al € C n | Z| lo ga aeO y Ma 4 * _ 2 O “ta _ The Wiltshire Archaeological _ and Natural History Magazine — Volume 101 2008 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 | 39 JAN cc03 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE VOLUME 101 (2008) ISSN 0262 6608 © Wiltshire Archaeological and Natural History Society and authors 2008 Hon. Editor: Andrew Reynolds, BA, PhD, FSA, FSA Scot, FRHistS. Hon. Assistant Editor: Stuart Brookes, BA (Hons), PGCTiLL, MA, PhD. Hon. Local History Editor: James Thomas, BA, PhD, FRHistS. Hon. Natural History Editor: Michael Darby, PhD, FRES Hon. Reviews Editor: Robert Clarke BA (Hons), Cert Ed, MIfL. Hon. Excavation and Fieldwork in Wiltshire Editor: Simon Draper, BA (Hons), MA, PhD. Finds Liaison Officer: Katie Hinds, BA (Hons), MA. 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: English Heritage for ‘An Iron Age and Romano-British settlement at Cleveland Farm, Ashton Keynes, Wiltshire’ by Andrew B Powell, Grace Perpetua Jones and Lorraine Mepham, and ‘Excavations at Atworth Roman villa, Wiltshire 1970-1975’ by Jonathan G P Erskine and Peter Ellis, and Avon Archaeological Unit for ‘Evidence of 17th century brick-making at Highworth, Wiltshire’ by Donna E Y Young. Addendum to Volume 100. On page 21, Early Tertiary Turtles in Wiltshire, the grid references for the old Hamptworth brickpit and disused Redlynch brickpit should have read ‘SU 233 230’ and ‘SU 202 288’ respectively. The journals issued to volume 69 as parts of The Wiltshire Archaeological and Natural History Magazine (Part A Natural History; Part B Archaeology and Local History) were from volumes 70 to 75 published under separate titles as The Wiltshire Natural History Magazine and The Wiltshire Archaeological Magazine. With volume 76 the magazine reverted to its combined form and title. The cover title ‘Wiltshire Heritage Studies’ (volume 93) and ‘Wiltshire Studies’ (volume 94 onwards) should not be used in citations. The title of the journal, The Wiltshire Archaeological and Natural History Magazine, remains unchanged. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the Society and authors. Cover illustration: 12 o’clock Drive, photograph by Graham Bathe 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 NATURAL HISTORY, ARCHAEOLOGY and LOCAL HISTORY The Win Green Mosasaur: the earliest British record? by Justin B. Delair Dry stone walls: a biological treasure? by John Presland An Iron Age and Romano-British settlement at Cleveland Farm, Ashton Keynes, Wiltshire, by Andrew B. Powell, Grace Perpetua Jones and Lorraine Mepham, with contributions by Stephanie Knight, Facqueline I. McKinley, Chris Stevens and Nicholas A. Wells, and illustrations by Rob Goller Excavations at Atworth Roman villa, Wiltshire 1970-1975, by fonathan G.P Erskine and Peter Ellis, with contributions by Fane Bircher, Gillian Campbell, H.E.M. Cool, Brenda Dickinson, Cathy King, Donald Mackreth, facqui Mulville, Jane Timby, and Susan Youngs Sarum cathedral as rebuilt by Roger, Bishop of Salisbury, 1102-1139: the state of research and open questions, by Malcolm Thurlby In search of the Bishop of Salisbury’s Manor in Potterne, by Neil D. McGlashan Church, war and plague: the evidence from medieval Etchilhampton, by Simon Phillips The Great Inclosure of Savernake with a note on Cross Valley Dykes, by Joanna Ramsay and Graham Bathe Evidence of 17th century brick-making at Highworth, Wiltshire, by Donna E.Y. Young Georgian gentlemen: the Canons of Salisbury in the long 18th century, by Graham Hendy Captain Augustus Montgomery, RN (1762-1797): Wiltshire’s forgotten aristocratic bastard, by Sheila R. Thomas Joseph Townsend 1739 — 1816: a well respected polymath, by Cheryl Underhill The Easton family of Salisbury: booksellers, printers and publishers c.1720-1839, by Robert Moody 18 51 130 141 149 158 176 193 213 226 237 NOTES and SHORTER CONTRIBUTIONS Some Hertfordshire finds in the Stourhead Collection, by Paul Robinson Idover and the Britons, by Andrew Breeze Kemble and the Britons, by Andrew Breeze A medieval silver finger ring from near Malmesbury: one ring, two merchants? by Nick Griffiths The quest for the earliest published image of Stonehinge (sic), by Stephen Allen The Jackson letters: a postscript, by Lorna Haycock WANHS Archaeology Field Group: recent activities and future plans, by James Gunter REVIEWS, edited by Robert Clark EXCAVATIONS and FIELDWORK in WILTSHIRE 2006, compiled by Simon Draper Highlights from the Portable Antiquities Scheme (PAS) in Wiltshire in 2005, recorded by Katie Hinds INDEX, by Philip Aslett 251 253 255 256 257 260 261 264 274 280 285 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 a summary of about 100 words. Please submit two copies of the text (with computer disk if possible) and clear photocopies of any illustrations to the editors at the Museum, 41 Long Street, Devizes, Wiltshire, SN10 INS. 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 (as at 1 January 2008) Chairman W A Perry, MSc Deputy Chairman D L Roseaman, BSc(Eng), CEng, MIMechE Hon Treasurer/Company Secretary Mrs W P Lansdown, FCA Other Elected Trustees Lt Col C Chamberlain C R Chippindale, BA, PhD, MIFA, FSA Miss K A Fielden, BA, D.Phil. J A Gunter, BA (Hons) N Harte, DSc (Econ), FRHistS, FSA Mrs V Knowles A Reynolds, BA, PhD, FSA, FSA Scot, FRHistS P R Saunders, BA, FSA, FMA, FRSA Mrs J Triggs Nominated Trustees C T Callow, BSc, ARCS, MBCS (Member, Devizes Town Council) A X T Green, BA, MA, AMA (Director, Salisbury & South Wiltshire Museum) A Molland (Member, Wiltshire County Council) Mrs J Swabey (Member, Wiltshire County Council) In attendance: T Craig BA, MA, AMA, DipMgmt (Wiltshire County Council Heritage Services Manager) OFFICERS Curator P H Robinson, PhD, FSA, AMA Assistant Curator Miss L Webb, BA, MA Sandell Librarian and Archivist Mrs L Haycock, BA (Hons), PhD, FRHistS, Dip ELH, Cert. Ed. Outreach Officer Miss A Siviter, BA, MA Volunteer Co-ordinator Mrs H Ault, BA Documentation Officer R C Watson, BA Wiltshire Archaeological & Natural History Magazine, vol. 101 (2008), pp. 1-7 The Win Green Mosasaur: the earliest British record? by Justin B. Delair A toothed skull of a previously unknown fossil animal was discovered in a chalk quarry near Maastricht, Holland in the late 18th century. With many post-cranial elements, the skull was excavated and taken to Paris. There, in 1812, Georges Cuvier figured and the described the specimen as a huge marine lizard but without name. In correspond- ence written to Cuvier during the early 1820s, William Conybeare proposed to name the lizard Mosasaurus. Cuvier referred to Conybeare’s authorship of Mosasaurus in print in 1824. Four years earlier, the local chalk had yielded two Mosasaurus-like vertebrae to Gideon Mantell near Lewes. Evidently aware in 1822 of Conybeare’s intention, Mantell could therefore record Mosasaurus as such from the Sussex chalk two years before Cuvier’s ‘official’ publica- tion of that name. An hitherto unpublished Rutter family memorandum now discloses the still older find of Mosasau- rus at Win Green, Wilthire, in 1818 - presently Britain’s earliest known discovery. Introduction In 1981 the late Aubrey Westlake of Godshill, Fordingbridge, presented the writer with a bound, but unpaginated, MS volume of miscellaneous geological memoranda written during the early 19th century by John Rutter (1796-1851) of Shaftesbury.' The volume had previously been owned by Aubrey Westlake’s geology-minded father, Ernest Westlake (1855-1922), into whose care it had passed when he married Rutter’s grand-daughter Lucy Anne in May 1890.’ This volume has recently been lodged in the library of the Dorset Natural History and Archaeological Society in Dorchester. The Rutter family: books and geology Before considering the memoranda of special relevance here, it will be useful to briefly note the connections former members of the Rutter family had with early 19th century geological enquiry generally and the associated popular literature. John Rutter was born in 1796 to Thomas (1741- 1800) and Hester (d.1806) Rutter of Bristol, the latter’s grand-father, Samuel Farley (d.1753), being a printer and book-binder there. Hester continued Samuel’s business activities following his demise in 1753,’ a point of some significance. John, therefore, was born into the world of book production, a business also prosecuted around 1797 at Lymington by another member of the Rutter family, Joshua Blake Rutter,* and by yet another, Clarence Rutter, at Shaftesbury during the 1820s and 1830s.° When aged 24, John married Anne Burchell Clarence of the Minories, London, and, twenty years later (1838), qualified as a solicitor.° By 1824 he and Anne were certainly living in Shaftesbury, where their fourth child, John Farley Rutter (1824- 1899) was born.’ John Farley Rutter seemingly entered the 4 Willowdene Close, Ashley, New Milton, Hants BH25 5BX 2 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 1 Reproduction of the record of the Win Green mosasaur discovery in the Rutter MS volume. family (?Clarence Rutter’s) printing and book- selling business at Shaftesbury, and, in July 1850, married Hannah Player Tanner (d. 1876) of Bristol, by whom he later had Lucy Anne. As mentioned above, Lucy Anne married Ernest Westlake in 1890, nine years before her father’s passing at Mere in neighbouring Wiltshire.® John Rutter’s personal interests antedating his marriage and his qualification as a solicitor are poorly recorded. Although conclusive proof is absent, it is very probable that it was he, rather than John Farley Rutter, who cultivated an interest in the geology recorded in the aforementioned memoranda; and that it was Anne (his wife), rather than Hannah, who was the ‘Mrs Rutter’ who exhibited various local Upper Greensand fossils at the Shaftesbury Literary Institute in the August of 1861.’ The Memoranda These consist of extracts from geological texts by other authors, various field notes (including several well sinkings), and a number of pencil drawings of local chalk and Upper Greensand fossils. Some original letters by contemporary ‘antiquaries’, e. g. Charles Hall (d. c.1859) of Anstey (Dorset), are also preserved at the volume’s hinder end. Although the great majority of the entries are written in a hand presumably that of John Rutter himself, a few indicate apparently different authorship. Not all entries are dated, but those that are the most recent were written in 1827. Collectively, the entries confirm that Rutter was familiar with the geological writings of Richard Pulteney (1730-1801),!° Joseph Townsend (1739-1816),'! James Parkinson (1755-1824),” and others, as well as with those instalments of James Sowerby’s multipart Mineral Conchology (1812- 1846) issued up to 1818;'? a date shown below to be of special significance. It can be reasonably assumed that, as a solicitor and a member of a family intimately involved with the production and selling of books, Rutter had access to publications of the calibre just mentioned. Furthermore, as a solicitor no doubt sometimes required to attend the county law courts at Salisbury and Dorchester, where geologically- minded individuals like the lawyer Charles Lyell (1797-1875) could be met, informed conversations on aspects of geology, as then perceived, could occur with beneficial results. The Win Green discovery Counting forwards from the MSs first page (unnumbered), the lower half of the 40" features drawings of two apparently associated fossil teeth accompanied by brief notes (Figure 1). The teeth are shown life-size, suggesting that they had been physically collected, while the notes (seemingly not in Rutter’s hand) state that they were found in Chalk at the base of a wall at Win Green in 1818. Their identity is queried as Mosasaurus. Win Green lies approximately 2 miles (3.2 km) south of Ludwell in Wiltshire and a like distance north of Ashmore in neighbouring Dorset. ‘Base of a wall’ is taken to signify foundations and, therefore, an THE WIN GREEN MOSASAUR: THE EARLIEST BRITISH RECORD? 3 excavation. Precisely which chalk zone yielded these teeth is unrecorded, a point exacerbated still further by the find-spot being similarly loosely recorded. Lower, Middle and Upper Chalk deposits are all present in, or near, the general Win Green locality (see Figure 1).'* On balance, the relevant zone will have been of late (Upper) Chalk age. The Win Green teeth Both teeth are strongly recurved and sharply pointed and exhibit the distinctive basal thickening typical of mosasaur laniaries. If accurately portrayed, the size of these teeth indicates that they were originally set in the jaws rather than on the palate (roof of the mouth) where mosasaurs had extra sets of teeth (Figure 2). Mosasaurs were large marine reptiles that flourished when the chalk deposits were laid down between 75 and 65 million years ago (Figure 3). In conforming so well with the known nosasaur record, the identity and geological horizon of the teeth as Fig. 2 Ventral view of a complete mosasaur (Platecarpus) skull showing the palatal teeth on the pterygoid bone in the roof of the mouth. After Williston stated in Rutter’s MS offer no grounds for doubting the one-time reality of the Win Green discovery. The alleged date of this find, 1818, and the main reason for undertaking this study, interestingly raises contentious literary discrepancies. These are considered below. An obscure fate Ernest Westlake’s still extant field notebooks record that, upon marrying Lucy Rutter, he incorporated Fig. 3 Late Cretaceous (Chalk) seascape depicting the probable life appearance of typical mosasaurs. 4 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 4 The celebrated ‘first’ mosasaur remains found at Maestricht. Length about 3 feet (Im). Note the striking similarity between the teeth in the jaws to those discovered at Win Green. The assitional palatal teeth are also well seen. After Buckland. into his own extensive collections John Rutter’s small one of fossils, or those specimens of it which then still survived.’ During the mid-20th century, the Department of Geology at Southampton University purchased the entire collection from Aubrey Westlake who had preserved it at Godshill. In the late 1990s, however, it was transferred to Oxford University, its present home. Exhaustive search of the collection at Southampton produced no trace of the Win Green teeth. It is possible, therefore, that (a) the teeth, if collected, were never an actual part of Rutter’s collection (the relevant note in the aforecited MS representing merely an ‘augmentary’ detail), or (b) they had been but, by 1890, had become lost and were thus never acquired by Ernest Westlake, hence no sign of them during the Southampton search. Furthermore, although it will be recalled (note 9) that other Rutter fossils (not mentioned by Westlake) were presented to the Natural History Museum in 1915, a search of the Rutter vertebrate fossils there, kindly undertaken on my behalf by the late Dr Alan Charig in 1982, revealed no trace of the teeth in question. Nor have the Win Green teeth been located in collections elsewhere. The Win Green discovery, therefore, is now known to us solely by the MS note and drawings featured here as Figure 1. A most famous fossil In 1770'° the toothed jaws of a large hitherto unknown animal (Figure 4) were found in Upper Chalk levels quarried at St Peter’s Mount, a suburb of the Dutch town of Maastricht in the Le, Oa Fig. 5 A typical mosasaur skeleton. This specimen (of Tylosaurus) is some 26 feet (8m) in overall length. The enlarged vertebrae near the end of the tail are believed to have supported a vertical tail fin. After Osborn. THE WIN GREEN MOSASAUR: THE EARLIEST BRITISH RECORD? 5 valley of the River Meuse. As the chalk there was already celebrated for its marine fossils (corals, echinoderms, etc.) it was quickly apparent that the toothed jaws belonged to some kind of undescribed marine creature of obvious scientific importance, A Dr Hoffmann, a French army surgeon who had previously collected fossils from the quarry, was duly notified of the find. On arrival, Hoffmann organised retrieval of the specimen, directing excavations so effectively that much of the animal’s post-cranial skeleton, until then unexposed, was successfully procured recumbent in a single block of chalk-rock. The skeleton conformed to a pattern unlike that of any other animal then known (Figure 5), and the creature’s identity soon became a matter of much speculation and some priority. At this juncture it should be stressed that, in 1770, the great sea saurians (ichthyosaurs and plesiosaurs) of Jurassic and Cretaceous times, now so familiar to us, had yet to be discovered and popularised. The late 18" century naturalists, therefore, had no firm conception that other equally large and impressive creatures had once co- existed with the Maastricht animal, a fact which, at the time, undoubtedly rendered the Dutch skull an intriguing novelty.” Without detailing here the subsequent legal wrangling over the rightful ownership of the fossil, suffice to say that, in 1795, following the French invasion of Holland, the by then famous skeleton was taken to Paris as a spoil of war and installed in the Museum of the Jardin des Plantes (now the Muséum national d’Histoire naturelle).!* Differing opinions In Paris the specimen soon attracted the attention of naturalists and geologists alike. The well known anatomist Peter Camper, for example, who had previously studied an allegedly similar jaw found elsewhere, which he concluded was that of an ancient whale related to the living Physeter, pronounced the Maastricht animal to be cetacean too.!? Faujas de Saint-Fond, then (1799) a professor of natural historyin Paris, reached sounder conclusions when, in acknowledging the specimen’s reptilian affinities, he identified it as a crocodilian.” Slightly over a decade later, the German savant M. Oppel thought the Maastricht skull resembled a lizard’s more than a crocodile’s, and accordingly referred it to the Squamata (lizards).”! In 1812 Georges Cuvier (1769-1832) not only confirmed Oppel’s classification but proved the Dutch reptile to have been some sort of marine saurian (unnamed) allied to the living monitor lizards.” The latter, however, being of terrestrial habits, differed greatly from the obviously sea-going Maastricht animal, a detail clearly emphasised by Cuvier’s excellent descriptions and figures of the Dutch specimen’s vertebrae. It was not until 1824 that the Dutch fossil was formally awarded the eminently suitable new generic appellation Mosasaurus (= Meuse - Roman Mosa, saurus = lizard),”? as the first of a varied family of specialised marine lizards of great size.” Many remains of these creatures would come to light in the world’s chalk deposits, to which they are confined, during the following century, and, indeed, continue to do so down to present times. Contradictory dates Although the official naming of Mosasaurus occurred in 1824, the name was actually used two years earlier by Gideon Mantell (1790-1852)” to identify two typical mosasaur vertebrae, now in the national collection, that had been discovered in the Upper Chalk at Lewes, East Sussex, in 1820.’° This discrepancy was briefly mentioned by the present writer in 2003.” Mantell was presumably cognizant in 1822 of Conybeare’s intention to name the Maastricht animal Mosasaurus; but, being evidently familiar with Cuvier’s earlier figures of the Dutch vertebrae, very probably cited the name Mosasaurus then more as a means of demonstrating that genus’s hitherto unnoticed presence in the British chalk, rather than as a deliberate nomenclatural pre-emptive measure. Such presumptions and actions (however plausible or well-intentioned) have nevertheless produced contradictory dates, with Mantell’s laying claim to being the earliest printed reference to the name. But even Mantell’s record that the first British mosasaur to be encountered dated from 1820 is now seen to be erroneous since, although he was apparently unaware of their discovery, the Win Green teeth were found two years before the Lewes vertebrae he highlighted in print. On the basis of the foregoing considerations, 6 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE therefore, Wiltshire now enjoys the distinction of having hosted the earliest known British find of mosasaur remains. Acknowledgements During admittedly sporadic and _ protracted enquiries, the writer has received encouragement and assistance from the following, whose participation (however sometimes peripheral) in this exercise is gratefully acknowledged: Professor Frank Hodson for unlimited access to the Westlake collection at Southampton, the late Dr Alan Charig of the British Museum (Natural History) for his search of the Rutter fossil vertebrates in the palaeontological collections, Dr Colin Forbes of the Sedgwick Museum, Cambridge, for drawing my attention to the early Elston and Conington discoveries, and Professor Denis Dean of the University of Wisconsin for helpful discussions about the earlier career and activities of Gideon Mantell. Mr Raymond Chapman has critically read the present text. References and Notes 1 W. Huggins-Wingate, The Family of Le Roter or Rutter (1966), privately printed, p.140. 2 J. B. Delair, Ernest Westlake (1855-1922), Founder Member of the Hampshire Field Club. Proceedings of the Hampshire Field Club 41, (1985), 37-44. 3 Huggins-Wingate, Op.cit. in note 1, pp. 38-39. J. Oldfield, Printers, Booksellers, and Libraries in Hampshire, 1750-1800. Hampshire County Council, Hampshire Papers 3, (1993), p.22. 5 E. Miles, Tisbury (Past and Present), 2nd.edn, (1920), p. 41. Huggins-Wingate, Op.cit. in note 1, pp. 40-41. Huggins-Wingate, Op.cit. in note 1, pp. 45-46. Huggins-Wingate, Op.cit. in note 1, pp. 46-47. Anon. A List of the Articles Exhibited in the Temporary Museum, at the Literary Institute, Shaftesbury, August 7th., 8th., and 9th. 1861. WANHM VII, (1862), 245-49; p. 249. NB. An imperfectly preserved tortoise from the Upper Greensand of Melbury Down (on the Wiltshire/ Dorset county boundary by Win Green) was among a collection of fossils of the ‘late John Rutter’ presented to the national collection by a Clarence E. Rutter in 1915, a benefactor clearly distinct from the Clarence Rutter resident in Shaftesbury during the 1820s and 1830s (note 5). This tortoise (specimen R.4214 B.M.), described in 1920 by C. W. Andrews, Notes on two Oo OND New Species of Fossil Tortoises, Annual Magazine Natural History 9 (5),145-150) as a new pleurodiran form Trachydermochelys rutteri, was not improbably that labelled ?7ronyx sp. among the Upper Greensand exhibits at Shaftesbury in 1861. 10 H. Pulteney (c.1801), Catalogue of Birds, Shells, and Rare Plants (with figures of Melbury Down fossils), in J. Hutchins, History of Dorset, 2nd.edn.(1813), 3 vols; see vol.3 (1% edn. 1773). 11 J. Townsend, The Character of Moses Established for Veracity as an Historian, recording events from the Creation to the Deluge (1813) 2 vols. 12 J. Parkinson, Organic Remains of a former World. An Examination of the Mineralized Remains of the Vegetables and Animals of the Antediluvian World; generally termed Extraneous Fossils (1804-1811) 3 vols. 13 J. Sowerby, Mineral Conchology (1812-1846) 7 vols. 14 C.R. Bristowe, C.J. Wood, I.P Wilkinson, and K.C. Dunham, Geology of the Tollard Royal-Tarrant Hinton district (1991). British Geological Survey Technical Report WA/91/20, onshore geology series 15 J. B. Delair, Ernest Westlake (1855-1922), geologist and Prehistorian with a Synopsis of his Field notebooks. Geological Curator 3 (2-3), (1991), 133-152; p. 151. 16 C. Lyell, Students’ Elements of Geology (1871) 1st edn; see pp. 359. Zittel, K. von, History of Geology and Palaeontology (1901), English translation of German original (1899) by M. M. Ogilvie-Gordon: see p. 107. N. Bardet and J. W. M. Jagt. Mosasaurus hoffmanni, le Grand Animal Fossile des Carriéres de Maestricht : deux siécles dhistoire. Bulletin Museé d’Histoire Naturelle (4th Series)18, (1996), 569-593. 17 This despite the discovery of a partial ‘plesiosaur’(?) skeleton at Elston, Nottinghamshire in 1719 (W. Stukeley, An Account of the Impression of the almost entire skeleton of a large animal in a very hard stone from Nottinghamshire, Philosophical Transactions of the Royal Society of London 30, 1719, 963-968, pl.), and of a 20 foot (6.8m) long skeleton of, probably, an ‘ichthyosaur’ at Conington, Huntingdonshire, during or before 1635 (R. James (1635) in T. Corser (ed.), Iter. Lancastrens, lines 324-326, (1845). As the first of these was identified at the time as a ‘crocodile’ and the latter as a ‘sea-fish’ or ‘whale’, the finds, though curious, were evidently not regarded as scientifically novel or especially important. 18 Understandably, a plaster replica of this historic specimen is one of today’s major exhibits at Maastricht’s Naturhistorisch Museum. 