] a log y e0 istor’ Archa H ire sh t Natural i THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL VAIS TORY?S OCIEIEY The Society was founded in 1853. Its activities include the promotion of archaeological and historical work within the County, and of the study of all branches of Natural History; the issue of a Magazine and other publications; excursions to places of archaeological and historical interest; and the maintenance of a Museum and Library. The annual subscription rates for membership of the Society are: Individual Mem- bership £2.50; Family Membership £4.00; Student Membership £1.50; Junior Member- ship £0.50. Enquiries about membership should be made to the Secretary of the Society, 41 Long Street, Devizes, Wiltshire. OFFICERS OF THE SOCIETY TRUSTEES E. C. Barnes, Esq., C. E. Blunt, Esq., 0.B.£., F.B.A., F.S.A., The Right Hon. The Lord Devlin, P.c., F.B.A., E. G. H. Kempson, Esq., M.a., Bonar Sykes, Esq. VICE-PRESIDENT Professor R. B. Pugh, M.A., F.S.A., F.R.HIST.S. COUNCIL C. E. Blunt, Esq., 0.B.E., F.B.A., F.S.A. (President and Chairman) R. E. Sandell, Esq., M.a., F.s.A., F.L.S. (Hon. Librarian; also representing the Wiltshire Record Society) K. H. Rogers, Esq., B.A., F.s.A. (Hon. Assistant Librarian) Miss I. F. Smith, B.a., pH.D., F.s.A. (Hon. Editor of W.A.M.) Maurice G. Rathbone, Esq., A.L.A., (ex-officio as County Archivist) H. de 8. Shortt, Esq., M.A., F.S.A., F.R.N.S. (ex-officio as Curator of Salisbury Museum) Group Captain F. A. Willan, G.B.E., D.F.c., D.L. (representing the Wiltshire County Council) Air Commodore H. Eeles, c.B., c.B.E. (representing the Wiltshire County Council) Professor R. J. C. Atkinson, M.A., F.S.A. R. S. Barron, Esq., M.A., B.SC. D. J. Bonney, Esq., B.A., F.S.A. N. J. Gordon Clark, Esq. Miss M. W. Evans, A.R.C.A. P. J. Fowler, Esq., M.A., F.S.A. Miss E. Dampier-Child, 0.B.£. Miss B. Gillam R. Macdonald Smith, Esq., M.A. Lieut.-Commander J. E. Manners, D.s.c. Mrs. J. M. Morrison R. P. de B. Nicholson, Esq., G.ENG., M.I.MECH.E. Miss S. F. Rooke H. Ross, Esq., B.A. J. M. Stratton, Esq. N. P. Thompson, Esq. Mrs. F. de M. Vatcher, F.s.A. SECRETARY AND TREASURER Brigadier A. R. Forbes CURATOR ASSISTANT CURATOR F. K. Annable, Esq., B.A., F.S.A., F.M.A. A. M. Burchard, Esq., M.A. HON. ARCHITECT HON. SOLICITOR D. A. S. Webster, Esq., M.A., F.R.ILB.A. H. G. Awdry, Esq., B.A. THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE Volume 66 1971 PART B: ARCHAEOLOGY AND LOCAL HISTORY PRINTED IN GREAT BRITAIN BY HEADLEY BROTHERS LTD 10Q KINGSWAY LONDON WC2B 6PX AND ASHFORD KENT THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE VOLUME 66 1971 Wall Bs ARCHAEOLOGY AND LOCAL HISTORY CONTENTS An ARCHAEOLOGICAL SURVEY AND PoLicy FOR WILTSHIRE: Part III, Neourruic AND Bronze AGE, by Stuart Piggott - - - - - - - - ARCHAEOLOGICAL RESEARCH IN THE Pewsey VALE, by N. P. Thompson, with contributions by F. K. Annable, I. W. Cornwall, R. A. Harcourt, E. Machin and R. F. Tylecote - . - - - - - - THe EXCAVATION OF PREHISTORIC AND ROMANO-BRITISH SETTLEMENTS NEAR DuRRINGTON WALLS, WILTSHIRE, 1970, by G. J. Wainwright, with P. Donaldson, I. H. Longworth and V. Swan - - - - - - WANSDYKE: EXCAVATIONS 1966 To 1970, by H. Stephen Green, with contributions from G. W. Dimbleby, J. N. L. Myres and D. J. Bonney - - - - CATHEDRAL FURNISHINGS OF THE RESTORATION PERIOD: A SALISBURY INVENTORY oF 1685, by Robert Beddard - - : - - - - - - Jerrry Hunr Pipes, 6y D. R. Atkinson - - - - - - - - NINETEENTH-CENTURY COUNTRY CARRIERS IN NORTH WILTSHIRE, by Alan Greening - NOTES: A Bronze Basal-looped Spearhead from Wilcot; Cricklade, the Roman River Port of Corinium; An Anglo-Saxon Brooch from Grafton; The Naish Hill Kilns: an Interim Report; Ice-houses at Lydiard Park and Southbroom House; Narrow-gauge Railway on Salisbury Plain; List of MS. Pedigrees in Cricklade Museum - - - - - - - - - - EXCAVATION AND FIELDWORK IN WILTSHIRE, 1970 - - o - = e OrrFicers’ REPORTS AND ACCESSIONS, 1970 - - - - - : : z ANNUAL GENERAL MEETING, 1971 - - - = 7 e = = = REVIEWS - - = = S = = < . = : = s o OBITUARIES - = é = = = = 2 = = . = cz INDEX TO VOLUME 66 - - = a = = = - = = ce PAGE 177 188 192 203 205 207 208 PLATES - - - - - - - - - - - - at end of volume ill a agiet? sae? & 048 wa a wd Spetiy | R28 loed Af = : : a. 7 a bat VA. u Stes oe ee = are | 2 @ Fo =e ; ee ee i i Vp 4 oi) n + ‘ t 7 : > nig a nM ] A, AN ARCHAEOLOGICAL SURVEY AND POLICY FOR WILTSHIRE PART III: NEOLITHIC AND BRONZE AGE by STUART PIGGOTT INTRODUCTION PARTS I AND II of this survey of Wiltshire archaeology and its problems were published in 1969,' and dealt with the Palaeolithic (Roe) and Mesolithic periods (Radley). This part continues chronologically from Part II and comprises the cul- tures marking the appearance of stone-using agriculturalists (conventionally the Neolithic period), which in their later phase overlap those of communities with the added technology of non-ferrous metallurgy, initially in copper and gold and soon in the copper-tin alloy of bronze, which in the Three Ages system began a ‘Bronze Age’. This in its turn continues until the introduction of iron-working and other cultural innovations and the opening of an ‘Iron Age’. In conventional radiocarbon dates the total span of this Part III is from the late fourth millennium B.c. to around the seventh or sixth century B.c.? The Victoria County History of Wiltshire, Vol. 1, Part 1 (1957) contains a Gazetteer of the relevant archaeological material to that date. Part 2, a comment and dis- cussion on the evidence available to 1970 by the writer and Professor B. W. Cunliffe, will be published shortly, and in view of this situation the following treatment will minimize the ‘survey’ aspect and concentrate on ‘policy’. Full bibliographical references to the Wiltshire evidence will therefore be found in these volumes. It is assumed that in the immediate future and for some time to come most excavations in the county will be rescue operations in advance of destruction, but nevertheless such operations can usually be made to serve the purpose of research. Small research excavations with limited objectives, as for instance the recovery of samples for C4 determinations, should not, however, be impossible to achieve with a minimum expenditure of resources. The desirability of publishing material in corpus or catalogue form with draw- ings of high technical standards is of course as great in the period dealt with here as in other phases of prehistory. The Society has given a lead in the publication of Annable and Simpson’s admirable Guide Catalogue to the Neolithic and Bronze Age collections in its Museum, and D. L. Clarke’s Beaker Pottery of Great Britain and Ireland (1970) publishes in corpus form the Wiltshire material as part of the wider survey. Such publications, however, inevitably need the provision of addenda et corrigenda by those with local knowledge and in touch with new discoveries. A periodical list, with illustrations, of Wiltshire beakers and associated finds, supplementing Clarke’s corpus, could be of considerable value. 47 CHRONOLOGY Since the initiation and application of isotopic dating to organic material containing radio-active carbon ( Ci4) our prehistoric chronology for the past 50,000 years has been directly or indirectly obtained by this method, itself dependent upon samples of charcoal, or other substances containing Ci, in measurable quantities, from closely defined archaeological contexts. Wiltshire has contributed its share to the general body of dates which now give a framework of absolute chronology to British prehistory, but far more are needed. The period now under review contains at least one major problem which radiocarbon dating might do much to resolve, the date and affiliations of the ‘Wessex Culture’, and this is further discussed below. It is in part involved in the recent demonstration that radiocarbon ‘years’, previously expressed as the equivalent of normal calendar years B.P., B.C., Or A.D., are not in fact coincident, but vary as a function of time. What are now being called ‘conven- tional’ radiocarbon dates can be corrected against true dates for a period up to about 6,000 years ago, and so alternatively expressed in adjusted form, and it has been suggested that confusion could be avoided by using the symbol ‘be’ for conven- tional Cy, dates , with the normal ‘B.c.’ reserved for those corrected by means of the curve worked out by Suess, or of course dates derived directly or indirectly from historical sources using calendar years.3 ‘This ingenious and convenient convention will be used throughout this paper. The theoretical situation outlined above can only be strengthened, refined and tested by the determination of far more radiocarbon dates from samples derived from accurately ascertained associations with characteristic archaeological material. From the late fourth millennium to the middle third millennium be we have a fairly respectable cluster of dates, but cumulative probability is increased by every additional determination, and in many periods we have only one or two, or no, radiocarbon dates throughout Britain. In any future policy of research a deliberate return to sites excavated in the past for the express purpose of obtaining samples for radiocarbon assay should be considered as a serious possibility, and it should be remembered that bone and antler in sufficient weight contain residual collagen with a Ci4 content, and may be used. One thinks in particular of selected bones from human burials with associated grave-goods in this context, and the work need not always involve excavation in the field, but in the reserve collections of museums. For a start, a census of extant Wiltshire material which could potentially provide samples for radiocarbon dating (with its inevitable destruction of the sample) could be of great help in planning such a campaign of dating on a larger scale. Specific problems involving radiocarbon dating will be taken up in the ensuing sections. MESOLITHICG AND NEOLITHIC The relationship of the hunting and gathering cultures of the Mesolithic to the agricultural communities constituting the earliest Neolithic is still uncertain. Radiocarbon dates suggest a considerable chronological gap (e.g. Mesolithic at Cherhill c. 5280 be and the earliest Wiltshire Neolithic at Horslip long barrow, ¢. 3240 be) but more evidence is needed, in the form of radiocarbon dates from Meso- 48 lithic sites, and demonstration of stratigraphical relationships—vertical, as at Cherhill, or horizontal, as inferentially at Downton. It is known that elsewhere in southern England Mesolithic cultures have a relatively longer survival, as for instance ¢. 4350 be (Oakhanger, Surrey) or even c. 3310 be (Wawcott, Berks.),4 the latter overlapping with the dates of two Wiltshire and one Berkshire long barrow.5 As Radley pointed out, ‘it is to the Wiltshire valleys that future fieldworkers must look for Mesolithic sites’ in Wiltshire, and as we shall see, this may well be the case for Neolithic settlements as well: Cherhill points the way.® ANCIENT SOILS AND LAND USAGE Following the lead of Evans,7 every opportunity should be taken to obtain information on the natural soil cover of the chalk areas before human interference, and so inferentially of its original vegetation. Buried soils beneath Neolithic monu- ments such as barrows or (if sufficiently large) the banks of causewayed camps or henges could give most valuable evidence on the contemporary environment, and Silbury Hill has shown dramatically in recent years what can result from the close cooperation of archaeologists with natural scientists in this field. The recovery techniques of flotation, sieving, etc., recently elaborated for plant and similar remains by the Cambridge research teams working on the origins of agriculture in the Levant and East Europe might well be applied here. There is an increasing body of evidence to suggest that human interference with the natural landscape began in Mesolithic times, and was of course increased by the agricultural activities of the first Neolithic colonists. Wiltshire has produced the earliest plough-furrows in Britain under the South Street long barrow, with a date before ¢c. 2750 be (which could be around 3500 B.c.), and this is a topic of European importance, as it raises the question not only of the first ploughs, but of the first draught oxen (and so of castration) and possibly of wheeled transport as well. Every effort should therefore be made to examine with the greatest care ancient land sur- faces exposed by excavation under Neolithic and later monuments, in collaboration with natural scientists.9 Evidence for early cultivation on soils other than the chalk (e.g. the greensand) should be looked for in particular. SETTLEMENTS The virtual absence of settlement sites of the Neolithic or earlier Bronze Age cultures in Wiltshire, as in much of southern England, has been frequently com- mented on, and generally attributed to the weathering, solution and erosion of the chalk downland.'° Such a process can, as at the Fussells Lodge long barrow, remove up to half a metre of the original surface, and so only the remains of pits or post-holes dug to over this depth can be expected from sites dating from around 3000-2500 be in such situations. The chance finds of pits containing Neolithic material at various points on the Wiltshire Downs should be pointers to possible settlement sites,!'! and Bradley’s recently published excavations of a Beaker period settlement on the Sussex Downs show what can be done by the patient recording of vestigial features and the use of techniques such as trend mapping of finds.1? The careful application of such 49 methods might well lead to the recovery of settlement plans on sites previously thought be be eroded beyond redemption: further excavation on the Easton Down settlement, extending Stone’s work of a generation ago, might be very rewarding. The Sussex site at Belle ‘Tout was in a valley on a chalk plateau and was origin- ally identified (in 1909) from surface finds of flints and sherds: this may give a clue to the likely siting of such settlements in Wiltshire, and indeed Easton Down occupies a similar position. Furthermore, the Belle Tout site is contained within a contempor- ary roughly rectangular earthwork of Toms’s ‘valley-bottom enclosure’ type. Such slight enclosures might profitably be searched for and should be possible to distinguish from the more conspicuous rectangular earthworks of medieval date as on Morgan’s Hill in Wiltshire or Handley Hill in Dorset.%3 On the other hand, the Cherhill site once again draws our attention to the fact that Neolithic and Bronze Age settlements need not always be sought for on the now- eroded chalk uplands, but may lurk unsuspected in the valley bottoms. Discovery here is likely, as at Cherhill, to be a matter of chance observation when the ground 1s disturbed by building, drainage or similar activities, but it is worth remembering that a large sherd of Ebbsfleet pottery and a Bos primigenius skull were recovered from the streambed there many years previous to the discovery of the Mesolithic and Neolithic settlements. It would be worthwhile therefore to collect any instances of valley-bottom finds of Neolithic and earlier Bronze Age date and where possible to follow these up by fieldwork. The Pewsey Vale must have been accessible and in- habited in late Neolithic times at least, as the new Marden excavations have shown us,'4 and it might even be possible to discover waterlogged or damp sites, with a consequent enrichment of finds of environmental significance. Marden and Durring- ton Walls also demonstrate that massive circular timber structures which need not all be interpreted as temples or ritual post-circles were being built in later Neolithic times, and could therefore appear in purely ‘secular’ contexts. The settlements and enclosures attributable to the Deverel-Rimbury culture will be commented on later, but in the meantime it is worth remarking in passing that the Rams Hill earthwork enclosure on the Berkshire Downs, with a flat-bottomed ditch and sherds of a cinerary urn of Longworth’s Primary Series, can hardly be unique in early second millennium Wessex, although its counterparts have so far eluded discovery,'5 and constitutes a challenge to field workers. CAUSEWAYED CAMPS Here, although numerous excavations have taken place, many problems are unsolved. There seems a possibility that the concentric-ditched type of Windmill Hill, Robin Hood’s Ball or Whitehawk in Sussex performed a function in Neolithic society different from such sites as Knap Hill, or Hembury in Devon, but this seems difficult to determine. If further excavations take place particular attention should be paid to the ditch fillings, in order to test afresh the thesis that in some instances at least deliberate infilling rather than natural silting took place:'® here differentia- tion of function might be inferred, as the Knap Hill ditch appears to have silted naturally, unlike those of Windmill Hill, Robin Hood’s Ball and Whitesheet. 50 The other aspect that might be pursued, with Belle Tout again in mind, is the further investigation of the interiors of the enclosures, in the hope of recovering the residual vestiges of structures. Where banks remain of sufficient height to have protected the ancient soil, further excavation on the lines of that carried out on a small scale under the Outer Bank at Windmill Hill might be rewarding for both archaeologists and palaeoecologists. HENGES Wainwright’s spectacular discoveries at Durrington Walls and Marden in Wilt- shire, and at Mount Pleasant in Dorset, show us both what surprises these monuments can still spring on us, and also how the internal features can only be recovered by excavation on a very large scale. Future policy, guided by this new knowledge of the internal complexities latent in these monuments, must take the form of urgently promoting their complete conservation. The preservation of surviving field monu- ments of all kinds is now becoming an increasingly pressing necessity, but the great henges, with their large areas (35 acres (14 ha) at Marden; c. 30 acres (12 ha) at Durrington) which contain features visible only on excavation, present the same problems as prehistoric field-systems and extensive settlement sites ill-defined on the surface when the archaeologists’ view-point has to be presented persuasively to the authorities behind the threat of destruction. LONG BARROWS Apart from their intrinsic interest as our earliest monumental structures, the long barrows by their very bulk cover more of the Neolithic land-surface of Britain than any other class of antiquity. At a very rough estimate the long barrows of Wiltshire could between them cover somewhere around 10 acres (4 ha) of fourth to third millennium bc landscape and in the ploughed-down condition of so many, this ancient surface is relatively accessible to excavation. If the lower part of the mound is composed of turfs, as seems to have been the case in many barrows excavated in the last century, this of course opens further opportunities for palaeoecologists. As experience has shown that the total excavation of such barrows is demanded if they are to be adequately understood, the opportunities for palaeoecological studies, and for the detection of other features such as cultivation or field enclosures beneath long barrows is very considerable. Although much new evidence has come to light, and many features found in earlier excavations made susceptible of interpretation in new terms, there are still many serious problems needing elucidation.'7 South Street and Beckhampton Road both show the extraordinary features of hurdle-work compartmentation and the lack not only of burials but apparently of any provision for them,'® and one wonders how unusual these barrows may be: the hurdle-work is partly paralleled at least in the Cotswold barrow at Ascott-under-Wychwood. Ashbee has claimed, initially on the evidence of Wayland’s Smithy and Fussells Lodge, that pitched-roof timber mortuary houses were normal to British ‘unchambered’ long barrows, and that collateral evidence can be indirectly gleaned from earlier excavation reports: this Al has been called into question’? and future excavators will have to examine any presumptive features of this type with more than usual care. Again, Wayland’s Smithy2° not only demonstrates the possibility that some barrows may embody structures of more than one period, but raises in fresh terms the long-standing problem as to the precise nature of ‘collective’ inhumation burial in wooden- or stone-chambered Neolithic tombs. Here an exact and informed study of the skeletal remains in situ is demanded, with the archaeologist working with a human anatomist in the field during the whole recording process, as at Millin Bay in Northern Ireland?! or at Ascott-under-Wychwood. All these problems are more than usually of an importance beyond Wiltshire, for as Simpson stressed,”? they are intimately connected with the relationships of the British Neolithic cultures to the Continent and of our own long barrows to those monuments we now know to be broadly contemporary on the North European plain.3 FLINT MINES AND WORKING SITES Easton Down and Durrington are at present the only flint-mines identified in the county: the former were found by field-work, the latter as a chance discovery in commercial digging. More may well exist and, with increased ploughing of the chalk, surface concentrations of the characteristic waste from working floors might be the best clue. We now know that Easton Down was active around c. 2500 be (between the Sussex mines beginning ¢. 2960 and Grimes Graves at c. 2000 be) and the demand for flint axes must have been continuous during the third millennium. Very recently the preliminary results of a new technique for the identification of flint-mine products by trace element analysis has been published, and Easton Down has a characteristic and statistically significant assemblage of elements distinguishing it from other centres of production in England and on the Continent.?+ Further work on these lines will certainly be illuminating. MATERIAL CULTURE: STONE AND FLINT Moving from monuments to artifacts excavated in the field or in the stores of museums, several lines of approach could profitably be followed in any policy of research. ‘Thanks to the work of the Sub-Committee of the South-Western Group of Museums over the past 34 years Wiltshire is fortunate in having the greater bulk of its implements of stones other than flints thin-sectioned and petrographically identified. Gaps, however, remain and should be filled, and new finds as they appear added to the corpus. The stone bracers of the Beaker cultures have never been subjected to any form of analysis, though superficially the stone used seems characteristic and consistent. Here some approach using non-destructive methods could perhaps be devised, or by minute sectioning of broken ends where they exist.25 Recent advances in flint identification have been noted above. More work needs to be done on the flint industries which range from the earliest Neolithic to the Deverel-Rimbury culture in Wiltshire. An indication of methods which could profitably be employed has been given in respect of the Windmill Hill and Avebury flint assemblages,*° and other simple statistical and allied methods are 52 now becoming familiar to archaeologists. In any such studies, while material from closed contexts would naturally form the framework, well-defined surface assem- blages such as that from King Barrow Ridge?7 deserve examination and indeed their homogeneity or heterogeneity could best be tested by simple mathematical pro- cedures. The Easton Down settlement has an important industry deserving fresh study on modern lines, as have Deverel-Rimbury sites such as Thorny Down. A thorough examination of second millennium flint industries might do much on the one hand to illuminate the late Neolithic situation, with its sudden appearance of petit tranchet derivatives apparently in constant association with Grooved Ware, and on the other, might lead to the identification of sites intermediate between this and Deverel-Rimbury. Stone long ago saw the importance of his flint assemblage from the Stockbridge Down (Hants) pit, associated with Overhanging Rim Urns, as a pointer in this direction, and in connection with the Mildenhall Fen industry with ‘Wessex’ Biconical Urns.?8 If industries could be defined, surface finds might lead to the recovery of settlement sites. There is another approach to flint industries, as yet hardly exploited in this country, the identification and interpretation of characteristic patterns of usage and wear on the working edges and faces of implements. Pioneered by Semenoy?9 and his pupils in the USSR it is now being developed by Charles MacBurney and Ruth Tringham among others, and opens important new avenues of approach, which had previously gone little further than the observation of ‘sickle-gloss’ resulting from cutting siliceous plant-stems. POTTERY New technological approaches also promise additional information from pottery beyond its typological classification. Peacock has recently shown how petrographical determinations of the composition of Neolithic sherds can show the distribution of pottery made of the gabbroic clay of Lizard Head in Cornwall as far east as Robin Hood’s Ball and Windmill Hill, and so has re-opened the whole question of the application of analytical methods to the identification of pottery-making centres throughout prehistory.3° More work could perhaps be done on these lines (which involves thin-sectioning and examination by a petrologist), and again some of the Grooved Ware problems might be tackled on a new front. Beakers are known in certain instances to incorporate ‘grog’ in their fabric in the form of crushed fragments of other beakers,3! a technique also known in certain areas of the Continent, but its occurrence in detail has never been worked out in any region. These and allied approaches might do much to give a new and valid dimension to pottery studies. The techniques of dating pottery or other fired clay by thermo-luminescence are still being developed, but extremely interesting results have already emerged, agreeing, incidentally, with the new calibrated readings of Ci, dates already referred to.32 The process demands sherds newly excavated and in their detailed context of surrounding soil, but such circumstances could be obtained in new excavations, and determinations of the TL (Thermoluminescent) date of British Neolithic sherds would not only be most interesting in themselves, but in comparison with other TL op) dates such as those of the Linear Pottery culture in Germany, the Netherlands and Czechoslovakia already obtained. Helbaek’s pioneer work on the impressions of seeds (including cereal grains) in prehistoric pottery in Britain dealt with material available to him up to 1952. Nearly 20 years later, with a considerable amount of new finds, we appear to have no further identifications from Wiltshire sherds of any period.33 This is clearly work that should be done, and in an Archaeological and Natural History Society collabora- tion between archaeologists and botanists skilled in seed recognition and taxonomy, could most appropriately be effected.34 For the Neolithic and Bronze Age in particular, much might be added to our knowledge of early agriculture by this means. ROUND BARROWS With the current advances in excavation techniques, recommendations on policy in rescue operations on round barrows become platitudinous. It is now obvious that total excavation is needed, that the burials and structures associated with them may be of several periods (even if not distinguishable in vertical stratigraphy in a much degraded mound), and that features in fugitive substances such as wood or turf demand acute observation and recognition. The importance of the old land surface for palaeoecological studies has already been stressed in the instance of long barrows, as has the use of turf in the barrow mound, and the desirability of obtaining suitable samples for C:, dating is obvious. A minor point with regard to inhumation burials may be worth watching for, and this is evidence that the burial might have been robbed of some (or all) of its grave-goods before the end of the funeral rituals and the completion of the barrow. Tomb-robbing of this kind certainly took place, for instance, in richly-furnished Hallstatt tombs in Germany and has occasionally been suspected in earlier contexts. Apart from rescue excavations on ploughed-out sites or those about to be des- troyed by other means, serious thought should be given to deliberate re-excavation in certain important barrows of the so-called ‘Wessex Culture’. This phase of British prehistory has been under deserved attack from several quarters—the nature of its genesis, its status as an archaeological culture, its foreign relationships and above all its absolute chronology.35 It must be remembered that we have no direct radiocarbon dates for the Wessex Culture, but the graves with halberd-pendants as at Manton (Preshute ra) and Wilsford 8, should approximate in date to the Continental metal- shafted halberds they so closely copy, and here Cr, dates of c. 1660 be are available: adjusted on the Suess curve already referred to this would mean a true date of c. 2000 B.C. This would effectually remove this phase of the Wessex Culture from any possible contact with early Mycenaean Greece at the historical computed date of ¢c. 1600— 1550 B.c., the radiocarbon equivalent of which would be c. 1275 be. Nor have we any reason to think of the Wessex Culture as really being divisible into distinct chronological phases, or of anything but a very short duration. This point cannot be argued fully here and partly depends on work unpublished or in progress, but it has for instance been cogently argued that all the fine ‘Wessex’ gold-work could well 54 have been produced by two master-craftsmen, and so is unlikely to span more than half a century.3® As controls we badly need direct Ci, dates for Wessex Culture burials, and the majority of the finest of these are the products of Colt Hoare’s excavations in Wilt- shire barrows. Hoare and Cunnington did not remove the skeletons from inhumation graves: the ‘tall and stout’ man in Bush Barrow should be there still, ready to yield a Ci, date from his collagen. Re-excavation of the classic barrows would enlighten us in many respects other than chronology—does the great bell-barrow, Amesbury 15, really contain a gable-roofed timber mortuary-house, as Hoare’s tantalizingly laconic account of oak timbers “diverging in an angular direction from the top of the barrow to the interment’ suggests? Any such re-excavations would have to be on a total scale, and would involve high standards of skill and a considerable expenditure, but if we dare to hope that long-term planning of research beyond the exigencies of rescue operations may one day be possible, a new attack on the Wessex Culture in the field might be planned in these terms.37 THE DEVEREL-RIMBURY CULTURE This cultural assemblage, though well defined and well represented in Wiltshire, still presents many uncertainties which might be elucidated by further research. In the first place, it seems likely that field-work (with air-photographs) and surface sherd-collecting could add to the number of settlement sites, open or ditch-enclosed. If the diagnostic features of a distinctive flint industry could be worked out, as sug- gested above, this might enhance the chances of new discoveries owing to its better survival-value on ploughed land. What is particularly important is to seek every trace of possible relationship between such sites, datable barrows, and ancient field-systems where vestiges still survive: relationship to boundary ditches, touched on again below, is again important. There has been a tendency in the past to regard Deverel-Rimbury enclosures of the type of South Lodge or Boscombe Down East in Wiltshire as cattle enclosures without internal structures. It is argued in greater detail in the forthcoming VC'H volume that it is extremely likely that Pitt-Rivers was unprepared for post-hole structures in his excavations, and his ‘trenching’ technique of examining the interior of his ditched enclosures would not only not have permitted of their recognition, but could have destroyed them in its very process. Thorny Down after all produced a mass of post-structures and pits, but the interior of the Boscombe Down enclosure was hardly touched, and the same goes for the Ogbourne group. What is needed is the careful stripping of the interior of more than one of these enclosures where this could still expose a reasonably undisturbed chalk surface. Until then their function must remain uncertain. Linear earthworks of the type originally studied by Hawkes at Quarley Hill in Hants,38 while certainly earlier than some hill-forts and later than a Deverel-Rimbury enclosure at Boscombe Down and probably the Woodhenge ‘Egg’, still remain an outstanding problem. Excavation is likely to yield negative results as in the past, 53 unless at a point where a relationship to a settlement or another earthwork can be established. The dating and duration of the culture still depends on very few fixed points. The earliest Ci, date, of c. 1380 be for Barrel and Globular Urn sherds in the Wils- ford Shaft, stands alone at present, the only other radiocarbon determination being that of c. 1180 be from the Shearplace Hill (Dorset) site, a settlement of more than one phase.39 ‘The bronze ornaments of ‘Somerset Hoard’ types which occasionally occur in Deverel-Rimbury contexts have on archaeological grounds been given a date around the 12th century B.c., and the Cy, date of the Drakenstein Culture settlement of Nijnsel in the Netherlands, of c. 1140 bc, would be appropriate here.4° Samples for C,, dating from Wiltshire Deverel-Rimbury sites are therefore very desirable: animal bones from earlier excavations should still be available. LATE BRONZE AGE Our knowledge of the final phases of the Bronze Age, c. 700-500 B.c. is in Wilt- shire as elsewhere almost wholly dependent on the bronzes themselves. We do not know how long something recognizable as the Deverel-Rimbury culture may have survived, nor what pottery was used before that conventionally within the Early Iron Age, c. 630 be at Longbridge Deverill Cow Down. There are hints, however, that settlement sites might be identified: the seven certain or presumptive pieces of Late Bronze Age metal-work at All Cannings Cross seem too many to be accidental, and remind one of the Yorkshire finds of Staple Howe and Scarborough, and nearer home, at Ivinghoe in Bucks.4" Identification of such sites would be of the greatest help in the understanding of our first archaeological material unequivocally associated with iron metallurgy, and ushering in a new phase of British prehistory. NOTE This paper is published with the aid of a grant from the Council for British Archaeology. 1 W.A.M. 64, 1969, 1-20. 2 For ‘conventional’ and other Cy4 dates see discussion below. 3 Outline of new position in Neustupny, Antigq. 44, 1970, 38; ‘BC’ and ‘bc’ dates in Suess, tbid., 92; Renfrew, Proc. Prehist. Soc. 36, 1970, 280. Former conventions such as the use of the ‘old’ half-life of C,, and the use of A.p. 1950 as ‘present’ of course continue. 4'The Wawcott date is 3310 + 130 be; BM-449 (Radiocarbon 13, 2, 1971, 173). 