronghlold which was named Cysseburi. Apparently the story does not belong to the older stratum of Abingdon tradition and has no historical value. However, this does not completely undermine the case for a possessor of the area by the name of Cissa (Gover et al. 1939, 334-5). S756: A.D. 958 or 968. King Edgar to Abingdon Abbey; grant of 72 hides (cassati) at Bedwyn, Wilts. Latin with English bounds. S688: AD 961. King Edgar to Abingdon Abbey; grant of 20 hides (cassati) at Burbage, Wilts. Latin with English bounds. S264: AD 778. Cynewulf, king of the Saxons, to Bica, comes and minister; grant of 13 hides (manentes) at Little Bedwyn, Wilts. Latin with bounds. (Sawyer 1968). Crawford differed in his interpretation of location of the gates, placing, for instance, hrammes dene as the valley running between Belmore Copse and Birch Copse(Crawford 1922, 295). These differences matter little as the line of the perambulation identified is roughly consistent. 288 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE References BIRCH, W. DE GRAY (ED.), 1885-93, Cartularium Saxonicum. 3 vols. London BURNE, A.H., 1953, Wansdyke West and South, WAM 55, pp.126-34 BURROW, I., 1981, ‘Hill-forts after the Iron Age: the relevance of surface fieldwork’ in G. Guilbert (ed.) Hill- fort Studies. Leicester: Leicester University Press COLLINSON, J., 1791 History of Somerset Vol 3, Privately published CRAWFORD, O. G. S., 1922. The Anglo-Saxon Bounds of Bedwyn and Burbage. WANHM 41, 281-301 CRAWFORD, O. G. S., 1953(a). The East End of the Wansdyke. WANHM 55, 119-125 CRAWFORD, O. G. S., 1953(b), Archaeology in the Field, London: Phoenix House EKWALL, E., 1928, English River-names. Oxford, Clarendon Press. FOWLER, P J., 2001, ‘Wansdyke in the Woods: An unfinished Roman Earthwork for a non-event’ in P. Ellis (ed.), 2001, Roman Wiltshire and After, Papers in Honour of Ken Annable, 179-98. Devizes: Wiltshire Archaeological and Natural History Society FOX, A. AND FOX, C. 1960. Wansdyke Reconsidered. Archaeological Journal 115, 1-48 GOVER, J.E.B., MAWER, A. AND STENTON, FM., 1939, The Place-names of Wiltshire. English Place-name Society. Cambridge University Press GRUNDY, G.B., 1919, Saxon Land Charters of Wiltshire. Archaeological fournal 76, 143-301 GRUNDY, G.B., 1920, Saxon Land Charters of Wiltshire. Archaeological Fournal 77, 8-126 HOARE, Sir R.C.,1819, The Ancient History of Wiltshire. London: Lackington, Hughes, Harding, Mavor & Jones HOOKE, D., 1998, ‘Medieval Forests and Parks in Southern and Central England’, in C. Watkins (ed.), European Woods and Forest: Studies in Cultural History, 19-32. Oxford: CAB International HOSTETTER, E. AND HOWE, T. H. (eds), 1997, The Romano-British Villa at Castle Copse, Great Bedwyn. Bloomington: Indiana University Press KEMBLE.J. M. (ed.), 1839-48, Codex Diplomaticus oevi Saxonici, 6 vols. London. LENNON, B., and CROW, P, 2009. LiDAR and its role in understanding the historic landscape of Savernake Forest. WANHM 102, 245-61 MAJOR, A. FE and BURROW, E. J., 1926, The Mystery Of Wansdyke. Ed Burrow And Co., Privately published Cheltenham PITT-RIVERS, A. H. L. F, 1892, Excavations in Bokerly and Wansdyke Dorset and Wilts., 1888-1891. Privately published RACKHAM, O., 1993, The History of the Countryside. London. J.M. Dent. RAMSAY, J. and BATHE, G., 2008. The great inclosure of Savernake with a note on cross valley dykes. WANHM 101, 158-75 REYNOLDS, A. and LANGLANDS, A., 2006, ‘Social Identities on the Macro Scale: A Maximum View of Wansdyke’, in W. Davies, G. Halsall and A. Reynolds (eds), People and Space in the Middle Ages, 300-1300, 13-44. Turnhout: Brepols. SAWYER, P. H., 1968, Anglo-Saxon Charters: an annotated list and bibliography. London: Royal Historical Society SAXTON, C. (author) and RAVENHILL, W. (Introduction), 1992, Christopher Saxton’s 16th Century Maps. Chatsworth: Chatsworth Library STUKELEY, W., 1776, Itinerartum Curiosum. London: Baker and Leigh TAYLOR, C. S., 1904. The Date of Wansdyke. Transactions of the Bristol and Gloucestershire Archaeological Society 27, 131-155 Wiltshire Historic Environment Record, Wiltshire County Council. The Electronic Sawyer. Catalogue of Anglo-Saxon Charters: http://www.trin.cam.ac.uk/chartwww/eSawyer.99/ eSawyer2.html. Consulted 1/1/07 Published Map Sources Andrews and Dury’s Map of Wiltshire 1773 Andrews and Dury’s Map of Wiltshire 1810. Published by William Faden. 1817 Ordnance Survey one inch “Old series”. (reproduced by timeline maps) 1880s Ist Ed. Ordnance Survey. Epoch 1. 25” 1880s 1st Ed. Ordnance Survey. Epoch 1. 6” 1920s 1st Ed. Ordnance Survey. Epoch 2. 25” Primary Sources Wiltshire and Swindon Archive (WSA) 15/11/07 A plan of the Manor of the Stock & Ford with the borough of Great Bedwyn. Dated 1725. WSA 1300/59C Roll of Henry Esturmy, warden of forest of attachments and inquisitions of vert and venison. 1332-1334. WSA 1300/65 14th Century Parchment Document. Perambulation of the Forest of Savernack. WSA 9/8/135/h Map of North Standen by George Wythers dated 1612. WSA x6/4 Map of Savernake Forest Estate. 1820 Forestry Commission. LiDAR survey of Savernake Forest, 2006. eee eeepereeeeesnee Wiltshire Archaeological & Natural History Magazine, vol. 103 (2010), pp. 289-308 Devizes Literary and Scientific Institution: a study of change over time by Elizabeth Gibb The Devizes Literary and Scientific Institution was founded in 1833. This account describes its foundation, notable events in its history and its gradual transformation from an educational establishment to a recreational club. An advertisement in the Devizes and Wiltshire Gazette in early July 1833 invited those inhabitants of Devizes and the surrounding districts interested in founding a Mechanics Institution, for the Diffusion of Useful Knowledge, to attend a meeting in the Town Hall. The aim of the Institution was to be the _ instruction of its members in science and general | knowledge. The mechanics of the town in particular were urged to attend.! Today a mechanic is associated with machinery but, when the first mechanics’ institutes were established in the early 1820s, the use of power- driven machinery had not been widely developed and so the term ‘mechanic’ then did not have the meaning it does today.’ Rather, the term referred to a skilled craftsman and tradesman, men who needed to keep up with the latest ideas, new inventions, new methods and materials in a period of rapid technological change. The meeting was held on 22 July, the Mayor | presiding. Captain Tayler RN, the main advocate of _ the proposal, addressed the meeting, pointing out the benefits a Mechanics Institution bestowed both on the individual and the community: that which would exalt the intellectual powers of one body, tended to diffuse its influence all around; and in proportion as man rose in knowledge he became a better and more useful member of the community. He went on to say that the lower classes undoubtedly had able minds but that they needed to be cultivated and stimulated into life. He pointed out that the world’s most renowned philosophers had humble origins and that the greatest scientific improvements had been made by ‘persons who moved in the lower class of society’, saying the development of ‘that stupendous machine, the steam engine’, had been by a mechanic. The first step, he said, would be to procure a library of useful and scientific works — but excluding any on politics. He emphasised that politics in any shape must not be allowed in the Institution; it was to be for the cultivation of the arts and sciences. Following his address it was decided unanimously that a Mechanics Institute should be formed in Devizes and a committee was appointed to draw up its rules and regulations, these to be presented to a general meeting the following month.? At that meeting the objects of the Institution were set out. The instruction of its members, in science and general knowledge, would be by means of lectures to be delivered at stated times, a library, museum and philosophical apparatus to which members would have access and the formation of classes amongst its members. A committee was elected to manage the new Institution and the following officers were chosen: President, T H S B Estcourt; Treasurer, Wadham Locke, MP; Honorary Secretary, The Rev Mr | 15 Castle Court, St. John’s St., Devizes SN10 1DQ 290 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE |. Mayo; Librarian, Mr Symes and as Curator, Joseph Sloper. Under the newly adopted rules and regulations the committee was to be made up of twenty-four subscribing members, of whom no less than one-third had to be working mechanics. Each working mechanic was to subscribe ls. quarterly, in advance; and all others 2s. 6d. Ladies, many of whom had attended the inaugural meeting and had enthusiastically supported the founding of the Institute, were offered access to the Library, Museum, and all lectures, readings and conversations on payment of 2s. 6d. quarterly. Visitors could attend a lecture, reading or conversation, by permission of a Committee member, on payment of Is. The library was to be open from nine o’clock in the morning to nine every evening and a certain number of books purchased by the Institution could be borrowed, but those volumes presented to the Institution to be excluded. The rules and regulations again reiterated that all controversies on religion and politics be strictly forbidden.’ It must have been at this meeting that the name ‘The Mechanics Institution’ was dropped in favour of ‘The Devizes Literary and Scientific Institution’, a title often adopted in less industrialised rural areas. Two weeks later, on 29 August 1833 at the Town Hall, the first lecture was delivered to the members, who it was said already numbered one hundred. It was given by Mr Burt, described as a mechanic, as the first of three talks on “The History and Principles of Architecture’. Thereafter the lectures continued at weekly intervals throughout the season, until 1 May the following year (Appendix 1). On 17 October the Institution opened its first premises, at the New Room in High Street, originally built for the Society of Friends, and where the fledgling library and museum were housed.° At the first Annual Meeting, held on 31 July 1834, it was reported that there were 310 volumes on the library shelves, all donated by well wishers. In addition 80 books had been purchased by the Institution to form a circulating library. Generous donations to the museum had been made also, including a pair of fine globes from Wadham Locke, MP° and a specimen tooth of a Narwhal, or Sea Unicorn, presented by W H F Talbot of Lacock Abbey.’ Philosophical apparatus, too, had been presented, an electrical machine, two galvanic troughs, and models to illustrate mechanical powers. The President, in his Address, reported that thirty-four lectures had been delivered by sixteen lecturers, many of whom were members willing to share their knowledge and talents with their fellows. Frequent readings had also been given. Three classes had been formed to study Mechanics, Geology and Mineralogy and Architectural Drawing and Geometry. He reported that there were 349 subscribers including around 24 ladies and several young boys. The Chairman commented that on many occasions it had been necessary to adjourn the meetings to a larger building in order to admit the numbers wishing to attend and it was clear that the Institution had already outgrown its premises.® Later that same year it was reported that the Institution: is in contemplation to erect a suite of rooms over the entrance to the Market-house which, if arrangements can be made, will be proceeded with in the course of a week or two.’ This plan, however, came to nothing and the Town Hall continued to be used for over-subscribed lectures. One exception to this arrangement took place in February 1835 when the skeleton of a whale captured at Plymouth was exhibited for several days in Devizes market-place. It reportedly measured 102 feet in length and had weighed 448,000 pounds (i.e. 200 tons). Dr Tomkins gave a demonstrative lecture to the members of the Devizes Literary and Scientific Institution and others on ‘The Natural History and Anatomy of the Whale’ in the room in which the skeleton was being exhibited. This space was formed by the junction of several large vans and although spacious, it was uncomfortably crowded by the attendance of some 200-300 persons."° Throughout the following year supporters continued to make donations to the Institution. At the second Annual Meeting held on 2 July 1835 it was reported that many valuable works had been presented to the library including 45 volumes of Rees’s Encyclopaedia and various reports of Committees of the House of Commons. The museum had received a fully fitted microscope, two boxes of mineral specimens, some South American birds, maps and plans; but unfortunately, owing to lack of space at their premises, it was still unable to accept a collection of fossils and minerals offered by member W W Salmon. Fewer lectures had been delivered than in the first year but, following the very successful series given by Dr Robert Brabant on ‘Chemistry’ illustrated by experiments, that had attracted large numbers of visitors, many of whom had travelled quite a distance, Dr Warwick had been engaged to give a course of lectures on ‘Chemistry’, ‘Magnetism’, ‘Electricity’, etc. commencing 17 q a : DEVIZES LITERARY AND SCIENTIFIC INSTITUTION September. Other lectures planned for the season were ‘The Science of Optics’, ‘The Physiology of Digestion’, ‘Natural History’, ‘Liberty’ and ‘Architecture’ by T H Wyatt, architect of the recently completed Assize Court in Devizes. Early in 1836, however, the Institution was plunged into controversy when several members requested that newspapers be provided in the reading room. A letter to the editor of the Devizes and Wiltshire Gazette expressed the danger seen by some in taking such a step: Newspapers are now, it is said, to be introduced; political observations and discussion will naturally follow; and the object of the Institution being thus frustrated, the inevitable consequence will be its downfall." A majority of the Committee considered that such provision would be of benefit to the members but it was decided to call a general meeting before any decision was made. At the meeting a letter was read out from Capt. Tayler, the original promoter of the Institution, in which he said: . if newspapers convey knowledge and amusement to the rich, I cannot conceive upon what principles the operatives should be deprived of their share of the benefit. [He considered] . . . that reading newspapers would prompt the Mechanic to higher reading. They had at present no means of obtaining the sight of a newspaper but by frequenting the public house. In his opinion the introduction of newspapers would improve the minds of the mechanics and the Institution’s funds, as the expenditure at this time exceeded the income due to a drop in membership from 340 to 180. He pointed out that there were on the library shelves a vast number of books containing the history of past ages and the prophecies of ages to come and that he could not see why an account of the present times should be excluded. Another supporter said that the most learned, the most intelligent of mankind, daily read the newspaper, even Statesmen _ derived information from them. The tables of the Institution at Bath were covered with them; they were introduced into our colleges; they were to be found in nearly all the houses of the clergy; and yet the mechanics were to be debarred from reading them .. .!” Speakers opposing the request believed that their admission would lead to contention and strife 291 within the Institution. The outcome of the debate was not clear-cut and it was decided to canvass the opinions of all the members before proceeding. A second general meeting, noisy and heated, was held on 3 March: ... those gentlemen who advocated the measure were listened to with patience and respect; but we sincerely regret to say that the most unbecoming and indecent conduct was observed towards some of those who expressed unfavourable opinions." Of those canvassed a majority of seven was in favour of the introduction of newspapers and so the request was granted. Four months later, in a letter to the editor of the Devizes and Wiltshire Gazette, the President of the Institution, the Rev C Lucas, observed that, despite the fears, political controversy had not followed their introduction. Without a doubt, however, the affair did affect the Institution and polarised political opinion within it. A conservative correspondent claimed some time later that ... the Radicals insisted on deviating from the original plan of this Institution by introducing newspapers and consequently politics; and that by their conduct, nearly all the conservative members felt themselves driven from the room .. .- now [the Institution] is looked upon as a mere party or radical affair .'* The opening lecture of the 1836-7 season was delivered by the President, Dr Brabant, at the Town Hall on 3 November on the subject of ‘Witchcraft’. Despite all the acrimony over the introduction of newspapers at the Institution, the event was well supported with over 200 persons attending, including Thomas Moore, the poet, and his family, who resided at Bromham. Other topics that season included ‘Ancient Customs’, “The History of Devizes Castle’ delivered by James Waylen, and a return visit by Messrs Brady. Readings and general discussions upon some particular subject took place alternately with the lectures. In November the subject for discussion was ‘The Devizes Improvements’, which included the town’s on-going modernization by the Improvement Commissioners, the Kennet and Avon canal and the new prison. The 1830s and 1840s was a period of great political and economic agitation which was reflected locally. Following the Liberal triumph over the Tories in the general election of 1831, under Earl Grey the electoral Reform Bill was introduced. This provided for the relocation of parliamentary 292 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE . seats — rotten boroughs such as Old Sarum, site of an Iron Age fort, had two MPs but flourishing industrial towns such as Manchester, Leeds and Sheffield had no representation at all — and for extension of the franchise, the right to vote. The Church and the Tories, in whose interests it was to keep the link between property and representation, vigorously opposed this Bill and there were riots in Bristol and elsewhere following its second defeat but it was carried at the third attempt. Devizes kept its two MPs and over 400 male householders paying £10 rent annually, i.e. the middle classes, were enfranchised. Many who were actively involved locally in the struggle for electoral reform went on to found the Devizes Literary and Scientific Institution — professional men such as Wadham Locke, Capt. Tayler, Robert Waylen, Dr Brabant and George W Anstie; Benjamin and Paul Anstie, James Randell, Samuel Whitchurch and Mr Knight, manufacturers and tradesmen of Devizes; and the Non-conformist minister the Rev Richard Elliott.» The passing of the Reform Act in 1832, however, did not satisfy everyone. There was a sense of betrayal throughout the country that universal suffrage had not been granted and out of this grew the Chartist movement — to achieve further electoral reform as set out in the People’s Charter. This listed six main aims — annual parliaments, universal male suffrage, equal electoral districts, the removal of the property qualification for MPs, a secret ballot and payment of MPs. The movement grew rapidly, fired by the current economic depression; by fears, particularly in the north of the country, of unemployment due to the introduction of machinery; and by anger against the new Poor Law which, by withdrawing parish relief for the able-bodied poor, gave assistance only when the whole family entered the workhouse. The Charter, drawn up in 1838, was perceived by the working classes as the answer to all their ills, and associations were quickly established in Bath, Bradford-on-Avon, Trowbridge and other manufacturing towns in Wiltshire, as well as in smaller communities such as Holt. By March 1839 there was a Devizes association that met at the Curriers Arms in Bridewell Street.'