19 Camper’s deliberations were published firstin England (P Camper, Conjectures relative to Petrifactions fount in St. Peter’s Mountain near Maastricht. Philosophical Transactions of the Royal Society of London 76, (1786), 443-56) and then in Germany (P. Camper, Muthmassungen tiber im St. Petersberge bey Maestricht gefundene Versteinerungen (1788). 20 B. Faujas de Saint-Fond, Histoire Naturelle de la THE WIN GREEN MOSASAUR: THE EARLIEST BRITISH RECORD? 21 22 23 24 Montagne de Saint-Pierre de Maestricht (1799). M. Oppel, Die Ordnungen, Familien, und Gattungen als Prodrom einer Naturgeschichte derselben (1811). G. Cuvier, Récherches sur les Ossemens Fossiles (1812). lst edn, 4 vols; see vol.3. W. D. Conybeare, in G. Cuvier.. Op. cit. in note 22, (1821-1824), 2nd edn; see vol. 3 (1824). NB: Although J. Morris, A Catalogue of British fossils (1854) 2nd. edn; p. 351, correctly gives the date as 1824, E. H, Colbert, The Dinosaur Book (1951) 2nd edn; p. 114, erroneously records it as 1828. Among the best known of these are Leiodon, Tylosaurus, Platecarpus, Taniwhasaurus, Plotosaurus, and Clidastes. More recently, a new, but different, Mosasaur found at Maastricht has been named Prognathodon. Some of these genera reached lengths exceeding 30 feet (9m), the largest perhaps attaining lengths of around 50 feet (W. E. Swinton, Fossil Amphibians and Reptiles (1965), 4th edn, p. 119). G.A. Mantell, Fossils of the South Downs (1822); see pp. 93, 108, 242-246, pls xxxiil, xli, appendix p.362f. G.A. Mantell, Petrifactions and Their Teachings (1851); see p. 197 fig. 44, 198. J. B. Delair, Wiltshire’s Contributions to Early Geological Sciences: Beginnings to the Founding of the Society, in J.H. Thomas and L. Haycock (eds.) Wiltshire Archaeological and Natural History Society. The First 150 Years (2003), 137-182; see pp. 157-8 and 177. Wiltshire Archaeological & Natural History Magazine, vol. 101 (2008), pp. 8-17 Dry stone walls: a biological treasure? by John Presland This article looks at the biological value of dry stone walls. Dry stone walls are defined and, after a general account of their value for wildlife, the flora are described by reference to a study in the parish of Winsley, covering both the species present and the development of the characteristic communities of plants and lichens. The question of whether this community can be regarded as specific to dry stone walls is investigated by describing a local comparison with the floras of mortared walls, reference to plant community types defined nationally and the small amount of previous literature on dry stone wall flora. The plants of Winsley mortared walls are listed and a comparison with the dry stone walls shows a marked difference in the two communities. Whether or not this is typical must await studies over a wider geographical range. Finally, measures are suggested for conserving dry stone walls and their flora and attention is drawn to an advisory booklet on the subject. Introduction The Cotswold area of Wiltshire, in the north and west of the county, is well supplied with dry stone walls. Dry stone walls date back to the Neolithic period, for example at West Kennet long barrow, and many of those in our present landscape have stood at least since the 18th Century, when large numbers were built by farm labourers following acts of parliamentary enclosure. Their main function is to enclose fields and, in doing so, they mark administrative, ownership and agricultural boundaries. Dry stone walls are thus a feature of human activity in the landscape and can throw light on human cultures. Stephens (2007, 2008), for instance describes the work of Billingsley (1793) in the Mendips; he described various kinds of Mendip dry stone wall and compared the costs and advantages of these with hedges - an insight into the concerns of the enclosure period. A landscape survey on the website www.wiltshire.gov. uk emphasises the association of dry stone walls with human settlements in “Type 16 Limestone Lowland’, which covers an area of Wiltshire including the location for the Winsley surveys described below. Dry stone walls have been described as a remnant of a ‘hand-made world’, without mechanisation or prefabrication. Their history is, therefore, of interest to archaeologists. Dry stone walls are also important from the point of view of the landscape and a number of studies have been carried out in this respect. The Archaeology Service of Gloucestershire County Council Environment Department, for instance, is currently undertaking an historical landscape assessment of the Cotswold Area of Outstanding Natural Beauty (CAONB), which extends into Wiltshire. The Cotswold assessment is intended to provide a rapid view of the historical character of the Cotswolds landscape, based largely on written and cartographic sources. It categorises the present landscape in terms of the date of origin of the dominant landscape form, generally derived from the predominant enclosure pattern, and the principal historical land use from which it was derived. Enclosure in the area began as early as the 14th century, but before that, during much of the medieval period, the area was dominated by open 175c Ashley Lane, Winsley, Bradford-on-Avon, Wiltshire BA15 2HR DRY STONE WALLS: A BIOLOGICAL TREASURE? field agriculture. Again, the “Type 16’ survey above gives an even more local perspective. It is clear that dry stone walls are of interest across arange of disciplines. This article looks at them from a biological viewpoint. They are a sanctuary for wildlife, both plant and animal, and therefore of particular interests. The nature of dry stone walls, information about their biological significance and an account of recent investigations into their flora and its conservation are all considered below. What is a dry stone wall? Dry stone walls are normally the dominant field boundaries in areas where hedges are difficult to grow and stone is easily available. Helpful descriptions are provided by the British Trust for Conservation Volunteers (1994) and Brooks et al. (1999). Such walls are free-standing and rest on a foundation of large stones set in a trench. Layers of stone are then built upwards, without mortar, with the largest stones at the bottom and the smallest at the top. All stones are laid level or dipping outwards, and with stones in successive rows overlapping each other rather than straight on top of each other. Two such horizontal structures are built leaning towards each other, with a cavity between them being filled with stones of varying size to give a solid structure. At intervals, long stones (throughs) are laid across the wall to bond the two sides together. A line of coping stones is commonly laid on top. The illustration in Figure 1 is from a BTCV (1994) pamphlet. topstones or coping stones é oI er a o a (Srasy N LOT ULL —— 3 V \7 <7 (/ wedge | trench (section) Fig. 1 Dry stone wall structure (© BTCV) Details of design vary from area to area. Sometimes, throughs are replaced by bonders (or 3/4 throughs), which, as one might expect, extend about 3/4 of the way through the wall, and should 9 be adjacent and overlapping to provide the strongest structure. Coping stones may be omitted (Figure 2), or laid flat across the top. Retaining walls, where only one side is visible, are not included in the definition. Fig. 2 A newly built dry stone wall without coping stones Dry stone walls and wildlife Dry stone walls function much like hedges. They provide shelter for farm animals and varied habitats and micro-climates for wild animals and plants. The top is windswept, but the bottom sheltered. Inside such a wall it can be dry, with perhaps with a trickle of water. Collectively, dry stone walls can be seen as a substantial linear nature reserve and important to preserve. A well-maintained dry stone wall has a variety of internal spaces which provide shelter for spiders, woodlice, springtails, millipedes, bees, snails, toads, slow-worms, adders, voles, field and house mice, shrews, bats, rats, rabbits, hares, squirrels, weasels, polecats and hedgehogs. Foxes may dig into a wall to prey on its inhabitants. Birds such as wheatears, 10 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE robins, redstarts, blue and great tits, pied and grey wagtails, house and tree sparrows, spotted fly- catchers, nuthatches and little owls may nest in its holes. Upright coping stones provide perches or viewpoints. Like hedgerows, dry stone walls provide protected ‘corridors’ for small animals to move between areas favourable to them. External faces of walls provide a surface on which plants and lichens can grow and this aspect forms the subject of this article. The terms ‘flora’ and ‘plants’ are used, for simplicity, to include lichens and fungi, which are now regarded as being in different kingdoms from either animals or plants. A local study The flora of dry stone walls has been little investigated. Almost all work on wall flora is of a generalized nature (Darlington 1981; Segal 1969; South Court Environmental Ltd 1994), and either makes no mention at all of dry stone walls or treats them as though they do not need to be distinguished from those bonded with mortar. The only directly relevant sources that have been located are those by Payne (1989) and Williams (1988). In view of the above, a local survey of the flora was undertaken in an area of the parish of Winsley in West Wiltshire, about 6 miles from Bath at the southern end of the Cotswolds. To the author’s knowledge this study represents the first such systematic exploration of dry stone walls in the Cotswolds and the first to look in any detail at the lichens, mosses and liverworts which belong to this kind of community anywhere. A full account is provided by Presland (2008). The type of stone is a major determinant of the type of plant community found on walls. Limestone walls, such as those in the Winsley area, usually have a relatively rich flora, from lichens, through mosses and ferns to the higher plants. The evidence available at present (Williams 1988; C. Farmer pers. comm. 2007) suggests that the flora of dry stone walls built of acid rocks is completely different, but awaits its own investigation of the detailed type described here. Study methods For recording purposes, the study area was divided into stretches of roads with convenient landmarks, each referred to as a site. Walls were often covered by Ivy, Clematis, hedge shrubs obscuring or markedly shading the wall-top, or were concreted on the top. Only walls which had at least the tops bare and unshaded bore plants of interest. Records were made of plants growing on the tops of the walls or well clear of the ground on the sides. This provided a list of plants in the flora of this environment and the number of sites in which each occurred. An account of the flora in terms of these findings is given below. Popular names of bryophytes are included, although they are not universally accepted by bryologists. They are not vernacular names, but have been assigned by interested botanists, presumably in an attempt to make study of this group more accessible. Crustose lichens were not included in the systematic recording, partly because of identification difficulties, although they occurred wherever a wall was bare and they are pioneers in this kind of environment. Results Crustose lichens which had been identified previously by lichenologists from specimens from the study area include Caloplaca aurantia, Caloplaca citrina, Lecanora campestris and Lecanora calcarea, all typical of oolitic dry stone walls (Figure 3). There are also one or more unidentified black species. Overall, recording of species occurred at 27 dry stone wall sites. Nothing of significance was observed at two sites, giving useful records from 25. A total of 40 species of plants and fungi was recorded, though this was not the total actually occurring. The number of species recorded per site ranged from zero to 16. Nineteen sites had 10 or more species recorded, while a further four had five to nine species. This unpromising looking habitat therefore hosted a significant number of species, though not all are typically wall plants. The most common individual species recorded are listed below in order of the number of sites in which they were observed and the percentages of the 25 sites at which they occurred. Homalothecium sericeum (Silky Wall Feather-moss) - 24 sites (96%) Tortula muralis (Wall Screw-moss) - 24 sites (96%) Grimmia pulvinata (Grey Cushion-moss) - 23 sites (92%) Bryum capillare (Greater Matted Thread-moss) - 22 sites (88%) Geranium lucidum (Shining Cranesbill) - 21 sites (84%) Schistidium apocarpum (Common Beard-moss) - 20 sites DRY STONE WALLS: A BIOLOGICAL TREASURE? Fig. 3 Crustose lichens: Caloplaca aurantia (above), Lecanora species (below) (80%) Sedum acre (Biting Stonecrop) - 18 sites (72%) Orthotrichum anomalum (Anomalous Bristle-moss) - 14 sites (56%) Saxifraga tridactylites (Rue-leaved Saxifrage) - 14 sites (56%) Porella platyphylla (Wall Scale-moss, but actually a liverwort) - 13 sites (52%) Geranium pyrenaicum (Hedgerow Cranesbill) - 10 sites (40%) Fig. 4 A liverwort (Porella platyphylla) 11 Fig. 5 Pioneer mosses - Tortula muralis (above) and Grimmia pulvinata (below) All the above were predominantly observed on wall tops, except Porella platyphylla, which was mainly on the sides. Other plants seen largely on wall tops were: Rhynchostegium confertum (Clustered Feather-moss) - 5 sites (20%) Xanthoria parietina (a greenish orange lichen) - 3 sites (12%) Erophila verna (Common Whitlow-grass) - 2 sites (8%) Sedum rupestre (Reflexed Stonecrop) - 2 sites, a garden escape, thoroughly naturalised some distance from any garden at one site, but possibly introduced deliberately at the other (8%) Encalypta vulgaris (Common Extinguisher-moss) - | site, but found at another some years previously (4%) Galerina pumila (a fungus which grows on mosses) - | site, but noted at another site previously (4%) Other plants found mainly on the sides of walls rather than the tops (a more protected location) were: 12 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE a Fig. 6 Later mosses: Bryum capillare (above) and Rhynchostegium confertum (below) Polypodium interjectum (Intermediate Polypody) - 5 sites, 3 on the sides and 2 on the tops of walls (20%) Umbilicus rupestris (Wall Pennywort) - 4 sites (16%) Asplenium ruta-muraria (Wall-rue) - 3 sites (12%) Ceterach officinarum (Rusty-back) - 1 site (4%) Phyllitis scolopendrium (Hart’s-tongue) - 1 site (4%) Also found were: Cladonia pyxidata (Cup-moss, actually a lichen) - 5 sites, occurring where crumbling or distortion of walls created more or less horizontal surfaces lower than the wall tops and therefore providing more shelter (20%) Centranthus ruber (Red Valerian) - 1 site, on top of the wall (4%) A number of other species occurred either once or twice, but are not typical wall plants. Development of the flora Unstructured observation locally and a study of literature relating to limestone rocks and walls more generally (Darlington 1981; Segal 1969; South Court Fig. 7 Later lichens: Cladonia pyxidata (above) and Xanthoria parietina (below) Environmental Limited 1994) enable an account to be written of the succession of vegetation on Winsley’s dry stone walls. The first species to occur, the pioneers, are crustose lichens, which have flat bodies that cling closely to the surface to minimise exposure to wind and cold. They are found on the tops and sides of walls accompanied by particular species of moss, mainly on the horizontal surfaces - Tortula muralis (Wall Screw-moss), Grimmia pulvinata (Grey Cushion-moss) and Schistidium apocarpum (Common Beard-moss). The growth of these pioneers breaks off little fragments of rock, which combine with dead bits of their bodies and windblown particles of dust to form a simple soil upon which more demanding plants grow - mosses and lichens, a few fungi and a variety of flowering plants and ferns. One of these latter is the greenish orange lichen Xanthoria parietina (Figure 7). Further mosses also appear at this stage - mainly Bryum capillare (Greater Matted Thread- moss) and Orthotrichum anomalum (Anomalous Bristle-moss). As such soil becomes more substantial, further mosses appear on wall tops, including the rest of DRY STONE WALLS: A BIOLOGICAL TREASURE? Fig. 8 Flowering plants: Shining Cranesbill (top) and Biting stonecrop (below) those listed above. The liverwort Porella platyphylla (Wall Scale-moss) and the lichen Cladonia pyxidata (Cup-moss) also arrive, but mostly on the sides. The fungus Galerina pumila can occasionally be seen growing on wall-top mosses. Various flowering plants and a range of ferns also invade, as listed above. 13 Fig. 9 Ferns: Polypody (top) and Wall-rue (below) Do dry stone walls have a different flora from mortared walls? The question arises as to whether the flora described for dry stone walls is specific to that kind of environment or just part of a wider community type. If it is unique to this environment, its conservation 14 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 10 Plants of mortared walls: Ivy-leaved toadflax (above), Yellow Corydalis (top right) and Maidenhair spleenwort (right) is perhaps of even greater value. To address this issue, the flora of mortared limestone walls, also in Winsley parish, was investigated. Full details of the investigation and its results are provided elsewhere (Presland, forthcoming), although it was not as systematic or comprehensive as the dry stone wall study, observations being reduced to those essential to make a useful comparison. Mortared walls are typically built with stones of a regular, rectangular shape held together by mortar. These principal differences between mortared and dry stone walls might be expected to result in differing floras. Mortar is a source of nutrients and water and provides anchorage for roots and rhizomes, while regularly shaped stones of mortared walls fit closer together, with effects on anchorage and availability of nutrients and water. In the study area, more that 40 species were found on dry stone walls and approximately 50 species on mortared walls. In particular, the following plants were judged to be locally frequent on the mortared walls but were not noted at all in the dry stone wall survey: Asplenium trichomanes (Maidenhair Spleenwort) Campanula portenschlagiana (Adria Bellflower) Cymbalaria muralis (Ivy-leaved Toadflax) Lepraria incana (a lichen) - though it did occur on one non-retaining non-mortared wall in the mortared wall study area. Parietaria judaica (Pellitory-of-the-wall) Pseudofumania lutea (Yellow Corydalis) Plants common on dry stone walls but absent from mortared ones were: DRY STONE WALLS: A BIOLOGICAL TREASURE? Geranium pyrenaicum (Hedgerow Cranesbill) Orthotrichum anomalum (Anomalous Bristle-moss) Sedum acre (Biting Stonecrop) Particularly striking is the observation that mosses are surprisingly rare, even Homalothecium sericeum, Tortula muralis, Grimmia _ apocarpa, Schistidium apocarpum, and Bryum capillare, all common on the tops of dry stone walls. None of them occurred on more than three short stretches of mortared walls. The older mortared walls, which almost alone hosted plants, were almost certainly constructed of the same stone as the dry stone walls, and might be expected to have similar floras. It is therefore of interest to have found that certain species occurred only on dry stone walls, while others were on only the mortared walls. It appeared, also, that certain species occurred at different frequencies in the two environments. Perhaps a clinching comparison emerges from listing all species judged to be locally frequent on mortared walls, which amount to 11, and then identifying the 11 species recorded at the greatest number of sites on dry stone walls. These are shown in Table 1. There was no overlap at all, which is a strong argument for regarding the two communities as different. Table 1: acomparison of species on mortared and dry stone walls in Winsley Top 11 number of sites on| dry stone Homalothecium sericeum Tortula muralis Grimmia pulvinata Bryum capillare Geranium lucidum Schistidium apocarpum Sedum acre on mortared Parietaria judaica Cymbalaria muralis Pseudofumaria lutea Centranthus ruber Erysimum cheiri Asplenium trichomanes Asplenium ruta-muraria Ceterach officinarum Phyllitis scolopendrium Lepraria incana Campanula portenschlagiana |Geranium pyrenaicum Has a dry stone wall community been identified? At least locally frequent Clearly a difference has been demonstrated between dry stone and mortared walls at a local level. Is this situation found often enough elsewhere to allow it to be generalised, or is it purely local? As mentioned above, there is reason to believe that the findings do not apply to walls of acid rocks. Comparison with 15 a study of largely limestone dry stone walls in the Mendips (Payne 1989), however, shows a high level of agreement with the Winsley findings (for details see Presland forthcoming). It is, of course, conceivable that wall vegetation generally is not distinctive, but simply reflects the flora of the areas in which the walls occur, a claim made by Darlington (1981). At Winsley, however, most of the plants growing on walls do not occur in other situations, or only in very similar ones, such as a roof accommodating a colony of Biting Stonecrop. Further, comparison of the local dry stone wall flora with national categories of plant vegetation (Rodwell 2000) shows that it cannot be identified with any of them (Presland in prep.). Studies are needed in a wide range of areas to establish whether this is a typical pattern or a local phenomenon. Conserving dry stone walls The importance of dry stone walls both for landscape and archaeological reasons and for plant and animal life has been noted above. While this investigation gave reassuring results, if not for the survival of walls, then certainly for the integrity of their plant communities, investigation in other areas is less encouraging. A survey in the South Cotswolds (Dry Stone Wall Campaign, n.d.) showed that over 20% of walls are in very poor condition and in danger of collapse, and that over 20% have fallen down already. Although the Cotswold Area of Outstanding National Beauty (CAONB) has at least 4,000 miles of such walls (CAONB, n.d.), it is easy to see how this habitat is at increasing risk. Walls can become covered with Ivy preventing other species from growing; native shrubs can grow and shade out typical wall plants; walls can completely or partially collapse, or be removed by farmers; decreasing agricultural use has resulted in a reduction of people possessing the skills to build and repair in dry stone; stones are sometimes removed from walls to furnish rockeries; and soil or compost are sometimes spread on the tops of walls, presumably to deter theft. There are indications that the importance of this habitat and the risks to which it is subject are beginning to be recognised, and government action has taken place. The Department for Environment, Food and Rural Affairs (DEFRA) supervises the Countryside Stewardship Scheme. The Scheme offers payments to farmers and land managers to improve the natural beauty and diversity of the countryside under ten year agreements. In its tenth anniversary 16 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 11 Mature to degraded walls (good for flora), with the moss Homalothecium sericeum (above) and with Biting Stonecrop (below) year, 2002, the scheme saw the completion of 1,000 miles of dry stone walling restoration under the scheme (DEFRA 2002). Applications this year for the Countryside Stewardship Scheme exceeded 3,000, with the level of interest and commitment by farmers and other land managers high. While training is available for those wishing to build or repair dry stone walls, it is not clear how far this takes account of adapting the work to maximise survival and colonisation of the relevant plants. These issues are pursued below. The first step is to conserve as much of the habitat as is reasonably possible. The DEFRA scheme should plainly continue. Other bodies should also be encouraged. The Cotswold Rangers, for instance, have repaired and restored a number of walls in Winsley. There remains a significant role for public education and, it is hoped that, landowners and farmers not yet won over can be so in the future. When making conservation decisions, it is important not to rebuild walls without considering the implications for their flora. The most varied flora is found on older walls not in their optimal mechanical state (see Figure 11) and these are best left until there is a strong likelihood of their collapse. Walls below about 2 feet in height tend not to support significant wall vegetation, and can therefore be rebuilt or restored with less concern regarding flora. Ways of countering threats to the habitat need to be implemented. The growth of Ivy and scrub should be controlled; concreting of wall tops and spreading of compost