5 Horslip, 3240 + 150 be (BM-180); Fussells Lodge, 3230 + 150 be (BM-134); Lambourn, 3397 -- 180 be (GX-1178). 6 W.A.M. 64, 1969, 20; Cherhill in zbid. 63, 1968, 107. 7 W.A.M. 63, 1968, 12. 8 Simmons in Ucko and Dimbleby (Edd.) The Domestication of Plants and Animals, 1969, 113; Smith, in Walker and West (Edd.) Studies in the Vegetational History of the British Isles, 1970, 81. 56 9 Cf. Evans, W.A.M. 61, 1966, 9; Evans and Fowler, Antig. 41, 1967, 289. Plough-marks under a Polish long barrow (Sarnowo 8) could be as early as c. 3500 bc; Bakker et al., Helinium 9, 1969, 224. 10 The situation was first demonstrated by Atkin- son, Antig. 31, 1957, 228. 1 For such Neolithic storage pits in general, Field et al., Proc. Prehist. Soc. 30, 1964, 352. 1 Bradley, Proc. Prehist. Soc. 36, 1970, 312. 3Cunnington, JW.A.M. 36, 1909-10, 590; Piggott, Proc. Prehist. Soc. 2, 1936, 229. ™4 Wainwright, Antig. 44, 1970, 56. 15 Piggott, Antig. Journ. 20, 1940, 465. The pot is Longworth no. 93, very closely comparable to his no. g2 from Barrow g of the Lambourn Seven Barrows group (Proc. Prehist. Soc. 27, 1961, 263). The ‘MBA’ sherds from a rectangular enclosure on Preshute Down (V.C.H. Gazeteer, 268, no. 182) might be a significant pointer. 16 Smith, Windmill Hill and Avebury, 17ff.; Palaeohist. 12, 1966, 469. 17 Ashbee, The Earthen Long Barrow in Britain, 1970, assembles evidence available up to 1967. 18 Smith and Evans, Antig. 42, 1968, 138. 19 Simpson, Antig. 42, 1968, 142; Ashbee and Simpson, ibid. 43, 1969, 43; Kinnes, ibid. 44, 1970. 318. 20 Atkinson, Antig. 39, 1965, 126. 21 Collins and Waterman, Millin Bay, 1955, with appendix by Morton and Scott, 59 ff. 22 Antiq. 43, 1969, 45- 23 Recently reviewed (with new Cryyq dates) by Bakker et al., Helinium 9, 1969, 3; 10, 1969, 209. 24 Sieveking et al., Nature 228, 1970, 251. 25 The problem is very similar to that arising with prehistoric jade axes: Campbell Smith, Proc. Prehist. Soc. 29, 1963, 133. 26 Smith, Windmill Hill and Avebury, 85-109; 236- 243. 27 Laidler and Young, W.A.M. 48, 1938, 150; Piggott, Neo. Cultures, 283; cf. Bradley, loc. cit. 344-358. : 28 Stone, Proc. Prehist. Soc. 4, 1938, 249. 29 Semenov, Prehistoric Technology, 1964. 3° Peacock, Archaeometry 10, 1967, 97; Proc. Prehist. Soc. 34, 1968, 414; Antiq. Journ. 49, 1969, 41 (Iron Age); Antig. 43, 1969, 145 (Neolithic). 31 Clarke, Beaker Pottery of Great Britain . . . 1970, I, 256. 32 Zimmerman and Huxtable, Antig. 44, 1970, 304. 33 The list of prehistoric plant remains from Wiltshire published in 1964 contains no identifica- tions from impressions on pottery since Helbaek (Grose and Sandell, W.A.M. 59, 1964, 58). 34 Some Wiltshire groups of sherds collected before 1952 probably need examining as well as later finds. Helbaek does not seem to have made an exhaustive examination of, for instance, the whole of the Woodhenge pottery, nor that from the Easton Down Beaker settlement. The latter would be particularly interesting in view of Arthur’s 30-odd identifications of seed impressions (from about 1,200 sherds) at Belle Tout (Proc. Prehist. Soc. 36, 1970, 363). 35 Clark, Antig. 40, 1966, 182; Renfrew, Annual Brit. School Athens 63, 1968, 277; Antig. 44, 1970, 199; Proc. Prehist. Soc. 36, 1970, 280. 36 Coles and Taylor, Antig. 45, 1971, 6-14. Mr. A. ApSimon tells me he had independently come to much the same conclusion. New work on Wessex Culture and late Beaker bronze daggers in terms of their metallurgy is now in progress by Mrs. B. Ottaway, Department of Archaeology, University of Edinburgh. 37 It is ironical that all the recent rescue excava- tions on Wessex barrows have so far hardly en- countered any very rich Wessex Culture burials except for that with a gold-covered shale cone on Ports Down, Hants (Proc. Hants F.C. G Arch. Soc. 24, 1967, 26). The Edmonsham (Dorset) barrow with cremation and Wessex Culture assembly of dagger, whetstone, tweezers and pin had a crema- tion-pyre, but no Crq date has been obtained from its charcoal (Proudfoot, Proc. Prehist. Soc. 29, 1963, 95): 38 Hawkes, Proc. Hants F.C. & Arch. Soc. 14, 1940, 136. 39 Avery and Close-Brooks, Proc. Prehist. Soc. 35, 1969, 345. There is now a Crq date for Itford Hill of 1000 be + 35 (GrN-6167, unpub.). 4° At this point on the Suess curve the discrepan- cies between ’bc’ and ‘B.c.’ dates become relatively slight. 41 Brewster, Excav. Staple Howe, 1963; Smith, Arch. 77, 1928, 179; Cotton and Frere, Records of Bucks. 18, 1968, 187. 57 ARCHAEOLOGICAL RESEARCH IN THE PEWSEY VALE by N. P. THOMPSON With contributions by F. K. ANNABLE, I. W. CORNWALL, R. A. HARCOURT, E. MACHIN and R. F. TYLECOTE INTRODUCTION SINCE THE AUTUMN of 1967 a number of discoveries of considerable archaeological interest have come to light within the Pewsey Vale. Thanks to the keen observation of Mr. Paul Bowerman, of White Hare Farm, Pewsey, who collected pottery and other artifacts from the surface in Black Patch Field, a mile south of the village, an Iron Age settlement and a pagan Saxon inhumation cemetery have been recognized and are now under active investigation (W.A.M., 65 (1970), 206). In 1968, following ploughing, Mr. Bowerman also brought to notice a previously unrecorded ditched enclosure which he had observed on the summit of Pewsey Hill, one and a half miles south of Pewsey. A limited excavation (described below) has since been carried out at this site, directed by Mr. N. P. Thompson under the aegis of the Archaeology Research Committee. In addition to these new discoveries, the combined evidence of earlier site records and subsequent finds within the area of the Pewsey Vale have made it apparent that throughout the Iron Age and the Romano-British and Saxon periods a widespread distribution of agricultural settlements existed in all probability along the Vale (Fic. 1), none of which has so far received much attention. Of possible Iron Age or Romano-British date is a second circular enclosure recently recognized near Milton Clump, one and a half miles to the north-east of Pewsey Hill. The existence of two substantial buildings of Romano-British date, previously suspected but not accurately recorded, has also been confirmed by means of geophysical survey and limited excavation carried out by members of the Archaeo- logy Research Committee with the help of Society volunteers. The buildings lie respectively in the parishes of Wilcot and Alton; see below for further details. At Wilton, in the parish of Grafton, near Great Bedwyn, the finding of a bronze ‘small long’ brooch of 6th century date was recently reported to the Museum (see Woltes, p. 178). Saxon burials are also known from Easton Royal and the Wood- bridge Inn, North Newnton. It is significant that, as with the Black Patch cemetery, all these find spots lie below the chalk escarpment close to well drained land admir- ably suited for agriculture. Future research may well pin-point the positions of new cemeteries and their implied settlements, whose investigation could shed a fuller light on the perplexing problems concerning the origins of early Saxon occupation in North Wiltshire and, indeed, within the Wessex region. 58 Oyuish \ O Draycot ; Farm O West Stowell Al Owitcot Al2 es |S nes ee All fa) Pewsey Hill One Mile Vee} Fic. 1 Location of prehistoric and Romano-British sites in the Pewsey Vale. 1. Stanchester: Romano-British building; 2. Draycot: Romano-British buildings; 3. Giant’s Grave: promontory fort; 4. Withy Copse: Romano-British midden; 5. Martinsell: hill-fort; 6. Broomsgrove: Romano-British kilns; 7. Milton Clump : Iron Age (?) enclosure; 8. Giant’s Grave: long barrow; 9. The Spectacles : Iron Age( ?) enclosures; 10. Pewsey Hill: Iron Age enclosure; 11. Denny Sutton Hipend: concentration of Romano-British material; 12. Black Patch: Iron Age settlement and Anglo-Saxon cemetery. 59 The outcome of this activity has been the initiation, almost of itself, of a pro- gramme of archaeological research within the Vale, directed towards accumulating evidence about the settlement pattern within this geographically sharply defined area, and the project is continuing under the sponsorship of the Archaeology Research Committee. A third season of excavation is planned for June, 1971, at the Black Patch site, where the burials, in particular, are threatened by deep ploughing. Elsewhere, it should be stressed, only very limited excavation is envisaged, and it is hoped that much of the fact-finding over the Vale will be by means of surface fieldwork and geophysical survey. If the consent of landowners is forthcoming, exploratory magnetometer surveys will be carried out at the recorded sites of Saxon burials to the east and west of the Black Patch cemetery, in the hope of locating new cemetery sites. A long-term programme of field walking is also planned to locate surface finds, predominantly pottery fragments and occupation debris; these, when plotted on six-inch maps, may give evidence of so far unrecorded settle- ments along the Vale. Additional resistivity work at the Stanchester Roman building is also necessary if its extent is to be determined, and the circular enclosure at Milton Clump is to be planned by resistivity survey and limited ditch sectioning. It would be equally worthwhile to survey the interesting pair of (?) Iron Age enclosures known as “The Spectacles’, a quarter of a mile to the east of Pewsey Down. Geologically, the territory covered by the project includes the edge of the chalk plateau and escarpment of the Marlborough and Pewsey Downs. A narrow clay strip lies at the foot of the hills enclosing the valley. Cereal crops are predominant on the downland chalk soil which, in the last century, was largely grass-covered, and on the steep escarpment pasture still remains. Most of the valley is arable, although from time to time there is some rotation to grass. The clay border is occasionally ploughed but more often than not it remains as permanent grassland. It should perhaps be noted that the predominance of cereal cultivation over the Pewsey Vale limits the carrying out of fieldwork to the late autumn and winter months. The Archaeology Research Committee has been fortunate so far in obtaining the willing and interested help of a number of younger Society members living within the Vale, or within reasonable travelling distance of the area; indeed, one important aim of the research project has been to promote interest, and to offer to members and helpers wishing to take an active part in the work of the Society an opportunity to obtain training in field archaeology techniques and excavation. Without the goodwill and co-operation of a number of farmers a great deal of the investigation now in progress could not have been started. The Research Com- mittee is greatly indebted to Mr. Paul Bowerman, not only for permission to excavate the Black Patch cemetery site, and the Pewsey Hill enclosure, but also for his considerable generosity in meeting the cost of a J.C.B. digger, when, on a number of occasions, 1t was used to facilitate topsoil clearance. The many archaeological finds he has made over his fields are now being catalogued by Mrs. E. Machin with his kind permission. Mr. Neil Swanton, of Draycot Farm, gave every facility to the Research Committee during the excavation of the Romano-British building on part of his farm, and the Committee is indebted to Sir Philip Dunn for his consent to the investigation of the Romano-British building in the Stanchester field. Such interested 60 co-operation on the part of farmers and landowners is both heartening and welcome; to others in the Vale who have additionally consented to field walking being carried out by Society volunteers, the Research Committee offers its thanks. The first of the Reports on the work of the Archaeology Research Committee within the Pewsey Vale follows below. Interim reports on the Iron Age and Saxon site in Black Patch Field will continue to appear in the annual summary of Excava- tion and Fieldwork in Wiltshire until the publication of the final report. F. K, ANNABLE AN IRON AGE ENCLOSURE ON PEWSEY HILL In 1968, following autumn ploughing, Mr. Paul Bowerman of White Hare Farm, Pewsey, informed the writer that he had observed a segment of an unrecorded circular ditch on the summit of Pewsey Hill. The site was later confirmed from aerial photographs taken by Mr. John Hampton, Royal Commission on Historical Monu- ments (England). With Mr. Bowerman’s permission, limited investigation of the site was under- taken for the following purposes: (1) By means of resistivity survey to trace the line of the enclosure ditch, and in particular the entrance area; (11) To section the ditch, including the ditch terminals, at several points to obtain dating evidence; (11) To record surface finds within the enclosure. TEE OLE Pewsey Hill, within Pewsey parish, lies one and a half miles south of Pewsey (FIG. 1) and forms part of the chalk escarpment along the southern edge of the Vale. The hill rises to a height of 650 ft. O.D., but to the west the ground falls sharply towards Denny Sutton Hipend, a chalk plateau forming an east-west promontory of Pewsey Down. The enclosure (SU 16755765) (FIG. 2) lies on the Upper Chalk and contains an area of about two acres; it is divided by a wire fence running approximately north to south. Some two-thirds of the interior, on the eastern side, is now arable and was first ploughed about ten years ago after an undetermined period as downland pasture; the remainder, partly on the escarpment, still survives under grass. A thin spread of chalk rubble over the arable field suggests the remains of a ploughed-out internal bank, and in low sunlight it is just possible to follow the line of the ditch in both ploughland and pasture. The existence of the ditch was confirmed at a number of points by resistivity survey, although continued search by this method failed to indicate the position of the enclosure entrance. Several Iron Age hill-forts and settlements are known in the vicinity. The hill- fort of Martinsell lies about three and a half miles to the north; Chisenbury Trendle is three miles to the south, and Casterley Camp four miles to the south-west. The two enclosures known as ‘Spectacles’ lie only a quarter of a mile to the east on the opposite side of a dry valley on Pewsey Down. Further enclosures le one and a half 61 1/4 MILE Reservoir {ip 0° Burne flints PEWSEY HILL ENCLOSURE “AAS Dew-pond? Fic. 2 Pewsey Hill enclosure: plan. 200 yj ee = OSI — = PH ali miles to the north-east near Milton Clump, and possibly on Wilsford Hill (Wilsford, North), some five miles to the south-west (Grinsell, 1957). Between the ‘Spectacles’ and the site under investigation ‘Celtic’ fields are clearly visible on an air photograph of 1946; one lynchet appears to come up to the enclosure ditch, but does not con- tinue inside. Farming establishments are known from the well-known site of All Cannings Cross six and a half miles north-west of Pewsey Hill, and, now, in Black Patch field, one mile to the west, where storage pits have been discovered. THE EXCAVATIONS, 1969 AND 1970 The Enclosure Ditch On the basis of resistivity surveys, four sections were cut across the ditch in order to verify its line and to obtain profiles and dating evidence. A mechanical excavator was additionally used to cut across the ditch at over a dozen points in the arable field, but as this was primarily to trace the ditch edges, the sections were only cut down to the top of the silting which invariably was indicated by a covering deposit of pseudo-mycelium. A uniformly V-shaped ditch was indicated by Cuttings I, III, IV, V, averaging 2°50 metres wide and 1-50 metres deep; the width of the ditch at base averaged 0-50 metres (FIG. 2; FIG. 3, Section 1). The ditch fill on the hillside (Cuttings I, IIT) consisted of a basal layer of chalk lumps overlaid by chalk rubble with some soil turning to a chalky hillwash; above was a chalky soil and the present turf line. Cutting II was laid out in the north-west segment of the ditch where it had been dug across the steepest gradient of the hillside. Here only 30 cms. of the ditch counterscarp had survived; the remainder had probably been washed downhill. The primary chalk rubble fill still remained up to the top of the counterscarp, covered by a chalky silt that perhaps represents hillwash. On the comparatively level ground of the arable field the ditch fill was consistent, but differed from that on the hill slope. The fill in Cutting IV and from the east and west ditch terminals (Cutting V) consisted of a primary layer of chalk lumps overlaid by a finer chalk rubble with some soil intermixed; above this an old turf line of dark soil extended across the ditch, and was capped in turn by chalky plough soil and topsoil (Fic. 3, Section 2). The mechanically dug sections all revealed the old turf line with identical top layers. In Cutting I a level berm extended from the inner edge of the ditch into the enclosure for 1-75 metres, terminating at the edge of a raised platform, 10 cms. higher than the level of the berm. The lowering of the chalk level across the berm may be explained as the result of weathering of the unprotected chalk; immediately beyond the step the enclosure bank would have served to protect the chalk beneath from erosion, thus preserving the original level. Unfortunately, the other cuttings were not extended sufficiently far into the enclosure to determine whether this feature was present elsewhere along the enclosure ditch. As noted above, attempts to trace the butt ends of the enclosure ditch by means of resistivity survey were unsuccessful; the terminals were finally located by trenching in the southern segment of the enclosure. Excavation showed the entrance to be slightly staggered, the eastern terminal projecting a little more to the north beyond the line of the western butt end. It will be seen from the plan of the entrance (Fic. 4) that approximately four metres west of the eastern edge of the latter the ditch makes a decided turn to the north- west before continuing its circuit of the hill summit. The Linear Ditch A linear ditch with U-shaped profile, go cms. wide and cut to a depth of 40 cms. into natural chalk, ran in a north-westerly direction from the north lip of the butt end of the eastern terminal for a distance of at least five metres into the enclosure interior. The ditch was sectioned in several places (FIG. 3, Section 4) and the fill found to consist of a homo- geneous mixture of chalk lumps and soil. It was difficult to be certain whether the ditch 63 Enclosure ditch: section 1 East Enclosure ditch: section Linear ditch: section [3] North South East West Linear ditch: section East ti Fic. 3 Pewsey Hill enclosure: ditch sections. Key to layers: (1) Modern plough-soil; (2) Earlier plough-soil: (3) Buried soil; (4) Chalk rubble and soil; (5) Chalk rubble and chalky silt; (6) Clean chalk rubble. had been deliberately filled or whether it had silted naturally. The full extent of the ditch within the enclosure was not traced. An irregularly cut ditch extending from the southern lip of the eastern terminal was sectioned at the point of junction with the enclosure ditch (Fic. 4; FIG. 3, Section 3). It is suggested that the higher ditch base on the western side of Section 3 represents a continua- tion of the U-shaped ditch to the north, as further indicated by its orientation and level. The eastern half of the ditch, with its gully similar to that of the enclosure ditch, may represent a re-cut of a date contemporary with the enclosure. The fill was similar to that of the U-shaped ditch. A further cutting (Fic. 3, Section 5), 7°5 metres south of the eastern terminal, revealed a single ditch similar in profile to the enclosure ditch, but on a smaller scale. The fill 64. “IJOURIJUD JO urd + IINSOPIUD [TET AdsMag JONVYLNI 4O NV1d FYNSOTINA TH AaSMAAd 65 included the turf layer over chalk lumps—chalk and soil silting; no pottery was found in the silting, but the turf layer contained Iron Age and Romano-British wares, as did the same layer in the enclosure ditch. There was, however, no evidence of re-cutting as in Section 3. Whether the larger and deeper later ditch had obliterated the U-shaped ditch or whether the latter did not continue much beyond the main enclosure, or had changed its direction, was not determined. It seems, therefore, a fair assumption that the enclosure ditch had cut through an earlier ditch of a pre-enclosure period, and that the occupants had re-cut the latter to serve as a field boundary. Thus we have some evidence of the continued use of existing field systems from pre-enclosure times up to the time of the Iron Age occupation. The 1946 aerial photographs, taken before the field was ploughed, also show a lynchet coming up to the enclosure ditch, but not into the enclosure. The lynchet is 90 metres east of the linear ditch (Fic. 2) and lies almost parallel to it. It was not sectioned. The Entrance The topsoil in an area across the entrance (FIGs. 2 and 4) was mechanically removed to the level of the weathered chalk, which was afterwards cleared manually down to the natural. Two holes, presumably gate-post sockets, were found to be set across the entrance gap, 3:2 metres apart from centre to centre. The eastern hole, oval in shape, appeared to have two post sockets, one 35 cms., the other 32 cms. deep, cut into solid chalk. The western hole was 24. cms. in depth. There was no evidence of sockets, but the diameter of the hole in relation to its depth would suggest that it might have been made for two posts rather than one. Neither hole showed any signs of re-cutting. With due allowance for the weathering and erosion of topsoil and chalk subsoil since the Iron Age, the addition of a further 30 cms. might give a truer indication of the original depth of the holes. The orientation of the sockets in the eastern hole, across the ditch line, may indicate that closure of the entrance was achieved by placing a hurdle between the two posts in each hole. This differs from the double post-holes at the entrance to Down Barn West enclosure (Fowler, Musty and Taylor, 1965, fig. 4), which are in line with the ditch and would suggest a normal gate. It will be noted from the plan (Fic. 4) that between the eastern post-hole and the edge of the ditch terminal there is a gap of 1-30 metres not covered by the entrance as represented by the two post-holes. There was, however, a noticeable rise in the chalk level (3 cms.) extending from the eastern post-hole towards the enclosure ditch. If this difference in levels results from the protection of the original chalk surface by a bank, the gap may indicate that the original enclosure bank extended right up to the gate-post. On the western side no similar rise in the chalk level was observed. Further irregular holes exposed in the subsoil to the north were considered to be natural solution holes. The Interior Immediately to the east of the fence line, within the north-east sector of the enclosure, was a circular banked depression about 15 metres in diameter. The bank remains were visible as a chalk spread over the arable, and subsequent excavation using the mechanical digger showed that on the north the bank overlay the enclosure ditch. The depression was probably a disused dew-pond of post-enclosure date. Calcined flints were found in quantity over some areas of the enclosure. Trend mapping was not possible as the field had recently been sown, but there was a noticeable concentration of fire-cracked flints towards the centre of the enclosure (FIG. 2). If these are taken as evidence of disturbed hearths they may indicate the approximate position of a hut site. The surface finds included a number of spherical flint hammer-stones, waste flint flakes, sherds of Iron Age pottery, and many chips of sarsen stones. A broken sarsen saddle quern, possibly re-used as a quern rubber, was also picked up. From the density of this 66 scatter of occupation material over some parts of the enclosure, it seems that ploughing has considerably disturbed the occupation level. Several stones foreign to the district were picked up. These were later identified as Liassic limestone, probably from East Somerset. This stone is frequently used in Wiltshire for road metalling. Mr. Bowerman confirmed that sludge had been spread on the field, and doubtless the stones had been conveyed to Pewsey Hill from Pewsey Sewerage Works. DISCUSSION Situated within one and a half miles of each other are four enclosures of similar size: Pewsey Hill, Milton Clump, and the two linked examples known as “The Spectacles’ (Crawford and Keiller, 1928, 224-6, pl. XLI). Geographically, it would not be unreasonable to add this group to Perry’s Hampshire group of enclosures of between one and three acres (Perry, 1969). All have some features in common with the eleven sites listed in the Gazetteer included in his paper. Milton Clump is roughly circular, with a linear ditch extending southwards from the enclosure and with a well-preserved ‘Celtic’ field system to the west. “The Spectacles’ are linked by a linear ditch, and to the north-west an uninterrupted ‘Celtic’ field system extends to the Pewsey Hill enclosure. As previously mentioned, a lynchet appears to come up to but not into the latter enclosure, and excavation revealed a linear ditch extending outwards from near the entrance. There is no visible evidence of pits in the four enclosures. Much of the interior of Pewsey Hill had been ploughed down to the chalk, and if pits were present it is probable that there would have been noticeable surface indications of their existence. The Iron Age pottery recovered from the lower ditch silting of the Pewsey Hill enclosure is presumably contemporary with occupation of the site. It seems to be a homogeneous group, and comments kindly made by Mrs. Sonia Chadwick Hawkes, based upon her analysis of comparable pottery from the Iron Age settlement at Cow Down, Longbridge Deverill (report forthcoming), suggest that it is characteristic of her earliest phase of Iron Age ‘B’, beginning not earlier than the 4th century B.c. By contrast, the sherds collected from the Mancombe Down enclosure were Early Iron Age ‘A’ (Fowler, Musty and Taylor, 1965, 55). Mancombe Down is of similar size and shape to Pewsey Hill and is also a hill-top enclosure. The presence of hammer-stones, sarsen flakes, calcined flints, sherds, animal bones (including those of dog), a fragmentary saddle quern and iron slag, indicate the use of the site for domestic and farming occupation. The density of occupation material west of centre of the enclosure may also indicate the position of a hut, but this can only be proved by excavation. The ditch on the plateau area was allowed to silt up naturally until stabilization had been reached: subsequently turf covered the remaining depression. Iron Age sherds prove occupation throughout this time. Some Romano-British sherds occurred in this turf line, and pottery in the plough-soil above was exclusively of this period. There was also a scatter of Romano-British wares over the surrounding field. A concentration of Romano-British material has been recorded at Denny Sutton Hipend, half a mile to the west, suggesting an occupation area. It is probable that the enclosure was ploughed over in the Romano-British period. 67 The suggestion has been made (Fowlcr, Musty and Taylor, 1965, 54) that these small enclosures, surrounding probably a single homestead, may have been depen- dent upon a larger and socially superior settlement, e.g. Mancombe Down on Baitles- bury hillfort. Within the Pewsey Vale we have All Cannings Cross, and now the Iron Age site within Black Patch Field (W.A.M., 65 (1970), 206), both possibly representing large permanent farming establishments situated at the edge of fertile valley land beneath the escarpment ridge. Did the cultivation of marginal land become necessary as a result of population increase, and do these downland enclosures represent satellite homesteads subsidiary to the main settlements, but established for convenience nearer to the working areas ? Little can usefully be said on this subject without further investigation and, in par- ticular, complete excavation of the interior of one of the enclosures. N. P. THOMPSON THE FINDS Pottery A total of ninety-five sherds was recovered from sections cut across the enclosure ditch. Some four or five examples were difficult to identify with certainty, but sixty-one sherds were considered to be Iron Age fabrics, and thirty-three Romano-British, with a single, small abraded fragment of a Samian cup, Drag. F. 33. The assemblage as a whole represented thirty-five vessels, nineteen of them being Romano-British. The Iron Age pottery was recovered almost exclusively from two stratified levels within the enclosure ditch: Layer 4, consisting of chalk rubble with an intermixture of soil, immediately overlying the primary filling of clean chalk lumps, and Layer 3—a chalky soil (? hillwash), but replaced in Cuttings IV and V, in the arable field, by an old turf line of dark soil extending across the ditch sections. The Romano-British wares, with one or two exceptions only, were recovered entirely from the upper layers. The Iron Age pottery is all hand-made, and of a coarse, sandy, dark-grey to brown or buff fabric, with one or two sherds a pale orange colour. A number contained minute inclusions of crushed flint or chalk; all were devoid of decoration, and no rims were recovered. Only one sherd appeared to show any attempt at external burnishing, and many fragments were abraded on the face and edges. There was also a single body sherd which could, perhaps, be considered as ‘haematite’ ware. The character of the Iron Age fabrics is so similar as to make it almost certain that all of them fall chronologically within a single phase. Comparison with wares from the sites of All Cannings Cross, Swallowcliffe Down, and Fyfield Bavant shows that the Pewsey Hill assemblage falls within this ceramic group. A number of fragments can be paralleled exactly at Swallowcliffe Down and Fyfield Bavant, and are also similar to some of the coarser types from All Cannings Cross. It is noticeable, however, that the pottery from the latter site is a more finely made and sophisticated product than that from the other two sites under comparison, and it is with these latter that Pewsey Hill should be more closely linked. The vessel types represented, incidentally, seem mainly to consist of situlate jars, and there is no single example of the smoothed wares and saucepan and pedestal forms of the later Iron Age phases. Mrs. Sonia Hawkes has kindly examined the Pewsey Hill assemblage, and has commented that the group shows further similarities to wares from Pit 21, Cow Down, Longbridge Deverill (report forthcoming), and assigned her classification to an Initial ‘B’ phase, not earlier than the 4th but not later than the 3rd centuries B.c. The Romano-British pottery is generally a harder fabric, but also sandy in texture, with a colour range of pale brown to dark grey; one or two sherds are fired a pale orange. The main vessel forms represented are jars with everted rims, and three rim fragments of the type occurred within the group. A single rim of a bead rim bow], and a larger example were also recovered, the fabric of the latter being reminiscent of the Savernake Forest kiln products, though not certainly manufactured at the kilns. One or two thicker sherds, representing large storage jars, were also present in the group. It seems pointless, on the evidence of this small group alone, to make any comment beyond the suggestion that the bead rims, and a single late type of flaring rim, indicate occupation somewhere in the neighbourhood of the enclosure between the late 1st and late 3rd centuries A.D. None of the pottery from the excavation was considered sufficiently diagnostic to be worth illustrating. F. K. ANNABLE Quern Fragment Part of a sarsen saddle quern (FIG. 5: 2) was found on the surface within the enclosure. It measures 17-8 cms. long, with a maximum width of 10 cms.; the height is 5-3 cms. The grinding surface is horizontal and the underside well smoothed. An Unfinished Chalk Loom Weight? This object (Fic. 5:1), found in the buried turf layer of the western ditch terminal, was submitted for analysis to Dr. I. W. Cornwall, together with specimens of Upper Chalk from the site and of Lower Chalk from an outcrop near Etchilhampton, in Pewsey Vale. At a later date it was found that beneath the topsoil at the foot of the escarpment of Pewsey Fic. 5 Pewsey Hill enclosure: (1) ?Unfinished chalk loomweight; (2) Saddle quern fragment. Scale: 4. 69 Down the Lower Chalk occurs in a plastic condition and can easily be hand moulded. It is probable that the object was made from this material. Dr. Cornwall writes as follows: “This specimen is a thick discoid of subrectangular section with a central tapering hole in each of the opposing, more-or-less plane, faces. There did not appear to be a clear passage through these holes, and, after the object was sectioned, it could be seen that the join had never been completed. The holes are not manifestly pierced by a rotating drill and cond have been formed by simple impression with a slightly tapered cylindrical tool with a rounded point, if the material were plastic when so treated. ‘In order to ascertain the nature of the raw material which made up the body of the object, a thin-section was prepared and compared with thin sections of Upper and Lower Chalk. The make-up of the object most closely resembled that of Lower Chalk but differed from it in the following features: ‘(1) Numerous small circular voids were present—similar features in the solid Chalk not being empty but filled with crystalline secondary calcite. The voids could be explained as air bubbles in an originally plastic mass. ‘(2) Orientated strings of grains could be seen, suggesting flow-structures in a not entirely homogeneous plastic mass subjected to moulding or modelling forces. ‘(3) More quartz-grains in the fine sand grade (0-2—0-06 mm.) than in the Chalk. ‘(4) The scattered glauconite-grains were all brown and somewhat weathered to limonite, unlike the fresh bright green glauconite in Lower Chalk. ‘(5) A general yellowish-brown colour pervaded the object due to finely divided limonite widely dispersed through the body of the material. ‘Conclusion: The object is an undoubted artifact—an unfinished loomweight ?—made by hand-moulding, and partly piercing, a plastic mass of somewhat weathered Lower Chalk mud, perhaps with some added fine quartz sand.’ The Animal Bones The collection recovered from the ditch sections, either in the buried turf layer or the silt below, was a very small one containing only seventy-five identifiable specimens. The species present, with the minimum number of each shown in brackets, were: domestic cattle (4), sheep (2), pig (2), horse (1) and dog (2). All were typical of the time and region, the farm animals small, the dogs fairly large and closely resembling those found at Cow Down, Longbridge Deverill, less than twenty-five miles away (Harcourt, R. A., in Hawkes, Sonia, The Iron Age Settlement at Cow Down, Longbridge Deverill, report forthcoming). The only complete bone, a bovine metacarpal, had a total length of 174 mm.; proximal articular surface, 53 mm.; midshaft diameter 31 mm.; and distal articular surface 57 mm. This was probably from a steer and indicates an animal of about 42 in. (106 cms.) shoulder height. Dog was represented by two nearly identical portions of right maxilla in which the carnassial (PM 4) was respectively 17-5 and 17-9 mm. in length, similar to those of a dog of 21-23 in. (53-58 cms.) in height. The hunting of wild animals was, on bone evidence, little practised in the Iron Age, so these animals may have been watchdogs or perhaps used for guarding cattle and sheep. The remaining species, sheep, pig and horse, provided only about a dozen specimens in all, none of them of particular note. R. A. HARCOURT Flints Four cores, one core-rejuvenator, 64 struck flakes, two used flakes, one flake with secondary working, 4 pounders and 13 fragments were found. None is ascribable to any particular period, except the pounders. These are of the type associated with Early pre-Roman Iron Age settlement sites. One is completely spherical, with entirely pocked surface and diameter of 6 cms. The others appear to be approaching that size and condition. All were found in the buried turf layer. There were also many calcined flints. EVE MACHIN a] 70 Tron Slag? Five small lumps of possible iron slag were found in the layer beneath the plough-soil in Cutting III, associated with Iron Age pottery. The specimens were sent to Dr. R. F. Tylecote, Department of Metallurgy, University of Newcastle, who has kindly commented as follows: ‘The samples are smithing or smelting slag from either iron or non-ferrous working. It is mildly magnetic, which tends to put it outside the range of non-metallurgical vitrified products, such as ash from vegetable sources, e.g. haystack, dung, wood, etc. Although a complete examination was not made, I am inclined to say that the slag is the product of iron smelting or smithing and, if so, it is of a primitive type and most likely pre-Roman as far as this country is concerned.’ TWO ROMANO-BRITISH BUILDINGS IN THE PEWSEY VALE DRAYCOT FARM, WILCOT (SU 14606320) That a Romano-British building existed somewhere on Draycot Farm is evident from a letter written in 1892 by G. E. Dartnell, son of the Rector of Huish (Goddard, 1928). He mentions that a building in a field near Huish had never been recorded. Mr. Newman, at that time owner of Draycot Farm, had, between 1882 and 1884, observed crop-marks in one of his fields. Out of curiosity he had dug there and found ‘four corner foundations of a room or building, with some pillars here and there in the middle. The walls ran out in other directions and the field had Roman tesserae lying about.’ It is of further interest that Mr. Dartnell also mentions a tradition in Huish that ‘a great city’ once stood in that field. Miss Dora Ibberson confirms the story: ‘Huish was once a big place with buildings in Great Mead’ (Ibberson, 1963). Great Mead is a Swampy pasture south of Huish and adjoins the field belonging to Draycot Farm to the south of the Huish-Wilcot parish boundary. The improbability that Great Mead was ever a building site was shown by the stratigraphy of a cutting dug in 1968. At the foot of the bank on the south side of the field probing had shown a layer of black material beneath topsoil. The small sample observed appeared to be charcoal, and, with the possibility of kiln sites in mind, a trial cutting was laid out across the bank. With the exception of the topsoil from which were recovered two paste imitation amethysts of Georgian date, the layers were undisturbed. Beneath the topsoil was a layer, 10 cms. thick, of black, decayed organic material lying on the clay subsoil. Great Mead is basin-shaped, and the springs emerging from the junction of the Chalk and Greensand just above Huish, would drain into the field. In the past, with less drainage, the field may have formed a large pond and this could explain the presence of the layer of decayed organic material. The tradition of occupation within the field may, however, have arisen out of the situation and general contours of the latter, for any buildings in the higher ground to the south, when seen from Huish, would appear to be rising from Great Mead. The Nap (Fic. 6), part of Draycot Farm, is a knoll of Greensand rising from the Ful te Great Mead HUISH Ph ‘Wall foundation located by probing 052 ye 9000 O00 excavated Rubble Fic. 6 Romano-British buildings on Draycot Farm, Wilcot. Limit of rubble =}! Huish-Wilcot boundary. Although it is little more than 8 metres above the surround- ing landscape it is prominent from all directions except the east. It is well drained and attractively situated and this site would have been ideal for a Romano-British farmstead. Although the late G. M. Young told Mr. Swanton, the owner of Draycot Farm, that he thought a villa would be found near his farm house (and this may explain Grinsell’s grid reference of SU 143628), it was decided, should a suitable opportunity arise, to investigate the Nap. The field had been under grass for at least a decade, but when in October 1968 it again came under cultivation, Mr. Neil Swanton kindly informed the writer and gave his consent to trial soundings in the area. Site I After ploughing, the soil over Site I showed a slight colour change with a few large flint nodules and some chalk lumps scattered across the field surface. Preliminary probing suggested the likelihood of foundations some 25 cms. below the surface, and an exploratory trench, 2 metres square (Cutting I, ric. 6) exposed at a depth of 25 cms. the corner of a building with well laid chalk rubble and flint foundation walls 74 cms. wide, extended by a flint footing 13 cms. wide. Two circular slots, each 25 cms. in diameter and 10 cms. deep, were set into the centre of one of the foundation walls, probably to support timber uprights. Within the limited area of the interior that was excavated a cobbled flint floor and gritty soil were found to lie directly above natural Greensand. A few nondescript sherds of Romano-British date and tile fragments of Pennant sandstone occurred inside the wall. Following the confirmation of the location of the site, an area 50 metres by 25 metres was surveyed using an electrical resistivity meter in the hope of tracing, if not individual foundation walls, at least the overall area occupied by the building. Although considerable sub-surface anomalies were recorded by the meter, the general pattern was unconvincing, and further extensions of the building were traced by probing, and afterwards corroborated by excavation. Small cuttings (II and III) revealed evidence of a flint footing, disturbed walling, and a large Pennant sandstone tile. The fourth cutting (IV) brought to light a wall of consolidated chalk rubble above flint footing similar in width and construction to that revealed in Cutting I. Site I In the autumn of 1969 Mr. Swanton informed the writer that his ploughmen had struck further wall foundations on the crest of the knoll. A resistivity survey was again carried out but with no better results than those obtained over Site I. In 1970, a trench 5 metres square was fully excavated, and the wall foundations of a single room were exposed. The walls were well constructed of chalk rubble, consolidated by puddled chalk. All had been carefully levelled on the top, but no post-sockets were visible. The west and south walls were 1-10 metres, and 0:75 metres wide respectively. (Only a portion of the north wall was excavated.) All except the east wall were 0-50 metres in depth. The east wall only remained to a depth of 6 cms., and it may possibly represent a threshold rather than a wall foundation proper. The room interior was without recognizable flooring and almost free of occupation material apart from a few iron nails, tile fragments and a small group of Romano-British coarse wares which included a flanged rim of a New Forest painted ware bowl. Ploughing disturbance had, however, reached to the level of the top of the foundation walls, and it is likely that any existing flooring or occupation debris had been dispersed in antiquity. The suggested extended ground plan (ric. 6) therefore was recovered by probing, and must be considered as tentative until excavation is carried out at the site at some future date. 13 STANCHESTER, ALTON (SU 13806185) Stanchester is the name of a field at West Stowell, in Alton parish, approxi- mately one mile south-west of the Romano-British building at Draycot. In 1930 the late G. M. Young recorded the discovery at Stanchester of two coins of Valentinian, pottery of Romano-British date, including body fragments of Rhenish ware, box flue-tiles, stone roofing-tiles and oyster shells (Young, 1931, 504-5). The coins and pottery are now deposited in Devizes Museum through the kindness of Lady Phipps. All this material was recovered from a single trench, but its precise location within the field was not recorded. In 1969 a resistivity survey was carried out (again with little real success) to locate the site of the building. In one area, however, probing suggested the possibility of buried foundation material and with the kind permission of Sir Philip Dunn a small exploratory trench was dug at a point 64 metres east of Firtrees plantation. Within the trench a sub- stantial chalk rubble wall of undetermined width was exposed, into which had been incor- porated a pillar of square tiles. The trench also contained a mass of fallen material, including decorated flue-tile fragments. Part of a bronze bracelet was also recovered. CONCLUSION The small scale investigations described above were carried out initially to locate exactly the sites of the two buildings and by means of geophysical surveys to trace their extent. Due seemingly to adverse soil conditions over the sites, the latter has not been possible, but further resistivity surveys will be attempted in differing weather conditions in the hope of obtaining more informative results. Failing this, large scale excavation is necessary, and this is not envisaged as part of the overall research project. Nevertheless, it is of considerable interest to have identified two substantial buildings of certain Romano-British date only a mile apart. In this potentially favourable agricultural area can we expect that other similar establish- ments exist across the Vale? If we are correct in assuming the survival into the Romano-British period of rural settlements of whose existence we know within the Initial Phase of the pre-Roman Iron Age, what is the relationship between the former sophisticated farming establishments, for this is surely what the buildings must represent, and the native farming villages? At least the discoveries made so far indicate the possibility of a widespread pattern of agricultural settlement, and con- tinued fieldwork and investigation must surely in time extend the limited picture existing at present. N. P. THOMPSON Acknowledgements I should like to thank Mr. F. K. Annable for his helpful criticism of the typescript and Dr. I. F. Smith for her practical help in improving the plans. Mr. F. K. Annable, R. Harcourt, Mrs. Eve Machin and Dr. R. F. Tylecote kindly provided the specialist reports. The objects illustrated were drawn by Mrs. Brian Jones. Special thanks are due to all the Society members who helped in the excavations and other work undertaken in Pewsey Vale. 74 BIBLIOGRAPHY Crawford, O. G. S., and Keiller, A. 1928. Wessex from the Air. Fowler, P. J., Musty, J. W. G., and Taylor, C. C. 1965. Some Earthwork Enclosures in Wiltshire. W.A.M., 60, 52-74- Goddard, E. H. 1928. Roman Building on Draycott Farm, near Huish. W.A.M., 44, 270. Grinsell, L. V. 1957. Victoria County History, Wiltshire, 1, Pt. 1, 122. Ibberson, D. 1963. Two Wiltshire Villages (Marlborough College Press). Perry, B. T. 1969. Iron Age Enclosures and Settlement on the Hampshire Chalklands. Archaeol. 7., 126, 36. Young, G. M. 1931. A Romano-British Site in Alton Priors. W.A.M., 45, 504-5. 19 3a THE EXCAVATION OF PREHISTORIC AND ROMANO-BRITISH SETTLEMENTS NEAR DURRINGTON WALLS, WILTSHIRE, 1970 by G. J. WAINWRIGHT With Pp. DONALDSON, I. H. LONGWORTH and v. SWAN SUMMARY Areas to the south and east of the Married Officers’ Quarters on Larkhill near Amesbury were investigated in advance of a tree-planting scheme. Evidence was recorded of Neolithic, Bronze Age and Iron Age settlement, but the bulk of the structural evidence, consisting of pits, post-holes, ditches and hollows, can be attributed to occupation in the 3rd and 4th centuries A.D. INTRODUCTION IN 1970 THE THEN Ministry of Public Building and Works was notified by the Ministry of Defence of the intention to plant a belt of trees to the south and east of the Married Officers’ Quarters at Larkhill near Durrington Walls. The plantations were proposed on either side of the Fargo Road which leaves the A345 near Wood- henge and runs west to skirt the southern boundary of the buildings (FIG. 1, PL. Ia, SU 147436). The plans involved the planting of two strips to the north and south of that road, 200 m. and 160 m. long respectively and 40 m. wide, in addition to a smaller area to the north of the quarters. An examination of the location plan (Fic. 1) will indicate that this project involved potentially important archaeological areas. Not only was there a record of Romano-British pottery from the field to the south of the Fargo Road, but the strips occurred immediately south-west of the Late Neolithic Durrington Walls enclosure excavated by the author between 1966 and 1968 (Wainwright and Longworth, 1971), and some 320 m. west of the enclosure surrounding the Woodhenge structure excavated by Mrs. Cunnington (1929). In addition, the fields surrounding these enclosures had produced structural evidence of settlement of which no trace had been visible on the ground. These comprise flint-mine pits east of the Stonehenge Inn and north of the Durrington Walls enclosure (Booth and Stone, 1952) and pits which produced Grooved Ware from the garden of a house south of Woodhenge (Stone and Young, 1948; Stone, 1949), which are presumably of Neolithic date. Of the Bronze Age are the linear barrow cemetery south of Woodhenge (Cunnington, 1929, 42-49), ploughed-out barrows south of the Durrington Walls enclosure (Hoare, 1812, 170; Stone, Piggott and Booth, 1954, 164) and a flat grave ‘above Durrington Walls’ (Cunnington, 1954, 234). Of the later Bronze Age is the ‘egg- shaped enclosure’ south of Woodhenge (Cunnington, 1929, 49 f.), whilst the Packway Enclosure west of the Stonehenge Inn is evidence of Iron Age settlement (Wain- 76 DURRINGFON WALLS 1970 x] ZF WMA Ae aiens EJ ecartnwork 77 wright and Longworth, 1971, chapter VII). As a result it was decided to investigate the areas which were to be planted with trees even though no indication of structures existed either on the ground or on aerial photographs. The structural remains thus recorded (FIG. 2) justified this course of action and reinforce doubts regarding the completeness of an archaeological record on chalk downland based solely on the evidence of visible earthworks. The proposed tree-planting was regarded as sufficient threat to any hidden structures on account of the disturbance caused by the planting operations and subsequent penetration by roots. The Ministry of Defence gave their permission for the excavation, which was under the immediate supervision of Mr. P. Donaldson. The author has pleasure in expressing his gratitude to Mrs. V. Swan for her report on the Romano-British pottery, to Dr. I. H. Longworth for his report on the Neolithic pottery, to his colleagues Miss S. Butcher and Mr. P. E. Curnow for their reports on the bronze leaf and coins respectively, to Mrs. B. Westley for her report on the animal bones, and to Mr. P. Sandiford for his report on the human skeletal material. Mr. Sandiford was also responsible for the photography during the excavation. The plans, sections, Grooved Ware and antler pick were drawn by Mrs. C. Boddington, the small finds by Mr. D. Neal and the Romano-British pottery by Messrs. F. Gardiner and J. Thorne of the Illustrators Section, Department of the Environment. Mr. J. P. Y. Clarke provided assistance during the preparation of this report. EXCAVATION A Drott Broo was used to clear the turf and plough-soil down to the surface of the underlying chalk. After the excavation had been in progress for a few days it was estimated that only one-third of the threatened area could be stripped with the available time and finances and therefore the area north of the Fargo road was trial-trenched in strips 25 m. long and 2 m. wide (this being suited to the width of the Drott’s bucket). The method proved to be satisfactory for the area in question (FIG. 2), the intervening baulks being removed wherever this was necessary. As the excavation progressed, the threatened area north of the quarters was abandoned on account of its use during the 1914—18 war for military installations that had removed the surface of the chalk. As a result, Site IT south of the Fargo Road was almost totally excavated save at its western extremity which was also disturbed by military installations. As the post-holes, pits, ditches and hollows were most closely grouped in the area of Site II, a policy of total excavation was in any case desirable. Slight structural evidence was recorded for Neolithic and Iron Age settlement and a Bronze Age presence was indicated by a socketed bronze knife. However, the bulk of the structural evidence, consisting of pits, post-holes, ditches and hollows, can be attributed to the Romano-British period. The whole area had been heavily ploughed for a considerable period and the excavated features were very eroded. THE NEOLITHIC SETTLEMENT The evidence for Neolithic settlement comprised four pits (7, 27, 47, 48) and a shallow ditch (29), which produced Late Neolithic pottery of Grooved Ware type, 78 IDUIRIRIING TOIN WWAULILS I@70) GENERAL PLAN OF EXCAVATIONS \ \ INN . N x [pitch ~ \ \ \S Es hupaetnes Ss => SS aN DS Se ae Hollow io : OD \ Nw (0) 50 100 036 9 1215 30 ee’ SCale in Feet ee nl Metres SITE Il SITE | DURRINGTON WALLS © 1970 SITE 1 : SOUTH 4 | SeWer pipe trench | a 4 ch? - és \ 5 fs e747 3 2 Qi8 — nl 0 5 10 35 Of ABA GS 10 Post-hole Pit ec ee eel Scale in Feet i eed Metres Fio, 5, [face pare 79 DURRINGTON WALLS 1970 SITEI : NORTH pypEencel™ 7. a : ; : ] * 7 ‘ Uy \ yy i / \2\ b -_ eee .o. | ADj i \O\ oe \\ | \o\ poe \ | 38 \O\ ce ae WE Fa ay ee ee heat \a\ ies —_— . > f 1s a SS ee 1 oo / \A\ ae = t Vay = i] a eee N ae \ —— ‘ t — \ \ on if / _ raed aa H / sie ie \ Nae d —— oor \ \ . / a ae . \ eacen a aN _ _— Za Je) ae 3 ‘ ss eee VeX > >> =--— A | \ Se epee i 1 a 1 1 8 i 1 \ ‘ ‘ boa — Y 1 AC; NEN J 38 aM | 4 aes TN . NY j | \ \ 47 * | Iz 38 HN | o ‘ 27 | ! SO | LY Nise Y | ===leme | (Om eee | (ken O° 5 10 15.20 Oniee a) 4e 5 | Pit Peed Scale in Feet ee Metres Fic. 3 together with artifacts of flint, antler and bone, animal bones and a marine shell. Three pits and the ditch were recorded in close proximity at the north end of Site I (27, 29, 47, 48; FIG. 3), whilst Pit 7 was located 98 m. south of this group (FIG. 5). PIT 7: a circular pit 1-25 m. in diameter and 26 cms. deep, filled with burnt flints and brown loam (FIG. 10). Pottery: nine sherds of Grooved Ware. Flints: five flakes. PIT 27: an irregular shallow pit 1-48 m. in diameter and 30 cms. deep with a fill of brown loam mixed with small chalk lumps (Fic. 10). Pottery : forty-seven sherds of Grooved Ware. | Flints: eleven flakes and one scraper. | Antler: three picks. DITCH 29: a narrow ditch which runs out of the excavation to the north and to the south rk) P ‘o1y solo;w = 994 Ul 9[00S — pl 49110 St) 3|O4-\SOd Le] OL SE Ol Ss O TALIS OZ6L STTVM NOLONIMUIAC (FIG. 3). It was traced for a distance of 15 m. and averaged 26 cms. wide and 20 cms. deep (FIG. 10), but because of erosion it was not possible to establish whether it was the base of a boundary ditch or whether it had originally supported a fence of posts. Pottery: ten sherds of Grooved Ware together with one sherd which may be from a Beaker. Flints : six flakes. PIT 47: a shallow pit 1-42 m. in diameter and 14 cms. deep with a fill of brown loam and chalk granules (Fic. 11). Pottery : seventeen sherds of Grooved Ware. Flints : twenty-one flakes. Shell: one limpet shell. pir 48: a well-defined pit 90 cms. in diameter and 60 cms. deep with vertical walls. The fill at the base was an ashy soil in which the finds were made and which merges upwards into a brown loam (FIG. 11). Pottery: seven sherds of Grooved Ware. Flints: thirty-seven flakes and one scraper. Antler: one antler pick. Bone: one awl and a fragment of a second; one polished rib fragment. The pits can probably be regarded as being for the disposal of rubbish, particu- larly in the case of Pit 7 with its fill of pot-boilers. Since sherds apparently from the same vessel were recovered from Pits 27 and 47, these pits at least are likely to have been contemporary. Similar pits producing Grooved Ware, antler picks, bone and flint artifacts, animal bones and marine shells have been recorded 600 m. to the south-east in the garden of a house called Woodlands (Stone and Young, 1948; Stone, 1949). Such pits are normally very productive in finds and the examples under discussion are no exception. The ditch (29) appears to define the limit of a small group of post-holes (39-40, 51, 53, 543 FIG. 3) but their relative age cannot be established on account of the lack of dating evidence from the latter. A total of 316 fragments of animal bones were recorded from the pits (Appendix I). These were very fragmentary and the total of seventeen individuals is a minimum figure only. Amongst the species represented, pigs are the most numerous (202 fragments, 8 individuals), followed by cattle (98 fragments, 4 individuals), sheep (5 fragments, 2 individuals), dog (2 fragments, 1 individual) and vole (1 individual). The presence of roe deer is indicated by an antler and pedicle joined to a skull fragment. The preponderance of pig is a normal feature in Late Neolithic contexts in southern England. They were of a small domestic variety with evidence of one boar, possibly wild. In his report on the pottery Dr. Longworth has indicated that of the 91 sherds of Late Neolithic pottery recorded from the ditch and pits, all but one are Grooved Ware of the Durrington Walls sub-style. The exception is P24, a sherd from the ditch that may be Beaker. Furthermore, the motifs on sherds P18 and P1g which include comb impressions may indicate some absorption of Beaker influences. At the time of writing one radiocarbon determination of 1523-+-72 B.c. (BM-703) was available for animal bone recorded from Pit 27* and it is of interest to compare this date with the series obtained from the large enclosure 20 m. to the east. The basal silt of the ditch of this enclosure produced radiocarbon dates of 1977-90 B.C., * A second determination of 1647 + 76 B.c. (BM-702) on an antler pick from the same pit is now available. 8I 2015-+90 B.c. and 2050-++90 B.c. (BM-398—-400) and contained only Grooved Ware of the Durrington Walls sub-style with no Beaker pottery. A date more compar- able with that from Pit 27 was obtained from the slower silts of the main enclosure ditch where charcoal from a hearth produced a determination of 1610-+120 B.c. (BM-285). At this level occurred sherds of rusticated, stroke and ridge decorated and biconical beakers, whilst the deposition of Grooved Ware had all but ceased. THE BRONZE AGE SETTLEMENT No structural remains of a Bronze Age settlement were recorded in the 1970 excavations and the evidence consists of a socketed bronze knife which was recorded from post-hole 162 (FIG. 19). ‘There seems little doubt that the find is a stray or had been acquired in Romano-British times as a curio, for pottery of that date was recorded from the post-hole together with point of a bronze pin. The socketed knife is a commonly recurring form for which Evans illustrates a good parallel from the Thorndon hoard in Suffolk, associated with socketed axes, a spearhead, a hammer, a gouge and an awl (Evans, 1881, 205-6, figs. 240-1). Similar knives occur in the Reach Fen hoard, Cambs." and the Feltwell Fen hoard in Norfolk’ with objects proper to the ‘Carp’s Tongue’ complex, which was intrusive into south-east Britain during a late phase of the British Late Bronze Age, probably during the 7th century s.c. Socketed knives of “Thorndon’ type have straight blades which are rarely sinuous and which normally expand slightly towards the hilt. The blade and handle always form a straight line at their junction (ude Hodges, 1956, 38, fig. 4). This type of knife has been found associated in many hoards in the British Isles and has a wide distribution with a concentration of find-spots in south- east England where it is a characteristic artifact of the Carp’s Tongue complex (vide Burgess, 1968, 39). An enclosure of Bronze Age date was excavated by Mrs. Cunnington 480 m. south-east of the find-spot of the knife (Cunnington, 1929, 49-51, pl. 45). The ‘egg- shaped’ enclosure surrounded a group of pits, one of which contained about 100 grains of charred barley, and sherds of a barrel urn were recorded from the surround- ing ditch. Between Woodhenge and this enclosure a pit containing a large barrel urn was recorded in 1951, and 100 m. west of this pit a pipe trench cut through a possible barrow circle which has not been investigated, although the base of a barrel urn and the rim of a globular urn were recorded from the base of the ditch (Stone, Piggott and Booth, 1954, 164-166). Pottery of the types associated with these structures has been found elsewhere, together with bronzes of the so-called ‘ornament- horizon’ which Dr. Smith has related to the Montelius III period of the Nordic Bronze Age, with a commencing date of around 1250-1200 B.c. (Smith, 1959, 144- 187). It seems unlikely therefore that the bronze knife is to be assigned to the settle- ment represented by these structures. THE IRON AGE SETTLEMENT Structural evidence for an Iron Age settlement was limited to two pits (21 and 23) in Site I (ric. 4). An adjacent pit (49) may be of Iron Age date but this is uncertain. ’ 82 PIT 21: a shallow pit 1-80 m. in diameter and 16 cms. deep (FIG. 10). Pottery: fourteen body sherds of hard, flint-gritted fabric. Flints: nine flakes; one unfinished leaf arrowhead or small knife and one fragmentary blade with flat edge retouch. Both implements are patinated white and are probably strays from an earlier occupation. Daub: two fragments. PIT 23: a circular pit 66 cms. in diameter and 48 cms. deep, the upper fill of which contained a number of sherds from Iron Age vessels (FIG. 10). Pottery: rim, base and body sherds of hard, flint-gritted fabric from two large pots with rounded shoulders and everted rims, the larger of which is decorated on the shoulder with oblique incisions. Flints : seven flakes and one hammerstone. Chalk: a probable weight with notched edges to facilitate suspension. A small group of animal bones comprising 24 fragments was obtained from the two pits (Appendix I). They include fragments of sheep, cattle, pig and horse. The evidence from pits 21 and 23 is not sufficiently extensive to enable any conclusions to be drawn regarding date and cultural affiliations. The two pots from pit 23 are of Early Iron Age type and are probably to be compared with the small cluster of Early Iron Age pits recorded in 1951 within the Durrington Walls enclosure which produced pottery of Little Woodbury type, three unfinished miniature cups of chalk and a chalk block with concave scraped surfaces (Stone, Piggott and Booth, 1954, 164). THE ROMANO-BRITISH SETTLEMENT The great majority of the excavated features can be assigned to a Romano- British settlement of late 3rd to 4th century date. Although a large number of post-holes, pits, gullies and hollows was recorded there was not one example of a coherent house structure and it may be concluded that the nucleus of the settlement was on higher ground to the west and is now covered by the buildings of the Larkhill Married Officers’ Quarters. A concentration of features was recorded in the centre of Site I (FIGs. 2, 5) but the majority of the structures occurred in Site IT (Fics. 2, 8). On account of the absence of well-defined structures the features have been described under site headings. Site I: North (rcs. 2-4) Features of Romano-British date were sparse in this area and were bounded to the north by a broad shallow ditch (38) 1-80 m. wide and 20 cms. deep (FIG. 10), which crossed the excavation from south-east to north-west (FIG. 3). Only one pit (34) occurred to the north of this feature which may have formed a boundary on this flank. The area south of this ditch is barren of settlement (FIG. 4.) with the exception of a scatter of post-holes (24-26; 30-33), a pit (35), and two short lengths of a gully (28, 52). Furthermore, these features did not produce any dating evidence and their attribution to the Romano-British settlement is inferential. Site I: South (FIGs. 2, 5-7) Two enclosures were recorded at the south end of Site I, one of which was associated with a corn-drying kiln. 83 DURRINGTON. WALES 1970 PLAN OF CORN DRYING KILN (SITE I 19/20) Section lines Chalk block Qa BBY S fe) 1 2 ee Scale in Feet — ee Metres Fic. 6 The Corn-Drying Kiln 19/20 (FIGs. 5-7; PLS. Id and IIa). An oval stoke-hole 1-20 m. deep, led into a T-shaped flue with an overall length of 5-30 m. This system heated a rectangular pit 2°80 m. by 2:60 m. with a clay floor and walls of chalk blocks set in clay. A dark ashy layer (layer 6) was preserved at the bottom of the fire-pit which extended along the floor of the flue (layer 18) and into its T-terminal where it thickened at the end of each arm of the T. The flue was lined with chalk blocks set in yellow clay and part of the semi- circular vaulting had been preserved. The pit over the flue housed a rectangular chamber 2°20 m. by 2:00 m., the sides of which were plastered with yellow clay lined with chalk blocks set in the clay matrix. There was some evidence that the end of the chamber furthest from the fire-pit had been lined with slabs of sandstone. The walls of the chamber had been subjected to heating and were reddened and cracked. The floor of the chamber overlying the flue had collapsed into that cavity, where the quantity of ash and nails (49 in all) suggested that a wooden framework may have been utilized. However, the nails could have come from a collapsed roof structure and the evidence near the base of the chamber walls is that the floor was composed of hard baked clay on a base of chalk lumps. Some ash on the floor of the chamber spilled down into the flue-pit when the former collapsed (layer 24). 84. Apart from the nails only a few finds were recorded from this structure and include a bronze bracelet fragment and spoon from layers (1) and (2) respectively. The evidence of the coarse pottery is that it was filled with rubbish in the 4th century a.p. The kiln is sited asymmetrically within an irregular and intermittent rectangular enclosure defined by a shallow ditch 90 cms. wide and 10-15 cms. deep (FIG. 5, 6, 22, 44), surrounding an area with maximum dimensions of 22 m. from west to east and 13 m. from north to south. Gaps in the ditch occur at the north-east and south-east corners of the enclosure but the presumed bank may have been continuous. The relationship of this enclosure to the kiln is uncertain but the only other features it surrounds are an alignment of four post-holes (15-18) and the terminal of a shallow ditch (46). However, to the south of the kiln, an arc of post-holes defining an area 6:20 m. by 6 m. is presumably associated with it in some way, perhaps as a windbreak or a temporary storage area for grain or fuel (36, 42, 43, 45, 55-9)- DURRINGTON WALLS 1970. SECTIONS OF CORN-DRYING KILN (SITEI 19/20) SECTION A-A SECTION B-B Chalk [a] Flint | Ash and charcoal Sandstone Fic. 7 Such T-shaped kilns or ovens are numerous in Romano-British contexts and have been discussed in detail by Goodchild (1943). Eleven such kilns have since been recorded at West Blatchington near Hove (Norris and Burstow, 1950) which were principally of T-shaped type, whilst simple ovens of 13th century A.D. date at Beere in Devon have also been interpreted as for drying corn (Jope and Threlfall, 1958). Good parallels for the the T-shaped kiln at Durrington were recorded from Woodcuts (Pitt-Rivers, 1887, pls. V and VIII, 29-31). Thirteen metres to the south of the corn-drying kiln was a second enclosure (FIG. 5; 1, 5, 8), defined by a ditch 1-10 m. to 1-30 m. wide and 30 cms. deep. The dimensions of this rectilinear enclosure are unknown as only its east angles were identified within the area of the excavation and its southern arm is incomplete, but it encloses an area of some 30 m. from north to south. Amongst other items it produced a coin of the period a.p. 337-341. Only two pits (2 and 14) were found within its confines, a third pit (50) was dug through its eastern arm and a fourth was recorded outside its north boundary (11). It was crossed by a lynchet system (Fic. 5, A-C), presumably of post-Roman date. The intervals between the three lynchets were 15 m. and 16-60 m. respectively but no pottery of post- Roman date was found in the plough-soil. The only other features recorded in Site I were a short length of gully (3) and the terminals of two more (4, 10) which have parallel alignments. These three gullies are presumably successive recuts for the boundary of an enclosure which lies to the west of the excavated area. Site II (FIGs. 2, 8) The investigation of the threatened area south of the road (Fic. 1) involved the exca- vation of a roughly trapezoid area 118 m. from east to west and 24 m. from north to south (ric. 8). A number of features was recorded within this area, all of which appear to be of Romano-British date. POST-HOLE SETTINGS: no structures were recorded which can be interpreted as the remains of buildings. The groups of post-holes do not form any coherent patterns and one must assume that they include replacement posts for temporary structures which had an agricultural rather than a domestic purpose. The absence of house structures provides evidence for the view that only the periphery of the settlement has been excavated. The principal groups are as follows (Fic. 8): 131-4, 223-30: linear arrangements of post-holes which do not form any coherent pattern. 139-48: two alignments of post-holes which meet to form an angle. Post-hole 141 is quadruple. Site II D: a scatter of post-holes which do not form any coherent plan. Many occur in pairs (221-222; 181-2; 198-9; 179-80; 162-167; 200-201) and may have held the upright supports for racks. Site II E/B: a group of post-holes enclosing a roughly rectangular area 12 m. by 5 m., the corners of the rectangle being represented by 96, 116, 187 and 21a. It is possible that this arrangement may represent the remains of a fenced enclosure, but the long sides of the rectangle are poorly defined. GULLY sysTEMs: the ditches or gullies recorded in the excavation appear to have been boundary indicators but are clearly of different dates and like the post-hole groupings do not form any coherent patterns. 61:50 cms. wide and 24 cms. deep (FIG. 11), this ditch runs out of the excavation to the east. 66, 67: two shallow gullies of which 67 runs out of the excavation to north and was cut by 66 (ric. 11). 79: an angle ditch 42 cms. wide and 12 cms. deep which was cut by gullies 70 and 124 (FIG. 12). 86 DURRINGTON WALLS 1970 SITE II [] Hollow Le ] Post-hole TO} pit () & 4(0) 50 Scale in Feet O11 2.3545 15 Metres [foce page 86 Fro. 8 DURRINGTON WALLS 1970 © sITE 11 SITE 110 - NORTH FACE SITE 110 : NORTH FACE 184 SITE IIB : NORTH FACE 70, 121, 122, 124: three gullies which are presumably recuts for the same enclosure (FIGS. II, 12). 