® The struggle to have the Charter made law, often violent with strikes and riots, finally ended in failure in 1848. Universal male suffrage was not granted until 1918 and it was another ten years before women over the age of 21 were given the right to vote. Simultaneously there was a growing demand for the repeal of the Corn Laws, strongly resisted by landowners and farmers. Under these laws the importation of cheap foreign corn was curtailed, with the result that the price of corn at home rose and so did the price of bread, causing particular hardship to the poorer classes. The Anti-Corn Law League was founded by businessmen who favoured free trade; they argued that the Corn Laws harmed not only the poor but also the entire economy. The Corn Laws were finally repealed in June 1848. An article in the pro-Tory Devizes and Wiltshire Gazette during this period exemplified the bitter animosity in the town against the reformers: The person selected last year to represent the radicals of this part of the country at the Anti- Corn law conference could not leave Devizes, it was said, until a subscription was entered into to provide him with a pair of breeches. This year a Mr Oriel is very pompously announced as the delegate ... and Mr Oriel, it is said, has made quite a sensation! He figures in the Sun with the letters A.C.L.D. after his name!! Only think of William Oriel, esq. A.C.L.D! Why people who know anything of Mr Oriel in this neighbourhood, know that he keeps a shop, a little bit of a shop on Devizes Green — and that he himself is of as small dimensions as his shop."” During these years Devizes was torn apart by political feuds and, as a newspaper correspondent commented, ‘its trade seriously injured as a result . . . peace would be most desirable’.'® To counter such bitter partisanship, and to gain a platform for their views, the local Liberals launched their own newspaper, the Wiltshire Independent, in November 1836. Several prominent members of the Institution were shareholders: Benjamin Anstie and Paul Anstie, snuff makers; George W Anstie, solicitor; Robert Valentine Leach, corn factor; James Randell, coal merchant; Samuel Whitchurch; and the Rev Richard Elliot, Congregational minister, amongst others. It would appear that the fortunes of the Institution were slow to recover following the rupture over the admittance of newspapers, for in 1838 a speaker commented: the Devizes Literary and Scientific Institution has remained stationary or rather fallen back, while our members (at Salisbury) have increased from 50 — 500. I used to read in the Gazette of your prosperity, now I seldom see your society mentioned = 4.) At the Annual Meeting of 1839 the President, the Rev C Lucas, in his Address took members to task DEVIZES LITERARY AND SCIENTIFIC INSTITUTION for employing visiting lecturers rather than finding them from among their number. He pointed out that by preparing lectures for the instruction of their fellow members, ‘they would be laying by a store of well-arranged information for themselves’ — ‘self improvement, the great object of Scientific Institutions’, as a later speaker remarked. There would be a financial saving, too, as it would appear that their financial position was far from secure. Tribute was paid to the patronage of the ladies whose presence, the Rev Johnson said, had done much to support the cause of scientific education. It was said by some that ladies ought not to have anything to do with science, in short they could not understand it, but he was sure the female mind could reap the same benefits as that of the male and he would like to see the study of Philosophy cultivated by the ladies. He also spoke on the desirability of holding classes on algebra and geometry within the Institution to facilitate the study of the sciences, as many works were wholly unintelligible without such knowledge. In this the Devizes Institution lagged behind others. A suggestion that prizes be awarded annually for the encouragement of literature and science had been under consideration, and at the annual meeting it was decided that three medals would be awarded, one for each of the following categories: The best collection of British plants to be made, named and arranged, according to the Linnaean system by any one member from the present time to the next annual meeting. The best essay on the Improvement of Time. A mechanical model which shall most tend to the improvement of the useful arts.”° The two Patrons, the President and the Vice- presidents, generously met the cost of the medal die which was exhibited at the Devizes and Wiltshire exhibition the following year. This impressive exhibition was mounted in the Town Hall, from12 August to 8 October, for the mutual financial benefit of the Literary and Scientific Institution and the Devizes Dispensary. The announcement of the planned event revived the ill-feeling of some Conservatives against the Institution, caused by the introduction of newspapers there four years previously, and one correspondent wrote that having been forced out it was, ‘too much to ask the Conservatives to go out of their way to support such an Institution, for the mere benefit of their political opponents’.”) That this Opinion was not held generally was demonstrated | by the magnificent response to the appeal for the 293 loan of exhibits. The catalogue lists 263 paintings, including works by Canaletto, Brueghel, Reynolds, Poussin, Vandyke, Sir Peter Lely, Murillo, Cuyp and three portraits by Sir Thomas Lawrence including the two full-length portraits of George III and Queen Charlotte that hang in the Assembly Room today. No fewer than 98 drawings were exhibited by such masters as Cotman, J M W Turner, Sir Thomas Lawrence and Gainsborough; and 125 engravings, some by Hogarth. Natural History exhibits included 128 specimens of British birds and 176 foreign ones. There were 403 examples of fossils, shells, minerals, coins, antiquities and curiosities, such as lamps from Herculaneum, a dagger made of lava from Mount Vesuvius, shoes from China, Turkey and India, and sugar cane. There were 277 sundry items including a section devoted to scientific apparatus, including a model of a gas works, an air pump, a small working orrery and a model of a steam engine. Exhibits were borrowed from across Wiltshire, from the Marquis of Lansdowne, T P Methuen, Esq of Corsham, J Neeld, Esq of Grittleton, and from John Britton, the antiquary, for example. Many prominent figures in Devizes also lent their support, among them T H Sotheron, MP, R V Leach of Devizes Castle, G E Sloper, W Cunnington and Thomas Moore. Other Literary and Scientific Institutions also sent exhibits; Bristol sent a dolphin’s head, the jaws of a shark, a porpoise, and several other objects, including the head of a New Zealander! ” The items exhibited by the Devizes branch give an interesting insight into the contents of their museum at that time (Appendix 2). The Institution then had a membership of 181, some 94 of whom were mechanics and 87 ‘other members’, a remarkable ratio. This was not matched anywhere else in Wiltshire and the Devizes branch was alone in having its own premises.” After almost ten years in the High Street the Institution removed, in late September 1842, to New Park Street, close to St Mary’s church. The ground floor premises, described as extensive and commodious, had been previously occupied by Miss Sibree’s school. The Census Return of 1841 shows that this three-storey building was between the Royal Oak and New Park Road, between the premises of whitesmith Joel Holloway and those of surgeon Charles Trinder. The three-storey brick building, housing nos. 44 and 45, next to the Royal Oak, has several early architectural features such as timber framing, visible on the side wall of the yard entrance, first floor oak beams, and fielded panelling 294 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE | in the interior of No 44, none of which would have survived the fire that was to engulf the Institution in 1843 if had been located there. Therefore, it must have been where No. 46 New Park Street stands today, adjoining no. 45. The present building, once occupied by Way’s Garage and later by Renelec, was built in the early 19th century, probably following the fire. The removal to the new premises was celebrated by a Social Tea Party attended by about 180 persons at 2s. per head, ‘amongst whom were several of the most wealthy, influential and talented inhabitants of the town’. Following an excellent tea and a musical interlude the President, the Rev Richard Elliott, took the Chair, with its motto ‘Live to learn that you may learn to live’, and addressed the company. He said he was pleased to find the future prospects of the Institution now looked brighter than they had been and referred to past differences that had arisen within the Institution, due in part to politics, class and religious opinion but which he trusted would all now be removed. He then presented the society’s medal, awarded by the judges, the Hon Member for the Borough, Mr Sotheron, the Rev C Lucas and Dr. Brabant, to Mr Maysmore, a bank clerk, the author of the prize-winning essay on ‘The Improvement of Time’. Many of the junior members then went to another room to watch experiments while those who remained heard a talk on Geology by William Cunnington. The Social ended with the singing of several songs followed by the National Anthem.” However, an account of the tea party published in the partisan Devizes and Wiltshire Gazette newspaper painted a very different picture, claiming that the mechanics: sour in temper and yellow and discontented or the unwashed as the radical members call them, huddled together in a different room from the rest of the company. . . After tea the unwashed were admitted among their more dignified brethren. It went on to claim falsely that many of Alderman William Cunnington’s spoons, loaned for the occasion, were missing. This offensive article, designed it was claimed to cause mischief within the Institution, resulted in a flurry of replies printed in the Wiltshire Independent newspaper” that attributed the Gazette’s spleen to the decision of the committee to vacate their late premises in High Street — the property of the Editor of the Gazette! This attack, as much personal as political, caused great disgust in the town and for once united Tories and Liberals in its condemnation. Three months later, in the early hours of 9 January 1843, disaster struck when the new premises were totally destroyed by fire. Other tenants in the building, owned by Miss Dyke Poor, were Mr Goldstone, a surgeon dentist on the first floor, who had moved there only a month earlier, and Mr Sinclair, who acted as caretaker, and his family on the second. It was when Mrs Sinclair got up to attend to her crying child that the fire was discovered and the alarm raised. Although the watchmen were quickly on the scene and a great number of townspeople gathered to give assistance, within an hour the building was engulfed in flames. Strong south-west winds fanned the flames and it took great effort to prevent the fire spreading, north to Mr Trinder’s mansion house, and with less success, south to Mr Holloway’s house which, despite efforts to contain the fire by opening the roof and cutting away the timbers, was partially destroyed. Unfortunately, the two town engines were found to be totally ineffective, due to poor maintenance, and an engine was hurriedly borrowed from Mr Tylee’s brewery. Mr Colston, of New Park, sent his engine and a third was sent for from Melksham. Meanwhile, the townspeople salvaged what they could from the three premises and fought the fire with buckets of water and wet blankets. There was a tragic sequel to this disastrous fire. Later that day several people were in the ruins checking that the fire was extinguished when the floor gave way. Most scrambled out of the debris but a young man called Brewer was trapped to his waist in burning bricks; he died later as a result. Another young man, named Powell, was also very badly burnt.” During the investigation that followed it was stated by Mr Sinclair that, having removed his wife and five children to a place of safety, he went into the ground-floor premises occupied by the Institution and found no fire in the news-room nor in the library behind it, although Mr Goldstone’s drawing-room, over the news-room, was well alight. It was concluded that a beam, which ran from the chimney of the newsroom under Mr Goldstone’s drawing room to Mr Holloway’s house, had caught alight with such dreadful consequences. Fortunately for the Institution a large proportion of its books and other property was saved. It found temporary refuge at Mr Leach’s premises in Wine Street, formerly the Temperance Coffee House, before re-establishing itself in August 1843 in premises vacated by the Chequer Auction Rooms in Exchequer Lane. (now known as The Chequer and occupied today by the firm Edwin Giddings).’’ a a Be DEVIZES LITERARY AND SCIENTIFIC INSTITUTION Settled in its new premises the Institution resumed its lectures, and in October a soirée, similar to that held the year before in their rooms in New Park Street, though on an extended scale, was organised at the Town Hall. About 350 people attended, nearly half of them members. After tea, while the room was being cleared, there was an exhibition of ‘fire cloud’ (unexplained) in the Court below, conducted and explained by William Cunnington, assisted by his brother and friends. He conducted other experiments in the Council Chamber during an interval in the evening’s musical entertainment. An Address was given on the usefulness of such societies and the President commented, ‘what institution can be more worthy of support than that which has for its object the enlargement of the human mind, the elevation of human character’.”8 The third anniversary celebration, in October 1844, followed the same format, though the musical entertainment provided was more ambitious than in previous years — Haydn’s Third Symphony, Mozart’s overture to “The Marriage of Figaro’, the overture to ‘The Barber of Seville’ by Rossini and Beethoven’s ‘Men of Prometheus’ overture. Philosophical and chemical experiments were again carried out in the Court at the Town Hall and much amusement was caused by a demonstration of the effects of laughing gas. Such was the general interest in matters scientific that the following month a series of philosophical and other experiments was exhibited in the Town Hall. The apparatus was constructed by a mechanic, much of it original in design, and on two evenings explanatory lectures were given. An advertisement in the local press informed the general public that the first would be on: ‘Electro Magnetism’. Illustrated by numerous, novel and striking experiments amongst which will be shown a new form of Electro Magnetic Power, giving motion to a model of the Aerial Machine, etc. etc. and a working model of the Atmospheric Railway’ and the second on ‘Electricity’ when a variety of beautiful experiments connected with the science will be introduced — The principles of The Rotating Magnet will be explained.” The Annual Soirée held on 2 Nov 1847 furthered this interest, a greater proportion of the evening being devoted to scientific subjects. After an address 295 by the Rev R Elliott, the Rev J H Johnson gave a talk on the history of the microscope, six examples of which were on show. William Cunnington gave a lecture on the ichthyosaurus, an extinct marine animal, the remains of which he exhibited. Various experiments were demonstrated — electrical, magnetic and chemical involving gun cotton, an explosive made by soaking cotton in nitric and sulphuric acids and used for blasting. Other demonstrations included Phantasmagoria, optical illusions; the Daguerreotype, an early photographic process, and Electric type. Light entertainment was provided by H Hersee who gave a lecture on the ‘Songs of the Present Century’, and read ‘The Steam Boat Excursion’ by Charles Dickens.” Again the evening was a great success with an attendance of 320 persons. At that time the Institution had 145 members, the maximum number that could be accommodated in their lecture room. To increase the number of members, and also the attendance at lectures, the Committee would have to find larger premises. A sub-committee was appointed to enquire, ‘into the expediency and practicability of removing the Institution to more central and commodious premises’.*! William Cunnington, wine merchant and Institution member, offered the use of the three rooms above his wine business in St John’s Street, two of which, when thrown together, would seat 220 persons. This proposal met with some opposition during discussion by the Committee and an amendment was proposed: that we do not remove from these premises: 1.Because the funds will not admit of such a removal; 2 Because the premises of Messrs Cunnington do not afford that extra accommodation which the Institution requires and consequently the removal to that place would not be permanent.” To liquidate the debt on the lease of the premises in the Chequer would cost about £35, the balance in hand being only £9 12s. 6d. The amendment was carried by eight votes to six. However, following this decision by the Committee, twenty-one members, who supported the proposed move, made their views known to the Secretaries, who called a Special General Meeting to discuss, ‘the suitability of the present premises to their needs now and in the future’. After much discussion the meeting was adjourned to enable the Secretaries to obtain information from the managers of The Savings Bank about funding the building of new premises. They were unable to assist the Institution: “The 296 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE - Managers of The Savings Bank could hold out no hope or expectation that premises would be erected from the funds of The Savings Bank’.** In February the proposal that the premises of Messrs. Cunnington be taken on a lease at an annual rent of £25 was carried unanimously, conditional upon the sum necessary for defraying the loss on the present lease and that the expenses of removal and alterations could be raised. The remainder of the Chequer lease was taken over by J E Hayward at £10 p.a. The cost of the removal from the Chequer to the rooms in St John’s Street would amount to £141 3s .5d. Money was raised through the annual auction sale of the periodicals, excepting the Mechanics Magazine. These included the Athenaeum, Tait, Chambers, Sharpe’s London fournal, Penny Magazine, Hoggs Weekly Instructor, People’s Journal, Illuminated Magazine and Punch. All the old newspapers, excepting the Economist, were sold as waste paper at 2d. per lb. The Bristol and Clifton Orpheus Society was invited to give a musical performance in Devizes for the Institution’s benefit and their Glee concert took place on the evening of Monday 25 September at the Town Hall.