or soil reduced; coping stones should not be overused; top stones with vegetation ought to be removed for rebuilding and replaced afterwards. A guide to conserving the flora of limestone dry stone walls is available (Presland 2007), based largely on the author’s studies in Winsley. It describes dry stone walls, gives an account of the flora of those built with limestone, shows, with accompanying colour photographs, how to recognise typical plants on limestone dry stone walls, and gives a systematic account of conservation measures. Where next? This account attempts to show the importance of dry stone walls both biologically and as important features of the historical landscape, and argues for increased efforts to conserve them. The study of dry stone walls provides a fruitful topic for further investigation, both in Wiltshire and elsewhere. Investigations and written accounts of their flora are continuing and progress may be tracked by visiting the website www.dry-stone-wall-flora.co.uk. References BILLINGSLEY, J., 1797, General View of the Agricultural DRY STONE WALLS: A BIOLOGICAL TREASURE? walls of the County of Somersetshire. Bath: Cruttwell BROOKS, A., AGATE, E. and ADCOCK, S., 1999, Dry Stone Walling. Doncaster: BTCV BRITISH TRUST FOR CONSERVATION VOLUNTEERS, 1994, How to Dry Stone Wall. Doncaster: BTCV COTSWOLD AREA OF OUTSTANDING NATIONAL BEAUTY, n.d., Cotswold Dry stone Walls: Conservation and Management. CAONB DARLINGTON, A., 1981, Ecology of Walls. London: Heinemann Educational DEPARTMENT FOR ENVIRONMENT, FOOD AND RURAL AFFAIRS, 2002, One Thousand Miles in Ten Years - another milestone for the countryside stewardship scheme: news release. London: DEFRA DRY STONE CAMPAIGN, n.d., Dry Stone Wails. Bristol: BTCV CREATE Centre PAYNE, R. M., 1989. The flora of walls in the Chew Valley. Somerset Archaeology and Natural History 133, 231-242 PRESLAND, J., 2007, Conserving the Flora of Limestone Dry Stone walls. Salisbury: Wiltshire Natural History Publications Trust PRESLAND, J., 2008. Dry stone walls in Winsley: their flora and its conservation. Wiltshire Botany 10, 23-28 PRESLAND., J., forthcoming, The flora of limestone walls: dry stone versus mortared. Submitted to BSBI News PRESLAND, J., in prep., Is there a limestone dry stone wall community? RODWELL, J. S., (ed), 2000, British Plant Communities Volume 5: Maritime Communities and Vegetation of Open Habitats. Cambridge: Cambridge University Press 17 SEGAL, S., 1969, Ecological Notes on Wall Vegetation. The Hague: Dr. W. Junk N. V. STEPHENS, C., 2007. The stone walls of Mendip Part 1. Waller and Dyker 10 STEPHENS, C., in press (2008). The stone walls of Mendip Part 2. Waller and Dyker 11 SOUTH COURT ENVIRONMENTAL LIMITED, 1994, What’s on a Wall? The Ecology of Walls. Northampton: South Court Environmental Ltd, 3" edn WILLIAMS, L., 1988, ‘Observations on the flora of wall habitats on Yell, Shetland’, in J. A. Fowler (ed.), Ecological Studies in the Maritime Approaches to the Shetland oil Terminal: Report of the Leicester Polytechnic to Shetland, August 1986 and Fuly 1987, 47-53. Leicester: Leicester Polytechnic Note: The booklet Conserving the Flora of Limestone Dry Stone Walls is available from either of the following: Dry Stone Walling Association of Great Britain (DSWA), Westmorland County Showground, Lane Farm, Crooklands, Milnethorpe, Cumbria LA7 7NH. Tel: 015395 67953. Website: www. dswa.org.uk. Email: information@dswa.org.uk. Summerfield Books, 3 Phoenix park, Skelton, Penrith, Cumbria CA11 9SD. Tel: 017684 84909. Website: www.summerfieldbooks.com. Email: info@summerfieldbooks.com. Wiltshire Archaeological & Natural History Magazine, vol. 101 (2008), pp. 18-50 An Iron Age and Romano-British settlement at Cleveland Farm, Ashton Keynes, Wiltshire by Andrew B. Powell, Grace Perpetua Fones and Lorraine Mepham with contributions by Stephanie Knight, Jacqueline I. McKinley, Chris Stevens and Nicholas A. Wells, and illustrations by Rob Goller Between 1984 and 1990, a sertes of archaeological investigations was undertaken in the area around Cleveland Farm, Ashton Keynes, Wiltshire (centred on NGR 40675 19450), in advance of gravel extraction. Fieldwork revealed evidence for occupation of the site from the Middle/Late Iron Age through to at least the 4th century A.D., with some indications of Early/Middle Saxon activity. Iron Age settlement comprised at least six small enclosures in the northern part of the site, interspersed with areas of open settlement. Environmental remains indicate a hedged landscape around the site, with little evidence for agriculture. Small scale iron smithing took place on the site, and two fragments of bronze-working moulds were also found. Local sources of supply predominate amongst the pottery and quernstone assemblages, but in the Late Iron Age more long- distance contacts are attested. There is little indication that the imposition of Roman rule was socially disruptive in the area, and settlement continued, after a shift in location to the south, through the Roman period, with a major restructuring in the 3rd or 4th century A.D. A range of structural remains, artefacts and environmental evidence illustrates the nature of the settlement at this period, with significant assemblages of pottery, coins, metalwork and animal bones, and with some preservation of waterlogged material. In contrast to the Iron Age, environmental evidence indicates the cultivation of spelt wheat and barley, both in the immediate vicinity of the site and in the wider landscape; the relatively high number of quernstones from this period could mean that the inhabitants were processing and redistributing cereals. Cattle still dominate the faunal assemblage, although there 1s some evidence for improvement of breeds, and perhaps for the export of hides from the site for processing elsewhere. Textile-working and other craft or industrial activities are only sparsely attested, although the quernstone assemblage includes intriguing evidence for ironworking. Coin copying may also have been taking place during the late Roman period, when coins also provide tantalising hints of a possible temple on the site. Personal items such as jewellery were common finds, and ‘luxury’ goods like imported pottery and glass suggest that the inhabitants of the site attained a relative level of affluence. A small number of inhumation burials, however, provided the only direct evidence of the population. Introduction a survey of the Upper Thames gravels (primarily from air photographs), undertaken in 1983-4 by Project background Richard Hingley on behalf of Gloucestershire and The area around Ashton Keynes was identified as Wiltshire County Councils. The survey, which one of national archaeological importance during expanded on the results of an earlier survey (Leech Wessex Archaeology, Portway House, Old Sarum Park, Salisbury SP4 6EB AN IRON AGE AND ROMANO-BRITISH SETTLEMENT AT CLEVELAND FARM 19 Cotswold Water Park (Ashton Keynes section) = oa Cerney \ Sf Wick Gloucestershire 195000: Cleveland Farm : 1 ages fo . ay LL a7? ES The Site a y 194000 / Ashton Keynes Wiltshire 4 493000 =) Ss = 9 = © o=S Key: (EH Features recorded during excavation/watching brief Feature taken from RCHME earthwork survey/cropmark plot Cleveland Farm Cottages Field D 194400 o 8 8 Oh, i=} 55 Seo oa Fle toy So fo oe £5 a3 oe E2 B83 At ofS 3 58 25 3 > ae a5 2) Eee es 5 25 gO 238 Sthe 3 8 5 No a2 z 2D ae gz 85 Sus og © | 2% oa a C2 =< 8 oe os a 8 fy ac) £3 RS so = 5 6 S as Ea Ou a & 2 =o Sin Zs a. Bs nz Scie a 38 re) $8 R82 af S as ae & 2 =s & 8 28 ag ca oo of ge 3s 2 o = a os Ee 2o 8 sa iat. Or Fig. 1 Site location and full extent of recorded archaeological features 20 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 1977), revealed cropmarks and earthworks covering c. 25 hectares adjacent to Bradley’s Pit, including a nucleated complex of enclosures and ditches south of Cleveland Farm, covering c. 5ha (centred on NGR 40675 19450)(Figure 1). As planning permission had already been granted for gravel extraction in the area around the farm, a programme of targeted archaeological investigations was undertaken between 1984 and 1990 in order to recover as much information as possible before the site was destroyed (Plate 1). The fieldwork, funded primarily by English Heritage, with contributions also from the developers E. H. Bradley Ltd and ECC Quarries Limited, was undertaken by (the Trust for) Wessex Archaeology with assistance from Reading University students and weekend volunteers. Funding for post-excavation analysis and publication was provided by the Aggregates Levy Sustainability Fund distributed by English Heritage. The site occupies a flat low-lying area c. lkm north of the River Thames, the river at this point being little more than a stream. The geology consists of drift deposits, including the first terrace gravels of the river and alluvium. Until gravel extraction commenced, the site was under permanent pasture (Fields B, C and D), resulting in a shallow (c. 0.1m thick) topsoil. Archaeological background Ashton Keynes is located within the boundary of the Cotswold Water Park, an area of 40 square miles of gravel quarries, now restored as country parks, nature reserves and lakes. A number of other archaeological sites are present in the immediate vicinity including Horcott Totterdown Lane, near Fairford (Pine and Preston 2004); Latton Lands (Powell and Laws forthcoming); Thornhill Farm, Fairford (Jennings et al. 2004) and those published as part of the Cotswold Water Park project: Claydon Pike, Somerford Keynes, Whelford Bowmoor and Stubbs Farm (Miles et al. 2007). The Iron Age settlement pattern in the Upper Thames Valley consists of open settlements with paddocks, enclosed farmsteads and short-lived, seasonal farmsteads such as Farmoor, located further down the Thames Valley (Lambrick 1992). A large banjo enclosure exists 3km to the west-north-west (Darvill and Locke 1988), and a bivallate hillfort at Ranbury Ring, c. 7 km to the north. The Cotswold Water Park sites lie in the hinterland of Roman Cirencester, and Cleveland Farm is less than 2.5km south-west of Ermin Street, the Roman road that connects Cirencester, located to the north-west, and Wanborough, to the south-east (Figure 2). The Iron Age and Romano-British settlement patterns in the Upper Thames Valley include evidence for periods of continuity but also disruption and landscape re-organisation. Middle Iron Age settlements include those at Claydon Pike and Thornhill Farm, although both sites produced little evidence of activity during the lst century B.C. During the first half of the Ist century A.D. a nucleated settlement was established to the south of the previous Middle Iron Age focus at Claydon Pike. The same period at Thornhill Farm saw a dramatic change with large rectilinear enclosures, a long, linear boundary and a loosely gridded enclosure system laid out. A major droveway put in place during the second half of the century suggests the movement of livestock on a relatively large scale (Jennings et al. 2004, 15). A radical reorganisation of the landscape can be seen at both sites in the 2nd century A.D. At Claydon Pike the enclosures were replaced by an aisled barn, aisled house and rectangular enclosures (Smith 2007, 25). The Thornhill Farm enclosures were superseded by a system of trackways, although there is no evidence of actual occupation during this period. During the 3rd to 4th centuries A.D. the major trackway at Thornhill Farm went out of use and was replaced by a number of linear boundaries (Jennings et al. 2004, 19). At Claydon Pike in the the 4th century A.D. the site was cleared and a modest villa constructed (Smith 2007, 25). Fieldwork The 1984 fieldwork (Bradley’s Pit) consisted of an earthwork survey and two evaluation trenches in the area south of the farm where features were most evident in the air photographs. Remarkably, many of the features identified survived as standing earthworks. Most archaeological sites in rural Wiltshire have suffered as a result of ploughing, with earthworks levelled and occupation levels eroded (Gingell 1976, 3). The preservation of earthworks at Ashton Keynes may be explained by the fact that the site was last ploughed in the 1940s and probably only occasionally before that (Wessex Archaeology 1984). The survey, covering c. 1.5 hectares, added detail to the cropmark data, identifying, for example, a number of slightly raised platforms and a series of low banks and ditches. The evaluation trenches were opened in order to confirm the date range and level AN IRON AGE AND ROMANO-BRITISH SETTLEMENT AT CLEVELAND FARM 21 a Gloucester (Glevum) Cirencester \w (Corinium) SX Totterdown Lane Q Latton Lands Somerford Key: Keynes © [1 Colonia > Tribal city © Town © Other Settlement A Kiln =— Roman road ==== Roman road (uncertain) s2sa Prehistoric track in contemporary use Lh White Walls Cleveland Farm Kempsford Stubbs Farm Roughground farm Whelford Bowmore O Claydon Pike Horcott Thornhill Farm OQ Cricklade Wanborough (Durocornovium) ‘A Whitefield Farm Base Information taken from 1994 Ordnance survey 1:625000 Roman Britain® map with the permission of the controller of Her Majesty's Stationary Office © Crown copyright. Fig. 2 The Romano-British context of Cleveland Farm, with sites mentioned in the text of preservation of the surveyed features, Trench A examining the edge of one of the apparent platforms and its surrounding ditch, Trench B being opened across the bank and ditch of an enclosure (Figure 5)(Wessex Archaeology 1984). Fieldwork resumed in July 1988 with a small- scale investigation of surviving trackways and boundary ditches to the east of Cleveland Farm (Field B). This was followed, in August, by further earthwork survey undertaken by the Royal Commission for Historical Monuments of England (RCHME), which indicated the full extent and complexity of the site (Coe et al. 1991, fig. 2). Between September 1988 and January 1989, work focused on the most northerly of the enclosures revealed in the air photographs (at NGR 4068 1945), most of whose interior was machine-stripped of topsoil. A roundhouse and other internal features, and nine ditch sections on the enclosure’s southern and eastern sides were hand-excavated, indicating a Middle to Late Iron Age date, while a further eight sections were machine excavated in an attempt to determine the line of the ditch on its northern and western sides (Wessex Archaeology 1989). Work continued through 1989 in the area to the south of the enclosure. The Ancient Monuments Laboratory (AML) conducted magnetometry and resistivity surveys in Field C, while Geophysical Surveys of Bradford Limited (GSB) undertook a limited resistivity survey in the southern part of Field D. Following this, a watching brief, covering c. Sha in Field C and the northern part of Field D (Coe et al. 1991, fig. 3), was carried out during topsoil stripping, with features being plotted using tacheometry. All visible archaeological features were planned at 1:200, and the physical relationships between some of extensive features determined. A sample of larger features, such as pits, sumps and ditches, was excavated to provide dating evidence and environmental data, and metal detecting was undertaken across the site, producing a large assemblage. At the same time, the main area of earthworks in the southern part of Field D was evaluated by a 25m grid of 1.4m wide machine trenches (totalling c. 1.2km)(Coe et al. 1991, fig. 4). This indicated two locations with high densities of artefacts in association with limestone rubble, the more 22 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE northerly of which, slightly raised, was examined in further detail over some 780m7. In the final season of fieldwork, in 1990, the southern part of the site, including much of the evaluated area, was subject to a watching brief covering some 2ha (Plate 1). This again revealed a high density of features that corresponded in many respects to the previously surveyed earthwork features. Metal detecting again produced a large metalwork assemblage. Cleveland Farm is clearly an important Iron Age and Roman site. However, only limited time and resources was available for the investigations. Given the scale of the archaeology a range of techniques was employed to maximise the level of data recovered. These included air photographic survey, earthwork survey, geophysical survey, evaluation, excavation, metal detection and watching brief. Accordingly, much of the site was recorded in plan only. A large finds assemblage was recovered, however. A number of specific problems have reduced the level of information that could be recorded, including the machining methods. During the 1990 watching brief the topsoil was stripped using a 360° excavator with a wide toothless bucket. In the previous year a-caterpillar box scraper had been used, with the result that most features were recognised only after both the topsoil and the subsoil had been removed, resulting in the loss of some of the shallower features. Only limited and localised phasing of the features is possible, and the development of the site Table 1: Finds totals by material type can therefore only be described in broad, general terms. The finds The archaeological investigations produced a large artefactual assemblage dominated by pottery but also including a wide range of other material types; there are significant collections of coins and other metalwork (consisting largely of metal-detecting finds), worked stone (primarily quernstones) and vessel glass. Object numbers were allocated to 4824 finds, including complete or nearly complete pottery vessels. Of these, 1449 are metal-detected finds, some 50% of which have been geo-referenced. Overall, 37% of the small finds can be located in plan. Waterlogged deposits in a small number of features have preserved organic remains comprising a small number of leather shoes (represented by bottom units), a bundle of coppiced twigs from Iron Age Enclosure 1, and a group of planks and stakes from Roman pit 3255. The date range of the assemblage extends from the Middle Iron Age to the post-medieval/modern period (with some earlier lithic material and a Bronze Age copper alloy pin), but the assemblage is predominantly Romano-British. Quantified records of all finds by material type within each context are held in the project archive, and summary totals by material type are presented in Table 1. Selected finds categories (ceramic building material, Iron Age pottery, quernstones, animal BRADLEYS PIT CLEVELAND FARM TOTAL Material Type Number Weight (g Number Weight (g Number Weight (g) Pottery 1438 3941 52,679 522,093 54,117 526,034 Ceramic Building Material |61 1560 1218 86,002 1279 87,562 Plaster/Opus signinum . - 8 Fired Cla : - 949 246 7312 841 Burnt Stone - - 1088 Worked Flint - - Burnt Flint ll 193 Glass 6 10 |Slag 9 35 Coins 110 - Metal 5 - Copper alloy 1 - Lead - - Iron 4 - Shale - - Wood - - Leather - - Worked Bone - - Human Bone - - 7 individuals |- + 355g - - redeposited Animal Bone 383 3772 30,129 Marine Shell : - 135 135 1038 AN IRON AGE AND ROMANO-BRITISH SETTLEMENT AT CLEVELAND FARM 23 bone) have been used as the basis for MA or MSc reports and theses (Universities of Sheffield and Southampton), and the results of these have been incorporated where appropriate in this report. Copies of the reports on pottery, quernstones and animal bone are held in the project archive, by permission of the authors. Further specialist finds reports were compiled as part of this latest stage of post-excavation work, and the results incorporated in this report; in the case of Iron Age pottery and animal bone these were updated versions of the existing MA report and MSc thesis respectively. These reports, together with selected existing catalogues, form part of the project archive, and are available on request (see end of report for details). The environmental material The calcareous gravel geology and the waterlogged nature of the clay-filled ditches at Cleveland Farm enabled the preservation of a suite of environmental data. The waterlogged and anaerobic clay deposits preserved limited amounts of pollen, waterlogged plant macrofossils and Coleoptera, as well as charred plant remains. A number of samples were taken during the various phases of archaeological investigation (1988- 90) for charred plant macrofossils, charred plant remains, waterlogged plant remains and pollen. The assessment of samples from the 1988-89 excavations represented a rapid examination of a small number of samples from an Iron Age ditch of Enclosure 1, while more detailed analysis has been undertaken for the Romano-British features excavated in 1989-90. Analysis of the charred plant remains formed the subject of an MSc thesis (University of Sheffield); a copy of this is held in the project archive, by permission of the author. A review of the results of this analysis, together with those of all other environmental assessments, is incorporated in this report, and is available on request (see end of report for details). Landscapes, settlements and chronology Possible long barrow While many of the features remain undated, most can be provisionally assigned to either the Iron Age or the Romano-British period on the basis of their form or location within the site. However, the arrangement of one group of undated features is closely suggestive of a small Neolithic earthen long barrow, comprising two c. 22m long slightly divergent ditches (811 and 812), 4m apart at the north-west and 5.5m apart at the south-east (Figure 3). Between them, towards the south-east (front) end, were two adjacent oval pits (629 and 630). However, sections through the ditches and the larger pit produced no dating evidence, nor any evidence such as human bone that might support this interpretation. Although there are no recorded long barrows in the area, Neolithic activity is indicated by a possible small causewayed enclosure at Down Ampneyc. 