184: the terminal of a ditch 1 m. wide and 56 cms. deep which runs out of the excavation to the north (FIG. 9). 154: a short length of broad shallow gully 1-40 m. wide and 30 cms. deep (Fic. 13). 182A: a short length of gully 30 cms. wide and ro cms. deep (FIG. 14). PITS: a scatter of pits throughout the excavated area produced the majority of the finds. It is clear that their terminal use had been for the disposal of rubbish. HOLLOWS: the hollows recorded throughout the excavated area were normally not more than 10-15 cms. deep and of irregular plan with a fill of chalky earth and flint nodules. They were frequently cut by gullies, pits and post-holes and two hollows (68 and 160) contained grave pits (215 and 232). Such features are normally termed ‘working- hollows’ and they may well have afforded some temporary shelter for domestic or agricul- tural tasks if one allows for the erosion of the chalk around their upper edges and the possibility of light wind-breaks. OVENS: a group of nine pits have been interpreted as temporary corn-drying ovens (68A, 71, 110, 123, 127-9, 136-7). In plan they are normally dumb-bell shaped or circular with a linguate extension, although more diffuse plans do occur (e.g. 137). Their size is variable, but a characteristic feature is the presence of a floor which shelves downwards into a pit which is occasionally lined with baked clay (e.g. 123, 127) or ash (e.g. 68A, 71). Occasionally the clay lining had become displaced and occurred as lumps of baked clay in the filling of the pit (e.g. 110, 136-7). Six of the ovens occurred in groups of three (127-9, 123, 136-7). There can be little doubt that these pits were used as ovens but there was no evidence as to the substance that was heated. BURIALS: two burials were found in shallow pits that had been dug through the floors of hollows 68 and 160 (215, 232, FIG. 15. PL. IIb). A detailed report on the remains has been prepared by Mr. P. Sandiford (Appendix IT). 215: an oval pit 22 cms. deep housed the remains of an infant of about twenty months. The interment was extended with the head to the south. 232: the remains of an infant of about three months with the head orientated north- west were recorded in an oval pit some 10 cms. deep. No grave goods were recorded with these burials. ECONOMY A total of 449 fragments of animal bone were examined from features of Romano- British date. Mrs. Westley has reported that these were very fragmentary so that it is difficult to estimate the number of individuals represented, but at least 19 animals are present. Amongst these, sheep (6 individuals, 243 fragments) and cattle (6 individuals, 141 fragments) are the most numerous, followed by dog (2 individuals, 11 fragments), horse (2 individuals, 27 fragments) and pig (2 individuals, 26 frag- ments). The presence of red deer was indicated by one worn antler fragment. The cultivation of cereals is indicated by the presence of rotary querns and corn-drying ovens. THE NATURE OF THE SETTLEMENT AND ITS DATE Prior to the construction of the Larkhill Married Officers’ Quarters and the 1970 excavations, the presence of a Romano-British settlement was indicated by the occurrence of sherds in the ploughsoil and by superficial traces of pits and gullies. In 1918 Mr. P. Farrer wrote: “There is a great deal of Romano-British pottery visible 87 1 Sandstone 1¢) 1 2 3 4 5 ce) 1 i ee Scale in Feet 38 —~ Metre Fic. 10 Durrington Walls, 1970. Sections of features. Sancsiaue bf QO 1 _———— Fic. 11 Durrington Walls, 1970. Sections of features. Metre we Ly 87 Be 89 93 oy 5 é ‘Ww 9 96 cues Sandstone ; : é “ar we ~~ 100 102 97 99 10 pee, 107 106 Sandstone am Scale in Feet - ——_—— Metre Fic. 12 Durrington Walls, 1970. Sections of features. go ee ‘ 5 a | §SCale in Feet ' ., 5 Ons ws Qe. Bees poeenre ie? : Do : > Sandstone 136 Sandstone Baked cla ; : peg QE: ae eeG)o%, 0 131 132 1 ns Metre Fic. 13 Durrington Walls, 1970. Sections of features. gl 164 165 166 171 Sandstone re) A 1 Scale in Feet Metre Fic. 14 Durrington Walls, 1970. Sections of features. A ao 232 0 1 2 3 4 Sante fo) 1 a ns SCQle in Feet — Metre Fic. 15 Durrington Walls, 1970. Sections of features. about one hundred yards south- west of the Walls, with many pits and trenches. This was evidently the site of a considerable settlement and the pottery extends for some distance southwards to a point at least 450 yards from the south-west corner of the work’ (Farrer, 1918, 101). Similarly, Mrs. Cunnington recorded that Romano-British pottery had been found over a considerable area south-west of Durrington Walls and west of Woodhenge. It was also found in the upper layers of the ditch at Wood- henge and on the surface thereabouts (Cunnington, 1930, 186). It is clear that the settlement extended for some distance to the west of the 1970 excavations and that a great proportion of it has been destroyed by the construction of the Married Officers’ Quarters on Larkhill. These buildings are on slightly elevated ground overlooking 93 the excavated areas to the east and south and it seems very likely that the focus of the settlement was in this area. Confirmatory evidence for this view is provided by the pottery which had been burnt and freshly broken before it was tipped into pits or pre-existing features such as the corn-drying oven. Included with this rubbish were fragments of building tile and occasional dressed stones, possibly from some demolished structure. The bulk of the coarse pottery forms a homogeneous group whose date range agrees well with that suggested by the coins within the late 3rd and 4th centuries A.D. There is, however, a small amount of residual 2nd century pottery present, including samian, and this is not so abraded as to be the result of manuring. Mrs. Swan has suggested that this pottery must have come from an earlier settlement in the vicinity beyond the limits of the 1970 excavations. THE FINDS NEOLITHIC The Pottery (Fics. 16, 17) by 1. H. LONGWORTH Some gi sherds of Late Neolithic pottery? were recovered from the ditch and pits, the majority (47) being recovered from Pit 27, 17 from Pit 47, 9 from Pit 7, 7 from Pit 48 with a further 11 from the Ditch. Of these, all but one are Grooved Ware of the Durrington Walls sub-style (Wainwright and Longworth, 1971), the exception being P24, a sherd from the Ditch which may be Beaker. Since sherds apparently from the same vessel were recovered from Pits 27 and 47, these pits at least are likely to have been contemporary. Though the collection is relatively small, several of the more characteristic features of the Durrington Walls sub-style are present: vertical plain and decorated cordons, grooved and incised panel filling, cordon junction emphasized by finger-tip disc, etc. Sherd P6 decorated with triple vertical cordons enclosing two narrow vertical panels of ladder pattern shows a variant on the use of this motif as already recorded inside the enclosure (Wain- wright and Longworth, 1971, P50-3). A further novelty is provided by sherd P18, which illustrates a new usage for a blunt-toothed comb in which the comb is impressed into the side of the groove giving a grooved and impressed effect. This together with Pig with point-toothed comb impressions may indicate some absorption of Beaker usages (Wain- wright and Longworth, 1971). If the interpretation of P28 as a piece of potter’s clay, accidentally fired, is correct, then this may be taken as further evidence for local pottery manufacture. Catalogue FIG. 16 Pr Rim sherd of compact slightly porous paste tempered with some grit and ? grog. Light brown both faces with dark grey core. Decoration: a zone of grooved horizontal lines broken by two diagonal grooved lines above short diagonal incised strokes. PIT 4.7 P2, 3 Two rim sherds of compact sandy paste. Greyish brown throughout. Decoration: diagonal incised strokes. PITS 27 and 47 P4 Rim sherd of compact sandy paste. Grey throughout. Decoration: remains of diagonal impressions. PIT 27 | P5 Rim sherd of compact paste. Grey throughout. Decoration: diagonal incised lines. DITCH 29 P6 Wall sherd of slightly porous paste tempered with a little grit. Brown externally, grey internally. Decoration: three vertical cordons, the outer two plain, the central cordon decorated with diagonal slashes, enclose horizontal incised strokes. These cordons separate vertical panels of horizontal and diagonal opposed, incised lines. PIT P7 Wall sherd of fairly compact paste tempered with some grit. Brown both faces with dark grey core. Decoration: vertical plain cordon separating horizontal stroke decoration and diagonal incised lines. Probably from same vessel as P6. PIT 27 P8 Wall sherd of compact paste. Grey to brown externally, grey internally. Decoration: vertical plain cordon on to which run opposed diagonal incised lines. PIT 27 Pg Wall sherd of fairly compact paste. Brown both faces with dark grey core. Decoration: vertical plain cordon on to which run diagonal incised strokes. DITCH 29 Pro Wall sherd of coarse paste tempered with some shell. Light brown externally, brown internally with dark grey core. Decoration: remains of vertical cordon with central groove, the edges carrying diagonal incised strokes. PIT 47 P11 Wall sherd of flaky soft paste tempered with shell, grog and grit. Light brown through- out. Decoration: remains of vertical decorated cordon with central groove, adjoining diagonal grooved lines. PIT 27 P12 Wall sherd of coarse paste tempered with a little shell. Brown both faces with dark grey core. Decoration: remains of vertical plain cordon. PIT P13 Wall sherd of coarse compact paste tempered with grit. Light brown both faces with dark grey core. Decoration: vertical cordon decorated with diagonal strokes with converging diagonal cordons decorated with ? transverse finger-tip impressions. PIT P14 Wall sherd of compact sandy paste. Light brown both faces with dark grey core. Decoration: line made with undefined impressions. DITCH 29 P15 Wall sherd of flaky paste tempered with some grit. Reddish brown both faces with dark grey core. Decoration: vertical plain cordon adjoining diagonal incised strokes. PIT 27 P16 Wall sherd of crumbly paste tempered with grit and grog. Brown both faces with dark grey core. Decoration: vertical plain cordon. PIT 27 P17 Wall sherd of compact sandy paste. Light grey to brown externally, dark grey internally. Decoration: opposed incised diagonal curvilinear strokes and remains of cordon. PIT 27 29 Fic. 16 Durrington Walls, 1970. Late Neolithic pottery. Scale, 1:2. FIG. 17 P18 Three wall sherds of coarse but fairly compact paste tempered with large quantity of shell grit and a little flint. Brown both faces with dark grey core. Decoration: remains of vertical grooved cordon on to which runs a filled triangle pattern executed with coarse toothed comb. PIT Pig Wall sherd of coarse but fairly compact paste tempered with some grit. Grey externally, brown internally. Decoration: vertical plain cordon adjoining diagonal lines of point toothed comb impressions. PIT 7 P20 Wall sherd of compact sandy paste. Light brown externally, grey internally. Decoration: groups of opposed diagonal incised lines. PIT 27 Wall sherd of fairly compact paste tempered with a large quantity of shell grit. Greyish brown externally, brown internally. Decoration: opposed diagonal incised lines, probably part of filled triangle pattern. PIT/27 P22 Wall sherd of compact coarse paste tempered with a large quantity of shell and some other grit. Brown externally and grey to brown internally. Decoration: groups of converging and opposed grooved lines. PIT 48 P23 Wall sherd of coarse paste tempered with shell and some grit. Brown to grey externally, dark grey internally. Decoration: remains of cordon with finger-tip disc and diagonal grooved lines. PIT P24 Wall sherd of fine compact paste tempered with a little fine grit. Reddish brown both faces with dark grey core. Decoration: remains of opposed groups of incised lines bordered by a single incised line, and plain zone. DITCH 29 P25 Base angle of compact slightly porous paste. Grey to brown both faces. Decoration: groups of opposed diagonal impressed lines, the stamp having a slightly serrated edge, some forming a vertical herringbone motif. PITE27 P26 Fragment of base angle, of coarse paste tempered with some grit. Reddish brown externally, grey core, internal surface lost. PIT 27 P27 Base angle of coarse paste tempered with a large quantity of shell. Light brown to grey externally, with dark grey core. Internal surface lost. Decoration: remains of vertical decorated cordon. PIT 47 P28 Fragment of potter’s clay, fired. PIT 7 P2 _ The Flints FEATURE 7: pit Five flakes FEATURE 27: pit Eleven flakes; one short end scraper (FIG. 18, 4) FEATURE 29: ditch Six flakes FEATURE 47: pit Twenty-one flakes OF P27 Fic. 17 Durrington Walls, 1970. Late Neolithic pottery. Scale, 1:2. FEATURE 48: pit Thirty-seven flakes; one scraper on a flake (Fic. 18, 5) Antler FEATURE 27: pit Three picks FEATURE 48: pit One pick (ric. 18, 6) Bone FEATURE 48: pit A rib fragment 25-8 cms. long, polished on one flat surface and ground to a point which is rather worn (ric. 18, 1). An awl 13+2 cms. long with a highly polished shaft which retains the articular surface (Fic. 18, 2). A bone pin or awl with the articular end and tip broken in antiquity (Fic. 18, 3). Fic. 18 Durrington Walls, 1970. Objects of bone, flint and antler from Late Neolithic contexts. Scale, 1, 6, 1:4; 2-5, 1:2. Shell FEATURE 47: pit One limpet shell BRONZE AGE FEATURE 162: post-hole A bronze knife-blade of sinuous profile, cast in one piece with its socket which retains a rivet hole for securing the haft. The blade is complete and strengthened by the provision of a central mid-rib (FIG. 19); it expands slightly towards its hilt and a straight line is formed at their junction. This type of knife is a commonly recurring form in Late Bronze Age hoards where it is associated with elements of the Carp’s Tongue complex of south-east Britain (vide Burgess, 1968, 39). Si) Fic. 19 Durrington Walls, 1970. Late Bronze Age knife. Scale, 1:1. IRON AGE The Pottery FEATURE 21: pit Fourteen body sherds of a hard flint-gritted fabric with a black core and orange-red exterior. Finger-wiping occurs on the exterior of some sherds. FEATURE 23: pit A large jar of hard, pink, flint gritted fabric, with well defined but rounded shoulder and short everted rim. The shoulder is emphasized by an applied cordon decorated with oblique incisions (FIG. 20). A smaller jar of similar fabric with rounded shoulder and short everted rim which is undecorated (FIG. 21, I). The Flints FEATURE 21: pit Nine flakes. A possible unfinished leaf arrowhead or small knife (Fic. 21, 2). A frag- mentary blade with flat edge retouch (FIG. 21, 3). A short end-scraper (FIG. 21, 5). FEATURE 23: pit Seven flakes; one hammerstone Chalk FEATURE 23: pit One chalk object with four notches worked into its edges and cut marks on one scraped face. Possibly a weight (FIG. 21, 4). Baked Clay FEATURE 21: pit Two fragments of daub ROMANO-BRITISH The Coarse Pottery by v. SWAN The bulk of the coarse pottery from the Romano-British settlement at Durrington falls into three categories: that produced comparatively locally, New Forest wares and vessels manufactured by the Oxfordshire kilns. Much of the pottery has a distinctly local character and must have been manufactured in Wiltshire. Unfortunately there are scarcely any closely dated well stratified pottery TOO W600 wz ile bi Mi Zip Fic. 20 Durrington Walls, 1970. Iron Age jar. Scale, 1:4. groups from the county and thus reliably dated true parallels are lacking. The almost complete absence of kiln groups has also added to the problem of assigning the pottery to any particular source. However, in the hope that analagous material may in the future be recognized, as many vessels as possible have been illustrated, though in some groups some of the sherds may ultimately be recognized as rubbish survival. Attention must first be drawn to the grog* tempered hand-made pottery from the site. This type of ware has already been discussed briefly in print (Cunliffe, 1964, 58) and is known to occur in 8.W. Wilts. and in central and Southern Hants. The range of vessels at Durrington (R4, 5, 13, 16, 25, 42, 51, 59, 67, 81, 95 and 101), many of them imitating black-burnished forms, corresponds closely to the known range of forms and the context in which they were found at Durrington substantiates a mid 3rd—4th century a.p. date. In view of the proximity of the Durrington settlement to the New Forest kiln sites, and the ease by which the pottery could have been transported from its source up the River IOI Durrington Walls, 1970. Iron Age jar; Fic. 21 worked flints; worked chalk. Scale, 1, 1:43 2-5; 1:2- Avon, it is not surprising that a great variety of New Forest forms and fabrics are present. The chronology of the New Forest industry has been the subject of much discussion, and the scheme drawn up by Sumner (1927, 81-2) and, following him, Hawkes (1938) of early (250-go aA.D.), middle (290-33 a.p.) and late (330 A.p.+) periods of production has been regularly used by many archaeologists for dating the pottery. However, such a scheme rests almost entirely on the typology of the New Forest products—the centres producing coarse wares having been judged earlier than those making the finer colour-coated and painted wares. Among the strongest arguments against the validity of this scheme has been the discovery of a considerable number of sherds assigned to the ‘middle’ and ‘late’ periods securely stratified in deposits associated with the New Forest kilns recently excavated at Rough Piece, Linwood, a centre producing coarse wares only, and normally considered to be among the earliest manufacturing sites. Furthermore, two probable late 3rd century contexts at Winchester (Cunliffe, 1964, 69 and 176-9) contain the fine colour-coated indented beakers, formerly considered to be exclusively 4th century in date. The flanged bowl from Crock Hill VII which Cunliffe (1965, 45) has used as evidence of an early 3rd century and possibly late 2nd century A.D. starting date for the industry, is a well-known mid-3grd to 4th century a.p. form, and thus a date earlier than the mid-3rd century A.D. for the foundation of the industry cannot be argued on present evidence. In the absence of a reasonably large number of dated stratified deposits in Wessex, it seems wiser, at present, to assume that the variation in kiln output is due, not to chronological factors, but to differing production emphasis, and that until there are published many more dated pottery groups on sites supplied by the kilns, all New Forest products lie within the range mid-3rd century to end of 4th century a.p. or later. In the following report probable dates for New Forest vessels have not been repeated for each example, but must be assumed to lie within this suggested range. Distinctive from the New Forest fabrics, there also occurs at Durrington a quantity of colour-coated and other vessels known to have been produced at about the same time by the kilns sited in the areas around Oxford. The dating of these Oxfordshire red colour- coated wares has been discussed quite recently (Brodribb, Hands and Walker, 1968, 52-3), and it would appear that large-scale production and distribution of the range of vessels imitating late samian forms was under way by the late 3rd century A.p., and continued, with increasing intensity, until the early 5th century A.p., at least. How soon these vessels gained a market in Southern Wiltshire cannot be estimated precisely, though this may well have been immediately, since the population had probably already been buying mortaria from the Oxfordshire area for almost a century. As with the New Forest wares, any sub- division of the chronology within the period suggested for individual Oxfordshire colour- coated forms must await the excavation of many more reliably dated deposits on sites supplied by the kilns. While the many varieties of New Forest vessels are familiar from the work of Heywood Sumner (1927), their relative importance to the Romano-British consumer has received little detailed discussion; still less has the frequency of the various products on Romano- British settlements been related to the competition which the New Forest potters faced from the rival contemporary potting establishments in Oxfordshire. Indeed many excavators have not distinguished the Oxfordshire red colour-coated vessels imitating samian forms from those produced in the New Forest and it is clear that their occurrence in New Forest fabric on archaeological sites has been grossly overestimated. The figures in the table below give some indication of the marketing success of the Oxfordshire potters at a site close to the New Forest kiln centres, particularly with regard to mortaria and ‘pseudo-samian’ vessels, and it is plain from recent work that further away from the New Forest, for example, in the northern part of Wiltshire, the numerical balance of such products is tilted entirely in favour of the Oxfordshire-produced material. In contrast, it may be noted that of the other New Forest products, those potted in the fine, very hard, well-fired, often grey colour-coated fabrics such as the lustrous purple indented beakers, flagons, and other small table-wares were of great importance at Durrington and in fact elsewhere in Wiltshire. 103 TABLE I TOTAL NUMBER OF SHERDS FROM SITE: 2,830 NEW FOREST WARES | OXFORDSHIRE WARES Total number | Total number of sherds of sherds Mortaria 3 Red colour-coated mortaria 15 Imitation samian 28 Other late 3rd-4th century A.D. Indented beakers 28 mortaria Flagons 9 | Imitation samian 45 Other sherds in hard fine colour- | Other red colour-coated vessels 6 coated fabric 48 Parchment ware excl. mortaria 15 No attempt has been made to assess percentage of New Forest coarse grey ‘kitchen’ wares present on the Durrington settlement. The fabric is not sufficiently different from other Roman ‘grey ware’ for it to be readily distinguishable in body sherds alone. Only when rim forms are present can a reliable statement be attempted, and it is likely that its distribution was even more localized than the other New Forest products. In the table above, the term New Forest ‘parchment’ ware has been used exclusively for sherds in the distinctive hard, white, very sandy fabric, with a rough, often slip-coated, surface. Vessel forms included mortaria, flagons, jars and bowls (as R49). Dating evidence In the following list each group of finds is described together, first the coins and samian,°> if any, and then the coarse pottery feature by feature (see FIGS. 22-27). SITE I FEATURE 5: Ditch Ri Hard coarse grey-black fabric, liberally tempered with small-medium irregular pieces of grey and orange grog (ire, R-B pottery), hand-made and burnished on exterior of vessel and interior of rim; ‘wipe’ marks under the rim. The fabric compares closely with that of vessels from Downton, Wilts., Roman villa where the same general form occurs in a deposit dated to post 268 A.p. and perhaps considerably later (Rahtz 1963, fig. 17, no. 10). Similar cooking pots were also present on a Sloden Inclosure New Forest kiln site (unpublished, 1966). Rez Mortarium, hard fine light red micaceous ware, deep reddish-orange matt slip, many translucent large-small pink and red quartz grits, thick base worn away at centre (through use), burnt. Oxfordshire kilns, late 3rd—early 5th century a.p. Also (not illustrated) sherd of New Forest indented beaker (as R63). A late grd—early 5th century A.D. deposit. FEATURE 8: Ditch Coin: 337-341 A.D. Samian: scrap, Central Gaulish, 2nd century A.D. (residual) Rg Mortarium, hard slightly sandy whitish fabric, small-medium translucent pink quartz grits, burnt. Oxfordshire kilns, late 3rd—4th century A.p. R4 Dish/bowl coarse hard grey granular slightly micaceous body with inclusions of fine white quartz sand and small pieces of orange and grey grog, surface black and burnished, hand-made. For same form in this fabric cf. Winchester (Cunliffe, 1964, fig. 21, no. 12). Late 3rd—4th century a.p. R5 Fabric as R4 but without slip, same general form as Gillam (1957), 321; probably late grd—4th century A.D. J Fic. 22 Durrington Walls, 1970. Romano-British pottery. Scale, 1:4. 105 R6 Wide-mouthed jar: very hard grey/buff sandy ware, buff surface, black matt slip, New Forest kilns; this form seems to have been produced at Sloden Inclosure (Sumner, 1927, fig. xxx, 12, and unpublished excavation by the writer, 1966), Ashley Rails (ibid., fig. xxx, 39), Rough Piece, Linwood (unpublished excavation by the writer, 1969), Pitt’s Wood (unpublished excavation by the writer, 1966) and Linwood North (Hawkes, 1938, fig. 3, no. 1). A comparable vessel occurs in a late 4th century A.D. level at Clausentum (Cotton and Gathercole, 1958, fig. 30, no. 14). R7 Very hard sandy pale grey fabric. Probably New Forest product. R8 Fine hard grey slightly micaceous body with light brown surface. Rg Base, probably of small beaker or jar, very hard fine grey ware, buff surface, matt black slip. New Forest kilns. Rio Large ‘rope-rimmed’ storage jar in very hard coarse grey sandy fabric tempered with flint grits, hand-made. New Forest product, produced at Ashley Rails (zbid., fig. x1, no. 3), Sloden Inclosure (ibid., fig. xvu, no. 10 and unpublished, 1966), Black Heath Meadow (vbid., fig. xxu, nos. 4-6). Rough Piece, Linwood (unpublished, 1969), Linwood North (Cunliffe, 1965, fig. 5, no. 23) and Crock Hill VII (zdzd., fig. 6, no. 12). The type occurs in a late 3rd century deposit at Winchester (Cunliffe, 1964, fig 119, NO. 4): FEATURE 19: Stokehole of corn dryer. Samian: Drag. 18/31 or 18/31 R, burnt, Central Gaulish, Hadrian—Antonine (residual). The burnt and freshly broken sherds of this same vessel occurred also in the filling of the corn-drying oven itself and in feature 64/65 and suggests that they were all open and infilled with rubbish at the same time. R11 Hard fine orange micaceous ware, grey core, brick red slip tending to be glossy, rim folded over, exterior surface of bowl slightly facetted, impressed band of demi-rosettes and tripartite band of rouletting, partly burnt. Oxfordshire kilns, imitates late East Gaulish samian form Ludowici Sp. (Ludowici, 1908-12, 245). The production of rosette and other stamped wares in the Oxfordshire potteries seems to have begun at least by the end of the 3rd century a.p. (Webster, 1968), but a widespread distribution may not have been achieved until well into the 4th century a.p. The form, fabric, technique and slip of R11 compare very closely with several vessels from a 4th century A.D. deposit in a well at Mildenhall (Cunnington, 1920, pl. IT). R12 Fabric as Rit, thin cream matt slip, brick-red painted band on rim and shoulder moulding and on interior; some of the fragments of this vessel were burnt subsequent to its breakage. Probable Oxfordshire product; the form occurs on the Headington kiln site (Case and Kirk, 1953, fig. 45, no. 8) in the white iron-free clay. But it seems most likely that the kilns producing mortaria such as Rroo, in identical fabric and slip, also manufactured such orange carinated bowls and slip-coated them in imitation of the widely distributed white Headington products. Late 3rd—early 5th century A.D. Rrg Fabric as R1. Hand-made, lightly burnished or smoothed on surface overall. Probably late 3rd—4th century A.D. R14 Hard coarse very sandy dark grey body, incised wavy line decoration, burnished on exterior above decoration and on interior of rim, burnt. R15 Very hard slightly micaceous granular black ware containing grains of white ? quartz sand (i.e. Gillam Category I fabric, Gillam, 1961, 126-7), hand-made, burnished on interior of rim and exterior of body above band of scored line decoration (cf. Gillam, 1957, no. 145). Late 3rd—4th century A.D. R16 Fabric as R4, black surface, hand-made, lightly burnished below grooves on exterior, ‘facet-burnished’ on interior, burnt. R17 Narrow-mouthed jar, hard buff sandy paste, burnished on exterior, burnt. Possibly New Forest product. Also (not illustrated) mortarium as Re and dish as R4. 106 FEATURE 20: corn-drying oven Samian: (see Feature 19) R18 Very hard dark-grey granular body containing white ? quartz sand, black surface, probably wheel thrown but ‘wiped’ on exterior, lightly burnished overall. Same general form as R5, probably late 3rd—4th century A.D. Rrg Small jug or flagon, fabric as R6, buff surface, thin white slip, burnt. New Forest kilns: similar forms were produced at Rough Piece, Linwood (unpublished, 1969). Also (not illustrated) mortarium as Ra, and flanged bowl in red colour-coated fabric imitating samian form Drag. 38, Oxfordshire product (late 3rd—early 5th century A.D.). The corn-dryer seems to have gone out of use and been filled up with rubbish some time in the 4th century A.D. FEATURE 46: gully R20 Bowl, hard very fine ware, orange/buff core, deep cream surface, traces of ‘wavy-line’ decoration in red/brown paint on internal bevel; New Forest product. This well-known form most frequently occurs in the New Forest very sandy white ‘parchment’ fabric, but, as here, was occasionally potted in the non-sandy fine cream ware which was used more frequently for the New Forest vessels imitating samian forms. Kiln sources at Pitt’s Wood (unpublished, 1966) and Ashley Rails (from Sumner’s excavations, unpublished in Salisbury Museum). A similar vessel occurs in a late 3rd—mid 4th century A.D. context at Winchester, Hants (Cunliffe, 1964, fig. 19, no. 27). Rai Fabric generally similar to that of R1, uneven surface, black on exterior, grey on interior of vessel, hand-made, ‘wiped’ on exterior of body and interior of rim, burnt. For a similar vessel in a late 3rd—mid 4th century a.p. level, cf. Cunliffe, 1964, figz 205 No. 22. This gully was probably infilled post a.p. 270. SITE IT FEATURE 63: pit R22 Mortarium in very hard sandy orange/buff ware, grey core, traces of cream slip, medium-small red/brown and translucent pink quartz grits. Oxfordshire kilns. Second— early 3rd century A.D. (residual). R23 Bowl, fabric as R11, thin matt orange slip, rim folded over, probably imitating a late samian form. Oxfordshire product, late 3rd—early 5th century a.p. R24 Bowl: hard creamy-buff very fine paste containing minute orange specks, red-brown matt slip, burnt. One of the range of New Forest vessels imitating samian forms. Also (not illustrated) sherd of New Forest indented beaker: a 4th century A.D. deposit. FEATURE 64/65: pit R25 as R13. R26 Bowl, fabric as R6, dark grey slip, lightly burnished on rim and interior. New Forest form produced by kilns at Sloden (Sumner, 1927, fig. xxx, no. 2 and unpublished, 1966) and Rough Piece, Linwood (unpublished, 1969). R27 Globular beaker (body sherd), fabric as Rg, dark grey slip, white painted decoration inside incised medallion. New Forest product. This form, and the execution of decoration within incised circles are most characteristic of the Crock Hill potteries (Akerman, 1853, fig. opp. p. 96, no. 4 and fig. p. 99, nos. 12 and 13); occurs in a late 3rd—mid 4th century a.p. deposit at Winchester (Cunliffe, 1964, fig. 20, no. 18). R28 Base of beaker or jar. Body and slip as R24. New Forest kilns. R2o Base of small beaker or jar, hard fine whitish-grey ware, purplish-black matt slip. New Forest source. Also (not illustrated) sherd of New Forest indented beaker. This pit was presumably open and filled the same time as Feature 19 (corn-dryer) as both contained freshly broken sherds from same vessel. Probably 4th century A.D. 107 108 ; =X 229 via! =e | Fic. 23 Durrington Walls, 1970. Romano-British pottery. Scale, 1:4. FEATURE 68: hollow Coins: 286-293 A.D. 364-375 A.D. Samian: Two scraps, Central Gaulish, 2nd century a.p. | Drag. 31, Lezoux, late 2nd century A.D. Rgo Flagon, fabric as R29, dark brown matt slip, burnt. New Forest ware. R31 Jar, body and slip as R24. New Forest product. R32 Narrow-mouthed jar, very hard grey sandy ware, burnished immediately below rim on exterior. New Forest vessel, closely comparable with products from Rough Piece, Linwood (unpublished, 1969). R33 Narrow-mouthed jar, fabric as R6, thin grey slip on exterior and just inside mouth. New Forest kilns, cf. products of Sloden (Sumner, 1927, fig. xxx, no. 15), Rough Piece, Linwood (zbid., fig. xxvu, no. 4 and unpublished, 1969), Linwood North (Cunliffe, 1965, fig. 5, no. 21), and Crock Hill VII (zd7d., fig. 6, nos. 17 & 18). A mid 4th century A.D. deposit at Winchester contains a close parallel (Cunliffe, 1964, fig. 21, no. 32). R34 Body sherd, probably of globular jar, hard grey-brown slightly sandy, fine paste, furrowed, burnt. R35 Body sherd, possibly globular beaker, fabric as R11, with deep orange core containing air ‘bubbles’, deep orange matt slip, band of incised rouletting. Oxfordshire product, late 3rd—early 5th century A.D. R36 Body sherd, paste as Rg, matt purplish/black slip, abraded on exterior, incised wavy line and ‘stabbed’ decoration. New Forest source. R37 Handle of ? jug, hard fine micaceous grey ware, pale orange surface. ? Oxfordshire. R38 Narrow-mouthed jar, fabric as R6, burnished on interior of rim, probably New Forest product. R39 Lid, body as R6, lightly burnished on exterior, burnt. Probably New Forest ware. R40 Fabric as R6, black matt slip on exterior, burnished on interior of rim and exterior of vessel. ? New Forest kilns. R41 Fabric as R32, burnished on interior of rim. New Forest source likely. R42 As Rt. R43 Fabric as R6, rim and interior of body smoothed. New Forest kilns: vessels in this form are known to have been produced at Rough Piece, Linwood (unpublished, 1969), where they have colander bases and also Linwood North (Cunliffe, 1965, fig. 4, no. 9) and Crock Hill VII (zbid., fig. 6, no. 4) where the profiles are incomplete, but where sherds of colander bases have been found associated. R44 Coarse hard very sandy friable grey ware, burnished on interior of rim and exterior of vessel. Perhaps a local product. R45 ‘Rope-rimmed’ storage jar. As Rio. R46 Wide-mouthed jar, body as R6, burnished on exterior of vessel. New Forest ware with parallels at Sloden (unpublished, 1966), Rough Piece, Linwood (unpublished, 1969), Sloden (Sumner, 1927, fig. xxx, no. 12), Ashley Rails (zbid., fig. xxx, no. 39), and Crock Hill VII (Cunliffe, 1965, no. 5). ‘The form occurs ina late 3rd—mid 4th century A.D. deposit at Winchester (Cunliffe, 1964, fig. 61, no. 14). R47 Wide-mouthed jar, fabric and slip as R6, vertical fold in rim visible, stabbed band of decoration on lower part of rim, New Forest kilns; generally similar rim forms com- bined with this stabbing technique are known from Pitt’s Wood (unpublished, 1966), Rough Piece, Linwood (unpublished, 1969), Black Heath Meadow, Linwood (Sumner, 1927, fig. XxII, nos. 10-15), Old Sloden (zbid., fig. xv, no. 15), Ashley Rails (zbid., fig. x1, nos. 8-12), and Linwood North (Cunliffe, 1965, fig. 5, nos. 28-30). R48 Bowl, body as R6, traces of pale grey slip. Possibly New Forest product. R49 Bowl, hard white very sandy paste containing minute red/brown specks, burnt. New Forest ‘parchment’ ware. This well-known form, often painted, was produced at kiln sites such as Ashley Rails (Sumner, 1927, fig. vim, no. 1 and fig. xxxIII, nos. 13 and 14), Islands Thorns (zbid., fig. xxxi, no. 5), and Pitt’s Wood (unpublished, 1966). residual 109 I1lo vy R63) ‘ R62 i “cal é + R45! : “Gi \GaNG ANG: | er MMU IMMAY R47 | = Fic. 2 Durrington Walls, 1970. Romano-British pottery. Scale, 1:4. § ato If Pp ) 4 R50 Fabric as R15, hand-made, lightly burnished overall. For same general type, cf. Gillam, 1957, no. 228; late 3rd—4th century A.D. R51 Dish, body as R4, hand-made, lighty burnished overall, ‘wiped’ on exterior, burnt. Cf. Gillam, 1957, type 330; late 3rd—4th century a.p. R52 Body sherd, fabric, slip, and paint, as R12, probably part of generally similar type of bowl, from same source. R53 Paste and slip as R24. New Forest ware, imitating samian form Drag. 38. Vessels such as this were produced at Ashley Rails (Sumner, 1927, fig. vu, nos. 5-10), Islands Thorns (zbzd., fig. xxxm, nos. 17-19) and Pitt’s Wood (unpublished, 1966). R54 Fabric and slip as R24, probably base of same general type of bowl. R55 Bowl; body and slip as R2, a common vessel imitating samian form Drag. 38 and produced by the Oxfordshire kilns such as those at Dorchester-on-Thames (Harden, 1936, fig. 15, nos. 14 and 15). On the basis of the stratified examples at Shakenoak, Oxon., it has been suggested that thicker coarser flanges as here, and such as occur in the latest town deposits at Dorchester (Frere, 1962, fig. 18, no. 220), tend to be later in date than the finer flanges which terminate in a small bead. R55 may therefore be tentatively dated c. 350 A.p.