* To meet the remaining shortfall a fund was set up to which almost fifty persons subscribed, including J H S Sotheron, Esq MP; T H S B Estcourt Esq MP; Walter Long, Esq MP; Joseph Neeld Esq MP; Messrs P Anstie, Snr and Jnr; J B Anstie; and W Anstie. The rooms in the past had served a variety of uses — a militia stores, a mess room for the officers of the Devizes Loyal Corps of Volunteers, a temporary Sunday School and Magistrates’ Offices. They were opened on 11 May 1848 offering the following subscriptions: To the Newsroom, lectures, Library, etc. 20s. p.a. To the lectures, Library and Museum 12s. p.a. Ditto for mechanics 6s. p.a. Ditto for juniors As. p.a. The rooms were open from 9am to 10pm, although the Newsroom was reserved for the sole use of the 20s. subscribers from 9am to 6pm, afterwards to all other subscribers. It must have been with great satisfaction when four months later, at the Annual Meeting, the chairman announced that the membership had grown to over 200, particularly as there had been an increase in the number of mechanics belonging to the Institution. It was reported that there were: 76 members paying 20s. 41 members paying 12s. 63 members paying 6s. (the mechanics) 29 members paying 4s. (Juniors under 16 years of age) 1 member paying 10s. Total 210 members paying £125 16s. The property of the Institution had also increased considerably, both by purchases made on its behalf and donations from private individuals. Sixteen lectures had been delivered during the period 11 November 1847 to 9 March 1848, and a chess class established that met on Tuesday evenings in the Newsroom. However, the financial situation was not so satisfactory for, after payment of the librarian’s £14 salary, the day attendant’s at £7 19s., payment to the boy who assisted in the museum 8s. ld. and rent of premises, the balance in hand was negligible. The following year it was reported that the membership had fallen back to 180, probably because some had been induced to join the society because of the novelty of the move to the new premises in St John’s Street. At the Annual Meeting the committee © of the Institution urged all its members to do: everything in their power to promote the object it has in view — that of diffusing sound and useful knowledge through the minds of the rising generation . . . to assist those whose efforts are directed to turn the rising tide of popular education into safe channels as the only way to improve the future character of the people and secure the continued safety of the state: an exhortation made perhaps with the recent overthrow of the monarchy and the proclamation of a Republic in France, and at home the Chartist strikes and riots, in mind. It is rather surprising, therefore, that the programme committee issued an invitation to Henry Vincent, a controversial figure, to give two lectures to the members during the forthcoming season and by doing so plunged the Institution once again into turmoil and dissension. Ten years previously he had been a leading figure in the Chartist movement, a fluent and impassioned speaker who had travelled the West Country speaking at Bath, Trowbridge, Bradford-on-Avon and Holt in support of the Charter. Such meetings had attracted large crowds and the inflammatory language and seditious speeches of some of the speakers and the display and discharge of firearms and other weapons had caused alarm. On 22 March 1839 Vincent, accompanied by about 150 weavers and others, had visited Devizes to obtain more signatures for the National Petition in support of the Charter but, following mud slinging SS re aeepERN eeepc nasa he DEVIZES LITERARY AND SCIENTIFIC INSTITUTION and stone throwing by a crowd of ‘drunken farmers, lawyers, clerks and parsons’, had been forced out of the market-place and had to adjourn to the Curriers Arms in Bridewell Street, the meeting place of local Chartists. He had returned on Easter Monday, ‘in defiance of the Tory persecutors’, leading it is said, 4,000 supporters wearing green ribbons marching 5 or 6 abreast with their banners and flags, some 20-30 of them, bearing various inscriptions such as ‘We will have our rights, come what may’, and ‘Liberty or blood’ and accompanied by a band. But, no sooner had the meeting commenced in the market place than a hostile crowd had set upon the Chartists with stones and bludgeons. Vincent, knocked senseless, and the other speakers sought refuge in the Curniers Arms once again and it was only after the arrival of the High Sheriff and magistrates, who dispersed the mob, that they were able to leave Devizes. Soon after, on 9 May, he had been arrested on a warrant from the Newport magistrates for attending a riotous assembly held there and at the Monmouth Assizes he was found guilty and imprisoned for twelve calendar months, despite evidence by the chief witnesses for the prosecution that Vincent had urged the people to disperse quietly and to keep the peace. Following his release from a second term of imprisonment on 31 January 1841 he had turned to democratic means to try and achieve the advantages he had hoped for from the Chartist movement, standing as an advanced radical. He had unsuccessfully contested Banbury in 1841, Ipswich in 1842, Tavistock in 1843, Kilmarnock in 1844, Plymouth in 1846, Ipswich again in 1847 and York in 1848. He lectured on social and historical questions and was an advocate of free trade and the education of the people.* A letter received from Henry Vincent stating his willingness to deliver the two lectures caused great alarm within the Institution and two members indicated that they would resign if he did so and, when approached, the Mayor refused the Institution the use of the Town Hall. When informed, Vincent offered to waive his fee of £10 and to give his lectures free. The committee accepted this very liberal offer and: ...1n order to conciliate and to concede to the wishes of those who, desiring to hear Henry Vincent, object to his coming in connexion with the Institution, it was agreed . . . that the Chairman, -R. Waylen, should apply to the Secretary of the British School for the use of their room... In case of refusal however, it was agreed unanimously that the lectures should be delivered at the 297 Institution lecture room. They also set out the subjects to be covered in the lectures: 1. The popular and progressive tendencies of the present age — Influence of the Press — Railways — Public Meetings, and the general intercourse of nations upon our character as a people — The point towards which the world is tending. 2. The duties of our age towards the ignorant and the vicious — The interest the rich have in the intellectual elevation of the poor— The interest the poor have in their own education — Popular and vulgar prejudices against the efficient instruction of the people — Responsibilities devolving on young men. However, these moves did not placate 29 members who had signed a memorial stating, ‘the introduction of Mr Vincent as a lecturer will be deeply injurious to the interests of the Society’, and the committee was forced to write and cancel the engagement. The signatories included Richard Biggs (President), William, Henry and Edward Cunnington, G E Sloper, (Snr. and Jnr.) and Sam Sloper, all very prominent members of the Institution. As the mid-point of the century was reached the Annual Report reflected on the great scientific progress made since its commencement, such as the invention of the steam engine and of the electric telegraph. Other landmark achievements, of particular local interest, had been the launch at Bristol of Brunel’s steam ship the Great Western; the photographic discoveries of W H F Talbot of Lacock; the opening of the Great Western Railway route, 118 miles in length, from Paddington to Bristol incorporating the two-mile Box tunnel; and the launch by Prince Albert of the Great Britain, the first screw propeller-driven iron ship, also constructed at Bristol. The forthcoming ‘Great Exhibition’, planned by Prince Albert and members of the Society of Arts , created keen interest within the Institution and Major Byng Hall was invited to lecture at the Town Hall on ‘The Exhibition of Industry of all Nations’. In 1851 a survey of the Mechanics and Literary Institutions in Wiltshire?” showed the membership of the Institution had recovered almost to the level reached four years previously, following its removal to St John’s Street (Table 1). It had a library second only to Swindon in size and, with Swindon and Salisbury, was one of only three branches that held classes for its members, 20 attending in Devizes. It was a seemingly very satisfactory report. 298 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE © Table 1. Mechanics and Literary Institutions in Wiltshire, 1851 ° ° Subscription Corsham LSI Chippenham LSI EF < =} S) (a) = —" N oO N cz N — RMIS alisbury LSI windon MI Trowbridge MI s 6d; 1s 6d Quarterl ‘Trowbridge MI_| 3s; 2s 6d; 1s 6d Quarter! SO (500 00m ae ives (RMIS = Reading and Mutual Improvement Society; MI = Mechanics’ Institution; LSI = Literary and Scientific Institution) Source: Hudson J.W. The History of Adult Education 1851 London. But the improved membership figure masked a worrying situation for, as the number of members at the lower rate of subscription had steadily increased, the number of 20s. subscribers over the four-year period had decreased, from 76 to 59, resulting in a significant loss of income for the Institution. In the 19th Annual Report of 1852 a deficit amounting to £14 5s.114d. was recorded. The expenses amounted to £120 12s. Librarian’s salary EN. 2253... 0d: Day Attendant £8 Ise Hae Year’s rent £25e) Os. 0d: Gas <7 Geall6s? © 0d: Coal £4 5s... 6d. Newspapers £330 0385 be sd: Expenses of lectures £24 Is... 0d. Purchases for Museum £1 25a Od: This debt was cleared after an appeal to the 20s. subscribers. The next year the Committee just managed to keep within its budget reporting a small balance of £2 4s. 8Yad. In an effort to reduce their expenditure the Committee approached the Society of Arts with a view to going into union. Literary and Scientific Institutions were admitted into union by an annual payment of two guineas from their funds and by the President becoming a member of the Society of Arts. As well as a perceived financial benefit it was hoped that the Institution would benefit from having access to interesting and effective speakers. The questionnaire the Committee completed states that they generally paid from 3—5 Guineas for a visiting speaker but were finding lecture expense a serious tax on their funds. By going into union it was agreed that the Society of Arts would negotiate between the Institution and the lecturers and charge a small annual contribution to defray actual expenses, the Committee to settle terms with the lecturer. One of the questions posed in the questionnaire was whether there was any opening for the establishment of a ‘drawing school’ or a ‘modelling school’ in connection with the Institution to which the reply was ‘we think not’. Later, however, a Devizes School of Art was established in 1863 by Mr Wittey, a former secretary of the Institution, and the Annual Soirée of 1866 at the Town Hall was held in conjunction with the Devizes School of Art, when drawings and painting by the pupils were exhibited.** After union the Institution also benefited by gifts of books and papers from the Society of Arts. In 1853 a considerable number of young men requested the formation of a grammar class and Mr Sinclair, the librarian, indicated his willingness to organise it.*? Another new class was established some time later for those interested in choral singing.*® Over the next ten years or so, however, the Institution continued to lose members and income. At the Annual Soirée held on 9 April 1863 the attendance was down on previous years. In his Address the President, Mr Ewart, MB. said: there was a question often asked regarding these Institutions — Why did they not attract a greater number of working men? Was it that the lectures were not commonly and sufficiently clear and plain? — allegations made in a recent article in the Quarterly Review. He observed that lectures should be well illustrated and the subject matter be more relative to members’ lives. As early as 1851 J. W.Hudson, in his History of Adult Education, had commented that in many Mechanics Institutes: the plain and easily understood discourses on the elements of science, and their application to the useful arts, illustrated by numerous experiments, have been abandoned and the preference shown for light literature, criticism, music and the drama... DEVIZES LITERARY AND SCIENTIFIC INSTITUTION This was precisely what had happened in the Devizes Literary and Scientific Institution, as can be seen by comparing the lecture programme for 1833-4 with that of 1850-51 (Appendix 3). By August 1863 the Institution had debts amounting to £40 and the number of members had fallen to about 133.4! A meeting was held to decide its future: whether it should be suffered to die the natural death that has been the ignominious termination of many similar societies or whether a vigorous effort should be made for its resuscitation? The decision was taken to try and revive it. A new committee was appointed and the constitution amended, one result of an expenditure review being a drastic cut in the librarian’s salary from £30 to £15! The outstanding debt was again cleared on application to the more affluent members. Following the reorganisation there was a satisfying increase in the number of mechanics and working class men becoming members, between 30 and 40, for, as the President Mr Merewether remarked, ‘it was such that the Devizes Institute had been originally called into existence, and their interests would in future be held paramount’.”” Was this the reason for its decline, that the mechanics felt it served mainly the interests of the middle classes? The subject matter of the lectures would seem to indicate this was the case. In newspaper accounts of the soirées and lectures much was made of the attendance of ‘the respectable inhabitants of the town’. Professional and business-men had played a leading role within the Institution since its foundation and their subscriptions had kept it economically viable. But by this patronage did the mechanics feel displaced? This may also be why the decision was taken to drop ‘Literary and Scientific’ from the society’s name and in future to call it The Mechanics Institution.” This change, however, did not last long and the Institution reverted to the name it had borne for the previous thirty years. However, over the next few years as the government gradually assumed responsibility for general and technical education, superseding voluntary efforts, a shift in attitude took place within the Institution, its role being perceived as more recreational than educational. In his Annual Address of 1869 the President, Mr Merewether, remarked: he regarded it as established not so much for the purpose of evoking literature from its members, or making them scientific, as for the 299 purpose of affording rational amusement to those who might be pleased to attend lectures and other sources of recreation. A series of Penny Readings was introduced which proved very popular, taking place alternately with the lectures, about once a fortnight at the Town Hall throughout the winter. In December an entertainment, ‘Jest and Earnest’, was given in the Assembly Room to a full house and the following week, sponsored by the Institution, ‘Professor Pepper’s Ghost’ was exhibited at the Corn Exchange. But in 1872, despite the number of subscribers remaining constant, the Institution again registered a deficit. This, it was said, was due to a loss on the lectures. Only two paid lecturers had been engaged that season but the audience on both occasions had been very small, entailing a considerable loss to the funds of the Institution. It was pointed out that for several years past the attendance at lectures and entertainments had fallen off so much that they had become a source of loss, instead of, as in former years, an addition to their income. The Committee therefore reluctantly decided that unless there was a considerable increase in the number of subscribers no more paid lecturers would be engaged.“ After 39 years the lecture seasons ended, although occasionally single lectures were to take place. In 1873 another fundamental change took place when the Institution donated much of its collection of ethnological specimens and geological and natural history objects to the newly established museum of the Wiltshire Archaeological and Natural History Society.” This dismantling of its Museum may have been due in part to pressure on space for at the Annual meeting of 1874 it was stated that, ‘the articles are exhibited to much greater advantage than they could possibly have been on these premises’.*® In 1876 the death occurred of the Right Hon T H Sotheron Estcourt and there was much discussion in the town as to the provision of a suitable memorial. The writer of a letter to the editor of the Devizes and Wilts. Gazette *’ proposed the erection of a building on the site of the Bear Club Charity School to house the School of Art and the Devizes Literary and Scientific Institution, pointing out that Mr Estcourt, its first President, repeatedly gave handsome sums to keep the Institution going. Another correspondent wrote in support: ...I, a careless boy often being decoyed by an interesting book or a game of chess to that Institution. No one can tell the large amount of 300 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE | good done here, . . . the tree which Mr Estcourt helped to plant has borne much good fruit but now a shift to better soil is required to restore it to its former vigour, for vigorous it once was when a wider interest was taken in its welfare. This suggestion however, was not adopted. Instead the memorial took the form of a stone fountain, surmounted by a statue of Mr Estcourt, erected in the Market Place in 1879. Despite the ending of the lecture seasons and the dismantling of the museum it would seem that there was no decline of scientific interest within the Institution and in the town, as a lecture given in 1878 in the Town Hall on “The Four Marvels of Modern Science, explanatory of the construction and use of the telephone, phonograph, microphone and phoneidoscope’ was reported as being very well attended.** And in February 1879 an impressive exhibition of 70 pieces of scientific apparatus was assembled and many demonstrations were carried out during the Annual Soirée held at the Town Hall.*? The introductory Address was given by the Rector of Devizes, the Rev J Hart Burges. A short paper followed this on the ‘Late Eruption of Vesuvius’ read by G F Rodwell, science master at Marlborough College. A second paper, on 80- ton guns was read by Mr Braham of Bath — with experiments! Other papers were read throughout the evening to a large and appreciative audience of members and friends.