ee ; . fon) zs. wwitshce Archaeological a ee ee . Ural mistory Magazine Volume 99 2006 ‘ at r The Wiltshire Archaeological and Natural History Magazine Volume 99 2006 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 99 (2006) ISSN 0262 6608 © Wiltshire Archaeological and Natural History Society and authors 2006 Hon. Editor: Andrew Reynolds, BA, PhD, FSA, FSA Scot. Hon. Assistant Editor: Stuart Brookes, BA (Hons), MA, PhD, PGCTiLL. Hon. Local History Editor: James Thomas, BA, PhD, FRHistS. Hon. Natural History Editor: Michael Darby, PhD, FRES Editorial Assistant: Lorna Haycock, BA (Hons), PhD, FRHistS, Dip.ELH, Cert Ed. Editorial Assistant: Elizabeth Blackwelder We acknowledge with thanks grants towards the cost of publishing specific papers in this volume from the following bodies: Crest Nicholson (South West) Limited for ‘Early Bronze Age Ring Ditches and Romano- British agriculture at Showell Farm, Chippenham. Excavations in 1999’ by Annette Hancocks, Defence Estates (Land) SW for ‘Excavations on the Old Ditch Linear Earthwork, Breach Hill, Tilshead’ by Vaughan Birbeck and ‘A Middle to Late Bronze Age Settlement at Dunch Hill, Tidworth’ by Phil Andrews, English Heritage for ‘Iron Age and Later Defences at Malmesbury: Excavations 1998-2000’ by Tim Longman and “The Romano-British Roadside Settlement of Whitewalls, Easton Grey, Wiltshire: recent fieldwork’ by Tony Wilmott and Deborah Shipp, and Michael Heaton for ‘Were House, Warminster: Britain’s oldest purpose- built fire station?’. 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 utles 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 Seciety and authors. Cover illustration: detail from South View of the Abbots House, Malmsbury (sic), watercolour by John Buckler, 1809, in the Society’s art collection (DZSWS:1982.712) Typeset in Plantin by Stuart Brookes and produced for the Society by Salisbury Printing Co. Ltd, Greencroft Street, Salisbury SP1 1JF Printed in Great Britain Contents ARCHAEOLOGY and LOCAL HISTORY A detailed re-examination of the petrography of the Altar Stone and other non-sarsen sandstones from Stonehenge as a guide to their provenance, by Rob A. Ixer and Peter Turner Early Bronze Age Ring Ditches and Romano-British agriculture at Showell Farm, Chippenham. Excavations in 1999, by Richard Young and Annette Hancocks, with contributions by Wendy 7. Carruthers, H.E.M. Cool, G.B. Dannell, Teresa Gilmore, Lorrain Higbee, Robert Hopkins and E.R. McSloy. Illustrations by Lorna Gray, Peter Moore and Bryony Ryder A Middle to Late Bronze Age Settlement at Dunch Hill, Tidworth, by Phil Andrews, with contributions by Michael Ff. Allen, Rowena Gale, Phil Harding, Pat Hinton, Emma Loader, Facqueline I. McKinley and Pippa Smith Excavations on the Old Ditch Linear Earthwork, Breach Hill, Tilshead, by Vaughan Birbeck, with contributions from Michael 7. Allen, Rowena Gale, S. Hall, Phil Harding, Emma Loader, Lorraine Mepham and M.S. Tite Iron Age and Later Defences at Malmesbury: Excavations 1998-2000, by Tim Longman, with contributions by Lisa Brown, Rod Burchill, Paul Davies, Rowena Gale, Fulie fones, David Mullin and Naomi Sykes The Romano-British Roadside Settlement of Whitewalls, Easton Grey, Wiltshire: recent fieldwork, by Tony Wilmott and Deborah Shipp, with contributions by Richard Brickstock, Andy Hammon, Ben Moore, Sarah Fennings, Michael Stone, Forn Schuster, and Jane Timby A Romano-British Prone Burial from Bratton, Wiltshire, by Stella Maddock and Pamela Mahon Eve of Wilton and Goscelin of St. Bertin, by D. I. Stroud A Possible Site for the Hospital of St John the Baptist and St Anthony at Old Sarum, Salisbury, by Andrew B. Powell Heraldry and Hostelries: The Evidence from Wiltshire, by R. Moody Thomas Stevens (1549-1619): Wiltshireman and Jesuit, by Daphne Tighe and James H. Thomas The Duke of Somerset’s Grand Mansion at The Brails, Great Bedwyn: A Review of the Evidence, by Graham Bathe 10 5] 79 104 165 190 204 213 221 230 236 NOTES and SHORTER CONTRIBUTIONS The Duke’s Vaunt: a recently discovered painting, by foan Davies and Graham Bathe Celtic Survival at Chicklade, by Graham Breeze The Lost Dedication of the Church at Winterbourne Dauntsey, by John Lush The Romanesque Corbel-Table at St. John’s, Devizes, and its Sheela-Na-Gig, by Alex Woodcock and Theresa Oakley South Wiltshire Young Archaeologists’ Club excavation at Baverstock, by Susan Clelland and Bill Snell Were House, Warminster: Britain’s oldest purpose-built fire station?, by Michael Heaton The International Bee Research Association (IBRA) Bee Boles Register online, by Penelope Walker The Archaeology Field Group: recent activities and future plans, by Jim Gunter and Susie Stidolph OBITUARIES EXCAVATIONS and FIELDWORK in WILTSHIRE 2004 INDEX, by Philip Aslett 246 248 249 250 254 256 259 259 261 264 271 The Wiltshire Archaeological and Natural History Society The Society was founded in 1853. Its activities include the promotion of the study of archaeology (including industrial 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. 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 asummary 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, SN101NS. 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, LE, 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: Ten offprints of each article will be given free (to be shared between joint authors). Offprints are not given for notes and shorter contributions. WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY SOCIETY BOARD OF TRUSTEES (from 1 January 2006) Chairman W A Perry, MSc Deputy Chairman D L Roseaman, BSc(Eng), CEng, MIMechE Other Elected Trustees Miss A Arrowsmith, BSc (Hons) Microbiology (Lon) Lt Col C Chamberlain Ms C Conybeare, MA, FMA J A Gunter, BA (Hons) Mrs V Knowles Mrs W P Lansdown (Hon. Treasurer/Company Secretary) J SS Stewart, BSc (Oxon), MB, ChB (Glasgow), FRCS (Ed). FRCS (Glasgow), FRCS Mrs J Triggs Nominated Trustees J N Fogg (Member, Kennet District Council) A Molland (Member, Wiltshire County Council) P R Saunders, BA, FSA, FMA, FRSA (Director, Salisbury & South Wiltshire Museum) Mrs J Swabey (Member, Wiltshire County Council) Mrs K J Walling, BA (Hons), PGCE (Member, Devizes Town Council) In attendance: T Craig BA, MA, AMA, DipMgmt (Wiltshire County Council Heritage Services Manager) OFFICERS Curator Assistant Curator Sandell Librarian and Archivist Outreach Officer Volunteer Co-ordinator Finds Liaison Officer (Wiltshire) Documentation Project Manager Documentation Officer P H Robinson, PhD, FSA, AMA Miss L Webb, BA, MA Mrs L Haycock, BA (Hons), PhD, FRHistS, Dip ELH, Cert. Ed. Miss A Siviter, BA, MA Mrs H Ault, BA Miss C A Hinds, BA (Hons), MA A Martin, BA (Hons), M.Phil, PhD Ms P Luscombe, BA (Hons), DMS Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 1-9 A detailed re-examination of the petrography of the Altar Stone and other non-sarsen sandstones from Stonehenge as a guide to their provenance by Rob A. Ixer! and Peter Turner? The Altar Stone at Stonehenge has a disputed origin and sandstones from both the Devonian (Old Red Sandstone) Cosheston Group and the Senni Beds in South Wales have been proposed as possible parent lithologies. A thin section of the Altar Stone together with those of four fragments of non-sarsen sandstone excavated at Stonehenge and attributed to the Cosheston Group sandstones have been re-examined. These detailed petrographical descriptions are compared/contrasted with earlier published accounts and with new descriptions of typical lithologies from the Senni Beds of South Wales. The Altar Stone is petrographically very similar to fine- to medium-grained, calcareous sandstones found in the Senni Beds and this is consistent with the suggestion they are the best candidates for being the parent rock. The Altar Stone is, however, very different lithologically from the four fragments; comparing their petrography with that published confirms that that they are probably not Cosheston Group sediments and that their degree of deformation may indicate that they are older than the Devonian. As their origin remains unknown it is still not possible to provenance them. None of the five samples originated from the Old Red Sandstone rocks around Milford Haven. Introduction A lithologically unremarkable, grey-green, micaceous sandstone is perhaps the most famous Welsh lithic export in the world. Stone 80 (numbering after Atkinson, 1979), namely the fallen ‘Altar Stone’ from Stonehenge, at nearly Sm long, is the largest bluestone (following the convention that uses the term to denote all non-sarsen lithologies) at the site, weighs over six tonnes and is the only sedimentary rock to be visible amongst the otherwise igneous bluestones (Figure 1). The geographical, as opposed to mythical, origin of this sandstone (alongside all the other stones at Stonehenge) has been a matter of speculation since the mid-18th century and is succinctly reviewed in Thorpe et.al. (1991, 119-124). William Stukeley (1740) thought the Altar Stone to be a Derbyshire marble, Phillips in Maskelyne (1878, 151) noted it was ‘a grey sandstone composed of quartz sand, silvery mica and dark grains (possibly hornblende)’ and suggested it was Devonian (Old Red Sandstone) or Cambrian in age and Davies on the same page in Maskelyne noted the nearest similar looking sandstones cropped out at Frome in Somerset. H. H. Thomas (1923, 244-245) too, believed the Altar Stone to be Devonian in age and his is the most complete, published, macroscopic and microscopic description of the Altar Stone to date. He suggested that the Altar Stone compared well with certain green, micaceous and calcareous Old Red Sandstones in South Wales, namely the Senni Beds that crop out in a band between Kidwelly and Abergavenny or the Cosheston Beds of Pembrokeshire close to Milford Haven. Thomas highlighted the difficulty of deciding between the two formations either in the field or in the laboratory. Thomas’s work has been the basis for subsequent 'Dept. of Geology, University of Leicester, University Road, Leicester, LEl 7RH ?School of Earth Sciences, University of Birmingham, Edgbaston, Birmingham, B15 2TT bo 30 ! 29 we oe) CII SS Pay 21D “2- ee ail S 450 4 46 neous Bluestones SS; ; 32 “Sys Pn Oz 72 160A_s608 324% 6 a & : 32e en BS a Oh: g3e 2a(| on 8 4a =e oe eine (se Wy o70a 620 22 \ ee 63e 33e 4 af BO iar g/l 33f8 ea \ 53 340 9 ard Mata aN 69 : 35 Ae 1 @ 689 2 (es ge a ® 6 938 36 120 % Bre” Ory 41? 409 oy 40c 1574 5 Soy Fig. 1. Plan of the central stone settings after Atkinson (1979) and modified from Thorpe et al. (1991). Sarsens are plain and the bluestones have a textured infill. The Altar Stone close to the centre is number 80. attempts to provide a provenance for the stone with most workers favouring the Senni Beds over the Cosheston Group. Later, R.G. Thomas in Thorpe et al. (1991, 152- 153) macroscopically described the Altar Stone as highly micaceous and without a pervasive spaced cleavage and, based on his extensive knowledge of the Cosheston Beds (Thomas 1978), noted that the Senni Beds were ‘the most likely parent unit for the Altar Stone’. He cautioned, however, that a thin section description was needed before any provenance could be given. Most recently, Kellaway (2002, 59) has opined that the Altar Stone derives from the Lower Old Red Sandstone (Senni Beds) but probably where the outcrop pattern widens, namely to the east, in the Brecon Beacons. Kellaway’s Brecon origin is informed by his views that an inferred Pliocene ice- sheet transported the Altar Stone to Wiltshire. A more cautious approach to the origin of the Altar Stone is given by Clough and Cummins (1988, 156, thin section 277). Here, the Altar Stone is left unprovenanced and described as a ‘micaceous sandstone’ distinguished from four other non-sarsen sandstone fragments found in the Stonehenge area, namely “Stonehenge Cursus’ (thin section 275), at ‘Stonehenge’ (276) and from Aubrey Holes 1 and 5 (thin sections 444 and 450 respectively). These latter THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE four are described as ‘sandstone (Cosheston Beds)’ by Clough and Cummins (1988, 157). Table 1 tabulates the five sections. “Total petrography’ using polished thin sections combined with high quality geochemistry has proved to be effective in describing and assigning fine-scale provenancing to lithics (Ixer 1996; Ixer et al., 2004), including other bluestones from Stonehenge (Ixer 1996; 1997a). Total petrography alone has also proved other ‘bluestones’ not to be associated with the site (Ixer 1997b, 279; 1997c, 1-2). Sadly these techniques (combined transmitted and reflected light microscopy and XRF chemical analysis) were not possible for the present samples as all the sections have a glass cover slip over them, are too valuable to reprepare and can only be investigated in transmitted light. In transmitted light it is only possible to identify transparent or translucent minerals hence much of the potential mineralogy of a rock, most especially that of the abundant, opaque iron and/or titanium oxides, cannot be determined. These oxides display a wide range of mineralogical and petrographical complexity and are used alongside other ‘heavy minerals’, in the routine, geological provenancing of sediments and metasediments. Cunnington (1884) identified five fragments of the Altar Stone, that he assumed resulted from dressing of the stone, amongst his loose finds, although they are now missing. Obtaining geochemical data using a non-destructive, portable XRF machine placed on the Altar Stone is possible but removing in situ material for petrographical and detailed geochemical characterisation is unlikely to be permitted. Therefore, the thin section labelled ‘277 Altar Stone Stonehenge’ in the Salisbury Museum Collection is likely to remain, for the foreseeable future, the only piece of the monolith Table 1 List of thin sections as given in Stone Axe Studies 2 eae. find Type of sample Rock type TS 275 Loose Sandstone | Fine Stonehenge cursus | fragment metasandstone TS 276 Loose Sandstone | Fine Stonehenge fragment metasandstone S277 Altar Stone Samples Calcareous collected in situ? | sandstone TS444* Aubrey Loose Sandstone | Fine Hole 1 fragment metasandstone TS450 Aubrey Loose Sandstone | Fine Hole 5 *Thin section 444 is designated as OU9 in (Thomas 1991) fragment metasandstone THE PETROGRAPHY OF THE NON-SARSEN STONES FROM STONEHENGE 3 Table 2 Modal analyses of the Altar Stone based on 750 counts Column | is the modal analysis, column 2 is the analysis recalculated without the calcite matrix and showing the rock to be a subarkose/sublitharenite based on the clas- sification of Folk (1974). 48.48 Potassium feldspar Calcite 35.50 Plagioclase White mica available for investigation. It is imperative then that it should be described as fully as possible and that this description becomes widely available. To this end a large number of standard (3"x 1") thin sections of lithics collected from Stonehenge were loaned out by Dr Vin Davis on behalf of the Southwest Group of Museums and Art Galleries Implement Petrology Committee and included the five non-sarsen siliciclastics 275, 276, 277, 444 and 450. The first three thin sections were described in detail by the present authors and the last two briefly. All five were compared with representative samples of Old Red Sandstone siliciclastics from the Senni Beds near Kidwelly loaned by Dr G. Owen. Both authors independently described the Altar Stone thin section and later its mineralogical composition was determined by point counting (750 points) (Table 2). Thin section mineralogical identifications were made following Kerr (1959) and Deer, Howie and Zussman (1992) and petrographical and lithological descriptions with reference to Folk (1974) and Adams et al. (1984). The Altar Stone Macroscopically Thomas (1923, 244-245) described the Altar Stone as ‘a fine-grained, pale sage-green, micaceous sandstone with a partly calcareous and partly siliceous cement and with prominent mica along its divisional planes’ (bedding/laminae). Microscopically he identified angular quartz, mica flakes, abundant chlorite and ‘a greenish mineral that might be glauconite in a fine textured calcareo- siliceous matrix’. A heavy mineral ‘residue’ from the Altar Stone was ‘exceedingly rich in angular, pink or colourless garnet’ some grains being idiomorphic; other heavy mineral grains were less abundant but included zircon, tourmaline, rutile and anatase. This account can be compared with the present description. Detailed petrography of thin section 277 Macroscopically the thin section shows the Altar Stone to bea fine-grained, buff-pale grey sandstone. Microscopically the rock is a very fine- to fine- grained, well-sorted sandstone with a mean grain size of 0.13mm and a maximum grain size of approximately 0.32mm. The clastic grains are angular to subrounded and the modal group is subangular. The grains predominantly show a high sphericity but a number of prolate grains with a width/length ration of <0.5 are also present. The rock has a homogeneous fabric but bedding is observed by thin, less than 0.5mm wide, opaque mineral-dominated, heavy mineral laminae (Figure 2) and by the orientation of thin phyllosilicate laths, mainly white mica and chlorite, lying parallel to those laminae. There is very little detrital, interstitial material. Authigenic kaolinite is locally present but the main matrix is a pervasive carbonate cement. A number of detrital grains float in the cement and show both pristine and corroded grain boundaries. The sandstone is only moderately compacted and the carbonate cement is probably pre- or syn- compaction. Mineralogically, both visual inspection and Fig. 2. Thin, heavy mineral laminae (east-west concentrations of black grains) alternate with quartz-rich bands. Angular to subangular quartz grains (white) are enclosed within carbonate (grey and speckled). An east-west orientated prismatic zircon lies above one of the heavy mineral bands (middle right hand edge). The average grain size of the quartz grains 1s 0.1 —0.1Smm. Plane polarised transmitted light. 4 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE determination of the modal composition of the Altar Stone by point counting show quartz to be the most abundant detrital clast accompanied by lesser amounts of (in order of decreasing abundance) fine-grained rock clasts, alkali and plagioclase feldspars and phyllosilicates including white mica and a number of optically different chlorite minerals. Accessory minerals are dominated by equant, opaque mineral grains accompanied by zircon, tourmaline and rutile with lesser amounts of apatite and trace amounts of possible amphibole and garnet. Glauconite was not recognised. Modal analysis (table 2, column 2) shows the rock is a subarkose/sublitharenite. Rounded to subangular, single quartz grains showing uniform extinction are present within the carbonate matrix (figure 3). Polycrystalline quartz clasts are rare but include some with strained extinction suggesting a metamorphic origin. Untwinned or simply twinned, tabular alkali feldspar shows slight alteration; this is seen as pale brown/orange, turbid cores or as replacement by fine-grained white mica. Microcline, if present, is rare. Lath-shaped, multiple twinned, sodic plagioclase (albite?) is fresh to slightly altered. Neither of the feldspar groups, nor quartz, show authigenic overgrowths. Muscovite laths lie parallel to the bedding but many show fine-scale kinking about quartz or feldspar grains, whilst others have splayed ends. Pale Fig. 3 Typical quartz-rich area at higher magnification. Angular to subrounded, exclusively monocrystalline quartz grains (white, uniform grey and black) float. in a carbonate cement (grey and speckled). The concentric bands around many quartz grains are an optical effect. The average grain size of the quartz grains is 0.1 -0.15mm. Crossed polarised transmitted light. to dark green or brown-green chlorite laths showing a range of interference colours are widespread, lie parallel to bedding and show fine-scale kinking. Other chlorite patches are fine-grained mosaics. Although very minor amounts of biotite altering to chlorite occur, interlayered muscovite-chlorite grains are absent. Accessory minerals are concentrated in heavy mineral bands dominated by opaque iron titanium oxide minerals and by TiO, minerals including orange-brown rutile. Lesser amounts of euhedral to rounded zircon, green-brown tourmaline, apatite laths and rare, blue-green and brown amphibole and colourless ?garnet accompany the opaque minerals. Rock clasts are the same size as the mineral clasts. Most are fine-grained, siliceous “cherts” that probably include acid volcanic lavas; minor amounts of quartz-chlorite or quartz-white mica phyllite are also present. There is no primary matrix. Although fine- grained kaolinite locally infills void spaces between clasts, the main matrix/cement is an untwinned carbonate, optically identified as calcite. The absence of mixed chlorite-muscovite grains may indicate that the sandstone is sub-greenschist in facies; it is essentially unmetamorphosed and has no tectonic fabric. This in turn suggests that rock is post Caledonian Orogeny in age and hence Devonian to Carboniferous (or younger) in age. Rock clasts and sparse garnet may indicate that the source area comprised first generation acidic rocks rather than high-grade metamorphics. The absence of authigenic overgrowths on quartz and feldspar is typical of many of the finer grained rocks from the Old Red Sandstone. However, the corroded grain margins indicate that early overgrowths might have been dissolved during calcite cementation. Four Sandstone Fragments from Stonehenge and the Aubrey Holes Thin sections 275 labelled ‘rock, Stonehenge Cursus’, 276 ‘rock, Stonehenge’, 444 ‘rock, Aubrey hole 1’ and 450 ‘rock, Aubrey hole 5’ in the Salisbury Museum Collection have been identified as sandstones from the Cosheston Beds (Clough and Cummins, 1988). Detailed petrography of thin sections 275 and 276 shows them both to be clast-supported, metasandstones and very similar to each other. THE PETROGRAPHY OF THE NON-SARSEN STONES FROM STONEHENGE 5 Briefer examination of sections 444 and 450 show them to be sufficiently similar to each other and to 275 and 276 that they require no further description. Thin Section 276 was initially described by Judd (in Gowland 1902) as ‘an argillaceous flagstone’ with ‘a mass of angular quartz grains of two sizes cemented with micaceous material. Small amounts of muscovite, chlorite and feldspar are present. Scattered fragments of volcanic material also occur’. K.C. Dunham in a letter (17/10/1947) to Dr Stone stated that it is ‘an impure sandstone, colouration due to chlorite. Also muscovite, hydrobiotite and clay mica are present. Fresh oligoclase is a conspicuous minor constituent.’ These descriptions can be compared with the present account that combines the results from sections 275 and 276. Detailed petrography of thin sections 275-and 276 Microscopically rounded to elongated, monocrystalline quartz grains have syntaxial overgrowths about them; elsewhere touching quartz grains display serrated edges due to pressure solution effects. Some quartz grains have patchy extinction suggesting a metamorphic origin. Simply twinned, alkali feldspar including perthite is more abundant than polysynthetically twinned microcline; plagioclase has clean, syntaxial overgrowths on dusty cores and is partly replaced by calcite (in 276). All feldspars alter to fine-grained, white mica. Large, detrital chlorite, muscovite and rare biotite (altering to chlorite) laths are present. However, much phyllosilicate is authigenic and includes rounded, interlayered, muscovite-chlorite grains and fine- grained, white mica. This mica lies along the main planar fabric of the sandstones or forms cross-cutting veinlets lying at a high angle to bedding, suggesting a syn- or post-tectonic age; elsewhere, much fine- grained white mica occurs as selvages/fringes about quartz grains. Angular, zoned zircon, orange-brown and green tourmaline, euhedral apatite, euhedral TiO, minerals and epidote plus opaque minerals including altered pyrite (in 275) are accessory minerals. Rock clasts include polycrystalline quartz, shale and metashale, ‘chert’/acid volcanics, chlorite-rich clasts and a little graphic granite. Rhombic carbonate forms a very minor, local cement or replaces plagioclase. Petrography of Rock 444 and OU9 Section 444, was not studied in detail in the present study as preliminary investigations showed it to be very similar to 275 and 276. However, another thin section of the same rock sample has been described by R.G. Thomas (1991) and later by J. Huggett (1993) as sample OU9; their detailed petrographical descriptions are very like the present one for sections 275 and 276. They too noted that most rock fragments are ‘chert’ but include low grade metamorphics namely slates/phyllites, that garnet is almost missing from the heavy mineral suite, that abundant white mica shows preferred orientations and that there is a pervasive spaced pressure solution cleavage. Huggett classified the sample as a micaceous sublitharenite and suggested that the minor amounts of carbonate were siderite whereas Thomas (1991, figure 16.153) plots the rock in the middle of the subarkose field. All the rocks have a marked tectonic fabric and are low-grade metasediments suggesting they suffered deformation during the Caledonian (or possibly Variscan) Orogeny. Although this does not eliminate Devonian/Carboniferous rocks, it also allows that either the sandstones are older Palaeozoic sediments or that they originated from outcrops present in the more tectonically deformed very west of Wales. Authigenic overgrowths on quartz and feldspars and the growth of neomorphic phyllosilicates are important characteristic of these sandstones and indicate a complex diagenetic and post-diagenetic history. Sediments from the Senni Beds The Lower Old Red Sandstone Senni Beds dominated by grey-green sandstones with lesser amounts of siltstones, conglomerates and calcretes are up to 380m thick and lie beneath the red-brown, unfossiliferous Brownstone Group. They crop out in South Wales from west of Kidwelly to the Welsh Marches and include widespread outcrops in the Black Mountains of Powys, Gwent and Hereford and Worcester. They are the lateral equivalent of the Cosheston Beds that are found to the west in southwest Dyfed. The Senni Beds notably carry plant remains and potentially this is an important macroscopic aid to identifying them and any artefact manufactured from them. Published petrographical descriptions of the rocks are rare and the first detailed sedimentological analysis of the Senni Beds is given by Owen (1995). Thin sections of typical sandstone, siltstone and cornstone lithologies present within outcrops of the Senni Beds from a coastal exposure on the estuary of 6 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE the River Taf at Craig Ddu were investigated in order to compare them with the Stonehenge samples. The outcrops are at the western end of the Senni Beds outcrop, closest to Milford Haven. The lithologies range from micritic limestones with minor amounts of muscovite and quartz (cornstones) to carbonate- poor, matrix-supported siltstones and carbonate- poor, clast-supported, very fine-grained sandstones and to fine- to medium-grained carbonate-cemented sandstones. Characteristically, all of the lithologies lack any significant signs of a tectonic fabric. Although the overall mineralogy including the heavy minerals suite is similar, there are significant differences between the petrography of the siltstones and sandstones from Craig Ddu. Only the coarser grained sandstones have significant carbonate (>10% by volume) and long muscovite laths lying along bedding planes, whereas only the fine-grained lithologies carry significant interlayered chlorite- muscovite grains. The fine- to medium-grained carbonate- cemented sandstones, as typified by thin section 86.31, resemble the Altar Stone and so are described in detail. Macroscopically the rocks are grey, micaceous, fine- to medium-grained lithic arenites with a calcareous cement (Owen 1995, 223), but are pale grey-green in thin section. Detailed petrographical description of Senni Beds sandstone 86.31 Microscopically the rock is a fine-grained, well- sorted sandstone with a mean grain size of 0.15mm and a maximum grain size of approximately 0.30mm. The clastic grains are angular to subrounded and the modal group is subangular. The grains are predominantly equant in shape, but a number of prolate grains are also present. The rock has a homogeneous fabric but thin, planar bedding is observed by less than 0.5mm wide, phyllosilicate-rich laminae that are muscovite- and green and brown chlorite-rich. There is very little detrital, fine-grained, interstitial material/clay and the main matrix is a carbonate cement. Quartz and feldspar grains float in this cement and show both pristine and corroded grain boundaries. Mineralogically, visual inspection shows quartz to be the most abundant detrital mineral accompanied by lesser amounts (in order of decreasing abundance) of fine-grained rock clasts, alkali and plagioclase feldspars and phyllosilicates including white mica and a number of optically different chlorites. Accessory minerals are dominated by equant, opaque mineral grains accompanied by zircon, tourmaline and rutile plus very rare amounts of garnet, sphene and anatase. Angular to subangular, single quartz grains showing uniform extinction and no authigenic overgrowths are present within the carbonate matrix. Polycrystalline quartz clasts include quartzite and stretched quartz suggesting a metamorphic origin for them. Untwinned or simply twinned, tabular, alkali feldspar, including some perthite, shows slight alteration; this is seen as pale brown/orange, turbid cores. Microcline, if present, is rare. Lath- shaped, multiple twinned, sodic plagioclase is predominately fresh and shows very rare, thin, authigenic overgrowths. Long muscovite laths lie parallel to the bedding and many show fine-scale kinking about quartz or feldspar grains. Pale to dark green or brown-green chlorite laths show a range of interference colours and are widespread. They too lie parallel to bedding and show fine-scale kinking. Biotite, if present, is very rare. A single interlayered muscovite-chlorite intergrowth was identified. Accessory minerals are concentrated in poorly defined heavy mineral bands. These bands are dominated by opaque iron-titanium oxides and TiO, minerals including orange-brown rutile. Lesser amounts of rounded, zoned zircon, dark green-brown tourmaline, sphene and subhedral garnet accompany the opaque minerals. A little authigenic anatase associated with carbonate infills void spaces. Rock clasts are the same size as the mineral clasts. Many are fine-grained siliceous ‘cherts’ that include acid volcanic lavas but minor amounts of quartz- white mica phyllite, graphic granite and foliated feldspathic lava (‘trachyte’) are also present. The matrix/cement is untwinned carbonate, probably calcite. The sandstone is unmetamorphosed and lacks any tectonic fabric. It lacks significant, authigenic overgrowths on its detrital quartz and feldspar clasts or any interlayered chlorite-muscovite grains. Its rock fragments and heavy mineral suite suggest that its source area comprised eroding acid igneous rocks rather than high-grade metamorphics. Lithologically, it is clear that this rock shares many characteristics with the Altar Stone and since carbonate-cemented sandstones are common in the Senni Beds, one of them may well be the source of the Altar Stone. In addition, although the fine-grained Senni Beds lithologies (very fine sandstones and siltstones) THE PETROGRAPHY OF THE NON-SARSEN STONES FROM STONEHENGE 7 share certain mineralogical characteristics with the four Stonehenge rock fragments, the lack ofa strong tectonic fabric in any of the rocks from the Senni Beds precludes them from being the source of those fragments. Discussion This paper represents the first detailed description of the Altar Stone for over eighty years and is in broad agreement with H.H. Thomas, other than his identification of abundant garnet and glauconite. Glauconite is a green, chlorite-like mineral and so, if present, has been subsumed under chlorite in the present description. The disparity over the amount of garnet is more significant and puzzling. Thomas noted significant amounts of garnet in his ‘heavy residues’ (Thomas, 1923, 244) but did not report garnet in his thin section description of the Altar Stone. Although trace amounts of garnet can be overlooked/underestimated in thin section the present study could not confirm significant amounts of garnet microscopically. The presence and amount of garnet is important as Thomas was struck by the coincidence between the garnet-rich nature of his Altar Stone ‘heavy residues’ and the unusually garnetiferous nature of the Cosheston Beds and it was the presence of these unusual amounts of garnet in both that led him to suggest the Cosheston Group might have been the origin of the Altar Stone. Without further sampling (this would require many grammes of Altar Stone to crush before separating the heavy minerals) the garnet problem must remain unresolved. The Altar Stone shares sufficient petrographical similarities with carbonate-cemented Senni Beds fine-grained sandstones to state that the microscopic description reinforces earlier, macroscopic identifications of the Altar Stone as probably originating from the Senni Beds. The dimensions of the Altar Stone are known and it should be possible to compare them to bed thicknesses and joint spacings within suitable, Senni Beds, calcareous sandstones. Were the Altar Stone anthropogenically transported to Stonehenge then coastal exposures would be a sensible first place to search but were natural agencies the prime mover then the search area is considerably widened. The exact geographical location of the source bed, whilst it would be very interesting to know, is perhaps not as significant as knowing that the source for the Altar Stone is far from either the Preseli Hills and its environs or Milford Haven and that the rock does not appear to have any innate macro- or microscopic characteristics that mark it as special. Petrographically this study allied with previous descriptions clearly shows that the four rock samples from Stonehenge (275, 276, 444 and 450) are low-grade metasediments and have the same lithology and provenance; their common grouping in Stone Axe Studies Two is confirmed. They are quite unlike either the Altar Stone (this is again in agreement with Stone Axe Studies Two (Clough and Cummins 1988)), or any lithology from the Senni Beds at Craig Ddu. The four rocks fragments differ from the Altar Stone and carbonate-cemented, Senni Beds sandstones as the four carry significant amounts of shaly rock fragments and mixed-layered muscovite-chlorite grains; they have abundant fine- grained, white mica along spaced cleavage and lack any carbonate cement. This suggests that the four rock samples originated from a different source area (geological provenance) and have suffered divergent diagenetic and post-diagenetic histories from the Altar Stone/carbonate-cemented Senni Beds lithologies. Although the fine-grained siliciclastics at Craig Ddu share more petrographical features with the four Stonehenge rocks, notably a lack of carbonate cement and presence of mixed layered muscovite-chlorite grains, there are significant differences. In particular their detrital rock clasts and metamorphic/tectonic histories as evidenced by their phyllosilicates are quite disparate. The origin of these fragments remains contentious. Thin section 275 has been matched to British Geological Survey Slide E8148 with a suggested source from the shore of Mill Bay, Milford Haven estuary some 0.75 miles from Cosheston (pers. com. V. Davis 1996). By contrast R. G. Thomas (1991, 152-3), an authority on the Cosheston Group, compared sample 444 with a ‘typical’ sandstone from the Mill Bay Formation (Cosheston Group), taken from the Mill Bay (west) section and listed a number of ‘fundamental dissimilarities’ in both mineralogy and fabric. In particular he noted the absence of pervasive pressure solution seams and stress-related authigenic white mica but the presence of high- grade metamorphic rock fragments and a wide- ranging heavy mineral suite dominated by abundant garnet in the Cosheston Group lithologies. Further, he stated that the Senni Beds were not the source of sandstone 444 either and concluded that ‘the compositional differences between these potential source formations (Senni Beds or Cosheston Group) 8 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE and OU9 (444) are obvious’ and (OU9) ‘definitely did not derive from the Cosheston Group or Senni Beds’. Rather he suggests that OU9 might be a more tectonically deformed Devonian sandstone from westernmost Pembrokeshire (southwest Dyfed) but more likely from a Welsh Silurian or older Lower Palaeozoic sandstone formation. Based on the degree of metamorphism/tectonism shown by the four sandstone fragments detailed, thin section petrography supports this view. Modern, detailed petrography suggests that none of the five sandstone samples have an origin within the Old Red Sandstone lithologies around Milford Haven. Therefore they cannot be cited as evidence for Milford Haven being the exit port for the bluestones as suggested by Atkinson (1979, 57 and108) and repeated by Hawkins (1970, 82) in their popular and influential paperback books. Conclusions The Altar Stone is confirmed as an unmetamor- phosed, carbonate-cemented sandstone most probably from the Devonian (Old Red Sandstone) Senni Beds of South Wales. The exact geographical origin for the artefact is far from the Preseli Hills and Milford Haven. The four other non-sarsen sandstone fragments collected from Stonehenge are not from the Senni Beds and probably not from the Old Red Sandstone Cosheston Group as has been suggested. Their tectonic fabric and metamorphism suggest that they may be from an older, more deformed, Lower Palaeozoic sandstone-bearing sequence. Without a knowledge of either their geological age or formation it is not possible to provenance them further. Acknowledgements The authors are grateful to Drs Vin Davis, Geraint Owen, Jenny Huggett and Olwen Williams-Thorpe for lending material and/or allowing permission to quote from their unpublished works. Dr Vin Davis is further thanked for his many acts of kindness and academic generosity and including perspicacious comments on an earlier draft, that allowed this study to be started and finished. The Constantine Palaeologos Research Fund is acknowledged for its usual assistance. Bibliography ADAMS, A.E., MACKENZIE, W.S. and GUILFORD, C., 1984, Atlas of Sedimentary Rocks under the Microscope. Harlow, Essex: Longman Scientific and Technical ATKINSON, R.J.C. 1979, Stonehenge. Harmondsworth: Pelican Books 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 CUNNINGTON, W. 1884. Stonehenge notes: the fragments. WANHM 21, 141-49 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 FOLK, R.L., 1974, Petrology of Sedimentary Rocks. Austin, Texas: Hemphills GOWLAND, W. 1902. Recent excavations at Stonehenge. Archaeologia 58, 37 -118 HAWKINS, G:.S., 1970, Stonehenge Decoded. Glasgow: Fontana Books HUGGETT, J. 1993. Stonehenge Project. Petrology of sandstones. Unpublished consultancy report to Open University Earth Science Department. 9% November 1993 6pp 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. 1996. Ore petrography and archaeological provenance. Mineralogical Society Bulletin, 113, 17-19 IXER, R.A. 1997a, ‘Detailed provenancing of the Stonehenge Dolerites using reflected light petrography: a return to the light’. in A. Sinclair, E. Slater, and J. Gowlett, (eds). Archaeological Sciences 1995.. Oxbow Monograph 64, 11-17 IXER, R. 1997b. Steep Holm “Bluestones”. Current Archaeology 151, 279 IXER, R. 1997c. Steep Holm “Bluestones”. Steep Holm Newsletter, New Year 1997, 1-2 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, FWenban-Smith and EHealy (eds). Luithics in Action. Lithics Studies Society Occasional Paper 8. 105-115. Oxford: Oxbow Books JUDD, J.W. 1902. ‘Note on the nature and origin of the rock-fragments found in the excavations made at Stonehenge by Mr Gowland in 1901 in W Gowland. (ed) Recent excavations at Stonehenge. Archaeologia 58, 106 -118 KELLAWAY, G.A. 2002. Glacial and tectonic factors in the emplacement of the Bluestones of Salisbury Plain. The Survey of Bath and District. 17. 57-71 THE PETROGRAPHY OF THE NON-SARSEN STONES FROM STONEHENGE 9 KERR, PF, 1959, Optical Mineralogy. Third Edition. New York, London: Mcgraw-Hill MASKELYNE, N.S. 1878. Stonehenge: the petrology of its stones. WANHM 17, 147-61 OWEN, G. 1994. Senni Beds of the Devonian Old Red Sandstone, Dyfed, Wales: anatomy of a semi-arid flood plain. Sedimentary Geology, 95, 221-35 STUKELEY, W., 1740, Stonehenge a Temple restored to the British Druids. London THOMAS, H.H. 1923. The source of the stones of Stonehenge. The Antiquaries Fournal 3, 239 — 60 THOMAS, R.G. 1991, ‘Petrography and possible provenance of sandstone sample OU9 (Aubrey Hole 1) and a comment on the Altar Stone’ Appendix 2 in R.S.Thorpe, O. Williams-Thorpe D.G.Jenkins and J.S.Watson (eds) The Geological Sources and Transport of the Bluestones of Stonehenge, Wiltshire, UK. Proceedings of the Prehistoric Society, 57, 152-3 THOMAS, R.G. 1978 The Stratigraphy, Palynology and Sedimentology of the Lower Old Red Sandstone Cosheston Group, South-West Dyfed Wales. Unpublished Ph.D thesis University of Bristol 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. 99 (2006), pp. 10-50 Early Bronze Age ring ditches and Romano- British agriculture at Showell Farm, Chippenham. Excavations in 1999 by Richard Young and Annette Hancocks with contributions by Wendy F. Carruthers, H.E.M. Cool, G.B. Dannell, Teresa Gilmore, Lorrain Higbee, Robert Hopkins and E.R. McSloy. Illustrations by Lorna Gray, Peter Moore and Bryony Ryder Early Bronze Age ring ditches and a late Ist to 2nd-century AD Romano-British field system with associated agricultural activity have been discovered at Showell Farm, Chippenham. INTRODUCTION Project and archaeological background (Figure 1) From July to September 1999 Cotswold Archaeology (then Cotswold Archaeological Trust) carried out an archaeological excavation for Crest Nicholson Properties, now Nicholson Estates, at Showell Farm, Chippenham, Wiltshire (centred on NGR: ST 907 712; Figure 1). The excavation was undertaken in anticipation of construction of a new business park. An Environmental Statement which accompanied the planning application included a cultural heritage assessment comprising two elements. The first, a desk-based assessment (CAT 1998; DPDS 1998), concluded that no previously recorded archaeological sites existed within the boundary of the site, although there was evidence for archaeological activity in its immediate environs. Flint assemblages testify to Mesolithic, Neolithic, and Early Bronze Age activity just beyond the limits of the site, primarily along the route of the new Chippenham bypass (Bateman and Enright 2000). During investigations around Showell Nurseries, to the east of Showell Farm, cut features of the above dates were also identified (OAU 1991; Anon 1993). Late Neolithic/Early Bronze Age activity at Showell Nurseries included pits, postholes, ditches and gullies. One pit contained Beaker pottery. Roman activity previously identified from cropmarks took the form of trackways, boundary ditches and gullies with domestic debris including pottery and animal bone dated principally to the late 1st to 2nd century AD (Anon 1993, 160). The results of the desk-based assessment prompted the Local Planning Authority to require a field evaluation comprising 30 trial trenches to determine the presence or absence of archaeological deposits within the proposed development area (Figure 1). Many of the evaluation trenches were devoid of archaeological evidence, but in the northern part of the site linear features of Roman date, together with a Bronze Age gully, were encountered (Figure 3, trenches 20, 29 and 30). Topography and geology The proposed development site comprised c. 25ha Cotswold Archaeology, Building 11, Kemble Enterprise Park, Cirencester, Glos. GL7 6BQ SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 11 Gloucestershire LAR HAM Oxfordshire e Chippenham Wiltshire Area C Geological division Kellaway clay (north) and Cornbrash (south) Showell Nursery : OH 00 Showell Farm site evaluation trench flint scatter ?terracing ?ridge and furrow excavated feature Fig. 1. Site location plan (1:7500) 12 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE of pasture to the south of Chippenham, immediately east of the London to Bristol railway line. The northern and eastern boundaries of the site were formed by the Chippenham to Lacock road, with the south-western boundary defined by the A350 Chippenham bypass. The ground slopes very gently towards the road to the east, although the north-eastern part of the site lies on a gentle rise at approximately 54-55m above OD overlooking the river Avon to the east and one of its tributaries, the Pudding Brook, to the north. The underlying geology across the southern half of the site is mapped as Cornbrash, with Kellaways Clay to the north (BGS 1990 and Figure 1). Excavation Methodology The Environmental Statement recommended that the area of principal archaeological interest be excavated prior to development. Crest Nicholson commissioned CA to undertake the excavation whilst the planning application was being considered by West Wiltshire District Council. In the event, and following completion of the fieldwork, planning permission was refused at Public Inquiry. Fieldwork focussed on two areas of archaeological interest identified from the evaluation (Figure 1). Area A was c. 120m by 80m where a flint scatter had been identified. Area B represented the site of a probable Romano-British farmstead occupying the higher land above the river Avon. Area C lay in different ownership and had not been surveyed in the first phase of work, but was evaluated by excavation of a further four trial trenches, 50m long by 1.5m wide, to clarify the extent of archaeological deposits on the site. Within Area A, aseries of 1m? test pits was hand excavated on a 25m grid across an area measuring 120m by 80m, including both of the recognised lithic scatters and their immediate surrounds. The test pits were excavated down to the underlying limestone brash at a depth of 0.25m. A 20% sample of material from each test pit was hand sieved on site through a 10mm mesh. In Area B, topsoil, subsoil and colluvium were stripped under archaeological supervision using tracked excavators with toothless ditching buckets. The evaluation had demonstrated that archaeological features only survived where they had been cut into the surface of the natural clay. The examination of features concentrated on recovering the overall plan, stratigraphic sequence, and the collection of an adequate sample of dateable finds. Archaeological deposits were sampled by hand excavation to the following levels: burials and cremations 100%; pits, postholes and other non-linear features 50% minimum; whilst linear features such as ditches and gullies had a maximum of 10% excavated (Figure 3). EXCAVATION RESULTS (Figures 2-7) Area A No archaeological features were identified during the excavation of the 28 test-pits located in Area A. A total of 37 worked flints was recovered as a result of hand sieving topsoil excavated within the test-pits. Area B Archaeological deposits in the north-western half of Area B were covered by up to 0.5m of topsoil and colluvium and did not appear to have been significantly truncated. Features in the south- eastern half of the site lay on a slight rise covered by between 0.25m and 0.35m of topsoil and subsoil. The depth of several linear features diminished by c. 0.2m from the north-west to the south-east parts of the trench demonstrating greater truncation on the slope. Dateable features fell into two periods. Other features produced no artefactual material or had no stratigraphic relationships with dateable features, but may be related by their form and location with features that yielded artefacts. The two periods were: Period 1: Early Bronze Age (2460 BC — 2030 BC) Period 2: Early Roman (Mid Ist century to later 2nd/early 3rd century AD) Period 1: Early Bronze Age; Figures 2 and 4 Two ring ditches were identified, 63m apart. Ring ditch 1 was shallow (maximum depth 0.1m) with an internal diameter of c. 10.5m. The ditch was not perfectly circular, with its western side markedly flattened. No internal features were identified. The ditch was generally u-shaped with a narrow, concave base and a north-east facing entrance 1.3m wide, with rounded terminals. The ditch was filled with silty sand. Early Bronze Age beaker-type pottery was recovered from the primary fill of ring ditch 1 (Figure 9.1 and 9.2) as was a broken Neolithic leaf- shaped arrowhead (Figure 8.2). An environmental SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 Area C 30 | \ GQ Se 1 / Fig. 2 Plan of site showing the location of cropmarks (1:2500) 13 14 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 2014 22904 a ‘2016 Fig. 3 Feature plan for Area B and evaluation trenches in Area C (1:1250) SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 15 sample <43>, taken from the unexcavated primary fill of ring ditch 1, contained a human heel bone. A single hazelnut shell and hulled barley from the primary fill of ring ditch 1 were sent for radiocarbon dating. Ring ditch 2 was heavily truncated with an internal diameter of c. 7m. The ditch appeared to have been recut once and had a relatively steep profile with a predominantly flat base. The ditch produced no dating evidence although it had been cut by a Roman fence line. The north-eastern portion of this feature survived to a maximum depth of 0.21m, whilst to the south-west it was much shallower (0.07 and 0.04m). No internal features were identified. The colluvium that sealed the ring ditch and the fence line contained a barbed and tanged flint arrowhead (Figure 8.4) amongst post-medieval finds. Dating Evidence Two radiocarbon dates were obtained, from a fragment of hazelnut shell (Wk-15448: 2460-2140 cal BC) anda hulled barley seed (Wk-15449: 2300-2030 cal BC) from the primary fill of ring ditch 1, while two Beaker fineware sherds (Figure 9.1 and 9.2) were recovered from the ditch. Together these suggest a late 3rd millennium BC date for ring ditch 1. Two unstratified barbed and tanged arrowheads and a single small unstratified ‘thumbnail’ scraper are characteristic of the 3rd millennium BC too (Figure 8.4 and 11.5) and could be contemporary with ring ditch 1. A further three flints exhibit characteristics of Early Bronze Age date and a broken leaf-shaped arrowhead is likely to be a residual Early-Middle Neolithic item. Period 2: Early Roman; Figure 5 The Early Roman activity sub-divided into five phases (a-e) based on stratigraphic relationships, form and distribution of features, fill characteristics, ceramic evidence and the analysis of other categories of artefactual material. Phase 2a c. AD 50-140/150 Parallel gullies (1, 2 and 3) were aligned north-east to south-west. Their southern terminals were rounded whilst their northern ends lay outside the excavation area. Gullies 2 and 3 lay 11m and 14m to the north- west of gully 1 respectively. Gully 1 had been recut ~ at its southern end. Two further gullies (4 and 5) lay at right angles to this alignment, with similarly rounded south-eastern terminals. The north-western end of gully 4 was truncated and the north-western terminal of gully 5 lay outside the excavation area. The form of gully 4 would indicate that it was constructed in two phases, with its south-eastern half being an extension to the original feature. All gullies were up to 1m wide with shallow u-shaped profiles and sandy clay fills. Gully 3 was cut by gully 4. The function of these gullies is unclear, but the regular layout of nos._1 to 5 suggests that they were used as boundary markers, possibly to separate small agricultural plots. The spatial relationship of gullies 2 and 3, and the stratigraphic relationship between gullies 3 and 4, suggest two phases of construction. It is possible that gully 1 defined the eastern edge of a trackway, with gullies 2 and 3 forming successive western edges. If this is the case the south-eastern extension to gully 4 may have been dug after this trackway went out of use. Equally it might represent an alteration to the trackway, which allied with the gap in gully 3 to the north-east, defined a new right- angled course to the north-west. A number of other features have been ascribed to Phase 2a. The majority are less regular gullies, although there are several large pits and one grave. The gullies either shared a common alignment with gullies 1 to 5 or else were cut by later features. The distribution of these features suggested that the agricultural plots originally extended across much of the excavation area, but that their remains have been lost through later truncation to the south-west and south-east. The fill of pit 456 produced large quantities of charcoal, burnt bone, a few emmer/ spelt wheat grains, a chaff fragment and several weed seeds. Immediately east of ditch 5 lay a heavily truncated, crouched burial (Grave 1) of an adult, although the sex could not be determined. A radiocarbon date was obtained on a sample of the left tibia from this burial that yielded an AMS date of Wk-15450: 170 cal BC to 60 cal AD, a Late Iron Age or Early Roman date (Table 8). Phase 2b c. AD 50-140/150 A further three linear features (6, 7 and 8) are ascribed to Phase 2b because they cut Phase 2a features and had different forms and alignments. Gully 6 was aligned north/south and cut gully 4. It was up to 1m wide and may have been recut once at its mid point. Fences 7 and 8 differed from the Phase 2a gullies in that they were shorter, deeper and steeper-sided; 6m and 3.3m long respectively, with a maximum width of 0.75m and filled with sandy clay. The differing alignment of linear features 6 to 8 suggests that they were not part of the Phase 2a system of plot division. 16 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE location of beaker sherd from evaluation location of radiocarbon dating samples WK15448 and WK15449 location of ring ditch 1 beaker sherd Section A location of human heel bone <43> Section A N 54.5m |- AOD Fig. 4 Area B, phase 1, ring ditch 1, plan and representative section (1:100 and 1:10) SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 7) Dating Evidence Four residual Iron Age sherds were recovered and grave 1 was radiocarbon dated to the Late Iron Age/Early Roman period (170 cal BC to 60 cal AD) suggesting an element of Iron Age activity in the vicinity. For the Romano-British phases of the site a series of key groups has been defined on the basis of their stratigraphic integrity, size and level of completeness. Dating each phase of activity has been determined by reference to pottery finds, but incorporating information from other artefact classes where appropriate. The overall date-span of the assemblage is narrow. Phase 2a/b Dating this phase to c. AD 50-140/50 is suggested largely by coarsewares (see pottery report). The virtual absence of Late Iron Age and Late pre- Roman Iron Age or Early Roman ‘Belgic’ type wares is significant and a key factor in suggesting a post- Conquest date. Coarsewares are instead dominated by ‘Romanised’ reduced sandy coarsewares and Savernake type wares which, on the basis of their occurrence at Cirencester, date to before c. 140/50 (Rigby 1982, nos 193-197). Limited support for such a date is provided by dateable non-ceramic finds from Phase 2a/b, which include a hinged iron brooch, almost certainly pre-Flavian in date (Figure 11.2), and a mould-blown glass fragment from gully 4 (Glass catalogue no. 1), dated to the third quarter of the Ist century AD. Samian from phase 2a/b (5 sherds) consists mainly of South Gaulish material of Flavian date. Phase 2c c. AD 140/150-200/210 (Figure 6) Ditch 1, aligned north-east/south-west, cut across three of the Phase 2a gullies (3, 4, 6) and Phase 2b gully 6 and fence 7. Ditch 1 had a rounded southern terminal, whilst its northern end lay outside the excavation area. Ditch 1 with its sandy clay fill appeared to have been recut and had a maximum width of 2.3m. A number of pits and postholes immediately to the north-west may represent activity within an enclosure partially demarked by the ditch. Posthole 270 produced the highest concentration of charred plant remains from the twenty samples taken on the site, with a ratio of grain to chaff to weed seeds - of 1:3:3 indicating cereal processing waste. Dating Evidence Dating is derived primarily from pottery, particularly the samian which includes Antonine forms as well as residual Flavian material (see pottery report). Elements within the coarsewares supportive of such a date include a large number of bag-shaped beakers of likely Antonine date and Black-Burnished ware or BB1 influenced forms dating to before c. AD 225. Non-ceramic finds include a coin of Domitian (AD 81-96; catalogue 1) and a brooch of Polden Hill form (catalogue no. 5), each of which would appear to be residual in its respective context. Phase 2d (Figures 6 and 7) This phase 1s characterised by the laying out ofa field system with a trackway on a common alignment. Trackway 1 was defined by three ditches 2, 3 and 4, much more substantial than the Phase 2a and 2b gullies, being up to 2.5m wide and 0.8m deep, and regular in form with a straighter alignment than the Phase 2c ditch (see Figure 6). The trackway ran east-west across the whole excavation area. In the western and central portions of the site, the trackway was defined by the parallel ditches 2 and 3, approximately 8m apart at the western edge of Area B. Towards the eastern edge of the excavation area, ditch 2 turned 90° to run northwards, where it was intersected by ditch 4, which formed the northern boundary of the trackway from this point. At the eastern edge of the excavation area, the trackway had widened to 18m. The relationship between ditches 2 and 4 was not defined and is entirely possible that the two are contemporary and that trackway 1 as originally conceived ran right across the excavation area. Equally, however, ditch 4 can be interpreted as a later addition to the system that formed an extension to the original trackway. If this was the case it is possible that ditches 3 and 5 originally mirrored the layout of ditch 2, and that ditch 3 was extended to the south-east at the same time as ditch 4. The land to either side of the trackway was sub-divided by ditches 6, 7, 8 and 10. Within the field system some discrete features were identified, including drying oven 1, well 1, and a number of gullies and shallow pits. Drying oven 1 lay to the south of trackway 1 and was T-shaped with a stoking area, a main flue and a cross flue (Figure 7). The base of the structure was formed of limestone slabs, many of which exhibited signs of scorching. The base of the stoking area had been renewed once. Although no superstructure survived, asubstantial deposit of clay overlying the feature to a depth ofc. 0.14m might indicate that the oven was originally capped by a clay roof. Charred plant remains, samples <16> and <17> from the north-western end of the flue, consisted of 99% and THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 18 (OSL:1) 9Z pun vz saspyd suimoys Ud] “J VAL § ‘SLT L aneib rs) = - aa borer Ihe aa ee és 1 Ajin6 zAyn6—-¢ Ajin6 WQS q¢ eseud (zz 0 e¢ OSeUd EEE 19 SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 LL Youp 7 \ OL Youp S Youp 9 ae | y youp (OSL°1) az 01 97 saspyd Summoys UD]_J “J VIAY 9 ‘BLT 2 Youp © YoUp yk ea ey sels recovered from the stoking area contained a much higher concentration of chaff than grain, suggesting that cereal processing waste was used as fuel. Well 1 was located in the northern corner of the excavation area, in the south-eastern corner of an enclosure formed by ditches 7 and 8. The well was cut 1.6m deep into the bedrock, with an oval top 3.2m by 2.5m and a sub-circular shaft with a maximum diameter of 2m. The well was filled with a mixture of clay cornbrash, suggesting deliberate backfilling. There was no evidence of a well lining. One inhumation and two cremations were found. Grave 2 was found to the north of well 1, disturbed with only the skull, arms and right femur surviving. This extended burial appeared to be of a young adult/adult male. A radiocarbon date could not be determined on the sample of bone submitted. Cremations 1 and 2 consisted of cremated bone and pyre debris buried in shallow, sub-circular pits, located 2.5m apart to the north of the Phase 2c, ditch 1. All three burials are undated and it is possible that they are Iron Age or Early Roman in date as grave 1. SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 21 Phase 2e The final Phase of Roman activity is the addition of a second trackway to the Phase 2d field system. Trackway 2 ran northwards from trackway | and was defined by Phase 2d ditch 8, and ditch 9 that had been newly dug to the west. Ditch 9 cut the Phase 2d east/west ditch 7. Although no evidence for the infilling of the portion of ditch 2 that separated the two trackways was observed, it would seem likely that this had occurred in relation to trackway 1, otherwise trackway 2 would have been a dead-end. During this phase a waterhole was dug immediately to the west of ditch 9, perhaps to replace well 1, which now lay within trackway 2 and is likely to have gone out of use at this time. This feature measured 4.6m by 3.9m, but was not fully excavated due to flooding, although it was at least 1.5m deep. Dating evidence Phase 2d/e: c. AD 150/60-200/210 Phase 2d/e could not be separated on the basis of the artefactual material from Phase 2c. Samian is largely unhelpful, comprising residual South Gaulish material and scraps of broadly 2nd century dateable Central Gaulish vessels. The appearance for the first time of small quantities of mortaria and amphora fabrics might be seen.as significant, although a cut-off point for all Phase 2 material of c. AD 240 is evident from the absence of common Late Roman ware types such as Oxfordshire colour-coated ware and distinctively late forms in Black-Burnished ware such as flanged bowls. Two brooches from ditch 7, of trumpet-headed (Figure 12.6) and enamelled ‘knee’ type (Figure 12.7), probably date to the first half of the 2nd century AD. Area C (Figure 3) Archaeological features were found in two of the four evaluation trenches. In trench | seven features, five of which contained pottery of Roman date, were found. These included a ditch 1008/1010, the fill of which produced 50 sherds of 2nd-century AD date, 30 animal bones and 3 fired clay fragments. This ditch was cut by pit 1012 and gully 1006, dated to the 2nd century AD. Gully 1006 was on the same _alignment as a further gully 1020 and ditch 1004. Gully 1020 lay c. 11m to the north-west of gully 1006 and was of similar dimensions. The alignment of these gullies and the distances between them are similar to the properties of gullies 1,2 and 3 in Area B, which may mean that the Phase 2a agricultural plots defined by these gullies continued into Area C. In all, 41 sherds of 2nd-century AD date were recovered from the fill of ditch 1004, along with 2 fired clay fragments. The remaining features in Trench 1 consisted of a north/south aligned gully 1018 and pit 1014 (not illustrated). Five sherds of 2nd-century AD date were recovered from the fill of the latter. Six archaeological features were uncovered in trench 2. Crossing the trench on a north-east/south- west alignment was ditch 2008. The fill (2007) of this ditch contained South-West White-Slipped ware mortaria of hammer-head and flanged rim forms of late 2nd to early 3rd-century date. The ditch was adjacent to and on the same alignment as a possible ditch terminal 2006. A single shallow posthole 2010 was apparent to the north-east of these two ditches. A further four postholes 2012, 2004, 2014 and 2016, on a linear arrangement, were apparent to the south-west. THE FINDS by E.R. McSloy Flint A total of 110 pieces of worked flint was recovered from evaluation and excavation. Pieces with secondary working amount to 16.4% of the group (Table 1). The bulk of the assemblage comprises flakes or blades, ‘shatter pieces’ and cores. The lithic material includes 37 pieces recovered as a result of sieving spoil through a 10mm mesh from 28 test pits in Area A (Table 1). A further 24 pieces are unstratified. The remaining 49, pieces or 44.5%, derive from hand excavated features mostly of Roman date. Four pieces, comprising unretouched flakes and a (residual) broken leaf-shaped arrowhead, are associated with Early Bronze Age Period 1 features. The material is largely in good or fair condition with very little edge damage or heavy ‘rolling’ evident. The unstratified material by contrast is poorly preserved. Blades or bladelets have suffered breakage and most pieces exhibit moderate edge damage. Patination, the discolouration of flint as the result of burial in a calcareous environment, was noted on a majority of pieces (57%). The discolouration varies from a light milky or yellowish mottling to a deep white. Material of likely Mesolithic date comprises a 22 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 8 Flint (scale 1:1) single obliquely-blunted blade (Figure 8.1) and seven bladelets, including broken pieces. Additionally, a further 15 blades or blade fragments probably date to this period or to the earlier Neolithic. A proportion of the flakes and undiagnostic scrapers may also be of this date. A single broken leaf-shaped arrowhead (Figure 8.2) and some blades probably date to the earlier part of the Neolithic. A broken, bi-facially worked tool (Figure 8.3) may represent a second leaf-shaped arrowhead. Alternatively, it may be from an ‘ovate’ form of knife dateable to the Neolithic or Early Bronze Age. The bulk of the flake debitage exhibits characteristics typical for the Late Neolithic and subsequent periods. Cores are multi-platform types and flake removals are for the most part of ‘squat’ proportions with length to breadth ratios close to 1:1. Striking platforms are mostly thick and most exhibit the pronounced bulb and other characteristics suggestive of ‘hard hammer’ percussion. Scrapers, retouched flakes and the piercer may date to this period or the Bronze Age. Barbed and tanged and tanged arrowheads (Figure 8.4 and Figure 8.5) and a single small ‘thumbnail’ scraper are characteristic of the Early Bronze Age period. A further three pieces, a scraper/knife, a domed discoidal scraper and a retouched flake exhibit shallow invasive retouch, often a characteristic of Beaker, lithic groups (Gibson 1982). A proportion of the debitage, cores and tools, which can be indistinguishable on typological or technological grounds from late Neolithic material, will almost certainly relate to this period. Discussion Mesolithic activity previously alluded to in the area (Walker 2000, 236) is evidenced by this group. A significant proportion of the probable Mesolithic component (8 pieces or approx 22%) including bladelets and an obliquely-blunted blade (flint cat. 1) was recovered from sieving of topsoil from the test pits. No sieving was undertaken of general excavated feature fills and the Mesolithic component may be underestimated as a result. The bulk of the worked flint dates to the Late Neolithic or Early Bronze Age and would seem to be associated with Beaker activity on the site indicated by pottery finds. So little Beaker material is dated nationally that this group associated with a radiocarbon dates is very useful, although the assemblage is not large and very little worked flint is stratified. Retouched pieces are well represented and, if this is not as the result of a collection bias, may be indicative of domestic use of the site. SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 23 Table 1: Worked flint summary Area | Phase Cores | Flakes | Blades | Tools | Total Unph. A (test mits) 4 24 7 2 37 1 - 1 2 1 4 2a/b - 4 2, 1 if 2c - 3 - 1 4 2d 1 12 4 2 19 Unph. 2 16 zi 11 36 B 2a/b - 1 - - 1 C 2d - 2 - - 2 7 63 22 18| 110 Illustrated pieces (Figure 8) 1 Area A, test-pit 6. Bladelet, obliquely blunted. Continuous fine abrupt retouch to blunted edge and one blade edge. Grey flint with slight milky patina. Surviving length 26mm. 2 Area B, ring ditch 1. Period 1. Broken leaf-shaped arrowhead of honey-coloured flint. Bi-facially worked with shallow invasive retouch. Lower portion only and not classifiable. Surviving length 26mm. 3. Unstratified. Broken leaf-shaped core tool on unpatinated black flint. Bi-facially worked with invasive retouch. May represent a large arrowhead Green types la/b (Green 1980, fig. 26 and 27), or an ovate/’laurel leaf’ knife. Surviving length 33mm, width 35mm. 4 Colluvium 97. Complete barbed and tanged arrowhead of Green’s ‘Sutton b’ type (Green 1980, 122). Well-made with shallow invasive retouch over most of surface. Grey flint with slight milky patina. Length 24mm, width 19mm. 5 Area A, test-pit 17. Broken tanged arrowhead of Green’s ‘Sutton a type (Green 1980, 122). Very crude. Fire damaged. Surviving length 20mm, width 15mm. Pottery Beaker pottery Four sherds, representing four individual vessels, weighing 62g were recovered. Fabric codes following the nomenclature of the PCRG (1997). Beaker Fineware — fine grog-tempered type (GR1l BKF) Handmade. Dark reddish brown surfaces and dark grey core. Fabric is hard with soapy feel and finely irregular fracture. Inclusions are well sorted, comprising medium (0.5-lmm) red-brown grog, sparse fine (0.1-0.3mm) colourless sub-angular quartz and sparse medium (0.3- 0.5mm) angular calcined flint. Beaker Coarseware — coarser grog-tempered type (GR2 BKC) Handmade. Mid or light reddish brown outer surfaces and dark grey core. Fabric is soft with a soapy feel and irregular fracture. Inclusions are poorly sorted, comprising coarse (1-2mm) sub-rounded dark grey or brown grog and sparse voids from burnt-out organic inclusions. Two Beaker fineware sherds (Figure 9.1 and 9.2) were recovered. The incomplete nature of the sherds prevents classification. The style of decoration, however, is consistent with Case’s (1993, 260-3) regional group D, for southern Britain, which essentially comprises his Style 2 Beakers (Case 1977). Successive bands of combing on sherd no. 2 in particular have affinities with other vessels from the region, including Berwick St John (Case 1993, fig. 16, no. 7) and Downton, Wilts (abid, fig. 18, nos. 1-7). The unstratified sherd comes from a large coarseware vessel (Figure 9.3). The rusticated paired fingernail impression decoration is typical for such vessels and characteristic of Beaker domestic assemblages (Gibson 1982). Beaker coarsewares are also known among Case’s (Case 1993, 262) Group D assemblages and include crow’s foot type fingernail- decorated vessels. The fourth sherd (not illustrated) is an undecorated base sherd identified on the basis of fabric alone, which was residual within Waterhole 1 along with Iron Age pottery. Illustration catalogue (Figure 9) 1 GRIBKE Ring ditch 1. Period 1. Fineware body sherd with three lines of square-toothed comb impressions probably forming border between decorated panels. Thickness 5mm. 2 GRI BKF Ring ditch 1. Period 1. Fineware body sherd. Three rows of oblique square-toothed comb impressions arranged in herringbone fashion. Thickness 6mm. 3. GR2BKC. Unstratified. Coarseware sherd. Lower body and base angle. Paired fingernail impressions. Tron Age Pottery Four handmade body sherds of quartz or organic- tempered fabrics are distinctly Iron Age as they are not like the wheelmade early Roman fabrics of Period 2. Diagnostic features of fabric or form are absent, precluding closer dating. All sherds are residual within Period 2 features. Quartz-tempered (QU) Handmade. Dark grey/black throughout or with reddish brown surfaces. Fabric is soft with a sandy or harsh feel and a finely irregular fracture. Inclusions are well-sorted, comprising fine/medium (0.2-0.3mm) colourless sub- angular quartz and sparse, sub-rounded red-brown clay pellet or grog (0.3-0.Smm). Organic-tempered (ORG) Handmade. Dark grey/black throughout. Fabric is soft with a soapy feel and a finely laminated fracture. No mineral 24 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 9 Prehistoric and Romano-British pottery (scale 1:3 and 1:4) SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 25 inclusions visible. Abundant linear voids from burnt-out organic inclusions. Roman Pottery with samian identifications by G.B. Dannell and description of Wiltshire imitation samian by Robert Hopkins A total of 5047 sherds of Roman pottery, weighing 42.42kg was recovered from 218 separate contexts. Minimum vessel count, calculated by quantifying sherd families, is 3313 and total estimated vessel equivalence (rim and base EVEs) is 62.68. The Roman pottery was quantified according to sherd count, minimum vessel count (sherd families), rim and base EVEs and weight per fabric. Fabrics were identified macroscopically or with the use of a (x4) hand lens and matched, wherever possible, against the National Roman Fabric Reference Collection (Tomber and Dore 1998). Table 2 summarises the quantities of all represented types. Attributes such as decoration use wear evidence, visible residues and any cross-context matches are recorded for each sherd family. Vessel form was recorded using a system of mnemonic codes which relate to the generic form class, jar, bowl, beaker etc, and specific elements, mostly based on rim morphology which are explained in detail in the archive. Illustrated pieces emphasise forms which appear to be characteristic of the local (North Wiltshire) pottery industry. The Roman material extends from c. AD 50 to no later than the first quarter of the 3rd century, with perhaps an emphasis on the middle years of the 2nd century AD. Poor surface preservation is a characteristic of this assemblage, most likely the result of burial environment. A small number of vessels are sufficiently complete for reconstruction of profile. Average sherd weight is low for a Romano-British group at 8.4g and suggestive of a well broken-up, dispersed assemblage. High levels of fragmentation may be an effect of the pattern of recovery which is characterised by retrieval from linear features (74% by sherd count). Only a small proportion of the pottery derived from pits or other discrete features might reasonably be expected to produce groups deposited over a short period of time. Truncation of the site overall has ensured that there is no horizontal stratigraphy surviving and all pottery is from negative features. The low average sherd weight suggests a gradual build-up of domestic detritus as the result of agricultural processes and the movement of people and livestock. More structured, single event rubbish disposal within the larger ditch fills may also be indicated in some instances. Convincing evidence for this survives from the recovery of certain groups of pottery which are discrete in terms of date, containing con-joining sherds or even substantially complete vessels. The occurrence of such groups and the overall quantity of pottery from the site are a likely indication that the main areas of domestic activity were not far from the area of excavation. Few comparable published assemblages are known from the immediate area, although a small assemblage was recovered from excavations along the course of the Littleton Drew to Chippenham gas pipeline (Timby 2000, 94-5) and Malmesbury villa (McSloy 2004). This material exhibits a similar range of ware types, dominated by material from the Wiltshire industries. More useful for comparison are the expansive, predominantly north-Wiltshire based coarseware assemblages from Cirencester (Rigby 1982a; Rigby 1982b) and Wanborough (Seagar Smith 2001). Fabrics Fabric descriptions are detailed in the site archive and follow methodologies employed for the National Roman Fabric Reference Collection (NRFRC; Tomber and Dore 1998). NRFRC codes are used where applicable. Represented forms are indicated in quantitative order. Local, grogged Savernake ware (SAV GT); Tomber and Dore (1998, 191). Soft grog-tempered fabric ?Local or Savernake variant (GR) Grog with quartz ?Savernake variant (QZ GR) Coarse grog with quartz ?Savernake variant (GR QZ) Grog with coarse flint ?Savernake variant (GR FL) Pale grey grog and organic tempered ware ?Savernake variant (GR CH) Local reduced Common greyware fabric (WIL GW) Grey throughout fabric (WIL GW1) Coarse greyware fabric (WIL GW2) Dense, gritty greyware fabric (WIL GW3) ‘Dense’ Greyware with oxidised (colour-washed) surfaces (WIL GW5) Greyware with coarse chalk inclusions (WIL GW6) ‘Dense’ Greyware with external white slip (WIL GW7) Fine greyware (GW F) Fine greyware with orange margins (WIL BS) Black sandy fabric, usually burnished (WIL BS) 26 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 2: Roman pottery. Quantification by fabric. + indicates NRFRC code used (Tomber and Dore 1998) Group Fabric Count Vessel No. Weight(g) EVEs %EVEs Imports BAT AMt 1 1 302 LEZ SA2i: 30 25 344 R25 LGF SAt Bi 37 285 2.19 Sub-total 68 63 931 3.44 55) Grog SAV GTT 428 301 13433 4.11 GR 47 30 1074 .60 GRE 4 4 113 04 GRQ 17 72 807 .28 WIL GW4 52 34 855 Sy) Sub-total 548 381 16282 5.55 8.9 Local WIL Bl 220 98 1226 3155 oxidised WIL B2 5 5) 40 0 WIL B3 24 17 289 2:29 WIL B4 18 12 244 53 WIL BS 14 5 122 93 WIL CC 2) 2 7 1.0 CRE 9 7 48 0 GLAZ 1 1 3 0 SA IM 12 1 108 ail OX M 1 1 5) 0 OX WS 37 33 222 1.86 OX WS1 8 3 67 0 OX WS2 l 1 32 0 WIL RS 2 2) 30 15 WIL OX 577 422 3098 8.88 WIL OX3 2 2 69 0 WIL OX4 ll 7} 69 .26 SOW WSt 4 4 172 13 Sub-total 948 618 5851 19.39 31.2 Local WIL BS 2476 1496 11578 22.67 reduced WIL BS1 111 37 811 oS) FL 6 2 39 0 1 EKG 1 1 56 0 GW 657 SWS) 4844 6.73 WIL GW6 47 2 448 “alls GW F 2 7) 26 10 GW WS 4 4 46 0 WIL GW1 50 32 383 2 WIL GW2 6 5 So aU WIL GW3 9 5 120 1.45 WIL GW5 13 7 145 05 ORG 3 1 46 0 QZ 3 3 24 0 QZF 4 4 24 0 Misc (burnt) 21 14 145 0 Sub-total 3413 2190 18787 32.48 51.8 Non-local British SOW BBIT | 70 61 564 1.67 Sub-total 70 61 564 1.67 Daf, Totals 5047 3313 42415 62.68 Coarse black sandy fabric, usually burnished (WIL BS1) Soft buff fabric (WIL B3) Fine black sandy fabric, usually burnished (WIL BS3) Soft orange-brown fabric (WIL B4) Black firing fabric with fine flint inclusions (FL) Common sandy oxidised fabric (WIL OX) Black firing fabric with quartz and fine flint inclusions Coarse oxidised fabric with prominent iron oxide (WIL (QZ F) OX3) Sandy-oxidised fabric with grog inclusions (WIL OX 4) Coarse flint (FL C) Pale orange mortaria fabric (OX M) Local oxidised Sandy-oxidised white-slipped flagon fabric with grog Sandy buff fabric (WIL B1) inclusions (OX WS1) Buff fabric with grog and organic inclusions (WIL B2) Sandy-oxidised white-slipped flagon fabric (OX WS2) SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 D7; Lead-glazed oxidised fabric (GLAZ) Red-slipped micaceous oxidised fabric (WIL RS) Local British Wilts Imitation Samian (WIL IMIT) Creamware (CRE) Gritty whiteware (GR WH) Local colour-coated ware (WIL CC) South West white-slipped ware (SOW WS); Tomber and Dore (1998, 192). Non local British South-West Black-Burnished ware 1 (SOW BB1); Tomber and Dore (1998, 129). NRFRC code SOW BBI. Lower Nene valley colour-coated ware (LNV CC); Tomber and Dore (1998, 118). Continental imports La Graufesenque (South Gaulish) Samian (LGF SA); Tomber and Dore (1998, 28). Lezoux (Central Gaulish) Samian (LEZ SA2); Tomber and Dore (1998, 32). Baetican Dressel 20 type amphora fabric. (BAT AM); Tomber and Dore (1998, 84-5). Finewares Imported finewares are restricted to quantities of South and Central Gaulish samian. A total of 67 sherds (629g) of Gaulish samian, representing at least 62 vessels was recovered ranging in date from the pre-Flavian to the Late Antonine periods, with the bulk of material probably dating to c. AD 70- 100. South Gaulish vessels, all of which pre date c. AD 110, predominate (64% according to vessel count) with the remainder deriving from Central Gaulish sources. Condition reflects that of the coarsewares. The majority (65%) of vessel sherds could be identified to form (Table 3). Of the remainder, three can be assigned to vessel class and 17 chips are unassignable. Few contexts contained more than one or 2 vessels. Ditch 1 (Phase 2c) is exceptional in producing six vessels, including complete profiles of Drag. 27 and Drag. 18/31 vessels. Some 37 South Gaulish vessels are represented. A single Drag. 24/5 cup is of pre-Flavian date. The remainder is almost certainly all Flavian. Much of this material occurs with 2nd-century or later material and can therefore be considered residual or as survivals in use. The bulk of all of the South Gaulish samian probably comes from La -Graufesenque. Forms (Table 3) are dominated by platters and cups of the period c. AD 70-110. No maker-stamped or (mould) decorated vessels were recovered. A total of 21 Central Gaulish samian vessels was recorded. The bulk of material (almost certainly entirely composed of Lezoux products) dates to the early to middle Antonine period. The preponderance of plain bowls, dishes and cups, seen with the Southern Gaulish group, is repeated. Decorated forms are restricted to a single chip of a Drag. 37 bowl. Of significance is the proportionally high occurrence of Drag. 18/31 dishes — forms which ceased production around AD 160. Later Antonine forms are generally rare, restricted to 3 vessels (Drag. 38 and 31R), two of which are unstratified. British finewares are dominated by oxidised wares, including a single Wiltshire imitation samian vessel (Figure 10.24), and white-slipped fabrics. A north-west Wiltshire source is assumed for Wiltshire samian on the basis of its distribution. The few sherds of this fabric identified over the last 25 years are within the zone identified by Anderson (1979b, fig.1). Accepted dating for material of this type is the late Ist to early 2nd century AD, based on the similarity of some recorded designs to pre-Flavian and Flavian samian and the recovery of sherds from a Flavian context at Sea Mills, near Bristol (Bennett 1985). If such a date is accepted the Showell Farm vessel is likely to be residual within a mid to later 2nd-century context, or perhaps represents a survival in use. Oxidised wares account for 31% of the total assemblage according to EVEs. Forms include ring- necked flagons (Figure 10.21) similar to examples in fabric 9 at Cirencester (Figure 9.7 and 9.15; Rigby 1982a, 154), a variety of carinated and round-bodied bowls, tankards in imitation Severn Valley Ware forms and cornice-rim beakers. The form of cornice- rim (‘bag-shaped’) beakers in Wiltshire of oxidised type can be compared with North Wiltshire colour- coated wares known from Cirencester (Rigby 1982b, D03) and Wanborough (Seagar-Smith 2001, 240). Evidence for a separate, though probably still local source, is provided by two vessels (Figure 9.8 and 9.9) with white-painted decoration, a technique not seen on the Cirencester or Wanborough examples. Fine, probably local, greyware types occur in small quantities and include a bowl imitating a Drag. 36 bowl (Figure 10.26). Coarser reduced fabrics WIL BS and WIL GW3 include tableware forms; for example beakers and platters (Figures 9.11 and 10.28) as part of their predominantly utilitarian repertoire. Single vessels of red-slipped (Figure 10.27) and lead-glazed fabrics are also likely to be local. THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 10 Romano-British pottery (scale 1:4) Coarsewares A single fragment of imported Dressel 20 amphora was recovered. Mortaria are poorly represented at Showell Farm. Fabric OX M is unsourced but likely to be local. Only two other vessels are recorded, both in South-West white-slipped ware (Figure 10.30 and 10.31). This fabric type is well known from Cirencester and almost certainly originates SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 29 in the North Wiltshire region, possibly close to the small town at Wanborough (Tomber and Dore 1998, 192). The coarseware assemblage is dominated, throughout Phases 2a-2e by black, sandy fabric WIL BS and its coarser variant WIL BS1. A link with Cirencester fabric 5 (and its coarse variant type 15) is demonstrable, not only from the similarity of the fabric (Rigby 1982a, 153), but also from a distinctive and reasonably diverse repertoire of forms. Similar material also occurs at Bagendon (Clifford 1961, fig 65 116D) and Kingsholm (Darling 1977, fig 6, 9). Cirencester fabric 5 is present in the earliest military phases in the town and continues in use until the end of the 2nd century AD. Flint-tempered fabrics represent only a small component within the total assemblage (Table 2). The use of calcined flint is common in the Berkshire/ north Hampshire region (Timby 2000, 239) in the early Roman period. It also occurs infrequently at Wanborough (Seagar Smith 2001, 250). Flint occurs as an inclusion in Savernake type wares and an East Wiltshire source is thought likely in this instance. Greywares form a major and constant component of the assemblage, seemingly increasing slightly over the period represented. No major greyware producing kilns of Early Roman date have been identified in the region, although it is widely believed that the Later Roman production sites near Swindon at Whitehill Farm and Purton (Anderson 1979a), must have precursors. The fabrics represented are broadly comparable to Cirencester fabric 17 (Rigby 1982a), and in some instances, types WIL GW 1 and WIL GW3, with Wanborough type 52 (Seagar-Smith 2001, 243). The likelihood is that this group of fabrics includes material from a variety of sources, all of which are likely to be reasonably ‘local’. The identification of a probable ‘second’, a necked jar with a distorted rim, might be regarded as evidence for a source nearby. The grog-tempered group contains a small quantity of material (type GR) resembling Late Pre-Roman Iron Age types known across southern parts of Britain, which clearly continues into the Early Roman period. Savernake wares form a significant part of the assemblage. The collection includes a number of apparent sub-types (Table 2) and accords with the variability of the ware described by Tomber and Dore (1998, 191) and noted by Rigby (1982a, 154). This variability may indicate a number of different North Wiltshire sources, with a concentration in the region of Mildenhall, c. 30km to the east of the site. The sole British traded ware present of any importance is Black-Burnished ware | (BB1), 2.7% of the total according to EVEs and of SOW BB1 type. Forms present comprise mainly everted-rim jars with a small number of bowls/dishes with flat or flat grooved rims. The forms represented, together with an absence of such common late form such as conical flanged bowls or jars with obtuse angled lattice decoration, suggest a date before c. AD 230. Forms The composition of the assemblage by vessel form is fairly typical for a rural site of this date. Jars predominate, with necked and everted-rim (BB1 type) forms most prominent among the sandy reduced wares and the larger, mainly neckless varieties dominant among the grogged (Savernake type) wares. The ‘Belgic’ derived jar form with a prominently rounded shoulder (Figure 9.16), characteristically here occurs with black-firing fabric WIL BS. Interestingly, a single example is also present in oxidised fabric WIL OX. Also noteworthy are two examples of (greyware) jars with ‘cornice’ rims (Figure 9.4) usually associated with beakers. Open forms, consisting of dishes/platters and bowls are well represented. Platters, including stepped (Figure 10.28) and curved-sided (Figures 9.18 and 10.27) forms are derived from Gallo-Belgic precursors. Platters are also dominant among the samian assemblage (Table 3). Bowls occur with greatest variety among the oxidised wares. Curved- sided and carinated forms such as Figure 9.6 and 9.15 are local varieties of common 2nd-century forms (see Gillam 1957, nos. 214-8). Other, straight-sided dish and bowl forms are influenced by Black Burnished wares. Figure 10.25 probably imitates samian form Drag. 36, with white painted decoration replacing the barbotine swirls. A second vessel, imitating the Drag. 36 bowl, occurs in fine greyware fabric GW F (Figure 10.26). Decoration is in this instance more faithful to the original samian prototypes. The drinking vessel component is significantly large at 15% total EVEs and is inflated by the identification of some vessels (mostly tankards) from diagnostic base sherds. This reflects the narrow date range of the group as a whole and the emergence of the bag-shaped and conical rimmed beaker as a popular form at this time. The form is derived from earlier 2nd-century continental finewares (Anderson 1980, fig. 9). The tankards present, including the substantially complete illustrated example Figure 10.32, are faithful copies of 2nd- century Severn valley forms (Webster 1976, fig. 7). 30 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 3: Pottery forms (summary) Class Description E BTO$ | redu.| _oxid. sam. total | % total Flagon Disc-neck 1.20 Ring-neck a Sub-total 131 1.31 3.10 Beaker/tankard | uncertain 30 /cup Beaker, bag-shaped (cornice or triangular rim) 58 1.82 indented | present 2?cup with bead rim 10 eG | Tankard, SVW type 2.61 [ Drag. 24/25 06 Drag. 27 .60 Drag. 33 45 Drag. 35 10 Drag. 46 a2 Sub-total 68 Swill 1.46 TDs) TZ. | Jar Neckless, everted rim, BB1 type 3.34 Neckless, other 1.08 0.31 0.41 Necked, medium-mouthed 0.65 8.07 1.36 Necked, narrow-mouthed Bu ‘Belgic’ type jar/bowl 1.0 alg) Large, storage jar size 3 Neckless with triangular rim Sub-total 2.46 | 15.84 1:9 20.2 47.8 Bowl uncertain 07 05 eal carinated NEPAL Curved sided with bead rim a2 17 Curved sided with plain rim Curved-sided, ‘s’-shaped rim 30 Curved sided with flat rim 10 Curved-sided, flat bifid rim 06 Flat rim (BBI type) 07 Flat rim with groove (BB1 type) ali] hemispherical, flanged 50 Imitating Drag 30 al Imitating Drag. 36 24 Drag. 31r a) Drag. 37 present Straight-sided, bead rim .05 Straight-sided, flat rim 18 Sub-total 6.37 2.65 .20 9.22 21.8 Dish/ Curved sided with bead rim 13 [ Shallow bowl _| Curved sided with plain rim 2 Curved sided, triangular rim | Flat rim with groove wlll Flat top rim BB1 type 44 Plain rim 10 Plain rim with groove below 02 | Triangular rim 07 Drag. 18/31 70 Drag. 18r 10 Drag. 42 OS Sub-total 0.86 13 85 1.84 4.4 Platter All coarseware 94 40 Drag. 18 93 a Sub-total 94 40 293 227. 5.4 Mortaria uncertain present flanged 07 Hammer-head 06 Sub-total 0.13 0.13 0.3 | Strainer-bowl | all IE present Total 42.22 | SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 31 To this material must be added cups which account for approximately 10% of the samian assemblage and a single coarseware vessel (Figure 9.19), probably imitating a Drag. 33 cup. Other forms are only sparsely present. Flagons are especially poorly represented given, the earlier Roman date of the deposits and when compared to other drinking-related forms. Chronology is likely to be a factor affecting the very poor representation of mortaria, which tend to be rare on many rural sites until the later 2nd to 3rd centuries. Dating: pottery by phase Dating evidence for the Romano-British phases (2a- e) is briefly outlined in the main body of the text. Overall the period represented is a narrow one and whilst specific date markers are apparent (below), differences in the general composition are subtle. The samian is of restricted value as dating evidence: South Gaulish material, most of which is Flavian in date, would seem to be largely residual, occurring in all phases; Central Gaulish samian, largely Antonine in date, occurs in smaller quantities, although it is most abundant in Phases 2c to 2d/e. The homogeneity of the assemblage is most likely a factor brought about a reasonably unvarying pottery supply in this period. Compositional differences are most marked when comparing Phase 2a with the succeeding Phases (2b-2e), which viewed overall appear reasonably homogeneous. The proportion of oxidised fabrics is lower for Phases 2a/2b compared with subsequent phases, more particularly for Phase 2a where the figure is less than 10% by count _ compared to in excess of 15% for each succeeding phase. Further comparisons between the Phase 2a and 2b assemblages reveal differences in the representation of certain fabrics including, most- markedly, Savernake ware. This type (fabric SAV GT) is most abundant in Phase 2a, representing approximately 16% of the assemblage (according to sherd count), a figure that is reduced to 8% for Phase 2b and 6% and 7% for Phases 2c and 2d respectively. Date markers- Phase 2a/b Four sherds of South Gaulish samian of Flavian date and a single Central Gaulish Drag. 33 cup dateable _to the Hadrianic to Antonine period were recovered from Gully 3, Phase 2a deposits. Non-sigillata finewares are also poorly represented. Flagons, tankards, shouldered jars and bowls occur among Phase 2a, but not in closely dateable vessel forms. An unusual bowl (Figure 9.6) is noteworthy and can be paralleled among examples in (reduced) Cirencester fabric 5 (Rigby 1982a, nos. 470 and 491), from Ist- century AD contexts, and from Wanborough in a more comparable oxidised fabric (Seagar Smith 2001, no. 122) from a phase corresponding to AD 60-65. Dating evidence for the black-sandy types which dominate the Phase 2a/b groups is largely reliant on material from Cirencester, where fabric 5 is closely comparable in terms of fabric and the range of forms. Among the forms most commonly represented at Showell Farm are a number of platters, shouldered jars/bowls (Figure 9.18 and 9.16) and other forms which can be readily paralleled among vessels known from Cirencester. Platters of types, which are ultimately derived from Gallo-Belgic forms, occur in four Phase 2b contexts. Such vessels compare with Cirencester examples from deposits immediately post-dating the military occupation and unlikely to much post-date c. AD 100 (Rigby 1982, nos 193-197). Other forms, including everted-rim jars imitating Black-Burnished ware vessels, certainly date to the middle or later years of the 2nd century (Rigby 1982b, C14). Phase 2c Black-sandy coarseware types continue to dominate Phase 2c and only subtle differences are apparent in the representation of other types. Samian is scarcely more abundant compared with the preceding phase. Significantly, however, Central Gaulish vessels of certain 2nd-century currency now predominate. Typically (earlier) Antonine in date are bag- shaped beakers with cornice rims (Figure 9.8- 11) which occur in oxidised and reduced ware fabrics. The form of these vessels is derived from continental colour-coated beaker forms (Anderson 1980, 6-9) current in the first half of the 2nd century AD. Such forms are copied in Britain at centres including Colchester, the Lower Nene Valley and North Wiltshire, probably from the early Antonine period. A date for this material of c. AD140-180 is favoured here. Black-sandy fabric WIL BS present in Phase 2c shows considerable influence from the south- western British Black-Burnished ware industries, and is unlikely to pre-date the early Antonine period (Holbrook and Bidwell 1991, 97-8). A dish in greyware fabric WIL GW1, with a flat grooved rim (Figure 9.14) imitates a BB1 form of probable later Antonine or later date. True (south-western) Black-Burnished ware occurs in small quantities in ditch 1, as everted-rim jars with acute-angled lattice decoration and, in one instance, a burnished wavy 32 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE line below the rim (Figure 9.13). Absence of specific forms of BB1, including conical flanged bowls or jars with obtuse-angled lattice, and the overall paucity of this ware type, probably indicates an upper limit for this phase of before c. AD 250. Phases 2d/e Phases 2d/e could not be separated on the basis of the artefactual material from earlier phases, with reduced coarseware types continuing to dominate. Black-Burnished ware forms include everted-rim jars with acute angled lattice and/or burnished wavy decoration on the rim and flat-rimmed bowls or dishes indicative of the late 2nd to early 3rd century. Occasional other forms such as disc-necked flagons (Figures 9.17 and 10.22) in fabric OX WS compare closely with examples from Cirencester and neighbouring sites (Rigby 1982b, no. 139; Timby 1999, no. 162) and are consistent with the date range proposed. Other vessels can also be attributed to this date bracket, for example South-West White- Shpped ware mortaria with hammer-head and flanged rim forms (Figure 10.30 and 10.31), both from (unphased) Area C, trench 2, gully fill 2007. Each vessel can be paralleled among finds from Cirencester (Rigby 1982b, nos. 4 and 48) where they are considered to date to the first half of the 3rd century AD. The absence of types post dating c. AD 240/50, which might include Oxfordshire colour-coated wares and flanged bowls or obtuse- lattice decorated Black-Burnished ware, is likely to be significant. Discussion The range of coarseware forms is entirely consistent with a primarily utilitarian assemblage. The relatively large drinking vessel component is probably skewed by recording methods but may also reflect the narrow chronology of the assemblage and the coincident production by local sources of certain beaker forms. Overall, the repertoire is not far removed from some Late pre-Roman Iron Age groups, with emphasis on cooking, storage and personal serving vessels. Specifically Roman vessel forms, such as flagons and mortaria are poorly represented, indicating that Roman culinary practices may not have been widely adopted. Evidence for use is relatively sparse, probably due, at least in part, to the poor surface survival. Eleven sherds preserved traces of burnt food residues, three had traces of external sooting and a further 11 exhibited white limey internal residues, probably resulting from heating or long- term storage of water. A Savernake jar (Figure 9.5) is notable for pitting and wear to the interior, possibly resulting from storage of acidic contents. Use as an impromptu strainer can be inferred for a single vessel with multiple (post-firing) perforations to its base. A perforation to the neck of a BB] jar (Figure 10.23) may have been to facilitate suspension (?for cooking). Other perforations on a fine greyware bowl (Figure 10.26) were probably for repair. Evidence of vessel repair was recorded on two Central Gaulish samian vessels, which have been repaired using shaped staple type rivets. A single South Gaulish Drag. 27 cup appears to have been trimmed at a point mid way down the wall, presumably to create a serviceable vessel after breakage. The low status of the assemblage is best demonstrated by the small amounts of amphorae and continental fineware imports. Samian represents 5.5% of the Roman assemblage (total EVEs) and only 1.2% of minimum vessel count. Such a figure is very much lower than for the equivalent period at Cirencester (Cooper 1998, 327-31) and consistent with rural sites in the region and beyond (Willis 1998). Illustration catalogue (Figures 9 and 10) The illustrated material is arranged according to Phase and with larger context groups presented together. Phase 2a/2b 4 WIL GW. Large jar with cornice-like rim. Gully 6. Phase 2b. 5 SAV GT. Large storage jar with bead rim with pitting to interior. Fence 8. Phase 2b. 6 WILBI1. Bowl, curved-sided, ‘s’-shaped and bifid rim. Gully 3. The s-shaped and bifid rim can be paralleled among examples in (reduced) Cirencester fabric 5 (Rigby 1982a, nos. 470 and 491), from lst-century AD contexts, and from Wanborough in a more comparable oxidised fabric (Seagar Smith 2001, no. 122) from a phase corresponding to AD 60-65. Phase 2a Phase 2c ditch 7 7 WIL BS. Jar with everted rim. 8 WIL OX4. Bag-shaped beaker with everted, grooved rim with traces of white-painted decoration. 9 WIL BI. Bag-shaped beaker with cornice rim with traces of white painted decoration. 10 WIL B1. Bag-shaped beaker with cornice rim. 11 WIL GW. Bag-shaped beaker with cornice rim. 12 WILBS. Jar with everted rim, imitating BB1. 13 SOW BB1. Jar with everted rim-BB1 type and burnished wavy line below rim. Phase 2c other 14 WIL GW. Dish with bead and flange rim. Ditch 1. SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 33 15 WIL OX. Carinated bow] with bifid rim. Ditch 1. 16 WIL BS. Shouldered jar/bowl. Pit 251. 17 OX WS. Flagon with disc-neck. Ditch 1. Phase2d/e ditch 7 and other 18 WIL BS. Platter. Probable scratched graffito. Ditch Te 19 WIL OX. ?cup possibly imitating Drag. 33, with bead rim. Ditch 7. 20 WIL B4. Hemispherical flanged bowl. Ditch 7. 21 WIL BS. Ring-necked flagon. Pit 620. 22 OX WS. Flagon with disc-neck. Ditch 7 23 SW BBI1. Jar with everted rim-BB1 type and repair or suspension hole. Ditch 6. 24 WIL IMIT. Bowl in imitation of samian form Drag. 30. The decoration is divided into a repeat of at least 4, probably 5 panels, separated by a wreath of vertical bifid leaves (?ears of wheat). The design is based on a saltire; a toothed wheel with eight spokes at the centre; straight wreaths of bifid leaves extending into each corner, dividing the panel into 4 quadrants. In the left and right hand segments, pairs of tendrils emanate from the wheel and drop to the bottom. The upper section has a trapezoid ornament with rounded corners infilled with dots. The lower zone has a triangular motif with its apex towards the wheel. It is clear from the variations in the repeated ornaments that these were incised freehand into the mould. Ditch 7. 25 WIL OX. Bowl imitating Drag. 36. with white painted decoration to flange upper. Ditch 7. 26 GW E Bowl imitating Drag. 36. with barbotine traces to flange. Repair hole. Ditch 7. 27 WILRS. Platter. Ditch 5. 28 WILBS. Platter. Ditch 7. 29 WIL OX. Curved-sided dish with bead-rim. Ditch 2. Unphased material 30 SOW WS. Hammer-head rim mortaria. Area C. Ditch fill 2007. 31 SOW WS. Curved-flange mortaria. Area C. Ditch fill 2007. 32 WIL BI. Tankard. Area B. Ditch 11. Ceramic building material Small quantities of Roman tile (6 fragments, weighing 1118g) were recovered from five contexts. A single tegula and box flue tile fragment were recovered from ditch 11 and a further two tegula were recovered from ditch 7, both within Phase 2d. The remaining fragments were non-diagnostic. All _ fragments are of a similar, fairly soft, pale orange fabric with common red iron oxide and rare quartz and calcareous inclusions. Although roof and flue tile is present, the quantities are too small and scattered to suggest that a substantial Romanised structure lay just beyond the excavated areas. Metalwork and fired-clay objects The published metalwork catalogue has been restricted to items of intrinsic interest or of specific chronological significance. Objects are described below according to functional category. A full object catalogue is included in the archive. Personal items (Figure 11) Cosmetic grinder — end-looped mortar (Figure 11.1) Use as cosmetic grinders, in combination with a crescentic ‘pestle’, has been suggested for this object class which appears to represent an insular British development (Stead 1986, 136; Jackson 1985, 165). A large number have been recorded (Jackson 1985), mainly from sites in eastern Britain, but extending into southern Wales and with examples from Cirencester and Wanborough. The elaborate zoomorphic decoration of some mortars indicates Late Pre-Roman Iron Age ancestry. Jackson (1985, 175-6) indicates that the class may continue into the 2nd century AD. 1 Unstratified. End-looped Mortar. Copper alloy, surviving length 51mm. Very corroded, with little of the original surface remaining. Boat-shaped, ‘U’-shaped section, with plain knob and (broken) loop terminal. Brooches Hinged iron type (Figure 11.2) Iron brooches with comparable hinged design are known from Silchester (Corney 2000, nos. 7-10), Baldock, Herts. (Stead 1986, nos. 128-30), and from the north-west Wiltshire villa site at Euridge Manor, Colerne (Mackreth 2000, nos. 7-9). The form of the bow on no. 2 is more arched and closer to Colchester derivatives known in iron, for instance from the Ditches, North Cerney, Glos. (Mackreth 1988, 45). The use of iron is likely to be a pre-Flavian characteristic (ibid). A predominantly Claudian/ Neronian date is indicated for hinged iron brooches from Silchester (Corney 2000, 322-5). 2 Pit 127. Phase 2. Hinged iron type. Iron, length 49mm. The bow is rectangular in section, curving to a solid catchplate. As far as can be discerned the pin is attached by means of a short, ‘rolled under’ tubular hinge. One-piece Nauheim derivative (Figure 11.3) The triple moulding to the bow of no. 3 is a feature of the most common form of a La Tene III type brooch. It occurs only rarely with the simple wire bow form of the Nauheim derivative class (Olivier 1988, 37). The ‘IT’ shaped head and the flattening of the bow below the moulding would appear to 34 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE be unusual. The Nauheim derivative is the most common brooch class for the Late Pre-Roman Iron Age to Early Roman period, c. AD 30-60, and occurs over the whole of southern Britain. 3. Ditch 3. Phase 2d. Nauheim derivative. Copper alloy, surviving length 34mm. Four coil spring with internal chord. Triple moulding to lower part of bow. The bow is ‘D-sectioned’, apparently cold-worked and flattened below the moulding and, at the head, to form short ‘wings’. The catchplate is missing. Colchester Derivative (Figure 11.4) The exact form of no. 4 is obscured by unusually heavy corrosion. The apparent method of attachment is by means of a bi-perforated central lug, a device which marks this out as belonging to the most common, developed form. The absence of cavetto mouldings indicates that it belongs to Crummy’s Colchester BB type (Crummy, after Hull, 1983, 12). Brooches of this general class are abundant in post- conquest contexts in sites across southern Britain and probably remain in currency into the Flavian period (Olivier 1988, 46). 4 Unstratified. Colchester derivative. Copper alloy. Length 38mm. Corrosion has removed almost all traces of the original surfaces. The pin and half of the ?10 coiled spring are missing. The bow appears to be ‘D-sectioned and plain. Polden Hill (Figure 11.5) Polden Hill brooches are distinguished by the method of retaining the spring. This is by means of closed wings, perforated to secure an axis bar and with the chord retained by a perforation on the upper face of the head. The Polden Hill class, whose distribution has a marked tendency to western areas of Britain, is, as with this example, frequently elaborately ornamented. The form may have evolved in the pre-Flavian period, but is most common in the Flavian to early 2nd-century period (Butcher 2001, 56). Side flanges at the head end of the bow are likely to be a later development in the class. However the general shape of this brooch, which is similar to two- piece Colchester forms, and the decoratively pierced catchplate, might suggest an earlier Flavian date. The form of the catchplate is similar to an example with an East Midlands provenance illustrated by Hattatt (1985, no. 374). 5 Gully 1. Phase 2c. Polden Hill. Copper alloy. Length 66mm. Complete, except for portion of pin and catchplate. Central bifid crest and side flanges are knurled. The wings each feature two double-grooved mouldings. Openwork decoration to the catchplate. Filing marks to underside of head. Trumpet-headed (Figure 12.6) No. 6 belongs to the plain trumpet form which describes examples which lack elaborate cast or enamelled decoration. Brooches of this form appear to be relatively common in the west of Britain and probably represent a regional group (Butcher 2001, 58). The disc-shaped terminal on no. 6 is prominent compared to other examples from the region. A Flavian to Trajanic date is likely for most examples of the class (Boon and Savory 1975, 54). 6 Ditch 7. Phase 2d. Plain trumpet-headed. Copper alloy. Length 52mm. Complete except for portion of pin and lower catch-plate. Pin is sprung, with an axial bar passing through the central, perforated lug. The axial bar may originally have extended above the head to form a loop. Large, complex waist moulding, the exact form of which is obscured by corrosion. This is balanced by a large disc-like foot moulding. Knee brooch (Figure 12.7) Knee brooches most commonly occur in early to middle 2nd-century AD contexts although the class probably continues in use into the early 3rd century AD. Crummy (1983, 14) asserts a distribution mainly in the west of Britain for the type. Unusually, no brooches of this type were represented among the otherwise large and varied group from Wanborough (Butcher 2001). A number of enamelled knee brooches with a similar spring arrangement are illustrated by Hattatt (2000, fig. 194). 7 Ditch 7. Phase 2d. Enamelled ‘knee’ brooch. Copper alloy. Surviving length 29mm. Pin, head loop, and lower part of catchplate missing. Square panel on bow with inner triangular panel in-filled with enamel appearing yellow green. Outer zone dark blue. Five coil spring with inferior chord mounted on axial bar passing through closed wing ends. “Top hat’ form foot terminal. Fragments (not illustrated) None of the brooch fragments listed can be identified to a particular class with any certainty and are not illustrated (n.i.). The short length of the hinged pin from ditch 7 probably indicates use with a plate brooch. n.i Ditch 7. Phase 2d. Bow brooch fragment. Copper alloy. Surviving length 24mm. Remaining portion of bow it triangular in section. Extends to a triangular catchplate and short, upturned foot. i. Colluvium 97. Bow brooch fragment. Copper alloy. Surviving length 31mm. Flat bow, tapering slightly to foot. Recess along length of back with double line of zigzag. 5 SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 50mm J Fig. 11 Small finds: personal items (scale 1:1) 35 36 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE n.i. Ditch 7. Phase 2d. Hinged brooch pin. Copper alloy. Length 29mm. Flattened end perforated for axial bar. n.i. Ditch 7. Phase 2d. Sprung brooch pin. Copper alloy. Length 40mm. Broken at first coil. Finger ring (not illustrated) The form of this ring and the use of incised decoration to the bezel are similar to an example of 2nd-century date from Wanborough (Hooley 2001, no. 91). The faceted form is widespread in the 3rd century (Phillips and Henig 2000). A number of rings, with or without a bezel but with polygonal outer hoop are also known from the locality, including an example in gold from Froxfield, Wiltshire (zbid.) and two in copper alloy from Wanborough (Hooley 2001, 92). n.i. Ditch 9. Phase 2d. Finger ring with expanded bezel and polygonal hoop. Copper alloy. Complete small ring. Hoop is rounded and slightly flattened — length (internal) 17mm, depth 16mm. Polygonal outer hoop is 9-sided, including top-most facet which serves as the bezel. Width tapers from bezel/shoulder to bottom of hoop. The bezel is rectangular with lines of zigzag above and below a horizontal groove and within a plain border. Tools Knife (not illustrated) A single knife blade was recovered of Manning’s (1985, 115) type 15. This is seemingly the most common of Roman forms and was current throughout the period. n.i. Ditch 3. Phase 2d. Tanged knife fragment (Manning type 15). Iron. Surviving length 75mm. The back is straight, the cutting edge straight and angled upwards to the (absent) tip. Tang (also broken) is square sectioned and central to blade. Crafts (Figure 12) ? Loomweight (Figure 12.8) The bar-like fired clay object no. 8 derives from a mid to late 2nd century ditch 7. It has a soft and pale brown fabric, with abundant chalk-like inclusions and frequent voids from vegetable tempering. In the absence of a perforation, identification is tentative. The closest parallel for this and a similar but more fragmentary example from the same context is from Baldock, Herts., where similar items are described as ‘tall rectangular’ form loomweights (Foster 1986, no. 689). Use as household or possibly thatch weights is also possible. 8 Ditch 7. Phase 2d. ?Rectangular formed weight. Fired clay. Bar-like, expanding towards upper (broken) end. Width at ‘base’ 60mm. Surviving height 95mm. Objects associated with hunting Slingshot (Figure 12.9) Two fired clay objects interpreted as sling missiles were recovered. Both examples are in a similar, hard, inclusionless fabric. Similar objects are known from Late Iron Age and earlier Romano-British contexts across Britain. Examples from the region include Glastonbury (Bulleid and Grey 1917) and Roughground Farm, Lechlade, Glos. (Palmer and Allen 1993, fig. 108, nos. 136-8). The size and comparative fragility of such items indicate use for hunting of game or possibly for defence against animal predators. 9 Pit 328. Phase 2c. Oval/bi-conical slingshot. Fired clay. Length 40mm. Essentially complete, but with some spalling to one side. n.i. Ditch 1. Phase 2c. ?Oval slingshot. Fired clay. Fragmentary. Largest fragment 16mm. Household items Vessel fragment with XRF analysis by Dr B. Gilmour (not illustrated) The form of this fragment is likely to have been a small bowl, although it is difficult to be certain given its small size. A colander from Colchester (Crummy 1983, no. 2029) has a similar rim form and is of comparable size. The presence of tinning on this fragment, however, probably indicates a more decorative use. n.i. Ditch 7. Phase 2d. Copper alloy with tinning. XRF reveals metal to be heavily leaded bronze. Vessel rim fragment — Diam. approx 140mm. Flat top rim from small ?bowl. Rim edge is bent-over rim. Traces of tnning below rim only. Thickness 0.4mm. Coins Catalogue 1 Ditch 1. Phase 2c. As of Domitian AD 81-96. Reverse: Fortuna with rudder and cornucopia (Poor preservation precludes RIC identification). 