1.3km to the east-north-east, as well as by Neolithic pits and finds of flint and pottery (Oxford Archaeology 2005). While the position of this structure within the site might suggest that it was contemporary with the adjacent Iron Age enclosures, there are no obvious parallels for such an arrangement of features within the context of an Iron Age settlement. The landscape setting Evidence for Iron Age settlement, comprising six small enclosures and interspersed areas of open settlement, was found predominantly in the northern half of the site, i.e. north of the main concentration of earthworks that reflects the area of the later, Romano-British settlement. While it is possible that the intensity of Romano-British activity, as indicated by the substantial nature of the earthworks, may have erased or concealed traces of earlier activity, the northerly distribution of Iron Age pottery across the site suggests that this was not substantially the case. The Iron Age enclosures are located within a series of boundary ditches. To the north of the enclosures, a long winding ditch (not numbered), recorded for c. 400m, appears to form the northern boundary of the Iron Age settlement. This ditch also appeared to be linked to an extensive complex of enclosures and boundary ditches to its north, visible as cropmarks. To the south of the settlement a long straight ditch (748) ran east-west for over 300m. At the west it ended at, and appeared to form either the southern side of, or an internal division within, the most westerly of the Iron Age enclosures (Enclosure 4). An un-numbered ditch that ran from the north- west of this enclosure created a wide funnel-shaped arrangement with the western end of the northern boundary. The eastern end of ditch 748 joined the northern side of the Romano-British trackway that continued eastward beyond the excavation area. While most of the Iron Age features lie north of 24 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE this ditch, several others, including a Late Iron Age roundhouse gully, lay to the south. Air photographic survey, however, revealed a clear distinction between the area north of the ditch, where the individual Iron Age enclosures were largely discrete elements within the landscape, and that to the south where the various features were clearly interconnected within a tightly knit, nucleated settlement. This raises the possibility that ditch 748 was laid as a formal boundary between the area of previous, now abandoned settlement, containing all the enclosures (if not quite all of the roundhouses), and a new, post- conquest settlement site constructed on both sides of the ditched trackway (from which another track ran south towards the river). Ditch 748 was not dated, but clearly relates to large-scale re-organisation of the landscape at the start of the Roman period involving laying out of a predominantly rectilinear array of field boundaries containing a new settlement. Other gullies or ditches may represent the changing layout of the settlement during the Middle and Late Iron Age. One, ditch (639), ran north-east to south-west for over 200m along the eastern edge of the site, while another (813), of similar length, ran on a slightly wavy curving line towards the north, where it petered out, the two ditches almost converging at the south. At its northern end, ditch 813 may have formed the eastern side of Enclosure 2, while towards the south it appeared to cut a 2m wide pit (658). The ditch produced only three sherds of Middle/Late Iron Age pottery. Ditches 639 and 813 were both cut by ditch 748. A Romano-British trackway, possibly following an earlier line, dog- legged around the settlement to the south. These features help place the Iron Age settlement within a wider landscape that contained trackways for the movement of livestock and long boundary/drainage ditches, possibly defining enclosed fields or areas of open pasture. The area of Iron Age settlement, including the small enclosures within it, was clearly closely associated with these features, which air photographs suggest were part of an extensively occupied landscape. The settlements The Iron Age enclosures (Figure 3) There were four clearly identifiable Iron Age enclosures (Enclosures 1-4), with a probable fifth being represented by an L-shaped ditch (823) anda possible sixth formed by ditch 816, although the latter lay only partially within the area of investigation. The enclosures vary in form and content and, while there are minor distinctions that can be made in their dating within the span of the Middle and Late Iron Age, it is not possible to ascertain to what extent they were contemporary or successive features. For this reason they are described below in the order that they were numbered. A number of features, such as the ring gullies of roundhouses, were located within enclosures; others lay outside or were positioned on the line of the enclosure ditches, indicating settlement pre-dating and/or post-dating enclosure construction. Although it has not been possible to securely identify stratigraphic associations, inferences can be made on the basis of their layout. Enclosure 1 The most northerly enclosure was sub-rectangular in form, measuring 38m east to west. The first interim report (Wessex Archaeology 1989) states that only its eastern, southern and western sides (ditch 113) were recorded, and that ditches exposed in the machine sections to the north were probably of later date, although possibly masking an earlier ditch. The north to south dimensions of the enclosure can therefore only be estimated, at 35m. It is clear from the air photographic and earthwork surveys, however, that at its north-west corner the enclosure abutted the long, northern boundary ditch (not numbered) which may have formed its northern side. The enclosure had at least one entrance, c. 3.6m wide, at the south-east corner, with another suggested by a possible ditch terminal at the north- west corner, and a third, possibly blocked, entrance suggested by a sharp narrowing of the ditch at the south-west corner. The ditch was extremely irregular in plan and profile, varying between 1.25m and 4m wide, and 0.6m to 1m deep, with moderately steep sides. Close to the western terminal of the south- eastern entrance there was a distinctive black organic fill at the bottom of the ditch, but elsewhere along the south side it contained a gravelly lower fill, then a layer of clay containing varying amounts of gravel with an upper fill of clay loam, often sealed by a clean layer of clay. There was clear evidence along its eastern side, however, that ditch had been cleaned out by a steep-sided, flat-bottomed re-cut, presumably to aid drainage, which also contained the black organic fill at its base. Waterlogged conditions at the base of the ditch led to good preservation of a range of environmental data, including plant macrofossils, Coleoptera and 25 AN IRON AGE AND ROMANO-BRITISH SETTLEMENT AT CLEVELAND FARM Kemyoes G eunsojoug g ainsojouz (mong suo] ajgissod osjp) quawmazuas uado pup sainsojIua asp UOLT ¢ ‘BLT qz eunsojouR oe) SoZ O09PEL ¢ einsojoug € eunsojoug SUI|INO A}IS 0} UONE{aJ Ul eINBy UIeEW jo easy aunyonsys }SOd BAY/NOY ft: Aanins FWHON Woy uaye} aunjeay aunjea}y payepuy anes) aby uo) ‘hey 26 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE charred plant remains, as well as animal bone and molluscs. Among the finds was a bundle of straight hazel twigs (or possibly willow), on average 4.8mm in diameter and c. 500mm long, found on the base of the ditch near the enclosure’s north-east corner. It is evident that the twigs were coppiced, probably from one to three year-old stools, indicating a level of woodland management. They may have been used as roofing materials, or for basket-making. The enclosure contained a roundhouse in its drier, north-west corner, the western half of which had been truncated during topsoil stripping. The roundhouse was defined on the eastern side by ac. 10m diameter gully, up to 0.45m wide and 0.2m deep, with a short length of concentric gully suggesting two phases of construction or repair. Ac. 3.5m gap in its eastern side may represent the roundhouse entrance. Despite the fact that a number of postholes were located within the interior of the structure, they do not form any clearly recognisable post-built structure. It is uncertain whether a length of straight gully cutting across this gap at a tangent to the circle was part of the roundhouse structure. A gully (163) that ran | approximately east-west across the roundhouse and the eastern half of the enclosure appears to be a later feature, as it cut the ditch fills. Several other features were located within the enclosure, mostly clustered in the south-eastern area. They include gullies 126 and 361, both c. 3.6m long, that ran inwards from the enclosure ditch, apparently cut by its inner edge. A third similar feature (124), 2.7m long, contained a nearly complete Middle Iron Age barrel-shaped jar (Figure 4, 1) as well as smithing slag and a little hearth lining. A small number of other features excavated in this area contained single fills of clean clay and were therefore thought to be of natural origin. Re-cutting of the enclosure ditch and two phases of roundhouse construction may be reflected in the 0 50mm Fig. 4 Iron Age pottery: (1) barrel-shaped jar, shelly fabric, Object Number (ON) 2, context 157, gully 124, Enclosure 1; (2) Dressel 1B amphora rim, context 723, pit 721 date range of the pottery and other finds recovered. The majority of the pottery from the ditch (c. 80% by weight) was of Middle Iron Age date, although the re-cut produced Middle-Late and Late Iron Age pottery. Features within the interior had a similar date range - a number of the postholes (as well as feature 124) produced Middle Iron Age pottery, while features associated with the roundhouse were weighted more towards the Middle-Late Iron Age. This may indicate that the construction of the roundhouse was related not to the initial phase of the enclosure, but to the subsequent re-cutting of AN IRON AGE AND ROMANO-BRITISH SETTLEMENT AT CLEVELAND FARM 27 its ditch (and possibly the blocking of one of its entrances). Other finds from the first phase of the enclosure ditch included a small fragment of a bronze mould recovered from the lower ditch fill, and a Nauheim brooch (Ist century B.C., Plate 2, Object 3) and shale armlet fragment (not illustrated) from the upper fill. The upper fill also yielded a fragment of shale bracelet. Two quern fragments in a very hard fossiliferous limestone were recovered from the upper fills of the ditch recut. Enclosure 2 Enclosure 2 lay 30m to the south of Enclosure 1. The arrangement of ditches suggests two distinct phases, whose chronological relationships, including those with a large penannular gully at the north of the enclosure (789), were not ascertained. Enclosure 2a comprised three straight sides at the south-east (840), south-west (788) and north-west (795), creating an area of 28m by 45m. The ditches were quite regular in plan and profile, 1.1m-1.5m wide and 0.4-0.5m deep. It is not certain whether the east side was left open, or if a fourth side was formed by the longer north-south ditch 813. Ditches 840 and 813 each produced three sherds of handmade Middle Iron Age pottery, although of those from ditch 813, one is a Late Iron Age wheel-thrown sherd and suggests a slightly later date for this feature. If ditch 813 was utilised as a fourth side, gaps of 8.3m at the south and 5.6m to the north would have formed entrance points, the southern entrance being partly blocked by a 4m length of ditch (652). A 2m wide pit (841), located on the line of the ditch at the south appeared to cut it and produced three sherds of Middle Iron Age pottery. The other undated phase (Enclosure 2b) appeared to be a larger, four-sided enclosure, with a straight ditch along the north side (792) located across the line of ditch 795 and, partly concealed by a hedge line, ditches along the north-west and south-west sides (not numbered). This enclosure presumably incorporated ditch 840 at the south-east, thereby creating maximum dimensions of 67m east-west by 43m north-south, with a 3.5m wide entrance at the south. It is unclear if ditch 813 formed a fourth side to this phase of the enclosure. Within both enclosure phases, abutting both ditch 813 and the northern side of Enclosure 2b, was a penannular ditch (789), although its stratigraphic relationship to neither ditch is recorded. The ditch was 1.80m wide and 0.85m deep and described a circle c. 18m in diameter with a 6m wide gap on the south-east side facing towards the south-eastern entrance of Enclosure 2a. Given its form and orientation it is likely that the ditch surrounded a roundhouse, although no recognisable structure could be discerned in the 11 postholes recorded within it. Both the ditch and the postholes produced considerable amounts of burnt limestone. Twenty- five sherds of Middle Iron Age pottery were recorded from the gully. The same fabric types were seen in the penannular gully, ditch 840 and ditch segment 652. Close to the penannular gully, and within Enclosure 2a, were a sub-square pit (796) c. 1.6m across, and two four-post structures, c. 1.5m and 2.5m square, of a type frequently interpreted as granaries. There were also a number of other postholes of varying size, whose distribution extended into the western part of Enclosure 2b, some forming possible pairs, perhaps supporting looms or drying racks, but others forming no recognisable structures. Also in Enclosure 2a were a number of short lengths of gully, one measuring 4m long with a small pit or posthole at its eastern end; similar combinations of features of unknown function were recorded in Enclosures 3 and 4, and one at the eastern end of the site. Enclosure 3 Enclosure 3 was 15m north-west of Enclosure 2a, but just 4.5m from Enclosure 2b. It was ‘shield-shaped’, almost square at the north-east but rounded at the south-west, measuring c. 38m by 28m, and with a 1.6m wide entrance midway along its south-east side. Its regular ditch (701) was 2m wide and 0.6m deep. A semicircular gully (742) formed the south- western arc of a circle c. 13m in diameter in the south-western half of the enclosure, probably the remains ofa truncated roundhouse. A clear terminal to the gully at the south-east, facing towards the enclosure entrance reflects the often typical orientation of roundhouse entrances. While there was a cluster of postholes and stakeholes within the arc, it formed no recognisable structure. Both the roundhouse gully and the enclosure ditch contained Middle Iron Age pottery (the latter also producing a presumably intrusive sherd of Ist/2nd century A.D. coarseware). Lying across the projected line of the roundhouse gully, and therefore unlikely to be contemporary with it, was a gully (743) with a posthole (757) at its end, comparable in arrangement to those recorded in Enclosure 2. The gully, which was 6m long, 0.5m wide and 0.7m deep, with large quantities of limestone from its upper fills, produced Late Iron 28 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Age pottery, while the posthole contained two sherds of Middle Iron Age pottery, and may instead have formed part of the roundhouse. A short length of gully (755) towards the north-east of the enclosure produced seven sherds of Late Iron Age pottery. Enclosure 4 The most westerly Enclosure (4) was rectangular in shape, measuring 54m by 27m, its long axis aligned north-south. The enclosure ditch (746) was generally uniform around its circuit, !.5m wide and up to 1.15m deep, but unbroken by any entrance. No structures were discernible among the various postholes, pits and lengths of gully (although one slightly curved length of gully had a projected diameter of c. 16m) recorded in the interior. Features were concentrated in the centre of the enclosure and appeared to be bounded to the north by a length of straight gully, apparently an internal division. Pottery from the enclosure ditch and a short length of straight gully (756) indicated a Middle Iron Age date. The enclosure was connected to a number of ditches or gullies, although its stratigraphic relationship to these was not established. As noted above, a ditch running from the north-west corner of the enclosure formed a funnel with the northern boundary ditch. As this enclosure was sited at the mouth of that funnel, it may have had some function related to the movement of livestock. Air photographs indicated that the ditch forming the southern end of the enclosure continued eastward by the long straight ditch 748 (above), although during the watching brief it became evident that the enclosure either extended some 15m further south to abut the trackway (and had been bisected by ditch 748), or that a small annexe had been added at that end at an unknown date. Other possible enclosures Towards the east of the site were two roundhouse gullies (764 and 770), 20m apart, possibly associated with an L-shaped ditch (823). The ditch had a rounded terminal at the south-east, but petered out at the north and had the appearance of forming two sides of an ‘open’ enclosure. Roundhouse gully 764 at the northern end of the ditch produced Middle Iron Age pottery and was 13.5m in diameter, with a wide gap on the south-east side. Gully 770 was 12m in diameter, probably with an east-facing entrance. Close to the roundhouses and within the area defined by the two arms of ditch 823 were a number of other small features including two postholes and an angled length of gully that may have formed a small structure outside the entrance to roundhouse 764. Three pits (631, 831 and 832) produced Middle Iron Age pottery, and one of them (831) a piece of briquetage. Pit 631 cut roundhouse gully 764, while a large pit (817) at the north end of ditch 823, also cutting roundhouse gully 764, produced sherds of Middle/Late Iron Age pottery, a fragment of bronze mould and another piece of briquetage. To the east of this group were a number of isolated postholes and lengths of gully, one of which (774) produced an almost complete smithing hearth bottom. To the north was a 35m long ditch (816) turning at either end to the north and north-east, extending beyond the 1989 excavation area; it had the appearance of the southern end of a sixth enclosure, although no such feature was visible in air photographs. Unenclosed Iron Age roundhouses and other features In addition to the five suggested roundhouses already described, a further possible ten were located during the watching brief across Fields C and D, one of which dates to the Middle Iron Age period (797) and two to the Late Iron Age (705 and 825). The earlier, Middle Iron Age structure is located in the northern part of the site, whilst the later structures are to be found further south, close to the line of boundary ditch 748. The structures were indicated by the presence of penannular gullies, thought to represent the drip gullies around roundhouses, of between 7m and 14m diameter. In most cases these were 0.2-0.25m wide and 0.1-0.15m deep and, where discernible, the entrances face between south and east, with the exception of south-facing Late Iron Age gully 825. Middle Iron Age roundhouse 797 lay immediately north of Enclosure 2 on the line of the northern end of ditch 813. The gully was 11m in diameter and almost complete apart from on the eastern side where the circuit was made up of two short lengths of gully with intervening postholes. A line of four postholes across the interior may have formed part of the structure. An arrangement of five posts in a small east-facing arc just 2m across clearly formed a structure, of unknown function, immediately south of the roundhouse. In the same general area, a very small oval penannular gully (not numbered), 4m by 5.5m, also lay on the line of ditch 813 as well as on the north- east side of penannular ditch 789 in Enclosure 2. It was too small to have been a domestic structure and is likely to have served an ancillary function. Immediately to the north of roundhouse 797 were two arcs of ditch (801 and 802) describing circles AN IRON AGE AND ROMANO-BRITISH SETTLEMENT AT CLEVELAND FARM 29 c. 15m in diameter. These lay at the northern edge of the 1989 excavation area and, although their function is unclear, it is possible that they are further roundhouses. South of Enclosure 2 was a cluster of at least five quite variable circular or oval structures defined by gullies, only one of which could be dated (to the Late Iron Age). This was a north-south aligned oval gully (825), measuring 13.5m by 9.3m with a posthole (826) at the centre and a south-facing entrance, which produced a Dressel 1B amphora spike. To its south, was a circular gully (657), c. 10m in diameter, with a wide south-east facing entrance. Also in this group was a small double ring gully (829), with a 7.5m diameter outer gully and a 5m diameter inner gully; it was cut through the centre by ditch 748 so that although the terminals of an east-facing entrance were recorded for the outer gully, none was visible for the inner gully. Other features were an oval gully, 6m by 7m with a gap at the eastern end, and two opposed arcs describing an oval c. 5m by 6.5m, as well as other lengths of curved gully, pits and postholes. Between Enclosures 3 and 4 was a pair of roundhouse gullies, 9m apart. Although only the larger (705), at c. 11m in diameter, could be dated (to the Late Iron Age), gully 735, measuring 9m in diameter, was very similar in form and is likely to have been broadly contemporary. Both had entrances facing approximately east, and both had central, stone-lined hearths. Each contained a few postholes, insufficient to indicate their construction, although 705 had a pair of postholes in the entrance, set just back from the line of the gully. Immediately outside the pair, on the line of the gully and therefore unlikely to be directly contemporary with the structure, was a 1.3m diameter pit (721) containing an interesting ceramic assemblage. This latter included three Dressel 1B amphorae rims (Figure 4. 2), as well as 16 sherds of grog-tempered pottery of which two were identified as belonging to a necked bowl/jar, and nine sherds of sandy ware. The pit also produced a smithing hearth bottom, several amorphous lumps of smithing slag, and fuel-ash slag. Dressel 1B amphora are relatively uncommon in the region and the presence of three vessel rims in a single pit is unusual. West of Enclosure 4, a cluster of postholes and short gully segments, three forming an arc describing acirclec. 7-8m in diameter, and others lying parallel to each other on its eastern side (possibly a porch), may represent one or more possible structures, although their form is unclear. One irregular gully (754) produced 121 Middle Iron Age sherds from two neckless ovoid jars, one lid-seated jar and one barrel- shaped jar. It may also be noted that a Bronze Age ‘Picardy’ pin (Hawkes 1942) was found unstratified by metal-detecting in this area. The Romano-British enclosures/compounds (Figure 5) During the early Roman period the settlement shifted to the south, to an area demarcated by ditch 748. Unfortunately it has not been possible to interpret the layout of the Romano-British settlement, despite the presence of an extensive network of features visible as cropmarks and earthworks. The archaeological features were plotted only by tacheometry during the watching brief and very limited excavation took place. Furthermore, a significant part of the area north of the trackway was recorded only in the trenches of the 1989 evaluation. With few stratigraphic relationships recorded, phasing relied mainly on pottery dating as outlined below. Interpretation is also complicated by the three methodologies by which the settlement was recorded. While air photographic survey, earthwork survey and watching brief revealed comparable and complementary overall layouts for the settlement, there were also significant differences. This is to be expected as, for instance, alluvial deposits sealed some features preventing their showing as cropmarks. Similarly, the apparent intensity and longevity of Romano-British settlement activity, as revealed by the watching brief, will have resulted in some earlier earthwork features being modified or levelled, so that the earthworks that survived are more likely to reflect the later phases of the site’s occupation. What the watching brief revealed, however, was evidence for repeated modifications to the layout of the Romano-British settlement as indicated by the many intercutting features, including ditches, gullies, pits and ponds, as well as continuity of use evident in multiple re-cuts recorded in the ditch sections. Many of the ditches appeared to form part of a system of fields and paddocks linked to the trackway located in the north-west of Field D. Some of the larger ditches formed sub-rectangular enclosures with rounded corners similar in size and shape to the Iron Age enclosures to the north, although here they appear integrated within the tightly-knit nucleated settlement. Two of these, flanking the north and south sides of the trackway, were examined by single evaluation trenches in 1984 (Figure 5, Trenches A and B). 