*° Over the next few years the Institution’s income failed to improve as billiard receipts, for many years an important mainstay of the finances of the Institution, fell. This was due to the installation of a billiard table, in conjunction with a reading-room, at the Oddfellows Hall, taking away not only some subscribers but a considerable number of players as well. The number of members at the Institution fluctuated throughout the 1880s, dropping to a low of 147 in 1887, with a consequent reduction of income. The situation was described by the Honorary Secretary at that year’s Annual meeting as, ‘anything but a satisfactory state of affairs’. In 1902 the books in the library were catalogued, listing a total of 2,337 volumes.*! There were over 120 volumes on Adventure and Travel, including books by Stanley, Dr Livingstone and Mark Twain, and slightly more on Art, Literature and Science, including works by Michael Faraday, Charles Darwin, T H Huxley, John Stuart Mill and Thomas Hobbes. There were 131 biographies and 59 books on Natural History, including 8 written by Richard Jefferies, the Wiltshire author and naturalist. There were books on History, Poetry and Drama and Sports and Pastimes, but it was works of Fiction, 1,715 volumes, that comprised the major part of the lending library. Reference books included Gillman’s Devizes Directory and Kelly’s Wiltshire. Ten newspapers were taken in daily and fifteen weekly, including Building News, Engineering, English Mechanic and Scientific American. Five weekly and eighteen monthly magazines are listed, two being of specific interest to the lady members, Woman at Home and Lady’s Realm. At the Annual Meeting 1902 it was stated that 138 new volumes had been added to the library during the past year but criticism was made of the choice of books, the speaker stating that there was too much fiction. He suggested that a more appropriate name for the Institution might be The Devizes Novel and Newspaper Club! Another argued that the Institution ought to be something more than the reading room and games club it had become. In March 1906 the old British School building in Northgate Street, which had housed part of the Technical Institute, was offered for sale. A Special General Meeting of the members was called to discuss its possible purchase. They were reminded that the present premises were cramped without room to expand. Furthermore, following the recent death of Mr William Cunnington, the owner of the rooms in the Old Town Hall, the future was uncertain as their tenancy was only an annual one. No structural alterations were needed to adapt the British School building for their purposes. There were four main rooms, one on the ground floor, 20ft x 22ft, was suitable for the library; another, 30ft x 27ft, communicating with this was suitable for the reading room and two corresponding rooms overhead were suitable for billiards, snooker, chess, draughts, etc. These two rooms could be made into one if required for lectures. There was also another small room 18ft x 11ft which, it was suggested, could be used as a committee room for some local societies. The spacious entrance lobby could hold 15-20 bicycles that, at the present premises, had to be left stacked against the outer doors in St John’s Street, much to the annoyance of passers-by. The British School building appeared eminently suitable and the resolution to proceed was carried unanimously, provided one half of the total cost could be met by donations, the rest to be raised on mortgage. Before the resolution was carried, however, there was some discussion as to the future role of the Institution. Already it was referred to informally as the Literary Institution, the term ‘Scientific’ having fallen into DEVIZES LITERARY AND SCIENTIFIC INSTITUTION disuse. In 1889 it had been acknowledged that: the Institute hardly met the original design for which it had been established with regard to scientific matters, but as there was in another part of town a very good school of Science and Art, he [the Honorary Secretary] took it that that school now met the purpose.” Was it to adopt a purely social role in its new home with the aim of attracting the young men of the town who had nowhere else to go? The discussion ended with a statement that as well as providing recreational activities there would also be opportunities for intellectual development. On 30 May 1906 the Institution purchased the Northgate Street premises for the sum of £400 but by the date it opened, because £222 had to be spent on alterations and adaptations and £178 on furnishings, the total cost had risen to over £800.°? On 1 January 1907, the Marquess of Bath officially declared the building open. When the public were admitted at 6pm to view the new premises they found on the first floor two large recreation rooms with raised seating equipped with billiard tables and games of chess and draughts; and two well-lit reading rooms, comfortably furnished with chairs and upholstered settees and warmed by coal stoves, and provided with all the leading daily papers and magazines, on the ground floor; and at the front of the building overlooking Northgate Street, was the library, ‘much more spacious than the “cupboard” which had housed it for nearly sixty years .. «, on which a considerable sum of money had been spent.™ In February a meeting was held in the large reading room at the Institution with the object of forming a Devizes Debating Society under the auspices of the Institution. A Debating Society had been tried before, in 1886, but had failed due, it was said, to the difficulty of finding interesting topics for discussion, as politics and religion were necessarily excluded. Lack of room had also been a problem. The new Society would have free use of the room for its meetings but would be run by its own elected chairman and committee, so leaving them with a free hand as to the choice of subjects, as the rule ‘no politics or religion’ which still applied at the Institution, would not then affect them. The annual subscription would be ls. with membership also open to ladies and anyone over the age of 16 years. Meetings would be held on alternate Thursday evenings throughout the debating season. Speeches had to be delivered without notes, except in the 301 case of those opening the debate for each side, who would have the option to do so.*® The first debate was on “The Powers of the House of Lords’ and others followed on ‘A Channel Tunnel’, ‘Sunday Observance’, and ‘Smoking’. The debates proved enormously popular, so much so, that that when the room was full, many people were content to stand. The topics chosen were relevant to the issues of the day ensuring active participation and lively debate. The following season, Mr F N Rogers MP took part in a debate on ‘Socialism’, opposing Mr Bevin, a member of the Social Democratic Federation who moved: that this meeting is of the opinion that the only remedy for the removal of the evils arising from poverty and unemployment lies in the ownership of the land and capital, and the instruments of production, distribution and exchange by the people.”’ Other debates were on ‘The Licensing Bill’, put forward to please the anti-drink lobby of the Liberal party, that proposed reducing the number of licensed premises by a third; ‘Compulsory Military Service’; and “The New Education Bill’, supporting militant Non-Conformist demands for completely secular education, but opposed by the Anglicans and the Roman Catholics who wished to protect their schools. The Roman Catholic view was presented by Father Valluet and the Non—Conformist view by the Rev J Day and the Rev J P Wiles. A novel variation was introduced one month when a mock Trial by Jury was held with a judge, counsel, court officials and witnesses, the subject being an alleged breach of promise of marriage: ... the procedure of a court of law was faithfully followed. The judge, robed and wigged, sat on his bench beneath a draped canopy; . the Clerk of the Court and the four bewigged barristers were at the green baize table beneath, at which accommodation was also found for the representatives of the press; and there was a witness box. The whole audience was the jury as it would be in the case of a motion at a debate. The ‘jury’ filled the room to suffocation and a number of would be jurymen and jurywomen were crowded out. Quite unexpectedly the voting was exactly equal and the Judge promptly accepted that as a verdict for the plaintiff with a “4d damages.°® The debate, Women’s Sphere of Work’, was also unusual in that it was a Ladies Night. Miss Hart 302 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE — put the proposal, ‘that women of the present day are trying to fill places which ought to be occupied by men’. Mrs B.H.Cunnington’s amendment was ‘ that there is no reason why a woman should not fill any place she can legally and efficiently occupy’. The vote, by both sexes in the audience, gave a substantial majority to Miss Hart. In his Address at the Annual Meeting of that year the President, Mr Colston, pointed out that ‘debates were not only generally instructive but counteracted the tendency to hold extreme views by showing those who held them that there was something to be said on the other side’. The Honorary Secretary, Mr F J Phipp, in his Report said that the first complete year in the new premises had been very successful, membership had risen slightly from 314 to 317; the library had issued no fewer than 11,578 volumes, its highest circulation in the history of the Institution. Commenting on the great success of the Debating Society, he said it was due in no little part to its excellent and hard-working Secretary, Mr.J J Slade. Nearly all items of receipts had increased with the result that they were able to reduce the debt on the premises out of revenue by £20 to £197. Under the auspices of the Committee of the Institution a public lecture was arranged for Friday 30 October 1908 on the ‘Coming of the Aeroplane’. It was to be ‘profusely illustrated by Novel, Unique and Striking Experiments, Models and Lantern Views’.*’ The Assembly Room of the Town Hall was very crowded because: ... The Literary and Scientific Institution, [had] fixed the price of admission low (with special terms for its members and schools), the idea being not so much to make a profit as to provide an instructive evening without incurring a loss.” Another popular lecture took place at the Institution early the following year, on 25 February, when Mr Victor Dickins, CB, gave his illustrated lecture on Japan, tracing the present Empire from medieval times. The slides showed, by limelight, reproductions of old Japanese drawings and photographs of Japanese scenery and buildings. The first debate of the 1908-1909 season was on ‘Co-operative Trading’ followed by a full capacity turnout of 200 people for a debate on “Tariff Reform’. Other topics were ‘The Newspaper Press’, ‘The Place of Music in Public Worship’, and ‘Women’s Suffrage’. Because it was anticipated that the last named debate would generate great interest it was decided to hold it at the Town Hall and, as a further departure from the usual procedure, the subject was to be debated by experts — one a representative from The Women’s Social And Political Union (WSPU) led by Mrs Emmeline Pankhurst and one from The Women’s Anti-Suffrage League. Unfortunately, the representative from WSPU arrived late, after her opponent had presented her case to the audience, putting her at a disadvantage being in total ignorance of what had been said. The voting was: For Against Ladies 35 104 Men 20 101 55) 205 Majority against 150 Source: Devizes and Wiltshire Gazette 14 Fanuary 1909. Other debates that season were on ‘The Roman Catholic Disabilities (Removal) Bill’, ‘The Daylight Saving Bill’, and ‘The Disestablishment of the Church of England’.°! In the following year the enthusiastic Honorary Secretary of the Debating Society, J J Slade, resigned and the debates lapsed. As a result a petition, bearing between 70 and 80 signatures of disappointed members, was presented to the Committee asking for their resumption. Fortunately a successor was found and the debates recommenced. That year’s Annual Meeting was followed by a lecture, a custom established some years previously in the old premises but that had fallen into abeyance. Mr Cunnington spoke on the ‘Antiquities of Wiltshire’ illustrated by some 60 lantern slides. During World War I free entry was granted to soldiers billeted in Devizes and to the wounded men at the hospitals. Their use of the billiard room offset a potential loss of income due to the absence of between 30 and 40 members serving at the Front. In 1917 it was reported that the receipts from billiards was the highest ever — £150 11s. 9d. As a consequence of the increased income from billiards over the war years the Committee was able to reduce the capital debt from £185 to £115.” Throughout the 1920s the Institution enjoyed a period of stability with membership averaging 225. New members who joined included a tailor, a school teacher (female), a hairdresser (female), a grocer, a professional musician and a journalist. Several lectures were given, including one on ‘Broadcasting’ in 1926 by J H A Whitehouse of the Broadcasting Company. Following this it was announced that members were welcome to introduce non-member friends to the Reading Room on Sunday evenings | | | | | | | DEVIZES LITERARY AND SCIENTIFIC INSTITUTION when a radio programme was being broadcast. In 1927 Captain Cunnington spoke on the ‘History of Devizes in the Civil War 1643-1645, The Siege of Devizes’, a well attended lecture that raised 45s. The following year a lecturer from the Western Electric Supply Co. spoke on ‘Electricity in the Home’. Electric lighting had been installed at the Institution in 1927 and the gas supply disconnected. In autumn 1927 the Institution was approached by the County Education Dept. with a view to housing a proposed Devizes Branch Library, to be opened within the next few months. Following a meeting, between the Committee and the Chairman and the Librarian of the County Library Committee, it was agreed that the County Library Committee would pay £20 p.a. to the Institution for the use of the premises and the distribution of the County Library books. The County Library was to have its own shelves, provided by them, the opening times to be the same as those of the library of the Institution and the same librarian to act for both. Initially, the libraries were to open Tuesday and Thursday 7.30-9.00 p.m., Saturday afternoon 3.00—4.00p. m. and Saturday evening 7.00-10.00p.m. In May 1928 the arrangement came into effect and worked surprisingly well and amicably for several years. At the Annual Meeting of 1931 the Secretary was able to give: one of the most satisfactory reports presented for many years. Over 200 names on the membership list . . . and a satisfactory balance sheet, the revenue for the past year showing an increase. . .. The library continues to be of good service, a sum of 25.17.4 was spent on the purchase of new books. We now have a well-equipped building and organisation whose advantages can be obtained by members for a very modest subscription and it behoves everyone to make these benefits as widely known as possible, especially among the younger generation of Devizes and neighbourhood. The hundredth birthday of the Institution found it, ‘vigorous, alert, financially sound, in a full career of usefulness and housed in premises which in brightness, comfort and equipment, are exceptional in such cases’. The library, of nearly 10,000 volumes, was being well used by the members, who numbered 207. Some 9,000 books had been issued during the preceding year, an increase of 2,000 on the year before. A wide range of daily and weekly newspapers and periodicals was also provided in the Reading Room. The purchase of a better billiard table the year before had contributed to an increase 303 in takings and the Institution at the close of that momentous year showed a credit balance of £49 14s.9d. Its future looked assured. The next few years, however, up to the outbreak of World War II, found the Institution again in decline and there was great anxiety as to whether it would survive. The membership dropped, year by year, until by 1938 it stood at 137.The loss of members was attributed to two causes. Firstly, many of the younger members were leaving the town in pursuit of better employment and secondly, the Devizes branch of the County Library was enjoying great success. In 1935 the Devizes branch had circulated more books than nearly all other branches in the county and the acquisition of larger premises became essential. In September 1936, after almost eight and a half years, the branch left the Institution and moved to St John’s Street. Those members who used only the library at the Institution left also. The remaining members were again urged to point out the benefits of membership to their friends and to encourage them to join. Membership, it was argued, could benefit those in the surrounding districts by giving them somewhere comfortable and pleasant to wait for their buses or trains, with the Library and Reading Room provided with the chief daily newspapers as well as periodicals and magazines, at their disposal. Discussions were held on how to attract the young and it was decided to allow young members between 16 and 18 years of age to use the billiard tables at an annual 6s. subscription. The Honorary Secretary, after consulting the Devizes headmasters, was to compile a list of ex-pupils still living in the town with a view to enrolling them. A cheap rate for billiards in the afternoon was introduced. Advertising was tried and a new sign, outlining the benefits of membership, was put up outside. Consideration was given to the holding of ‘guest’ nights to publicise the activities at the Institution, also to the formation of a rifle-club and the provision of table tennis. In a bid to attract women and girls a Ladies Billiard Room was created, but it was not a success and was soon abandoned. With a reduced membership there was, of course, a corresponding reduction in income, both from loss of subscriptions and of income from billiards. Each year the billiard takings fell with the exception of 1938.The removal of the Devizes Branch Library in 1936 also meant a loss of income from rent. In 1937 a debit of £36 was reported at the end of the financial year followed by a further loss of £30 in 1938. Economies were put in place: there were cutbacks in the amount spent annually on books for the library; 304 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE © the provision of newspapers and periodicals was reviewed; and Mr Whatley, caretaker and librarian, had his salary reduced by £6. The outbreak of war added to the Institution’s difficulties. As men were called up for active service the membership gradually fell to below 80. The committee decided, however, that it was their duty to carry on for the duration of the war, after which the Institution’s future would have to be carefully considered. Some financial relief was gained by the demand for accommodation from those involved in the war effort. The newspapers and