2 Unstratified Late Roman bronze coin of Gloria Exercitus type, dateable to AD 330-5 3. Unstratified George II halfpenny dateable to AD 1740- 54 Glass by H.E.M. Cool (not illustrated) All four of the fragments of glass are of Roman date. The mould blown fragment from the primary fill of gully 4 can be dated to the third quarter of the Ist SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 37 0 50mm = Fig. 12 Small finds: personal items and crafts (scale 1:1) 38 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE century AD. It probably came from a cylindrical beaker. Similar rows of pellets between bordering ribs occur on a blue/green beaker from the Neronian site at Kingsholm (Price and Cool 1985, 46, no. 18). On the Kingsholm vessel one of the pellet rows is bordered by a zone of lattice decoration and the tip of a V-shaped moulding seen on the Showell Farm piece might hint that a similar decoration was also present on this vessel. This is an unusual discovery because Ist-century mould blown vessels are not normally found on rural sites. The form of the vessel from the upper fill of ditch 7 cannot be identified, but the colour and ribbed decoration would be consistent with a lst to mid 2nd-century AD date as a variety of common forms of that date have this combination of features (see for example Price and Cottam 1998, 78, 137 and 152). There is also one fragment (test-pit 26) from a blue/green prismatic bottle, a very common Ist to mid 3rd-century form (Price and Cottam 1998, 194- 9), frequently found on rural sites. The remaining blue/green body fragment cannot be more closely dated than to the Ist to 3rd century. Catalogue 1. Gully 4. Phase 2a. Body fragment. Deep blue; mould blown. Straight side. Moulded decoration consists of a band of circular pellets (3 extant) between two horizontal ribs; the tip of a V-shaped moulding remains beyond one of the ribs. Dimensions 23mm x 17mm, wall thickness 1.5mm 2. Ditch 7. Phase 2d. Body fragment. Light green. Two ribs. Dimensions 16mm x 14mm, wall thickness 2mm 3. Area A, Test pit 26. Prismatic bottle; body fragment. Blue/green 4. Unstratified. Body fragment; blue/green Metallurgical residues A very small quantity of metallurgical residues (208g) was recovered from three contexts assigned to Period 2. This material would appear to be of Table 4: Radiocarbon dating results very similar composition, characteristically dense with low vesicularity. A fragment from gully 575 features a smooth, convex underside and is presumed to represent a fragmentary smithing hearth bottom. The material represents limited evidence for early Roman ferrous metalworking activity, most likely in the form of small-scale smithing. BIOLOGICAL REMAINS Radiocarbon Dating by Annette Hancocks Four radiocarbon measurements were obtained from samples at Showell Farm. Three samples were subjected to AMS dating. A further sample (Wk-15451), from Grave 2 proved to be unsuitable for analysis as the bone contained insufficient nitrogen. Samples were processed during 2004 at the University of Waikato Radiocarbon Dating Laboratory. For details of methods see University of Waikato Radiocarbon Dating Laboratory Website (2004). Results and Calibration The results are conventional radiocarbon ages (Stuiver and Polach 1977) and simple calibrations of the results are given in Table 4. All have been calculated using the calibration curve of Stuiver et al (1998) and the computer program OXCal 3.9 Bronk Ramsey (2003). Date ranges cited in the text are those at 95% confidence level unless otherwise specified. Ranges are derived from the probability method (Stuiver and Reimer 1993). Charred Plant Remains by Wendy fF. Carruthers Soil samples were taken for the recovery of Faber Type | Context | Period | Description | Material used exis oh Ge doieouinlelieoiane (at2 Wk 15448 | AMS |422. | Ring Ditch 1 | Hazelnut 3822+/-38 | 2460-2140 cal BC Wk 15449 | AMS | 420, 1 Ring Ditch 1 | Hulled Barley | 3768+/-38 | 2300-2030 cal BC Wk 15450 | AMS | Skeleton | oq | Grave 1 Human left tibia 2040+/-37 | 170 cal BC-60 cal AD Wk 15451 |RM | $Keleton [oq | Grave 2 Bone n/a <0.5%N_ | n/a SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 39 environmental remains and processed using standard methods of flotation (flot mesh = 250 microns, residue sieved to 500 microns). Charred plant remains recovered from the samples are listed in Table 5. Nomenclature and most of the habitat information is taken from Stace (1997). Charred plant remains were scarce with concentrations of less than six fragments per litre (fpl) occurring in twelve of the twenty fully analysed samples, and were not particularly well preserved; many were encrusted with fine silt making identification difficult. The samples concerned were primarily from gully, ditch and pit fills. A few concentrated deposits of burnt cereal processing waste and processed grain were recovered from fence 8, posthole 270 and drying oven 1. The results are discussed in more detail below by period. Period I Early Bronze Age ring ditch 1 Of the two samples assessed from ring ditch 1, only one <43> produced charred plant macrofossils consisting of a few fragments of hazelnut shell (Corylus avellana) and several poorly preserved cereal grains. Both emmer/spelt wheat (Jiiticum dicoccum/spelta) and barley (Hordeum sp.) were being cultivated, although the barley was too poorly preserved to tell whether it was hulled or naked. Barley was the most frequent cereal (22 grains compared with 3 wheat grains), as is typical of most Bronze Age sites. However, the results from a single small sample are inadequate to determine the importance of the different crop plants. Since no chaff or weed seeds were present, these remains _ probably derived from burnt domestic food waste. A fragment of hazel nut and a hulled barley seed were subsequently submitted for radiocarbon dating (Table 4). Period 2 Romano-British features Phase 2a Several samples examined from Period 2 features produced fairly low levels of background charred plant material (< 6 fpl). These derived from four miscellaneous features (a gully and three pits) and gully 4. The assemblages consisted of a few cereal grains, sparse chaff fragments and several weed seeds. Gully 4 <47> contained some dilute cereal - processing waste (2.5 fpl). The presence of two sheep’s sorrel seeds (Rumex acetosella) and a sedge nutlet (Carex sp.) suggest that some acidic and damp soils may have been cultivated. Many of the more clayey soils in the Chippenham area are seasonally waterlogged, particularly in the Avon valley. It is possible that these remains became charred amongst hay from damp meadows along the valley bottom. Most of the soils in the Cotswolds are calcareous, but some areas of sandy soil exist to the south-east of Chippenham. Samples 12-15 from pit 456 produced large quantities of charcoal, burnt bone, a few emmer/spelt wheat grains, a chaff fragment and several weed seeds. The cereal remains may represent all that survived from an offering of grain, waste that had been used as tinder, or redeposited material in the backfill of the pit. The main component of each of the four samples was dock seeds (Rumex sp.), amounting to 47 nutlets (69% of the charred remains). Docks grow in a wide variety of disturbed soils, including cultivated land and wastegrounds. They are particularly frequent on nutrient-enriched soils. Phase 2b Sample <1> came from fence 8. A small concentration of cereal processing waste (CPW) was recovered from this sample (18.3 fpl), comprising a few cereal grains, frequent hulled wheat chaff fragments and one or two weed seeds. The ratio of grain to chaff to weed seeds was 2:58:1, confirming the interpretation as cereal processing waste. Both emmer (Triticum dicoccum) and spelt wheat (T. spelta) were represented, although spelt was by far the most abundant cereal identified. Since spelt glume bases are more robust than emmer chaff there may be some taphonomic bias in operation. However, spelt wheat had taken over as the principal cereal being cultivated by the Romano-British period (Greig 1991). As on most sites of this period, small quantities of barley were present in three of the nine samples. It is likely that this crop was primarily used for fodder, so it may be under-represented in charred plant assemblages. Phase 2c Posthole 270 (sample 4) produced the highest concentration of charred plant remains out of all twenty samples discussed (87.3 fpl). The ratio of grain to chaff to weed seeds was 1:3:3, indicating that cereal processing waste is present. Although the number of grains being lost with the waste appears to be high, experimental work has shown that grains are far more readily preserved than chaff (Boardman and Jones 1990), so this preservation bias needs to be taken into account. As in the Phase 2b sample of cereal processing waste, spelt chaff was significantly more frequent than emmer chaff, and barley was present in small quantities. However, weed seeds were much more 40 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 5: Charred Plant Remains KEY: Ee auure types : C = cremation; D = ditch; F = fence; G = gully; DO = drying oven; P = pit; PH = posthole; RD = ring ditc ; : Habitat Preferences : A = arable; C = cultivated; D = disturbed/waste; E = heath; G = grassland; H = hedgerow; M = marsh/bog; R = rivers/ditches/ponds; S$ = scrub; W = woods; Y = waysides/hedgerows; a = acidic soils; c = calcareous soils; n = nutrient-rich soils; o = open ground; d = damp soils; y = dry soils; * = plant of economic value Sample} 43 47 | 12 13 14 15 34 1 2 49 4 Context | 699 | 702) 457 | 459 | 461 | 463 | 352 | 111) 148 | 704) 271 Feature no. & type RD 1 | G4 | P456 | P456 | P456 | P456 | P308 | F8 | P147| F8 | PH270 TAXA Phase 1 2a_| 2a 2a 2a 2a 2a_| 2b | 2b | 2b 2c Cereals : | Triticum dicoccum/spelta (emmer/spelt wheat grain) 3 3 2 Z 1 1 10 Triticum dicoccum/spelta (emmer/spelt wheat grain, sprouted) Triticum sp. (wheat grain NFI) 2 1 1 1 Hordeum sp. (hulled barley grain) 1 1 Hordeum sp. (barley grain) 22 2 6 1 1 Avena sp. (wild/cultivated oat grain) Avena/Bromus sp. (oat/chess grain) 1 [ 1 Indeterminate cereals of 9 1 1 13 1 4 36 Chaff : Triticum spelta L. (spelt glume base) 1 | 2 59 23 Triticum spelta L. (spelt spikelet fork) Al Triticum spelta L. (spelt rachis fragment) 1 Triticum dicoccum (emmer glume base) 3 4 Triticum dicoccum (emmer spikelet fork) eae 2 Triticum dicoccum / spelta (emmer / spelt glume base) 26 1 3 4 4 107 | 126 Triticum dicoccum / spelta (emmer / spelt spikelet fork) 4 6 Triticum dicoccum / spelta (emmer / spelt rachis fragment) Hordeum sp. (barley rachis frag.) Z 1 Avena sp. (oat awn fragment) + | ee Cereal sprout 1 Weeds : il| Chenopodium album L. (fat hen seed) CDn Atriplex patula/prostrata (orache seed) CDn 1 1 Chenopodiaceae embryo 1 Stellaria media (chickweed seed) CDn Montia fontana ssp. chondrosperma (blinks seed) Gd Corylus avellana L. (hazel nut shell frag.) HSW* | 4 | 1 Polygonum aviculare L. (knotgrass achene) CD Fallopia convolvulus (L.) A.Love (black-bindweed achene) 1 1 2 AD Rumex acetosella agg. (sheep’s sorrel achene) EGa 2 Rumex sp. (dock achene) CDG 8 14 20 7 6 4 Malva sylvestris. (common mallow nutlet, with capsule l frag.) DG 1B. Maiva sp. (mallow nutlet) DG Rosa sp. (rose seed) HSW* Trifolium/Lotus sp. (clover/trefoil) DG 2 Vicia/Lathyrus sp. (c.2mm, small seeded weed vetch/tare) 1 1 1 2 l l CDG =| 4 Linum catharticum L. (purging flax seed) Gy (é Galium aparine L. (cleavers) CDH 1 1 + Galium palustre L. (common marsh-bedstraw) GAPMF Plantago laceolata L.(ribwort plantain) Go Odontites verna/Euphrasia sp. (red bartsia/eyebright) CD 1 Tripleurospermum inodorum (L.)Schultz-Bip. (scentless 3 2 mayweed achene) CD Rhinanthus sp. (yellow rattle etc. achene) G f Primulaceae (scarlet pimpernel etc. seed) Eleocharis subg. Palustres (spike-rush nutlet) MPd Carex sp. (trigonous sedge nutlet) MPd | Carex sp. (lenticular sedge nutlet) MPd 1 Bromus sect. Bromus (chess caryopsis) ADG 4 1 2 3 Lolium perenne-type grass caryopsis B) 1 Z 3 Long seeded grass caryopsis Poaceae (small seeded grass caryopsis) CDG 2 9 Zi Total charred remains: | 36 | 75 | 28 24 8 Sample size (litres): | 30 | 30 5 5 5 10 | 10 4 Fragments per litre: | 1.2 | 2.5 | 5.6 | 48 | 1.6 | 2.7 | 16 | 1.1 | 0.6 | 18.3] 87.3 Ww }00 w i—) _ i—) SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 4] Sample Context Feature no. & type DO1 DOI 20 366 DO1 677 P675 All others TAXA Phase 2d 25 samples Cereals : Triticum dicoccum/spelta (emmer/spelt wheat grain) 14 Triticum dicoccum/spelta (emmer/spelt wheat grain, sprouted) Triticum sp. (wheat grain NFI) Hordeum sp. (hulled barley grain) Hordeum sp. (barley grain) Avena sp. (wild/cultivated oat grain) Avena/Bromus sp. (oat/chess. grain) Indeterminate cereals 21 61 15 13 17 Chaff : Triticum spelta L. (spelt glume base) 15 14 Triticum spelta L. (spelt spikelet fork) Triticum spelta L. (spelt rachis fragment) Triticum dicoccum (emmer glume base) Triticum dicoccum (emmer spikelet fork) Triticum dicoccum / spelta (emmer / spelt glume base) 19 20 Triticum dicoccum / spelta (emmer / spelt spikelet fork) Triticum dicoccum / spelta (emmer / spelt rachis fragment) Hordeum sp. (barley rachis frag.) Avena sp. (oat awn fragment) Cereal sprout Weeds : Chenopodium album L. (fat hen seed) CDn Atriplex patula/prostrata (orache seed) CDn Chenopodiaceae embryo Stellaria media (chickweed seed) CDn Montia fontana ssp. chondrosperma (blinks seed) Gd Corylus avellana L. (hazel nut shell frag.) HSW* Polygonum aviculare L. (knotgrass achene) CD Fallopia convolvulus (L.) A.Love (black-bindweed achene) AD | Rumex sp. (dock achene) CDG Rumex acetosella agg. (sheep’s sorrel achene) EGa 23. Malva sylvestris. (common mallow nutlet, with capsule frag.) DG Malva sp. (mallow nutlet) DG Rosa sp. (rose seed) HSW* Trifolium/Lotus sp. (clover/trefoil) DG Vicia/Lathyrus sp. (c.2mm, small seeded weed vetch/tare) CDG Linum catharticum L. (purging flax seed) Gy Galium aparine L. (cleavers) CDH Galium palustre L. (common marsh-bedstraw) GAPMF Plantago laceolata L.(ribwort plantain) Go Odontites verna/Euphrasia sp. (red bartsia/eyebright) CD Tripleurospermum inodorum (L.)Schultz-Bip. (scentless mayweed achene) CD Rhinanthus sp. (yellow rattle etc. achene) G Primulaceae (scarlet pimpernel etc. seed) Cf. Eleocharis subg. Palustres (spike-rush nutlet) MPd Carex sp. (trigonous sedge nutlet) MPd Carex sp. (lenticular sedge nutlet) MPd Bromus sect. Bromus (chess caryopsis) ADG Lolium perenne-type grass caryopsis | Long seeded grass caryopsis Poaceae (small seeded grass caryopsis) CDG 15 Total charred remains: 89 88 10 10 10 Sample size: Fragments per litre: 8.9 8.8 9:9. 74 42 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE frequent in this sample, comprising at least six taxa. In addition to some common weeds of cultivated and disturbed soils, several seeds of common mallow (Malva sylvestris) were recovered. This native plant can be eaten as a vegetable and the seeds make a tasty snack (Mabey 1972). Mallow appears to have been popular in Roman times, since it is much more frequently recovered from assemblages of this date. A mallow seed was also present in a Phase 2b sample 34 from pit 308. Seeds tentatively identified as perennial rye grass (Lolium perenne-type) were remarkably frequent in this single Phase 2c sample (59, plus 57 less well-preserved ‘long-seeded grass’ seeds). This large-seeded weedy grass is particularly frequent in Late Iron Age to Late Roman cereal processing waste deposits in central and southern England (Carruthers 2002); along with chess (Bromus sect. Bromus) which is abundant on some sites of this period, it was tolerated to some extent (Campbell 2000). Alternatively it may simply have been too difficult to eradicate from spelt crops that were being cultivated more intensively during this time. Phase 2d Two samples from ditches 7 (sample 45) and ditch 11 (sample 5) produced low concentrations (3.4 fpl and 2.4 fpl) of cereals, chaff and weed seeds in roughly equal proportions of grain to contaminants. These charred remains were probably washed into the ditches in run-off from fields that had been manured with domestic waste. Alternatively, some burnt material may have been deposited as waste, or in situ burning of vegetation may have taken place. Two of the taxa identified are indicative of damp soils, such as might occur in a ditch; sedge and cf. spike-rush (cf. Eleocharis subg. Palustres). Domestic waste is equally likely to contain these taxa if hay and bedding were being burnt. The most productive samples from this Phase came from a drying oven 1. Six samples from this feature were examined, two of which produced deposits primarily composed of grain, and four containing cereal processing waste. Two samples (16 and 17; Table 5) came from context 526, an ashy deposit recovered from the top end of the T-shape flue (north-west end). These samples consisted of 99% and 98% grain, respectively, the few identifiable grains being almost equally emmer/spelt wheat and barley (30:28 grains). It is difficult to tell whether the increased recovery of barley from this feature is due to more barley being grown during this phase, or preservation factors. Barley is more likely to become charred in an oven than in a domestic hearth, since it requires parching if it is going to be dehusked or malted, and possibly prior to storage. Too few ‘non-oven’ features were examined from this phase to make comparisons with the Phase 2 features, although ditch 7 (sample 45) did produce more barley than emmer/spelt grains. Two samples derived from the stoking end of the oven, with sample <21> from directly above a stone slab (366) and sample <20> from alongside the slab. A further two samples <22> and <23> derived from above the stone slab and from the primary fill of the oven, below the stone slab. Most of the charred plant remains came from above and beside the stone slab (9.9, 10.9 and 12.4 fpl compared with 2.8 fpl from 23). All four samples (20-23) contained more chaff than grain or weed seeds, indicating that they represented the burnt remains of cereal processing waste used as fuel. Further confirmation of this was the presence of some hay meadow/grassland taxa, including purging flax (Linum catharticum), yellow rattle (Rhinanthus sp.), marsh bedstraw (Galium palustre) and ribwort plantain (Plantago lanceolata). This, too, had probably been used as tinder. Almost all of the well-preserved glume bases in the cereal processing waste were from spelt wheat, with only one being identifiable as emmer wheat. This could indicate a further reduction in the cultivation of emmer in favour of spelt, or poorer conditions of preservation in the oven favouring the survival of the more robust remains. A total of three oat grains were recovered from the 45 samples submitted for assessment, including one from sample 21. Itis not possible to determine whether oats were growing as weeds or a minor crop, as floret bases were not preserved. Since sample 21 consisted of cereal processing waste perhaps the former is more likely. However, all three were found in Phase 2d samples, so it is likely that oats were being introduced as a crop at this time. In her review of the archaeobotanical evidence from Romano-British corn driers, van der Veen (1989) found that in 45% of cases chaff had been used for fuel. The crop plants recorded from all deposits samples were as follows: e in 72% at least some spelt was identified; e in 41% barley was present; ° 36% produced both spelt and barley, or unspecified ‘wheat’ and barley. Van der Veen (ibid.) suggests that such drying ovens would have performed a variety of functions, including drying prior to storage and drying prior to husk removal and malting. Only three of the SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 43 twenty-one ovens included in the review produced definite evidence of malting, i.e. large quantities of sprouted spelt grains. At Showell Farm all four cereal processing waste samples produced some sprouted emmer/spelt grains, and samples 20 and 21 contained a few detached sprouts. Unfortunately few of the grains in the six oven samples were well- enough preserved to determine whether or not they had sprouted, so it is not possible to confirm that the sprouting was deliberate, but this in itself could indicate that malting had taken place, as distortion of the grain occurs during the growth of the sprout. The grain in the T-end of the oven did not show signs of sprouting, so it seems that the final use of the oven had probably not been for malting. There is evidence for the use of spelt in the production of malt during the Roman period (Hillman 1982), but by the medieval period barley became the preferred cereal. A single sample (37) was taken from pit 675. A surprisingly high concentration of charred plant remains (25.5 fpl) was recovered from the small soil sample, including grain, chaff and weed seeds at a ratio of 5:1:3. A rose seed was also recovered from the sample. Conclusions The Bronze Age assemblage from ring ditch 1 and the Romano-British assemblages from various features were all typical of the periods covered. Whilst barley was probably the main crop during the Bronze Age, spelt wheat was the principal crop recovered during most of the Roman period, with some emmer continuing in cultivation for at least the earlier period, and barley being a constant though minor crop. Oats may have come into cultivation in Phases 2d/e, perhaps being grown as a fodder crop, like barley. Oats and barley (‘dredge’) have also been used to produce malt in the medieval period. Unlike some larger sites, particularly those with military connections, no bread-type wheat was recorded. This could indicate that the settlement was fairly Table 6: Age, sex, preservation and pathology. low status. The Phase 2d drying oven 1 produced evidence for drying spelt wheat and barley, with the suggestion that the oven had sometimes been used to produce malt. The only evidence for other foods was a trace of hazelnut shell from the Bronze Age ring ditch 1; the common mallow from the Phase 2a pit 308 and the Phase 2c posthole 270, and a rose seed from the Phase 2d pit 675. Human remains by Teresa Gilmore The human remains consisted of two inhumations and two cremation deposits. In addition, a single isolated heel bone was recovered (Table 6). The two inhumations have been subject to fragmentation. Skeleton 407 (Grave 2; Phase 2d) was the least well preserved, with moderate loss of integral structure and severe loss of cortex. Skeleton 662 (Grave 1; Phase 2d) had the best quality of bone including preservation of the cortex. Table 5 shows the age, sex and pathological information gained from the skeletal remains. Three adults are represented (one Probable Male, two of unknown sex). Pathological changes could only be noted from one individual (Skeleton 662, Grave 1). No diagnosis of conditions leading to pathological changes was undertaken at this level of analysis. Cremated bone was weighed and measured from five environmental samples, representing two deposits (Table 7, Cremations | and 2). The state of oxidation and hence the quality of cremation varies from white/grey fragments to black/grey fragments. The remains appear to represent a minimum number of five individuals (two inhumations, two cremations and a single isolated heel bone). The heel bone is of interest as it derives from ring ditch 1 and raises the possibility of a disturbed Early Bronze Age funerary deposit. The only pathology present was N/A = sex not determinable due to juvenile bone. Unknown = Appropriate criteria not present to assign Ranier Phase | Feature Age 699 ak Ring ditch | | Adult eal 662 2a Grave | | Adult | U E 407_ [Pret Grave 2 | Young Adult / Adult | ?M sex. ?Male / ?>Female = Demonstrated sexing traits not fully male or female iises | Preservation | patnolocy <1% [F ragmentary heel bone only from sample <43> | Dental caries in lower PMI. Calculus on buccal | 359 ees | and lingual crown surfaces on loose teeth 15 % 4 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 7: Weight, recognisable fragments and approximate dimensions of cremated bone samples. Cremation Context | Phase | Weight (g) | Represented anatomy Fragment size (largest) - mm R. Patella, cranial vault, distal humerus | 441 2c 102 long bones, vertebral neural arch ” | 32 by 25 Cranial vault, L. Mastoid Process, L. patella, 1 443 2¢ 394 1“ metacarpal, petrous portion, long bones, | 50 by 18 tooth root 1 445 I 160 Een bone, cranial vault, hand phalanx, 35 by 14 1 447 2c 58 Long bone, cranial vault 30 by 14 2 107 Ic 188 eepee © humeral head, cranial vault, 34 by 12 calculus on the lingual and buccal surfaces of the loose dentition of grave | with one tooth exhibiting a case of dental caries along the gum line on the distal surface. Animal bone by Lorrain Higbee The entire assemblage was subjected to assessment by rapid scanning (after Davis 1992) and the recording of the following information; species, skeletal element, age related features, completeness for biometric analysis, as well as more general observations on butchery, taphonomy and pathology. This information was entered into a database and is available in the site archive. The majority of the assemblage is well-preserved. Only a small number of bones were recorded as poorly preserved and these tended to be undiagnostic fragments or fragments which could only be identified to general size categories. Only a small proportion of bones were recorded with canid gnaw marks and this did not generally affect identification. Charred and calcined bones was also recovered but generally too fragmented to identify. A total of 312 fragments of animal bone, the majority of which are from deposits assigned to Period 2, Phase 2d/e (later 2nd to early 3rd century) was recovered. Only 181 (58%) could be identified to species (Table 8). A small proportion of the bones are sufficiently complete enough to provide more detailed information (see archive). The majority of the remaining 131 fragments are either diagnostic non-countable bones (ibid.,) or bones which could only be assigned to a general size category (e.g. “cattle-sized”); a small proportion are undiagnostic fragments greater than lcm in size (anything smaller was not counted). Cattle mandibles are well represented and butchery evidence noted on scapulae in the assemblage is characteristically Roman. The most obvious example is a scapula from ditch fill 1011 in trench | of Area C which bears butchery marks consistent with processing the joint to produce cured (or salt) beef. Such marks include trimming around the glenoid cavity, removal of the spina, and nick marks along the margo thoracalis. This pattern of butchery has been recorded on scapulae from a number of Roman assemblages (Maltby 1985; 1989) and probably represents brined or cold-smoked joints (Dobney et al 1995; Dobney 2001). Seventeen horse bones were recovered, mostly as isolated anatomical elements scattered across the site. Loose teeth and pelvises are common and both immature and adult individuals are represented. One of the adult horse bones, a metatarsal from the primary fill of ditch 7 is complete and gave an estimated shoulder height of c. 12 hands (lateral length measurement following von den Driesch 1976 converted using Kiesewalter’s factors in von den Driesch and Boessneck 1974). Sheep/or goat is represented mostly by mandible fragments and loose teeth. Age data based upon tooth eruption and wear suggests that most sheep were culled as adult individuals (age category G following Payne 1973; 1987; 1988). Less common species include pig, dog, chicken, amphibian and Table 8: Number of specimens identified to species (or NISP) includes both hand-recovered bone and bone from sample residues. Species Bbaee Rpase ae Undated | Total Cattle 16 3 41 2, 62 Sheep/Goat 25 18 36 1 80 Pig 3 2 5 - 10 Horse 1 2 14 = ell) Dog 1 - 4 - 5 Chicken 1 - 2 - 3 Amphibian - - 3 - 3 Rodent - - 1 - 1 Total 47 25 106 3 181 SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 45 rodent (vole). All microfauna (i.e. amphibian and rodent) bones were recovered from sample residues and probably represent natural fatalities. Given the size of the animal bone assemblage no meaningful interpretation can be ascertained, but the overall make-up, nature and character would appear to be typical of occupation of small-scale domestic rural settlement DISCUSSION by Annette Hancocks and Richard Young Mesolithic The earliest evidence from the site relates to Mesolithic activity in the form of worked flint recovered as a result of topsoil sieving. The volume of material found, and its location, on a terrace beside a river is suggestive of a short term camp with no evidence for long-term occupation. The Mesolithic activity forms part of a broader pattern of flint debris in the area identified during earlier work in the vicinity (Walker 2000, 236), and the Southern Cotswolds generally (Saville 1984). 3rd millennium BC Structural evidence dated to the late 3rd millennium BC comprised two ring ditches, one of which (ring ditch 1) is more correctly described as a pennanular ditch. These are interpreted here as funerary or ritual monuments rather than roundhouses given that a human heel bone, two sherds of Beaker pottery, a few charred plant remains and a broken leaf-shaped arrowhead were recovered from the infilling of ring ditch 1. The lack of associated domestic activity also supports the funerary interpretation. The position of the ring ditches, on a promontory overlooking a tributary of the river Avon adds further substance to their interpretation. The location is comparable with that of other well-documented barrows (Powesland 2004, 82). Ring ditch 1 was the more complete of the two identified, with a probable north-east facing entrance and dates to the late 3rd millennium BC. Pennanular ring ditches surrounding round barrows occur ‘mainly in Dorset and Southern Wessex, although a scatter is known elsewhere. However, there is no central grave and associated pit associated with ring ditch 1. The pit is usually deeper than the ditch. The north-east entrance to the mid-summer rising sun may also be significant, and this monument could represent a small Class I single entrance henge. A broadly comparable hengiform ring ditch was found at Shorncote in the Upper Thames Valley. Here a multi-phased pennanular ring ditch with north- facing entrance was excavated (Hearne and Adam 1999, fig. 9 and 33). The single entranced henge is similar to a number of hengiform enclosures of Neolithic date found on the gravel terraces of the Thames and Avon valleys (Loveday 1989, 51-85) and similar examples have recently come to light at Barrow Hills, Radley (Barclay and Halpin 1999, 35-43) and Gravelly Guy, Oxon (Lambrick and Allen 2004, 45-51). At Barrow Hills, a continuous hengiform ring ditch 611 had an external diameter of 6.5m and two samples of modified antler bone produced radiocarbon dates of 2900-2200 cal. BC and 2600-2000 cal. BC respectively and were found in association with plain grooved ware, struck flint of Later Neolithic character and a fragment of Group IV stone axe. A subsequent cremation was radiocarbon dated to 2040-1640 cal. BC. The Gravelly Guy ring ditch was 10m in diameter with a narrow north-west facing causeway, marginally smaller than the Showell Farm example, but Late Neolithic/Early Bronze Age pottery, struck flint and charcoal were recovered. The latter was radiocarbon dated to 1870-1520 cal. BC. Although poorly preserved, ring ditch 2, at Showell Farm, is best considered to be the quarry ditch surrounding a small round barrow. Such monuments are typical of the southern Cotswolds and this example may be part of a small cemetery (See figure 2; Darvill 1987, 95-9). The lack of internal features within the ring ditches is likely a result of truncation by modern ploughing. Radiocarbon dating confirms the late third millennium date of ring ditch 1, a date supported by the recovery of Beaker pottery. The funerary monument interpretation is somewhat tenuously supported by a single aerial photograph taken in February 1949 which suggests that the excavated ring ditches possibly form part of a barrow cemetery (RAF/541/222/4072; Fig. 2). If the anomalies on the aerial photographs have been interpreted correctly further ring ditches should have been observed during the evaluation (trench 12) and throughout Area B. The absence of these features might be explained as a consequence of severe truncation on this part of the site over the last fifty-five years by modern ploughing, but the failure of the excavated ring ditch to show as a cropmark on the photograph may cast some doubt on the validity of the interpretation. It may also bea result 46 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE of localised geological conditions, Kellaways clay not being conducive to the production of cropmarks. It is pertinent to remember that the identification of barrows, whether as earthworks or ring ditches, is always slightly hazardous. Darvill and Grinsell (1989, 49) have noted in Gloucestershire, for example, that less than 50% of putative round barrows that have been excavated proved to be genuine barrows, while the figure was less than 20% for ring ditches. Late Neolithic and Bronze Age settlement has previously been suggested both to the north and the west of Showell Farm. Evaluation at Showell Nurseries revealed pits, postholes, gullies and ditches dating to this period (OAU 1991, Anon 1993, 159), while investigations along the Chippenham bypass noted a concentration of flint of the same date and sherds of Bronze Age pottery (Bateman and Enright 2000, 233-54). These discoveries testify to an increasingly dense pattern of Late Neolithic and Early Bronze Age activity in the Chippenham area. Tron Age At Showell Farm there is limited evidence for Iron Age activity with four sherds of pottery recovered in total, although a Middle Iron Age pit and ditch was recorded at nearby Showell Nurseries (OAU 1991). The skeleton in grave | was also radiocarbon dated to the Late Iron Age or Early Roman period. Roman The Romano-British evidence comprises a series of intercutting fields and trackways that do not respect the earlier ring ditches of Period 1. The latter features must have been ploughed flat by the time the Roman agricultural field system was laid out in Phase 2d. Activity started c. AD 50-100 with no evidence for continuity from the late pre-Roman Iron Age (in contrast to the Upper Thames Valley). The site was abandoned by the early 3rd century as there is an absence of later pottery, although a 4th- century coin was recovered from the topsoil. The low status of the ceramic assemblage is demonstrated by the small amounts of amphorae and continental fineware imports (McSloy above). This is consistent with rural sites in the region (Willis 1998). The majority of the ceramics, two tegula and small finds evidence were recovered from ditch 7. This concentration of domestic waste may imply that the main occupation area was centred a short distance to the north-west of this feature. The results suggest that the Romano-British activity represents a small-scale farmed landscape with settlement foci to the north-west of the excavated Roman field system. The system appears to form part of an early Roman managed agricultural landscape of fields, paddocks, enclosures and droveways relating to a mixed agricultural regime incorporating stock management, with the occasional outlying feature such as the drying oven. Most drying ovens are known on or adjacent to villa sites, mainly in the 4th century. However, more are known in lesser rural settlements during the earlier Roman period. The T-shaped form of the Showell Farm drying oven is typically 2nd century in date and typical too of the form of oven previously recorded in this region. A close parallel is the drying oven from Woodcuts 4 (Dorset), where the main flue was c. 2.50m, with a short cross flue with rectangular ends (Morris 1979, 159). The charred plant remains recovered from the oven are indicative of the processing of spelt wheat. This form of drying oven is seen to be multi-functional, both in the roasting of germinated grains for the production of malt and the parching and drying of grain in preparation for consumption and storage. T-shaped drying ovens are common in Wiltshire at both villas and farmsteads (Walters 2001, 133-5). Although charring is clearly a very biased form of preservation, the diet at Showell Farm appears to have been fairly limited and rural in nature. No evidence for bread wheat, imported fruits and spices, or orchard crops was found. The limited animal bone assemblage comprised cattle, sheep and horse and adds little to our knowledge of economy or diet. The crouched inhumation was the principal (archaeologically visible) burial rite in the pre- conquest period in Wiltshire and this pattern continues, with some modifications into the Roman period (Whimster 1981, 195 and Philpott 1991, 6). This fits in well with the radiocarbon date obtained on the skeleton from grave 1 at Showell Farm, although rare extended inhumations have been found in pre-conquest or early post-conquest contexts (ibid, 53). The two cremations identified (Cl and C2) conform to the change in burial practices after the Roman conquest in AD 43 and would appear to be classified as unenclosed due to the lack of both a durable cinerary container and grave furniture, although any organic container may have subsequently perished (ibid, 45). Cremations are known in Wiltshire for this period (ibid, 55.) The dispersed nature of the burials and their association with the trackways and ditches at the edge of Romano-British activity seems to represent a more SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 47 informal pattern of burial commonly identified at the fringes of Romano-British rural settlements (Pearce 1999, 151). What has not been discovered is the associated agricultural buildings and possible villa around which this agricultural landscape was formed. A few tantalising glimpses suggest that a villa-type building might lie in close proximity, with a small amount of tegula and box flue tile and small-scale smithing debris recovered. Artefacts suggest that use of this agricultural landscape was restricted to the lst and 2nd century AD. It is possible that the fields were abandoned due to a major re-organisation of the agricultural landscape associated with the rise of villa estates, as may have occurred at the villa near Malmesbury (Hart et al 2005), but if this was the case at Showell Farm the villa house did not lie sufficiently close to the excavated area for later Roman material to be spread across it. It seems likely that the settlement focus lay to the north-west of the field system and may even have been destroyed by the construction of the railway in the 19th century. Previously unknown Romano-British field systems in the area around Chippenham have come to light during development works over the last few years (Bateman 2000, 90-104). There is increasing evidence for Romano-British agricultural activity in the North Wiltshire clay vale and Chippenham areas, but to date no villas and farms have been excavated. Acknowledgements ‘We should like to thank Mark Stanley-Smith of Nicholson Estates and Roy Canham, Wiltshire County Council Archaeologist for their assistance during the course of this project. The fieldwork was directed by Alan Thomas with the assistance of Jo Williams and Kate Cullen. The fieldwork was managed by Neil Holbrook and Dawn Enright, with post-excavation managed by Annette Hancocks. Neil Holbrook edited and commented on this report. The discussion was greatly enhanced by the comments of Professor Tim Darvill and Dr Joshua Pollard. The illustrations are by Lorna Gray, Peter Moore and Bryony Ryder. The site archive will be deposited with Chippenham Museum. Bibliography ANDERSON, A.S. 1979a. The Roman pottery industry of north Wiltshire. Szoindon Archaeological Society Report 2 Swindon, 1-4 ANDERSON, A.S. 1979b. A type of moulded imitation Samian and its military connections. Britannia 10, 109-19 ANDERSON, A.S. 1980, A Guide to Roman Finewares. Swindon: VORDA Research Series 1, Highworth ANON, 1993. Excavation and fieldwork in Wiltshire: Chippenham. WANHM 86, 159-60 BARCLAY, A. and HALPIN, C. 1999, Excavations at Barrow Hills, Radley, Oxfordshire. Volume I: The Neolithic and Bronze Age Monument Complex. Oxford Archaeological Unit Thames Valley Landscapes Volume 11 BATEMAN, C. 2000. Excavations along the Littleton Drew to Chippenham Gas Pipeline. WANHM 93, 90-104 BATEMAN, C. and ENRIGHT, D. 2000. Excavations of Bronze Age and Romano-British sites along the Chippenham Western Bypass A4 to A350 link. WANHM 93, 233-7 BENNETT, J., 1985, ‘Appendix 3: Moulded Coarse Ware’, in J.Bennett, Sea Mills The Roman Town of Abonae Excavations at Nazareth House, 65-7. City of Bristol Museum & Art Gallery Monograph No. 3. Bristol: City of Bristol Museum & Art Gallery BGS (British Geological Survey), 1990, Bath, England and Wales Sheet 265, solid and drift geology, 1:50,000, Keyworth, Nottingham: British Geological Survey BOARDMAN, S. and JONES, G. 1990. Experiments on the Effects of Charring on Cereal Plant Components. Journal of Archaeological Science 17, 1-11 BOON, G.C. and SAVORY, H.N. 1975. A silver trumpet brooch from Carmarthen. Antiquaries Journal 40, 41-60 BRONK RAMSEY, C., 2003, OxCal Program v3.9 BULLEID, A., and GREY, H.G., 1917, The Glastonbury Lake Village ii Glastonbury Taunton: Glastonbury Antiquarian Society BUTCHER, S.A., 2001, ‘Brooches of copper alloy’ in A.Anderson., J.S.Wacher., and A.PFitzpatrick, The Romano-British ‘Small Town’ at Wanborough, Wiltshire, 41-69. Britannia Monograph No. 19. London: Society for the Promotion of Roman Studies CAMPBELL, G., 2000, ‘The charred plant remains’, in B.Cunliffe and C.Poole (eds), Danebury Environs Programme: The Prehistory of a Wessex Landscape. Volume 2 — Part 3, Suddern Farm, Middle Wallop, Hants, 1991 and 1996, 193-8. English Heritage and Oxford University Committee for Archaeology Monograph No. 49. Oxford: Institute of Archaeology CARRUTHERS, W. J., 2002, “The charred plant remains’ in R.Atkins and A.Mudd (eds), An Iron Age and Romano-British settlement at Prickwillow Road, Ely, Cambridgeshire: Excavations 1999-2000. Proceedings of the Cambridgeshire Antiquarian Society XCII, 44-8 CARRUTHERS, W.J. forthcoming a ‘The charred plant remains’, in L.Coleman and A.Hancocks, Four Multiperiod Sites by the Carrant Brook and River Ishbourne, Gloucestershire and Worcestershire: Excavations on the Wormington to Tirley Pipeline 2000. Cotswold Archaeology Occasional Paper 2 48 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE CASE, H., 1977, “The Beaker culture in Britain and Ireland’, in R.Mercer (ed), Beakers in Britain and Europe: Four Studies, 71-101. British Archaeological Report No. 26, Oxford: British Archaeological Reports CASE, H. 1993. Beakers: deconstruction and after. Proceedings of the Prehistoric Society 59, 241-68 CAT (Cotswold Archaeological Trust) 1998, Land at Showell Farm, Chippenham, Wiltshire: Cultural Heritage Assessment. Client Report No. 98960 CAT (Cotswold Archaeological Trust) 1999a, Land at Showell Farm, Chippenham, Wiltshire. Cultural Heritage Assessment: Technical Appendix, Archaeological Evaluation. Client Report No. 991001 CAT (Cotswold Archaeological Trust) 1999b Showell Farm, Chippenham, Wiltshire: Archaeological Excavation, Interim Report. Client Report No. 991103 CAT (Cotswold Archaeological Trust) 1999c, Showell Farm, Chippenham, Wiltshire: Archaeological Evaluation (Area C). Client Report No. 991094 CAT (Cotswold Archaeological Trust) 2002, Land Between New Terrace and Marina Drive Staverton, Trowbridge, Wiltshire: Archaeological Evaluation. Client Report No. 02005 CLIFFORD, E., 1961, Bagendon, a Belgic Oppidum. Cambridge: W. Heffer and Sons Ltd COOPER, N., 1998, ‘The Supply of Pottery to Roman Cirencester’, in N.Holbrook, Cirencester: The Roman town defences, public buildings and shops, 324-50 Cirencester Excavations V. Cirencester: Cotswold Archaeological Trust CORNEY, M., 2000, ‘The Brooches’ in M.G.Fulford and J.R.Timby (eds), Late Iron Age and Roman Silchester: Excavations on the site of the forum-basilica 1977, 1980-86, 322-37. Britannia Monograph Series No. 15. London: Society for the Promotion of Roman Studies CORNEY, M., 2001, ‘The Romano-British Nucieated settlements of Wiltshire’ in PEllis (ed), Roman Wiltshire and after: papers in honour of Ken Annable, 5-38. Bath: Wiltshire Archaeological and Natural History Society CRUMMY, N., 1983,.Colchester Archaeological Report 2: The Roman small finds from excavations in Colchester 1971-9. Colchester: Colchester Archaeological Trust Ltd DARLING, M.J., 1977, ‘Pottery in Early Military sites in Western Britain’ in J.Dore and K.Green (eds), Pottery studies form Britain and beyond, British Archaeological Review Supplementary Series 30, 57-99 DARVILL, T-C. 1987, Prehistoric Gloucestershire. Sutton Publishing DARVILL, T.C. and GRINSELL, L.V. 1989. Gloucestershire barrows: supplement 1961-1988. Transactions of Bristol and Gloucestershire Archaeological Society 107, 39-105 DAVIS, S.J.M., 1992, A Rapid method for recording information about mammal bones from archaeological sites. Ancient Monuments Laboratory Report No. 19/92 DOBNEY, K., 2001, ‘A place at the table: the role of vertebrate zooarchaeology within a Roman research agenda for Britain’, in S.James and M.Millett (eds), Britons and Romans: advancing an archaeological agenda, 36-45. CBA Research Report 125, York: CBA DOBNEY, K.M., JAQUES, S.D., and IRVING, B.G., 1995, Of butchers and breeds: report on vertebrate remains from various sites in the City of Lincoln. Lincoln: Lincoln Archaeological Studies No. 5 DPDS, 1998, ‘Cultural Heritage’, in North Wiltshire Business Park, Showell Farm, Chippenham. Environmental Statement. DPDS Consulting Group FOSTER, J., 1986, ‘Fired Clay Objects’, in I.M.Stead and V.Rigby (eds), Baldock. The Excavation of a Roman and Pre-Roman Settlement, 1968-72, 168. Britannia Monograph Series No. 7. London: Society for the Promotion of Roman Studies GIBSON, A.M., 1982, Beaker Domestic Sites: a study in the domestic pottery of the late third and early second millennia BC in the British Isles British Archaeological Report No 107. Oxford: British Archaeological Reports GILLAM, J.P. 1957. Types of Roman coarse pottery vessels in northern Britain. Archaeologia Aeliana (4th series) 35, 180-251 GREEN, H.S., 1980, The flint arrowheads of the British Isles Oxford. British Archaeological Report No. 751i. Oxford: British Archaeological Reports GREIG, J.R.A., 1991, ‘The British Isles’, in W.Van Zeist, K.Wasylikowa and K.Behre (eds), Progress in Old World Palaeoethnobotany, 299-334. Rotterdam: A.A. Balkema HART, J., COLLARD, M.C. and HOLBROOK, N. 2005. A New Roman Villa Near Malmesbury Wiltshire. WANHM forthcoming HATTATT, R., 1985, Iron Age and Roman Brooches. Ipswich: Anglia Publishing HATTATSR., 2000, A Visual Catalogue of Richard Hattatt’s Ancient Brooches. Oxford: Oxbow Books HEARNE, C.M. and ADAM, N. 1999 Excavation of an Extensive Late Bronze Age Settlement at Shorncote Quarry, near Cirencester, 1995-6. Transactions of the Bristol and Gloucestershire Archaeological Society 117, 35-73 HILLMAN, G., 1982, ‘Evidence for spelting malt’, in R.Leech Excavations at Catsgore 1970-1973. A Romano- British village, 137-40. Western Archaeological Trust Monograph 2. Bristol HOLBROOK, N., and BIDWELL, PT., 1991, Roman finds from Exeter. Exeter Archaeological Reports 4 Exeter: Exeter City Council and University of Exeter HOOLEY, A.D., 2001, ‘Copper alloy and silver objects’ in A.Anderson., J.S.Wacher., and A.P Fitzpatrick, The Romano-British ‘Small Town’ at Wanborough, Wiltshire, 75-116. Britannia Monograph No. 19. London: Society for the Promotion of Roman Studies JACKSON, R. 1985. Cosmetic sets from Late Iron Age and Roman Britain Britannia 16, 165-192 LAMBRICK, G., and ALLEN, T. 2004, Gravelly Guy Stanton Harcourt Oxfordshire. The Development of a Prehistoric and Romano-British Community. Oxford Archaeology Thames Valley Landscapes Monograph SHOWELL FARM, CHIPPENHAM. EXCAVATIONS IN 1999 49 No. 21 LOVEDAY, R. 1989, “The Barford Ritual Complex: Further Excavations (1972) and A Regional Perspective’ in A.Gibson (ed) Midlands Prehistory, 51-84. British Archaeological Report British Series No 204. Oxford: British Archaeological Reports MABEY, R., 1972, Food for Free. London: Collins MACKRETH, D.F, 1988, ‘The brooches’ in S.D.Trow Excavations at Ditches hillfort, North Cerney, Gloucestershire, 1982-3 Transactions of the Bristol and Gloucestershire Archaeological Society 106, 43-51 MACKRETH, D.F, 2000, ‘Brooches’ in L.Luckett, Investigation of a Roman villa site at Euridge Manor Farm, Colerne. WANHM 93, 220-5 MALTBY, J.M., 1985, ‘Assessing variations in Iron Age and Roman butchery practices: the need for quantification’, in N.J.R.Fieller., D.D.Gilbertson., and N.G.A.Ralph., (eds), Palaeobiological investigations: research, design, methods and data analysis, 19-31. British Archaeological Report International Series 266. Oxford: BAR MALTBY, M., 1989, ‘Urban-rural variations in the butchery of cattle in Romano-British Hampshire’, in D.Serjeantson and T.Waldron (eds), Diet and Crafts in Towns, 75-106. British Archaeological Report British Series. Oxford: BAR MANNING, W.H. 1985, Catalogue of the Romano-Bnitish Tron Tools, Fittings and Weapons in the British Museum. London: British Museum Publications Ltd MCSLOY, E.R. 2004 ‘Pottery’ in J.Hart, M.Collard and N.Holbrook A new Roman villa near Malmesbury Cotswold Archaeology typescript report 04180 MORRIS, P, 1979, Agricultural Buildings in Roman Britain. BAR British Series 70, Oxford OAU (OXFORD ARCHAEOLOGICAL UNIT) 1991, Archaeological Evaluation: Milbourne Farm and Showell Nurseries, Chippenham, Wiltshire. Unpublished client report OLIVIER, A., 1988. ‘The Brooches’ in T.W. Potter and S.D.Trow, Puckeridge-Broughing, Herts. The Ermine Street Excavations, 1971-1972 Hertfordshire Archaeology 10, 35-53 PALMER, A., and ALLEN, T., 1993, ‘Fired Clay Objects’ in T.G. Allen, T. Darvill, L. Green, and U. Jones (eds), Excavations at Roughground Farm, Lechlade: a prehistoric and Roman landscape, 165-6. Thames Valley Landscapes: the Cotswold Water Park, Volume 1. Oxford: Oxford Archaeological Unit PAYNE, S. 1973. Kill-off patterns in sheep and goats: the mandibles from Asvan Kale. Anatolian Studies 23, 281-303. PAYNE, S. 1987. Reference codes for wear states in the mandibular cheek teeth of sheep and goat. Journal of Archaeological Science 14, 609-14 PAYNE, S., 1988, ‘Animal bones from Tell Rubeidheh’, in R.G.Killick (ed), Zell Rubeidheh: an Uruk village in the Jebel Hamrin, 98-135. Iraq Archaeological Reports 2, Warminster: Aris and Phillips PEARCE, J., 1999, ‘The dispersed dead: preliminary observations on burial and settlement space in rural Roman Britain’ in PBaker et al. (eds) Proceedings of the Eighth Annual Theoretical Roman Archaeology Conference, Leicester 1998 (1999), 151 PCRG (PREHISTORIC CERAMIC RESEARCH GROUP). 1997, The Study of Later Prehistoric Pottery: General Policies and Guidelines for Analysis and Publication. Salisbury: Prehistoric Ceramics Research Group Occasional Papers | and 2 PHILLIPS, B. and HENIG, M. 2000. A gold finger ring from the Rudge Romano-British villa site, Froxfield, Wiltshire. WANHM 93, 239-40 PHILPOTT, R., 1991, Burial Practices in Roman Britain. A survey of grave treatment and furnishing A.D. 43-410. BAR British Series 219, Oxford POWESLAND, I., 2004, The Later Prehistoric Landscape of the Bristol Avon Region. Submitted as a Phd thesis to Bristol University PRICE, J., and COOL, H.E.M., 1985, ‘Glass (including glass from 72 Dean’s Way)’ in H.R. Hurst (ed), Kingsholm,. 41-54. Gloucester Archaeological Report No. 1. Cambridge: Gloucestershire Archaeological Publications PRICE, J., and COTTAM. S., 1998, Romano-British glass vessels: a handbook. CBA Practical Handbook in Archaeology 14. York: Council for British Archaeology RAWLINGS, M. and Fitzpatrick, A.P. 1996. Prehistoric sites and Romano-British settlkement at Butterfield Down, Amesbury. WANHM 89, 1-43 RIGBY, V., 1982a, “The Coarse Pottery’ in J. Wacher, and A. McWhirr, (eds) Cirencester Excavations I: Early Roman Occupation at Cirencester, 153-204 Cirencester Excavation Committee. Gloucester: Alan Sutton RIGBY, V., 1982b, ‘The Coarse Pottery’ in A. McWhirr, L. Viner, and C. Wells, Cirencester Excavations Volume IT: Romano-British Cemeteries at Cirencester, Microfiche Section C13-D10. Cirencester Excavation Committee. Gloucester: Alan Sutton SAVILLE, A., 1984, Archaeology in Gloucestershire. Cheltenham: Cheltenham Art Gallery and Museum and Bristol and Gloucestershire Archaeological Society SEAGAR SMITH, R.H., 2001, ‘The Coarse Pottery’ in A.S. Anderson, J.S. Wacher, and A.P. Fitzpatrick, The Romano-British ‘Small Town at Wanborough, Wiltshire, 232-301. Britannia Monograph No. 19. London: Society for the Promotion of Roman Studies STACE, C., 1997, New Flora of the British Isles. Second Edition. Cambridge: Cambridge University Press STANSFIELD, J.A. and SIMPSON,G. 1990. (Reprint) Les Potiers de la Gaule Central. Revue Archeologique Sites Hors-serie No. 37 STEAD, I.M., 1986, ‘Brooches’ in I.M. Stead and V. Rigby, Baldock. The Excavation of a Roman and Pre-Roman Settlement, 1968-72 Britannia Monograph Series No. 7, 109-25. London: Society for the Promotion of Roman Studies 50 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE STUIVER, M. and POLACH, H.A. 1977. Discussion: Reporting of 14C Data. Radiocarbon 19, no. 3, 355-63 STUIVER, M. and REIMER, PJ. 1993. Extended 14C database and revised CALIB 3.0 14C Age calibration program. Radiocarbon 35(1), 215-30 STUIVER, M., REIMER, PJ., BARD, E., BECK, J.W., BURR, G.S., HUGHEN, K.A., KROMER, B., MCCORMAC, VAN DER PLICHT, J. AND SPURK, M. 1998. INTCAL98 Radiocarbon Age Calibration, 24000-0 cal BP. Radiocarbon 40(3), 1041-83 TAYLOR, A.F and WOODWARD,PJ. 1985. A Bronze Age barrow cemetery, and associated settlement at Roxton, Bedfordshire. Archaeological Fournal 142, 73-149 TIMBY, J. 1999 ‘Later Prehistoric and Roman pottery’ in A.Mudd, R.J.Williams and A.Lupton, 1999 Excavations alongside Roman Ermin Street, Gloucestershire and Wiltshire: The archaeology of the A419/A417 Swindon to Gloucester road scheme, 320-62. Oxford: Archaeological Unit TIMBY, J. 2000. ‘Pottery’ in C.Bateman, Excavations along the Littleton Drew to Chippenham Gas Pipeline WANHM 93, 94-5 TOMBER, R., and DORE, J., 1998, The National Roman Fabric Reference Collection: A Handbook. London: Museum of London Archaeology Service UNIVERSITY OF WAIKATO RADIOCARBON DATING LABORATORY. 2004. Operating Procedures http:// www.radiocarbondating.com/operatingprocedures (accessed on 23rd December 2004) VAN DER VEEN, M., 1989 Charred Grain Assemblages from Romano-Period Corn Driers in Britain. Archaeological Fournal 146, 302-19 VON DEN DRIESCH, A. 1976. ‘A guide to the measurement of animal bones from archaeological sites’. Peabody Museum Bulletin 1, Cambridge Mass., Harvard University VON DEN DRIESCH, A. and BOESSNECK,J. 1974. Kritische anmerkungen zur widerristhohenberechnung aus Langenmassen vorund fruhgeschichtlicher Tierknochen. Saugetierkundliche Mitteilungen 22, 325-48 WALKER, G. 2000. ‘Worked Flint’ in C.Bateman and D.Enright, Excavations of Romano-British sites along the Chippenham Western Bypass A4 to A350 Link WANHM 93, 236 WALTERS, B., 2001, ‘A perspective on the social order of Roman villas in Wiltshire’ in PEllis (ed), Roman Wiltshire and after: papers in honour of Ken Annable, 127-46. Bath: Wiltshire Archaeological and Natural History Society WEBSTER, PV. 1976. Severn Valley ware: A Preliminary Study. Transactions of the Bristol and Gloucestershire Archaeological Society 94, 18-46 WHIMSTER, R., 1981, Burial Practices in Iron Age Britain. BAR British Series 90. Oxford WILLIS, S., 1998, Samian pottery in Britain: Exploring its distribution and Archaeological Potential. Archaeological Fournal 155, 82-133 Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 51-78 A Middle to Late Bronze Age Settlement at Dunch Hill, Tidworth by Phil Andrews with contributions by Michael F. Allen, Rowena Gale, Phil Harding, Pat Hinton, Emma Loader, facqueline I. McKinley and Pippa Smith Excavations and a watching brief on an earth trackway at Dunch Hill near Tidworth produced surprisingly well preserved evidence for Middle to Late Bronze Age settlement, broadly spanning the 14th—8th centuries BC, spread over a distance of 370m. The earliest feature was a single, small pit which contained Late Neolithic Grooved Ware along with some Beaker pottery. Middle Bronze Age features comprised a cremation burial and a ditch probably representing part of a field system. Late Bronze Age remains included four roundhouses, two four-post structures, several small pits, and a series of fence lines — perhaps defining animal enclosures. The quantity and range of finds were limited, but environmental remains provide important information on the development and use of the landscape. INTRODUCTION In 1995 Wessex Archaeology undertook the excavation of approximately 4600m2 of an earthen trackway at Dunch Hill near Tidworth (SU 2050 4860). A further 2900m2 was subject to a watching brief. A small amount of additional information was recorded during the digging of a military slit trench in 1998. The excavation was undertaken prior to the construction of a stone track from Tidworth to Larkhill designed to protect both the immediate environment and archaeological sites in the vicinity from the repeated passage of heavy military vehicles. The excavated section of trackway lies on the south-west facing slope of Dunch Hill (Figures 1 and 2), rising from approximately 120m OD at the west to 155m OD near the crest at the east. It is well- drained and overlooks a seasonally flooded coombe (Bourne Bottom) at the head of the Nine Mile River, a tributary of the Avon, a kilometre or so to the west. The geology comprises Upper Chalk with occasional patches of Clay-with-flints. Much of the surface of the chalk on the slope was cut by periglacial features comprising a series of small, closely spaced, parallel cracks running down the slope, filled with light brown clay. Most of the trackway was unmetalled and consisted of a thin layer of loamy soil, typically 0.1-0.3m thick, containing a variable amount of chalk pieces and directly overlying the chalk. At the extreme east end there was some gravel and brick hardcore which had been used to consolidate the surface. The topsoil on the sloping part of the track was prone to water erosion, and in places this had created shallow gullies, sometimes deepened into larger ruts by vehicle traffic. Archaeological Background Towards the eastern end and immediately to the north of the excavated area are two extant Bronze Age round barrows (Figure 1), both of which (SMR No. SU24NW 604/SPTA 1950 and SMR No. SU24NW 605/SPTA 1954) lie within a Scheduled Area (SM Wessex Archaeology, Portway House, Old Sarum Park, Salisbury SP4 6EB 52 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE ‘g [= Earthworks © Round barrow 48 iors / 0 1km ny St) All contours in metres OD | : “Cremation burial 439 Excavation/Watching Brief Trench Fig. 1. Dunch Hill: Site and trench location plan showing principal excavated features A MIDDLE TO LATE BRONZE AGE SETTLEMENT AT DUNCH HILL, TIDWORTH 53 Fig. 2. Dunch Hill: Site partially stripped with round barrow (SMR 604) in foreground W110140) which also comprises the field to the west (see Figure 17). To the north of the eastern half of the excavated area, and within the Scheduled Area, is part of a so-called ‘Celtic’ field system, surviving as slight earthworks and clearly apparent on aerial photographs (see Figure 17). A little more than 100m to the south are the partly extant remains of a Late Bronze Age linear ditch, running approximately north-east to south-west (Figure 1), the line of which can also be traced on aerial photographs (Bradley er al. 1994, fig. 23). Earlier fieldwork at Dunch Hill formed an important component of the Linear Ditches Project which was initiated to study prehistoric land divisions on Salisbury Plain (Bradley et al. 1994). At Dunch Hill this included fieldwalking which produced a scatter of Late Bronze Age pottery, worked and burnt flint, and the recovery of pottery, also dated to the Late Bronze Age, from a number of tree-holes to the south of the area subsequently excavated in 1995 (Bradley et al. 1994, 31-2, fig. 46; see Figure 17). Excavation of a small trench to the north of the 1995 excavation area produced a relatively large quantity of Late Bronze Age pottery, worked flint, burnt flint and animal bone from a deposit interpreted as a midden that sealed a ditch and a hollow-way, both thought to be of Late Bronze Age date (Bradley et al. 1994, 49-50, figs 45 and 46; see Figure 17). At the same time, three small features exposed in the side of a track adjacent to the 1995 excavation area were investigated, two of which contained Late Bronze Age pottery (Bradley et al. 1994, 53). In January 1995 an archaeological evaluation was carried out at Dunch Hill which established that Late Bronze Age remains survived beneath the existing track which it was proposed to upgrade to a hard surface track (Wessex Archaeology 1995a). Asa result of this evaluation, Wiltshire County Council Museums Service advised the Defence Land Agent (now Defence Estates) that further archaeological investigation was desirable prior to the proposed works. The subsequent excavation was undertaken between June and August 1995 (Wessex Archaeology 1995b). Methods An area approximately 220m long and 12m wide along the line of the proposed hard surface track ‘was stripped of topsoil (Figure 3). All archaeological features other than ditches were fully excavated; approximately 50% of the one ditch was excavated anda mollusc column taken. A representative sampie of the remaining features, mostly tree-throw holes, were 50% excavated. Almost all archaeological features and a small number of tree throw-holes were sampled for environmental analysis and artefact recovery. Further machine stripping immediately to the east of the previously excavated area was subject to a watching brief and exposed archaeological features were excavated. Further lengths of track stripped to the west (280m by 10m) and east (S3m by 1.8m) were also subject to a watching brief, but no archaeological features were observed. A single archaeological feature was observed and recorded in an army slit trench to the south of the area. Fig. 3 Dunch Hill: Central section of site after stripping 54 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Pit 482 ARCHAEOLOGICAL DEPOSITS Archaeological remains comprised entirely negative features, with no stratified deposits above the natural chalk. Although the surface of the chalk had been subject to a certain amount of damage as a result of rainwater run-off and wheel rutting, features were generally well preserved. These appeared to have suffered comparatively little truncation, except towards the north-east and south-west ends of the site, but only small quantities of finds were present and no artefacts were recovered from the overlying layer of heavily disturbed topsoil. Three phases of prehistoric activity were identified, assigned to the Late Neolithic/Early Bronze Age, Middle Bronze Age and Late Bronze Age respectively. Many features, particularly post- holes, contained little or no dating evidence, but have been assigned a phase on the basis of their certain or inferred association with other dated features. Few intercutting features were encountered. Two radiocarbon dates provide a more secure chronology for the sequence, which is otherwise almost entirely dependent on a small assemblage of pottery. Prehistoric features comprised up to 164 post- holes, nine pits, one ditch and one cremation burial. In addition to these, 33 features were identified as of natural origin, most representing tree-throw holes, in a few cases cut by Late Bronze Age features. More recent features were parts of two military slit trenches and a sheep burial. Late Neolithic/Early Bronze Age A small oval pit, 482, was the only feature assigned to this period (Figure 4). It was approximately 0.2m deep, flat-bottomed, and the lower half was filled with light greyish brown silty clay which contained six sherds of Late Neolithic Grooved Ware, five sherds of Beaker pottery and three flint end scrapers. Middle Bronze Age Two features assigned to this phase lay close together towards the east end of the site. Cremation burial 439 lay less than 10m to the south of the smaller of the two barrows (Figure 1) and comprised a shallow, heavily truncated pit approximately 0.7m in diameter and 0.2m deep containing cremated human bone and several large sherds of two Deverel-Rimbury bucket urns, including part of Fig. 4 Dunch Hill: Plan and section, Late Neolithic/Early Bronze Age pit 482 a base (Figures 5 and 6). Charcoal from cremation burial 439 provided a radiocarbon date of 1450-1210 cal BC (AA-33134, 3065+50 BP). Less than 2m to the north-east was shallow scoop 463; two smaller features, scoop 512 and possible post-hole 498, also lay in the vicinity, but none contained any finds or cremated human bone and they need not have been contemporaneous. Ditch 402 was a slightly curvilinear feature, up to 0.7m wide and 0.4m deep, which ran north-west to south-east for approximately 9m across the entire width of the excavated area (Figures 5 and 7). It was filled with a homogeneous greyish brown silty loam which produced 11 sherds of Middle Bronze Age pottery. A column sample for land snail analysis was taken from one of the three excavated sections. Late Bronze Age Most of the excavated features are assigned to this phase and all lay to the west of Middle Bronze A MIDDLE TO LATE BRONZE AGE SETTLEMENT AT DUNCH HILL, TIDWORTH 55 Key: Chalk [30 years old; insufficient evidence remained to suggest the sex of the individual. Tentative evidence for the remains of an infant/juvenile were also recovered —a deciduous molar root branch — though the retention of the molar into adulthood cannot be ruled-out. Most of the bone was buff-white in colour (oxidised; Holden er al. 1995), a fragment of fibula 64 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE shaft and a tooth root showing darker hues (brown, grey) indicative of incomplete oxidation. Bone elements from each skeletal area were recovered and there is no evidence to suggest any deliberate selection of specific bones for burial. ENVIRONMENTAL EVIDENCE The land snail evidence by Michael Ff. Allen The environment of the Late Bronze Age chalkland landscape in the general vicinity of Dunch Hill has been presented in the analysis of the Linear Ditches Project (Entwistle 1994a). A column of seven contiguous samples was taken at close (0.05m) intervals from one of the sections of Middle Bronze Age ditch 402 in order to present detailed information and attempt to provide some refinement of the interpretations presented by Entwistle (1994a). In addition, spot samples were taken from Late Neolithic/Early Bronze Age pit 482 and three probable tree-throw holes. All samples were processed by standard methods (Evans 1972). The identifications by the writer are tabulated (Table 2) and presented in histograms of relative abundance (Figure 15) in the same format as those from the Linear Ditches Project (Entwistle 1994a) to facilitate direct comparison. The ascription of shade-loving, catholic and intermediate species follows Evans (1984) as used by Entwistle (1994a) but, in the histogram, other shade-loving species include Clausilia bidentata, and also Carychium tridentatum and Punctum pygmaeum which occur in very low numbers. The results can be compared with Entwistle’s analysis of the Late Bronze Age ditch LDP 052 (Entwistle 1994a, 116-8, fig. 69 and table 57) which lies approximately 800m to the south-east of the Dunch Hill site. Tiee-throw holes Assessment of the flots showed that where snails survived, all samples contained typically open country species (Vertigo cf. pygmaea, Pupilla muscorum, Helicella itala and Vallonia species). These features are, therefore, unlikely to represent, or provide evidence of, the earlier (i.e, Neolithic) wooded nature of the local landscape. Late Neolithic/Early Bronze Age pit 482 Two spot samples from this feature produced assemblages that were predominately open country; a very small shade-loving component comprising Discus rotundatus and individual specimens of Carychium tridentatum, Aegopinella nitidula and Clausilia bidentata were also present, and here probably reflect ungrazed grassland (Cameron and Morgan-Huws 1975). Pupilla muscorum is dominant, if not super-abundant, in both assemblages. The main difference between the assemblages 1s depicted in the strong divergence of the A2 (probability e £ e . , ie og? ra a e Rd s i Fe i Leg gee a ad oy oe © o eo oon : Ay fi ND w Re) £ 3 RA Naa oe S § e x PF PP Pw vad $ Fe PAE SS wt QF GE SF Ditch 402 (Geom (he =e L (EAE (Ee L L L LE Leet bob ] | 1+ + + + l + + +++ 4 + + fj + + | ae + | ap ct l | | + + ++ 6+ + + lil -ll-: tz [i - ee: -l1 tl -tl ito td Sel ed am EA erie ey oh es Spee lege ee ee) Teh a al idee pete a me Fig. 15 Dunch Hill: Land snails A MIDDLE TO LATE BRONZE AGE SETTLEMENT AT DUNCH HILL, TIDWORTH 65 Table 2. Dunch Hill: Land Snail Data (nr = not recorded) Phase | L. Neolithic Feature pit 482 Sample 1094 1093 Context 481 480 Depth(cm) spot — spot Middle Bronze Age Ditch 4062 1104 1103 1102 1101 1100 1099 1098 401 401 401 401 401 401 = 401 30-35 25-30 20-25 15-20 10-15 5-10 0-5 basal top Wt 1000 1000 600 nr 650 800 800 900 1000 Pomatias elegans (Miller) 3 11 3 7 15 17 5 15 19 Carychium tridentatum (Risso) 3 1 - 1 - - - - - Cochlicopa lubrica (Miller) - - - - - - - - Z Cochlicopa lubricella (Porro) l 1 1 + l - + Z - Cochlicopa spp. 8 3 1 5 2 3 5 3 Truncatellina cylindica (Férussac)) - - ] 6 2 12 11 8 1 Vertigo pygmaea (Draparnaud) 4 5 1 4 - 1 Z. 2 - Pupilla muscorum (Linnaeus) 96 55 123 85 17 LSi, = 115) ~Vl0L 75 Vallonia costata (Miller) 5 1 39 31 42 39 4] 31 35 Vallonia excentrica Sterki 33 27 35 15 23 32 29 21 9 Punctum pygmaeum (Draparnaud) - 1 - 1 1 - - - - Discus rotundatus (Miller) 2 2 - = - 3 = = = Vitrea crystallina (Miller) - - - - - - 1 - Aegopinella nitidula (Draparnaud) 1 . - . - 1 - - - Limacidae 1 6 3 6 1 5 6 5 2 Cecilioides acicula (Miller) 2 1 - - - - - - - Clausilia bidentata (Str6m) 1 - - 1 - 7 5 - 3 Cernuella virgata (Da Costa) - 1 - - - - - - - Helicella itala (Linnaeus) 10 9 13 8 12 17 14 12 7 Trichia hispida (Linnaeus) 10 6 1 - - - 1 - 3 Helicigona lapicida (Linnaeus) - + - . - a - a . Cepaea spp. - - - - + 1 - - - Cepaea/Arianta spp. 1 1 1 - - - 1 - i Taxa 14 14 11 12 9 13 13 9 11 TOTAL 179 130 222 170 116 «295 236 198 157 Molluscs per kilo 79 300 370 - 173.369) 3289 4.220. 157 Shannon Index POU E85) ESS) eRe 72 eG 60s e600 57 Brillouin Index VAST SIEZOE, L282 e483 9160 2 WS4e5 E52" 152" 31-47 A4 DOA E325) 75 8 2220 3168) ee2 513") 2334 2.39) + 2:40 A2 0.67 0.76 0.63 0.68 0.78 0.68 0.70 0.70 0.70 of non-lethal interspecific encounters) and A4 (ratio of interspecific to intraspecific competition respectively) indices (Hurlbert 1971; Figure 15), indicating an increased complexity of vegetation mosaic within the vicinity. Middle Bronze Age Ditch 402 The shallow 0.4m deep ditch contained no vacuous chalk rubble, and was infilled with a fine uniform calcareous deposit. In this respect it is similar to LDP 052 (and LDP 027). The infilling is, therefore, taken to represent a continuous process of erosion and the molluscan evidence indicates little significant change in the ecological setting of the ditch during infilling (cf Entwistle 1994a, 116). Very open, dry, calcareous conditions are depicted throughout the sequence; Pupilla muscorum is dominant, but Vallonia costata is significant whereas it was almost absent in the Late Neolithic/ Early Bronze Age samples. The assemblages from ditch 402 compare closely with those published by Entwistle from LDP 052, and with low diversity indices are typical of a restricted range of habitats and unstable local conditions such as arable or possibly short grazed grassland with bare loose soil patches (Pomatias elegans and Pupilla muscorum). Most of the few shade-loving species present are known to occur in grassland habitats, and the presence of a Clausilia bidentata throughout the sequence (Table 2) may indicate some woody or scrub refugia nearby. The lack of Evans’s ‘Punctum Group’, except for single occurrences of Punctum pygmaeum, indicates that no significant vegetation colonisation of the ditch occurred, except perhaps of a short, grassland sward. The A2 and A4 ranges are very similar to those from LDP 052 (A2 ranges 66 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE between 0.67 and 0.78 and A4 between 1.75 and 3.68) which Entwistle (1994a) argues are indicative of uninterrupted sedimentation from the exposed margins of the ditch. Close sampling did reveal one significant fluctuation at 20-—25cm where proportions of Pomatias elegans, Vallonia excentrica and Helicella itala all rise to the detriment of the relative abundance of Pupilla muscorum. This may represent a minor infill of loose earth, but this is not reflected in the fractionated residues weights or proportions. Charcoal by Rowena Gale Charcoal from Late Neolithic/Early Bronze Age pit 482, Middle Bronze Age cremation burial 439, and post-holes and pits associated with the Late Bronze Age roundhouses and a four-post structure 338 was identified to genus level. With the exception of the sample from post-hole 45 (roundhouse 57), most charcoal was sparse and poorly preserved. Samples were prepared for examination using standard methods. Fragments measuring >2mm in cross-section were sorted from each sample. These were fractured to expose fresh transverse surfaces and sorted into groups based on the anatomical features observed using a x20 hand lens. Representative fragments from each sample were selected for detailed study at high magnification. These were fractured to expose the tangential and radial planes, and examined at magnifications of up to x400. The anatomical structure was matched to prepared reference slides. When possible the maturity (ie, heartwood/ sapwood) of the wood was assessed but most fragments were too small to record details of growth patterns. The results are summarised in Table 3. The anatomical structure of the charcoal is consistent with the taxa or groups of taxa given below. It should be noted that the anatomical structure of some related taxa cannot be distinguished with any certainty, for example, members of the Pomoideae (Crataegus, Malus, Pyrus and Sorbus) and Salicaceae (Salix and Populus). Classification follows that of Flora Europaea (Tutin, Heywood et al. 1964-80). Aceraceae. Acer sp., maple Caprifoliaceae. Sambucus sp., elder Corylaceae. Corylus sp., hazel Fagaceae. Quercus sp., oak Oleaceae. Fraxinus sp., ash Rosaceae. Subfamily Pomoideae which includes Crataegus sp., hawthorn; Malus sp., apple; Pyrus sp., pear; Sorbus spp., rowan, service tree and whitebeam. These taxa are anatomically indistinguishable. Subfamily Prunoideae. Prunus spinosa, blackthorn. Subfamily Rosoideae which includes Rosa sp., briar, and Rubus sp., bramble. These taxa are anatomically similar. Salicaceae. Salix sp., willow and Populus sp., poplar. These taxa are anatomically similar. Late Neolithic/Early Bronze Age The charcoal from pit 482 included hazel (Corylus), ash (Fraxinus), hawthorn group (Pomoideae) and oak (Quercus) heartwood. Middle Bronze Age The charcoal from cremation burial 439 consisted of small fragments of ash sapwood which is likely to have been from residues of the pyre fuel. Late Bronze Age Charcoal was examined from post-holes on opposite sides of roundhouses 57 and 158 (post-holes 35 and 45, and 128 and 138 respectively), from post-hole 199 on the eastern side of roundhouse 196, and from internal post-hole 231. The charcoal from all contexts except post-hole 45 was comminuted and sparse. Oak was common to all samples and, in addition, maple (Acer) and member/s of the hawthorn group were identified from post-holes 199 and 138 respectively. The large volume of charred oak from post-hole 45 (roundhouse 57) is likely to be from the post, burnt in situ. Although the charred remains of the post had fractured, the wide growth rings in the resulting numerous fragments of sapwood could be seen to be consistent with rapid growth. The low ratio of heartwood to sapwood suggested that heartwood had barely started to develop when the oak pole/ trunk was cut and converted into the post, which infers that the post was probably relatively thin. Charcoal from two pits, 136 and 152, both within roundhouse 158, was sparse. Oak was identified from both, and pit 136 also included other taxa but the charcoal was too degraded to identify with certainty. Those provisionally identified included elder (Sambucus), blackthorn (P spinosa) and bramble/ briar (Rosaceae). Four-post structure 338 produced scant charcoal. Oak, the hawthorn group and possibly maple were identified. A MIDDLE TO LATE BRONZE AGE SETTLEMENT AT DUNCH HILL, TIDWORTH 67 Table 3. Dunch Hill: Charcoal Feature Context Sample Acer Corylus Fraxinus Pom Prunus Quercus Sal Sam Rosoid Late Neolithic/Early Bronze Age Pit 482 481 1094 - 4 2 10 - 3h - - - Middle/ Late Bronze Age Cremation 439 438 1087 - - 18s - - - - 2 = Pit 193 245 1023 - - - 2 - - 23r - - Roundhouse 57 Post-holes 35 36 1010 - - - - - 30 - - - 45 46 1042 - - - - - 137s, h - - - Roundhouse 158 Post-holes 138 139 1004 - - - 2 - 2s - - - 128 129 1040 - - - - - ?1 - - - Pit 152 153 1054 - . - - - 5 - - ~ 136 137 1012 - - - 3 22 2 - 21 21 Roundhouse 196 Post-holes 199 200 1028 3 - - - - 5 - - - 231 232 1063 - - - : - 3 - - - Four-post structure 338 Post-holes 178 177 1007 ?l . - 1 - 14 - - = 180 179 1051 - . - 1 - 6 - - = Key: Pom = Pomoideae; Sal = Salicaceae; Sam = Sambucus; Rosoid = Rosoideae; r = roundwood (diameter <20mm); s = sapwood; h = heartwood; The number of fragments identified is indicated. Charcoal from pit 193, to the south-east of four- post structure 515, consisted of tiny fragments. Three narrow pieces of stem/twig (diameters Imm, 2mm and 3mm) were tentatively identified as willow/poplar (Salix/Populus); the charcoal was -semi-vitrified (heated until the cell walls become plastic and fuse together, thereby losing recognisable features). Small fragments from a member of the hawthorn group were also identified. Discussion It is probable that the charcoal examined from post-holes associated with the roundhouses and four- post structure derived mainly from the structures themselves rather than representing fuel debris. All charcoal from post-holes in roundhouse 57 was exclusively oak, and oak was the most common species in post-holes associated with the other structures. The probable remains of an oak post, burnt in situ (post-hole 45), testifies to the use of oak in the construction of roundhouse 57. The charcoal samples from all but post-hole 45 were sparse and poorly preserved, and the relatively narrow range of taxa identified probably reflects this paucity of charcoal. Interestingly, the charcoal from Middle Bronze Age cremation burial 439, probably representing pyre debris, consisted entirely of ash sapwood. Unfortunately, very little material was available for identification and it would be dangerous to speculate too far as to the preferential use of ash in favour of other species for the construction of the pyre. It is worth noting, however, that in many Bronze Age cremation burials, oak appears to have been the preferred fuel, although often used with other species. Examples of such have been recorded by the author from several sites elsewhere including Durrington Down Round Barrow, Salisbury Plain (Gale 1990). From the purely practical aspect, the high calorific value of ash wood and the added bonus that ash is one of the few woods which burns efficiently when green (Edlin 1949), may have been contributory factors in its use at Dunch Hill. Although pollen was not preserved at Dunch Hill, several woodland taxa were identified from the charcoal analysis. From the Late Neolithic/ Early Bronze Age period these included hazel (charred nutshells were also recorded), ash, oak and member/s of the hawthorn group. From the Middle and Late Bronze Age periods these 68 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE included maple, ash, member/s of the hawthorn group, oak and possibly willow/ poplar, elder and bramble/briar. Oak occurred in almost every sample but the remaining taxa were more sporadic. The apparent absence of some common species almost certainly has more to do with the artefactual use of the wood by the inhabitants of the settlement, combined with inadequate soil conditions for long- term preservation, rather than the distribution of species in the environment. For example, hazel was present in the single Late Neolithic/Early Bronze Age pit, but was not identified from Late Bronze Age features. Since hazel characteristically flourishes on chalkland and quickly regenerates when cut, it seems very unlikely that it would have been eradicated from the Late Bronze Age landscape in the vicinity of the site. The analysis of charcoal from Dunch Hill augments similar studies from other sites on Salisbury Plain. These include: the Mesolithic deposits in the car park at Stonehenge which included pine (Pinus) (Allen 1995) and Neolithic and Bronze Age contexts from sites around Stonehenge including Coneybury and Coneybury Henge, Robin Hood’s Ball, King Barrow Ridge, North Kite, Durrington Down and Fargo Wood. A wide range of taxa included oak, blackthorn, the hawthorn group, hazel, maple, ash, hornbeam (Carpinus), elm (Ulmus), purging buckthorn (Rhamnus cathartica), dogwood (Cornus) and yew (Taxus) (Gale 1990). The Late Bronze Age double linear ditches at Sidbury included ash, oak, elder and hazel (Gale 1994) while the Late Bronze Age linear ditches and other features at Breach Hill included hazel, oak, the hawthorn group, willow/ poplar and pine (Gale 2006). Excavations at the Romano-British village at Chapperton Down, also undertaken for Defence Estates, included holly Ulex), alder (Alnus), hazel and oak (Gale in prep.). Charred plant remains by Pat Hinton Samples of 10 litres were processed by standard floatation methods with the flots retained on a 0.5mm mesh and residues on 1mm mesh. Fifteen samples were selected for analysis (Table 4) and the flots and items from their residues were examined by stereo microscope at x 7-40 magnification. All examined samples contained very small amounts of charred plant remains comprising mainly cereals in the form of poorly preserved grains and small fragments, a few seeds, stem parts, and fragments of Corylus avellana (hazel) nut shells. Several seeds remain unidentified. Neolithic/Early Bronze Age Both layers in pit 482 contained a few fragments of charred starchy material, mostly resembling cereal, of which one, in context 480, was of sufficient size to suggest part of the high ridged back of Triticum dicoccum (emmer). This context also contained one seed of Plantago lanceolata (ribwort plantain), and both contexts included fragments of hazel nut shell representing no more than one or two nuts in total. Middle Bronze Age The sample from cremation burial 439 included cereal fragments and one grain of Hordeum sp. (hulled barley). It is difficult to decide whether the asymmetry of this grain is natural, and therefore indicative of 6-row barley, or whether it is distorted by charring. Apart from a seed of Galium aparine (cleavers) the only other item appears to be either part of a tuber or of a corm-like basal node of Arrhenatherum elatius (false oat-grass or onion couch). Although the structure of this item is not entirely typical of the latter, and unfortunately it is somewhat compressed, swollen stem nodes of this grass are not infrequently found in Bronze Age cremation deposits, and are usually interpreted as part of the pyre fuel (eg, Clapham 1995). Late Bronze Age Samples from three of the roundhouses include barley, with one grain from roundhouse 196 probably of 6-row form, a single oat grain in the same sample, and a glume base of emmer in roundhouse 158. Wild plant seeds are of typical field weeds and/or grassland plants. Stem nodes of onion couch were found in two post-holes (in roundhouses 57 and 158). Slightly larger numbers of cereals and seeds were recovered from two post-holes in four-post structure 338. These included small quantities of cereal fragments with barley (one certainly appearing to indicate the askew form of 6-row barley) and very small fragments of Triticum chaff (three glume bases and one rachis node) in post-hole 162. Post-hole 182 also contained barley and one badly degraded wheat grain, probably emmer. Both post-holes included seeds of characteristic field weeds. Discussion The rather sparse remains provide only a glimpse 69 A MIDDLE TO LATE BRONZE AGE SETTLEMENT AT DUNCH HILL, TIDWORTH 6 S [ezeyssedj [JOYS- “"] vurzjaen snjuoy UIZIEW POOM = I 2 <3 = - = E = - - - - - - ssid - * 1OPUT aRIDROg NOD UOTUO/SseIB 180 ISTP] - 7 : . c ; I 7 dsr E s Te 7 . Wee (-p)snuja Ro aS I 2 = é a = = = > = - - - - - aspas — ‘ds xan ; I I € I F Selj+p = : = : = I = = SIDABIIO - “J auundp wniypoH urejueyd 7 : = 7 : . _ : : : : ; 7 I IIOMQII - “7 01Dj0a9IUD] OSDIUD] opeysiysiu 3 I F i 2 ; 7 ‘i i y E : F 5 7 yor - 7] wnisiu unuvnjos 2 = I = = - - - - - - - - - : xepy sursind - wnsiinyin9 wnury “jd = I = = I - - - - - - - - - - I9AO[N/YSIPIM - wnyofuy /odvripayw 2 = = = = - = = 2 - - I - - - yaaa — “ds v1914 “Jd : = = I = = = : = - - - - - - ward Aajasied - “"] sisuaain saunydy I - = = 2 E = - = - - - - - - JOpul aRadeUOSAOg PeeMpOIgQUe(qsS i I . 7 . z : % z 7 c : 2 - : DAOT YW (J) snpnejoauos nidopjn.q 2 5 = 2 = - : Self+]| - - - - - - - sseiIZ10UY -""J anjndiav wnuos Og é 3 I 3 = - : = - - - - - - uordures — ‘ds auajig “Jo Ppo2eMyAIIYS 1918918 5 I ; ; i _ 5 : ¢ z : ‘ a ; ; /WOUTUIOD - D1Ia/daUu/DIpau DiUD]IA1S 2 I = : 2 < = : I - - - : - - 1oojasoos - ‘ds wnipodouayy) sdnoi911ng : S 7 I c = : i % e : Fi ” snsoqng/suadas/siuop snjnounuvy pueyjsseiy/s[qeiy SEE SES AS SUE Ses Se $0 é s'0> a) S02) = Si0 =a S05 I S00 = (Tex) “ssery “epur etfeasaD : 5 o = I = : = : - - - - - - s1e0 — ‘ds nuaap Z Z : z ¢ = = = = € I - I - - Aayreq payyny — “ds wnapsoxy E Z 2 = = = 7 = = = - - - - - qeay — “ds wnguy I € : : : : 2 I . F _ : A ; saseq ouinys : : IOUIUID - WéNII0I1p JO WNIT, EAN) ST Or Ol Or Or Ol Or Ol Or Or Ol Ol Or Or SI SOUT) E[OAS|COIES esol JZIOT S90 ¢€901 -8zOI SzOr 8001 SOI ve0l 9001 SOOT e0t JZ80I | vE0T Ve60l ajdures 697 I9T vET CEC 007 cy vE esl eel r6l Sel 6b7 ser 18h 08h 1x91U07) c8T c9T £7 T&T 661 Ip £& cE! 6&F Heil ful Z1Cu 914 9104 ZOU GEL ASIA ATG LET OA OS Ettde ve C8b UL CEUTA | “1I90J ISO “S010 “SO ISO ASOT -180¢f 4D BRE ee sod-p 96[ asnoypunoy LS asnoypunoy SSI asnoypunoy adnIInitg aay a2Uorg a0 T/2PPIN __ VE A/OPN 210T poten sUIvUIOY 1URTg peeYyyD “[[IH young ‘p 2921 70 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE of the agricultural and natural background of the settlement. The small amounts of cereals and hazel nuts in the Late Neolithic/Early Bronze Age pit represent a common occurrence in charred plant assemblages of this period, and ribwort plantain is a plant often linked with early clearance and cultivation. The contents of the Middle Bronze Age cremation burial sample, although scanty, are also not untypical. Traces of possible cereal processing waste and the grass stem fragment, perhaps gathered dry for tinder, may represent part of the pyre fuel, and similar material has been recorded at Stonehenge (Clapham 1995), the Rollright Stones (Robinson 1988) and Hurst Park, Surrey (Hinton 1996). The charred remains in the post-holes of the three Late Bronze Age roundhouses are likely to have come from hearth sweepings which included fragments of fuel and waste from domestic food preparation. However, all that can safely be deduced from these charred remains is that burning occurred in the vicinity. The four-post structure suggests different activity and, apart from slightly greater numbers, the cereals and wild plant seeds suggest cereal preparation, or possibly storage. Cereals in the Late Bronze Age settlement included emmer and possibly other wheat species and hulled barley, most probably all 6-row, as concluded by Carruthers (1994) from similarly sparse remains from the Brigmerston Down linear ditch (LPD 052). The one oat grain at Dunch Hill is not specifically identifiable and may have been present only as a weed contaminant. Evidence of field conditions and surrounding environment at this time is also scanty. Wild plant seeds are those of common field weeds, with some evidence in plants such as Aphanes arvensis (parsley piert) and Linum catharticum (purging flax), of well drained calcareous soils. A few seeds such as the sedge, and the stem nodes of onion couch suggest grassland, possibly damp, but these plants could equally well have occurred in a cultivated area. Animal bone by Pippa Smith Only 29 of the 156 bones recovered were identifiable to species. The bones which have survived are those more robust bones (including teeth) which are more resilient to post-depositional processes and the assemblage has obviously been badly affected. Very little can be said about the economy of the site based on what is an unrepresentative sample of the original assemblage deposited. Sheep is the only species represented in the late Neolithic/Early Bronze Age pit (2 teeth), and cow in Middle Bronze Age ditch 402 (1 jaw fragment; the vole in cremation burial 463 is likely to be intrusive). The mammal bones from Late Bronze Age contexts comprise mostly sheep (13 frags, mostly teeth and foot bones) with cow (2 femur fragments), pig (1 tooth) and horse (2 teeth) also present. There are single occurrences of bird and amphibian. Roe deer is represented by a single metatarsal in a post-hole which, although undated, is most likely to be Late Bronze Age. The small Late Bronze Age assemblage is broadly comparable with that recovered from the midden deposit (Bradley et al. 1994, table 16). Detailed quantification is not given, but almost 200 bone fragments (including ribs) are recorded, with cattle, sheep/goat and pig identified. Radiocarbon dating by Michael Ff. Allen Material was carefully selected in order to attempt to date both cremation burial 439 and the construction of roundhouse 57. In each case charcoal, preferably sapwood or young lived species, was chosen to ensure the radiocarbon results reflected the event that it was hoped to date. It was thought that the cremation burial and round-house construction both belonged to a broad phase of Middle to Late Bronze Age activity, with cremation burial 439 the earlier of the two. It was also thought that the use M. Stulver and R.S. Kra eds. 1986 Radicarbon 28(2B): 805-1030; OxCal v2.15 cub r:4 ad;12 prob{chron} OxA-3046 242070BP | | | | | | | | Gu- a1e4 2940+50BP ii | er AA- 33134 30652508 nlf | 2000BC 1500BC 1000BC 500BC AD Calibrated date Fig. 16 Dunch Hill: Radiocarbon dates A MIDDLE TO LATE BRONZE AGE SETTLEMENT AT DUNCH HILL, TIDWORTH Al Table 5. Dunch Hill Radiocarbon dates Context description Material lab no result BP calibrated date (2 sig) cal BC Midden (LDP 081A) context 3 bone (unident.) OxA-3048 2420+70 770-390 Post-hole 45 (roundhouse 57) charcoal: oak sapwood GU-8184 2940+50 1320-1000 Cremation burial 439 __ charcoal; ash from pyre —_ AA-33134 3065+50 _ 1450-1210 of roundhouse 57 might be broadly contemporary with the Late Bronze Age midden at Dunch Hill excavated as part of the Linear Ditches Project and dated to 770-390 cal BC (OxA-3048, 2420+70 BP) (Bradley et al. 1994, table 22). The results were calibrated using OxCal v 2.15 and the data set provided by Stuiver and Kra (1986), and all calibrated date ranges are quoted at two standard deviations (95% confidence), with the end points rounded outwards to 10 years (Mook 1986). The calibrated ranges from cremation burial 439 and roundhouse 57 overlap, confirming that they both belonged to the same broad phase of (late) Middle—(early) Late Bronze Age activity on the site. However, the radiocarbon result from the discrete cluster of animal bone in the midden is clearly later, and the radiocarbon probability distribution shows no overlap with the other two results (Table 5 and Figure 16). DISCUSSION The 1995 excavation at Dunch Hill demonstrated, perhaps surprisingly, that archaeological features were generally well preserved in what, on initial appearance, was an unpromising environment. Beneath a heavily used, unsurfaced military track even small features such as post-holes survived to depths of up to 0.35m within the excavated area, with only localised damage mainly resulting from rainwater run-off. This survival might be attributed to the broad width of the track and the fact that few wheel ruts had developed. However, it is clear that some truncation of archaeological features had occurred, partly as a result of grading of the track, and no stratified deposits survived overlying the natural chalk. The main archaeological value of the 1995 excavation was that it provided a substantial transect across one of the principal sites investigated as part of the Linear Ditches Project (Dunch Hill, comprising LDP 081; LDP 081A; LDP 087; LDP 098;and LDP 103. Bradley et al. 1994, fig. 46. See Figure 17). This project had been undertaken in order to better understand the date and function of the prehistoric land divisions known as linear ditches and establish their relationship to the network of ‘Celtic’ fields on Salisbury Plain. Earlier work at Dunch Hill included surface artefact collection and very limited excavation but demonstrated the presence of a Late Bronze Age settlement, one of relatively few identified in the study area which covered this part of Salisbury Plain (Bradley et al. 1994, fig. 71; McOmish et al. 2002, 71, fig. 3.20A), and the only one so far to be excavated under modern conditions. Late Neolithic-Early Bronze Age The earliest archaeological feature recorded on the 1995 excavation was a shallow pit which contained small quantities of Late Neolithic Grooved Ware and Beaker pottery along with some worked flint including three end scrapers. No Neolithic pottery was found on the Linear Ditches Project (Raymond 1994, 71), although some Beaker pottery is recorded including two sherds from Dunch Hill (Bradley et al. 1994, table 23). No material of Neolithic or Early Bronze Age date was identified amongst the five worked flint assemblages examined which included one from Dunch Hill (Whitehead 1994, 91-100). Grooved Ware and Beaker pottery, however, has been found in a pre-barrow horizon at Snail Down, approximately 3km to the north of Dunch Hill, where evidence for structures survived in the form of stake-holes (Bradley et al. 1994, 20; Thomas 2005). The pit at Dunch Hill produced no grain or chaff and the charcoal, which included hazel, ash and oak, suggests the possibility of surrounding woodland. However, the land snail evidence from the earliest archaeological features (pit 482 and the tree-throw holes) shows that the former woodland had been cleared and open, probably grazed, grassland had already become established by the later Neolithic. The presence of sheep teeth in the pit tends to confirm local grazing. Evidence of localised later Neolithic clearance is not uncommon on Salisbury Plain, but as yet there are still few instances of evidence of earlier Neolithic clearance, with obvious major monument exceptions, and that of Mesolithic clearance near Stonehenge (Allen 1995). The activities associated with pre-Late Neolithic clearance of the woodland locally are not reflected in 72 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE _/ LDPo81 /, wa x 1995 W.A. excavation Lire) Vn} wy 4 a a WEOHY MOO ag WA pan © v2 ROR AY x \ Oo A hy" ANS We Ss LDP103 ~# A een - B RNS Ss Ses SY B £ oe Tt a,8 SO Pottery sherds:— 1-5 6-10 11-15 16-20 21-25 26-30 LDP 103; no pottery Se Lynchets 0 eee ‘ . ¥ “ x ' Scheduled Area Fig. 17 Dunch Hill: Plan, 1995 excavation and Linear Ditches Project sites in the vicinity (after Bradley et al. 1994, fig. 46) A MIDDLE TO LATE BRONZE AGE SETTLEMENT AT DUNCH HILL, TIDWORTH 73 the archaeological evidence (features or artefact) and have only been detected from palaeo-environmental evidence. The main density of Early Bronze Age activity is likely to have been in the fertile lowlands, away from the barrow cemeteries. Nothing is recorded of any antiquarian investigations or finds from the two barrows at Dunch Hill, but their presence alone might provide an indication of activity from the Beaker period throughout the Early Bronze Age. Middle Bronze Age At Dunch Hill the sequence of barrow-related activity might be extended through into the Middle Bronze Age by cremation burial 439 which lay less than 10m to the south of the smaller of the two barrows (SMR No. 605). It is likely that this represents a secondary burial, comprising two vessels and possibly two individuals, like those found in the barrows explored by Colonel Hawley on Brigmerston and Figheldean Down (Hawley 1910, fig. 21) and at Milston Down (Grinsell 1957, 183: Milston 2 and 23). The radiocarbon date of 1450-1210 cal BC (AA-33134, 3065+50 BP) from charcoal associated with cremation burial 439 provides a secure Middle Bronze Age date for the Dunch Hill example. Ditch 402 has also been assigned to the Middle Bronze Age. The alignment of this feature suggests either that it runs immediately to the south-west of the two, earlier barrows, or (and perhaps more likely) that these formed an integral part of a Middle Bronze Age field system. A similar ditch was found in the Linear Ditches Project excavation at Dunch Hill (Bradley et al. 1994, 49-50, fig. 45), aligned approximately north-east to south-west, and this possibly formed part of the same field system. Parallel and immediately to the north of this ditch was a hollow-way or trackway, approximately 2m wide, and defined by a pair of shallow wheel ruts. Both the ditch and hollow-way were sealed by a Late Bronze Age midden, and though neither feature was clearly demonstrated to be of Middle Bronze Age date this seems most likely. The only datable find was a single sherd of Deverel-Rimbury pottery from the ditch, and some later pottery might be expected to ~ have been present if either feature had been of Late Bronze Age date. Ditched boundaries and trackways are acommon feature of Lowland Zone Bronze Age field systems, delineating individual fields as well as blocks of fields (Fowler 1983), and the limited evidence from Dunch Hill suggests the existence of a similar a field system there. The land snail evidence from ditch 402 indicates that it was cut into a long-established open downland and infilled continually in an open dry calcareous environment typical of such field systems. The assemblages are probably representative of loose earth (Parable) contexts or close cropped, grazed, short grassland with bare patches of soil. Little change can be detected during the infill which probably represents a relatively short period of time. No vegetation regeneration was detected, nor evena longer grassy sward over the ditch which might have marked this feature out in its later, infilled, stage. Although there is evidence from some areas of southern England that woodland was still a prominent component of the downland until the Late Bronze Age (Entwistle 1994a, 101), this is not the case at Dunch Hill. Here, as in much of the Salisbury Plain area investigated through projects such as the Wessex Linear Ditches Project (Entwistle 1994a) and other linear ditches near Amesbury (Allen 1991), Earl’s Farm Down (Allen and Wyles 2004) and colluvial deposits at Figheldean (Allen and Wyles 1993; Allen 1992), it is evident that clearance was widespread by the Middle Bronze Age at least, often starting in the later Neolithic. Middle Bronze Age activity appears to have been concentrated in certain areas (eg Cranborne Chase and the Marlborough Downs) and there is comparatively little evidence for occupation from the Salisbury Plain Training Area, though this may reflect the less detailed investigations undertaken here. It may also in part be because evidence for occupation has been obscured by later fields and lynchets, as appears to be the case at Dunch Hill, and because settlements were often unenclosed and therefore have left little or no trace, with earthwork enclosures forming a characteristic but perhaps minor element in the landscape. Certainly, there are several poorly dated enclosures in the area which, on the basis of their size and form are likely to have been Middle Bronze Age settlements (Roy Entwistle pers. comm.; Bradley et al. 1994, fig. 15). However, there may have been a real paucity of Middle Bronze Age occupation sites in the area, coupled with a low level of arable activity, and widespread reorganisation of landuse on the chalk may not have taken place until the Iron Age. Late Bronze Age Prior to the Linear Ditches Project (Bradley et al. 1994) there was relatively little evidence for Late Bronze Age settlement in the Salisbury Plain Training Area and the date and function of the linear ditches and ‘Celtic’ fields were poorly understood. 74 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE The results of this project have significantly increased our knowledge of the Late Bronze Age landscape of the area, and the subsequent excavation at Dunch Hill in 1995 has allowed the detailed investigation of part of one of the settlements. The value of this work has been increased further by the fact that it can fitted into a longer sequence of activity on this site which began prior to the Late Neolithic period and may have continued into the Iron Age and, perhaps, Romano-British periods. Subsequent work, including fieldwalking, has identified the sites of at least four further probable Late Bronze Age settlements within 3km or so of Dunch Hill (Roy Entwistle pers. comm.). Whether there was any continuity of activity between the Middle and Late Bronze Age at Dunch Hill remains uncertain despite the overlap in radiocarbon dates from cremation burial 439 and roundhouse 57. The ditches and the hollow-way assigned to the Middle Bronze Age appear to have silted up and fallen into disuse by the Late Bronze Age, but some activity may have continued and it is quite probable that the Early Bronze Age barrows provided a focus for the later settlement. Roy Entwistle provides a summary of the impression formed of the Middle Bronze Age/Late Bronze Age settlement sequence in the Dunch Hill/Brigmerston Down area based on the work of the Linear Ditches Project: ‘The “classic” Deverel-Rimbury ceramic phase was associated with small, enclosed settlements, probably existing alongside an emergent tradition of open settlements which pre-figured the larger, unenclosed settlements of the Late Bronze Age. Almost without exception, surface spreads of Late Bronze Age ceramics, representing unenclosed settlement, occurred in conjunction with smaller quantities of Deverel-Rimbury pottery. Significantly, the Middle Bronze Age pottery from these sites differed somewhat from the “classic” Deverel-Rimbury styles, and had more in common with ceramics of the Late Bronze Age. It may be that these trends are evidence of a general progression from enclosed to unenclosed settlement, with the latter becoming dominant towards the end of the Deverel-Rimbury period’. in this postulated sequence, the settlement at Dunch Hill could be seen as belonging to a transitional phase between the Deverel-Rimbury and Late Bronze Age periods, although any such dating is to some extent hampered by our poor understanding of the absolute chronology for the conventional period classifications. The Linear Ditches Project defined an ‘Upper Study Area’, and within this identified a ‘Northern Core Territory’ and a contiguous ‘Southern Core Territory’, both delineated by Late Bronze Age linear ditches (Bradley et al. 1994, fig. 10). These territories lay between the Bourne Valley and the Nine Mile River and formed the most coherent pattern of linear ditch territories within this area. The site at Dunch Hill lay just to the north of the Northern Core Territory (see Figure 17) and, as a result of work undertaken as part of the Linear Ditches Project and the 1995 excavation, is the most intensively investigated Late Bronze Age settlement in the area. The 1995 excavation provided a transect across the area studied by the Linear Ditches Project and revealed part of an open settlement comprising roundhouses, four-post structures, pits and several fence lines possibly defining small stock enclosures (Figure 17). There was no intercutting of features or overlapping of structures suggesting, perhaps, that they were broadly contemporaneous and belonged to a single, relatively short-lived phase of occupation, although this seems to be contradicted by the radiocarbon dating evidence (see above). Very few finds were present as a result of the truncation of deposits overlying the natural chalk, and dating of this phase of activity relies on the small quantities of post-Deverel-Rimbury Plain Ware pottery recovered from several of the post-holes and shallow pits which is not closely datable, and a single radiocarbon date of 1320-1000 cal BC (GU-8184, 2940+50 BP) from the charred remains of a post in roundhouse 57. This evidence and the results of the earlier, Linear Ditches Project investigation suggest that the Late Bronze Age settlement at Dunch Hill should probably be assigned to the later second and early Ist millennium BC, with occupation ceasing around the beginning of the 8th century BC. It would, therefore, as suggested above, have been established towards the end of the Middle Bronze Age and possibly continued throughout much of the Late Bronze Age. The Late Bronze Age pottery from the site has been identified as belonging to the final phases of the Plain Ware ceramic tradition (Raymond 1994, 86), but only three sherds of All Cannings Cross pottery were present, all from the earlier excavation of a midden deposit during the Linear Ditches Project work (Bradley et al. 1994, table 15. see below), indicating that the site had been virtually abandoned by the A MIDDLE TO LATE BRONZE AGE SETTLEMENT AT DUNCH HILL, TIDWORTH 75 Early Iron Age. Four roundhouses lay within the excavated area, spread out over a distance of approximately 250m at intervals of between 50m and 100m. Two of these roundhouses, 57 and 346, were very similar in plan comprising circular arrangements of post-holes approximately 7m in diameter, representing roof supports inside the hut, with a single, central post- hole. Roundhouse 196 may also have been similar, but was less well preserved, and no central post-hole was identified. Roundhouse 196 also contained some evidence for an inner ring of posts, albeit rather irregular, and the possibility that roundhouse 346 was at least partly rebuilt is suggested by the presence of several double post-holes. Roundhouse 196 enclosed a small pit containing sherds of two jars, assumed to be contemporary with the structure. No other internal features were identified in round- houses 57, 196 and 346, and none of the post-hole rings included pairs of larger post-holes which might indicate the positions of doorways. No roundhouse at Dunch Hill had post-holes marking porches, which are characteristic of Middle—Late Bronze Age roundhouses elsewhere (e.g. Gingell 1992, figs 4 and 5), but it is possible that in some cases these may have lain to the south, outside the excavated area. Similar structures to roundhouses 57, 196 and 346, with and without porches, have been excavated at Easton Lane, Hampshire, where they are dated to the Middle Bronze Age (Fasham et al. 1989, figs 44 and 54), and Balksbury Camp, Hampshire, where they are dated to the early Iron Age (Wainwright and Davies 1995, fig. 19) The Late Bronze Age examples at Dunch Hill seem to fit this chronological sequence. The paucity of finds means that little can be deduced about differences of use between or within individual roundhouses at Dunch Hill, with the possible exception of roundhouse 158. Roundhouse 158 was the least regular of the four structures and its plan was somewhat obscured by a scatter of other post-holes in this part of the site which were generally absent elsewhere. A shallow, possibly internal, pit was present and at least three shallow scoops filled with burnt flint lay in the immediate vicinity, including one (104) found during the Linear Ditches Project (LDP 097; Bradley et al. 1994, 33). The burnt flint may have derived from food preparation (cooking) and its presence hints at the possibility of a different ‘function for roundhouse 158. Other roundhouses are likely to have lain outside the limits of the excavated area and two or three slight ‘platforms’ interpreted as possible house sites were noted during the Linear Ditches Project (Bradley et al. 1994, 50), although these were not discernible at the time of the 1995 excavation. The two four-post structures, 338 and 515, lay close together between roundhouses 57 and 196. Such structures, common on sites from the Middle Bronze Age through to the Iron Age, are often interpreted as granaries, although other uses such as hay stores or excarnation towers have been postulated. The plant remains from the post-holes included small amounts of charred grain, generally absent from other features on the site, and this would support their interpretation as granaries at Dunch Hill. A single two-post structure was identified, to the south of roundhouse 158, and these also are common on Bronze Age and Iron Age sites where they are sometimes interpreted as racks. The three small pits which lay together close to the four-post structures contained few finds and no carbonised grain, and their function(s) remains uncertain. These, as well as the four-post structures, apparently lay within the settlement. The fence lines cannot be dated with certainty, though a Middle or Late Bronze Age date is probable. They clearly formed part ofa coherent group, aligned parallel or at 90° to Middle Bronze Age ditch 439 to the north-east (as well as to the general slope of the ground), but did not extend beyond Late Bronze Age roundhouse 346 to the south-east. On balance, a Late Bronze Age date is preferred here for this group, and they may have formed a series of small stock enclosures on the edge of the settlement. Fence lines have also been identified on the Middle Bronze Age site at Easton Lane, Hampshire (Fasham et al. 1989, fig. 53), and the Late Bronze Age site at Black Patch, Sussex (Drewett 1982, fig. 4), but these latter arrangements were less regular and seem to have been more directly related to structures, while at Dunch Hill they appear to have been separate. Although relatively few finds were recovered during the 1995 excavation, the earlier Linear Ditches Project work identified in a small trench (LDP 081A, which measured 9m by 2m) what was interpreted as a Late Bronze Age midden less than 100m to the north-west (Bradley er al. 1994, 49-50; see Figure 17). This midden, approximately 0.2m thick and 5m across (Bradley et al. 1994, fig. 45), sealed a Middle Bronze Age ditch and associated trackway (see above), and the excavated section produced almost 300 sherds of Late Bronze Age Plain Ware, approximately 250 fragments of animal bone and some worked flint as well as burnt flint. Animal bone from the midden produced a relatively 76 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE late radiocarbon date of 770-390 cal BC (OxA-3048, 2420+70BP; Bradley et al. 1994, table 22): ‘although the presence of Plain Ware throughout this deposit, and the absence of Iron Age ceramics, would favour a date at the beginning of this rather broad date range’ (Raymond 1994, 88). Relatively large quantities of Late Bronze Age pottery were also recovered from fieldwalking of the surrounding area (LDP 081), with smaller quantities from the tree-hole survey to the south (LDP 103), anda further 44 sherds from very limited excavation in the edge of the track (LDP 087). It seems almost certain that this material derives from the Late Bronze Age settlement excavated in 1995, and that the midden represents a more discrete dump of debris, perhaps from a single household of later date than roundhouse 57, the only example determined by radiocarbon dating. While the nature of some of the surface artefact collection was unsystematic, it nevertheless helps to broadly define the Late Bronze Age settlement at Dunch Hill. The extent to the north-east and south-west has been demonstrated by the structural evidence revealed in the 1995 excavation, and the extent to the north-west and south-east perhaps by the greater concentrations of pottery which may represent debris being disposed of on the edge of the settlement. If this assumption is correct then it indicates an unenclosed settlement covering an area of perhaps 4 or 5 hectares. There was a marked fall-off in pottery towards the linear ditch which lay less than 100m beyond the probable southern extent of the settlement while the excavated section of this ditch produced very little pottery (LDP 098. Bradley et al. 1994, 55). This may reflect the later insertion of this boundary between the settlement at Dunch Hill, outside the ‘Northern Core Territory’ defined by the linear ditch, and the settlement at LDP 080 (identified from fieldwalking) less than 1km to the south, the nearest site within this territory (Bradley et al. 1994, 128). The discovery in 1980 of an unaccompanied crouched inhumation less than 30m to the south of the ditch may provide evidence for burial on the edge of the territory (SPTA 1908. Bradley et al. 1994, 128). Certainly, the ceramic evidence from the Linear Ditches Project has demonstrated closer links between Dunch Hill and Milston Down (LDP 112) 3km to the south, both of which lay outside the ‘Northern Core Territory’, than between Dunch Hill and site LDP 080 within this territory but less than 1km away (Raymond 1994, 72). The environmental evidence from Dunch Hill indicates an open landscape with evidence, albeit slight, for both arable and pastoral agriculture. The charred plant remains include wheat and field weeds | and the four-post structures are likely to represent granaries. Sheep are the most common (or better preserved) animal bones, but little can be deduced from the small numbers present. It is suggested, however, that the fence lines may have demarcated stock enclosures. Iron Age—Romano-British Three sherds of All Cannings Cross pottery, all from the midden deposit excavated during the Linear Ditches Project work (Bradley et al. 1994, table 15), provide the only evidence for occupation continuing in to the Early Iron Age. Between the 8th and the 5th centuries BC there were changes to the system of linear ditches and open settlements were abandoned, indicating ‘a radical restructuring of previous territorial arrangements’ (Entwistle 1994b, 122), which, in the vicinity of Dunch Hill, saw a shift of activity towards Sidbury Hill approximately 2km to the north-east. At Dunch Hill, a series of low lynchets have been recorded immediately to the north of the 1995 excavation, forming part of a large ‘Celtic’ field system stretching across Dunch Hill Plantation and Tidworth Golf Course, and apparently extending across the linear ditch at Dunch Hill (Entwistle 1994b, 122, fig.10; Figure 17). It is possible that elements of this field system have taken their alignments from earlier, Middle Bronze Age ditches (note the correspondence between ditch 7 and the later lynchet on LDP 081A; Bradley et al. 1994, figs 45 and 46), although most may have originated in the Iron Age, with their final form reflecting Roman activity (Entwistle 1994b, 129). Unfortunately, the 1995 excavation shed little light on this possible sequence, as any former lynchets had been destroyed within the area covered by the track. However, Dunch Hill would provide an excellent site on which to test the validity of this proposed sequence given the work already undertaken and the quality of the results obtained. ACKNOWLEDGEMENTS The project was commissioned by Jane Hallett (Defence Estates) and funded by Headquarters, Salisbury Training Area. Duncan Coe of Wiltshire County Council Museums Service wrote the brief and monitored the excavation. Wessex Archaeology would particularly like to thank Jane Hallett and Arthur Kendrew of Defence Estates for their A MIDDLE TO LATE BRONZE AGE SETTLEMENT AT DUNCH HILL, TIDWORTH 77 assistance throughout the excavation including arranging an army helicopter for aerial photography of the site. Post-excavation and publication has been made possible by Ian Barnes of Defence Estates. This report has benefited from advice provided by various individuals including Roy Canham, Duncan Coe, Ros Cleal, Andrew Lawson and Frances Raymond. We are particularly grateful to Roy Entwistle for reading and providing constructive criticism of this text, especially the discussion, and for making available the results of recent unpublished work in the eastern part of the Training Area. The project was managed for Wessex Archaeology until 1996 by Ian Barnes and subsequently by Roland Smith. The excavation was directed by Phil Andrews and Andrew Powell, supervised by Dave Murdie with a team comprising Kevin Ritchie, Paul Pearce, Richard May, Jez Fry, Joe Whelan and Dom Barker. Ian Barnes recorded the archaeological remains exposed during the digging of a military slit trench in 1998. Michael J. Allen (Environmental Manager) has managed the programme of environmental analysis and Lorraine Mepham (Finds Manager) has edited the finds report for publication. S. E. James prepared the illustrations and Elaine Wakefield took the aerial photographs. Sarah Wyles and Hayley Clarke undertook the sorting of material for environmental analysis and Hayley Clarke extracted the snails. Full details of all contexts are contained in the site archive which will eventually be deposited with Wiltshire Heritage Museum, Devizes. BIBLIOGRAPHY ALLEN, M.J., 1991, ‘The environmental evidence (A303)’, in Bellamy, P, The investigations of the prehistoric landscape along the route of the A303 road improvement between Andover, Hampshire and Amesbury, Wiltshire 1984-7. Proceedings of the Hampshire Field Club & Archaeological Society 47, 41-5 ALLEN, M.J., 1992, Products of erosion and the prehistoric land-use of the Wessex Chalk, in Bell, M.G. and Boardman, J.