-early 5th century a.D. R56 Fabric as R35, brick red matt slip, probably base of a bowl in the series imitating samian forms, Oxfordshire kilns late 3rd—early 5th century A.D. FEATURE 75: hollow R57 Extremely hard fine grey ware, traces of lustrous black slip on interior. New Forest ‘bag’ beaker, identical to those produced at Sloden (unpublished, 1966). R58 Wide-mouthed jar, hard sandy slightly micaceous pinky/orange fabric, buff surface, burnished on exterior, hand-made. Possibly local product. R59 Lid, fine light-grey, slightly micaceous ware, tempered with sparse fragments of dark grey grog, black exterior surface and rim lightly burnished, burnt. Also (not illustrated) black burnished flanged bowl, cf. Gillam, 1951, form 228; body sherd of New Forest vessel in the white sandy ‘parchment’ ware, and a late 3rd—4th century A.D. Oxfordshire cream-coloured mortarium, cf. Dorchester-on-Thames (Frere, 1962, fig. 19, no. 247). Group post ¢. 270 A.D. FEATURE 74: pit R60 Hard pinky-brown micaceous gritty ware, black surface, heavily ‘facet-burnished’ on exterior of vessel, hand-made. In general Durotrigian tradition. Also (not illustrated) sherd of red colour-coated ‘imitation samian’ bowl, Oxfordshire kilns, late 3rd—early 5th century A.D. FEATURE 126: pit R61 Bowl/dish with lid seating, fabric as R32, burnished except underside of rim. ? New Forest product. R62 Storage jar, body similar to R32, burnished on exterior of vessel and interior of rim. FEATURE 150: hollow Coin: 337-341 A.D. R63 Indented beaker, rim chipped, paste as R57, lustrous ‘plum-coloured’ slip. New Forest kilns, e.g. Sloden (unpublished, 1966), Ashley Rails (Sumner, 1927, fig. m1, no. 3) and Pitt’s Wood (unpublished, 1966). A late 3rd century deposit at Winchester contained such a vessel (Cunliffe, 1964, fig. 19, no. 68). R64 Fabric as R24, red/brown slip, burnt. One of the range of New Forest vessels imitating the late East Gaulish samian forms; cf. same general type at Ashley Rails (Sumner, 1927, fig. Iv, no. 1), and Pitt’s Wood (unpublished, 1966). R65 Fabric as R32. New Forest kilns, cf. R6. R66 Body as R32. Probably New Forest, where this general form occurs on most of the known kiln sites. III es Us mt i ti Mh; a“ R84, Fic. 25 Durrington Walls, 1970. Romano-British pottery. Scale, 1:4. R67 Fabric as R1, burnt, orange surface, with creamy-orange slip, hand-made, ‘wipe’ marks on exterior. Local imitation of black-burnished form 228 (Gillam, 1957). Similar vessels occur in late 3rd and 4th century a.p. deposits at Winchester (Cunliffe, 1964, fig. 61, no. ro and fig. 21, no. 12). Also (not illustrated) sherd as R35 (late 3rd—early 5th century A.p.). FEATURE 151: hollow R68 Paste as R15 (Gillam, 1961, Category I), surface burnt off. 3rd—4th century a.p. form. R69 Coarse hard orange body, liberally tempered with grey and red grog; probably wheel thrown. Also dish, Gillam, 1957, form 329 (not illustrated). FEATURE 156: hollow R7o Storage jar, fabric as R1o, hand-made, holes pierced through neck, stabbing on rim and knife-trimmed ‘rippling’ effect on body. New Forest source such as Old Sloden (Sumner, 1927, fig. xtv, no. 6) and Linwood North (Cunliffe, 1965, fig. 5, no. 35). R71 Hard coarse pink ware, liberally tempered with orange grog, hand-made, probably local. Also (not illustrated) an imitation of black-burnished form, Gillam, 1957, no. 228. Hollow filled post c. 250 a.p. FEATURE 162: post-hole R72 Beaker (rim missing), fabric as R24, thick pale orange-brown matt slip, two zones of incised rouletting, grit (from kiln) adhering to base. Probably New Forest product. FEATURE 185: pit R73 Jug/flagon, body as R6, traces of part of vertically burnished area on neck. New Forest kilns; e.g. Sloden Inclosure Kiln II (Sumner, 1927, fig. xvi, nos. 8. 9 and 15), Lower Sloden (ibid., fig. xx1, no. 2), Linwood North (Cunliffe, 1965, fig. 4, no. 1) and Rough Piece, Linwood (unpublished, 1969). R74 Fabric as R15, cf. Gillam, 1957, form 319; 3rd or 4th century A.D. Also (not illustrated) narrow-mouthed jar as R6 sherd of Oxfordshire red colour- coated ware, and fragment of mortarium in New Forest sandy ‘parchment’ ware. A post ¢. 270 A.D. pit group. FEATURE 186: hollow R75 Mortarium, paste as R3, small-medium translucent white and pink quartz grits. Oxfordshire kilns, late 3rd—early 5th century A.D. R76 Mortarium hard slightly sandy cream body, pink core, many small-medium translucent red/brown, pink and white quartz grits. Oxfordshire kilns, late 3rd—early 5th century A.D. R77 Flagon or jug, hard, fine, orange, slightly sandy fabric, grey core, cream matt slip. ? South Midlands source. R78 Narrow-mouthed jar, hard sandy orange ware, traces of pale orange slip. Probably New Forest product, closely comparable with vessels produced at Crock Hill VII (Cunliffe, 1965, fig. 7, nos. 22 and 23) and at Sloden (unpublished, 1966). Also (not illustrated) sherd of New Forest indented beaker. FEATURE 188: pit Coins: 337-++ A.D. 341-348 A.D. R79 ‘Pinched lip’ jug, fabric as R57, uneven purplish/brown slip overall. New Forest kilns, cf. Sloden Inclosure (Sumner, 1927, fig. xvi, no. 24, and unpublished, 1966) and Crock Hill (Bartlett, 1873, opp. 320; Wise, 1863, 218). R80 Fabric as R6, burnished on exterior below rim, and on interior of neck and rim, white slip on exterior and interior of rim. Probably New Forest product. rg R81 Bowl/dish, paste as R4, thick dark grey core, black surface, hand-made, burnished overall. For form cf. Gillam, 1957, type 228; mid-3rd to 4th century A.D. FEATURE 207: pit Samian: Fic. 26: Drag. 31, East Gaulish, very thick base worn and burnt, stamped FIRM/// NVSF, by the Rheinzabern potter FIRMANVS. Late end or early 3rd century a.p. (residual). Also (not illustrated) two sherds of Oxfordshire red colour-coated ‘imitation samian’, post 270 A.D.; and small sherd of jar or bowl in fine slightly sandy orange fabric, containing a trace of very fine mica, thick orange/brown transparent lead glaze covered with fine cracks and partly chipped off the body. The fabric and glaze exactly match some sherds from Armsley, Hants, recently published (Musty, 1969) and dated to the late 1st or 2nd century a.p. The occurrence of this ware at Durrington, together with a new find from the Roman settlement at Stratford-sub-Castle, Wilts. (Arch. Rev., 1969, 49), and the examples cited by Musty, suggest a greater concentra- tion of this fabric in S. Wiltshire than elsewhere. Indeed, the scarcity of this lead-glazed fabric on Roman sites in the rest of Wiltshire, may weaken the case for its suggested Mendip source and imply a more southerly origin.°® ARN IVY SF Fic. 26 Durrington Walls, 1970. Potter’s stamp. Scale, 1:2. FEATURE 208: pit R82 Narrow-mouthed jar, hard orange/buff sandy ware, brown core, burnished on rim. Possibly of New Forest manufacture, cf. Sloden (Sumner, 1927, fig. xxx, no. 13), and Ashley Rails (7bid., fig. xxxi, no. 38). R83 Hard fine orange/buff fabric, grey core, dark brown matt slip. New Forest product, probably a globular beaker of the same general type as those produced at Ashley Rails (Sumner, 1927, fig. 111, nos. 4 and g) and Sloden (unpublished, 1966). R84 Jar, orange hard fine sandy body, thin dark grey core, cream matt slip on exterior. This distinctive fabric and slip combination is known from at least 25 sites scattered over Wiltshire, though only three examples come from the Southern part of the county and on every site the relative amount is quite small. Flagons seem to be the most common form though small jars and beakers also occur. The kiln source is at present unknown; it is probably unlikely to be local and may perhaps lie somewhere in the South Midlands.’ Also (not illustrated) sherds of 4 New Forest indented beakers. FEATURE 210: hollow R85 Narrow-mouthed jar, paste as R32. New Forest source, generally similar vessels were produced at Rough Piece, Linwood (unpublished, 1969). R86 Fabric similar to R84, but slightly coarser and containing small inclusions of quartz grit. Thick white slip. ? South Midlands source. Also (not illustrated). Sherd of mortarium in New Forest white sandy ‘parchment’ ware; sherd of red colour-coated Oxfordshire ‘imitation samian’; sherd of Oxfordshire mortarium (fabric and slip as R1oo), and flanged pie dish (Gillam Form 228) in Gillam’s (1961) Category I fabric. FEATURE 213: pit Coin: 388-402 A.D. Samian: Drag. 31, Central Gaulish, Antonine (residual). R87 Fabric and slip as Ro. Upper portion of rather thick indented beaker; New Forest kilns, see R63. 114 102 Durrington Walls, 1970. Romano-British pottery. Scale, 1:4. R88 Body as R6, lightly burnished on rim and exterior and of vessel. ? New Forest product. R89 As R53. Rgo Fabric and slip as Re. Oxfordshire kilns, perhaps part of a bowl such as those produced at Dorchester-on-Thames (Harden, 1936, fig. 15, nos. 23 and 25). Late 3rd-early 5th century A.D. Rg1 Cooking pot, hard brown slightly sandy ware, grey core, dark grey surface smoothed and ‘wiped’. Rog2 Jar, hard orange sandy fabric, burnished on upper side of rim. ? New Forest product. Ro3 Jar, hard, slightly sandy orange ware, with minute inclusions of red grog, burnished on interior of rim and exterior of body. Rog4 Storage jar, fabric as R1o, hand-made, but probably wheel-finished. New Forest kilns, cf. Ashley Rails (Sumner, 1927, fig. x1, no. 2), Sloden (zbzd., fig. xx, no. 9, and un- published, 1966) and Rough Piece, Linwood (unpublished, 1969). Ro5 Bowl, fabric as R1, hand-made, burnished on surface. Ro6 Paste similar to R15, i.e. Gillam (1961) Category I, hand-made, burnished on surface, part burnt orange, cf. Gillam (1957) Form 228. This proportionately deeper form tends to be more common in the 4th century a.p. than the proportionately shallower vessels of the same general rim type which are dated from the middle of the grd century A.D. Rog7 Fabric as Gillam (1961) Category I, hand-made, burnished on surface, 3rd—4th century A.D. form. FEATURE 231: pit Ro8 Body as R24, but part burnt grey, traces of red-brown matt slip. One of the range of New Forest vessels imitating late samian forms and finish, such as those produced at Ashley Rails and Pitt’s Wood. Rgg Fabric as R32. ? New Forest kilns. The following sherds are included for their intrinsic interest only, as they occurred in ploughsoil, and presumably originated from either the more superficial occupation layers associated with the settlement, or from the surfaces of pits and hollows. Coin (Area IC) 392-395 A.D: R1ioo (Area IIB) Mortarium, paste as Rr2, thick grey core, cream matt slip, translucent pink and red quartz grits on interior. Oxfordshire product, cf. Dorchester-on-Thames (Frere, 1962, fig. 19, nos. 238-9). Post 250 A.D. Rio (Area IIB) Fabric as Rr, burnt orange-buff on exterior, surfaces smoothed or lightly burnished. For a similar vessel in a late deposit, cf. Winchester (Cunliffe, 1964, fig. 21, no. 15). Rroz (Area IIE) Mortarium, body as R4g (i.e. New Forest sandy ‘parchment’ ware), traces of white slip, burnt, many multi-coloured flint grits on interior. Cf. Islands Thorns (Sumner, 1927, fig. xxx, no. 11) and Sloden (unpublished, 1966). The Coins by Pp. E. CURNOW 1 Carausius 286-93 RIC V ii 880 (5c) 68 2 GLORIA EXERCITVS 337-41 ‘Trier Constans LRBC I 133 p. 150 3 Irregular 3 a 337+ Chelinien Aa Cis nec: jh Ch. eG spaes 8s 4 PAX PVBLICA 337-41 ‘Trier Helena S31) RLOBNP: 8 5 VICTORIAE DDAVGG QNN 341-8 : Constans 5) 25 Onn: 188 6 SECVRITAS REIPVBLICAE 364-75 Arles Valens sn Ueeapes 68 7 VIRTVS ROMANORVM 392-5 Trier Eugenius R/C X Trier ro6d I(3) (clipped siliqua) 8 SALVS REIPVBLICAE — 388-402 =: Arcadius sn Chagoewetcw 2g 116 ul | | |, ‘al, nT i || at || il hl le ul Ii ila HU Fic. 28 Durrington Walls,*1970. Romano-British iron-work. Scale, 1:2. 117 Iron FEATURE 5: ditch A rectangular washer with a central rivet hole (Fic. 28, 8). FEATURE 64/5: pit A simple strip with a single rivet hole. FEATURE 68: hollow An iron ring (FIG. 28, 9). A fragmentary latch-lifter. FEATURE 68B: pit The tip of a knife-blade (Fic. 28, 5). FEATURE 137: pit The tip of a knife-blade with wood traces preserved on both faces (FIG. 28, 2). FEATURE 152: pit The much corroded tip of a possible sickle (Fic. 28, 4). FEATURE 153: hollow A flat knife-blade with straight sides, a rounded tip and short tang which is centrally positioned (FIG. 28, 1). FEATURE 156: hollow A flat strip, rectangular in section and tapering towards one terminal (Fic. 28, 7). A fragmentary handle, distorted at one terminal (Fic. 28, 3). A narrow strip curved back upon itself at both terminals. FEATURE 185: pit A horse-shoe fragment with one nail preserved in position (FIG. 28, 6). An unidenti- fiable object, probably part of a knife blade. FEATURE 210: hollow An unidentifiable object, possibly part of a binding strip. In addition, a number of nails was recorded from the features: those found in the plough- soil were discarded as being of uncertain age. A basic division can be made into carpentry nails and clamps on the one hand and hob-nails which were probably used to stud footwear. On this basis a total of 124 nails and 69 hob-nails were recorded. Of the latter, 60 were recorded from Pit 207 and presumably represent the decayed remains of a sandal. Of the carpentry nails, 49 were recorded from the corn-drying oven (19/20) and the remainder were fairly evenly distributed amongst 28 features as follows 1(2); 2(1); 5(11); 8(2); 38(2); 60(2); 61(1); 62(3); 63A(1); 64/65(3); 68(13); 69(5) 3 72(1); 74(1) 78(1)3 84(1); 123(2)5 126(2); 137(2); 139(1); 152(1); 160(2); 185(2); 186(1); 188(7); 207(1); 210(4); 231(1)- Bronze IIB (2) A pin with a round head (FIG. 29, 5). FEATURE 2: pit An unidentifiable object. FEATURE 5: ditch A short length of contorted wire—possibly scrap. FEATURE 19(1): stokehole of corn-dryer A fragment of a wire bracelet with a ring around the bow and transverse incisions (FIG. 29, 6). 4th century A.D. FEATURE 20(2): corn-drying oven A much worn and corroded spoon with a rectangular tang which was presumably intended for insertion into a wooden handle (Fic. 29, 4). 4th century A.D. FEATURE 151: hollow A fragmentary bracelet composed of two twisted rods (FIG. 29, 3). 118 8 ‘ 10 9 CMU, Fic. 29 iG Durrington Walls, 1970. Romano-British bronze-work ( 10, I: 1-8), bead (9), spindle-whorl (10). Scale, 1-9, 1:13 Be 11g FEATURE 162: post-hole The point of a pin (FIG. 29, 8). FEATURE 188: pit A bronze object of uncertain function (FIG. 29, 2). A conical bronze rivet cap (Fic. 29, 7). Part of a thin bronze plaque with diagonal fluting to represent a leaf or feather.® Complete examples have been found which are thought to have formed detachable adornments for ritual crowns: a single feather and a crown came from a hoard of religious objects at Cavenham, Suffolk (Layard, 1925, 261-3), and a feather from the temple site at Lydney, Glos. (Wheeler and Wheeler, 1932, pl. xxrx, 137, p. 90). The same type of decoration occurs on larger plaques bearing votive inscriptions, such as those in silver from Barkway and Stony Stratford (British Museum, 1958, FIG. 31, 9, p. 62), or the figure of a deity such as that from Maiden Castle (Wheeler, 1943, pl. XXIXB, 131). These are thought to have decorated shrines (cp. Toynbee, 1964, 329), so that to whichever class the present object belongs, a religious purpose is virtually certain. Where datable, the examples quoted seem to belong to the 4th century A.D. (FIG. 29, I). Stone FEATURE 2: pit A fragment of perforated roofing tile. FEATURE 8: ditch A fragment of perforated roofing tile. FEATURE 19(1): stoke-hole of corn-drying oven A whetstone. FEATURE 19(6) A whetstone (FIG. 30, 2). FEATURE 19(8) A fragment of perforated roofing tile. FEATURE 46: ditch A hammerstone. FEATURE 68A:: pit A fragment of a lower stone of a large rotary quern. Two nails preserved in its upper surface may indicate that the grinding face was covered with a finer material (FIG.30, 4). A fragment of the lower stone of a rotary retaining part of its central perforation (FIG. 30, 1). A fragment of a rotary quern. A fragment of a possible saddle quern with one very smoothed and worn surface (FIG. 30, 3). Flint The flint artifacts from the ploughsoil were largely dispersed by the machine and waste flakes only were recorded from eight features as follows: 1, 2, 5, 22, 72, 177, 184 and 217. Glass FEATURE 188: pit A small spherical blue bead (Fic. 29, 9). Lead FEATURE 69: hollow Two molten lead fragments. Iron Slag Iron slag was recorded from features 69, 72, 125, 155 and 207. 120 = Baked Clay FEATURE 68: hollow A pottery spindle-whorl (FIG. 29, 10). FEATURE 2: pit Fragments of daub. Tile Tile fragments, occasionally with keying incisions on one face, were recorded from features 8, 68A, 156 and 160. Oyster Shells A number of oyster shells as recorded from the ploughsoil and from features 2, 5, 8, 60, 69, 123, 151, 185 and ato. APPENDIX I The Animal Bones from Durrington Walls, 1970 by B. WESTLEY, B.SC., F.Z.S. Seven hundred and eighty fragments of bone were identified from several boxes of material. Most of the long bones had been chopped into two, and subsequently much broken up by erosion and decay, and there were many splinters and shaft fragments not deter- minable. Only five main long-bones remain entire and their overall lengths are given below. These are of sheep and cattle. For each species, the minimum possible number of individuals represented is given, though in the case of this material, I hardly feel such calculation is worth serious attention. For example, small fragments of the right and left limbs of an animal may or may not be a possible match. It is unlikely that they are, if they come from a wide area which may represent a long time-span, yet for the purpose of ‘minimum possible’ calculation, they must be reckoned as one. The material is so fragmentary that scarcely any of the pieces articulate, or can be seen to match, left and right, and it seems probable that many more individuals are present than the figures suggested below. In my opinion it is not possible to estimate, from bone fragments, the number of animals on a site, though the proportions of species relative to each other may be indicated. The bones fall into three groups, ‘Late Neolithic’, ‘Iron Age’ and ‘Romano-British’, so the fauna are listed separately, as follows: Late Neolithic (316 fragments) CARNIVORA Dog Tooth I, upper carnassial (medium) Mandible 1 fragment, condyle. Very large Possibly a wolf, but specific determina- tion impossible ARTIODACTYLA Pig Skull 16 mandible fragments, small and most without teeth Tush 3, one a large boar Humerus 18 distal and shaft fragments Radius 9 fragments, proximal and distal Ulna 3 proximal fragments. One boar? Ribs 20 fragments Pelvic 8 Femur 4 5 Distal and shaft Tibia 30 35 All parts Tarsal 13 5 Teeth 78 loose. 16 deciduous Total 202. At least 8 individuals UEP He} RODENTIA PERISSODACTYLA ARTIODACTYLA CARNIVORA PERISSODACTYLA Roe deer Sheep Cattle Vole Horse Pig Sheep Cattle Dog Horse Antler and pedicle, joined to skull fragment Humerus Radius Scapula Tibia 1 fragment, shaft I bb) oP) I 29 29 2 fragments, distal At least 2 individuals Skull Teeth Ribs/Vertebrae Humerus Radius Hind limb 4 fragments. One frontal, with incom- plete horncore, measuring 193 mm. circumference at thickest point near base 8 adult, little worn 65 (6 unfused, immature) 4 fragments, distal 6 fragments (3 proximal, 3 distal) 11 small fragments Total 98. At least 4 individuals Remains of one individual Tron Age (24 fragments) Tooth Rib Tooth Teeth Metacarpal Metatarsal Radius Tibia 1 cheek tooth 1 (possible) 1 deciduous 9 1 shaft fragment I ”° 39 I I eee, 29 be) h} At least two individuals Teeth Sacrum Metacarpal Phalanx (2) Phalanx (3) 3, fragmentary 1 small fragment 1 fragment I, entire 2, entire, making hoof Possibly one individual Mandible Maxilla Atlas Axis Radius Metacarpal Teeth Romano-British (449 fragments) 2 fragments, left, with M* and M? 2, left, adult. Medium terrier-size 1 fragment 1 fragment 1 fragment 1 fragment 3 loose Total 11. At least two individuals Skull Maxilla Mandible Teeth Radius Scapula 1 fragment I 29 I 9 21 loose 2 fragments, proximal and shaft 1 fragment Total 27. At least two animals, one old and one perhaps 6 years. 123 ARTIODACTYLA Pig Occiput 1 fragment Mandible 7 fragments, some teeth, newly adult Maxilla 2 fragments Teeth 11 loose. Mature, little worn Vertebrae Ortiis Tibia I ; Proximal, unfused Phalanx I, entire Total 26. A minimum of 2 individuals Red deer 1 antler fragment, worn Sheep Horncore 2 fragments Mandible 11 fragments Teeth 170. Two deciduous, and others mature, unworn Humerus 8 distal and shaft fragments Radius g fragments Metacarpal 5 fragments and 2 entire, 127 and 121 mm. length Vertebrae 6 fragments Pelvis 3 Be Femur i Distal and shaft Tibia 4 shaft fragments Metatarsal 15 fragments, all parts Phalanges 3 entire Tarsal Total 243 fragments. A minimum of 6 individuals Cattle Horncore A basal fragment, 157 mm. circumfer- ence at thickest point Mandible 6 fragments Teeth 51 (5 deciduous). Others adult, unworn Ribs, vert. 8 fragments Humerus I proximal fragment, immature Radius 3 fragments, proximal and distal Ulna 7 proximal fragments Metacarpal 2 entire, 186 and 188 mm. overall 11 fragments Tibia 8 fragments, distal and proximal Femur 8 small fragments Metatarsal 10 fragments and one entire, 218 mm. Tarsals 16 Phalanges 9 Total 141. At least 6 individuals Discussion The earliest fauna is typically Late Neolithic, with a great preponderance of pig bones. They comprise two-thirds of the whole, and there is only 2 per cent of sheep. The pigs are of a small domestic variety, with evidence of one boar, perhaps wild. The cattle appear to be of medium size but unfortunately yield nothing for measurement as a comparison with the later cattle. ‘The remains appear fairly large and the stout horncore (193 mm. maximum circumference) indicates something certainly larger than the ‘Celtic ox’. The environment was no doubt more wooded than at present, since the pigs are so numerous, and this is confirmed by the presence of deer. It is perhaps a little surprising that there are no birds and little evidence of wild animals; fox, badger and hare being absent, and also red deer. The ‘Iron Age’ group is a small one of only 24 fragments and no conclusions can be reached as to size and type of animal. 124 The ‘Romano-British’ group shows a major change in husbandry, with sheep now forming two-thirds of the whole and pig reduced to 7 per cent. The cattle proportions remain about the same, a third of the whole. The horse is small, but the cattle appear larger than the familiar Romano-British type found on most sites. This is confirmed by measurement of the three complete metapodials. I have kept a register of whole long-bones of prehistoric cattle that have passed through my hands, and from this it appears that an average Romano-British metacarpal would be about 174 mm., whereas our present speci- mens measure 186 and 188 mm. overall. Similarly, an average metatarsal from my list is 199 mm., whereas the specimen from Durrington is 218 mm. The sheep are of a primitive type, horned and slender, though here again they are larger than usual if one can use two single bones as a comparison. The average sheep metacarpal on my list is 110 mm., whereas our measured examples are 127 mm. and 121 mm. All the food-bones have been chopped or split in antiquity. APPENDIX II The Human Skeletal Remains from Durrington Walls, 1970 by PETER SANDIFORD, B.SC. A. SKELETON F 215 An infant of about 20 months. Unfortunately these remains have suffered from a great deal of post-mortem damage. This is probably the result of the proximity of the burial to the surface. The largest fragment of bone, a complete tibia, is only 10 cm. long. All parts of the body were represented in the following proportions: 106 skull fragments A complete mandible 12 upper limb long bone fragments 7 lower limb long bone fragments 8 phalanges 10 metacarpal/metatarsal fragments 4 fragments of scapula 55 rib fragments 2 pelvic fragments 31 vertebral fragments Despite the extensive breakage a good deal of information was obtained from the nandible and maxillary fragments. The state of eruption of the teeth (all deciduous) was as follows: Lower dentition, left side Central incisor — fully erupted Lateral incisor — just appearing Canine — about to erupt but growing medially First molar — fully erupted but crown deflected medially at an angle of 56 degrees to the vertical Second molar = not yet erupted Lower dentition, right side Central incisor — fully erupted Lateral incisor —- just appearing Canine _ about to erupt First molar — fully erupted but crown deflected medially at an angle of 50 degrees to the vertical Second molar — not yet erupted 125 Upper dentition, left side Central incisor — fully erupted, slight wear observed Lateral incisor — half erupted Canine — just appearing, no deflection First molar — fully erupted and vertical Second molar — eruption imminent Upper dentition, right side Central incisor — fully erupted, slight wear Lateral incisor —= half erupted Canine —- just breaking through First molar — fully erupted and vertical Second molar —_ eruption imminent The most usual time for the eruption of deciduous canines is between 16 and 20 months, and for second molars between 20 and 24 months. Since the upper second molars seem about to erupt this would place the age of the individual at about 20 months. These figures assume adequate dietary provision and the age at death would be extended slightly if this had not been available. There is no evidence from any of the other remains to suggest that this might have been the case. Even at this early stage of development there is evidence of overcrowding in the lower jaw, reflected in the peculiar angle of growth of the molars and left canine. No abnormalities were observed either in the rest of the skull or in the post-cranial remains. ‘Their small size and lack of epiphyses confirmed the dental evidence, especially the presence of the frontal (metopic) suture which has normally disappeared after the second year. The pelvic fragments contained the sciatic notch which was in both cases extremely wide. Even in an infant 20 months old the sciatic notch dimorphism is significant and here would point to the individual having been female. However, with no means of confirmation this must be regarded with caution. There are no indications as to the cause of death. B. SKELETON F 232 An infant of about 3 months. These fairly complete but extremely well broken up remains consisted of fragments as follows: 26 long bone fragments 2 pelvic fragments 42 skull fragments 2 mandibular fragments 8 tooth crowns 12 metatarsal/metacarpal fragments 4 phalanges 1 clavicle 4 scapula fragments 34. rib fragments 19 epiphyseal fragments 49 vertebral fragments ane overall size of the bones is very small inde ed; the skull for example is on average only 0-5 mm. thick and the right humerus is only 6-5 cm. long. They obviously belonged toa eae immature individual. The mandible is in two separate halves with no sign whatsoever of the symphysial fusion which begins to take place in the first year of life. The body of the bone is just a thin shell with no clear partitioning of the sockets. Eight tooth crowns were recovered, four molars and four incisors. There were no complete teeth with both crowns and roots. 126 Compared to the adult form the coronoid process is very large and projects well above the head of the mandible. Also the ascending ramus is swept back so much as to be almost horizontal. All this is typical of the early post-natal mandible. It is clear that the individual must have died within a few months of birth. At death no teeth had erupted so the infant must have been less than 6 or 7 months of age. ‘The incomplete mandibular development plus the lack of any sign of root formation would indicate an age of about three months. The other remains support the conclusions drawn from the dental and mandibular evidence in that their size and state of development are consistent with an infant of three months. The sciatic notch on the pelvis would point to the possibility of the individual having been male but as with F 215 this is unconfirmed by other features. No evidence of abnormality was found and there were no indications as to the cause of death. BIBLIOGRAPHY Akerman, J. Y., 1853. An account of excavations on the site of some ancient potteries in the western district of the New Forest. Archaeologia, 35, 91-9. Arch. Rev., 1969. Archaeological Review, 4, for 1969. Bartlett, J. Pemberton, 1873. The Ancient Potteries of the New Forest, Hampshire. Archaeol. J7., 30, 319-24. Booth, A. St, J., and Stone, J. F. S., 1952. A trial flint mine at Durrington, Wiltshire. Wiltshire Archaeol. Natur. Mist. Mag., 54, 381-8. British Museum, 1958. Guide to the Antiquities of Roman Britain. Brodribb, A. C. C., Hands, A. R., and Walker, D. R., 1968. Excavations at Shakenoak 1. Burgess, C. B., 1968. The later Bronze Age in the British Isles and north-western France. Archaeol. 7., 125, 1-45. Case, H. J., and Kirk, J. R., 1953. Archaeological notes 1953. Oxoniensia, 17/18, 224-6. Cotton, M. A., and Gathercole, P. W., 1958. Excavations at Clausentum, Southampton, 1951-5 4- Cunliffe, B., 1964. Winchester Excavations 1949-60. Vol. I. Cunliffe, B., 1965. Report on the excavation of three pottery kilns in the New Forest, 1955. Proc. Hampshire Fid. Club, 23, Pt. 1, 1-45. Cunnington, M. E., 1920. Notes on the pottery from a well on the site of ‘Cunetio’ (Mildenhall), near Marlborough. Wiltshire Archaeol. Natur. Hist. Mag., 41, 153-9. Cunnington, M. E., 1929. Woodhenge. Cunnington, M. EM 1930. Romano-British Wiltshire. Wiltshire Archaeol. Natur. Hist. Mag., 45, 166-216. Cunnington, R. H., 1954. The Cunningtons of Wiltshire. Wiltshire Archaeol. Natur. Hist. Mag., 55, 211-36. Evans, J., 1881. The Ancient Bronze Implements, Weapons and Ornaments of Great Britain and Ireland. Farrer, P., 1918. Durrington Walls or Long Walls. Wiltshire Archaeol. Natur. Hist. Mag., 40, 95-103. Frere, S., 1962. Excavations at Dorchester on Thames, 1962. Archaeol. F., 119, 114-49. Gillam, J. P., 1957. Types of Roman coarse pottery vessels in northern Britain. Archaeol. Aeliana, 4th ser., 35, 180-251. Gillam, J. P., 1961. “The coarse pottery’, in Steer, K. A., Excavations at Mumrills Roman Fort 1958-60. Proc. Soc. Antiq. Scot., 94, 86-132. Goodchild, R. G., 1943. T-shaped corn-drying ovens in Roman Britain. Antiq. 7., 23. 148-153. Harden, D. B., 1936. Two Romano-British potters’ fields near Oxford. Oxoniensia, 1, 81-102. Hawkes, C. F. C., 1938. An unusual find in the New Forest Potteries at Linwood, Hants. Antiq. 7., 18, 113-36. Hoare, R. C., 1812. The Ancient History of South Wiltshire, Volume 1. Hodges, H. W. M., 1956. Studies in the Late Bronze Age in Ireland: 2. The typology and distribution of bronze implements. Ulster 7. Archaeol., 19, 29-56. Jope, E. M., and Threlfall, R. I., 1958. Excavation of a medieval settlement at Beere, North Tawton, Devon. Medieval Archaeol., 2, 112-40. Layard, N. F., 1925. Bronze crowns and a bronze head-dress from a Roman site at Cavenham Heath, Suffolk. Antiq. F., 5, 258-65. Ludowici, W., 1908-12. Rémische Ziegel-Gaber Kat. IV. Meiner Ausgrabungen in Rheinzabern. Musty, J., 1969. Roman glazed pottery from Armsley, Godshill, Hants. Antiq. 7., 49, 128-30. Norris, N. E. S., and Burstow, G. P., 1950. A Prehistoric and Romano-British site at West Blatchington, Hove. Sussex Archaeol. Collect., 89, 1-56. Pitt-Rivers, A. H., 1887. Excavations in Cranborne Chase, near Rushmore, on the Borders of Dorset and Wiltshire, Vol. 1. Rahtz, P. A., 1963. A Roman Villa at Downton. Wiltshire Archaeol. Natur. Hist. Mag., 58, 303-41. Smith, M. A., 1959. Some Somerset hoards and their place in the Bronze Age of Southern Britain. Proc. Prehist. Soc., 25, 144-87. Stone, J. F. S., and Young, W. E. V., 1948. Two pits of Grooved Ware date near Woodhenge. Wiltshire Archaeol. Natur. Hist. Mag., 52, 287-306. Stone, J. F. S., 1949. Some Grooved Ware pottery from the Woodhenge area. Proc. Prehist. Soc., 15, 122-7. Stone, J. F. S., Piggott, S., and Booth, A. St. J., 1954. Durrington Walls, Wiltshire: recent excavations at a ceremonial site of the early second millennium B.c. Antiq. F., 34, 155-77- 127 Sumner, H., 1927. Excavations in New Forest Roman Pottery Sites. Toynbee, J. M. C., 1964. Art in Britain under the Romans. Wainwright, G. J., and Longworth, I. H., 1971. Durrington Walls: Excavations 1966-68. Reports of the Res. Comm. of the Soc. of Ants., xxviI. Webster, G., 1968. A sherd of pottery from Cirencester. Antig. F., 48, 102-3. Wheeler, R. E. M., and Wheeler, T. V., 1932. Report on the Excavation of the Prehistoric, Roman and Post-Roman site in Lydney Park, Gloucestershire. Reports of the Res. Comm. of the Soc. of Ants., rx. Wheeler, R. E. M., 1943. Maiden Castle, Dorset. Reports of the Res. Comm. of the Soc. of Ants., xm. Wise, J. R., 1863. The New Forest: Its History and Scenery. t Inventaria Archaeologica. 3rd set: GB 14-18. Grave Groups and Hoards of the British Bronze Age, ed. M. A. Smith. GB 17. 2 Inventaria Archaeologica. 6th set: GB 35-41. Late Bronze Hoards in the British Museum, ed. M. A. Smith. GB 35. 3 Excluding fragments less than } in. square. 4 As used by potters the term ‘grog’, here and throughout, indicates tempering of a ceramic nature. 5 The writer wishes to thank Mr. G. Dannell for _ N oD 4 his help with the samian, and Mr. B. R. Hartley for reporting on the samian stamp. 6 Since this report was written a quantity of this ware has been published from Chichester and a local production centre suggested there (Down, A., and Rule, M., Chichester Excavations, I (1970), 77-79; Pls. 15-17). 7 The author is grateful to Dr. G. Webster for this suggestion. 