magazines were transferred to the Small Billiard Room and the Reading Room was let to the Red Cross Hospital Supplies Dept. The Committee Room was rented by the Ministry of Works and Buildings (Mines Dept.) and the West Wilts.Conservative Benefit Society rented the Small Billiard Room one Saturday afternoon each month for the purpose of collecting contributions. Monthly membership was offered to the men serving in His Majesty’s Forces based in Devizes and it was reported in 1939 that 90 men had taken up the offer. In 1942 this was also extended to members of the U.S.A. forces. Asa result the billiard and snooker tables were very well patronised. There was a steady drop, however, in the number of books issued from the library due to competition from the Devizes branch of the public library, which was, of course, free. Additions to the library continued to be made, usually novels, while a large quantity of books in poor condition was sold as waste paper. The railings outside the Institution were removed as scrap metal salvage, and the bricks outside the front door were painted white to avoid accidents in the blackout. In 1942 it was agreed that the Billiard Room should be used as a Rest Room in the event of air raids if no other accommodation was available. The price of anthracite for the stoves rose and there was increasing difficulty in getting supplies. Because of the increased billiard takings, and the rentals from the various rooms amounting in 1942 to £66 10s., the Institution in that year was able to show a small credit balance. However, it was recognised that once the war was over and the rented rooms given up there would be a return to their financial problems unless changes were made. At the AGM in 1943, the Chairman of the Committee, Mr London, stated that the maintenance costs of the present building were too high, given the income, and the most practical solution would be to acquire a central site in the town and to build premises suitable to their needs using the proceeds from the sale of the present property, a suggestion which won the unanimous support of those present. In the months following two sites were considered, one behind the General Post Office in Castle Lane, owned by Mrs Read of the Castle, and the other in the field next to the Bowling Club, owned by Mrs Gundry of Hillworth Cottage. In 1945 an offer was made for the plot of land behind the Post Office but it was not accepted, nor was a second much increased one. Following the end of the war the Dept. of Mines quit the Institution and the Red Cross gave up the large room they had rented and took instead a small room at the front of the building. The large room vacated by them was let to Miss I Abbott for use as a school, the tenant to find her own labour and fuel. Finding the latter was no easy matter in post- war Britain as the fuel shortage grew more acute. Lack of coal forced many industries throughout the country to close down or work a four-day week and many householders had to queue for supplies. The Committee was forced to apologise to members for reduced room temperatures. In early 1947 the misery was compounded by the coldest winter since 1880-81 and there was fuel at the Institution sufficient only to heat the Billiard Room. Those using the Reading Room did so in some discomfort. During 1946 the number of books issued by the Institution’s library fell to 735 and its days were numbered. The decision to dispose of the books and library fittings was taken in September 1948. The library was advertised for sale in several local newspapers as being of approximately 3,000 volumes, fiction and non-fiction, some being of archaeological interest. There is no record of the fate of a rare book that had been presented to the Institution in December 1934 by Mr Edward Coward, a vice-president. This was a copy of John Gerard’s Historie of Plants (1597), dedicated to Sir William Cecil, Lord Burghley, and Secretary of State to Elizabeth I. It had nearly 1,400 pages illustrated with well-executed woodcuts. Certainly it was not in the library when Mr C W Pugh, of the Wiltshire Archaeological Society, visited, as he found nothing of interest to them.“ Disposal of the library, preceded as it was by the abandonment of the lecture season and the dismantling of the museum, finally completed the transition of the Institution from an educational establishment to a recreational club. In a bid to attract young members a table-tennis table was purchased in 1948 and the quarterly membership fee for 16-18 year olds was set at 2s. A number of young people of both sexes joined and | | | | | | | | | | DEVIZES LITERARY AND SCIENTIFIC INSTITUTION played in the newly formed Devizes Table Tennis League in competition. Also, that year a Devizes and District Snooker League was created which stimulated interest and the Billiard Room receipts showed a satisfying increase. The Institution, however, was still dependent on its rentals in order to maintain a healthy financial situation. The Red Cross left at the end of 1948 buta new tenant was found. Also in 1948 Miss Abbott gave up the room she rented for the kindergarten school, but it was taken immediately for the same purpose by Mrs P J Kerr who ran it until 1951. Expenditure was heavy during this period due to the installation of a new heating system, redecoration of the billiard and table-tennis rooms and repair work carried out on the guttering and the roof, which had been found to be in bad condition. At the AGM on 22 November 1951 members heard that, ‘a considerable sum more than income’ had been spent and that the income from lettings had ceased. In the interest of economy several newspapers and periodicals were discontinued — the Sphere, the Western Daily Press, the Daily Graphic, the Daily Telegraph, Woman’s Fournal, the Good Housekeeping Magazine, Punch and the Illustrated London News. Investments were realised to clear the bank overdraft of almost £400 but there was an urgent need to improve the Institution’s income. A licensed bar on the premises was suggested and following discussion at a Special General Meeting held in December the idea was adopted. Advice was sought from Wadworth and Co. and an estimate for the work to be undertaken obtained from Gaiger Bros. An application for a licence was submitted and as soon as this was granted work began on the fitting out of the bar at a cost of £41 15s. It was officially opened on 22 March 1952 by the President, W Ireland, with representatives from other clubs in the town attending. The Annual Report that year stated that the bar had turned out a great success. Many new members had been admitted and membership was continuing to grow. A telephone had been installed for the members’ use and the following year, on the occasion of the Coronation of Queen Elizabeth II, a television receiver was provided which was greatly appreciated by the members, their families and friends. Exchange visits with other social clubs were arranged, Anstie’s Social Club and the Sergeants Mess of the Wiltshire Regiment, for example. The Social Club at Messrs R Rendell and Sons was invited for an evening of Darts, Snooker and Table-Tennis — though their members were reminded that no ladies were allowed in the bar! However in 1955, with just 305 92 fully paid up members, this rule was modified to allow admittance of wives or lady friends into the Bar or Billiard room on Saturday and Sunday evenings in order to encourage the members to use the club more frequently and to attract new ones. It was also decided to provide a shove halfpenny board. But by November it was reported that the Institution had an overdraft amounting to £513. Economies were put in place but, because of the high interest rate payable on the overdraft, the following year the club was forced to sell more of its securities to liquidate it. Over the next few years the financial situation gradually improved following the introduction of raffles, tombola sessions, fruit machines in the bar area, the renting out of the table tennis room to other groups and the letting of rooms to tenants such as the WCC Education Dept., the Conservative Association and the Football Supporters Club, so that by 1962 there was a modest credit balance at the bank. By September 1965 this had grown to £2,200 and the Committee felt confident enough to consider making improvements to the forecourt, the entrance hall and the bar. A skittle alley was installed on the top floor, formally opened on 21 March 1966 by Major Bartholomew of Wadworth and Co., who had given the project financial assistance. This, too, raised income for the Institute when it was hired out to local groups and clubs. With this new financial stability it was decided in 1974 to install gas-fired central heating. As wellas skittles, snooker, table-tennis, cribbage, and darts continued to be popular. Social activities widened to include outings to race meetings at Ascot and Goodwood, to Weymouth and London. Social evenings were organised including one held in the Town Hall in 1976 at Christmas with a band and buffet for about 160 persons. This was followed by a children’s party held in the New Year. Both these popular events were held for many years before ceasing about 1997. Also in 1976 several members expressed a wish that a holiday saving scheme be organised and 23 members joined up. Earlier that summer the Committee had voted to proceed with extensive alterations to the premises. The Institute had never looked in better health and when the Chairman Mr J H Jennings died two months later his successor W W Whitelegg, in paying tribute, said that, ‘the success which the Institution enjoys today, its strong financial position and its friendly atmosphere were due in large measure to the endeavours of Mr Jennings’. The work, undertaken by builders Downing Rudman, commenced in 1977. The existing stairway was removed and new 306 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE - access made to the first floor; a fire escape was installed; a boiler room and a beer store created; the bar resited and a ladies toilet installed. The total cost was £15,292, a considerable investment but one that began to show a profitable return almost immediately. At the Annual Meeting the following year the accounts showed a net profit for the year of £7,805 compared with £3,689 the previous year, including ‘a spectacular increase in amusement machine receipts from £8,981 to £11,697’. Throughout the 1990s the annual membership was about 355 reducing to around 300 in the early years of 2000. Despite such healthy numbers, however, social activities had to be curtailed for a time because of lack of support and to avoid potential financial losses. The most pressing problem was the lack of members willing to serve on the Committee; in 2003 they numbered just four! Happily in 2008, the committee is up to full strength and the Institution is flourishing financially and socially. Though changed in form and despite several fluctuations in its fortunes, after 175 years the Literary and Scientific Institution continues to serve Devizes. Appendix 1: lectures delivered 1833-1834 1833 29 Aug. Mr Burt The History and Principals of Architecture 5 Sept. Rev J. Mayo (Sec.) The Rise and Progress of Horticulture 12 Rev R. Elliot The Use of the Globes 19 Rev I. Allen The Nature and Properties of Water 26 Rev C. Lucas The Antiquities of Avebury 3 Oct. Capt. Tayler The Compass and the General Management of a Man-of-War 10 Mr H. Butcher Geology 17 The President Inaugural Lecture 24 Rev I. Allen The Nature and Properties of Water (second lecture) 31 Mr Burt Architecture (second lecture) 7 Nov. Rev R. Elliot On the Use of the Globes (second lecture) 14 Rev I. Allen The Nature and Properties of Water (third lecture) 21 Rev G. Majendie The History, Antiquities and Statistics of the County of Wilts. 28 Col. Hewett The Winds 5 Dec. Dr. Tomkins Animal Mechanics 12 Mr Paul Anstie Origin and Progress of Mechanical Knowledge 19 Mr.H. Butcher Geology (second lecture) 26 Dr. Brabant Chemistry 1834 2Jan. Dr. Tomkins Animal Mechanics (second lecture) 9 _ Dr. Brabant Chemistry (second lecture) 16 Dr. Brabant Chemistry (third lecture) 23 Rev I. Allen Nature and Properties of the Air of the Atmosphere 30 Mr Salmon Uses, Application and Pleasure of the Study of Natural History 6 Feb. Capt. Tayler The Ocean. 13 Mr Trinder The Nature and Properties of Animal Life 20 Mr Burt Ecclesiastical Architecture (third lecture) 27 Mr H. Butcher Geology (third lecture) 6 Mar. Mr Mackay Analogy of Attractive Forces 13 Mr Mackay Electricity Di Dr Tomkins Animal Mechanics (third lecture) 10 Apl. Rev I. Allen Air of the Atmosphere (second lecture) 17 Rev I. Allen Air of the Atmosphere (third lecture) 22 Col. Hewett Poor Laws 1 May Col. Hewett Poor Laws (second lecture) Readings 1833 26 Aug. Capt. Tayler Nature and Properties of Sound 16 Sept. Mr Fowler Recovery of Persons Apparently Drowned 30 Capt. Tayler Gravitation 11 Oct. Mr H. Knight (jnr) Aurora Borealis 28 Mr Fowler Nature and Effects of Poison 4 Nov. Rev C. Lucas The Almanack 2Dec. MrJ.SloperA Passage from The Beauties of Wiltshire, the Geology of the County 16 Mr H. Butcher Franklin’s Electrical Discovertes-from The Library of Useful Knowledge 30 Rev Mr Lucas Antiquities, Customs and Pastimes of Christmas Day 1834 7 Jan. Rev Mr Lucas Explanation of the Perpetual Almanack 13 The Sec. A Passage from Readings in Science 10 Mar. Col. Hewett A Plan of Moving Houses as adopted in America 17 Col. Hewett The Corporal’s Narrative 31 Mr H. Knight On Dreams Appendix 2: the museum of the Devizes Literary and Scientific Institution: items on show at the 1840 Exhibition Collection of Cape birds, white hare, small saw-fish, head of a walrus, globe Diodon, slow worm, sea-horse DEVIZES LITERARY AND SCIENTIFIC INSTITUTION etc, two guanas, four heads of the hornbill, trunk fish, nest of Vespa Britannica, five-rayed starfish, horn of the narwhal, touracoo, etc., two Indian arrows, Indian crossbow, palm leaf, model of the main mast of a 74-gun ship, calabash, two ancient French cuirasses, crusader’s sword, Indian bow, Indian paddle, string of an Indian bow, ship’s log (Massey’s Patent), model of an Indian canoe, Indian bow and arrows, set of Indian arrows, shawl and cap of South Sea island chief, fossil of Sil’cified (calcified?) wood — Isle of Portland, slab of Ammonites communis, pair of Chinese slippers, ancient English shoes, dies of the Devizes Institution medal, impressions from same, madrepora, Mauritius, Madrepora (2), air-pump, safety lamp, facsimile of an ancient Greek inscription, plan of a Roman tessellated pavement at Whatley, Somerset, perpetual almanack. Subsequent additions to the Museum made by 1854 Model of St. Mary’s Church made by Mr J. Presto, piece of oak from a peat bog nr. Newbury, pair of shark’s jaws, toadfish and a mantis, minerals and fossils, a bust of Scott, two pieces of coral, set of fossils bought from the Yorkshire Natural History Society, specimen of cryolite, three cases of birds —a heron, a Canada goose and a kingfisher, skin of a serpent, framed painting of The Shipwreck Mariners Asylum on the Goodwin Sands. Appendix 3: lectures delivered 1850-1851 1850 10 Oct. Rev R.Elliott History and Objects of Mechanic Institutions 14 Oct. Mrs Balfour The Study of Biography 24 Oct. ditto 7 Nov. Rev Henry Jolly Don Ouixott 14 Nov. John Sheppard Alexander the Great 21 Nov. Major Byng Hall The Great Exhibition Town Hall 29 Nov. Mr Bolter Ventriloguism 1851 16 Jan. J. Player The Progress of Ancient Discovery 30 Jan. P Klitz and Son Library and Material Source 13 Feb. Rev C. Stanford Egypt and the Pharaohs (1) 27 Feb. R. W. Biggs LLD The Laws of England 20 Mar. Rev C. Stanford Egypt and the Pharaohs (i1) 307 References Primary Sources Held at the Wiltshire Archaeological and Natural History Society Library : Rule Book 1833 First Annual Report Catalogue of the Devizes and Wilts. Exhibition 1840. Devizes Literary and Scientific Institution. Words of Glees sung at the amateur Glee Concert. Sept. 25 1848 Descriptive Catalogue of some of the Scientific Apparatus at the Devizes Literary and Scientific Institution Soiree. Feb. 13 1879 Catalogue of Books forming the Library 1902 Material deposited by The Devizes Literary and Scientific Institution at the Wiltshire and Swindon History Centre (WSA 3678): Proceedings of the Literary and Scientific Institution, Devizes, Oct.1847 — Oct. 1854. Minute Book, 4 Nov. 1926 — 11 July 1935 Minute Book, 12 Sept.1935 — 5 Oct. 1950 Minute Book, 2 Nov. 1950 — 5 Mar 1962 Minute Book, 4 April 1962 — 2 Jan. 1967 Minute Book, 7 Dec. 1970 — 4 June 1973 Minute Book, 2 July 1973 — 1 Nov. 1976 Register of Members, 1851 — 1879. References 1 Devizes) and Wilts.) G(azette) 11July 1833 Kelly Thos. A History of Adult Education from the Middle Ages to the Twentieth Century. (p112) 3. D.and WG. 25 July 1833 4 Walts.) T(racts) 27.2; D.and WG. 15 Aug. 1833. 5 D.and WG. 3; 10 October 1833 6 7 8 N D.and WG. 12 Sept. 1833 D.and W.G. 13 Feb 1834 WT. CXXXI, no. 21; 9 D.and WG. 9 Oct. 1834 10 D.and WG. 19 Feb; 21 Feb; and 5 March1835 11 D.and WG. 21 Jan 1836. 12 D.and WG. 4 Feb.1836 13 D.and WG. 10 Mar.1836 14 D.and WG. 23 July 1840 15 D.and WG. 17 May1832; 5 July 1832. 16 Pugh R.B.Chartism in Somerset and Wilts. WT. 158.10. 17 D.and WG. 17 Feb.1842. In 1834 William Oriel was a committee member of the Institution . 18 D.and WG. 13 July 1837 19 D.and WG. 8 Feb. 1838 20 D.and WG. 31 Oct.1839 21 D.and WG. 23 July 1840 22 Catalogue of the Devizes and Wilts. Exhibition 1840. 23 Report on the State of Literary, Scientific and Mechanics THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE © Institutions in England. 1841 Wilts(hire) Indep(endent). 