(eds), Past and Present Soil Erosion; archaeological and geographical perspectives. Oxford: Oxbow, 37-52 ALLEN, M.J., 1995, ‘Before Stonehenge’, in Cleal et al. 1995, 41-64 ALLEN, M.J. and WYLES, S.E, 2004, ‘Land Mollusca’ {Earl’s Farm Down], in Cleal, R.M.J., Allen, M.J. and Newman, C., An archaeological and environmental study of the Mesolithic and later prehistory of the Avon valley between Durrington Walls and Earl’s Farm Down. WANHM 97, 218-48 ALLEN, M.J. and WYLES, S.F, 1993, ‘The land-use history; the molluscan evidence’, in Graham, A. and Newman, C., Recent excavations of Iron Age and Romano-British Enclosures in the Avon valley, Wiltshire. WANHM 86, 45-50 BARRETT, J.C., 1980, The pottery of the Later Bronze Age in lowland Britain. Proceedings of the Prehistoric Society 46, 297-319 BIRBECK, V., this volume, Excavations on the Old Ditch Linear Earthwork, Breach Hill, Tilshead, 79-103 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 of the Prehistoric Society 46, 217-95 BRADLEY, R., ENTWISTLE, R. and RAYMOND, F, 1994, Prehistoric Land Divisions on Salisbury Plain: the work of the Wessex Linear Ditches Project. London: English Heritage Archaeological Report 2 CAMERON, R.A.D. and MORGAN-HUWS, D.L., 1975, Snail faunas in the early stages of a chalk grassland succession. Biological Journal of the Linnean Society 7, 215-29 CARRUTHERS, W., 1994, ‘Carbonised cereal remains’, in Bradley et al. 1994, 119 CLAPHAM, A., 1995, ‘Charred tuber of Arrhenatherum elatius var. bulbosum (Wild) St Amans (Onion couch grass) from Aubrey Hole 5’, in Cleal et al. 1995, 461-2 CLEAL, R. and RAYMOND, F. 1990, ‘The prehistoric pottery’, in Richards 1990, 233-46 CLEAL, R.M.J.. WALKER, K. and MONTAGUE, R., 1995, Stonehenge in its Landscape: the twentieth-century excavations. London: English Heritage Archaeological Report 10 DREWETT, P, 1982, Later Bronze Age downland economy and excavations at Black Patch, East Sussex. Proceedings of the Prehistoric Society 48, 321-400 EDLIN, H.L., 1949, Woodland Crafts in Britain. Cambridge: Batsford ENTWISTLE, R., 1994a, ‘The environmental setting of the linear ditches system’, in Bradley et al. 1994, 101-21 ENTWISTLE, R., 1994b, ‘Settlements, territories, and ‘Celtic’ fields:the changing role of boundary earthworks’, in Bradley et al. 1994, 122-36 EVANS, J.G., 1972, Land Snails in Archaeology. London: Seminar Press EVANS, J.G. 1984, Stonehenge — the environment in the Late Neolithic and Early Bronze Age and a Beaker burial. WANHM 78, 7-30 FASHAM, PJ., FARWELL, D.E. and WHINNEY, R.J.B., 1989, The Archaeological Site at Easton Lane, Winchester. Winchester: Hampshire Field Club Monograph 6 FOWLER, PJ, 1983, Farming in Prehistoric Britain. 78 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Cambridge: Cambridge University Press GALE, R., 1990, ‘Charcoals’, in Richards 1990, 252-3 GALE, R., 1994, ‘Charcoal identification’, in Bradley et al. 1994, 119 GALE, R., 2006, ‘Charcoal’, in Birbeck, V. this volume GINGELL, C., 1992, The Marlborough Downs: a Later Bronze Age landscape and its origins. Devizes: WANHM Monograph 1 GRINSELL, L.V., 1957, Archaeological gazeteer. Victoria County History of Wiltshire, 1.1, 21-279 HAWLEY, W., 1910, Notes on barrows in south Wiltshire. WANHM 36, 615-28 HINTON, P, 1996, ‘Charred plant remains’, in Andrews, P., Hurst Park, East Molesey, Surrey: Riverside settlement and burial from the Neolithic to the Early Saxon Periods, in Andrews, P. and Crockett, A. Three Excavations along the Thames and its Tributaries, 1994; Neolithic to Saxon settlement and burial in the Thames, Colne and Kennet Valleys. Salisbury: Wessex Archaeology Report 10, 95-9 HOLDEN, J. L., PHAKEY, P. P and CEMENT, J. G., 1995, Scanning electron microscope observation of incinerated human femoral bone: a case study. Forensic Science International 74, 17-28 HURLBERYT, S.H., 1971, The nonconcept of species diversity: a critique and alternative parameters. Ecology 52, 577-86 LONGWORTH, I.H., 1971, ‘The Neolithic pottery’, in Wainwright, G.J. and Longworth I.H. Durrington Walls: Excavations 1966-1968. London: Reports of the Research Committee of the Society of Antiquaries of London 29, 48-155 McKINLEY, J.I., 1994, The Anglo-Saxon Cemetery at Spong Hill, North Elmham Part VII: The Cremations. East Dereham: East Anglian Archaeology 69 McKINLEY, J.I., 1997, Cremated human bone from burial and pyre-related contexts, in Fitzpatrick, A.P. Archaeological Excavations on the Route of the A27 Westhampnett Bypass, West Sussex, 1992. Volume 2: the cemeteries. Salisbury: Wessex Archaeology Report 12, 244-52 McMINN, R.M.H. and HUTCHINGS, R.T., 1985, A Colour Atlas of Human Anatomy. New York: Wolfe Medical Publications McOMISH, D., FIELD, D. and BROWN, G., 2002, The Field Archaeology of the Salisbury Plain Training Area. Swindon: English Heritage MOOK, W.G., 1986, Business meeting: Recommendations/ resolutions adopted by the 12th International Radiocarbon Conference. Radiocarbon 28, 799 MORRIS, E.L. 1994, The Analysis of Pottery. Salisbury: Wessex Archaeology Guideline 4 RAYMOND, F.,, 1994, ‘The pottery’, in Bradley et al. 1994, 69-90 RICHARDS, J., 1990, The Stonehenge Environs Project. London: English Heritage Archaeological Report 16 ROBINSON, M., 1988, ‘The significance of tubers of Arrhenatherum elatius (L.) Beavu. from site 4, cremation 15/11’, in Lambrick, G. The Rollright Stones. London: English Heritage Archaeological Report 6, 102 STUIVER, M. and KRA, R.S., 1986, Calibration issue, Proceedings of the 12th International Radiocarbon Conference. Radiocarbon 28, 805-1030 THOMAS, N., 2005, Snail Down. Devizes: WAHNS Monograph 3 TUTIN, T.G. and HEYWOOD, V.H. et al., 1964-80, Flora Europaea 1-5. Cambridge: Cambridge University Press WAINWRIGHT, G.J. and DAVIES, S.M., 1995, Balksbury Camp, Hampshire: Excavations 1973 and 1981. London: English Heritage Archaeological Report 1 WESSEX ARCHAEOLOGY, 1995a, Dunch Hill, Tidworth: Archaeological Evaluation. Unpublished client report WESSEX ARCHAEOLOGY, 1995b, Excavations at Dunch Hill, Tidworth: Interim Assessment Report and Proposals for Analysis and Publication. Unpublished client report WHITEHEAD, R., 1994, ‘The analysis of five flint assemblages from the Upper Study Area’, in Bradley et al. 1994, 91-100 2 between Copehill Down and Breach Hill. Rapid Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 79-103 Excavations on the Old Ditch Linear Earthwork, Breach Hill, Tilshead by Vaughan Birbeck with contributions from Michael Ff. Allen, Rowena Gale, S. Hall, Phil Harding, Emma Loader, Lorraine Mepham and M.S. Tite Excavation and earthwork recording were undertaken of a Scheduled Monument at Breach Hill, Tilshead, where a substantial area of c. 1.57ha had been inadvertently stripped of topsoil during construction of a military track. Stripping had destroyed the previously upstanding bank of the earthwork and, in places, truncated the upper fills of the ditch. A smaller ditch parallel to the large ditch lay to the north of the bank. Excavations produced very little artefactual dating evidence for the construction/primary silting of either ditch. A Mid-Late Bronze Age or later date is indicated for the monument, however, by comparison with similar monuments on Salisbury Plain. Optically stimulated luminescence (OSL) dating was also used. Other deposits included a pit or tree- throw hole containing Late Mesolithic flintwork, a subsoil area which contained large quantities of worked flint, burnt flint and Beaker pottery, two Late Bronze Age or Early Iron Age features, a separate subsoil which produced evidence of Romano-British activity, and several undated features. The site comprised an area approximately 545m from east to west, averaging 27.5m wide, centred on SU 015 465, between 138m and 141m OD, immediately below the crest of the north-facing slope of Breach Hill, c. 2.5km south-west of Tilshead. The underlying geology comprised weathered Upper Chalk, overlain by Clay-with-flints deposits mainly towards the western end of the site, although small patches were observed across the remainder of the area. INTRODUCTION An area of approximately 1.57ha had been inadvertently stripped of topsoil over a ditch and bank earthwork, known as the Old Ditch, at Breach Hill, Tilshead (Scheduled Monument WI 10109/1; SMR reference SU04NW615), on Salisbury Plain Training Area, during construction of a track survey noted the known ditch and a previously unsuspected second smaller ditch running parallel. The bank, extant before stripping, was situated between the two ditches. Stripping had destroyed the bank and in places truncated the upper fills of the ditches, though in some areas they survived to their full depth. Wessex Archaeology undertook excavation and earthwork recording to determine ARCHAEOLOGICAL BACKGROUND No excavation had previously been undertaken on the monument disturbed by topsoil stripping. the function, character and date of the monument and to recover information about the contemporary environment, while assessing the level of destruction sustained. AC Archaeology conducted work in 1995 on the monument within 20m east of the Scheduled area during monitoring of the track construction. This work comprised two machine trenches over the Wessex Archaeology, Portway House, Old Sarum Park, Salisbury SP4 6EB 80 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Breach Hill > + \ \ 4 +k =WAExcavations | | Fig. 1 Breach Hill: Site location plan southern ditch, the upper surface of which was cleaned and recorded. It was noted that the upper fill had been deposited relatively recently, as plastic and glass were noted within it, although no fill was removed during this exercise. Archaeological evaluations before construction of the Southern Range Road (Gifford and Partners 1997) and upgrading of Track 58 (Wessex Archaeology 1996a), both to the west of the excavation area, did not find any continuation of the monument. It must be concluded that it terminates within c. 200m of the western limits recorded at Breach Hill. The monument comprises a ditch and bank that runs along the north-facing slope of the ridge of Breach Hill, slightly below the crest, with the ditch on the southern, upslope, side of the bank. The stripped area was situated at the western end of the monument, which extends for c.5km to the east (Figure 1, W9418). From the eastern end of the stripped area, it is visible on the ground as a low bank with a very shallow ditch to the south, and continues eastwards towards Old Ditch long barrow, investigated in 1865 by John Thurnam (1869). Here the larger southern ditch appears to run along the southern side of the barrow quarry ditch. It then continues, though only as a cropmark, across a dry valley and up a north-west facing slope towards a second long barrow (White Barrow). In the base of the dry valley, cropmark evidence suggests that it forms the southern boundary of a small enclosure. The ditch then changes direction, passing to the south of White Barrow and continues east-north- east, across the valley of the river Till and up the north-west facing slope of Orcheston Down, heading towards a large round barrow (Silver Barrow). A little below the crest of the hill the earthwork changes direction to the north-east and continues along the north-west facing ridge towards West Down. It continues north-eastwards, following the 135m contour across the north-west facing slope of West Down, for c. 2km, apparently terminating at a “T” junction with a similar linear earthwork, NE616. Further linear ditches in the vicinity survive as earthworks or cropmarks (Figure 1) or have been recorded during recent excavations and evaluations (Wessex Archaeology site codes W251, W9409 and W2024). The Breach Hill ditch (NW615), along with NE616, NW616, NW614 and NW634 (see Figure 1), appear to enclose a large area of downland potentially EXCAVATIONS ON THE OLD LINEAR EARTHWORK, BREACH HILL, TILSHEAD 81 part of a larger system of land divisions. The Tilshead parish boundary follows the course of the Breach Hill ditch for much of its length. The Breach Hill ditch may also form the northern boundary ofa further area with the southern boundary represented by ditch NE636, c. 1.2km to the south. The survey work of the Wessex Linear Ditches Project on the eastern side of the Salisbury Plain Training Area has highlighted the density and importance of such features (Bradley et al. 1994). This work showed that many linear ditches were probably established in the Middle to Late Bronze Age and maintained throughout the later prehistoric period. Their relationship with other contemporaneous and later landscape features was evidently complex. The function of these boundaries is a matter of conjecture; one interpretation is that they form a network of ‘ranch’ boundaries superimposed on the landscape, using earlier monuments such as long barrows as reference points. METHODS Damage to a scheduled monument is undesirable, but in this case provided a rare opportunity to examine the structure of a major prehistoric earthwork. The work was undertaken in order to address questions of survival, form, chronology, maintenance, environmental setting and function. Bulldozer tracks, residual spoil and the crusted surface were machine-stripped over the entire area and a contour survey undertaken. Two spoilheaps from the stripping were subjected , to systematic surface collection. The surface of the stripped area was cleaned by hand. Any artefacts recovered were recorded by 10m grid square, based on the site grid. Excavation then proceeded with a 10% sample of each linear feature excavated by hand and deep homogeneous deposits removed in spits of 0.10m. All other archaeological features were fully recorded and a sample of each excavated. Two 15m wide sections and three 5m wide sections were excavated across the two known ditches and the area of the former bank (Figure 2), with all deposits removed stratigraphically by hand. The excavation produced few datable finds and so the fills of both ditches were sampled for optucally stimulated Luminescence (OSL) dating, the results of which are presented below. In addition, the secondary fill of the southern ditch was spot sampled in an attempt to archaeomagnetically date its deposition. The material contained a weak remnant magnetisation of low stability and random orientation and no date could be calculated. Sampling for land snails and carbonised (charcoal and plant) remains was undertaken. Two columns of snail samples were taken from the ditches, a single column of contiguous samples being taken from each ditch at the most appropriate location. Columns from the ditches were related as closely as possible to sections sampled for archaeomagnetic and optical dating. Bulk samples were also taken from the same sections. To augment the research of the Wessex Linear Ditches Project (Bradley et al. 1994) and Allen (Wessex Archaeology 1988), samples were also taken from two putative Holocene tree hollows sealed beneath the bank and thus pre-dating the linear ditches. RESULTS In addition to the Middle-Late Bronze Age linear earthwork three further phases of archaeological activity were recorded (Figure 2). These comprised Mesolithic activity, represented by an assemblage of diagnostic flintwork recovered from a single feature; Early Bronze Age activity represented by large quantities of pottery, flintwork and burnt flint recovered from a subsoil deposit, and Romano- British activity represented by finds from a second subsoil deposit. Several undated features were also excavated. Unstratified finds recovered during hand cleaning and collection from the spoilheaps indicated slight concentrations of worked flint in the area around slot 6000 and around the Early Bronze Age subsoil, which was subjected to more intensive collection. The general overall distribution of worked and burnt flint was higher towards the western end of the site, in the area of the Clay-with- flints. Relatively high concentrations of worked and burnt flint were also noted in the Romano-British subsoil in this area, as were all but one of the seven sherds of pottery recovered (two Late Bronze Age, one Late Iron Age/Roman and four Roman). Phase 1: Mesolithic A single feature, 1537, was dated to the Late Mesolithic. Other material of this date was found in later features, of which the only diagnostic item was a single burnt crested piece of flint, recovered from a Romano-British or later subsoil deposit 4m to the north of the pit. THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE (LBA/EM) o—1001 1517" 1513__ > -1500 1519: 1511 1539 ~~ Fig. 2 Breach Hill: Site plan showing all excavated features EBA subsoil Romano-British subsoil Modern disturbance Undated archaeological feature WSS NS SES DSS EXCAVATIONS ON THE OLD LINEAR EARTHWORK, BREACH HILL, TILSHEAD 83 Feature 1537 lay towards the western end of the excavation area (Figure 2), and was cut through Clay-with-flints into the underlying chalk. It was sub-circular, c.1.90m in diameter and 0.48m deep with steep, slightly irregular sides and an irregular base (Figure 3). The fills comprised yellowish brown silty clay with varying quantities of flint and chalk inclusions, probably the result of natural silting. A large quantity of worked flint, including a microlith, a microburin and other diagnostic material, was recovered, most (240+ pieces) from the two upper fills. Flintwork included refitting pieces and microdebitage, indicating a dump of Late Mesolithic knapping debris. While this feature may be man-made, its irregular form suggests a natural origin, possibly a tree throw. Phase 2: Early Bronze Age Beaker pottery, burnt and worked flint were recovered from a subsoil deposit that continued beyond the northern limit of excavation (Figure 2, bottom). This area was cleaned by hand, and finds were collected by 1m grid square. The deposit was then sampled using 18 1m sq. test pits. The subsoil comprised dark yellowish brown silty clay with abundant flint inclusions that appeared to have slumped into the top of a large irregular hollow (feature 1064) and numerous smaller depressions, all of which seemed to be natural in origin. Feature 1064 (Figure 4) comprised avery irregular, curving hollow, approximately 4.2m long, 0.40-0.90m wide and 0.10—0.25m deep. The fill contained slightly lower quantities of flint nodules towards the base than the rest of the subsoil deposit but was otherwise indistinguishable. Although this could be a man-made feature, its very irregular form suggests that it is of natural origin. Large assemblages of Early Bronze Age pottery, worked flint and burnt flint were recovered and their general distribution within the subsoil (Figure 4) mirrors that of the finds recovered from the hand cleaning. The greatest concentrations were recovered from the deeper parts, in particular in and around feature 1064. A marked concentration of worked flint and burnt flint was noted in test pits 1026 and 1065, towards the eastern end of the deposit, where its depth did not exceed 0.08m. Environmental samples were taken from the subsoil (context 1068) and from the fills of feature 1064 (contexts 1069 and 1070). Small quantities of charcoal, including hazel, oak and, in context 1068, pine, were recovered, although charred plant remains Pit 1537 Me Flint Silty clay Fig. 3 Breach Hill: Plan and section of feature 1537 and mollusca were almost completely absent. Phase 3: Middle-Late Bronze Age and Early Iron Age The third phase of activity is represented by the scheduled ditch and bank monument (Figures 2 and 5) and by the smaller northern ditch. The only other feature dated to this phase was a single post- hole towards the western end of the excavation, although undated features in the vicinity may belong to this phase. The southern ditch A length ofc. 450m of the southern ditch was exposed and this varied in width between 4.0m and 7.0m and was 1.40-1.80m deep with a slightly irregular ‘V’ shaped profile (Figure 5). The sequence of fills was broadly similar in each of the five excavated sections. The primary fill comprised very abundant medium to large chalk rubble fragments within a light yellowish brown silty clay matrix, varying from 0.40m to 0.55m thick. The texture, colour and composition of this fill varied only very slightly along the length of the ditch. The only artefacts recovered were four undiagnostic flint waste flakes and a well-made end scraper on a cortical flake. ?) At the western end of slot 3000 (Figure 2 84 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Worked flint ‘ % Numbers: ® 25 — @ 2-50 @ 51-7 | @ SS 5 5 - i 2 i & 101-125 & 126-150 Pottery Burnt flint Weight in gms: A 1-200 A 201-400 A 401-600 A 601-800 A 801-958 Space ee Fig. 4 Breach Hill: Distribution of artefacts within Early Bronze Age subsoil and feature 1064 EXCAVATIONS ON THE OLD LINEAR EARTHWORK, BREACH HILL, TILSHEAD Mollusc column 85 Section through northern and southern ditches (slot 4000) “ig. S 86 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE the primary fill was overlain by a very pale grey calcareous silty clay deposit with sparse flint inclusions, which was probably loessic in origin. This was in turn overlain by a yellowish brown silty clay secondary fill with abundant small to medium chalk and sparse subangular flint inclusions. This was present in all excavated slots and varied between 0.10m and 0.25m in depth. A small assemblage of undiagnostic flint waste flakes were the only finds recovered from this deposit. Overlying the secondary fill in all slots was an irregular narrow deposit of large, subangular flint nodules within a dark brown silt loam matrix, containing a large assemblage of worked flint, mostly unretouched waste flakes and a single sherd of Late Bronze Age pottery, in slot 6000 (context 6007). Overlying this was a dark brown silty loam with common chalk and sparse flint inclusions, from which small quantities of Romano-British coarseware pottery were recovered. This deposit was characterised by a fine ‘spongy’ texture and contained small, widespread patches of burnt material and very few inclusions. All fills above the flint deposit, in general the upper 1m, were found to be quite heavily disturbed by burrowing activity and in places by modern military activity. A sample taken from this deposit in slot 4000 (context 4014), the only slot where this deposit was relatively undisturbed by rabbits, was dated by OSL to AD 790-960 (1514a, 1190+90 before 1996). The northern ditch Approximately 365m of the northern ditch was exposed and this varied in width from 1.5m to 2.5m and was 0.45—0.80m in depth with an irregular ‘V’ shaped profile (Figure 5) and base. The sequence of fills encountered in the northern ditch was broadly similar in each of the five excavated sections. The primary fill comprised abundant small to medium (0.01—-0.10m in diameter) chalk rubble fragments within a very dark grey silty clay, 0.10-— 0.40m in depth. The texture, colour and composition of this fill varied only very slightly along the length of the ditch. Only undiagnostic flint waste flakes were recovered. The secondary fill, in all sections, was similar to that in the southern ditch but with more subangular flint inclusions. It was 0.01—0.12m in depth. The only artefacts were undiagnostic flint waste flakes, five flake cores, a retouched flake and a single piece of copper alloy casting waste, from ditch section 4000 (context 4012, ON 7001). This latter piece probably represents seepage from a two-piece mould although it is typologically undatable. Overlying the secondary fill in all slots, as in the > southern ditch, was an irregular deposit of large, subangular flint nodules within a dark yellowish brown silty clay matrix. Large quantities of worked flint were recovered from slot 6000 (context 6005), mostly large core preparation flakes with a few flake cores; lesser quantities were recovered in other slots. In less severely truncated areas the flint deposit was overlain by a dark yellowish brown silty clay tertiary fill, with occasional lenses of charcoal. A small assemblage of Romano-British pottery was recovered from this deposit which had been disturbed by rabbit burrows and military activity. Overlying this deposit was remnant topsoil and chalk rubble deposited during bulldozing. In one excavated section, slot 5000, a possible recut was noted in the primary fill, suggesting that the ditch had been maintained, although for how long is uncertain. Bank material The removal of topsoil by bulldozer destroyed all traces of the central bank. Only in slot 6000 was a possible trace of bank material noted in the form ofa 0.04m thick layer of dark yellowish brown silty clay with moderate chalk inclusions (context 6011). No finds were recovered and the environmental sample taken from this deposit showed that it contained introduced snail species and was therefore not a sealed deposit. Two tree hollows (2033 and 4020) between the ditches, presumably once sealed below the bank, were excavated in slots 2000 and 4000. Although of natural origin, samples were taken for mollusca and charcoal in order to examine the pre-monument environment. Dating evidence Little artefactual dating evidence was recovered from the lower fills of either ditch. The date of 2355-1715 BC (1514c, 4030+320 BC before 1996), obtained by OSL dating, from the primary fills of the northern ditch suggests that it was constructed in the Early Bronze Age. Even the most recent date in this range is considerably earlier than the Middle Bronze Age date generally assumed for the establishment of the Linear Ditches system (Bradley et al. 1994), and must be viewed with caution. It is perhaps significant that no Early Bronze Age material was recovered from slot 2000 despite its location only c.10m from the concentration of Early Bronze Age ee a EXCAVATIONS ON THE OLD LINEAR EARTHWORK, BREACH HILL, TILSHEAD 87 pottery and flint. Overall, a Middle to Late Bronze Age date is suggested for the construction of the main monument based on the single sherd of Late Bronze Age pottery recovered from the flint deposit in slot 6000 and by a date of 1225-745 BC (1514b, 2980+ 240 before 1996) obtained by OSL dating from the secondary fill of slot 4000 (context 4017). Other features Pottery of Late Bronze Age or Early Iron Age date was recovered from post-hole 1001 and feature 1018. Isolated post-hole 1001 at the western end of the site, did not appear to be structural and no other datable features were encountered in the immediate vicinity, although several undated pits, hearths and post/stake-holes in the same general area may be related. Two sherds of Late Bronze Age pottery were recovered during hand cleaning in this area. Two very small, abraded sherds of Late Bronze Age pottery, small quantities of burnt flint and a small assemblage of worked flint were recovered from a probably natural feature, 1018, an irregular, - sub-rectangular feature, 2.90m long 1.0m wide and up to 0.45m deep with steep, irregular sides and base. These finds are assumed to be residual. Phase 4: Romano-British Romano-British activity is represented by pottery from the upper ditch fills and from a well sorted, gravelly subsoil at the western end of the site (Figure 2). It seems probable that these finds represent agricultural activity in the general area, possibly manuring with domestic waste. All the Romano- British pottery recovered was heavily abraded and mixed with earlier or later material. Undated features Sixteen excavated features contained no datable material: eight pits (1003, 1007, 1009, 1021, 1300, 1519, 1525 and 1532), four post/stake-holes (1511, 1513, 1517 and 1539), three probable hearths (1500, 1503 and 1504) and a single narrow, rectangular, slot (1327). Of these, the four small circular pits (1003, 1007, 1009 and 1300), large, irregular pit (1021) and feature 1327 clustered to the south-west of slot 4000 (Figure 2). The remainder lay towards the western end of the excavation area. FINDS Flint by Phil Harding The total flint assemblage (Table 1) shows that unstratified or disturbed contexts account for almost 50% of the collection. Only the assemblages from Mesolithic feature 1537, Early Bronze Age subsoil (feature 1064 etc.), and from the tertiary fill of both ditches were large enough to be subjected to more detailed examination. Mesolithic Feature 1537 This feature contained 261 pieces of worked flint of which 32% are blade/lets. Additional features of blade technology included four blade/let cores, one of which refitted to a broken blade (Figure 6, 1), two crested pieces and three rejuvenation tablets. There was also a broken backed bladelet (Figure 6, 2) and a proximal microburin (Figure 6, 3). Analysis of the unbroken material showed that the blades ranged from 88mm to 14mm (mean 37mm) long and 32-6mm (mean 14mm) broad. They included parallel and slightly convex edged pieces from opposed and single platform cores which were flaked using soft hammers. Platform abrasion prepared the core for the removal of blades with narrow butts, 57% measured up to 1mm wide. This was substantiated by a high frequency of linear, crushed and punctiform butts which accounted for 36% of the total. The cores averaged 54mm long and included three examples with opposed platforms. They were all made from weathered nodules with incipient thermal fractures except for one which had been introduced from a gravel source. Most of the cores were rejected when the flaking angle of the striking platform had increased to a point beyond which blades could be removed successfully. One core showed signs of failed preparation and was probably unproductive. Thirty chips were recovered. This microdebitage is undoubtedly under-represented as the feature was not included in the sampling strategy for the excavation. A burnt crested piece found in an artefact sample taken from the adjacent sorted horizon of the natural soil profile is the only probable artefact of blade technology found beyond the pit. Early Bronze Age subsoil Flint from this area provided the largest group of material from the site. Unbroken flakes form 88 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 1: Flint totals by feature type (context numbers in brackets) Category Spoil 10m sq. Features slot 2000 slot 3000 slot 3000 slot 4000 — slot 5000 slot 5000 slot 6000 TOTAL (1-110) (500-690) (1000- (2000- (S) (N) (4000- (S) (N) (6000- 1999) 36) (3000-74) (3500-47) 23) (5000-406) (5450-550) 13) 1 Blade cores 1 1 3 - - - - - - - 5 2 Bladelet cores - - 1 - - - - - - 1 3 Flake cores 3 1 22 - 4 2 4 - 7 9 52 4 Broken cores 1 - 10 1 2 1 - 1 16 5 Blades 2 10 59 2 6 6 8 I) 2 13 110 6 Broken blades 4 3 43 3 3 2 4 1 8 71 7 Bladelets - 1 18 - 1 - 1 21 8 Broken bladelets 1 3 33 1 1 - - 1 - 2 42 9 Flakes 92 69 430 45 117 74 145 33 32 218 1255 10 Broken flakes 29 30 421 24 77 43 61 14 10 107 816 11 Crested pieces - . 2 - - - - - - - 2 12 Rejuvenation - - 4 4 13 Microliths - - 1 1 14 Microburins - . 1 - 1 15 Chips - 1 88 - 4 1 6 100 16 Scrapers 2 3 28 - 11 3 1 - 48 17 Other tools 2 3 8 1 2 1 - 1 3 21 137 125 1172 77 228 132 224 52 By 367 2566 No. Contexts 35) 43 73 9 31 18 12 18 9 7. an unusually high proportion of the total and outnumber broken pieces at a ratio of 2:1. There is also a marked absence of miscellaneous, broken knapping waste. The cores comprise poorly prepared flake cores made from small nodules of weathered flint. They included undiagnostic single platform, opposed platform and multi-platform cores although a larger opposed platform flake core with a flat, thermally fractured back is typical of many Late Neolithic prepared cores. The analysis of 109 flakes from adjacent contexts 1060, 1068 and 1079 represented 58% of all flakes from test-pits in the subsoil. They were also associated with the greatest density of scrapers. The results showed that the flakes were predominantly 20-40mm long and were elongated or squat in plan proportions, which accords well with a domestic Beaker-industry from Dean Bottom on the Marlborough Downs (Harding 1992). Butts were occasionally abraded and generally plain. The retouched tools, which formed 5% of the assemblage, included 24 end scrapers made on flakes, one made on a blade and two side scrapers made on flakes (Figure 6, 4-11). Four were broken. There were also two knives, a broken tip of a barbed and tanged arrowhead, a hammerstone and two miscellaneous retouched flakes. The scrapers were well made, averaging 34mm long, 32mm broad and 9mm thick — dimensions which are identical with those from Dean Bottom (Harding 1992). Scraper blades averaged 39mm in length (35mm at Dean Bottom) and were formed by regular, continuous, direct semi-abrupt retouch which often extended beyond the distal end onto the edge. The results of this analysis confirmed that, although the industry does not come from a sealed context, it compares well with domestic Beaker flint industries and appears to contain no residual or intrusive material. The ditches Small collections of undiagnostic waste flakes were found in the primary and secondary ditch fills with a well made end scraper on a cortical flake, a retouched flake and a flint hammerstone. An accumulation of large, cortical core preparation flakes, which outnumber cores 40:1, was found in the flint deposit overlying the secondary fills of both ditches in section 6000. They probably represent knapping waste which was dumped into the silted ditch hollow. The two small groups share a broadly similar technology and are similar stratigraphically although some heavily patinated pieces may be residual. A sample of unbroken flakes from contexts 6005 and 6007 was analysed, — representing 75% of the available flakes. The results confirmed that the flakes represent by-products of flake core preparation, with blades and broken blades accounting for only 6% of the combined total. This proportion is well within what may be produced accidentally. The flakes have broad plain butts and were probably removed by powerful blows of the hammer. No attempt was made to prepare the core edge beyond trimming to tidy the edge. The sample from the flint deposit in the northern ditch includes EXCAVATIONS ON THE OLD LINEAR EARTHWORK, BREACH HILL, TILSHEAD 89 Fig. 6 Flint tools from feature 1537 (nos 1-3) and Early Bronze Age subsoil 4-11). 1) Blade core and refitting broken blade (context 1509/1510); 2) broken backed bladelet (context 1508); 3) proximal microburin (context 1509); 4-11) scrapers (contexts 1060, 1068, 1075) proportionally larger flakes, more cortex, lower flaking angles and more irregular edges than the sample from the flint deposit in the southern ditch which suggests that the former comprises purely core preparation flakes and the latter a proportion of by-products from blank production. This activity probably represents a relatively small scale knapping episode which utilised locally available surface flint that may have accumulated during field clearance. The tools comprise an end scraper and two flakes with edge retouch. Prehistoric pottery by Emma Loader and Lorraine Mepham The small assemblage of pottery comprises 339 sherds (1324g). Out of this total, 81 sherds of Late Iron Age/Romano-British pottery, one sherd of medieval pottery and one sherd of modern pottery are not considered further here. The prehistoric pottery has been analysed in accordance with the standard Wessex Archaeology recording system for pottery, which follows nationally recommended guidelines (Morris 1994; PCRG 1997). Three broad fabric groups were identified on the basis of dominant inclusion types: flint-tempered fabrics (Group F), grog-tempered fabrics (Group G) and organic-tempered fabrics (Group V). These groups were then subdivided into five separate fabric types based on the range and coarseness of the inclusions present. The pottery has been fully quantified and full details are in the archive. Fabric totals are given in Table 2. Beaker pottery A total of 228 sherds (674g) of Beaker pottery was recovered, all from the subsoil area associated with Table 2: Fabric totals by chronological period Fabric No. sherds Weight (g) BEAKER Gl 26 65 G2 202 609 sub-total 228 674 LATER PREHISTORIC F] 5 10 F2 9 $3 Vi \4 18 sub-total 28 81 LIA/R-B 81 551 Post-Roman 2 Is TOTAL 339 90 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 7 Beaker pottery from Early Bronze Age subsoil. 1) Rim, G2; finger-pinched decoration. PRN (Pottery Record Number) 14 (context 1044); 2) G2; finger-pinched decoration. PRN 15 (context 1044); 3) G1; ?fingertip impressions. PRN 20 (context 1053); 4) Rim, G1; square-toothed comb impressions. PRN 12 (context 1035); 5) G1; square-toothed comb impressions. PRN 33 (context 1070) feature 1064. The group includes comb-decorated ‘fineware’ Beakers as well as coarser, finger-pinched ‘rusticated’ examples. The sherds have suffered severe abrasion (mean sherd weight 3.0g), and this has made any estimate of the minimum number of vessels virtually impossible, although a combination of fabric, rim form, colour variations and decorative schemes suggest at least five (Figure 7). Two fabrics are represented. Both are grog- tempered; the most common inclusion in Beaker pottery in this area (Cleal 1995a, 190). Fabric Gl has a high content of quartz inclusions in a fine clay matrix, whereas G2 has few other inclusions, although rare glauconite was noted in several sherds. Glauconite was also noted in Beaker sherds from Dean Bottom on the Marlborough Downs, where it was considered to possibly derive from the Greensand, which outcrops to the north and west of Salisbury Plain (Cleal 1992, 62). Gl Fine; moderate, well sorted, subrounded quartz <0.25mm, iron oxides <0.25mm. Oxidised exterior and interior surface — margins and cores black. Soft, soapy texture. G2 Abundant grog, very rare angular flint <0.5mm, sparse subrounded quartz <0.25mm. Soft, soapy texture. Possible glauconite noted in several sherds. Oxidised. Sherds in fabric G1 tend to be thinner walled, and just over half are decorated (61.5% by number of sherds). The most common decorative motif comprises square-toothed comb impressions (8 examples; Figure 7, 4, 5), although finger-pinching (one example), ?fingertip impressions (one example; Figure 7, 3) and other indeterminate impressions, possibly finger-pinching (four examples) were also ° observed. Two rims are present (eg., Figure 7, 4). These sherds are representative of finer vessels, but overall forms are impossible to reconstruct given the fragmentation of the assemblage. Sherds in fabric G2 are from thicker-walled, coarser vessels. Again, most are decorated (84.7%); the commonest identifiable technique is finger- pinching (65 examples; Figure 7, 1, 2), although a few have square-toothed comb impressions (6 examples). A further 96 decorated sherds carry indeterminate impressions, possibly finger-pinching. Two rims in this fabric probably derive from the same large, thick- walled vessel (eg, Figure 7, 1). Again, reconstruction of overall form proved impossible. This small group, then, contains the very abraded remains of at least five Beaker vessels, three in the finer fabric G1 (two comb-decorated and one with Pfingertip impressions) and two rusticated (finger-pinched) vessels in the coarser fabric G2. In the absence of reconstructable profiles these vessels cannot be assigned to specific Beaker groups (Clarke 1970), steps (Lanting and van der Waals 1972) or styles (Case 1977), although the decorative schemes seen here are more likely to belong to Case’s Middle or Late Style Beakers. This seems to be a purely domestic assemblage. General parallels for this group can be found amongst the Beaker material from Stonehenge and its environs (Cleal 1995b; Cleal and Raymond 1990). At Stonehenge itself a similarly high proportion of rusticated sherds was noted, although these were considered to derive from just two or three large vessels (Cleal 1995b, 356), as may be the case here. The close parallels between the flintwork from the same subsoil layer and a group from Dean Bottom on the Marlborough Downs have already been discussed above, and the domestic Beaker assemblage from that site is also comparable, with a mixture of comb- decorated and rusticated (fingernail-impressed) vessels (Cleal 1992, figs 45-8). Later prehistoric pottery A small number of sherds (28; 81g) are of later prehistoric date. Again, condition is poor, and sherds are heavily abraded (mean sherd size 2.9g). Three fabric types were identified: Fl Rare, poorly sorted coarse calcined flint T T € 2 9 € I Z - - = = + 5 5 e I = ¢ s I s : : = € iE - - G I - - - - te - = 5B is I I - - - I - 2 . S € LE E 5 = I 2 = F I ; I a 8 tI I > S fb 6 I S T 9 LI € €I z I I s a I I 2 = I I I I 5 - I I - - = = I G (a z 2 I I Ie x a Z Z C OO0T | OOOT OOOT | OOOT ODOT SI 0€ (a €S £9 SSO) 02 =s0€:|/22é7— “seS 6/ +00r | OL0b/Z10r SIOr S708 | #708 €Z08} 7Z08 1708 ‘Waq | Arepuoses Areurtid Z00h YOU WIYLION 69 09 «SIT OZ st 7e Cir OT Or Il or €1 IT (al ZI (al = I (¢ I I S 5 5 + = = . = = . + 3 = _ z I = I = Il jé Il 8€ (4 ee (0) Ge.) ev tI (al IZ 87 IZ 9€ 6£ tI I € = I z 3 I = > = I = c IT € € iL tZ 9E LE € = 5 = = 2 I = I = ss = b I v Il = - - - T - - I - I ( I IT 9 9 L T I I : I £ I S I tI 6 IT 0¢ bE LE 89 Gl L C 8 9 SI T€ 0s 87 S IT 9€ Si LY 89 IZ cI A € 6 Or SI v 8 2 (a S (al 61 (43 ST (XE ( = = T € + iS 4 I - I 5 I iS C S = 5 + + S - O000T | OOOT OOOT ODOT | OOOT IT 6l LT SE (a4 cS c9 -0 SG) Olt a lCl|p SSS. sche SsGS II/ S00F T10¢ el0r F10r tr08 €b08 7rO8 IP08} 0F08 6£E08 8£08 | 9E08 Arena Arepuosas LO0r YU WIayINOS aay ozuolg 21e] — plw col 8 I 48 L be COT L 9 as = LI I v I I 2 Z : G c 9 I Il S a I I ar OOST OOST c6 “78 LOT a) 810r vEO8 E€08 Arewid cE08 8 = ky Sol O16 7, € Bite Ore ES oar Sm hae aay Stam ee ae ace et os So aang wae eee Gee ay est, Disa cops ee - - - - I (é - el i ee Oe ae eee G.Uk ae We aa ae ee WN i era oe ee en ee ae ee ee ae We big ce = ae ee ee ee Re EP bs oe ee - - - I - vé - Se ee ae Se ge eee Ee a ee ee ee I - - - - - - ae ee ee ee Ee ee cup Gra Cle eae Q00T O00T O00T]O0OT O00T O00T OS 0E;| 0G 06 = fe cS sSI- = =0 S| (Eel “01 302 «Zt s/ PEO? PEO S07 CCOV €70b 610P Lp08 OSO8 6F08| 008 6208 8208 L708 €€07 MOTTOYIAN polepun ZZOP MOTOY2AN TVLOL BxeL ‘dds njupup/vanday ‘dds vanday (JOIN) stsuaqsoy vanday (snoeuUuly) vpi2dn] puosiyazy (snaeuury) vpidsiy viysuy (Snaeuul]) von vpjaIaFT (JaOg) D7IaSLaIUL JO DJNpIpuDD (wIQNg) vIPIUapig DYISNDID (nse -UOW) DIDUIUD] DUIPO]YI0D (49][NY) 0jNI19n saproyiIzay aeploeulry (J9T[NYY) sntuvjJa9 snpryrkxC, ae -Iedeiq) vjnpinu vyjaudosap (1aply) vind vjaudosap (WIONS) stowuDy DaineosaNy (PUNIAISIAA\) 7190.9U0) Salil (J9][NYP) 0170189 vas14A (291M) vprongjed vumir4 (49][NYP) snwpunjos snosiqy (pneu -Iedeiq) wnanwsdd wnioung (aa][NP) vinosqgo vu (JIT[NW) vipapnsv vjnuryunrp ‘dds piuoyjny TYIAS vIUujUaIxa pIU0]]DA (19][NY) VIVIs09 IUCN (snaeuUry) wnioosnu vjpidng “dds 0811124 (pneuiedeiq) eaewisAd 0522194 Jan vpjisnd os11a4 ‘dds ndooy09 (49TTNW) VIugn vdor17Yy909) ‘dds wmyotivy (OSSTYy) wnwiuepiy wniyreing (Ja][NP) sunsaja smwwog DAN (wo) yidaq 1x91U09 adures afdures ainie2J ase ZOO PUL [OOF SPYSUP pur “ZO “EEOT MOTTOYIaN Woy ep OsN]TOUI pur] :¢ a1qRL. EXCAVATIONS ON THE OLD LINEAR EARTHWORK, BREACH HILL, TILSHEAD 93 PP eo or es Depth in cm Ds — ab om + sh + + + + ogg te Ay ae SS ee tt es + + xe r r re re Fig.8 Snail diagram competition diversity indices (A4) (Hurlbert 1971). The Shannon (#”’) and Brillouin (HB) indices were also calculated (Evans 1983), and in addition, small air-dried subsamples (<2mm) were taken from each sample for magnetic susceptibility to assist in the characterisation of the deposits. Tree hollows Two tree hollows sealed beneath the bank and old land surface were sampled to provide some indication of the pre-ditch environments. Tree hollow 4020 (Figure 5), which contained pine charcoal, produced few shells throughout, although where shells were present in the upper fills they were dominated by shade-loving species (Table 3) and included rupestral taxa (Acanthinula aculeata and Clausilia bidentata), A second tree hollow (2033) contained only just sufficient shells for analysis. This feature was crescentic in plan and contained two main fills; a chalky rubble (2035 and 2036) and a silty clay with chalk pieces (2034) typical of many tree throw hollows (see Allen 1995b, fig 69; Macphail 1987; Macphail and Goldberg 1990). Samples were taken from the main fills, but unfortunately not from the primary chalky weathering fill (2036). Shell numbers are low (Table 3), but are predominantly shade-loving species, including the rupestral species Acanthinula aculeata. The only occurrence of the rare Vertignid, Vertigo pusilla, from the site came from this feature, albeit as a single specimen; it is a relatively rare species typical of woodland environment, especially moss and dead leaves. The occurrence of this species is severely restricted by the onset of (Neolithic) forest clearance (Evans 1972, 142) and thus is more common in the earlier part of the post-glacial, for example the ancient woodland assemblages from Balksbury (Allen 1995b) and a subsoil hollow at Avebury (Evans 1972, 272). These assemblages may, therefore, indicate the presence of shadier shrubbier conditions, if not woodland, before the construction of the bank. 94 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Old land surface The buried soil (6011) formerly preserved beneath the bank existed as a thin veneer of dark soil, surviving predominantly in shallow hollows or as isolated patches. It was only 40mm thick and was a dark yellowish brown (10YR 4/6), silty clay with few very small to small subrounded chalk inclusions over chalk and moderate magnetic susceptibility (39). It was, however, heavily disturbed by modern interference (machine stripping and bulldozing) as well as limited evidence of root and worm activity. One sample was assessed from this remnant buried soil but contained very few shells, and although the assemblage was dominated by open country species (Helicella itala and Vallonia spp.), the presence of Introduced Helicellids (a medieval or later introduction to Britain) indicate that mixing and intrusion must have occurred. No further analysis was undertaken; nevertheless, the lack of any shade-loving species, in contrast to the tree hollows, tends to indicate the presence of established open country conditions prior to burial. The ditches Columns of contiguous samples were taken through the deposits of both ditches; the primary fills were uniform and comprised loose medium to large chalk rubble ina highly calcareous (light yellowish brown 10YR 6/4) silt clay matrix, secondary fills were silty clays to very silty calcareous deposits with common to abundant chalk pieces. Occasionally silty, almost stonefree, deposits occurred in the southern ditch. No evidence for any stabilisation or buried soils at the top of the secondary fills could be seen in either ditch, but in both ditches a central inverted cone of flint gravel was present, presumably having eroded and rolled into the middle of the ditch when the sides were largely stable. The highly silty nature of the secondary fills was questioned during excavation and a spot sample from one (context 3067) was taken to examine the possibility of it being waterlain. The sample contained only three shells in the flot, all of which were land snails. Although the absence of fresh/ brackish water snails, or even marsh snails, does not preclude the possibility of this deposit having been waterlain, a disturbed bulk sample was examined in the laboratory, and its well sorted fine-medium texture and buff colour (brownish yellow 1OYR 6/6) is typical of aeolian, reworked, loessic deposits such as those seen in the Y Holes at Stonehenge (Cornwall 1953). Northern Ditch (section 4002): The shallow (c. 0.75m) ditch produced shell-poor assemblages with no ~ sample containing more than 60 molluscs per kilogram. Shell numbers generally decreased up the profile in contrast to normal expectations and virtually all the shells excepting Limacidae plates were recovered from the flots. Although these low numbers hinder detailed interpretation, the trends are particularly significant in view of the results from the southern ditch. The primary fills (context 4015) contained relatively high proportions of shade-loving species, particularly Discus rotundatus and Carychium tridentatum with Zonitidae. These species decrease and are virtually absent in the secondary fills. The assemblages tend to suggest that the ditch was cut into, or initially existed (30 years or so) within an environment with at least a modicum of shade, perhaps provided by scrub (Discus rotundatus) and long grassland (Carychium tridentatum), set within a generally open dry calcareous environment (Pupilla muscorum and Vallonia excentrica). Although some of this shade could have been afforded by plants growing within the ditch, this seems unlikely in the coarse chalk rubble primary fills; further the assemblages do not represent typical rock-rubble habitats (cf. Evans and Jones 1973). The secondary fills (context 4012) show a change to assemblages largely dominated by open country species, ultimately to the total exclusion of any shade-loving species. They are characterised by the prevalence of Vallonia excentrica over Vallonia costata with Pupilla muscorum and indicate a very dry calcareous downland. Low shell numbers prohibit the distinction between grazed grassland and arable contexts, though the former is favoured, indicating a change in the local ditch, or adjacent, micro- environments. The Romano-British tertiary fill contained too few shells to make any interpretation except to say that all three species recovered are open country suggesting no major change (i.e., vegetation regeneration) during its deposition. Southern Ditch (section 4001): The main ditch was V-shaped and about 1.45m deep with the bank to the north separating it from the northern ditch, presumably also separated by a berm and possibly a revetment too as no bank spill was recorded in the ditch sequences. Shell numbers were much higher through these ditch deposits. The primary fills (context 4018), in contrast with the northern ditch, are dominated by open country species (c. 90%) only declining in relative terms with the increase EXCAVATIONS ON THE OLD LINEAR EARTHWORK, BREACH HILL, TILSHEAD 95 Table 4: Land mollusc summary data, species diversity and magnetic susceptibility results sample context depth (cm) taxa total A2 Ditch 4001 8044 4005/11 0-11 10 69 0.84 8043 4011 11-19 11 60 0.86 8042 4011 19-27 10 115 0.81 8041 4013 27-35 13 210 0.86 8040 4014 35-42 11 251 0.86 8039 4014 42-52 12 B52 0.84 8038 4014 52-62 12 412 0.86 8037 4016 spot 8 102 0.72 8036 4016 62-72 12 120 0.78 8035 4017 72-82 7 87 0.62 8034 4018 82-92 7 44 0.74 8033 4018 92-107 6 11 0.73 8032 4018 107-122 I 52 0.57 Ditch 4002 8025 4009/4 5-15 5 6 0.78 8024 4012 20-30 9 39 0.85 8023 4012 30-42 it 59 0.81 8022 4015 42-53 16 26 0.92 8021 4015 53-63 11 49 0.85 treehole 2033 8048 2034 0-15 8 28 0.79 8047 2034 15-30 6 13 0.77 4050 2035 0-15 4 4 0.75 8049 2035 15-30 5 16 0.79 treehole 4020 8030 4022 1-10 9 16 2.05 of the catholic species Trichia hispida, indicating that this ditch was cut into a pre-existing open dry downland. The secondary fills show a persistence of open country conditions, being dominated by the catholic species Trchia hispida accompanied by the Vallonia species, Pupilla muscorum and Helicella itala. Only when the ditch sides become relatively stable, and flint gravel had rolled into the middle of the ditch (context 4016), is there any evidence of any shade-loving species. A minor increase in the Zonitidae population and the record of Acanthinula aculeata give a presence of shade-loving species at 2.5% and probably indicates the slightly shadier micro-habitats created by the flints and any associated local grassy vegetation in the ditch. In the layers above this (context 4014) which contain no flints, the shade-loving species are diminished and wholly open country habitats prevail once more. Stable conditions both in the ditch and the local environment are indicated by the high shell numbers and significant increase in magnetic suscepubility; this is not reflected in the A2 and A 4 indices (Table 4), but the Shannon and Brillouin indices do reflect this with a rise in diversity indices and results in excess of 2.00. These assemblages probably represent open dry short-grazed grassland and arable conditions. Ad ee pee MS 5.92 1.99 1.79 80 6.94 Ae 1.87 82 4.50 1.87 1:73 108 6.13 2.10 2.00 72 6.28 2.08 2.00 35 5.19 2.03 1.96 31 6.27 2.10 2.04 23 De gfp) 1.55 1.44 23 3.65 1.78 1.65 22 172 E27, 1.18 12 3.19 1.55 1.36 6 4.00 1.54 1.09 8 1.38 e277. 1.10 7 14.0 1.56 0.98 4] 7.05 2.08 1.73 28 4.82 1.95 1.71 25 222 2.64 2.03 24 6.30 2.07 1.79 23 4.64 1.79 1.47 35 5.00 1.61 1.19 28 - 1.39 0.80 6 3.29 1.39 1.10 3 SYA 0.85 9.91 55 Open conditions prevail throughout the tertiary deposits (contexts 4013, 4011 and 4009), with the highest Shannon diversity index on context 4013. An increase in Helicella itala to the detriment of Pupilla muscorum, suggests drier conditions locally, possibly tillage. Charcoal by Rowena Gale Charcoal was sparse or absent in most of the excavated contexts at Breach Hill. Small deposits of charcoal were present in Beaker, Early Bronze Age and Mid-Late Bronze Age features and some undated tree hollows. Environmental data were sought from the charcoal and compared to results from other sites in the area. Bulk soil samples were processed by flotation with flots retained on 0.5mm mesh residues on Imm. The flots and residues were scanned for charcoal, which was sparse. Samples were prepared for examination using standard methods. Fragments from each sample were fractured to expose fresh transverse surfaces and sorted into groups based on the anatomical features observed using a x20 hand lens. Representative fragments from each 96 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE sample were selected for detailed study at high magnification. These were fractured to expose the tangential and radial planes, supported 1n washed sand, and examined using a Nikon Labophot microscope at magnifications of up to x400. The anatomical structure was matched to prepared reference slides. When possible the maturity (i.e. heartwood/ sapwood) of the wood was assessed. The results are shown in Table 5. The anatomical structure of the charcoal was consistent with the taxa or groups of taxa given below. It should be noted that the anatomical structure of some related taxa can not be distinguished with any certainty; for example, members of the Pomoideae (Crataegus, Malus, Pyrus and Sorbus) and Salicaceae (Salix and Populus). Classification follows that of Flora Europaea (Tutin, Heywood et al. 1964-80). Broadleaf species: Corylaceae. Corylus sp., hazel Fagaceae. Quercus sp., oak Rosaceae. Subfamily Pomoideae which includes Crataegus spp., hawthorn; Malus sp., apple; Pyrus sp., pear; Sorbus spp., rowan, service tree and whitebeam. These taxa are anatomically indistinguishable. Salicaceae. Salix sp., willow and Populus sp., poplar. These taxa are anatomically similar. Conifers: Pinaceae. Pine (Pinus) Table 5: Charcoal Pre-monument tree hollows Charcoal was examined from the two pre-monument (undated) tree hollows. Feature 2033, context 2034, included hazel (Corylus). Feature 4020, context 4022, included pine (Pinus) and short pieces of narrow diameter (2mm), probably root, with distorted structure (?vitrified), possibly indicating exposure to temperatures exceeding 800°C. The material was tentatively identified as willow/ poplar (Salix/ Populus). The presence of roots 1n a tree hollow would not be unusual but the presence of burnt roots in a context where in situ burning was not recorded is more difficult to interpret. Although it is feasible that the stump of the tree was destroyed by burning, this event would have left traces in the surrounding soil. Alternatively, if these roots derived from the fallen tree, roots from the exposed stump may have been used nearby for fuel. Beaker Three samples from the fill of feature 1064 were examined. Hazel was common to all three samples whereas oak (Quercus) occurred in contexts 1068 and 1069, with heartwood in the former. A sample from context 1068 also included a member of the hawthorn group (Pomoideae) and pine; a sample from context 1070 also included willow/ poplar. Middle to Late Bronze Age Despite the general paucity of charcoal in ditches 4001 (southern) and 4002 (northern), small quantities were extracted. From 4001, the basal fill included a member of the hawthorn group (context 4018), while Feature Context Sample Corylus Pomoideae Quercus Salicaceae Pinus Beaker ?hollow 1064 1068 8055 3 lh - 1 1069 8056 1 2 2 = 1070 8057 iL - 1 = Mid — Late Bronze Age Main linear ditch (southern) 4001 4017 8081 1 - - - 4018 8082 - - - - V-shaped linear ditch (northern) 4002 4012 8072 2 - - 1 4015 8073 - - - 1 Late Bronze Age/Early Iron Age posthole 1001 1000 8000 9 - - - Undated Tree hollows 4020 4022 8074 - - ?3root 3 2033 2034 8084 2 - - - Key: h = heartwood The number of fragments identified is indicated EXCAVATIONS ON THE OLD LINEAR EARTHWORK, BREACH HILL, TILSHEAD 97 a lens of flint in the secondary fill included hazel (context 4017). Charcoal from the secondary fill (context 4012) of the northern ditch 4002 consisted of hazel and pine (Pinus); the basal fill (context 4015) also included pine. Late Bronze Age/Early Iron Age The sample from possible post-hole 1001 included large (up to 10mm in length), well preserved fragments of hazel. Discussion Small quantities of charcoal were found in features from Beaker to Late Bronze Age/Early Iron Age periods, from a post-hole and the two ditches. Charcoal deposits were also present in a natural feature 1064 and two probable tree hollows 4020 and 2033. The use of the site from the Mesolithic period was indicated by flint artefacts and appears to have continued through the succeeding millennia, although in what form is unclear. The origin of the charcoal is unknown but, by association, was linked with the occupation of the site. Despite the paucity of the charcoal the following taxa were identified: hazel, oak, willow/ poplar, member/s of the hawthorn group and pine. These compare with hazel, oak, ash and elder recorded from the double linear ditches at Sidbury Hill, sited on the Plain to the north-east of Breach Hill (Gale 1994). The presence of pine in the Beaker and Middle to Late Bronze Age periods is of considerable interest. The natural distribution of pine in the post-glacial period was much wider than it is today and included lowland England. As the climate warmed and broadleaf woodland began to dominate southern Britain, pine woods retreated north and in the present day wildwood is confined to Scotland. It is generally thought that pine was absent in southern Britain from early prehistoric times (Rackham 1986). Pine was recorded in Mesolithic features at Stonehenge (Gale in Allen 1995a) and from dated Mesolithic features at Strawberry Hill, Salisbury Plain (Allen 1992) and the identification of pine charcoal from Breach Hill confirms its continued existence in the area until at least the Middle to Late Bronze Age. DISCUSSION Mesolithic by Phil Harding The flint from Breach Hill, although often poorly provenanced and in small quantities, has provided valuable information of activity at the site, much of it pre-dating the construction of the linear ditches. The assemblage from feature 1537 is particularly significant. It is indistinguishable from Early Neolithic blade material but the presence of diagnostic Mesolithic artefacts suggests that the contents are earlier. The broken backed bladelet, which may be more suggestive of a Late Mesolithic date, could be residual or an isolated hunting loss; however microburins are by-products of on-site tool manufacture and are unlikely to have been taken away from the knapping site. Refitting material also confirms that flaking was taking place at the site. Evidence of Mesolithic activity is poorly represented on the chalk throughout southern England but, where it occurs, it is often restricted to areas of tertiary gravels or Clay-with-flints (Gardiner 1988) as at Breach Hill. Large parts of Salisbury Plain, now existing as unploughed downland, provide few opportunities for discovering evidence of Mesolithic activity. A possible core and butt of a broken tranchet axe were found on the surface of Strawberry Hill, West Lavington on the north scarp of Salisbury Plain (Wiltshire County Council SMR ST9OSSE 052) associated with Neolithic and Early Bronze Age artefacts. A linear ditch approximately 100m to the south, sealed by colluvium in the base of a coombe, produced radiocarbon dates of 9350+ 120 BP and 6820+120 BP (OxA-3040 and 3041) (Allen 1992), although no artefacts were found. Given the context and associations of the artefacts it is possible that they have been misidentified. The core is described as ‘possible’ and the tranchet axe, which lacks its diagnostic tranchet end, may be a broken, crudely flaked, Neolithic axe. Evidence of occupation is also meagre on the southern part of Salisbury Plain. Mesolithic artefacts, predominantly large core tools, are more common (Wymer 1977) towards the river valleys. Radiocarbon dates of 8880+120 BP (GU-5109), 9130+180 BP and 8090+ 140 BP C(HAR-455 and 456) (Allen and Bayliss 1995) were obtained from three post-pits in the Stonehenge car park although no artefacts were present. Richards (1990, 263) found no diagnostic blade/let debitage during extensive 98 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE fieldwalking from the Stonehenge Environs and only a single microlith from Fargo Road, north of the monument. He concluded that extensive Mesolithic activity was unlikely to have taken place on Salisbury Plain but noted the discovery of Mesolithic material including blades and microliths in the River Avon valley below Durrington Walls. He stressed that the strong relationship with the river continued into the Early Neolithic period. Breach Hill lies in the heart of Salisbury Plain, but is within easy reach of local watercourses. It is located on the watershed between the headwaters of the River Till, 3km to the east, and the Chitterne Brook 1km to the west, which flow south to the River Wylye. These watercourses are all likely to have flowed throughout the year during the Mesolithic and to have attracted game. They also contained deposits of flint valley gravel (BGS sheet 282) suitable for tool manufacture. Wymer (1977) listed 12 Mesolithic findspots from seven parishes in the Wylye valley and an axe from Imber at the headwaters of the Chitterne Brook. Most were single axes, picks or pebble maceheads; however a nucleated scatter including tranchet axes, blades, cores, microliths and a scraper was found at Stockton, downstream from the confluence of the Wylye with the Chitterne Brook. Blades of ‘Mesolithic character’, of which two refit, have also been recovered from a pit containing Late Bronze Age pottery at Boreham Farm Bungalow, immediately south of Battlesbury Camp (Wessex Archaeology 1999). Breach Hill therefore lies within an existing area of Mesolithic activity but provides a valuable addition to the distribution of artefacts and man -made features of that period on the central part of Salisbury Plain. It remains a strong probability that Mesolithic occupation was small scale and low density but suggests that occasionally groups of hunters used the river valleys as potential routeways across the plain to the Vale of Pewsey. Early Bronze Age by Vaughan Birbeck The probable number of vessels represented by the pottery assemblage and the wide range of worked flint, including unretouched flakes and cores in addition to tools, recovered from the Early Bronze Age subsoil make it unlikely that this spread of material represents a disturbed funerary deposit. The lack of definable domestic features such as pits, post-holes or hearths, along with the lack of any environmental material usually associated with settlement, apart from charcoal, makes interpretation of this deposit uncertain. The pottery is probably of domestic rather than funerary type and many sites interpreted as Beaker settlements are no more than protected artefact scatters, even where the contemporaneous land surfaces survive (Healy 1998). Apart from funerary monuments, Beaker activity in the area around the site is represented by rare topsoil finds and occasional small, apparently isolated pits, such as that found at Willis’s Field Barn, approximately 7km to the west (Wessex Archaeology 1999). Excavations at Downton, Wiltshire (Rahtz 1962) uncovered a series of irregular hollows, some of which were similar in form to feature 1064, along with a hearth and several post-holes, all associated with Beaker pottery. It is possible that the subsoil seals a similar group of features, most of which lay beyond the northern limit of excavation. A concentration of Beaker pottery and worked flint has been recovered from ‘the upper levels of a periglacial feature’ at Barrow Pleck, Cranborne Chase, Dorset, where they were associated with a barrow cemetery (Barrett et al. 1991, 118). It was suggested that these had been deposited on a land surface and had been protected by slumping within the natural feature. Middle-Late Bronze Age linear ditch earthwork by Vaughan Birbeck and Michael F. Allen The primary aim of the project was to examine the present condition of the monument, its original form, the date of its construction, whether it was maintained, its environmental setting and _ its original function. These questions may now be addressed. Within the area of excavation the monument survives as two parallel ditches; all traces of the bank have been completely destroyed. To the east of the excavation area it survives as a visible ditch and bank earthwork or as a cropmark. Originally the form of the monument within the area of excavation probably comprised two parallel ditches, the larger to the south (upslope), with a bank, presumably flanked by berms, between the two. The complete destruction of the bank within the excavation area means that questions regarding EXCAVATIONS ON THE OLD LINEAR EARTHWORK, BREACH HILL, TILSHEAD 99 its height and form, the presence or absence of berms and its position between the ditches cannot be usefully addressed. The worked flint assemblage recovered from the flint deposits in slot 6000 demonstrates that both ditches were almost certainly partly silted up when the assemblage was deposited. The sequences of fills are very similar in both ditches, suggesting that they probably silted up over the same period. The parallel alignment of the two ditches suggests that either the northern ditch was extant at the time of the monument’s construction, which given the almost complete lack of evidence for maintenance or recutting seems unlikely, that they were constructed at the same time or that the northern ditch was excavated along the base of the extant bank. Without secure dating evidence, the imprecise nature of the OSL dates, and the lack of artefactual data, it was hoped that some indication of the relative order or contemporaniety of the two ditches might be gleaned from the molluscan assemblages. Indeed, there are differences between the two in terms of shell numbers, species representation and overall interpretation which might be taken to suggest that they could have chronological significance. This supposes that both ditches are sampling the same environment as both occur within metres of each other. However, if they were contemporaneous, the configuration of the extant bank separating the two might result in the molluscan faunas representing different locales and thus possibly different environments; the northern ditch sampling the adjacent large terrain to the north, the berm and northern side of the bank, and the southern ditch - sampling the southern bank and berm and terrain to the south. This being the case it is quite possible for two contemporaneous molluscan assemblages to sample different environments and, if the land- use on either side of the bank differs, for this to be reflected in the assemblages. On archaeological grounds, and despite the OSL dating evidence, it seems unlikely that the two ditches were successive in view of their symmetry and the almost identical fill sequences that suggest a single build for the earthwork. It is possible that the smaller northern ditch acted as a marker for the larger southern ditch, although it is unlikely that the northern pre-dated the southern by the 500 or more years suggested by the OSL dates. It is equally possible that the northern ditch post-dates the southern ditch and was excavated along the northern side of the bank. From the land snail evidence, if the shadier habits represented by the northern ditch assemblages were an earlier environment it seems unlikely that the old land surface (despite the problems with the land snail assemblage, discussed below) would be completely devoid of all shade- loving elements especially the more robust taxa such as Clausiliidae. On balance it is felt that it is more likely on archaeological grounds that the ditches are broadly contemporaneous and that the palaeo-environmental evidence does not detract from this argument. A Middle to Late Bronze Age date is therefore suggested for the construction of the monument which had become stabilised by vegetation by 1225-745 BC (1514b, 2980+ 240 before 1996). Only very slight evidence for maintenance or recutting of the ditches was found, despite excavation methods specifically designed to identify and record such evidence. The flint deposit overlying the secondary fill in both ditches may be in a small recut, but no evidence of maintenance was discovered apart from possible traces of a recut noted in the primary fills of the northern ditch in slot 5000. The position of the monument, just below the crest of the north- facing slope, may indicate that it was intended to be visible from within the northern land unit. The irregular form of the northern ditch could suggest that it was dug along the northern edge of an existing bank, possibly to provide chalk to emphasise the visibility of the bank after vegetation had become established on it, although this is conjectural. OSL dating of the fill immediately above the flint deposit to AD 790-960 (1514a, 1190+90 before 1996) indicates that the monument probably survived as a substantial earthwork in the Middle Anglo-Saxon period; however, this was probably due to stabilisation by vegetation rather than deliberate maintenance. The apparent lack of maintenance of the monument may perhaps be explained by the possibility that its original function as a land-use division (see below) became redundant a relatively short period after its construction. The parish boundary running along the monument may suggest that it survived as a land division throughout the later prehistoric and Romano-British periods. What is more likely, however, is that the parish boundary was established along a conveniently placed, but long disused, earthwork. Evidence from the undated, but pre-bank, tree hollows (2033 and 4022) indicates woodland on Breach Hill at some time before construction of the earthwork. Immediately prior to construction there is evidence that this woodland cover had 100 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE been cleared and replaced with either open scrub or an established short-grazed grassland. Here the evidence from the two ditches differs. Nevertheless, regardless of which of these environments existed at the precise moment of construction, it is evident that the woodland had been cleared and that more open conditions prevailed. This is not surprising in view of the evidence for Mesolithic activity on site, the Neolithic long barrows in the immediately proximity (Figure 1), and the Early Bronze Age activity. The buried soil preserved beneath the bank would have provided definitive evidence of the nature of this environment, and although lost in the area investigated, it presumably survives under several kilometres of the earthwork elsewhere. In view of the entirely open nature of the very limited assemblage from the buried soil (although not all of this assemblage may be contemporaneous with the buried soil), and the very open nature of the assemblages from the southern ditch, it is assumed that an open and dry, grazed, grassland sward existed immediately prior to construction of the earthwork. This fits with evidence elsewhere in the area. The later Bronze Age linear ditch at Copehill Down was built, and existed, within open downland (Wessex Archaeology 1988), and is supported by the detailed sequence from the northern edge of Salisbury Plain (3km to the north) at Strawberry Hill (Allen 1992; 1994). The ditches divide two areas of different land- use, and thus a pattern of contemporaneous land-use can be inferred. Initially that to the south (i.e. the land extending towards Copehill Down) seems to be one of open grazed grassland and thus part of a managed or farmed landscape. In contrast, that to the north seems to be less well maintained, comprising longer grassland and even some shrubs, possibly reflecting ungrazed or rough pasture. Through time, however, the landscape which the earthwork divides became more unified with similar open grazed or tilled downland being seen to both south and north in the secondary fills (presumable of Iron Age date). The later (tertiary) fills of broadly Romano-British or later date reveal very open conditions, typical of farmed and maintained downland. There is, therefore, no evidence of abandonment here, or at least no evidence of vegetation regeneration that might be taken as abandonment or disuse of the local landscape; indeed, quite the contrary. The later fills of the ditches indicate more stable conditions in the ditches, but within a harsher drier downland environment of tillage and short-grazed grassland. In the Wessex Linear Ditches Project, much emphasis was placed on defining the environment and land-use of the later prehistoric land division but it was not possible in this excavation to define the land-use that the ditch system bounded, if indeed these earthworks were land-use boundaries (see also McOmish et al. 2002, 51-65, for a recent discussion of linear ditches). The examination of a molluscan sequence from a single ditch and bank is likely to reflect only one of those environments or at best a mixture of both the bounded land-uses which may make palaeo-ecological interpretation very difficult. This aspect has been tentatively suggested from other linear ditches on the Salisbury Plain (Allen unpub.). The opportunity to examine the nature of the linear earthwork as a boundary dividing areas of differing land-use is only made possible in palaeo-environmental terms because of the rare configuration of the bank having a ditch on both sides and the sampling and analysis of both ditch sequences. In the later Bronze Age, much of Salisbury Plain in the immediate area at least was open downland set aside for arable or pasture. The linear ditch crossing Copehill Down, although containing relatively sparse snails, 1s comparable with the analysis here, as is the evidence from, for instance, Earl’s Farm Down (Allen and Wyles 2004). Despite a relatively modest analytical programme, we can tentatively suggest that the linear earthwork that crosses Breach Hill may not just have divided the landscape, but also provided division in land-use. Areas to the south, i.e. around Copehill Down, contain a large number of field systems as recorded by aerial photographs and in the SMR (Figure 1). It is from this aspect that we have evidence of more open conditions, and grazed and tilled land from the earliest ditch fills. In contrast, the land unit between Breach Hill and the NW61]14 linear ditch, approximately 1.8km to the north, appears largely devoid of lynchets and field systems. The field systems, however, are not necessarily contemporary with the two monuments; indeed NW614 has a complex relationship with a number of field systems towards its western end (not shown in Figure 1) so this remains speculative. The northern ditch faces this aspect from which evidence of a less managed, longer ungrazed or rough pasture and scrub downland is seen in the early stages of ditch infill. It is therefore possible that the Breach Hill linear earthwork defines different territories of which the portions sampled, at least, were under different land-use regimes during and shortly after its construction. EXCAVATIONS ON THE OLD LINEAR EARTHWORK, BREACH HILL, TILSHEAD 101 The topographic position of the monument, on the north facing slope just below the crest ofa ridge, suggests that it was intended to be seen from within the northern, apparently less intensively managed, land unit. The nature of the activities represented by these differing land-uses and how they inter-relate is beyond the scope of this investigation. If the original function of the monument was that of land-use division, the environmental evidence indicates that this function soon became redundant. The snail evidence from the secondary fills which, by comparison with the results obtained from experimental earthworks, may have begun to form as little as 20-30 years after construction (Bell et al. 1996), indicates a similar open grazed or tilled downland to both the north and south. Whilst the monument does not appear to have been maintained it probably continued to function as a boundary. The alignment of the present parish boundary is a matter of interest in this respect. Other features Pottery of Late Bronze Age or Early Iron Age date was recovered from a single post-hole (1001, context 1000). This did not appear to be an element of any discernible structure and no other datable features were encountered in the vicinity, although several undated pits, hearths and post/stake holes were recorded in the same general area and may be contemporaneous. Romano-British material was recovered from a subsoil/sorted horizon, confined to the western end of the site, possibly representing a manuring scatter. ’ Romano-British pottery was also recovered from the upper (disturbed) fills of both the northern and southern ditches, often along with later material and is presumably residual. ACKNOWLEDGEMENTS The project was commissioned by the Defence Estates (Ministry of Defence) for whom Jane Hallett and Keith Maddison acted; Duncan Coe of Wiltshire County Council’s Library and Museum Service prepared the brief and monitored the excavation on behalf of Defence Estates, whilst Amanda Chadburn monitored the work for English Heritage. Discussions and advice concerning the archaeological deposits, finds and environmental sampling strategy have been provided by a number of individuals including Mike Allen, Phil Andrews, Duncan Coe, Ros Cleal, Roy Entwistle, Andrew Lawson, Lorraine Mepham and Frances Raymond. The excavation was managed for Wessex Archaeology by Ian Barnes, Roland Smith and Lorraine Mepham, and directed by Vaughan Birbeck with the assistance of Andy Crockett, Nick Wells, Paul Pearce and Steve Tatler. The excavation team comprised Dom Barker, Andy Bates, Steve Campbell- Curtis, Jez Fry, Trevor George, Chris Heatley, Barry Hennessy, Astrid Hudson, Emma Loader, Richard May, Phil McMahon, Kevin Ritchie, Joe Whelan and Gary Wickenden. Assistance with the sorting of snail samples was provided by Hayley Clark and with the identifications by Sarah E Wyles, who also undertook preliminary scanning of flots and residues for charcoal. Archaeomagnetic dating was conducted by GeoQuest Associates Ltd. Optical dating was undertaken by the Oxford University Research Laboratory. The illustrations for this report were prepared by S.E. James. Full details of all contexts are contained in the site archive which will eventually be deposited with Wiltshire Heritage Museum, Devizes. BIBLIOGRAPHY AITKEN, M.]J., 1976, Thermoluminescence age evaluations and assessment of error limits: revised system. Archaeometry 18, 233-8 AITKEN, M.]J., 1985, Thermoluminescence Dating. London: Academic Press ALLEN, M.J., 1992, ‘Products of erosion and the prehistoric land-use of the Wessex Chalk’, in Bell, M.G. and Boardman, J. (eds), Past and Present Soil Erosion; archaeological and geographical perspectives. Oxford: Oxbow Books, 37-52 ALLEN, M.]J., 1994, The land-use history of the southern English chalklands with an evaluation of the Beaker period using environmental data; colluvial deposits as cultural indicators. Unpublished Ph.D thesis, University of Southampton ALLEN, M.J., 199Sa, ‘Before Stonehenge’, in Cleal er al. 1995, 41-64 ALLEN, M.J., 1995b, ‘Land molluscs’, in Wainwright, G.J. and Davies, $.M., Balksbury Camp, Hampshire; excavations 1973 and 1981. London: English Heritage Archaeological Report 4, 92-100 ALLEN, M.J. and BAYLISS, A., 1995, ‘The radiocarbon dating programme’, in Cleal ef al. 1995, 511-2 ALEN, M.J. and WYLES, S.F, 2004, Land Mollusca, in Cleal, R.M.J., Allen, M.J. and Newman, C., An archacological and environmental study of the Neolithic and later prehistoric landscape of the Avon Valley between Durrington Walls and Earl’s Farm 102 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Down. WANHM 97, 218-48 BARRETT, J.C., 1980, The pottery of the Later Bronze Age in lowland Britain. Proceedings of the Prehistoric Society 46, 297-319 BARRETT, J., BRADLEY, R. and GREEN, M., 1991, Landscape, Monuments and Society: the prehistory of Cranborne Chase. Cambridge: University Press BELL, M., FOWLER, PJ. and HILLSON, S.W. 1996, The Experimental Earthwork Project 1960-1992. York: Council for British Archaeology Research Report 100 BRADLEY, R., ENTWISTLE, R. and RAYMOND, F, 1994, Prehistoric Land Divisions on Salisbury Plain: the work of the Wessex Linear Ditches Project. London: English Heritage Archaeological Report 2 CASE, H.J., 1977, ‘The Beaker culture in Britain and Ireland’, in Mercer, R. (ed.), Beakers in Britain and Europe. Oxford: British Archaeological Report S26, 71-84 CLARKE, D.L., 1970, Beaker Pottery of Great Britain and Treland. Cambridge: Cambridge University Press CLEAL,R.M.J., 1992, ‘The Neolithic and Beaker pottery’, in Gingell 1992, 61-71 CLEAL, R.M.J., 1995a, ‘Pottery fabrics in Wessex in the fourth to second millennia BC’, in Kinnes, I. and Varndell, G., (eds), Unbaked Urns of Rudely Shape — Essays on British and Irish Pottery. Oxford: Oxbow Monograph 55, 185-94 CLEAL, R.M.]J., 1995b, ‘The prehistoric pottery’, in Cleal et al. 1995, 349-67 CLEAL, R.M.J., Walker, K.E. and Montague, R. 1995, Stonehenge in its Landscape, the Twentieth Century Excavations. London: English Heritage Archaeological Report 10 CLEAL, R.M.J. and RAYMOND, EF, 1990, ‘The prehistoric pottery’, in Richards 1990, 233-46 CORNWALL, I., 1953, Soil science and archaeology with illustrations from some British Bronze Age monuments. Proceedings of the Prehistoric Society 19, 129-47 ENTWISTLE, R., 1994, ‘The environmental setting of the linear ditches system’, in Bradley et al. 1994, 101-21 EVANS, J.G., 1972, Land Snails in Archaeology. London: Seminar Press EVANS, J.G., 1983, ‘Molluscan analysis’, in Evans, J.G., and Smith, I.E, Excavations at Cherhill, North Wiltshire 1967. Proceedings of the Prehistoric Society 49, 101-9 EVANS, J.G., 1984, Stonehenge — the environment in the Late Neolithic and Early Bronze Age and a Beaker burial. WANHM 78, 7-30 EVANS, J.G. and JONES, H., 1973, Subfossil and modern land-snail faunas from rock-rubble habitats. Journal of Conchology 28, 103-29 GALE, R., 1994, ‘Charcoal identification’, in Bradley et al. 1994, 119 GARDINER, J.P, 1988, The composition and distribution of Neolithic surface flint assemblages in Central Southern England .Unpublished PhD thesis, University of Reading GIFFORD and PARTNERS, 1997, SPTA Southern Range Road: Evaluation Report. Unpublished client rep. 7428.4R GINGELL, C., 1992, The Marlborough Downs: A Later Bronze Age Landscape and its Origins, WANHS Monograph 1 HALL, S. and TITE, M.S., 1996, Optically stimulated luminescence (OSL) dating of the linear earthwork, Breach Hill, Wiltshire. Unpublished MS HEALY, F, 1998, ‘The Early Bronze Age settlement gap’, in Woodward, A. and Gardiner, J. (eds), Wessex Before Words: some new research directions for prehistoric Wessex. Salisbury: Council for British Archaeology Wessex Forum for Archaeolology in Wessex, 41-2 HARDING, PA., 1992, ‘The flint’, in Gingell 1992, 123-33 HURLBERT, S.H., 1971, The nonconcept of species diversity: a critique and alternative parameters. Ecology 52, 577-86 LANTING, J.N. and van der WAALS, J.D., 1972, British Beakers as seen from the Continent. Helinium 12, 20-46 MACPHAIL, R.L., 1987, The soil micromorphology of tree subsoil hollows. Circaea 5, 14-17 MACPHAIL, R.I. and GOLDBERG, P,, 1990, ‘The micromorphology of tree subsoil hollows: their significance to soil science and archaeology’, in Douglas, L. (ed.), Soil Micromorphology: a basic and applied science. Amsterdam: Elsevier, 431-40 McOMISH, D., FIELD, D. and BROWN, G., 2002, The Field Archaeology of the Salisbury Plain Training Area. Swindon: English Heritage MORRIS, E.L. 1994, The Analysis of Pottery. Salisbury: Wessex Archaeology Guideline 4 MORRIS, E.L., 2000, ‘Pottery fabrics’, in Lawson, A.]J., Potterne 1982-5: Animal Husbandry in Later Prehistoric Wiltshire. Salisbury: Wessex Archaeology Report 17 PCRG, 1997, The Study of Later Prehistoric Pottery: general policies and guidelines for analysis and publication. Prehistoric Ceramic Research Group Occasional Papers 1/2, revised edition RACKHAM, O. 1986, The History of the Countryside. London: Dent RAHTZ, PA. 1962, Neolithic and Beaker Sites at Downton, near Salisbury, Wiltshire. WANHM 58, 123-7 RAYMOND, EF, 1994, ‘The pottery’, in Bradley er al. 1994, 69-90 RICHARDS, J.C., 1990, The Stonehenge Environs Project. London: English Heritage Archaeology Report 16 THURNAM, J., 1869, On Ancient British Barrows: part 1, long barrows, Archaeologia 42, 161-244 TUTIN, T.-G., HEYWOOD, V.H. et al., 1964-80, Flora Europaea, 1-5. Cambridge: University Press WESSEX ARCHAEOLOGY, 1988,. Excavations at the SPTA FIBUA Site, Copehill Down 1987-8. Unpublished client report WESSEX ARCHAEOLOGY, 1995, SPTA Tracks 28, 43A EXCAVATIONS ON THE OLD LINEAR EARTHWORK, BREACH HILL, TILSHEAD 103 &54D, Archaeological Evaluation. Unpublished client rep. 39404 WESSEX ARCHAEOLOGY, 1996a,. SPTA Track 58, Breach Hill, Tilshead, Archaeological Evaluation. Unpublished client rep. 42046 WESSEX ARCHAEOLOGY, 1996b, SPTA Track 21E and 21K Tilshead, Archaeological Evaluation. Unpublished client rep. 42024 WESSEX ARCHAEOLOGY, 1999, Salisbury Plain Training Area Southern Range Road: Stage 1 Archaeology Contract. Unpublished client rep. 44897.1 WYMER, J.J. 1977, Gazetteer of Mesolithic Sites in England and Wales with a Gazeteer of Upper Palaeolithic Sites in England and Wales. London: Council for British Archaeology Research Report 20 APPENDIX: OPTICALLY STIMULATED LUMINES- CENCE (OSL) DATING by S. Hall and M.S. Tite This report is a summary of the detailed report produced by the Research Laboratory for Archaeology and History of Art, University of Oxford. The full details can be found in archive (Hall and Tite 1996). Four sediment samples were taken from three open sections; two from ditch 4001 (the southern ditch in slot 4000), one from ditch 4002 (the northern ditch in slot 4000) and one from the southern ditch in slot 3000. The samples were pre-treated to remove calcium carbonate, organic material and clays. One sample (1514d from context 3062, ditch 3001) contained insufficient quartz in the sample for viable analysis. The luminescence signal was Table 6: OSL results measured using an Elsec optical dating system. The luminescence signals from the polymineral grains of each prepared sample were measured by exposure to infrared light, which stimulates a luminescence signal only from feldspars. Measurement followed normal procedures adopted by the laboratory (see archive) and included a preheat test at 160°C. The measurement procedure for dating involved constructing additive alpha and beta growth curves. The annual dose rate received by the sediments during burial was calculated using a portable gamma spectrometry, inductively coupled plasma mass spectrometry (ICPMS) and potassium flame photometry. The final dose rate used for the age calculation was a combination of these sets of measurements: the total gamma dose rate and the beta dose rate due to uranium and thorium was determined by on-site gamma spectroscopy and the remaining beta dose due to potassium was determined by flame photometry. The alpha dose was determined using the results of the ICPMS analysis. The saturation water contents of the sediments were also measured at the time of sampling and found to be between 65 and 80% of saturation values. As a result a value for water content of 80+20% was used to correct the beta and alpha dose rates absorption of energy by water. The errors quoted in association with the age estimates take into account both systematic and random errors (at a 68% confidence level) in OSL measurements, dose-rate measurements and calibrations radioactive sources and equipment. The calculation of these errors follows the method of Aitken (1976; 1985). The largest contribution to the errors for these samples was that related to the dose rate and the assumed value of the water content of the sediment during the burial period. The results are presented in Table 6. Ditch — Context Lab. ref — Palaeodose a-value ‘G Dose-rate Age years (before meas- Age range ry) (mGy/yr) urement = 1996) (— 1995) _ 4001 4014 1Sl4a 1.43 + 0.04 0.06 1.283 + 0.39 1,115 + 90 AD 790-960 4001 4017 1S14b 2.53 + 0.06 0.08 0.85 + 0.25 2,980 + 240 1225-745 BC + 2355-1715 BC 4002 4015 I5l4c 6.73 0.12 0.08 3001 3062 1Sl4d - 1.67 + 0.54 4,030 + 320 ol. 99 (2006), pp. 104-164 Wiltshire Archaeological & Natural History Magazine, v s5\ See ‘a ITT W ia LY = oe q O at = nhl ees 7, OH F sent n 2 = Bo & = Dj 2 oO a ry) 3 th ie iA a pn i) 2 os d PA\ eee \ ) me ‘ 8) 5 1_| Z 5 orf oA \ (4 Ey 4 r ci \ \8 oO ML] S| 1 Ip si mee < ANF = acl aN B, “yy LO \ ey "| c) S ‘ R r Yo\in 0 2 QT \ Us oO O i A lan, scale 1:3000 ion p Fig. 1 Site locat IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 105 Iron Age and later defences at Malmesbury: Excavations 1998-2000 by Timothy Longman with contributions by Lisa Brown, Rod Burchill, Paul Davies, Rowena Gale, Julie Fones, David Mullin and Naomi Sykes Malmesbury occupies a steep sided roughly triangular promontory on the north bank of the Bristol Avon. The excavation of ten trenches, between 1998 and 2000, at intervals along the line of the eastern circuit of the historic town wall, revealed evidence of human activity from the prehistoric period to the twentieth century. Extended occupation during the Iron Age was represented by several constructional phases of ramparts and ditches of amultvallate hillfort, upon which the modern town 1s situated. The original Early Iron Age inner rampart was probably a clay bank, revetted in front with timber. Three postholes, possibly associated with such a timber revetment, were sealed beneath a later layer of rampart material. A ditch, set in front of a berm less than 0.50 m wide, was probably contemporary with the earliest rampart. The foundations of a 3.72m wide limestone wall, which originally would have probably stood some 7-8 metres high, and dated from the Middle Iron Age, surmounted the latest phase of inner earthen rampart. The outer defences consisted of a further series of successive earth banks surmounted at various times hy timber and stone ramparts. A second, outer, ditch was excavated in front of the outer bank and contained a large quantity of collapsed burnt masonry derived from one of the earliest phases of rampart. Ceramic evidence and radiocarbon dates suggest that the ditch originated in the Early/Middle Iron Age. In general terms, most of the layers of rampart material were separated by buried soils and contained considerable quantities of burnt debris, including oak charcoal, heat affected clay and limestone masonry, some of which was vitrified. Most of the stonework belonging to the inner revetment wall (within Trench 10) had been robbed as evidenced by several large quarry pits backfilled with domestic rubbish, including sherds of Anglo-Saxon pottery. The town is mentioned in the Burghal Hidage, the early 10th century list of fortresses defending the coasts and ‘frontiers of Wessex. There was evidence of a substantial earth and stone bank dating from the 10th or 11th century on the line of the inner Iron Age defences. In the late 11th or early 12th century, two successive stone buildings were erected at right angles to the inner line of the town defences, although their function 1s unclear. Building I was constructed some distance back from the wall permitting the presence of a walkway, or intramural lane, between it and the town defences. Building 2 was positioned further to the south-east and later demolished for the construction of another major defensive embankment. A stone retaining wall was later added to the rear of the embankment and may be associated with the construction of the early 12th century town wall. In the early 12th century the town defences were improved by the construction of a substantial stone wall following the inner line of the Iron Age and Anglo-Saxon ramparts. There was archaeological evidence of the original structure as well as the subsequent improvements and modifications that were made in the late 12th and/or the early 13th century and in the 15th century. A deep occupation layer with considerable quantities of butchered animal bones and 13th/14th century pottery sealed the medieval inner rampart and was later truncated by landscaping work associated with the construction of the nearby Great Western Railway line in the 1870s. It appears that land adjacent to the railway cutting was heavily disturbed, a fact that probably partly accounts for the absence of post-medteval archaeology, including evidence of the town’s 17th century Civil War defences. Bristol and Region Archacological Services, St. Nicholas Church, St. Nicholas Street, Bristol BS} LUI 106 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE INTRODUCTION The town of Malmesbury (Figure 1) occupies a steep sided roughly triangular promontory on the north bank of the Bristol Avon in north-west Wiltshire. The area investigated (Figure 2) (centred at NGR ST 9353 8731) comprised a 200 metre long section of the eastern circuit of the historic town wall and a narrow parallel strip of land known as the ‘Nun’s Walk’. The Nun’s Walk, a footpath which rises gradually from approximately 80m OD (at school steps) to 83m OD (at church steps) is bounded to the north by St Joseph’s R. C. Primary School, west by the back gardens of properties (including Malmesbury Library and St Aldhelm’s R. C. Church) fronting on Oxford Street and Cross Hayes Lane, and east by two large paddocks that slope steeply down to a narrow millstream. The geology comprises Cornbrash and shelly oolitic limestone (Acton Turville Beds) and clay of the Great Oolite Series from the Jurassic period. Only a limited amount of prehistoric evidence had been recorded before the 1998-99 excavations; a single flint tool (WSMR 150), from the Late Neolithic/Early Bronze Age, was found during a watching brief at the rear of 26-30 Backhill in 1997 (Pilkington 1997). In 1805, Mr. J. A. Moffat noted that an excavation near the East Gate in the late eighteenth or early nineteenth century ‘at a small distance from the base of this wall (the town wall) about six feet beneath the surface of the earth, a substance has been discovered, which has supposed to be vitrified matter, and has been imagined that the place was formerly encompassed with a vitrified bank or vallum’ (Moffat 1805). The remains have been interpreted as a timber and stone box rampart burnt in situ (Haslam 1976). Early in the nineteenth century (Gents. Mag. 1831) during excavations of the town wall (NGR ST 9325 8705), near Kings’ Wall, several Iron Age or Romano-British loom weights were found ‘n considerable quantities’ in a layer of red earth overlaid with clay. The red earth and stones, apparently, all showed signs of burning (WSMR 301). The possibility of a hillfort at Malmesbury has been theorised for two hundred years or more. Indeed, in 1976 Jeremy Haslam postulated that the steep topography of the hill upon which the modern town is situated has its origins as an Iron Age hillfort (Haslam 1976). Measured survey by Mark Corney (Bristol University) of the Nuns’ Walk area in 1998 led him to conclude that an earth bank at least 4m wide and Im high could indeed belong to a hillfort (Corney 1998). In its regional context, Malmesbury may be regarded as an outlier of the multivallate hillforts on the higher ground of the Cotswolds. One entrance on to the promontory is at Abbey Row, on the northwest side, from the west along a narrow limestone ridge. As yet no excavation has confirmed this. In addition, the route of the ‘Nun’s Walk’ from the north-east curves around the eastern perimeter of the Iron Age defences, possibly indicating another former entrance near the southern end of the promontory, just north of the medieval South Gate. Limited Romano-British archaeology has been recorded, but the little that has been suggests settled occupation of the hillfort. Mr. G. Esdaile recorded a Romano-British building in the abbey grounds in 1887 at NGR ST 934 873 reporting that ‘a hypocaust was recently removed’ (WSMR 300). During an excavation by Julian Richards in the abbey grounds in 1997 at NGR ST 9334 8732 several sherds of late Romano-British pottery (WSMR 308) were retrieved from a sealed context. Other finds from the period comprise a Romano- British coin (follis of Valentinian I) (WSMR 303) found in Westport (NGR ST 9297 8743) and a scatter of pottery and a coin (imitation centenionalis) (WSMR 306) found during excavations at Postern Mill, Burnivale (NGR ST 932 872) in 1986-87. The early history of West Saxon settlement in Malmesbury is generally uncertain. There are references to a nunnery c.603 (Knowles & Hadcock 1957). It has been suggested that the name Malmesbury is possibly derived from Maeldulbh, the name ofan obscure Irish hermit or monk who is reputed to have settled in the town c.637 (Brakspear 1930). Medieval abbey records refer to the early English settlement of ‘Ingleburne’, the name by which the nearby Tetbury tributary of the River Avon was known until at least the 15th century, but by around AD 675 the settlement was known as ‘Mealdumesburg’. At about this time Aldhelm, a nephew of the West Saxon King Ina, established a monastery (after c.965 the Benedictine abbey) in the town (Knowles & Hadcock 1957). It is generally thought that Malmesbury had acquired defences by the late 9th century, when it is recorded as one of the fortified towns in ‘the Burghal FAiidage’, established by King Alfred to counter the threat of a Danish invasion. It is also likely that the street system and main property boundaries were laid out at this time. The early 10th century ‘Burghal Hidage’ records 1,200 hides assigned to the burh indicating defences between 1,650-1,870 yards 107 IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 — 2? —_- = \\ JT EY), Fone \ \ ~ == Ss wie il ZN Gate imal ~ —— ie Oh a i i a = p=] S Sules ~~ val town defences and archaeological (SMR) Saxon/medte mentioned in the text é line of the Fig. 2 Plan of Malmesbury showing th 108 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Trench 10 Trench 6 Trench 4 Trench 2 FB __ footbridge IC Inspection cover © O tree (to scale) 50 Metres eS ee After Mark Corney, 1998 Fig. 3 Trench location plan | | IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 long (Elrington 1991). Malmesbury was supposedly granted its first borough charter by King Athelstan (924-39) (Cat. No. 454; Sawyer 1968), but this document is a late 13th or 14th century forgery (Kelly, 2005). Tait has described it as ‘the most obvious of post-Conquest forgeries’ (Tait, 1936). By the late 10th century there was a mint in the town and it was probably the most important moneying centre in Wiltshire in the early 11th century. There have been limited finds of archaeological material from the Anglo-Saxon period, prior to the 1998-99 excavations. Work at Postern Mill, Burnivale (NGR ST 932 872) in 1986-87 found evidence of industrial activity (WSMR 474). Late Saxon pottery in a dark soil, human burials and a stone wall (WSMR 400) were found during excavations in the Market Place (NGR ST 932 874) in 1988 and 1993. Another excavation in 1993, at the rear of the library (NGR ST 9346 8715), revealed an earthen rampart, pottery sherds and a clay spindle-whorl (WSMR 404) (Duck and Toombs 1993). The building of the early 12th century town wall is usually attributed to Roger, Bishop of Sarum (d.1139). The original circuit of the wall was generally assumed to respect the line of the West Saxon defences. Bishop Roger is also thought to have been responsible for the building of a motte and bailey castle in the town c.1118. Documentary research places this structure in the area later occupied by the abbey cloisters and the grounds of the later Bell Hotel (formerly Castle House). This has yet to be confirmed by archaeological excavation. It is also thought that Roger Bore initiated the rebuilding of the abbey church early in the twelfth century, although it was not completed until c.1160-70. During the civil wars between King Stephen and the Empress Matilda in the mid-12th century the castle and walled town were considered an important stronghold and control changed hands several times between 1139 and 1153. In 1215 King John granted the lordship of the manor, which the crown had seized after 1066, back to the abbey, and the following year he gave permission for the abbot to demolish the castle. It was only from the 12th century that English boroughs began to acquire charters defining their privileges, and Malmesbury itself does not appear to have been granted borough status until the later 13th century, being summoned to Parliament as a ‘market town’ (villa mercatoria) in 1275, and then resummoned as a borough in 1295 (Kelly, 2005). The town wall was extant in 1283-4 when arrangements were made for its repair (Ballard, 109 1906). In the late 13th century the walled part of town was called ‘Bynport’ to distinguish it from the suburb of Westport — the name was still in use in the sixteenth century. At this time there were at least four gates: the East Gate across the Oxford road (Holloway), Wyniard Gate was a small gate at the southern end of Silver Street, the South Gate at the southern end of the High Street, and the Postern Gate at the junction of King’s Wall and Burnivale. A fifth, west, gate probably at Abbey Row, at or near the site of the castle, was mentioned in the earlier 13th century (VC.H. 1991). Although the rebuilt walls on the line of the historic defences date only from the mid-19th century or later, the plan of the town within is probably little different today from what had been established by the late 13th century. The eastern and south-eastern sections of the line of the town wall are marked by the eastern boundary wall of plots behind Cross Hayes Lane and Silver Street, and by the southern boundary wall of the plots behind Ingram Street. The western circuit is indicated by the western boundary wall of plots between the lane called Burnivale and Gloucester Street, and the lane known as King’s Wall and the High Street. To the north the wall may survive as the garden wall of the Bell Hotel; east of the Bell Hotel the town wall was probably also the outer wall of the abbey buildings that extended to the edge of the steep slope above a tributary of the River Avon. Many excavations have recorded archaeological material from the medieval period in the town. Excavation near the Market Cross (NGR ST 9331 8725) (WSMR 460) recorded a substanual boundary ditch (probably the boundary between the town and the abbey), which had been infilled by the late 12th century and several human burials that probably lay in the former burial ground of St Paul’s Church. Excavation in the same area in 1993-94 recorded a late 13th/early 14th century floor surface (WSMR 477 and in 1996 two human skulls. Excavation in 1993 behind the library (NGR ST 9334 8715) found part of the 12th century town wall (WSMR 479) (Duck and Toombs 1993). By the mid-1l6th century the town wall was in poor condition. When John Leland visited the town in 1542 he recorded that ‘Jn the toun be 4 gates by the names of est, west, north and south, ruinus al. The walles in many places stand up: but now very feble’. It has been assumed that Parliamentary forces undertook the alterations and repairs that were carried out on the walls during the Civil War (1642- 49) (Figure 4) after their final capture of the town 110 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 4 Part of the ‘Bird’s Eye View”: an illustration of the town defences during the Civil War c.1646 in May 1644. From then until the late summer or autumn of 1646 a garrison of possibly as many as 1,000 men was kept in the town. By then it had been decided to slight the town’s defences and so render them useless to the Royalist forces should the town fall to them again. Two of the town gates survived intact into the late 18th century. The East Gate was partly demolished for road widening in 1778 and the Postern Gate was demolished at about the same time, but by about 1800 the town wall had largely disappeared. In Holloway the south jamb of the medieval East Gate survives, and the west side of the South Gate may be reflected in the rounded plan of a house at the junction of the High Street and King’s Wall. A large part of the circuit of the town wall has statutory protection as a Scheduled Ancient Monument (Wiltshire 881). It was noted in the early 1990s that several sections of the wall were deteriorating while other sections had actually collapsed. Organisations at both national and local level drew up plans to record, monitor and, if necessary, rebuild parts of the wall. This resulted in 1994 in Keystone Historic Buildings Consultants conducting a detailed survey of the extant circuit of the historic town wall. As a result of this survey and further work by Fondedile (Foundation and Consolidation Specialists) and Locks Geotechnical in 1995, Stansell (Conservation) Ltd was contracted in 1998 to undertake the consolidation and reconstruction of the section of wall between St Joseph’s R. C. Primary School and St Aldhelm’s R. C. Church (Figure 1). — Bristol and Region Archaeological Services (BaRAS) was contracted to carry out archaeological monitoring, which comprised a watching brief and the excavation of nine trenches. The objectives of the work were twofold: (1) to identify the cultures associated with the periods of occupation of the site; and (2) to investigate the structural history of the site. The project commenced on 25 August 1998 and was completed on 15 June 1999. The programme of excavation work in March- April 2000 was undertaken as a direct consequence IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 111 of the 1998-99 project. Due to the garden of St. Joseph’s School having been used for storage and as the location of temporary site offices, the local authorities paid for the laying out and planting of a new garden. Local parties suggested that because of the findings of the earlier excavations, the opportunity should be taken to investigate the archaeology of the garden area by digging a single trench, prior to landscaping work. BaRAS drew up proposals for excavation and English Heritage funded the project, which was approved in February 2000. The objectives were threefold: (1) to further investigate the complexity of the Iron Age defences; (2) to determine the nature and intensity of Iron Age occupation; and (3) to attempt to clarify the extent of the Anglo-Saxon defences. The excavation took place from 6 March to 7 April 2000. The archive of records and finds from the excavations will be deposited with Devizes Museum under accession number DZSWS 2005/109. EXCAVATIONS IN 1998-99 | Period Summary | Period I, Phase I: Late Neolithic/Early Bronze Age (c.2500-1800 BC) flints. Period II, Phase I: Early Iron Age (c.800-500 BC) inner and outer ramparts/ditches and postholes. Period I, Phases H-VIIT; Middle Iron Age (c.500-100 | BC) inner and outer ramparts (including inner revetment wall) and occupation layers. | Period III, Phases I-V: Late Anglo-Saxon (Late 10th/11th century) inner and outer ramparts, | occupation layers and rubbish pits. | Period IV, Phases I-X: Medieval (late 11th/15th century) Buildings | and 2, occupation layers, inner and outer ramparts and town walls. Period V, Phases I-IV: Post-medieval (16th century to date) paths, walls and occupation layers. Trench 1 Period IV: Medieval: | At the base of the trial pit a thick layer of possibly | redeposited mixed stiff yellow clay and limestone | fragments (103) contained a single sherd of 14th/15Sth century pottery. Period V: Post-medieval: Sealing the medieval horizon was a 19th-century (Phase III) black ashy deposit (102), probably a bedding layer for limestone slabs that lay directly above. This flagstone floor comprised several square limestone slabs (101), which in turn lay beneath a layer of dark brown loamy topsoil (100) (Phase IV). Trench 2 (Figures 5 & 6) Initially a trial pit, this trench was also excavated to assess the structural condition of a section of the extant town wall. Only 18th and 19th century archaeology (Period V: Phases IT & IIT) was recorded within a 1m high earth bank that had built up against the east face of the wall. A layer of orange brown clay (204) lay at the bottom of the pit and was sealed by a mid-brown silty clay subsoil (201) containing two sherds of late 18th/early 19th century pottery. The subsoil had been truncated by a small sub-rectangular pit (202) filled with a mix of black ash and fragments of plaster (203). Several sherds of 18th/19th century pottery were recovered from the pit fill, which was sealed by a thick deposit of dark brown loamy topsoil (200). This trial pit was later considerably enlarged (Figure 5) for wall reinstatement and a much more complex series of deposits was revealed, with finds ranging in date from the Late Neolithic/Early Bronze Age to the post-medieval period. This section of the town wall, which was in a very ruinous state, had to be almost totally dismantled, leaving just the foundations. Behind the internal face of the wall was a 2.5m high section (Figure 6), some 23 metres long, through deposits that had accumulated since the Early/Middle Iron Age. The ground surface, in the back gardens of properties west of the wall, was 3.0m above the Nun’s Walk. Period IT: Iron Age Phase I: Early Iron Age: At the base of the section was a layer of mid-brown silty clay with charcoal flecks and several slabs of limestone masonry (235). It was sealed beneath context (234), a layer of orange-brown silty clay containing charcoal and limestone rubble (Rampart I). This deposit produced a small sherd of Early Middle Iron Age pottery. Sealing that deposit was context (217), a deposit of yellowish-orange clay, containing the base of a possibly residual small Early Iron Age vessel (c.800-600 BC). Above this lay a former turf-line composed of mid-brown clayey THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 112 Z youasy fo uousas surdvf-1svq 9 ‘su pe A [igcoaaee te oS OP Os tain u See | pre ave S bpe— x oO o % > 2° ope 2 ee A 622 hee aes | 3N nee 110M UMoL ay? Suimoys 7 youasy fo ud] § ‘Buy £ ro) S hd & & Se pe ee 3. : g bares Bee i 2 2 ave : egigg 5 --- vee IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 113 silt with lenses of yellow clay and charcoal, plus several limestone slabs (215). The top of the possible rampart was recorded at 82.10 metres OD. Phase IIT; Middle Iron Age: Sealing the earlier rampart was context (231), an orange brown silty clay identical to context (214); both were sealed by context (213). This layer contained several lenses of yellowish-red and light yellowish-brown clay as well as inclusions of charcoal and limestone rubble (Rampart III) and a single sherd of Middle Iron Age pottery, indicating that these layers probably represent phases associated with the later hillfort defences. The top of the rampart (Rampart III) was recorded at a height of approximately 82.40 metres OD. Period IT: Anglo-Saxon: Visible in the east facing section were two undated, but possibly Anglo-Saxon rampart deposits. These comprised two similar layers of orange-brown silty clay (228) and (227), containing charcoal, heat affected clay and fragments of limestone. The top of the rampart was recorded at a height of 83.25 metres OD. There were no datable finds. Period IV: Medieval Phases VI-VI: 12th/13th century Two lengths of limestone wall, contexts (223) and (224), which could be remnants of the original early twelfth-century town wall (Phase VI), were truncated by a construction cut (236), associated with a later phase of the medieval wall. Sherds of 12th/13th-century pottery were recovered from the later wall (219) during cleaning. The exposed length of wall foundation could be seen to comprise five distinct abutting sections (218), (219), (220), (221), and (222). Context (237), a deposit of dark brown sandy silt, had built up against the latter wall. Stratigraphically, the succeeding deposits (238) and (243) probably derive from building work and ‘trample’ apparently contemporary with the late 12th/early 13th-century reconstruction of the town wall (Phase VII). Context (238) was a mixed lens of dark greyish-brown silt, small fragments of limestone and orange-brown clay. Deposit (243), a layer of orange-brown clay, was sealed by a possible medieval soil horizon, 0.45m thick, composed of dark brown clayey silt (212). A group of deposits had slumped down the front of the exposed section at the north-east end of the trench, probably when an earlier length of wall collapsed. These deposits (contexts 240-242, and 244- 247) have been assigned to this period, although this is tentative due to limited dating evidence. Period V: Post-medieval: The upper few layers of post 18th-century (Phases III-IV) build-up comprised dark brown loam (210), a thin lens of limestone and brick rubble (211) and limestone wall foundations (230), (232) and (233) produced no datable finds. Trench 3 (Figures 8, 9 & 10) Initially excavated, like Trenches | and 2, as a lm sq. trial pit to assess the condition of a section of the town wall, only 19th-century deposits were recorded. The trench was later extended across the earth bank at a right-angle (east-west) to the standing wall. The new trench measured 2m x 3m. Period IT: Iron Age Phase I: Early Iron Age: The earliest deposit recorded was a thin lens of stiff reddish-orange clay (308) that sealed the limestone bedrock and formed the primary fill ofa ditch (Ditch I), context (309), at the western end of the trench. The fill contained no datable finds. Above context (308) a further five ditch fills were recorded (307), (306), (305), (310) and (312). Ditch 1 proved to be steeply cut (Figures 7, 8, 10 & 11) and was 0.8 m deep. Its full width (E-W) was not determined but it was aligned roughly north/south and cut into the limestone bedrock. The bottom of the ditch was recorded at 79.18 metres OD. The secondary ditch fill, context (307), consisted Fig. 7 View southeast across the line of the outer ditch in Trench 3 114 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE © ‘ bedrock = aa section Fig.9 \ 79.04m \ \ ditch fil ae \ a i ee ee /309) I i] bedrock | 79.9. | ' i 79.75M -_—_ - — - zt 2m SSS] Fig. 8 Plan of Trench 3 Fig. 9 East-facing section of Trench 3 almost entirely of medium to large sized slabs of oolitic limestone masonry (some burnt), that represent the collapse or demolition of a revetment wall (Rampart II) in the Early/Middle Iron Age. Apart from the rubble, the context comprised a reddish-orange clay with charcoal flecks and small nodules of yellow clay. A calibrated radiocarbon date of 800-250 cal. BC was obtained from a fragment of a burnt oak timber, possibly from a post. Sealing context (307) was a 0.4m deep dark 80m x Fig. 10 North-facing section of Trench 3 brown silty soil (306) that contained what appeared to be vitrified masonry, as well as large fragments of burnt wood. A single sherd of Early Iron Age pottery (c.800-600 BC) was recovered. Ditch fills were the result of two major phases of deposition. Above fill (306) were a further three orange-brown and yellowish-brown silty clays (305), (310), and (312), associated with an eroded/collapsed early phase of the Iron Age defences. The first of these, fill (305) was a deposit of orange-brown silty IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 115 Fig. 11 East-facing section in Trench 3 showing fills within the outer Early/Middle Iron Age ditch 309 clay sealed by a lens of pale brownish-yellow clay ‘with flecks of charcoal and heat-affected clay (310). The last of these deposits, fill (312), was a 0.3m deep horizon of orange-brown silty clay with inclusions of charcoal, heat-affected clay and burnt limestone, but no datable finds. Sealing the ditch fill was a succession of Iron Age rampart deposits consisting of orange-brown and yellowish-brown silty clays (304), (303), (302) and (311) (Figure 9). These deposits had suffered erosion and slumped down the bank. They contained small inclusions of charcoal, heat-affected clay, slabs of limestone and, in (304), several lumps of vitrified slag-like material, but no pottery or other datable finds. Context (304) contained part of a burnt wooden post (later identified as oak); a calibrated radiocarbon date of 770-400 cal. BC was later determined, suggesting a Middle Iron Age date for the structure of which the post had formed part. Period V: Post-medieval: Above the Iron Age deposits were a series of 19th-century (Phase III) layers. Sealing the latest prehistoric deposit was a 0.75m wide footpath composed of loosely compacted yellowish-orange gravel (313). It was aligned slightly to the west of the modern Nun’s Walk path. Above this was a mid- brown clayey silt subsoil (301) that had slumped down the bank and covered the footpath. A layer of dark brown loam (300) produced several sherds of 18th/19th-century pottery and sealed the latter deposit. Trench 4 (Figures 12, 13 & 14) Excavated in two stages, the first trench (NGR ST 9352 8729) was