30 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Key: RN Round and possible rectangular structures f) Other feature Feature taken from RCHME survey Area of main figure in relation to site outline Track Way | f Trench B Trench A Fig. 5 Area of 1990 watching brief highlighting possible roundhouses and rectangular structures AN IRON AGE AND ROMANO-BRITISH SETTLEMENT AT CLEVELAND FARM 3] Trench B The larger sub-rectangular enclosure flanking the north side of the trackway was only partially exposed during the watching brief, but is visible in air photographic and earthwork surveys. It had a bank surviving to a height of 0.2m built up of material excavated from a 1m deep internal ditch from which large quantities of animal bone were recovered. The ditch had been recut to aid drainage with the upcast dumped in the interior and a substantial post of uncertain function set in its base. Earthwork survey suggested an entrance off the trackway at the south- east corner, while air photographs indicated a series of possible internal divisions within the enclosure. Trench A A smaller enclosure on the opposite side of the trackway was defined by a 1.1m deep ditch that was subject to periodic flooding and had been recut three times. There were postholes, stakeholes and a possible ditch or gully in the interior of the enclosure. Two of the enclosure ditch recuts produced small quantities of Romano-British pottery, dating no earlier than the early 2nd century A.D. Two 3rd century A.D. coins were also recovered from one of the ditch recuts; other finds comprised limestone fragments, pieces of dressed stone and a few tile fragments. The Romano-British structures (Figure 5) Five possible structures were recorded in 1990 in the area of the main earthworks; three of them south of the trackway. Their location south of ditch boundary 748 clearly reflects a southward shift in settlement fromthe Middle Iron Age to the Romano-British period. The watching brief revealed a roundhouse located to the north of the trackway and within the sub-rectangular enclosure exposed in Trench B. The structure was represented by two gullies, the inner one (5139) cut at the north by an otherwise concentric outer gully (5153), suggesting two phases of construction or an episode of repair. Together they described a semicircle, and while it is reasonable to infer that the structure was a roundhouse c. 7m in diameter, a line of four postholes running from the southern terminal of gully 5139 across the open face of the semicircle could indicate a D-shaped structure, or an internal division within a roundhouse. The roundhouse is undated but spatially related to the network of Romano-British features in Field D, indicating the continuation of roundhouse construction into the Roman period. Immediately to the north-west was a small rectangular structure, 12m by 10m. Possible roundhouse 6118 also lay north of the trackway and was recorded during the 1989 evaluation and 1990 excavation. This undated structure was c. 7m in diameter and represented by three short lengths of gully with a 4m wide entrance at the south-east, set back within which was a large posthole. Despite this structure’s relatively small size, there was a 1.5m wide stone-built hearth (6110) off-centre within it, raising questions about its viability as a domestic structure. There was considerable variation in the possible roundhouse structures recorded south of the trackway, of which two are dated as Romano-British. One (not numbered) consisted of a penannular ditch c. 9m in diameter with a 4.5m wide gap on the east side; no internal features were recorded. A late Roman penannular brooch was recovered by metal detecting from the pond at the eastern side of this ditch. A further possible structure (6303) consisted of a 5m curved gully, the upper fill of which contained a substantial quantity of large stones among which were 25 sherds of mid 3rd/4th century A.D. coarseware pottery. The stones appeared to have fallen into the gully and were not considered to bea wall, although a further small spread of stones c. 8.5m to the south-west was interpreted as representing the opposite side of the structure. Between them was a group of at least seven flat stones which, although three were burnt, showed no evidence of in situ burning as might be expected with a hearth, and they may instead have been a post pad for a central timber. Interpretation of these features as a roundhouse remains highly tentative. The most complete evidence for a Romano- British roundhouse lay in a small oval ditched enclosure south of the trackway towards the west of the settlement. The 1989 evaluation revealed a slightly raised area with a high density of artefacts in association with limestone rubble. During machine stripping large quantities of artefacts were recovered from the turf and upper part of the topsoil. In parts, the topsoil overlay natural gravel, but elsewhere it overlay a 0.1m thick layer of ‘dirty gravel’ that sealed archaeological features and contained further quantities of artefacts and rubble. Removal of this layer revealed several incomplete overlapping circular gullies representing at least three phases of roundhouse construction. One (3134), represented by two lengths of gully on its south and north- west sides, was c. 7.5m in diameter and produced Romano-British pottery of lst/2nd century A.D. 32 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE date. This was succeeded by a larger structure (3132), with a more complete gully c. 9.5m in diameter that produced pottery of late 3rd/early 4th century A.D. date. A 4th century coin was recovered from this gully (364-375 A.D.). The third structure (3225), whose chronological relationship to the other two could not be ascertained, would also have been c. 7.5m in diameter. Within these arcs were a number of small pits and postholes, none of which could be associated by phase, although several appeared to either cut or be cut by lengths of gully. A sequence of three roundhouses is suggested by the relatively long time span of the pottery from gullies 3134 (1st/2nd century A.D.) and 3132 (late 3rd/early 4th century A.D.). Given the substantial changes in the layout of the settlement indicated by intercutting features, dating should be treated with caution; the later finds may be derived from the layer of ‘dirty gravel’ that sealed them. Two smaller sub-circular structures (3385 and 6357) c. 3m and 5m across were probably too small to have been roundhouses. Both were cut by small rectangular enclosures/compounds (see below). Five-post structures (Figure 3) In addition to the two four-post ‘granaries’ between 1.5m and 2.5m square in Iron Age Enclosure 2, three other square structures lay within the open areas of the site. These were slightly larger in size and hada fifth central posthole. One (not numbered), c. 3.2m square, lay just north-east of Enclosure 2, while the other two (621 and 749), both c. 4m square, were located to its south and south-west respectively. It is possible that they fulfilled the same function as the fonr-post structures, their larger size requiring the support of an extra post. The two to the south both cut the fills of the silted up ditch 748, suggesting an early Romano-British date at the earliest and, given their open locations, they may have served another function. If ditch 748 was constructed as a formal boundary between the areas of old and new settlement, as suggested above, the positioning of these structures over it may have had some symbolic significance. Small rectangular enclosures (Figure 5) During the 1990 watching brief several square or rectangular enclosures formed of small ditches or gullies were recorded. Certain of these were similar in appearance and dimensions, measuring 10-13m across, some apparently open on one (often the northern) side, while others had small internal compartments. The most southerly example (3384) produced Romano-British pottery of 2nd/3rd century A.D. date, measured c. 11m long and 8.6m wide and was open along its long north side. In the south-east corner was a small compartment measuring 4.2m by 2.2m internally. The ditch of the most easterly enclosure (6334) produced Ist/2nd century A.D. pottery. Although there is insufficient information to determine the function of these enclosures, their rectilinear form, comparable in overall dimensions to the larger of the roundhouses, suggests Romanised domestic structures. It 1s perhaps significant that at least four of rectangular structures were constructed either adjacent to or in the same locations as roundhouses. A number of similar, slightly larger, features are suggested by the arrangements of rectilinear gullies/ ditches to the north, in one case abutting, and in another overlapping, ditch 748 (Figure 3). Other features Possible footings for part of a stone wall were recorded c. 30m east of feature 3384 adjacent to an area of laid stone, 1.2m in length and 0.6m wide, possibly indicating a hearth. No clear evidence for a structure was located. A possible late 4th/early 5th century Liebenau type brooch (Object 9676, Plate 4) was located in an area of stone rubble to the immediate north of these features. Immediately south of the overlapping Romano- British roundhouses a spread of limestone rubble, aligned north-east/south-west and edged at the north-west by lines of pitched and vertical slabs, perhaps represents a paved area on the edge of a pond to the south. A possible pit or well (3255; Figure 5) of 3rd/4th century A.D. date lay on the southern edge of the trackway; wooden planks and stakes located within may have formed a lining. The chronology of the settlements The chronological framework of the site depends largely on the Middle Iron Age to late Romano- British pottery assemblage and is limited by the predominance of less closely datable coarsewares. Refinement is in part supplied by the coins and metalwork, although such material is rarely from stratified deposits (less than 5% of the brooches, for example, and only 19% of the coins). This has frustrated attempts to impose chronological order on the complexity of features recorded. Overall, the evidence suggests a focus of Middle Iron Age settlement in the northern part of the site, disparate and fairly low level activity in the early Roman AN IRON AGE AND ROMANO-BRITISH SETTLEMENT AT CLEVELAND FARM 33 N=1142 British Average 200 + 50 -100 Fig. 6 Coins by period period (1st to mid 2nd century A.D.), with a marked increase in the spatial extent and level of activity in the later Roman period (3rd/4th century A.D.). Chronological evidence for Middle Iron Age settlement comes almost entirely from the ceramic assemblage, although a La Téne I brooch of the 4th century B.C. was recovered during metal detecting (Object 1400, Plate 2). Comparisons with ceramic collections from the region, such as Groundwell Farm, Wiltshire and Ashville and Watkins Farm, Oxfordshire (Gingell 1982; De Roche 1978; Allen 1990), enable the Cleveland Farm assemblage to be dated to the period c. 400-100 B.C., when calcareous fabrics were being superseded by sandy wares in the Upper Thames and Cotswold region. Late Iron Age and Conquest period pottery is also present, although in relatively small quantities - four wide-mouthed, cordoned bowls appear to be the regionally preferred form from the ‘Belgic’ repertoire; one bead-rimmed vessel; and at least three Dressel 1B amphorae (Edgeley-Long 2002). Alongside the pottery are 12 pre-Conquest brooches (1st century B.C. to mid Ist century A.D.), including La Téne III, Nauheim and Nauheim-derivative, Colchester, Langton Down and Nertomarus types (Plate 2); of these only one (a true Nauheim type) was stratified in the upper fill of the Enclosure 1 ditch. Coins and pottery appear to show low level activity from the mid Ist century A.D. at least until the mid 2nd century. Only 45 coins were found dating from the end of the Iron Age to the early 3rd century, while amongst the ceramic assemblage Savernake wares, so common in early Romano-British contexts at Wanborough, Wiltshire, for example (Seager Smith 2001), are scarce. Square or prismatic glass bottles, utilitarian forms common on later Ist and early 2nd century A.D. sites, are likewise notable by their absence here. In contrast, brooch use at the site had its floruit from the mid Ist century to the 2nd century A.D. (62 brooches, including Hod Hill, Colchester derivatives, T-shaped, trumpet-headed and headstud types: Plate 3). Moreover, the typology of the quernstones indicates an emphasis on the early Roman period (55 examples, compared to 17 from the late Roman period), although, confusingly, of the 52 quernstones from dated contexts, 46 came from contexts dated to the 3rd/4th century A.D. (on pottery grounds), a situation perhaps explained by reuse of quernstones. More extensive use of the site from the middle of the 2nd century A.D. is evidenced by the presence of Black Burnished ware from the Wareham/Poole Harbour region of Dorset. The Samian assemblage has an emphasis on the 2nd to early 3rd century, with Eastern Gaulish products well represented. Other imports dating from the later 2nd to 3rd century were identified in small quantities - Central Gaulish and Trier black-slipped wares. Ceramic evidence from the later Roman period is more forthcoming in the form of Oxfordshire and, to a lesser extent, New Forest finewares and mortaria, and Dorset Black Burnished ware in characteristic late Roman forms such as dropped flange bowls and flared rim jars. Coin evidence shows an emphasis on the period from A.D. 238 onwards, reflecting a more widespread use of coinage across the Empire at this time. Four hundred coins from Cleveland Farm date from between A.D. 238 and the end of the 3rd 34 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE century, the vast majority struck between A.D. 260 and 296 (Figure 6). From A.D. 318 the volume of coinage increased again and, by the middle of the 4th century, large numbers of small bronze coins were produced, a pattern again echoed at Cleveland Farm. Spatial analysis of the coins indicates a higher proportion of late 3rd century radiates in Field C, the eastern part of the Romano-British settlement, while Field D, to the west, shows a higher proportion of early/mid 4th century nummi. Whilst the detritus in which coins were lost would have been periodically swept away and dumped elsewhere - a phenomenon recognised at Cirencester (Guest 1998, 265-8), it is possible that the distribution of coins represents the shifting of particular activities around the site in the 3rd and 4th centuries A.D., or perhaps a shift in the focus of the settlement. The end of the occupation of Cleveland Farm is difficult to pinpoint. Late Roman pottery types extend the date range up to the end of the 4th century A.D. if not into the early 5th century. The possible late 4th/early Sth century Liebenau type brooch was recovered from the southern part of the site, but was not well stratified. Was the site continuously occupied into the early Saxon period? Certainly Early/Middle Saxon pottery is present, just under 100 sherds of organic-tempered, sandy and calcareous fabrics dated broadly to the 5th to 8th centuries. In three contexts such sherds were associated with ‘latest Roman’ wares. An Anglo- Saxon cast saucer brooch was found unstratified at the eastern edge of Field C (Object 712, Plate 4). Environment, economy and society Farming and environment Around 60 environmental samples were taken from pits, gullies, ditches and other features in order to retrieve waterlogged and charred plant remains that might shed light on the Iron Age and Romano- British economy and environment. Eighteen samples were taken from Iron Age contexts. Many of these focused on Enclosure 1, with seven samples from the enclosure ditch, one from the roundhouse gully and one from pit 127. Samples were also extracted from pits located on the enclosure ditches, roundhouse gullies and linear boundaries (pits 631 and 817, cutting gully 764; pit 841, cutting ditch 840, Enclosure 2; pit 658, cut by ditch 813) and pit 721, located in the entrance to roundhouse 720. These samples produced relatively little evidence for agriculture - a few charred remains of hulled wheat, emmer or spelt (Triticum dicoccum/spelta) and probable six-row barley (Hordeum vulgare sl), as well as a number of weed seeds that probably grew among the crops, such as black medick (Medicago lupulina), buttercup (Ranunculus acris/repens/bulbosus), cat’s tails (Phleum sp.) and some fragments of brome grass (Bromus sp.). Despite the finding of waterlogged material in association with Iron Age houses, ditches, pits and enclosures, plants indicative of human activity were not recovered. The exception was a bundle of straight woody stems or rods, either of hazelnut (Corylus avellana) or willow (Salix sp.), recovered from the ditch of Enclosure 1. It is possible they were used for thatching or basket-weaving - rods may be placed in shallow water to make them more pliable. Certainly fragments of willow basket were recovered from the Glastonbury Lake Village (Coles and Coles 1986) and it might be assumed that such crafts were acommon domestic activity within Iron Age Britain. Plant remains did, however, provide rich evidence for the localised landscape, indicating a hedged environment similar to that at Mingies Ditch, Oxfordshire (Allen and Robinson 1993), but in contrast to Thornhill Farm and Claydon Pike where, despite waterlogged evidence being available, there was little to no evidence for scrubland species and hence hedges (Robinson 2004, 141). The ditches clearly contained standing water, although it is probable that the site was subjected to occasional flooding which allowed colonisation by various molluscan as well as plant species. The remaining samples were taken from the area south of ditch 748. They include three samples from timber-lined pit 3255; four samples from a possible occupation deposit located to the north of the trackway, in the area of structure 5153/5139; three from ditch 1182; one from a pit located to the north-east of the same structure; one from a pit found to the north of structure 6303; two from an occupation deposit to the east of structure 6118 and three from the hearth associated with this structure; and single samples from two pits to the south of the pond and three pits near the eastern edge of Field D. In contrast to the Iron Age samples, the Romano-British contexts were comparatively rich in charred plant remains (e.g. ditch 1182, Table 2), the predominance of spelt wheat and probably hulled, six-row barley being a common feature of Upper Thames Valley sites (Robinson and Wilson 1987), and the dominance of spelt wheat being a common AN IRON AGE AND ROMANO-BRITISH SETTLEMENT AT CLEVELAND FARM 3 Table 2: Charred plant remains Nn Feature ditch 113 _|ditch 113|ditch 113205 337 1182 {1182 Feature Type Enc. | Enc. 1 Enc. 1 at aan ditch ditch Phase 1A 1A IA IA RB [RB [RB Sample Vol. (Ltrs) 4 10 10 10 Imm |>lmm |>lmm 125 1000 60 50 Cereals Hordeum sp. (grain) ares Hordeum sp. (rachis frag) 1 Cereal (grain) 4 Triticum spelta/dicoccum (grain) 1 + + + Triticum spelta/dicoccum (sprouted grain) ++ ++ ++++]++4+4+]++4+4+4+ Triticum spelta (glumes) 2 Triticum spelta/dicoccum (glumes) Cereal culm node LS 2 ee eee Ranunculus sp. Rumex sp. cf.1 + + Medicago lupilina 1/2 charred 1 Anthemis cotula ~ +a ae Coe ou Pane Sn Avena sp. (awns) fos Bromus sp. 2 1 1 Phleum sp. Avena/Bromus sp. 1 Key: + = 0-10, ++ = 11-50, +++ = 51-100, ++++ = >100 feature of many Roman sites across Southern England. Coriander, plum and apple were all also found. While the remains of hazelnut and sloe are likely to represent natural elements of the local vegetation it is also probable that they contributed to the diet in both periods. The poor representation of weed seeds may be explained by the crops having been stored as semi-clean spikelets (Stevens 2003). The weed species indicate a relatively broad range of soil types under cultivation during the Roman period, with wetland species indicating cultivation in the immediate vicinity of the site, but with heavier clay and drier, sandy soils also being exploited The animal bone assemblage exhibits no sudden or dramatic changes during the period of the site’s occupation. Cattle remained dominant during the Iron Age, although almost equalled by sheep/goat (Table 3) and remained so in the Romano-British period. There was, however, some increase in the size of animals in the Roman period, probably as a result of introducing or improving breeds. In addition, a number of new species appear, such as domestic fowl and donkey, some of which were probably Roman introductions. Animal husbandry practices display continuity from Iron Age to Romano-British, although a higher proportion of older cattle in the Roman period may indicate retention of these animals for traction (Figure 7). With increasing population in the post-Conquest period more land would need to be productive and draught cattle thus became more important (Sykes 2007). In addition, Romano-British cattle slaughtered for meat were killed at a slightly younger age than their Iron Age THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Table 3: Iron Age species and bone elements by unadjusted RFCs (Restricted Fragment Counts) (*over-represented: proximal and distal parts counted separately) le Scapula Humerus* 0 ay 5 9 6 | 2 2 4 aa Ulna Metacarpal* Pelvis Femur* Mi wle | RL nM wlw]e | wle ol ae (@) CoO; MR] wl] w A }oo;n oo i) = = DN a @ @ Third phalange 100 90 ite = —~— Iron Age 80 . ---@--- Romano-British 0-1 1-8 8-18 18-30 30-36 young adult old adult senile months months months months months adult Fig. 7 Cattle survival from toothwear analysis by period AN IRON AGE AND ROMANO-BRITISH SETTLEMENT AT CLEVELAND FARM 37 counterparts, perhaps because Romano-British husbandry techniques resulted in faster maturing individuals, or because consumers favoured younger meat. A more convincing difference between phases is the smaller proportion of butchered horse bones in the later phases of occupation, indicating a move away from consumption of horse flesh (Table 4). An interesting feature of the Romano-British assemblage was the low proportion of head and foot bones, elements that may be removed, sometimes with the skin, during primary butchery, and deposited elsewhere due to their low meat content. It is possible that primary and secondary butchery was undertaken in different parts of the site, although there were no indications of such specialisation; alternatively, the hides may have exported elsewhere for processing. Table 4: Animal bone - taphonomic factors by phase % % (ec a sheepieae| 13 | DA / wy OP ¥ LTH} fA oS / eS ESS Rei ZO OES IFTIOOONSSS . TA —eNve@, UI YOU x Bu; SOD. DOS SONS ESS F4 Kiln Room 19 ae te) 2M Ls ~ N Fig. 14 North wing, corn driers Room 19; view west EXCAVATIONS AT ATWORTH ROMAN VILLA, WILTSHIRE 1970-1975 65 Period 5B, Late 4th/?5th century In the north wing, flagstone floors were noted in Rooms 28 and 36 and a sub-floor in Room 29. These were laid after the Period 5A floors had become heavily worn and tread layers had accumulated on them (Mellor and Goodchild 1942, 64). A stone filled pit had later been cut through the floors (Figure 11: 5.3). The presumably unused hypocaust chamber in Room 29 had been infilled before the floor was laid. This fill was examined in 1937/8 when it was described as a mass of ash and refuse. Large numbers of blue, red and white tesserae were also recorded. The Period 3 Room 19, and the Period 4 added room at the north-east corner, Room 19A, were converted to agricultural use with two similar corn-driers set parallel with each other separated by the Period 3 wall (Figures 13 and 14). This latter was curtailed Fill Oven Fig. 15 North wing, oven Room 37; 1:50 short of the east wall of the range at a 0.38m wide gap marked by two large dressed blocks. The entrance gap was hollowed by use (F1) continuing to the exterior, suggesting that access to the corn driers was from there. Both T-shaped corn-drying ovens were of similar dimensions and plan. Their stone flues were a single stone thick and they had been trench built into the natural clay. The stoking pits were roughly circular, measuring 2m in diameter. No floor levels survived or the upper parts of the kilns. Although the corn-driers were not located in 1937/8, traces of a cement floor were recorded in the north part of the room. Outside the building to the north were deposits of rubbish and four pits, F2, F7, F9 and F10. On the opposite corner of the building to Rooms 19 and 19A, an oven was inserted into one corner of Room 37 constructed of mortared stone with a T-shaped flue and a stoking pit surrounded by a rough stone setting (Figure 15). A pit and stakehole were also recorded. The baths in the east wing were extensively modified with the furnace room, 9, turned into a stone-floored storehouse, and the earlier furnace room, Room 7, recommissioned to heat Rooms 11 and 12. The hypocaust in Room 10 was now the site of a corn-drying oven serviced from the apsidal end and linked with the Room 9. The rebuilding work was characterised by the extensive reuse of building stone, architectural fragments and large floor tiles In Room 28, in the 1937/8 excavations, the flagging sealed a Theodosian coin of 379-95 although this was argued by the excavators to have been originally lost on, not below, the floor. The Room 29 sub-floor contained 4th-century pottery. The fills of the oven in Room 37 contained four coins (Cat. nos 34, 109, 183 and 192) the two latest dated A.D. 364-78. A coin of 321-2 (Cat. no. 80) and a copper alloy awl (Figure 22.47) came from the northern oven in Room 19A. Pit F7 to the north of the building contained a fine silver Room 37 66 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE nail-cleaner (Cat. no. 29), a blade (Cat. no. 123) and 4th-century pottery. Pit F9 produced a metal strip (Cat. no 62) and a fragment of later 4th or early 5th- century vessel glass (Cat. no. 43). Pit F10 revealed a fragment of glass reused as a blade (Cat. no. 70) and 4th-century pottery. Period 6 The presence of small finds of medieval and post- medieval date, and of post-medieval pottery, may be an indication of later occupation rather than dispersed material. The only later feature was found in Room 35 where a roughly circular stone-lined pit had been cut through infill and destruction layers in the room’s south-east corner. Extensive stone robbing including the removal of some foundation stone remained undated but may be associated with the medieval finds. Subsequent destruction and levelling was witnessed by deposits of rubble where spared by the plough. Finds Coins, by Cathy King The report and a catalogue can be found on the accompanying CD-ROM. Brooches, by Donald F. Mackreth All are made from a copper alloy, unless otherwise stated. Each item is described by the catalogue number, trench number and context number. This text was written in May 1991 and revised in April 2001. Colchester derivatives 1. Figure 16. 