3 Nov.1842 Wilts. Indep.10 Nov. 1842 D.and WG. 14 Jan.1843 Gillman’s Directory 1859; Wilts Indep.10 Aug.1843. Wilts. Indep.26 Oct.1843 Wilts. Indep.21 Nov.1843 D.and W.G. Advertisement 28 Oct. 1847 Minute Book 27 Dec.1847 ibid 7 Jan. 1848 ibid 31 Jan.1848 ibid 14 Feb. 1848 WT. 104.15 Pugh R.B. Chartism in Somerset and Wilts. WT: 158.10; D.and W.G. 28 Mar.1839; 4 Apr.1839; O(xford) D(ictionary) of N(ational) B(tography). Hudson J.W. The History of Adult Education 1851 London. D and WG. 15 Nov. 1866. Annual Meeting 1853 Oct. 6 Gilman’s Devizes Public Register and Almanack, 1858. Register of Members 1851 — 1879. D. and WG. 19 Nov. 1863. D .and WG. 13 Aug; 24 Sept; 15 Oct; 19 Nov; 26 Nov. 1863. Devizes Advertiser Oct. 31, 1872 Wilts. Archaeological and Natural History Magazine. X1V 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 pl23. Wilts. Indep. Nov. 5, 1874 D. and W.G. 30 Nov; 7 Dec.1876 D. and W.G. Oct. 3, 1878 W.T.104.11 Cuttings Vol.10 p94. WT. 68.16 D.and WG. 31O0ct.1889; The Government School of Science and Art was based in the Assize Court from 1867 to 1906. Report of Annual Meeting 1908 D.and W.G.27 Dec; 3 Jan 1907 ibid 14 Feb 1907 ibid 21 Feb.1907 Foundation meeting ibid. 16 Jan. 1908 ibid 20; 27 Feb.1908 ibid 22 Oct. 1908 ibid 5 Nov 1908 : ibid 12 Nov. 26 Nov. 1908; 14 Jan.11 Feb. 11 Mar. 25 Mar. 8 Apr. 1909 Wilts Gazette Nov.15 1914;Nov.11 1915; Nov.8 1917. ibid 19 Oct. 1933 ibid 6 Dec.1934 Wiltshire Archaeological & Natural History Magazine, vol. 103 (2010), pp. 309-26 Notes and Shorter Contributions Early Bronze Age Burial Remains from Boscombe Down Airfield, Amesbury, Wiltshire by Andrew Manning, Jacqueline I. McKinley and Alistair Barclay A watching brief during the excavation of a new fire hydrant at Boscombe Down Airfield, Amesbury (centred on NGR 417855 140200; Figure 1) uncovered a small number of undated but probably prehistoric features and a burial sealed by a flint cairn, subsequently radiocarbon dated to the Early Bronze Age. The Airfield lies within a particularly rich archaeological landscape. The Wiltshire Sites and Monuments Record contains over 60 entries for the Airfield alone, including an Iron Age and early Romano-British settlement and associated cemetery, prehistoric and later field systems, boundaries and barrow sites. Intensive archaeological investigation in recent years has confirmed a similar picture to the west of the Airfield (Figure 1) (Wessex Archaeology 2008; Rawlings and Fitzpatrick 1996). - The new, 3.7km, fire hydrant looped around the western runway. The narrow 0.80m wide pipe trench contained a small number of poorly dated, frequently disturbed, archaeological features in the western section. A shallow, oval grave (feature 40; 1.13m by 0.62m and 0.23m deep) was badly disturbed by the trench digging in the southern section (Figure 1). It contained the remains of an adult male, 35-45 years of age, below a cairn of large flint nodules. The only finds from the grave fill were four possibly later Neolithic—Early Bronze Age flint waste flakes. A radiocarbon date of 1750-1610 cal BC (NZA 28700, 3379+ 30 BP) was obtained from a sample of human femur shaft. Other identified features included shallow ditches (8, 31, 64, 67 and 90), all 1-2m wide and 0.3—0.9m deep. A sixth substantial ditch (38), 4m wide and 1.6m deep, appears to be a previously recorded Wessex Linear Ditch that apparently survived as an earthwork until at least 1918 (Colin Kirby pers. comm.). All were aligned north-east to south-west with the exception of east-west aligned ditch 8 (Figure 1). In addition, two shallow flat-based pits (5 and 10), approximately 1m in diameter and 0.3m in depth were found. Two Early Bronze Age pottery sherds, possibly from a Collared Urn, were recovered with later finds from a tree throw (25). The human remains from feature 40 Bone from four contexts was subject to analysis; all derived from the inhumation grave. The bone is in poor condition being moderately-heavily eroded and root marked (grades 3-4), with the loss of much of the trabecular bone. The remains are also heavily fragmented. Approximately 45% of the skeleton Wessex Archaeology, Portway House, Old Sarum Park, Salisbury SP4 6EB 310 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 000914 QOOLLY 000814 New Barn Barrow Group 442000. ~~ Earl’s Farm Down Barrow Group ~ . Newton Barrow Group 141000 ag SE pe OG. ee | Amesbury Archer—-O | Boscombe Down 440000 ute Boeke . eee ep mentee: PES ESL AS eC MA SETAE Uo RE CS 1S RI" Site of Early e Bronze Age SS burial Digital Map Dale © (2004) xy 2 Digna Map Company 4200000 | | w Archaeological feature on hydrant line i Pit Circle ® Barrow © Prehistoric burial(s) Collared Urn associated roundhouse 0 | 1km _— $s ———| ; e Windon ebrsiol © @ Chippenham a Devizes Bain «. ® . Trowbridge Amesbury @ Salisbury ®@ Reproduced from the Ordnance Survey 1:25000 Explorer ® map (2007) with the permission of the controller of Her Majesty's Stationery Office. © Crown copyright all rights reserved. Ordnance Survey Licence number 100028190 R CO nt Soe GOERS > SORA DION & Fig. 1 Site location and sites mentioned in the text NOTES AND SHORTER CONTRIBUTIONS was recovered, most from the in situ deposit, but small proportions of elements from all skeletal areas were recovered from the spoil. Although it is clear that some of the bone was not recovered, some fragmentation and bone loss — particularly of the skull — had occurred prior to its recent discovery, probably as a consequence of the shallow surviving depth of the grave and the subsequent use of the land for arable agriculture. The overlying cairn of flint nodules did not protect the bone from damage/ disturbance, and its weight may have added to the pressure on the bone. The degree of erosion to the bone was recorded using the writer’s system of grading (McKinley 2004, fig. 7.1-7). Age was assessed from the stage of skeletal development (Scheuer and Black 2000), and the patterns and degree of age-related changes to the bone (Buikstra and Ubelaker 1994). Sex was ascertained from the sexually dimorphic traits of the skeleton (Bass 1987; Buikstra and Ubelaker 1994). Measurements were taken where possible (Brothwell and Zakrzewski 2004), but the data was insufficient to enable the calculation of many skeletal indices (Bass 1987). Non-metric traits were recorded in accordance with Berry and Berry (1967) and Finnegan (1978). The remains represent those of an adult male, c. 35-45 years of age at death. It was not possible to calculate any skeletal indices for this individual other than the platymeric index of the right femur (demonstrating the degree of anterior-posterior flattening of the proximal femur: 77.2; platymeric range); however, the bones do indicate a relatively large, robust individual. No teeth or parts of the supportive structure were recovered. With the exception of slight destructive lesions in the surface of the Ist sacral body, no degenerative lesions were observed in any of the spinal (five lumbar and min. five lower thoracic) or 18 extra-spinal joints recorded. Slight bony callusing and enthesophytes were observed in three left metacarpal shafts, suggestive of some form of trauma resulting in hyperflexion of the hand (a fall or a blow to the palmar surface). Fine- grained active periosteal new bone was observed in the visceral surface of several rib shafts (only left side recovered), in places overlying lamellar (healed) new bone. The lesions are indicative of a recurring chronic lung infection. Slight fine-grained new bone was also recorded in the right superior-lateral ilium, adjacent to the crest; probably related to an infection in the overlying soft tissue (the bone is close to the | surface at this point). 311 Discussion The discovery of Early Bronze Age burial remains is unsurprising given the presence of several barrow cemeteries in the vicinity of the Airfield and other burials and features recently found in the wider Boscombe Down area (Figure 1). The grave lay at the eastern edge of a group of at least 15 barrows, probably an extended dispersed linear cemetery (Clarke and Kirby 2003). Unfortunately, many of these barrows were destroyed without detailed archaeological investigation and existing evidence points to a broad Late Neolithic and Bronze Age temporal range. Two other major barrow Groups (Earls Farm Down and New Barn Down) are located to the north of the Airfield. Recent archaeological work on Boscombe Down has revealed a substantial Late Neolithic—Early Bronze Age mortuary landscape. To date, the remains of 36 prehistoric inhumation burials and cremation- related deposits have been found. These include both single and multiple burials, found within or close to barrow sites and associated features. Among them are the Amesbury Archer’ and “Boscombe Bowmen’. Most were accompanied by Beakers, with a relatively small group of burials with Food Vessels/Collared Urns dating towards the end of the Early Bronze Age (2150-1600 cal BC). The grave is therefore broadly contemporaneous with the Wessex 2 grave series (Needham 1996, 2005) as well as two roundhouses associated with Collared Urn sherds found close to the south-western edge of the airfield (Wessex Archaeology 2008). Acknowledgements The project was funded by Earth Tech Engineering Ltd. Wessex Archaeology wish to acknowledge the help and support of Ian Phones and Graham Allan of Earth Tech, Colin Kirby of QinetiQ and Helena Cave Penney of Wiltshire County Archaeological Service. The project was managed by Andrew Manning, with fieldwork undertaken by David Godden, James Wright and Neil Fitzpatrick, and post-excavation work by Lorraine Mepham and Jacqueline McKinley. Figure 1 was prepared by Rob Goller. The project archive is currently held by Wessex Archaeology at Old Sarum Park, Salisbury, Wiltshire under the project code 62260 and in due course will be deposited with the Salisbury and South Wiltshire Museum. 312 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE — References BASS, W. M., 1987, Human Osteology. Columbia: Missouri Archaeological Society BERRY, A.C. and BERRY, R.J. 1967. Epigenetic variation in the human cranium. Journal of Anatomy 101:2, 261-379 BROTHWELL, D. and ZAKRZEWSKI, S., 2004, ‘Metric and non-metric studies of archaeological human remains’, in M. Brickley and J.I. McKinley (eds), Guidelines to the Standards for Recording Human Remains, 24-30. Bradford: British Association for Biological Anthropology and Osteoarchaeology/ Institute for Field Archaeology BUIKSTRA, J.E. and UBELAKER, D.H.., 1994, Standards for Data Collection from Human Skeletal Remains. Arkansas: Archaeological Survey Research Series 44 CLARKE, B. and KIRBY, C. 2003. A newly discovered round barrow and proposed dispersed linear cemetery at Boscombe West. WANHM 96, 215-18 FINNEGAN, M._1978. Non-metric variations of the infracranial skeleton. Journal of Anatomy 125:1, 23- 37 McKINLEY, J.I. 2004, ‘Compiling a skeletal inventory: disarticulated and co-mingled remains’, in M. Brickley and J.I. McKinley (eds), Guidelines to the Standards for Recording Human Remains, 13-16. Bradford: British Association for Biological Anthropology and Osteoarchaeology/Institute for Field Archaeology NEEDHAM, S.P, 1996, ‘Chronology and periodisation in the British Bronze Age’, in K. Randsborg (ed.), Absolute Chronology: archaeological Europe 2500- SOOBC, 121-40. Acta Archaeologica 67 NEEDHAM, S.P. 2005. Transforming Beaker culture in north-west Europe; process of fusion and fission. Proceedings of the Prehistoric Society 71, 171-217 RAWLINGS, M. and FITZPATRICK, A.P 1996. Prehistoric sites and a Romano-British settlement at Butterfield Down, Amesbury. WANHM 89, 1-43 SCHEUER, L. and BLACK, S., 2000, Developmental Juvenile Osteology. London: Academic Press WESSEX ARCHAEOLOGY, 2008, Boscombe Down Areas A—D, Amesbury, Wiltshire: (Including other minor works), Post-excavation Assessment Report; unpublished client report 65531.01 Cricklade without any Britons by Richard Coates Andrew Breeze, in this journal (WANHM 101, 315- 7), dismisses previous attempts to argue for a British Celtic or Brittonic origin for the name of Cricklade (NGR SU 1093). He is surely right to do this, and I will not repeat his arguments, but his own solution, in which he suggests a new Brittonic source for the Crick- part, must also be rejected. I recant my own previous very tentative suggestion, made following earlier literature (e.g. Thomson 1976), that it might be Brittonic (Coates and Breeze 2000, 340). Breeze suggests that the name contains Old English geldad ‘river-crossing’, in a compound with a form of the ancestor of Welsh crych ‘wrinkled; rippling’, used as a name for a stretch of water (presumably in the Thames). This account does not stand up. There is no phonetic reason to believe that the British sound /X*/ in that word, or its product, Brittonic /x/ (now spelt in Welsh), would be replaced by the Old English sound /k/, spelt or , when the Anglo-Saxons assumed control of this area. Contrary to Breeze’s claim, Old English did have a ‘guttural’ spirant /x/, represented by the letter - it was not glottal at the end of a word (Hogg (1992, 34) - and this is what appears in Old English luh ‘pool’, copied from the Brittonic ancestor of Welsh Jlwch. If his claim is therefore rejected, it cuts the ground from under the idea that the present name of Cricklade is a partial Anglicization of a Brittonic one involving the ancestor of crych. The second element is undoubtedly Old English gelad ‘river-crossing’ (Gover et al. 1939, 42). The recorded Old English manuscript spellings of the first element are contradictory: Crecca, Creocc, Creacc, Creca, Cric, Crac, Croc, Crec, Creca, Crec, and these, and spellings found on coins, are thoroughly discussed by Carroll and Parsons (2007, 108-12). Interpreting this range is difficult. appears only in two medieval (12th- and 13th-century) copies of Anglo-Saxon manuscripts, and could be ignored if it were not for the predominance of University of the West of England, Bristol, BS16 1QY ee — NOTES AND SHORTER CONTRIBUTIONS on coins of the period between about 980 and 1044. All the other forms suggest a front vowel, with the majority implying /e/, whether short or long. It is hard to know what to make of the variation between and . Carroll and Parsons cut the knot as regards the vowel by suggesting that the place was known alternatively by the name *Croc, interpreted as ‘the [sharp] bend’, and by the dative-locative case- form of the same word, *Créc ‘at the [sharp] bend’, the bend being either one in the Thames, or, more likely, the diversion of Ermin Street to approach the site of Cricklade’s Alfredian fortification. This is a neat solution, but there is a difficulty: from the (ultimately dominant) dative-locative form, one would expect the pronunciation /t|/, eventually spelt , as opposed to /k/, eventually spelt , for the consonant before the <1>, but of this there is no unambiguous trace, as Carroll and Parsons acknowledge (2007, 111). Granting the general plausibility of their solution, it is worth also pursuing another. The earliest attempt at explaining the name, as Breeze reminds us, involved an appeal to the Old English word Crécas ‘Greeks’, and led to a 15th-century belief that a university had been founded at Cricklade by disoriented Greek philosophers. This idea has since been dismissed with derision, but in a roundabout way it can be shown to have a possibly correct kernel. In German, there is a word Krieche, meaning the bullace, a species of plum, Prunus domestica ssp. institia (LL) Poiret (see Kluge 1967, 404; cf. also Dutch kriek ‘black cherry’ or in earlier texts apparently “bullace’, van Wijk 1949, 348). It is recorded in early medieval times in the Old High German compound noun kriahboum, criehboum or chriehpaum meaning “Krieche” tree’. This OHG word corresponds precisely to the singular form of Old English Crécas; the bullace was understood to be the Greek plum, and is said to have appeared as prunum grecum in an early glossary by Trochus (1517); but cf. also Meinecke (1998, 116-7) for the possibility that this designation ‘classicized’, i.e. misrepresented the sound of, some ordinary German word. A diminutive of the word appears in the name of steirisches Kriacherl [traditional Styrian] plum brandy’, from Austria. There is no record of a counterpart of this term being in use in Anglo-Saxon England, and if there were we would suspect it to be a term used by the literate (monks) and not by ordinary people. But light comes from a direction with surprising implications. The word bullace was probably copied from French (cf. Old French beloce) in the Middle Ages, though its ultimate origin is not known for certain. 313 We do not know what the tree or fruit were called in Old English. But there are words in widely separated English dialects which are very suggestive. In Essex, the fruits were known as crickseys, and in Cumberland the tree was known as crex (Grigson 1975, 176; see also Thiriold 1851, 451, who gives crex as meaning ‘white bullace’ in Cambridgeshire, Gepp 1969, 35 on the Essex word, and cracks and crex in Wright 1898). These are ancient-looking designations which have no obvious etymology unless one relates them to the German and Dutch words just mentioned. They provide fairly direct evidence that in Old English bullaces were, or could be, known as Crécas ‘Greeks’ (or perhaps alternatively Creccas, if the short vowels in the dialect words are anything to go by, and that might account for the consonantal variation in the ancient spellings). The original plural may have been taken in the tree-name as a singular, “a creeks”, as it were, and provided with a new plural ‘creekses’ for the fruits. Bullaces were known as christians in Somerset and crystals in Devon (Grigson 1975); it may not be too far-fetched to guess that these are rationalizations of the name which appears in Essex dialect. However it got to England, the word seems to have been adopted in the long run by ordinary speakers of the language. In the light of Meinecke’s comment on the German counterpart, we should recognize the possibility that the word involved was not actually the one meaning ‘Greek(s)’, but an original English one identical or nearly-identical in sound and meaning ‘bullace’. Armed with this information, we can suggest that the name of Cricklade is fully English, Créca-gelad (perhaps Crecca-gelad, for reasons that are not clear), and meant ‘river-crossing where bullaces grow’. If this is correct, it provides evidence that bullaces were known in England in the first millennium, but does not settle the question of whether they are native or introduced, nor the question of whether they were wild at Cricklade or cultivated by the clerics of the minster church for their well-known medicinal properties. It must be acknowledged that this solution does not address the question of the frequent in Anglo-Saxon spellings of the name on coins. It seems rather a cop-out to attempt simply to play down the importance of coin-spellings, but it is not hard to come up with persistent but apparently non-accidental anomalies in the coin-record, such as the inexplicable absence of everything before the first in five coins from Hertford with otherwise different legends (1017x1023; Carroll and Parsons 2007, 166), or the numerous coins from Canterbury 314 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE > inscribed which have no counterpart in the manuscript record (Carroll and Parsons 2007, 76-8). These considerations might be taken as undermining the significance of the in the coin-spellings of Cricklade. Given the various delicate and unresolved matters raised in this note, it is probable that the last word has not been said about Cricklade. But it is most unlikely to be even partly Brittonic. References BREEZE, A. 2008. Cricklade and the Britons. VANHM 101, 315-317 CARROLL, J. and PARSONS, D.N. , 2007, Anglo-Saxon mint names, vol. 1. Nottingham: English Place-Name Society COATES, R. and BREEZE, A. with HOROVITZ, D., 2000, Celtic voices, English places. Stamford: Shaun Tyas GEPB E., 1969, Essex dialect dictionary. East Ardsley: S.R. Publishers [originally London: Routledge & Sons (1923)] GOVER, J.E.B., Mawer, A. and Stenton, EM., 1939, Celts and the Wylye by Andrew Breeze Wiltshire, Wilton, and the village of Wylye (NGR SU 0037) are all ultimately named after the river Wylye, which rises near Kingstone Deverill and flows 22 miles east-south-east to join the Nadder near Wilton. Although the Wylye’s name is accepted as Celtic, its meaning has been unclear. Hence this note, giving a new answer to an old question; one so old that it was asked in the 1660s by John Aubrey (1626-97), Wiltshire squire and antiquarian, when he gave first place to the Wylye in a description of the county’s rivers: I shall begin with the river of Wyley-bourn, which gives name to Wilton, the shire town. The mappe-makers write it Wyley fluvius, and joyne a British and a Saxon word together; but that is a received [standard] errour. I doe believe that The place-names of Wiltshire. Cambridge: Cambridge University Press GRIGSON, G., 1975, The