located on the results of a geophysical survey of an earthwork believed to be the site of a probable late medieval tower known as ‘the Cat Tower’, shown on a mid-17th-century illustration of the town defences called ‘the Birds Eye View’ (Figure 4). At the base of the trench was a possible medieval embankment (405) at least 2m wide and 1m high, aligned roughly north/south, and composed of small, crushed fragments of limestone. Sealing this was 2 greyish-brown silty soil (404) that produced a single sherd of undated pottery. Above it were a series of greyish-brown silty soils (403), (402) and (401), containing 18th and 19th century ceramics. Sealing the latest of these was a modern topsoil (400). Excavation stopped when no trace of the medieval tower was revealed, but the trench was later reopened and extended (Figure 10) to the west as far as the east side of the steps that allow access between the Nuns’ Walk and St Josephs’ School. Period III: Anglo-Saxon Phase I: Late Anglo-Saxon: The earliest deposit recorded was a dark brown clayey silt (415) that contained a single sherd of 11th-century pottery. This buried soil was sealed by a probable rampart deposit, context (409) (Rampart 1X), composed of crushed limestone and containing a few sherds of 10th/early 11th-century pottery. Period IV: Medieval: Sealing the probable late Anglo-Saxon rampart was a buried soil that consisted of dark brown clayey silt (408) containing 14th/15th-century pottery. The latter deposit lay beneath context (405), another rampart deposit (Rampart X) of heavily compacted, crushed limestone fragments, with only a single undated sherd recovered. Towards the western end of the trench a 0.4m THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 116 pb youasy fo uongsas surspf-yinog pl ‘sy Wi 0 f SOv p youady fo unjey ZI Sy Ae|o pue auojseusl| payisodapai J. —_—_,> Zt 614 uoljoas Re s}oo1 304) \}) 0 —— oo | ADVGONOS sOuoo) ajas0Uu090 |;]EM IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 117 thick orange-brown silty clay (406) was recorded, with inclusions of charcoal and very small fragments of limestone which sealed the natural limestone bedrock in that area. Two sherds of 15th/16th- century pottery were recovered. Period V: Post-Medieval: A lens of yellow clay (414) sealed context (406). Probably contemporary with that deposit was a layer of reddish-brown clayey silt (413). Neither deposit (Phase I) produced datable finds. This latter deposit was sealed by both a path (410) and a wall (411). The 0.8m wide footpath was composed of yellowish gravel. This probably 19th-century feature (Phase III) was bordered to the east by the foundations of a narrow limestone wall and to the west by a line of pennant slabs (412), set on end, forming the edge of the path. Both footpath and wall were sealed beneath modern topsoil (400) (Phase IV). Trench 5 (Figures 15 & 16) This trench was excavated in the grounds of St Joseph’s School, on the line of the historic town wall. A length of wall, context (S02), had collapsed and as part of its restoration a new foundation trench, initially measuring 2m x 0.9m, was excavated. After several extensions, for further wall reconstruction, the trench finally measured 12m x 1.5m by 1.8m deep. A sondage, c.lm sq, was excavated towards the north-east end of the trench. Period IT: Iron Age Phases IV-VII; Middle Iron Age: At the base of the sondage was the surface of a rampart (81.32m OD) (Rampart IV) composed of a mix of heavily compacted small crushed fragments of limestone and yellow clay (531). Above was a former turf line comprising lenses of greyish-brown and yellowish-brown clayey silts (530), (533), (S32),(529) and (528). Several sherds of Middle Iron Age pottery were retrieved from contexts (528) and (530). Sealing the buried soil was a succession of redeposited orange-brown and yellowish-brown silty clays with limestone masonry rubble inclusions, specifically contexts (507), (523), (525), (526) and (S27), which constituted another probable rampart (Rampart V). The deposits contained charcoal and reddish heat-affected clay. The surface of the feature was recorded at a height of 81.72 metres OD. Eighteen sherds of Middle Iron Age pottery were recovered from contexts (523), (525), (526) and (527). These deposits were in turn sealed by an old turf line, of yellowish-brown clayey silt (506), that contained Middle Iron Age pottery. Sealing this turf line was a suff orange-brown clay (S19). This possible rampart deposit (Rampart VI), the surface of which was recorded ata height of bottom slab of wall 81.05m n ee = — 2 Fig. 1S South-facing section of Trench S THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 118 trac) ye Eets 9 youasy, fo uoiaras Surspf-1svqy [su ¢ youasy fo uorzas surspf-Isvqy OT Buy BES 2S = ges | ses 4 ves \ v \ _ \ ——SSEOET = — | sauojsbeyy IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 119 82.09 metres OD, contained sherds of Middle Iron Age pottery. The rampart deposit (519) was partially sealed by contexts (515) and (538). These were the earliest of a 0.5 m deep horizon of thin buried soils. In addition to (515) and (538) the deposits comprised contexts (517), (537), (536), (542), (541), (535), (505), (540) and (511). Deposits (505), (S15) and (517) contained sherds of Middle Iron Age pottery. A calibrated radiocarbon date of 400-110 cal. BC was obtained from a sample of burnt oak timber from context (505). Within the buried soil horizon were thin lenses of red heat-affected clay (510), charcoal (512) and greenish-yellow clay with burnt limestone rubble inclusions (524). Sealing context (511) was a deposit of mid brown clayey silt (508), representing a former turf line. It contained several sherds of Middle Iron Age pottery. Above context (508), at a height of 82.66 m OD, was the latest (Rampart VII) of four Iron Age ramparts recorded in this trench (Figure 15). It was composed of yellowish-orange clay with small limestone inclusions (504). Ten sherds of Middle Iron Age pottery were recovered from the deposit. Sealing the latter rampart deposit was a mid brown clayey silt (S03), which represented an old turf line. It was sealed by a medieval deposit (501). Period IV: Medieval: A possible 12th century rampart (Rampart XI) (Phase VI), composed of large limestone rubble in a matrix of stiff brownish-yellow silty clay (501) was sealed by the topsoil (500). The latter deposit contained two probably residual 14th/1Sth century sherds. Trench 6 (Figure 17) This trench was excavated along the line of a dismantled section ofa dry stone wall, which formed the eastern side of the flight of steps between the Nun’s Walk and St Josephs’ School. When the wall (601) was taken down, an east facing section over 11 metres long that rose in height from 0.8 metres at the southern end to 3 metres at the northern end, was exposed behind the internal wall face. After the section had been cleaned, a series of deposits, which lay in front (east) of the outer Iron Age defences, was revealed. All the accumulated layers (Figure 15) had been eroded from higher up the slope, originating from dump banks on the line of the outer Iron Age rampart. Similar deposits were recorded in Trenches 5 and 7. Period IT: Iron Age Phases I-IV: Middle Iron Age: Sealing the natural limestone was a lens of orange- brown silty clay (607). There was no sign of the Early Iron Age outer ditch previously recorded in Trench 3. Above the earliest recorded deposit was a layer of orange-brown clayey silt with small limestone inclusions (606) containing a sherd of undated pottery. Partially sealing the previous deposit were thin lenses of dark brown silty clay (620) and greenish-yellow clay (619) and (621). These probable hill-wash deposits were sealed beneath a buried soil comprising two deposits, a mid-brown clayey silt with limestone rubble inclusions (604) and a reddish-brown clayey silt (618). Sealing the latter deposit was a succession of probable Iron Age rampart deposits, consisting of orange-brown silty clays (617), (615) and (613), separated by two thin lenses of heat-affected red silty sand (616) and (614). The probable rampart deposits contained inclusions of reddish heat-affected clay, charcoal and limestone rubble. It is likely that all this material originated higher up the slope and was gradually eroded over ume. None of the deposits provided any datable finds. Period IV: Medieval: Sealing the latest Iron Age horizon was a dump layer of mid-brown clayey silt with limestone rubble inclusions (612), possibly dating from the late medieval period. It had built-up against the south- facing elevation of the east/west aligned wall (622), which could be a surviving remnant of the medieval bastion, the ‘Cat Tower’. Above the latter deposit was a dark brown clayey silt (611) sealed by a mid-brown clayey silt (610). There were no datable finds. Period V: Post-medieval: A layer of dark brown topsoil (600) formed the upper part of a deep post-medieval soil containing several sherds of 18th and 19th century pottery. Trench 7 (Figures 18 &19) A partially collapsed section of the town wall located at the top of a 2.5 metre high earth bank was dismantled prior to reconstruction. A trench measuring c.5.5m x 0.9m was excavated on the site of the dismantled wall. An east facing section (Figure 120 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE S N | undergrowth | | | i | 700 N Fig. 18 East-facing section of Trench 7 16) c.2.6m high was exposed behind the internal face of wall (702). Period IT: Iron Age Phases II-VI: Middle Iron Age: Near the southern end of the trench a Im sq. sondage was excavated. At the base was a layer of yellowish-brown clayey silt (715) sealed by a thin lens of charcoal (714). Above this was a 0.28m thick soil horizon of mid-brown clayey silt (713) containing several large limestone masonry slabs from a collapsed or demolished revetment wall (Plate 3). Some stones showed evidence of burning, while the soil contained charcoal and charred cereal grains. Four sherds of Middle Iron Age pottery were retrieved and a date of 400-110 cal. BC was obtained from carbonised barley. The latter deposit was sealed by a brownish- orange clay (712) that lay beneath a 0.12m deep layer of compacted small, crushed fragments of redeposited limestone (711) (Rampart IV), the surface of which was recorded at a height of 81.38m OD. Deposits (711) and (712) seem to constitute a single phase of rampart construction. Above the latter was a subsoil of orange-brown silty clay (710), c.0.lm deep sealed by another former turf line, a 0.16m thick mid-brown clayey silt (707) which contained three sherds of Middle Iron Age pottery and a single undated sherd. The succeeding deposit was rampart layer (709) (Rampart V), a yellowish-orange clay. Above was a lens of brownish-grey clayey silt (708). There were no datable finds from either deposit. Above was another probable rampart deposit (Rampart VI) of yellowish-orange clay and limestone rubble (706) containing reddish heat-affected clay and charcoal. The top of the feature was recorded at a height of 82.14m OD. Above (706) was a subsoil of orange-brown clayey silt (705), itself sealed by a possible turf line of greyish-brown clayey silt (704), containing a single undated sherd. Sealing context (704) a 0.4m thick deposit of orange-brown silty clay (703) yielded no datable finds. The top of the possible rampart feature (Rampart VII) was recorded at a height of 82.84m OD. Period V: Post-medieval: A layer of post-18th century topsoil (700) (Phase IT) sealed the latest in a succession of probable Iron Age rampart deposits. IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 12] E Ww Fig. 20 Limestone masonry from a collapsed or demolished revetment wall on the line of the outer bank (Middle Iron Age) in Trench 7 Trench 8 (Figures 21, 22 & 23) This trench was excavated on the line of the town wall, about 5 metres to the south of Trench 7, where another section of wall had recently collapsed. The trench measured 8.6 m x 0.9m and was later used as a foundation trench for the rebuilt wall. An east facing section (Figure 20) c.2.4m high, was exposed behind the internal face of wall (802). The deposits were almost identical to those in Trench 7. Period IT: Iron Age Phases II-VIIT; Middle Iron Age: Halfway along the trench a lm sq. sondage was excavated. At the base the surface of a probable embankment/rampart (Rampart IV) (Figure 23 & 24) was recorded, composed of small, heavily compacted crushed fragments of redeposited limestone (812). Sealing this was a former turf line (811), a lens of greyish-brown silt with charcoal flecks. It was located at a height of 81.1 metres OD. The earlier rampart was sealed by context (808) (Rampart V), a 0.6m thick deposit of yellowish- orange clay with limestone rubble containing reddish heat-affected clay and charcoal flecks. Two sherds of Middle Iron Age pottery were retrieved. A layer of yellow-brown clayey silt (807), c.0.3m Fig. 21 North-facing section of Trench 8 122 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 23 Surface of Middle Iron Age outer rampart, in Trench 8, composed of crushed fragments of limestone thick sealed this deposit and also contained sherds of Middle Iron Age pottery. That deposit was in turn sealed by a lens of hazel charcoal (809) from which a date of 410-170 cal. BC was obtained. Above was 0) im ———— EEE == Fig. 24 South-facing section of Trench 8 another former turf line consisting of a 0.14m deep brownish-grey clayey silt (806). Above the latter soil was a probable rampart deposit of orange-brown silty clay and limestone rubble (805) (Rampart VII). Above this was the former turf line, a brownish-grey clayey silt (804). The surface of the feature was recorded at a height of 82.82m OD. The latest bank (Rampart VIII) comprised two deposits, a layer of dark brown clay and limestone rubble (803), beneath a lens of yellowish-orange clay (810). The surface of the feature was recorded at a~ height of 83.02m OD. There were no finds. Period V: Post Eighteenth-Century: Above the latest of four possible phases of Iron Age rampart lay 1.4 metres of post-18th century topsoil and masonry rubble (800). IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 123 N wall 904 > —_ ee 905 OO = = Sa 83.04m gos 5914: Ss topsoil 900 wall 902 _ 903_ — —s —_——= Grew 905 S31 LIL, fea 906 nx a 907 IS 907 = 911 a ee cS SS he Oe te Be wall 915 Fig. 25 East-facing section of Trench 9 ~ Trench 9 (Figure 25) _ The trench measured 5.1m x 0.9m and was excavated on the line of the town wall prior to reconstruction of a length of wall collapse. An east facing section (Figure 21) c.2.7 metres high was revealed in the exposed embankment. Undated: The earliest deposit was a layer of orange-brown silty clay (910), which could have formed a rampart. A single undated sherd was recovered. This layer was sealed by a thin greyish-brown silt (909) with inclusions of charcoal and small limestone fragments. It contained one undated sherd. A similar deposit (912) was recorded in the section nearby and represents the same phase of deposition. Sealing deposit (909) was another possible rampart layer composed of orange-brown silty clay (908) but with no datable finds. This deposit also resembled context (911). Both layers, contexts (908) and (911) were sealed by a mid-brown silty soil (907), some 0.2m thick, which is probably a former turf line. A 0.24m deep layer of orange-brown silty clay (906) sealed context (907). It contained charcoal and small fragments of limestone. This possible rampart deposit contained no datable finds; above was a former turf line composed of mid-brown silty soil (905). Period IV: Medieval Phases VI-VII; 12th/13th century: It seems likely that contexts (903), a mixed lens of orange-yellow clay and limestone fragments and (904) a layer of limestone masonry rubble, both of which seal (905), probably derive from stone building debris and ‘trample’ and are contemporary with a phase (Phase VII) of late 12th/early 13th century reconstruction of the town wall recorded on the adjacent library car park site (Duck and Toombs, 1993). Above these deposits a dark brown silty soil (902) contained charcoal and very small fragments of limestone. No finds were recovered. At the base of the trench was a north-south aligned limestone wall (915), approximately 1m wide on the line of the town wall. The date of the 124 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE structure is uncertain but it could be part of the 12th century wall (Phase VI); a date suggested for a wall recorded on the neighbouring library car park site (Duck and Toombs 1993). This probably means that the two adjacent stone walls (913) and (914) are contemporary with wall (915). Period V: Post-medieval: Sealing a layer of yellowish-brown clayey silt and limestone fragments (901) was a dark brown silty loam with fragments of limestone and brick rubble (900). There were no ceramic finds from either context. Interpretation Period I: Late Neolithic/ Early Bronze Age: A small assemblage of unstratified and residual late Neolithic/ early Bronze Age flint tools and flakes represent the earliest (Phase I) activity on the site. The only diagnostic blade was a flint knife (Figure 30), recovered from Trench 2, which is similar to known examples from Beaker burials and domestic sites from the period. No features or deposits were recorded. Period IT: Iron Age: The stratified rampart deposits, revetment walls and ditch, along with contemporary datable ceramic finds and several radiocarbon dates, provide evidence of the development and occupation of the site spanning the period from possibly as early as the eighth century to the early second century BC. Four major phases (Phases I — IV) of activity have been identified within the period. The existence of a hillfort on the site in the Early Iron Age was suggested by the excavation of two significant features; firstly, a steeply cut ditch in Trench 3 cut into the limestone natural, and secondly, the collapsed remains of a possibly contemporary revetment wall in the bottom of the same ditch. This secondary fill was composed of considerable quantities of loose slabs of oolitic limestone, some of which were burnt, and fragments of oak charcoal both of which probably derived from a rampart that was destroyed, at least partially, by fire and then collapsed or demolished into the ditch. The only datable ceramic find recovered from the fill was a single sherd dated to c.800-600 BC. Two radiocarbon dates were obtained from samples of oak charcoal, one from the secondary ditch fill of 800-250 cal. BC and one of 770-400 cal. BC from the soil sealing the ditch fill. There was no evidence of an Iron Age counterscarp bank beyond the outer ditch. Limestone masonry from a revetment wall was also recorded in Trench 7 near the top of the earth bank. But it is likely that this collapsed structure is somewhat later than the structure that was the source of the masonry recorded in the fill of the ditch in Trench 3. Middle Iron Age pottery and a radiocarbon date of 400-110 cal. BC were obtained from the soil horizon sealing the collapsed wall in Trench 7, whereas an Early/Middle Iron Age date was suggested for the masonry in Trench 3. The development of a deep soil horizon over the rubble of the Middle Iron Age revetment wall recorded in Trench 7 could suggest a prolonged period of abandonment of part or all of the hillfort. Alternatively it may indicate a period of neglect and that the site was still occupied, albeit on a reduced scale. It does appear though that the former eastern defences were left unrepaired for a considerable time. Period IIT: Anglo-Saxon: Evidence of West Saxon burh defences from this period was, at best, fragmentary. Probable late Anglo-Saxon defences (Phase VII) were recorded in only one trench (Trench 4), which was located east of the line of the outer Iron Age ditch. It is also possible that several undated deposits in Trenches 2, 5, 7 and 8 may also be attributed to this period. The defences in Trench 4 comprised a single rampart, the surface of which was composed of small fragments of crushed limestone. No postholes were recorded to suggest a palisade. Period IV: Medieval: Evidence of an earthen rampart (Phase VI), possibly dating from the 12th century, was recorded in Trench 5 to the rear (west) of the line of the town wall. This structure may represent a raised intramural walkway that gave access to those guarding the defences. In Trenches 2 and 9 were deposits resembling building debris, possibly from medieval reconstruction work (Phase VII) on the town wall. Such activity probably dates from either the late 12th/early 13th century or from the late 13th century, when there are documentary references to rebuilding work. This appears to tally with ceramic and stratigraphic evidence. Possible structural evidence of late 12th/early 13th century reconstruction of the town wall was also recorded in Trench 2. Part of this later medieval wall was found to abut two other sections of wall, which IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 125 could be remnants of the original early 12th century town wall. Close dating is problematic largely due to limited ceramic finds, the lack of identifiable mortar and the fact that there had been so much subsequent disturbance, largely as a result of successive phases of rebuilding. In Trench 4 evidence was recorded of reconstruction (Phase VIII) in the later medieval period of the late Anglo-Saxon outer earthwork defences east of the town wall. This work may have been contemporary with improvements made to the town wall itself. No evidence of medieval stone fortifications was recorded in the trench and a subsequent examination of the extant town wall suggests that the so-called ‘Cat Tower’ (see above), might lie just north of the school steps. An examination of the mortar found in the wall only confirmed that it had been re-pointed in the late 19th century. Period V: Post-Medieval: No evidence was recorded of features or deposits associated with the 17th century Civil War town defences. It is known, owing to a contemporary illustration, that the defences were extensive, but their slighting in 1646 by Parliamentary forces and several subsequent phases of rebuilding work have erased traces of the circuit between Silver Street and Holloway. EXCAVATION IN 2000 Trench 10 (Figures 26 — 34) ‘The trench was aligned north/south in the school garden, a few metres east of, and parallel with the east end of the school building. It was located some 7 metres north west of the extant town wall and measured 10m x 5m (Figure 27). Period IT; Iron Age Phase I: Early/ Middle Iron Age: The earliest features were three postholes and a ditch (Ditch 1) (Figure 26), all of which truncated the limestone bedrock. Only one of the three postholes (1056) contained finds, a single sherd of Middle Iron Age pottery from fill (1055). There was no trace of the earth rampart, probably indicating that it had been levelled by later activity. The earliest recorded phase of defences would appear to have comprised an 1046 ! | “€ | oaemte Pee cay (1063 | ! 80.39m | ule rw | Peal | | y - | 1054 WS : | Ww 1046 go. 15m_ | S EaNG | 18 \4 — 1 laa se —— t \ ae Ly section Fig.23 Fig.26 Plan showing sondage excavated into the inner Early Tron Age ditch in Trench 10 earthen bank (Rampart 1, Phase I) faced with timber and in front of which was a narrow berm, perhaps 0.5m wide, beyond which was a steeply cut ditch. Phases II-VI: Middle Iron Age: Sealing the fill of the original ditch was a rampart deposit (1053) (Rampart 2). It had been truncated by a recutting of the ditch (1061) (Ditch 2) that was filled with a mid-brown silty clay (1062). Neither the earlier rampart (Rampart 2) nor the ditch fill produced datable finds, although stratigraphically both features appear to date from the Middle Iron Age. Sealing the fill of the recut ditch was a redeposited brownish-yellow silty clay (1051). A small quantity THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 126 OL Youasy ur youp asp uosy ung sauur ays fo aur] ay) ssormp uo0rjdas SuIIM{-1ISV 6 Sb] Wg yooipeq 9b0L W ce] OL youaty fo uoysas Suianf-1svq] LZ Buy OL youasy, fo uonsas surspf-YylwoN 97 ‘su Wg 0 yooupeq [szor] [sgor] e901 [1901 IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 127 0) 2m Fig. 30 Plan of Trench 10 showing remains of the heavily robbed Middle Iron Age rampart wall 1044 of Iron Age pottery was recovered from the latter rampart deposit (Rampart 3). Probably contemporary with the earth bank were two postholes (1049) and (1065). No datable finds were recovered from either feature but it seems likely that both are contemporary with the rampart. A layer of redeposited brownish-yellow silty clay (1037) constituted the remains of another rampart (Rampart 4) and sealed the two postholes (1049) and (1065). This dump deposit contained several sherds of Middle Iron Age pottery. Context (1037) was post-dated by both posthole (1042) and wall (1044). The Middle Iron Age rampart wall (1044) (Rampart 5) (Figure 30 & 31) had been Fig. 31 The inner rampart wall (Middle Iron Age) in Trench 10 heavily robbed in antiquity, at least in the vicinity of Trench 10. Enough survived to determine with a reasonable degree of accuracy its original width and probable height. It was built of limestone slabs and measured 3.72m wide and may well originally have stood some 4 - 5 metres high. Contemporary with this impressive structure was a small posthole set in front of the east-facing wall elevation, perhaps evidence of timber scaffolding during the wall’s construction. The fill of posthole (1042) was sealed by context (1023), whereas wall (1044) was truncated by pit (1045) which contained two fills (1040) and (1047); both produced sherds of Middle Iron Age pottery. Context (1040) also contained briquetage. The secondary fill, context (1040), of pit (1045) was truncated by pit (1039). The fill (1038) contained several sherds of Iron Age pottery. It appears likely that both could have been robber pits — dug to remove stone from the remains of the Middle Iron Age wall (1044) for re-use elsewhere. Fill (1038) was then sealed by late Saxon deposits. Period IIT: Anglo-Saxon Phases I-V: Late Anglo-Saxon Sealing the Iron Age stratigraphy was a buried soil (Phase I) (1023) composed of a very dark grey loam with small, medium and large fragments of limestone. This deposit was tentatively interpreted initially as a sub-Roman ‘dark earth’, but the presence of sherds of late Saxon pottery seems to preclude this. A late Saxon rampart (Rampart 6), context (1020), constructed largely from limestone rubble, sealed deposit (1023). The rampart (Phase I1) lay on a similar alignment to the Middle Iron Age defences and measured at least 1.75m wide and was recorded 128 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 80.84m 1014 Fig. 32 Plan of Trench 10 showing late Anglo-Saxon pits Fig. 33 (right) Plan of Trench 10 showing part of the late Anglo-Saxon rampart at a height of 81.46m OD. Truncating rampart 6 (Figure 31) were pits (1026) and (1028) (Phase III). The primary fill of pit (1026) was a pale yellow-brown clay (1024) containing several sherds of late Saxon pottery. Sealing the latter fill was a thin layer of charcoal (1032), which contained four sherds of undated pottery (possibly Saxo-Norman). Above the latter deposit was a thin layer of mid-brown clayey silt with charcoal flecks (1031). This deposit was sealed by a lens of reddish-orange heat affected clay (1030). The om upper fill of pit (1026), context (1029), was composed IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 129 of yellowish-brown clayey silt with small limestone inclusions. There were no datable finds recovered from the upper three fills. Pit (1028) contained four fills. The primary fill was a 0.32m thick deposit of mid-brown clayey silt (1036). This was sealed by a deposit of greyish-brown clayey silt (1035). In turn, sealing the latter fill was a 30mm lens of charcoal (1034). None of these fills contained datable finds. The upper fill, context (1025), consisted of a greyish-brown clayey silt, some 0.32m thick, which contained several sherds of late Saxon pottery. The latter deposit was sealed by the 12th century (Phase IV) deposit (1014). Period IV: Medieval: Phases I-VIIT; 11th - 14th century The upper fill of the late Saxon pit (1026) was sealed by a mixed horizon of clay and rubble (1015). Pit (1027) truncated the late Saxon rampart (1020) and contained a single fill (Phase I), a deposit of mid brown clayey silt (1022) from which several sherds of Saxo-Norman pottery were recovered. The latter fill was sealed beneath a soil (1014), some 0.34m thick, which consisted of a greyish- brown clayey silt with limestone rubble inclusions. The large ceramic assemblage recovered dates the deposit to the 12th century (Phase II). A linear cut feature (1016) containing wall (1017) (Building 1) (Phase III), truncated the medieval occupation layer (1014). A thin lens of brownish- grey clay and limestone rubble (1015) and a layer of limestone demolition rubble (1012) were truncated by construction cut (1021). The latter construction cut contained a foundation (1018) of the west wall of a stone building (Building 2, Phase IV). Also overlying context (1015) was an 8mm thick mixed lens of cinders, charcoal and heat-affected clay (1010). Presumably the remnants of a bonfire, it contained several unburnt sherds of Saxo-Norman pottery. Part of a laid stone floor (1009) of limestone paving was bedded on the ashy deposit. Contexts (1009) and (1018) were both sealed by buried soil (1068). Other demolished remains of the latter stone building (Figure 29) survived, specifically the base of a 0.75m wide wall (1013) and other foundations (1003). Sealing the latter contexts was a soil horizon (Phase V) (1068). This occupation layer was composed of a dark brownish-grey clayey Fig. 34 (left) Plan of Trench 10 showing the remains of tve medieval structures (Buildings 1 & 2) 130 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE silt with inclusions of charcoal, heat-affected clay and small limestone chippings, but no datable finds. An embankment constructed from a mix of stiff brownish-yellow clay and limestone rubble (1007) overlay soil horizon (1068). The medieval rampart (Rampart 7, Phase VI) contained several sherds of 11th/ 12th century pottery. Sealing the remnants of a cobble surface (1019) was a thin yellowish-brown soil horizon (1005), which covered the southern half of the trench. Several sherds of 12th century pottery were recovered from this deposit. The latter deposit was overlain by a narrow limestone wall (1002) and truncated by a small shallow pit (1004) filled by a dark brown loamy soil (1008). Both the wall and the pit (Phase VII) produced ceramics of 12th/ 13th century date. Sealing the fill of the latter pit and the wall was a layer of yellowish-brown subsoil (1001) (Phase VII), some 0.5m deep, which contained many sherds of 13th/ 14th century pottery. Period V: Post Medieval: Above the latest medieval deposit was the modern dark brown topsoil, context (1000) (Phase IV). Interpretation Period I: Late Neolithic/Early Bronze Age: The earliest (Phase I) activity was represented by a small assemblage of unstratified and residual Late Neolithic/Early Bronze Age flints. Period IT: Iron Age: Activity during the Iron Age period comprised successive ramparts on the inner line of the defences of a multivallate hillfort. The sections disclosed five constructional phases in this period.The earliest recorded defences (Phase I) probably comprised a timber revetted earthen rampart and a ditch, although there was no trace of the rampart itself, which had probably been removed by subsequent activity. Surviving negative features consisted of three postholes, probably associated with a timber revetment, and a ditch (not bottomed), set in front of a berm less than 0.5m wide. One post-hole contained a single sherd of pottery dated to c.400 BC. The next period of activity was represented by the remnants ofa second earthen rampart and a recut ditch. The rampart deposit partially sealed the fill of the earlier ditch and associated post-holes but contained no datable finds. A third earthen rampart superseded the latter defensive work. This clay dump sealed the fill of the earlier recut ditch. In addition, the rampart was cut by two postholes, possibly indicating a contemporary timber palisade. Three sherds of Iron Age pottery were retrieved from the rampart deposit. The next phase involved the construction of a fourth rampart, probably almost identical to the two partially surviving earlier structures. Several sherds of Iron Age pottery were retrieved from both the rampart deposit and the occupation layer that sealed it. A substantial 3.72m wide limestone wall replaced the latest earthen rampart and was set in a construction trench cut into the truncated remains of the previous embankment. Several sherds of Middle Iron Age pottery were retrieved from within the wall itself. Contemporary with this impressive structure was a single posthole, adjacent to the front (east-facing) elevation of the wall, which could indicate that timber scaffolding was used during construction. Period IIT: Anglo-Saxon: Two successive occupation layers were recorded, the earlier of which produced pottery indicative of Late Anglo-Saxon activity. These dark brownish-grey clayey silts sealed the Iron Age stratigraphy. The latter of these occupation layers was, in turn, sealed by a clay and limestone rubble rampart (Phase II), or collapsed wall, that lay on a similar alignment to the Middle Iron Age hillfort defences. This defensive work probably formed part of the 10th or early 11th century defences of the West Saxon burh, and was subsequently truncated by two pits, the fills of which contained Late Saxon pottery. Period IV: Medieval: The earliest medieval contexts comprised a pit and two occupation layers of greyish-brown soils that contained much late 11th/ 12th century pottery. The latter of these occupation layers was truncated by a construction cut associated with the south and west walls of a stone building (Building 1, Phase III) of unknown function in the late 11th/ early 12th century. That structure was soon replaced with a more substantial stone building (Building 2, Phase IV). The latter building was soon demolished and its foundations sealed beneath an occupation layer of dark brownish-grey soil. That deposit, which contained no datable finds, was in turn sealed beneath a rampart (Rampart 7, Phase VI) composed IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 131 of clay and limestone rubble that contained 11th/12th century pottery. Because of their close proximity and the fact that the two deposits are almost identical it is very likely that this feature is the same embankment as one recorded in Trench 5 (Rampart XI). A thin soil (former turf line) was subsequently recorded across the northern half of the trench that developed after the construction of the embankment. A small pit containing four sherds of late 11th/12th century pottery was recorded truncating the buried soil. The bank or rampart was probably intended as araised intramural walkway contemporary with the town wall, documented as having been built in the early twelfth century by Bishop Roger. To prevent slumping of the clay and limestone rubble bank a retaining wall was later built parallel to the rampart, at the base of the slope. 12th/13th century pottery was retrieved from among the remnants of this wall. The rampart and retaining wall were sealed beneath a layer of yellowish-brown subsoil (Phase VIII), some 0.50m deep containing many sherds of 13th/14th century pottery. Period V: Post-Medieval: There were no traces of mid-17th century Civil War defences probably due to slighting of the walls by the Parliamentarians in 1646, and subsequent levelling/ clearance work in the 1870s during the construction of the nearby railway line. SPECIALIST REPORTS CERAMIC FINDS by Lisa Brown and Rod Burchill Introduction The pottery assemblage from the two sites gives a combined total of 1,366 sherds weighing 15,470gms. Preliminary assessment by R. Burchill concluded that whilst demonstrating an ability to provide a chronology for the general development of the sites, the assemblage was limited in its research potential and only the Later Prehistoric pottery warranted further analysis (see Brown below). The Later Prehistoric Pottery by Lisa Brown Introduction The excavations produced 104 sherds of later prehistoric pottery from six of the ten trenches (none was present in trenches 1, 4, 6 and 9). The majority of sherds came from trenches 5 (46) and 10 (29). The condition of the assemblage was generally very fragmentary and 30% of sherds were heavily abraded. Much of the earliest Iron Age pottery had been redeposited in Middie Iron Age or later deposits. Methodology All sherds were examined using a hand lens and binocular microscope at x10 and x20 power and assigned to a fabric category based on the dominant inclusion present. A fabric record sheet has been compiled for each fabric category (available in archive). All featured sherds are illustrated. The entire collection was recorded following the current guidelines for analysis of later prehistoric pottery (PCRG 1992). Fabrics Eight fabric types within three fabric groups were defined. Due to the limited scale of the intervention and the modest size and poor condition of the assemblage none of the fabrics has been subjected to petrological analysis but more substantial assemblages of similar type recovered from future work in the region would benefit from such analysis. The most common fabric group, representing 65% of the total, includes five varieties of calcareous inclusions. A smaller group of shell-tempered wares forms 30% of the assemblage and the remaining 5% is made up of a mere three sherds of predominantly sand-tempered fabric. A further 73 fragments of fired clay were identified as briquetage. The majority of fabrics reflect the underlying geology of the site which, located on the south- eastern edge of the Cotswold Hills, is Upper Jurassic rubbly limestone (cornbrash) and Middle Jurassic shelly oolite clay. It is highly likely that fabrics S1 and C2-CS5 were produced in the immediate locality or in a geologically similar area. Similar fabric types have been recognised within assemblages in the Bath area including Batheaston and Mells (Morris, forthcoming and Brown, forthcoming) and to the north in the region around Gloucester (Vince and Goudge forthcoming). Two burnished saucepan pot sherds in glauconitic sandy ware Q] resemble vessels known throughout Wiltshire and Hampshire which are likely to have originated from Cretaceous Upper Greensand deposits. A source 10 km to the 132 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE west of Salisbury at Compton Chamberlayne has been proposed for similar wares from a number of sites (Williams 1984). The presence of small fragments of shell in this fabric suggests a Jurassic source but shell-bearing and Upper Greensand deposits often occur in the same area (Chatwin 1960). Calcite-tempered ware Cl could originate from a similar source since calcite matrices are known in Greensand, but alternative sources to the south of Bristol and in South Wales are also possible (Peacock 1969). Calcareous group Cl Moderate to common quantity of moderately sorted sub-angular crushed calcite under 2mm in size along with sparser inclusions of calcite limestone rock of similar size. The inclusions are set ina fine, silt grade sandy clay which may be glauconitic. 4 sherds. C2 Abundant ill-assorted inclusions of limestone rock, rich with fossil shell and oolites in a non- sanded clay matrix. Shell pieces measured up to 8 mm, angular grey limestone fragments up to 3mm. 31 sherds. C3 Lightly sanded clay containing abundant finely crushed fossiliferous limestone rock including rare oolites and platey shell. Inclusions range from < 2 to 5mm in size. 30 sherds. C4 Fine, silt grade slightly micaceous sandy clay containing sparse small inclusions of limestone rock, including fossil shell and oolites. 1 sherd. C5 Finely sanded micaceous clay (slightly coarser than C4) with moderate to common oolite inclusions and rarer additional fossil limestone rock fragments < 2mm. | sherd. Shelly group S1 Smooth, hard, lightly sanded slightly micaceous clay, with a hard, almost brittle quality, containing moderate to common quantity of shell fragments up to 3mm in size, rarely up to 8mm. The shell is probably fossil shell. 32 sherds. Sandy group Q1 Fine glauconitic sandy fabric containing an abundant quantity of subrounded quartz grains 2mm in cross-section were considered for species identification. The charcoal was prepared for examination using standard methods (Gale and Cutler 2000). The wood structures were examined using a Nikon Labophot-2 microscope at magnifications up to x400 and matched to reference slides. When possible, the maturity of the wood was assessed (1.e. heartwood/ sapwood). Results The charcoal analysis is summarised in Table 6 and discussed below. Group names are given when anatomical differences between related genera are too slight to allow secure identification to genus level. These include members of the Pomoideae (Crataegus, Malus, Pyrus and Sorbus) and Salicaceae (Salix and Populus). Where a genus is represented by a single species in the British flora this is named as the most likely origin of the wood, given the provenance and period, but it should be noted that it is rarely possible to name individual species from wood features, and exotic species of trees and shrubs were introduced to Britain from an early period (Godwin 1956; Mitchell 1974). Classification follows that of Flora Europaea (Tutin, Heywood et al 1964-80). The anatomical structure of the charcoal was consistent with the following taxa or groups of taxa: Aceraceae. Acer campestre L., field maple Aquifoliaceae. Ilex aquifolium L., holly Betulaceae. Alnus glutinosa (L.) Gaertner, European alder Corylaceae. Corylus avellana L., hazel Fagaceae. Fagus sylvatica L., beech; Quercus spp., oak Oleaceae. Fraxinus excelsior L., ash Rosaceae. Subfamilies: Pomoideae which includes Crataegus spp., hawthorn; Malus sp., apple; Pyrus sp., pear; Sorbus spp., rowan, service tree and whitebeam. These taxa are anatomically similar; one or more taxa may be represented in the charcoal. Prunoideae: Prunus spinosa L., blackthorn. Salicaceae. Salix spp., willow, and Populus spp., poplar. In most respects these taxa are anatomically similar. The ray type sometimes allows the taxon to be named; however this feature is not always a reliable indicator and has not been used in this instance. Tiliaceae. Tilia sp., lime Period II: Iron Age A radiocarbon date of 2460+50 BP (MALM 3) was obtained from handpicked oak (Quercus sp.) from Trench 3, sample 12 (304), and a somewhat similar date of 2370+50 BP (MALM 4) from sample 20 (307), also oak. Pottery dated to 800-600 BC was sealed in a deposit overlying context 307. These results indicate Early to Middle Iron Age origins for the major defences, which included ramparts and ditches with timber palisades. Charcoal deposits were examined from Trenches 3 and 10. Trench 3 The trench consisted of layers of brownish red and yellow silty clays underlying the topsoil. Charcoal was abundant in contexts 304, 305 and 306 (a 25% subsample was examined from sample 18 (306)) and appeared to represent structural debris from the destruction and burning of the ramparts. Associated debris included burnt limestone masonry and daub. IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 147 Table 6: Charcoal from Iron Age, Anglo-Saxon and medieval contexts (Trenches 3 and 10) Key: h = heartwood and unknown maturity; r = roundwood (diameter <20mm); s = sapwood (diameter unknown) Trench 3 Early Iron Age eit LM SE] RM a ee) eee ee eee ee ee eee 305 | Fill of ditch 1 eri lesion 306 | Fill of ditch 1 Pea Been aoe aeons ee [595 nc a a DS (a Trench 10 Early Iron Age [ie i a a a a Middle Iron Age Upper fill of pit Ss 1040 | 1045 pen Messer Clayey silt between [4 Wel stones of wall 1044 Bae 2a eae [rete ie Deposit sealing 1041 layer 1037. Predates inner rampart wall 1044 Layer of re-depos- ited silt, tuncated 1037 | prior to construc- tion of inner rampart wall 1044 Fill of posthole, which comprised 12) 1050 the remnants of umber post Primary fill of pit wor SS sin apace Late Anglo-Saxon — 12s |g ee es eee er 1023 | layer truncated by 24 2h, 8r several A/S pits Pit fill - thin layer of charcoal forming : 1032 secondary fill of pit 29h; 4s 3 1026 - lens Occupational layer sealed beneath early 3} 1014 | medieval Build- ing | Charcoal lens and fired clay depos- 1010 | ited as sub-base for stone floor of Building | Fill of small 1008 rubbish pit L004; contained burnt debris 55h, 8s/r 148 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Discoloured snail shells were indicative of in situ burning. Charred plant macrofossils (cereal grains and weed seeds, Jones, this paper) and other burnt organic material were sparse. Large fragments of handpicked charcoal (sample 12) from context 304, a layer of burnt destruction debris, was thought to be from the charred remains of a post, perhaps from the palisade. Although a small piece of oak (Quercus sp.) heartwood was present, the bulk of this charcoal was identified as lime (7ilia sp.), probably at least 80mm in radius (when charred). The bulk soil sample 19, from the same context also included numerous large fragments from wide roundwood or poles of maple (Acer campestre) and oak (Quercus sp.). The association of burnt fragments of large wood or timber and masonry rubble strongly supports the suggested origin - as from wooden building components. While oak (particularly) and maple would have offered strong, durable timbers for the palisade construction, the use of lime, which is soft and perishable, seems unlikely in an outdoor context and was, perhaps, used for some other purpose. Burnt debris from the destruction layer also filled the steeply cut ditch 1, which contained the burnt layer 304. Charcoal from contexts 305 and 306 also comprised large fragments from oak (Quercus sp.), field maple (Acer campestre) and lime (Tilia sp.) (Table 6). Oak charcoal from the secondary fill (307) of ditch 1 (identified by Vanessa Straker) provided the radiocarbon date indicated above. Trench 10 Small fragments of charcoal from context 1057, the fill of a posthole dated to the early phase of this period, were mostly identified as oak (Quercus sp.) heartwood, but also included alder (Alnus glutinosa), blackthorn (Prunus spinosa) and a member of the hawthorn/ Sorbus group (Pomoideae). Charred plant macrofossils included cereal grains and a sloe stone (Prunus spinosa) (Jones, this volume). Charcoal was examined from both the primary and upper fills of pit 1045. The basal fill (context 1047) contained a mixture of oak (Quercus sp.) charcoal and an abundance of fired clay. By association this could implicate the charcoal as fuel debris from the firing process but, in the absence of secure evidence, this suggestion remains speculative. Oak (Quercus sp.) was also recorded in the upper fill (context 1040) together with blackthorn (Prunus spinosa) and willow (Salix sp.) or poplar (Populus sp.). Charcoal recovered from context 1050, the fill of a posthole, was recorded as the remains of a timber post. This interpretation was supported by the identification of a quantity of oak (Quercus sp.) . heartwood and the absence of other charred refuse, e.g. cereal and weed seeds, which could implicate fuel debris. Sample 11, from a layer of redeposited silt (context 1037) truncated prior to the construction of the inner rampart wall 1044, consisted of oak (Quercus sp.) heartwood and sapwood, probably from fairly wide roundwood. Other plant remains were sparse (Jones, this volume). Context 1041 also predated the construction of the wall 1044 and related to a deposit sealing layer 1037. Associated charcoal included ash (Fraxinus excelsior), beech (Fagus sylvatica), blackthorn (Prunus spinosa), a member of the hawthorn/ Sorbus group (Pomoideae) and oak (Quercus sp.). A relatively large volume of charcoal was recovered from context 1067, the clayey silt from between the stones of the rampart wall 1044. The charcoal here was predominantly oak (Quercus sp.) and included large fragments measuring up to 25mm radially, suggesting an origin from fairly substantial poles, trunks or branches. In addition to oak, small quantities of blackthorn (Prunus spinosa) and willow (Salix sp.) or poplar (Populus sp.) were recorded. The origin of the charcoal in this sequence of contexts is difficult to interpret especially when occurring in layers, e.g. contexts 1037 and 1041. These deposits may represent scatters or accumulations of spent fuel (particularly context 1041, which included a wide range of species), although the remains of burnt structural components could also have given rise to these deposits. Similar parameters apply to context 1067. Period III: Anglo-Saxon Further defences were constructed throughout the Late Anglo-Saxon period. Charcoal was examined from an occupation layer and two pits. Trench 10 Context 1023, from an occupation layer truncated by several Anglo-Saxon pits, produced very friable charcoal from which ash (Fraxinus excelsior), oak (Quercus sp.) and blackthorn (Prunus spinosa) were recorded. Charred cereal grains were also present. Charcoal was also examined from the fills of two pits. Taxa identified from context 1032, from the secondary fill of pit 1026, included oak (Quercus sp.), blackthorn (Prunus spinosa), willow (Salix sp.) or poplar (Populus sp.), the hawthorn/ Sorbus group (Pomoideae) and holly (lex aquifolium). Charcoal was abundant both in the upper fill (context 1025) of pit IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 149 1028 anda lens of charcoal (context 1034) within the same pit. Both samples predominantly consisted of oak (Quercus sp.) but also included a member of the hawthorn/ Sorbus group (Pomoideae). Charred cereal grains, weed seeds, sloe stones (Prunus spinosa) and hazel nutshell (Corylus avellana) were also recorded from pit 1028 (Jones, this paper). Period IV: Medieval Trench 10 A relatively wide range of taxa was identified from an occupation layer (context 1014) sealed beneath medieval Building 1. These included oak (Quercus sp.), blackthorn (Prunus spinosa), beech (Fagus sylvatica), willow (Salix sp.) or poplar (Populus sp.), the hawthorn/ Sorbus group (Pomoideae), ivy (Hedera helix) and probably hazel (Corylus avellana). Charcoal (context 1010) included in the thin lens of material deposited as a sub-base for the stone floor of Building 1, consisted of oak (Quercus sp.), willow (Salix sp.) or poplar (Populus sp.), blackthorn (Prunus spinosa) and probably beech (Fagus sylvatica). Although the origin of the charcoal is unknown, the presence of charred cereal grains and weed seeds suggested fuel debris. The fill (context 1008) ofa small rubbish pit 1004 contained a large amount of burnt debris but mostly charcoal — almost certainly from fuel residues. The charcoal was predominantly oak (Quercus sp.), but also contained small amounts of blackthorn (Prunus spinosa), beech (Fagus sylvatica) and hazel (Corylus avellana). ‘Discussion The excavation of Trenches 3 and 10, both close to the town wall, exposed stratified deposits dating from the Early/Middle Iron Age to the post-medieval period. Until the Late Anglo-Saxon period the ramparts, of which the earliest parts date from the Early Iron Age, formed the major defensive system of the town. Associated postholes suggest that timber palisades were constructed. The earliest contexts from which charcoal was examined related to debris in the burnt layer within ditch | (period IT) of Trench 3, thought to represent the destruction of the ramparts and palisade. It was not clear from the charcoal whether it derived from structural clements or some other source but the frequency of large fragments from wide roundwood or poles from a relatively narrow range of taxa tended to suggest artefactual or structural use as a distinct possibility. In addition, its close association with fallen, burnt masonry also inferred structural use. However, while both oak (Quercus sp.) and field maple (Acer campestre) would have offered timber of suitable strength and durability for outdoor use (e.g. palisades), lime (Tilia sp.) would have quickly perished. It may be significant that neither field maple nor lime was identified from contexts in Trench 10, which with the possible exception of context 1050, probably derived from fuel debris. Charcoal also occurred in numerous contexts in Trench 10, usually in association with charred cereal grains and weed seeds (Table 6). When deposited in pits, e.g. pit 1045, 1026 and 1032, it is probable that the charcoal derived from fuel debris and, similarly, when scattered across occupation layers, e.g. contexts 1023 and 1014. The origin of the charcoal from the Iron Age layers 1037, 1041 and from between stones in wall 1044 (context 1067) may also be attributable to fuel debris or, possibly (although burnt soils were not recorded) deposits from burnt structural timbers and wooden components burnt im situ. Charred oak (Quercus sp.) fragments recovered from an Iron Age posthole (context 1050), however, could be interpreted as the remains of an oak post. Apart from fired clay in the Period II Iron Age pit 1045, which may be indicative of pottery-making or kiln refuse, there was no evidence of industrial activity and it must be assumed that most fuel debris resulted either from domestic hearths or from burning rubbish. Oak (Quercus sp.) was present, sometimes abundantly, in every sample examined while other taxa occurred only sporadically. The latter included blackthorn (Prunus spinosa), the hawthorn/ Sorbus group (Pomoideae), willow (Salix sp.) and/ or poplar (Populus sp.), ash (Fraxinus excelsior), beech (Fagus sylvatica), hazel (Corylus avellana), alder (Alnus glutinosa), ivy (Hedera helix) and holly Ulex aquifolium). Fuel and woodland resources Oak appears to have supplied the main fuel requirements from the Iron Age until the medieval period. The charcoal was too comminuted to assess the type of fuel used (i.e. narrow roundwood, poles or cordwood), although the high percentage of oak heartwood suggested that much of it may have been fairly wide roundwood (e.g. diameter >20mm) or cordwood. Oak firewood provides a long-lasting, high-energy heat source (Edlin 1949; Porter 1990) and the selection of fuel was clearly biased in favour of oak. Other species may have been used as and when available and/or as kindling. It is interesting 150 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE that neither maple (Acer campestre) nor lime (Tilia sp.) occurred in the fuel debris, although both were present in the layer of destruction debris in Trench 3. Although lime does not make top-grade firewood, evidence of its use in domestic ovens in the 14th — 15th century was recorded from dwellings in Bristol (5 Welsh Back) (Gale, unpub); ancient coppice stools of lime and other species which provided local firewood are still extant in woodland in the Avon Gorge and probably date from the medieval period (Lovatt 1987). The presence of ivy in the Anglo-Saxon context 1014 may be incidental, perhaps introduced as fuel while still adhering to its host tree (e.g. oak). Records of ivy in fuel deposits, however, are relatively rare and it is worth considering alternative sources of origin, since small pieces of stem or wood could have found their way into the hearth through various routes. For example, as bare stems (kindling) from browse - ivy foliage is known to have been gathered and used in Europe since prehistoric times (Troels-Smith 1960); more traditional uses for the wood (from which damaged or discarded pieces may have been thrown onto the fire) have included rope-making (from the twisted stems), wood carving (Usher 1974) and, perhaps more significantly, as firesticks - Theophrastus (5.9) states that ivy was especially effective. In addition, ivy has long associations with folk belief and ritual (Grigson 1958). With the exception of willow (Salix sp.), poplar (Populus sp.) and alder (Alnus glutinosa), the species named above provide efficient fuel (Edlin 1949; Porter 1990). The apparent paucity of hazel (Corylus avellana) in the fuel residues from Iron Age and Anglo-Saxon contexts is curious, since it clearly grew at the site (identified from both charcoal and plant macrofossils) and would have provided an excellent source of fuel, especially through the use of fast- grown coppice stems. Traditionally, coppiced hazel has provided the mainstay of domestic firewood (Edlin 1949; Anon 1956) but it was also valued for a wide range of other uses including wattle-work, poles and charcoal-making. It could, therefore, be argued that hazel stocks were reserved for other uses. Its apparent paucity in the charcoal deposits at St Joseph’s School parallels the findings from Roman, early medieval and post-medieval deposits at Shapwick in the Somerset Levels (Gale, unpub). Although the large volumes of charcoal recovered from Shapwick indicated the predominant use of oak (Quercus sp.), a wide range of other species were also identified including field maple (Acer campestre), alder (Alnus glutinosa), birch (Betula sp.), hazel (Corylus avellana), beech (Fagus sylvatica), . alder buckthorn (Frangula alnus), ash (Fraxinus excelsior), the hawthorn/ Sorbus group (Pomoideae), willow (Salix sp.) and/ or poplar (Populus sp.), elm (Ulmus sp.), ivy (Hedera helix) and/or elder (Sambucus sp.), dogwood (Cornus) or Viburnum and probably lime (Tilia sp.). While there are clearly differences between the largely wetland environment in the Somerset Levels and the landscape around Malmesbury, firewood at Shapwick would have been drawn mostly from the higher ground on the edge of the Polden Hills. Numerous excavations in that region have recorded the wide distribution and use of coppiced hazel for constructing trackways and for other activities (Coles and Orme 1980), which could infer a greater relevance for these purposes than as fuel. The extensive construction of defensive ramparts topped with wooden palisades during the Iron Age suggests that the settlement at Malmesbury was of some significant size. Local woodlands would have been the main source for many of the materials essential for the maintenance and stability of the settlement, e.g. fuel (firewood and charcoal), timber and wood for construction, hurdle-making, basketry, tanbark, wood tars and, probably, leaf fodder and pannage. The initial building of the defences and palisades would have required substantial and durable timbers and it is probable that oak provided the bulk of this (the wide distribution of oak in the Malmesbury region was confirmed by its dominance in the fuel residues). But even with the use of oak heartwood, it would have been necessary to replace these posts relatively frequently. Evidence from the destruction layer and ditch fill (Trench 3) suggested that field maple (Acer campestre) and lime (Tilia sp.) may also have been used for construction in this area of the defences (see discussion above). Brash, trimmings and cordwood from the conversion of timber trees or poles would have supplied fuel and small wood. Hedge-cuttings could have provided a further source of fuel and may account for the slight frequency in the charcoal deposits of blackthorn (Prunus spinosa) and, possibly, hawthorn (Crataegus sp.). Historically, both species were favoured for hedgerows, especially for enclosing livestock (Edlin 1949). Environmental evidence The site was based on a natural hill partially enclosed by the River Avon and its tributary, thereby affording a range of ecological zones in the IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 151 immediate environment. Although evidence from the charcoal analysis is undoubtedly biased towards the preferential selection of oak (Quercus sp.) for economic uses (mostly fuel, in this instance), it is probable that oak formed the dominant woodland cover in the region. Other large woodland trees identified from the charcoal included beech (Fagus sylvatica), ash (Fraxinus excelsior), field maple (Acer campestre) and lime (Tilia sp.) all of which probably occurred within oak woodland or in discrete stands. These woodlands may also have supported hazel (Corylus avellana) and holly (Ilex aquifolium). Blackthorn (Prunus spinosa) and members of the Pomoideae such as hawthorn (Crataegus sp.) would be more likely to have colonised woodland margin, scrub and cleared areas. Wetter soils within woodland or along the river-banks or low- lying meadows would have supported alder (Alnus glutinosa) and willow (Salix sp.), with poplar (Populus sp.) in damp meadows. While a significant number of woody species has been recorded from the charcoal, it is unlikely that these represent the entire range of woody taxa growing in the vicinity of the site. The edaphic and climatic conditions would have been favourable for species such as elm (U/mus sp.), elder (Sambucus sp.) and numerous other shrubs. The charcoal was too comminuted to establish the use of coppiced or pollarded wood, but in view of the large-scale construction work during the early part of the Iron Age and the ongoing demand for timber, small wood and fuel, it is probable that woodlands were managed from an early period. If coppicing was not already established at the site -in the Early Iron Age, the reduction or felling of immature trees for the palisades would almost certainly have initiated coppice type growth (fast- grown stems) from the remaining root stock. By the Saxon period woodland management probably came within the remit of the abbey. Apart from ivy and blackthorn all the species identified respond well to coppicing (Rackham 1976). Conclusion This report includes the analysis of charcoal from Trenches 3 and 10 from Iron Age, Anglo-Saxon and medieval features associated with the defensive walls of the setthement. The charcoal has been attributed as mostly from hearths although Iron Age structural remains were implicated in Trench 3 (oak (Quercus sp.), field maple (Acer campestre) and lime (Tilia sp.)) and possibly an oak post from the same period in Trench 10. Fuel supplies throughout these periods consisted predominantly of oak (Quercus sp.), although other species identified included blackthorn (Prunus spinosa), the hawthorn/ Sorbus group (Pomoideae), willow (Salix sp.) and/ or poplar (Populus sp.), ash (Fraxinus excelsior), beech (Fagus sylvatica), hazel (Corylus avellana), alder (Alnus glutinosa), ivy (Hedera helix) and holly (lex aquifolium). Despite lack of evidence from the charcoal, it was argued that woodland would have been managed from an early period to provision the settlement. CHARRED PLANT REMAINS by Fulie Fones Introduction Bulk samples were taken from Trench 10 deposits dating from the Early Iron Age, through to Late Anglo-Saxon and medieval phases. Following recommendations made as part of the environmental assessment of the charred plant remains, eleven samples were chosen for full analysis. Methodology Samples processed as part of the assessment programme varied in size from 3 to 56 litres. They were wet sieved in the School of Geographical Science at Bristol University to a mesh size of 0.5mm for the residue and 250 microns for the float. Both the residue and float were dried before examination. The floats were sorted and all plant remains extracted and counted. Full details are shown on Table 7. Grain and chaff determinations are based on Jacomet (1987) and nomenclature and habitat information for the weeds are based on Stace (1991). The bulk of the plant remains were preserved by charring, although a few mineralised seeds and fly pupae occurred in two samples from Late Anglo- Saxon (context 1025) and medieval (context 1008) deposits. Preservation of the charred remains was variable. While many of the cereal grains could be determined as wheat (7iiticum sp), with both hulled (Triticum dicoccum/spelta threshing wheat (7riticum aestivum s.l. - bread wheat - emmer/spelt) and free- type) identified from the better preserved grains, many were in a poor condition, being distorted and with pitted and abraded surfaces. While in some cases these were still recognisable as wheat, many were too deformed, which accounts for the 152 Table 7: Charred and mineralised plant remains THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Early Anglo-Saxon Iron Medieval Age Context No | 1057 | 1041 1067__| 1040 1023 | 1032 | 1025 | 1034 | 1014 1010 1008 Sample No | 14 10 9 8 7 19 6 20 3 2 it Sample size (litres) | 22 24 14 18 22 4 20 3 45 56 16 Size of residue (kg) | 6.7 Sl 4.65 | 49 ue) 1 ESA 0.75 | 10.2 14.2 3.7 - Size of float (ml) | 25 60 100 10 22 46 185 150 29 10 70 CHARRED PLANT REMAINS Grain Avena sp Oat 4 2 if 20 25 3 9 9 30 # Hordeum sp Barley 10 20 26 18 26 15 10 1 7 12 # Hordeum sp (Hulled) Barley {3 3 VW 8 i HES) # | Hordeum sp (Hulled/Straight)_| Barley 2 4 [ 1 B # Hordeum sp (tail grain/hulled) | Barley i Me # Triticum dicoccum spelta Hulled wheat 141 25 28 50 4 # Triticum aestivum s1 Free-threshing wheat 6 5) 292 231 42 51 27 11S # cf Triticum sp Wheat 97 13 9 26 13 76 79 10 31 9 47 # Cereal indet 73 18 29 39 24 99 98 20 23 16 26 # Total: | 325 85 99 149 WAL 503 446 76 129 68 230 Chaff | Hordeum sp Barley 1 1 # (rachis internode base) nf ie Triticum spelta (glume base) _| Spelt wheat 1 2 # Triticum dicoccum spelta Hulled wheat i 1 \2 # (glume base) Triticum sp Free threshing wheat 1 88 28 13 1 1 # (tough rachis internode) Triticum sp Wheat 2 # (rachis internode base) ees. Triticum sp (glume base - ¢ f. Bread/club wheat 2 1 # hexaploid type) L HE Cereal indet Culm frags 2 | # Cereal indet Embryo area 2 5) # Total: | 1 0 1 3 3 94 36 13 0 1 1 Context No | 1057_ | 1041_| 1067 | 1040 1023 | 1032 | 1025 | 1034 | 1014 1010 1008 Sample No | 14 10 9 8 ul 19 6 20 3 2 1 Weeds Habitat RANUNCULACEAE Ranunculus Buttercup 1 DG acris repens/bulbosus pon FAGACEAE Quercus spp (bud) Oak 2 _ | HSW BETULACEAE | Corylus avellana L Hazel If lf If 8f 8f lf 6f 4f lf HSW (nut frags) CHENOPODIACEAE Alriplex spp Orache 1 1 If 1 4 CDn Chenopodium album L. | Fat-hen ] 1 CDn POLYGONACEAE il Fallopta convolvulus Black-bindweed 1 1 | | cD (L.)A.Love Persicaria lapathifolia Pale Persicaria i [ if Cdow (L.)Gray | Rumex acetosella L Sheep's Sorrel 1 Ho,CG, a, i sandy Rumex spp Dock 2 1 3 1 1 5 1 DG MALVACEAE Malva spp Mallow 1 BRASSICACEAE a) Brassica Sinapis spp Mustard/Rape/ 4 1 1 CD# Cole etc | se Raphanus raphanistrum ssp | Wild Radish ale 2f cD raphanistrum (pod frags) ROSACEAE Prunus spinosa L Sloe 2f il 1 HSW FABACEAE Lathyrus Vicia spp Pea/Vetch 1 1 1 1 2 3 2f 2+6f DG Trifolium Medicago spp Clover/Medick 3 | in 1 4 DG IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 Table 7: (continued) 153 fo Context No | 1057 | 1041 | 1067 | 1040 [1023 | 1032 | 1025 | 1034 | 1014 CSE TY a (a rT 19/6 20 | 3 Vicia faba L. -Cevieore Bete} | ff SOLANACEAE }--—|- | fames Hyoscyamus niger L. Henbane [aT | eae) [i aa |e VERBENACEAE SS Bara an rae 2 aA ea a LAMIACEAE mar Ph A Prunella vulgaris L. [Selfheal si [tera] IZUSNTAGINACEARSaS ieee oan i pclaoge leneolarat--__{ Ribwor Plantain _}_f____f {| | SCROPHULARIACEAE | Odontites/Enuphrasia spp Bartsia/Eyebright eS 1 3 cD Veronica beccabunga L. Brooklime a aaa) 1 BMPR RUBIACEAE me ! | CHSo VALERIANACE AE [cee Valerianella dentata Narrow-fruited a aa 1 cD L.)Pollich Cornsalad aaa | ASTERACEAE ee ee Anthemis cotula L. | Stinking Chamomile | | | if 7 15 2 36 3 CDh | Centaurea spp | Knapweed | TT ! 1 1 CDG Tripleurospermum inodorum | Scentless Mayweed (es RE) i I 1 cD _)Schultz-Bip (RS RE Cultivated/Arable D: Disturbed E Heath/Moor, G: Grassland H > Hedgerow M> Marsh P. Ponds, ditches - stagnant/slow flowing water R: Rivers, streams S: Scrub W: Woodland a acidic c: calcareous d° dry soils bh: heavy soils no: nitrogenrich soils o: open habitats p: phosphate rich soils ww. wet/damp soils # cultivated planVof economic importance fe fragments 154 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE relatively high proportion of grains determined as unidentifiable (Cereal indet). The poor preservation is thought to relate to moisture content of the grain and high burning temperatures, rather than post- depositional erosion, although in some contexts, such as occupation layers, increased grain fragmentation is seen when compared to material in sealed features such as pits. The same is true for barley (Hordeum sp) and oat (Avena sp) grains, although occasional grains were very well preserved. The cereal chaff, especially the wheat rachis internodes were mostly too fragmentary to determine which form was present, but in contrast many of the weed seeds were in good condition. The charred plant remains The bulk of the charred cereal remains from Malmesbury came from wheat and although much of the grain from all phases was in poor condition, some examples were well preserved. Most of the grain from the earlier phases (Early and Middle Iron Age) was identified as glume wheats from the characteristic slender, parallel-sided grains showing the low, rounded dorsal profile characteristic of hulled wheat. The presence of two well-preserved glumes bases from one of the Iron Age pit fills (1040), showing a prominent primary keel and strong venation pattern on the surface of the glume confirms the presence of spelt (Ziiticum spelta). The other glume bases from the earlier phases were poorly preserved and have been identified as Triticum dicoccum/spelta (emmer/spelt) type. Much of the well-preserved grain from the later phases, showed the short rounded profile of free-threshing bread wheat, associated with tough rachis internodes also characteristic of free-threshing wheat. There were only three well preserved examples of hexaploid (Triticum aestivum/compactum - bread/club wheat) glume bases, as described by Moffett (1991), from the Late Anglo-Saxon period. Remains of barley were generally less frequent than wheat and preservation of grains was poor. Many were deformed or the surface had been lost, although some hulled forms were noted from all phases. The presence of a few straight grains suggests that both the two and six-row form may have occurred. Only single examples of barley rachis internode bases were found from the Late Anglo- Saxon period (contexts 1032 and 1034). The few poorly preserved oat grains are likely to have been present as a weed contaminant to the cereals. There was no supporting chaff to suggest whether these were the wild or cultivated variety. Brome caryopses similarly occurred in all phases at low abundance and are likely to have been from crop weeds. Evidence of legumes was present in low abundance, in all phases, although there was only a single celtic/horse bean (Vicia faba) from a medieval pit fill (1008). Preservation of legumes in a charred form seems to be less common than cereals, as they are less likely to come into contact with fire in the preparations for cooking. Unfortunately preservation of all the remaining legumes was poor, being in a fragmentary state and having lost the hilum, necessary for identification to species. They have all been classified as pea/vetch (Lathyrus/Vicia). Results Period IT: Iron Age Early Iron Age (Context 1057/sample 14) The only sample from this phase (Phase I) came from a mid brown silt with charcoal flecks taken from a posthole fill. As well as frequent charcoal fragments there was a high concentration of fairly well preserved wheat grains from the 25ml float. Although many showed signs of damage from charring, they retained the characteristic form of hulled wheat grains. 141 grains were therefore identified as Triticum dicoccum/spelta, with a further 97 of the more poorly preserved individuals classified as Triticum sp. The only cereal chaff was a single wheat glume base, but was too fragmentary to assign to species. The few barley and oat grains were also in poor condition and many other grains were classified as Cereal indet. due to their fragmentary nature. The small assemblage of weeds included individual examples of black bindweed (Fallopia convolvulus), barren brome (Anisantha sterilis) and bartsia/eyebright (Odontites/Euphrasia), common arable weeds. Middle Iron Age (Context 1041/sample 10, context 1067/sample 9 and context 1040/sample 8) Ten samples were originally taken from pit, posthole and ditch fills from this period. However, only three of the samples associated with the rampart wall (contexts 1040, 1041 and 1067) were fully analysed. The grain recovered from context 1041, a layer of mid brown-grey silt, which predated the inner rampart wall (1044), was poorly preserved, although it was possible to say that both hulled wheat and the more rounded grains of bread wheat were present. Several well-preserved grains of barley IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 155 still retaining their outer glumes were identified as hulled barley. There was no cereal chaff and only a single orache (Atriplex sp), a weed of disturbed soils. Other weeds included small grass caryopses (Poa/Phleum species -meadow-grass/cat’s tail) and vetch (Lathyrus/Vicia) typical of grassy habitats, which may have grown on the margins of arable fields. There was also a single hazelnut (Corylus avellana) fragment. A layer of dark brown-grey silt (context 1067) that occurred between the stones of the inner rampart wall (1044), also contained poorly preserved wheat grains, identified as a hulled variety, with a single spelt wheat glume base confirming the presence of spelt. There were examples of both oat and barley, with some well preserved examples of barley grain identified as the hulled variety, with two straight grains suggesting either 2 or 6 row forms could have been present. Arable weeds included fat-hen (Chenopodium album) and pale persicaria (Persicaria lapathifolia), with species more typical of grassy conditions such as selfheal (Prunella vulgaris), meadow-grass/cat’s tail and vetch. Sample 8 (context 1040) came from the dark brown clay-silt upper fill of pit 1045, which contained abundant fired clay and charcoal fragments. There were some well-preserved examples of hulled wheat from this feature, plus two spelt wheat glume bases. It was possible to identify hulled barley from eleven grains with four straight grains noted. However much of the grain was classified as cereal indet. due to its poor condition. A similar small suite of arable and grassland type weeds occurred, as in the earlier phase, plus a single hazelnut fragment. ‘Other features sampled from this phase (Phase VII) included pit, ditch and posthole fills. These produced few charred remains including wheat, barley and oat grains but were not fully analysed. Period III; Late Anglo-Saxon (Contexts 1023/sample 7, 1032/sample 19, 1025/ sample 6 and 1034/sample 20) Four samples were taken from Late Anglo-Saxon deposits. Context 1023 (sample 7) was an occupation layer truncated by several pits. As well as charcoal fragments this deposit contained frequent wheat and barley grains, a few oats plus a range of weed seeds. The grain from here was fragmented and generally in poor condition, although examples of the more rounded form typical of bread wheat were better preserved. Occasional hulled type grains and two spelt type glume bases suggest the continued presence of this wheat variety. The single tough rachis internode recorded is more typical of free threshing bread wheat forms. There are slightly more grains of barley than wheat here although oats sull form a minor component. Some of the weed species are again suggestive of grassy habitats and include buttercup (Ranunculus acris/repens/bulbosus ) and heath-grass (Danthonia decumbens). A fragment of stinking chamomile (Anthemis cotula), a typical cornfield weed of heavier soils, also occurred. The other three samples were from pit fills. A . thin layer of charcoal forming the secondary fill of pit 1026, (context 1032) contained a more significant quantity of bread wheat type grains, with many of the 292 counted in good condition, although 76 grains were still identified as c.f. Triticum sp. There were many tough rachis internode fragments plus two possible hexaploid type glume bases, which were unfortunately fragmented. Barley and oats were less common here. Arable weeds included cleavers (Galium aparine), orache, stinking chamomile, plus ribwort plantain (Plantago lanceolata), more typical of grassy places. There were also eight hazelnut fragments plus a charred sloe (Prunus spinosa) stone. The upper fill of pit 1028 (context 1025) included abundant charcoal and butchered animal bone. Charred cereals were again predominantly bread wheat type grains, with tough rachis internodes with one hexaploid glume. There were few barley grains, mostly pitted, with surfaces lost and oat grains were also in very poor condition. The few weeds included orache, brome and bartsia/eyebright. The pit fill also produced a number of mineralised seeds, including a cherry (Prunus avium) stone. Other mineralised seeds included some of the same weed species already recovered in a charred form, such as orache and dock (Rumex sp), with the addition of the cornfield weed thorow-wax (Bupleurum rotundifolium). There were also a few mineralised fly pupae, including one example with the fly beginning to emerge from its pupae, showing the excellent preservation. A lens of charcoal within pit 1028 (context 1034) consisted primarily of sizeable charcoal fragments. Grain was less abundant here and was all of bread wheat type, with tough rachis internodes also present. There was only a single barley grain, plus one small fragment of a barley rachis internode base. Only a few weed seeds occurred. Period IV: Medieval (Context 1014/sample 3, 1010/sample 2 and context 1008/sample 1) A 12th century occupation layer (context 1014), sealed beneath Building 1, contained frequent bread 156 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE wheat grains with some barley and oat, although no chaff was found. The small seed assemblage included the arable weeds fat-hen, stinking chamomile and brome, with sheep’s sorrel (Rumex acetosella) and clover/medick, suggesting grassy habitats. A single spike-rush (Eleocharis palustris/uniglumis) may indicate areas of damper ground. A thin lens (context 1010) of charcoal and fired clay deposited as a sub-base for the stone floor of Building 1, produced frequent bread wheat, a few poorly preserved barley and oat grains, with the only cereal chaff, a single wheat tough-rachis internode. Arable weeds included many of the species recovered from the other late Saxon and medieval samples, with the addition of scentless mayweed (T7iipleurospermum inodorum), narrow-fruited cornsalad (Valerianella dentata) plus over 30 stinking chamomile achenes. The residue from this sample also included animal bone, fired clay and eggshell fragments. The fill (context 1008) of a small rubbish pit contained a large quantity of burnt debris with the sample float producing an abundance of charcoal fragments. Wheat grains dominated although again were in fair to poor condition. Oats were more frequent than barley and a further field crop may be indicated by the presence of two cotyledons of celtic bean (Vicia faba). Other seeds were in good condition and included a similar assemblage of arable weeds as in other samples. Grassland species include clover/medick (Trifolium/Medicago), vervain (Verbena officinalis), with vetches and grasses. Some degree of mineralisation had also occurred with a single orache seed and a mineralised fly pupae. Cultivated plants The most common cereal crop from all phases of occupation at Malmesbury is wheat. A change can be seen from hulled wheat, likely to be spelt in the earlier Iron Age phases, to free-threshing bread wheat in the Anglo-Saxon and later phases. Spelt (Jitticum spelta) was the principal wheat crop in England in the Iron Age and Roman periods with hulled six-row barley (Hordeum vulgare) (Greig 1991). Bread/club wheat (Triticum aestivum/ compactum) is recorded as a minor crop from the Neolithic, although it seems to occur increasingly from the Iron Age (Dark 2000). Continuity is shown at Malmesbury by the occurrence of occasional bread wheat in an assemblage of predominantly hulled wheat forms in one of the Middle Iron Age contexts (1041) and the occurrence of spelt type grains, with free-threshing wheat in one of the Late Anglo-Saxon layers. Spelt was favoured for its excellent baking and . milling properties and it seems likely that harvested crops were brought into the hillfort at Malmesbury during the Iron Age for processing and subsequent re-distribution or for storage as seed-corn. By the end of the Anglo-Saxon period, bread wheat (Triticum aestivum) was one of the principal crops in England, a trend also shown to occur at Malmesbury, with rye (Secale cereale), barley and oats also present (Greig 1991). Spelt was becoming more rare, although it occurs sporadically, as in one of the Malmesbury contexts. The presence of a few hexaploid (Jiiticum aestivum/compactum) wheat rachis internodes from two of the Late Anglo-Saxon contexts (1032 and 1025) suggests that the bread wheat/club wheat form was cultivated. Bread wheat was favoured for its bread making qualities, with its high gluten content producing a lighter bread (Percival 1921). Barley was consistently present throughout all phases examined at Malmesbury, second in importance to wheat, but appears to become less significant from the Anglo-Saxon period onwards. Both barley and oats were widely cultivated as spring crops for use as food or fodder. However, there is no evidence from any oat chaff to confirm whether the grains found here are from a cultivated variety. They may have been growing as the wild form (Avena fatua) as a crop weed, along with brome (Bromus sp), stinking chamomile (Anthemis cotula) and cleavers (Galium aparine) also found in these deposits. The single celtic/horse bean from the medieval pit fill (1008) may have come from part of a field or garden crop, such crops providing an important addition to the staple diet. Other vetches present could also have been growing amongst field crops, or in the cereal fields. All legumes would have been valued for their nitrogen fixing abilities, in helping to improve the soil, and may have been grown as part of a crop rotation system. Other food plants There is little evidence for other possible food plants, although there are occasional finds of hazel nut (Corylus avellana) fragments from all phases, plus two sloe (Prunus spinosa) stone fragments from an Early Iron Age posthole fill (1057) and a whole sloe stone from an Anglo-Saxon pit fill (1032). The hazelnuts and sloe/blackthorn are likely to have been collected from local hedgerows or scrub, although it is interesting that Rowena Gale (see Charcoal Report) has also identified blackthorn charcoal from IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 157 the same Iron Age posthole. The sloe stone may therefore simply have been charred with the wood. Hazel charcoal only occurs sporadically, although hazel must have grown locally, perhaps used for wattles and poles rather than firewood. Another fruit, wild cherry (Prunus avium), came from a Late Anglo-Saxon pit (1025) and together with other weed seeds was preserved by mineralisation. Arable weeds Many of the weeds recovered from contexts at Malmesbury occur very sporadically, often as individual occurrences and were present in all deposits examined. Some of these weeds are common species of disturbed and arable ground and while today these would not be considered as crop weeds, they are commonly recovered in the archaeobotanical record in contexts associated with charred cereals. Their occurrence in a charred form suggests they were growing with the crops, or on field margins or adjacent to boundary ditches, (which may have been wet), and would all have been gathered at harvest. The presence of oat and brome has already been mentioned, and these may have grown with other weeds such as orache (Atriplex spp), docks (Rumex spp),black-bindweed (Fallopia convolvulus) and bartsia/eyebright (Odontites/Euphrasia). There are several species that only appear from the Anglo-Saxon period onwards as cornfield weeds associated with medieval cereal assemblages. These include stinking mayweed (Anthemis cotula), thorow-wax (Bupleurum rotundifolium) and knapweed (Centaurea sp). Although the single knapweed -achene was poorly preserved, this may have been a cornflower (Centaurea cyanus), often found with the other “cornfield weeds.” Although there is evidence for cornflower from the Roman period from several sites, it does not appear to become common until c.1200AD (Grieg 1989). Certain species with specific soil preferences suggest cultivation of marginal ground. Heath- grass (Danthonia decumbens) occurs at Malmesbury in one of the Late Anglo-Saxon contexts (1023) and prefers damp acid soils. Spike-rush (Eleocharis palustris/uniglumis), a plant of damp places, also frequently occurs in the archaeobotanical record in a charred form in association with crop remains. Its presence at Malmesbury in a medieval deposit (1014), with other sedges (Carex spp) in (1023) and (1010) and rush (Juncus spp) in (1010) from the late 11th century onwards, may hint at the utilisation of more marginal areas for cultivation. Comparison with other sites The charred plant remains from the excavation at Malmesbury can be compared with evidence from other sites. Cereal cultivation seems to have been widespread in Iron Age Britain and in areas close to Malmesbury communities appear to have practised a mixed farming economy with sites at Uley Bury and Crickley Hill providing evidence of cereal | cultivation. At the small nucleated village of Claydon Pike interpretation of the charred cereal remains suggested that although the people who lived there consumed cereals they did not grow their own crops. Enclosure of land surrounding the site by ditches and possibly hedges, with other environmental evidence, suggested an open grassland landscape, more suggestive of pastoralism (Darvill 1987). A site at Bishop’s Cleeve, near Cheltenham, revealed evidence of occupation from the Roman to the medieval period, with later occupation thought to be associated with low status peasant dwellings. Samples were taken from pit and ditch fills, with both waterlogged and charred preservation (Jones 1998). All the staple crops of the medieval period are represented with bread/club wheat and river/ macaroni wheat (Triticum turgidum/durum), barley, oats, rye, as well as peas (Lathyrus/Vicia/Pisum) and beans. Evidence of the staple crops was also provided from 11th to 12th century features at Hillesley, near Wotton-under-Edge, Gloucestershire. Although documentary sources from here have shown that crops of hemp, flax, beans and vines were grown for the manor of Hillesley, no evidence was forthcoming for these in the archaeobotanical record (Jones 1998). At Keynsham, to the south of Bristol, samples were examined from deposits associated with the 12th century Abbey (Hunter 1996). Apart from the usual cereal crops there was evidence of the use of celtic bean and peas. Legumes were seen as an important crop at Eckweek (Carruthers 1995), a medieval farmstead near Bath where activity occurred from the 9th to the 14th century. There is little evidence from any of these medieval sites, including Malmesbury, for the importation of more ‘luxurious’ foodstuffs in contrast to the wide range of cultivated and imported fruits and nuts, such as figs, grapes, peach and almond, recovered from urban sites of this period, normally preserved in a waterlogged form (Greig 1983). Conclusion The main cereal in use at Malmesbury from Early 158 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Iron Age contexts is wheat, firstly hulled wheat, likely to be spelt, with bread wheat becoming more dominant from the Anglo-Saxon period onwards. Barley is also present as a second cereal, although at times is poorly represented. While many of the weed species are present in all periods, suggesting some continuity, the presence of species of typical “cornfield weeds” such as stinking chamomile, thorow-wax, and possible cornflower are associated particularly with the free-threshing wheat. The presence of weed species that grow in more marginal, damp soils may suggest an intensification of cultivation on to more marginal land. Some caution is, however, needed in the interpretation of weed assemblages, as although it is suggested that weeds from different habitat groups could all be linked to cereal cultivation, they may just as easily have originated from flooring, dung, hay, or use as fodder. Also the charred remains are found in a secondary context and interpretations would be more secure if they were associated with contexts such as storage pits or the remains of a burnt granary. MOLLUSCA by Dr Paul Davies Introduction Six molluscan samples from six contexts in three trenches were examined. The mollusca were picked from the floats and residues of large (14-70 kg) samples processed through a flotation tank with 250u (float) and 1mm (residue) mesh sizes. Trench 3: Context 304 (burnt destruction layer): Middle Iron Age Context 305 (ditch-fill): Middle Iron Age Context 306 (ditch fill): Early/Middle Iron Age Trench 5: Context 505: Middle Iron Age Trench 7: Context 707 (buried soil): Middle Iron Age Context 713 (buried soil): Middle Iron Age All of the sampled contexts from Ditch 3 contained heat-affected clay, although none was found in the samples from either Ditch 5 or Ditch 7. Results and interpretation Mollusca recovered are listed in Table 8. Nomenclature follows Kerney (1999). Cecilioides Table 8: Mollusca. Number in brackets represents num- ber of individuals burnt; thus 4 (4) represents 4 burnt out of a total of 4 recovered Context 304 305 306 505 707 713 Terrestrial Carychium tridentatum - 13i6 3D 3 - - Succineidae - 1 - - - - Cochlicopa lubrica 2 7 4 3 - - Vertigo antivertigo - 3 (1) - - - - Vertigo pygmaea 6 (1) 4 2, 1 1 - Pupilla muscorum - 4 - - - - Lauria cylindracaea 1 2 1 3 - - Vallonia costata 27 ~105:12)) “35 13 1 1 Vallonia excentrica 1 - - - - - Ena obscura - I - 1 - - Punctum pygmaeum - - - 1 - - Discus rotundatus 12 23 7 a4 3 1 Vitrina pellucida - 2 - - - - Vitrea sp. - 2 - - - - Nesovitrea hammonis 1 1 - - - - Aegopinella nitidula - 1 - fe - - Oxychilus cellarius 4 3 - 7 1 ] Oxychilus alliarius - - - - - 1 Cecilioides acicula 3} 14 23 64 83 23 Clausilia bidentata 2 - 1 - 1 - Trichia hispida 9 23 de 9 1 Freshwater Valvata cristata ] 2 (2) - - - - Lymnaea truncatula 1 10 (8) 2 - - - Lymnaea peregra 2 (2) - - - Gyraulus laevis - 2 (2) - - - - TOTAL 70 225. 792) 123) 199), 28 acicula, a burrowing species, is discounted from interpretation (Evans 1972). Three samples (Contexts 304, 305, 306) contained freshwater mollusca, represented by fragmentary apical pieces. Many of these were burnt, and the freshwater species are confined to the samples that were noted as containing heat-affected clay in an earlier assessment (see above). It is likely, therefore, that the freshwater species were brought in with the clay. A similar origin is likely for Vertigo antivertigo, a species associated with river margin areas (Kerney 1999). The terrestrial faunas are broadly similar, with Vallonia costata, Vertigo pygmaea, and Trichia hispida the predominant characteristic ‘suite’. Together with Cochlicopa lubrica these are indicative of dry grassland-type conditions. It is noticeable, however, that all the samples contain small numbers of more shade-requiring mollusca, particularly Carychium tridentatum, Discus rotundatus and Oxychilus spp. All are species found in longer grassland (see, for IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 159 example Cameron and Morgan-Huws 1975, Davies et al 1996). Additionally, Clausilia bidentata and Ena obscura are even more strongly shade-loving, occurring in woods, hedges, at the base of walls and other similarly shady places where they often climb vertical surfaces (such as tree trunks and walls) (Kerney 1999), although Clausilia has also been found living in tussocky grassland (Cooke and Shotton 1978). Given the nature of the site the samples as a whole indicate a general environment of short- turved grassland, with some areas of longer, more shaded, grassland associated with the ditches and/or palisades. It is worthy of note that with the exception of a small number of Vallonia excentrica in Context 305 anda single Vertigo pygmaea in Context 304, the majority of the terrestrial species recovered show no indication of burning, indicating that fire did not directly affect the surfaces they occupied. DISCUSSION The earliest finds are an unstratified assemblage of late Neolithic/early Bronze Age flints. Their presence indicates that the site was the focus for activity/occupation during those periods, but of what nature is unclear. The flint knife found in Trench 2 is similar to examples from Beaker burials, hinting at burial mounds on the hilltop, although many Beaker burials are found in flat graves. Most contexts relate to defensive features dating from the Iron Age, the Late Anglo-Saxon period and the medieval period. While only a minority of the contexts produced datable finds, their distribuuon through the stratigraphy provided sufficient data to interpret and phase the development of the site. The date and construction of the Iron Age defences are principally of interest to our understanding of the development and layout of the hillfort and its possible abandonment in the early 2nd century BC. Malmesbury is a multivallate hillfort with natural defences on three sides and a level approach on the fourth (northwest) side. The excavated eastern circuit was constructed just above and below the 80m contour. The interior appears quite flat, although the accumulation of material behind the outer rampart and the presence of the modern town have probably contributed to this appearance. Externally the sides of the hill slope steeply; largely dictating the position of the ramparts. Strong defences were probably required on the northwest side (Abbey Row — west gate entrance), where the main entrance may have lain at the eastern end of a natural causeway. This would be similar to Uley Bury (Gloucestershire) where the hillfort has steep slopes on all sides except for a narrow neck of land at the same height as the fort, linking it to an adjacent plateau. Could the Nun’s Walk and its termination at the rear of St Aldhelm’s R.C. Church indicate the site of a second entrance on either the eastern or southern side of the hillfort? Malmesbury certainly belongs to a group of forts sited with defensive possibilities in mind. This ‘hill-top’ class of hillfort normally exceeds 1.6 to 2.2 hectares, averaging about 4.4 to 6.1 hectares but occasionally reaching as much as 22 hectares: Malmesbury covers an area of about 9 hectares. They are invariably strongly defended and, where excavated, usually show evidence of occupation over a considerable period leading in many cases to sporadic strengthening and extension of the fortified circuit — at Malmesbury there is certainly evidence of the former. While geomorphology is a significant factor in the siting and form of hillforts in general, geological considerations, particularly the availability of good building stone can sometimes have a direct effect on the actual structure and perhaps on the form of the hillfort defences: at Malmesbury this is exemplified by the widespread use of limestone for walling. Excavations across Britain have revealed two principal methods of rampart construction, firstly a simple dump of earth and rubble (‘glacis’ type) and secondly a rampart constructed of two rows of vertical timbers tied together with horizontals (‘box rampart type) and infilled with material quarried from the ditch or ditches. In areas where stone is abundant, such as in the Cotswolds, vertical facing of front and rear elevations was common, sometimes with the ends of timbers protruding through the face of the stonework. There are several variations on these two principal techniques. Excavation has shown that the former is not necessarily the earlier. At Malmesbury the phase I and II defences appear to be of the ‘box rampart’ type on the outer rampart with the ‘glacis’ type forming the inner rampart. Post-holes associated with ‘box ramparts’ have been recorded at a number of sites in southern England. At Malmesbury the ‘glacis type’ were replaced by further dump banks with one phase of bank (rampart IV) having a rammed limestone surface. The inner rampart was finally surmounted in the Middle Iron Age by a massive stone wall, maybe as much as 4 160 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE — 5m high. In many Iron Age hillforts the banks or ramparts are between 7.6m and 10.6m wide and between 1.5 — 2.0m in height above present ground level — the original Malmesbury ramparts would have probably been 2.7 — 3.0m high. With revetted ramparts a further 1m could be added for breastwork, so the structure would be between 3.9 — 4.2m high. At Malmesbury the inner rampart wall of the Middle Iron Age hillfort would have been even more impressive. Supplemented by a ditch this would have represented a fermidable obstacle. Rowena Gale identifies oak as a major component in layers of burnt debris sampled in Trenches 3 and 10. Along with the presence of large amounts of heat affected stonework this evidence supports arguments for the existence of both timber palisades and probable ‘box ramparts’ at various periods during the Early and Middle Iron Age and that some were destroyed by fire, or at least fire was a contributory factor in their destruction. These dates are supported by ceramic evidence (see Brown) and radiocarbon dates. Related to this is the recovery, particularly in Trench 5, of quite large quantities of ‘vitrified stone’. Similar material was recorded in Malmesbury near the East Gate in the early 19th century (Moffat 1805), as well as having been recorded on sites in north-east Scotland and in northern and western England. At several hillforts in Scotland, including Craig Phadrig (Highland) and Dun Lagaidh (Highland) it has been observed that stonework had been fused into a glassy mass as a result of intense heat (about 1000° Celsius). The stone used in north-east Scotland is rich in silica that melts when subjected to intense heat (Oram 1996). The presence of this ‘vitrified’ building material at Malmesbury suggests that the rampart was either timber laced or part dry-stone wall with a timber superstructure. Burning of the timber appears to have caused the stonework in the rampart to become both discoloured and to fuse together and to other burnt debris. Excavation of ten separate trenches proved problematic with regard to determining which features were anything more than broadly contemporary. The one real exception to this was Rampart IV, identified as having a surface composed of fragments of crushed redeposited limestone in a matrix of yellowish-orange clay. This deposit was recorded in Trenches 5, 7 and 8 and the level of the surface in each trench was closely comparable. It seems likely that this represents a single phase of rampart construction, broadly datable to the Middle Iron Age. An inner rampart and the western edge of an inner ditch recorded in Trench 10 were probably contemporary with features recorded in 1998-99, strongly suggesting that the hillfort was multivallate from an early stage of development. Evidence from southern Britain, including Maiden Castle (Dorset) and South Cadbury (Somerset), for the period from the 4th century BC to the Roman conquest shows that a proportion of forts were continuously, or at least frequently, inhabited on a large scale. However, no evidence of Iron Age domestic structures was recorded at Malmesbury. The lack of domestic structural remains from this period does not necessarily prove that the site did not fulfil a similar function as the densely occupied and developed hillforts, such as Maiden Castle and Danebury. The lack of data is due to the small scale and location of the excavations. Little is known of interior hillfort layout, but limited excavation work in the vicinity of the abbey in the late 1880s recorded the site of a possible Roman building with a hypocaust and a late Roman horizon was recorded nearby by Julian Richards in 1997. This may suggest the site of a high status Roman building, perhaps a temple or a basilica and potentially a continuity of religious practice from the pre-Roman invasion period. When possible the opportunity should be taken to conduct excavation within the interior of the hillfort, focusing principally on the areas of the west gate (Abbey Row) and the abbey. Archaeological evidence recovered to date is tantalising. The existence of a major Iron Age hillfort is now proven, which may have continued to be occupied for much of the Roman period. It is only possible to speculate as to the interior arrangements as well as the pattern and intensity of occupation. Palaeoenvironmental material, principally charred plant remains (see Jones) and animal bones (see Sykes), suggest a variety of cultivated food plants and livestock in the surrounding area during the Early and Middle Iron Age. Probably the single most interesting animal find was that of the mandible of an adult cat (Felis catus) in a stratified Iron Age deposit as it was previously considered that the domestic cat was not introduced into Britain until after AD 400. While no stratified deposits or features from the Late Iron Age or Romano-British periods were recorded during the recent excavations, residual Romano-British pottery was found in later deposits on site and has been recorded elsewhere around the town. The Anglo-Saxon structure recorded in Trench 10 is probably associated with late 10th/early IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 161] 11th century burh defences. This feature appears contemporary with the rampart recorded immediately east of the Nun’s Walk in 1998-99. While the ‘Burghal Hidage’ implies the existence of a fort by the early 10th century, there was no archaeological evidence of defences or the presence of datable deposits from the late 9th/early 10th century. The recording of successive late 11th/early 12th century stone buildings within the study area is of interest when one considers how intensively land within the early medieval town was occupied. Were the buildings utilised as storehouses for agricultural produce? i.e. as barns or outhouses, or were they cottages? Was the land part of the abbey estates? These are questions, which at present, cannot be easily answered. The position of the south wall of the earlier building (Building 1) may indicate that an intramural walkway lay between it and the town defences. The two buildings were probably only occupied for a few years as pottery found in the deposits sealing the later of the two was contemporary with that found in contexts immediately beneath the earlier one. The area was then cleared to make way for a clay and limestone rubble rampart, which was either contemporary with the early 12th century town wall or the later 12th century improvements to the town wall defences. A layer of subsoil that contained domestic refuse, including many 13th/14th century pottery sherds and animal bones, sealed the earth/stone embankment. The presence of this waste may indicate that the defences, at least in this area, had by then become neglected. Other than topsoil there was a total lack of stratified post-medieval features and deposits, and no trace of the apparently sophisticated mid-17th century Civil War defences. It is documented that the defences were slighted in 1646 and late 19th century landscaping probably erased any other surviving traces. ACKNOWLEDGEMENTS Bristol and Region Archaeological Services (BaRAS) would like to thank Roy Canham, County Archaeological Officer for Wiltshire County Council, Tom Cromwell of English Heritage, David Faulkner, Historic Buildings Officer for North Wiltshire District Council, Dr Andrew Reynolds, Editor of the Wiltshire Archaeological and Natural History Magazine (WANHM), the BaRAS site team of Pete Insole, Stuart Whatley, Chris Zukowski, Patrick Watson, Andy King, Alison Kennen, Dave Mullin (also for the Lithics report), Kim Watkins, David New and Sam Smith, Reuben White and his team from Stansell (Conservation) Ltd., Paul Davies, Rowena Gale, Julie Jones, Rod Burchill, Naomi Sykes and Lisa Brown for the specialist reports, Ann Linge for the illustrative work (except the Iron Age pottery), Vanessa Straker, SW Regional Adviser for Archaeological Science to English Heritage, Margaret Brooks of English Heritage Contract Conservation, Dr Alan Hogg of the Radiocarbon Dating Laboratory, University of Waikato, New Zealand, Mark Corney of Bristol University and John Bowen, Mayor of Malmesbury and local historian. Many thanks also to Mr J. Longstaff (Headteacher) and the staff and pupils of St Joseph’s R. C. Primary School, Malmesbury for their interest and co-operation. BIBLIOGRAPHY ANON 1831 Gentlemans’ Magazine, 101 (Pt 2) ANON 1956 The Utilisation of Hazel Coppice. Forestry Commission Bull. No. 27 APSIMON, A., RAHTZ, P A. and HARRIS, L. G. 1958 The Iron Age A Ditch and Pottery at Pagan’s Hill, Chew Stoke, Proceedings of the Bristol University Speleological Society 8, 97-105 BALLARD, A. 1906 The Walls of Malmesbury. English Historical Review 21, 98-105 BIRELL, J. 1982 Who Poached the Kings Deer? A Study in 13th Century Crime. Midland History 7, 9-25 BOESSNECK, J. 1969 ‘Osteological Differences between Sheep (Ovis aries Linné) and Goat (Capra hircus Linné)’, in Brothwell, D., Higgs, E. and Clark, G. (eds) Science in Archaeology, 331-58 BOURDILLON, J. 1993. ‘Animal Bones’, in Graham, A. H. and Davies, S. M. Excavations in Trowbridge, Wiltshire, 1977 and 1986-1988: The Prehistoric, Saxon and Saxo- Norman Settlements and the Anarchy Period Castle. 127- 136. Wessex Archaeological Report 2 BRAKSPEAR, Sir H. 1930 Proceedings at Meetings: Malmesbury. Archaeological Journal 87, 453-62 BROWN, L. 1991 ‘Iron Age Pottery’, in Cunliffe, B. and Poole, C., Danebury, An Iron Age Hillfort in Hampshire, vol § The Excavations 1979-88; The Finds, 277-319 BROWN, L. n.d. ‘The Iron Age Pottery’, in Powlesland, I., forthcoming thesis BROWN, L. and CUNLIFFE, B. 1984 ‘Iron Age Pottery’, in Cunliffe, B., Danebury, An Iron Age Hillfort in Hampshire, vol 2 The Excavations 1969-78: The Finds, 231-331 BROWN, L. and CUNLIFFE, B. 1987 ‘The Later Prehistoric and Roman Pottery’, in Cunliffe, B., 162 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Hengistbury Head, Dorset, vol 1: The Prehistoric and Roman Settlement, 3500 BC — AD 500, 205-307 CAMERON, R.A.D. and MORGAN-HUWS, D.I. 1975 Snail Faunas in the early stages of a Chalk grassland succession. Biological Journal of the Linnean Society 7, 215-29 CARRUTHERS, W. 1995 Charred plant remains from the medieval farmstead at Eckweek, Avon. Ancient Monuments Lab Report 27/95 CENTRE FOR HUMAN ECOLOGY. 1995. Animal Bone Metrical Archive Project (ABMAP). Draft Report on the Project Phase for English Heritage. Centre for Human Ecology, Department of Archaeology, University of Southampton CHATWIN, C. P. 1960 The Hampshire Basin and Adjoining Areas COLES, J. M. and ORME, B. J. 1980 Prehistory of the Somerset Levels COOKE, P and SHOTTON, EW. 1978 Molluscs from grass tussocks at Upton Fields Nature Reserve. Proceedings of the Coventry Natural History and Scientific Society 5, 55-56 CORNEY, M. 1998 The Nun’s Walk, Malmesbury (Wiltm Report 1998.035) COY, J. 1980. ‘The Animal Bones’, in Haslam, J. ‘A Middle Saxon Iron Smelting Site at Ramsbury, Wiltshire’. Medieval Archaeology 24, 1-68 CUNLIFFE, B. 1991 Iron Age Communities in Britain CUNNINGTON, M. E. 1923 The Early Iron Age Inhabited Site at All Cannings Cross, Wiltshire CURRIE, T. C. 1993. ‘The Animal Bones’, in Currie, C. K. ‘Excavations at the Postern Mill Site, Malmesbury, 1986-87’. WANHM 86. 72-74 DARK, PR 2000 The Environment of Britain in the First Millenium A.D. DARVILL, T. 1987 Prehistoric Gloucestershire DAVIES, P, GALE, C.H. and LEES, M. 1996 Quantitative studies of modern wet-ground molluscan faunas from Bossington, Hampshire. Journal of Biogeography 23, 371-77 DAVIS, S. J. M. 1981. Late Pleistocene and Holocene equid remains from Israel. Zoological Journal of the Linnaen Society 70, 289-312 DAVIS, S. J. M. 1992. A rapid method for recording information about mammal bones from archaeological sites. HBMC Ancient Monuments Laboratory report 19/92 DOWDEN, W. A. 1962 Little Solsbury Hill Camp. Proceedings of the Bristol University Speleological Society 9, 177-82 DUCK, J. and TOOMBS, E. 1993 Excavations on the Town Defences, Malmesbury Library (Wiltm Report 1993.072) EDLIN, H.L. 1949 Woodland crafts in Britain. EGAN, G. and PRITCHARD, EF. 1991 Medieval Finds from Excavations in London: 3: Dress Accessories c.1150- c.1450, 50-123 ELRINGTON, C.R. (ed) 1991 Victoria County History of the ~ County of Wiltshire ENGLISH HERITAGE 1991 Management of Archaeological Projects EVANS, J.G. 1972 Land snails in archaeology FEACHEM, R.W. 1992 Guide to Prehistoric Scotland FINDLAY, D.C, COLBORNE, G.J.N, COPE, D.W, HARROD, T.R, HOGAN, D.V and STAINES, S.J. 1984 Soils and their use in South-west England. Soil Survey of England & Wales. Bulletin No. 14 GALE, R. 2002 ‘The Charcoal’, in Longman, TJ. Jron Age and later defences at Malmesbury: Excavations 1998-2000. Forthcoming report GALE, R. and CUTLER, D. 2000 Plants in Archaeology GETTY, R. 1975. Sisson and Grossman’s the Anatomy of the Domestic Animals GILCHRIST, R. & MYTUM, H. C. 1986. Experimental archaeology and burnt animal bone. Circaea 4 (1): 29-38 GODWIN, H. 1956 The History of the British Flora GRANT, A. 1979. ‘The Animal Bones (Citizen’s House)’, in Cunliffe, B. W- (ed.) Excavations in Bath, 1950-1975. CRAAGS: Committee for Rescue Archaeology in Avon, Gloucestershire and Somerset (Excavation Monograph 1), 60-70 GRANT, A. 1982. ‘The use of tooth wear as a guide to the age of domestic ungulates’, in Wilson, B., Grigson, C., and Payne, S. (eds) Ageing and Sexing Animals from Archaeological sites. British Archaeological Reports, British Series 109. 91-108 GRANT, A. 1987. Some Observations on Butchery in England from the Iron Age to The Medieval Period. Anthropozoologica, Premier Numéro Spécial. 53-58 GRANT, A. 1988. ‘The Animal Resources’, in Astill, G. and Grant, A. (eds.) The Countryside of Medieval England. 149-187 GREIG, J. 1983 Plant foods in the past: a review of the evidence from northern Europe. Journal of Plant Foods. 5. 179-214 GREIG, J. 1989 The early history of the Cornflower (Centaurea cyanus L.) in the British Isles. Palaeoethnobotany and Archaeology. International Work-group for palaeoethnobotany. 8th Symposium, Nitra-Nove Vozokany, 1989. Archaeological Academy of Sciences GREIG, J. 1991 ‘The British Isles’ in van Zeist, W, Waslikowa, K and Behre, K.E Progress in Old World Palaeoethnobotany. 299-334 GRIGSON, G. 1958 The Englishman’s flora HAGEN, A. 1998. A Handbook of Anglo-Saxon Food: processing and consumption HAMBLETON, E. 1999. Animal Husbandry Regimes in Iron Age Britain. British Archaeological Reports, British Series 282 HAMILTON, J. 2000. ‘The Animal Bones’, in Cunliffe, B. and Poole, C. (eds) The Danebury Environs Programme. The Prehistory of a Wessex Landscape Volume 2, part 2. Bury Hill, Upper Clatford, Hampshire 1990. English Heritage and Oxford University Committee for IRON AGE AND LATER DEFENCES AT MALMESBURY: EXCAVATIONS 1998-2000 163 Archaeology, Monograph No 49, 67-73 HARCOURT, R. A. 1979. ‘The Animal Bones’, in Wainwright, G, J (ed) Gussage All Saints, an Iron Age Settlement in Dorset. Department of Environment Report 10, 150-160 HASLAM, J. 1976 The Archaeological Potential of Wiltshire Towns HILLMAN, G. 1984 ‘Interpretation of archaeological plant remains: The application of ethnographic models from Turkey’, in van Zeist, W and Caspaire, W.A. (eds) Plants and Ancient Man. 1-41 HUNTER, K. 1996 ‘Plant macrofossil analysis at Keynsham Nursery’ in unpublished report INGREM, C. 2000 ‘Appendix D: Animal Bones’, in Longman, T.J. Post-Excavation Assessment Report of the Excavation at St. Joseph’s Primary School, Malmesbury, Wiltshire. Unpublished report for English Heritage JACOMET, S. 1987 Prahistorische Getreidefunde. Eine Anleitung zur Bestimmung prahistorische Gerstenund Weizen-Funde JACOMET, S_ 1989 Prahistorische Getreidefunde: a guide to the identification of prehistoric barley and wheat finds. Botanical Institute of the University of Basel, Department of Taxonomy and Geobotany JONES, J. 1998 ‘Charred plant macrofossils’, in Longman, T.J. The Excavation of an early medieval field system at Fillesley Farm, Hillesley, Gloucestershire. Forthcoming Bristol and Gloucestershire Archaeological Society report JONES, J. 1999 ‘Charred plant remains’, in Vallender, J. Tron Age occupation at Guiting Manor Farm, Guiting Power, Gloucestershire. Unpublished report for Gloucestershire County Council JONES, J. 2000 ‘Assessment of charred plant remains’, in Longman, T.J. Post-Excavation Assessment Report of the Excavation at St Joseph’s Primary School, Malmesbury, Wiltshire. Unpublished report for English Heritage JONES, J.and SMITH, D. 2002 ‘The Ecofactual Evidence’, in Enright, D. & Watts, M. A Romano-British and medieval settlement at Stoke Road, Bishop’s Cleeve, Gloucestershire. Bristol & Gloucestershire Archaeological Society Report No. 1, 49-67 JONES, M. 1978 ‘The Plant remains’, in Parrington, M. The excavation of an Iron Age settlement, Bronze Age ring ditches and Roman features at Ashville Trading Estate, Abingdon, Oxon. 1974-76. CBA Research Report 28. 93-110 KELLY, S. E. (ed) 2005 Charters of Malmesbury Abbey. Anglo-Saxon charters xi. Joint Committee on Anglo- Saxon charters KERNEY, M.P. 1999 Atlas of the land and freshwater molluscs of Britain and Ireland KNOWLES, D. and HADCOCK, R. N. 1957 Medteval Religious Houses of England and Wales KRATOCHVIL, Z. 1973. Schddelkriterien der Wild- und Hauskatse (Felis silvestris silvestris Schreb. 1777 und Fs. f catus L. 1758) Acta Sci. Nat. Acad. Sci. Bohemoslov., 7 (10): 1-50 LAWSON, A. J. 2000 Potterne 1982-85: Animal Husbandry in Later Prehistoric Wiltshire. Wessex Archaeology Report 17 LEGGE, A. J. 1992. Excavations at Grimes Graves, Norfolk 1972-1976. Fasicule 4. Animals, Environment and the Bronze Age Economy LEVINE, M. 1982. ‘The use of crown height measurements and tooth eruption sequences to age horse teeth’, in Wilson, B., Grigson, C., and Payne, S. (eds) Ageing and Sexing Animals from Archaeological Sites. British Archaeological Reports, British Series 109. 91-108 LEVITAN, B. 1982. ‘The Faunal Remains’, in Leach, P. Ilchester Vol 1: Excavations 1974-5 Western Archaeological Trust Excavation Monograph 3, 269-283 LOVATT, C.M. 1987 The historical ecology of Leigh Woods, Bristol. Proceedings of the Bristol Naturalists’ Society, 47, 3-19 LUFE, R-M. 1982. A Zooarchaeological Study of the Roman North-West Provinces. British Archaeoloogical Reports, International Series 137 MALTBY, M. 1987. The Animal Bones from the Excavations at Owslebury, Hants. An Iron Age and Early Romano- British Settlement. Ancient Monuments Laboratory Report 6/87 MALTBY, M. 1993. ‘Animal Bones’, in Woodward, P. J., Davies, S. M. and Graham, A. H. (eds) Excavations at the Old Methodist Chapel and Greyhound Yard, Dorchester, 1981-1984. Dorset Natural History and Archaeological Society Monograph Series: Number 12. 315-340 MITCHELL, A. 1974.A Field Guide to the Trees of Britain and Northern Europe MOFFAT, J. M. 1805 History of Malmesbury and its Abbey MOFFETT, L. 1991 The archaeobotanical evidence for free- threshing tetraploid wheat in Britain. Palaeoethnobotany and Archaeology. International Work-group for palaeoethnobotany. 8th Symposium, Nitra-Nove Vozokany, 1989. Archaeological Academy of Sciences. Nitra MORRIS, E. L., n.d. ‘The Iron Age Pottery from the Batheaston By-pass’, in forthcoming report O’CONNOR, T. P. 1989. ‘What Shall we Have for Dinner? Food Remains from Urban Sites’, in Serjeantson, D. and Waldron, T. (eds) Diet and Crafts in Towns: The evidence of animal remains from the Roman to Post- Medieval periods. 