288, 16.4, west of wall 108 topsoil. The seventeen-coil spring has an axis bar which passes through the lower of two holes in a plate behind the head of the bow, the chord passing through the upper hole. Each wing has two sunken beaded ridges. The junction of the bow with each wing is masked by a slight moulding rising from the wing. The plate behind the head is carried over the top to run as a beaded ridge down the upper part of the bow. The top of the ridge has cross-cuts on either side of a flat face into which is cut a saltire. The rest of the bow is plain and tapers to a pointed foot. The catch- plate is fretted to form an open-work cross with a pierced circle above and below. The return is defined on each side by a groove and has an incised chevron between these. The writer has only noted one parallel for the basic decorative scheme coupled with the spring-fixing system (France and Gobel 1985, 78, fig. 40.55), the usual one being the Polden Hill system in which the spring is mounted between pierced plates at the ends of the wings. The buried beaded mouldings on the wings and the beaded ridge on the upper bow (e.g. Romilly Allen 1904, pl. opp. p. 102) are the chief features of the group which also has big piercings in the catch-plate. Although the group is still undated, it gave rise to another resembling the present piece, including the mouldings masking the junctions of the wings with the bow (Goodyear 1969, 114, fig. 4, B). The distribution of the early type tends to favour the lower Severn Valley while that of the more developed type runs further up the valley and has outliers across central southern England. Very few have been published and fewer still have any dating: Bagendon, Glos, c. 50-60 (Clifford 1961, 173, fig. 31.5); Baginton, Warks, The Lunt, before c. 75 (Hobley 1973, 66, fig. 19.7); Nettleton, lst century (Wedlake 1982, 123-5, fig. 52.36A); Camerton, after 90 (Wedlake 1958, 218, fig. 50.8); Nettleton, 2nd century (Wedlake 1982, 123, fig. 52.36); Wilcote, Oxon, 3rd century (Brodribb et al. 1971, 110, fig. 47.67). As the developed form is the fore-runner of a major family belonging to the later 1st century and the 2nd, the preferred dating here is that given by the examples from Bagendon and the Lunt: before A.D. 75, possibly before A.D. 66/70. Figure 16. 3442, 55.2, courtyard topsoil. The spring, now missing, had been mounted like that on the last brooch. Each wing is plain except for a groove at the end. The plate behind the head rises as a crest above the wings. The bow is very thin and has two cross-cuts at its foot. This brooch belongs to a family made in the area centering on south Northamptonshire. Its characteristics are the very thin bow and the overall proportions including the way in which the plate holding the spring system over sails the wings. The only variations here are the grooves across the foot and at the end EXCAVATIONS AT ATWORTH ROMAN VILLA, WILTSHIRE 1970-1975 67 7 ANTONY Wi") Fig. 16 Brooches; scale 1:1 of the wings. The only parallel known to the writer for the small size of the present example comes from Cirencester (Corinium Museum, no accession number). The dating is not good, due mainly to the excavations producing them remaining unpublished: Bancroft, Mausoleum, 1/25-50 (Mackreth 1994, 9, fig. 131.9); Quinton, Northants, 50-60 (Friendship-Taylor 1979, 135-7, fig. 63.471); Bancroft villa. late 3rd—early/ mid 4th century (Mackreth 1994, 290, fig. 135.41); Wilcote, Oxon, 300-360 (Hands 1998, 49, fig. 18.32). Such a sharp break in the dates should mean, in default of better evidence, that the second half of the Ist century is preferred; the others must have been residual in their contexts. 3. Figure 16. 2025, 45.2, courtyard topsoil. The axis bar of the hinged pin is housed in the wings. Each of these has a series of angular ridges, the one at the end being more prominent than the others. The bow is divided into two parts by two cross-mouldings. The upper bow has two mouldings on each side, the outer one being short. Down the centre is a thin ridge. The lower bow has a V incised at the top and a deep groove down each side. At the foot is a small projection. There is some evidence for tinning. The group to which this belongs is predominantly south-western in its distribution. The use of extra mouldings at the head of the bow has recently been discussed (Mackreth 1992a, 53-4) and for hinged-pin brooches the present THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE dating points to the end of the Ist century and the 2nd up to c. A.D. 150/175 when all British bow brooches had ceased to be made. In the present case, the moulding across the middle of the bow may reflect the knop on the Trumpet type, in which case, the indicated date-range is appropriate. However, a reasonably close parallel dates to after c. A.D. 65 (Maxfield 1991, 82, fig. 23.1) and may suggest that to tie the present example to the Trumpet with its more extended range might be unwise. 4225, 67.1, courtyard topsoil. Only the head of the bow is left. The axis bar for the hinged pin is housed in very short and plain wings. On the head is part of a cast-on loop with a cut-out at its base on each side. The top of the bow is flat and the front has a step down each side and the tops of two rectangular cells for enamel, now missing. The probable full form of the brooch has the cells for enamel confined to the upper part of the bow, the lower bow having a central arris, a divided lenticular boss on each side at the top, and tapering to a broad foot with a projecting moulding (e.g., Dudley 1967, 34, figs. 12 and 13, 17, 19-21, 24, 25, 27, 29, 31). The dating is: Caerleon, A.D. 130-180 (Wheeler and Wheeler 1928, 162, fig. 3.13). A version having lozenge cells with infilling triangles on the upper bow is, as yet, undated. There is a version in which the lower bow is longer and divided by two more lenticular mouldings (e.g., Dudley 1967, 32, fig. 11.12, 14, 15), and this is also undated. All three should have come from the same workshop and the distribution of its products is mainly in Wiltshire, Dorset, Somerset and the nearer parts of Gloucestershire. The large number from Nor’Nour distorts the pattern, but a sea trade would perhaps be the best explanation for the spread of the type along the south coast and for isolated examples in eastern England. The date-range is likely to be the 2nd century up to c. A.D. 150/175. Figure 16. 3408, 45.2, courtyard topsoil. Only the lower bow and catch-plate survive. The front has a central arris and a small projecting foot. Not enough survives for the family to be identified. All that can be said is that the brooch was made before A.D. 150/175. Strip 6. 4070, 67.2, courtyard topsoil. x-ray 7026. Iron. Only the rolled-under head with part of the pin survives; the rolled-under section is as wide as the top of the bow. Almost exclusively at home in the south- west, Gloucestershire, Somerset, Wiltshire and parts of Dorset, there are some outliers, but not enough to suggest a secondary production area. The material is commonly iron, the pin invariably hinged and the head of the bow overwhelmingly rolled-under. The dating is: Maiden Castle, before A.D. 43 (Wheeler 1943, 252, fig. 85.35); Bradley Stoke, mid-lst century or before (Mackreth 1993, 51, fig. 7.55); Gussage All Saints, mid-lst century (Wainwright 1979, 108, fig. 82.1049); Puckeridge, Herts, Claudian ?— A.D. 70 (Partridge 1979, 40, fig. 6.13); Waddon Hill, Stoke Abbot, A.D. 50-60 (Webster 1960, 97, fig. 7.16); Longthorpe, Claudian-Neronian (Dannell and Wild 1987, 85, fig. 21.2); Nettleton, later 1st into early 2nd century (Wedlake 1982, 120, fig. 50.2); Ilchester, 3rd—4th century? (Leach 1982, 245-7, fig. 116.24); Barton Court, Oxon, late or post-Roman (Miles 1986, 5:D4, fig. 102.2, no. 2). Late La Téne 7. 4068, 67.1, courtyard topsoil. x-ray 7042. Iron. Most of the bow and part of the spring, with the chord now being separate, survives. The bow has a thin rectangular section; only the top of the catch-plate is present. 4095, 70.1, South-East Building topsoil. x-ray 7030. Iron. A fragment of a similar brooch, the section of the bow being much narrower. 7014, 64.1, west wing topsoil. x-ray 7044. Iron. Complete four-coil-internal-chord spring, but no pin. The bow is roughly oval in section and the x-ray suggests that there may be a groove down the middle. The catch-plate is small as is the whole brooch which is no more than 29mm high. None of these is a Nauheim or Drahtfibel proper; for that, one or more would have had to have a framed catch-plate. However, as all three are made of iron, it is not wise to draw too many conclusions on the differing size of section. One might say that the last is likely to be later than the other two, but no stress is laid on this. When dealing with the dating, the parallels are all in iron, all 4th-century examples have been omitted: Braughing, c. 25 BC-AD 20 (Partridge 1979, 103, fig. 30.3); Verulamium, King Harry Lane, pre-conquest (Stead and Rigby 1989, 386, fig. 177.440.2), before A.D. 45 (ibid., 310, fig. 112, 134.2); Bagendon, A.D. 43/45-47/52 (Clifford 1961, 167, fig. 29.1); Kelvedon, Essex, before A.D. 65 (Rodwell 1988, 67, fig. 53.4); Longthorpe, Claudian-Neronian (Dannell and Wild 1987, 85, fig. 21.1); Leicester, mid—late lst century (Connor and Buckley 1999, 261, fig. 119.18, 20); Chichester, 43—late Flavian (Down 1989, 186), late 1st-mid/late 2nd century (ibid. 186); Wilcote, Oxon, 150-175 (Hands 1998, 53, fig. 19.44); London, early 2nd century (Wilmott 1991, 118, fig. 81.301); Neatham, Hants, late 2nd-early 3rd century (Millett and Graham 1986, 118, fig. 82.301). The dates given show that there is a distinct fall off in the late Ist century and most had probably passed out of use by 75/80. Trumpet variety 10. Figure 16. 265, 15.1, Room 37 topsoil. The bilateral spring is mounted on an axis bar held by two pierced plates. On the head is a cast-on tab with a depression in the centre of each side suggesting that it had been intended to be a loop. The trumpet head is offset from the top by a groove, has a line of white metal solder round the top, another down the middle and spot of solder on each side. The lower bow forms the wings of an insect, there being a D-shaped cell for enamel on each wing and another on the body between. A mid-blue enamel survives in the latter, but the wings only have a discoloured green. The enamelled cells are surrounded by white metal solder. The foot-knob is the head and has a projection for its proboscis. This is the commonest form of insect brooch, the trumpet along with the pin-fixing arrangement showing that it was made in Britain. The full decorative scheme would have had beaded white metal, probably silver, wire along the soldered strips with a rosette on each side of the head. Although easily recognised, numbers of the type are not great with the result that the dating is thin: Caerleon, 160-230 (Brewer 1986, 172, fig. 55,13). An alternative design, lacking the trumpet head, comes from South Shields and is dated to the middle of the 2nd century (Miket 1983, 113, fig. 72.87). In general, it should be expected that the trumpet head here should be roughly coeval with the main Trumpet series, but not starting as early: 2nd century up to c. A.D. 175 with some lasting possibly into the EXCAVATIONS AT ATWORTH ROMAN VILLA, WILTSHIRE 1970-1975 69 3rd century. However, the specialised form of the brooch may have given rise to a different floruit, but, without more dated examples, this cannot be established. Crossbow 11. 1024, 12.9, Room 36 stratified floor, PSA. The foot only, the front is plain and the slot for the pin is on the left instead of the right which is the almost universal position for the catch. There is no trace of decoration and there is no trace of gilding, but this is not to say that it had not once been so finished. The distinctive feature is the pin-slot on the left-hand side. Invariably brooches have the catch for the pin on the right and it is only amongst the Crossbows that the reverse may be found. In default of any other feature, the dating of those Crossbows with the same feature is presented in the hopes that it will reflect the possible range here. 1. Richborough, no date (Cunliffe 1968, 91, pl. 32.77). 2. Winchester, Lankhills, 350-390 (Clarke 1979, 260, fig. 32.532); South Shields, late 4th century and later (Bidwell and Speake 1994, 179, fig. 7, 1, 5) and no date (Allason-Jones and Miket 1984, 100, 3.45). 3. Caernarvon, late 4th century (Casey and Davies 1993, 166, fig. 10.2, 12); Corbridge, no date (Snape 1993, 54, fig. 11,85); Wroxeter, no date (Shrewsbury, Rowley’s House Museum, X16). 4. South Shields, no date (Allason-Jones and Miket 1984, 106, 3.86). 5. Richborough, after A.D. 340 (Cunliffe 1968, 80 described under 81, pl. 33.80); Winchester, Lankhills, 350-380 (Clarke 1979, 260, fig. 32.24); Ickham, Kent, (unpublished, sf. 51); Lydney, no date (Wheeler and Wheeler 1932, 78, fig. 13.26). 6. Caernarvon, early—mid 4th century (Casey and Davies 1993, 165, fig. 10.1, 8); Catterick, excavations J. S. Wacher (Lab. No 3641); Corbridge, no date (Snape 1993, 55, fig. 11.86); near Peterborough, no date (Peterborough Museum,-L297). Most in this group have decoration which has a tendency to be C-shaped. 7. C-ornament Winchester, Lankhills, 350-90 (Clarke 1979, 260, fig. 32.278); Cirencester, 2 examples (unpublished: Cir. 60.A.XIII (2) sf. 185, AE [35]; Cir. 71.S sf. 358 Tr. IX (5)); Brancaster, Bunkle Collection, no date (Hinchliffe and Green 1985, 200, fig. 86.10); near Bury St. Edmunds, no date (Hattatt 1987, 267-8, fig. 83,1269). 8. C-ornament, Winchester, Lankhills, 370-390 (Clarke 1979, 261, fig. 32.447), 390-400 (ibid., 262, fig. 32.587); Richborough, no date (Bushe- 70 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 17 Brooches; scale 1:1 Fox 1926, 44, pl. 18.19). These are guaranteed to be late 4th century. The numbers which mark each section above are ultimately derived from the divisions perceived by Keller (1971) to be successive stages in the development of the Crossbow based on coin-dated graves in Southern Bavaria. His sequence has much to commend it, but the surprising feature here is that a purely functional aspect of a brooch, which side has the pin slot, should overwhelmingly show a date-range in the second half of the 4th century and beyond. One only is potentially earlier, the Caernarvon example in group 6, and that is not guaranteed. It is hard to believe that this dating is purely chance. If it has to be accepted, tight divisions on the style of decoration on particular Crossbows are useless. The fragment from Atworth almost certainly dates after A.D. 350. Penannulars 12. Figure 17. 2007, 44.1, courtyard topsoil. The ring has a circular section and was forged from rolled sheet metal. Each terminal is returned along the top surface of the ring and has two cross-grooves. Such brooches were easy and cheap to make and their light weight almost certainly meant that they were cheap to buy. Because of this, great subtlety in design should not be expected and the presence of one or two or three grooves on the terminals may not be chronologically EXCAVATIONS AT ATWORTH ROMAN VILLA, WILTSHIRE 1970-1975 71 13. significant: one groove, Hod Hill, before A.D. 50 (Brailsford 1962, 12, fig. 11.E11; Richmond 1968, 117-19); North Cerney, Ditches, Claudian- Neronian (Trow 1988, 51, fig. 24.26), Nettleton, 4th century (Wedlake 1982, 133, fig. 55.81); two grooves, Bagendon, Glos, A.D. 43/45—47/52 (Clifford 1961, 184, fig. 36.9); Wilcote, Oxon, Claudian (Hands 1998, 34, fig. 25.26); Waddon Hill, Stoke Abbott, Dorset, three examples, A:D. 50-60 (Webster 1960, 97, fig. 7.21, 22; Webster 1981, 62, fig. 59.13); Longthorpe, Cambs, c. A.D. 45-60/65 (Frere and St Joseph 1974, 46, fig. 24.15); Halstock, Dorset, 2 examples, before A.D. 60 or after c. A.D. 110/20 (Lucas 1993, 79, fig. 14.22—3); Prestatyn, A.D. 70s—160 (Mackreth 1989, 98. fig. 40.28); Verulamium, A.D. 80-150, (Wheeler and Wheeler 1936, 210, fig. 45.39); Caerleon, Hadrianic-Antonine, (Wheeler and Wheeler 1928, 166, fig. 14.23); Leicester, early- mid 2nd century (Clay and Pollard 1994, 143, fig. 74.18), early 3rd, (ibid., 145, fig. 75.28); Wilcote, Oxon, later 3rd century (Brodribb et al. 1971, 110, fig. 47.71); three grooves, Longthorpe, Cambs, c. A.D. 45-60 (Frere and St Joseph 1974, 46, fig. 24.14), Claudian-Neronian (Dannell and Wild 1987, 87, fig. 1.12); Bancroft Mausoleum, mid 4th century—early 5th (Mackreth 1994, 296, fig. 134.31). Only those having plain rings with circular sections are included and a few with idiosyncratic terminals have also been left out. The indicated date-range appears to be almost entirely lst century AD. Figure 17. 211, 7.1, corridor 21, north wing, topsoil. The ring has a circular section and was forged from rolled or folded sheet metal. Each terminal is returned along the top surface of the ring and only the left-hand one retains traces of the design: a concave surface on each side between a cross-groove at each end. The design is relatively clear, but can be confused with one in which there is a flute where there are concave faces here. The dating is: Bagendon, Tiberian—early Claudian (Clifford 1961, 184, fig. 36.8); Maiden Castle, A.D. 25-50 (Wheeler 1943, 264, fig. 86.8); Hod Hill, two examples, before A.D. 50 (BM 1958, 22, fig. 12.49; Brailsford 1962, 13, fig. 11.E18; Richmond 1968, 117-19); Colchester, A.D. 49-61, (Hawkes and Hull 1947, 327, fig. 59.7); North Cerney, Ditches, before mid-2nd century (Trow 1988, 50, fig. 24.25); Worcester, A.D. 150- 200 (Mackreth 1992b, 77, fig. 38.6); Dorchester, 14. 15: A.D. 150-250 (Woodward et al. 1993, 123, fig. 62.48); Canterbury, A.D. 175—300/320 (Blockley et al. 1995, 982, fig. 410.127); Marshfield, Avon, after A.D. 370 (Blockley 1985, 148, fig. 46.35); Nettleton, Wilts, with coin of 388 (Wedlake 1982, 133, fig. 55.80). Again, only those with plain rings and circular sections have been used. The date-range once favoured the Ist century, but more recent publications have pointed to the floruit having lasted into the 2nd. However, the sites producing the later dating have early occupation and so may all be residual. Those with 4th-century contexts were almost certainly residual finds. Figure 17. 3426, 59.1, west of west wing, topsoil. Made in the same way as Brooch 8, the ring has a circular section and the terminals are similarly returned along the top surface. The design of the terminals is worn and is best seen on the inside. Each had a set of three nicks on each side and the left-hand one has a wedge-shaped hollow at the end. The terminal design is unclear: it could either be grooved like Brooch 7 or a badly preserved pseudo-zoomorphic. In the first case, the dating should be like Brooch 7, in the second, probably 4th-century (see Brooch 10). Figure 17. 3478, 57.1, west wing, topsoil. The brooch is cast. The ring has a circular section and is continuously ribbed along the top. Each terminal is in the same plane as the ring and is zoomorphic: there is a pair of well defined ears and an eye-brow ridge above a short snout. The ribbed ring and the terminals cast in one with the ring should be sure signs that this brooch is Late Roman at the earliest. Like many Late Roman finds, the dating is not as good as it may be and the following list looks at various elements beginning with ribbed rings excluding two types of terminals which form the next two elements: pseudo-zoomorphic/zoomorphic terminals turned back along the ring, and zoomorphic terminals cast in with the ring. Ribbed rings: Halstock, after A.D. 110/20 (Lucas 1993, 79, fig. 14.24); Camerton, A.D. 150-200 (Wedlake 1958, 232, fig. 54.61); Catterick, with 4th-century pottery (Kilbride-Jones, 1980a, 150, fig. 52.14); Shipham, Star Villa, Avon, A.D 2325-350 (Barton 1964, 84, fig. 7.3); Bancroft villa, mid 4th—early 5th century (Mackreth 1994, 302, fig. 137.55); Great Horwood, Bucks, A.D. 72 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 350-425 (Waugh 1965, 64, fig. 2.5); Chichester, mid-4th-early 5th century (Down 1989, 192); Colchester, c. A.D. 350-«. 450 (Crummy 1983, 18, fig. 16.103); Ichester, later than the 4th century? (Leach 1982, 247, fig. 117.29). Pseudo or real zoomorphic terminals turned back along the ring: Moulton, Northants, late 3rd—4th century (Hunter and Mynard 1977, 134, fig. 19.262); Milton Keynes, early—mid-4th century (Marney and Mackreth 1987, 133, fig. 41.18); Bradley Hill, Somerton, 4th century (Leech 1981, 214, fig. 16.7); Verulamium, late 4th century (Kenyon 1934, 259, fig. 12.12). Cast-in terminals: Caernarvon, late 3rd—early/mid 4th century (Casey and Davies, 166, fig. 10.2, 16); Brancaster, 4th century (Hinchliffe and Green 1985, 42, fig. 28.2); Caernarvon, early—mid 4th century (Casey and Davies 1993, 168, fig. 10.2, 17); Barton Court, Oxon, late Roman (Miles 1986, 5: D10, fig. 103.7); Caernarvon, late 4th century (Casey and Davies 1993, 168, fig. 10.2, 19); Birdoswald, after 383 (Kilbride-Jones 1980a, 147, fig. 52.2); Wilcote, Oxon, late 4th century (Brodribb et al. 1973, 108, fig. 109.180); Feltwell, Norfolk, c. 400 or later (Gurney 1986, 30, fig. 22.2). The message seems to be clear; all could be later than the mid-4th century and some contexts could be later than A.D. 400. The only exception to this is the ribbed ring from Camerton which must be wrongly dated. The same site produced another which could have been 2nd century, but on the evidence set out above should not have been. The cast-in terminal should be later than the pseudo-zoomorphic one. The metalwork, by Jane M. Burcher, with contributions from Susan Youngs Overview The assemblage from Atworth is the only group from a large villa complex in the Bath and west Wiltshire area to be published in recent years. This fact alone underscores its importance for regional Romano- British studies. However the lack of comparable local assemblages also presents problems of comparison and quantification. Although a number of other villa sites in the region have been investigated, these excavations all date from the 19th or earlier 20th century and very little of the material has survived or been analysed and published in detail. At Box, only 2.5 km from Atworth, the list of finds from Brakspear’s investigations (1904) is clearly incomplete and the most recent work so limited in scope that it provided little comparative material (Hurst et al. 1987). None of the small finds from the very large villa complex at Truckle Hill, North Wraxall, investigated in the mid 19th century, survive. In general, the Atworth metalwork indicates a cross-section of Roman life and is typical of a rural civilian site spanning the Ist to 4th centuries. Male artefacts (nos 37-41 and probably most of the tools) and female artefacts (most of the jewellery) are represented. A handful of fine quality finds (in particular the nail-cleaners nos 29, 33 and 34) suggests residents of relatively high status. The two styli (nos 44 and 129) indicate literacy, the locks and keys imply that items of value were kept in the villa and the military equipment belonged to a man holding rank in society. The majority of the finds are of a domestic nature although the work of the villa is also represented by finds of tools. The spatial analysis of the Atworth metalwork (Tables 1 and 2) suggests that all the buildings were multi-functional, providing both residential and working space. The metalwork indicates earlier Romano-British activity on the site. There are eight objects which can confidently be assigned a Ist or 2nd-century date (nos 11, 18, 30-2, 42, 74 and 108). This accords with the mainly late 1st and 2nd-century dates for the brooches and the evidence of the samian that demonstrates considerable occupation of the site at this time. The dating of the military metalwork (nos 37-41) and other objects possibly including the spoons (nos 42-3) and finger ring (no. 12) span the 2nd and 3rd centuries and can be paralleled by the late 2nd-3rd-century vessel glass and 3rd century or earlier cast window glass. In its final phase the villa is a relatively substantial, well appointed complex, and can be compared with the large group of similar sites that lay in the outer hinterland of Roman Bath. The greatest number of datable metal finds come from this period. However, the metalwork indicates continuity of occupation from at least the late 1st century and does not indicate any change of status or function as the villa evolved. A particularly interesting feature of the metalwork is the suggestion that activity continued on the site into the very latest Roman period. A number of copper alloy objects including the bracelets (particularly nos 4-7), finger rings (nos 13-15), pins (nos 21-23), ear-ring (no. 26), and EXCAVATIONS AT ATWORTH ROMAN VILLA, WILTSHIRE 1970-1975 73 possibly nail-cleaners (nos 33 and 34) fit comfortably into the 4th century and even post-date A.D. 350. Two of these late Roman objects (nos 4 and 14) have been intentionally flattened, perhaps in preparation for recycling and indicating small-scale industrial activity at a very late period. Fourth-century material at Atworth should be viewed in the context of late and post-Roman material at Bath (Cunliffe 1988; Bircher 1999), around Bristol, as at Henley Wood (Watts and Leach 1996); Cadbury Congresbury (Rahtzet al. 1992); and possibly Cadbury Tickenham and Stantonbury (Burrow 198]) and in the Calne area to the east (Youngs 1995). Moreover, the site lies close to the main Roman road from Cunetio to Bath. It has been suggested that the line of this road forms a continuation of the East Wansdyke, a probable 5th-century boundary, although a later date has been argued (Eagles 1994; Bonney 1972; Reynolds and Langlands 2006). Evidence of activity at Atworth in the late-Roman period is therefore potentially of great importance. Objects of silver and copper alloy A total of 144 finds of copper alloy and one of silver (no. 29 below) were recovered from the 1970-5 excavations. In addition 20 objects or fragments and one silver object from the 1937/8 excavations were available for re-examination (nos 6, 22, 23, 36, 41, 79 and 80 below, together with nos A3, A6-10, A25, A35-6, A44 and A51-3 in the archive report). The shallow stratigraphy of the site meant that all but 20 of these objects derived from the top two layers. Therefore little can be said concerning the phasing of the site from the objects just as little dating evidence for the objects can be inferred from the context. The criteria for publication have been 1) diagnostically Roman material, and 2) material recovered from the deeper stratigraphy. A selection of the more interesting non-Roman material has also been included. Unless stated otherwise, the published objects are assumed to be Roman. The remainder (numbered Al—A79) has been consigned to a full archive catalogue under the categories: token, buttons, miscellaneous (almost certainly modern) objects, wires and fine bars, penannular links and rings, washers or perforated discs, binding strips, sheet fragments and corroded lumps. L = length W = width, D = diameter, Th = thickness, Ht height, ext = external, int = internal Jewellery Bracelets 1. 