Englishman’s flora. St Albans and London: Hart-Davis, MacGibbon [original text of 1958] HOGG, R., 1992, A grammar of Old English, vol. 1. Oxford: Blackwell KLUGE, EF, 1963, Etymologisches Worterbuch der deutschen Sprache, 19th edn by Walther Mitzka. Berlin: Walter de Gruyter MEINECKE, E. 1998, ‘Fruchtbaume’, in J. Hoops, H. Beck, H. Steuer, and D. Timpe, (eds), Reallexikon der germanischen Altertumskunde, 115-21, 2nd edn. Berlin: Walter de Gruyter THIRIOLD, C. 1851. Crex, the white bullace. Notes and Queries series 1-3 [84 of the continuous series], 451 THOMSON, T-R. 1976. The name Cricklade. Cricklade Historical Society Bulletin 2:1, 9-11 TROCHUS, B., 1517, Vocabulorum rerum promptuarium (4 vols). Leipzig van WIJK, N., (ed.), 1949, Franck’s etymologisch woordenboek der nederlandsche taal, 2nd edn. The Hague: Martinus Nijhoff WRIGHT, J., 1898, English dialect dictionary, vol. 1. Oxford: Oxford University Press the ancient and true name was Wy, as the river Wy in Herefordshire, which signifies ‘vaga’ [= wanderer]: and so this river Wye, which is fed with the Deverill springs, in its meandrous winding, watering the meadows, gives the name to the village called Wyley, as also Wilton (Wyleg-ton); where, meeting with the upper Avon and the river Adder, it runnes to Downton and Fordingbridge, visiting the New Forest, and disembogues into the sea at Christ Church in Hantshire (Aubrey 1847, 27). Aubrey’s comment on the Wye (called ‘Gwy’ by Welsh-speakers) alerts one to disarray amongst modern place-name scholars. English ones explain it as ‘water’, citing ancient European roots (Watts 2004, 706). But Welsh scholars take it as ‘winding University of Navarre, 31080 Pamplona, Spain 25) RE ees Geet ee eens ene nas Se NOTES AND SHORTER CONTRIBUTIONS one’, comparing the twisting Gower Peninsula of Glamorgan (Owen and Morgan 2007, 498). This note sides with them and Aubrey himself by explaining Celtic toponyms in England from ones in Wales. The Swedish philologist Ekwall did likewise. Noting that the Wylye was recorded as Wileo in 688, Guilou by Asser (d. 909), and Wilig in 901 and 943, he took it as a namesake of the Gwili in Carmarthenshire, explaining both from Welsh gwil, which ‘seems to mean “trick” or the like’. He concluded that the name meant ‘tricky river, 1.e., one liable to floods or the like’ (Ekwall 1936, 515). Appearing in an Oxford reference book, this opinion was naturally influential. Jackson refined Ekwall’s phonology, though remaining silent on his proposed meaning. He derived the form from (unrecorded) British Uilisa, with the initial as a light semi-vowel ‘w’ and the remaining vowels all long (Jackson 1953, 341-2). Others repeated Ekwall’s explanation. Wilton was thus ‘farm on the river Wylye, i.e., tricky stream’ (Field 1980, 189). Still less reliable is a popular dictionary (Room 1988, 395). “The river’s own name is of Celtic origin and comes from a word that is ultimately related to English “guile” and “wily”. This means “tricky” and refers to the river’s habit of flooding unpredictably.’ More cautious is the explanation (Mills 1991, 374) ‘possibly meaning “tricky stream’, 1.e., one liable to flood’. Yet the new Cambridge dictionary now rejects all this (Watts 2004, 706). It describes the Celtic element Uil as of obscure origin and meaning, though also known from the rivers Gwili of Wales. It goes on (more dubiously) to see the Wylye as namesake of the rivers Weil of Germany and Vilia (which flows by Vilnius, called after it) of Belarus and Lithuania. But no place for wily or tricky rivers of any kind. There is no early Welsh word gwil ‘tricky’. Ekwall was perhaps misled by gwil ‘mare, filly’, unattested before the fourteenth century, and apparently derived from English wild (Geiriadur 1950-2002, 1661). However, the Cambridge volume shows itself unaware of Welsh views on the rivers Gwili. There are four of these streams, three in Carmarthenshire and one in Ceredigion. The first are found near Abergwili (SN 4321), in the region (SN 5707) east of Llanelli, and near Llan-crwys (SN 6245). The Ceredigion one is near Braenog (SN 5357). Gwylain is the former name of a stream, now the Carrog (SH 4657) or ‘torrent’, between Snowdonia and Caernarfon Bay. ‘Two of these five rivers are known early. That at -Abergwili, near Carmarthen, is recorded in the twelfth century as Guili; the other, near the border S15 with Glamorgan, appears in the twelfth-century Book of Llandaff as Camguili ‘bending Gwili’. The form is related to Middle Welsh gwyl ‘modest; kind; joyous’, with original stress on ‘w’, but later shifting. A link with the Wylye of Wiltshire has been ruled out (Thomas 1938, 147-8). Yet we shall challenge that. More recently, Abergwili near Carmarthen has been explained as ‘mouth of (the river) Gwili’, itself understood from Middle Welsh gwyl as ‘kind one, generous one’, with the original falling diphthong becoming a rising diphthong and finally the mere ‘i’ of the modern form (Owen and Morgan 2007, 8). Though we shall modify it below, an explanation of the rivers Gwili of Dyfed as ‘kind’ or ‘generous’ helps with the Wylye of Wiltshire, where it is clear that early Wileo or Wilig (in Aubrey’s phrase) ‘joyne a British and a Saxon word together’, the final element (variously spelt) being Old English ea ‘river’. The problem is in the phonology of the first element. It has also not been understood how early Welsh gwyl relates to its supposed Old Irish cognate fial ‘modest, noble; generous, welcoming’. This leads to some technical discussion. Let us seek an analogy. Middle Welsh gwydd ‘face’ and Old Irish fiad ‘presence’ are derived from (unattested) Common Celtic weidos, from an Indo- European root meaning ‘see’ (also found in Greek eidos ‘form’ or eidolon ‘idol, false god’, and Latin video ‘I see’). This implies that early Welsh gwyl and Irish fial similarly go back to a (reconstructed) Common Celtic root uei-. But how would that give the Wil- of Anglo-Saxon documents? The answer is indicated by Jackson. At an early date the diphthong ei become a long close e in all the Celtic languages, eventually breaking to give Welsh wy and Irish Za. Yet the early West Saxon dialect of Old English had no long close e. In Celtic place-names borrowed at that date it was hence substituted by long 1, as in Old English Zemis or Thames from Late British Tamesa (Jackson 1953, 331). This allows us to link Middle Welsh gwyl with Wileo, Wilig, and so on. The 6th-century English settlers of the Wiltshire area would hear the first element of the hydronym as Uel-, with long e. Because early West Saxon had no long e, they pronounced the form as Wil-, with longi. Hence (nearly fifteen centuries later) modern ‘Wylye’, rhyming with ‘highly’. As for Guilou in Bishop Asser’s life of King Alfred, this shows the normal sound development in Old Welsh of initial w- to gw-, which occurred in the 8th century, and so a hundred years before Asser was writing (Jackson 1953, 390) and long after the English settlement of Wiltshire. On the basis of the 316 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE © above, we can thus propose the original British name. It would not be the Uzlisa put forward by Jackson, but Uelouia, which would give Asser’s Guilou. When the West Saxons heard it in the 6th century, they substituted long: for a Late British long e, and also English ea ‘river’ for the original Celtic termination. Hence their Wileo, the river Wylye of Wilton and Wiltshire. This analysis rules out totally the sense of ‘tricky’ or ‘winder’ or ‘flooder’, where reference books have repeated error. Nor is it is credible that the Wylye has the same name as the Weil of Germany or the Vilia of Vilnius in Lithuania. Yet we must also disagree with the meaning ‘kind one, generous one’ proposed by Welsh scholars for the rivers Gwili. This is not the earliest sense of gwyl, which originally meant ‘modest, unassuming, unobtrusive’, and later ‘mild, tender, gentle, gracious’ or ‘glad, joyous’. The last is shown by a 10th-century gloss guilat ‘merry, cheerful’ in an Oxford manuscript on Latin hilaris (Geiriadur 1950-2002, 1760). Yet it was not the oldest meaning. The point is proved by the Old Irish cognate fial ‘decorous, becoming, seemly; modest, chaste; well-bred, honourable, noble (of persons); generous, hospitable’, where the original sense is clearly the one shared by Welsh and Irish. In the light of this, the Wylye would have been a stream that the Britons thought of as modest, unassuming, unobtrusive, seemly. We can go further. The Celts regarded rivers as divine, as proved by the Brent ‘exalted one’ of Middlesex or the Dee ‘goddess’ of Chester or Aberdeenshire (Coates and Breeze 2000, 144, 268). So the nymph of the Wylye, whom the Britons no doubt worshipped, would have had a chaste and somewhat reserved kind of divinity, which shunned attention. She thereby contrasted with other rivers. In the north-west, the Mama ‘mother’ (now the Medlock ‘meadow stream’) that gave her name to Mamucium or Manchester was clearly seen as fertile and a bringer of fertility. If we are correct in seeing Adiunus ‘one who is desired’ as the old name of the stream by the fort of Portus Adiuni (corrupted as ‘Portus Adurni’) or Portchester, this southern stream was also seen in terms of love. Still more certain is the Sark ‘lover’ of the Anglo-Scottish border, where the sense 1s shown by Middle Breton serch ‘concubine’, and even by Old Cornish les serchoc ‘loving herb’ or burdock, a plant with burs that stick or cling to one, like a fond sweetheart (Geiriadur 1950-2002, 3228; Breeze 2004a, 2004b, 2008). The Wylye rises near Warminster, passes through a long chalk valley, and meets the Nadder by the lawns of Wilton House. It is not a violent or tumultuous stream. A name meaning ‘modest one’, even ‘chaste one’, would thus suit it, either because its course is quiet, or (more probably) because in British times it was surrounded by woods and so was glimpsed rarely, or a combination of the two. The sense ‘modest, unobtrusive’ is strongly implied by the Welsh streams. Not all of them have quiet or placid currents, yet they all flow along narrow dells or valleys, still often thick with trees, as with much of the Gwili of Abergwili, with woods that conceal or protect her modesty. If palaeobotanists can confirm that in pre-Roman times the Wylye ran through a wooded valley (long since cleared for meadowland), this hypothesis would be confirmed. So we may conclude by taking the Wylye as the namesake of the rivers Gwili in Wales, where the forms relate to Middle Welsh gwyl and (more distantly) Old Irish fial, and the meaning is ‘modest, assuming, unobtrusive’, very probably because the course of these rivers was hidden by woods. We can thus feel more certain of how the local Britons thought of the Wylye. They regarded it in a personalizing way, which was also tempered with awe, since the Celts thought rivers were divine. If a Romano-British shrine to the Uelouia or Wylye were ever discovered, it will be a spot where men and women revered one who was a modest and even chaste nymph, not easily seen, and not a fertile mother or (still less) an alluring demi-goddess known for wildness and fickleness. References AUBREY, J., 1847, The Natural History of Wiltshire. London: Wiltshire Topographical Society BREEZE, A. C. 2004a. Manchester’s ancient name. Antiquaries Journal 84, 353-7 BREEZE, A. C. 2004b. Portus Adurni and Portchester, Hampshire. Studia Celtica 38, 180-3 BREEZE, A. C. 2008. Brittonic place-names from south-west Scotland, part 8: Sark. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society 82, 49-50 COATES, R. A., and BREEZE, A. C., 2000, Celtic voices, English places. Stamford: Paul Watkins EKWALL, E., 1936, The concise Oxford dictionary of English place-names. Oxford: Clarendon Press Geiriadur Prifysgol Cymru, 1950-2002. Caerdydd: Gwasg Prifysgol Cymru JACKSON, K. H., 1953, Language and history in early Britain. Edinburgh: Edinburgh University Press MILLS, A. D., 1991, A dictionary of English place-names. | | | | | | | | NOTES AND SHORTER CONTRIBUTIONS Oxford: Oxford University Press OWEN, H. W, and MORGAN, R., 2007, Dictionary of the place-names of Wales. Llandysul: Gwasg Gomer ROOM, A., 1988, Dictionary of place-names in the British Isles. London: Bloomsbury Cunetio (1960) Hoard by Sam Moorhead and fohn Kent’ This small hoard of late 3rd to mid-4th century coins was found in the excavations conducted by Ken Annable at Cunetio (Black Field, Mildenhall, SU/216695), a few miles east of Marlborough, in 1960.! The hoard was located about a foot above the surface of large limestone paving at the west gate of the fortified town. Mr Annable sent the coins to the late John Kent at the British Museum who sent them back in two batches on February 2nd and 3rd, 1961. Kent identified the coins on the envelopes, although I have revised and modified various descriptions. There are 32 coins, 3 fragments which probably come from different coins, and 3 fragments of copper alloy which are probably not coins. The pieces range from two radiate pieces of c. 269-85’ to poor quality small module FEL , TEMP REPARATIO ‘falling horseman’ copies of the period AD355 to the early AD360s. Because there are no Valentinianic pieces, _ Table 1: Breakdown of the coins c. 269-85 30-41 GLORIA EXERCITVS@2 35-41 GLORIA EXERCITVS( 30-41 GLORIA EXERCITVS(or2 30-40 VRBS ROMA 30-40 CONSTANTINOPOLIS 37-41 PAX PVBLICA S @) 4 ‘e) a pb ley) Ss) Oo > < @) (@) ® Z Z ° Us nr na D S wlwlaela|< BENG @ |) =I410 =. | eg BS Ie Fle ig | f|0 aie Z S\Z > < Sls ey) a o| > tT ie) Uo Mn nN on oe fragments (one or more coins non-coin fragments Total Uo ]uo]uo Tuo [uo [uoTuo BR ~ co & WIN] |N] | — Oe) — °“uU Wile | Pm (Ca feat Ilo 2] i) “~uU 317 THOMAS, R. J., 1938, Enwau afonydd a nentydd Cymru. Caerdydd: Gwasg Prifysgol Cymru WATTS, V. E. (ed.), 2004, The Cambridge dictionary of English place-names. Cambridge: Cambridge University Press coins which have been found in large numbers at Cunetio, it is probably safe to date the deposit of this hoard to pre-AD364.? The monetary reforms of AD348 resulted in the demonetisation of earlier bronze coins. However, a few regular and several barbarous issues do survive, as evinced by this hoard. The reformed coinage was short-lived, being apparently withdrawn by decree in AD354, leaving only the small module FEL TEMP REPARATIO issues in circulation.* This resulted in a shortage of small change, thus precipitating the massive output of FEL TEMP REPARATIO copies of the later AD350s and early AD360s, of which there are at least four in this hoard (nos 23-6). This cache does represent the dire nature of small change prior to the large output of bronze coinage under the House of Valentinian (AD 364-78). arbarous | Uncertain | Total 4 Department of Portable Antiquities and Treasure, The British Museum, Great Russell Street, London WC1B 3DG 318 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Catalogue Gallic Empire Radiate, uncertain emperor, c.269-74 (1) No. | Reverse |Obverse Obverse |Ref: | Ref: Oty bust RIC |Cunetio [fsneinenen(Z Ff Barbarous Radiates, c.275-85 (1) bust RIC |Cunetio Zo WNOls ee ees unclear |bearded right House of Constantine (26) GLORIA EXERCITVS, 2 soldiers and 2 standards, 330- 35 (1) (No. |Mint / Mint-mark |Ruler___——_—[Ref: RIC__|Qry| 3. |Uncertain; //? Uncertain; 1 probably a son of Constantine Barbarous GLORIA EXERCITVS, 2 soldiers and 2 standards, 330-35 (1) ea ears ooraee RIC VI 4.7 |“Lyons”; -//PLG |Constantine I |c.f. 237 GLORIA EXERCITVS, 2 soldiers and 1 standard, 337- 4] (2) VII II as Augustus ee ae as Augustus Barbarous GLORIA EXERCITVS, 2 soldiers and 1 standard, 337-41 (1) [No. |Mint / Mint-mark |Ruler __|Ref: RIC VIT|Oty | on lypecas eon on lee GLORIA EXERCITVS, 2 Soldiers and 1 or 2 standards, 330-41 No. |Mint / Mint-mark Ref: RIC Oty | [827 Wacentatn. ||| a WUmcerain st [semen Inns VRBS ROMA, wolf and twins, 330-37 (2) No. |Mint / Mint-mark Ref: RIC VIT\Oty | (Qe0) Ievonsse))/ Sele Gill eaien en nt |247 mc meemeel om PCONST “Trier”; TRS \O Barbarous VRBS ROMA, wolf and twins, 330-40 (1) No. _|Mint / Mint-mark | Ruler Ref: Oty | Tey eee OE ee NT | Barbarous CONSTANTINOPOLIS, Victory on prow, 330-40 [No. |Mint / Mint-mark |Ruler ___—*([Ref: RIC __|Oty| 115.2 |‘Trier’; //TR Se ear PAX PVBLICA, Pax standing left, 337-40 VII S](branch ?PAX PVBLICA, Pax standing left, 337-40 mark 07.4] Uncertain; ?//25| ?Mlelena sacle VICTORIAE DD AVGGQ NN, 2 Victories, 347-8 mark 19. |D//TRS 20. _|Trier; (?)//TRS 182ff ?)// stantius II Barbarous VICTORIAE DD NN AVG ET CAE, 2 Victories with wreath on column, 2c. 355-early 360s No. |Mint / Mint-mark |Ruler =| Ref: RIC VIII |Oty | 22.'°|‘Lyons’; //.