13-25. British Archaeological Reports. British Series 199 O'CONNOR, T. P. 1992. Provisioning Urban Communities: a topic in search of a model. Anthropozoologica 16, 101- 106 ORAM, R. 1996 Scottish Prehistory PAYNE, S. 1973. Kill-off patterns in sheep and goats: the mandibles from Asvan Kale. Anatolian Studies 23, 139-47 PAYNE, S. 1985. Morphological distinctions between the mandibular teeth of young sheep, Ovis and goats, Capra. Journal of Archaeological Science 1, 139-47 164 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE PAYNE, S. and BULL, G. 1988. Components of variation in measurements of pig bones and teeth, and the use of measurements to distinguish wild boar form domestic pig remains. Archaeozoologia 2, 27-66 PEACOCK, D. P. S. 1969 A Contribution to the study of Glastonbury Ware from South-Western Britain. Antiquaries Fournal 49, 41-61 PERCIVAL, J 1921 The Wheat Plant PILKINGTON, J. 1997 Archaeological Watching Brief at Malmesbury Town Wall, 26-30 Backhill, Silver Street, Malmesbury, Wiltshire (Unpublished BaRAS Report) PORTER, V. 1990. Small woods and hedgerows PREHISTORIC CERAMICS RESEARCH GROUP (PCRG) 1997 The Study of Later Prehistoric Pottery: General Policies and Guidelines for Analysis and Publication, Prehistoric Ceramics Research Group Occasional Paper 1/2 RACKHAM, O. 1976 (1990 edition) Trees and Woodland in the British Landscape RAHTZ, P. A. and BROWN, J. C. 1958 ‘Blaise Castle Hill, Bristol 1957’. Proceedings of the Bristol University Speleological Society 8, 147-71 REYNOLDS, P 1979 Iron Age Farm. The Butser Experiment ROBINSON, R. 1984. ‘Cat’, in Mason, I., L. (ed) Evolution of Domesticated Mammals. 217-226 SAWYER, P. H. 1968 Anglo-Saxon Charters: An Anotated List and Bibliography SERJEANTSON, D. 1996 ‘The animal bones’, in Needham, S. R. and Spence, A. (eds) Refuse and Disposal at Area 16, East Runnymede. Runnymede Bridge Research Excavations, Volume 2, 194-223 SERJEANTSON, D. and LOCKER, A. n.d. Fish Consumption in Medieval Southern England: the archaeological evidence. Unpublished paper to ICAZ Fish Conference SPENCER, B. 1983 ‘Limestone-tempered pottery from South Wales in the Late Iron Age and Early Roman Period’. Bulletin of the Board of Celtic Studies 30, 405- 19 STACE, C 1991 New flora of the British Isles SYKES, N. 2001. The Norman Conquest: A Zooarchaeological Perspective. PhD thesis, University of Southampton TAIT, J. 1936 The Medieval English Borough: studies on its origins and constitutional history THEOPHRASTUS Enquiry into Plants, trans. A. Hort, 1961, Loeb ed TROELS-SMITH, J. 1960 Jvy, mistletoe and elm: Climate indicators — fodder plants, Geological Survey of Denmark, IV, (4), 4 TUTIN, T. G., HEYWOOD, V.H. et al. 1964-80 Flora Europaea, 1-5 USHER, G. 1974 A Dictionary of Plants used by Man™ VINCE, A. and GOUDGE, C. n.d. Pottery Type-series for Gloucester. Forthcoming report VON DEN DRIESCH, A.1976. A guide to the measurement of animal bones from archaeological sites. Peabody Museum Bulletin 1 WAINWRIGHT, G. 1970 An Iron Age Promontory Fort at Budbury, Bradford-on-Avon, Wiltshire. WANHM 65, 108-66 WILLIAMS, D. FE 1984 ‘Petrological Examination’ in Cunliffe, B. Danebury, An Iron Age Hillfort in Hampshire, vol 2 The Excavations 1969-78: The Finds, Microfiche 8, D13-E3 WILTSHIRE COUNTY COUNCIL Wiltshire Sites and Monuments Record (WSMR) Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 165-189 The Romano-British roadside settlement of White- walls, Easton Grey, Wiltshire: recent fieldwork by Tony Wilmott and Deborah Shipp with contributions by Richard Brickstock, Andy Hammon, Sarah fennings, Ben Moore, Michael Stone, forn Schuster, and fane Timby The Romano-British roadside settlement of Whitewalls, Easton Grey, near Malmesbury, runs along the Fosse Way for a distance of almost Ikm, on both sides of the bridging point of the Sherston branch of the river Avon. Settlement extends uniformly some 50m back from the road frontage on both sides. There is evidence for pre-Roman agricultural use, though development really began with the imposition of the Fosse Way on the landscape, and the establishment of the river crossing. There 1s some evidence that the settlement began in the Ist century, though most finds material derives from the 2nd-4th centuries. Though the settlement did not develop a sophisticated plan-form, it was not notably poor, and benefitted from the products of trade. It boasted at least one large stone building (possibly a mansio). It prospered up to the end of the 4th century, and there 1s small scale evidence hinting at early post-Roman continuity. Introduction _ The Romano-British roadside settlement of Whitewalls is located on the Fosse Way, approximately mid-way between Bath and Cirencester, at the crossing of the Sherston branch of the River Avon lkm south east of the village of Easton Grey, Wilts (centred at ST 890871) and 4km to the west of Malmesbury (Fig. 1). The site occupies parts of a series of fields known as Park Hill, Linehams, Lobley, Fosse Knowl and Foxley Park (Colt-Hoare 1821; Shipp 2000), as well as Whitewalls Wood, to the north of Park Hill, which is bounded to the north by the main B4040 road between Malmesbury and Sherston and to the east by the former POW camp of Easton Grey (Thomas 2003). Part of the site was scheduled as an Ancient Monument in 1992 (SAM no 12046), the scheduling covering the fields to the north of the river and to the west of the Fosse Way and the limits of visible earthworks within English Heritage, Fort Cumberland, Eastney, Portsmouth, POL6 9PY Whitewalls Wood (Figs 2 and 3). Most of the site is under the plough, but the fields known as Fosse Knowl and Linehams, as well as the valley area of Foxley Park, are not cultivated. For a considerable time the portion of the site nearest the river has been used for the illicit recreational use of 4 x 4 vehicles, including a part of the southern end of the scheduled area. Despite the landowner’s attempts to protect the site, in Spring 2003 continued 4 x 4 activity had caused the ground to become very heavily rutted, and suspected Roman stonework was visible in the base of the ruts. As a first step towards the mitigation of this problem, the Inspector of Ancient Monuments, Mr P McMahon, requested the English Heritage Centre for Archaeology to undertake an evaluation intended to establish the nature of the archaeology, the level of threat and the amount of damage which had already taken place. The evaluation took place in August 2003 following the preparation of a project 166 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Swindon _| Malmesbury ®Easton Grey Fig. 1 Site location plan design (Wilmott 2003). It has an importance out of all proportion to its scale, as it was the first modern excavation to take place anywhere on the site. Other works have recently taken place, including archaeological watching briefs and non-invasive survey. In 1997, the Royal Commission on Ancient and Historic Monuments undertook an earthwork survey within the scheduled area between Whitewalls Wood and the river (in Linehams field) in response to a request by English Heritage for information to assist in the management of the site (Kirkham 1997). In 2000, Deborah Shipp (2000) undertook fieldwork on the site as part of her GCSE A-level course in Archaeology, including aerial photographic transcription and fieldwalking. In addition a small magnetometry survey was undertaken by L Chadd as part of an MA course for Bristol University. It was decided that it would be appropriate to take advantage of the opportunity provided by the 2003 excavation to publish all of these works in a single report in order to allow a statement to be made of current knowledge of the site. The Site From the north the Fosse Way descends a moderately steep gradient towards the river. A modern track diverges east of the Roman line, which is visible on the hill slope as a hollow-way and on the flood THE ROMANO-BRITISH ROADSIDE SETTLMENT OF WHITEWELLS, EASTON GREY 167 388000 388500 White Walls Park Hill Linehams ~187500 389000 Lobley Field Extent of Finds Extent of Dark Earth Deposits Magnetometer Survey Grids 500m This map is based upon the OS map by English Heritage with the permission of Ordnance Survey on behalf of The Controller of Her Mojesty’s Stationery Office, Crown copyright. All rights reserved Unauthorised reproduction infringes Crown Copyright and may lead to prosecution of civil proceedings. Liconce Number 100019088 Fig. 2 Site extent as defined by dark earth spreads and fieldwalking finds, with location of magnetometry grid. plain of the river as a clear earthwork bank which intersects with the river some 20m downstream of the small present-day bridge. On the southern side of the Avon the slope is rather steeper and shorter, rising to the well-named Foss Knowl, beyond which the ground levels out. The floodplain les slightly below 80m OD, while the highest point, on Foss Knowl reaches 95m OD. In geological terms the site hes close to the junction of the Oxtord Clay to the east and the 168 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Great Oolite to the west (Geddes 2000). Available building stone comprises Middle Jurassic limestone, Cornbrash and Forest Marble (Geological Survey, Sheet 251). Observations of the natural substrate during the 2003 excavation suggest that the excavated site is on the clay, though over the broader area both clay and oolitic stone have been encountered during works (Messrs Bailey, pers comm). Soils developing over this geology comprise shallow, well-drained calcareous pelosols of Evesham Associations (Soil Survey of England and Wales, 1983). In the valley of the river, the flood plain is covered by thick deposits of alluvium, which have concealed and protected the underlying Roman archaeology (below). The site is little known and has never been systematically explored. More unusually it seems to have escaped the notice of most of the noted antiquaries. Collinson (1791, 100) wrote the first comprehensive description: ‘bn the 264-5 parish of Easton Grey [the Fosse Way] communicates its name to the remarkable eminence of Foss-Knoll; and here stood the ancient city of Whitewalls, denominated without doubt from the appearance of the walls with which it was environed. This city was of considerable extent, and not only its masoned rampires, but the remains of four gates and a series of ruined foundations of buildings within its area have been disclosed. A vast quantity of Roman coins, particularly of Vespasian, Trajan, Adrian, Antoninus Pius and Faustina authenticate the Antiquity of this curious, but little noted spot.” In the second volume of his Ancient Wiltshire, Colt-Hoare (1821, 100-101) presented the first plan of the site. He showed earthworks (Fig 2) lying to the east of the Fosse Way on the north side of the river and notes that: “the irregularities of the surface and the numerous Roman coins etc, which chance and the plough have discovered, prove that this spot was a large and important point on the Foss ........ It possesses no agger of circumvallation, but in many places the cavities and appearance of the soil mark the place of ancient residence ...... A short distance to the north within a copse there is a small square camp which probably had some connection with it”. The extreme contrast between these two descriptions only thirty years apart suggests that Collinson was indulging in exaggeration, particularly concerning the “masoned rampires”, unless, of course, he was reporting an older source which has not survived. Corney (2001) follows Grinsell (1957, 68) and Goddard (1913, 249) in referring to a tesselated pavement, and missattribute the reference to Collinson. The Roman name for the site is unknown. Colt Hoare suggested it was the Mutuantonis of the Ravenna Cosmography, now known to be erroneous (Rivet and Smith 1979, 476-7). Prior to the works reported upon here, the only recorded archaeological intervention on the site was in 1931, when Passmore (1933) observed the installation of a water pipe for the Bristol Corporation. At a point about half-way between the modern Malmesbury to Sherston road and the river, a good surface of the Roman Fosse Way was exposed 1 ft below present level. The old road surface consisted of stone slabs resting on 1 ft of rubble, and the depth was increased down slope in order to carry the road over wet ground to a presumed bridge. Finds included pottery from the Ist to 4th century including a mortarium stamped by DOINUS, a potter of Brockley Hill, Herts, who worked from AD 70-110 (Castle 1972; Green 1980, 49), and “a pretty little bronze nail cleaner”, now in the Devizes museum. Coins recovered included those of Titus, Gallienus, Licinius and Constantine I. Of these a group of silver denarii found in 1920 (Passmore 1921), consisting of one of Julius Caesar, one of Mark Antony, four of Vespasian and three of Domitian, may have been a small hoard. A sketch plan and section of the course of the pipeline survives in the Devizes Museum (Fig 4). Passmore records 1*'- 4" century pottery from the pipe trench where it ‘was carried across the site of the Roman station’, Grinsell (1957, 68) adds that the finds came from ‘beside the Roman road for nearly 100yds NE of the stream’ and notes parts of a human skeleton and Roman glass. In 1997 cathodic protection along the water pipeline was replaced. This involved the re-excavation of backfill, but one Roman coin and five sherds were recovered from spoil (WAM 1998, 138) Chance finds have been the most frequently noted aspect of the site. Colt-Hoare was the first to record a remarkable relief sculpture which was found in a nearby stone quarry (Colt-Hoare 1821, 100). The relief was surmounted by a pediment and depicted four figures. Within the pediment was an inscription which read CIVILIS FECIT (RIB 99). According to THE ROMANO-BRITISH ROADSIDE SETTLMENT OF WHITEWELLS, EASTON GREY 169 388000 388500 389000 White Walls Park Hill Lobley Field Aerial Photography Colt Hoare Earth Works RCHME Earthwork Survey 500m This map is based upon the OS map by English Heritage with the permission of Ordnance Survey on behalf of The Controller of Her Majesty's Stationery Office, Crown copyright. All nights reserved. Unauthorised reproduction infringes Crown Copyright and may lead to prosecution or civil proceedings. Licence Number 100019088 Fig. 3 Site extent as defined by earthwork survey and aerial photograph transcription Colt-Hoare’s original assessment, which is followed that the seated figure wears a crested helmet, and by Passmore (1934), Cunliffe and Fulford (1982, is actually a depiction of Mars. This interpretation 34) and Henig (2001, 114), these represent a seated seems more reasonable. Robinson believes that the mother goddess with three attendants. Robinson relief represents Mars with three female attendants, (2001, 161) however follows Wright’s idea (RIB) or possibly mothers, citing the importance of the 170 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Celtic Mars in the Cotswolds. The head of a female deity sculpted in the round was also recovered, and remains in the Devizes Museum (Cunliffe and Fulford 1982, 37). This small group of sculpture is thought to have come from a shrine on the Fosse Way. (Cunliffe and Fulford 1982; RIB 99). Griffiths (1982, 58) notes the discovery of an early military bronze martingale from the site, as well as military metalwork of 4th century date (Griffiths 2001, 68), while Broomhall and Henig (1981) have published an iron signet ring with intaglio decorated with the head of an eagle. Moorhead (2001) has noted that 374 Roman coins have been found, mostly by metal detectorists, reported and catalogued. Finally, Corney (2001) has recently published a brief summary of the site, which includes consideration of the RCHME earthwork survey and aerial photographic transcription. Non-invasive survey by Deborah Shipp and Tony Wilmott The description of the results of the surveys are presented field-by-field, and are summarised in Figs 2 and 3. Park Hill Park Hill lies within the scheduled area. It is a flat plateau above the valley of the Avon to the north, though at its southern edge it includes the lip of the river valley. It is defined to the east by the line of the Fosse Way, which is preserved in the line of the modern by-way at this point. Earthworks recorded by Colt Hoare in this area have long been ploughed out, and crop marks visible in aerial photography bear no clear relationship to these earthworks (Fig 3). They appear as a series of linear features broadly aligned on the Fosse Way. The crop marks seem to reflect a number of small enclosures among larger field boundaries. The alignment on the road suggests a Roman or later date. A spread of dark earth occupies the eastern 45-50m of the field along the edge of the Fosse Way (Fig 2). Within this band Roman material is visible following ploughing. No field-walking took place here as Scheduled Monument Consent would have been required. Linehams This field, part of the scheduled area, lies to the south of Park Hill, and to the north of the area excavated in 2003. It comprises the northern edge of the river valley; a moderately steep south-facing slope. This is the only part of the site where well-preserved earthworks survive (Kirkham 1997). A series of roughly parallel lynchets of Celtic fields are the earliest features discernible, and these follow the contours of the slope, perhaps enhancing natural river terraces. On the steeper part of the valley side these lynchets survive to a height of 1.5m. Cutting into or overlying the lynchets and fields is a series of about a dozen sub-rectangular hollows measuring 6-8 x 3-4m and a number of larger terraced platforms. Kirkham suggests that these represent building hollows and possibly garden plots or domestic compounds (Fig 3). There is little doubt that these are the traces of Roman buildings and enclosures relating to ribbon development along the Fosse Way. It is significant that the soil observed in molehills over this area is of a dark colour like the dark band along the edge of the Fosse Way in Park Hill field. It is also significant that the earthworks which seem to represent building platforms and compounds extend approximately 50m back from the western edge of the Fosse Way, the same distance as the band of dark earth in Park Hill. An authorised excavation was observed here in 1978, when an area was opened on the side of the Fosse Way earthwork. A wall was encountered which was built of reused sandstone roofing tiles. The field boundary to the east follows the modern track. This lies somewhat to the east of the line of the Fosse Way, which appears within the field as a raised bank against the side of which a hollow- way has developed. On the flood plain, features have been masked by alluvial deposition, and by the development of minor hollow-ways converging on the river crossing during this deposition. There are a number of small hollows, possibly formed by the robbing of materials from the road. The final feature is a sinuous ditch with a bank on its southern, downhill side. This feature, 8m wide in total, runs from the hollow-way to the river edge, cutting through all other earthworks, and may be a late (possibly medieval) field boundary. Limited magnetometry survey was undertaken on three areas; the shallow slope to the north of the late field boundary, within the earthworks on the steep slope, and on the flood plain. The flood plain survey was undiagnostic, apparently reflecting the deep alluvium. On the slope features reflected the visible earthworks. The earthworks were also reflected in the northern sample, though this survey further demonstrated that the occupation reflected by the earthworks was limited to a band 50m deep THE ROMANO-BRITISH ROADSIDE SETTLMENT OF WHITEWELLS, EASTON GREY 171 N A Footings of pier eral Made ground Sand Fig. 4 Sketch plan and section of pipeline course (redrawn from Passmore archive) from the edge of the road. ' The water pipeline laid in 1930 runs broadly down the line of the modern track on the eastern side of the field, though it does cross the track to run across the flood-plain in Linehams field to the east of the line of the Fosse Way. It then emerges from the river bank, and is carried across the river on concrete piers. The section (Fig 4) reveals the profile of the valley and flood plain. It shows that the flood plain largely comprises made ground “chiefly flat stones in which remains were found”. Adjacent to the river, the section reveals a deposit of “sand” within which is buried stonework described as the “footings ofa pier”. The site of this supposed pier is well to the east of the modern bridge, and the stonework is unlikely to be the remains of a Roman bridge as implied in the accompanying note. The most likely place for a Roman crossing is the point where the embankment across the flood plain which marks the line of the Fosse Way meets the river. This is approximately 5m west (downstream) of the modern bridge. Here the river bank is wholly shrouded with thick brambles, though stonework of indeterminate date is partially visible. Traces of large dressed stone in the river bed may be the remains of a ford, though the level at which the road meets the river would imply that the road would have been bridged in the Roman period. Earthworks of differing post-Roman tracks converging on the crossing suggest that this has changed form and location over the post-Roman centuries, culminating in the present day bridge. Lobley The field known as Lobley is the counterpart of Park Hill and Linehams to the east of the modern track. The northern part of the field is comparatively flat, falling sharply into the Avon valley. Approximately 25m from the beginning of the slope a linear bank is visible running along the contour of the slope. This forms a flat terrace on the uphill slope. Aerial photography shows two groups of crop marks. The 172 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE N / 7. : 45 “a 75 AD, Wee é Aan Ag Lobley / P Field 7 by %0 i gr / i, 4 5p / & j 4 85 V6 A70 & / é 46 ki 2 A75 oi L E ae n DN Ravel Number of Finds 0) 100 m fe OO Fig. 5 Plan of fieldwalking lines in Lobley field northern group are distinctively off the alignment settlement. The same alignment appears to the north of the Fosse Way, and comprise two long ditches of the river, and may reflect pre-Roman landscape and a group of rectilinear enclosures indicative of layout and even settlement, like the early lynchetted THE ROMANO-BRITISH ROADSIDE SETTLMENT OF WHITEWELLS, EASTON GREY 173 fields which survive as earthworks in Linehams field. This impression may be enhanced by the results of random fieldwalking over these earthworks, which produced three flints, including a Mesolithic blade, six Roman sherds, a post-medieval sherd and a nail. To the south of the field the earthworks recorded by Colt Hoare have been ploughed out; however they do seem to have been aligned north-south and east-west, and do not appear to be laid out with respect to the Fosse Way. Vestiges of these earthworks appear to be reflected in aerial photography, as are at least two apparent stone buildings. These buildings occupy earthwork enclosures, and may reflect the ploughed-out version of the earthworks still visible in Linehams field. Random field walking in this area produced 44 Roman sherds, oyster shell, a Carausian coin, and fragments of building material including pennant sandstone roofing, and ceramic building material. There seems little doubt that these crop marks represent structures of Roman date. In 1977 a rapidly arranged watching brief was carried out on a water pipe trench along the edge of the track to the west of the field. Extensive terraced Roman buildings were observed, matching the earthworks to the west of the track. Finds included 3rd- and 4th-century Roman pottery, Pennant sandstone and ceramic roofing tiles. Roman material lay on the surface adjacent to the modern track, and a programme of structured fieldwalking was undertaken to explore this (Fig. 5). A line 1m wide was walked along the field edge, and 1m wide strips perpendicular to the field edge, ‘spaced at Sm intervals, were also walked (Fig. 5). The result was that Roman finds extended some 50m eastwards from the western field boundary on the modern track. Concentrations of material varied within this area, but centred on the visible terraced area mentioned above (Shipp 2000, 39). Finds included a large quantity of pennant sandstone roofing material, ceramic roofing tiles and box-flue tle, all implying the presence of stone buildings, possibly with at least one heating system. The most frequent class of find was Roman pottery of which 2,159 sherds were found. Most of the material by far comprised either undiagnostic reduced and oxidised wares or products of the known Wiltshire pottery industries. Of continental imports, the majority comprised samian ware, though there was also a small assemblage of amphora sherds. Diagnostic wares include south western Black Burnished wares, Oxfordshire products, Severn Valley wares, a small number of Verulamium Region sherds, and colour coated wares from the Nene Valley and the New Forest. In addition, there is a small but significant group of sherds of late Roman shelly ware suggesting occupation after c. 360-70. The overall picture is of occupation lasting from the late Ist to the late 4th century across the whole area from which finds were recovered. Foxley Park Although earthworks were drawn in this area by Colt Hoare, these are now ploughed flat (Fig. 3). Aerial photography has revealed a double-ditched feature, interpreted by Corney (2001, 26) as a trackway. This deviated at 90 degrees to cross the line of the Fosse Way. The fact that the feature is well off the alignment of the Fosse Way suggests that it predates the Roman road, and field walking along its line was very unproductive in terms of finds. By contrast, and as in Lobley field, a strip of dark earth up to 50m wide was Visible along the field edge (Fig 2). Fieldwalking of this area showed that the boundary of the dark earth and that of recovered finds was very similar. Finds consisted predominantly of Roman pottery (96 sherds), with a spread of sandstone roofing material and Roman ceramic building material. The pottery range was similar to that from Lobley. Fosse Knowl The surface of the field known as Fosse Knowl at first appears featureless, though there are some slight earthworks which are visible in low, late afternoon sun. It is possible that these are the vestiges of the earthworks recorded by Colt Hoare (Fig. 3). Aerial photography within this field shows firstly a continuation of the probable pre-Roman track from Foxley Park. This is integral with a trapezoidal enclosure which may also be part of pre-Roman land use. Over these features are a range of crop marks presenting the appearance of a classic roadside setlement (Corney 2001, 25). These seem to be a series of ditched compounds or properties, within at least one of which is a stone building. These features, like the earthworks in Linehams field, are laid out perpendicular to the Fosse Way, and occur over an area over 50m deep from the road. Fieldwalking has again produced a range of Roman pottery (194 sherds) and sandstone and ceramic roofing materials extending in a strip up to 76m broad along the road side (Fig. 2). 174 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 388900 / / ee TTenen A or Trench D@ Trench F* Trench B 0 a This map is based upon the OS map by English Heritage with the permission of Ordnance Survey on behalf of The Controller of Her Majesty's Stationery Office, Crown copyright. All rights reserved Unauthorised reproduction infringes Crown Copyright and may lead to prosecution or civil proceedings. Licence Number: 100019088 Fig. 6 Location of 2003 excavation trenches The 2003 excavations The location of the trenches excavated in 2003 was determined by the scale of wheel-rut damage to the site. The site was first surveyed, and the deepest and most consistent ruts were mapped for future reference. The areas of the site most prone to damage and most heavily rutted consisted of the entrance to the area utilized as a 4 x 4 track, and the three main routes to the river. The original intention was to position trenches in each of these areas, though it became clear that one of the river access routes had been protected by the deposition of modern make-up material. Trench A was located in the entrance, and Trenches B and C on the river access routes. Trench D was opened in the area of deepest rutting on the flood plain between the other trenches (Fig 6). All excavation was undertaken by hand, down to the top of extant archaeology, which was then recorded. In areas where damage was apparent, the archaeology was excavated down to the latest intact level. Archaeological recording followed the methods laid out in the CfA Recording Manual. Stratigraphic description by Ben Moore Trench A Anarea 5 x 5m was opened up initially, and the deep wheel-ruts in the surface (2, 3, 4) were emptied and planned. Due to the unexpected depth of deposits, however, the excavated area was reduced to 3 x 3m (Fig. 7). Beneath a thin skim of top-soil (6) were two successive alluvial layers (7, 8), 130 and 140mm in depth respectively. Underneath these was a thin (60mm) layer of black silty material containing much charcoal, which covered the entire trench (10). Concentrated areas of stone were visible within THE ROMANO-BRITISH ROADSIDE SETTLMENT OF WHITEWELLS, EASTON GREY 175 rs eo” A Trench DO (Trench B Trench C River Avon x \ Y Wheel ruts A - 4 waA ea 2-7 — TrenchA Fig. 7 Plan of Trench A this deposit, which also contained a large number of Roman sherds and two mid-4" century Roman coins (Nos 1, 5). Pottery dating suggests deposition c. 360- 400+. When the black material was removed, the stone concentrations were identified as structures. In the south west corner of the trench was a north- West to south-east aligned wall constructed of thinly bedded oolitic limestone brash (13). The north east side had very regular facing and the core was a rubble in-fill of angular stones. At right-angles to this was a slightly less substantial and more ruinous north east- south west aligned wall (14) of similar material and width (620mm), but with no regular facing visible at this upper level. A fair number of burnt stones were noted in this feature as well as signs of subsidence in the centre, perhaps suggesting a covered channel. The two walls may have been butted, but this relationship was not seen. Only the top courses of 176 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE the walls were uncovered, so the depth of survival cannot be assessed (Fig. 8). The two walls defined a room, within which was a lower dark burnt deposit (11) containing quantities of pottery, two mid-4" century coins (Nos 2, 4) and numerous hob- nails. Again pottery indicates a deposition date c. 360-400+. Beneath this, in the angle of the walls was a rubble deposit. It is possible that this marked the destruction of the stone building. The burnt material was cut by a linear, straight-sided, flat bottomed timber slot (15) 250mm wide and 220mm deep, which was filled with a deposit of grey-brown clay-silt, very gritty and containing frequent angular stones Fig. 8 Trench A showing excavated Roman walls and the depth of ruts (12) which may have originated as packing created by illicit 4 x 4 vehicle use (Fig 9). Ww 77.91m O.D. n Fig. 9 South-east facing section in edge of Trench A 177 THE ROMANO-BRITISH ROADSIDE SETTLMENT OF WHITEWELLS, EASTON GREY — So Track __ Trae ver A Trench D® - @a— TrenchB Trench C OA River Avon H 103 tmen, Trench B N eds ae ie Lt cic ‘Ka ce ed hd nce roe ie - Unexcavated \ \ \ \ \ \ \ \ \ \ ae - oe \ co le \ oe \ a ‘ 2 ; . a pe Trench D Fig. 10 Plans of Trenches B, C and D 178 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE E 77.88m O.D x el nx Fig. 11 North-west elevation of wall in Trench B Trench B This trench was sealed by layers of top-soil (100) and sub-soil (101), 180mm and 160mm deep respectively. Beneath this was a distinct alluvial layer (103) 300mm deep. Two further alluvial horizons were encountered beneath this. All three layers of alluvium contained a few Roman pottery sherds, as well as some later material. It was in layer (103) that a spread of stones was first noted and, around 100mm deeper, lay the top of a surviving wall (105). The wall ran north- east to south-west across the trench and was 800mm wide. A 1m wide sondage was excavated against the north-west side of the wall to discover the height of the surviving wall and any occupation layers present. Both the wall and the collapsed or tumbled stone surrounding it continued down through the second layer of alluvium (104), which was 200mm deep, and the wall survived to a maximum height of 580mm. It had an irregular finish and random coursing. No bonding or core material was present. The wall sat on a layer of compacted alluvium (106) with a high clay content, 180mm deep. Below this layer of alluvium lay a paved surface (107) of flat but subangular stones packed closely together in a random arrangement (Figs 10-12). Small sherds of Roman pottery were found amongst the stones during cleaning. The space of time between the construction of this surface and the building of the wall is difficult to determine, and it is not known whether the surface continues under the wall or respects it in some way. Trench C Removal of topsoil (200) revealed a dump of modern building/ demolition material (201) used as a form of river management. This material was a maximum of 800mm deep. Below this was the latest layer of alluvium (202) at 200m deep, which was cut by the line of the riverbank which existed prior to the dumping of the modern material. This sloped at an angle of about 45 degrees. Another layer of alluvium (203) was removed, uncovering a third (204) with a high clay content within which the top of a wall (206) was revealed. This wall was excavated to a depth of 300mm, and is c. 500mm wide (Fig 10). It is randomly coursed and no bonding material was visible. In plan it could be observed that the wall had Fig. 12 Stone wall and crushed stone surface in Trench B under excavation ——————— THE ROMANO-BRITISH ROADSIDE SETTLMENT OF WHITEWELLS, EASTON GREY Fig. 14 Stone wall and floor, Trench D 180 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE two phases, each on a slightly different alignment although this could be due to coilapse. More collapse material (205) was uncovered at the bottom of the original river bank slope (Fig. 13). Trench D In Trench D below 100mm of topsoil (300), four distinct layers of alluvial material were excavated. The first three layers (301), (302) and (303) were similar in composition (clay silt), but each layer contained more finds than the one above. Late Roman coins were found in both layers (302) and (303) (Nos 6, 7). Within (304), a more clayey deposit than those above, the remains ofa severely truncated north-west to south-east wall (305) were uncovered. This deposit contained a mid-4" century coin (No 3). The wall seems to have been heavily robbed and lacked any clear facing stones. Directly to the north east of the wall a paved surface (306) was found. It appears very similar to (107) in Trench B, although here the wall sat directly on the surface. The deterioration of the wall and the consequent tumble (307) clearly took place during the accumulation of the lowest layer of alluvium (304) (Figs10, 14). The Finds Coins by Richard Brickstock The excavation produced only seven coins, all of 4th-century date (Table 1). The lack of any earlier coins is the result of the excavation strategy; as has been shown, previous finds from the settlement of Whitewalls encompass a much wider range of dates. Of the seven coins, five are regular (i.e. official) issues of very common types of mid- fourth-century small change: Constantinian GLORIA EXERCITVS and VRBS ROMA of the 330s-early 340s; and Valentinianic SECVRITAS Table 1 Roman coins from the 2003 excavation catalogued in chronological order (continued over) 1 CONSTANTINEI Date: 330-31 Obv VRBS ROMA Denom: - Cat: 7LG 242, HK 184 Rev Wolfand Twins PLG Wear: SW/SW Axis: 12 Diam: 17.5 mm Wt: 2.0¢8 Mint: LG Off: P Context A,010 (Sf No 411) 2 ‘CONSTANTINE? Date: ‘330-37’ Obv [VRBS ROMA] Helmeted head, 1. Denom: - Cat: c.as 7TR 522, HK 51 Rev Wolf and Twins Wear: ?SW/SW Axis: 5 Diam: 8.5 mm Wt: 0.4¢8 Mint: - Off: - Context A, 011 (Sf No 421) Cataloguing conventions: Wear statistics: Catalogue references: 3. probably ‘CONSTANTINE TY Date: ‘330-37’ Obv 2?[VRBS ROMA] Helmeted head, 1. Denom: - Cat: c.as 7TR 522, HK 51 Rev - Wear: C/C Axis: ? Diam: 7.5mm Wt: 0.2 g Mint: - Off: - Context D, 304 (Sf No 432) Notes: (or, possibly, FEL TEMP REPARATIO copy of 350s) 4 CONSTANTIUS II Date: 337-4 Obv FLIVL CONSTANTIVS A[VG] Denom: - Cat: 8TR 82, HK 126 Rev GLORI-A EXER-CITVS 1 std. TRSbr. Wear: SW/SW Axis: 1 Diam: 16.0 mm Wt Lig Mint: TR Off: S Context A, 011 (Sf No 417) Conventions are as outlined in Brickstock 2004 a question mark implies uncertainty due to corrosion products. The Roman Imperial Coinage (RIC), vols. 1-10, edd. Mattingly et al., 1926-1994, unless otherwise stated: HK and CK refer to Late Roman Bronze Coinage, parts | and 2, respectively, Carson, Hill and Kent, 1960. THE ROMANO-BRITISH ROADSIDE SETTLMENT OF WHITEWELLS, EASTON GREY 181 Table 1 continued 5 HOUSE OF CONSTANTINE Date: 335-41 Obv - Denom: - Cat: as 8TR 57, HK 107 Rev [GLORIA EXERCITVS] 1 std. Wear: C/C Axis: 6 Diam: 15.0 mm Wt: 10g Mint: - Off: - Context A, 010 (Sf No 412) Notes: (or, possibly, VICTORIAEDDAVGGQNN of 346-48) 6 VALENTINIAN I Date: 364-67 Obv DN VALENTINI-[ANVS PFJAVG Denom: - Cat: CK 484 Rev SECVRITAS [REIJPVB[LICASE] Wear: ?W/W Axis: 12 Diam: 19.5 mm We: 18¢ Mint: AR Off: I OF/I//CONP Context D, 302 (Sf No 409) 7 HOUSE OF VALENTINIAN Date: 364-78 Obv Bust diademed, r. Denom: - Cat: as CK 478 Rev [SECVRITAS REIPVBLICAE] Wear: 2W/W Axis: 12 Diam: 16.5 mm Wt: 0.9¢ Mint: - Off: - Context D, 303 (Sf No 410) REIPVBLICAE of the 360s-370s. The remaining two are Constantinian copies, one certainly (No 1) and one probably (No 2) of VRBS ROMA type. The latter two coins provide a point of interest in this small assemblage, on account of their very small module (8.5 mm and 7.5 mm respectively). Copies of Constantinian types of the 330s are very common on many Romano-British sites and appear to belong, for the most part, to the early 340s when there appears to have been a hiatus in the supply of regular small change. The British reaction to this shortage was the wholesale production of copies | of the current types, VRBS ROMA, GLORIA EXERCITVS and CONSTANTINOPOLIS. Many copies are of reasonable size and execution, but others are of remarkably small size and very poor quality, the latter perhaps belonging relatively late in the sequence of production. A similar sequence of copying occurred in the 350s, when the FEL TEMP REPARATIO ‘Falling horseman’ types provided the prototypes, and it is possible that No 3, corroded and virtually illegible, belongs to that episode rather than the earlier one. Small-module copies are found on by no means all Romano- British sites, but tend to be concentrated, perhaps not surprisingly, in sites such as market towns, where a lack of small change might be most keenly felt. Their presence at Whitewalls is therefore not entirely unexpected, though one might anticipate that a larger-scale investigation would also provide examples of larger-module copies and_ other common later-Constantinian types. Pottery by FJ R Timby A total of 1,641 sherds of pottery, weighing 7 kg, were recovered. These dated almost exclusively to the Roman period with a few accompanying medieval and post-medieval pieces (Table 2). In addition a small quantity of ceramic building material (CBM) and one possible fragment of crucible were present. In general terms the assemblage was in very poor condition, reflected in the overall average sherd weight of just 4g; exceptionally low for Roman material, which is generally well-fired and quite robust. This has obvious repercussions on fabric and form analysis and thereby the dating of small groups. The assemblage was sorted into broad fabric groups based on inclusions present, the frequency and grade of such inclusions and the firing colour. Known regional or traded wares were coded following the system advocated for the National Roman reference collection (Tomber and Dore 1998). More local wares were grouped and coded in a similar manner (Timby 2003, Appendix 1). The sorted assemblage was quantified by sherd count and weight for each recorded context. Dating in many cases has had to remain quite vague as many contexts produced undiagnostic local wares or just single or very small sherds which could date to anytime after the earliest Known production date. Roman pottery accounts for the bulk of the assemblage, some 1,631 sherds. In terms of date the material appears mainly to range from the 2nd to later 4th century. If the assemblage is broken _— 82 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 2 Pottery spot dating for 2003 excavated contexts Context | Roman | Med | Pmed | Total No | Total Wt | Date sam__| amp | fwimp | BB1 | Oxford ers | 1 0 0 0 0 0 1 2 45 a ae ee ee ee a a 0 2 26 67 C18-20th fOr | OO! le) Oe | Oe Oe Sale| 0 0 5 28 @2763 7 3 0 0 10 0 0 68 20515 NC22C3 8 0 0 0 Oriel less Oe Opa] Orel 4 late C3-C4 9 1 2 0 5 74 C4 10 9 3 2 32 0 in 15 5 l 14 0 12 1 1 1 8 0 100 4 0 0 0 0 0 13 101 0 0 0 0 2 0 2 103 4 0 0 0 39 1 157 | late C12-13th 104 0 ed a ee ae ee ee 3 Roman p22 062 Ale On al On a0 al Onl SO De =| 15 0. aon 9 Roman OF. 0 On OO) ao so 0 |e | ites | 200. | ol of 0 Fo] o | o | 0 fo] 4 | 4 | © 63 “)Pinedimodern™| | 200 | oO or To | lo oo | 0 so ae iedimicdern| 204 fF oe Oo 0 too 0 tee et eae Romane 2 0 0 2 0 0 2 o[ o | 6 | 14 [late C3-c4 302 0 2 0 7 0 5.26. O50. 400 eae 303 0 3 0 19 2 15 9] 0 130 304 0 1 0 3 0 0 19 0 23 TOTAL | 39 17 4 101 14 124 | 1332| 1 1641 | 6964 | © | down according to source, approximately 3.6% are continental imports, 14.5% regional imports and the remaining 81.9% local, or undetermined. A number of continental imports are present in the group, both amphorae and finewares. The amphorae largely comprise sherds of Spanish Dressel 20 olive oil types. A single sherd from (303) with a volcanic black sand fabric is probably from Italy, and possibly from a Dresssel 2-4 wine amphora. The finewares, some 43 sherds in total, are mainly sherds of Central Gaulish samian with single sherds of Argonne colour-coated ware, Central Gaulish black-slipped ware, Cologne colour-coat and Moselle black slipped ware. In terms of date these wares span the 2nd to early 3rd centuries but are often found associated with later contexts suggesting curation. The samian includes examples of Dragendorff form types 33, 31 and 37. One sherd from (11) has a repair rivet hole. The regional imports include wares from the Poole Harbour and South-west black burnished industries, products from the Oxfordshire industries, (whitewares, whiteware mortaria and colour-coated wares), New Forest and Lower Nene Valley colour-coated wares, vessels from the Wiltshire industries (North Wilts grey and oxidised ware, Savernake ware) and late Roman shelly ware. The late Roman shelly ware is particularly useful chronologically indicating occupation after c. 360-370. Two sherds are present in this group, one from context (10) the other from (11). Similarly the New Forest colour-coated wares and some of the Oxfordshire colour-coated sherds indicate occupation in the 4th century. These were associated with contexts (8-12) inclusive and (303). The Savernake ware along with some of the more local wares also indicate occupation going back to at least the 2nd century. Included within the local wares are sherds of South-west oxidised and white- slipped ware and a few sherds of Severn Valley ware alongside various grey micaceous and sandy wares of unknown source. Nine degraded fragments of ceramic building material weighing 31 g were noted from contexts (10, 300, 302 and 303) probably of Roman origin, but too abraded to identify to type. Two small fragments from (11) may be crucible or just vitrified fired clay. The post-Roman component of the assemblage was quite small with just one medieval and nine post-medieval-modern sherds. The medieval sherd came from (103) and is a green glazed jug or pitcher from the Minety kilns, Wiltshire. The post- medieval/modern sherds mainly appear to have come from topsoil contexts. Other finds by Forn Schuster Table 3 quantifies the main categories of finds by THE ROMANO-BRITISH ROADSIDE SETTLMENT OF WHITEWELLS, EASTON GREY 183 Table 3 Quantification of small finds from 2003 excavation (Animal bone, Pottery and Fired Clay not included) Tota of which stratified Coins 7, 5 Iron 106 44 Lead Waste 1 Glass 19 2 Slag 5 Metal oxide 1 nodule Stone 22 10 CBM 13 (plus 9, cf. pottery rep.) Wood 1 material (excluding animal bone and pottery/ fired clay). This also demonstrates how much of the material was stratified. It remains difficult, especially with iron objects from the colluvial deposits, to distinguish between Roman and “modern” finds in all instances. Tron objects Iron objects have been identified by x-radiography. The relatively large number of such objects recovered justifies discussion based upon functional categories as defined by Crummy (1983). Category 1: Personal Adornment and Dress: there are only 23 hobnails (Manning Type 10) and one cleat, which might have been used to strengthen the heels or soles of boots (Manning 1985, 131), belonging to this category. A further cleat was too corroded to allow certainty as to its function. Category 4: Household utensils and furniture: three tacks, two with flat polygonal heads and one with a domed, hollow head, may have been used as box fittings or possibly in upholstery work. Category 10: Tools: a chisel with a square shaft, tapering to the rectangular point, and a battered, slightly trapezoidal head as well as a hinge and part of one arm, or jaw, of tongs were found in the colluvial layers of Trench D. As neither was found in sealed contexts, it is impossible to date them securely to the Roman period. Category 11: Fasteners and Fittings: a flat thin cramp and acramp-like fitting are the only objects in this category that are not nails. The fitting has been included here because the stems on either side of its longish oval body are worked in a way that is found e.g. in rivet stems of belt fittings. Alternatively, it may have been used as a cleat for boots. Manning Nail Type Type 1b Type 2 Type 7 Type 8 Type 10 Indeterminate Modern Fig 15 Histogram of nail types from the 2003 excavation Seventy nails, other than hobnails, were found, of which 63 are likely to be Roman. They have been preliminarily sorted according to the typology suggested by Manning (1985, 133 ff. Fig. 32); however, their length has not been considered in this process. Fig 15 compares the number of nails by Manning type. For ease of comparison, this graph also includes types included in categories 1| and 4. As is the case with the nails included in the Owmby Nails Pilot Study (Graham and Cox 2001), the majority of the diagnostic nails are of Type 1b and 10. Type 1b nails will in most cases have had some use in a building, the longer ones for joining structural timbers, the smaller ones for fixing roofing or cladding. Category 15: Metal Working: five pieces of iron working slag were recovered from the probable occupation layer (11) in Trench A, as were three pieces of iron bar which may have been scrap for re-use. One had both ends deliberately broken off. This layer also produced one piece of vitrified clay (possibly a crucible) as well as many of the iron nails, hobnails and tacks. Stone objects Among the stone objects are two pieces of a collar-like object of as yet unknown use and date. Both pieces fit together but were found in different contexts (100) and (101) in Trench B. The remainder of the stone objects were Pennant sandstone roofing slates, some of which featured nail-holes. Five pieces of flint were examined, all of which were unworked. Glass by Sarah Fennings Window glass was recovered from 5 contexts, but most of this was modern. Two fragments from the 184 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 4. Trench A: Identified specimens by taxa and context Taxa/Context A005 Cattle (Bos taurus) Sheep (Ovis aries) Sheep/Goat (Ovis aries/Capra hircus) 3 Pig (Sus scrofa) Equid (Equus sp.) Red deer (Cervus elaphus) Mouse (Mus/Apodemus spp.) Chicken/Guinea fowl/Pheasant (Gallus gallus/Numida meleagris/Phasianus sp.) Duck (Anas sp.) Raven (Corvus corax) Eel (Anguilla anguilla) Large mammal Medium mammal topsoil in Trench B, one of which had a fire-rounded edge, the other pale blue, might date to the 19" century. Contexts 11, 12, and 303 all contained a single small fragment of vessel glass. These fragments were all from thin walled vessels and were made from shades of very pale natural bluish-green glass, but were not all the same colour. A small fragment of glass was recovered from context 203. This was very dark blue in colour and had one fire-polished surface. It was not from a plain vessel body, but it is too small to ascertain its shape. There is nothing to indicate that these pieces are not Roman. Dark or deliberately coloured Roman glass is more common in the 1* century AD. Vertebrate remains by Andy Hammon The evaluation produced a small vertebrate assemblage of 1,141 fragents (Hammon 2003). Hand-retrieval produced 500 fragments, while the remainder derived from sample residues (washed over a4mm mesh). All fragments were generally well preserved; characterised by fragments with minimal exfoliation or abrasion to their original surfaces. Fragmentation was moderate. A total of 139 fragments were identified to species or taxonomic group. Trench 1 was the only area in which intact stratigraphy was sampled, and these deposits (contexts 010 and 011) produced the largest quantity of bone (Table 4). A007__A009. A010. AOll A012 Total i 7 21 l 30 l l 7 1 10 3 1 9 40 3 59 l 1 5 4 ll 1 2 3 l 1 4 4 1 l 1 1 l l 1 1 l l 2 1 l 6 l 22 83 10 125 The assemblage was dominated by the major domesticates. Sheep/goat predominated (only sheep was positively identified), followed by cattle, pig and then equids. Single specimens of red deer, duck and raven were recovered from Trench 1. A range of smaller species was also recovered, which reflects the comprehensive sampling strategy. This included microfauna (e.g. mole and mice) and fish (eel). The assemblage appears to represent typical Romano-British material: species composition, body part distributions, mortality data, butchery evidence, etc. The vertebrate assemblage is too small to attempt any detailed reconstruction of the site economy and husbandry practices, though the range of species encountered combined with anatomical distributions, fragmentation and butchery evidence would indicate that the majority of bones derive from normal butchery and kitchen waste. The animal bone assemblage essentially represents occupation debris. Discussion by Tony Wilmott As a Roman small town, the site of Whitewalls fits into Burnham’s (1987, 157) third category of the limitations of evidence on such sites: ‘Those small towns where both aerial photography and excavation have been extremely limited, and which are consequently only known from the distribution THE ROMANO-BRITISH ROADSIDE SETTLMENT OF WHITEWELLS, EASTON GREY 185 of debris during fieldwork.....’, though it might be argued that the work reported in this article does take us a little beyond this. The non-invasive work, review of earlier references and finds, and the limited excavation now enable us to state with some confidence the nature and extent of the Whitewalls site. Itis clear that the sheltered river valley attracted human activity from an early date, and fieldwalking has demonstrated human presence during the Mesolithic period. Earthworks and cropmarks which differ in alignment to the Roman Fosse Way suggest pre-Roman, possibly Iron Age settlement, and this may be supported by the chance find of two Iron Age coins. Both are silver Dobunnic coins and consist of an ‘irregular’ issue, and a regular coin of BODVOC. This is the only site where both types occur together (Robinson 1977; Boon 1983). It is possible that the early alignments persisted into the Roman period. No evidence for habitation has been found, though activity is attested by the lynchets of early fields and crop marks of enclosures. It has been suggested that there might have been an early Roman fort on the Whitewalls site. Cunliffe (1973, 440) suggests that such an installation ‘might be expected’ and Webster (1980, 151, 161) places it on his map of conquest Britain as a ‘possible fort site with slight evidence’ on the basis of the early military martingale mentioned above (Griffiths 1982) and early pottery identified by Mrs Swan. There is, however, no evidence beyond these objects. If there was a fort, it would almost certainly not have occupied the narrow, steep river valley, but the flat land on either side. The construction of the Fosse Way as a major route is the key to the development of the settlement. The road descends the hill slope on the north side of the valley, and appears as a substantial earthwork bank running across the flood plain to the river crossing, probably originally a small bridge. Excavation has shown that since Roman times alluvium has been deposited to considerable depth on the flood plain, and it seems likely that this accumulated against the sides of the earthwork, suggesting that the Fosse Way was built on quite a substantial embankment, possibly in order to lessen the slope on the south side of the river, rather than to protect the road against flooding. The evidence of earthworks, aerial photography, the spread of dark soil and the distribution of Roman artefacts are consistent in suggesting that Whitewalls was a linear setthement, with actual occupation concentrated on the Fosse Way frontages; an example of Burnham’s (1987, 162) Group II small towns. Occupation seems to have been limited to roadside strips 50-70m in width. Further development along side-streets, often the next stage in such settlements, does not seem to have occurred here. This does not necessarily mean that the setthement was particularly poor either materially or in terms of building types; Burnham gives, as a typical example of the type, the site of Letocetum (Wall, Staffs), which stretches up to 1.5km along Ryknield Street, rarely extends more than 100m from the street frontage, and which possesses a well appointed mansio with a bath-house. At Whitewalls the earthwork evidence suggests the existence of buildings within enclosures on both sides of the road. The made ground recorded by Passmore on the flood plain consisting of “chiefly flat stones in which remains were found” seems consistent with the findings in the excavation, where the walls were found to be constructed of flat stones, and it seems likely that the pipeline was cut through the remains of Roman stone buildings buried in alluvial deposits. The idea that the settlement comprised stone-built roadside buildings is confirmed by fieldwalking finds, which included both pennant sandstone and ceramic roofing materials and also flue ules which suggest the presence of heated rooms in some buildings. The excavation shows that the river’s edge was used during for a large stone-built structure. As Figs 7 and 10 show, all of the walls found during the excavation stand at right angles to each other and therefore probably belong to the same structure, though the nature and plan of this structure could not be established. It is probable that the supposed ‘pier’ found by the river in 1930 (Fig 4) was actually part of this building. Suggestions made during the excavation were that this might have been a bath-house, or possibly a mansio sited in a sensible position adjacent to the the Fosse Way and the river, right on the river crossing. The fact that Trench B showed that a stone wall had been constructed across an existing stone surface suggests that the building was of more than one phase. In Trench A, a room seems to have been defined by a pair of stone walls set at right angles. Within this room was a deposit of stone and burnt material, which may have represented the destruction of the building. There is a possible alternative interpretation as one of the walls may not have been structural, and the possibility exists that this was in fact part of a channelled hypocaust system. The kind ot settlement nucleated roadside 186 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE represented by Whitewalls is a type now known to be relatively common in Roman Britain. There are several similar sites along the Fosse Way in Somerset, including Camerton (Wedlake 1958) and Fosse Lane, Shepton Mallet, which is perhaps the most comprehensively published example (Leach 2001). Fosse Lane, Shepton Mallet is, like Whitewalls, sited adjacent to a river crossing (here the Sheppey) on the Fosse Way. Extensive excavations on this site concentrated on the rear parts of plots which fronted on the Fosse Way. As at Whitewalls, an earlier alignment which ran at an angle to that of the Fosse Way was present. This was perpetuated by a side street in the early phase of the site’s development, and remained an aspect of its plan even after a major reorganization in the early 2nd century. The reorganization consisted of the imposition of a system of unitary boundaries comprising drystone walls and ditches, which defined rectangular compounds laid out at right- angles to the Fosse Way (ibid, 308). Within this framework differing styles of buildings, including timber buildings, stone-founded aisled halls, and a later trend for mortared stone buildings, developed in a variety of ways, presumably dependent upon the wealth and status acquired by their owners. To the rear of the Fosse Way frontage there were separate properties containing stone-founded buildings. The maximum period of development and prosperity came during the late 3rd century, and throughout the 4th century. It seems likely that the type of development seen at Shepton Mallet was also the case at Whitewalls. Though the finds assemblages recovered are small, they indicate a history spanning the entire Roman period. Pottery types derived from fieldwalking are similar to those from controlled excavation. Though these are predominantly 2nd — 4th century in date, a small quantity of Ist century material has been found (Shipp 2000, Webster 1980). Late Roman occupation is shown by the presence of late Roman shelly ware from both fieldwalking and excavation. This date range is confirmed by the evidence of coinage. Moorhead (2001) has quantified coin finds from the site by period. This list, together with material published by Passmore (1921, 1933), a single coin derived from fieldwalking, and coins from the 2003 excavation is the basis for Fig 16. The pattern of coin finds from Whitewalls is typical of the general Wiltshire pattern with very few finds before the mid- 2nd century (Moorhead 2001, 89). Moorhead quotes Whitewalls as one of the sites with a higher proportion of early coins than the 110— 100 90— 80— 70— 60— fo] Coin Numbers 50 te i ") | | | | | a | | | B E UW ee le ie i | \ ies f { = | i ii! — } ea \ Geer -. i} eer | a | | {| | Se it =} \ if 1 Due Go “401112 13-14 15 16 17 18 19 20 21. Periods (Reece 1973) Fig. 16 Histogram of coin finds from Whitewalls norm for the county, citing its position on the major route as the reason. Moorhead’s statistics, however, are based on surface and detected finds, while the coins published by Passmore doubles his numbers of coins dating to pre- AD 43 and to the Flavian period. The increase in coin loss c. 260 is common to other large Wiltshire sites, and is general for Romano-British urban sites (Reece 1987, 93). Coinage loss from this point broadly mirrors the national patterns within towns of all sizes. The pottery from the excavation has proved a useful initial indicator of the range and quality of the occupation if the presence or absence of fine and specialist wares reflects the status of the site, although the larger urban settlements received a more varied range of material when compared to rural sites. The quality, range and type of the Roman assemblage from Whitewalls is indicative of moderately high status with access to various traded wares. In particular, samian accounts for 2% by weight, 2.3% by count which is quite typical for rural settlements in the West but the presence of the other finewares and the Italian amphora give added status. This could be the result of the location of Whitewalls on the Fosse Way linking Cirencester with Bath. The proportion of Dorset black burnished ware is very high accounting for 49.4% by weight which may be linked to trade with Cirencester which was also receiving significant quantities of this ware. In summary, the assemblage is entirely consistent with what one might expect THE ROMANO-BRITISH ROADSIDE SETTLMENT OF WHITEWELLS, EASTON GREY 187 from a Roman small town established from the 2nd century and located on a major routeway. The animal bone suggests a diet in which the major domesticated species were augmented with wild game and river fish. Although small in scale, the 2003 excavation provided hints as to the longevity of the site. In Trench A, the black material (011) contained pottery, almost 70% of the total recovered, dating to c. 360- 400+ and coinage dating to c. 340. The deposit also contained several pieces of smithing slag and a possible crucible fragment. This layer was cut by a linear, stone-filled and -lined slot. This appears to have been a timber slot, which respected the stone walls of the room. This was sealed by a further layer (10) which contained pottery of the same date as (11); the second largest group of material recovered. Such evidence has been found elsewhere in Wiltshire. At Castle Copse, Great Bedwyn, the evidence was reminiscent to the author (TW) of findings on that site, where late Roman/ early post- Roman timber features were built within rooms ofa former grand house. Here stakeholes were cut into the mortar bedding for a mosaic after the tesserae were removed. Contemporary or later were parallel timber slots, which respected the alignment of the stone walls of the Roman building. Associated finds indicated that these former high-status buildings were used for metal working in the very latest Roman or earliest post-Roman period. This occurred before the robbing of the stone building (Hostetter and Howe 1997, 91-2). The presence of a coin of Arcadius and late 4th- early 5th century Overwey ware together with an unusual late-4th century antler composite comb from terminal Roman deposits on the south wing of the villa further suggests continuity of occupation on this site. At Shepton Mallet, already cited as a similar Fosse Way settlement to Whitewalls, there was also some indication of early post-Roman continuity, though this evidence was greatly enhanced by the presence of burials from which radiocarbon determinations could be derived (Lbeach 2001, 287). Relatively litth is known of the historic landscape context of the site. To the east, it has been suggested that the hill upon which Malmesbury stands was an Iron Age hillfort (Haslam 1984), an idea given more credence following survey work by Corney (1998) and apparently confirmed in recent excavation (Wilts C.C. 2000, 8). Roman occupation can be suggested in Malmesbury on the basis of the discovery of a hypocaust floor in 1887 (Arch J 1887, 53) and of Roman material in a variety of locations within the town (Currie et al 1993; Wilts C.C. 2000, 8-9). In addition, recent work by the Cotswold Archaeological Trust (Colls et al, 2002) identified a Romano-British complex to the north- east of Malmesbury at NGR ST944884. Evaluation suggested that this was a villa site, occupied from the 2nd to the 4th century. The site seems to have been well appointed, with tesserae from mosaic pavements recovered with other building materials such as stone roof tiles. The excavated remains consisted of the surrounding agricultural field systems, enclosures and trackways. Only 1.4km to the south east of the site, on the southern side of the river is the important 6th-7th century settlement of Cowage Farm, Foxley (Hinchliffe 1986), a high status early medieval site which seems to include an early church. The existence of this important site so close to Whitewalls fuels interesting questions concerning the nature of the post-Roman transition in this part of Wiltshire. Conclusions This report demonstrates how major advances in knowledge of a site like Whitewalls can be made through small scale work. An outline of the site’s history and development can now be offered. It seems that the Avon valley was occupied in the pre- Roman Iron Age, though in the immediate vicinity of Whitewalls this took the form of agricultural activity attested by cultivation lynchets, boundaries and trackways. It is possible that this landscape was associated with the hill fort settlement at Malmesbury. The site developed as a result of the imposition of the Fosse Way as a major route, and its position on a minor river crossing. It is possible that a short-lived fort occupied the site during the conquest period, though the evidence for this has yet to be found. The site probably began to develop quite early as a roadside setthement comprising buildings ranged along the roadside. The settlement never developed a more complex plan form, though its structures may have been fairly sophisticated, including stone buildings, possibly with heated rooms. The site was not a poor one, and benefitted from the use of a wide variety of pottery types. It probably derived its prosperity from its position on a route between the major centres of Corinnam and Aquae Sulis, and it is possible that its facilities included a mansto, possibly represented by the building excavated in 2003, The small town clearly 188 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE existed within a broader Romanised landscape, with at least one other Roman complex to the east, north of Malmesbury, and Roman settlement at Malmesbury itself, though its position within the local and regional settlement hierarchy cannot be determined. The site seems to have continued to prosper up to the end of the 4th century, and the evidence of the late timber slot within the building excavated in Trench A might indicate continuity into the 5th century. If so, interesting questions can be raised concerning the place of Whitewalls in the immediate post-Roman history of the area, and the development of the early medieval landscape exemplified by Cowage Farm Foxley, and the early development of Malmesbury. Acknowledgements Thanks are due to the owner of the site Mr J. Tremayne, and to the Inspector of Ancient Monuments, Phil McMahon. Deborah Shipp wishes to thank landowners Mr Robin Ingall, Mr David Willis, Major Turner and Mr Herbert Irvine; the staff at the SMR, NMRC, Malmesbury library, Devizes museum, and Mr Paul Robinson; and also Mr Michael Stone, Mr and Mrs Cragg-Barber, Mr and Mrs Eckett, Mrs Zoe Hawkins, Mrs Jane Pilbeam, Mr Alex Shipp, Mrs Jean Slater and Mr Peter Smith. Messrs Bailey allowed the use of their premises for the excavation base which was set up by the English Heritage logistics team of David Webb and Philip Wimbleton. The 2003 excavation was directed by Tony Wilmott, and thanks are due to the excavation team; Ben Moore (supervisor) who also compiled the stratigraphic narrative, Tracey Clark (finds and environmental supervisor), Liz Chambers, Sarah Cross, Anna Doherty, Ruth Duckworth, James Moore, Geoff Morley, and Catherine Rackham. On-site advice on the nature of the soils and geology was given by Matthew Canti. Finds work was co-ordinated by Jorn Schuster, and specialist work was undertaken by Jane Timby, Richard Brickstock, Sarah Jennings, Phil Harding, and Andy Hammon. Neil Holbrook supplied information on the work of the Cotswold Archeological Trust. Illustrations were drawn by Chris Evans of the English Heritage Centre for Archaeology Graphics Studio. References ARCH. J., 1887. Archaeological Journal, 44 BOON, G. C. 1983. A Dobunnic Note. Seaby Coin and Medal Bulletin 778, 145-6 BROOMHALL, D. and HENIG, M. 1981. An iron signet ring from Easton Grey. WANHM 76, 167-8 BURNHAM, B. C. 1987. The morphology of Romano- British small towns. Archaeological Journal 144, 156-90 CASTLE, S. A. 1972. A kiln of the potter Doinus. Archaeological Fournal 129, 69-88 COLLINSON, J. 1791. The History and Antiquities of the County of Somerset, Vol 1. London COLLS, C., HART, J., and COLLARD, M. 2002. Easton Grey to Minety High Pressure Natural Gas Supply Pipeline, Wiltshire; archaeological recording summary report. Cirencester: Cotswold Archaeological Trust (unpublished) CORNEY, M. 1998. Nun’s Walk, Malmesbury: an analytical archaeological earthwork survey. unpublished report 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. WANHS: Devizes CRUMMY, N. 1983. The Roman small finds from excavations in Colchester 1971-9, Colchester Archaeological Report 2. Colchester: Colchester Archaeological Trust CUNLIFFE, B. W. 1973. The Period of Romanisation. Victoria County History; Wiltshire vol I, Part 2, 439-52. London: VCH CUNLIFFE, B. W. and FULFORD, M. G. 1982. Bath and the rest of Wessex, Corpus Signorum Impertui Romanorum I, Fascicule 2. Oxford: University Press CURRIE, C. K., CURRIE, C. T., LOFT, M. A. and ROBINSON, P. 1993. Excavations at the Postern Mill site, Malmesbury, 1986-87. WANHM 86, 58-74 ELLIS, PB 2001 Roman Wiltshire and after: papers in honour of Ken Annable, WANHS: Devizes GEDDES, I. 2000. Hidden depths, Wiltshire’s Geology and Landscapes. Swindon GODDARD, E. H. 1913 List of Prehistoric, Roman and Pagan Saxon Antiquities in the County of Wiltshire Arranged under Parish. WANH™M 38, 153-378 GRAHAM, K. and COX, D. 2001. Owmby Nails Pilot Study: Owmby, Owmby-by-Spital, Lincolnshire, CfA Report Series 26/2001, Portsmouth: English Heritage GREEN C. M. 1980. ‘The Roman Pottery’, in D. M. Jones, Excavation at Billingsgate Buildings, lower Thames Street, London, 1974, London and Middlesex Archaeological Society Special Paper 4, 39-80. London: LAMAS GRIFFITHS, N. 1982. Early Roman military metalwork from Wiltshire. WANHM 72, 50-59 GRIFFITHS, N. 2001. ‘The Roman army in Wiltshire’ in P. Ellis (ed), Roman Wiltshire and after: papers in honour of Ken Annable, 39-72. WANHS: Devizes GRINSELL, L. V. 1957. ‘Archaeological gazetteer’, Victoria County History; Wiltshire vol I, Part 1, 21-279. London: THE ROMANO-BRITISH ROADSIDE SETTLMENT OF WHITEWELLS, EASTON GREY 189 VCH HASLAM, J. 1984. "The towns of Wiltshire’, in J. Haslam (ed) Anglo-Saxon towns in Southern England. Chichester: Phillimore HENIG, M. 2001. ‘Art in Roman Wiltshire’ in P. Ellis (ed), Roman Wiltshire and after: papers in honour of Ken Annable, 107-126. WANHS: Devizes HINCHLIFFE, J. 1986. An early medieval settlement at Cowage Farm, Foxley, near Malmesbury. Archaeological Journal 143, 240-59 HOSTETTER, E. and HOWE, T. N. (eds). 1997. The Romano-British villa at Castle Copse, Great Bedwyn, Bloomington: Indiana University Press KIRKHAM, G. 1997A survey of earthworks at Easton Grey, Wiltshire, Royal Commission on Historic Monuments (England) survey report 1997.076, unpublished LEACH, P. V. 2001Fosse Lane Shepton Mallet 1990; excavation of a Romano-British roadside settlement in Somerset, Britannia Monograph Series 18, London: Roman Society MANNING, W. H. 1985. Catalogue of the Romano-British iron tools, fittings and weapons in the British Museum. London: British Museum Press MOORHEAD, T. S. N. 2001. ‘Roman coin finds from Wiltshire’ in in P Ellis (ed), Roman Wiltshire and after: papers in honour of Ken Annable, 85 - 106. WANHS: Devizes PASSMORE, A. D. 1921. Notes on Roman finds in north Wiltshre. WANHM 41, 389-95 PASSMORE, A. D. 1933. Roman Remains from Easton Grey. WANH™M 46, 270-2 RIVET, A. L. E and SMITH, C. 1979. The Place-names of Roman Britain. London: Batsford ROBINSON, P. 1977. A local Iron Age Coinage in Silver and perhaps Gold in Wiltshire. British Numismatic Journal 47, 5-19 ROBINSON, P. 2001. ‘Religion in Roman Wiltshire, in P Ellis (ed), Roman Wiltshire and after: papers in honour of Ken Annable, 146-158. WANHS: Devizes SHIPP, D. 2000. An archaeological field study of the earthworks on the Fosse Way at Whitewalls Wood, Easton Grey, Wiltshire. unpublished GCSE A-level dissertation, Wiltshire College, Chippenham, unpublished SMITH, R. E 1978. Roadside Settlements in Lowland Roman Britain, Brit Archaeological Reports, 157, Oxford: BAR SOIL SURVEY OF ENGLAND AND WALES, 1983 Legend for the 1:250,000 Soil Map of England and Wales: a brief explanation of the constituent soil associations (Harpenden) THOMAS, R. J. C. 2003. PoW camps; what survives and where. English Heritage Conservation Bulletin, 44, 18-21 TOMBER, R. and DORE, J. 1998. A National Roman Fabric Reference Collection: a handbook. London: Museum of London Archaeology Service/English Heritage/British Museum WANHM, 1998. Excavation and Fieldwork in Wiltshire in 1997. WANHM 92, 134-141 WEBSTER, G. 1980. The Roman Invasion of Britat., London: Batsford WEDLAKE, W. J. 1958. Excavations at Camerton, Somerset. Bath WILMOTT, T. 2003. Easton Grey, Wiltshire: project design for evaluation. English Heritage Centre for Archaeology, unpublished. WILTSHIRE COUNTY COUNCIL. 2000. Malmesbury; an extensive urban survey, Wiltshire County Council, unpublished Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 190-203 A Romano-British prone burial from Bratton, Wiltshire by Stella Maddock! and Pamela Mahon? In October 1955 human bones were unearthed beside watercress beds at Luccombe, Bratton in Wiltshire (reported variously as NGR 923519 and 92235204). Subsequent excavation revealed the remains of a single adult prone inhumation with the arms flexed behind the torso. Only the upper half remained, but the burial position has led to speculation of disrespectful disposal, perhaps the result of a crime or execution. Possible associations with local Anglo-Saxon battles had been postulated. However radiocarbon dating has established that death occurred during the Romano-British period. Other burials are rumoured in the vicinity, but as yet no other skeletons have been discovered. INTRODUCTION Project background In 2001, Mr Hugh Seymour of Luccombe Mill, Bratton approached the Society (WANHS) with a request for further investigation of a burial excavated on his land in 1955 and reports of other burials in close proximity. The WANHS Archaeology Field Group (AFG), with the support of the Archaeology Committee, agreed to undertake further research and this article describes the results to date. Location The burial site is in a narrow valley at the lower (south-western) end of Luccombe Bottom, a large amphitheatre-like combe that cuts into the north- facing chalk escarpment of Salisbury Plain, near the village of Bratton in the vicinity of NGR 92215205. The precise location of the burial site is uncertain as there is a difference between the grid reference published by the excavator in 1955 (NGR 923519) and another from an Ordnance Survey Field Investigator’s recording on the National Monuments Record (NMR) in 1968 (NGR 92235204). The site is on the upper greensand, close to Luccombe Springs, in a wet area now overgrown with trees and scrub and the remains of watercress beds. The area around the burial site is named “Danes Ley’ on the 1:2500 Ordnance Survey (OS) map. Land immediately to the south-west of the site rises very steeply to a field currently laid to pasture. The area around the site has considerable atmosphere, being watered by a number of hillside springs, both seasonal and permanent, whose output forms two streams that unite into the main stream now called Stradbrook. The boundary between Bratton and Edington parishes follows this stream very close to the burial site (Figure 1). THE BURIAL Discovery in 1955 In October 1955, men employed by Ronald Pearson of Baynton House, Coulston, unearthed human bones whilst working at his watercress beds at Luccombe, Bratton. The then Assistant Curator of Devizes Museum, Ken Annable, went to the site on October 28th 1955 and excavated the remains of a single inhumation of which only the upper half remained, the rest having been inadvertently destroyed by Mr Pearson’s workmen. 'Emms House, Emms Lane, Bratton, Westbury, Wiltshire BA13 4SA ?Saxholm Dale, Bassett, Southampton, SO16 7HA A ROMANO-BRITISH PRONE BURIAL FROM BRATTON, WILTSHIRE 191 A e Malmesbury Wiltshire © Melksham e@ Westbury @ Tisbury Edington Bratton @ Luccombe Springs # Burial site of BRA 495 Fig. 1. A shows the location of the area around Luccombe within Wiltshire. B shows the inferred location of the burial at Luccombe which is close to the village of Bratton Description , According to Annable (1955, 190), the skeleton was *...In an extended position and oriented NW/SE with the skull to the NW. It lay just on top of the Greensand subsoil at a depth of 3ft 8ins below surface Coyil Socnonee The skull lay face downwards in the soil and ... the position of the upper arm bones clearly indicated that at death both arms were positioned behind the back. ...... No visible trace of wounds to skull or bones was evident. The skeleton was that of a male of mature age. A single thick sherd of hard, buff fabric was the only associated find. This was situated at the distal end of the left humerus.’ (Figure 2). However, the photograph clearly shows the pottery sherd to be at the distal end of the right humerus as the skeleton is prone in position. It is also unclear if the sherd was within the same grave infill as the skeleton, or outside the margin of the burial. In 1955, both skeleton and sherd were deposited at the museum in Devizes where they have remained. The skeleton is now designated as “BRA 495” and the sherd has since been identified as Romano-British. Annable’s account also notes that since his excavation a further two skeletons ‘.....have turned up within a foot or so of the original discovery. These still remain in the ground though not in imminent danger of destruction.’ In May 2003, and again in May/June 2005, members of the AFG carried out several small investigations of the potential site hoping to find evidence of the 1955 excavation and/or further burials but nothing conclusive was found. Documentation relating to these investigations, which includes a limited resistivity survey, is currently held by Stella Maddock (AFG). Locating the alleged additional burials is difficult for a number of reasons. Firstly, the terrain has changed significantly since the original excavation in 1955. Although a few of the wooden sleepers delineating the edges are visible, the watercress beds are now overgrown with trees and 192 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE other vegetation; the flow of water from the springs ~ through the area has altered as it is no longer used to feed the beds; beech trees have been planted on the slope above and are now mature; and the profile of the banks and slope has been altered by the build-up of humic material from vegetation and soil slippage caused by natural erosion and badger workings above (Figure 3). As a result, it is difficult to ascertain where the men maintaining the watercress beds in 1955 are likely to have been digging when they found the skeleton and how that could fit with the depth, level, and orientation of the burial as reported by Annable. Secondly, the grid reference given by Annable in his account and the one recorded later by the OS Field Investigator, (probably based on a position indicated by the landowner), conflict. There is now no first-hand memory of the location of the burials(s). The burial is recorded on the County Sites and Monuments Record (SMR) as three inhumation burials (ST95SW552) and on the NMR as ‘Seven inhumation burials, date uncertain’ (ST95SW21). Information about the four additional burials is attributed to Lady Seymour, the then landowner, but could not be corroborated by the OS Field Investigator in 1968. Fig 2. Photograph taken before excavation of BRA 495 in situ. Note the position of the skeleton, with its prone position and arms flexed behind the torso. The piece of buff pottery, identified as Romano-British, is alongside the lower end of the right humerus (arrow). This photograph was taken by Ken Annable and given to Lady Seymour, mother of Hugh Seymour and reproduced with his permission. Fig 3. The site as it is today, overgrown with trees down the steep slope leading to the abandoned watercress beds at the bottom left of the photograph. Photograph by Stella Maddock. A ROMANO-BRITISH PRONE BURIAL FROM BRATTON, WILTSHIRE 193 THE SKELETON Osteological Report by Pamela Mahon Condition of the skeleton in November 2002 In 2002, Mr Seymour commissioned an osteological report of BRA 495 and these are the findings. The skeletal elements of BRA 495 were generally in good condition, but many of them had not been cleaned. Clay was caked over many bones, which had to be carefully removed before examination. There were several cut marks on long bones, which were almost certainly caused by the sharp spades used by the workmen in 1955. Where the femora had been truncated, sharp cuts across the bones bear testimony to this post-depositional disturbance. The bones had been saved damage by tree roots and gnawing by animals, probably because of the depth at which they had been buried. Representation of skeletal elements On examination it was assessed that 70% of the skull, facial bones and mandible were present, most of the teeth in situ. The vertebral column was present in its entirety, although there was damage to the spinous processes which had been uppermost, closer to the surface, and therefore more likely to be affected by pressure from the weight of earth above it. Parts of the pelvic girdle and upper femora were also missing, as was the lower left arm and hand. The right arm was complete with the exception of most of the hand bones and scapula. Approximately 80% of the ribs remained, but the sternum was absent. Both clavicles were present. No lower femora, lower leg bones, foot or toe bones were represented. In all only 55% of the total skeleton was present (Figure 4). Small bone representation was poor and repackaging of the skeleton in newspaper bundles, carried out in 1975, may have led to further bone element loss. The quality of the photograph taken at the time of discovery renders it impossible to equate how much has been lost since the bones were lifted (Figure 2). The bones were cleaned (dry brushing), measured, annotated as BRA 495, photographed, rebagged and boxed to prevent further loss. Some long bones were reconstructed using water-soluble glue, prior to repacking, where there had clearly been a post-depositional break. This Right -) LY ~~ ~~, fab LS —L Anterior view of BRA 495 Fig 4. The skeletal elements of BRA 495. The shaded areas represent bone recovered, approximating 55% of the entire skeleton. enabled measurements to be taken for the estimation of stature. Approximately 60% of the cranium had been reconstructed at an earlier date using a hard adhesive. In places this was more resilient than the bone, allowing further cracks to appear. The new reconstruction had been allowed to dry in an unsupported position and had unfortunately contorted before it was set. Therefore many cranial measurements were not possible due to the deformation. Assessment of sex, age, stature estimation, dentition, pathology, environmental stresses and other findings Sex of the indretdual: From the pelvic girdle it could be seen that this individual was male (Bass 1995), 194 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE although the skull and mandible were gracile and feminine in comparison. The skull and mandible did not display the classically ‘male’ appearances of robust muscle insertions and angular jaw line, but overall the size of the skeleton indicated a male individual. Age at death: All of the bone epiphyses present were fully fused, indicating an adult of more than 25 years at time of death (Brothwell 1981). By assessing tooth wear a more specific age between 25 and 35 years can be estimated (Brothwell 1981). By pelvic features, such as the auricular surface (Lovejoy et al 1985), the age estimation can be narrowed down to 31-35 years of age at death. Stature estimation: 1.70 — 1.72m (5ft 6in — Sft 7in) from Trotter and Gleser 1952 and 1958 (using right humerus and right radius). Dentition: The teeth were well represented with only 5 teeth of the full adult complement of 32 being missing. One was lost ante-mortem (upper right 8, wisdom tooth) and two missing with a portion of missing maxilla (upper left 7 and 8). The lower left canine and upper left incisor were post-mortem losses. The teeth appear moderately worn, with areas of exposed dentine. The pattern of wear indicates a coarse diet, high in carbohydrates. This is shown by the moderate to advanced calculus on upper and lower lingual and buccal surfaces (Dobney and Brothwell 1987). Within the upper left 6 there is a large cavity involving half of the crown. The anterior portion of the root of this tooth is exposed through an open cavity within the bone of the maxilla, caused by a root abscess (Figure 5). Pathology and environmental stresses: Tooth abscess, wear and calculus suggest a ‘farinaceous diet’ (Annable 1955, 190), full of soft starchy, carbohydrate components such as bread. Minimal pathology was noted in the remainder of the skeleton. Within the spine there were 2 thoracic and 2 lumbar vertebrae which displayed Schmorl’s nodes, indicating ruptured intra-vertebral discs, which can result from heavy stresses on the spinal column (Brothwell 1987). These could have been caused by a heavy lifting occupation, such as building or farming. No arthritis, bone infection, indicators of interrupted growth, evidence of malnutrition or bone disease were noted. A small bony growth (an exostosis) was present on the medial end of an upper left rib. This may be due to an over-developed muscle insertion caused by high physical upper body stresses (Rogers Anterior “ Cig oj Cavity in tooth crown i y Right Fig 5. The maxilla (upper jaw) of BRA 495 showing the bone sinus created by an abscess at the root of the upper left 6 pre- molar (A). This tooth has an extensive cavity (B). Photograph by Pamela Mahon. et al 1997; Larsen 1997). The Schmorl’s nodes in the mid to lower spine would also indicate this type of activity. The bones of the right upper limb were generally longer than those of the left, which showed increased bone mass in comparison to the right. There is also a very rugose muscle insertion in the left humerus at the site of the biceps muscle (bicipital groove), also indicating heavy use of this limb. The effect was not so apparent on the right humerus. This can be seen in some skeletons where one arm is favoured for physical tasks over the other. The more used arm develops larger bone mass than the bones of the opposing side, which lengthen in compensation (Steele and Mays 1995; Latimer and Lowrance 1968). It may be hesitantly suggested that this individual could have been left-handed. Other findings: Several more recent areas of damage to the skeleton are visible. For example on the remaining portions of the upper femora, fresh, lighter slice marks probably made by a trowel or A ROMANO-BRITISH PRONE BURIAL FROM BRATTON, WILTSHIRE 195 Femoral head Cut surface Fig 6. Photograph showing the detached femoral head of the femur, with its sliced lateral margin, probably caused by digging around the skeleton. The circular facet seen to the left of the centre 1s the fovea, where a strong ligament attaches the femoral head to the acetabulum (socket joint) within the pelvis. Photograph by Pamela Mahon. shovel can be seen cutting into the darker, degraded bone cortex. The head of the left femur had also been sliced in a vertical direction (Figure 6). After cleaning, it could be discerned that no healing had taken place as earth was impacted into the small cavities of the internal cancellous bone. Thus it is most likely that the slicing of the femoral head must have taken place after death and not as a result of excavation damage. Therefore, this clean and unusually placed cut must have happened after the body had been placed in the ground. Unfortunately, the missing part of the femoral head was not amongst the assemblage recovered. No other evidence of peri-mortem trauma to BRA 495 was discovered. Further technical analysis After the osteological report, further investigation in the form of radiocarbon (C14) dating was arranged to ascertain the time period from which this individual came. Mr Hugh Seymour kindly funded radiocarbon dating, which took place in the University of Waikato Laboratories in Hamilton, New Zealand. In order to date the skeleton between, 100 and 200g of dry bone is required for the laboratory. Here, 103g of bone was cut from a broken piece of humeral shaft, measuring 3cm in length. In addition, 11 rib fragments were harvested and sent for analysis. Results The Waikato Laboratory ran contamination and preservation assays on the samples to assess the quality of bone gelatin. The percentage of carbon (C) and nitrogen (N) and the ratio of C:N were calculated on the pretreated bone of BRA 495. Modern collagen which contains gelatin has about 43% carbon and 16% nitrogen, and should have a C:N value of about 3.2. Most well preserved archaeological bone protein ranges between 11 and 16%N, with an average 35%C anda C:N ratio range of 3.1-3.5 (van Klinken, 1999). The samples from BRA 495 show 42% of carbon, 15.9% of nitrogen and a value of 3.1 for C:N ratio. This indicates that BRA 495 has quality indicator values that are very similar to modern bone gelatin and that contamination was minimal. Stable isotopes of bone gelatin (deltal3C and deltalSN) are commonly used to determine the types of food a person has been living on and were also calculated for BRA 495. Different foods have different d13C and d15N values and these are passed on to the bone protein of the animal that eats the particular food. The results for stable isotopes of BRA 495 were d13C = -19.44 and d15N = 9.32 which suggest primarily a terrestrial C-3 diet. Generally, human bone gelatin d13C values of -11 per ml indicate a diet composed almost entirely of marine protein (or a diet that has a lot of C-4 plants in the diet), while -21 indicates a mainly terrestrial diet (-16 indicates a 50/50 diet). The information gathered from these bones suggests a small marine component to the diet. The presence of marine food and the possibility of freshwater input (ie. freshwater fish) or C-4 plants such as grains/sugarcane, can sometimes make a significant difference to the Cl4 date (10- 20 years in some cases). As the sample comes from Wiltshire, there is a small possibility that fresh water fish could influence the Cl4 date obtained. Further refinement of the test results indicated that marine protein was not a major constituent in the diet of BRA 495, though the d15N value of 9.32 does suggest that the individual consumed a reasonable amount of animal protein. All the evidence suggests BRA 495 did not live close to the coast, at least during the last 10 years of life. C14 radiocarbon dating From the Cl4 data output (Figure 7) it can 196 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 2200BP r Wk15985 : 1847451 BP 68.2% probability f B80AD ( 4.8%) 100AD 2000BP 120AD (63.4%) 240AD, 95.4% probability 1800BP S50AD (89.6%) 260AD 280AD ( 5.8%) 330AD | 1600BP ae ey Radiocarbon determination 200CalIBC CaIBC/CalAD 200CalIAD 400CalAD — 600CalAD Calibrated date Fig 7. The calibrated output from the C14 radio-carbon dating indicating the 94 % probability that BRA 495 died within the range SOAD and 260AD. Further testing narrowed this range to 103AD+S1 years. The calibration has been reproduced by kind permission of Waikato Laboratories, Hamilton, New Zealand. that the preliminary age suggests the individual died between 100AD and 400AD. It was estimated that there was a 94% probability that death occurred between 50AD and 260AD. After further counts had been undertaken, a date range of 1847+51 BP was calculated, where BP (before present) refers to 1950, indicating a date of 103AD+51 years, putting the date firmly in the Roman occupation of Britain. HISTORY OF THE BURIAL SITE AREA by Stella Maddock Documented history As part of the project, the land use, ownership, and place-name history of the area around the burial site were researched using the resources of the Wiltshire and Swindon Record Office (WSRO), the County Local Studies Library in Trowbridge, the WANHS Library in Devizes, and the National Monuments Record Centre in Swindon. A summary of this research is presented here. The wooded area around the stream where the burial was discovered is named Danes Ley on the 1:2500 OS map. Other sources (eg. Crawford c.1930) suggest that the name Danes Ley (and earlier versions of the name) covered a wider area to include what is now the field immediately above the site to the south-west. ‘The Place Names of Wiltshire’ (Gover et al 1939, 145) states that ‘DANES LYE is Denelyghe, Deanelighe t. Eliz LRMB. “Valley ° clearing,” v. denu, leah’. The earliest documented reference to the site is acharter dated AD 968 in which King Edgar grants land at Edington to the nunnery at Romsey (Sawyer 1968, 244). The boundary of the land granted is the same as the current boundary between Edington and Bratton parishes and also the boundary of the Hundred and the ancient parish of Westbury as described in a Perambulation in 1575 (Hoare 1829, 54-57; WSRO MS 540/320). The Charter boundary begins ‘Erest on Mulbourne Ewelmen’ and Grundy (Grundy 1920, 80-84) translates this as ‘First to the (great) Springs of Mill Bourne’ and notes that aewielm implies a spring of large volume. The number of springs and their volume, even today when Wessex Water is extracting water, fit this description. Apparently, before the water works were in place, ‘the spring in Luccombe Bottom used to leap out of the hillside, six feet in the air’ (George 1997). Mil(1)bourne is the earlier name for the stream, which is now called Stradbrook, and was also the name of the hamlet that grew up along its banks. ‘The Place Names of Wiltshire’ (Gover et al 1939, 146) refers to the AD 968 charter under ‘Melbourne’ and gives the etymology of ‘on Mulbourne Ewelmen’ as: myln, burna = millstream and ewelmen is OE aewielm, ‘source’. Examination of the boundary described in the Charter and the 1575 Perambulation of the Hundred of Westbury, suggests Ewelmen was the area where the waters from the several springs met and flowed together. O.G.S. Crawford noted this spot on his map of c.1930. It is very close to the burial site. At Domesday, Westbury was a royal vill or township with six mills (unlocated) but the three areas of what is now Bratton — Bratton, Millbourne and Stoke — are not mentioned by name (Darlington 1955, 118). The Edington Cartulary incorporates various charters, deeds and other legal documents that shed light on land ownership in Bratton in the early 13" century (Stevenson 1987). The Milbourne/ Melbourne area of Bratton seems to have been particularly associated with the Dancy family (variously spelt, including Dansy, de Anasya, Danesy) whose manor was at Dilton. They appear to have rented out a mill and land in the Milbourne area to various members of the de Dene (variously spelt Dene, Dena, Deona, de Dena, de Deone) family (Stevenson 1987, 68, 71, 87; WSRO MS 929/53, 12; Morrison 1992, 13). A ROMANO-BRITISH PRONE BURIAL FROM BRATTON, WILTSHIRE 197 406 #16 Fig 8. Area around Luccombe from Ordnance Survey Ist edition (1887) scale 1:2500 (WSRO 45/6) showing the position of the suggested burial site for BRA 495. Reproduced by kind permission of the Wiltshire and Swindon Record Office. A deed of land transfer of this period concerns ‘Deonalegh’ (WSRO MS 929/53, 39). In 1332/33 (6 Ed. IID) an Indenture between Isabella (de Dene) and Nicholas Fitzwaryn refers to ‘half a parcel of land at _ “locland” viz. the southern part; .....and in “la Deone Legh” four parcels as they are bounded ...... >and other lands which once belonged to Walter de Deone and Sarah de Deone, and Agnes, Isabella’s mother in Mulebourne, Bratton and Westbury (Stevenson 1987, 74). “Locland’ might be Luccombe, and this reference to La Deone Legh is noted by Crawford on his map (Crawford c.1930). A relationship between person and place name is strongly suggested but which came first is unknown. The name resurfaces in the 16" century in the form ‘Deane Lee’. In the Chancery Proceedings for 1571, a dispute over land between the Whitakers (Wheatacres) and the Paties refers to “.....the reversion of one close of land with the appurtenances commonly called Deane Lee ... (WN and Q Volume III 1899, 157-9). A perambulation of the Hundred of Westbury undertaken in 1575 is documented in the form of a description of the boundary of the Hundred transcribed by Richard Harris (WSRO MS 540/320). This transcription is included in Colt Hoare’s ‘Modern Wiltshire’ (Hoare 1829, 54-57). The original source is not cited by Harris or Hoare. The part of the boundary that runs close to the site is described as follows: *... and from the Flood Hatch by the Ditch between Eddington field and the Mill Close on the River and up it to the Meeting of the Two Springs Anciently Called the EWELM (or Melm) then southward by a Strait line to a Stone Among the Springs lately renewd. And so up the Steep Hill to a Stone at the Crest of the Fill newly Set up .... The Whitaker family appears to have maintained an interest in the area around the burial site during the 17" and 18" centuries, the name being variously written as ‘Deanleigh’, ‘Deanly’ and ‘Dean Leigh Gate’ (Reeves 1978, 22; Reeves and Morrison 1989, 198 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE xv, lvili, map; WSRO MS 1195/22). Early in the 19" century Thomas Jarvis built a cloth mill at Luccombe (Rogers 1976, 231). The Enclosure Map and Award for Westbury dated 1808 does not have any information on Danes Ley, (the field or the wooded area), but the rest of the area around the springs is listed as a Public Watering Place named ‘Dean Lye Bottom Watering Place’ (WSRO EA/76). Colt Hoare’s account of King Alfred and the Battle of Ethandun (AD 878) in volume 1 of Ancient Wiltshire’, published in 1812, refers to ‘ ....a fine perennial spring called Lockham, near which the residence of the Danes is still commemorated in the name of a field called Danes Ley.’ (Hoare 1812, 64). This appears to be the first use of the name ‘Danes Ley’. The 1842 Tithe Map and Apportionment for Bratton shows several parcels of land in the area of the burial site ‘owned’ by the executors of Thomas Jarvis and occupied by Samuel Brent (who also had the mill) but the name Dean Lye (or similar) has disappeared (WSRO TA Westbury). The area immediately around the burial site is described as Lockham Plantation, the field above as Lockham Close (arable) and there is an ‘Orchard open to Lockham Close’ which is a strip of land forming the northern boundary of Lockham Close. By 1859 ‘Lockham Plantation’ has reverted to ‘Deanly Wood’, described as a grove or coppice and ‘Lockham Close’ to ‘Deanly’, described as a close of arable formerly meadow (WSRO 2699/23). The Ordnance Survey 25 inch 1* edition map of 1887 shows the area around the burial site as wooded and names it Danes Ley, the name used by Colt Hoare (Figure 8). Luccombe Mill was demolished at the end of the 19 century but the adjacent mill owner’s house, also built by Jarvis, remains and is called Luccombe Mill. This house, with the surrounding land including the burial site, eventually came into the ownership of Sir Horace and Lady Seymour during the 20" century. Trees were planted on the land in the valley around the burial site. In 1941/42 plans were drawn up by the then Air Ministry for a scheme to supply water to Keevil Airfield using water from Luccombe Springs (WSRO K4 140/4). The original springs were ‘capped’ with mini pumphouses, shafts dug, and the water piped to a pumping station close to the site. This station pumped water extracted from shafts close to the spring heads to a water treatment plant on the side of Picquet Hill. In the early 1950s R.H. Pearson of Baynton House, Coulston constructed watercress beds on and around the site and grew watercress for sale, as he had previously done at Baynton House (Pearson 1955). The outline _ and layout of the watercress beds can be seen on an aerial photograph dated September 1955, just one month before the discovery of the skeletons (RAF/82/1239 V0032). How long Pearson continued with this business after the discovery of the skeletons in October 1955 is not certain but possibly he abandoned his watercress venture in Bratton at some point between 1956 and 1958. Apart from further minor works by Wessex Water (and its predecessors), such as the construction of a metal bridge close to the site, the area around the site is much as it was in 1955. However, the natural vegetation has been allowed to grow up and the watercress beds to gradually disappear, although some of the wooden sleepers separating the beds are still visible. Local legend From at least the early years of the 19" century, a connection was made by some people between the ‘Danes’ in the name of Danes Ley for the area around the burial site and the Danes who fought King Alfred in AD 878 at the Battle of Ethandun. Although it is fairly certain that this battle took place somewhere on the downs above Bratton and Edington (eg. Peddie and Dillon 1981; Hodgkin 1952; Stenton 1947), details of the battle are limited and legends abound. As already mentioned above, Colt Hoare connects the area around the burial site with the battle and calls it Danes Ley (Hoare 1812, 64). Whether this was as a result of talking to local people or simply his own imagination is not clear. Colt Hoare continued to refer to Danes Ley in ‘Modern Wiltshire’ in his account of the Hundred of Westbury (Hoare 1829, 45). The first draft of Colt Hoare’s account of the Westbury Hundred was written by Richard Harris and Harris’s handwritten notes read: ‘Near Bratton in the parish of Eddington in a Valley called to this day Lucomb Springs and Dane Leys or Lays, Supposed to be the Very Spot of Ground where the Danes were encamped - (within sight of Bratton Castle) - when Alfred the Great came upon them in the Night — they suspecting no danger — and routed them from which they fled to Bratton Castle their Fortress and at which place they held out the Siege 14 days. Dane Lays is a pleasant valley cover’d in with the Hills — and at which place are the A ROMANO-BRITISH PRONE BURIAL FROM BRATTON, WILTSHIRE 199 Steeple Ashton 1 + + Vest Ashton Westbury * 7 _ a + + % Burial site * Other Romano-British burials Erlestoke . Springs , . A Romano-British building / occupation site + Romano-British pottery, coins or metalwork ---- Boundary of Westbury Hundred Fig 9. Sketch map by Stella Maddock, showing the location of Romano-British finds within a 4km square around the burial site. Scale 1:25000. Information from the County Sites and Monuments Record, with kind permission of Wiltshire County Council Archaeology Service. Traces of a Camp to this day — Viz'. Fireplaces Parade — and where the Great Tent and others were pitched with a Good Supply of Water and fire Wood etc.’ (WSRO MS 540/320). Other local stories about the area have been documented. One is about the sarsen stone that lies in the ditch ofa pillow mound in Luccombe Bottom. It has a reddish tinge to it and is known locally as the Blood Stone. According to the legend, Danes were beheaded on the stone when they fled after the battle (Morrison 2000, XI-XII). One local historian and long-term resident says that when the skeleton(s) was discovered in 1955, local people were not surprised as they always knew there were skeletons there and had assumed they were Danes. No documented reference to this knowledge has been found. All these stories added intrigue and interest to the research that was done before the radiocarbon dating analysis. The documented history of the site described above suggests, not surprisingly, that the origin of the name Danes Ley is unrelated to the Battle of Ethandun. And, perhaps disappointingly, the results of the C14 dating prove the skeleton(s) pre-dates the battle by at least 600 years. WIDER CONTEXT The site lies in a landscape that displays much evidence of earlier human activity. Within 2km of the site there are Neolithic long barrows at Tinhead (NMR ST9SSW13) and within Bratton Camp. Several Bronze Age bowl barrows (eg. NMR ST95SW4 and ST9SSW$87) lie within 1km of the site and the Iron Age hillfort of Bratton Camp is approximately 2km to the west of the site. 200 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE The nearest recorded Romano-British finds, including two inhumations (NMR ST95SW37 and ST95SW10) are about 1km away from the burial site to the north-west within the village of Bratton. One of the inhumations was an east—west burial with iron nails, suggesting a coffin, and sherds of Samian and coarse ware. Other Romano-British finds suggest the location ofa possible villa less than 1km from the site on the western edge of Edington (NMR ST95SW35) and is listed in a gazetteer of Roman villas (Scott 1993, 201). Additionally, an excavation in 1974/75 found evidence of Romano-British occupation within Bratton village (NMR ST95SW48) (Luckett 1987, 125) (Figure 9). Westbury, about 3 miles from the site, was important to the Romano-British as a source of iron. It is conjectured the Romans built a fort there and that Westbury was a ‘small town’ (Corney 2001, 35; Moorhead 2001, 88). A collection of Romano-British material salvaged from ore extraction diggings at The Ham, near Westbury railway station, suggests a substantial settlement, possibly with industrial activity based on the iron ore deposits (Corney 2001, 34). Several wells on the same site that contained human and animal skulls could have functioned as ‘ritual shafts’ associated with religious activity (Corney 2001, 35; Robinson 2001, 158). At Wellhead, near Westbury, an area quite similar in topography and geology to Luccombe, there was another prosperous Romano-British settlement with industrial activity (Rogers and Roddham 1991-511): Traces of medieval settlement earthworks to the west of Bratton church were identified as part of the Royal Commission on Historic Monuments of England (RCHME) Salisbury Plain Training Area project (NMR ST95SW 132). In Luccombe Bottom, within 500m of the burial site there are two mounds recorded on the NMR (ST95SW85 and ST95SW86). According to the record, both were originally interpreted as bowl barrows (Cunnington c.1926) and listed as ‘Edington 8’ and ‘Edington 9 (Grinsell 1957, 174; Goddard 1913, 250). However, later an OS Field Investigator interpreted them as pillow mounds. There is documentary evidence for rabbit farming having taken place in Luccombe Bottom at least as early as the late fourteenth century (Rogers 1965, 243). Also close by, on the slopes of Edington Hill, Patcombe Hill, White Cliff, and Longcombe Bottom, are the remains of field systems and lynchets of uncertain date, although some are thought to be late Iron Age/early Romano-British (eg. NMR ST95SW83, ST95SW82, ST95SW81, ST95SW80, ST95SW79, ST95SW27, ST95SW133). There are references to two possible mass | burial sites within a few kilometres of the burial site, neither of which have been investigated or interpreted. In 1814, workmen quarrying chalk on the south side of Bratton Camp allegedly told Dr. Seagram of Bratton House they had found bones near the surface as they dug (Morrison 1990, 31; Morrison 2000, X-XI). In the late 19" century, some dewpond diggers reportedly found another shallow mass burial on Chapel Down at Imber. A folk memory of this burial existed as the place was called the ‘Carrion Pit’, later corrupted to ‘Carpit’ (Morrison 1993, 7; Sawyer 2001, 59). DISCUSSION The circumstances surrounding the inhumation of this individual may never be known, or indeed if BRA 495 was the only the burial at this location. Further fieldwork will be needed to investigate the possibility of other skeletons on the site. What can be ascertained is that the skeleton belonged to an adult male, aged between 31 and 35 years of age, of medium height and build. This man was physically fit, despite a tooth abscess, perhaps caused by a diet full of starch and coarse foods. His upper body and lower spine showed signs of heavy work. This especially involved his left arm, where muscles had developed more than their right-sided counter-parts. From the photograph taken at the time of excavation (Figure 2), it can be seen that the bones were still articulated, indicating that the body was buried intact. As the bones were still in approximate anatomical situs, it can be inferred that the body was either placed in a grave pit, covered over securely with earth, or wrapped in material which has now disintegrated, ensuring the bones stayed in place after the soft tissue had decomposed. If the bones had been disarticulated and spread about, it could have indicated that the body was unwrapped or that it was placed or thrown onto the ground without being covered. Unceremonious deposition of the body in the ground may be indicative of haste, concealment or the result of an execution. So the circumstances surrounding the deposition of BRA 495 can only be speculated. As this is a lone burial and not part of a cemetery group, the other burials still uncorroborated, it is difficult to be certain about the nature of the inhumation. Lone prone burials possibly indicate haste and use existing features A ROMANO-BRITISH PRONE BURIAL FROM BRATTON, WILTSHIRE 201 Fig 10. Photograph of the skull interior of BRA 495. The black arrow points to the stained ring of settled exudate, indicating the prone but tilted angle of the head after death. Photograph by Pamela Mahon. such as ditches, rather than graves within a cemetery setting (Philpott 1991, 71). No grave goods, coffin, clothing or evidence of ceremony were found. Features of note were firstly the prone position of the remains, with the head slightly relaxed to one side (Figure 10), evident by the rim of levelled exudate within the skull vault, which indicates the angle of settlement. Secondly, the position of the arms behind the torso suggest the wrists could have been bound. A high proportion of the hand and wrist bones are missing, as these were in the upper part of the grave which was disturbed first and closest to the ground surface. Prone burials of Romano-British and Anglo- Saxon date are not uncommon within cemeteries but are suggestive of criminal executions (Philpott 1991). The orientation of the burial and the position of the arms re-enforce the theory of execution although no evidence of peri-mortem trauma could be seen on the skeletal elements excavated and all of the cervical vertebrae were intact. The body was articulated, showing no sign of post-mortem mutilation. The actual cause of death is therefore not apparent from the skeleton but execution cannot be ruled out. Tacitus refers to a form of punishment in which cowards and shirkers are drowned in miry swamps under a cover of wattle hurdles (McKinley 2003, 15), which is another possible scenario for BRA 495. On the evidence of the C14 dating we can now discount any connection between these burials and the Battle of Ethandun. It is sull not possible to make assumptions about the circumstances surrounding the death of the individual whose skeleton has been examined or of the others reportedly still buried at the site. The ‘further two’ skeletons noted by Annable as having been left in the ground and possibly four more noted on the NMR, have not as yet been located. No documentary evidence of their precise location, whether they too were prone or showed evidence of violent death has been found. It may not be safe to assume they are of the same or similar date to BRA 495 so we do not know whether we are dealing with an event which resulted in a mass burial or single burials over time. The location of the burial on a boundary — hundred and parish — may be significant. Hundred boundaries may very tentatively have their origins in Romano-British territorial divisions (Clarke 2004, 93). If Westbury was the centre of a Romano-British territorial division, it is possible that the extent of the later hundred followed its boundary, meaning the burial site was on a significant boundary in the Romano-British period. This might be a relevant factor in explaining the burials, suggesting the victims were criminals or outcasts, deliberately buried at the outer limits of a local territory. The other potentially relevant feature of the burial site’s location is its watery nature and proximity to powerful springs. The site may have had ritual significance as a liminal place where human sacrifices might have been offered. Studies of bog burials in Europe, show that the Iron Age practice of human sacrifice continued well into the Roman era, the victims of this practice often exhibiting markers of physical impairment (Taylor 2003, 14). There is insufficient evidence as yet to determine whether the location of the burial(s) is one with ritual or territorial significance, or neither. Finding and examining any remaining skeletons in the ground would be an important step towards understanding the nature of the site, the people buried there, and more about the early Romano-British period in this part of Wiltshire. Acknowledgements Grateful thanks are due to everyone who has supported, helped and advised with this project in a variety of ways, notably Hugh and Elizabeth Seymour, Wendy Smith, Robin Holley, Heather Ault, members of the WANHS Archaeology Field Group, Dr Paul Robinson, Alison Maddock, Jean Morrison and the Wiltshire County Council Archaeology Service. Thanks are also due to Dr Alan Hogg (Director) and Dr Fiona Petchey (Deputy Director of the University of Waikato Radiocarbon Dating 202 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Laboratory in Hamilton, New Zealand for their fast turn-around of the samples and their expert advice on interpretation of the results. Bibliography ANNABLE, FE K., 1955, Burials at Bratton and Longbridge Deverill, WAHNM 56, 190-191 BASS, W. M., 1995, Human Osteology: A laboratory and Field Manual (4th Edition). Missouri Archaeological Society, USA BROTHWELL, D. R., 1981, Digging Up Bones (3" edition). London : British Museum Natural History Series CLARKE, B., 2004, An early Anglo-Saxon cross-roads burial from Broad Town, North Wiltshire. WANHM 97, 89-94 CORNEY, M., 2001, ‘The Romano-British nucleated settlements of Wiltshire’, in P. Ellis (ed), Roman Wiltshire and after. Devizes: WANHS, 5-38 CRAWFORD, O.G.S., c.1930, Annotated Ordnance Survey 6” to 1 mile map sheet XLV NW. Held at WANHS Library, Devizes CUNNINGTON, M.E., c.1926, Annotated Ordnance Survey 6” to 1 mile map sheet XLV NW. Held at WANHS Library, Devizes DARLINGTON, R.R., 1955, ‘Domesday Survey’, in R.B.Pugh and E.Crittall (eds), A History of Wiltshire. The Victoria History of the Counties of England, 2. London: Oxford University Press, 42-168 DOBNEY, K., and BROTHWELL, D. R., 1987, A method for evaluating the amount of dental calculus on teeth from archaeological sites. fournal of Archaeological Science, 14, 343-351 GEORGE, H., 1997, The footpaths of Edington. Privately printed GODDARD, E.H., 1913, A list of prehistoric, Roman, and Pagan Saxon antiquities in the county of Wiltshire. WANHM 38, 153-378 GOVER, J.E.B. et al, 1939, The Place-names of Wiltshire. English Place Name Society. Cambridge: Cambridge University Press GRINSELL, L.V., 1957, ‘Archaeological Gazetteer’, in R.B.Pugh and E.Crittall (eds), A History of Wiltshire. The Victoria History of the Counties of England, I, 1. London: Oxford University Press, 21-279 GRUNDY, G.B., 1920, The Saxon land charters of Wiltshire (second series). Archaeology fournal, 77, 8-126 HOARE, R. Colt, 1812, Ancient Wiltshire, vol.1. London: William Miller HOARE, R. Colt, 1829, History of Modern Wiltshire, vol.3. London: J.Nichols HODGKIN, R.H., 1952, A history of the Anglo-Saxons. Oxford: O.U.P. LARSEN, C. S., 1997, Bioarchaeology : Interpreting Behaviour From the Human Skeleton. Cambridge University Press, Cambridge, England LATIMER, H. B., and LOWRANCE, E. W., 1968, Bilateral Asymmetry In Weight and Length of Human Bone, American Journal of Human Evolution, 152, 217-224 LOVEJOY, C. O. et al, 1985, Multifactorial determination ° of skeletal age at death: a method and blind tests of its accuracy. American Journal of Physical Anthropology, 68, 1-14 LUCKETT, L., 1987, Manor Farm, Bratton. ST 91255132. WANHM, 81, 125-130 McKINLEY, J. I. 2003, A Wiltshire ‘bog body’?: Discussion of a fifth/sixth century AD burial in the Woodford valley. WANHM, 96, 7-18 MOORHEAD, T.S.N., 2001, ‘Roman coin finds from Wiltshire’, in PEllis (ed), Roman Wiltshire and after. Devizes: Wiltshire Archaeological & Natural History Society, 85-105 MORRISON, J., 1990, Turnpike Roads. Bratton History Association Journal, 1, 31-33 MORRISON, J., 1992, Bratton Woollen Mills Part 1. Bratton History Association Fournal, 3, 12-15 MORRISON, J., 1993, Imber. Bratton History Association Journal, 4, 2-9 MORRISON, J., 2000, The Battle of Ethandun AD 878. Privately printed PEARSON, R.H., 1955, Baynton House. London: Putnam PEDDIE, J. and DILLON, P, 1981, Alfred’s defeat of the Vikings. Ethandun Memorial Trust PHILPOTT, R., 1991, Burial Practices in Roman Britain : a survey of treatment and furnishing AD 43-410. Oxford : British Archaeological Reports, British series 219 REEVES, M., 1978, Sheepbell and ploughshare. Bradford on Avon: Moonraker Press REEVES, M. and MORRISON, J. (eds), 1989, The Diaries of Jeffery Whitaker 1739-1741. Trowbridge: Wiltshire Record Society Vol. 44 ROBINSON, P, 2001, ‘Religion in Roman Wiltshire’, in PEllis (ed), Roman Wiltshire and after. Devizes: Wiltshire Archaeological & Natural History Society, 147-164 ROGERS, B. and RODDHAM, D., 1991, The excavations at Wellhead, Westbury 1959-1966. WANHM, 84, 51-60 ROGERS, J., SHEPSTONE, L. and DIEPPE, P, 1997, Bone formers: osteophyte and enthesophyte formations are positively associated. Annals of Rheumatic Diseases, 56, 85-90 ROGERS, K., 1965, ‘Whorwellsdown Hundred’, in E.Crittall (ed), A History of Wiltshire. The Victoria History of the Counties of England, 8, London: Oxford University Press, 193-263 ROGERS, K., 1976, Wiltshire and Somerset Woollen Mills. Edington, Wilts: Pasold Research Fund Ltd. SAWYER, PH., 1968, Anglo-Saxon Charters: an annotated list and bibliography. London: Royal Historical Society SAWYER, R., 2001, Little Imber on the Down. Salisbury: Hobnob Press SCOTT, E., 1993, A Gazetteer of Roman Villas in Britain. Leicester Archaeology Monographs No.1. Leicester: A ROMANO-BRITISH PRONE BURIAL FROM BRATTON, WILTSHIRE 203 University of Leicester School of Archaeological Studies. STEELE, T. J. M., and MAYS, S. A., 1995, Handedness and Directional Asymmetry in the Long Bones of the Human Upper Limb, Jnternational Journal of Osteoarchaeology, 5, 39-49 STENTON, FE, 1947, Anglo-Saxon England. Oxford: OUP STEVENSON, J.H. (ed), 1987, The Edington Cartulary. Devizes: Wiltshire Record Society Volume 42 TAYLOR, A., 2003, Burial with the Romans. British Archaeology, No.69 March, 14-19 TROTTER, M., and GLESER, G. C., 1952, Estimation of stature from long bones of American whites and Negroes. American Journal of Physical Anthropology, 10(4), 463-514 TROTTER, M., and GLESER, G. C., 1958, A Re- Evaluation of Estimation Based on Measurements of Stature Taken During Life and of Long Bones After Death. American Journal of Physical Anthropology, 16, 79-123 van Klinken, G. J., 1999, Bone collagen quality indicators for palaeodietary and radiocarbon measurement. Fournal of Archaeological Science, 26, 687-695 WN and Q 1899 Records of Wiltshire parishes — Bratton. Wiltshire Notes & Queries, 3, 154-61 WSRO 2699/23 (part of the Jean Morrison collection) 1859 ‘Lifehold Property in Bratton’, a survey of tenancies of the Marquess of Bath. WSRO EA/76 1808 Enclosure Award for Westbury WSRO K4/140/4 1938-49 Documents related to the Edington and Steeple Ashton Water Supply Scheme WSRO MS 540/320 c.1820 Draft by Richard Harris of Dilton Marsh of the article on the history of Westbury which subsequently appeared in Hoare’s ‘Modern Wiltshire’ but containing some material not printed. WSRO MS 929/53 c.1945 Historical notes on the village of Bratton by students of St. Boniface College, Warminster. Unpublished typescript WSRO MS 1195/22 c.1699 Terrier of Mr. Ballard’s freehold at Bratton WSRO TA 1842 Tithe Award for Westbury Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 204-212 Eve of Wilton and Goscelin of St Bertin at Old Sarum c.1070-1078 by Daphne Stroud In about 1080 Goscelin of St. Bertin, a Flemish monk who had served in the household of Bishop Hereman of Salisbury, wrote a treatise of spiritual advice to his former pupil Eve who had left Wilton nunnery to become an anchoress at Angers in France. He prefaced his work with an account of their relationship during the years of Bishop Hereman’s episcopate at Old Sarum (c. 1070 -1078). This paper examines Goscelin’s account in detail and considers some connections with other evidence about this little-documented period of the history of Salisbury diocese. The Liber Confortatorius The Liber Confortatorus (The Book of Encouragement) is a treatise on the spiritual life in the form of a letter written by the monk Goscelin of St Bertin c.1080-1082 to Eve, anun of Wilton Abbey in Wiltshire who had recently become an anchoress at Angers in France. It survives in a single copy, made in the late 12th century, apparently at the Abbey of St. Sauveur-le-Vicomte in Coutances, Brittany and is now in the British Library (MS Sloane 3103). It remained virtually unknown to scholars until the 1930s when André Wilmart (1934, 1938) published an account of its contents. An edition of the text was published by C.H. Talbot in 1955, and a translation by Monika Otter in 2004. Its subject matter as a book of spiritual guidance has been studied in recent years by a number of scholars. Where they have commented on Goscelin’s factual account of his association with Eve almost all published works have accepted the chronology proposed by Wilmart which relied on his identification of two church dedications attended by Eve and Goscelin as those of Wilton Abbey (3 Oct. 1065) and Westminster Abbey (28 Dec. 1065). This dating would mean that Goscelin’s emotional attachment to Eve started when she was a child of five or six and gives an unjustifiably sinister slant to their story. The contention is that Wilmart’s interpretation was mistaken, that the church dedications were not those of Wilton and Westminster and that the association of Eve and Goscelin started later, when Eve was well in her teens (Appendix 1). This paper offers a history of the association using this revised dating, based as closely as possible on Goscelin’s own account, and related wherever possible to other evidence about Old Sarum and Wilton in the 1070s. At the outset of his Book, Goscelin sets out the circumstances in which it was written - his nurturing of Eve’s vocation in the nunnery of Wilton under the patronage of Bishop Hereman of Old Sarum, and his own deep and barely controlled love for his pupil; the lack of understanding of Hereman’s successor Osmund and of some members of the Wilton community; his own departure from Old Sarum in face of the ensuing scandal and the devastating blow when he heard that Eve had left Wilton to become an anchoress in France and that the separation which he had expected to be temporary was to be permanent. Goscelin claims that when the Norman Bishop Osmund, ‘with step-fatherly barbarity,’ (Lib. Conf. 29, 24) dismissed him from Old Sarum ‘he preferred 35 Upper Street, West Harnham, Salisbury, Wiltshire SP2 8LS EVE OF WILTON AND GOSCELIN OF ST BERTIN AT OLD SARUM .1070-1078 205 to be humiliated by alien arrogance rather than do too little for charity’s sake’ (Lib. Conf. 26, 16-7). That is, for the sake of peace in the community of Wilton, he did not attempt to attack the scandal- mongers in return. Later, however, after his agonised acceptance of the permanent loss of his beloved, he believed that he had earned the right [Liceat me nunc] to give in the Liber Confortatorius an account of the association which would serve to provide ‘an everlasting memorial’, a memorial which would demonstrate to posterity that Eve had pursued her vocation in ever-increasing holiness, and that any sinfulness in the association had lain solely in Goscelin’s own barely contained passion. Goscelin’s narrative covers a period of at most seven or eight years. Information about the lives of the two participants before and after this time of intense experience has to be sought either from passing references elsewhere in the Liber Confortatorius or from independent sources. Eve: parentage and early life For Eve’s early life at Wilton the only reliable source is the Liber Confortatorius itself. For her later years when she lived as a deeply respected anchoress at Angers the information comes mainly from the verses written after her death by the scholar ‘Hilary’ (Fuller 1929). These purport to give details of her early life as well as of her years at Angers. Hilary, however, was writing 50 years or more after Eve’s childhood and must have been using second or third hand information. He cannot be regarded as a reliable source for her life at Wilton and where his evidence conflicts with Goscelin’s it must be rejected. Eve’s date of birth is known only approximately from the information given in the Liber Confortatorius, but must fall between 1058 and 1062 (Appendix 1). Of her parentage Goscelin gives more exact information. Her father was a Dane and her mother came from Lorraine. The family, of honourable standing, had come to England during Eve’s childhood in the 1060s. Eve herself was born abroad and came to England as an immigrant (advena); here she ‘emerged’ as English (Lib. Conf. 41, 19-22). The Latin word used - ‘emersisse’ - has overtones of a butterfly emerging from a chryalis, and implies an improvement in status. By the time Goscelin knew them, the family were well-to-do land-owners living apparently in the vicinity of Wilton. He does not indicate how this position had been achieved, but it is clear that Eve’s Danish father had acquired a land-holding in England in the decade of the Norman Conquest and was eager that his daughter should be accepted as English. Many of the motley band of adventurers who fought with William the Conqueror at Hastings were in just that position and there must be a strong possibility that Eve’s father was one of them. When she left for Angers Eve left behind her ‘loving parents, dear kindred’ as well as her companions in the nunnery (Lib. Conf. 37, 10-3). Goscelin opens his account of his association with Eve at the point when he became her teacher, when she would have been about 12, and does not indicate how she came to be at Wilton. However, two comments he makes in the course of his narrative imply that she was not a child oblate (a child given for life to the nunnery by her parents, often at a very early age). He says that, during the period before her profession, she ‘gave books and performed every exercise of charity’ (Lib. Conf. 28, 4-6). If Eve had been a child oblate she would have had no possessions and could have made no gifts. He also says that, in contrast to her later willing acceptance of the status of one of Christ’s poor, she had once been affronted if someone called her a nun, suggesting that she was at one stage living in the nunnery without having consecrated status (Lib. Conf. 104, 1-3). The word Goscelin uses (indignabaris) suggests that Eve felt strongly about her social position. She was presumably one of the well-born girls who were sent to Wilton for their education. Goscelin and Bishop Hereman Goscelin himself was a scholar and historian of some distinction. William of Malmesbury says of him that he was ‘distinguished by his literary and musical craftsmanship .. . none in his homage to English saints’ (De Gestts Regum ii, 389). He had been a monk of St Bertin, at St Omer in Flanders and came to England to join the household of Bishop Hereman of Sherborne in after Bede second to Dorset, 35 miles from Wilton, probably between 1058 and 1060. In the early 1070s Hereman moved from Sherborne to Old Sarum, in accordance with the decrees of the Church Council of London, 1075, which required sees to be moved from villages to larger towns. It is probable that, in the case of Old 206 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Sarum, the Council gave retroactive authority for an existing situation and that the move was in fact made in 1070-72. Goscelin, by then a long-serving member of Hereman’s household, came with him. The new see was extremely poor. Hereman had only been able to establish himself at Sherborne in 1058 by uniting its resources with those of the old bishopric of Ramsbury and using the Benedictine abbey church as his cathedral. Now he had to start again, building a new church and community buildings on a bleak, waterless hill-top. Progress was inevitably slow; Hereman does not appear to have recruited personnel for a new monastic establishment, and the stone church, planned on a much smaller scale than other Norman cathedrals, had not got far beyond the foundations when he died in 1078. The establishment of a well-endowed chapter of canons and the completion of the church had to await the arrival of his Norman successor, Osmund. Meanwhile his household had to be lodged somewhere. The position must have been particularly galling as Hereman and Goscelin both relished the relative comfort and cultivated atmosphere of the great Flemish monastery of St Bertin where they had lived together in the 1050s. Hereman, according to William of Malmesbury’s prejudiced and hostile account, ‘was used to servants and feeding on luxuries (De Gestis Pontificum, 83) and missed in England the comforts he had known during his life on the continent. Goscelin himself was repelled by the initial squalor of his lodgings when he came to England. At Wilton, three miles from Old Sarum, there was a wealthy nunnery, a royal palace, and a small township. Hereman may possibly have lodged there during the early years of his Salisbury episcopate, only moving up to Old Sarum when there were suitable buildings to occupy. This would explain the remarkable degree of intimacy with the life of Wilton nunnery which the Liber Confortatorius reveals in respect of Hereman as well as Goscelin. Hereman came to the nunnery apparently informally, and not infrequently - whenever our Bishop visited, says Goscelin, writing of Hereman’s practice of calling Eve by her professed name at the nunnery; it was at Hereman’s urging that Goscelin undertook to write his Life of St Edith of Wilton - perhaps by way of repayment for hospitality received, since he regarded his delay in carrying out this work as negligent and blameworthy. In the ‘Life’ Goscelin recounts the story of a nun from Wilton with whom he personally gossiped when she was staying at the Bishop’s residence at Old Sarum on some matter of nunnery business (Vita S. Edithe, 298). He seems | to take her presence there as a matter of course, requiring no particular explanation. Eve and Goscelin Phase I Goscelin himself seems to have taken part in the general life of the nunnery, joining in the services of the abbey church and feasting with the sisters on major festivals. He also taught some of the younger nuns. There was nothing in Eve’s birth or parentage that would have marked her out for special attention in that aristocratic society, and at first she was presumably merely one among a number of pupils. Goscelin, however, seems to have quickly recognised her spiritual qualities and under his guidance Eve herself changed her former worldly attitude. ‘You remember,’ he wrote, ‘how I first aroused you from your state of childish illiteracy, confident that I could easily correct so devout a soul.’(Lib. Conf. 28, 2). [The word ‘infantiam’ which Goscelin uses here has its root meaning of ‘without words’; that is, he had found Eve ignorant of Latin, the only language in which she could start to study the spiritual life]. Eve did not immediately become a nun. At this stage Goscelin praises her for the virtues appropriate to a laywoman. She ‘gave the books we wanted’ - no doubt she directed her parents charities - and ‘performed every exercise of charity’. Having left the facilities of Sherborne behind Hereman must have been in dire need of books for study as well as for the performance of the liturgy. His successor, Osmund, was able to establish his own scriptorium and build up a renowned collection of books, many of which survive in Salisbury Cathedral Library today. Hereman had to be content with more modest acquisitions, bought second-hand from other religious houses. Only a handful of manuscripts which might have come to Old Sarum during his episcopate have survived. Two of these, MS 96 and MS 173, are of particular interest. Both manuscripts were produced on the continent in the 10th century and both are thought to have been in use in other English religious houses before coming to Salisbury. MS 96, written in a large, clear hand, would have been suitable for reading aloud in a monastic refectory; but the text is very corrupt. MS 173, a smaller volume suitable for individual reading, presents a somewhat untidy appearance, and also contains a number of errors EVE OF WILTON AND GOSCELIN OF ST BERTIN AT OLD SARUM .1070-1078 207 and omissions. Wealthy monasteries in post- Conquest England might well have wished to replace such volumes with more handsome and accurate texts. The older manuscripts would still have been welcome in poorer communities such as Old Sarum. MS 96, Dialogues of Gregory the Great, recounts the lives and deaths of the saints and martyrs of the early Church, with edifying instances of divine intervention, particularly in monastic life; one entire section deals with St Benedict, the founding father of western monasticism. MS 173 contains St Augustine’s Soliloquies, a meditation on the knowledge of God, and St Isidore’s Synonyms, a mixture of linguistic and doctrinal items with a pastoral slant, calling for the repentance of sinners. Both manuscripts provide the sort of basic texts which would have been required by the new community to be founded at Old Sarum; they also reflect some of the themes - the martyrs and the contemplative life - elaborated by Goscelin in the Liber Confortatorius. Of his own feelings at this time Goscelin says simply, ‘Up ull then my love for you had been bearable, to all appearances that of a good Christian’ (Lib. Conf. 28, 6-7). However, with Eve’s profession at the age of about 15, this all changed... “Then came the day when you came fearfully to your nuptials with the Lord, the last but one of fourteen virgins... you put on the consecrated garments and made your pledge of constancy to God.’ Watching Eve’s trembling and humble offering of herself Goscelin was overwhelmed, pierced to the very depths of his being by his sudden realisation of his own passion. ‘I was touched by the dews of Heaven and wept with a passion of tears.’ (Lib. Conf. 28, 7-14). Of Eve’s emotions as she made her lifelong commitment under Goscelin’s influence nothing is known, but from this point on he praises her not for the generosity appropriate to a laywoman but for her devotion to her duties as a nun: her silence, her self-control, her devout psalmody (Lib. Conf. 28, 14-16). Eve and Goscelin Phase II After the Norman Conquest, Pope Alexander II issued canons (the Penitentiary after Hastings) imposing a tariff of penances on all those who had fought at Hastings on the side of the invaders and had wounded or killed a fellow Christian (Morton 1975, 381). The penances could, at the discretion of the bishop, be commuted to the building or endowing of a church. It is not known how widely these canons were enforced, but in Salisbury diocese at least efforts at compliance were apparently made over a considerable period. The canons, issued c. 1070 and probably reaching Old Sarum soon after, were copied after Hereman’s death into the Salisbury Pontifical (British Library MS. Cotton Tiberius Ci, ff. 111v-112v,), a service book containing rites peculiar to a bishop, which was used by both Hereman and his successor Osmund. The manuscript also contains an otherwise unknown sermon for the dedication of a church written partly in Old English and consequently suitable for delivery to a rustic audience. Goscelin believed that the life of contemplation which he envisaged for Eve required an even deeper personal commitment than that of her formal profession and he was anxious that this inward dedication should be made under the aura of sanctity furnished by the great sacraments involved in the dedication of a church. His lengthy description of Eve’s attendance at two such dedications lies at the heart of his apologia for their two lives. Some time after her profession he arranged for her to accompany him to the dedication of a country church by Bishop Hereman. This was an exceptional privilege for a professed nun and seems to call for more explanation than the account in the Liber Confortatorius offers. Was Eve’s father perhaps the local landowner responsible for the building of the church in question (Goscelin does not name it) as a penance under the terms of the papal penitential for his part in the battle of Hastings? The service of consecration was immensely long, comprising as it did at least three three-fold circumambulations of the outside of the church by clergy and congregation, as well as the long and complex rites within. In the course of these lengthy proceedings Goscelin would have had ample opportunity to speak to Eve while continuing to take his full part in the proceedings as a member of the bishop’s retinue. He spelt out the commitment which would be required of her, and the self- offering which was to be her life’s goal; in the Liber Confortatorius he repeats the urgent admonitions he then made: ‘.. . ask one thing only of the Lord, namely that, wounded by holy love, you should desire Christ alone and that with your whole heart and mind you should seek Him alone as your dower .. .” (Lib. Conf. 28, 18-23). At the climax of the consecration, when the bishop returned from anointing the outside walls with chrism and censed the congregation as the verse ‘Let my prayer be 208 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE set before Thee as incense’ was chanted, Goscelin urged his beloved to use the sanctified moment to make her life’s commitment in the spirit of holy love: ‘This shows how you should enter the house of the Lord. Not with coldness, not with an empty heart, but with the burning incense of love, with the sweetly-smelling glow of virtue . . .’ (Lib. Conf. 28, 28-30). After the ceremony there was a banquet. Eve’s parents were present and ‘through the fondness of her father’, she was permitted to sit with her mother among the laypeople. The fact that her father was in a position to ask and obtain such a concession for his daughter implies that he was a man of some standing and local influence and strengthens the case for regarding him as the donor of the new church. Goscelin, however, was worried by Eve’s presence at the banquet; it might break the aura of holiness under which she was to live throughout the three-day period of the ceremony [Day 1 Vigil: Day 2 Consecration and banquet: Day 3 Open air address by Bishop to people]. He sent her a dried fish, the standard fare of fast days, with a succinct warning to mortify the flesh by abstinence from rich foods: ‘Smoked fish, Christ suffered.’ (Lib. Conf. 28, 33-35). A second ceremony of dedication followed. Goscelin’s use of the impersonal form ‘transitus est’ [We/they went on to] implies that the second ceremony was attended by the same congregation as the first and took place shortly after. It may well have been in the course of a single journey by the bishop. For a second time Goscelin ‘won the reward’ of seeing his beloved as a three-day ‘Mary’ - a reference to the Mary, sister of Martha, who sat at the feet of Christ and was the exemplar of the contemplative. The congregation dispersed, but Eve had indeed received the gift Goscelin had implored her to pray for - a heart pierced by the wounds of God’s love. It was a first fruit, a foretaste of her long, dedicated life of prayer. Eve and Goscelin Phase III After the dedications Goscelin continued as Eve’s spiritual director. He was teaching her on a one to one basis now, and there was an exchange of writings. The fact that there was occasion for such exchanges, in an age when almost all teaching was oral, implies that by then Goscelin was not living at Wilton and his visits, although frequent, were not daily. Eve must have made good progress; in the Liber Confortatorius he sets out a formidable list of the works of the Latin Fathers - especially St_ Augustine and St Jerome - that she should study, and assumes that she is capable of reading them on her own. ‘Do not give up if you do not understand a passage, but work at it, go over it again, re-read it until you understand it fully, for nothing is so difficult that it cannot be made out by trying and everything can be overcome by an extra effort.’ (Lib. Conf. 81, 5-8). Goscelin makes no attempt to conceal the intensity of the emotional bond between himself and Eve - indeed he can almost be said to flaunt it in his determination to demonstrate the difference between Eve’s sublimated, sacred love and his own profane and sinful passion. ‘Your breast, pierced with the arrows of the Lord, knows what groans, what revolts, what weary longings, what empty days and nights, what unattainable love of a mind inflamed by passion, I bewailed.’ (Lib. Conf. 29, 6-10). However, in the aristocratic community of Wilton, devoted to the seemly performance of the daily Hours and to polite conversation rather than to deep erudition and an intense spirituality, there were inevitably those who cast doubts on the association. Goscelin vigorously rebuts their approach. ‘Those who forbid study because they fear it may puff the students up are mistaken, since people may be the humbler the more learned they are... On the contrary those who do so should themselves learn through studying how they may guard humility, the guardian of all the virtues, and trample underfoot the barbarism of pride and the bragging of the undisciplined.’ (Lib. Conf. 82, 1-5). Bishop Hereman, however, fully approved of Eve’s way of life; Goscelin tells a slightly puzzling story to demonstrate this. ‘Whenever our Bishop, your father in God, embraced you and addressed you as “Eve, mother of the living,” I made a play on the name, saying, “That Eve was the mother of the living; this Eve will be the daughter of the living.”’ (Lib. Conf. 29, 18-20). The point of recalling these incidents appears to be to stress that the bishop himself made no objection to Goscelin’s rather daring word-play ‘daughter of the living’, which implied that Eve was one of the elect, the blessed of God. Perhaps some of the sisters took exception to such presumption, and brought it up as a weapon against the couple later. By stressing that it was just ‘word play,’ a literary conceit, Goscelin may have hoped to lessen his offence. Hereman died on 20 February 1078 and his successor Osmund had reached Old Sarum by EVE OF WILTON AND GOSCELIN OF ST BERTIN AT OLD SARUM .1070-1078 209 early summer. Goscelin records the ending of his association with Eve in one bitter sentence: ‘After the death of our father I came more often to find consolation with you in our common bereavement, until there arose a king who knew not Joseph and he who remains faithful to you was driven by viperish envy and step-fatherly barbarity to journey far away.’ (Lib. Conf. 29, 22-25). Osmund was Salisbury’s most famous bishop, respected in his life-time and revered as a saint by the local people after his death. The only fault that William of Malmesbury ascribes to him in an otherwise laudatory account of his life is that, being outstanding for his own chastity, ‘he punished more sternly in others sins which he did not find in himself.’ (De Gestis Pontificum, 83). It is clear from the prologue to the Book of Encouragement that when he arrived at Old Sarum in 1078 he found gossip and scandal about Eve and Goscelin rife. He was the last man to tolerate any suggestion of ‘incontinence’ in a member of his own household and Goscelin was dismissed. The ‘viperish envy’ of which he speaks no doubt refers to those among the sisters of Wilton who had already shown their lack of sympathy with Eve’s spirituality; and some of the younger nuns - perhaps among the thirteen who were professed with Eve - may well have resented the favouritism Goscelin had shown to his most outstanding pupil. But the main culprit in Goscelin’s eyes was Osmund, whose duty it was to act as a father in God, but who showed a total lack of understanding, acting as a Pharaoh, the type of secular oppressors, or as a barbarian whose lack of learning and inward thought made him incapable of comprehending the life of the spirit. Epilogue With Goscelin’s departure from Old Sarum in 1078 his direct narrative comes to an end; later developments are mentioned only when they are needed for the didactic purposes of his Book. He spent the next two or three years travelling between various monasteries in England. He still thought of himself as Eve’s spiritual director and assumed that she would continue to live the life they had both contemplated for her, ‘a useful vessel in the house of the Lord, the dove of the dovecote, not the solitary turtle-dove’. Or, if she preferred the life of a solitary, it would at least be in her own country where he could still watch over her (Lib. Conf. 36, 30-34). He remained in touch with Wilton and was preparing to visit the nunnery when a letter reached him at Peterborough. Eve had left Wilton to embrace the life of an anchoress, an ‘enclosed one’, at Angers in France. A ‘blessed lady’ had apparently prepared an anchorhold there and offered Eve support for her life as an anchoress. It was not unusual for wealthy laypeople to support an anchorhold as a form of religious offering, but Eve’s acceptance of the offer and precipitate departure from Wilton were highly unorthodox. In later centuries, when the procedure for enclosure had become more formal, a long period of preparation, often many years, would have been required to ensure that the postulant had a true vocation. In the less precedent-bound 1080s Eve’s hasty departure seems to have been accepted without opposition by the authorities concerned. How the ‘blessed lady’ came to make her offer and how Eve made her long journey to Angers is not known. Goscelin was devastated by the news of Eve’s flight. The opening pages of the Liber Confortatorius are incoherent with his grief. Then he steadied himself and turned to the task of writing his long Book of Encouragement for Eve and consolation for himself. He spent the rest of his life in a number of monasteries in England, where he wrote the lives of many local saints. He eventually settled in Canterbury where he died in or after 1107. Eve achieved her goal of a life dedicated to the contemplation of God. She reached Angers and spent the rest of her chosen life as an anchoress there. In her later years she shared a double anchorhold at the church of St. Eutrope with a male anchorite, Hervey. After her death a certain Hilary, a scholar who lived at Angers in the 1120s, wrote a poem about her in which he records for posterity her local fame and the holiness of her life. Appendix 1 Chronology (a) Eve’s Date of Birth Two passages in the Liber Confortatortus have a bearing on the question of Eve’s date of birth: (a) ‘I first came to the bishop... when I was a young man (adolescentulus) and you a child Gnfantulay (Lib, Conf. 102.11); and (b) ‘You are a young girl . . . (in) your twenty-year-old youth’ (adolescentula es... hu vicenna etatula)(Lib. Conf. 70.12). At (a) Goscelin 210 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE is referring to his arrival in England to join the household of Bishop Hereman; at (b) he describes the devil tempting Eve after her enclosure, at the time the Liber Confortatorius was written. To deduce a firm date from these passages the following would need to be known: 1. When Goscelin came to join Hereman. Hereman came to England in 1058, but Goscelin may not have joined him until later. 2. When the Liber Confortatorius was written. It is usually dated 1080-1082 on the basis of a reference to the marriage of King Canute of Denmark (succeeded 1080) to the daughter of the Count of Flanders (Lib. Conf. 41.7). 3. What latitude Goscelin allowed himself in referring to Eve’s twenty-year-old youth. Juggling with these variables can yield a date anywhere between 1058 and 1062. (b) The Church Dedications as a basis for dating the Goscelin/Eve association The generally accepted dating for Goscelin and Eve’s association 1s based on the identification of the two church dedications in Goscelin’s account as the consecrations of the Abbey churches of Wilton and Westminster on 3 October and 28 December 1065 respectively. The identifications were first made by André Wilmart, who says of them, ‘La prémiere fut celle de Wilton ... Par suite, la seconde ne peut étre que celle de Westminster.’ (Eve et Goscelin II, p. 58 n.5; p. 59 n.1). He offers no evidence in support of these bald statements and he was mistaken both for reasons arising from the text itself, and because of the distortion of the story that would result if Eve is assumed to have been a child of only five or six at its inception: 1. Goscelin does not name either church or give any evidence pointing to Wilton or Westminster. Some explanation of the choice of these two venues is clearly required, and none is given. 2. Goscelin says that he ‘arranged’ [fecz ut] for Eve to attend the first dedication. As a professed sister of Wilton nunnery she would have attended the dedication of the community’s own abbey church without any ‘arrangement’ by an outsider. Goscelin’s use of this phrase appears to rule out any possibility of the church in question being Wilton Abbey. 3. William of Malmesbury introduces Goscelin simply as a monk of St Bertin, who came to England with Hereman, Bishop of Salisbury’ (De Gestis Regum, 389). Goscelin himself says that he was for many years a member of the familia of Bishop Hereman - ‘cui laterales adhaerebamus’ (Historia translations S.Augustini 14). There is no evidence for his having held any other appointment during Hereman’s lifetime. So long as Hereman’s see was at Sherborne, 35 miles from Wilton, Goscelin could not have been simultaneously a member of his familia and a teacher at Wilton. This would only have become possible when the see was moved to Old Sarum, 3 miles from Wilton in 1070-1072. 4. It is clear from Goscelin’s description of Eve’s profession - which he says took place before the church dedications - that it was the clothing ceremony of a group of young women who were capable of making their own vows, not the offering of a child oblate by her parents. Eve would have been about 15 at the time. On the Wilton/Westminster theory she would have been 6 or 7. 5. Goscelin’s words of admonition (in Latin) at the first dedication ceremony (‘.. .ask one thing only of the Lord, namely that, wounded by holy love, you should desire Christ alone . . . Strive with the Lord until you triumph...’) which he confidently expected Eve to remember when she read the Liber Confortatorius at the age of twenty, are totally inappropriate for, and probably incomprehensible to, a child of six. Finally, the general tenor of the whole work is a factor which should not be ignored. To read the account as a reprehensible obsession with a child who was at its outset perhaps five years old distorts the whole gist of the Liber Confortatorius. The logical outcome of this approach is seen in Latzke - ‘under the eyes of the abbess, the Norman (sic) cleric was able to exert the overwhelming influence over the little girl . . . (which he) wished to develop into an exemplar of ‘agonista Dei’ whereby he drew her body and soul into his power’(Latzke 1984, 136). On the contrary, the work is what Goscelin claims it to be - his account of his training of a devout mind for the spiritual life and his struggle to prevent his own passion intruding on this process. In his preface Goscelin warns potentially hostile readers against misinterpreting his story in words which may as well apply to the 21° as to the 11"-century reader - ‘May insidious whispers, the evil-seeing eye, the slyly-pointing finger, the muck-raker and the prurient sniggerer be far from the purity of this prayer; ... he who does not like such things need not read it; let him leave what is not written for him to those to whom it belongs.’ (Lib. Conf. 102). EVE OF WILTON AND GOSCELIN OF ST BERTIN AT OLD SARUM .1070-1078 211 Suggested chronology 1060. Birth of Eve. Goscelin joins Hereman at Sherborne. 1066. Eve aged 6. Norman Conquest. 1067. Eve aged 7. Eve’s family settle in Wiltshire. 1071. Eve aged 11. Goscelin with Hereman to Old Sarum/Wilton. Starts to teach the young nuns at Wilton Abbey. Eve sent there for her education. 1075. Eve aged 15. Eve’s profession. 1077. Eve aged 17. Eve and Goscelin attend church dedications. 1078. Eve aged 18. Death of Hereman. Goscelin dismissed from Old Sarum by Bishop Osmund. 1080. Eve aged 20. Eve leaves Wilton for Angers. Goscelin writes Liber Confortatorius. Appendix 2 Goscelin’s narrative. Liber Confortatorius pp. 28-9. Translation. The Birth of Love You remember, my dearest soul, how I first aroused you from your state of illiteracy, confident that I could easily correct so devout a soul. Even so bees choose to make honey from apple juice. I conquered you with my teaching; you conquered me with your gifts. You gave the books I desired, you exalted our St Bertin with great affection, you performed every exercise of charity. Up till then my love for you had been bearable, to all appearances that of a good Christian. Then came the day when you came fearfully to your nuptials with the Lord, the last but one of fourteen virgins bearing candles glittering like the stars and the lamps of Heaven. While the great congregation solemnly watched and the wedding hymn of wondrous grace, ‘I am given in marriage to him whom the angels serve and he has wedded me with his ring,’ was sung, you put on the consecrated garments and made your pledge of constancy to God. Your humble deportment, your trembling approach, your countenance suffused with colour as though from the burning throne of God seated above the cherubim, whose fear is wisdom, pierced me to the very depths of my being. I was touched by the dews of heaven and wept with a passion of tears. Your unbroken silences, your careful self- control, your constant psalmody, your teacher’s report of your piety, only enflamed my desires the more. I arranged that you, yourself now dedicated as an individual, should as a next step attend the dedication of a church, as I wanted you to take advantage of such great sacraments. As you read this ponder with your utmost attention the words which I then poured into your ear, speaking as if I were in travail: ‘Cry aloud in the presence of the Lord’, that is you should earnestly ask one thing only of the Lord, namely that, wounded by holy love, you should desire Christ alone and that with your whole heart and mind you should seek him alone as your dower. “Thus, I said, ‘thus, O soul most dear to me, implore your Lord for the everlasting gift of love: “Give yourself to me”. This is all that you should ask for. In order to receive him alone, this alone you should ask. Cry, wail, knock that it should be opened to you. Strive with the Lord until you triumph; use force against the kingdom of heaven that you may enter. The grace of the Lord 1s offered freely to you; take care that it be not offered to you in vain.’ Later, when our bishop brought the smoking censer chanting, ‘Let my prayer be set before Thee as incense’, what words did I then pour into your breast? ‘This shows how you should enter the house of the Lord. Not with coldness, not with an empty heart, but with the burning incense of love, with the sweetly-smelling glow of virtue. Of such are the many sacrifices given to the angel with the golden thurible, namely innumerable vows and sighs of holy minds which, carried by the angel’s hand, ascend to God as aromatic smoke from the surge and outburst of tears.’ And when, through the fondness of your father, you were at the banquet with your mother, I sent you a fish and charged you with these words to mortify the flesh: ‘Smoked fish; Christ suffered.’ We proceeded to another ceremony of dedication. Once more I was rewarded by being with you - a Mary, by the grace of the Lord, of three days’ duration. It was not the garments of purple, not the bejewelled adornments, not the cloth of gold, not the rows of girls superb in their outward appearance on which I fixed my gaze; it was on you that I gazed, who were as the king of chastity desired, even as the daughter of Jerusalem, clothed in black but beautiful in faith. The holy ceremony over, the congregation dispersed; the darts which had been implanted from on high remained; your wounded heart was faint with love. The frequent missives, long and short, which I wrote brought Christ to you, nor did your own most chaste letters lack serious content. I came very 212 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE often to talk with you; you longed for our meetings as much as I did. Your breast, filled with the arrows of the Lord, knows what groans, what revolts, what weary longings, what empty days and nights, what unattainable love of a mind inflamed by passion I bewailed, so that sometimes when I came in all haste to speak to you as an ambassador of holy salvation I returned having achieved nothing. When we talked together you imbibed the holy admonitions in silence which was more eloquent than speech. Whatever advice I gave, I saw the result in your actions, not your replies. Do you remember you dreamt that I was feeding you, as it were in real life, on the whitest of bread, and that as you ate you filled your mouth with golden dough which you had gently collected in your lap? Think on the bread of life and the word of God which is more to be desired than gold, and from this gold strive to make for yourself an everlasting necklace, like Mary keeping and pondering all these words of the Lord in your heart. Whenever our bishop, your father in God, embraced you and addressed you as: “Eve, mother of the living,’ I made a play on the name, saying, “That Eve was the mother of the living; this Eve will be the daughter of the living.’ Your whole way of life, the place where you lived and your studies were to me by the grace of the Lord redolent of nectar and balsam. It would be tedious to pursue it all; being far away makes no difference and you will not have forgotten. After the death of our father I came more often to find consolation with you in our common bereavement, until there arose a king who knew not Joseph and he who remains faithful to you was driven by viperish envy and step-fatherly barbarity to journey far away. References Primary texts [Goscelin of St. Bertin] The Liber Confortatorius of Goscelin of Saint Bertin. Ed. C. H. Talbot. Studia Anselmiana 37; Analecta Monastica, 3rd series, 1955, 2-117 {Goscelin of St. Bertin] Vita S. Editha. La légende de Ste. Edith en prose et vers par le moine Goscelin. Ed. A. Wilmart. Analecta Bollandia, Ivi, 1938, 5-101,265- 307. [Goscelin of St. Bertin] Historia translationis S. Augustini. Patrologia Latina clv. Aiilarui versus et ludi. Ed. J. B. Fuller. New York: Holt, 1929 [William of Malmesbury] De gestis pontificum Anglorum. ed. N.E.S.A. Hamilton, Rolls Series 1870, 83 [William of Malmesbury] De gestis regum Anglorum ii. ed. W. Stubbs, Rolls Series 1887-9, 389 Other Works BARLOW, F. 1992. The Life of King Edward who rests at Westminster. 2nd edn. Oxford LATZKE, T. 1984. Robert von Arbrissel, Ermengard und Eva. Mittellatinisches Fahrbuch 19, 116-54 MORTON, C. 1975. Pope Alexander II and the Norman Conquest. Latomus Tome xxxiv Fasc. 2 , 362-82. British Library Pamphlet No. 2267 (contains text of Canons promulgated after the Battle of Hastings). Translation in Councils and Synods with other Documents Relating to the English Church I, ed. D. Whitelock, M.Brett, C.N.L.Brooke [Oxford, 1981], ii. 583-4) OTTER, M. 2004. Goscelin of St. Bertin. The Book of Encouragement and Consolation [Liber Confortatorius]. Library of Medieval Women. Cambridge (contains a translation of the text) WILMART, A, 1934, 1938. Eve et Goscelin. Révue Bénédictine 46, 414-38; 50, 42-83. Clarendon. Appendix C. Goscelin of Saint-Bertin and his Works 131-149 Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 213-220 A Possible Site for the Hospital of St John the Baptist and St Anthony at Old Sarum, Salisbury by Andrew B. Powell A ditched enclosure, containing the remains of a medieval stone building and a number of associated graves was revealed in 2002 during archaeological fieldwork at Old Sarum, to the north-east of Salisbury, in advance of the replacement of a water pipeline. The site has been tentatively identified as the hospital of St John the Baptist and St Anthony, possibly originating as a leper hospital dating from the mid-12th century. INTRODUCTION In 2001/02 Wessex Archaeology undertook an ar- chaeological excavation, commissioned and funded by Wessex Water, in advance of the replacement ofa 4.5km long water pipeline to the north of Salisbury, between the reservoirs on Camp Hill to the west and Castle Hill to the east. During topsoil stripping at the base of Castle Hill, the remains of a substantial stone building, with associated graves, were revealed within a ditched enclosure, centred on NGR SU 145 327 (Figure 1). Prehistoric and Roman finds from the excavation are reported in a separate paper (Powell et al. 2005). - The building lies on the south side of Ford Road some 500m east of the Iron Age hillfort at Old Sarum. Ford Road follows the line of the Roman road be- tween Sorviodunum (Old Sarum) and Venta Belgarum (Winchester) and is one of four Roman roads that converged at the eastern entrance to the hillfort. The hillfort earthworks were strengthened in the Saxon period, and again following the Norman conquest when a castle and cathedral were built within them, the site becoming the centre of the medieval town of Saresbyri (Stroud 1986). Abandonment of Old Sarum began in the early 13th century when the bishopric was moved in 1219 to the newly founded city of Salisbury to the south. Archaeological finds on and around Castle Hill, including Iron Age features, dense spreads of Romano-British material and medieval structures, have shown that from the date of the hillfort’s construction, the area outside its eastern entrance was a focus of continuing activity (Stone and Algar 1956; Musty 1959; Musty and Rahtz 1964; Powell et al. 2005). Archaeological finds and documentary sources give indications of the extent of the extra- mural settlement (Wiltshire SMR No. SU13458) (Figure 1), although the layout of the medieval town has yet to be fully established. John Leland, the ‘King’s Antiquary’ to Henry VIII, wrote in 1540 of there having been ‘a fair suburbe’ outside the eastern gate, although by the time he visited ‘ther is not one house neither within Old Saresbyri, nor without it, inhabited’ (Musty and Rahtz 1964, 131). METHODS The pipeline was laid in a 1.5m wide trench within a 10m wide easement that was machine stripped of top- soil and subsoil under archaeological direction; a full description of the methodology for the programme of archaeological works is given elsewhere (Powell et al. 2005). The remains of the excavated masonry building were exposed during topsoil stripping at the point on the south side of Ford Road where one branch of the pipeline ran south to the reservoir on Castle Hill, and another ran east towards Ford Because of the potential significance of the find, tt Wessex Archacology, Portway House, Old Sarum Park, Salisbury SP4 6EB 214 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE ey : // Area shown below Pipeline route Previously suggested locations for the hospital Extra-mural medieval settlement 133000 132000 A36 a : 0 1 2km —— SESE \\ —— eee 7 133000 Philli i NONE bpery 4 \ ty Win) ra fn ry 4a, ‘ . FAT ee y jaar ce WTB APTEVA TLV VAN A a \ iy, : as igen eeu ter ast, Lali 5 oH TH rer veges WU \ \ a Old Sarum Se nse ss Sie | 0 100 200 300m . eS \ \ we Ls Fig. 1 Location of the site in relation to Old Sarum A POSSIBLE SITE FOR THE HOSPITAL AT OLD SARUM, SALISBURY 215 fs, BN Bes a =" N ‘ . _ x N aoe 1. SS o< s caer \ 7 < / 7 : Area shown below —) S 2 8 =k _ = > ad c 6209 Saale = 1 >) 2 anal zs a 1G eee ; > p . z — gl} ; 132710 ai fcaeee er a _—~ 6251 | 5. PR | Wiehe ~ i 6250 ML. \ wh a SS oa — eel Flint ay Ma SS _ cobbling re a ee Sy) ES ama a : j——~— 6252 Li) ea eee {i Section shown in figure 3 + 6249 132700 ay 0 o) 10m ———— = == 2 6247 1 Fig. 2 The excavated enclosure ditches 216 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE N 81 90m0D Key Oo Chalk @p Flint 6251 iL wy 6239 0 0.5 — $j Fig. 3 Section across enclosure ditches 6249, 6250 and 6251 was decided that archaeological investigation should be limited to hand cleaning of the surfaces of the exposed features and deposits within the enclosure, sufficient for a detailed plan to be made, with only sample excavation of the enclosure ditches. Once this had been done, protective measures were installed and this section of the easement was reinstated so as to preserve the remains in situ. The pipeline was diverted around the site along the centre of the Ford Road. RESULTS The following description of the features recorded is necessarily constrained by the limited extent of the archaeological works. The Enclosure On the east, north and west sides of the building the enclosure was represented by between three and five parallel ditches, indicating a number of phases of construction, probably linked to the expansion of the site. Ten sections were excavated through the ditches at the north-west corner of the enclosure (Figure 2). The sampled features were all medieval in date, although they also produced small quantities of prehistoric and Roman finds. The innermost, and possibly the earliest, ditch on the west side (6249) was recorded running north- south for c. 6m before turning at a right angle to the east, where it was almost completely cut on the north side by a later ditch (6250, below). Ditch 6249 was 2m wide and 0.5m deep with concave sides and a wide flat base. It appears to have silted up naturally, although one fill (6181) may be material eroded from a possible bank to the west. The ditch produced medieval pottery, pig and sheep/goat bone, a silver cut halfpenny of Henry II (1154-89) and iron nails. A rounded pit (6252), at least 1m across and 0.4m deep with a steep concave east side, cut by the inner edge of the ditch, also produced medieval pottery and animal bone from its upper two fills. Up to 1.5m outside ditch 6249 to the west, there was a shallow gully (6188), possibly marking a hedge-line, also cut across at the north by ditch 6250. It was c. 1m wide and up to 0.2m deep. Its single fill (6189) produced three medieval sherds. A further 1m to the west were two overlapping ditches, the outer one (6251) representing a widening and re-cutting of the inner one (6250) (Figure 3). Ditch 6250 was recorded running north for 10m be- fore turning ata right angle to the east, cutting across gully 6188 and ditch 6249. It may have extended some 22m to arounded terminal. Its full dimensions could not be ascertained, although it was at least 2m wide and 0.7m deep, with a moderately steep inner side. It produced two sherds of medieval pottery, ceramic building material (CBM), a fragment of iron sheet or plate, a piece of lava (probably a quern fragment) and cattle and sheep/goat bone. Ditch 6251 was up to 2.6m wide and 1m deep with straight, moderately steep sides and a narrow flat base, although the inner side, which cut the fills of ditch 6250, was slightly shallower and concave. An initial silting layer was overlain by material ap- parently deriving from the outside, possibly from an external bank. A layer of material eroded from the fills of ditch 6250 produced the majority of the finds, comprising five medieval sherds, and pig and cattle bone. The upper 0.6m of the ditch appears to have filled in gradually. Along the north side of the enclosure this ditch diverged northwards from A POSSIBLE SITE FOR THE HOSPITAL AT OLD SARUM, SALISBURY 217 ditch 6250 to a terminal some 3m outside, but on the same line as, the inner terminal. The possible entrance to the east of the ditch 6251 terminal had a ().8m wide linear feature (not excavated) lying across it. An opposing terminal was not recorded — if such existed, it underlay surviving deposits and features associated with the building. The outermost ditch on the west side (6247) lay 2.5m outside ditch 6251. It was 1.25m wide and (0.55m deep with steep convex sides and a 0.35m wide flat base. Only a single fill was recorded, and this was disturbed by the roots from an existing hedge. This deposit produced a moderate quantity of finds, including medieval pottery, four pieces of incised slate, animal bone (cattle, pig, sheep/goat and one bird long bone), oyster shell, 0.8kg of CBM (some glazed) and a piece of shelly limestone — probably building material. Along the north side of the enclosure, a gap of up to 1.2m separated ditch 6251 from the outermost recorded ditch (6209) along Ford Road, which continued eastwards past the terminals of ditches 6250 and 6251. It was up to 2.6m wide and 1.1m deep, with convex sides, steep on the south side and near vertical towards the base, but shallower on the north side, and with a flat 0.55m wide base. Nine fills were recorded, the material in the lower five all appearing to derive from the outer side of the ditch, and containing medieval roof tile and animal bone (sheep/goat and possibly cattle). The overlying layer filled the resulting hollow on the south side, followed by two layers deriving from the inner side of the ditch. The upper part of the ditch may have been recut at this point, as the upper levels of these fills appear to have been truncated, the subsequent _ layers of natural infill producing further medieval pottery and CBM as well as bone and shell. To the east of the building, the inner (6304) of the three parallel ditches (none of which was exca- vated) was cut by two of the graves in the graveyard (see below). It was up to 1m wide, and separated by a gap of 0.7m from a second, 2m wide ditch (6305). The outer, 1.8m wide ditch (6310) lay a further 2m to the east (Figure 4). Other than suggesting a simple sequence of ditches, from the inner to the outer, involving the gradual enlargement of the enclosure to the east and west, it is not possible to identify which ditches were contemporaneous, other than perhaps the inner ditch on the east and west sides (6304 and 6249). The frontage of the enclosure along Ford Road appears to have increased from 63m to at least 78m over the period of occupation. How far the enclosure extended to the south is not known. The relatively steep slope of Castle Hill to the rear, and the fact that the ditches at the west curve slightly inwards, suggest that it may not have extended far back from the road. A subsequent geophysical survey of the site (Bartlett 2003) detected a number of resistivity anomalies and linear magnetic anomalies that may indicate drains or watercourses associated with the building. The latter include two parallel magnetic anomalies run- ning parallel to Ford Road, some 50m to the south, suggesting the possibility of a similar sequence of enlargement at the back of the enclosure. The Building Only the northern part of the building was exposed within the pipeline easement, its main axis ly- ing approximately east-west, and parallel to Ford Road (Figure 4). Certain anomalies detected by the geophysical survey may indicate other structural remains to the south (Bartlett 2003). The surviving walls of the building were constructed from a mix- ture of chalk blocks and flint nodules and at least two phases of construction were identified. The most substantial part of the structure was recorded at the east, where there was a substantial flint-faced chalk wall, 5.7m long, running north— south. At the north it was 2.2m wide, but at the south narrowed on its outer face to just 0.7m wide. It then turned west for 1.8m, where a length of less substantial flint wall continued the line west for a further 2.4m. At the north-east corner of the larger wall was a greensand quoin, adjacent to a bedding of mortar containing in situ traces of tiling. The wall foundation running west from this corner had been robbed for some 10m before dog-legging slightly to the south and continuing for a further 4m as a surviving wall. Here it encountered a north-south wall, and in the south facing corner formed by these two walls there was a tile-lined feature. The space defined by these walls may represent the original structure. Near its centre was a large area of disturbed floor tiles. Where the tiles overlay two of five graves, subsidence into the grave fill had protected them from subsequent plough damage. The western wall of this structure ran north for some 4.4m from the southern edge of the easement, and after a gap of 2.5m continued for 1.5m to join an east-west wall recorded along the north edge of the easement (see below). Running through the gap in this wall, and extending 15m east and 6m west, was a 1.2—2m wide surface of compacted redeposited chalk. 218 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Area shown below Robbed wall ) Cut feature Other S S t oO wo © + ae _ é + vt 132720 Flint cobbling 0 10 20 30m _ 8 g < N < os | os | =a Fig. 4 The building foundations, the cemetery, and the eastern enclosure ditches Running west from the north-south wall, on the same line as the robbed foundation, was a 9m long wall, with another running south from its mid point (an open well, just outside the easement, would have cut this latter wall). To the west was a surface of flint cobbles, almost 5m long and continuing south beyond the easement. Immediately west of the cobbles were two short linear features, possibly beam slots, the western one cut by a circular feature 5m in diameter which extended up to the terminal of ditch 6250. The east—west wall along the north edge of the easement was recorded running from beyond the easement at the north-east, westwards for some 34m, although robbed in places and elsewhere lying just outside the easement. At the west it lay outside the enclosure formed by ditch 6251; its relationship to the outermost ditch (6209), however, could not be ascertained (Figure 2). It may therefore represent a late boundary wall replacing earlier bank and ditch A POSSIBLE SITE FOR THE HOSPITAL AT OLD SARUM, SALISBURY 219 boundaries along the road frontage. A 3m length of east—west wall was recorded just south of it in the western part of the building. The Graveyard Over 30 east—west-aligned graves were exposed by topsoil stripping within the enclosure. Some of the graves appear to have been shallow as a number of skulls were exposed at their western ends. Five graves lay within what may have been the original structure, at least two of them below a tiled floor, with most of the rest to the east and south-east. At least eleven graves at the south-east were arranged in two rows, some of their ends overlapping; the easternmost graves in both rows cut the fill of the inner enclosure ditch on the east side. The Finds Most of the pottery recovered comprised locally made Laverstock-type fabrics, including both coarse- and finewares. These had a relatively long currency — the coarse variants were present in 11th and 12th century levels at Old Sarum (Musty and Rahtz 1964), and continued in use into the 13th century in the new city, with a gradual tendency towards the finer variants. Vessel forms indicate that there is a potential 12th century component to the assemblage in the form of glazed sherds, some with combed decoration, which probably derive from tripod pitch- ers. There was also one ‘West Country’ or inturned dish, a 12th/13th century form, and one jar rim. The presence of fineware sherds, however, presumably _representing glazed jugs in the Laverstock tradition, extend the date range well into the 13th century and possibly beyond. The assemblage of CBM consists mainly of flat roof (peg) tiles, some partially glazed on the lower part of the upper surface. A few ridge tiles were also identified, partially glazed with knife cut crests. Most were in irregular, poorly mixed fabrics containing varying frequencies of quartz sand and prominent iron oxides. Such fabrics are commonly found in south-east Wiltshire and west Hampshire. One po- tential source, known from documentary evidence, was at Alderbury, 5km to the south-west of Salisbury, which was in operation from at least the mid 14th to late 1Sth century (Hare 1991, 89), although evidence from Salisbury indicates that ceramic roof tiles were in use from the earliest period of the city in the early 13th century. The bulk of the assemblage came from the cleaning layer within the building. A silver coin, a cut halfpenny of Henry II (1154— 89), was found in the innermost of the enclosure ditches (6249). A few pieces of medieval window glass and window lead were also recovered during cleaning, from immediately east of the thick wall at the eastern end of the building. DISCUSSION Because of the limited investigation of this site, interpretation of these features must remain provi- sional. One possible, and indeed likely, interpreta- tion is that the structure was the medieval hospital of St John the Baptist and St Anthony, which is referred to in documentary sources. Although its location has yet to be conclusively identified, it seems clear, from a review of the evidence by Musty and Rahtz (1964, 132), that the hospital was located in the eastern suburb of Old Sarum. The hospital received a grant of royal protection in 1231] and was still in existence in 1535 when the Valor Ecclesiasticus, Henry VIII's survey of ecclesiastical property in England and Wales, described ‘the Hospital or Free Chapel of St John, near the castle of Old Sarum’ as owning half an acre of pasture and 15 acres of arable. In 1540, Leland, having recorded that the eastern suburb of Old Sarum had been recently deserted, mentioned that a chapel, which belonged to the former parish church of St John, still stood on the site (RCHM(E) 1980, 12). The absence of late medieval material from the finds assemblage tempers this interpretation. The hospital may have started as a leper hos- pital that is known to have existed in 1195. Leper hospitals frequently had marginal locations on the outskirts of towns, as was the case with this build- ing. Because of the social stigma attached to leprosy — lepers were considered to be the ‘dead among the living’ — leper houses were set up with the aim of segregating those suffering from the disease. In 1175, the English Church Council decreed that lepers should not live among the healthy, and in 1179 the Lateran Council at Rome insisted that leper com- munities should have their own priests, churches and cemeteries (Richards 1997). Leprosy is present in British populations from the post-Roman period (Roberts and Cox 2003, 218), but reached its peak in 12th century, declining rapidly in the 14th, so that such institutions had a relatively short lifespan, being then often converted to more general use. Not all medieval hospitals, how- ever, were primarily for the sick (many in fact refused 220 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE admission to the sick), providing instead hospitality and shelter for travellers and pilgrims; this may be reflected in the site’s roadside location. A number of other locations within Old Sarum have previously been suggested for the hospital (Fig- ure 1). In 1843, Benson and Hatcher placed it some 300m to the west (at SU 142 327), in the position later recorded on the 25 inch OS map (Benson and Hatcher 1843). There appear to be no substantive reasons for this location, although the floor levels of two associated timber buildings with 12th—13th cen- tury pottery were subsequently recorded in a pipeline trench at that location (WANHM 1973, 136). In 1933, the remains of a stone building with fragments of 13th century carved stonework were found by Stone and Charlton north of the Old Castle Inn (at SU 1414 3261) which they suggested may have been part of the church of St John (Stone and Charlton 1935). They also found many skeletons, the same area producing a large number of graves between 1834 and 1960 and representing a large cemetery. One grave contained a paten and chalice (Akerman 1855) indicating the grave of an ecclesi- astic. Others were mass graves, with superimposed skeletons suggesting rapid burial, perhaps due to plague. None of the skeletons, however, showed signs of leprosy, and the cemetery may represent the main burial grounds of the borough (RCHM(E) 1980, 12). A combination of features at the site described above support its identification as the hospital of St John the Baptist and St Anthony. These include its roadside location set some distance from the main focus of settlement, the maintenance of a clear boundary around the site, and the presence of a burial ground within the enclosure. It is not possible at present to determine whether any part of the building was a church or chapel; in many medieval hospitals the sick were laid in the body of the church, which functioned as the infirmary hall, with the chancel as their chapel (Platt 1978). The presence of graves within the eastern part of the building, with its substantial eastern wall and frag- ments of window glass and lead, make that structure a possible candidate for the chapel. Acknowledgements The archaeological investigations were commis- sioned by Wessex Water. Wessex Archaeology are grateful for the help and support of Neil Callard, Liz Clarke, Steve Dodwell, Andy Gale and Dave Jones of Wessex Water during the project. The field- work was monitored on behalf of Wiltshire County Council by Roy Canham, Wiltshire County Council » Archaeological Officer. The project was managed for Wessex Archae- ology by Jonathan Nowell. The excavation was directed by James Chapman with the assistance of Neil Fitzpatrick, Becky Fitzpatrick, Hannah Mar- riot, Kirsten Egging, Russell Gant and Laura Cassey. Rob Goller prepared the illustrations. Bibliography AKERMAN, J.Y., 1855, Old Sarum, in Notes of antiquar- ian researches in the summer and autumn of 1854, Archaeologia 36, 175-86 BARTLETT, A.D.H., 2003, Old Sarum ‘Chapel Site’, Wiltshire: Report on Geophysical Survey 2003. English Heritage Centre for Archaeology Report 60/2003 BENSON, R. and HATCHER, H., 1843, History of Mod- ern Wiltshire: Old and New Sarum, or Salisbury, vol. 6. London HARE, J.N., 1991, The Growth of the roof-tile industry in Later Medieval Wessex, Medieval Archaeology 35, 86-103 MUSTY, J.W.G., 1959, A pipe-line near Old Sarum: prehis- toric, Roman and medieval finds including two twelfth century lime kilns, WANHM 57, 171-191 MUSTY, J.W.G. and RAHTZ, PA., 1964, ‘The suburbs of Old Sarum’, WANHM 59, 130-54 PLATT, C., 1978, Medieval England: A Social History and Archaeology from the Conquest to 1600 AD. London: Book Club Associates POWELL, A.B., ALLEN, M.J., CHAPMAN, J., EVERY, R., GALE, R., HARDING, P, KNIGHT, S., MCKINLEY, J.I. and STEVENS, C., 2005, Excavations along the Old Sarum water pipeline, north of Salisbury, WANHM 98, 250-280 RCHM(E), 1980, Ancient and Historical Monuments in the City of Salisbury, Volume 1. London: HMSO RICHARDS, P, 1997, The Medieval Leper. Suffolk: Boydell & Brewer ROBERTS, C. AND COX, M., 2003, Health and Disease in Britain: From Prehistory to the Present Day. Stroud: Sutton Publishing STONE, J.ES. and ALGAR, D.J., 1956, Sorviodunum, WANHM 56, 102-26 STONE, J.ES and CHARLTON, J., 1935, Trial excavations in the east suburb of Old Sarum, Antiquaries Journal 15, 174-92 STROUD, D., 1986, The site of the borough at Old Sarum 1066-1226: an examination of some documentary evidence, WANHM 80, 120-126 WANHM, 1973, Wiltshire archaeological register for 1972, WANHM 68, 126-39 Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 221-229 Heraldry and Hostelries: the evidence from Wiltshire by R. Moody In 1877, the Rev. A C. Smith presented to the Wiltshire Archaeological and Natural History Society a paper entitled “Some Account of the Tavern Signs of Wiltshire and their Origin”. It is likely that Smith would have been aware of some fifty-six establishments displaying the arms of the families who at that time controlled and exercised very considerable influence over large areas of the county. In the intervening years there has been a mass exodus of most of these families from their lands. In this paper an account is given of the establishment and subsequent departure of the majority of these families with an overview of the arms of the families that can nevertheless still be seen. In conclusion, consideration is given to the reasons why these families were either unable or unwilling to retain ownership of their ancestral properties. In August 1877 the Rev. A.C.Smith presented to the Wiltshire Archaeological and Natural History Society at its meeting in Warminster a paper entitled “Some Account of the Tavern Signs of Wiltshire and their Origin”. In this paper he divided the signs into nine categories, the first of which he entitled “Heraldic”. In the course of his paper Smith wrote: The Heraldic of right occupies the first place in my catalogue not only because it ranks first in priority of time ... but also because in Wiltshire it still holds the first all pointing to the great Wiltshire families which flourished when these severai inns were established...and I venture to say that if the aggregate area of the combined domains of these large landed proprietors, whose arms still decorate our sign-boards, could be reckoned, it would be found that a very considerable portion of our county was still in their hands: while some few which, in the vicissitudes of families, have dwindled and decayed, the coat of arms, swinging over the village inn, alone remains, a melancholy memorial of bygone greatness. place in priority of numbers. Thus, by way of example, we have the Bath Arms, at Warminster; the Ailesbury Arms at Marlborough; the Lansdowne Arms, at Calne; the Pembroke Arms at Wilton; the Suffolk Arms at Malmesbury: and not only in regard to noble lords, but in like manner we have the Long Arms, at Steeple Ashton, Keevil and Wraxall; the Wyndham Arms, at Dinton; the Goddard Arms, at Swindon and Clyffe Pypard; the Penruddocke Arms, at Barford St.Martin; the Phipps Arms at Westbury; the Benett Arms at Tisbury;... “Fair Orchard’, South Lidcombe, East Harptree, Bristol BS40 6BL Smith found no less than fifty-six heraldic sign-boards, this number probably taken from the then current edition of Kelly’s Directory which helpfully provided a complete list of all hotels, inns and taverns throughout the county, although only fifty-five establishments that would be expected to display a non-royal coat of arms are listed in the 1875 directory.* The total may be divided the total into two categories. The first comprises the fifteen signs bearing the arms of the six great aristocratic landowning families - the Brudenell-Bruces of Tottenham House, the Herberts of Wilton, the Howards of Charlton Park, the Petty-Fitzmaurices of 222 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Bowood, the Pleydell-Bouveries of Longford Castle and the Thynnes of Longleat who between them in 1873 controlled some 140,000 acres in Wiltshire. Their history is too well known to be repeated here and they still largely flourish notwithstanding the very large reduction in the total area of their land. The second category comprises the remaining signs bearing the arms of the landed gentry and of those aristocratic families whose land holdings were similar to, and in some cases smaller than, those of their non-aristocratic neighbours. The purpose of this paper is firstly to identify each of these remaining families, to show when they were either first established or purchased land in their respective vicinities as a result of which their arms were adopted by usually the chief or only inn on or near their estates. It will also note the subsequent disposal of their properties; such disposal in several cases resulting in the replacement of their arms altogether. The inns bearing the arms of the remaining families that Smith would have found in the list mentioned, have had varied histories, as will be shown. 1. The ARUNDELL Arms in Ansty. In 1973 this inn was renamed The Maypole Inn’ and became a private house in 1997. The manor of Ansty was purchased by Sir Matthew Arundell in 1594 and remained in his descendants’ hands until the estate was forfeited by Henry Arundell as a royalist. Henry, Baron Arundell recovered the manor at the Restoration and it then formed part of the Wardour estate. However, the title became extinct on the death of the 16 Lord Arundell in 1944 whereupon the estate passed to his kinsman, R.J.A.Talbot, who assumed the name of Arundell. He subsequently sold all his land in Ansty other than some woodland that passed to his son R.J.R.Arundell.° 2. The ARUNDELL Arms in Tisbury. This inn, closed in the 1960s, and is now a private house. In 1540 the manor of Tisbury was granted by the Crown to Sir Thomas Arundell who, in 1547, bought Wardour manor and castle. The manor of Tisbury remained in his family (with a short break between 1552 and 1554 and another in the middle of the 17" century) and passed with the Wardour estate® until , after a number of sales in the first half of the 19" century, nearly the whole of the remainder of the family’s land in Tisbury was sold as part of the Wardour estate in c. 1946.’ 3. The BEACH Arms in Keevil. This inn was closed at the end of the last century and is now a private house. The Manor of Keevil was purchased by — William Beach in 1681. His descendant Henrietta Maria, daughter of William Beach who died in 1790, married Michael Hicks of Beverstone Castle, near Tetbury in Gloucestershire.’ The family was thereafter known as Hicks Beach and in 1911 the estate was sold off in lots by W.A. Hicks Beach.’ 4. The BECKFORD Arms in Fonthill Gifford. There is a sign bearing the Beckford arms outside the inn at Fonthill Gifford. An estate at Fonthill was purchased in 1745 by William Beckford, who was to be Lord Mayor of London in 1763 and 1770.'° Following his death the estate passed to his son, William, who in 1823 sold it to John Farquhar.'! 5. The BENETT Arms in Tisbury. There is a sign bearing the arms of Benett of Norton Bavant outside the inn at Tisbury. In 1752 the Pythouse estate was purchased by Thomas Benett of Norton Bavant whose mother was a daughter of John Benett whose family had been in possession of the property for at least 200 years until sold in 1669. His descendant John Benett purchased large areas of land in the neighbourhood in the first half of the 19" century.’? Much of this land was sold prior to 1985, with Pythouse and the surrounding park being sold to Mutual Households Association Ltd in 1958.'? However, some land remains in the ownership of a descendant of John Benett. 6. The BENETT Arms in Semley. There is a sign bearing the arms of Bennett of Pythouse outside the inn at Semley. Land in Semley formed part of the Pythouse estate in 1580 when it belonged to a family called Pytt alias Bennett in the 1563 Herald’s Visitation.'* Further, in 1623 Thomas Bennett purchased Callis Place and in 1631 granted it to his third son William in whose family it remained untii acquired by Thomas Benett of Pythouse in 1796! following the death of John Bennett in1772.!° John Benett of Pythouse bought 441 acres of Semley manor between 1806 and 1820 and in 1839 his land in Semley extended to over 1,000 acres.'’ The Benett family’s land in Semley subsequently formed part of the Pythouse estate, much of which was sold by the trustees of Evelyn Fane-Benett-Stanford and by his kinsman Sir HERALDRY AND HOSTELRIES: THE EVIDENCE FROM WILTSHIRE (Horace) Anthony Rumbold, Bt. between the 1950s and the 1970s.!® The BENETT Arms in East Hatch, West Tisbury. Land in East Hatch formed part of the Pythouse estate (see 6 above) in 1585.'? In c.1882, Vere Fane-Benett-Stanford, the owner of Pythouse, invited the householders of East Hatch, West Hatch and Newtown to indicate in a book at the School House at Newtown and at Mr.Street’s Hatch Farm their opinion as to the desirability of the inn continuing as a public house. A majority must have voted to see the inn closed as no mention is made of it in the 1885 or later editions of Kelly’s Directory.” The BOLINGBROKE Arms in Hook Lydiard Tregoze. There is a sign bearing the St. John arms outside the inn at Hook. The Lydiard Park estate passed into the hands of the St. John family when in 1420 Margaret Beauchamp, whose first husband was Oliver St. John, succeeded her brother as heir to their father’s estate.’! Henry St. John was created Viscount Bolingbroke in 1712. The mansion was generally the secondary residence of the St. John family. In 1920 just over 1000 acres of the estate were sold and in 1930 a further 1,800 acres. Following the death of Mary Emily Elizabeth, Viscountess Bolingbroke in 1940 about 750 acres were sold by her executor including Lydiard Park and 147 acres to Swindon Corporation.” The BURDETT Arms in Ramsbury. The arms of Burdett are displayed outside the inn which has the happy distinction of once again bearing the name of Burdett having for a while been known as The White House. Ramsbury Manor was inherited in1800 by Sir Francis Burdett from his aunt Elizabeth Langham Jones.” Following the death of Sir Francis Burdett in 1951 Ramsbury Manor and Park farm were sold.” . The CALLEY Arms in Hodson. There is a sign bearing the Calley arms outside the inn in Hodson. The manor of Chisledon, which included Hodson, was purchased by William Calley, citizen and draper of London, in 1619.~ The manor thereafter remained in the possession of the Calley family and, following the death of Miss Joan Calley in the 1970s, the estate passed to her nephew Sir Henry Langton, who changed his name to Calley. After his death the estate passed to his nephew. Burderop Park is ll. 2: 13. 223 no longer the residence of the descendants of the Calleys although they continue to own land in the neighbourhood.” The CALLEY Arms in Wanborough. There is a sign bearing the Calley arms outside the inn in Wanborough.”’ It does not appear that members of the Calley family held land in Wanborough but no doubt their position as large landowners in this part of the county resulted in their name being given to this inn. The CHURCHILL Arms in West Lavington. There is a sign bearing the Churchill arms outside the inn in West Lavington. The Duke of Marlborough purchased the manor c.1766 and in the early 19" century it was settled on his second son who was created Baron Churchill.” In 1875 Lord Churchill offered 1,113 acres of land for sale’ and in 1905 the manor house and most of the land then belonging to the 3“ Baron (created Viscount Churchill in 1902) was sold to H.T. Holloway.” The CLARENDON Arms in Wootton Bassett. This inn no longer exists having been closed in 1910, converted into a house and demolished in 1969.*! In 1676 the manor of Wootton Bassett, with most of the land in the area, was purchased by Laurence Hyde, later Earl of Rochester, and on his son’s death the estate passed to Charlotte Villiers whose husband succeeded to the Earldom of Clarendon. He died without a male heir and his estates were inherited by his granddaughter whose husband was created Earl of Clarendon in 1776.The estate at Wootton Bassett ran with the Earldom of Clarendon until 1866 when it was sold to Sir Henry Meux.*” . The CLEVELAND Arms in Ashton Keynes. This inn, formerly known as The New Inn was renamed The Cleveland Arms when the owner succeeded to the Dukedom of Cleveland*® but was closed early in the 20" century. In c.1848 the Ashton Keynes estate and the Lordship of the Manor were purchased by Lord Harry Vane, later the 4° Duke of Cleveland. The Duke of Cleveland left the estate to A.W.H. Hay-Drummond who in 1913 offered the whole of it for sale by auction. Part was sold, with the remainder being disposed of by private treaty thereafter. In 1917 the Lordship of the Manor was purchased by A.W.Bowley.* 224 15. iG: Le 18. 19. THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE The CODRINGTON Arms in Nettleton. The name of this inn was changed in the closing years of the 20" century and shortly thereafter became a private house. In 1843 Sir C Bethel Codrington was the owner of land at Nettleton and West Kington.*” The Codrington family were long seated at Dodington Park over the county boundary in Gloucestershire. The FREKE Arms in Swanborough There is a sign bearing the Freke arms outside the inn in Swanborough. In 1605 the manor of Hannington was acquired by Sir Thomas Freke.** The estate remained in the family until sold by the widow of Ambrose Denis Hussey-Freke in 1921 to Claude B Fry.*? However, in 1957 Hannington Hall was repurchased by a member of the Hussey-Freke family in whose hands it remains."” The GODDARD Arms in Clyffe Pypard. There is a sign bearing the Goddard arms outside the inn in Clyffe Pypard. The manor of Clyffe Pypard was purchased by John Goddard of Aldbourne in 1530.4! On the death of Horatio Nelson Goddard in 1900 the Clyffe Pypard property passed to his daughter Frances the wife of William Wilson” following which part of the estate was sold with a further sale in the 1940s (including the contents of Clyffe Pypard Manor in 1940).* In 1992 the Manor House was offered for sale and so passed out of the hands of the descendants of the Goddards.** The GODDARD Arms in Swindon . There is a sign bearing the Goddard arms outside the inn in Swindon. The manor of Swindon was purchased by Thomas Goddard of Upham in Aldbourne in 1562. After the death of Fitzroy Pleydell Goddard in 1927 his widow continued to live at the Lawn (formerly known as Swindon House and built on the site of the medieval manor house) until 1931. In 1943 the house and grounds were sold to Swindon Corporation who demolished the house in 1952.*° The GROSVENOR Arms in Hindon. In 2001 the name of this inn was changed to the Angel Inn by which description it had been known in 1830.’ According to Colt Hoare, Earl Grosvenor purchased half of the borough of Hindon formerly belonging to William Beckford*. It is likely, therefore, that this purchase gave rise to the naming of the Grosvenor Arms mentioned in a trade directory of 1842.4? The manor was 20. 2 D2: 23% purchased by Richard Grosvenor, Marquis of Westminster in the 1850s°° and eventually ~ passed with the Fonthill Abbey estate to Sir Michael Shaw-Stewart, Bt. whose wife was a Grosvenor.*! In 1922 his son W.R. Shaw-Stewart sold 193 acres in Hindon, Chicklade and Fonthill Gifford’; in 1925 he put 2062 acres of land on the market and the manor was thereupon broken up. The GROVE Arms in Berwick St.John. Between 1899 and 1905 (probably following the sale of Ferne in 1902) the name of this inn was changed to The Talbot Inn. In 1658 Upton Farm was purchased by Thomas Grove of Ferne.» In 1792 Thomas Grove of Ferne bought Lower Bridmore Farm and in about 1842 Manor Farm was also purchased.*° However, in 1902 Ferne House and park and most of the family’s land in Berwick St John were sold by Sir Walter Grove Bt. to Major A.H.Charlesworth*’ with Easton Farm being sold by Sir Walter Grove in 1920 to William Follett.** The LONG’S Arms in South Wraxall. There is a sign bearing the Long arms outside the inn in South Wraxall. Robert Long had a house in Wraxall in 1429 and in 1448 Wraxall became the principal seat of his family. Wraxall Manor remained in the hands of the family until 1919 when the estate was broken up and most of it was sold off..? However, the manor house and 830 acres were excluded from the sale but were sold by Viscount Long’s daughter, Mrs Sara Morrison, in 1966. The LONG’S Arms in Steeple Ashton. There is a sign bearing the Long arms outside the inn in Steeple Ashton. The manor of Steeple Ashton was purchased by Walter Long of Whaddon who died in 1672 and the manor house and farm by Richard Long of Rood Ashton in 1799. Most of the land of the Longs in Steeple Ashton was sold in the 1930s. The manor house was for a time in the occupation of the present Viscount Long of Wraxall who no longer resides there. The LONG’S Arms in North Bradley. There is a sign bearing the Long arms outside the inn in North Bradley. In 1426 Robert and Margaret Long were parties to a deed conveying the manor of North Bradley to feoffees and thereafter the manor descended in the family of Long of Wraxall and later in the family of HERALDRY AND HOSTELRIES: THE EVIDENCE FROM WILTSHIRE 24. 25: 26. 27. 28. Long of Draycot Cerne and finally to the first Earl Cowley. He sold the estate to C.P Moore the tenant of the Manor Farm who in 1879 sold it to Walter Hume Long of Rood Ashton. He owned it until the break up of the Rood Ashton estate in 1911 when Manor Farm, with other property in North Bradley with some chief rents and such manorial rights as belonged to him as Lord of the Manor of North Bradley, was sold.°! The LOPES Arms in Westbury. In 1989 the name of this inn was changed to The Westbury Hotel. Heywood House was purchased in 1789 by Gaisford Gibbs, a Westbury clothier whose daughter Susannah married Abraham Ludlow.” On the death of their son Henry and his wife the estate passed to their nephew Henry Charles Lopes who became 1*' Baron Ludlow of Heywood inl897. In 1920 Heywood House, with about 1,368 acres of land, was offered for sale.°’ His son, the 2" Baron, died in 1922 and in 1924 Heywood House with 225 acres was sold by mortgagees to Captain Walter Shaw.” The LUDLOW Arms in Westbury. See 24 above. There is a sign bearing the Ludlow arms outside the inn in Westbury. The MALET Arms in Newton Tony. There is a sign bearing the Malet arms outside the inn in Newton Tony. Wilbury House was purchased by Sir Charles Malet Bt. c. 1803. In 1925 c.1,000 acres of land were sold by Sir Harry Malet and at about the same time Wilbury Park, with a further c.764 acres, were sold to Major Despencer Robertson.” The MALMESBURY Arms in Salisbury. The name of this inn was changed to The Hogshead c. 1985. In the 1790 enclosure of Fisherton Anger (the area of Salisbury in which this inn is to be found) James Harris 1‘ Baron Malmesbury, later created Earl of Malmesbury, was allotted 103 acres within the parish® which in thel9"" and early 20" centuries was largely developed for industrial and housing purposes. The METHUEN Arms in Corsham. There is a sign bearing the Methuen arms outside the inn in Corsham. Corsham House (as it was then called) was purchased by Paul Methuen in 1745." Although the present Lord Methuen no longer lives at Corsham and the title is now separated from the ownership of the estate, a member of he) 30. ie 225 the Methuen family still lives at Corsham Court and owns a considerable amount of land in and around the town of Corsham. The MOMPESSON Arms in Oare. This inn, formerly known as The Green Man, was renamed The Mompesson Arms after its acquisition by a member of the Rogers family of Rainscombe, one of the family having intermarried with the Mompessons. Soon after 1889 this inn was renamed The White Hart.°* Rainscombe was purchased in 1582 by Henry Pyke whose descendant married Henry Rogers who died in 1721. Rainscombe remained in the Newman Rogers family until the estate was sold partly after the first World War, the remainder being sold in 1940.” The NORTHEY Arms in Box. There is a sign bearing the Northey arms outside the inn near Box. In the 18" century William Northey purchased considerable property in Box including, in 1726, Hazelbury Manor.” In 1912 G.E.Northey sold the Ashley Manor Estate comprising some 900 acres’! and the Northey family continued to live at Cheney Court until it was sold in the 1960s. Hazelbury Manor remained in the possession of the Northey family until 1919 when it was sold by Sir Edward Northey to G.J. Kidston.” The PENRUDDOCKE Arms in Barford St Martin. There is a sign bearing the Penruddocke arms outside the inn near Barford St. Martin. Compton Chamberlayne was purchased by Sir Edward Penruddocke in c. 1598.’ Compton Park remained in the Penruddocke family until 1930 when it was sold by the executors of the will of Charles Penruddocke.” . The PETERBOROUGH Arms in Dauntsey. There is a sign bearing the Peterborough arms outside the inn in Dauntsey. The manor was granted in 1691 to Charles Mordaunt who, in 1697, became Ear! of Peterborough. The manor descended to Charles, Earl of Peterborough who died in 1814 whereupon it passed to his niece Jane who conveyed it to her son Mordaunt Fenwick. He sold it in1853 to C.W. Miles. . The PHIPPS Arms in Westbury Leigh. There is a sign bearing the Phipps arms outside the inn in Westbury Leigh. Henry and Nicholas Phipps acquired property in Westbury towards the end of 226 34. 35. 36. Bye THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE the 16" century” and in the succeeding centuries members of the Phipps family purchased large areas of land in Chalford and Westbury Leigh, eventually occupying both Chalcot House and Leighton House. In 1888 Richard Leckonby Hothersall Phipps sold the Leighton House estate to William Henry Laverton.” The Chalcot estate eventually passed to Charles Bathurst Hele Phipps who died in 1960” and in 1968 Chalcott House was offered for sale.” The POORES Arms at Charlton . Between 1920 and 1921 this inn was renamed The Charlton Cat.*° In 1784 the manor of Charlton was purchased by Edward Poore and thereafter descended with the manor of Rushall which had been bought by him in 1749. Both manors remained in the family until sold in 1838 by Sir Edward Poore to Welbore Agar, Earl of Normanton.*! The SEYMOUR Arms in East Knoyle. There is a sign bearing the Seymour arms outside the inn in East Knoyle. This inn was at one time known as The Benett Arms® (see 5 and 6 above). Between 1843 and 1862 large areas of land in East Knoyle were purchased by members of the Seymour family.*’ In 1877 Alfred Seymour sold most of this property apart from Knoyle House, and his remaining land was eventually sold in 1948 following his daughter’s death in 1943." The SOMERSET Arms in Maiden Bradley. There is a sign bearing the arms of the Dukes of Somerset outside the inn in Maiden Bradley. At the dissolution of the monasteries, Thomas Seymour became the owner of the Priory and the land belonging to it at Maiden Bradley. On his death the property passed to the 1*' Duke of Somerset, the Lord Protector.* In 1951 and 1953 numerous properties in and near Maiden Bradley were offered for sale® and in the following years there were further sales of large areas of land. However Maiden Bradley House remains in the ownership of the Seymours and is occupied by the present Duke of Somerset and his family. The SOMERSET Arms in Semington. There is a sign bearing the arms of the Dukes of Somerset outside the inn in Semington. By 1800 the manor of Semington belonged to the Duke of Somerset.*” In 1925 the outlying portions of the Duke of Somerset’s estate, including land at Semington and The Somerset Arms, were offered for sale.* 38. The WELLESLEY Arms in Sutton Benger. There is a sign bearing the Wellesley arms outside the inn in Sutton Benger. Sir Robert Long owned the manor in1576 and the estate then descended in the Long family until Catherine Long married William Pole-Tylney-Long-Wellesley, later Earl of Mornington. Following the death of the 5" Earl of Mornington in 1863 the estate passed to William, Earl Cowley and remained in his family until the estate was broken up and sold in the 1920s.*° 39. The WYNDHAM Arms in Dinton. There is a sign bearing the Wyndham arms outside the inn in Dinton. The name of this inn was changed to The Nadder Arms” before becoming a private house c. 1960. However, another Dinton inn formerly known as The East End Inn has now been renamed The Wyndham Arms. In 1689 the Dinton Park estate was purchased by William Wyndham. The estate descended in the Wyndham family until sold in 1916 to Bertram Erasmus Phillipps.”! Smith’s paper appeared very shortly after the publication of the comprehensive, if not always entirely accurate, 1873 Return of Owners of Land in the county which revealed that at that time almost 111,000 acres of land were owned by 24 of the families dealt with in this paper - the names of the remaining families not appearing in the return. The relevant extract is shown below. Table 1 Landholdings of the landowners in 1873 Name Acreage Name Acreage Arundell 6037 Grove 7178 Beach 7199 Long 13829 Benett 4357 Lopes 1469 Bolingbroke 3382 Malet 2828 Burdett 3958 Malmesbury — 1079 Calley 3444 Methuen 5542 Churchill 3760 Northey 2614 Cleveland 2377 Penruddocke 3201 Codrington 2191 Phipps 3925 Freke 3911 Seymour 6398 Goddard 4863 Somerset 5824 Grosvenor 5824 Wyndham 5734 Total acreage 110,924 Source: Return of Owners of Land Wilts 1872-3, (London 1874), Ixxii (2), Wilts.,pp.1-29. HERALDRY AND HOSTELRIES: THE EVIDENCE FROM WILTSHIRE 227 Conclusion In 1877, thirty-two families lent their names to the forty inns that have been considered. Many of these families occupied substantial houses on their estates but as a result of an almost continuous process of sales, however, only Hannington Hall (the Frekes), Corsham Court (the Methuens) and Maiden Bradley House (the Dukes of Somerset) continue in the occupation of members of the families who were in possession in 1877. In a few cases (e.g. the Calleys and the Benetts) descendants of the original families continue to own land in their respective neighbourhoods but the general picture is of an almost total disappearance of the landed proprietors mentioned by Smith - save for the descendants of the major landed aristocracy. To what should be attributed this decline? In the case of the Beckford and Arundell families it was simply bad management and extravagance which resulted in overwhelming debts that could not be sustained. The financial profligacy of William Beckford is well known. In the 18th century the estates of the Arundells were of such a size that the family would have been considered one of the great landowning families of the kingdom but by the 1870s their land holdings had been reduced to a little over six thousand acres with few, if any, other resources and the palatial Wardour Castle to maintain. The underlying cause of the decline of the majority of the landowning gentry in the 19" century, however, was the agricultural depression that began in the early 1870s and continued until almost the end of the century. A dramatic reduction _ in the price of wheat, barley and oats, caused by the market being flooded with cheap imports from North America and a similar reduction in the price of meat as a result of imports from Australasia, the United States and Argentina had a disastrous effect on farmers’ incomes. As a result, many were unable to pay their rents and in some cases farms were abandoned. When, in 1882, the trustees of Sir Richard Colt Hoare’s will sought a private Act of Parliament to enable them to sell heirlooms to raise funds to make improvements to the property settled by the will, it was said on their behalf that the rental income of the settled estates had fallen by £2,700 per annum since 1880 and that the rents so reduced were in many cases either not paid punctually or not in full.”*In order to prevent a total loss of income, landlords had no alternative but to reduce rents. For instance, Sir Michael Hicks Beach was compelled to reduce the rent from his Wiltshire estate by 40 per cent between 1873 and 1898” and the Rev. Edward Northey on his more modest estate by 52 per cent between 1873 and 1894”.A further consequence of such reductions was a dramatic slump in the value of agricultural land thus adversely effecting the ability of landowners to borrow money on its security. A reduction in the value of land could be expected to have led to a ready supply of purchasers. However, this was not by any means the case as by the second half of the 19" century the political and social advantages of owning a country estate had largely disappeared. Further general pessimism about the future prosperity of agriculture persuaded many capitalists to invest their money elsewhere.” Another contributory factor to the declining fortunes of the county’s smaller landed proprietors of the county was the higher income and capital taxes introduced during and after the First World War, coupled with another agricultural depression in the 1920s. Unlike most of the great landowning aristocracy, the majority of the resources of the old gentry families were tied up in their landed estates, most of which would have been heavily encumbered. The only solution was the sale of their estates. The Grove family of Ferne provides an excellent example of the difficulties faced by many of the families who have been considered. This family was established on its estate in 1563 and possessed over 7,000 acres in the county in the 1870s but by 1894, however, was in such a parlous financial state that Sir Thomas Grove’s bank stopped payment of his son’s quarterly allowance and Equitable Life threatened to foreclose on the mortgage of his Winterslow estate if overdue interest was not paid by return of post!”° In order to raise much needed cash, in 1892 he sold his Cornish estate and in 1896 two family portraits by Romney” having in the previous year been compelled to agree to reductions in several of his tenants’ rent. Sir Thomas Grove died in the following year intestate and his son, as administrator of his estate, put Ferne on the market. No purchaser could immediately be found and so the house and the sporting rights were let’’ until the house and estate were eventually sold in 1902. In 1910 Walter Long decided to sell most of his land in the county and the reason he gave encapsulates the feelings that must have been in the minds of many of those landowners who had not already disposed of their estates. He reached his decision, he said, because the policy of the Government towards large landowners: compels all of us who are interested in land 228 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE to most carefully consider our position. We who are owners have done our best to act as if in partnership with our tenants and have not been governed by purely mercenary considerations. A change, however, is coming over the scene, and those of us who do not possess other sources of income must regulate our affairs accordingly.” By this he meant, of course, disposing of property which, in the 16" century would have been an absolute necessity for any aspiring gentleman to possess but by the end of the 19" and the beginning of the 20" century was a burden to be disposed of, if at all possible. With the sole exception of the Hussey-Frekes who repurchased Hannington Hall in 1957, therefore, not one of the gentry families whose arms would have been displayed outside the forty inns in 1877 are today in possession of their ancestral homes although, as has been shown, in a few cases some land remains in their ownership. Such of their coats of arms, therefore, which continue to decorate the sign boards of the inns of the county remain as a melancholy memorial of the bygone greatness of more than a few families who, in their heyday, would have been well known, by reputation at least, to every man, woman and child in their respective localities but who are now almost entirely forgotten. Notes and References ' A.C.Smith, ‘Some Account of the Tavern Signs of Wiltshire and their Origin’, Wiltshire Archaeological and Natural History Society Magazine, Vol. 17 (1878), p.306. For a complete descriptions of the arms, see Sir B.Burke, A Genealogical & Heraldic Dictionary of the Landed Gentry of Great Britain & Ireland (2 vols.1863), Sir B.Burke, A Genealogical & Heraldic Dictionary of the Peerage & Baronetage (1859), Burke’s Landed Gentry (1858), Fairbairn’s Crests of the Families of Great Britain & Ireland (1986) and A.C. Fox-Davies, Armorial Families (6 edn.1910). 2 R.Kelly (ed.), The Post Office Directory of Wiltshire (London 1875) pp.726-730. The Coot’s Arms in Martin was one of the inns included in the directory. However, as Martin was transferred from Wiltshire to Hampshire at the end of the 19" century, it will not be treated here as a Wiltshire inn. The Victoria County History of Wiltshire {hereafter V.C.H.Wilts.] Vol. 13, ed. D.A.Crowley (1987), p.94. 4 Tbid., p.95. > Ibid., p.95. 6 — [bid., p.208. Ibid., p.221. Ibid., Vol. 8, ed. R.B.Pugh (1965), p.252. Ibid. Tbid., Vol.13, p.160. Ibid., p.161. Ibid. p.211. VC.H. Wilts., Vol.13, p.212. Ibid., p.71. Wiltshire and Swindon Record Office [hereafter W.R.O.], 413/361, codicil to the will of Thomas Benet (sic) 1797. VC.H. Wilts., Vol.13, p.74 Ibid., p. 70. Ibid., pp. 70-71. TIbid., p. 211. The author is indebted to Mr. Rex Sawyer for bringing to his notice this report which apparently appeared in the Salisbury and Winchester nl. VC.H.Wilts., Vol. 9, ed. R.B.Pugh (1970), p.79. Ibid., p.85. and WANHS Library, Devizes S.C. 19b. no.18, sale particulars. For an account of the decline of the St.John family see Brian Carne, The Decline and Fall of the St.Fohns of Lydiard Tregoze (Friends of Lydiard Tregoze Report No 34), pp. 34-55. B. Croucher, The Village in the Valley-A History of Ramsbury (Ramsbury 1986) p. 121. Croucher, op.cit., p.157. VC.H.Wilts., Vol.9, p.10. ex. info. Ms. J.L.Mactaggart. This is one of the inns that changed its location since 1875 (another being The Malet Arms at Newton Tony) VC.H.Wilts., Vol.7, ed.R.B.Pugh (1953), p.200. WANHS Library, Devizes, S.C. 1, no. 9, sale particulars. VC.H. Wilts., Vol. 7, p.201. ex. info. Wootton Bassett Historical Society. VC.H.Wilts., Vol. 9, p.191. M.Paterson & E.Ward, Ashton Keynes - A Village with no History (Chirkbank 1986) p.95. Ibid., p.38. Tbid., pp. 39-40. Ibid., p.42. W.R.O. 1002/TH: plan of the land in Nettleton owned by Sir C Bethel Codrington. C.S.Fry, Hannington (Gloucester 1935), p.25. Fry, op.cit., p.47. ex.info. Mrs.A.F Hussey-Freke. VC.H. Wilts., Vol. 9, p.28. Ibid. WANHS Library, Devizes S.C. 29b, no. 17, sale particulars. Ibid. S.C. 50, no. 27, sale particulars. VC.H. Wilts., Vol. 9, p.120. Ibid. J.Pigot, Commercial Directory, (1830), p.802. James Everard Baron Arundel and Sir Richard Colt Hoare, The History of Modern Wilts Hundred of Dunworth HERALDRY AND HOSTELRIES: THE EVIDENCE FROM WILTSHIRE 229 64 68 and Vale of Noddre (6 vols. 1829), Vol.4, p.24. J.Pigot, Commercial Directory, 1842. VC.H.Wilts., Vol.11, ed.D.A.Crowley (1980), pp.99- 100. Ibid., p.100. WANHS Library, Devizes, S.C. 33, no. 70, sale particulars. VC.H. Wilts., Vol. 11, p.100. The Grove family crest is a talbot or collared dog. VC.H. Wilts., Vol.13, p.22. Ibid. Ibid. Ibid., p.23. Ibid.,Vol.7, pp.22.-23. Tbid.,Vol. 8, p.202. Ibid., p.221. Ibid., p.163. WANHS Library, Devizes S.C. 16, no. 14, sale particulars. The house and park failed to sell at the auction and only 31 of the 76 lots found purchasers. WANHS Library, Devizes, S.C. 19, no. 33, sale particulars. VC.H. Wilts., Vol.15, p.147. Ibid. ,Vol. 6, ed. E.Crittall (1962), p.186. Lord Methuen, Corsham Court, (Corsham 1993, reprinted with additions 1999), p.26. D.Ibberson, Ziwo Wiltshire Villages (Marlborough 1963), p. 60. VC.H. Wilts., Vol. 10, ed. E.Crittall (1975), p.130. G.J.Kidston, A History of the Manor of Hazelbury (London 1936), p.245. WANHS Library, Devizes, S.C. 28g, no. 35, sale particulars. Kidston, ibid. Country Life, 13 Aug.1910, p.230. WANHS Library, Devizes, S.C. 29a, no. 1, sale particulars. 80 VC.H.Wilts., Vol.14, ed. C.R.Elrington (1991), p.68. Ibid., Vol.8, p.155. Ibid., p.156. Ibid., p.155. WANHS Library, Devizes, Wiltshire Cuttings Vol. 23, p.381, sale advertisement. (Gillman’s) Devizes and District Almanack and Directory for 1921 (Devizes 1921), p.131. VC.H.Wilts., Vol.10, p.35. Ibid., Vol.11, p.83. Ibid., p.86. Ibid. H.Kitching, A History of Maiden Bradley (Maiden Bradley 1997), p.19. WANHS Library, Devizes S. C. 46, nos. 19 & 20, sale catalogues. VC.H.Wilts., Vol. 8, p.205. WANHS Library, Wiltshire Cuttings Vol.11.18 & S. C. 18, no.2, sale particulars. VC.H.Wilts., Vol.14, p.224. Ibid., Vol.8, p.31. Ibid., p.28. Salisbury and Winchester Journal, 10 Aug. 1882. V.C.H.Wilts., Vol.4, ed. Elizabeth Crittall (1959), p.105. Ibid., p.106. For a full account of the decline of the English landowning families see EM.L.Thompson, English Landed Society in the Nineteenth Century (London 1963), pp.238-327. Desmond Hawkins, The Grove Diaries: The Rise and Fall of an English Family 1809-1925 (Wimborne 1995), p.260. Ibid., p.263. Ibid., p.324. Thompson, op.cit., p.322. Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 230-235 Thomas Stephens (1549-1619): Wiltshireman and Jesuit by Daphne Tighe! and fames H. Thomas? This paper examines the life and activities of Thomas Stephens (1549-1619). A Wiltshireman by birth, he entered the Society of Fesus and became the first Englishman to sail round the Cape of Good Hope to India. There he performed in a Sterling fashion, bringing ecclesiastical and allied works to the native population of the Salcete peninsula, to the south of Goa. The rural parish of Clyffe Pypard, four miles to the south of Wootton Bassett, seems an unlikely birthplace for someone who was reputed to be the first Englishman to travel to India round the Cape of Good Hope. Even today, despite being caught between the sprawl of Swindon and the noise of RAF Lyneham, Clyffe Pypard has an air of remoteness about it. In the sixteenth century it would have been very isolated and given over to agriculture. Thomas Stephens was the son of Thomas, a prosperous merchant, and Jane his wife. He was born at Bushton in the parish of Clyffe Pypard in 1549, where the Stephens family was reasonably important. In 1533 the monastery of Saint Swithin, Winchester granted a lease of Bushton Manor to Richard Stephens, Thomas, his son and Richard, eldest son of Thomas, in which the Stephens bound themselves to keep everything in repair and to live continuously in the house. There was a chapel attached to the Manor, complete with plate and vestments, while there was also a requirement that they should provide the Steward of the Monastery with meat and drink, ‘good beds stuffed with wool’ and provender for horses on two days each year. In 1549, after the Monastic Dissolution, Receiver of the Crown Sir John Thynne surveyed the Manor in which Richard and Thomas Stephens lived, consisting of House and Chapel, named fields and common grazing for 200 sheep. Thomas Stephens (senior) was assessed in the Lay Subsidy Rolls of 1571 on goods valued at £15 on which he paid 25s. It would appear that by 1576 he had left Bushton and was living in London where he was still residing in 1581.! Thomas Stephens’ older brother, Richard, went to Winchester College appearing in the register of scholars. Aged 13, he entered in 1553 as Richard Stephens of Bushton. He then studied at New College, Oxford from 1559 and served as Secretary to both John Jewell (1522-1571), Bishop of Salisbury and Archbishop Matthew Parker (1504-1575) of Canterbury. Later, in 1573, Richard Stephens crossed the Channel and entered Douai Seminary, was ordained and became a Professor of Theology. He died in 1586, aged 46.’ Thomas Stephens of Bushton entered Winchester College in 1564 and subsequently attached himself to Thomas Pounde, a sometime Elizabethan courtier.’ After his fall from grace, however, Pounde withdrew from Court in 1569, vowing to become a priest and enter the Society of Jesus. The Society, with its high intellectual standards and complete Catholic commitment, appealed to countless able men in the years after 1550. For two years, between 1571 and 1573, Thomas Pounde and Thomas Stephens, in the guise of his servant, travelled the country in an effort to persuade Catholic gentry to preserve the old faith. This was a risky pursuit and they were ‘address? *School of Social, Historical and Literary Studies, University of Portsmouth, Milldam, Burnaby Road, Portsmouth PO1 3AS THOMAS STEPHENS (1549-1619): WILTSHIREMAN AND JESUIT 231 both imprisoned in Ludlow for a short time. The experience convinced them that it was time to leave the country and offer themselves to the Society of Jesus. However, before they could escape, Pounde was thrown into prison by Edwin Sandys, the Bishop of London, leaving Stephens to make his way alone to Rome. On 20 October 1575, at the age of 26, Thomas Stephens was received as a novice at Sant’ Andrea, Father Fabio de Fabiis acting as his novice-master.* Stephens was one of several new admissions to the Society of Jesus, as Father Robert Persons S.J. subsequently recalled: And soe upon ye yeare of ye Jubily wch was 1575 mett and entered at Rome diverse at one time, as by name E Persons, f. Henry Garnett, f. Wm Weston, f. W. Holt, all Oxford men and afterwards employed in ye Mission of England. FE Ths Stephens also, now a painfull and fruitfull labourer in ye east Indies and mr. John Lane mr of art of Corpus Xti College in Oxford and of great expectation, that dyed afterwards at Alcala INYO PAIN sence Pounde asked Stephens to win him the privilege of admission to the Society and this was granted on a second petition dated 4 November 1578, while Pounde was a prisoner in the Tower of London and Stephens was studying philosophy at the Roman College. Pounde was subsequently released and went to Rome. In September 1580 he risked life and limb returning to England, sending letters to Mr. Tripps, to the Privy Council and to Sir Christopher Hatton, Elizabeth I’s godson. Endorsed ‘Certain Papistical reasons set down for the withdrawing of men to come to the church’, they were forwarded to London by the Sheriff of Wiltshire for urgent consideration.® Stephens meanwhile also studied physics and mathematics under Christopher Clavius (1538-1612), the Bavarian-born Jesuit who has been described as the ‘Euclid of the Sixteenth Century’.’ At this time Clavius was serving on Pope Gregory XIIV’s commission to revise the calendar. The revision was accepted by Italy, Spain and Portugal in 1582 but England saw it as a Popish plot and did not accept the Gregorian calendar until 1752. In 1579 Stephens was given leave to join the mission of Portuguese East India. He had already created a sound impression upon some of his contemporaries, occasioning Father Robert Persons, S.J. to observe in March of that year: Missions since your departure hath byn dyvers of importance namely two or three to the East Indyes, wherehence, and especially from Japonia, we have had letters of great consolation which are newly printed. And amongst these Missions was one English-man, that is Thomas Stevens a yong man of great fervour and a reasonable talent, whome I would have diverted towards a certayne North India, but the lett you shall know after.” Stephens reached Lisbon at the end of March 1579 and became one of the twelve Jesuit missionaries who set sail on 4 April in a fleet of five ships. He travelled, with five other Jesuits, in the Sado Lourenco and his experiences on the voyage are recorded in a long letter to his father which was published in Richard Hakluyt’s The English Voyages. As it is the first account in English of a voyage round the Cape of Good Hope, it is worth reproducing in Thomas Stephens’ own words:” After most humble commendations: These shall be to crave your dayly blessing, with like commendations unto my mother; and withall, to certifie you of my being: according to your will and my duety. I wrote unto you taking my journey from Italy to Portugall, which letters I thinke are come to your hands, so that presuming thereupon, I thinke I have the lesse need at this ume to tell you the cause of my departing, which neverthelesse in one word I may conclude, if I do but name obedience. I came to Lisbon toward the end of March, eight dayes before the departure of the shippes, so late that if they had not bene stayed about some weighty matters, they had bene long gone before our comming: insomuch that there were others ordained to goe in our places, that the kings provision and ours also might not be in vaine. Nevertheless our sudden comming tooke place, and the fourth of April five ships departed for Goa, wherein besides shipmen and souldiers, there were a great number of children'’ which in the seas beare out better than men, and no marvell, when that many women also passe very well. The setting foorth from the port I need not to tell how solemne it is with trumpets, and shooting of ordinance, you may easily imagine it, considering that they go in the maner of warre. The tenth of the foresayd moneth we came to the sight of Porto Santo neere unto Madera, where an English shippe set upon ours (which was then also alone) with a few shots, that did no harme, but after that our ship had layed out her greatest ordinance, they straight departed as they came. The English shippe was very faire and great, which I was sorry to see so ill occupied, for she went roving about, so that we saw her againe at the Canarian Iles, unto which we came the thirteenth of the sayd monceth, and good leisure we had to woonder at the high mountaine of the Iland Tenerif, for we wandred betweene that 232 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE and great Canaria foure dayes by reason of contrary windes; and briefly, such evill weather we had untill the foureteenth of May, that they despaired to compasse the Cape of Good hope that yeere. Neverthelesse, taking our voyage betweene Guinea and the Ilands of Capo Verde, without seeing of any land at all, we arrived at length unto the coast of Guinie, which the Portugals so call, chiefly that part of the burning Zone, which is from the sixth degree unto the Equinoctiall, ..... In this coast .. we spent no lesse than thirty dayes, partly with contrary windes, partly with calme. The thirtieth of May we passed the Equinoctiall with contentation, directing our course aswell as we could to passe the promontory, but in all that gulfe, ..... we found so often calmes that the expertest mariners wondred at it. .... You shall understand, that being passed the line, they cannot straightway go the next way to the promontory: but according to the winde, they draw alwayes as neere South as they can to put themselves in the latitude of the point, which is 35 degrees and an halfe, and then they take their course towards the East and so compasse the point. But the winde served us so, that at 33 degrees we did direct our course toward the promontory of Good hope. You know that it is hard to saile from East to West, or contrary, because there is no fixed point in all the skie, whereby they may direct their course, wherefore I shall tell you what helps God provided for these men. .... the neerer we came to the people of Afrike, the more strange kindes of fowles appeared, insomuch that when we came within no lesse than thirty leagues (almost an hundred miles) and sixe hundred miles we thought from any Iland, as good as three thousand fowles of sundry kindes followed our ship; some of them so great that their wings being opened from one point to the other, contained seven spannes as the Mariners sayd. A marvellous thing to see how God provided, so that in so wide a sea these fowles are all fat and nothing wanteth them. .... And to speake somewhat of fishes in all places of calme, ... there waited on our ship fishes as long as aman which they call Tuberones, they come to eat such things as from the shippe fall into the sea, not refusing men themselves if they light upon them. ... The Mariners in time past have eaten of them, but since they have seene them eate men their stomacks abhorre them. ... There is another kind of fish as bigge almost as a herring, which hath wings and flieth and they are together in great number.... With these and like sights, but alwayes making our supplications to God for good weather and salvation of the ship, we came at length unto the point, so famous and feared of all men: but we found there no tempest only great waves, where our Pilot was a little overseene: ... he thinking himselfe to have winde at will, shot so nigh the land that the winde turning into the South, and the waves being exceeding great, rolled us so nere the land, that the ship stood in lesse than 14 fathoms of water, no more than sixe miles from the Cape, which is called Das Agulias, and there we stood as utterly cast away, for under us were rocks of maine stone so sharpe, and cutting, that no ancre could hold the ship, the shore so evill that nothing could take land, and the land itselfe so full of Tigers, and people that are savage and killers of all strangers, that we had no hope of life nor comfort, but onely in God and a good conscience. Notwithstanding, after we had lost ancres, hoising up the sailes for to get the ship a coast in some safer place, or when it should please God, it pleased his mercy suddenly, where no man looked for helpe, to fill our sailes with wind from the land and so we escaped, thanks be to God. And the day following ... we also fell a fishing, and so many they tooke that they served all the ship for that day, and part of the next. And one of them pulled up a corall of great bignesse and price. For there they say (as we saw by experience) that the corals doe grow in the manner of stalks upon the rocks in the bottome and waxe hard and red. The day of perill was the nine and twentieth of July. And you shall understand that, the Cape passed, there be two wayes to India: one within the Ile of S. Laurence, which they take willingly, because they refresh themselves at Mosambique, a fortnight or a moneth, not without great need, and thence in a moneth more land in Goa. The other is without the Ile of S. Laurence, which they take when they set foorth so late, and come so late to the point, that they have no time to take the foresayd Mosambique, and then they goe heavily, because in this way they take no port. And by reason of the long navigation and want of food and water, they fall into sundry diseases, their gummes waxe great, and swell, and they are faine to cut them away, their legges swell and all the body becommeth sore, and so benummed, that they can not stirre hand nor foot, and so they die for weaknesse, others fall into fluxes and agues, and die thereby. And this way it was our chance to make, yet though we had more than one hundred and fifty sicke, there died not past seven and twenty; which losses they esteemed not much in respect of other times. Though some of ours were diseased in this sort, yet, thanks be to God, I had my health all the way, contrary to the expectation of many, God send me my health so well in the land, if it may be to his honour and service.... After, about the eleventh degree, the space of many dayes, more than ten thousand fishes by estimation followed round about our ship, whereof we caught so many, that for fifteene dayes we did eate nothing els, and they served our turne very well: for at this time THOMAS STEPHENS (1549-1619): WILTSHIREMAN AND JESUIT 233 we had neither meat nor almost any thing els to eate, our navigation growing so long that it drew neere to seven months, where as commonly they goe it in five, 1 meane when they saile the inner way. But these fishes were not signe of land, but rather of deepe sea..... And we that thought that we had bene neere India, were in the same latitude neere Zocotoro,'! an Ile in the mouth of the Red sea. But there God sent us great winds from the Northeast or Northnortheast, whereupon unwillingly they bare up toward the East, and thus we went tenne dayes without seeing signe of land, whereby they perceived their errour, for they had directed their course before alwayes Northeast, coveting to multiply degrees of latitude, but partly the difference of the Needle, and most of all the running seas, which at that time ran Northwest, had drawen us to this other danger, had not God sent us this winde, which at length waxed larger, and restored us to our right course....The first signe of land were certaine fowles which they knew to be of India; the second, boughes of palmes and sedges: the third, snakes swimming on the water, and a substance which they call by the name of the coine of money, as broad and as round as a groat, woonderfully printed and stamped of nature, like unto some coine. And these two last signes be so certaine, that the next day after, if the winde serve, they see land, which we did to our great joy, when all our water (for you know they make no beere in those parts) and victuals began to faile us. And to Goa we came the foure and twentieth day of October, there being received with passing great charity. The people be tawny, but not disfigured in their lips and noses, as the Moores and Cafres of Ethiopia. They that be not of reputation, or at least the most part, goe naked, saving an apron ofa span long, and as much in bredth before them, and a lace two fingers broad before them girded about with a string and no more: and thus thinke them as well as we with all our trimming. Of the fruits and trees that be here I cannot now speake, for I should make another letter as long as this. For hitherto I have not seene a tree here, whose like I have seene in Europe, the vine excepted, which neverthelesse here is to no good purpose, so that all the wines are brought out of Portugall. The drinke of this countrey is good water, or wine of the Palme tree, or of a fruit call Cocos. And this shall suffice for this ime. If God send me my health, I shall have opportunity to write to you once againe. Now the length of my letter compelleth me to take my leave: and thus I wish your most prosperous health. From Goa the tenth of November, 1579. Your loving sonne Thomas Stevens. This letter brings home vividly the hazards of navigation before it was possible to calculate longitude accurately. It also highlights the problems of provisioning ships for an unknown length of journey and, although only ‘seven and twenty’ died from what was presumably scurvy, it must have meant that there was more food available for the survivors. After his long, dangerous and eventful voyage, Stephens found himself at Goa, Portugal’s Asian capital, a large and wealthy city replete with harbour and ship building facilities.'