4220 79.1, courtyard topsoil. Small fragment of a 2-strand cable bracelet. 2nd to 4th century. L llmm. 2. 233 11.2, north wing topsoil. Possible bracelet fragment, broken at both ends. A beaten solid bar of round section is slightly facetted, curved and tapers towards one end. L 56mm, max D 2.5mm. 3. Figure 18.2936 45.1, courtyard topsoil. Fragment of a strip bracelet with ribbed decoration. One terminal of the lapped and riveted joint is extant. This is not acommon type and a good parallel is published from Orton Hall Farm in Mackreth 1996, 96, fig. 62.39. Other examples cited range in date from c. 150 to post A.D. 350 L 30mm, max W 4mm, Th 120cm high) Type II (60cm - 120cm) Known Coppice Other Locations Boundaries (not known coppice) Type III (< 60cm high) 6 | 1 | 0 7 6 2 17 Fig. 6 Table showing 39 bank cross-sections classified by Type and location or purpose. Note that four cross sections falling into Type II (numbers 19-22), and which do not bound known coppices, are sited on a natural ridge along the western boundary of the Inclosure, as shown in Fig 1, which accentuates their vertical dimension. 4 26 12 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 15 123 45 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 9 123 45 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 33 12.3 4 5 6 7 8 9 10 M1 12 13 14 15 16 17 18 19 20 21 29 12.3 4 5 6 7 8 9 10 If 12 13 14 15 16 17 18 19 20 21 Fig. 7 Sample cross sections of the Great Inclosure boundary bank and ditch system. Vertical axis in centimetres, horizontal in metres. For comparative purposes all cross sections are shown with the inside of the Inclosure to the left; hence compass orientations vary. The inner bank top of each cross section is at the 6m mark with the outer ditch to the right of this. Type I: Profiles 4 and 9 both located in Shovel Bottom. Type I: Profile 26 located at Ashlade Coppice, and 33 at Bagden Coppice. Type III: Profile 15 between Sawpit Drive and Pigsty Bottom, and 29 near Amity Oak (neither known to be coppices). Discussion Six out of the seven examples of Type I cross- sections form part of a single massive structure, totally unrelated in scale to all other sections of the route, crossing Shovel Bottom at SU229660. The greatest vertical dimension of 1.56m was recorded at sample point nine. The system in Shovel Bottom was also significantly wider than any other part of the Inclosure bank, reaching 16m at its greatest. The profile in this location comprises a bank, ditch and counterscarp. There is no evidence that the great dyke in Shovel Bottom is contemporary with the remainder of the Inclosure bank. This structure crosses a valley of exposed chalk and alluvium, but does not continue on reaching clay-with-flints, gradually petering out at its eastern extent and stopping abruptly in a head-wall at the western end. Although structurally unrelated to the rest of the Inclosure, it appears similar to a system in Great Lodge Bottom (SU212671) which exhibits comparable characteristics. Type II cross-sections show a clear relationship with the location of known coppice boundaries at Woolslade, Ashlade, Bagden, 164 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Fig. 8 Bank between the Grand Avenue and Entrance of Savernake Lodge, composed of gravels which have eroded except where secured by tree roots. and a small stretch between the Grand Avenue and the former entrance to Savernake Lodge (see Figure 5). In addition, cross-sections 19-22, which fall within this category, are sited along the ridge shown in Figure 1, where natural topography may have contributed to overall scale. Type III cross-sections are the smallest earthworks still discernible; nearly all are un-related to known coppice boundaries. A short section of bank - between the area marked ‘Lodge Gates’ on Figure 1 and the eastern end of the Shovel Bottom Dyke - may not conform wholly with the above pattern. North of the Grand Avenue this bank is inconsistent in structure, with short sections 0.83m high intermittent with stretches which are indiscernible. Such variation might be attributable to the exceptionally gravelly soils, which are easily eroded except where secured by tree roots (see Figure 8). It is also the only section of bank which runs in a straight line (for 400m). This might suggest it represents the final section constructed, where a route was needed to link up with the Shovel Bottom structure. However, these two anomalous features may indicate that the bank here was somehow distinct, perhaps related with the Shovel Bottom Dyke into which it grades. Identification and Correlation with Documentary Sources Many deer parks were established in the Savernake area in medieval and post medieval times. These included Chisbury Park, for which a licence was granted in 1261 (Close Roll, 1259-61, 355), Savernake Great Park which was established by 1575 (WSRO 9- 22-204), Brimslade which was created between 1585 and 1595 (WSRO 1300-90 and 1300-87), Wolfhall or Sudden Park in existence by 1536 (WSRO 9-1- 30), and, based on place-name evidence, probably The Brails by the 12th century (PRO E 146-2-22). However there are no extant records of parks which could be attributed to the central Inclosure of Savernake. Furthermore, the continuous external ditch suggests an intention to keep animals outside the bank, and not within, as normally expected by THE GREAT INCLOSURE OF SAVERNAKE WITH A NOTE ON CROSS VALLEY DYKES 165 a park pale. John de Havering’s licence to enclose land in Lytelferme and Kyngesheth in 1302 required him to ‘enclose them on that part towards the foreign lands there lately disafforested with a great dyke 6 feet deep and 7 feet broad, with a hedge, and make the crest of the dyke so that the king’s deer cannot get out of the close but can enter without hindrance’ (Cal Pat Rolls, 30 Edward I). The licence permitted the inclosure of 421 acres (as measured by the ‘Forest perch’), an area broadly comparable to the Inclosure. Lytelferme probably relates to the ancient settlement of Littleworth at Timbridge, and to Little Frith, both less than 1km from the edge of the Inclosure bank.” However, the licence implies that de Havering’s land was to be bounded on one side only, and the Great Inclosure bank, for most of its course, is small in comparison to the specifications in the licence. The consistent external ditch reveals that construction of the Great Inclosure was an attempt to exclude animals, either domestic stock or deer, suggesting that cultivation or woodland management was the primary intention. Archival material from the Savernake Estate, now deposited with the Wiltshire and Swindon Record Office, suggests that the Great Inclosure was a feature under construction in Tudor times (albeit incorporating an older cross valley dyke along its course), and was further employed for landscaping purposes during the later 18th century. Phase I: The Great Inclosure in Tudor Times Until the great period of agricultural inclosure, much of Savernake Forest was common land. Commons were available to people from neighbouring settlements (‘borderers’), which, from 16th to 18th century records, were known to include Bedwyn, Stock, Burbage, Easton, Durley, St Margaret’s (Marlborough), Crabtree, Chisbury, Wilton, Stitchcombe and Puthall (WSRO 1300-320, WSRO 1300-87, WSRO 9-22-2211). Savernake Forest, like other commonable lands, provided distinctly different benefits for the right-holders and land-owner. The former derived largely agricultural and domestic products. The Forest provided, more than anything else, pasturage for their animals. Lopped or pollarded trees also supplied browse, and could provide nutrition even outside the growing season. Heath and fern were harvested for animal bedding, whilst fern was also sometimes used in the Forest for thatching. Coppice poles, which had to be purchased rather than taken as right, were partly used to make fence hurdles, and faggots of poles, with furze and thorn, together with deadwood, provided domestic fuel. In contrast, the landowner derived benefit from the deer and the prestige associated with them, and from constructional timber and sale of wood products, including coppice and bark for the tanning industry. The landowner’s interests suffered from the activities of the borderers, whilst they, in contrast, found that the deer, which they were not entitled to take, competed for browse and trespassed into their arable lands. Although acorns and beech mast provided fodder for pigs, any large, unlopped and spreading trees could be a nuisance to the commoners by shading out the lush grazing, whilst they had no authority to take timber anyway. Indeed, in most cases borderers were forbidden from taking any oak or ‘great trees’ without authorisation, even on their own holdings. An awareness of timber shortfalls led to the introduction of legislation to protect woodlands throughout Europe from the late 15th century. However it was not until the Statute for the Preservation of Woods was passed in 1544 (Henry VIII), that attempts were made to address timber shortfalls and its economic implications at a national level. The introduction gives emphatic expression to the parlous state of the woods. It states ‘The King perceiving the great decay of timber and woods universally to be such that unless speedy remedy be provided, there is great and manifest likelihood of scarcity and lack, as well of timber for making, repairing and maintaining of houses, as for ships for the necessary relief of the whole community.’ The Act required that, during clearance, coppices should retain 12 uncut standard trees per acre. These had to be at least 10 inches in diameter before they might be felled. Young coppices had to be enclosed to protect them from grazing animals for at least four years after clearing; those over 14 years old had to be enclosed for six, and older coppices for seven years. An Act of 1558 (Elizabeth I) extended the closure period to nine years. Once standard trees were a certain height, they were out of reach of both deer and domestic stock, and could continue unhindered. In ideal conditions this would lead to coppice-with-standards woodland, in which great trees formed an intermittent canopy over an understory of coppiced hazel and other poles. It is clear that, at Savernake, the interests of the Earl of Hertford as landowner, and those of the borderers, were diametrically opposed, establishing barriers related to status and class, part of the 166 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE traditional antagonism between lord and peasant found throughout much of Europe. This antipathy, which persisted for centuries, was part of a land-use system later described in the House of Commons Journal as ‘a perpetual struggle of jarring interests, in which no party can improve his own share without hurting that of others’ (House of Commons, 1788). At Savernake the opposing interests caused conflict which erupted into threats and brinkmanship. The paucity of trees on Hertford’s extensive estate was thrown into sharp relief immediately when he established residence in Wiltshire in 1569, and obtained confirmation of news he had received whilst in prison, that his family manor house of Wolfhall was probably beyond repair. Resolving to build a new mansion near Tottenham Lodge, he enquired about his main sources of timber. Previously his father, the first Duke of Somerset, had intended to use about 2000 great oaks at The Brails for constructing a grand house overlooking the Pewsey Vale. Upon his execution, The Brails had been transferred to the Earl of Pembroke (WSRO 1300-176) who lost no opportunity in felling the oaks and using the timber in the building of Ramsbury Manor House. However, Sir John Thynne, replying to Hertford’s enquiry about other sources in 1574, gave stark news: not only had certain family woodlands been seized by the Crown following his father’s downfall, but remaining sources had been degraded through abuse and unauthorised felling. He wrote: ‘Jouching the woods [from which] your father meant to have his timber for building, you have none of them; for he meant to have had most of it out of the Forests of Chute and Bradon, and some out of The Broyle and other places thereabouts. Your Lordship must now reserve your timber in all places so as you may have some part in one place and some in another as it may be to serve your turn. Give orders to [your steward] Tutt that no trees anywhere [may be] sold that would serve either for long or short timber: it 1s a bad tree that will not serve for short timber, for if it would rise to 5 or 6 feet of short timber it would serve for some purpose and save the cutting of longer timber. Under the colour [pretence] of felling scrub, good timber may be sold. Therefore give order to all your woodwards and those that have the doing with your woods that no trees be sold, and then you shall be sure not to be deceived, or otherwise your under-woodwards would please their friends and cut down your best trees under colour of felling scrub [pretence of scrub clearance], as you have experienced in your Forest’ (Longleat House/ Seymour Papers/Volume V, fos 56, 57d). Despite having an estate of thousands of acres, in 1575 Hertford was in the absurd position of having to purchase some 640 oaks growing in Stock Woods and its adjacent grounds and hedgerows, from his neighbour Anthony Hungerford, at a cost of £322, almost certainly for incorporation into his new mansion at Tottenham. He acquired them as marked standing timber, with ability to fell them as required over the next five years, together with the bodies, lops, tops and other commodities when felled, but not the ‘rymes’[rind] or bark, which were reserved to the seller, presumably for the tanning industry (WSRO 9-26-512). Hertford took steps to address some of the abuses in the Forest. Keepers were forbidden from cutting down any tree or wood for fuel, except those ‘marked by the Lord’s marking axe’ (WSRO 1300-87). However, in due course he also sought to establish more reliable sources, and mindful of the losses throughout his and neighbouring estates, resolved to investigate options for restoration. Reinstating the woods would not be achievable without evicting commoners from some part of the Forest, and because of uncertainty over his right to do this, legal advice was sought. The original documentation from this time had become much ‘eaten and decayed’ by the 18th century, and has presumably since disintegrated, but in 1718 the readable contents were fortunately preserved in a transcription by William Longueville, in a project summarising legal disputes in the Forest for the benefit of Charles Lord Bruce (WSRO 1300-1484). The records reveal attempts to tackle chronic deterioration and abuse of the woodlands, coupled with commoners’ traditional determination to guard their rights jeaiously. Letters were sent to Mr Richard Wheeler instructing him ‘to give the Earle of Hertford satisfaction of the lawfulness of his right to inclose the coppices of Savernack’. They speak of ‘the lamentation of the people for the great decay and spoils of woods now that much wood ground 1s converted into pasture as at Southgrove, the Broyle, and Clatford [all part of the former Royal Forest of Savernake], Chilton Parke [Chilton Foliat, Berks], Balston Parke [Hannington, Hants] and that Great Woods were cutt downe, as in Chute forest, Fortesbury, Stockwood and in many other places. And that his own Great Wood in Savernack 1s very leere [empty] and old, and his underwood is much decayed’. The Earl commanded his officers to consider how the situation might be addressed. They reported that ‘the Redd deer and Valler [Fallow] deer were great annoyances and caused destruction to the underwood. And the borderers do not only oppress the said Forest with more cattle than in former times, but many of them connive to bring cattle from foreign townships [ie those without true THE GREAT INCLOSURE OF SAVERNAKE WITH A NOTE ON CROSS VALLEY DYKES 167 rights] which ought not to be allowed. And they also doe pasture old oxen and such ould cattle without hetching [tethering?] and spoil the first spring of wood as well of coppice, which were formerly preserved’. In 1594 Hertford instructed his officers to find a suitable location for inclosing wood in some convenient parts of the Forest. He resolved to take his deer out of the open places, and to inclose a portion ‘which in former time hath ever been coppice ground and inclosed’, citing Henry VIII’s statute of 1544 for the preservation of woods. The inclosure was to be ditched and hedged with ‘quickplants’ [probably hawthorn or blackthorn]. It is clear that the commoners’ capacity to undermine the success of this venture was recognised by the Estate from the beginning. The borderers were assembled at the Great Lodge in Savernake, and asked whether they would agree to a bargain, whereby the Earl would remove his deer, so that they would no longer trespass in their corn, and he would inclose sufficient ground to allow restoration of the woodlands. The records imply an intention to eliminate deer from the whole open Forest, an enormous enterprise of dubious likely success. Only through reduction of the main Forest population could arable farmers outside the Forest benefit from reduced grazing. Over ten square kilometres of the western Forest had been inclosed as Savernake Great Park by 1575, and it may have been envisaged that deer from the Open Forest could have been driven into the Park. The chances of preventing escapes however from the extensive perimeter fence would have been fruitless, especially as there were reports of local villagers ripping up the paling (WSRO 1300- 87). Apparently, Hertford’s officers believed that all or most of the borderers ‘held themselves well dealt with’, and all but two of them “desired that his Lordship _ would proceed’. These claims were later refuted, but nevertheless Hertford decided to implement his plan, based on legal assertions. His lawyers argued that, as he was Lord of the soil of the whole Forest, he could keep what number of deer and cattle of his own and also accept rents for such numbers of agisters”? cattle as he pleased. His officers contended that it would benefit everyone if he kept as common land more than half of the Forest, whilst keeping them free of deer or his own cattle, and also denying access to other agisters’ cattle, except for those borderers who had anciently depastured animals in the Forest at certain rents. Instructions were given to create a hedge and ditch to include ‘Ashlade Coppice, Hurlemore Coppice, Bagden Coppice, Bagden Rayles, Buthenbed Coppice, Wolslade Coppice and Lye Coppice alias Lewdens Lye Coppice, and for a small quantity of void ground lying between the said coppices’. Of the coppices listed above, all but two can be located within the bank and ditch system described in this paper (See Figure 1), thus positively identifying the Great Inclosure as the feature constructed at this time. Of the other two coppices mentioned, Buthenbed cannot be traced. However, Hurlemore seems to be listed erroneously here, unless there were two coppices with that name, since it lay 5km to the west, between Manton Coppice and Granham, and was grubbed in the 18th century. It was claimed that the length of the Inclosure bank was ‘in circuit two and three quarter miles, and not above one eighth part of the open ground of the Forest’. Reference to modern maps indicates that the bank is 3.47 miles long by statute measurement. In 1598, four years after the original instruction to create an inclosure, there was ongoing disquiet and uncertainty, with Hertford seeking legal confirmation of his authority. It was claimed that ‘his Lordship’s purpose is to sow and plant the void ground of the coppices with many oaks, ashes, akerns [acorns], nuts and other seeds, to increase the same to thick and strong wood; and not to suffer any cattle to be kept in the coppices to the hindrances of the wood, which for the present 1s a great charge to his Lordship, and will be a great benefit to the inhabitants in future time’. It appears that Hertford countenanced the possibility of cattle being re-admitted to the area once the woodland was established, and may have been attempting to bargain further. It was envisaged that once the hazel was established, then the commoners could bring ‘such great cattle as shall not hurt the woods, or be compelled to tether their cattle, as is [common practice] in other forests and [formerly] used in this Forest, although not since the death of the Duke of Somerset (1552]; as may appear by the woods which have been fair, and are now mean’. It was claimed that overall the Inclosure would provide great benefits to the commoners. It was also contended that ‘in former time no man was suffered to keep any cattle or do any act prejudicial to the rest, as may appear by the presentments made in Swanimote Courts of the Forest, whereof many presentments and precedents remain in his Lordship’s Evidence House at Wolphall’. Following continued dissent, in the light of the revised proposals, the Earl directed any Borderers who still ‘conceived any dislike’ to give notice to his officers. Shortly afterwards, Hertford’s steward Alexander Tutt, wrote to Richard Wheeler from Bedwyn, updating him concerning long-running 168 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE battles with commoners in Savernake, and also in Bentley Wood and other parts of the Estate (WSRO 1300-104). He said that he had been viewing the great bounds newly made in the Forest, and hoped that the ‘needless fault finders and repiners’ [complainants] should not have their desires, so that the new venture could be profitable to his Lordship. He also hinted at threats received from commoners, and the possibility that they could yet successfully oppose their eviction, claiming that ‘for any to forbid [interfere with the activities of] the workmen is I think very unadvisedly done, for the soil being his Lordship’s, no one having now common within the same for any manner of cattle but only [those of] his lordship, he may well inclose it for himself leaving if the worst happen sufficient gaps for their cattle to come in at their time commonable which 1s from Holyrood day in May to St Martin’s day [November]. In all likelihood, Hertford had some legal argument in his favour. The right of ‘approvement’, the principle whereby the Lord of the Manor was entitled to inclose any land surplus to the requirements of the commoners, was enshrined in English statute since 1285 (‘The Statute of Westminster’), and was not repealed until the Commons Act of 2006 overturned some of the oldest surviving legislation in England (Stationery Office, 2006). However, irrespective of legal authority, it appears that local opposition was too great for the scheme to succeed. In a general rant against destruction of the Estate woods generally the steward bewailed ‘these commoners do marvellously murmur and grudge’, and continued: ‘this Realm will, in Short time, rue the waste of it, and yet how lamentable a case 1s it to see the perverse disposition of the people’. It is likely that the grand plan to establish woodland in the Great Inclosure never enjoyed any success. The annual Courts known as Forest Views, whose proceedings