8 IT am indebted to Miss S. Butcher for providing the report on this object. WANSDYKE, EXCAVATIONS 1966 TO 1970 by H. STEPHEN GREEN with contributions from G. W. DIMBLEBY, J. N. L. MYRES and D. J. BONNEY SUMMARY Excavation was undertaken at Red Shore (SU 117648) and at New Buildings (SU 193665). At Red Shore, Celtic fields (certainly in use in the late rst century A.D.) underlie the Wansdyke. After the cessation of arable farming, the area became pasture land until the construc- tion of the dyke. The dyke, west of SU 135653 near Shaw House, appears to have cut across pre-existing estate boundaries which have become the present parish boundaries. The dyke consists of bank, ditch and counterscarp (all primary) ; the bank is of dump construction with an unbroken slope into the ditch and exhibits evidence of only one period in its build. There is evidence that the causeway at Red Shore, where the Ridgeway passes through the Wansdyke, 1s an original feature. The dyke eastwards from Shaw House to New Buildings differs both in scale and construc- tion from the sector across the chalk downs. This part of the dyke was constructed through wooded country ; it seems to have, in part, respected earlier estate boundaries and, at New Buildings, to have lain across disused agricultural land. It is possibly of later date than the main construction. The date of the Wansdyke cannot be put closer than a strong probability of its lying between A.D. 450 to 500 or 550 to 600. The suggestion is made, however, that the hypothesis which best fits the available evidence is that the Wansdyke, in both Somerset and Wiltshire, JSunctioned as a single defensive frontier between the dates, recorded in the Anglo-Saxon Chronicle, of A.D. 577 and 592. It is suggested that it was erected by the Cerdicings against Ceawlin, who was most probably a prince of the Thames Valley Saxons. It 1s possible that the section of the Wansdyke from Morgan’s Hill eastwards to Shaw House was of earlier origin. The view of the Somerset Wansdyke as an early 7th century political frontier seems no longer tenable. INTRODUCTION THIS STUDY IS CONCERNED with certain aspects of the Wansdyke only and does not attempt to embrace all earlier work. Certain vexed questions, such as whether the earthworks east of Savernake Forest form part of Wansdyke, have not been reconsidered here. The following works are basic to the study of the Wansdyke: Pitt-Rivers, A., Excavations in Bokerly and Wansdyke (1892). Fox, C. and A., Wansdyke Reconsidered, Archaeological Journal, 115 (1960), 1-48. Myres, J. N. L., Wansdyke and the Origin of Wessex, in Roper, H. R. T-. (ed.), Essays in British History (1964), 1-27. Tratman, E. K., The Iron Age Defences and Wansdyke, in Rahtz, P. A., Barton, K. J., and Tratman, E. K., Maes Knoll Camp, Dundry, Somerset, Proceedings of the University of Bristol Spelaeological Soctety, 10 (1962-1965), 11-15. 129 PART ONE THE EXCAVATION I. RED SHORE (FIG. 1) 1a. Section I In 1966 a partial transverse bank section was cut (FIG. 2) in steps in the eroded side of a trench cut through the Wansdyke by the American Army in 1942 or 1943. The section revealed a layer of redeposited topsoil high in the bank (hereafter known as the ‘mull layer’, a designation referring to its characteristics as a soil) and, at the time of excavation, a pos- sible palisade trench was suspected following the line of the bank and cutting the southern end of the mull layer. Further study has indicated that this interpretation was erroneous. The layer of topsoil (mull) was then interpreted as evidence for a second phase of construc- tion, being possibly the result of ditch cleaning.* The mull produced a samian sherd of form 35/6 and a piece of iron slag. The Red Shore pollen analysis derives from samples taken by Professor G. W. Dimbleby from the buried soil profile in this section. The pollen analysis is described and discussed in Parts Two and Three. WANSDYKE Red Shore Excavations 1966 to 1970 N d s% / BELL BARROW a iH? RED SHORE “(x Section Il I Section | ! 0) 100 Metres ed 100 Yards Ditch~ ! nee eee Bank 1 Main Section Mu, Probable Celtic Field Lynchets Fic. I Location of excavations. The bank structure at this point is almost identical to that in the main section (FIG. 4) and only points of difference will be noted here. These include the mull layer, which erosion has almost completely removed from the main section, and a natural gully filled with clean grey soil in the clay-with-flints subsoil. A similar feature (but filled “with brown soil and flints) was found below the counterscarp and is shown on FIG. 5. 1b. Section IT A trench 3-60 m. by 0:90 m. was cut on the crest of the bank east of the 1966 section. The bank was here well preserved and it is quite clear from the stratification (FIG. 3) that 130 POSITION OF SECTION J Red Shore 1966 Section | natural gully Fic. 2 the palisade trench suspected in 1966 does not exist. The symbols used on Fic. 3 are explain- ed in Fics. 4 and 5. The top of the bank at this point consists principally of clay-with-flints, or clay-with-flints mixed with chalk and presumably derived from the junction of the natural clay-with-flints with the Upper Chalk. There is also one large tip of chalk with a certain amount of humus admixture. 1c. The Main Section (FIGs. 4-6) (i) Subsoil This consists of clay-with-flints approximately 60 cms. thick overlying Upper Chalk. Resting on the clay-with-flints is a layer of flints stained black, probably by manganese. This layer, which seems to be a natural feature, is normally no more than 8 cms. thick as under the counterscarp bank (FIG. 5). However, its thickness may have been modified by prehistoric and early Romano-British agriculture.’ Lom POSITION OF SECTION II Section Fic. 3 (i) The Buried Soil The buried soil (layer g) is of great interest and is discussed further in Parts Two and Three. It consists of a stone-free zone varying in thickness from approximately 5 to 8 cms. Where the bank material sealing this layer consisted of redeposited clay-with-flints, layer 9 was blue-grey in colour on first exposure and slowly became brown on exposure to the air; its texture in the latter instance was of ‘clay’. Where the overburden was of chalk (as at the tail of the bank), the buried soil was brown in colour and crumbly in texture. Below this layer (but only where the overlying bank consisted of clay-with-flints) was a thin red iron pan line and below this a thin black line resulting from humus accumulation: both these lines were c. 1 cm. thick. Below the red and black lines and resting directly on the subsoil was the layer of black-stained flints discussed above and here designated layer 2. The thickness of this layer varied from 3 cms. to 30 cms. It was full of charcoal fragments and very small and weathered sherds of pottery. The only identifiable sherds are of Ist century A.D. Savernake ware (identified by Mr. F. K. Annable). This layer is very thin under the tail of the bank but becomes extremely thick on the lip of the ditch: it seems to be piling up against a lynchet downhill, which must have been destroyed by the digging of the ditch. It is of interest in this context that a similar thick layer (27 cms. thick, con- taining much charcoal but no pottery) was revealed in a soil profile trench, cut for the purpose of pollen sampling, in a bell barrow at SU 11736488 just over 100 m. north-east of the main section.* (iii) The Bank (FIG. 4) The front, rear and upper layers of the bank were excavated: the central area has been reconstructed on the basis of three auger holes marked A, B and C on Fic. 4. The bank was of simple dump construction being 9-50 m. wide with a height of just under 2 m. No evidence of palisading or revetment was discovered and there was no evidence for the earlier existence of more than a small berm between bank and ditch. The core of the bank was formed of clay-with-flints with a few tips of flints derived from layer 2. It is of interest that the turf which must have been stripped from the surface of the ditch did not form the core of the bank but was added only late in its construction as the mull layer. It must have been placed temporarily to the rear of the bank since, otherwise, it would have been necessary to carry it over the newly dug ditch. The same procedure seems to have taken place at Brown’s Barn,> although here the turf was added at an earlier stage in the 132 Mull Layer WANSDYKE Z Gf, or wri ruts Z (1M BANK), sees BLACK PAN LINE RED SHORE vite WED PAN LINE DQ chan Bank g OTHER SYMBOLS SHOWN ON DITCH SECTION Tate 3 DAV AAG Auger Holes A, 8,C Natural Clay-with— Finis Fia. 4 (face page 132 ceunterscarp notural cloy-with -flints chalk 0 notural clay-with-flints 5 p B ‘\* Loose som WANSDYKE Ss. Jee cholk See re eer \ ee as} Yi z— weathered chalk 40.0 SMALL FLINTS RED SHORE LE iste «fh 900 if Soi02 SMALL CHALK ee 44% compact soit é the aS FINE EARTHY BAND Ditch 0 1 2 3 4 5 6 7 8 9 10 Feet 0 1 2 3 Metres Ric. 6 Uface page 133 construction. We may note an earlier (Iron Age) parallel to this practice at the hill fort of Ladle Hill, Hampshire.® The purpose of reserving the turf for later use (at Brown’s Barn) seems to have been the stabilizing of a large bank of loose chalk. The same purpose would not apply at Red Shore but it is quite possible that the turf was stripped along the whole course of the dyke before construction began or that an earlier procedure was being mechanically followed. Tips of clay-with-flints high in the bank (Fics. 3 and 4) indicate that the ditch was not dug vertically, with possibly a long stretch excavated simultaneously. It seems rather that the ditch was dug to a sloping or stepped end face and that the bank material was derived both horizontally from the section of ditch immediately in front of it and from the section a little further on where excavation of the upper layers was still! taking place. It is worth noting that this procedure could mean that any original crossing through the Wansdyke would be set obliquely, since the ditch would have to be dug (albeit not to its full depth) to a point slightly beyond the end of the bank to maintain its full height. An original crossing might then be set obliquely with the ditch, on one side of it only, becoming progressively shallower close to the causeway. It may be significant (although only excavation can give a final answer) that the crossing where the Ridgeway passes through the dyke at Red Shore itself is of precisely this form. Furthermore, if this hypothesis be correct, this would indicate west to east progression in construction. The red and black pan lines found below the buried soil (layer 9) were present also on the base of the inverted turf (shown in section on Fic. 4) dropped on the old land surface at the front of the bank (presumably whilst on its way, as part of a load, for temporary stacking at the rear of the future bank site). The presence of these lines at the base of the turf, in the same sequence as in the buried soil profile, supports the hypothesis (see below) that they formed in the buried soil at some time prior to the dyke’s construction. Since the turfis inverted (an inference based on the sequence in the buried soil), the vertical sequence of the lines is black/red and not red/black as at the junction of layers 2 and g. The same black/red sequence is found slightly higher in the front of the bank (Fic. 4) where it must have formed in the bank material after deposition. This line presumably formed at an interface of some kind, possibly one resulting from trampling. The significance of the pan lines will be discussed further in Parts Two and Three. There has been humus penetration in the upper layers of the bank, a phenomenon also observed in the bank of the causewayed camp on Knap Hill 1-25 kms. south-south-east of Red Shore.’ Slip from the bank extends into the field to the south. (iv) The Ditch The following numbered layers appear in the section (FIG. 6) : 1. Modern humus. 2. Black-stained flints in brown soil (probably a natural feature derived from the subsoil). . Rainwash, consisting of clean loose soil with small chalk pellets. Large flint nodules with a relatively small amount of humus. Clean compact soil. Possibly a soil developing on the secondary silt. . Identical to layer 4. . Primary chalk silt consisting of relatively small angular pieces of chalk weathered from the south side of the ditch. There is a band of fine earth crumbs at its base, another in the centre and a third resting on top of the layer. 8. Clean red clay derived from the clay-with-flints subsoil. The ditch was always V-shaped and at its centre was originally 3-90 m. deep. The depth of silt in the centre is only 1-50 m. The southern face of the ditch was probably steeper originally and chalk silt from it forms the primary infilling (layer 7). This consists of thick coarse winter silt bands and fine earthy summer silt bands comparable to those revealed in the ditch of the experimental earthwork on Overton Down.’ From the number of such bands we may deduce that the primary silt is unlikely to have taken much more £33 than three years to form. The fact that a band of summer silt lies directly on the ditch floor suggests that construction, of this sector at any rate, took place in spring or summer. Erosion does not seem to have altered significantly the northern ditch face. Its surface, which slopes gently in comparison to the steep southern face, is weathered and seems to have been long exposed but it has not contributed to the primary silt at all. The primary silt from the north face consists only of a collapsed block of clay-with-flints from the upper ditch edge (layer 8). Deposits of rainwashed material in the ditch (layer 3) contribute to the secondary silt, but of especial interest are layers 4 and 6 which consist of flint nodules with humus admixture. These layers seem to derive from the northern side of the ditch and are best interpreted as flints encountered during ploughing of land north of the dyke and thrown into the ditch as a result of field clearance. If this interpretation is correct, it should be noted how close layer 6 lies to the primary silt. In consequence, it is suggested that the divisive and disruptive effects of the Wansdyke (if such there were) can only have been of a few years’ duration; a possible minimum of five years is suggested. (v) The Counterscarp (FIG. 5) The counterscarp bank survives today to the height of only a few centimetres and has a maximum width (probably spread by ploughing) of 11 metres. Excavation revealed the old soil profile (layers 9 and 2) exactly as found under the bank (but without the pan lines) and with pottery fragments in the flinty layer 2. The section shows quite clearly (as can also be observed when the field is freshly ploughed) that the part of the counterscarp nearest to the ditch must originally have consisted of clay-with-flints (although now humi- fied) and that the part furthest from the ditch consists purely of chalk. The interpretation must be that the counterscarp is a primary feature of the Wansdyke (and not, for example, the result of ditch cleaning): it was constructed, successively, of the immediate subsoil (clay-with-flints) followed by chalk as the ditch deepened. The preservation of the buried soil under so shallow a layer is of some interest, as is also the preservation of the buried soil under the modern ploughsoil on the southern edge of the main bank, as shown in FIG. 4. The significance of this will be discussed in Part Three. id. Exposures at SU 11356468 Near the probably Iron Age earthwork known as Eald Burgh, just over 400 metres west-south-west of Red Shore, erosion of the northern face of the bank has clearly revealed the mull layer and the buried soil in longitudinal section. The mull layer (high in the bank as at Red Shore) is here about 23 cms. thick. The buried soil consists of a stone-free zone (equivalent to layer g) 7-5 cms. thick lying directly on top of a thin flinty band (equivalent to layer 2) 6 cms. thick. The subsoil consists of a layer of clay-with-flints 15 cms. thick on Upper Chalk. 1e. Resistivity Surveys Resistivity surveying by A. J. Clark indicated that the causeway at SU 113647 Is quite definitely a secondary feature. Resistivity survey at Red Shore produced inconclusive results, which is to say that it gave no reason to suppose that the causeway at Red Shore is not original: however, the presence of a very thick layer of clay-with-flints on the chalk at Red Shore introduces an element of doubt as to whether a negative result is necessarily significant. At the causeway site at SU 113647 there is a layer of only 15 cms. of clay-with- flints on chalk. However, it should be said that the evidence of resistivity survey combined with the evidence for procedures of construction discussed above do independently suggest that the Red Shore causeway is an original feature. 2. NEW BUILDINGS (FIG. 7) The excavation took place at SU 19306648 close to the present end of Wansdyke near Savernake Forest. 134 Sl ST far modern pipe trench 3 Metres 10 Feet COUNTERSCARP surface counterscarp 10 Metres 3 Feer WANSDYKE NEW BUILDINGS 1967 Fic. 7 Partial bank section. A trench 1:50 m. by 1-00 m. was cut into the face of the dyke on the west side of a gap of uncertain age through the bank. The partial transverse section, whose northern edge was 3 metres from the apparent junction of bank with ditch (now completely filled in by ploughing), revealed the following stratigraphy (Fic. 7) : . Red clay derived from the subsoil (layer 6). . Redeposited topsoil (derived from layer 5). . A band (1 to 2°5 cms. thick) of red clay streaked with white or yellow in parts. . A thin black band approximately 2-5 mm. thick. . The buried soil profile (of pH 6-0). Its top 10 cms. shows evidence of bleaching. . Red clay subsoil. Dub o nd The partial section (dug primarily for the purpose of obtaining pollen samples) shows that the bank at this point has little in common with the Red Shore section, in that it shows no evidence of topsoil being reserved for later use in construction. However, a clay bank would need no such stabilizing layer. The surface of the buried soil, on Professor Dimbleby’s interpretation, lies c. 2 cms. below the base of layer 4. Layers 3 and 4 would appear to be similar to the red iron pan line with humus accumulation below at Red Shore. At New Buildings, the pan lines have formed at the interface of the buried soil (possibly on top of a thin layer of earth trampled during the first stages of construction) with the base of the bank. Of interest were lumps of multicoloured clay (red and white or red and yellow) contained within layer 1 and shown hatched in Fic. 7. My original hypothesis interpreted layer 3 as the (probably trampled) buried soil surface from which these lumps were derived. This hypothesis is not tenable in the light of the pollen analysis and the colours must relate to changes which have taken place since construction. The processes of oxidation and reduc- tion would seem to be involved but the precise mechanics of the colour change are not easy to explain. 3. FINDS The finds derive from Section I and from the main section. Mull layer (Section I): (a) Samian rim sherd with barbotine leaf form 35/6. Kevin T. Greene, of University College, Cardiff, has examined the sherd and comments that it is South Gaulish in origin and probably of Flavian date (i.e. c. 80-100 a.p.). The diameter of the vessel from which the sherd is derived is 10 cms. (b) Fragment of iron slag. Layer 2 (Main Section): (a) Three sherds identified by F. K. Annable as being, with little doubt, a product of the Savernake Forest or Broomsgrove kiln sites. He comments as follows: “Three body sherds. Fabric somewhat sandy, containing heavy concentration of 135 Fic. 8 Wansdyke. Iron penannular brooch. Scale, 1:1. minute crushed chalk, flint and probable pottery inclusions. Pale grey core and surface. Sherds probably from same vessel, a wide-mouthed bowl. The Savernake products have been dated to the late Ist century A.D.’ (b) 41 fragments (total weight 53 grams) of pottery, all body sherds except for one possible everted rim which, if correctly identified, would belong to a Roman cooking jar. The pottery fragments are of variable colour but the fabric is otherwise identical: very crumbly with minute flint inclusions. Fragments of pottery of this type come also from: Layer 2 (Section I): Four sherds (weight 2-5 grams) ; Mull layer (Main Section): Nine sherds (weight 9 grams). Layer 3 (Main Section: on slope into ditch): Seven sherds (weight 14 grams). These sherds are probably derived by weathering from the mull layer or from layer 2. Layer 2 (Main Section: below counterscarp): Two small sherds (weight 2 grams). This group of pottery with crumbly, flint-gritted fabric is far from easy to date. Important negative evidence for a date later than the Early Bronze Age is afforded by the soil-profile trench excavated in the bell-barrow nearby. Here the excavation of 1-50 cubic metres of pre-barrow soil produced no pottery of any kind, but the excavation of just one cubic metre of layer 2 from below the bank in the main section produced 41 fragments of this pottery. F. K. Annable has examined the pottery and comments that he believes it to be Roman, although he is far from certain. The final find (ric. 8) is that of a probable iron penannular brooch (corrosion leaves some doubt about the identification) from the topsoil on the southern tail of the bank, 1°20 m. north of the junction of the bank with the modern field to the south, from the main section. The type has a long life from the Iron Age onwards and cannot be closely dated. I am indebted to Leslie Alcock for discussing this find with me. PART TWO POLLEN ANALYSIS by G. W. DIMBLEBY Serial samples were taken through the buried soil at Red Shore and at New Buildings. Moderate frequencies of pollen were found at the former site, but the profile at New Build- ings was much less rich. Nevertheless, it was possible to construct a pollen diagram for both sites (FIG. 9). Red Shore A series of samples was taken from the red clay subsoil up through the buried soil and into the base of the overlying bank. The top of the buried soil was a stone-free layer about 3 in. thick, clearly the result of earthworm sorting. Subsequently this layer had become somewhat bleached, and iron and 136 ‘sBurpying MoNy pue o10Yg poy ye oyApsueAA “WreIsvIp UdT[Og 6 ‘OL Ud{]Og penop % wei d sureig os of Ol ry g zi Ql pe SONI TING MIN (I ( esapay x hoy 9 | wa}}og Je19, CO” wean d sureay os oc o1! @ or ezzec’ JHOHS O44 sn6Bey x ent x G) > d ~ 2D er a0” ‘ SOR PHP? GX Pwo? Oe Ree Cy i 2 SEES rh Seve L Sore? > + a Sh oS S ce) b ¢ a5 @ Pe" wf ae < RS e < 3 < 2 we ew » wa & oY’ 92s ? 6 A eS » < ~f CN 2 2 y ¥ » \ ) \ @° @ oe » & } 2 C r) »Y 6 & > P02 2 ee os JYAQGSNVM 137 humus accumulation was visible at its base, indicating a change of soil genesis towards a thin iron-pan soil. Pollen was present in countable amounts throughout this stone-free zone, but below this it was present, if at all, as the merest trace. Pollen frequency was low in the lowest sample of the overlying bank, but immediately beneath this, in the surface of the buried soil, the frequency was 15 times greater, clearly indicating that this was the true surface of the soil. The pattern of distribution showed a steep falling off of total frequency down the profile, in a manner characteristic of a soil in which soil faunal mixing is not taking place. This is to be expected in a soil of the thin iron-pan type, and confirms the change in the nature of the soil inferred from the profile. It is also apparent from the diagram that the only pollen types present in abundance—grasses, Liguliflorae, plantain and bracken, all show the same decrease in frequency with depth; there are no groups whose curves might indicate greater age, and woody species are, in fact, noticeably lacking. Again this would accord with a change in the soil, from a relatively base-rich soil in which pollen would not be preserved, to one of lower pH, at least at the surface, which permits pollen preservation but at the same time would eliminate the mixing species of earthworm. If we are correct in regarding this pollen record as fairly coeval, it should therefore give us a reliable picture of the vegetation at the time of construction of the embankment. Grasses, ribwort plantain and Liguliflorae are most abundant, clearly indicating grassland conditions. ‘There was probably some bracken in the vicinity, but this was not dominant on the site itself. There seems little doubt that the land-use was pastoral. Apart from the high grass and plantain percentages, there is consistent occurrence of devil’s-bit scabious (Succisa), buttercup (Ranunculaceae) and black knapweed (Centaurea nigra), all plants of rough pasture. On the other hand, there is no record of cereal pollen and weeds of arable land (e.g. Rumex) are virtually absent. The entire analysis suggests open pastureland with no wooded country nor arable land near enough to influence the pollen rain at this point. New Buildings A similar sampling pattern was adopted here, together with one or two spot samples to answer specific questions. In particular, a sample was taken from a thin clay band at what was interpreted as the surface of the buried soil (38 cm.). This sample proved to be very poor in pollen and could not therefore have been at an exposed surface. The sample beneath this (37-35 cm.) was also rather poor in pollen and is best interpreted as part of the make-up of the embankment (a on the diagram). Below 12 cm. the samples did not contain countable amounts of pollen, but between 12 and 35 cm. there was just enough pollen to make counting possible. On the face of it, therefore, this zone could reasonably be regarded as the A horizon of a buried soil. This soil does not show so much evidence of acidification as that at Red Shore, and this is probably why the pollen frequencies are very much lower. On the other hand, though some slight development of a stone-free zone has taken place at the top of the buried soil, this is nothing like so well developed as at Red Shore and might suggest that undisturbed conditions have not prevailed here for very long. The pollen diagram at New Buildings is much more complex than the one from Red Shore. In the pollen-rich zone (a relative term) there are clearly three distinct layers, indicated b, c and don the diagram. There is no pattern which can be recognized as charac- teristic of an undisturbed soil profile; indeed, the indications are that at the point of sampling the A horizon has been worked by artificial agency. Briefly, the three zones have the following pollen characteristics. The lowest, d, is dominated by grasses, but has also high values for hazel, oak and ribwort plantain. It contrasts with 6 above, which has a reduced dominance of grass pollen and high values of fern spores (Dryopteris type). In this layer, too, the representation of oak and hazel remains high. Grasses are again more abundant, but the fern spores do not reach the high values they did in c. 138 139 i { ry depanng VS ates eeee S2IYVONNOG i HSI8¥d ol dIHSNOILV139 JNAGSNVM ———————— ee ne It is a moot point whether layer b is to be regarded as a true soil surface. In a soil such as this, in which earthworms were probably still active at the time of burial, one might not find a pollen-rich surface sample which would identify the surface. There are, however, certain groups of pollen which occur in this zone and not significantly below. For instance, cereal pollen occurs in each of the three samples of this zone, Rumex is represented, and birch shows a marked increase at the top of this layer. On balance, I would regard this as the surface on which the embankment was built, but it was probably not a surface of long standing. Comparing the pollen analysis with those from Red Shore, two major differences stand out. Firstly the representation of woodland is high at New Buildings and was virtually nil at Red Shore. This is not to say that the New Buildings site was wooded; it was open, but there was abundant woodland nearby. Secondly the character of the two sites them- selves seems to have been different. Whereas the Red Shore site was quite clearly pasture, the characteristic plants of pasture are much less conspicuous at New Buildings. There is some evidence of arable farming here, however, which was lacking at Red Shore. Indeed, one has seriously to ask whether the soil heterogeneity which the pollen diagram shows could not in fact have been due to ploughing, though if this were so, I would suspect that there had been only one episode of cultivation and not prolonged tillage. One reason, amongst others, for saying this is that cereal and Rumex pollen has not been worked into the lower layers. The evidence of the spread of birch and perhaps of hazel and bracken immedi- ately before the earthwork was constructed might suggest that cultivation was soon aban- doned, and this phase may have been enough to provide the settled conditions necessary for the development of the slight stone-free zone which could be distinguished in the section. The colonization by woody species is of course to be expected where there is woodland nearby. SNAIL ANALYSIS A soil sample from the mull layer excavated in 1966 (Section I) was submitted to Dr. J. G. Evans for analysis for mollusca. He records the following details of the sample: pH 7-4-7°8. Munsell colour: 10YR 3/2, very dark greyish brown. Texture: clay loam; a few pieces of angular flint, chalk lumps; non-calcareous. A total of six complete snails was found. Dr. Evans comments that he suspects ‘most, if not all of the mollusca to be modern intrusions through worm or root holes’. Further samples have been analysed, both by Dr. Evans and the writer, from the buried soil below the southern edge of the bank in the main section at Red Shore and from both the buried soil and the mull layers in the exposures at SU 11356468 near Eald Burgh. These samples produced no mollusca. The buried soil at New Buildings likewise contained no mollusca. PART THREE 1. THE ENVIRONMENTAL EVIDENCE ta. Red Shore The interpretation of the pollen analysis as showing very open pasture land at Red Shore is not in doubt; nor does the identification of the surface of the buried soil present any problem. A question which does remain open, however, is whether the soil-type had undergone a change from ‘relatively base rich’ to one of lower pH before the Wansdyke’s construction. The soil colour of the buried soil cannot be significant for it changed from a blue-grey to a dark brown some time after exposure. 140 Professor Dimbleby has commented (in litt.) that layer 9, where sealed by clay, appeared gleyed (i.e. reduced) and that it is in this condition that ferrous iron will move most freely in solution. The problem is one of whether the iron and humus accumulation between layers 9 and 2 took place before or after the construction of the bank. Following a suggestion of mine that the pan lines formed after construction in the anaerobic conditions induced by the thick core of clay-with-flints in the bank, Dimbleby has suggested (in litt.) that possibly ‘the calcareous environment under the chalk (part of the bank) has led to an increase of biological activity, so producing a brown soil and redistributing the iron in the profile’. It should be emphasized that the only variation lies with the bank material and not with the parent material below. However, there is clearer evidence for a pre-construction soil change than that of the ambiguous pan lines. Dimbleby has further pointed out that, in contrast to the strong archaeological evidence for pre-dyke cultivation (certainly taking place in the late 1st century A.D.), there is no evidence of agricultural activity from the stratification of the pollen or from the content of the pollen record. Dimbleby stresses that the preserved pollen has started to accumulate as the soil became more acid, presumably as pasture took over from arable. Another point of some interest is the preservation of the buried soil profiles below the counterscarp bank and under the tail of the main bank on its southern side. Ancient soils preserved under such a shallow overburden are rare and Professor Atkinson has dealt at some length with their destruction by earthworms.’ In this instance it may be possible to refer their preservation to the absence of the mixing species of earthworm on an acidic subsoil. One may, however, refer to the remarkable preservation of the Neolithic land surface almost directly below the modern plough- soil on the site of the Lambourn long barrow, Berkshire, located on pure chalk.”° 1b. New Buildings The interpretation of the New Buildings section is dependent on the pollen analysis, which is quite consistent with the archaeological evidence. The pollen analysis attests the ancient presence of Savernake Forest (evidenced earlier by the location within the forest of Roman pottery kilns which would have needed a plenti- ful supply of timber). The Wansdyke appears to have been built across derelict land at this point. This might explain its straightness near New Buildings compared with its right-angle bend in West Woods to follow what was probably a pre-existing estate boundary. 2. ESTATE BOUNDARIES One of the remarkable features of the Wansdyke is the way in which modern parish boundaries bear almost no relation to it (FIG. 10). Desmond Bonney has pointed out" that parish boundaries may have very early origins and might even be following the lines of estate boundaries as early as the 5th century A.D. (if not earlier). The behaviour of the Wansdyke does strongly suggest that it postdates the boundaries which it crosses: furthermore, the effect of the Wansdyke as a physical obstacle cannot have been very long-lived or it is likely that these boundaries would have re-established themselves on different lines. This interpretation is completely I41 consistent with the archaeological inference as to the transitory importance of the Wansdyke, based on a study of the ditch silts. The pollen analysis lends its support also, for the Wansdyke was clearly not constructed across a derelict no-man’s-land at Red Shore but across open pastured downland. Desmond Bonney (im litt., 1968) makes the following comment: ‘The non- coincidence of Wansdyke with the boundaries of land-units, which I believe could be very early and which were subsequently fossilized as parishes, certainly prompts the view that Wansdyke is later—otherwise why is it not made use of when the boundaries are so near it in some places? ‘But the argument is strengthened, to my mind, by the behaviour of the parish boundaries in relation to the Roman road between Mildenhall and Bath. The road coincides with parish boundaries all the way from a point on the Avebury/Bishops Cannings boundary to the west Wiltshire border near Bath except for some ? mile near Chittoe where the parish boundary diverges northwards (? because of the presence of Verlucio and lingering territorial rights appendant to it). There can be no question which is the earlier—the Roman road must pre-date the boundaries which follow its very straight line, though by how much is a nice question. It is significant, I think, that where Wansdyke and the Roman road meet on Morgan’s Hill the Bishops Cannings/Calne parish boundary follows the road and not the dyke. The former is not a particularly well-marked feature (at least, compared to Wansdyke), though I presume it has remained in fairly constant use since Roman times, and I would scarcely have expected the parish (or earlier) boundary to have followed it in preference to Wansdyke if the latter had also been in existence when the boundary was first laid out.’ The fact that the Wansdyke east of Shaw House does respect what may be a pre-existing estate boundary on the edge of West Woods for a distance of about 1 kilometre together with the evidence of its reduction in size suggests that here it was not primarily a military obstacle but rather a political boundary. The evidence of pollen analysis at New Buildings suggests that its straighter stretches lie across land which had fallen derelict or was at any rate no longer used for agriculture. However, it should also be pointed out that the smaller scale of the eastern end of the dyke may equally be related to the environment, since a large dyke is not likely to have been a military necessity in a woodland area. It is of interest that the Wansdyke is of larger size on the eastern side of Clatford Bottom (SU 165662) where the immediate subsoil reverts from clay to chalk. It is of some importance that the Somerset Wansdyke is likewise ignored by parish boundaries, except for a short stretch on Odd Down on the southern boundary of Bath. 3. HISTORICAL CONSIDERATIONS The following is a summary of conclusions concerning the function and struc- tural history of the East Wansdyke based on archaeological evidence, pollen analysis and the study of parish boundaries. The Wiltshire Wansdyke consists of two portions joining at (approximately) SU 135653. The Wansdyke west of this point was a military barrier. It was constructed across open pastureland without regard for then 142 existing estate boundaries. Within a few years it had lost its military purpose and land north of the dyke near Red Shore was under plough. The Wansdyke east of SU 135653 was probably constructed primarily as a territorial boundary,” although it may have had a secondary function as a military obstacle. On purely archaeological grounds the terminus post quem for the construction of the Wansdyke (based on finds in the bank in Pitt-Rivers’s section at Brown’s Barn) is some time during the ard century A.D. The best assessment of the dating of the Wansdyke is that by Dr. J. N. L. Myres.%3 It should be noted, however, that the hypothesis that the Wansdyke was of transitory significance fits in very well with the Foxes’ suggestion that it was a West Saxon construction by Ceawlin in the late 6th century.'4 Dr. Myres has offered the following comments (7m litt., 1970): ‘I am very interested to know that you have evidence to corroborate the Foxes’ idea that East Wansdyke had a very short life as an effective frontier. That certainly strengthens the case for associating it with the formation of the historic Wessex in the 6th century . . . I don’t think I have much to add to what I said there (reference 14) about the variety of situations between 550 and 600 in which such a frontier would have made temporary sense. I thought then, and still think that it may have been rash of the Foxes to have plumped for one particular possibility among this welter of alternative situations. The references in the Anglo-Saxon Chronicle relating to Ceawlin’s career are so tantalisingly inade- quate to settle the political history. As to their historicity, | am sure they come from bits of a personal saga relating his achievements as a kind of epic tragedy. As such they preserve just notes of the more dramatic incidents in a career that clearly caught the imagination of the West Saxon people as that of a really significant hero. I don’t think there is any other period than the second half of the 6th century in which the situation they indicate makes sense historically. We know quite indepen- dently from Bede that Ceawlin was the one big figure between Aelle of Sussex in the late 5th century and Aethelbert of Kent at the end of the 6th: and if we can follow Gildas over the roughly 50 years of peace that followed Mons Badonicus c. 500 (as I think we must, for he was contemporary witness for this), then there isn’t room for a conquering warlord like Ceawlin before about 550 which is just when the Anglo- Saxon Chronicle brings him on stage. One can’t, of course, be certain about the actual years given by the Anglo-Saxon Chronicle for the individual battles, for exact dates hardly ever occur in sagas. All one can say is that in the 7th century, or whenever the Ceawlin saga was first used for making annals, they were a great deal nearer to the events than we are, and could well have had sources which are now lost for getting pretty close to a correct absolute dating for the important incidents.’ At the present time, the Wansdyke may be dated only on historical grounds supplemented by the evidence of archaeology. Possible historical occasions for the construction of both the Somerset and Wiltshire Wansdykes will now be considered. 1. Late Roman The Wansdyke here finds an immediate parallel in phase B at Bokerly Dyke, Dorset, which is dated to the late 4th century.!5 It is comparable both in scale 143 and construction, although its overall length is only about four miles: it also cuts a Roman road. (It should be noted that the available evidence is insufficient to give a firm dating for the final phase of Bokerly (phase C) which may well be of late 5th or early 6th century date and may be designed to prevent a Saxon advance from the Salisbury area.)!© An essential difference lies in the fact that the Roman road cut by Bokerly phase B was reinstated, whereas the road cut by Wansdyke was not. Perhaps the crucial evidence lies in the fact that, although the Wiltshire Wansdyke is clearly a major earthwork designed to prevent attack from the north, it does not cut the Roman road from Cunetio to Sorviodunum which runs through Savernake Forest. This strongly suggests that this route had long passed out of use and was by then completely overgrown. 2. Late 5th Century In this case, the Wiltshire Wansdyke would be a British work designed to defend ‘the native rulers of Wiltshire . . . against attack from the early Saxon settlers on the Upper Thames’.!7 There is a possibility, as Myres points out, that the Wansdyke could be the work of Ambrosius Aurelianus who defeated the Saxons at Mount Badon (whose site is unknown) sometime around A.D. 500. The fact that the modern place-name of Amesbury seems to contain the name Ambrosius would support this. 3. Mid 6th Century The Anglo-Saxon Chronicle records the advance of the Cerdicings from the Salisbury area, beginning at 552 with a victory at Old Sarum and followed in 556 with a victory at Barbury Castle. It is possible that the Wansdyke marks a temporary halt in this advance. This time would make perfect sense for the construction of the dyke from Morgan’s Hill to Shaw House and the evidence from Red Shore of the Wansdyke’s transitory significance at this point would be in support of this inter- pretation. 4. Late 6th Century The Wansdyke may have been associated with the reign of Ceawlin, the second Bretwalda. Myres deals with a maze of possible circumstances of construction.!§ The large number of these possibilities arises mainly from the uncertainty as to whether Ceawlin was a Thames Valley or a Wiltshire prince. The Foxes’ hypothesis that the Wansdyke was constructed by Ceawlin is dependent upon the latter case. However, following the publication of Kirby’s penetrating study of early West Saxon history,'® we can see that this is difficult of acceptance. Kirby’s analysis leaves little doubt that Ceawlin was a prince from the Thames Valley and not from Wiltshire: Ceawlin’s genealogy, in fact, seems to be a politically-motivated invention of the later 7th century. If we accept Kirby’s thesis, the number of possible historical occasions for the construction of the Wansdyke is somewhat reduced and it is easy to find a historical context for the construction of both East and West Wansdyke, which would both 144 have been parts of a Cerdicing frontier designed for defence against Ceawlin. The most likely terminus post quem for the construction is 577, after Ceawlin’s defeat of the British principalities based on Gloucester, Cirencester and Bath. It is an important point that the topography of the Somerset Wansdyke leaves no room for doubt that the dyke must have been built by people not in control of Bath.?° A terminus ante quem for both the Somerset and Wiltshire dykes is provided by two facts: firstly, the siting of the battle of 592 at Wodnesbeorg (and the Woden name of the latter) seem to imply the prior existence of the Wiltshire Wansdyke, for which there is no known later historical context;?! secondly, the non-relation of parish boundaries to the Somerset Wansdyke implies that this frontier was as temporary as its Wiltshire counterpart and this latter fact would seem to rule out the Foxes’ suggestion of its being a frontier between Cynegils of Wessex and Penda of Mercia dating from A.D. 628. In proposing a late 6th century date for the Wansdyke, we must retain the Foxes’ suggestion of its dedication to Woden in order to explain its name. Myres*? finds it difficult to accept the view that East and West Wansdyke were not parts of a single frontier system. If we accept that they were and if we rely on known historical events as a basis for dating the earthwork, then it must follow, on the most economical hypothesis, that the Wansdyke functioned as a frontier in the late 6th century at some time between the recorded dates of 577 and 592. This does not deny the possibility that the major part of East Wansdyke was of late 5th century or middle 6th century origin. However, this part (from Morgan’s Hill to Shaw House) with its easterly extension, and with the Somerset Wansdyke, would have functioned as a late 6th century Cerdicing defence line against impending invasion by Ceawlin. The invasion eventually came via the Ridgeway to Red Shore where Ceawlin was defeated in 592. ACKNOWLEDGEMENTS The research was carried out as a private venture with financial help from the Society of Antiquaries, which I most gratefully acknowledge. My thanks are due to New College, Oxford, and to Mr. A. G. Stratton for permission to excavate at Red Shore and for help in many ways. The finds from Red Shore have been generously donated to Devizes Museum. For permission to excavate at New Buildings, my thanks are due to Mr. E. Locke. I wish to express my thanks to the following for the loan of tools and equipment: Dept. of Archaeology, University College, Cardiff; Wiltshire Archaeological Society; Rendell’s and Chivers & Sons, both of Devizes; and to Mr. Smith of Honeystreet. Individua! thanks for help in many ways go to Leslie Alcock, Professor R. J. C. Atkinson, Dr. W. H. Manning, Lady Aileen Fox, F. K. Annable, Alan Burchard, Collin Bowen and Desmond Bonney. I am most grateful to the following for help in particular ways: Dr. J. N. L. Myres (for advice on historical matters); Desmond Bonney (for his contribution to this report and for reading the report through in typescript); F. K. Annable and Kevin T. Greene (for advice on the finds) ; Professor G. W. Dimbleby and Dr. J. G. Evans (for assis- tance with palaeoenvironmental problems); and to A. J. Clark (for resistivity surveys undertaken in 1966). For assistance in the field, I must thank the many volunteers who took part, including, in particular, J. Wyn Evans, A. J. Ingham, Michael Bishop (assistant director in the 1967 season), John Dawson, and to George Drew and the members of the T45 Bury Wood Club. My thanks go also to Mr. J. E. Hancock of Bristol for permitting me to see his collection of aerial photographs of the Somerset and Wiltshire Wansdyke. My especial gratitude is due to my wife, Miranda, for considerable help both in the field and in the writing of this report. 1Jn this case the mull layer would result from the clearing of collapsed turves such as those shown lying in the ditch of the Overton Down experi- mental earthwork in Proc. Prehist. Soc. 32 (1966), plate XXIII, nos. 2-3, between pp. 336-7. 2In W.A.M. 63 (1968), 110, it was stated incor- rectly that this layer had produced cereal pollen. 3 O’Kelly, M. J., Some Soil Problems in Arch- aeological Excavation, 7. Cork Hist. Archaeol. Soc., 56 (1951), 29-44. : f ; 4The results of this excavation will form the subject of a separate report. 5 Pitt-Rivers, A., Excavations in Bokerly and Wansdyke (1892), Section II, opposite p. 262. 6 Piggott, S., Ladle Hill—an unfinished hill-fort, Antiquity, 5 (1931), 474-85. 7 Connah, G., Excavations at Knap Hill, Alton Priors, 1961, W.A.M., 60 (1965), 5. 8 Proc. Prehist. Soc., 32 (1966), 316 and fig. 1, P- 314. 9 Atkinson, R. J. C., Worms and Weathering, Antiquity, 31 (1957), 219-33. 10 Wymer, J. J., Excavations of the Lambourn Long Barrow, Berkshire Archaeol. F., 62 (1965-66), 1-16. 1 Bonney, D. J., Pagan Saxon Burials and Boun- caries in Wiltshire, W.A.M., 61 (1966), 25-30. 2 As suggested by the Foxes in Fox, C. and A., Wansdyke Reconsidered, Archaeol. F., 115 (1960), 23. 13 Myres, J. N. L., Wansdyke and the Origin of Wessex, in Roper, H.R. T-, (ed.), Essays in British History (1964), 1-27. 14 Fox, C. and A., op. cit., 39-43. 146 15 Hawkes, C. F. C., Britons, Romans and Saxons round Salisbury and in Cranborne Chase, Archaeol. F., 104 (1947), 73-6: but see also Rahtz’s reconsideration in Archaeol. F7., 118 (1961), 76-9. 16 Myres, J. N. L., op. cit., 14. 17 ibid., 16. 18 jbid., 18-26. 19 Kirby, D. P., Problems of Early West Saxon History, English Historical Review, 80 (1965), 10-29. 20 Tratman’s reconsideration of the relationship of the Wansdyke to Maes Knoll Camp, Somerset, in Proc. Univ. Bristol Spelaeol. Soc., 10 (1962-5), 11-15, is of importance since it is clear that this hill-fort was included within the area of Wansdyke. Aerial photographs, taken by Mr. J. E. Hancock of Bristol, leave little doubt that the Wansdyke cuts through the eastern defence of the camp and continues through the interior to a point where it cuts the northern defence line, approximately 65 metres west of the north-east angle of Maes Knoll Camp; it is uncertain whether the dyke follows the northern defence line of the camp to terminate in the massive Maes Knoll Tump, which cuts across the neck of the spur on which the camp lies, but it seems likely that this is the case. It is interesting to speculate whether the inclusion of two Somerset hill-forts (Maes Knoll and Stantonbury) on the line of Wansdyke is in any way related to the evidence for the post-Roman occupation of hill-forts in Somerset (for example at Cadbury Congresbury or South Cadbury Castle). 21 Myres, J. N. L., op. cit., 20. 22 Myres, J. N. L., op. cit., 7. CATHEDRAL FURNISHINGS OF THE RESTORATION PERIOD A SALISBURY INVENTORY OF 1685 by ROBERT BEDDARD IN WALTER POPE'S LIFE Of his friend and patron, Seth Ward, bishop of Salisbury from 1667 to 1689, we meet with a brief account of Salisbury cathedral. After recalling the Bishop’s part in repairing and decorating the fabric of the great church, Pope went on to praise the cathedral for its high standards of maintenance and the comeliness of its worship during his episcopate. ‘I have seen’, he wrote, ‘many Metropolitan Churches, but never any, nay, not that glorious Fabrique of St. Peters at Rome, which exceeds the imagination of all those who have not beheld it, was kept so neat as this in his time: Nay, the Sacrifice therein was as pure; there might be heard excellent Preaching, and Divine Service celebrated, with exemplary Piety, admirable Decency, and Celestial Music.’* The inventory printed below not only furnishes us with a commentary on that ‘admirable Decency’ which Ward’s biographer singled out as characteristic of worship at Salisbury cathedral, but also allows us to appreciate something of the ritual which accompanied the revived Anglican liturgy after two decades of puritan proscription. Dated 23 March 1684/5, the inventory belongs to the very beginning of James II’s reign and to the latter part of Seth Ward’s long episcopate.? It is written in a uniform hand on one piece of stiff paper, indented at the top and folded at the bottom.3 It is taken from a neglected volume of papers preserved among the Tanner MSS. in the Bodleian Library at Oxford. The volume comprises a miscellaneous collection of documents and correspondence relating to the diocese of Salisbury, and represents, in the main, the business transacted between Bishop Ward and Dean Pierce and the then archbishop of Canterbury, William Sancroft (1677—1691).4 The inventory is headed, ‘A note of the communion plate and other goods belonging to the Cathedral Church, in the keeping of the Vestry Sexton’, and endorsed, ‘Goods of the Church’. In all twenty-two items are listed, though some contain two or more individual pieces. A monetary value is attached to each item; in the case of the initial eight entries (mostly dealing with silver) two separate valuations are given, the first marked ‘as new’, and the second ‘as it is old’. This double evaluation affords us a sight of the very considerable rate of depreciation which the Caroline Dean and Chapter had to face over the years, and reminds us of the high costs of wear and tear in generations well before our own. The altar and choir of Salisbury were, as the inventory amply shows, well provided for with respect to plate and other furnishings. On both counts Salisbury compares favourably with the metropolitical cathedral of Christchurch, Canterbury, 147 for which two inventories survive from the Restoration period, one from January 1662/3 and another from November 1689.5 Although the Canterbury inventories are admittedly fuller, in that they include a great many items scattered throughout the cathedral buildings (including books as well as furniture), the goods listed as specifically belonging to the offictum sacriste are more or less the same as those entrusted to the vestry sexton at Salisbury.® In both foundations there is evidence of the con- certed effort made by pious churchmen to make good the ravages of time and the destruction of puritanism. Broadly speaking, the Salisbury inventory covers two types of goods: plate and articles of furnishing. Of the former there were two silver gilt patens, two silver flagons (certainly a pair to judge from their approximate weights), a pair of silver gilt candlesticks, one large silver bason, two covered bowls and a single ewer, besides an assortment of sconces made from base metals. Together, these pieces supplied the regular apparatus of divine worship and, more particularly, the needs of the com- munion service. It is important to note in this context, and perhaps all the more so because no mention of chalices is made in our inventory,7 that it was customary in the Restoration Church to include flagons among the sacred vessels. They were not treated merely as receptacles to keep the communion wine in, but were actually used instead of chalices. There was, of course, proper authority for this custom. The rubric governing the manual acts of the celebrant in the 1662 Book of Common Prayer is quite explicit on the matter. In the prayer of consecration the priest is directed ‘to lay his hand upon every vessel (be it Chalice or Flagon) in which there is any Wine to be consecrated’. The restoration of the cup to the laity at the time of the Refor- mation had led to the widespread use of flagons for the communion of the people.® Not surprisingly, therefore, some of the best examples of Anglican church plate to have come down to us from the past are flagons, which continued to be made up until the middle of the last century, when there was a general return to the use of chalices under the influence of the Tractarian and Ecclesiological movements. 9 Fortunately, even at this distance of time, a number of the 1685 pieces can be identified by reference to J. E. Nightingale’s standard work on the church plate of Wiltshire.t° For example, two of the items came from the Hyde family, a county family noted for its loyalty to the Stuarts. The standing patens, inscribed ‘ex dono Jacobi Hyde’, date from 1661," and were almost certainly given by Dr. James Hyde, who from 1662 to 1691 was Principal of Magdalen Hall, Oxford. The altar candle- sticks, which date from 1663 and are still one of the chief treasures of the cathedral, were the gift of Sir Robert Hyde, lord chief justice of the Common Pleas and recorder of Salisbury. He had distinguished himself in the Interregnum by sheltering the fugitive Charles II after the battle of Worcester.'+ Both brothers were linked more nearly to the cathedral by a third brother, Dr. Alexander Hyde, who, at the Res- toration, was reinstated in his dignities at Salisbury before being preferred, in turn, to the deanery of Winchester and bishopric of Salisbury. He owed his advancement to the patronage of his first cousin, Lord Chancellor Clarendon.'s The flagons are much earlier. They are dated 1610'® and were presented by Dr. John Barnston, prebendary of Bishopstone and a benefactor of Brasenose College, Oxford. Barnston, who died in 1645, also lived to be ‘outed of his spiritualities’ in the Puritan Revo- 148 lution.!7 Mercifully his flagons survived the horrors of iconoclasm. Finally, the large silver bason or alms dish dates from 1672, when the Dean and Chapter, aided by a donation of £20 from Dr. John Selleck,'® prebendary of Ruscomb, renewed ‘an old basin of ye Church’.t9 Clearly on occasions of necessity, when adequate provision by way of benefaction was unforthcoming, the ecclesiastical authorities, in common with the colleges of the ancient universities, were by no means averse to using old plate for the making of new.?° The inventory shows that the high altar and sanctuary were equally well endowed with frontals, hangings, drapes, carpets and cushions; all of them made from the choicest of available materials. In the same way that the design of church plate was affected by the evolution of domestic silver, the fabrics employed in church furnishings reflected those favoured by secular fashion—velvet, silk, ‘tishue’ and mohair,?" along with that perennially prestigious stuff, cloth of gold, the tried favourite of countless kings and innumerable prelates.27 The Restoration Church was lavish in its use of such fineries, especially when it came to setting off the sanctum sanctorum. The years following 1660 saw a rapid return to pre-war cathedral usage, a form of ceremonial beloved of Andrewes, Laud, Wren and Cosin, and one which owed much to the liturgical practices of the Chapel Royal.?3 It was, in fact, part of a far greater revival: a revival of that distinctive Anglican cult of ‘the beauty of holi- ness’ which, during the desolate period of Cromwellian rule, had been kept up by the faithful remnant of the Church of England in exile,24 and which, from 1665 onwards, was to find dramatic expression in the ecclesiastical architecture of Wren and his associates in the decorative arts.25 Once again altars were richly dressed. Besides the ‘fair white linen cloth’ (here represented by the three damask cloths) enjoined by the rubric,”° the altar was to be covered, in accordance with Canon LXXXII of 1604, with a ‘carpet of silk or other decent stuff’.27 The inventory, as one would expect, lists a ‘tishue carpett for the communion table’, that is a cover of the finest cloth, probably interwoven with gold and silver thread.2? There was also a cushion of the same, which presumably served as a rest for the Bible or service book. In addition there were gorgeous hangings for the altar, of crimson velvet decked with gold lace, and a resplendent dorsal cut from cloth of gold, which was as likely as not reserved for high days and holidays, thereby perpetuating the traditional distinction between festal and ferial usage.?9 The value of the hangings alone was put at £130 in 1685 and, when new, they were estimated to have been worth the princely sum of £170. The resulting blaze of colour at the east end of Salisbury, in splendid contrast to the sombre grey tones of the stone and marble, must have had the effect of focusing the attention of the casual visitor no less than that of the devout worshipper. Celia Fiennes, who was no great shakes as a churchwoman, if not indeed an outright dissenter, thought the appearance of the altar worth a sentence on her visit to the cathedral, circa 1684: “The Communion table hangings and the books are all of crimson velvet with gold fringe, 2 large Candlestickes gilt with great white tapers in them, a large bason to receive the offerings in . . .°3° If such sumptuousness evinced the liberality and grandiose tastes of churchmen, it also bore witness to the renewed emphasis placed on the Eucharist itself within the doctrine and spiritual economy of the restored Church.3! It was an emphasis E49 which largely came to transcend the differences of churchmanship in the period (it united the best of the old-fashioned high churchmen with the best of the new-styled latitudinarians), and one which was officially fostered by successive primates. Archbishop Sheldon’s circular letter of 4 June 1670 called upon all cathedral and collegiate churches to celebrate the Lord’s Supper not only on feast days but on Sundays too, whenever practicable. He expressly referred to the superior status of cathedral usage in the liturgical life of the Church at large. ‘Our cathedrals are’, he declared to Bishop Ward, ‘the standard and rule to all parochial churches of the solemnity and decent manner of reading the liturgy and administring the holy sacraments. 3? Later on, in the early 1680s, Archbishop Sancroft, backed by a sizeable body of supporters in the hierarchy, took steps to encourage weekly celebrations in the greater churches of the land in an attempt to increase popular reverence in the towns and cities.33 It chimed in well with the efforts of the government to secure the magistracy of the corporations into the hands of dependable Anglicans. In a sense this movement, seen from a pastoral and political point of view, was yet another facet of Sancroft’s campaign to rejuvenate the Anglican establishment.34 The presence of the Salisbury inventory among his papers suggests his fatherly oversight of a cathedral which had fallen victim to internal strife, an unhappy state of affairs which was only terminated by his instituting a metropolitical visitation in the summer of 1686.35 But to return to our exegesis of the inventory. Further scrutiny of the items reveals signs of that other central office of Anglican worship—the sermon. The Restoration Church, as has long been recognized, exhibited an enthusiasm for hearing sermons which at times amounted to a veritable mania. Even intelligent church- goers of the calibre of John Evelyn the diarist could digest prodigious quantities of preaching. The age of the later Stuarts was par excellence an age of mammoth preach- ments and of long-winded parsons, capaciously ensconced in ornate pulpits that towered above the congregation, and which were themselves capped by huge sounding- boards. Just as the altar was the appropriate centre of the sacramental rites, the pulpit was the focal point of the evangelical ministry, the preaching of the Word of God. In a ceremonious age it, too, was not without the trappings of ceremony. Salis- bury cathedral had two sets of pulpit furniture, each comprising a fall (or hanging frontal) and a matching cushion. The latter, as nowadays, was ‘placed in the middle of the pulpit’ for the convenience of the preacher; Count Lorenzo Magalotti tells us this was exactly the case at Exeter when he visited the cathedral in 1669 as part of the suite of the Grand Duke Cosmo of ‘Tuscany.3¢ At Salisbury one set was of cloth of gold; the other, scarcely of less magnificence, was of ‘wrought velvett’. There were five more cushions, including one which was embroidered with the royal arms. Finally, as a reminder of one of the regular pomps of provincial life, in which the authorities of church and state met publicly and solemnly together, the inventory records a carpet and cushion of ‘silke and silver’ which was spread before the assize judges when they attended cathedral prayers in their formalities at the opening of the assize. The arrival of the King’s judges from London was, needless to say, a major event in the calendar of any c cathedral city, occasioning a flurry of unaccus- 150 tomed activity, political and social as well as strictly judicial. After all diocesans still looked to the secular power for support in the continual battle against the forces of disobedience within their jurisdiction. When spiritual sanctions failed, obstinate dissenters, recalcitrant laity and troublesome clergy might yet heed the monitions of the King’s itinerant judges. The hierarchy took care to husband the opportunities aiforded by the circuit to invoke the backing of the Court. Salisbury and its bishop were no exception to this rule, where, after joining in worship earlier in the day, it was ‘usuall on the Assize Sunday in the evening for the judges to honour the bishop with a visit at his palace, and to accept a collation from him’.37 Then, as now, there was a tendency to socialize the working relationships of those in positions of power and responsibility. Any departure from the normal pattern was immediately commented on, as happened in August 1685, when only three of the judges responded to Bishop Ward’s invitation to supper and Lord Chief Justice Jeffreys’s abstention broke the unison of church and state. Apparently he preferred the company of the ultra-Tory Dean, the factious Dr. Pierce. To one onlooker at least Jeffreys’s conduct appeared ominous as casting an implied slight on the Bishop and giving countenance to the Dean in his untoward proceedings against his canonical superior.33 However, this episode was merely a minor distur- bance of due procedures compared with what was to come later on; for in the reign of the Roman Catholic James II even the customary attendance of the assize judges at cathedral prayers could not be relied on, still less the habitual exchange of pleasantries from and to the diocesan, given that monarch’s preference for papist and nonconformist appointments in the state. It was but another sorry aspect of James’s wilful disruption of the traditional Anglican ordering of society which was eventually to lose him the allegiance of the bulk of his subjects. In conclusion, the Salisbury inventory preserves the names of three hitherto unknown donors which we may add to those which we have already enumerated. These are Bishop Earle’s widow, a certain Captain Davy and Dr. Drake. The first of these, Bridget Earle, the relict of Seth Ward’s predecessor but one in the see, gave the set of pulpit furniture made from cloth of gold and valued at a round £50. The set was presumably presented in memory of her deceased husband, Dr. John Earle, the celebrated author of Muicrocosmographie.39 The Bishop died quite suddenly on 17 November 1665 in University College, Oxford, while attending the Parliament which had been summoned thither by Charles II in order to avoid the plague- infested capital. His body was not borne back to Salisbury but was laid to rest (with every mark of respect for the entire Court was then resident at Oxford) in Merton College Chapel, near the altar, after prayers had been read by Dr. Richard Bayly, president of St. John’s and dean of Sarum.4° The gift of the pulpit fall and cushion was a suitable means of commemorating the Bishop in his own cathedral church. Incidentally, Mrs. Earle is otherwise known to us only from her dogged efforts to get some modicum of relief from the Cromwellian Committee for Plundered Ministers,#* and from her forwarding of sums of money to the Stuart Court in exile from the loyal Cavaliers who stayed at home; neither the Dictionary of National Biography nor Anthony Wood’s compendious Athene Oxonienses mentions the Bishop as having been married.43 151 The second donor, Captain Davy, is an even more shadowy figure than Mrs. Earle. He gave the cathedral two cushions ‘wrought with silver and gold’. He belonged, it is safe to presume, to the local family of that name, one of whose mem- bers entered Middle Temple in 1637.44 Almost certainly he was the same Captain Davy who we find recorded in the churchwardens’ accounts for 1674 at St. Thomas’s parish church in the city.45 With the third donor we are on firmer ground. This was Dr. Richard Drake, chaplain-in-ordinary to Charles II, a prebendary of Sarum and rector of Wyke Regis in Dorset.4° He it was who edited and translated Lancelot Andrewes’s Preces Private and other Latin devotions.47 From 1661 till his death in 1681 Dr. Drake was also chancellor of the cathedral.48 His benefactions were altogether of the homelier sort as befitted a cleric who was involved in the daily round of worship and service in the church: a couple of surplices, the obligatory choir vestment from the re-establishment of Anglican orthodoxy onwards, and a damask tablecloth for the altar. Evidently, then as now, the cathedral was not without generous friends, both great and small. Although infinitely less spectacular than the vast riches of ornaments, vestments and furnishings that Salisbury had possessed in medieval times and before the break with Rome, there was more than an adequate supply of plate and furniture in the Restoration period, and much of it was, as we have seen, of considerable magnificence.49 The total value of the cathedral goods in 1685 was reckoned at £350, and this sum allowed for a depreciation of at least a further £150! This was no mean amount by contemporary standards. Thus it can be seen that the Salisbury inventory is more than a chance archival relic of limited antiquarian appeal; it is a social document of historical importance, for it allows us to glimpse a worshipping community, and, in doing so, to appreciate something of its religious beliefs and liturgical practices, its material prosperity and aesthetic tastes. Men, we must needs recall, reveal themselves by their spending as much as they do by their getting; and that is a basic psychological fact which no historian can afford to ignore. March 23th, 1684 A note of the communion plate and other goods belonging to the Cathedral Church, in the keeping of the Vestry Sexton. first valew present valew lit istered hieyse ed 2 silver patens gilt, weighing 30 ounces and $ 1314 6 9 9 oO 1 silver bason gilt, weighing 80 ounces and 4 36 4 6 D4. 3 0 2 silver bowles with covers gilt, weighing 52 ounces and 4 23.12 6 15 15 0 1 silver yewer, weighing 37 ounces and } Lic, 3450 9 6.3 1 silver flagon, weighing 58 ounces r7 *St,0 14 10 0 1 silver flagon, weighing 57 ounces 4 17185120 14.5 76 1 paire of silver candle sticks gilt, weighing 119 ounces and 4 53 15 6 S567 40 The velvett hangings for the alter with gold lace; the cloth of gold behind the communion table with gold fring 170.0, 0 130 0 O The cloth of tishue carpett for the communion table; a cloth of tishue cushion for the communion table 25 0 O A pulpitt cloth of gold and a cushion of the same, given by Bishop Earle’s relict 50 0 O 152 A pulpit cloth and cushion of wrought velvett 5 0 O A carpet of silke and silver, usually layd before the judges, with a cushion of the same 5 0 0 2 velvet cushions wrought with silver and gold, given by Captain Davy 210i 0 One mohair tishue cushion o 6 8 One white satten cushion with the King’s armes wrought in gold and silk on it I 0 0 One old silke cushion which wants mending 0 2 6 Three damaske table clothes, whereof one was given by Doctor Drake I 10 oO Two damaske napkins 0 3.0 One pair of brass candlesticks Of pea One large brass sconce 4.07 0 Two lesser sconces of pewter I 10" 0 Two surplices given by Doctor Drake 2 0 0 500! 