? Having survived the voyage and written his filial letter Stephens had done much, probably without realising it, to reveal pertinent maritime information whilst also alluding to Portuguese frailties, thereby indirectly aiding that country’s potential enemies."’ Within six months of his arrival Padre Estevam, as he was styled, was ordained and sent to Salcete, south of Goa, which had a population of 80,000 Indians spread over 55 villages, with numerous temples devoted to the Hindu cobra-goddess Santéry Jesuit missionaries had been operating there since 1560. When Stephens arrived the mission had six churches and 8,000 Christians, but the missionaries were far from safe and in 1583 four Fathers and many native Christians were massacred. Stephens recovered the bodies and buried them in his church." Further details of Stephens can be gleaned from the correspondence of English merchant John Newbery. Leaving England in February 1583, Newbery and three companions had travelled overland from Aleppo to Babylon and then, early in September, been imprisoned in Ormuz,'° accused of being Spanish spies. Newbery was then sent with the other three, Ralph Fitch, William Betts/Leeds and James Story, to Goa where they were imprisoned for a month. Newbery’s and Fitch’s letters refer to Stephens. As Newbery remarked: had it not pleased God to put into the minds of the archbishop and other two Padres or Jesuits of S. Pauls colledge to stand our friends, we might have rotted in prison.... adding that one was Bruges-born Father Mark and ‘the other was borne in Wiltshire in England, and is called Padre Thomas Stephens’. Fathers Mark and Stephens interceded for them: These did sue for us unto the Viceroy and other officers, and stood us in as much stead, as our lives and goods were woorth: for if they had not stucke to us, if we had escaped with our lives, yet we had had long imprisonment The Portuguese authorities offered freedom after 234 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 14 days’ imprisonment at a price — 2,000 ducats. The priests ‘found sureties for us, that we should not depart the country without the licence of the Viceroy’.!° Story, a painter, entered the Society of Jesus but soon left; the other three broke bail on Whitsunday 1584 and fled. Only Fitch survived to return to England seven years later after an adventurous journey.!’ Some years later, in 1609, Stephens did a similar good turn to four Englishmen, a Dutchman and three Frenchmen as the Viceroy had imprisoned all foreigners who had come to India on non-Portuguese ships. One Frenchman, the famous traveller Pyrard de Laval, wrote glowingly of the assistance that they had received from the Jesuit Fathers, including Thomas Stephens, adding that ‘They could not have done more for us had we been their own brothers’.'® Three of the Englishmen he helped may be identified as Edward Beck, William Hudson and Thomas Davis who arrived in Lisbon en route for England in the summer of 1611. They had arrived in a carrack, having been ‘cast away in the East Indies’ aboard the Hector and Ascension. Also aboard the vessel was the Archbishop of Braga, ‘who was lately Viceroy’. Beck, a Jesuit, ‘can say something to the purpose if he be dealt withall accordingly’ and ‘gave out he was recommended by Thos. Stevens, an English Jesuit in Goa, to Henry Fludd, the Jesuit at Lisbon’.”” Helping fellow countrymen in Goan distress was, however, only part of Stephens’ duties. For four decades he laboured to bring the Catholic faith to native Goans in the Salcete mission. Early in February 1589 he was appointed spiritual coadjutor in Goa, and in 1596 was made Rector of the College there.”” He also served as parish priest in a number of communities.*! Blessed with a robust constitution, a keen mind and a caring nature, Stephens worked tirelessly on behalf of his missionary charges. He became involved in the running of the properties in his jurisdiction and was not above, at times, misappropriating revenue to his houses which should have gone elsewhere. He also took an interest in cartography, sending a map of Goa and the Salcete peninsula to Europe in the hope that there would be a ‘better understanding of this region’. It was, however, to be his work in native languages for which Stephens was to be chiefly remembered. In this respect, he followed a strong and strict Jesuit characteristic — conversion via mastery of native tongues, an approach that had served the Order well with, for example, the Guarani Indians of South America. Stephens mastered the native Konkani tongue,” composing a grammar in that language for use by his fellow Portuguese missionaries and a catechism. The work for which Stephens is famous, however, is his Purdna in the Marathi language, nearest of linguistic kin to Konkani. Hindu Puranas were sung epic poems and Stephens realised the potential for spreading the Gospel by these means, as well as satisfying a desire of his new converts. Accordingly, he composed a poem in two parts, the first based on the Old Testament from the story of the Creation to the prophecies of Christ’s coming, while the second recounted Christ’s life. The stories were carefully selected and the narrative was full of colour and included questions by a Brahmin with answers from the Padre Guru. The work was hailed as a masterpiece, readings from it taking place on feast days and after Mass.” While Stephens wrote the Purana in the Marathi language, he was forced to transcribe it to print the work in Roman characters. Completed in 1614, the first edition of the Purana was published in Latin type at Rachol in 1616 during his lifetime, further editions following in 1649 and 1654. An edition in Roman characters was published in Mangalore in 1907. It was Thomas Stephens’ ambition to print the work in Marathi but the technical problems defeated him. However, his work was remembered; in 1801 a visitor to the villages noted that prayer was combined with musical recitals from the Purana. By 1873 there were no longer any printed copies extant but it was still being read in churches. In 1926 it was still being chanted in the churches and in the villagers’ fields and homes.” In 1619 Francis Xavier, founder of the Society of Jesus, who had died in 1552, was beatified in Rome. In the same year, Thomas Stephens died in Goa and was buried beside him. Humble, strong, driven by total commitment, faith and conviction, Stephens was outstanding for his missionary and linguistic work. For four decades he promoted the Catholic faith against the rigours and climes of a very different environment from the Wiltshire one in which he had been raised. His brethren remembered him as an example to all and a letter to Philip III of Spain spoke of him as ‘an English priest of outstanding holiness’. With very good reason he has been described by one scholar as ‘the father of Christian Literature in India’.”” Thus he should be remembered in both Goa and Wiltshire as a very remarkable man. THOMAS STEPHENS (1549-1619); WILTSHIREMAN AND JESUIT 235 APPENDIX PRINCIPAL LITERARY WORKS BY THOMAS STEPHENS (1549-1619). Purana Christao (1616) — sung epic poem linked to the Old Testament and Christ’s Life. Further editions 1649, 1654, 1907, 1956, 1965, 1996.?’ Doutrina Cristam (1622) — first printed book in Konkani. Catechism of the Christian Doctrine. Facsimile edition, Lisbon, 1945. Arte da Lingoa Canarim (1640) — first printed grammar of Konkani — in Portuguese. Second edition 1857. References 1 nN 10 Canon F. Goddard, ‘Notes on Clyffe Pypard and Broad Town’, W(iltshire) A(rchaeological) (and) N(atural) H(istory) M(agazine) XLIV (1929), pp. 148- 50; H. Chitty, “Thomas Stevens, Primus in Indis’, WA.N.H.M., XXXII (1875), pp. 221, 222-223. Thynne (d. 1580), built Longleat House, probably from his own plans, between 1567 and 1579, and is noted in Oxford D.N.B. The move to London by Thomas Stephens (senior) may have been explained by his wife Jane’s death in 1575: Oxford D.N.B., Vol.52, p. 482. T.F. Kirby, Winchester Scholars, (1888), p.132; G. Schurhammer, ‘Thomas Stephens 1549-1619’, The Month (April 1955), p. 198. P. McGrath, ‘Winchester College and the Old Religion in the Sixteenth Century’, in R. Custance (ed.), Winchester College — Sixth-centenary Essays, (O.U.P. 1982), p.262 and n. Pounde used a number of aliases, including the names Gallop and Wallop. Schurhammer, p. 200. Revd. J.H. Pollen (ed.), The Memoirs of Father Robert Persons, (Catholic Record Society Miscellanea), II (1906), pp. 191-192. Calendar of State Papers Domestic 1547-80, p. 676. D. Steel, Marking Time: the Epic Quest to Invent the Perfect Calendar, (2000), p. 163. Father Persons to Father William Good, after 19 March 1579: L. Hicks (ed.), Letters and Memorials of Father Robert Persons, S.J., 1 (Catholic Record Society 1942), pp. 5-6. Quondam Canon of Wells in Somerset, Good had left England at Elizabeth I’s accession and joined the Society of Tournai early in June 1562. Subsequently he visited Ireland, Rome, Sweden and Poland. He died at Naples on 5 July 1586: Ibid., p.5. R. Hakluyt, The English Voyages, (10 vols. 1927-1928), IV, pp. 233-240. On the basis of this allusion it is possible that Stephens and his associates sailed with the ‘Orphans of the King’, marriageable batches of orphan girls sent out at Crown expense, a Portuguese practice that started in 1545: C.R. Boxer, Women in Iberian Expansion Overseas, 16 19 1415-1815, (1975), p. 66. Rocky, virtually waterless and located off Cape Guardafui, the island of Socotra was taken by the Portuguese in 1507. Intended as a base for attacks into the Red Sea and for intercepting spice cargoes bound for Jedda and Suez, it was held for a short while only. Subject to the rule of the Mahri Sultan of Qishn in Southern Arabia from c. 1480, Socotra was to become noted for its production of aloes, in demand as a purgative: J.H. Parry, The Age of Reconnaissance: Discovery, Exploration and Settlement 1450-1650 (1973 edn.), p. 184; J.H. Arrowsmith-Brown (ed.) Prutky’s Travels in Ethiopia and Other Countries (Hakluyt Society, 1991), pp. 386-7. Parry, p.184. ‘Golden Goa’, with poor soil, water- related problems and the spread of malaria, was held by the Portuguese between 1510 and 1760 when it was abandoned in favour of the nearby community of Panjim: M. Shennan, The European Dynamic: Aspects of European Expansion, 1450-1715 (1976), p. 78; C.R. Boxer, The Portuguese Seaborne Empire 1415-1825 (1973 edn.), pp. 133-4. R. Hall, Empires of the Monsoon: A History of the Indian Ocean and its Invaders (1996), pp.284-S. Schurhammer, pp. 201-3 for detailed background and coverage. Located at the mouth of the Persian Gulf, the Arabian entrepot port of Ormuz was captured and fortified with a Portuguese garrison in 1507: Shennan, p. 78. Goan citizen André de Tavora stood surety for them: Schurhammer, p. 203. Hakluyt, III, pp. 271-81, 327. Fitch’s journey took him to Burma and Siam. Schurhammer, p. 205. Hugh Lee to Earl of Salisbury ; Lee to Thomas Wilson, 26 June/ 6 July 1611: Calendar of State Papers Colonial, East Indies, China and Japan 1513-1616, p. 225. Braga is located to the north-east of Oporto. The carrack, or nao, was a heavily built but lightly gunned merchant ship, characterised by large, well-developed castes fore and aft: Boxer, Portuguese Seaborne Empire, p. 209. T.M. McCoog, S.J., English and Welsh Jesuits 1555- 1650, Pt. II, (Catholic Record Society 1995), p. 304. A coadjutor was an assistant to a Bishop. The communities were Bassein, Benaulim, Loutolim, Margao, Mormugao, Navelim and Rachol: Oxford D.N.B., Vol. 52, p. 483. Konkani is written in five different scripts — Roman, Kannada, Malayalam, Perso-Arabic and Devanagari, the first having the oldest literary tradition from the sixteenth century, many words of the basic vocabulary being derived from Portuguese: http://india-seminar com. Oxford D.N.B., Vol. 52, p. 483 Schurhammer, p. 209. Schurhammer, p. 207 Manohar Sar Dessai, A History of Konkani Literature (New Delhi, 2000), p. 34, quoted http://india-seminar com. Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 236-245 The Duke of Somerset’s Grand Mansion at The Brails, Great Bedwyn: A Review of the Evidence by Graham Bathe Since its discovery in 1874, a conduit trench in Bedwyn Brail has been interpreted as evidence that the Tudor mansion under construction for Sir Edward Seymour in 1548-9 was to be sited near Hill Barn. Written accounts by those undertaking the research, however, reveal that the house was not where it was anticipated from surviving documents. An examination of all data indicates that the description of the house-site given in contemporary letters, and the view described from it, do not accord comfortably with the accepted location near Hill Barn. Documentary sources and surface features provide a conundrum involving three possible locations, which archaeological survey could resolve. Background and Introduction The Rev Canon J.E. Jackson, when reviewing the extensive Seymour and Thynne archives at Longleat in the 1870s, made the startling discovery that Sir Edward Seymour, Duke of Somerset, had embarked upon construction of a massive mansion and park near Great Bedwyn during 1548-9 (Jackson 1875, 150). Letters and accounts made detailed reference to progress and the workforce employed. No traces of the house were known to survive, and no knowledge of this project appeared to be held by historians or local inhabitants. Edward Seymour had become the most powerful, and only survivor, of three remarkable children raised at Wolfhall, near Burbage, at the start of the 1500s. His brother Thomas, who became Lord Seymour of Sudeley, and had attempted to court both Lady Jane Grey and Princess Elizabeth, married the widow of Henry VIII, Catherine Parr. Their sister Jane Seymour had become third wife of Henry VIII, and died shortly after giving birth to a son, the future Edward VI. Through his will, King Henry had made arrangements for a ruling Council of Regency, comprising sixteen men with equal power, to serve jointly until his son was old enough to rule. However, Edward Seymour, drawing on his blood links to Queen Jane and the boy King, contrived with Secretary of State Sir William Paget to ensure that, contrary to instructions laid down in Henry’s will, he was given supreme power. He would rule as Lord Protector after Henry’s death in January 1547/8, whilst his nephew Edward VI was in his minority (Jordan 1968, 51). He effectively acted as if he were the King. Members of the Council awarded themselves favourable positions, Edward Seymour becoming the Duke of Somerset, whilst Royal wealth, obtained by expropriation of the monastic lands, was redistributed amongst them. In due course, tensions generated from Seymour’s crushing ambitions led his peers to conspire his downfall, and he was executed in January 1551/2 for felony (Jordan 1970, 92). In the late 1540s, Edward Seymour had started building some of the most opulent private mansions in England. He had converted Syon House, Middlesex, a great monastic property granted him by the Crown, and the arrangement of the surviving building with its three stories and four square turrets, and long gallery of 40m, largely date from this era. He also commenced building (what became known as) Somerset House in London. This required the clearance of London landmarks from 2.5ha of land, including the church of St Mary Byeley in Densome, Woodgreen, Fordingbridge, SP6 2QU THE DUKE OF SOMERSET’S GRAND MANSION AT THE BRAILS, GREAT BEDWYN The Duke of Somerset’s Park Boundary and Mansion Site at the Brails Mill aver the feld was dygged within the woode 1 km rowards ut thupper corner of the broyle eee | nedowe grounde from bushells mylne unto the veric backside Perambulation commences here; continues clockwise in boxes |The Park takes in a} pece between croughton and phewarens Crofton % Freewarren (ey Wilton bushell's close he medowe on the other side of the ponde towards grafton leving space rom thens to the ponde hed next to Wilton where the pale standeth ‘or a wey whiche the tenants of Wwillton shall have from theire village 0 theire comon and so on the oderside uppe the ‘lose to the upper ende of the medowe where we beganne 237 trayte thorowe the wood and copice 0 the corner of the woude ence the grete p He which he falling of he hill on the farther heyoude the springs and so forth in svde towards Ramphreis House Castle = Copse Harding . (Ramfrey’s House) ~ > at the corner where the chief spring is where the conduite house shall fande enile of the lane ht from the broile Dykes [ea Geology: Reading Beds =] ABC (CF) Park Boundary details recorded 1548-9 London Clay Woodland (2005) Possible Mansion Locations Fig. 1 Composite Map of The Brails, based on 1792 Benjamin Haynes Map, and Enclosure Award of Wilton, with geology and land features, incorporating details from documents 1548-9, and park perambulation from John Berwick’s letter to Sir John Thynne, 1548 (Longleat House/Thynne Papers/Volume II, f 30) le Strand, an Inn of Chancery, the former London houses of three bishops, and much of the fabric of St John of Jerusalem and cloisters of St Paul’s. The Search for Seymour’s Mansion at The Brails: A Historiography Jackson’s discovery revealed that Seymour had also commissioned a seat at the Brails, together with a park stretching from Bedwyn to Wilton. His local properties at Wolfhall and Tottenham were perhaps not befitting the richest noble in the land, and work commenced in 1548 on the new mansion. Given the scale of his other building projects, it is not surprising that his new property was scheduled to be of massive proportions. Two hundred and eighty men were engaged, out of a projected requirement of four hundred (Longleat House/Thynne Papers Volume II, f.60). No designs of Seymour’s mansion at The Brails survived, although Jackson found contemporary letters giving site details and place-names, and describing a conduit trench 1600 feet long and 15 feet deep (487m x 4.5m), constructed to collect water for the house. Deducing that the trench should be traceable and other features identified, Jackson visited the site in company with the Reverend George Stallard in the spring of 1874, and ‘walked about the hills and fairly identified the outline of the proposed park’ (Jackson 1875, 151). They could not 238 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE locate the trench, but a few days afterwards Rev Stallard revisited, apparently in the company of local historian Maurice Adams, and announced that he had found it in Bedwyn Brail. Being overgrown with brushwood it was not easily accessible but once found it could not be mistaken (Adams 1905). It was measured, and found to be 1,598 feet long. The house was deduced to be at the southern end. The findings were incorporated into Jackson’s substantive paper on Wulfhall and the Seymours, read before the Wiltshire Archaeological and Natural History Society in September 1874, prior to publication in Volume 15 of its journal, where a map interpretation of the park boundaries were given showing the ‘supposed site of Protector Somerset’s House’ (Jackson 1875, 140). Stallard traced no local folklore associated with the site. In May 1874 he wrote to Jackson ‘I find that the old people do not call this Conduit the ‘Wansdyke’ but ‘the moat’ or ‘the ditch” (Wiltshire Archaeological Notes 1932). He also described the old Cow Common, within which the conduit was located, and which appeared to relate to pasture referred to in the letters. He explained ‘There were three gates by which the cattle were driven in, one on the road coming up from Wilton — the mounds on either side of the road remain to this day — another on the road to Bedwyn, and a third at the top of the Common near the end of the Conduit called Tidcombe Pond-gate’ (ibid.). See Figure 1. Stallard was frustrated at being unable to check his interpretations against certain Ailesbury archives. In October 1874 he explained that he had delayed contacting Jackson ‘in the hope of obtaining some information from the papers at Savernake House, but [I] have not been able to get it. I do not despair that some day I may with the help of a friend be able to look over some of these’ (ibid.). It appears that, after reading the 1548-9 letters, Jackson, Stallard and Adams had rather anticipated that the house might be located further west near Wilton Brail. The unexpected discovery of the conduit in Bedwyn Brail raised questions, and residual doubts kept surfacing. Even after Jackson’s paper had been read to the Society, debate ensued, and Stallard opined that ‘Jf the Conduit was in the farther [Bedwyn] Brail, it could have been of no_use [author’s emphasis] if the House was to have been built in the higher [Wilton] Brail, for there is a long dip and rising between the two. I think that the solution of the difficulty will be found in placing the site of the new House at the end of the farther [Bedwyn] Brail near the Conduit House’ (ibid.). By December 1874 he was ‘satisfied that I have found the site of my Lord Grace’s Palace, which Green Drive in Brail Brail @ oo ee Conduit Present Pond Common Low Road at present to Nowhere Supposed Site Land raised to height of hedge. Whole meadow before him and view of Grafton and Wolfhall in distance. Fig. 2 Sketch Map (redrawn) included in letter from George Stallard to Canon Jackson, December 1874 fully answers to the description in the Agent’s letter. Of the water course and conduit there can be no possible doubt [author’s emphasis] and the House must have been in front of the nether [Bedwyn] Brail not as we supposed in the first [Wilton] Brail — How could the water be carried across the people’s common to so high a position as the one Brail is above the other?’ (ibid.). Later that month his resolve was strengthened. He wrote, ‘I am satisfied to the intended situation for the House. You will notice the pond immediately behind it, the meadow is made up to the hedge, you see Grafton and I think Wolfhall from there’ (ibid.). His sketch map of the same month, whilst difficult to relate to modern ground features, reveals further thinking here (Figure 2). Adams may have remained sceptical about Stallard’s conclusion. His slim volume on Wolfhall and Tottenham states pointedly, ‘The Duke had built himself a palatial residence in London on the banks of the Thames — since known as Somerset House — and he now contemplated erecting another on the summit of Dodsdown, on the verge of Bedwyn, or rather Wilton, Brail {emphasis added]. The site selected is not far from the present brick-kilns, and from it a fine and extensive view westwards over the Vale of Pewsey is obtained’ (Adams c.1900). Notwithstanding implied dissent, the identification of the house site with the southern end of the conduit became received wisdom. Although some 30 years after discovery of the conduit trench Adams (1905) had maintained that ‘nothing was done as regards the actual building of the Palace’, some reports countered this. Jordan (1968), in his seminal work on Edward VI, claimed that ‘traces of the foundation remain’. Ailesbury (1962) went further, and stated THE DUKE OF SOMERSET’S GRAND MANSION AT THE BRAILS, GREAT BEDWYN 239 that, at the time of Somerset’s fall from power, ‘the walls of the mansion had already begun to rise’. Geological and Topographical Constraints The location of the house and its water source were dependent on the local geology, which governs land form, agricultural use, the eruptions of springs and the distribution of semi-natural habitats at The Brails. Poorly fertile Eocene deposits of the Reading Beds and London Clay overlie the Upper Chalk, producing two ridges aligned SSW-NNE. These difficult and marginal lands have generally been used for rough grazing (including common pasture), coppice or woodland, interspersed with attempts at arable farming, after which the land has reverted again. Both ridges have been exploited for brick clay and had kilns constructed on them. The soils are complex, and water trickles from several springs at the junction of Eocene deposits even near the hill tops, where the conduit trench was dug. The current woodlands known as Wilton Brail and Bedwyn Brail occupy the Eocene exposures almost exactly (Figure 1). The location is traversed by a number of earthworks, some of which may be derived from deer enclosures, perambulation features of The Brails as an outlier of Savernake Forest, banks connected to the Bedwyn Dyke, or post-medieval field boundaries. The Haynes map (WRO 9/27/59) shows the two ridges and dip between them, bounded by a dyke, as ‘Cow Down’ in 1792. Subsequently, those parts which were not planted with woodland were known as Wilton Common. The extensive common pastures of The Brails may be adapted from earlier deer parks of which the Bedwyn area had many in 1231 (Watts 1996, 88). A Review of Evidence Relating to the Mansion House and Park A. Documentary Evidence The main evidence relating to works at The Brails, describing the mansion house, park land bounded by a great ditch, a lined conduit and proposed conduit- house stems from thirteen letters in the Thynne and Seymour Papers of the Longleat Archives (see References). Seven of these were written by Bryan Teshe, supervisor of works at The Brails, to Sir John Thynne, Seymour’s Steward. The other six are from Sir John Berwick, Seymour’s Receiver General and Ranger of Savernake Forest, to Thynne. No less than seven were written during June 1549, when work was at its height. Parts of these are transcribed in Jackson (1875) [where page 179, line 22 for ‘broke’, read ‘broile’|. There is also a detailed account of monies paid daily to labourers, carters, carpenters, masons and blacksmiths deployed in preparatory work, by John Hutchins, bailiff of Wolfhall, until February 1549. No later accounts have survived. The documents frequently mention Dodsdown; whilst recently the name has been specifically associated with the vicinity of Dodsdown Brickworks in Wilton Brail; this term was formerly applied over an extensive area of upland, stretching some 1.5km from Hill Barn to near Crofton. Finally a Letter Patent of 1552 ordered the disposal of usable building material after the house construction was abandoned. Key extracts from these documents are as follows: Dimensions of Demesne Land Although apparently precise measurements are given in the letters, figures do not tally, nor correlate closely with the interpreted perambulation (Figure 1). ¢ The perimeter of ‘the demaynes about my lords graces house at the broyle ende ...[is] in compas 111 myles saving cx lug’ (Longleat House/Thynne Papers/Volume II, f.30). [Taking an average lug of 16% feet, this gives a perimeter of 2.65 miles (4675 yards). A circle with this circumference would have a diameter of 0.84 miles (1488 yards), and area of 359 acres]. e Area: ‘there ts of ... grounde within the compass S85 acres’ (ibid.). Perambulation A reconstruction of the park boundary, and the extracts from Berwick’s letter of 1548 (Longleat House/Thynne Papers/Volume II, f.30) on which this is based, are given in Figure 1. Because of the profusion of ditches in the area, many of which may have been modified for new purposes, alignments can rarely be identified with certainty. Ramphrie’s house, which is mentioned twice in the letters, relates to Harding Farm, occupied by Ramfrey Simons (WRSO 9/14/223 240 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Hedging and bounds Much of the early effort went into bounding the park, grubbing and carting hedge-plants to the bank where they were set. e¢ 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). e ‘there be hedginge of the bowndes xxiiti men whiche have hedgd bothe sides of the quickfrithe and now within these u days will mak an end’ (Longleat House/Thynne Papers/Volume II, f.60). e ‘.. there is in the woodes cleaving of wood and felling of rowe [deadwood] for the hedginge’ (ibid.). e ‘the diche about the boundes 1s in a great forwardness and wilbe finished within these x days all saving that pece between croughton and phewarens’ (Longleat House/Thynne Papers/Volume II, f.41). Pond A pond and park pale were constructed at Wilton, taking 1823 man-days (Longleat House/Seymour Papers/Volume IX, ff.92-145). ¢ ‘The ponde ... will meatly well holde water (Longleat House/Thynne Papers/Volume II, f.41). > Land enclosed The boundary included woodland, meadow, common pasture and tenant’s arable land. e There was clear intention to avoid part of an area of common pasture: ‘my lords grace would have had the whole Broyle taken in/ and so to have compasside by the bottom nexte Ramphreis house/ that coulde not have been, for then the tenants of Willton should have no maner of comon for their Rudder beasts [horned cattle] in that syde which would have ben to their utter undoings ...if the wholle woode and bottome aforsaide shoulde be taken from them/ then they could kepe none’ (Longleat House/Thynne Papers/Volume II, f.30). e ‘My lords grace must [decide whether] the tenants whose landes we have taken yn may sow theire barlie croppe there nowe at our ladie day nexte/ for other wise they must be drevyn to sow theire awne/ within the inclosure’ (ibid.). ° Demesne land included ‘109 acres 8 lug of wood grounde [and] 476 acres 3 yerds 27 lug of medowe and other ground’ (ibid.). Location of house e The house was to be ‘at the broyle ende’ (Longleat House/Thynne Papers/Volume II, f.30). ¢ There was a letter signed off ‘from my lord’s graces workes, the hill’ (Longleat House/Thynne Papers/ ° Volume II, f.68). ° ‘we never had above four score and tene men at the top of the hill’ (in margin: ‘my lordes house’) (Longleat House/Thynne Papers/Volume II, f.60). Style of House and Ground Terrain The construction was to be a court house, on or adjacent to a slope. e ‘The courte nex the gate is apointyd to be raysed on the lower syd and abated on the upp syd the same wilbe this night or tomorrow by noone... redy finished’ (Longleat House/Thynne Papers/Volume I, f.100). e *...the inner court [1s] raysed 111 foote’ (ibid.). ° ‘ther remaynithe 11 foote for the lightes into the courts to serve the ix foote story’ (ibid.). e “... the garding on the lowe syd of the house where the ground dyd fall vu foote shalbe raysed to this heghth within thes xii dayes and made levell in all places’ (ibid.). ¢ TI would fayne know my lords graces pleasure for the well in the base courte’ (Longleat House/Thynne Papers/Volume IT, f.68). Scale of the house and foundations ¢ Twill trace and set forthe my lordes graces house according to the plat ... and so be ready to lay the foundacons’ (Longleat House/Thynne Papers/ Volume II, f.41). ° ‘we shall lacke stone for the foundacons for the vi foote walls’ (Longleat House/Thynne Papers/Volume II, £.60). e¢ ‘The brickmakers have cast as muche earthe as will make xx hundrethe thousande bryke by estimacon’ [2 million, in addition to quarry stone and rubble] (Longleat House/Thynne Papers/Volume II, f.41). e “...the houses aboute my lordes graces lodgings ... dugd round to the depth of thre foote’ (Longleat House/Thynne Papers/Volume IT, f.100). View from the house This is given in great detail from one letter, providing the best available information on the house location: ° ‘we have enlarged the wholle grounde as ye shall pceive/ first the medowe on the other side of the ponde towards grafton is enlarged so highe as the furlong goithe leving space for a wey whiche the tenants of willton shall have from theire village to theire comon where we poyntyde thother ponde to be made betwene bothe the felds in the bottom/ so that the medowe THE DUKE OF SOMERSET’S GRAND MANSION AT THE BRAILS, GREAT BEDWYN 24) [shall] rise as highe as the hedge where I would had it gon at the fyrst/ saving the wey aforsaid so that nowe my lords grace shall stand at the place where his house shalbe and have the whole medowe in his iey/ where before he should scarcelie have sene it/, but have lokde over it/. I doute not his grace shall lyke it well in that poynte’ (Longleat House/Thynne Papers/Volume II, f.30). Conduit The letters describe a conduit that had been excavated, with an intention to line it and build a conduit house at the main spring. There was significant debate about the adequacy of the water source: e ‘469 man-days were spent dyggyng springs at tytcombe gate nere dodsdownes busshe’ (Longleat House/Seymour Papers/Volume IX, ff. 92-145). e “.. the conduite 1s a thousand v1 hundrethe foote long and for the most pte thereof xv foote depe’ (Longleat House/Thynne Papers/Volume IT, f.100). * ‘tomorrow [26 June 1549] having ... prepared roughelayers [we] shalbegin to laye the vault which shall gether the water’ (ibid.). ° ‘the vault of the conduite beinge made within a thousand foot long will take much stuff [=stone]’ (Longleat House/Thynne Papers/Volume II, £.60). * ‘we have made the pavurs and althinges necessary for the conduit head’ (ibid.). ¢ ‘we shall have bryke and lead redye ... to begin the head and the vaut’ (ibid.). e “.. within a shorte space I doubt not to have water running in the base court or at least within i foote whear it is best to make the morter heapes because the water runneth there and the lyme kilne not farroff’ (ibid.). ¢ ‘Ther ts at the conduite and hathe bene all this year il score men and sometimes more’ {6000 man-days recorded in bailiff’s account] (ibid.; Longleat House/Seymour Papers/Volume IX, ff. 92-145). * ‘bryan hath advised you by his lre sent by this berer the springs on Wilton Hyll doth pve better then I thought they wold’ (Longleat House/Thynne Papers/Volume II, £98). ¢ ‘..whereas [Mr Hind the plumber] appointed a diche to be cast at the place where the water of the trenche dyd avoyd and to run up the hill to serche upon the very top we have don so and have found a goodly spring hable to fyll a pype of an inche over within a hundreth foote long and vt foote depe and we be lyke to have more {meaning that the water supply could continuously filla 1" pipe with a fall of 6%] And I doubt not but and if the same and all thother that comithe out of the trenche were congregate together it would mayntayn a pype of 11 inches over with as much water as it can avoyd and by estimacon it is a foot above the other. The springes in the pitt at the trenches end is clensyd but little or nothing will cum from them it is drye and nothing is there mor that sokethe from the ground on every syd and wyll not fyll xx gallons in 1 dayes at most’ (Longleat House/Thynne Papers/Volume II, f.83). * Aconduit house was to protect the main source: ‘the conduite house shall stande ... where the chief spring is’ [text re-ordered] (Longleat House/ Thynne Papers/Volume I, f.30). e There is reference to slumping and infill, some relating to Tottenham, and some possibly to the conduit: ‘...the wet wether hathe bene so muche and the bankes be falind againe of the diche upon wylton common whear the cley is about a v1 or vit loogge and our men is upon the same to amend it againe whiles the qickfrithe 1s grene’ (Longleat House/Thynne Papers/Volume II, f.68). Building Materials and Advancement of works Extensive examination and sampling had taken place for a suitable quarry of stone locally at Wilton, Tidworth, Martinsell, Shalbourne and further afield at Hasslebury (Hazlebury near Box) and Purbeck. Bricks and lime were in preparation and being transported, and lead channels and piping were being constructed for carrying the water. Workmen had been sent from Devon, and (unreliable) masons had been brought over from France. ¢ 159 man-days were spent making scaffolds and trenches at Dodsdowne and Topnam. [The date for this - September 1548 - would be too early for starting the house proper and must be associated with ancillary works] (Longleat House/Seymour Papers/Volume IX, ff. 92-145). e [there are ...] ‘xx carters occupied about the carriag of row, sand and water to the brickmaker’(Longleat House/Thynne Papers/Volume II, f.60). ¢ Consideration was given to using flint for building: ‘thoughe my lordes grace myndythe [to use quarry stone] J think it better [to make the outer walls] of flint for ther will aryse above v (c) [S00] lode of flint out of the dykes and other places which cum frecost’ (Longleat House/Thynne Papers/Volume II, f.72). ¢ ‘at Tidworth ther is about un ( ¢ ) [400] ton of chalke redy drawne’ (Longleat House/Thynne Papers/Volume II, £68). 185 man-days were 242 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE spent ‘dyggyng chalk at Boroughethe’, and a payment made ‘for ferrett lyme to meate the house at Dodsdownebusshe’ (Longleat House/Seymour Papers/Volume IX, ff. 92-145). e ‘The brickmaker hathe set fyer upon his kiln alredye and by the tyme Mr Hind cum he shall have bricks enoughe for his conduite’ (ibid.). ¢ “Sf Mr Hind were heare ther is lead wood and sand and all other thinges ready for him at the parsonage at Bedwin’ (ibid.). ° ‘all the devenshire men we have are in hand withal and have been thes i11 wekes thextremety of wether 1s so muche and the slewthe of the people together that muche worke can not be had at so few mens handes’ (Longleat House/Thynne Papers/Volume II, f.83). e “.. ye sent us such a lewd companye of Frenchmen mason as I never sawe the lyke ...the worst condicyoned people that eny I sawe and the dronkenst for they wyll drynke more in one day then thre dayes wages wyll comto and then lye lyke bests on the flore not able to stande ... And now they be [departed] and stolen away ... yndeptyd to dyvers folks yn theys ptes’ (ibid.). Cessation of the Project Two letters, both dated 25 June 1549, from Teshe to Thynne, and Berwick to Thynne, when the project was advancing rapidly, provide the last surviving word on construction. A final defensive letter from Berwick to Thynne of 23 August 1549 tangentially mentions Mr Hind the plumber, but otherwise begs excuse for the inadequacy of men, horses and harnesses he had sent, and might relate to a period of civil unrest (Longleat House/Thynne Papers/ Volume II, f.129). There is no record of whether work formally ceased, was held in abeyance or was curtailed by outside forces. Difficulties with the French masons, the re-deployment of the workforce following insurrection, problems with water supply, or Somerset’s increasingly precarious role as Protector, may have all played their part. Somerset was first arrested in October of 1549. Sir John Thynne was also seized, and under cross examination by Council in November he stated that he disliked the great expenses of the Duke of Somersett’s buildings, and often wished he would have left them’ (Longleat House/Seymour Papers/Volume IV, ff.127- 128). These sentiments did not dissuade him from building Longleat House later! After the execution of the Duke of Somerset, letters patent of Edward VI in 1552 granted to William Herbert, Earl of Pembroke, certain property Fig. 3 The Conduit Trench in Bedwyn Brail, holding stagnant water in May 2005 formerly belonging to the Duke, including ‘all those channels, leaden pipes, and bricks in or about our parcel of land called Doddysdowne adjacent to the wood called Le Broyle’ (WRSO 1300/176). B. Archaeological Evidence Although the ends of the conduit trench are insufficiently distinct to permit exact measurement and correlation with written records, it remains an archaeological feature unequivocally identified with Seymour’s mansion. An assessment of the Water and Sewage Industries of the Monuments Protection Programme, prepared for English Heritage, confirms the uniqueness and archaeological importance of this feature (Trent and Peak Archaeological Unit, 2000). It has never been excavated. The trench still holds stagnant water on occasions (Figure 3). Extensive lengths of the bank and hedge, constructed as part of the park boundary, probably also survive. THE DUKE OF SOMERSET’S GRAND MANSION AT THE BRAILS, GREAT BEDWYN 243 Grafton Site of The Heath (Common Pasture) 2 Mi (Ado Track (right-of-way) me + Cont tae RET cs A On as ht Brookland’s Furlong ‘a3 . Wilton ee ; yee = Water ™ Soe RS ah Oe ee =~ AP NT a Pee 4 Fig. 4 Photograph near the Roman road northwest of Wilton at SU267621, relating to a description given in John Berwick’s letter of November 1548 ‘the medowe on the other side of the ponde towards grafton is enlarged so highe as the furlong goithe leving space for a wey whiche the tenants of willton shall have from theire village to theire comon where we poyntyde thother ponde to be made betwene bothe the felds in the bottom ...so that nowe my lords grace shall stand at the place where his house shalbe and have the whole medowe in his iey’ (Longleat House/Thynne Papers/Volume I, f 30) Location of Seymour’s Mansion —a Review of Options Jackson and Stallard placed the house at the southern end of the trench near Hill Barn (Site A, Figure 1). In the absence of any excavation to determine gradient, amidst the undergrowth and debris it is impossible to determine the intended line of fall. Further, the Eocene strata here dip to the north, which militates in favour ofa northern flow under natural conditions (Joan Davies, pers. com.). There remains moderate local support (as expressed during a 50-strong field meeting of the Bedwyn History Society) for the theory that the house could have been located at the northern end of the conduit trench (Site B). Clear terracing to the east of this site, at SU285625, might be associated with works at this time. In 1874 Stallard asserted that a conduit located ‘just to the rear of the House would have been the most natural position’ (Wiltshire Archaeological Notes, 1932). However, the function of the trench may have been primarily to gather water, and not necessarily to transport it directly to the house. Conduits were constructed or adopted from existing structures widely to carry water to properties in Tudor times. Elsewhere these did not involve trenches, and water was carried in pipes from the source, often protected from animal fouling and degradation by a conduit house constructed over the spring. Sir William Sharrington had an involvement in many properties where this method was employed to provide domestic water. At Sudeley Castle, an acquisition of Somerset's brother Thomas Seymour (a confederate of Sharrington’s), a pipe carried water from a source some 1.2 km from the property. A conduit house constructed over the Castle’s water source at St Kenelm’s Well, shows features strongly characterisuc of Sharington’s style. At Lacock Abbey, a former monastic property which Sharrington converted following his acquisition in 1540, the water supply was piped some |.Skm from Bowden Hull, down 244 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 5 Aerial Photograph, 26 April 1976, NMR 921/402, at Grid Ref. SU2662, showing unidentified square structure, 39m across, and other ground features (centre). Roman road extends mid left to bottom RH corner (where a small reservoir is visible); from the same corner a public right-of-way extends as a dark line northwards to top mid-right of photograph. © Crown copyright. NMR the slope and under the river before rising again to reach the house. A conduit house on Bowden Hill is of similar design to that at Sudeley (Clark-Maxwell 1913, 178). There is no reason why water gathered in the trench at Bedwyn Brail could not have been piped, and contrary to Stallard’s contention, this might indeed involve traversing a dip in the ground, before rising again, provided that the end point was lower than the source. Knowledge gleaned from Sharrington’s various properties would have been available to workers at The Brails; the bailiff’s account for November 1548 specifically records an entry of 5/- paid ‘to a servant of Sir William Sharington’s who cam to vyede the springs nere dodsdowne busshe’ (Longleat House/Seymour Papers/Volume IX, ff. 92-145). Hence, there is no pre-requisite for the mansion to be close to the conduit trench. Of particular importance is the detailed description of the view from the house in 1548 (see above). To concur with the identification of the house at the southern end of the trench, Stallard concluded that the meadow referred to would have been located on the top of the hill, next to the house, rather than a more usual riverside or valley setting, — which would also be in keeping with the reference to a pond between fields in the bottom. Further, it is clear from his sketch map (Figure 2) that he interpreted the phrase ‘the medowe [shall] rise as highe as the hedge’ as meaning that soil would be spread to raise the ground level to the height of the adjoining hedge. Again this is implausible; it would require immense labour and effectively be futile. The letter claims that following these works, the view from the house would be improved, an unlikely scenario from simply raising the height of the land. The bailiff’s records include no such works. A more plausible interpretation is that the meadow ground has been enlarged (laterally) up to a hedgeline. This meadow might then come within the view of a house on a brow opposite. The meadow is described as being on the other side of a pond ‘towards Grafton’. This also presents difficulties. There is no surviving pond which would appear to conform to this location near Hill Barn. Further it would be more appropriate to describe the meadow as being ‘towards Wilton’, since this is closer. (Contrary to Stallard’s assertion, the Tudor letters do not state that Grafton or Wolfhall were visible from the site). They make clear reference to leaving a way for the tenants of Wilton to reach their common. Stallard considered that this was the same common mentioned in other letters as a pasture for ‘rudder beasts’. However, there was other common land in the vicinity of Wilton. Site C Figure 4 is a photograph from a site near the Roman road northwest of Wilton at SU267621. It is 1.4km from the southern end of the conduit trench. The valley is now occupied by Wilton Water, although a smaller pond was present at this location in the 18" century (Andrews and Dury 1773; Haynes 1792). Looking beyond the pond towards Grafton a bottom field is visible, formerly called Brookland’s Furlong (Haynes 1792). Along the edge of this is an extant right-of-way from Wilton to an area of former common pasture called The Heath (shown in Figure 1). Hence this photograph shows a view which conforms closely with that described in 1548. Figure 5 is an aerial photograph taken in April 1976 showing a soil mark adjoining the Roman road, near the view-site described above (labelled C on Figure 1). This shows the inner and outer walls of a building 39m across, within which is lighter ground. Further substantial features are THE DUKE OF SOMERSET’S GRAND MANSION AT THE BRAILS, GREAT BEDWYN 245 also visible. The remains are unidentified, and may be wholly irrelevant to this study, but they could be construed as a courtyard house, with corner turrets and a possible gate to the west. The Sites and Monuments Record describes the corner features as ‘apparent finials’ (site SU26SE, 665). The site looks out across other Seymour properties at Tottenham and Savernake. Any speculative linkage of Seymour’s mansion with Site C presents new difficulties. First, the ground level here is approximately the same as that of the trench. It is unlikely that water could ever have been brought to this site from The Brails for domestic purposes, although it might have been brought sufficiently close to assist building works. Second, there is a tumulus remaining as an earthwork immediately adjoining this site (SAM 30452). It is unlikely that this could have survived unless the works ceased at a very early stage. Conclusions Some 130 years after records were unearthed relating to Seymour’s mansion, its location at The Brails remains unresolved. It has emerged that the widely published Site A at the southern end of Bedwyn Brail was not unquestioningly accepted by those undertaking original research in this area. There is no documentary or archaeological evidence to support this site, or at alternative Site B at the northern end of the trench. Site C meets the descriptions and views given in documentary sources, although would have had considerable difficulties concerning water supply. Archaeological work is required to resolve this matter. First, survey and sample excavation are necessary to develop an understanding of the conduit, its depth, direction of flow, the location of the conduit house, and possible arrangement of pavers. Second, the soil mark at site C requires geophysical survey to elucidate its relevance to this work. A labour-force of 280, providing an effort that (incomplete) records show exceeded 11,600 man-days, must have left a trace which can be determined today. Acknowledgements I would like to record my grateful thanks to Mr Richard Charles, owner of The Brails, for his encouragement, enthusiasm and for permission to visit, and to Mrs Joan Davies, for her unstinting graft and challenging debate. References Original Sources Wiltshire and Swindon Record Office (WSRO) WSRO 9/14/223, Indentures Harding Farm, c.1597- c.1637. WSRO 9/27/59, Benjamin Haynes’s Plan of Wilton in Great Bedwyn, 1792. WRSO 1300/176, copy Letter Patent, Edward VI, 1552. Longleat House Archives Seymour Papers, Volume IV, John Thynne Draft Answers to Council, ff. 127-128. Seymour Papers, Volume IX, Money disbursed by John Hutchins, Bailiff of Wolfhall, March 1548 — February 1549, ff. 92-145. Thynne Papers, Volume II, Letters John Berwick to John Thynne, 1548-9, ff. 30, 45, 62, 88, 98, 129. Thynne Papers, Volume II, Letters Bryan Teshe to John Thynne, 1549, ff. 41, 60, 68, 72, 83, 87, 100. WANHS Library, Devizes Manuscript Letters Rev G. Stallard to Canon J.E. Jackson, May-December 1874, (collated in) Wiltshire Archaeological Notes, 1932 (WANHS Library, Devizes). National Monuments Record, Swindon Aerial Photograph NMR 921/402, SU2662/1, 26 April 1976. Printed and Published Sources ADAMS, W.M., undated cl1900, Wolfhall and Tottenham, Marlborough: Maurice & Co. ADAMS, W.M., 1905, Savernake in the Vale, Chapter X Wilton, Marlborough Times. ALILESBURY, Marquis of, 1962, A History of Savernake Forest, Devizes. ANDREWS and DURY, 1773,A Map of Wilshire taken from an Actual Survey. CLARK-MAXWELL, W.G., 1913, Sir William Sharrington’s Work at Lacock, Sudeley and Dudley, Archaeological Journal 70, no. 278, 174-182. JACKSON, J.E., 1875, Wulfhall and the Seymours, WANHM 1S, 140-207. JORDAN, W.K.,1968, Edward VI: the Young King, Cambridge, Mass: Harvard UP. Jordan, W.K.,1970, Edward VI: the Threshold of Power, London: George Allen & Unwin TRENT and PEAK ARCHAEOLOGICAL UNIT, 2000, Report of the Monuments Protection Programme, Water and Sewage Industries, English Heritage WATTS, K., 1996, Wiltshire Deer Parks, An Introductory Survey, WANHM 89, 88-98 Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 246-260 Notes and Shorter Contributions The Duke’s Vaunt: a recently discovered painting by Joan Davies! and Graham Bathe? Fig. 1 Picture labelled on reverse ‘Duke of Somerset’s Tree’ with the apparent artist’s name G Martin The Duke’s Vaunt is one of the large, ancient, named oaks in Savernake Forest. Local folklore asserts that the tree was the pride of Edward Seymour, Duke of Somerset, Lord Protector and Hereditary Warden of Savernake Forest who became owner of the For- est in 1548. It was Edward’s sister Jane Seymour who married Henry VIII and was the mother of Edward VI. A nineteenth-century painting depicting the Duke’s Vaunt, and labelled on the reverse “Duke of Somerset’s Tree” with the (apparent) artist’s name “G Martin”, has come to light following an auction in 2005 of six small pictures of Savernake. Whilst artistically undistinguished, it represents the only 'Ballard’s Piece, Forest Hill, Marlborough SN8 3HN ’Byeley in Densome, Woodgreen, Fordingbridge, SP6 2QU NOTES AND SHORTER CONTRIBUTIONS DwKEe's Vane. Fig. 2 Pen and watercolour drawing of the Duke’s Vaunt by John Stone, 1801 known faithful (un-stylised) rendering of this tree in its heyday (Figure 1). The Duke’s Vaunt, although hollow and decayed, was still alive in October 2005, situated in the eastern part of Savernake Forest on the western edge of Birch Copse at grid reference SU 2389 6646. The tree and known history is described alongside other great and named Savernake oaks in Oliver and Davies (2001, 24-46). The picture does not bear a date, but would appear to have been painted shortly after a known P . on : “t Tea Fig. 3 Sketch of the Duke’s Vaunt made in 1893 by IZM. Adams 247 pen and watercolour drawing made in 1801 by John Stone. The latter, held in Devizes Museum, is probably the original drawing for a published copper engraving by FE Cary (Stone 1802, 497) (Figure 2). A later book, Adams (c1905) reproduces a sketch of the same tree made in 1893 (Figure 3). All three pictures show a characteristic massive bulge on the trunk on the left-hand side of an opening into the hollow tree, and were probably drawn from the north-west. A tree marked as the “Duke’s Font” on an estate map of Savernake of 1786 (WSRO 1300/360) is shown in an identical position to the Duke’s Vaunt on Ordnance Survey maps since the six-inch edition of the 1880s. The tree was a boundary of Crabtree Common in 1778, when William Farmer’s Common was described as going “from Baden Coppice railes across to Puttal Horse Leaze Gate and so to Duke Vaunt, being the old oak at the end of Birch Coppice.” (WSRO 9/22/221). The sheep in the painting presumably derive from the Common. The part of the Forest known as Crabtree Common in 2005 was smaller, lying south of Crabtree Cottages and not extending as far as the Duke’s Vaunt. In the late twentieth century it became difficult to reach and find the Duke’s Vaunt, over-crowded within a dense plantation of Douglas Firs. In winter 2004-5 the Forestry Commission cleared the closest firs and removed a large overhanging beech. This exposed the structure of the two main living branches. Comparison of these with the pictures suggest that the bulge is now totally lost, and unrelated to another deformity which appears in the photograph of the Duke’s Vaunt in Oliver & Davies (2001). References Original Sources Wiltshire and Swindon Record Office (WSRO) WSRO 9/22/221, Survey of Savernake Forest and Savernake Great Park, with details of common grazing rights in the Forest, 1778-c.1819. WSRO 1300/360, A Plan of the Forest of Savernake and of Tottenham Park, 1786. Published Sources Adams W.M., undated c1905, Sylvan Savernake, Maurice & Co Marlborough Oliver J., & Davies J., 2001, ‘Savernake Forest Oaks’, WANHM 94, 24-46. Stone J., 1802, ‘Savernake’, Part I, 497 The Gentleman's Magazine, 248 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Celtic survival at Chicklade by Andrew Breeze The village of Chicklade (ST 9134) lies fourteen miles west of Salisbury on the A 303. It appears as Cytlid in a letter of c. 912 describing cattle-rustling at nearby Fonthill (Whitelock 1979, 544-6). So the name survives in an accurate early text. It has been taken as a hybrid, with a Celtic first element corresponding to Welsh coed ‘wood’ (there are dense woods north of Chicklade), and an Old English second one from hilid, meaning ‘gate’ or perhaps ‘slope’. Yet the absence of h in attestations casts doubt on this. Coates and Watts thus consider the second element ‘obscure’, and it was also obscure to later generations of the English, who corrupted the second element to lade ‘place where water is crossed’, even though Chicklade is far from anything resembling a ford (Coates and Breeze 2000, 340; Watts 2004, 136). A new approach may hence be possible. Many Welsh adjectives are formed by adding -/yd to nouns: llychlyd ‘dusty’ from llzoch ‘dust’, chweinlyd ‘flea-ridder’ from chwain ‘fleas’, gwaedlyd ‘bloody’ from gwaed ‘blood’, rhydlyd ‘rusty’ from rhwd ‘rust’ (Morris-Jones 1913, 257), and so on. This formation is ancient. The British-Latin name of Carrawburgh (NY 8571), on a moorland stretch of Hadrian’s Wall, was Brocolitia ‘heathery place’, which would give Welsh gruglyd ‘heathy, heathery’ (Rivet and Smith 1979, 284-5), a word still used in South Wales. (The existence of Welsh gruglyd, not previously mentioned in this context, lets us rule out explanations of Brocolitia as ‘place abounding in badgers’.) These analogies allow us to explain Cytlid from a Brittonic adjective meaning ‘woody, abounding in woods’, describing Chicklade’s ancient surroundings. Though the form is not otherwise known, we need not doubt its existence. It would be synonymous with Welsh coediog ‘wooded, abounding with trees, silvan’, which is well attested, and which accounts for the name of Chideock (SY 4292) near Bridport in Dorset (Coates and Breeze 2000, 294). Like Chideock, it might originally have been a stream-name. If it was, it would refer to the winterbourne flowing east from Chicklade and then south past Fonthill to the river Nadder. But we cannot be certain of this, because -lyd is rare or unknown in Welsh hydronyms. Nevertheless, we can Private now be sure that Chicklade is a pure Celtic form. This revised etymology will strengthen the evidence for Celtic survival in Wiltshire, in addition to nearby Pertwood (ST 8935) by ‘Upper Pertwood Bushes’, with a first element corresponding to Welsh perth ‘bush’, and Fonthill, with an element paralleling Welsh ffynnon ‘spring’ (Coates and Breeze 2000, 339, 340). The English occupied west Wiltshire in the later sixth century. After that, Selwood on the Wiltshire-Somerset border separated the territories of Briton and Saxon for a hundred years. (Asser in the ninth century still knew Selwood by its Old Welsh name Coit Maur ‘great wood’, suggesting its political importance.) The wooded terrain east of Selwood was presumably unattractive to English settlers. Hence, it seems, the survival of Brittonic names and perhaps British communities. If Chicklade is correctly explained as a Celtic term meaning ‘wooded, abounding in trees’, it deepens the Celtic colouring of south-west Wiltshire. Sixth- century archaeological remains there are likelier to be from Celtic communities (no doubt servile ones, with new English masters). It also offers a task for researchers in Wiltshire archives. The work of Pamela Russell on Lancashire records has brought to light many minor Celtic toponyms (of fields, streams, and ditches) around Liverpool. From now on, archaeologists may be more on the lookout for Celtic traces around Chicklade, and historians for minor Celtic toponyms in post-Norman records of the area. Together, these varied kinds of evidence reinforce the case for British survival in Anglo-Saxon Wiltshire. References COATES, R. A., and BREEZE, A. C., 2000, Celtic voices, English places. Stamford MORRIS-JONES, J., 1913, A Welsh grammar. Oxford RIVET, A. L. E, and SMITH, C., 1979, The place-names of Roman Britain. London RUSSELL, P B., 1992. Place-name evidence for the survival of British settlements in the West Derby Hundred, Northern History 28, 25-41 WATTS, V.E. (ed.), 2004, The Cambridge dictionary of English place-names. Cambridge WHITELOCK, D. (ed.), 1979, English historical documents. London 2™ edn. NOTES AND SHORTER CONTRIBUTIONS 249 The Lost Dedication of the Church at Winterbourne Dauntsey, Wiltshire by Fohn Lush The old churches at Winterbourne Dauntsey and Winterbourne Earls were demolished in 1867, and were replaced by a new church to serve both parishes. All three churches are well documented (RCHME, 1987, 220-223). The new church was built on the Winterbourne Earls side of the boundary between the two parishes, and retained the dedication to St Michael of the former church at Winterbourne Earls. The dedication of the former church at Winterbourne Dauntsey was not perpetuated. It is known that by 1405 it was dedicated to St Edward (Timmins, 1984, 26-27) but, with the passage of time, the tradition has been lost as to whether this referred to St Edward the Martyr or to St Edward the Confessor. The standard work on English church dedications is by Frances Arnold-Forster (1899). It is based primarily on the Clergy List for 1896, and considers the dedications of 14,000 Anglican churches and chapels in England. It is an impressive work, but the author herself did not regard it as definitive. With regard to churches dedicated to St Edward, if Arnold- Forster was unable to find positive evidence in favour ofa dedication to Edward the Martyr, she assumed, as for the church at Winterbourne Dauntsey, that _ the dedication was to Edward the Confessor, but she presented no evidence to support this assumption. This note provides that evidence. Edward the Martyr’s life is well summarised in The Oxford Dictionary of Saints (Farmer, 1997, 155-156). He was the son of King Edgar by his first wife Authelfled, and became king in 975, aged about 13. In 978, having been out hunting, he visited his stepmother, Edgar’s second wife Ailfthryth, at Corfe and, while drinking from a proffered cup, was stabbed by an attendant. Edward’s horse bolted, dragging him by one foot which was caught in a stirrup, and he died. He was initially buried without ceremony at Wareham but, following reports of miracles at his grave, he was translated to a new shrine at Shaftesbury Abbey in 980. He became the object of a widespread cult in the early 11" century, 2 Penman Close, St. Albans, Herts, AL2 3DJ and there are still churches or chapels dedicated to him. In portraits, he is often shown carrying a cup and a dagger, in allusion to the manner of his death. Edward the Confessor’s life has been well documented (Barlow, 1997). He was born in about 1004 or 1005, spent much of his life in Normandy during the reigns of Cnut and his sons, and became king in 1042 following the death of Harthacnut. He married Edith, daughter of Earl Godwin of Wessex, in 1045, but died childless on 5 January 1066. The lack of a clear successor was one of the causes of the Norman Conquest later that year. Edward was buried in his newly-rebuilt Westminster Abbey, his shrine became the focus of pilgrimages, and miracles are reported to have occurred there. He was canonised in 1161 by Pope Alexander III, but it was not until 1163 that his remains were translated to a magnificent new shrine in Westminster Abbey. For this occasion, Abbot Aelred of Rievaulx was commissioned to write a new Life of St Edward, the Vita Sancti Edwardi Regis et Confessoris (Twysden, 1652, cols.369-414), and the date of his translation, 13 October, rather than the date of his death, is now observed as Edward’s feast day. Aelred based his Life of St Edward on earlier Lives, but enhanced it by including accounts of new miracles. In particular, Aelred related the story of how Edward, while at the consecration of a church to St John the Evangelist, for whom Edward had a particular affection, was approached by a beggar asking for alms. Having no money to hand, Edward gave the beggar a valuable ring from his finger. Some time later two pilgrims, having become lost on their way to Jerusalem, encountered a procession which included a dignified elder, who asked the pilgrims whence they came, and who was their king. When told, the elder revealed that he was St John the Evangelist who, in the guise of a beggar, had received a ring from Edward. He asked them to return the ring to Edward, with the message that Edward would soon meet St John again, in paradise. The ring 250 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE and the message were duly delivered, to Edward’s delight, and he hastened to have Westminster Abbey completed before he died. This legend has defined Edward’s iconography, so that he is most often shown bearing a ring, sometimes in company with a pilgrim. Winterbourne Dauntsey church is first mentioned in three Salisbury Charters of 1253-59 (Jones, 1879, 373), so it must have been built by about 1250. During the destruction of the church in 1867, a series of fine wall paintings was discovered, which have been dated to about 1250-1275 (Tristram, 1950, 220-222), so they are essentially contemporaneous with the construction of the church. The paintings depicted scenes from the life of Christ, and appear to have extended right round the church. That the paintings survived at all is almost certainly because they had long since been covered by whitewash and plaster, which had to be painstakingly removed to reveal them. Not surprisingly, they were in very variable condition, and it is almost certain that some had been completely destroyed. Photographs were taken of those which had survived, before they too were destroyed. The best preserved of all the paintings was the one depicting the Adoration of the Magi. This was painted on the eastern splay of the westernmost window in the north wall of the church (RCHME, 1987, 25). The Virgin is seated on a high throne on the left, with her right hand supporting the infant Christ on her lap. He is holding a book in his left hand, and giving a blessing with His right hand. One of the three kings is kneeling in the centre of the painting, and the other two are standing behind him. All three kings wear crowns and garb in 13" century style. In the present context, the most important ° figure is that of the kneeling king. He is offering a gift to the Christ Child, not gold, frankincense or myrrh as usually depicted, but a large gold ring. It seems to be beyond doubt that this painting in the former church at Winterbourne Dauntsey depicts Edward the Confessor as one of the Magi, as a conscious contemporary reference to him as patron saint of the church, which was built and decorated at a time of great interest in England’s canonised king. Bibliography ARNOLD-FORSTER, F, 1899. Studies in Church Dedications: or, England’s Patron Saints, London BARLOW, E, 1997. Edward the Confessor, New Haven and London: Yale University Press FARMER, D. H., 1997. The Oxford Dictionary of Saints, Oxford University Press JONES, W. H., 1879. Fasti Ecclesie Sarisberiensis, or, a Calendar of the Bishops, Deans, Archdeacons and Members of the Cathedral Body at Salisbury, from the earliest times to the present. Salisbury: Brown & Co RCHME, 1987. Royal Commission on Historical Monuments (England), Churches of South-East Wiltshire, London: HMSO TIMMINS, T. C. B., 1984. The register of John Chandler, Dean of Salisbury 1404-17, Wiltshire Record Society, Vol. 39 (1983), Devizes TRISTRAM, E. W., 1950. English Medieval Wall Painting (The Thirteenth Century), London: OUP. TWYSDEN, Sir R., Bt., 1652. Historiae Anglicanae Scriptores X, London The Romanesque Corbel-table at St John’s, Devizes, and its Sheela-na-gig by Alex Woodcock! and Theresa Oakley’ The church of St John the Baptist in Devizes has long been noteworthy for its Romanesque architectural sculpture. In the 1930s A. W. Clapham suggested that the tower was unique among English churches for possessing, on its interior (above the crossing arches), probably the only example of a blind arcade with triple intersecting arches in the country (Clapham 1934, 108, 133). Yet just as important are the carved ‘alexwoodcock888@yahoo.co.uk ’sigillum@v21.me.uk NOTES AND SHORTER CONTRIBUTIONS 25] Fig. 1. The sheela-na-gig and male exhibitionist figure on the north side of the chancel (photograph by Alex Woodcock) corbel-tables on the exterior of the chancel and east end of the north and south transepts, that on the south side almost fully enclosed by a late medieval chapel but essentially undisturbed, that on the north partially destroyed by a chapel of similar date. Both sequences display a wide range of carving styles and subjects including, on the north side of the chancel, the figure of a sheela-na-gig (Fig. 1). Definitions of what constitutes a sheela-na-gig differ, but typically these are images of naked, bald, female figures which gesture towards or emphasize their genitalia. Sheela-na-gigs can be found throughout the medieval architectural sculpture of Great Britain and Ireland although some authors have suggested they also exist in other parts of Europe, such as Norway (Krtiger 1994) and France (Weir and Jerman 1986). They are frequently interpreted as fertility figures or didactic images to warn against the sin of lust (Cherry 1992). Approximately 140 images are extant, many of which are in Ireland. There, in situ images occur most frequently on tower houses with fewer examples on churches; in Britain, the vast majority occur on churches, although few are found in corbel tables. In the south of England examples can be found in Dorset at Studland (Lundgren and Thurlby 1999) and Worth Matravers (Woodcock 2002); in Somerset at Stoke Sub Hamdon (Ashdown 1993) and in Devon on the south tower of Exeter Cathedral (Woodcock 2005, 114). If finding a sheela-na-gig within a 12""-century corbel-table is a relative rarity, then it is even more unusual to find one partnered with a male figure, as at Devizes. One other example of such a pairing is currently known in Britain at Whittlesford (Cambridgeshire) carved upon a stone above a window on the north side of the tower, although there is possibly another, carved upon a corbel, at St Peter’s, Northampton. The church of St John at Devizes was built within the inner bailey of the Norman castle during the 1160s and consisted of a nave, chancel, crossing tower with circular stair turret on the northwest corner, and north and south transepts Modifications over the next two centuries included the addition of chapels or apses to the east walls of both transepts and the building of a north porch and north aisle to the nave, features that were 252 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE subsequently incorporated into or superseded by the fifteenth-century addition of north and south chapels and enlargement of the nave (Pugh 1975, 289). This phase of activity also witnessed the removal of the Romanesque west front, addition of crenellations and pinnacles to the tower, and rebuilding of the lower part of the tower staircase. Formerly external walls, the south side of the chancel and the east-end of the south transept were largely enclosed in the later medieval period by the construction of a chapel. Only one corbel remains on the exterior east face of the south transept; the corbel-table that runs along the two interior walls now forms part of a roof but is otherwise in its original situ. The style and subject matter of this sequence are very similar to those corbels on the exterior of the north and south sides of the chancel at the other medieval church in Devizes, St Mary’s. Both churches date to the same later twelfth-century period and were likely to have been constructed by similar teams of masons. From south to north on the former east face of the south transept, the series runs as follows: Exterior: 1) Beast head with open mouth lined with teeth, pointed ears and large eyes. Interior: 2) Open mouthed beast head with teeth. 3) Bundle of sticks tied with a rope. 4) Paired human and beast heads. 5) Beast head with wide open mouth. 6) Plain. 7) Beast head with open mouth and split tongue. 8) Bundle of sticks tied with a rope. The sequence continues from west to east along the former external south face of the chancel: 1) Beast head. 2) Bundle of sticks tied by a rope. 3) (Partially obscured by the insertion of a fifteenth-century arch) Lion with its head turned backwards and its tail curved upwards through its hind legs. 4) Paired human and animal heads. 5) Beast head with wide open mouth. 6) Former beast/cat head — the corbel has been cut away leaving only its outline. 7) Human head with open mouth. 8) Paired human and beast heads, each angled out from the wall by forty-five degrees and from each other by ninety degrees. 9) Beast head with protruding tongue. 10) Beast head with open mouth. 11) Plain. 12) Bundle of sticks tied by a rope. 13) Plain. Two corbels also survive on the west face of the south transept (north to south): 1) Bundle of sticks tied together by a rope. 2) Beast head with striated face. The sequence is distinguished by the concentration of generic beast heads — feline in overall appearance and characterized by large almond-shaped eyes, sharply pointed ears and an open snout with or without protruding tongue and/or dagger-like teeth. Whilst there is often great variation in the execution of the beast head template at a given site (for example, Romsey Abbey), in this sequence at Devizes the carvings tend towards uniformity and simplicity as evident in the rounded, bulbous snouts and general smoothness of the finish. Corbels carved in a similar style can be found on the north transept of Winchester Cathedral (c1100), in the choir at Chichester Cathedral (cl 100-25) and on the north transept of Romsey Abbey (c1120-1150). Similar motifs may also be found at some of these sites; the bundle of sticks tied together by a rope, for example, is common at Romsey while the paired heads (in various combinations of human and beast), angled out from the wall at forty-five degrees, can be found throughout the corbel-tables at both Romsey and Winchester. A number of wooden corbels carved with human heads, and likely to belong to the twelfth century, may also be connected with St John’s. Whilst their provenance is unknown the style of carving has been compared with stone corbels at Chichester Cathedral, which would substantiate this stylistic network further (Stalley 1970, 228-9). If the corbel-table on the south side of the chancel can be easily contextualized in terms of comparison with examples on other Romanesque buildings in the region, the corbel-table that remains upon the north side of the chancel and east face of the north transept is quite different. Many corbels are hidden (or have been destroyed) by a fifteenth-century chapel built against the second bay of the chancel and the east face of the north transept. Of the six that remain on the easternmost bay of the chancel, and the three on the east face of the north transept, the carving is detailed and the subject matter unusual and original. Unlike those on the south side of the chancel, repetition of imagery is avoided. On the north side of the chancel, east to west, the sequence runs as follows: NOTES AND SHORTER CONTRIBUTIONS 1) Abstract squatting figure with oversized head and large almond-shaped eyes. The hands rest upon the knees and appear to pull open the mouth, out of which the tongue protrudes. 2) Exhibitionist couple. On the viewer’s left a female figure (sheela-na-gig) squats, her genitals clearly visible. Her forearms and hands are missing from below the elbow. On the right a male figure, his body slightly angled away from the female, holds his erect penis in his right hand. 3) Two beast heads with staring eyes and grinning, teeth-filled mouths. 4) Beast head and a human figure with its left arm bent underneath the body, the hand resting beneath the chin. Both figures bite upon different ends of the same length of rope. 5) Beast head with wide snout and open mouth lined with teeth. Small arms and human hands appear from each side of the figure, and the hands grip the lower jaw on each side of the mouth. 6) An embracing couple inside the open jaws of a beast head with bulging eyes. East face of the north transept, south to north: 1) Beast head, possibly that of a boar, with its tongue extended to lick the up of its snout. 2) Head and upper body of an unidentifiable animal, whose forelimbs hold the sides of the face. Something appears to protrude from the mouth (two human legs?). 3) Exhibitionist figure with beast head, holding its legs upward to expose the anus. Patrons are difficult to identify. Bishop Roger of Salisbury, one of the most active sponsors of Romanesque architecture in the region, is unlikely (Stalley 1971, 81-83). From the Conquest until 1157 the land occupied by the castle and its precincts was church property belonging to the see of Salisbury and its manor at Bishops Cannings (Stone 1919, 418). However, following the eighteen-year dispute over ownership begun by King Stephen’s capture of Devizes in 1139, the cathedral renounced its claim on the town and castle in return for compensation elsewhere. Ecclesiastical patronage for St John’s, built during the 1160s, is therefore doubtful. Yet the rich architectural decoration at this site indicates patronage of considerable wealth. With ownership of the castle and town assumed by the crown the building of two churches in Devizes (St John’s and St Mary’s) may represent a consolidation of that ownership as well as an expression of secular power. 253 If the specifics of patronage remain conjectural, the presence of a sheela-na-gig raises interesting questions. The church of St John, within the castle bailey and possibly built by a secular authority, would not appear to have provided the most ideal situation for a didactic image warning against the sin of lust, especially when there is a more accessible church nearby. The inclusion of a sheela- na-gig within the architectural sculpture at this site suggests that it was at least considered favourable by patrons, if not positively chosen to represent a deeper meaning. Just what this might be, however, is elusive. Seeking definitive meanings for the images found among medieval architectural sculpture often leads to unsatisfactory conclusions; for example, the interpretation of sheela-na-gigs as fertility goddesses, or that they have a particular relationship with the north side of medieval buildings. The north side of a church is often understood to be more appropriate for certain images, an idea perhaps derived from the folklore associated with the north or ‘devil’s’ door. A symbolism is assumed based upon perceived programmes of imagery: thus ‘the north was associated with evil, darkness, the Jews and the Crucifixion; whereas the south was associated with good, light, the apostles and the Resurrection’ (Graves 1989, 309). Whether oversimplified or, indeed, irrelevant, the concept of different zones of meaning articulated by various media in medieval churches colours archaeological work — many nunneries had their cloisters on the north, rather than the south side (the layout for monastic sites) (Gilchrist 1994, 128-149) for example, which suggests a possible ascribed or symbolic association between the north side of medieval ecclesiastical buildings and the feminine. However, this view cannot be endorsed for medieval marginal sculpture. The Devizes sheela-na-gig, located on the north side of the chancel, is a case in point. Recent (and continuing) work suggests that there is no significance regarding the side of the building on which sheela-na-gigs are sited: in Britain, 37% are found on the north side and 41% on the south; in Ireland, only 9% are found on the north side while 31% appear on the south and 28% on the east (Oakley 2005, 30, Fig. 16). If the north side is significant at all, then so is the south and thus a simple association of the north with female imagery, eviland darkness cannot, with sheela-na-gigs at least, be substantiated. To understand the significance of sheela-na-gigs for medieval audiences they must be firmly set within 254 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE their Christian context. Space prohibits discussion here, but a significant feature, we suggest, appears to be their use as objects of status. The Devizes example, sited upon a richly ornamented church within the inner bailey of what would have been a noteworthy 12"-century castle, clearly supports such a hypothesis. Bibliography ASHDOWN, P, 1993, Two sheila-na-gigs at Stoke-sub- Hamdon. Somerset Archaeology and Natural History 137, 69-74 CHERRY, S., 1992, A Guide to Sheela-na-gigs. Dublin: National Museum of Ireland CLAPHAM, A. W., 1934, English Romanesque Architecture after the Conquest. Oxford: Clarendon Press GILCHRIST, R., 1994, Gender and Material Culture: The Archaeology of Religious Women, London: Routledge GRAVES, C. P, 1989, Social space in the English medieval parish church, Economy and Society 18, 297-322 KRUGER, S., 1994, Sheela-na-gig — provokasjon — konflikt — forlgsning. SPOR: Fortidsnytt Fra Midt-Norge 18, 26-30 LUNDGREN, K. and THURLBY, M., 1999, The Romanesque church of St. Nicholas, Studland (Dorset). Proceedings of the Dorset Natural History and Archaeological Society 121, 1-16 OAKLEY, T. C., 2005, Lifting the Veil: A New Study of the Sheela-na-gigs of Britain and Ireland. Unpublished PhD thesis, University of Southampton PUGH, R. D., 1975, ‘The borough of Devizes’, in E. Crittall (ed.), The Victoria County History of England: Wiltshire, Volume 10, 225-314. Oxford: Oxford University Press for the Institute of Historical Research STALLEY, R. A., 1970, Wooden corbel heads at Devizes Castle, Wiltshire. Archaeological Journal 127, 228-229 wonen nn nnnnn=e 1971, A twelfth-century patron of architecture: A study of the buildings erected by Roger, Bishop of Salisbury 1102-1139. Fournal of the British Archaeological Association 3" series 34, 62-83 STONE, E. H., 1919, The early Norman castle at Devizes. WANHM 40, 417-429 WEIR, A. and JERMAN, J., 1986, Images of Lust: Sexual Carvings on Medieval Churches. London: Batsford WOODCOCK, A., 2002, A sheela-na-gig at Worth Matravers. Proceedings of the Dorset Natural History and Archaeological Society 124, 110-111 we eeeennnneee- 2005, Liminal Images: Aspects of Medieval Architectural Sculpture in the South of England from the Eleventh to the Sixteenth Centuries, British Archaeological Reports (British Series) 386, Oxford: John and Erica Hedges South Wiltshire Young Archaeologists’ Club excavation at Baverstock by Susan Clelland and Bill Snell Thanks to funding from the Council for British Archaeology and support from Wessex Archaeology, members of South Wiltshire Young Archaeologists’ Club (SWYAC) were able to get a taste of real archaeology during the summers of 2003 and 2004 (Figure 1). Using information provided by the National Monuments Records Centre, Swindon in a ‘Local Studies Resource Pack’ and research carried out at the Wiltshire and Swindon Record Office, Trowbridge, YAC members found documentary evidence of a possible 18th century drovers’ inn on farmland adjacent to a drovers’ lane at Baverstock, Dinton, Wiltshire. Permission was kindly granted by Mr Stephen Combes of Manor Farm for a supervised excavation on his land to evaluate the potential survival of the New Inn and any possible ancillary structures on either side of the extant Ox Drove. Andrews’ and Dury’s map of 1773 clearly shows the location of a number of structures attributed to the New Inn, approximately one third of a mile west of the eight mile (from Sarum) milestone on the Ox Drove (Figure 2). The Ox Drove is well documented Wessex Archaeology, Portway House, Old Sarum Park, Salisbury SP4 6EB NOTES AND SHORTER CONTRIBUTIONS Fig. 1. Members of the Wiltshire YAC taking part in excavations on the site of the New Inn at Baverstock. Photo Wessex Archaeology in K.G Watts’ excellent 1990 study Droving in Wiltshire: The Trade and its Routes (Wiltshire Life Society) which, in part, acted as the inspiration for the project. No archaeological remains were found in the trenches along the northern side of the Ox Drove, but two heavily truncated wall foundations were identified in trenches to the south. An east-west aligned flint and cob wall with a possible north— south return at its western extent was recorded. The wall, 0.8m in width and surviving to a height of approximately 0.15m, was butted on both sides by a rammed chalk surface. A further north-south aligned truncated wall foundation was recorded 9.6m to the west. It was of a similar flint and cob construction with occasional greensand and limestone slabs. A rammed chalk surface was recorded against its eastern side. A number of sub-square, heavily truncated postholes or postpads and a circular posthole were identified to the west of the wall. Working on the assumption that the buildings illustrated to the north of the Ox Drove had been dismantled prior to changes in field usage, the 2004 excavation concentrated attenuon on the south side of the track. A further four trenches were opened to try to find further structural remains. The westernmost of these trenches was located at the corner of the Ox Drove and the north-south track shown on the 1773 map. The remains of a wall were found, comprised of roughly shaped flint, 255 closely packed within a friable chalk and lime mortar and interspersed with occasional large greensand blocks. It was thought that this surviving wall footing represented the north-western corner of a building frontage and the layer of local sub-angular stone, lain roughly on the inside edge of the wall was the remains of an associated surface. Rammed chalk formed a useable clean frontage to the structure. A flint and cob wall was found 15m further east, running parallel to the Ox Drove. To the south of this wall, a remnant slab surface apparently formed an associated internal floor. An extensive cobble area was uncovered towards i he 4 { ¥ . ¢ Fig. 2. Andrews’ and Dury’s 1773 map showing The New Inn, Baverstock the mid-point of the site, showing several phases of use. In some areas where the original cobble surface had become worn, further cobbles or re-used brick and tile had been overlain as part of ongoing maintenance and repair. A north-south aligned linear depression was recorded on the line of the north-south wall identified in 2003. This is thought to be the remains ofa robbed-out wall foundation, and would make the structure 15m wide. It seems likely that the Young Archaeologists have helped to uncover the remains of the western- most structure illustrated on the 1773 map tentauvely identified as the main building of the New Jnn complex, along with traces of the western edge of the second building. 256 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Were House, Warminster: Britain’s oldest purpose- built fire station? by Michael Heaton Description Were House was situated behind and south of the east end of George Street in Warminster, within a triangular plot adjoining the west side of the culverted River Were, centred on NGR 387238 145096. The ‘house’ comprised a rectangular, two-storey building, 14.75m x 9.65m, divided into two unequal cells either side of a 3.10m wide central corridor. The solid brick walls were of hand-made stock bricks in Flemish bond with narrow (<10mm) joints, overhand pointed and filled with a white sand/ lime mortar. The central bay of the main, eastern, elevation was broken forward slightly to create a pseudo-Classical portico and behind it, within the building, the central corridor was spanned by a steel-braced segmental brickwork arch supporting a hearth and stack on the first floor. There were no hearths serving the ground floor. The solid ground floor of stable ‘setts’ lay at external ground level; the first floor was of band-sawn softwood beams supporting joists and tongue and groove boards, fully ceiled with lath and plaster. The shallow-pitched, hipped and valieyed roof of band-sawn, softwood, kingpost trusses was covered with metamorphic slates. The main, eastern elevation, was pierced by three cart-width door openings on the ground floor, the southernmost spanned by a pulvinated timber frieze. The only pedestrian door opening served the first floor via an external timber staircase running up the north elevation to the north-west corner. There were no ground floor window openings in the main (east) or north elevations and no window openings in the main elevation at all: all original window openings in that wall having been formed blind. The first floor was lit, originally, from the north only. The internal walls were, or had been, finished in softwood beaded plank panelling to entablement level, on which were visible the shadows of saddle rests and coat hooks. The panelling sat on a Bath stone skirting and was surmounted by a rendered entablement in three-coat lime work, lined-out to replicate ashlar work with a simple ‘run’ cornice. The internal door openings were framed in ‘carp’s tongue’ softwood architraves and iron ‘eye’ rings were fixed at head-height to the internal walls of the central corridor. The northern cell had been fitted out with six stable stall partitions formed by cast iron ‘swan neck’ rails let into the brickwork, surmounted by decorative brackets formed in the entablement plaster. The north wall and the internal walls at skirt level were pierced by monolithic wall vents of Bath stone, comprising six radial vents cut into a square block. History The site was held from the Longleat estate by the parish in 1780 (WRO: G16/995/25), but a survey drawing of that date is too worn for a building to be discerned. The exact outline, however, was recorded on the 1838 Public Health Survey of Warminster (WRO. 628/20) within boundaries unchanged from the 1780 Longleat survey of the town. The survey used the same Schedule reference numbers as the 1780 survey, the reference number ‘815’ straddling the boundary between this site and the elongated plot behind (i.e. west of) it, indicating both are held by the same person, in this case a ‘Field and buildings’ owned and occupied by one George Curtis. The 1886 First Edition of the Ordnance Survey 25" scale survey records the outlines of the site and the principal buildings within it exactly as they were in 1838, with the addition of walled enclosures and penticed outshot structures against the north end (toned differently). By the time of the 1901 revision of the First Edition the building was annotated ‘Fire Engine House’. A parish-based fire service existed in 1867 when Furlong House, 61 East Street, Warminster, Wiltshire BA12 9BZ NOTES AND SHORTER CONTRIBUTIONS v uy Cornice Bench, coat hooks and saddle rests Panelled Stable stalls 4) \ V/ Floor level vents Broken forward t Inserted? Inserted Ground floor plan Inserted Pulvinated timber frieze NY Inserted Broken forward East 01 5m the Board of Health, formed in 1867, reported that the town’s three fire engines, which were as satisfactory as ‘their age would allow’, did not need to be replaced. (Crittall, 1955). Where the fire engines were housed in 1867 is not apparent, but by 1882 acommittee formed to establish a volunteer fire brigade (WRO: G16/760/85) was lobbying the Board of Magistrates for funding to build a ‘new fire station’ and was inviting designs from local architects. Sites in Weymouth Street and The Close were proposed and defeated on numerous occasions between 1882 and 1905. Nonetheless, the Warminster Voluntary Fire Brigade was established by deed on 14 February 1887 (WRO: G16/210/1-3) and its I I ! | Bees SER lee eee ee 257 Vent | 4} Vent p E (Were House fe Ws +t eee 2 ee. Ww , ae 4 Vent i ieee | ¥ oa Vent Inserted Wall vents North Were House, Warminster Committee resolved on 16'" August 1887 to provide a ‘proper engine house’. The resolution was aired again on 26'" October 1887, when it was declined by the Board, and again on 8" August 1901 when it was considered that ‘the ime had come when a more convenient and suitable Fire Station than the one in George Street should be provided’. As late as November 1904, a site in Weymouth Street was sull being considered for combined civic and fire station purposes. In 1905, new facilities were built in The Close (presently the St John Ambulance HQ) and the George Street premises, it is believed, became a UDC depot shortly after 258 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Discussion There was clearly a fire service in Warminster before 1867 and stationed at Were House by 1882. The building, though decorated, is not of domestic origin. It has no internal staircase, no ground floor heating, virtually no services other than the 20% century electrics and no alternative (rear) ground floor access. Furthermore, the blind windows in the main elevation indicate that it was never intended to be inhabited, whilst the wall vents show that problems of air circulation were anticipated in its design. As late as 1886, the River Were was, to all intents and purposes, the town sewer (Critall, 1955): It is highly unlikely that a private dwelling of this size would have been sited next to it, but it would provide a convenient source of — albeit intermittently foetid - water. These features, together with the external first floor staircase, suggest it was designed specifically for non-domestic and non-commercial use. The architectural pretension of the pseudo- portico, blind windows and the decorative internal details suggests a civic function, whilst the stables and the cloak room indicate it was designed to house horses and a company of men. There are few building types which meet these criteria, although a Fire Station is one of them. Were House was built before 1838, while brickwork became widely used in the town from c. 1807 onwards, mainly for domestic and commercial buildings, with the larger civic buildings (cf. Law Courts and Literary Institute of 1838) being built in Bath stone ashlars. The George Street terrace (1807) and the terraces at the Portway/High Street/ George Street junction (1825) are of well-executed hand-made brickwork with gauged arches similar to Were House. From the mid-19" century onwards in Warminster, as elsewhere, machine-made bricks become the material of choice for commercial and other non-domestic buildings. Fire Stations, as subjects of architectural and historical study, are a chiefly 20" century phenomenon, inextricably linked with the development of local government. Voluntary parish brigades and insurance company brigades operated from the 18" century onwards, sometimes in tandem. The first public brigade was established _ in Edinburgh in 1824, followed shortly by the London Fire Brigade. Both of these comprised large bodies of men and equipment serving major cities, but operating from ad hoc facilities. Purpose-built fire stations began to appear from the middle of the 19 century along with the other trappings of local government (public sewers, water supplies etc,) the majority being of late 19" or early 20" century date. Until 1913, all fire engines were pulled by horses and all fire stations essentially resembled stables; i.e. long thin buildings with hay lofts over and usually, separate office buildings, but always opening onto principal thoroughfares. There are 38 ‘Listed’ fire stations in England (English Heritage, 2004). All but nine are of 20" century construction and of those, five are situated in London or other cities. Of the remaining four, two are in the south-west and only one is in Wiltshire —at Calne. That building is a simple single-bayed building of wholly vernacular agricultural style, dated to the late 19" century, which may simply bea converted agricultural building. As a purpose-built, non-vernacular, fire station of mid-19" century date serving a provincial non-borough market town, Were House was apparently unique. References Original sources in the Wiltshire and Swindon Record Office (WSRO) WSRO 628/20: Cruse Survey of Warminster, 1838. WSRO G16/995/25: Longleat Survey of Warminster, 1780. WSRO G16/210: Minute books of the Committee of the Warminster Voluntary Fire Brigade 1886 — 1928. WRO G16/760/85: Design for a new Fire Engine House at Weymouth Street, 1882. WRO G16/70/18: Design for a new Fire Engine House at The Close, 1905 Published sources - Crittall, E., 1965, Warminster, in Victoria History of Wiltshire, Volume 9, London English Heritage, 2004: www.imagesofengland.org.uk NOTES AND SHORTER CONTRIBUTIONS 259 The International Bee Research Association (IBRA) Bee Boles Register online by Penelope Walker Until the introduction of modern hives in the late 19th century, beekeepers in Britain and Ireland kept their bees in traditional hives (skeps) made from wicker and, later, coiled-straw. Before sugar was introduced, the honey was much valued as a sweetener and was also used to make mead; the beeswax was very important for making candles for the church. Most skeps of bees were kept in the open, but some beekeepers built special structures to protect the skeps from the weather. The commonest type of structure surviving in the UK is a wall containing a row of recesses (bee boles), but other types have also been recorded: alcoves, bee shelters, bee houses and winter storage buildings. The bee bole walls in Wiltshire were described in an article by A. M. Foster in this magazine in 1986. Recordings and photographs of these and all other structures known in England, Ireland, Scotland and Wales are held in the IBRA Register. It contains 1370 records which are now accessible online in the new IBRA Bee Boles Register database at: http://www.ibra.org.uk/beeboles. Searching is easy, and selected records and images can be viewed. In addition, a list of relevant publications is provided. The site will be of special value to local 33 Kingsway, Gerrards Cross, Bucks SL9 8NX historians who are interested in our beekeeping heritage, in gardens and in vernacular buildings. It could also be useful for school or college projects. As the voluntary Curator of the IBRA Register, I organized the conversion of the paper record forms and photographs into this online database. Its main purposes are to make the information and images easily accessible to anyone interested, and to encourage conservation of the structures as well as to achieve further recording. New records are welcome, and the database will be updated regularly. For contact details, see the website, or tel. 029 2037 2409 (IBRA). Financial contributions towards the cost of the project were received from: Awards for All (to the English Bee Boles Society); Cadw, Cardiff; Historic Scotland, Edinburgh; and the Eva Crane Trust. Note. In December 2006, two demonstration bee boles are being built by members of the Dry Stone Walling Association at the Bath and West Showground. They can be visited by the public at any time (normally free, but on show days, there is an entry charge to the ground). Reference Foster, A.M. 1986. Bee boles in Wiltshire. WANHM 80, 176-183 The Archaeology Field Group: recent activities and future plans by Fim Gunter and Susie Stidolph Through the efforts of Wendy Smith, Secretary to the Archaeology Committee, the Archaeology Field Group (AFG) was re-formed in the summer of 2004 with 26 members of the Society. Membership was initially restricted in order to test the viability of the group but by January 2005 sufficient projects 260 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE had developed to widen participation to all members of the Society and numbers have now more than doubled. Emphasis during the last year has been to improve and extend the skill-base of members, and activities now range from desk-top research to excavation. Training sessions on various archaeological techniques, including surveying and fieldwalking, have been held. Besides gaining experience from our own projects, members are encouraged to take part in field activities run by other organisations in the county. Opportunities for field experience for the group’s members have been arranged with the Stonehenge Riverside Project, led by Mike Parker Pearson (Sheffield University), Josh Pollard (Bristol University) and Colin Richards (Manchester University), and the Potterne Medieval Burials Project led by Teresa Hawtin (Sheffield University). Volunteers from the Group helped FOAM (the Friends of Ancient Monuments) with scrub clearance of barrows on Normanton Down, near Stonehenge. Involvement with these projects provides members with access to cutting-edge archaeological techniques, such as 3D laser scanning, rarely available to amateur groups. Members of the AFG recently assisted Bernard Phillips on a rescue excavation near Calne following a request for help by the County Archaeologist. The group intends to continue to meet such demands which provide wide-ranging experience and help to develop a strong and positive reputation for the group in the county. The group also participated in fieldwalking at a large site near Calne. The field in question had not been cultivated for at least 300 years and was expected to produce good results. The project was assisted by Jim Balch, a local metal-detectorist, who provides Katie Hinds, the county’s Finds Liaison Officer, with a constant stream of finds. The results were most impressive. Roman tesserae and samian ware, medieval pottery, and bone objects with incised decoration were found. These latter items were identified by Josh Pollard as Earliest Bronze Age, very similar to finds from the Potterne Bronze Age midden. The Tilshead Project The major project currently being undertaken by the group is at Tilshead on a rectangular bank and ditch enclosure on land belonging to former long- standing WANHS member, Tom Blake, who sadly passed away during 2005. It was anticipated that the enclosure was probably part of a sheep holding area associated with an adjoining trackway to Market Lavington. The work included desk-based research, earthwork and geophysical survey culminating in a series of evaluation trenches dug over the Easter and August long weekends. Initially, trenches across the bank and ditch seemed to confirm a post-medieval date — especially the variety of animals buried in the ditch over recent centuries, including dogs, sheep and a whole horse dated by its shoes to the Victorian period. A second ditch, however, traced inside the first bank, produced a surprising range of artefacts, showing that the vicinity of the ditch at least had been used over a period of some 5000 years. The earliest find was of a sherd identified as Early Neolithic. Fragments of rare Beaker pottery were also found. An entrance way into the enclosure is dated to the Iron Age by the find of a child’s bracelet and the object is currently in the laboratory awaiting conservation. An animal tooth from the enclosure has been radio-carbon dated and appears to be 2,100 years old, with other bone fragments providing dates of 2300BP and 3200 BP. These dates span the Late Bronze Age and Iron Age. Geophysical survey results also hint at circular structures (round houses?) inside the enclosure supporting a Bronze Age or Iron Age date. Three sherds of Roman pottery were also found, including two pieces of samian ware. Other finds included a few worked flints, worked bone fragments, Georgian glass, anda WW 1 military badge, confirming re-use of the site over several millennia. Aerial photographs revealed that the enclosure may be one of a number in and around Tilshead, several of which were seen on the ground during a more recent field survey. Further fieldwork is planned. As we become increasingly active, new equipment is required. The group has been reliant so far on equipment owned by individual members or that which has been donated or borrowed from elsewhere. It is intended in the near future to buy resistivity equipment to broaden the group’s capabilities encouraging greater numbers of WANHS members to participate in fieldwork. The support of the Society’s members is greatly appreciated in these exciting new developments and any donations will be gratefully received and put to good use. Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 261-263 Obituaries Derek Douglas Alexander Simpson (1938-2005) Derek Simpson was born in Glasgow and attended Perth Academy and Carlisle School before entering Edinburgh University to study Archaeology under Stuart Piggott and Richard Atkinson in 1955. The enthusiasm and knowledge of the former ‘master’ in particular had a great influence on Derek — an influence which he would later pass on in turn to his own students — and he and Stuart remained good friends for the rest of Stuart’s life. Derek’s training excavations included some exceptional sites — Stonehenge, Wayland’s Smithy and West Kennet —and at the former site he often recalled finding a bottle containing the cremated remains of an Arch- druid. Fearing some druidical hex, Derek always said that this would be his posthumous publication. It will not be alone. Following his graduation in 1958, Derek had a brief but productive spell at the Museum of this august Society in Devizes with the then curator Ken Annable. The famous Devizes Catalogue emanated from those years and Derek also made a substantial contribution to the Wiltshire vglume of Pevsner’s Buildings of England for which he was rightly acknowledged. His undergraduate dissertation Food Vessels in SW Scotland was prepared for publication during his time in Wiltshire. There were lesser- known episodes at Devizes such as the dropping of a Collared Urn on day one of the new job. ‘Well’ Derek shrugged ‘it’s not as if it was a new one’. I heard this from Derek himself but later learned that a similar fate had befallen a necklace of jet disc beads each of which, unfortunately, was narrower than the gaps between the floor boards. I checked the authenticity of this potential slander with Derek himself ina jovial moment at his home in Co. Down. Derek blushed. The story’s true. Perhaps realising that museum work was not for the quite so ham-fisted, Derek took up a teaching post in the new Dept of Archaeology at Leicester along with Charles and Stanley Thomas, and John Wacher. There, as well as having an influential effect on many undergraduates, many now employed in archaeology, Derek’s fieldwork and publication portfolios grew. There was West Overton with Isobel Smith (alas also passed away this year), Raigmore, Croft Moraig with Stuart Piggott, Pitnacree with John Coles, Kaimes, Skendleby with John Evans (also alas recently passed on), King’s Newnham, Grendon, Seamer and, of course, the Beaker settlement at Northton. The list is not exhaustive. From the Leicester years sprang works such as Studies in Ancient Europe (with John Coles) the first Festschrift for his old friend and teacher Stuart Piggott and Settlement and Economy in Neolithic and Early Bronze Age, Britain and Europe the result of a conference at Leicester following the Northton excavations. There was also, of course /ntroduction to British Prehistory (with Vincent Megaw and others). Derek always referred to this as ‘The Idiots’ Guide: Written for Idiots by Idiots’. But despite this self- deprecation, The Idiots’ Guide went on to numerous reprints and has passed through the hands of thousands of undergraduates. Praise indeed. In 1984 Derek was offered the Chair at Queen’s University, Belfast. There had not always been a great deal of co-operation between British and Irish Archaeologists largely as a result of the less than prefect political situation between the two countries, Derek saw an Ireland of people: no borders, and an Irish colleague from Dublin described Derek’s arrival in Ireland to me as ‘A breath of fresh air to Irish archaeology’. Certainly with the present generation of British and Irish archaeologists, the relationship has never been better. Doubtless this ts partly a result of the improving political climate but I also like to think that Derek played a part in this. He was a great unifier. He was active in fieldwork = Lyles 262 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Hill, Ballygalley, Dun Rhuad, Island McHugh — he sat on numerous national committees in both the North and the Republic, he re-assessed Irish henges with Tom Condit and continued his artefact interests with his study of Irish and British antler maceheads. He served as co-editor of Ulster Fournal of Archaeology and went on to foster a new generation of research students through his post-excavation programmes: Derek made a serious contribution to Irish archaeology. Unfortunately Derek’s health had been deteriorating for a number of years. He retired in 2001 but did not really have the health to enjoy his short retirement. Alison Sheridan and I were privileged enough to edit a Festschrift to Derek. It was a pleasure to see how fondly the contributors remembered Derek and we were inundated with Prof. John G. Evans (1941-2005) The untimely death from cancer of John G. Evans came as a great shock to friends and colleagues. Recently retired as Professor of Archaeology at Cardiff University, John was an internationally respected scholar, renowned for his important research on late Pleistocene and Holocene environmental and vegetational change, particularly on the chalklands of southern England. Following a PhD at the Institute of Archaeology, London, John joined the staff of the Department of Archaeology at Cardiff University (then University College Cardiff). He remained at Cardiff throughout his academic career, building up a vibrant research group in environmental archaeology. John was a pioneer in the scientific application of molluscan analysis (the study of land snail shells assemblages) to detect environmental change; his work at key sites such as Durrington Walls, Marden, Mount Pleasant and Ascott-under-Wychwood in the 1960s and early 1970s doing much to enhance our understanding of late Mesolithic and Neolithic environments in southern Britain. However, he was no mere specialist, nor parochial in his research interests: his written worked ranged widely in time and space, he undertook fieldwork and excavations throughout Britain, and possessed a deep theoretical understanding. His most recent thinking, seen in Environmental Archaeology and the Social Order (2003), offers of papers. Derek was LIKED. The book was , presented to him at a small gathering in Belfast in January this year and he was thrilled to receive it. He was ashy and private man. He had a series of shields and he rarely dropped all of them. Those of us who were fortunate enough to know the man behind the shields knew a warm, and genuine individual with a cracking sense of humour. A true gentleman, a sincere, loyal and dependable friend. Derek died unexpectedly but peacefully at his home in Co. Down 3 days before his 67" birthday. He leaves a widow, Nancy, two children by his first marriage and two step-children. He also leaves many who will fondly remember him. ALEX GIBSON sought to create a new agenda for environmental archaeology, here with the environment as something through which people explored and created social worlds, a focus on the texture of landscapes and social relations. A prolific scholar, his publications included Land Snails in Archaeology (1972), The Environment of Early Man in the British Isles (1975), Land and Archaeology (1999) and Environmental Archaeology (1999, with Terry O’Connor), along with numerous academic papers. Between 1975-1994 he was Assistant Editor then Editor of the Proceedings of the Prehistoric Society. John had a close attachment, through decades of fieldwork, with the Wiltshire chalklands and the Avebury landscape in particular. He worked on Avebury region long barrows (notably South Street with its early 4" millennium BC ploughmarks) with Paul Ashbee and Isobel Smith, at the important and somewhat enigmatic Mesolithic and Neolithic occupation site at Cherhill, and exploring environmental change and sediment sequences in the Kennet Valley. In the late 1980s and early 1990s, Alasdair Whittle’s programme of excavations on Neolithic sites in the Avebury region provided another opportunity for fruitful collaboration. John had insatiable enthusiasm for all things ecological and ancient. I recall undergraduate fieldtrips led at a pace, with unexpected marches across bogs, rock climbing or scrabbling into caves, hunting for snails or flints, digging soil pits, impromptu visits to interesting medieval churches (the appropriate volume of Pevsner always at NOTES AND SHORTER CONTRIBUTIONS hand), and so on. Even ‘conventional’ lectures and practicals could be memorable when led by him, such as an indoor demonstration of flint knapping in which part of Cardiff Archaeology Department’s teaching collection of lithic implements became a source of raw material when flint nodules ran out. There are many similar anecdotes that former students could tell. There is no doubting that he could sometimes be cruel, or display a harsh tongue, was occasionally marginally dangerous (indoor flint knapping a case in point), and often frightening to 263 undergraduates who first met him. Yet despite, or perhaps because of his maverick character, John was incredibly inspiring as a teacher, converting many students to post-graduate study and/or a career in archaeology and environmental science. Because of his genuine interest and support for other people’s work, whether they were students or colleagues, he built up an intensely loyal following. His legacy will live on for a long time to come. JOSHUA POLLARD Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 264-270 Excavation and Fieldwork in Wiltshire 2004 Ashton Keynes East Barn, Church Farm (ST 0418 9424); Medieval A watching brief by Cotswold Archaeological Trust (CAT) was undertaken during groundworks within the moated site of Church Farm. The site was heavily truncated, but an undated stone-lined drain was identified and two sherds of 11th to 13th century pottery were recovered. Bratton Land at Court Lane (SU 391000 152300); Medieval and Post-Medieval Wessex Archaeology (WA) was commissioned by Ashford Homes Ltd. to undertake an archaeological excavation following an earlier evaluation by AC Archaeology suggesting archaeological remains of medieval date. Excavation in advance of housing construction identified a building, probably a barn or similar structure relating to arable cultivation, and an enclosure dating to the 13th and 14th centuries. Although no domestic structures were recorded, rubbish pits suggest occupation nearby, indicating a small farmstead exploiting the surrounding countryside. Activity appears to have declined during the 14th century, although a few pits and a ditch of early post-medieval date were recorded. The site appears to have been terraced, probably in the 18th century for the creation of orchards for cider production. Bulford St Leonard’s Church (SU 416583 143728); Post- Medieval and Modern ASI Heritage Consultants hand-excavated four trenches against the north and south walls of the nave and chancel. The work revealed 20th century drains and underpinning to a depth exceeding 600mm and extending c. 400mm out from the wall face. Human remains and coffin fittings were revealed, in situ, at depths of c. 450mm on the south side of the Chancel, but all other trenches revealed cemetery soils over undisturbed natural ‘C’ horizon subsoils. Calne Blacklands (SU0190 6830); Early Iron Age and Romano-British Bernard Phillips and others carried out excavation and recording during construction of a silage pit. The work revealed linear ditches, walls, cobbling and inhumations relating to a probable Romano- British farmstead dated by finds to the 2nd to 4th century AD. Much well preserved pottery datable to the 6th-7th centuries BC, perhaps midden material, suggest an adjacent or underlying Iron Age settlement. Intervention by Roy Canham, the County Archaeologist, and willing cooperation by the tenant farmer resulted in a large part of the site being preserved. Land adjacent to the Folly Miller, Quemerford (SU 400900 169750); Modern A sub-rectangular plot of approximately 700 square metres, bounded to the east and west by residential properties and to the north by open ground, was subject to an evaluation by WA. This open ground comprises a Scheduled Monument (Wiltshire 28997) and includes the remains of part of a medieval settlement south of Quemerford Farm and either side of a track, the origins of which may be contemporary with the settlement. A single 15m long trench was excavated. A ditch or trench and two post-holes, were encountered. Modern and post-medieval pottery, clay pipe fragments and brick and tile fragments indicate a 19th or 20th century origin for these features which were cut into natural Kimmeridge Shale and sealed below probable buried topsoil, in turn sealed by modern overburden. Despite the close proximity of earthworks, probably representing house platforms and property boundaries, within the field immediately to the north of the site, no medieval features or deposits were identified. EXCAVATION AND FIELDWORK IN WILTSHIRE 2004 235-243 Quemerford (SU 011 697); Post-medieval Oxford Archaeology (OA) carried out a field evaluation on behalf of Midas Homes Ltd, which revealed horticultural and agricultural soils sealing post-medieval features. No further archaeological features or significant deposits were encountered. Castle Eaton Castle Eaton Farm, (SU 1451 9577); Medieval Archaeological evaluation by CAT identified four ditches, three of which were medieval, the other undated. One of the medieval features may represent a trench within which structural posts were set, and may therefore be associated with early settlement of the village. Cherhill/Calstone Oldbury Castle (SU 040 693); Iron Age Oldbury Castle is the largest and probably the most complex of all the Iron Age hillforts on the Marlborough Downs. A request from the National Trust to English Heritage’s Archaeological Survey and Investigation team for analytical earthwork survey in advance of repair works gave an opportunity to add detailed understanding of the earthworks to knowledge previously gained through geophysics and aerial photography. The first recognisable constructions on the hill, apart from barrows, are linear ditches. A possible example (to the south-east) was newly identified by this survey and appears to form the alignment for the rampart and ditch crossing the hillfort. If this identification is correct this feature must be, as earlier commentators have suggested, a primary element of the fort. The first phase hillfort occupied the angle between two linear ditches joined by a curved, east-facing facade. The hillfort was subsequently ‘developed’ by extending the enclosure to the south-west and by the construction of an outer rampart and ditch (though this is unfinished) on the east and south. Why the hillfort should originally have been constrained by the pre-existing linears is a quesuon of great interest. It is not merely for convenience; a huge amount of labour was invested in creating the earliest ramparts, which could therefore have been constructed anywhere on the hill. The implication is that the linear ditches had significance for the hillfort builders. This is true of other hillforts laid out over the junctions of linears, such as Quarley Hill (Hampshire). It seems as if hilltops where linear ditches met were significant places in the later Bronze Age and that this significance was 265 encapsulated by enclosure in the early part of the Iron Age, although the linears themselves lost their importance. The original builders went to great trouble to provide a massive regular facade on the eastern side of the fort with an elaborate central entrance. Principal entrances facing east are a common feature of southern British hillforts. An impressive approach and an element of symbolism were at least as important as considerations of defence. Cricklade Blackwell’s Garage, Abingdon Court Lane (SU 1018 9378); 2Anglo-Saxon and Medieval Archaeological excavation by CAT revealed a series of medieval ditches. Their alignment and date by which they went out of use may suggest that they relate to a former street plan of Cricklade dating to the earlier medieval or possibly middle Anglo- Saxon periods. Land to the rear of 10 Calcutt Street (SU 1018 9362); Medieval and Post-Medieval Archaeological excavation by CAT revealed medieval rubbish pits within the backlands of a property fronting onto Calcutt Street. These were sealed by a thick accumulation of post-medieval dumped deposits probably resulting from land improvement in advance of building work. Devizes Land at St Fohn’s Court (SU 0042 6127); Post- Medieval Archaeological evaluation by CAT identified two pits and a further feature of unknown function. Ali three features date to the 17th to 18th centuries and were sealed by substantial deposits associated with 19th to 20th-century landscaping. St Fohn The Baptist, Long Street (SU 0050 6120); Post-Medieval Observations by ASI Heritage Consultants during groundworks occasioned by refurbishment of the heating system within the nave of the church revealed the upper 300mm-450mm of the sub-floor deposits intensively disturbed by construction of a existing heating system and pew platforms in the late 19th century. Deposits below had also been intensively disturbed by multiple grave excavation, but individual vaults and the foundations of earlier seating plans survive. 