(R)LG| Magnentius |c.f. Lyons 121 Barbarous FEL TEMP REPARATIO, soldier advancing left, spearing fallen horseman, c. 355 — early 360s Mint / Mint- mark lr ia totype; //... iad Sno ee prototype; - ?Barbarous FEL TEMP REPARATIO, soldier advancing left, spearing fallen horseman, c. 355 — early 360s Uncertain proto- |??Constantius II’ type; - ?’Constantius IP’ Uncertain proto- type; - NOTES AND SHORTER CONTRIBUTIONS Uncertain (6) Barbarous 4th century, probably c. 330-early 360s [No. |Mint | Mint-mark |Ruler__|Ref: RIC__|Qty oa a 292 Barbarous 3rd or 4th century Now Mint / Mint-mark |Ruler _|Ref: RIC Oty | De ae 3 Fragments, probably 4th century No. Mint | Mint-mark |Ruler___|Ref: RIC__|Qty| 32.3 eee eae SR (ev Illegible, 3rd or 4th century No. |Mint / Mint-mark |Ruler ——_—s*[ Ref: RIC Oty | ES a eae See 33 24/9 3 copper alloy fragments, probably not coins ie Mint / Mint-mark Ref: RIC Oty | 243] ee Ee ee Endnotes 1 ‘Excavation and Fieldwork in Wiltshire’, WANHM 58 (Sept 1961), p. 35; for more information about Cunetio, see M. Corney, ‘The Origins and Development of the ‘Small Town’ of Cunetio, Mildenhall, Wiltshire’, Britannia XXVIII, pp. 337-350 2 Note that the largest hoard of radiate coins (54,951) from Britain was found at Cunetio in 1978, see E. Besly & R. Bland, The Cunetio Treasure (British Museum Press, 1983) 3 There were 41 bronze coins (out of a total of 80 to 101) of the House of Valentinian in the Cunetio Well hoard, see T. S. N. Moorhead, ‘A Reappraisal of the Roman coins found in J. W. Brooke’s excavation of a late- Roman well at Cunetio (Mildenhall), 19127 WANHM 90 (1997), pp. 42-54; 109 of the 871 coins recorded 24 25 319 as site-finds from Cunetio belong to the House of Valentinian, see T. S. N. Moorhead, ‘Roman coin finds from Wiltshire’, in P Ellis (ed.), Roman Wiltshire and after (Wiltshire Archaeological and Natural History Society), pp. 88 & 100 Codex Theodosianus 1X.23.1, see J. B C. Kent, Roman Imperial Coinage VIII (Spink), pp. 64-5 Fragment, c. 6 by 16mm; probably Victorinus or the Tetrici; possibly barbarous Diameter 11mm Obv leg. [ ]-TINVS AVG (as opposed to TINVS MAX AVG); diameter 13mm Diameter 14/15mm Fragment, 4-7mm; reverse leg ITV and right-hand soldier visible Only half the coin survives Kent suggested ‘Lyons’ but mm very unclear; diameter 12/13mm Diameter 11mm The likelihood is that this coin was struck at Trier Kent suggested that this might be a PAX PVBLICA issue; it is very unclear; only a fragment survives Only about half the coin survives Obverse: D N MAGN-ENTIVS; (V)/MT/X; diameter 15mm 22 diameter 13mm; 23 diameter 13/14mm 24 diameter 8mm; 25 diameter 5mm; these coins are sometimes referred to as ‘minims’ Reverse appears to have soldier left; 10/11mm Diademed bust clear on obverse, but reverse very unclear; diameter 6/8mm which suggests it is a falling horseman copy Diameter 10/11mm; obverse legend [ |AVG 29 diameter 7/9mm; 30 diameter 4/5mm; the likelihood is that these two pieces are 4th century, probably c. 355-early 360s These fragments probably come from barbarous 4th- century coins, again possibly from c. 355-early 360s Fragment, diameter (11)/16mm Two folded pieces, c. 6/1lmm and one fragment c.4/7mm; they appear too thin to have been coins shield inscribed diameter 320 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE: Late medieval/early post-medieval burial remains from Boscombe Down Airfield, Amesbury, Wiltshire by Facqueline I. McKinley and Andrew Manning A watching brief during the construction of a new accommodation block and associated services in early 2008 at Boscombe Down Airfield, Amesbury (NGR 418000 140300; Figure 1) uncovered a section of a Wessex Linear Ditch and a shallow burial subsequently dated by radiocarbon to the late medieval/early post-medieval period. The remains of the inhumation burial (Figure 1) were discovered during topsoil stripping for a new soakaway. The grave (10012) lay 0.35-0.45m below the current ground level and was largely cut into subsoil, with the base of the grave just clipping the top of the natural chalk. Only the legs, in a probable supine position, were recorded in situ, in a roughly north-south alignment; the rest of the remains were recovered from the machine-excavated spoil. There was no evidence for a coffin and no finds were recovered from the grave other than a small sherd of residual Romano-British pottery. In view of the uncertain dating, a fragment of tibia was submitted for radiocarbon dating and returned the unexpected result of cal. AD 1460-1640 (2 sig. 94.5% NZA- 30656; 438+20 BP). The only other identified feature comprised a section of a substantial ditch (12003), 4.46m wide and 0.70m in depth, running approximately north-west to south-east. This feature appears to be a part of a previously recorded Wessex Linear Ditch system (Wiltshire SMR 749), which has been traced for at least 2km to the northwest, at least as far as the prehistoric settlement at Southmill Hill (Wessex Archaeology 2005). The Human Remains The degree of erosion to the bone was recorded using the writer’s system of grading (McKinley 2004, fig. 7.1-7). Age was assessed from the stage of skeletal development (Scheuer and Black 2000), and the patterns and degree of age-related changes to the bone (Buikstra and Ubelaker 1994). Sex was ascertained from the sexually dimorphic traits of the skeleton (Bass 1987; Buikstra and Ubelaker 1994). Full details are held in the archive. The circumstances of the grave’s discovery resulted in fragmentation of the bone and incomplete recovery of the skeletal remains (c. 65%). Although elements from all skeletal areas excepting the left upper limb were collected, other than a few of the foot bones none of the elements is complete. Most of the breaks to the bone are fresh, some showing no join, and it is likely that the missing elements were lost in excavation. The one possible exception could be the left upper limb. The grave lies in an area of downland — with a long history of use as sheep pasture, but it is possible that it may have been subject to ploughing at some stage in the 19th and/or 20th century (VCH Wiltshire XV 13). Given the very shallow depth of the grave such activity could have preferentially damaged/removed all or parts of the left upper limb, especially if it lay slightly higher that the rest of the body. In this respect it may be significant that parts of the left facial bones are also missing. The bone is in relatively good condition with some surface root etching/erosion (grade 2-3). The remains represent those of a young (c. 23-5 yr) adult male. No stature estimate could be made but the individual was clearly of a large and robust build. There is no indication of dental hypoplasia (stress-related lesions in the tooth enamel) and other possible stress indicators are limited to slight lateral bowing in the proximal portion of one tibia which could represent residual evidence of mild childhood rickets. Increased vascularity in the endocranial surface of the distal parietals and upper occipital suggests that the individual suffered from head-lice, the scratching associated with which increases the blood supply to the affected area (Capasso and Di Wessex Archaeology, Portway House, Old Sarum Park, Salisbury SP4 6EB NOTES AND SHORTER CONTRIBUTIONS 321 Reproduced from the Ordnance Survey 1:25000 Explorer @® map (2007) with the permission of the controller of Her Majesty’s Stationery Office. © Crown copyright all rights reserved. Ordnance Survey Licence number 100028190 Amesbury : | , 442000 exe Z Re a a a Amesbury @ rN The Site) 7S — Grave 10012 141000 000917 OOOLLY = big bo Ditch 12003 ~ Ss: oe It ms | : . sof ->=x~ Parish boundary ° ee Historic road wee WVessex linear (from excavation) «»«« Wessex linear (Wiltshire SMR % 749, from cropmark) 0 1km | "rs er ——| Figure 1 Site location and sites mentioned in the text Tota 1998). There is slight dental calculus (calcified plaque) on all the surviving teeth. The individual had suffered an abscess in the right mandibular molar socket, infection from which had spread to the surrounding buccal supportive structure but had healed by the time of death. A small defect in the superior dorsal margin of one of the three cervical vertebrae recovered suggests an acute trauma or possibly repetitive strain to the posterior common ligament in this area of the spine. Further evidence for trauma is represented by a well-healed depressed fracture, evident as an oval depression (27x12mm, c. 2mm deep) in the right parietal, adjacent to and almost parallel within the Sagittal suture 48mm dorsal to the bregma. It is possible such an injury could have been sustained as a result of a heavy fall on to a hard object, but it is more likely to have been caused by a deliberate blow with a blunt instrument by an assailant. Discussion The airfield and surrounding area is known to contain a substantial number of archaeological features and monuments, including a significant number of burial remains, recovered from barrow sites and Iron Age and Romano-British cemeteries. Recent small-scale improvements to the airfield continue to add new elements to this important archaeological landscape (Manning et al. forthcoming). It is worth noting that the location and singularity of this grave, together with burial position, orientation and paucity of grave goods, would, without radiocarbon dating of the bone, have led to the assumption that the remains were those of a Romano-British individual buried, as they frequently were in rural settings in marginal/boundary locations. This case thus further emphasises the importance of obtaining dates for otherwise undated burial remains. The grave lies on one of the highest points of a downland plateau to the south-west of Amesbury and in close proximity to the boundary between the parishes of Amesbury and Boscombe. The nearest known routeway (Idmiston to Amesbury) dates to at least the 17th century (VCH, Wiltshire XV, 16) 322 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE: and originally lay c. 750m to the south-west, with the cross-roads between Amesbury, Salisbury, Allington and the London road c. 1.25km to the north-east (Figure 1). There is no evidence for medieval or post-medieval buildings in the immediate vicinity, the only known land-use within this temporal range being that of sheep pasture. Non-churchyard burials of the late medieval/ early post-medieval date are rare. There was a general horror of not being buried in consecrated ground; few people were deliberately excluded and, even then, attitudes varied across the temporal range and, sometimes, on the inclination of the individual ecclesiastical authority (Cherryson et al. in prep.; Daniell 1997; 104-9; Gilchrist and Sloane 2005, 73). Those debarred from churchyard burial could include individuals killed in a tournament or duel, heretics, witches, excommunicants and those convicted of murder or treason. In some cases specific areas within the churchyard might be set aside for such burials and sometimes specific cemeteries were established or accepted places of deposition adopted, as was the case for unbaptised infants and criminals (guilty of lesser crimes; Cherryson et al. in prep.; Daniell 1997, 104-9; Gilchrist and Sloane 2005, 71-7). Those guilty of specific acts were often treated in a prescribed fashion and disposed of in specific places, such as the burial of suicides and witches at cross-roads; though some of the former are known to have been interred in open fields (MacDonald and Murphy 1990, 48; and in prep.). Victims of circumstance might also be buried in unconsecrated ground; the numbers dying in battle or as a result of epidemics or famine might become so great as to be unsupportable in a normal parish churchyard or be too far distant from one (Cherryson et al. in prep.; Gilchrist and Sloane 2005, 71-7). Those drowned at sea might be buried close to the shoreline; those suffering accidental death (presumably unattended) were sometimes buried where the body was found; while murderers usually concealed their victims in a range of locations (Cherryson et al. in prep.; Daniell 1997, 108; Gilchrist and Sloane 2005, 72-3; Pollard 1999; Reynolds 2009, 47-9). The location and singularity of the burial from Boscombe Down exclude this individual from many of the categories discussed. The most likely interpretations for the young man from Boscombe Down include murder, singular loss in a small martial skirmish or accidental death. The shallow depth of the grave, age and sex render him suitable for any of these categories. There is no peri-mortem trauma to aid interpretation, but such does not always leave skeletal traces and may have been present in the missing c. 35% of the skeleton missing in analysis. Two other situations might also be considered. Tudor England ‘lived in terror of the tramp’ (Bindoff 1966, 293-4). Extensive enclosure for sheep pasture, particularly in the first half of the Tudor period, and widespread hardship resulting from years of poor harvest led to the surplus population, previously employed in arable agriculture roaming ‘the countryside in search of food and work’ (ibid. 22-3). Such marginalised vagabonds, often travelling in variously sized groups, would not necessarily always have held to the roads and would have had to move outside their parish of origin. The parochial system at this time entitled those living within the parish to be buried within the parochial churchyard, a right not extended to strangers who might be banished to a marginal location if included and might be denied even that if someone did not come forward to pay the necessary fees (Cherryson et al. in prep.). Were one of such a number to die from want or illness, his fellows might have buried him wherever they lay before moving on. A further group of excluded individuals were the gypsies, who first appear in official English records in the early 16th century (http://www.bbc.co.uk/kent/voices/romany_roots. shtml). They would also have fallen foul of the parochial system and are known to have a tradition of ‘wayside’ burial (Dawson 2000). Although evidence from the 18th—20th centuries suggests they tended to bury near traditional stopping grounds often close to routeways, they also chose what were perceived as permanent landmarks as suitable spots. The Boscombe Down burial at lay on the parish boundary, at a high point and within view of a series of Bronze Age barrows; such may have been considered a sufficiently appropriate landmark. Acknowledgements The project was funded by Debut Services (South West) Limited - contractors to Defence Estates - and Wessex Archaeology wish to acknowledge the help and support of Warren Hurst and Darrin Kelly of Debut Services, James Read of Morgan Ashurst, Bob Clarke of QinetiQ and Helena Cave Penney of Wiltshire County Archaeological Service, who monitored the project. The authors are also i NOTES AND SHORTER CONTRIBUTIONS grateful to John Chandler and particularly Annia Cherryson and Sarah Tarlow, for advice and access to forthcoming publication material. The project was managed by Andrew Manning, with fieldwork undertaken by Daniel Tarrant and Simon Reames and post-excavation work was by Lorraine Mepham, Rachael Seager Smith and Jacqueline McKinley. The illustration was prepared by Rob Goller. The project archive is currently held by Wessex Archaeology under the project codes 66980-3. In due course the archive will be deposited with the Salisbury and South Wiltshire Museum References BASS, W. M., 1987, Human Osteology. Columbia: Missouri Archaeological Society BINDOFFE S.T., 1966, Tudor England. Harmondsworth: Pelican BUIKSTRA, J.E. and UBELAKER, D.H., 1994, Standards for Data Collection from Human Skeletal Remains. Arkansas: Archaeological Survey Research Series 44 CAPASSO, L. and DI TOTA, G., 1998. Lece [sic] buried under the ashes of Herculaneum. The Lance, 351 (9107), 992 DANIELL, C., 1997, Death and Burial in Medieval England. 323 London: Routledge DAWSON, R., 2000, Gypsy Wayside Burials. Privately published GILCHRIST, R. and SLOANE, B., 2005, Requiem: the Medieval Monastic Cemetery in Britain. London: Museum of London Archaeological Services MACDONALD, M. and MURPHY, T.R., 1990, Sleepless Souls: Suicide in Early Modern England. Oxford: Oxford University Press MANNING, A., McKINLEY, J.I. and BARCLAY, A., forthcoming. Early Bronze Age Burial Remains from Boscombe Down Airfield, Amesbury, Wiltshire, WANHM McKINLEY, J.I., 2004, ‘Compiling a skeletal inventory: disarticulated and co-mingled remains’, in M. Brickley and J.I. McKinley (eds), Guidelines to the Standards for Recording Human Remains, 13-16, Bradford: British Association for Biological Anthropology and Osteoarchaeology/Institute for Field Archaeology POLLARD, T., 1999, ‘The drowned and the saved: archaeological perspectives on the sea as grave’, in J. Downes and T. Pollard (eds), The Loved Body’s Corruption, 30-51. Glasgow: Cruithne Press REYNOLDS, A. 2009. Anglo-Saxon Deviant Burial Customs. Oxford: Oxford University Press SCHEUER, L. and BLACK, S., 2000, Developmental Fuvenile Osteology, London: Academic Press VICTORIA COUNTY HISTORY (VCH), Wiltshire, XV, Amesbury, 13-21 WESSEX ARCHAEOLOGY, 2005. Boscombe Down Phase V Excavations, 1995; Silbury Hill: Stukeley’s bridle and other finds, some reports from 1771 and 1850 by Brian Edwards This note provides details of two hitherto overlooked accounts of artefacts associated with Silbury Hill. These were a spur reported in 1771, and a possible dagger and a fragment of copper alloy, said to have derived from an otherwise unreported vertical excavation that took place in 1850. After the death of William Stukeley (1695-1765) the equestrian bit or bridle, that he recorded as having been recovered from Silbury Hill in 1723,! was purchased by Gustavus Brander (1720-1787),’ an enthusiast in natural history and antiquities. Brander was a wealthy London merchant with renowned scientific connections and a Trustee of the British Museum, a Fellow of the Royal Society and of the Society of Antiquaries of London. He later supplied information about Stukeley’s bridle for Richard Berenger’s The History and Art of Horsemanship (1771), an unexpected addition to his description being a spur (Figure 1), also stated as having been recovered from Silbury Hill: ‘An ancient British or Roman spur (the rowel wanting) and bitt : both dug out of a large burrow, in the road from London to Bath, called Silbury Hill ; and supposed by the antiquaries, to have The Wiltshire Archaeological and Natural History Society, 41 Long Street, Devizes, Wiltshire, SN10 INS 324 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE | D,... qe ( hn be ‘ Fig. 1 The spure and Stukeley’s bridle, after the image in Richard Berenger’s The History and Art of Horsemanship (1771) been the tumulus or place of burial, of the British- king Coel whose daughter Helena was the mother of Constantine the Great.’ ? The spur could have been purchased along with the bridle and thereby mistakenly identified as having been recovered from Silbury Hill, but Brander seems unlikely to have been misguided on this point, not least as he would have been familiar with Stukeley’s account. Indeed the image of the bit in Berenger’s book appears to be based on a simplified version of Stukeley’s drawing of it in Abury, redrawn from the published account rather than the artefact itself. Stukeley could perhaps have been unaware of this spur; it could even have been recovered after his death. It is a long spur of c. 1400-1500. It is not stated if it was made of iron or copper alloy, but the former is perhaps more likely. The rowel is missing and it appears from the drawing that it might have been quite small. The sides of the spur taper forward into a right-angled bend and then turn again, ending in two figure of eight terminals. These are depicted almost certainly incorrectly on the drawing as horizontal to the sides but are more likely to have been, in fact, vertical (cf. Ellis 1995 no 351).* It could be that like Stukeley’s ‘bridle’ the spur has been redrawn in a simplified form.° A report appeared in the I/lustrated London News in June 1850, of an excavation that took place within a year of the Archaeological Institute’s horizontal tunnel having reached the centre of Silbury Hill in 1849.° ANTIQUARIAN DISCOVERIES AT SILBURY- HILL. (To the Editor of the Illustrated London News). I take this opportunity of communicating to you a discovery of British antiquities lately dug up in Silbury-hill, near Marlborough, Wiltshire. This celebrated mound was excavated a short time ago by order of the Archaeological Society of London, but at that time no remains of anything whatsoever were discovered. I have since had reason to suppose that the above learned society were wrong in merely excavating the lower part of the hill, and that if a shaft were to be sunk from the top of the mound right down to the bottom there would be a chance of some important discoveries being made. I recently made an excursion to Silbury, provided with the necessary implements, intending to prove my theory of the top being more likely to produce something than the base. We had dug for, I should think, eight feet down, when the spade of one of my assistants struck against something, which returned a ringing sound, and, on digging it out, discovered it to be that interesting specimen of a dagger (an illustration of which is appended). Taking courage at our success, we renewed our exertions, and were rewarded in a few minutes by the sight of the bit, which is made of brass, and in which the hole through which the chain of the “bulla” is passed is exemplified. The dagger is [two and one half] inches long, and the bit [three] inches and about another [two thirds] of another inch. I was prevented from communicating this most important discovery to you by particular circumstances which called me away from my friend’s house that same evening, and I left my antiquities behind me; but on my return to that part of the country about a week ago, I formed the resolution of acquainting you with the particulars, which I am now putting in practice. ARCHAEOPHILUS. NOTES AND SHORTER CONTRIBUTIONS MEAN nde chars