exist for twelve of the years from 1594 to 1606, do not mention the Inclosure, under any name. This silence may itself be telling, for elsewhere where barriers such as hedges and pales had been erected, the Court cites examples of them being subjected to unauthorised damage and theft. It is possible that the Great Inclosure bank was not surmounted by anything worth stealing in its early years. The small bank would certainly never have been deer proof, and its primary purpose may have been just to define the area where woodland protection was seen as paramount. When a deer census was taken in 1611, by over a hundred Borderers, they counted 266 deer in the Forest, and there is no indication that any large scale attempt had been made to remove them (WSRO 1300-87). The Forest Views record minor trespasses by commoners within those coppices encircled by the Great Inclosure bank. For example, in March 1599 Thomas Hide was reported for ‘oppressing the Forest with one hundred sheep which have been kept within Ashlade and Lewedons Lye, ever since St Martin’s Day’. This was probably a habitual practice, for he was reported for the same offence in 1605 and 1610. Proceedings from the Forest Views actually record very few incidents inside the coppices of Bagden, Ashlade and Woolslade, either before or after the Inclosure bank was constructed at the end of the 16th century, and it is unclear to what extent commoners’ animals, rather than deer, really did contribute significantly to preventing strong growth of the coppices. However, Hertford’s intention was also to plant up the open areas between these coppices. These extensive grasslands would have been very attractive to domestic stock, and Hertford’s attempts to bargain with the commoners indicate recognition on both sides that their rights extended fully into these areas. The records may suggest that, when it was a Royal Forest and in the King’s hands, a clear distinction was made between the commons on privately held manors fringing the Forest, as at Durley, Chisbury and Burbage commons, where township boundaries were carefully denoted and trespasses reported, and the King’s land of the Forest itself, where such divisions may have been less vigorously maintained. After the transfer of Savernake from the Crown, initially the distinction between Forest and manorial commons remained, and for this reason it had been the borderers in general who had been summoned to consider the proposals at the Great Lodge. However, in due course it would become treated like the surrounding manors, and distinct township boundaries for the different commons were specified for the centre of the Forest. None of these later included the land in the Great Inclosure. The distinction between township boundaries may have been determined as early as the 17th century, since the record of Thomas Hide’s trespass in 1610 states ‘he hath pastured and kept his sheep about Ashlade Coppice whereas they ought to go but unto the northside of the Coppice and [only] so far westward as the Way leading from Bedwin to Mildenhall’. The evidence suggests that the estate reverted to trying to protect individual, small coppices rather than a single great woodland block. It is doubtful whether the plan to plant up the voids between coppices was ever implemented. A survey of all coppices in Savernake and Tottenham taken in 1716, shows that those within the Great Inclosure were THE GREAT INCLOSURE OF SAVERNAKE WITH A NOTE ON CROSS VALLEY DYKES 169 fort Fig. 9 The Great Inclosure bank at Woolslade Coppice. The bank runs A-B-C; from B to C (and from B to D) the bank is substantial, and represents a former coppice boundary, whereas from A to B, outside the former coppice, the earthwork itself (marked here by fallen trees) is small. modest in size, and probably no larger than they had been before the Inclosure had been contemplated. Woolslade was recorded as 72 acres, and Ashlett [Ashslade] and Bagden each at 60 acres (WSRO 1300- 301). This total of 192 acres falls well short of the 540 acres which had been incorporated within the Inclosure. The retention of the coppices, as distinct from the land between, explains the difference in . bank boundaries described above. In the former coppice areas, banks have been maintained and remain prominent, whereas outside these areas they are modest, and represent the original line defined in Tudor times (Figure 9). There is also evidence that these coppices were sparsely wooded, with little or no improvement upon the impoverished stands that Hertford had sought to replenish. Indeed, when the Great Inclosure was again examined in the 18th century, for landscaping purposes as described below, woodland cover within this area was poor, and Bagden itself described as extensive but ‘thin of wood’ (WSRO 1300-1928). Phase II: The Great Inclosure during landscape enhancements of the 18th Century A letter prepared in 1767 reveals that the Savernake Estate was re-considering the Great Inclosure during the period when Lancelot ‘Capability’ Brown was advising on landscaping (WSRO 1300-1917). The letter from Charles Bill, steward to Thomas Brudenell-Bruce, (later) Earl of Ailsbury, is probably a copy of one sent to Brown: 13 March 1767 Besides the usual business at Tottenham Park I had an additional reason for desiring to see you there at your next visit, my Lord having desired me to shew you the interior or inclosed part of the Forest where we are now making some improvements and whether I have the pleasure of seeing you there or not will endeavour to give you the best idea I can of the place, and what his Lordship wishes you to do concerning it. 170 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Ashlett Fine Level Lawn wanting little improvement BONG o @ Kes Lye Baad Lodge UN | : 8 g Q 3\ A : shoe k Re Mr Rolt’s House S LAN oe S é eae wae 8 % 8 XS @ ae? 4? Bagden Grounds & PORN y Ga” 8a PO Ny Qt g ge gh in lease to Mr Rolt \ Pa, Ss J@ Lewdens x, x @ “a. 9 sie WW, Lye Os Q Yo N28 3 : ae PAREN 3 2 ee” SYS = 2 4, g NOX = nO 4 “& gQ RN “Ab G Q CIN Zz Q Q So @ ON X Ga” ate? Woolslade < Q Q yt a7 Q ge %- 8 Q, Pd qt Rat a of & Thy Q a Sea Gg” & i 7 1 rary Ppnyagey eile Fig. 10 Map of Luton Lye redrawn from sketch in hand of Charles Bill, probably prepared 1767 (WSRO 1300-1928). The cross in “Marlborough Walk” (now known as the Grand Avenue), which may have been drawn on the original map as a later amend- ment, represents the location of Eight Walks, apparently constructed before 1775. The Place at present consists of four quarters or except one or two small openings or avenues. divisions viz Woolslade, Lewdens Lye, Ashlett and The improvement now making at Lewdens Lye by Bagden, the whole containing several hundred acres uniting open and woody parts seems to point out further and if entire would make one fine circle of about 5 miles improvements that might be made, by uniting Lewdens round - I have drawn a sketch or eyeplan of it, which Lye with Woolslade and Lewdens Lye with Ashlett though very imperfect may give you a better idea of it (perticularly by the latter) or at least by making some than I can by description. Woolslade is divided from larger openings into them - and some larger openings Lewdens Lye by a coppice running between them, and might also be made in Bagden Coppice (and some a pale standing upon an high bank - which effectively corners rounded or cutt off) instead of some narrow divides and makes two things of them. Lewdens Lye is walks or avenues that are cutt through it. divided from Ashlett by Bedwin Road (which causes These are the principal points his Lordship desires but little interruption there) by a ditch and large bank to have your opinion upon - and though he thinks he has and by a fine grove of young trees intermixt with thorns business enough upon his hands for the present nearer and coppice-wood, which shut out the sight of Ashlett home, the improvement in this part of the Forest, which from Lewdens Lye and makes two entire things of them, consists of fine lawns beautifully interspersed with wood, THE GREAT INCLOSURE OF SAVERNAKE WITH A NOTE ON CROSS VALLEY DYKES 171 will require little more than the moving of wood and improving the land - if a plan should be necessary it may be taken according to your directions. As to Bagden Grounds Mr Rolt is improving and laying them open, but though they lie in the circle, they cannot at present be united with the other parts. C Bill Note (another hand): Bagden Coppice is thin of wood ° and always will be by the injury it sustains from the deer - so that it would be no loss to convert such parts as may be proper into lawn - the whole is finely ornamented with large oaks and birch trees. A sketch map in the hand of Charles Bill, almost certainly the one referred to in the letter, is redrawn in Figure 10 (WSRO 1300-1928). This shows a large part of the area as open lawn. Works were clearly advanced. Woolslade was described as ‘chalked and improved into fine grass ground’. A great pit cut into one of the very few exposures of Upper Chalk in the Forest, facing Woolslade and east of Shovel Bottom at SU231661, was presumably the source for this. This pit is shown with a building on the estate map of 1786. Lewdens Lye is described as having ‘open and rough or woody parts improving to lay down fine lawn’, and Ashlett as having ‘fine level lawn wanting little improvement’. A cross marked on The Grand Avenue, probably added to the map subsequently, shows where Eight Walks became established, sometime between 1767, and 1775 when a document makes reference to the ‘Star in Lewdon’s Lye’, suggesting that it had already been constructed (WSRO 9-1-510).* It was certainly present by 1786, when shown on the map given in Figure 3. It appears incongruous that the rigidly straight avenue system of Eight Walks, created by the Estate, was contemporary with the design of serpentine routes, to visit features of interests, and gain expansive views, under the guidance of Lancelot Brown. However, it is clear that the Great Inclosure was still considered as a discrete entity when these further landscaping works were contemplated. The map shows it as paled or hedged for most of its boundary, with gates along the main routes. The Marlborough Walk (Grand Avenue) had been established. Hence in the 18th century, the bank delineated an area where park landscaping was the primary consideration. In the last part of that century, the keeper’s house at Bagden was replaced by a large and imposing stately home known as Savernake Lodge, and much of the Great Inclosure therefore formed the open landscaped setting to the grand mansion. Brown’s observations made during a visit on 17 and 18 March 1767 are recorded in a note by Charles Bill (WSRO 1300-1916): ‘Little Ashlett - Thinks Firs may grow amongst the bushes and furze but recommends it to be ploughed and cleaned, because small plants will then be sufficient, and one part may be made a nursery for the other, as it must in all events be fenced to do any good. Lewdens Lye, Ashlett, Woolslade: There are already two walks or openings between Lewdens Lye and Ashlett - They may be cleaned and made something larger and the corners rounded off but does not approve of destroying much coppice wood: - approves much of the business now executing in Lewdens Lye, and advises great care to be taken not to lay it down in grass, until we are sure all the heath is killed; as to Woolslade the corners may be rounded off at the entrance into the coppice out of Woolslade on each side the walk which is all that is necessary. Soul Bottom (sic). Advises a communication by a gate from the upper end of Woolslade Bottom into Shoul Bottom and to clean the Bottom and the rising ground some distance upon each side to show the oaks with an intent to make a ride up the bottom communicating with the several rides and walks in that part of the forest on account of its being so fine a valley’. It appears that, by the later 18th century, through a bizarre overturn of priorities, the Great Inclosure’s function had been reversed. Instead of an area of ‘thick and strong wood’, standing prominently amidst overgrazed heathland, it was retained as a grand park landscape, kept deliberately open, even when the surrounding land was smothered in woodland plantations in the 19th and 20th centuries. Cross Valley Dykes It seems likely that the Great Inclosure constructed in Tudor times took advantage of an existing great double bank and ditch system in Shovel Bottom (Figure 11). The difference in scale between the Inclosure boundary and the cross valley dyke is seen on a LiDAR image from this area given in Figure 12, and from the plan in Figure 13. The dyke straddles the valley obliquely here, and hence its construction would have been more resource- intensive than if aligned at right angles. This bank also changes direction: towards its western end, its course re-aligns from 254° to 294°, although there is no obvious explanation for this. At the junction of these otherwise largely straight sections, there is a gap in the bank on the southern side (see Figure 13). Assuming that this is not a designed gateway THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 172 looking eastwards 11 The cross valley dyke in Shovel Bottom, Fig. Sy SS Inclosure bank runs from with gap in structure and abrupt change in angle to west (D), north at top of image. ith orientated w g. 12 LiDAR image of southern stretch of the Great Inclosure, west (A) to east (B) Fi Cross Valley Dyke shown lower centre (C to F), 10n © Forestry Commiss Pit to east may have been the source for chalking Woolslade mid valley (E) ” “gateway and 1 THE GREAT INCLOSURE OF SAVERNAKE WITH A NOTE ON CROSS VALLEY DYKES aso yt si AN ang VV 8 ] = FF Tu Lf Deprrnanrrr rc ve / Aan yuu (Alii © aru cs WOW / / ull a mine // ssa aa Cy neg AE yy MEE / Aree z he i wo 0) nae cw 009 : Wi a Be sai eal Nee nn we / / é vyyg e Beat\\\) NAS nn p= 4H] Miting=i e e sped 70 =, TTA 1 hen " s cllllyayy nS une «wo Sol aMee a paca Udldee mary WO Fig. 13 Plan of Shovel Bottom cross valley dyke, showing change in angle and gap, with an abrupt head wall to west, followed by the minor bank of the Great Inclosure beyond, and gradual petering out and merging with Inclosure bank to east. feature for people or animals, it may simply be an artefact of the construction method, providing a route along which spoil could be removed from the inner ditch. However, the failure to close this gap would clearly have reduced the effectiveness of the barrier presented by the dyke. The cross valley dyke in Shovel Bottom resembles a feature in Great Lodge Bottom (see Figure 14). In both cases dykes extend across a small valley, and cease. Both have counterscarps, with the south-western or southern bank usually larger. There is no evidence that they continued or connected with other features. Two less prominent cross valley earthworks in Savernake may be analogous features. One is located further along Shovel Bottom at $U227657. A second example is at SU216666 in New Pond Bottom, another chalk valley which has tumuli to the north and south of the earthwork. These dykes have not been dated. At Shovel Bottom there is a tumulus in the same valley 400m NNE, and there may be signs of field systems on LiDAR images immediately south of Eight Walks. It is possible therefore that these valleys have been occupied and cultivated since prehistory, and that the dykes form part of a pre-historic or early historic landscape. What is remarkable at both Shovel Bottom and Great Lodge Bottom, is that the dykes are confined closely to exposures of chalk in the valleys. As soon as the clay-with-flints is reached, they cease. To this extent they replicate at the local scale, observations previously made for the Wansdyke generally, which extends across chalk, but has gaps along its course or becomes inconsistent across clay (Fox and Fox, 1960). It has Fig. 14 LiDAR image of Great Lodge Bottom (A to B), orientated with north at top of image. Apparent cross valley dyke near centre (C to D). Image also shows (unrelated) braided track traversing bottom left corner (E to F), and apparent equidistant, boundary-marker pits mid top. © Forestry Commission. 174 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE been conjectured that the Wansdyke was a defensive structure which intermittently entered impenetrable woodland. The notion that pre-historic woodland could have been so dense that it was impassable to a military force has little credence in current ecological thinking. It seems unlikely that the small cross valley dykes at Savernake, which may have been surmounted by a wooden stockade derived from nearby trees, could also have connected impenetrable woodland in this way. Whilst such dykes could have denoted territorial boundaries or might have regulated wheeled traffic in the valley floor, they seem functionless in military terms. It is therefore conjectured that the dykes were constructed to demarcate and possibly protect valuable arable land in the chalk valleys, whilst semi-natural habitat on the clay-with-flints adjoining was used as a common pool resource, probably for grazing, the provision of wood products, and hunting. Acknowledgements We would like to thank Ben Lennon of the Forestry Commission for access to LiDAR images for Savernake, for providing accurate measurements from map-work, and for making substantial contributions to interpretations generally; Dick Greenaway for archaeological and surveying guidance; and Joan Davies for providing ongoing ideas and suggestions. References Original Sources Wiltshire and Swindon Records Office (WSRO) WSRO 9-1-30, Act of Parliament concerning the lands of Sir Edward Seymour, Viscount Beauchamp, in Wiltshire, 1536 WSRO 9-1-509, Sketch Map for proposed realignment of Savernake Lodge entrance, 18th C. WSRO 9-1-510, Mr Bloxham’s list of Place-names in Savernake Forest, 1775 WSRO 9-22-204, Bargain and sale of land adjoining Savernake Great Park, from Sir Edward Baynton to Sir Edward Seymour, 1575 WSRO 9-22-221, Survey Savernake Forest and Great Park with notes on common land, cl778 WSRO 9-26-512, Sale of 210 oak trees in Stock woods, 1575 WSRO 2027-H, Map of Savernake Forest and Great Park, 1843 WSRO 1300-87, Savernake Forest Court Book, 1577- 1609 WSRO 1300-90, Survey of properties of Edward Earl of Hertford in Wolfhall, Tottenham Park, Crofton, Sudden, Crofton in Great Bedwyn, Bowden and Westcourt in Burbage, Brimslade in Savernake, Easton, Wootton Rivers, 17th century. WSRO 1300-104, Letter from Alexander Tutt (steward to Earl of Hertford) to Richard Wheeler, 1597/8 WSRO 1300-176, Copy Letters Patent granting Earl of Pembroke, property formerly of Edward, Duke of Somerset. Copy of original of 1552. WSRO 1300-301, Accounts of Coppices, 1716. WSRO 1300-320, Volume on commons and rights in Savernake Forest, undated, 18th C. WSRO 1300-360, Map of Savernake and Tottenham, 1786 WSRO 1300-363, Letter, John Ward concerning line for new Road from Great Bedwin to Marlborough, 1795. WSRO 1300-1484, Volume, compiled by William Longueville, in 1718, containing extracts of legal cases 16-18th centuries. WSRO 1300-1916, Charles Bill’s account of Lancelot Brown’s advice, 17-18 March 1767. WSRO 1300-1917, Copy of letter from Charles Bill, probably to Lancelot Brown, describing proposed landscaping at Lewdon’s Lye, Savernake, 13 March 1767 WSRO 1300-1928, Sketch Map in hand of Charles Bill relating to proposed landscaping, undated but apparently referred to in WSRO 1300-1917 of 1767. WSRO X6-4, Map of Savernake Estate, 1820. Public Records Office Close Roll, Licence to Matthew de Columbarius to take deer for stocking Chisbury Park, 1259-61, 355. Calendar Patent Rolls, Licence to John de Havering to enclose 421 acres in Savernake, 30 Edward I. E 146-2-22, Inquisition into Bounds of Savernake Forest and its bailiwicks, 28 Henry II — 1 John. Longleat House Seymour Papers Volume V, fos 56, 57d, Letter John Thynne to Earl of Hertford, Jan 27 1574/5. Printed and Published Sources CRAWFORD, O. G. S. 1920. The Anglo-Saxon Bounds of Bedwyn and Burbage. WANHM 41, 281-301 CROWLEY, D.A.1999. VCH Wilts. XVI Kinwardstone. Oxford: Oxford University Press FOX, A. and FOX, C. 1958. Wansdyke Reconsidered. Archaeological Fournal 115, 1-48 GREAT BRITAIN. COMMONS ACT 2006: Elizabeth II. Chapter 26, London: The Sationery Office HOUSE OF COMMONS, 1788, Fournal 43, p. 561 (Debate on land use) MAJOR, A.F. and BURROW, E.J. 1926. The Mystery of the Wandsdyke. Cheltenham: Ed. J. Burrow WATTS, K. 1996. Wiltshire Deer Parks: An Introductory Survey, WANHM 839, 88-98 The form inclosure is employed generally in this paper (in contradistinction to enclosure), conforming with contemporary documents cited, and being the legal term for the removal of lands from wastes and commons, cf simply bounding an area. Watts (1996) conjectures that de Havering’s inclosure was at Grafton. However, place-name evidence suggests Littleworth as more likely. In this context, meaning graziers turning out cattle for payment. This might be discretionary, or recognised as custom and practice, implying a right (although distinct from a ‘right of common’ for which no payment could be demanded). THE GREAT INCLOSURE OF SAVERNAKE WITH A NOTE ON CROSS VALLEY DYKES 175 Notes This document conflicts with the evidence above. It gives dates ‘respecting Tottenham Park and Savernake Forest collected from various sources but chiefly from old memoranda in the handwriting of Mr Bloxham and Mr Ward and from oral information given to the latter by Mr Philip Pearce about 50 years ago. Mr Bloxham wrote an account of Forest names and dates of plantations in 1775 — and Mr Ward continued the account during his stewardship’ . It claims that ‘the star centring at Lewdens Lye was made in 1733’. This reported recollection cannot be reconciled with the map and notes in Charles Bill’s hand, which appear unequivocal. Wiltshire Archaeological & Natural History Magazine, vol. 101 (2008), pp. 176-92 Evidence of 17th century brick-making at Highworth, Wiltshire by Donna E.Y. Young with contributions by Mark Corney, Alejandra Gutierrez, Mark Noel and Sarah Newns Archaeological investigations in advance of residential redevelopment close to the centre of the historic town of Highworth, near Swindon, revealed evidence of a brick-making industry that operated during the mid-late 17th century. A series of associated pits reflected successive phases of clay extraction over the period AD 1600-1800 providing raw material for the manufacture of bricks, some fired in a lozenge-shaped kiln. Other finds provided evidence for Romano-British, Anglo-Saxon and medieval activity either on the site or nearby. Introduction Archaeological investigation by the Avon Archaeological Unit Ltd was undertaken as part of the planning process on a site previously occupied by 1960s shops and carparking at Glebe Place, Cherry Orchard, Highworth, prior to residential development by Barratt Bristol Ltd. The historic town of Highworth is situated some 9km to the northeast of Swindon and occupies a prominent location at c. 130m O.D. on the crest of a low hill, part of a range extending through the Vale of North Wiltshire and overlooking the Thames Valley and Vale of the White Horse. The town lies on Jurassic Corallian Cuesta, a sequence of Coral Rag and associated limestone, underlain by sands, grits and loams and largely overlain by light soils. This report concerns a small archaeological excavation undertaken in January 2006 on a plot of land close to the heart of the town (NGR SU20139264, Figure 1), where the geology comprises heavier clay soils, possibly associated with a localised outcrop of Red Down Clay that interrupts the sequence in the area. Excavation was carried out in the light of a preceding evaluation and was designed to record a sample of the archaeology within the footprint of one of the proposed residential buildings. Background to the excavation Preliminary investigations included documentary research (Etheridge 2004) and archaeological trial excavation (Young 2004). SMR (HER) records highlighted the rich archaeological potential of the town and its environs, extending back into prehistory, while documentary sources indicate that the settlement was established by the time of Domesday Survey of 1086, Highworth representing one of only three hundreds named in the Wiltshire survey (Darlington 1955, 177). The development site lay on the northern edge of the Highworth Conservation Area, which encompasses the historic centre of the 13th century medieval planned town. It has been suggested that elements of the central Avon Archaeological Unit Ltd, Avondale Business Centre, Woodland Way, Kingswood, Bristol BS15 LAW EVIDENCE OF 17TH CENTURY BRICK-MAKING AT HIGHWORTH, WILTSHIRE 177 Hs Hs Speer el area Boundar ts ALTER ~>~--_ 10 13 SSAA ; ic lebe Piss =---. / TON R ; lace Entre Nous ie ne. Simon 1 Deg ; \ Published by The Wiltshire Archaeological _and Natural History Society ISSN 0262 6608