266 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Draycot Cerne St Fames Church (ST 935 787); Medieval and Post- Medieval OA carried out an archaeological watching brief commissioned by Andrew Townsend Architects on behalf of the Churches Conservation Trust in advance of repair works to the roof and pews. Monitoring revealed 18th and 19th century works to support pews and several earlier features in the wall elevations pre- dating 14th and 19th century alterations. Durrington Durrington Manor (SU 415600 144900); Prehistoric, Tron Age, Romano-British, Medieval and Post- Medieval ASI Heritage Consultants undertook recording during groundworks associated with construction of new dwellings within the former curtilage of Durrington Manor. The works revealed chalk quarry pits and garden features of late 18th to 19th century date containing a wide range of residual prehistoric, Romano-British and medieval artefacts, indicating the presence nearby of significant archaeological deposits of late Iron Age to Romano-British date. Hanging Langford Yew Tree Cottage (SU 402963 137039); Modern Two trenches machine-excavated by ASI Heritage Consultants within the footprint of a proposed residential development revealed agricultural soils in the southern half of the site and extensive disturbance across the northern half of the site caused by chalk quarrying in the late 19th or early 20th century. Hilperton Land west of Hilperton (ST 867 585); Prehistoric, Romano-British, Medieval and Post-Medieval OA carried out a watching brief and field evaluation for John Samuels Archaeological Consultants on behalf of RPS Group Plc. Fieldwork revealed a scatter of probably imported flint across the length of the southern half of the site, a Romano-British farmstead in the centre of the site, medieval activity at the northern end and extensive patterns of 19th century land improvement across the whole area. Land north of Paxcroft Farm (ST 88234 59535); Late Neolithic/Bronze Age, Roman and Anglo-Saxon Archaeological evaluation by CAT revealed evidence for a possible Late Neolithic or Early Bronze Age ring ditch, two enclosures of prehistoric or Romano- British date, two possible Anglo-Saxon sunken- feature buildings and a number of undated ditches. The enclosures seem likely to have surrounded small settlements associated with a field system represented by the undated ditches. The sunken- feature buildings appear to represent a later period of settlement also likely to have had associated field systems. Idmiston ‘Byford’, Porton (SU 418875 136625): Post-Medieval ASI Heritage Consultants machine-excavated three trenches across the footprint of a chapel recorded by the Tithes Survey of 1841 within the curtilage of the house known as ‘Byford’. The work revealed natural soils sealed beneath enhanced garden soils and collapsed cob walling of 18th century date. No deposits indicative of a medieval chapel were revealed. Latton Latton Lands (SU 085 961); Iron Age, medieval, post- medieval OA monitored the last phase of gravel extraction by Cotswold Aggregates to the north-west corner of the quarry and directly south of the old road A419 (the Roman road Ermine Street). This work investigated an area of 2 ha and revealed features including posthole structures, pits and waterholes, animal and human burials and linear boundary ditches. Certain features were early Iron Age, but most were of middle to late Iron Age date. Further evidence for extensive medieval ridge and furrow and post- medieval activity was also recorded. Significant finds included a perfectly preserved carved wooden ring and, from a large waterhole, an impressive quantity of animal bone, middle Iron Age pottery and a saddle quern. Metalworking debris was also recovered from a pit. The remains are consistent with those recorded in previous excavations of 2001 to 2003. Taken as a whole, the site reveals a diverse and changing landscape that includes a wider range of archaeological periods than anticipated. Liddington Great Moor Leaze, Wanborough (SU1960 8275); Medieval Following machine re-cutting of a drainage ditch, Bernard Phillips and others recovered numerous EXCAVATION AND FIELDWORK IN WILTSHIRE 2004 medieval sherds and much iron slag associated with stone surfaces. Malmesbury Tower House, Twatley Farm (ST 8986 8718); Modern A photographic survey of Twatley Manor Farm buildings was undertaken by CAT during renovations; Richard K. Morriss produced a subsequent detailed description. The buildings are Grade II Listed, and comprise two separate ranges built c. 1930, probably by architect Septimus Warwick for Herbert Choplin Cox. Most of the buildings were constructed with an inner core of brickwork and an outer face of coursed limestone rubble with dressed but irregular limestone quoins. The roofs, despite their various shapes and forms, are all covered with locally derived stone tiles laid to graded courses. Marlborough 95, London Road (SU 419300 169300); Prehistoric to Roman WA undertook an archaeological evaluation prior to the proposed construction of a new house, garage and associated services and specifically aimed at assessing the extent ofa buried soil containing Late Glacial or Early Mesolithic flint artefacts, which had been identified during a previous archaeological evaluation on a neighbouring property (Coombe End House). One trench, measuring 13m x 2m, was machine excavated to the top of a colluvial deposit, identified in the evaluation at Coombe End House. Several modern features were identified at this level, all of which were cut through this deposit. No deposits similar to those recorded on the neighbouring property were encountered. The colluvium directly overlay deposits of brickearth, which were found to extend across much of the trench. A test-pit at the northern end of the trench revealed a shallow pit of Romano-British date, which may belong to a similar phase of occupation as a pit recorded during the archaeological evaluation of the neighbouring property. Marlborough Mound (SU 18372 68651); Post- medieval WA was commissioned to carry outan archaeological investigation of the spiral path on the Marlborough Mound in the grounds of Marlborough College. The work followed on from a preliminary programme of archaeological investigation of a mid 17th century brick belvedere on the south-east facing slope of 267 the mound and formed part of a programme of work to inform restoration proposals relating to the 17th and 18th century garden features upon the mound. A single trench measuring c. 6m by 1m was hand-excavated across the spiral path, on the same path level as the belvedere previously investigated, and between it and the modern concrete steps to the summit on the south side of the mound. The remnants of the gravel pathway built (or at least first recorded) in 1664 were revealed. The path was bounded on the inner slope by the remains of what appeared to be a post and plank timber revetment. A non-structural cosmetic facade of mortar and flint nodules appeared to survive as facing for the more substantial timber revetment. The path itself comprised loosely compacted flint and stone rubble, with no evidence surviving for metalling or any other form of surfacing. No evidence was recovered along the outer edge of the path for the line of a ‘quickset hedge’ (elsewhere identified as yew) in the form ofa bedding trench or such; it is therefore probable that more recent erosion has removed such evidence. Land at Savernake Campus, St John’s School and Community College, Chopping Knife Lane (ST SU 2046 6887); Post-Medieval Archaeological evaluation by CAT was undertaken in the immediate vicinity of the Scheduled Ancient Monument of Forest Hill, a probable Iron Age hillfort, close to the Roman settlement of Cunetio. The Roman road from Winchester to Cirencester had been conjectured to pass through the site, on the basis of nearby earthwork remains and from aerial photography. In addition, a second Roman road, that from Bath to Speen, had been extrapolated to run immediately north of the site. Trenching revealed no evidence of foundation material, metalling or flanking ditches associated with either Roman road. A post-medieval ditched field boundary was noted, together with extensive terracing and dumping associated with construction of the school playing fields. Proposed A345 Access Road for St Fohn’s School, Granham Hill (SU 186 679); Late Neolithic/Bronze Age Archaeological evaluation by CAT revealed two ditches yielding several worked flints of broadly Late Neolithic/Bronze Age date, and a possible pit base. The ditches conceivably flanked a trackway extending downhill from a putative ring-ditch (Wilts SMR No. 620). 268 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Potterne Eastwell House (SU 39440 15800); Post-Medieval Archaeological investigation by ASI Heritage Consultants of walls and associated surfaces exposed during construction of a temporary car park suggest the remains of a dovecote, probably of 17th century construction, demolished before the end of the 18th century and definitely before 1839. The results indicate, also, that the architectural history of Eastwell House is more complex and spatially extensive than the limited documentary sources suggest. Further groundworks might reasonably be expected to reveal similar and associated deposits. Land west of Little Mill at Mill Lane (SU 399350 158580); Medieval Monitoring of groundworks was undertaken by WA. A short section of wall was recorded within a recently excavated pipe trench, probably representing footings, as the stone appeared to be unmortared. To the south of the pipe trench the ground was observed to rise toward the hill top, upon which lies the site of the residence of the Bishops of Salisbury at Potterne. The position of the wall suggests that it may have formed part of the northern boundary to the residence, perhaps spanning the area to the west where a stream is present, possibility part of a gate or gatehouse structure. The smooth finish to the stone blocks suggests more than the remains of a boundary wall per se. A group of unprovenanced finds, potentially associated with the wall, include limestone roof tiles, one of which contains a peg hole. The limited nature of the observed section of wall does not allow its identification as part of a roofed structure and allows little, other than its actual existence, to be ascertained. Salisbury 52-54 Endless Street (SU 414502 130317); Post- Medieval An evaluation carried out by WA to the rear of the property involved the excavation of a series of test- pits cut through 0.25m of tarmacadam and gravel hoggin. Interleaved layers of garden soils comprising rich dark brown silty clay loam and demolition layers comprising bricks, gravel, chalk, slate, CBM and other general building debris were recorded up to a depth of 1.2m in all test-pits. Natural valley gravels were not encountered. Two brick walls of Victorian date and three possible medieval walls constructed of chalk, flint, limestone and green sandstone were encountered in one test-pit. 20, North Street (SU 413940 130090); Post-medieval WA undertook an evaluation on land currently used as a car park bounded by both residential and commercial properties. The underlying geology of the area is Valley Gravel. One evaluation trench was excavated towards the north of the site and a series of three geo-technical pits were observed. Finds comprised post-medieval and modern pottery, including fragments ofa Victorian jug, undiagnostic Verwood ware and ceramic drain pipe, along with fragments of modern ceramic building material. The area seems to have been a garden until its modern development. The investigation provided useful information concerning the archaeological and historical development of Fisherton Street and Fisherton Anger as it seems to have lain beyond the limits of both settlements. Shrewton Karrick House, Tanners Lane (SU 068 442); Modern OA carried out a field evaluation on behalf of Mr and Mrs Manning that revealed extensive deposits of modern garden soils, an earlier worked soil and alluvial clay and gravels overlying natural clay. No further archaeological features or deposits of significance were encountered. Stanton Fitzwarren Land Adjacent to The Rookery, Church Lane (SU 179 901) OA carried out a field evaluation on behalf of Mr J. Goodheart that revealed no archaeological deposits or features as the site had been heavily truncated by the construction of garden buildings. Stanton Park (SU1731 9004); Romano-British Observation by Bernard Phillips during the replacing of existing fence posts revealed fragments of terracotta roofing and box tile, pieces of Forest of Dean sandstone roofing slabs and building stone on this scheduled villa site. Staverton Land between New Terrace and Marina Drive (ST 8580 6035); Iron Age and Romano-British Further geophysical and evaluation work by CAT followed initial evaluation in 2002. Geophysical survey revealed a square enclosure 40m in width in the northern part of the site with geophysical anomalies indicative of settlement activity within. A probable east/west orientated trackway was identified at the south-eastern corner of the enclosure. The EXCAVATION AND FIELDWORK IN WILTSHIRE 2004 evaluation indicated that the enclosure and trackway dated to the Roman period. A small number of pits and postholes of possible Iron Age date were also identified. Sutton Veny Longbridge Hill (ST ST8910 4095); Iron Age/ Neolithic A cutting by Bernard Phillips into the hillside at the side of a track exposed an apparent crouched inhumation located on a terrace of an ancient field system. Swindon A419 Blunsdon Bypass (SU 139 904); Mesolithic and Bronze Age CAT undertook fieldwalking in the vicinity of a previous evaluation trench that identified prehistoric features and flint flakes. A further 20 fragments of flint of Mesolithic to Bronze Age date were recovered. Holy Rood Church, The Lawns, Old Town (SU1603 8363); Romano-British, Anglo-Saxon, Medieval, Post- Medieval The excavation of a trench for a drainage pipe on the north side of the chancel by Bernard Phillips and Mogs Boon revealed demolition debris and the eastern wall of a former vestry. Artifacts recovered include Romano-British and Anglo-Saxon pottery and medieval fired clay roof tile fragments. Hreod Parkway School (SU1285 8587); Post- Medieval Supervised machine cutting and subsequent excavation by Bernard Phillips and Mogs Boon revealed part of a late 18th to mid-20th century farm complex evidently known as North Leaze. An extensive cobbled yard was bordered on one side by a long stone walled stable block with an internal brick drain. The other side had inset two cast-iron cisterns, one linked to a steel pipe. Kingsdown Crematorium (SU1712 8902); Mesolithic, Neolithic, Bronze Age, Post-Medieval Oservation and salvage during the cutting of an access road and car park by Bernard Phillips and Mogs Boon revealed much Mesolithic, Neolithic and Bronze Age flint working waste and tools, including arrowheads, scrapers, borers, knives and microliths. Post-medieval pottery sherds, demonstrate later farming activity. 269 Land at Groundwell Farm (SU 151 890); Medieval Archaeological evaluation by CAT revealed a large enclosure or boundary ditch that yielded a single sherd of 13"- to 1S5th-century Minety ware. No structural remains were encountered predating the construction of the present Groundwell Farm in the 18th-century. Nanonal Trust Central Office Project, Kemble Drive (ST 1430 8500); Modern CAT undertook a watching brief within the area of the former Great Western Railway Works. Although demolition and landscaping works have entirely removed upstanding structures such as walls and machinery, other structural remains in the form of ground-level features such as floors and below ground remains including drains, machine plinths and wall footings survived. These were all associated with the former Railway Works, 1.e. the Rolling Mills, Central Boiler Station, Points and Crossings Shop and Iron Store, and date to the 19th and 20th- century. Pipers Way (SU 162 825); Iron Age and Romano- British OA carried out a field evaluation on behalf of Dwyer Plc. that revealed little of archaeological significance. The site was heavily affected by landscaping which included both levelling and mounding of the existing topography. In the southern area evidence of late Iron Age/early Romano British field boundaries is almost certainly associated with the previous evidence of Roman activity to the east. Survival of archaeology in this part of the site is probably due to the relative lack of major landscaping work. Triangle Site (SU 1750 8840); Iron Age Archaeological evaluation by CAT revealed ditches, pits and postholes on a raised plateau on the eastern part of the site, and on level ground in the southern part. Most of the dated features belong within the Early to Middle Iron Age period with evidence of domestic activity from nearby settlement identifiable from concentrations of animal bone and pottery. The majority of features, consisting of truncated pits and ditches were undated, although their association with the dated features suggests an earlier lron Age date for these also; the ditches were probably field boundaries. Whitworth Road, Rodbourne Cheney (SU 1417 8716); Romano-Brittsh Wall foundation trenching for a house extension 270 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE revealed three extended W to E inhumations apparently part of a larger cemetery site. Excavation by Bernard Phillips and Mogs Boon recorded two burials of elderly males and a young female buried prone. A few small sherds from the grave fills and overlying soil indicate that the burials are of 3rd to 4th century date. Tilshead Salisbury Plain Training Area, ISO village (SU 401500 145300), Prehistoric-modern WA undertook an archaeological evaluation at the Fighting in Built Up Areas (FIBUA) complex on Copehill Down, Salisbury Plain Training Area. The Army Training Estate has expanded this complex to include an area of low rise buildings. This has been achieved by the use of ISO (International Standards Organization) steel containers placed in a street pattern located to the immediate north- west of the present FIBUA. The site area is 1.03ha. Previous archaeological fieldwork conducted on the FIBUA site in 1987-8 ahead of the original build revealed a low density of features dating to the Neolithic. During the ISO Training village Phase 1 observations no features of archaeological interest were noted. Eighteen machine dug trenches totalling 700m in length were excavated, principally along the centre and behind the existing streets formed between rows of containers and also where further street layouts have been proposed. Clay-with- flint with a few areas of exposed Upper Chalk was encountered. Toward the southern side of the site a large quantity of hardcore had been used to infill a wet boggy area of ground. A small assemblage of artefacts was recovered during machine digging of the trenches from the topsoil and these included a selection of prehistoric flint and one sherd of Romano-British pottery. No archaeological features were recorded and the low concentration of finds suggests there is very low probability of occupation on the site. Warminster Were House, George Street, (ST 872 451); Post- Medieval OA carried out a field evaluation on behalf of Hadfield Associates. A late post-medieval well and building foundations, possibly associated with Were House, were revealed. An organic-rich clay/peat layer was discovered at the base of both evaluation trenches representing an earlier period when the area had been exposed to occasional flooding. The work revealed no evidence relating to the medieval development of Warminster. West Overton George Bridge (SU1329 6844); Post-Medieval Recording of the construction of a replacement bridge span by Bernard Phillips and Mogs Boon revealed a late 18th century brick built inverted siphon incorporated in the bridge structure and an earlier brick fabricated water culvert. Wootton Bassett Land at Park Grounds (SU 0575 8375); 2?Modern Archaeological evaluation by CAT confirmed the presence of an earthwork previously identified on an aerial photograph. A clay bank visible as an earthwork prior to the evaluation was exposed. Although undated, this feature was cut by agricultural plough furrows, the fills of which contained post-medieval pottery; its function and date remain uncertain. Wroughton Barbury Castle (SU1453 7638); Neolithic, Bronze Age The cutting of a shallow ditch to prevent illegal vehicle access onto farmland cut through a crouched inhumation of possible Neolithic/Bronze Age date. Salvage excavation by Bernard Phillips and others also recovered a Neolithic flint scraper possibly from the grave fill. Yatesbury Yatesbury Manor Farm (SU 065 721); Medieval or Post-Medieval OA carried out a field evaluation at Yatesbury Manor Farm, which revealed two parallel shallow ditches corresponding to the locations of two linear cropmarks identified during initial baseline assessment by OA. Both ditches were investigated, although no dating evidence was retrieved. Historic cartographic sources suggest that the ditches are likely to represent former field boundaries or ploughed headlands/lynchets of late medieval or post-medieval date. Wiltshire Archaeological & Natural History Magazine, vol. 99 (2006), pp. 271-283 Index by Philip Aslett NOTE: Wiltshire places are indexed or referenced under civil parish. Page numbers in italics refer to figures. Abergavenny (Monmouthshire), | ABMAP (Animal Bone Metrical Archive Project), 138, 144 abscesses, 194, 200 Abury see Avebury AC archaeology: evaluations, 264; excavations, 79-80 Acanthinula aculeata (snail), 91, 93, 95 Acer spp. (maples), 66 Acer campestre (field maple), 146, 148, 149, 150, 151 Aceraceae (maple family), 66, 146 Adams, M. E., 3 Adams, Maurice, 238 Adams, W. M., 247 Adrian, 168 Aegopinella nitidula (snail), 64 Ailfthryth (945-1000), 249 Aelred of Rievaulx (1110-69), Vita Sancti Edwardi Regis et Confessoris, 249-50 aerial photography, 45-6; Bratton, 198; Figheldean, 53; Great Bedwyn, 244-5; Marlborough, 267; Oldbury Castle, 265; Tilshead, 260; Whitewalls, 170, 173, 185, see also cropmarks Aithelflad (fl. 963), 249 AFG see Archaeology Field Group (AFG) Africa, 232 Agar, Welbore Ellis, 2nd Earl of Normanton (1778-1868), 226 agricultural depressions, 227 agriculture: Iron Age, 100; Romano- British, 10-50, 100, see also cultivation Ailesbury, Marquis of, 238-9 Ailesbury family: archives, 238; arms, 221 Air Ministry, 198 Aitken, M. J., 103 Alcala (Spain), 231 Aldbourne, Upham, 224 Alderbury, roof tile production, 219 alders, 151; charcoal, 68, 146, 148, 149; as fuel, 150 Aldhelm (c. 639-709), 106 Aleppo (Syria), 233 Alexander II, Pope (d. 1073), 207 Alexander III, Pope (d. 1181), 249 Alfred, King (849-99), 106, 198 All Cannings: All Cannings Cross, 74, 76; pottery, 134 Allen, Michael J., 81; note on land snails from Breach Hill, Tilshead, 91-5; note on land snails from Dunch Hill, Tidworth, 64—6; note on linear ditch earthwork at Breach Hill, Tilshead, 98-101; note on radiocarbon dating at Dunch Hill, Tidworth, 70-1 almonds, 157 Alnus spp. (alders), 68 Alnus glutinosa (European alder), 146, 148, 149, 150, 151 Amesbury: Coneybury, 68; Coneybury Henge, 68; Earl's Farm Down, 73, 100; Fargo Road, 98; Fargo Wood, 68; King Barrow Ridge, 68; linear ditches, 73; Normanton Down, 260, see also Stonehenge amphibians, bones, 44-5, 70 amphorae, 21, 46, 173, 186; Dressel 2-4 wine type, 182; Dressel 20 type, 27, 28, 182 Anderson, A. S., 27 Andrew Townsend Architects, 266 Andrews and Dury, map (1773), 254, 255 Andrews, Phil, report on excavations at Dunch Hill, Tidworth, 51-78 Angers (France), 204, 205, 209, 211; St Eutrope’s Church, 209 Anguilla anguilla (eel), 145 animal bone see bone, animal Animal Bone Metrical Archive Project (ABMAP), 138, 144 animal economy, 142-4 animal husbandry, 143 Anisantha sterilis (barren brome), 154, 155 Annable, Frederick Kenneth (1922-2002), 190, 191, 192, 201, 261 Ansty: Arundell Arms, 222; Maypole Inn, 222 Anthemis cotula (stinking chamomile), 155, 6, antlers, worked, 187, 262 Antoninus Pius, 168 Antony, Mark, 168 Aphanes arvensis (parsley piert), 70 apple trees, charcoal, 66, 96, 146 Aquae Sulis, 187, see also Bath Aquifoliaceae (hollies), 146 Arcadius, 187 archaeological damage: off-roading, 165, 173-4, 176; ploughing, 45, 170, 173 Archaeology Field Group (AFG): activities, 259-60; fieldwork, 190-203 archaeomagnetic dating, 81 architecture, Romanesque, 250-4 Argentina, imports from, 227 Argonne (France), 182 aristocracy, landed, 227 Army Training Estate, 270 Arnold-Forster, Frances, 249 Arrhenatherum elatius (false oatgrass/onion couch), 68, 70 arrowheads, 12, 15, 21, 22, 45, 269; Early Bronze Age, 88 artefacts: Romano-British, 266; medieval, 266 Arundell, Henry, 3rd Baron Arundell of Wardour (c. 1606-94), 222 Arundell, John Francis, 16th Lord Arundell of Wardour (1907-44), 222 Arundell, Sir Matthew (fl. 1594), 222 Arundell, Reginald John Richard (1931- ), 2??? Arundell, Sir Thomas (d. 1552), 222 Arundell family, 227; arms, 222 Ascott-under-Wychwood (Oxfordshire), 262 ash, post-medieval, 111 ash trees, 151; charcoal, 66, 67-8, 71, 97, 148, 149; uses, 150 Ashbee, Paul, 262 Ashford Homes Ltd, 264 ashlars, 258 Ashton Keynes: Church Farm, 264; Cleveland Arms, 223; New Inn, 223 ASI Heritage Consultants: excavations, 264, 266; investigations, 265, 268; recording work, 266 Athelstan, King (c. 895-939), 109 Atkinson, Richard John Copland (1920-94), 2, 8, 261 Atriplex spp. (oraches), 154, 155, 157 Australasia: imports from, 227; New Zealand, 38, 195, 196 Avebury, 93; barrows, 262; South Street, 262; West Kennet, 261 Avena spp. (oats), charred grains, 154 Avena fatua (wild oat), 156 Avon Gorge, 150 Avon, River (North), 12, 45, 105, 109, 134; crossings, 165, 171; floodplain, 167, 171, 174; gravels, 136 Avon, River (South), 51, 98 Avon Valley, 39, 45, 170, 187 Awards for All, 259 axes: Mesolithic, 97, 98; Neolithic, 97; stone, 45 Babylon, 233 badges, military, 260 Bagendon (Gloucestershire), 29 Balch, Jim, 260 Baldock (Hertfordshire), 33, 36 Balksbury Camp (Hampshire), 75, 93 Ballygalley (Ireland), 262 banks: Anglo-Saxon, 105; medieval, 115 BaRAS see Bristol and Region Archaeological Services (BaRAS Barford St Martin, Penruddocke Arms, 221, 225 barley, 43; charred, 39, 42, 68, 120, 154, 155, 156; cultivation, 156, 157, 158; grains, 20; hulled, 15, 39, 68, 70, 154; prices, 227 barns, 264 Barrow Hills (Oxfordshire), 45 barrows, 45-6, 265; Neolithic, 100, 199; Bronze Age, 51-3; Early Bronze Age, 73, 74; Middle Bronze Age, 73; bowl, 200; long, 80, 81, 100, 199, 262; round, 45, 51-3 Bath family, arms, 221 Bath, 131, 142, 157, 165; Aguae Sufis, 187; Roman roads, 186, 267 Bath and North East Somerset wv Batheaston; Camerton; Eckweck; Keynsham; Little Solsbury Hill; Pagans Hill Bath and West Showground (Somerset), 59 272 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Bathe, Graham: note on Duke’s Vaunt, Savernake Forest, 246-7; paper on Somerset’s Mansion at The Brails, Great Bedwyn, 236-45 Batheaston (Bath and North East Somerset), 131 Beach, William (1719-90), 222 Beach family, arms, 222 beakers, Romano-British, 17, 27 beans, 154, 156, 157 beast heads, 252, 253 Beck, Edward (fl. 1611), 234 Beckford, William (1709-70), 222 Beckford, William (1759-1844), 222, 224, 227 Beckford family, 227; arms, 222 Bede, Venerable (673-735), 205 Bedwyn History Society, 243 bee boles, 259 beeches, 151, 247; charcoal, 146, 148, 149; uses, 150 beehives: modern, 259; traditional, 259 beeswax, 259 Belfast, 261, 262 belvederes, 267 Benedictines, 106, 206 Benett, John (d. 1852), 222 Benett, Thomas (d. 1754), 222 Benett, Thomas (d. 1797), 222 Benett family, 227; arms, 221, 222-3 Bennett, John (fl. 1700), 222 Bennett, John (d. 1772), 222 Bennett, Thomas (fl. 1623), 222 Bennett, William (fl. 1631), 222 Bennett family, arms, 222 Berkshire: pottery, 29, 134, see also Speen Berwick, Sir John (fl. 1548), 237, 239, 242, 243 Berwick St John: Easton Farm, 224; Ferne, 224, 227; Ferne House, 224; Grove Arms, 224; Lower Bridmore Farm, 224; Manor Farm, 224; pottery, 23; Talbot Inn, 224; Upton Farm, 224 Betts{/Leeds], William (fl. 1583), 233-4 Betula spp. (birches), 150 Betulaceae (birch family), 146 Beverstone Castle (Gloucestershire), 222 Bible, Old Testament, 234 biological remains, 38—45 Birbeck, Vaughan, report on excavations at Breach Hill, Tilshead, 79-103 birch trees, 150 birds, bones, 70, 145, 217 Bira’s Eye View of Malmesbury (c.1646/1648), 110 Bishops Cannings, 253 Bishop’s Cleeve (Gloucestershire), 157 Black Mountains, 5 Black Patch (East Sussex), 75 blackthorn trees, 150, 151, 155, 156; charcoal, 66, 68, 148, 149, 150 blades: Mesolithic, 22, 87, 97, 98, 173; Neolithic, 22; Late Neolithic/Early Bronze Age, 135, 136; Bronze Age, 136; Early Bronze Age, 88, 89; flint, 21 Blaise Castle (Bristol), 134 Blake, Tom, 260 bluestones, 1, 2,8 Boards of Health, 257 Boessneck, J., 138 Boessneck, L., 44 Bolingbroke family, arms, 223 bone: animal, 44-5, 46, 160, 260, 269 (Late Neolithic/Early Bronze Age, 70; Middle Bronze Age, 70; Late Bronze Age, 53, 60, 61, 70, 71, 75-6, 260; Iron Age, 138-45, 260; Roman, 10, 21; Romano-British, 184, 187, 200; Anglo- Saxon, 138, 155; medieval, 105, 138- 45, 216, 217; post-medieval, 138-45; 13th/14th century, 161; analysis, 144; fused, 138; taphonomy, 138-9); burnt, 15, 44, 139; charred, 139; gnawed, 139; human, 43-4, 168 (Early Bronze Age, 15, 45; Middle Bronze Age, 54, 63-4; Romano-British, 190-203; medieval, 109, 220); radiocarbon dating, 20, 45, 46, 190, 195-6, 199, 260; worked, 260, see also cattle bones; pig bones; sheep/goat bones; teeth bone objects, 144; Early Bronze Age, 260; incised, 260 Boon, Mogs, 269, 270 Bore, Roger, Bishop of Sarum see Roger of Salisbury (d. 1139) borers, 269 boundaries, 80-1, 240, 270; Roman, 10, 187, 201; post-medieval, 267; burials near, 201; hundred, 201; parish, 81, 99, 201 boundary markers, 15 Bourne Valley, 74 Bowley, A. W., 223 bowls: late prehistoric, 132, 133; Romano- British, 27, 29, 31, 32 Box: Ashley Manor Estate, 225; Cheney Court, 225; Hazlebury, 241; Hazlebury Manor, 225; Northey Arms, 225 bracelets, Iron Age, 260 Bradford-on-Avon, Budbury, 133 Bradley, R., 91 Braga, Archbishop of, 234 Brahmins, 234 brambles, charcoal, 66, 68 Bratton, 196; Blood Stone, 199; Bratton Camp, 199, 200; Bratton Castle, 198; Bratton House, 200; Church, 200; Court Lane, 264; Danes Ley, 190, 196, 197-9; Deanly Wood, 198; inhumations, 190-203; La Deone Legh, 197; Lockham Close, 198; Lockham Plantation, 198; Longcombe Bottom, 200; Luccombe, 190, 197, 197, 198, 200; Luccombe Bottom, 190, 196, 199, 200; Luccombe Mill, 190, 198; Luccombe Springs, 190, 198; Milbourne/Melbourne, 196; Patcombe Hill, 200; Tithe Map and Apportionment, 198; White Cliff, 200 Brecon Beacons (Powys), 2 Breeze, Andrew, note on Celtic survival at Chicklade, 248 Brent, Samuel, 198 briars, charcoal, 66, 68 brick, 241, 242, 264 brickearth, 267 bricks: 19th century, 258; machine-made, 258 Brickstock, Richard, note on coins from Whitewalls, Easton Grey, 180-1 brickwork, 258; Flemish bond, 256 bridges, 171, 270 Bridport (Dorset), 248 Brillouin index, 93, 95 Bristol, 132, 134, 150, 157; pottery, 135, see also Blaise Castle; Sea Mills Bristol Corporation, 168 Bristol and Region Archaeological Services (BaRAS): excavations, 104-64; watching briefs, 110 Bristol University, 106, 166, 260; School of Geographical Science, 146, 151 British Geological Survey, 7 British-Latin language, 248 Britons, 248 Brittany (France), 204 Brixton Deverill: Pertwood, 248; Upper Pertwood Bushes, 248 Brockley Hill (Hertfordshire), 168 Brocolitia (Northumberland), 248 Brokenborough: Tower House, 267; Twatley Manor Farm, 267 Bromus spp. (bromes), 156 Bronk Ramsey C., 38 bronze objects, 168, 170 brooches: Late Iron Age/Early Romano- British, 34; Romano-British, 17, 21, 33-6; Colchester derivatives, 34; hinged iron type, 33; iron, 17; knee, 34; La Tene type, 33; Nauheim derivatives, 33-4; Polden Hill type, 17, 34; trumpet-headed, 34 Broomhall, D., 170 Brown, Lisa, note on pottery from Malmesbury, 131-5 Brownstone Group, 5 Brudenell-Bruce family, arms, 221 buckles, Late Anglo-Saxon, 137 buckthorns, 150; charcoal, 68 building materials: Romano-British, 173; medieval, 217; 16th century, 241-2; brick, 241, 242, 264; debris, 124, see also bricks; ceramic building materials (CBMs); masonry; tiles buildings: Romano-British, 46, 106, 160, 165, 168, 170, 173, 185, 186, 187; medieval, 111, 130, 161, 213, 217-19, 220, 264; 11th/12th century, 105, 161; 16th century, 236-45; 18th century, 254-5; 19th century, 256-8; 20th century, 267; modern, 269; barns, 264; Listed, 258, 267; sunken-featured, 266; timber, 148, 149, 187, 220, see also churches; conduit houses; roundhouses; villas; walls Bulford, St Leonard’s Church, 264 Bull, G., 138 Bupleurum rotundifolium (thorow-wax), 155,157 Burbage, Wolfhall, 236, 237, 238, 239, 244 Burchill, Rod: note on pottery from Malmesbury, 131-5; note on small finds from Malmesbury, 136-8 Burdett, Sir Francis (1770-1844), 223 Burdett, Sir Francis (1869-1951), 223 Burdett family, arms, 223 Burghal Hidage, 105, 106-9, 161 burh defences, 109, 124, 130, 160-1 burials: animal, 266; bog, 201; crouched, 15, 269, 270; mass, 200; near boundaries, 201; prone, 190-203, 270, see also cemeteries; cremation burials; graves; inhumations Burnham, B. C., 184-5 butchery, 143; medieval, 105; carcass processing, 138, 139, 144; marks, 44, 138; waste, 145, 184 CA see Cotswold Archaeology (CA) Cadw (Wales), 259 Caesar, Gaius Julius (c.100-44 BC), 168 Caffres, 233 Caledonian Orogeny, 4, 5 calendar, Gregorian, 231 Calley, Sir Henry (née Langton) (1914—?), 223 Calley, Joan, 223 Calley, William (fl. 1619), 223 Calley family, 227; arms, 223 Calne, 260; fire stations, 258; Jolly Miller, 264; Lansdowne Arms, 221; Quemerford, 264-5; Quemerford Farm, 264 Calne Without: Blacklands, 264; Bowood, 221-2; Oldbury Castle, 265 Cambrian, sandstones, 1 Cambridgeshire see Whittlesford Cameron, Esther, 136 Camerton (Bath and North East Somerset), 186 Canary Islands (Spain), 231-2 candles, 259 Canham, Roy, 264 Canterbury (Kent), 209, 230 Canute IV, King of Denmark (c. 1043-1086), 210 INDEX Cape Verde Islands, 232 Caprifoliaceae (elders), 66 carbohydrates, in diet, 194 Carboniferous, 4, 5 Cardiff (Wales), 259 Cardiff University, Department of Archaeology, 262, 263 Carex spp. (sedges), 157; seeds, 39 caries, 44 Carlisle School (Cumbria), 261 Carmarthenshire see Craig Ddu; Kidwelly; Taf, River Carpinus spp. (hornbeams), 68 Carrawburgh (Northumberland), 248 Carruthers, Wendy J., 70; note on charred plant remains from Showell Farm, Chippenham, 38-43 cartography, 234 Cary, F., 247 Carychium tridentatum (snail), 64, 94, 158 Case, H., 23 Case, H. J., 90 Castle Eaton, Castle Eaton Farm, 265 castles, Norman, 213, 251 CAT see Cotswold Archaeological Trust (CAT) cathedrals, Norman, 206, 213 cats: bones, 138, 140, 141, 160; introduced to Britain, 140, 141, 160 cattle, teeth, 140, 144 cattle bones, 138; Middle Bronze Age, 70; Late Bronze Age, 70; Iron Age, 140; Roman, 44, 46; Romano-British, 184; Anglo-Saxon, 141-4; medieval, 141-4, 216, 217 cattle-rustling, 248 CBMs see ceramic building materials (CBMs) Cecilioides actcula (burrowing snail), 158 Celtic language: survival, 248; toponyms, Celts, 170 cemeteries, 45, 201; medieval, 219, 220; barrow, 73, see also graves; inhumations Centaurea spp. (knapweeds), 157 Centaurea cyanus (cornflower), 157 Centre for Archaeology (CfA), 165; Recording Manual, 174 Cepaea spp. (snails), 91 ceramic building materials (CBMs): Roman, 33; Romano-British, 173, 181, 182; medieval, 216, 217, 219; post- medieval, 268, see also bricks; ules ceramics see pottery cereal processing waste (CPW), 17, 39, 43 cereals: charred, 39, 68, 69, 120, 146, 149, 151; cultivation, 156, 157-8; prices, 227, see also barley; oats; rye; wheat Cervus elaphus (red deer) see deer CfA see Centre for Archaeology (CfA) Chadd, L., 166 chains, medieval, 136 chalices, 220 chalk quarries, 266 chapels, 219, 266 charcoal, 15, 81, 86, 93, 95-7, 146-51; undated, 123, 130; prehistoric, 96; Mesolithic, 68, 97; Neolithic, 68; Late Neolithic/Early Bronze Age, 45, 66, 67, 68; Beaker, 96, 97; Bronze Age, 68; Early Bronze Age, 83, 98; Middle Bronze Age, 54, 66, 67-8; Middle/Late Bronze Age, 96-7; Late Bronze Age, 57, 66-8; Late Bronze Age/Early Iron Age, 97; Iron Age, 146-8, 151, 156-7; Early Iron Age, 111, 114, 115, 154; Middle Iron Age, 113, 117, 119, 120, 121, 122; Anglo-Saxon, 113, 146, 148-9, 151, 155; medieval, 117, 129, 146, 149, 151, 155, 156; environmental evidence, 150-1; oak, 66, 67-8, 71, 83, 96, 97, 105, 124, 146, 148, 149; radiocarbon dating, 57, 70, 73, 74, 146 Charlesworth, Albany Hawke (1854— 1914), 224 Charlton, J., 220 Charlton (North), Charlton Park, 221 Charlton (South): The Charlton Cat, 226; manor, 226; Poores Arms, 226 charters, 196; borough, 109; Salisbury, 250 Cheltenham (Gloucestershire), 157 Chenopodium album (fat-hen), 155 Cherhill, 262; Oldbury Castle, 265; Yatesbury, 270; Yatesbury Manor Farm, 270 chert, 136 Chichester Cathedral (West Sussex), 252 chicken, bones, 44-5, 145 Chicklade, 224; Celtic survival at, 248 Chideock (Dorset), 248 Chippenham: bypass, 10, 12, 46; Showell Farm, 10-50; Showell Nurseries, 10, 46; soils, 39 Chippenham—Lacock road, 12 Chiseldon: Burderop Park, 223; Calley Arms, 223; Hodson, 223 chisels, 183 Chitterne Brook, 98 Christ, Jesus, 234, 250 Christianity, and sheela-na-gigs, 253-4 Christians, massacres, 233 Church Council of London, 205-6 churches: Norman, 251; medieval, 251-2; dedications, 249-50; excavations, 264; Romanesque, 250-4, see also chapels Churches Conservation Trust, 266 Churchill, John Winston Spencer, 6th Duke of Marlborough (1822-83), 223 Churchill, Victor Albert Francis Charles Spencer, Ist Viscount Churchill of Wychwood (1864-1934), 223 Churchill family, arms, 223 cider production, 264 Cirencester (Gloucestershire), 165; Corinium, 187; cosmetic grinders, 33; pottery, 17, 25, 27, 28-9, 31, 32; Roman roads, 186, 267 cisterns, 269 civil wars (1642-51), 105, 109-10, 125, 131, 161 Clapham, Alfred William (1883-1950), 250-1 Clarendon family, arms, 223 Clausilia bidentata (snail), 64, 65, 91, 93, 159 Clausiliidae (door snails), 99 Clavius, Christopher (1538-1612), 231 clay, fired, 21, 33-6, 63, 149 clay pipes, 264 Clay-with-flints, 51, 79, 81, 83, 270 Claydon Pike (Gloucestershire), 157 clays, 106 clearances: Mesolithic, 71; Neolithic, 71-3; Middle Bronze Age, 73, 99-100 cleats, 183 cleavers (plants), 68, 155, 156 Clelland, Susan, note on excavations at Baverstock, 254—5 Clergy List (1896), 249 Cleveland family, arms, 223 Clough, T. H. McK., 2,7 clover, 155, 156 Clupea harengus (herring), 145 Clyffe Pypard, 230; Bushton, 230; Bushton Manor, 230; Clytte Pypard Manor, 224; Goddard Arms, 221, 224 Cnut (c. 994-1035), 249 Coates, R. A., 248 coats of arms, 221-9 cob, walls, 255, 266 cobbles, 255, 264, 269 Cochlicopa lubrica (snail), 158 cocoa, 233 273 Codrington, Sir C. Bethel (fl. 1843), 224 Codrington family, arms, 224 coed (wood), 248 coediog (wooded), 248 coffins: fittings, 264; timber, 200 coins: Iron Age, 185; Roman, 17, 36, 46, 106, 168, 170, 173, 174, 175, 180-1, 186, 187; medieval, 216, 219; find patterns, 186 Coit Maur (Selwood), 248 colanders, 36 Colchester (Essex), 31, 34, 36 Colerne, Euridge Manor, 33 Coles, John, 261; ed. Studies in Ancient Europe (1968), 261 Collared Urn, 261 Collingbourne Ducis, Snail Down, 71 Collinson, John (c. 1757-93), 168 Cologne (Germany), 182 Colt Hoare, Sir Richard see Hoare, Sir Richard Colt (1758-1838) Combes, Stephen, 254 combs, 187 Compton Chamberlayne, 132, 225; Compton Park, 225 Condit, Tom, 262 conduit houses, 238, 241, 243, 244, 245 conduits, 236, 238, 241, 242, 243-4 Constantine I, 168, 180-1 Cool, H. E. M., note on glass from Showell Farm, Chippenham, 36-8 copper alloy objects: Middle Iron Age, 136, 138; Roman, 33-6 copper alloys, casting waste, 86, 91 corbel-tables, 250-4 corbels, 251, 252 cores: flint, 21, 22, 86, 87, 88-9, 97, 98, see also flintwork Corfe (Dorset), 249 Corinium, 187, see also Cirencester corn driers, 42—3 Cornbrash, 12, 106, 131, 168 Corney, Mark, 106, 168, 170, 173 Cornus spp. (dogwoods), 68, 150 Cornwall, 227 Corsham: Corsham Court, 225, 227; Corsham House, 225; Methuen Arms, 225 Corylaceae (hazel family), 66, 96, 146 Corylus spp. (hazels), 66, 96 Corylus avellana (hazel), 39, 68, 146, 148, 149, 150, 151, 155, 156 Cosheston Group, 1-2, 4-5, 7-8 cosmetic grinders, 33 Cotswold Aggregates, 266 Cotswold Archaeological Trust (CAT), 10, 187; evaluations, 265, 266, 267, 268-9, 270; excavations, 265; fieldwalking, 269; surveys, 267; watching briefs, 264, 269 Cotswold Archaeology (CA): evaluations, 10; excavauions, 10—S0 Cotswolds, 39, 45, 131, 134, 189, 170 Cottam, S., 38 Coulston, Baynton House, 190, 198 Council for British Archacology, 254 Council of Regency, 236 County Down (Ireland), 261, 262 Coutances (France), Abbey of St Sauveur- le-Vicomte, 204 Cox, Herbert Choplin, 267 CPW (cereal processing waste), 17, 39, 43 Craig Ddu (Carmarthenshire), 6, 7 Craig Phadrig (Highland), 160 cramps, 183 Cranborne Chase, 73; Barrow Pleck, 98 Crataegus spp. (thorns), 66, 96, 146, 150 Crawford, Osbert Guy Stanhope 1886-1957), 196 cremation burials; Bronze Age, 68; Middle Bronze Age, 54, 55, 62, 63-4, 66, 67, 70, 73; Middle Late Bronze 274 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Age, 51; Iron Age/Roman, 20, 43-4, 46-7; fuels, 67, 68, 70; radiocarbon dating, 70, 71 Crest Nicholson Properties, 10, 12 Cretaceous, 131-2 Cricklade: Abingdon Court Lane, 265; Blackwell’s Garage, 265; Calcutt Street, 265 Crickley Hill (Gloucestershire), 157 Croft Moraig (Perth and Kinross), 261 cropmarks, 10, 13, 45-6, 80, 98, 170, 171-3, 185 crucibles, 182, 183, 187 Crummy, N., 34 cultivation: cereals, 156, 157—8, 160; and weeds, 158 culverts, 270 Cumbria see Carlisle School Cummins, W. A., 2,7 Cunetio, 267 Cunliffe, Barry (1939- ), 169, 185 Cunnington, William (1813-1906), 2 cups, 31 Curtis, George, 256 Cytlid (Chicklade), 248 Dancy family, 196 Danebury (Hampshire), 160 Danes, 106, 198-9, 205 Dannell, G. B., 25 Danthonia decumbens (heath-grass), 155, 157 Darvill, T. C., 46 Das Agulias (South Africa), 232 dating: archaeomagnetic, 81, see also optically stimulated luminescence (OSL); radiocarbon dating Dauntsey: manor, 225; Peterborough Arms, 225 Davies, Joan, 247; note on Duke’s Vaunt, Savernake Forest, 246-7 Davies, Paul, note on molluscs from Malmesbury, 158-9 Davies, Thomas, 1 Davis, S. J. M., 138 Davis, Thomas (fl. 1611), 234 Davis, Vin, 3 de Dene, Isabella, 197 de Dene family, 196-7 de Deone, Agnes, 197 de Deone, Sarah, 197 de Deone, Walter, 197 Deer, W. A., 3 deer: bones, 70, 145, 184; enclosures, 239 deer parks, 239 Defence Estates, 53, 68 defences: Iron Age, 104-64; Anglo-Saxon, 107, 109, 111, 124, 125, 130, 148, 159, 160-1; medieval, 7/07, 119, 159; 12th century, 105; 12th/13th century, 105; 15th century, 105; 17th century, 105, 110, 131, 161, see also burh defences; earthworks; ramparts dentition see teeth Derbyshire, marble, 1 Despencer-Robertson, James Archibald St George Fitzwarenne (1893-1942), 225 Devizes: capture, 253; Castle, 251, 253; Long Street, 265; St John the Baptist Church, 250-4, 265; St John Court, 265; St Mary’s Church, 252, 253, see also Wiltshire Heritage Museum (WHM) Devizes Catalogue, 261 Devizes Museum see Wiltshire Heritage Museum (WHM) Devon, 241, 242, see also Exeter Cathedral Devonian, sandstones, 1, 4, 5, 8 dewponds, 200 Dialogues of Gregory the Great, 207 diet: Romano-British, 187, 194, 195, 200; Anglo-Saxon, 144; medieval, 144 Dilton Marsh, 196 Dinton: Baverstock, 254-5; Dinton Park, 226; The East End Inn, 226; Manor Farm, 254; The Nadder Arms, 226; New Inn, 254-5; Ox Drove, 254-5; Wyndham Arms, 221, 226 Discus rotundatus (snail), 64, 94, 158 diseases, 232 dishes: Romano-British, 32; medieval, 219; West Country, 219 ditches: undated, 270; Neolithic, 260; Late Neolithic/Early Bronze Age, 10, 46, 61, 266, 267; Bronze Age, 91; Early Bronze Age, 10-50, 86-7; Middle Bronze Age, 54, 55, 60, 61, 64, 65-6, 73, 75, 76; Middle/Late Bronze Age, 51, 79-103; Late Bronze Age, 53, 64, 68, 70, 74, 100, 265; Iron Age, 105, 124, 130, 265, 266, 269; Early Iron Age, 111, 113-15; Early/Middle Iron Age, 105, 125; Middle Iron Age, 125-7; Roman, 10; Early Roman, 15-21; Romano-British, 17-21, 264; medieval, 109, 216-17, 219, 265, 269; modern, 264; linear, 53, 68, 70, 71, 73-4, 76, 86, 91, 97, 265, 266; ring, 10-50, 267, see also gullies; linears; pits; Wessex Linear Ditches Project docks, 39, 155, 157 Dodington Park (Gloucestershire), 224 dogs, 140; bones, 44—5, 260 dogwoods: charcoal, 68; uses, 150 Doinus (potter), 168 Domesday Book, 196 Domitian, 17, 36, 168 donkeys, bones, 138 Dore, J., 29 Dorset: barrows, 45; pottery, 186, see also Bridport; Chideock; Corfe; Gussage All Saints; Maiden Castle; Mount Pleasant; Poole Harbour; Purbeck; Shaftesbury Abbey; Sherborne; Studland; Verwood; Wareham; Woodcuts; Worth Matravers Douai (France), Seminary, 230 dovecotes, 268 downland, 73, 80-1, 94, 95, 97, 100, 101 Downton, 98; pottery, 23 drains: undated, 264; post-medieval, 269; 20th century, 264 drovers, 254-5 drownings, 201 Druids, 261 Dry Stone Walling Association, 259 Dublin, 261 ducks, bones, 145, 184 Dun Lagaidh (Highland), 160 Dun Rhuad (Ireland), 262 Dunham, K. C., 5 Durrington: Durrington Down, 68; Durrington Down Round Barrow, 67; Durrington Manor, 266; Durrington Walls, 62, 98, 262 Dutchmen, 234 Dwyer Plc, 269 Dyfed, 5, 8 dykes, 239 earths see soils earthworks, 239, 270; Middle Bronze Age, 73; Middle/Late Bronze Age, 79-103; Iron Age, 105; Romano-British, 168, 170, 171-3, 185; Anglo-Saxon, 99, 213, see also ditches; enclosures; hillforts; motte and bailey; mounds; ramparts East Indies, 234; Portuguese, 231 East Knoyle: The Benett Arms, 226; Knoyle House, 226; Seymour Arms, 226 East Midlands, 34 East Sussex see Black Patch; Hastings Easton Grey, 165; Fosse Knowl, 165, 167, 168, 173; Foxley Park, 165, 173; geology, 167-8; Linehams, 165, 166, 170-1, 173; Lobley, 165, 171-3; Park Hill, 165, 170, 171; prisoner of war camp, 165; Whitewalls, 165-89; Whitewalls Wood, 165, 166 Easton Lane (Hampshire), 75 Eckweek (Bath and North East Somerset), 157 Edgar, King (c. 942-975), 196, 249 Edinburgh (Scotland), 258, 259 Edinburgh University, 261 Edington, 190, 196, 198, 200; Edington Hill, 200; Luccombe Bottom, 190, 196, 199, 200; Luccombe Springs, 190, 198; Picquet Hill, 198; Tinhead, 199 Edington Cartulary, 196 Edith of Wessex (c. 1029-75), 249 Edward VI, King (1537-53), 236, 238, 242, 246 Edward the Confessor (c.1004/5—1066), 249-50 Edward the Martyr (c. 963-978), 249 eels, bones, 145 Egan, G., 136 elder trees, 151; charcoal, 66, 68, 97; uses, 150 Eleocharis palustris (common spike-rush), 42, 156, 157 Eleocharis uniglumis (slender spike-rush), 156, 157 Elizabeth I, Queen (1533-1603), 231, 236 elms, 151; charcoal, 68; uses, 150 Elsec optical dating system, 103 emmer wheat, 39, 42, 43, 68, 151, 154 Ena obscura (snail), 159 enclosure maps, 198, 237 enclosures, 157, 240; prehistoric, 266; Neolithic, 45; Middle Bronze Age, 73; Late Bronze Age, 51, 74; Iron Age, 260, 265; Romano-British, 170, 185, 187, 266, 268-9; medieval, 213, 216-17, 219, 269; 13th/14th century, 264; bank and ditch, 260; deer, 239, see also hillforts; settlements English Bee Bole Society, 259 English Church Council, 219 English Heritage, 111, 242; Archaeological Survey and Investigation Team, 265; fieldwork, 165-89; surveys, 265, see also Centre for Archaeology (CfA) engravings, 247 Entwistle, Roy, 64, 65, 66, 74, 91 environmental change, 262 Environmental Statements, 10, 12 Eocene, 239, 243 Equitable Life, 227 Ermine Street, 266 Esdaile, G., 106 Essex see Colchester Ethandun, Battle of (878), 198, 199, 201 Ethiopia, 233 Euphrasia spp. (eyebrights), 154, 155, 157 Europe: bog burials, 201; sheela-na-gigs, 251 Eva Crane Trust, 259 Evans, John G. (1941-2005), 64, 261; The Environment of Early Man in the British Isles (1975), 262; Environmental Archaeology (1999), 262; Environmental Archaeology and the Social Order (2003), 262; Land and Archaeology (1999), 262; Land Snails in Archaeology (1972), 262; obituary, 262-3 Eve of Wilton (be. 1060), 204-12 Evesham Associations, 168 Ewelmen, 196, 197 executions, 190, 200, 201, 242 Exeter Cathedral (Devon), 251 Fabiis, Fabio de (fl. 1575), 231 Fagaceae (trees and shrubs), 66, 96, 146 INDEX Fagus sylvatica (Beech), 146, 148, 149, 150, 151 Fallopia convolvulus (black bindweed), 154, 157 families, heraldry, 221-9 Fane-Benett-Stanford, Evelyn, 222-3 Fane-Benett-Stanford, Vere (d. 1894), 223 farming see agriculture farms: 18th/20th century, 269; abandoned, 227 farmsteads, 264; Romano-British, 12, 46, 264, 266 Farquhar, John (1751-1826), 222 Faustina, 168 Felis catus (domestic cat), 140, 141, 160 fence lines, Late Bronze Age, 51, 57, 59-60, 61, 74, 75, 76 Fenwick, Jane, 225 Fenwick, Mordaunt (fl. 1853), 225 Festschrifts, 261, 262 Sfynnon (spring), 248 FIBUA (Fighting in Built Up Areas) complex, 270 field systems, 173, 266, 269; Bronze Age, 73; Middle/Late Bronze Age, 51; Late Iron Age/Early Romano-British, 200, 269; Romano-British, 10-50, 187; post-medieval, 239; Celtic, 53, 76, 170; Lowland Zone Bronze Age, 73; ridge and furrow, 266 fields: Romano-British, 46, 185; boundaries, 270; Celtic, 71, 73-4 fieldwalking, 74, 76, 166, 167, 172, 173, 185, 186 Figheldean: barrows, 51-3; Bourne Bottom, 51; clearances, 73; Dunch Hill, 51-78; Dunch Hill Plantation, 76; Figheldean Down, 73; Robin Hood’s Ball, 68 Fighting in Built Up Areas (FIBUA) complex, 270 figs, 157 fir trees, 247 fire brigades: insurance companies, 258; volunteer, 257, 258 fire engines, 257; horse-drawn, 258 fire stations, 256-8; Listed, 258 firesticks, 150 firewood, 146, 149-50 First World War see World War I fish, 187; bones, 145; eel bones, 145; freshwater, 195; marine, 232, 233 fishing, 145 Fitch, Ralph (c. 1550-1611), 233-4 Fitzwaryn, Nicholas, 197 flagons, Romano-British, 27, 31, 32 flakes, flint, 21, 61, 83, 86, 87-9, 269; Late Neolithic/Early Bronze Age, 124, 135-6, see also flintwork Flanders, Count of, 210 Flanders, 205, 206 flax, 157 Flemish bond, 256 flies, pupae, 151, 155, 156 flint nodules, 86, 263 flints, 95; as building materials, 241; burnt, 53, 57, 60-2, 75, 79, 81, 83, 87; knapped, 83; knapping, 89, 263; scattered, 12, 266; sources, 136; unworked, 183 flintwork, 12, 87, 260; prehistoric, 270; Late Glacial/Early Mesolithic, 267; Mesolithic, 10, 21-2, 45, 81-3, 87, 97-8, 173, 269; Late Mesolithic, 79, 83; Neolithic, 10, 12, 22, 269; Late Neolithic, 22, 45, 88; Late Neolithic/ Early Bronze Age, 22, 45, 61,71, 111, 124, 130, 135-6, 159, 267; Beaker, 88; Bronze Age, 269; Early Bronze Age, 10, 15, 21, 81, 83, 87-8, 98; Middle/Late Bronze Age, 86, 99; Late Bronze Age, 53, 60-2, 75; knapping debris, 83, 88; waste, 83, 86, 269, see also axes; blades; cores; flakes, flint; hammerstones; knives; scrapers; tools flooding, 185 floors: medieval, 129, 149; post-medieval, 111; 13th/14th century, 109; 19th century, 256, 265 Flora Europaea (1964-80), 66, 96, 146 Fludd, Henry (fl. 1611), 234 flying fish, 232 FOAM (Friends of Ancient Monuments), 260 Folk, R. L., 3 folklore, 253 Follett, William (fl. 1920), 224 Fondedile Foundations, 110 Fonthill Gifford, 224; Beckford Arms, 222; cattle rustling, 248; Fonthill Abbey, 224 food: cultivated, 157, 158, 160; imported, 157; marine, 195; waste, 145, see also diet fords, 248 Forest of Dean, 268 Forest Marble, 168 Forestry Commission, 247 forgeries, 109 forts, 105; Roman, 185, 200, see also hillforts Fosse Way, 165, 166, 168, 170, 171, 173; construction, 185; settlements along, 186, 187 Foster, A. M., 259 France, 241; sheela-na-gigs, 251, see also Angers; Argonne; Brittany; Coutances; Douai; La Graufesenque; Lezoux; Lorraine; Moselle; Normandy; St Omer Frangula alnus (alder buckthorn), 150 Fraxinus spp. (ashes), 66 Fraxinus excelsior (ash), 146, 148, 149, 150, 151 Freke, Sir Thomas (fl. 1605), 224 Freke family, 227; arms, 224 Frenchmen, 234, 242 Friends of Ancient Monuments (FOAM), 260 Frome (Somerset), | Froxfield, finger rings, 36 fruit, 157 Fry, Claude B., 224 fuels, plant, 67, 68, 70, 146, 149-S0, 151 Fulford, M. G., 169 Gale, Rowena, 156-7, 160; note on charcoal from Breach Hill, Tilshead, 95-7; note on charcoal from Dunch Hill, Tidworth, 66-8; note on charcoal from Malmesbury, 146-51 Galium aparine (goose grass), 68, 155, 156 Galium palustre (marsh bedstraw), 42 Gallienus, 168 game, 187; poached, 145 gardens, 266; landscaped, 267 Garnett, Henry (1555-1606), 231 gatchouses, 268 gates, 109, 110, 268 geese, bones, 145 geochemistry, 2 geophysics: Old Sarum, 217; Oldbury Castle, 265; Staverton, 268-9; Tilshead, 260 George II, King (1683-1760), 36 Germany see Cologne Gibbs, Gaisford (fl. 1789), 225 Gibson, Alex, obituary by, 261-2 Gifford and Partners, 80 Gilmore, Teresa, note on human remains from Showell Farm, Chippenham, 43-4 Gilmour, B., 36 Glasgow (Scotland), 261 275 glass: Roman, 17, 36-8, 168, 183-4; medieval, 219; Georgian, 260 Glastonbury (Somerset), 36 Gleser, G. C., 194 Gloucester, pottery, 131 Gloucestershire: barrows, 45; pottery, 134, see also Bagendon; Beverstone Castle; Bishop’s Cleeve; Cheltenham; Cirencester; Claydon Pike; Dodington Park; Hillesley; Kingsholm; Lechlade; North Cerney; St Kenelm’s Well; Shorncote; Sudeley Castle; Tetbury; Uley Bury; Wotton-under- Edge Goa (India), 230, 231, 233, 234 goats: bones, 44, see also sheep/goat bones Goddard, Edward Hungerford (1854- 1947), 168 Goddard, Fitzroy Pleydell (d. 1927), 224 Goddard, Horatio Nelson (d. 1900), 224 Goddard, John (fl. 1530), 224 Goddard, Thomas (fl. 1562), 224 Goddard family, arms, 221, 224 Godwin, Earl of Wessex (c. 1001-1053), 249 gold objects, rings, 36 Good Hope, Cape of, 230, 231-3 Goodheart, J., 268 Goscelin of St Bertin (c. 1035-c.1107): Liber Confortatorius (1080-82), 204-12; Life of St Edith of Wilton, 206 Gover, J. E. B., et al, The Place-names of Wiltshire (1939), 196 Gowland, W., 5 Grafton, 238, 244; Brookland’s Furlong, 244; The Heath, 244; Wilton, 237, 238, 240, 241, 243, 244; Wilton Hill, 241; Wilton Water, 244 Gran Canaria (Canary Islands), 232 granaries, 75 Grant, A., 138 grapes, 157 grasses, 155, 156, 157 grasslands, 64, 65, 68, 70, 71, 73, 94, 95, 100, 158-9 Gravelly Guy (Oxfordshire), 45 gravels, extraction, 266 graves, 265; Early Roman, 15, 20; Romano-Briush, 15; medieval, 213, 217, 219, 220; mass, 220, see also cemeteries; inhumations graveyards see cemeteries Great Bedwyn: Bedwyn Brail, 236-45; Bedwyn Dyke, 239; Castle Copse, 187; Cow Down, 239; Crofton, 239; Dodsdown, 238, 239, 240, 241, 242, 244; Dodsdown Brickworks, 239; Harding Farm, 239; Hill Barn, 236, 239, 243, 244; Seymour's Mansion, 236-45; Tidcombe Gate, 238, 241; Wilton Brail, 238, 239; Wilton Common, 239 Great Oolite Series, 106, 167-8 Great War see World War I Great Western Railway (GWR): construction, 105; Works, 269 Greensand, 90, 134, 191; slabs, 255 Gregory XIII, Pope (1502-85), 231 Grendon (Northamptonshire), 261 Grey, Lady Jane (1537-54), 236 Griffiths, N., 170 Grinsell, Leslie V. (1907-95), 46, 168 Grossman, James Daniels, 138 Grosvenor, Richard, 2nd Marquis of Westminster (1795-1869), 224 Grosvenor, Robert, 2nd Earl Grosvenor and Ist Marquis of Westminster (1767-1845), 224 Grosvenor family, arms, 224 Grove, Sir Thomas (d. 1897), 227 Grove, Thomas (fl. 1658), 224 Grove, Thomas (fl. 1792), 224 276 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Grove, Sir Walter John (1852-1932), 224 Grove family, 227; arms, 224 gruglyd (heathy/heathery), 248 Grundy, G. B., 196 Guarani Indians (South America), 234 Guinea, 232 gullies: Late Neolithic/Early Bronze Age, 10, 46; Bronze Age, 10; Roman, 10; Early Roman, 15; Romano-British, 17, 21; medieval, 216, see also ditches Gunter, Jim, note on the activities of the Archaeology Field Group, 259-60 Guru, Padre, 234 Gussage All Saints (Dorset), 140 Gwent, 5 GWAR see Great Western Railway (GWR) Hadfield Associates, 270 Hadrian’s Wall, 248 Hakluyt, Richard (c. 1552-1616), The English Voyages (1598-1600), 231-3 Hall, S., note on optically stimulated luminescence dating, 103 Hambleton, E., 140 Hamilton, J., 140 hammerstones, Early Bronze Age, 88 Hammon, Andy, note on vertebrate remains from Whitewalls, Easton Grey, 184 Hampshire: pottery, 29, 131-2, 134, 219, see also Balksbury Camp; Danebury; Easton Lane; New Forest; Owslebury; Quarley Hill; Romsey; Silchester; Winchester Hancocks, Annette: note on radiocarbon dating of biological remains, 38; report on excavations at Showell Farm, Chippenham, 10-50 Hannington: Hannington Hall, 224, 227, 228; Manor, 224 Harding, Philip: note on flint finds from Breach Hill, Tilshead, 87-9; note on Mesolithic activity at Breach Hill, Tilshead, 97-8; note on worked and burnt flint from Dunch Hill, Tidworth, 61-2 Harris, James, 1st Earl of Malmesbury (1746-1820), 225 Harris, Richard, 197, 198-9 Harthacnut (c. 1019-42), 249 Haslam, Jeremy, 106 Hastings (East Sussex), 205, 207 Hattatt, R., 34 Hatton, Sir Christopher (1540-91), 231 Hawkins, G. S., 8 Hawley, William, 73 hawthorns, 151; charcoal, 66, 67-8, 96, 97, 146, 148, 149; uses, 150 Hawtin, Teresa, 260 Hay-Drummond, Arthur William Henry (1862-1953), 223 Haynes, Benjamin, 237, 239 hazel trees, 150, 151; charcoal, 66, 67-8, 71, 83, 96, 97, 122, 146, 148, 149; as fuel, 150; uses, 150, 157 hazelnuts, 156; charred, 154, 155; shells, 15, 39, 43, 67, 68, 148 hearths, 87; Late Bronze Age, 70 Heaton, Michael, note on Were House, Warminster, 256-8 Hector and Ascension (ship), 234 Hedera helix (ivy), 149, 150, 151 hedgerows, 150 Helicella itala (snail), 64, 66, 94, 95 Helicellids, introduced, 94 hemp, 157 henges, Irish, 262 Henig, M., 169, 170 Henry II, King (1133-89), 216, 219 Henry VIII (1491-1547), 213, 219, 236 heraldry, 221-9 Heralds Visitation of Wiltshire, 1563, 222 Herbert, Sir William, 1st Earl of Pembroke of the second creation (c. 1501-70), 242 Herbert family, arms, 221 Herefordshire, 5 Hereman, Bishop of Salisbury (d. 1178), 204, 205-6, 207, 208, 210, 211 herring, 232; bones, 145 Hertfordshire see Baldock; Brockley Hill; St Albans Hervey, 209 Heywood: The Ham, 200; Heywood House, 225 Hicks Beach, Sir Michael Edward, 9th baronet, and Ist Earl St. Aldwyn (1837-1916), 227 Hicks Beach, W. A., 222 Hicks Beach family, 222 Hicks, Michael (1760-1830), 222 Higbee, Lorrain, note on animal bone from Showell Farm, Chippenham, 44-5 Highland see Craig Phadrig; Dun Lagaidh; Raigmore Highworth: Freke Arms, 224; Swanborough, 224 Hilary (fl. 1125), 205 Hillesley (Gloucestershire), 157 hillforts: Iron Age, 105, 106, 113, 124, 130, 132-3, 156, 159-61, 187, 199, 213, 265, 267; development, 160; entrances, 265, see also enclosures Hilperton, 266; Paxcroft Farm, 266; Whaddon, 224 Hind, Mr, 241, 242 Hindon, 224; Angel Inn, 224; Grosvenor Arms, 224 Hinds, Katie, 260 Hindu, 233, 234 Hinton, Pat, note on plant remains from Dunch Hill, Tidworth, 68-70 Historic Scotland, 259 hives see beehives hlid (gate/slope), 248 Hoare, Sir Richard Colt (1758-1838), 170, 173, 224, 227; The Ancient History of South and North Wiltshire (1812; 1821), 168, 198; finds, 168-9; History of Modern Wiltshire (1822-44), 197, 198 hollow ways, 53, 73 Holloway, Sir Henry Thomas (1876- 1951), 223 holly trees, 151; charcoal, 68, 146, 149; uses, 150 Holocene, 81, 262 Holt, William (1545-99), 231 Hopkins, Robert, 25 Hordeum spp. (barley), charred grains, 39, 68, 154 Hordeum vulgare (barley), 156 hornbeams, charcoal, 68 horncores, 144 Horningsham: Longleat, 222, 236, 239, 256; Longleat House, 242 horse harnesses, medieval, 136, 137 horses: age estimation, 138; bones, 44, 46, 138, 140, 144-5, 260; in Iron Age, 140; shoes, 260; teeth, 70, 140, 144-5 hospitals: medieval, 213-20; hospitality, 220; leper, 213, 219 hostelries, 221-9 Hounslow see Syon House Howard family, arms, 221 Howie, R. A., 3 Hudson, William (fl. 1611), 234 Huggett, J., 5 hunting, 145; objects, 36; rights, 145 Hurst Park (Surrey), 70 Hussey-Freke, Ambrose Denis, 224 Hussey-Freke family, 228 Hutchins, John, 239 Hyde, Laurence, Earl of Rochester (1641-1711), 223 hypocaust systems, 106, 160, 173, 185, 186 IBRA (International Bee Research Association), 259 ICPMS (inductively coupled plasma mass spectrometry), 103 Idmiston: ‘Byford’, 266; Porton, 266 Ilchester (Somerset), 142 Ilex spp. (holly trees), 68 Tlex aquifolium (holly), 149, 150, 151 Imber, 98; burials, 200; Carpit, 200; Chapel Down, 200 implements see tools Ina [Ine], King (d. 728), 106 India, 230, 232, 233, 234, see also Goa; Mangalore; Rachol; Salcete Peninsula inductively coupled plasma mass spectrometry (ICPMS), 103 Ingleburne, 106 inhumations: ?Neolithic, 124; 2Neolithic/ Bronze Age, 270; Beaker, 124, 136, 159; Early Bronze Age, 43; Late Bronze Age, 76; Iron Age, 266; Late Iron Age/Early Romano-British, 15, 17, 46; Iron Age/Roman, 20, 43-4; Romano-British, 190-203, 264, 270; Anglo-Saxon, 109, 200; medieval, 109; post-medieval, 264, see also cemeteries; graves inns, 221-9; drovers’, 254-5 Institute of Archaeology, 262 insurance company fire brigades, 258 International Bee Research Association (IBRA), 259; Bee Boles Register, 259 International Standards Organization (ISO), 270 Internet, bee bole register online, 259 Iran see Ormuz Ireland: archaeologists, 261-2; bee boles, 259; henges, 262; sheela-na-gigs, 251, 253, see also Ballygalley; Belfast; County Down; Dublin; Dun Rhuad; Island McHugh; Lyles Hill iron, sources of, 200 iron objects: Roman, 170, 183; Late Anglo-Saxon, 137; medieval, 136, 216 Island McHugh (Ireland), 262 ISO (International Standards Organization), 270 Italy, 231, see also Rome ivy: charcoal, 149; as fuel, 150; uses, 150, 151 Ixer, Rob A., paper on petrography of non- sarsen sandstones from Stonehenge, 1-9 Jackson, John Edward (1805-91), 236, 237-8, 239, 243 Jackson, R., 33 Jacomet, S., 151 Japan, 231 jars, 134; late prehistoric, 132, 133; Late Bronze Age, 57; Iron Age, 134; Romano-British, 29, 31-2; 12th/13th century, 219 Jarvis, Thomas, 198 Jennings, Sarah, note on glass from Whitewalls, Easton Grey, 183-4 Jerusalem, 249 Jesuits, 230-5 jet objects, 261 Jewell, John (1522-71), 230 jewellery see brooches; gold objects; rings Jews, 253 John, King (c. 1167-1216), 109 John Samuels Archaeological Consultants, 266 Jones, Elizabeth Langham, 223 Jones, Julie, 146; note on charred plant remains from Malmesbury, 151-8 Jordan, W. K., 238 INDEX Judd, J. W., 5 jugs, 134; medieval, 182; 13th century, 219; 13th/14th century, 134; 19th century, 268 Juncus spp. (rushes), 157 Jurassic, 132; clays, 106; Middle, 131, 168; Upper, 131 Kaimes (City of Edinburgh), 261 Keevil: Beach Arms, 222; Long Arms, 221; Manor, 222 Keevil Airfield, 198 Kellaway, G. A., 2 Kellaways Clay, 12, 46 Kelly’s Directories, 221, 223 Kennet Valley, 262 Kent see Canterbury Kerney, M. P, 158 Kerr, P F, 3 Keynsham (Bath and North East Somerset), 157 Keystone Historic Buildings Consultants, 110 Kidston, George Jardine (1873-1954), 225 Kidwelly (Carmarthenshire), 1, 3, 5 Kiesewalter’s factors, 44 kilns, 239, 242; Romano-British, 29; refuse, 149 Kimmeridge Shale, 264 King’s Newnham (Warwickshire), 261 Kingsholm (Gloucestershire), 29, 38 Kirkham, G., 170 knives, 269; Late Neolithic/Early Bronze Age, 22, 124, 135, 136, 159; Early Bronze Age, 88; Roman, 36; medieval, 136, 137, see also blades Knook, Willis’s Field Barn, 98 Konkani language, 234 Kra, R. S., 71 Kratochvil, Z., 138 La Graufesenque (France), 27 Lacock: Bowden Hill, 243-4; Lacock Abbey, 243-4 lade (water crossing), 248 Lancashire, 248 land divisions, 91, 99, 101, see also boundaries land improvements, 266 land ownership, 227-8 land values, 227 land-use, 100-1 landowners, 227-8 landscapes: prehistoric, 91; Mesolithic, 262; Neolithic, 262; Late Bronze Age, 64, 68, 74, 91; pre-Roman, 173, 185, 186 Lane, John, 231 Langton, Sir Henry (1914-2), 223 Lansdowne family, arms, 221 laser scanning, three-dimensional, 260 Lateran Council, 219 Lathyrus spp. (sweet pea/vetches), 154, 155, 157 Latton, Latton Lands, 266 Latzke, T., 210 lava, 216 Laval, Pyrard de (1578-1621), 234 Laverstock: Ford Road, 213, 217; pottery, 219 Laverton, William Henry (fl. 1888), 226 Lay Subsidy Rolls (1571), 230 lead objects, 241 Lechlade (Gloucestershire), Roughground Farm, 36 Legge, A. J., 138 legumes, 156, 157; charred, 154 Leicester University, 261 Leland, John (c. 1506-52), 109, 213, 219 lepers, 213, 219 leprosy, 219, 220 Letocetum (Wall, Staffordshire), 185 Lezoux (France), 27 Licinius, 168 Liddington, Great Moor Leaze, 266-7 Limacidae (keelback slugs), 94 lime, 241 limes, 148, 149; charcoal, 146, 148; uses, 150, 151 limestone, 117, 168; burnt, 115, 120; masonry, 105, 117, 120, 121, 123-4, 159; oolitic, 106, 114, 124, 135; rubble, 117, 119, 121, 122, 123, 124, 129, 130-1, 160, 267; shelly, 217; slabs, 111, 113, 114, 120, 124, 127, 255 Linear Ditches Project see Wessex Linear Ditches Project linears: Middle/Late Bronze Age, 79-103; Roman, 10; Early Roman, 15, see also ditches Linge, Ann, 136 Linum catharticum (purging flax), 42, 70 Lisbon (Portugal), 231, 234 lithics, 135-6; analysis, 2, 3, see also flintwork Little Solsbury Hill (Bath and North East Somerset), 134 Littleton Drew-Chippenham gas pipeline, Liverpool, 248 Loader, Emma: note on pottery from Breach Hill, Tilshead, 89-91; note on pottery from Dunch Hill, Tidworth, 62-3 local government, 258 Locks Geotechnical, 110 Lolium perenne (rye grass), 42 London, Bishop of, 231 London, 223, 230, 262; British Library, 204; fire stations, 258; Inns of Chancery, 237; St. John of Jerusalem, 237; St Mary le Strand, 236-7; St. Paul’s Cathedral, 237; Somerset House, 236-7, 238; Tower of London, 231; Westminster Abbey, 204, 210, 249-50 London Clay, 239 London Fire Brigade, 258 London-Bristol railway line, 12 Long, Margaret (fl. 1426), 224 Long, Richard (fl. 1799), 224 Long, Richard Gerard, 4th Viscount Long (1929- ), 224 Long, Robert (fl. 1426/9), 224 Long, Sir Robert (fl. 1576), 226 Long, Walter (d. 1672), 224 Long, Walter Francis David, 2nd Viscount Long (1911-44), 224 Long, Walter Hume, Ist Viscount Long of Wraxall (1854-1924), 225, 227-8 Long family, 226; arms, 221, 224-5 Longleat Archives, Seymour Papers, 239 Longman, Timothy, report on excavations at Malmesbury, 104-64 loom weights, 36, 59, 63; Late Iron Age/ Early Romano-British, 106 Lopes, Henry Charles, Ist Baron Ludlow (1828-99), 225 Lopes, Henry Ludlow, 2nd Baron Ludlow (1865-1922), 225 Lopes family, arms, 225 Lorraine (France), 205 Lower Nene Valley, 31, 182 Lower Old Red Sandstone, 2, 5-7 Lower Palaeozoic, 8 Ludlow, Abraham, 225 Ludlow, Henry, 225 Ludlow, Susannah (née Gibbs), 225 Ludlow family, arms, 225 luminescence dating see optically stimulated luminescence (OSL) Lush, John, note on lost dedication of church at Winterbourne Dauntsey, 249-50 277 Lydiard Tregoze: Bolingbroke Arms, 223; Hook, 223; Lydiard Park, 223 Lyles Hill (Ireland), 262 lynchets, 73, 76, 170, 187, 200 Lyneham, 230 maccheads, 98; antler, 262 McKinley, Jacqueline I., note on cremated human bone from Dunch Hill, Tidworth, 63-4 McMahon, P, 165 McSloy, E. R., note on finds from Showell Farm, Chippenham, 21-36 Madagascar, 232 Maddock, Stella, report on prone burial at Bratton, 190-203 Madeira (Portugal), 231 Maeldulbh, 106 magnetometer surveys, 166, 170 Mahon, Pamela, report on prone burial at Bratton, 190-203 Maiden Bradley with Yarnfield: Maiden Bradley House, 226, 227; Priory, 226; Somerset Arms, 226 Maiden Castle (Dorset), 160 Malet, Sir Charles Warre (c. 1753-1815), 225 Malet, Sir Harry Charles (1873-1931), 225 Malet family, arms, 225 Malmesbury family, arms, 225 Malmesbury, 165, 188; Abbey Row, 106, 109; Backhill, 106, 136; Bell Hotel, 109; borough charters, 109; Burnivale, 106, 109; Bynport, 109; capture, 109-10; Castle, 109; Case House, 109; Cat Tower, 115, 119, 125; Cross Hayes Lane, 105, 109; defences, 104-64; East Gate, 106, 109, 110, 160; Gloucester Street, 109; High Street, 109, 110; historical landscapes, 187; Holloway, 109, 110,125; Ingram Street, 109; King’s Wall, 106, 109, 110; Market Cross, 109; Market Place, 109; Nun’s Walk, 106, 111, 115, 119, 159, 161; origin of name, 106; Oxford Street, 106; Postern Gate, 109, 110; Postern Mill, 106, 109, 138; pottery, 25; Roman villa, 47; St Aldhelm’s R. C. Church, 106, 110, 159; St Joseph’s R. C. Primary School, 106, 110, 111, 115, 117, 119, 134, 135, 150; St Paul's Church, 109; Silver Street, 109, 125, 136; South Gate, 106, 109, 110; Suffolk Arms, 221; Town Wall, 134, 135; Westport, 106, 109; Wyniard Gate, 109 Malmesbury Abbey: cloisters, 109; grounds, 106 Malmesbury Library, 106, 109 Malmesbury-Sherston road, 168 Maltby, M., 138 Malus spp. (apples), 66, 96, 146 Malva sylvestris (common mallow), 42, 43 mammals, bones, 145 Manchester University, 260 Mangalore (India), 234 Manning, Mr, 268 Manning, Mrs, 268 Manning, W. H., 36, 183 mansions, Tudor, 236-45 manstos, 165, 185, 187 maple, as building material, 148 maple trees: charcoal, 66, 68, 146, 148, 149; uses, 150, 151 maps: Andrews and Dury (1773), 254, 255; Haynes's (1792), 2 Savernake estate, 247 Marathi language, 234 Marden, 262 Mark, Father, 233 Market Lavington, 260 markets, 143 278 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Marlborough: Ailesbury Arms, 221; Chopping Knife Lane, 267; Coombe End House, 267; Forest Hill, 267; London Road, 267; Mound (Mount), 267; St John’s School and Community College, 267 Marlborough College, Mound (Mount), 267 Marlborough Downs, 73, 88; flints, 136; hillforts, 265; pottery, 90, 91 Mars (god), 169-70 Martin, G., ‘Duke of Somerset’s Tree’, 246-7 martingales, 170, 185 Maskelyne, N. S., 1 masonry: burnt, 105, 149; limestone, 105, 117, 120, 121, 123-4 masons, 241, 242, 252 massacres, Christians, 233 Matilda, Empress (1102-67), 109 Mealdumesburg [Malmesbury], 106 meat, 143, 144, see also butchery meat cleavers, 144 Medicago spp. (medicks), 156 Megaw, J. V. S., ed. Introduction to British Prehistory (1992), 261 Mells (Somerset), 131 Mendips, 134 Mepham, Lorraine: note on copper alloy casting waste from Breach Hill, Tilshead, 91; note on pottery from Breach Hill, Tilshead, 89-91 metal detectors, 170, 260 metal objects, see also bronze objects; copper alloy objects; gold objects; iron objects; lead objects metallurgical residues, 38 metalwork: Roman, 33-6, 38, 170, 183; debris, 266; military, 170, see also blades; bronze objects; copper alloy objects; gold objects; iron objects; knives; lead objects; nails; scrapers; tools Methuen, Sir Paul (1672-1757), 225 Methuen family, 227; arms, 225 Meux, Sir Henry (fl. 1866), 223 mice, bones, 184 microscopy, 62, 68, 96, 146 Midas Homes Ltd, 265 middens: Bronze Age, 260; Late Bronze Age, 53, 61, 70, 71, 74-6; Early Iron Age, 264; medieval, 149, 265 Middle Jurassic, 131, 168 Middlesex, 236 Milbourne, 196 Mildenhall, 29; Cunetio, 267 Miles, C. W. (fl. 1853), 225 Milford Haven (Pembrokeshire), 1, 6, 7, 8 military badges, 260 military structures, slit trenches, 54, 61 Mill Bay (Pembrokeshire), 7 Mill Bay Formation, 7 Millbourne, 196 mills, 196, 198 Milston: Brigmerston Down, 70, 73; Milston Down, 73 mineralogy, 2 Minety, pottery, 182, 269 Minimum Number of Elements (MNE), 138 Minimum Number of Individuals (MNI), 138, 141-2 mints, 109 missions, Jesuit, 230-5 MNE (Minimum Number of Elements), 138 MNI (Minimum Number of Individuals), 138, 141-2 Moffatt, J. A., 106 Moffett, L., 154 moles, bones, 184 molluscs, 86, 99, 100, 158-9; analyses, 262; freshwater, 158, see also snails Mompesson family, arms, 225 monasteries, 106, 206, 207, 209; dissolution, 226, 230 monks, 210; Flemish, 204 Monmouthshire see Abergavenny monuments, funerary, 45-6 Monuments Protection Programme, Water and Sewage Industries, 242 Moody, R., paper on heraldry and hostelries, 221-9 Moore, Ben, note on stratigraphy at Whitewalls, Easton Grey, 174-80 Moore, C. P (fl. 1879), 225 Moorhead, T. S. N., 170, 186 Moors, 233 Mordaunt, Charles, 3rd Earl of Peterborough (1658-1735), 225 Mordaunt, Charles Henry, Sth Earl of Peterborough (1758-1814), 225 Morrison, Sara Antoinette Sibell Frances (née Long) (1934— ), 224 Morriss, Richard K., 267 mortar, 125, 256 mortaria, Roman, 21, 28, 31, 32, 168 mortars: Late Iron Age, 33; end-looped, 33 mosaics, 187 Moselle (France), 182 motte and bailey, Norman, 109 mounds, 267; pillow, 200 Mount Pleasant (Dorset), 262 Mozambique, 232 Mulbourne, 196 Mulebourne, 197 Mullin, David, note on lithics from Malmesbury, 135-6 Musty, J. W. G., 219 Mutual Households Association Ltd, 222 Mutuantonis, 168 Nadder, River, 248 nail cleaners, bronze, 168 nails, 173; Roman, 175, 183, 200; Late Anglo-Saxon, 137; medieval, 136, 216; hobnails, 175, 183 National Monuments Record Centre, 196, 254 National Monuments Record (NMR), 190, 192, 199, 200, 201 National Roman Fabric Reference Collection (NRFRC), 25, 181 National Trust, 265; Central Office Project, 269 necklaces, jet, 261 Nene Valley, pottery, 173 Nesovitrea hammonis (snail), 91 Nettleton: Codrington Arms, 224; West Kington, 224 New Forest (Hampshire), pottery, 173, 182 New Sarum see Salisbury New Zealand, 38, 195, 196 Newbery, John (fl. 1583), 233 Newman Rogers family, 225 Newton Tony: Malet Arms, 225; Wilbury House, 225 Nicholson Estates, 10 Nine Mile River, 51, 74 NISP (Number of Identified Specimens), 138, 141-2 NMR (National Monuments Record), 190, 192, 199, 200, 201 Norman Conquest (1066), 145, 205, 207, 211, 213, 249 Normandy (France), 249 North America, imports from, 227 North Bradley: Long’s Arms, 224; manor, 224-5; Manor Farm, 225 North Cerney (Gloucestershire), Ditches, 33 North Tidworth, 241; Sidbury, 68; Sidbury Hill, 76, 97 North Yorkshire see Seamer Northampton, St Peter’s Church, 251 Northamptonshire see Grendon Northey, Sir Edward (1868-1953), 225 Northey, Edward William (fl. 1873), 227 Northey, G. E. (fl. 1912), 225 Northey, William (fl. 1726), 225 Northey family, arms, 225 Northton (Isle of Harris), 261 Northumberland see Brocolitia; Carrawburgh Norton: Cowage Farm, 187, 188; Foxley, 187, 188 Norton Bavant, 222 Norway, sheela-na-gigs, 251 NRFRC (National Roman Fabric Reference Collection), 25, 181 Number of Identified Specimens (NISP), 138, 141-2 nunneries, 204, 205, 206, 209, 210 nuts, 157 OA see Oxford Archaeology (OA) oak, 150; as building material, 148; burnt, 115, 119; charcoal, 66, 67-8, 71, 83, 96, 97, 105, 124, 146, 148, 149; uses, 151, 160 Oakley, Theresa, note on Romanesque corbel-table at St John the Baptist Church, Devizes, 250-4 oaks, 246-7 oats, 43; charred, 154, 155, 156; cultivation, 156, 157, 158; prices, 227; remains, 68, 70 obituaries, 261-3 O’Connor, Terry P, 144; Environmental Archaeology (1999), 262 Odontites spp. (bartsias), 154, 155, 157 Odstock, Longford Castle, 222 off-roading, archaeological damage, 165, 173-4, 176 Ogbourne St Andrew: Barbury Castle, 270; Dean Bottom, 88, 90 Old Red Sandstone, 1, 4, 8; Lower, 2, 5 Old Sarum, 254; abandonment, 213; cathedral, 213; chapel, 219; Goscelin of St Bertin at, 204-12; hillfort, 213; Hospital of St John the Baptist and St Anthony, 213-20; St John’s Church, 219, 220; Sorviodunum, 213 Oleaceae (olive family), 66, 146 Oliver, Jack, 247 g optically stimulated luminescence (OSL), 79, 81, 99; dating techniques, 103 oraches, 154, 155, 156, 157 orchards, 264 Orcheston: Orcheston Down, 80; Silver Barrow, 80; West Down, 80 Ordnance Survey, 247; Field Investigator, 190, 192, 200; First Editions, 256 Ormuz (Iran), 233 OSL see optically sumulated luminescence (OSL) Osmund, Bishop of Salisbury (c. 1050-99), 204-5, 206, 207, 208-9, 211 Otter, Monika, 204 ovens, 17, 20, 42, 46 Overwey (Tilford, Surrey), 187 Owen, G., 3, 5-6 Owmby Nails Pilot Study, 183 Owslebury (Hampshire), 140 OXCal (computer program), 38 Oxford, 231; New College, 230 Oxford Archaeology (OA): evaluations, 265, 266, 268, 269, 270; watching briefs, 266 Oxford Clay, 167-8 Oxford Dictionary of Saints, The (1997), 249 Oxford University: Corpus Christi College, 231; Research Laboratory for Archaeology and History of Art, 103 Oxfordshire: pottery, 32, 173, 182, INDEX see also Ascott-under-Wychwood; Barrow Hills; Gravelly Guy; Radley; Wayland’s Smithy Oxychilus spp. (snails), 158 oyster shells, 173, 217 Pagans Hill (Bath and North East Somerset), 134 Paget, William, lst Baron Paget of Beaudesert (1505-63), 236 paintings, 246-7; wall, 250 palisades, 149, 150, 151, 159 Parker, Matthew (1504-75), 230 Parker Pearson, Mike, 260 Parliamentarians, 109-10, 125, 131 Parr, Catherine (1512-48), 236 Parsons [Persons], Robert (1546-1610), 231 Passmore, Arthur D. (c.1873-1958), 168, 169, 185, 186 pastoralism, 157 patens, 220 paths: post-medieval, 111, 115, 117; spiral, 267 Patie family, 197 Payne, S., 138 pear trees, charcoal, 66, 96, 146 Pearson, R. H., 198 Pearson, Ronald, 190 peas, 157 Pembroke family, arms, 221 Pembrokeshire, 1, 8, see also Milford Haven; Mill Bay; Preseli Hills penances, 207 Penruddocke, Charles (d. 1930), 225 Penruddocke, Edward (fl. 1598), 225 Penruddocke family, arms, 221, 225 perambulations: 1548, 237, 239; 1575, 196 Persicaria lapathifolia (pale persicaria), 155 Persons, Robert (1546-1610), 231 perth (bush), 248 Perth Academy (Perth and Kinross), 261 Perth and Kinross see Croft Moraig; Pitnacree pestles, 33 Peterborough family, arms, 225 Peterborough, 209 petrography: sandstones, 1-9; total, 2 Petty-Fitzmaurice family, arms, 221-2 Pevsner, Sir Nikolaus (1902-83), 262-3; Buildings of England (1954-71), 261 pews, 266; 19th century, 265 Pewsey, Martinsell, 241 Pewsey, Vale of, 98, 238 Philip III, King of Spain (1578-1621), 234 Phillipps, Bertram Erasmus (fl. 1916), 226 Phillips, Bernard, 260, 264, 266-7, 269, 270 Phillips, Professor, | Phipps, Charles Bathurst Hele (d. 1960), 226 Phipps, Henry, 225-6 Phipps, Nicholas, 225-6 Phipps, Richard Leckonby Hothersall (fl. 1888), 226 Phipps family, arms, 221, 225-6 Phleum spp. (cav’s tails), 154, 155 picks, flint, 98 pig bones, 44—S, 138; Iron Age, 140; Romano-British, 184; Anglo-Saxon, 141-4; medieval, 141-4, 213 Piggott, Stuart (1910-96), 261 pigs: husbandry, 143; teeth, 70 d prea 249 inaceae (pines), 96 ine trees, 247; charcoal, 68, 83, 93, 96, 97 inus spp. (pines), 68, 96, 97 pipelines, 241, see also water pipelines isum spp. (peas), 157 pitchers, medieval, 182 Pitnacree (Perth and Kinross), 261 pits, 87, 101; Late Mesolithic, 79; Early Neolithic, 91; Late Neolithic/Early Bronze Age, 10, 46, 54, 64—5, 66, 70, 71; Middle/Late Bronze Age, 51; Late Bronze Age, 51, 57, 60-1, 74, 75, 98; ?Iron Age, 269; Iron Age, 154, 266, 269; Middle Iron Age, 46, 127, 154; Early Roman, 15; Romano-British, 17, 33, 267; Anglo-Saxon, 138; Late Anglo-Saxon, 111, 128-9, 148, 155, 157; medieval, 130, 131, 154, 156; post-medieval, 264, 265; quarry, 105, see also ditches; middens; postholes plague, 220 plant remains, charred, 17~20, 38-43, 46, 68-70, 76, 81, 148, 151-8, 160 Plantago lanceolata (ribwort plantain), 42, 68, 70, 155 plants: cultivated, 156; food, 156-7, 160 platters, 29, 31; Gallo-Belgic, 29, 31 Pleistocene, 262 Pleydell-Bouverie family, arms, 222 Pliocene, 2 ploughing, archaeological damage, 45, 170, 173, 217 Poa spp. (meadow-grasses), 154, 155 Polden Hills (Somerset), 17, 34, 150 Pole-Tylney-Long-Wellesley, Catherine (née Tylney-Long) (fl. 1812), 226 Pole-Tylney-Long-Wellesley, William, 4th Earl of Mornington (1788-1857), 226 Pole-Tylney-Long-Wellesley, William Richard Arthur, Sth Earl of Mornington (1813-63), 226 Pollard, Josh, 260; obituary by, 262-3 pollen, 67 Pomatias elegans (snail), 65, 66, 91 Pomoideae (apple/medlar/quince subfamily), 66, 96, 146, 148, 149, 150, 151 ponds, 240, 244 Poole Harbour (Dorset), pottery, 182 Poore, Edward (fl. 1784), 226 Poore, Sir Edward (fl. 1838), 226 Poore family, arms, 226 Popish plots, 231 poplars, 151; charcoal, 66, 67, 68, 96, 97, 146, 148, 149; as fuel, 150 Populus spp. (poplars), 66, 67, 96, 146, 148, 149, 150, 151 porticos, pseudo-Classical, 256, 258 Porto Santo (Madeira), 231 Portugal, 231, 233, see also Lisbon; Madeira Portuguese, 233-4 postholes, 87, 97, 101; Late Neolithic/ Early Bronze Age, 10, 46; Middle Bronze Age, 54; Middle/Late Bronze Age, 83; Late Bronze Age, 51, 57-60, 66, 68, 70, 74, 75; Late Bronze Age/Early Iron Age, 101; ?Iron Age, 269; Iron Age, 105, 130, 159, 266, 269; Early Iron Age, 111, 149; Early/ Middle Iron Age, 125; Early Iron Age, 156; Middle Iron Age, 127, 130; Romano-British, 17, 21, 39; Anglo- Saxon, 156; modern, 264, see also pits; stakeholes posts, Middle Iron Age, 115, 148, 149 potassium flame photometry, 103 Potterne: Eastwell House, 268; finds, 260; Little Mill, 268; Mill Lane, 268; pottery, 91, 134 Potterne Medieval Burials Project, 260 potters, 168 pottery: undated, 115, 119, 121; prehistoric, 24; late prehistoric, 89, 90-1, 131-4; Early Neolithic, 260; Late Neolithic, 51, 54, 62, 71; Late Neolithic/Early Bronze Age, 45, 62, 71; Beaker, 10, 15, 22, 23, 45, 51, 54, 62, 71, 79, 83, 89-90, 98, 260; Bronze Age, 46; Early Bronze Age, 12, 81, 83, 279 87, 98; Middle Bronze Age, 54, 62-3; Late Bronze Age, 53, 57, 59, 61, 63, 74, 76, 81, 86, 87, 91, 98; Late Bronze Age/ Early Iron Age, 91, 101; Iron Age, 17, 23-4, 46, 124, 127, 130, 131, 269; Early Iron Age, 76, 87, 111, 114, 264; Early/ Middle Iron Age, 111, 133-4, 146; Middle Iron Age, 113, 117, 119, 120, 121-2, 124, 125-7, 130, 134, 266; Late Iron Age, 29; Late Iron Age/Early Romano-British, 81, 89; Roman, 10, 21, 25-33, 81, 173, 174, 260; Romano- British, 17, 24, 25-33, 86, 87, 101, 106; 160, 168, 175-6, 178, 181-2, 186, 187, 191, 200, 269, 270; Anglo-Saxon, 105, 109, 269; Late Anglo-Saxon, 115, 127, 129, 130, 134, 135; Saxo-Norman, 129; medieval, 89, 115, 130, 131, 134, 135, 181, 216, 217, 219, 260, 264, 269; post-medieval, 134, 135, 173, 181, 264, 268, 269, 270; 12th century, 129, 130; 12th/13th century, 113, 130, 131, 219, 220; 13th/14th century, 105, 130, 161; 14th/15th century, 111, 119; 15th/l6th century, 117; 18th century, 115, 119; 18th/19th century, 111, 115; 19th century, 115, 119; modern, 134, 135, 264, 268; beakers, 17; Black Burnished ware, 17, 27, 29, 31-2, 173, 182, 186; calcite-tempered ware, 132; chalk inclusions, 62; coarse flint, 26-7; coarse greywares, 25; coarsewares, 17, 23, 25, 27, 28-9, 32, 86, 200, 219; colour-coated ware, 27, 32, 173, 182; continental, 27; decoration, 90; Dressel type, 27, 28; fine greywares, 25-6; finewares, 23, 27, 31, 90, 182, 186, 219; flint inclusions, 62; flint- tempered, 29, 62, 89, 91; greywares, 25-6, 29, 32, 182; grog inclusions, 62; grog-tempered ware, 25, 29, 62, 89, 90; Grooved Ware, 51, 54, 62, 71; imported, 181-2; Laverstock type, 219; limestone-tempered ware, 134, 135; organic-tempered, 23-5, 89; Overwey ware, 187; Oxfordshire ware, 32, 173, 182; oxidised, 27, 182; post-Deverel-Rimbury type, 73, 74, 91 (Plain Ware phase, 63, 74, 75-6); quartz-tempered, 23, 135; Samian, 25, 27, 29, 31, 173, 186, 200, 260 (Central Gaulish, 21, 27, 31, 32, 182; South Gaulish, 17, 21, 27, 31, 32; Wiltshire imitation, 27); Savernake ware, 17, 25, 29, 31, 32, 182; Severn Valley Ware, 27, 173, 182; shell inclusions, 62; shell- tempered ware, 62, 132, 134, 135, 173, 182, 186; South-West white-slipped ware, 21, 28-9, 32; Verwood type, 268; whitewares, 182, see also bowls; ceramic building materials (CBMs); clay pipes; cups; dishes; jars; jugs; kilns; tiles; urns Pounde, Thomas (fl. 1569), 230-1 Powell, Andrew B., report on excavations at Hospital of St John the Bapust and St Anthony, Old Sarum, 213-20 Powys, 5, see also Brecon Beacons Prehistoric Society, 262 Preseli Hills (Pembrokeshire), 7, 8 Price, J., 38 prisoner of war camps, 165 Pritchard, F,, 136 Privy Council, 231 Proceedings of the Prehistoric Soctety, 262 Prunella vulgans (seltheal), 155 Prunoideae (stone fruit subfamily), 66, 146 Prunus avium (wild cherry), 155, 1$7 Prunus spinosa (blackthorn), 66, 146, 148, 149, 150, 151, 185, 156 Public Health Survey of Warminster (1838), 256 280 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Pudding Brook, 12 Punctum pygmaeum (snail), 64, 65, 91 Pupilla muscorum (snail), 64, 65, 66, 94, 95 Puranas, 234 Purbeck, Isle of (Dorset), 241 Purton, pottery, 29 Pyke, Henry (fl. 1582), 225 Pyrard, Francois (Pyrard de Laval) (1578-1621), 234 Pyrus spp. (pears), 66, 96, 146 Pytt family, 222 Quarley Hill (Hampshire), 265 quarries: chalk, 266; gravel, 266; stone, 241 quarry pits, 105 Queen’s University (Belfast), 261 Quercus spp. (oaks), 66, 96, 146, 148, 149, 150, 151 querns, 266; medieval, 216 quoins, 217, 267 rabbits, farming, 200 Rachol (India), 234 radiocarbon dating, 15, 17, 22, 54, 70-1, 76, 97; bone, 20, 45, 46, 190, 195-6, 199, 260; charcoal, 57, 70, 73, 74, 145 Radley (Oxfordshire), 45 RAF Lyneham, 230 Rahtz, P A., 219 Raigmore (Highland), 261 railways, construction, 47, 131 ramparts: undated, 123; Iron Age, 105, 124, 130, 159, 160, 265; Early Iron Age, 111, 149; Middle Iron Age, 111, 113, 117-19, 120, 121-2, 125-7, 154; Anglo-Saxon, 105, 113; Late Anglo- Saxon, 111, 115, 127-9; medieval, 111, 115, 119, 130-1; box type, 159; construction, 159; glacis type, 159 Ramsbury, 138, 206; Burdett Arms, 223; Ramsbury Manor, 223; White House, 223 Ranunculus acris (meadow buttercup), 155 Ranunculus bulbosus (bulbous buttercup), 155 Ranunculus repens (creeping buttercup), 155 Ravenna Cosmography, 168 ravens, bones, 184 RCHME see Royal Commission on the Historical Monuments of England (RCHME) Reading Beds, 239 red deer see deer Red Sea, 233 rents, 227 resistivity surveys, 260 Restoration, 222 Return of Owners of Land (1873), 226 revetments, timber, 267 Rhamnus cathartica (purging buckthorn), 68 Rhinanthus spp. (yellow rattle), 42 Richards, Colin, 260 Richards, Julian (1951— ), 97-8, 106, 160 ridge and furrow, 266 Rigby, V., 29 rings, 249-50; finger, 36; iron, 170 ritual shafts, 200 rivets, 183 roads: Roman, 165, 166, 168, 213, 243, 244-5, 266, 267, see also trackways Robert I, Count of Flanders, 210 Robinson, Paul, 169 rodents, bones, 45 Roger of Salisbury (d. 1139), 109, 131; and Romanesque architecture, 253 Rogers, Henry (d. 1721), 225 Rogers family, 225 Rollright Stones (Warwickshire), 70 Roman Catholics, 234 Romanesque style, 250-4 Rome (Italy), 219; Sant’ Andrea, 231 Romney, George (1734-1802), 227 Romsey (Hampshire), 196; Romsey Abbey, 252 roofing materials: sandstone, 173, 183, 185, 268; terracotta, 268 rope-making, 150 Rosa spp. (briars), 66 Rosaceae (roses and fruit trees), 66, 96, 146 roses, seeds, 43 Rosoideae (rose subfamily), 66 roundhouses: Middle Bronze Age, 75; Late Bronze Age, 51, 56, 57-9, 61, 63, 66, 68, 70, 74, 75; Late Bronze Age/Early Iron Age, 260; Iron Age, 75; radiocarbon dating, 70-1 rowans, charcoal, 66, 96, 146 Royal Commission on the Historical Monuments of England (RCHME): Salisbury Plain Training Area Project, 200; surveys, 166, 170 Royalists, 110 RPS Group Plc, 266 rubbish pits see middens Rubus spp. (brambles), 66 Rumbold, Sir Horace Anthony Claude (1911-83), 223 Rumex spp. (docks), 157; seeds, 39, 155 Rumex acetosella (sorrel), 39, 156 Rushall, manor, 226 rushes, 157 Russell, Pamela, 248 rye, cultivation, 156, 157 Ryknield Street, 185 sacrifices, human, 201 St Albans (Hertfordshire), Verulamium, 173 St Augustine (d. 604), 208; Soliloquies, 207 St Benedict (c. 628-690), 207 St Edward the Confessor (c.1004/5—1066), 249-50 St Edward the Martyr (c. 963-978), 249 St Isidore (c. 560-636), Synonyms, 207 St Jerome (c. 347-420), 208 St John, Henry (fl. 1712), 223 St John, Mary Emily Elizabeth (née Howard), Viscountess Bolingbroke (d. 1940), 223 St John, Oliver, 223 St John family, arms, 223 St John Ambulance, 257 St John the Evangelist, 249-50 St Kenelm’s Well (Gloucestershire), 243 St Laurence Island [Madagascar], 232 St Michael, 249 St Omer (France), St Bertin’s Abbey, 205, 206, 210 St Paul Malmesbury Without: Rodbourne Cheyney, 269-70; Whitworth Road, 269-70 Salcete Peninsula (India), 230, 233, 234 Salicaceae (poplars and willows), 66, 96, 146 Salisbury, bishops of, 230, 268 Salisbury, 132, 206, 213, 248; Castle Hill, 217; Castle Hill Reservoir, 213; diocese of, 204; Endless Street, 268; establishment, 213; Fisherton Anger, 225, 268; Fisherton Street, 268; The Hogshead, 225; Malmesbury Arms, 225; North Street, 268; Old Castle Inn, 220; pottery, 219; See of, 253, see also Old Sarum; Sorviodunum Salisbury Cathedral, Library, 206 Salisbury Charters (1253-59), 250 Salisbury Museum see Salisbury and South Wiltshire Museum Salisbury Plain, 53, 63, 90, 190; barrows, 67; charcoal, 67, 68, 97; clearances, 71, 73; earthworks, 79, 100; field systems, 71; linear ditches, 91,97, 100; ° Mesolithic activity, 97-8; pottery, 91 Salisbury Plain Training Area (SPTA), 73-4, 79, 269; Robin Hood’s Ball, 68; Southern Range Road, 80; Track 58, 80 Salisbury Pontifical, 207 Salisbury and South Wiltshire Museum, 2-3, 4-5 Salix spp. (willows), 66, 67, 146, 148, 149, 150, 151 Sambucus spp. (elders), 66, 150, 151 SAMs see Scheduled Ancient Monuments (SAMs) sandstones: petrography, 1—9; roofing materials, 173, 183, 185, 268 Sandys, Edwin (c. 1516-88), 231 Santéry (god), 233 Sao Lourenco (ship), 231 Saresbyri, 213, see also Old Sarum sarsen stones, 199 Sarum see Salisbury saucepan pots: late prehistoric, 132; Middle Iron Age, 134 Savernake, 245; estate map, 247; Savernake House, 238; Tottenham, 237, 238, 241, 245; Tottenham House, 221 Savernake Forest, 239; Baden Coppice, 247; Birch Copse, 247; Crabtree Common, 247; Crabtree Cottages, 247; Duke’s Vaunt, 246-7; oaks, 246-7; pottery, 17, 25, 29, 31, 32, 182; Puttal Horse Leaze Gate, 247; William Farmer’s Common, 247 Saxons, 248 Scheduled Ancient Monuments (SAMs), 79, 110, 165, 264, 267 Schuster, Jorn, note on small finds from Whitewalls, Easton Grey, 182-3 Scotland: bee boles, 259; hillforts, 160; wildwood, 97, see also Craig Phadrig; Croft Moraig; Dun Lagaidh; Edinburgh; Glasgow; Kaimes; Northton; Perth Academy; Pitnacree; Raigmore scrapers: Neolithic, 270; Late Neolithic/ Early Bronze Age, 54, 61, 71; Beaker, 22; Early Bronze Age, 22, 88, 89; flint, 15, 22, 83, 269 sculpture, Romano-British, 168-70 scurvy, 233 Sea Mills (Bristol), 27 Seagram, Dr, 200 Seamer (North Yorkshire), 261 Secale cereale (rye), 156 Sedgehill and Semley: Benett Arms, 222-3; Callis Place, 222 sedges, 157 sediments, 5—7 seeds: charred, 39-42, 68; mineralised, 151 Selwood (Wiltshire-Somerset borders), 248 Semington, Somerset Arms, 226 Senni Beds, 1-2, 3, 5-8 Serjeantson, D., 138 service trees, charcoal, 146 settlements: Mesolithic, 45, 262; Neolithic, 262; Late Neolithic, 46; Beaker, 98, 261; Bronze Age, 46; Middle Bronze Age, 73; Middle/Late Bronze Age, 51-78; Late Bronze Age, 70, 71, 73-4, 76; Iron Age, 105, 111, 185, 264; pre-Roman, 185; Romano- British, 46, 165-89, 200, 267; Anglo- Saxon, 106, 187; medieval, 200, 264; Celtic, 248; Deverel-Rimbury type, 74; English, 248; linear, 185-6, see also castles; enclosures; towns; villages Severn Valley, pottery, 27, 173, 182 sewers, 258 } INDEX Seymour, Alfred (fl. 1877), 226 Seymour, Edward, Ist Earl of Hertford and Duke of Somerset (c.1506-1552), 226, 236-45; and Duke’s Vaunt, 246-7 Seymour, Edward Adolphus, 11th Duke of Somerset (1775-1855), 226 Seymour, Sir Horace, 198 Seymour, Hugh, 190, 192, 193, 195 Seymour, Jane, Queen (c. 1509-37), 236, 246 Seymour, John Michael Edward, 19th Duke of Somerset (1952-— ), 226 Seymour, Lady, 192, 198 Seymour, Thomas (c. 1508-49), 226, 236, 243 Seymour family, 238; archives, 236, 239; arms, 226 SFBs (sunken-featured buildings), 266 Shaftesbury Abbey (Dorset), 249 shafts, ritual, 200 Shalbourne, 241 Shannon species diversity index, 93, 95 Shapwick (Somerset), 150 Sharrington, Sir William, 243-4 Shaw, Walter William (1868-1927), 225 Shaw-Stewart, Sir Michael Hugh (1854-1942), 224 Shaw-Stewart, Walter Richard (1861— 1934), 224 sheela-na-gigs, 250-4; distribution, 251, 253; interpretations, 251 sheep: burials, 54; culling patterns, 143-4; teeth, 70, 71, 140, 143-4 sheep/goat bones, 44, 46, 260; Late Bronze Age, 70, 76; Iron Age, 140; Romano-British, 184; Anglo-Saxon, 141-4; medieval, 141-4, 216, 217; differentiation, 138 Sheffield University, 260 Sheppey, River, 186 Shepton Mallet (Somerset), 186; Fosse Lane, 186 Sherborne (Dorset), 205, 206, 210 Sheridan, Alison, 262 Sherston, 165 Shipp, Deborah, report on fieldwork at Whitewalls, Easton Grey, 165-89 Shorncote (Gloucestershire), 45 Shrewton: Karrick House, 268; Robin Hood's Ball, 68; Tanners Lane, 268 Silchester (Hampshire), 33 Silurian, 8 Simons, Ramfrey, 239, 240 Simpson, Derek Douglas Alexander (1938-2005): ed. Introduction to British Prehistory (1992), 261; ed. Studies in Ancient Europe (1968), 261; Food Vessels in SW Scotland, 261; obituary, 261-2; Settlement and Economy in the Neolithic and Bronze Age, 261 Simpson, Nancy, 262 Sisson, Septimus, 138 Sites and Monuments Record (SMR), 192, 245 skeps, 259 slag, 115, 183 slate, 217 slingshot, 36 sloes see blackthorn trees Smith, Rev Alfred Charles (c.1823—99), 221, 222, 226 Smith, Isobel (1912-2005), 261, 262 Smith, Pippa, note on animal bone from Dunch Hill, Tidworth, 70 Smith, Wendy, 259 smithing, 38, 47 SMR (Sites and Monuments Record), 192, 245 snails: burnt, 148; introduced species, 86; land, 54, 64—6, 71, 73, 81, 91-5, 99, 100, 101, 158, 262; water, 94 Snell, Bill, note on excavations at Baverstock, 254—5 Society of Jesus, 230, 231, 232 Socotra [Zocotoro] (Yemen), 233 soils, 39, 265, 268, see also subsoils Somerset, Dukes of, 226, 227 Somerset family, arms, 226 Somerset: Roman settlements, 186, see also Bath and West Showground; Frome; Glastonbury; Ilchester; Mells; Polden Hills; Selwood; Shapwick; Shepton Mallet; South Cadbury; Stoke Sub Hamdon Somerset Levels, 150 Sorbiodunum see Sorviodunum Sorbus spp. (rowans/whitebeams), 66, 96, 146, 148, 149, 150, 151 Sorviodunum, 213, see also Old Sarum South Africa see Das Agulias; Good Hope, Cape of South America, 234 South Cadbury (Somerset), 160 South Newton, Camp Hill Reservoir, 213 South Tidworth, 241; Dunch Hill, 51-78; Tidworth Golf Course, 76 South Wales, 132, 248; sandstones, 1, 5-7, 8 South Wiltshire Young Archaeologists’ Club (SWYAC), excavations, 254-5 South Wraxall: Long’s Arms, 221, 224; Wraxall Manor, 224 Southwest Group of Museums and Art Galleries, Implement Petrology Committee, 3 Spain, 231, see also Canary Islands Spanish, 233 spectrometry, 103 Speen (West Berkshire), 267 spelt, 39, 42, 43, 46, 151, 154, 155; cultivation, 156 Spencer, Francis Almeric, lst Baron Churchill of Wychwood (1779-1845), 223 Spencer, George, 4th Duke of Marlborough (1739-1817), 223 spindlewhorls, clay, 109 springs, 201, 241 SPTA see Salisbury Plain Training Area (SPTA) stables, 256, 258; post-medieval, 269 Stace, C., 151 Staffordshire: pottery, 135, see also Wall stakeholes, 87, 101, 187; Late Neolithic/ Early Bronze Age, 71, see also postholes Stallard, George, 237-8, 243, 244 Stansell (Conservation) Ltd, 110 Stanton Fitzwarren: Church Lane, 268; The Rookery, 268; Stanton Park, 268 Staverton: Marina Drive, 268-9; New Terrace, 268-9 Steeple Ashton: Long’s Arms, 221, 224; manor, 224 Steeple Langford: Hanging Langford, 266; Yew Tree Cottage, 266 Stephen, King (1096-1154), 109, 253 Stephens, Jane, 230 Stephens, Richard (1) (fl. 1533), 230 Stephens, Richard (2) (1540-86), 230 Stephens, Thomas (2) (1549-1619), 230-5; Arte da Lingoa Canarim (1640), 235; Doutrina Cristam (1622), 235; Purdna Christdo (1616), 234, 235 stews, 144 Sudolph, Susie, note on the activities of the Archaeology Field Group, 259-60 Stockton, 98 Stoke Sub Hamdon (Somerset), 251 Stone, Dr, § Stone, J. FE S., 220 Stone, John, 247 stone objects, Roman, 183 281 stone quarries, 24] Stonehenge, 260; Altar Stone, 1-9; Aubrey Holes, 2, 4-5; bluestones, 1, 2, 8; charcoal, 68, 97; clearances near, 71; Cursus, 2, 4; excavations, 261; flintwork, 97~8; plant remains, 70; pottery, 90, 91; sandstone petrography, 1-9; Y Holes, 94, see also Amesbury Stonchenge Riverside Project, 260 stones: burnt, 62, 175; sarsen, 199; vitrified, 105, 114, 160 stonework: medieval, 220, see also flintwork; hammerstones; querns Story, James (fl. 1583), 233-4 Stradbrook, 190, 196 Suraker, Vanessa, 146, 148 Street, Mr, 223 Stroud, Daphne, paper on Eve of Wilton and Goscelin of St Bertin at Old Sarum, 204-12 Studland (Dorset), 251 Stuiver, M., 71 Stukeley, William (1687-1765), Stonehenge (1740), 1 subsoils: Early Bronze Age, 81, 87-8; Romano-Briush, 81, 101 Sudeley Castle (Gloucestershire), 243, 244 Suffolk family, arms, 221 sunken-featured buildings (SFBs), Anglo- Saxon, 266 sunrise, mid-summer, 45 Surrey see Hurst Park; Overwey surveys: magnetometer, 166, 170; resistivity, 260; Tilshead, 260; Warminster, 256; Whitewalls, 166, 170-3 Sutton Benger: Draycot Cerne, 225, 266; St James Church, 266; Wellesley Arms, 226 Sutton Veny, Longbridge Hill, 269 Swan, Mrs, 185 Swindon, 196, 254; Blunsdon Bypass, 269; Goddard Arms, 221, 224; Great Western Railway Works, 269; Groundwell Farm, 269; Holy Rood Church, 269; Hreod Parkway School, 269; Kemble Drive, 269; Kingsdown Crematorium, 269; The Lawn, 224; The Lawns, 269; manor, 224; National Trust Central Office, 269; North Leaze, 269; Old Town, 269; Pipers Way, 269; Swindon House, 224; Triangle Site, 269; Whitehill Farm, 29 Swindon Corporation, 223, 224 SWYAC (South Wiltshire Young Archaeologists’ Club), 254—5 Sykes, Naomi, note on animal bones from Malmesbury, 138-45 symbolism, 253, 265 Syon House (Hounslow), 236 Syria, 233 Tacitus (c. 5S—120), 201 tacks, 183 Taf, River (Carmarthenshire), 6 Tait, J., 109 Talbot, C. H., 204 Talbot, Reginald John Arthur Arundell (d 1953), 222 tankards, Romano-British, 27, 29, 31 taxes, capital, 227 Taxus spp. (yews), 68 teeth: abscesses, 194, 200; and ageing, 140, 143—4, 194; animal, 44, 70, 71, 260; cattle, 140, 144; horses, 144-5; human, 43—4, 63, 64, 194; sheep, 70, 71, 140, 143-4; wear, 138 tegulac, Roman, 33, 46, 47 Tenerife (Canary Islands), 231-2 Teshe, Bryan, 239, 242 tesscllation, Roman, 168, 187, 260 Tetbury (Gloucestershire), 222 282 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE textile industry, 144 Thames, River, 238 Thames Valley, 45 thatch weights, 36 Thomas, Charles, 261 Thomas, H. H., 1-2, 3, 7 Thomas, James H., paper on Thomas Stephens (1549-1619), 230-5 Thomas, R. G., 2, 5,7 Thomas, Stanley, 261 Thorpe, R. S., 1, 2 Thurnam, John (1810-73), 80 Thynne, Sir John (d. 1580), 230, 237, 239, 242 Thynne family: archives, 236, 239; arms, 222 Tidworth—Larkhill military track, 51 tigers, 232 Tighe, Daphne, paper on Thomas Stephens (1549-1619), 230-5 tiles: Roman, 33, 47, 268; medieval, 217, 219, 268, 269; post-medieval, 264; modern, 267; box-flue, 33, 47, 173, 185, 268; floor, 217; ridge, 219; roof, 219 Tilia spp. (limes), 146, 148, 149, 150, 151 Tiliaceae (linden family), 146 Till, River, 80, 98 Tilshead: Breach Hill, 68 (Old Ditch, 79-103); Copehill Down, 79, 91, 100, 270; enclosures, 260; Old Ditch long barrow, 80; parish boundary, 81; White Barrow, 80 Tilshead Project, 260 timber: in buildings, 148, 187; uses, 149-50 Timby, J. R., note on pottery from Whitewalls, Easton Grey, 181-2 Tisbury: Arundell Arms, 222; Benett Arms, 221, 222; manor, 222; Wardour, 222; Wardour Castle, 227 Tite, M. S., note on optically stimulated luminescence dating, 103 Tithe Award maps, 198 Titus, 168 Tomber, R., 29 tongs, 183 tools: Mesolithic, 98; Late Neolithic, 22; Late Neolithic/Early Bronze Age, 105, 124; Early Bronze Age, 88, 89, 98; Roman, 36, 183; flint, 269; lithic, 263, see also axes; blades; hammerstones; knives; scrapers toponyms, 248 towers, medieval, 115 towns: Roman, 184-5, 187-8, 200; medieval, 213 trackways, 173, 260, 267; ?Middle Bronze Age, 73; Roman, 10, 17, 21, 46, 187, 268-9; Early Roman, 15; ?medieval, 264; construction, 150; military, 51, 71, 79, see also hollow ways; paths; roads Trajan, 168 tree-throw holes, 53, 54, 62, 71, 79, 81, 83, 86, 99; charcoal, 95, 96, 97; land snails, 64, 91, 93 Trichia hispida (snail), 91, 95, 158 Trifolium spp. (clovers), 156 Tripleurospermum inodorum (scentless mayweed), 156 Tripps, Mr, 231 Triticum spp. (wheat), 155; charred grains, 68, 151 Triticum aestivum (bread wheat), 151, 154, 155, 157; cultivation, 156 Triticum compactum (club wheat), 154, 157; cultivation, 156 Triticum dicoccum (emmer wheat), 39, 42, 43, 68, 151, 154 Triticum durum (macaroni wheat), 157 Triticum spelta (spelt), 39, 42, 43, 46, 151, 154, 155, 156 Triticum turgidum (river wheat), 157 Trotter, M., 194 Trowbridge, 196; Castle, 138 tuberones, 232 tumuli see barrows Turner, Peter, paper on petrography of non-sarsen sandstones from Stonehenge, 1-9 Uley Bury (Gloucestershire), 157, 159 Ulmus spp. (elms), 68, 150, 151 Ulster Fournal of Archaeology, 262 United States (US), imports from, 227 University College Cardiff, 262, see also Cardiff University University of Oxford see Oxford University University of Southampton: Centre for Applied Archaeological Analysis, 138; Centre for Human Ecology and Environment, 138 University of Waikato (New Zealand), Radiocarbon Dating Laboratory, 38, 195, 196 Upper Chalk, 51, 79, 239, 270 Upper Greensand, 131-2 Upper Jurassic, 131 Upper Thames Valley, 45, 46 urns: Middle Bronze Age, 54, 62, 63; bucket, Deverel-Rimbury type, 54, 62, 63; collared, 261 Valentian I, 106, 180, 181 Valerianella dentata (narrow-fruited cornsalad), 156 Valley Gravel, 268 Vallonia spp. (snails), 64, 94, 95, Vallonia costata (snail), 65, 94, 158 Vallonia excentrica (snail), 66, 94, 159 Valor Ecclesiasticus (1535), 219 van der Veen, M., 42-3 Vane, Harry George Powlett, 4th Duke of Cleveland (1803-91), 223 Variscan Orogeny, 5 venison, poached, 145 Venta Belgarum (Winchester), 213 Verbena officianalis (vervain), 156 vertebrates, remains, 184 Vertignids, 93 Vertigo antivertigo (snail), 158 Vertigo pusilla (snail), 91, 93 Vertigo pygmaea (snail), 64, 158, 159 Verulamium (Hertfordshire), pottery, 173 Verwood (Dorset), pottery, 268 Vespasian, 168 vessels: late prehistoric, 132-3; Late Neolithic/Early Bronze Age, 62; Beaker, 90; Early Bronze Age, 98; Late Bronze Age, 63; Iron Age, 134; Early Iron Age, 111; Romano-British, 27, 29, 32, 36; glass, 36-8 vestries, 269 vetches, 154, 155, 156 viburnums, 150 Vicia spp. (vetches), 154, 155, 157 Vicia faba (broad bean), 154, 156 villages, Romano-British, 68 villas, Romano-British, 25, 46, 47, 187, 200 Villiers, Charlotte, 223 Villiers, Thomas, lst Earl of Clarendon of the Villiers family (1709-86), 223 vines, 157 Vitrina pellucida (snail), 91 voles, bones, 45, 70 von den Driesch, Angela, 44, 138 WA see Wessex Archaeology (WA) Wacher, John, 261 Wales: bee boles, 259; cosmetic grinders, 33; sediments, 5, see also Cardiff; Craig Ddu; Kidwelly; Pembrokeshire; Powys; South Wales; Taf, River Walker, Penelope, note on the International Bee Research Association Bee Boles Register, 259 Wall (Letocetum, Staffordshire), 185 wall paintings, 250 walls: Iron Age, 124; Early Iron Age, 114; Middle Iron Age, 105, 124, 127, 159-60; Romano-British, 175-6, 178, 179, 185, 264; Anglo-Saxon, 109; medieval, 111, 129, 130, 131, 217-19, 252, 268; post-medieval, 111, 113, 117; 12th century, 105, 109, 161; 12th/13th century, 113, 123-5; 13th/14th century, 131; 18th century, 255, 266; 19th century, 109, 256, 268; modern, 269; cob, 266; collapsed, 119-20, 121, 124, 149; flint and cob, 255; limestone, 105, 130, 159 Wanborough: brooches, 34; Calley Arms, 223; cosmetic grinders, 33; finger rings, 36; Great Moor Leaze, 266-7; pottery, 25, 27, 29, 31 WANHS see Wiltshire Archaeological and Natural History Society (WANHS) Wareham (Dorset), 249 Warminster: Bath Arms, 221; Battlesbury Camp, 98; Board of Health, 257; Boreham Farm Bungalow, 98; The Close, 257; fire stations, 256-8; George Street, 256, 257, 258, 270; High Street, 258; Law Courts, 258; Literary Institute, 258; Portway, 258; pottery, 134; Public Health Survey (1838), 256; St John’s Ambulance HQ, 257; surveys, 256; Were House, 256-8, 270; Weymouth Street, 257 Warminster Urban District Council, 257 Warminster Voluntary Fire Brigade, 257 Warwick, Septimus, 267 Warwickshire see King’s Newnham; Rollright Stones watching briefs: Ashton Keynes, 264; Dunch Hill, 51, 53; Easton Grey, 166; Malmesbury, 110; Swindon, 269 water pipelines, 168, 170-1, 185; Old Sarum, 213, 216, 220 water supplies, 242, 243-4, 245 waterholes, 21; Iron Age, 266 Watts, K. G., Droving in Wiltshire: The Trade and its Routes (1990), 255 Watts, V. E., 248 Wayland’s Smithy (Oxfordshire), 261 Webster, G., 185 weeds: arable, 157; charred remains, 76; and cultivation, 158; seeds, 15, 17, 39-42, 68, 70, 146, 149, 151, 155 weights, bone, 144 Wellesley, Henry Richard Charles, Ist Earl Cowley (1804-84), 225 Wellesley, William Henry, 2nd Earl Cowley (1834-95), 226 Wellesley family, arms, 226 wells: Romano-British, 20, 200; post- medieval, 270 Welsh language: adjective formation, 248; Celtic elements, 248 Welsh Marches, 5 Were, River, 256, 258 Wessex: barrows, 45; defences, 105 Wessex Archaeology (WA), 254; evaluations, 53, 79, 264, 267, 270; excavations (Bratton, 264; Old Sarum, 213-20; South Tidworth, 51-78; Tilshead, 79-103); monitoring, 268; recording system, 62, 89; watching briefs, 51, 53 Wessex Linear Ditches Project, 53, 64, 71, 73-4, 75, 76; activities, 81, 100; Northern Core Territory, 74, 76; sites, 72; Southern Core Territory, 74; study area, 91; Upper Study Area, 74 INDEX Wessex Water, 196, 198, 213 West Ashton, Rood Ashton, 224, 225 West Berkshire see Speen West Lavington: Churchill Arms, 223; Strawberry Hill, 97, 100 West Overton, 261; George Bridge, 270 West Sussex see Chichester Cathedral West Tisbury: Benett Arms, 223; East Hatch, 223; Hatch Farm, 223; Newtown, 223; Pythouse, 222, 223; West Hatch, 223 West Wiltshire District Council, 12 Westbury, 196, 200, 225-6; Chalcot House, 226; Chalford, 226; Enclosure Map and Award, 198; The Ham, 200; Leighton House, 226; Lopes Arms, 225; Ludlow Arms, 225; Phipps Arms, 221, 225; Wellhead, 200; Westbury Hotel, 225; Westbury Leigh, 225-6 Westbury, Hundred of, 197, 198; boundary, 201 Weston, William (c. 1550-1616), 231 wheat, 43; charred, 39, 42, 46, 76, 151, 154; cultivation, 156, 157, 158; grains, 15, 20; hulled, 155, 156; prices, 227, see also emmer wheat; spelt Whitaker family, 197-8 whitebeams, charcoal, 146 Whittle, Alasdair W. R., 262 Whittlesford (Cambridgeshire), 251 WHM see Wiltshire Heritage Museum (WHM) Wilcot: The Green Man, 225; Mompesson Arms, 225; Oare, 225; Rainscombe, 225; The White Hart, 225 wild plants see weeds William I, ‘the Conqueror’ (1027-87), 205 William of Malmesbury (de. 1143), 205, 206, 209, 210 willows, 151; charcoal, 66, 67, 68, 96, 97, 146, 148, 149; as fuel, 150; uses, 150 Wilmart, André (1876-1941), 204, 210 Wilmott, Tony, report on fieldwork at Whitewalls, Easton Grey, 165-89 Wilsford cum Lake, North Kite, 68 Wilson, Frances (née) Goddard (fl. 1900), 224 Wilson, William (fl. 1900), 224 Wilton, 205, 208, 209, 211; nunnery, 204, 206, 209, 210; Pembroke Arms, 22] Wilton Abbey, 204, 210, 211 Wiltshire, Sheriff of, 231 Wiltshire Archaeological and Natural History Magazine (WANHM), 259 Wiltshire Archaeological and Natural History Society (WANHS), 190, 221, 238; Archaeology Committee, 190, 259; Library, 196, see also Archaeology Field Group (AFG); Wiltshire Heritage Museum (WHM) Wiltshire County Council: Archaeological Officer, 260, 264; Finds Liaison Officer, 260; Local Studies Library, 196; Museums Service, 53 Wiltshire Heritage Museum (WHM), 111, 168, 170, 261; collections, 191, 247 Wiltshire Society see Wiltshire Archaeological and Natural History Society (WANHS) Wiltshire Studies see Wiltshire Archaeological and Natural History Magazine (WANHM) Wiltshire and Swindon Record Office (WSRO), 196, 254 Winchester (Hampshire): Roman roads, 267; St Swithin, 230; Venta Belgarum, 213; Winchester Cathedral, 252; Winchester College, 230 windows, glass, 219 Winterbourne: churches, 249-50; 283 Winterbourne Dauntsey, 249-50; Winterbourne Earls, 249 Winterslow, 227 wire, Late Anglo-Saxon, 137 wood carving, 150 Woodcock, Alex, note on Romanesque corbel-table at St John the Bapust Church, Devizes, 250-4 Woodcuts (Dorset), 46 woodland, 67, 71—3, 93, 97, 99-100, 248; resources, 150 wool, 143-4, 145 Wootton Bassett, 230; Clarendon Arms, 223; manor, 223; Park Grounds, 270 Worcestershire, 5 World War I, 227; military badges, 260 Worth Matravers (Dorset), 251 Wotton-under-Edge (Gloucestershire), 157 Wright, R. P, 169 Wroughton, Barbury Castle, 270 WSRO (Wiltshire and Swindon Record Office), 196 Wylye, River, 98 Wylye Valley, 98 Wymer, J. J., 98 Wyndham, William (fl. 1689), 226 Wyndham family, arms, 221, 226 Xavier, Francis (1506-52), 234 Yemen see Socotra yews, charcoal, 68 Young, Richard, report on excavations at Showell Farm, Chippenham, 10-50 Zone system, 138 Zonitidae (snails), 94, 95 Zussman, J., 3 284 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Publications of the Wiltshire Archaeological and Natural History Society Recent and some earlier issues of the Magazine may be purchased from the Society. Volumes 93-97 are available at £15 per copy. For details of earlier volumes apply to the Curator. Other WA&NHS publications may also be purchased, as follows: Annable, EK., and Simpson, D.D.A., Guide catalogue of the Neolithic and Bronze Age collections in Devizes Museum, [viii] 133pp, plates, casebound, 1964 (reprinted), £15.00 (+ £3 p&p) Dillon, Patrick (ed.), Mammals in Wiltshire, xii, 1S6pp, paperback, 1997, £7.50 (+ £1.50 p&p) Ellis, Peter (ed.), Ludgershall Castle: excavations by Peter Addyman 1964-1972, ix, 268pp, ill, A4 paperback, 2000 (WANHS Monograph Series 2), reduced to £5.00 (+ £4.50 p&p) Ellis, Peter (ed.), Roman Wiltshire and after: papers in honour of Ken Annable, xii, 240pp, ill, casebound, 2001, £19.95 (+ £4.50 p&p) Hatchwell, Richard, Art in Wiltshire, from the Wiltshire Archaeological and Natural History Society Collection, xii, 154pp, coloured ill, casebound and paperback, 2005, £40.00 casebound; £30 paperback (+ £3.00 p&p)) Haycock, Lorna, Devizes in the Civil War, 24pp, ill, paperback, 2000, £2.95 (+ £1.00 p&p) Thomas, James H. (ed.), Wiltshire Archaeological and Natural History Society: the first 150 years, xxxiv, 246pp, ill, case- bound, 2003, £12.00 (+ £3.95 p&p) Thomas, Nicholas, Snail Down, Wiltshire: The Bronze Age Barrow Cemetery and Related Earthworks, in the parishes of Collingbourne Ducis and Collingbourne Kingston, Excavations, 1953, 1955 and 1957, 2005, £25 (+£4.50 p&p) During 2006 the Society plans to publish a report on the excavation of the Saxon cemetery at Blacknall Field, Pewsey between 1969 and 1975. \) \\ WILTSHIRE HERITAGE -MUSEUM GALLERY SUcol ae. LIBRARY. ane Published by - . The Wiltshire Archaeological and Natural History Society ISSN 0262 6608