351 9 II This is the valuation of each particular as it is old, but the whole valuation, as new, is above 500}. [endorsed] Goods of the Church. DEAN PIERCE’S NOTE ON THE INVENTORY In his collection of Miscellanea Dean Pierce has entered a copy of the 1685 inventory. It begins as follows: “An inventorie of the communion plate and other the choisest goods belonging to the Cathedral Church of Sarum, now in the keeping of the vestry sexton, Edward Ghillo, or his mother, the widow Ghillo, in his behalf; for which security ought to be given by the 11th statute, de rebus Ecclesie conservandis, and by 57th, de thesauro et fabrica’.5° Edward Ghillo or Gillow had succeeded his father in the offices of verger and vestry sexton; he was temporarily suspended from his places on 26 March 1685 for a combination of bad behaviour and wilful neglect of his duties, refusing to perform them ‘as his father, Tom Ghillo, had don before him’.s: At the end of his transcript Pierce also noted that ‘this inventorie was made by Mr Frome, Mr Naish and Mr George Low in the presence of the Dean and Chapter; and these values were taken by Mr Naish and the said Mr Low, goldsmith, about the 18th of October 1685’, that is over half a year after the list had been compiled.s: George Frome was a notary public and acted as ‘registrarius’ to the Chapter ;53 Thomas Naish was a lay vicar of the cathedral.s4 The items and valuations of Pierce’s copy tally exactly with those of the list which we print above from the Tanner MSS. 1The Life of the Right Reverend Father in God Seth, Lord Bishop of Salisbury, and Chancellor of the Most Noble Order of the Garter . . . Written by Dr. Walter Pope, Fellow of the Royal Society (London: Printed for William Keblewhite, at the Swan in St. Paul’s Church-yard. 1697), pp. 62-3. 2 The double dating of the year stems from the contemporary practice of reckoning the beginning of the calendar year from 25 March and not from 1 January. The date of the inventory is given as 23 March 1684, which is to say 1685. 3 Bodleian Library, Oxford: Tanner MSS. 143, f. 225. All manuscript citations refer to the Bodleian Library collections unless stated to the contrary. Abbreviations have been extended, and punctuation modernized. 4 Much of the material contained in Tanner MSS. 143 has been used in my William Sancroft, as Archbishop of Canterbury, 1677-1691 (unpublished Oxford D.Phil., dissertation), chapter iv: The Use of the Metropolitical Authority. 5 Inventories of Christchurch Canterbury, edit. J. Wickham Legg and W. H. St. John Hope (West- minster 1902), pp. 266-88. 153 6 Ibid., pp. 274-7, 278-83. ‘Sexton’ is, of course, a variant of sacristan. 7 However, the cathedral does appear to have owned a pair of early 17th century chalices, though they have come down to us in a repaired state; see J. E. Nightingale, Old Church Plate in Wilts., W.A.M., xxi (1884), p. 376. 8 J. Gilchrist, Anglican Church Plate (London, 1967), Pp. 93 ef seqq. 9 Ibid., pp. 76 et seqq. ro J. E. Nightingale, The Church Plate of the County of Wilts. (Salisbury, 1891). " [bid., p. 9. 2 For Dr. James Hyde, see J. Foster, Alumni Oxonienses: The Members of the University of Oxford, 1500-1714 (4 vols., Oxford, 1891-1892), II, 781. According to Nightingale his relations with the cathedral were of long standing, he having been baptized there in 1617 (The Church Plate of the County of Wilts., p. 9). He was also a member of the Chapter there in 1663, see Clarendon MSS. 80, 149. 3C. Oman, English Church Plate, 597-1830 (London, 1957), p- 243 and pl. 141a. Nightingale, op. cit., p. 10 and pl. 3. 14 For Lord Chief Justice Hyde, see Dictionary of National Biography, cited below as D.N.B. His death on 1 May 1665 was a loss to the Church authorities, see Clarendon MSS. 81, f. 265: Bishop Earle to Lord Chancellor Clarendon, Salisbury, 25 May [1665]. He was buried in the cathedral. 15 Dr. Hyde succeeded Bishop Earle, and was consecrated on 31 December 1665 in New College Chapel, Oxford. W. Stubbs, Registrum Sacrum Anglicanum (2nd edit., Oxford, 1897), p. 124. Tanner MSS. 45, f. 47: Bishop Henchman of London to Dean Sancroft of St. Paul’s, Wadham College, Oxford, 13 December 1665. 16 Nightingale, The Church Plate of the County of Wilts., p. 10. 17 For Dr. Barnston, see D.N.B. 18 For Dr. Selleck, see Foster, Alumni Oxonienses, IV, 1332. 19 It is inscribed ‘cum substantia honora Dominum’ around the rim, and on the back is written, ‘It weighs 804 ounces, whereof 20 /. was the gift of Dr. Selleck, Prebendary, A.p. 1672, the rest was an old basin of ye church.’ Nightingale, The Church Plate of the County of Wilts., p. 10. 20 E.g. E. A. Jones, Catalogue of the Plate of Oriel College Oxford (Oxford, 1944), p. XX. 2 These fabrics and furnishing materials were, for example, conspicuous among the lots of furniture advertised in the forthcoming sale of goods from Lord Grey’s Soho house in King’s Square, St. Giles, London, in December 1683 and January 1684. See Brit. Mus., Addit. MSS. 6308, ff. 43v, 44. 2E.g. cloth of gold was one of the items of luxury presented to Charles I] by the Russian ambassadors in January 1662. See ibid., ff. 37v, 38. 23 One of the most cogent arguments advanced against the inroads of puritanism on the ceremonies of the Church of England was that which appealed to the liturgical forms of the Chapels Royal. In one of William Sancroft’s commonplace books we find J. Hayward’s protest against the indecency of ‘Geneva tables’. These were, he argued, contrary ‘to the intent of the Church and State at the 154 Reformation’, for, he asked, ‘how is it that from that day to this altars continue as they were in the King’s and Queen[’s] housholds, and in most cathedralls too . . . places likelyest to understand the true intent and meaning of the Church and State’. Sancroft MSS. 11, p. 64. 24R. S. Bosher, The Making of the Restoration Settlement (London, 1951), p. 59- 25 Sir John Summerson, Architecture in Britain 1530 to 1830 (5th edit., London, 1969), pp. 122 et seqqg. Margaret Whinney and Oliver Millar, English Art 1625-1714 (Oxford, 1957), pp. 155 et seq. 26 This direction comes in the ‘General Rubricks of the Administration of the Lord’s Supper or Holy Communion’ set out before the Order of Holy Communion in the Book of Common Prayer. 27 The Canons reiterated much of the liturgical practice of the Elizabethan Church which had been regulated by a sequence of royal and primatial injunctions. Hayward had referred to Queen Elizabeth’s injunction appointing ‘comely and decent tables covered’ to be erected in place of demolished altars. Sancroft MSS. 11, p. 64. 28 For ‘tissue’, see the Oxford English Dictionary. 29 That the cathedral at Lincoln under Bishop Fuller kept up this ceremonial distinction with regard to its altar frontals, see J. Wickham Legg, English Church Life from the Restoration to the Trac- tarian Movement (London, 1914), pp. 136-7. 39 The Journeys of Celia Fiennes, edit. C. Morris (London, 1949), pp. 6-7. For the dating of her visit to Salisbury, see Morris’s ‘Introduction’, ibzd., p. XX. 3 J. Wickham Legg, English Church Life, pp. 21 et seqq. For an indication of the frequency of cele- brations, see also N. Sykes, From Sheldon to Secker (Cambridge, 1959), pp. 26-30. 3? Tanner MSS. 282, f. 65: transcript headed by Sancroft, ‘Archbishop Sheldon’s circular letter to cathedrals that the residentiaries should in their own persons perform Divine Service on Sundaies and Holidaies at least’. He appends a further note, ‘Endorsed, For the Right Reverend Father in God, my very good Lord and Brother, Seth, Lord Bishop of Sarum.’ Jbid., f. 65v. 33 See ‘The Epistle Dedicatory’, addressed to Archbishop Sancroft, prefaced to Dean Patrick’s A Treatise of the Necessity and Frequency of Receiving the Holy Communion . . . Begun upon Whitsunday in the Cathedral Church of Peterborough (1684), reprinted in The Works of Symon Patrick, edit. A. Taylor (9 vols., Oxford, 1858), pp. 3 et segg. Sancroft’s efforts culminated in his Articles of July 1688, especially article VI. For these, see E. Cardwell, Documentary Annals (2 vols., Oxford, 1839), p. 323. 34 See my study, The Commission for Ecclesias- tical Promotions, 1681-84: An Instrument of Tory Reaction, The Historical Journal, x (1967), 11-40. 3s The consequences of the controversy between Dean Pierce and Bishop Ward on the cathedral services were dire. The picture given by the Bishop of Bristol, Sir Jonathan Trelawny, is very different from that given us by Dr. Pope's Life of Seth Ward: ‘By reason of the Deane’s supporting the choire against the Bishop there is a scandalous neglect in the performance of the service. The day I rested in that town the singingmen refused to sing an anthem which was then desired by the Bishop’s nephew and Canon Hill; and in the afternoon the organist (which they say happens often) was absent and the prayers performed without the organ. I cannot suppose this as done to me, being a stranger to them, but wholy intended to the Bishop to whom I made my visit as being his friend.’ Bishop Trelawny to Archbishop Sancroft, Trelawne, 1 June 1686: Tanner MSS. 30, f. 51. 36 Magalotti’s description of the service at Exeter Cathedral is quoted in English Historical Documents 1660-1714, edit. A. Browning (London, 1953), pp. 417-18. 37 Tanner MSS. 143, f. 117: Dr. Robert Wood- ward, chancellor of the cathedral, to Archbishop Sancroft, Salisbury, 3 September 1685. 38 Toc. cit. Jeflreys and the others were on the infamous ‘Bloody Assize’ through the West of England, following Monmouth’s rebellion. They left Salisbury on 31 August, that is three days before Woodward wrote to Lambeth. G. W. Keeton, Lord Chancellor Jeffreys and the Stuart Cause (London, 1965), p. 320. C. Chenevix Trench, The Western Rising (London, 1969), p. 237. 39 For Bishop Earle, see D.N.B. 4° Tanner MSS. 45, f. 47: Bishop Henchman to Dean Sancroft, Oxford, 13 December 1665. The best account of the funeral ceremonies is by the Oxford historian, Anthony Wood. The Life and Times of Anthony Wood, antiquary, of Oxford, 1632- 1695, described by Himself, edit. A. Clark, Oxford Historical Society, 5 vols. (Oxford, 1891-1900), I, 62, 66. Earle had enjoyed indifferent health even before his translation to Salisbury, on which occasion he had told Clarendon, ‘if I cannot long sitt in that church, I shall be glad to lye there in the company of my frends under ground, which are as many, I beleeve, as there are above it’. Clarendon MSS. 80, f. 200: Bishop Earle to Lord Chancellor Clarendon, 15 September 1663. A former fellow of Merton, the Bishop was interred, according to Wood’s ‘ichnography of Merton Coll. choire’, to the side of the altar and the monument erected nearby. A. Bott, The Monuments in Merton College Chapel (Oxford, 1964), pp. 74-75, 120 and plan 2. I am indebted to Dr. J. R. L. Highfield, the Lib- rarian of Merton College, for this reference. 41 A. G. Matthews, Walker Revised, being a revision of Fohn Walker’s Sufferings of the Clergy during the Grand Rebellion 1642-60 (Oxford, 1948), p. 372. # I have a signed testimonial of Edward Proger, esquire of Charles II’s Bedchamber, in support of John Fotherley’s petition for recompense for monies advanced to the King in exile. One sum of ‘fifty peeces in gold’ was delivered in August 1655 to Dr. George Wilde’ later bishop of Derry), ‘which hee sent to His Majesty by Mi£stris Earles, wife to Doctor Earles then Clerke of the Closett .. .’. It is dated 6 June 1664. The petition and supporting testimonies were referred by Secretary Morrice (Whitehall, 23 December 1663) to Lord Treasurer Southampton, who made his report to the King on 4 July 1664. 4 Athene Oxonienses. An Exact History of All the Writers and Bishops who have had their Education in the University of Oxford. To which are added the Fasti, or Annals of the said University, edit. P. Bliss (4 vols., London, 1813-1820), III, 716-19. It is not sur- prising that Mrs. Earle was missed for she rarely appears in the body of any letter; she usually inhabits the complimentary postscripts of her husband’s letters, e.g. Clarendon MSS. 80, ff. 200, 214. There are, however, passing references to her in two other letters from leading royalists in exile, see Clarendon MSS. 58, f. 279v: [Secretary Nicholas to Sir Edward Hyde], Brussels, 4/14 September 1658; Clarendon MSS. 70, f. 206: Sir Charles Cotterell to Henry Hyde, Breda, 30 March 1660. 44H. A. C. Sturgess, Register of Admissions to the Honourable Society of the Middle Temple (3 vols., London, 1949), I, 133: Richard, son and heir of John Davy of Salisbury, deceased. 45H. J. F. Swayne, Churchwardens’ Accounts of S. Edmund and S. Thomas, Sarum 1443-1703 (Wilts. Record Society, Salisbury, 1896), p. 342: 20 May 1674. 46 For Dr. Drake, see J. and J. A. Venn, Alumni Cantabrigienses, part 1: from the earliest times to 1715 (4 vols., Cambridge, 1922-1927), II, 64. 47 D.N.B., see under Andrewes. 48 Tanner MSS. 143, f. 265: Dean and Chapter of Salisbury to Archbishop Sheldon, Chapter House, Salisbury, 24 October 1670. 49 Compare the inventory of jewels and goods of Salisbury made by the Treasurer, Thomas Robert- son, in 1536, printed as appendix no. IT in Night- ingale, The Church Plate of the County of Wilts., pp. 235-45; also J. E. Jackson, Wiltshire Chantry Furniture, W.A.M., xxii (1885), pp. 318-29; C. Wordsworth, Ceremonies and Processions of the Cathedral Church of Salisbury (Cambridge, 1901), pp. 160-84. s°’The volume, which is kept in the Bishop’s Registry in Wren Hall, is labelled on the spine, ‘Miscellanea collected by Dr Thomas Pierce, Dean of Salisbury, 1675-1691’. The copy occurs at p. 127. st See “The manyfold cause of Edward Ghillo’s suspension’, ibid., pp. 137-138. 52 Ibid., p. 127. 53 He witnessed the proclamation made in the Cathedral on 5 October 1684, ibid., p. 123; cf. pp. 136, 138 et seqq. 54 See ‘The act for the admission of Mr Naish’ into a vicar’s place, ibid., pp. 135-136. ae) JEFFRY HUNT PIPES by D. R. ATKINSON THE HUNT FAMILY of pipemakers are well known in the south-west of England by the numerous examples of their pipes, marked in most cases with the full name stamped on the heel, incuse. Their earliest products date from about 1640 typologically and the latest examples were produced by Thomas Hunt at Marlborough probably up to about 1710, these having a spur and the name stamped on the stem. Documentary references to the family are few, and most of those which can be confirmed originate from Bristol, where contemporary records are relatively plentiful. Bristol from the second half of the 17th century rapidly became second only to London as a pipemaking centre,' and between 1650 and 1750 developed into the leading exporter of clay pipes, the trade being mainly with the rapidly developing American continent? (this trade was largely lost to Scotland in the 19th century).3 Pritchard4 records Flower, John, Jeffry and Thomas Hunts as working at Bristol in the mid-17th century, but cannot trace the origins of the family. Later in the century there were other makers of the same names practising their trade both in the city and elsewhere, and it is obvious that sons having completed apprenticeships moved to nearby West Country towns and established the trade there. One such was the Thomas Hunt, first mentioned at Marlborough in 1667,5 and later to become a leading citizen of the town. The Thomas whose earlier style pipes occur at Bristol and elsewhere but which have not been found at Marlborough was probably his father or perhaps an uncle. We know, for instance, that there were at least three successive John Hunts who became freemen at Bristol between 1651 and 1694, one being the son of Flower and another, probably the first, a brother.® Another Hunt, William, produced pipes in the Taunton area from c. 1670-90 marked with his initials or full name.7 Among these numerous and somewhat confusing makers of this prolific family the least seems to be known of Jeffry. Pritchard, describing the plentiful examples of — his pipes found in Bristol between 1898-1906 states that he ‘probably worked there, but does not appear to have taken up his freedom’. This maker is possibly the son of the man, also Jeffry, who died in 1690 aged g1 and who was a freeholder in 1665.‘ Pritchard also tells us that another Jeffry Hunt was the grandson and ‘also took his freedom of this city’. These statements are somewhat contradictory and have led to confusion more recently. K. Marochan, who has studied the Bristol pipe industry and has been concerned in many excavations since the war involving pipes in the city, says that Jeffry Hunt certainly was not a freeman of Bristol in 1651, and neither does he appear in W. J. Pountney’s list of Bristol pipemakers compiled from the 156 BrRist OL CHA ny NEL VGELOUCESTER Bee. BRIDGWATER 1°) O TAUNTON Leta Ff CHANNEL Fic: 1 Sketch map showing places in the South-West from which Jeffry Hunt pipes have been recorded. DISTRIBUTION OF JEFFRY HUNT PIPES SO FAR RECORDED Bristol Taunton Southampton Marlborough Salisbury Devizes Bath St. Ebbs, Oxford Blandford Bridgwater Shaftesbury Calne Gloucester Westbury Malmesbury Nunney Castle Bratton Wookey Hole Downside Abbey Early types, 1 and 2, numerous Early types, 1 and 2 Early type, 1 Early types, 1 and 2, numerous Early and middle types, 1, 2 and 3 All types, 1, 2, 3, 4 and 5 Early and middle types, 1, 2 and 3 Early type, 1-2 Early type, 1-2 Early type, 1-2 Early type, 1-2 Type I Type 2 Type 2 Type 2 Type 2 and 3 Type 2 Type 3 Type 3 157 freedom rolls. On the other hand F. G. Hilton Price, describing Bristol pipes and pipemakers in 1901,9 states that he was!* The questions we are, therefore, left with, are as follows: Was the Jeffry Hunt who died in 1690 the first pipemaker of the name in the city? Was the man stated to have been (or not to have been) a freeman in 1651, his son, and was the third person of the name Pritchard mentions rather vaguely as a freeman, the grandson? If so, there may have been three Jeffry Hunts, all pipemakers, between c. 1650 and 1690. The only reference outside Bristol to a Jeffry Hunt at this period is of a memorial stated to have once been in the church at Norton St. Philip, Somerset, recording the death of Edward Hunt, son of Jeffry Hunt, in 1656. No trade is (apparently) men- tioned, so this really can be discarded as there is no proof at all that it is the same person. The same names occur frequently in 17th century records and it is easy to be misled. According, for instance, to G. Kempson of Marlborough, there were three Thomas Hunts living in the town in the late 17th century, one of whom was the pipemaker, the others being an innkeeper and a bookseller respectively.t° All occur in different contexts in contemporary borough records. A look now at the pipes attributable to Jeffry Hunts and their distribution may help to unravel the tangle. The most significant feature is the very wide area over which they occur. It is reasonable to assume that the products of most small-town makers were sold either in the place of manufacture or, at most, in the immediate vicinity. Apart from odd specimens carried by travellers, a radius of at most 10-20 miles is about the normal area of distribution, except in the few cases of makers who had a lengthy period of working and gained such a reputation that their trade contacts led to sales over a wider area, even in the face of competition by local makers. Examples of this are Richard Greenland I and ‘Thomas Hunt, whose pipes of the period c. 1650-70 and c. 1660-1700 respectively have an unusually wide distribution in the West Country. The Jeffry Hunt pipes, however, have an even wider distribution than this, as a look at the list of find-places will verify. There must be numerous other examples not recorded or known to the writer, which would make an even fuller picture. There is a wider variety of bowl types and marks, too, than is generally realized, and even now there is probably a lot more to learn. All the types do not occur in the same areas (except at Devizes), but the earliest have the widest distribution, while the latest appear so far from one area only. The inference, therefore, is of a Bristol maker (like many others) in the mid-17th century selling his pipes fairly far afield before the establishment of the industry in many country towns, later to be succeeded by a son or grandson of the same name who practised the trade, possibly even using the same marks, elsewhere. Unfortunately no West Country town has yet produced a documented reference to a Jeffry Hunt, pipemaker. It is not wise to theorize on the basis of the incidence * Since this paper was written I have heard by letter from Mr. Iain C. Walker of Ottawa, Canada, that Jeffry Hunt definitely does not appear in the F reedom Book for 1651. Mr. Walker has been studying the history of the Bristol pipemakers in great detail and his results will be published shortly in a new paper on the subject. Mr. Walker has generously ‘supplied me with all the information he has on the Hunt family at Bristol but this throws no new light, unfortunately, on the Jeflry Hunts. 158 of pipes alone, as conclusive evidence to the contrary has a habit of turning up after one has theorized in print. An example of such rashness occurred when, some years ago, I suggested that the pipes of Richard Greenland II were made at Bath, so many examples with stem marks having been found there. Subsequently it was revealed™ that the maker did in fact work at Devizes between at least 1688 and 1736, the pipes being equally common there, though not found at the time of the original publication. The last Jeffry Hunt pipes, however, which have the largest bowls of the series (type 5), stamped with the largest of all the known marks (D), have so far been found at Devizes and its immediate vicinity only, and, although I have no other evidence on which to base the supposition, it is a distinct possibility that this was indeed where the last Jeffry Hunt worked. He probably died about 1695, or ceased working at that time, for these pipes belong to the late ‘heel’ period which ended in Wiltshire about the time of the death of Charles II, when contemporary makers began to adopt the new design with the spur, larger bowl, and stamped their marks on the stem.* The following conclusions may therefore be drawn: 1. The first Jeffry Hunt, who died in 1690 aged gt, was not a pipemaker at Bristol, or if he was he did not mark any of his pipes with his full name. 2. His son produced the first pipes, with the full name-stamp, about 1650, probably at Bristol, and achieved a wide distribution. Whether or not he became a freeman of the city must remain for the present an unanswered question. 3. The grandson, who Pritchard says did become a freeman, at some later date moved away from Bristol and set up his business in the country, perhaps at Devizes, where he worked until some time in the last 20 years of the 17th century. The writer will be very pleased to receive any further information which may help to throw new light on the Jeffry Hunts. Bowl types (see FIG. 2) 1. This is the earliest type seen and is probably not previous to c. 1650. The bowl is small, usually with a line of milling round the lip, but occasionally without. The heel is out- standing, round and not pinched in below the base of the bowl. The stem is usually thinner than shown and has a wide bore. Marks A and F. 2. A distinctive bowl shape favoured by many of the Hunt and Howell families in the mid-17th century. The bowl is strikingly overhanging and always has a _ very pronounced line of milling, taller than usually found, round the lip. Outstanding round heel not pinched in below the bowl. Medium stem with a wide bore. Marks A and B. 3. A larger bowl with no striking features in its shape. Usually without milling. Thick stem and bore, c. 1660-70. Marks C and G. 4. Large bowl, unmilled, with overhanging lip parallel to the line of the stem. Front and sides of heel are pinched in below the bowl. Thick stem and bore. Type favoured by some Marlborough makers, c. 1680. For parallels see Thomas Simes and Henry Marshal (Shaftesbury) and Thomas Hunt and Edward Mells (Marlborough). Mark C. 5. The most unusual and distinctive Jeffry Hunt type and probably the last made, c. 1680. Curvaceous outward leaning bowl, unmilled. Outstanding heel, pinched in all round below the bowl. Found only at Devizes with some identical bowls of Richard Greenland II.*? Mark D. * This date may possibly be extended slightly for a stem has recently been shown to me, found in a garden at East Knowle, bearing a stamp which, though incomplete, appears to read IEF/HUNT. 159 FR Few : Yr A 8 - Pp 1 tad 1 DA G F ¢ Fic. 2 Jeffry Hunt pipes: bowl types and maker’s marks. Marks. All incuse stamped on the base Seven different marks are shown but it is appreciated that there may well be other varieties. ‘he initials I H also occur on West Country pipes and may have been used by one or more of the Jeflry Hunts. A. This is the commonest form and has a wide distribution. Found on bow! types 1 and 2. Tall letters, partly serif. B. Avery similar mark, but the ‘Y’ of Jeffry and “T” of Hunt are cut smaller than the other letters. Seen only on bowl type 2 and much less common than A. C. The letters of Jeffry are smaller than in A and there is a colon instead of a dot between the two words. Occurs on bowl types 3 and 4. D. The largest mark seen and unlike any of the others it has a dotted border. Bowl type 5 only. 160 E. An early mark similar to A but has thicker more distinctly serif letters and the ‘N’ in Hunt is reversed. Seen only on heel with thick stem and very wide bore, bowl type unknown. F, G. This is an abbreviated form and is found with or without the dots beside the H. It is also said to be found with H' for Hunt. Occurs on bow] types 1, 3 and possibly others. ™ English Clay Tobacco Pipes, by Adrian Oswald. British Archaeological Association, 1967. 2 Reference should be made to the numerous excavation reports published in America since the war, which frequently illustrate Bristol pipes. 3 The Clay Tobacco Pipe—its place in English Ceramics, by Adrian Oswald, 1969. 4 Trans. Bristol Gloucestershire Archaeol. Soc.. XLV (1923). 5 Clay Tobacco Pipes of Marlborough, by D. R. Atkinson, Wiltshire Archaeol. Natur. Hist. Mag., 60 (1965); reprinted with additions 1969. 6 Pritchard. 7 Taunton Castle Museum collection. 8 Though it is possible that these two are not the same person. 9 Clay Tobacco Pipes of the Seventeenth Century found in Bristol, by F. G. Hilton Price, Archaeol. 7., 1gOl. z ° Thomas Hunt, the pipemaker, was an educated man and owned property in Marlborough, however, so it is not impossible that he sold books and kept an inn as well as running a pipe factory. Pipemakers usually belonged to the poorest classes as Oswald has pointed out, but instances are recorded of pipe- makers who carried on other trades as well. ™ By G. Kempson of Marlborough. 1 This is a very unusual and beautifully made type of bowl and examples marked Richard Green- land and Jeffry Hunt from the same site are so identical that they appear to be from the same mould. 161 NINETEENTH-CENTURY COUNTRY CARRIERS IN NORTH WILTSHIRE by ALAN GREENING A GREAT DEAL OF STUDY has gone into the surprisingly comprehensive network of stage-coach services operating in England before the advent of the railways, and rather less into the almost equally comprehensive system of goods transport by stage-waggon. Little seems to have been done by way of research into the ordinary local village carriers, people operating in a very small way to and from local market centres; yet a study of these operators can shed interesting light both on the relative importance of local markets and also on their overlapping spheres of influence. I have gathered information relating to the carriers of North Wiltshire, with particular reference to the markets at Devizes, Marlborough and Swindon, for the years 1831, 1865 and 1899. Although a great deal of local knowledge was required by the village carrier, very little documentation was involved and it was seldom necessary for him to advertise, unless he was starting up in business or changing his route. Such advertisements are few and far between, by contrast with those of the pre-railway stage operators, whose advertisements are frequently to be found in the pages of such papers as The Salisbury and Winchester Journal. A perusal of The Devizes Advertiser for 1865, for instance, yields only one or two such references; yet on market days (Thursdays) in that year, no less than forty-one carriers worked in and out of the town. Occasional news items can, however, be helpful: one finds reports of such things as turnpike toll disputes involving carriers and their traffic. ‘Chatty’ reminiscence books by local authors are also sometimes useful, with tit-bits of information of a personal nature. But by the far the most fruitful sources of information are the directories, although their usefulness varies, depending on the reliability of their local corre- spondents. I have relied mainly upon Harrod’s Postal and Commercial Directory and Kelly’s, supplemented by purely ‘local’ directories. It is possible to cross-check entries between the ‘local’ directories and the others and wonder which of them last checked the information they published, if indeed they ever seriously checked it at all; while it is by no means unusual to find entries for different parishes in the same directory that contradict each other and village entries sometimes omit carriers listed in detail under the neighbouring market town. It has to be admitted that information about purely local carriers is sadly lacking when one turns to 1831. For this particular year I consulted Pigot & Co.'s directory. The same applies to all pre-railway directories I have consulted. They concentrate, in the main, upon the stage network, for which there is no lack of information. It may be that the type of person likely to purchase directories in, say, Regency England, had no very great interest in the local market carriers, although these must 162 have been numerous. It does, however, seem possible that the general increase in the prosperity of the countryside in mid-Victorian times did lead to a considerable increase in the numbers of this type of operator. At all events, in 1831 the local men are beneath the notice of the directory compilers to any marked degree; as late as 1848 they are only sketchily covered, but by 1865 they are recorded in great detail. Thus the data here included for 1831 (Table 1) is of little use for the purpose of direct comparison with the later dates. What it does show is a picture of a system of transport then at its peak but the death-knell of which had already sounded. Table 2, by contrast, presents a very different picture: market-day traffic at Devizes in 1865. Devizes, of course, was the premier market of the northern and central parts of Wiltshire and second only to Salisbury in the whole county. By contrast the markets of Marlborough and Chippenham were much less impressive, while Swindon’s—despite the very rapid growth of the twin towns—very small beer indeed. Nevertheless, similar patterns, with pronounced market-day ‘peaks’, can be shown for them and for any other market centres, and I have shown the incidence TABLE I CARRIERS FROM DEVIZES, MARLBOROUGH AND SWINDON, 1831 (Based on Pigot & Co.’s Wiltshire Directory, 1831) A. DEVIZES Population (1831) 4,562 Market-Day: Thursday Carrier’s Name From Destination Frequency Giddings Black Swan Bath Friday Giles and Hooper* Bell London Mon. and Thurs. Giles and Hooper* Bell Frome Thurs. and Sun. Harding King’s Arms Melksham Thursday Hillier and May Wm. Baster’s, London Wed., Thurs., Sat. St. John Street Hillier and May Wm. Baster’s, ‘Trowbridge and Thurs. and Sat. St. John Street Bradford Lapham, Thomas The Green Salisbury Monday Longstreet, Josepht Pelican Chippenham ‘Thursday Miles Black Horse, Salisbury and Mon. and Thurs. St. John Street Southampton Mitchell ‘Goes through daily’| Bath Daily Moody Black Horse Lavington(s) Thursday Powell Black Swan Salisbury Thursday Purdue Black Swan Andover Thursday Rix, Samuel Ownhouse, BathRd.| Bath Tuesday Salmon, or Selman, Richd{ | Pelican Corsham Thursday Salter Black Swan Chippenham ‘Thursday Smith, Chas. King’s Arms Oxford Thursday Smith Black Swan Westbury and Thursday Warminster Smith White Swan Bath Monday Spackman, Stephen§ Black Horse Calne Thurs. and Sat. Stratton, James Wm. Baster’s, Melksham Wed. and Sat. St. John Street * Frome Carrier + Chippenham Carrier t Corsham Carrier § Calne Carrier 163 B. MARLBOROUGH . Population (1831) 3,426 Carrier’s Name From Destination Frequency Alexander, Geo. Bell and Shoulder Swindon Saturday of Mutton, Kingsbury Street Bedford, Richd., and Bell and Shoulder Wroughton Saturday Cawley, David of Mutton Clark, Thos. Bear and Castle Wootton Bassett Saturday Clench, Chas. Bell and Shoulder Newbury Saturday of Mutton Cook, Wm. Green Dragon Pewsey Saturday Fowler, Saml. Rose and Crown Ramsbury Wed. and Sat. George, John St. Margarets Salisbury Monday Gibbs, Chas. Bear and Castle Hungerford Tues. and Sat. Hammond and Reeve | Own Warehouse, London Tues., Thurs. & Sat. Hammond and Reeve Hillier and May Hillier and May Selwood, Isaac Smart, Chas. Stroud, Jeremiah Tanner, Chas. and John* High Street bed ”° Own Warehouse, Marsh 9 be) Green Dragon Bear and Castle Bear and Castle Royal Oak Bath and Bristol London Trowbridge and Bradford (Bath, Bristol ?) Wootton Bassett Great Bedwyn Reading Chippenham Tues., Thurs. & Sat. Daily, except Sat. and Sun. Tuesday Saturday Saturday Saturday Thrice weekly ‘Besides the above, there are Waggons passing every day to and from London and the other places mentioned.’ * Chippenham Carrier. C. SWINDON Population (1831) 1,742 Carrier’s Name From Destination Frequency Alexander, Wm. _- Cirencester Monday ss es — Highworth Wednesday f », (George) —- Marlborough Saturday Basings, John —_* London Wed. and Sat. Gale, Edward Bell, High Street Malmesbury Tuesday 29 > 3 (38 > Newbury Thursday Jefferies, Robert ft Goddard Arms Wootton Bassett Wednesday Smith, Chas. — Devizes Wednesday — Oxtord Friday ” ” * William Morris, Swindon Fifty Years Ago (More or Less) (1885), 228, refers to a warehouse or ‘tallot’ and stables on the south side of the Market Square in what is now Old Swindon, where Messrs. Brown & Plummer’s Wine and Spirit Store (formerly Old Swindon Town Hall) now stands. Basings also worked from Wootton Bassett. + Jefferies was a Wootton Bassett carrier whose main route, according to Morris, was to and from Bristol. Morris’s days are at variance with the information in Pigot’s Directory; although Morris, of course, is only reminiscing at a much later date. 164. Carrier’s Name TABLE II CARRIERS AT DEVIZES MARKET (THURSDAYS), 1865 Departing from Serving Ashley Axford Bailey Beaven Brown Bush Cooke Cruse Davis Draper* Eatwell Eatwell Franklin Fay George Goddard Hawkins Hayden James Jordan Long Lucas Martin Martin Maslen Matthews Mead Miles Mizen Mortimer Pepler Sawyer Spackman (Calne) Spackman, Peter Smith Stockham Sexton Simpkins Tackle Crown White Swan Crown Crown Crown Crown Bear Pelican Crown Elm Tree Black Swan Crown Elm Tree Crown Black Swan Cross Keys Crown Black Swan Pelican Bear Black Swan Black Swan Three Crowns Black Swan Three Crowns White Bear White Swan Black Swan Black Bear Bear Lamb Black Swan Crown Crown Crown Pelican White Bear Black Swan Black Swan Orcheston, Shrewton, Berwick St. James Market Lavington, Shrewton, Salisbury Potterne, Littleton & West Lavington Little Cheverell Lavington, Shrewton & Salisbury Compton Bassett Conock, Manningford & Pewsey West Lavington & Imber Conock, Charlton, Rushall, Manningford & Wilsford Potterne, West & Market Lavingtons & Easterton Chippenham & Bath Rushall Tilshead, Shrewton & Salisbury (apparently via War- minster) Rushall, Upavon & Enford Chitterne & Heytesbury Wootton Rivers & Marlborough Upavon, Everleigh, Collingbourne & Tidworth Figheldean & Salisbury Calne (Omnibus) Swindon & Highworth West Lavington, Tilshead, Shrewton, Winterbourne Stoke, Stapleford, Wilton & Salisbury Seend, Semington, Hilperton & Trowbridge Chirton, Marden, Beechingstoke, Woodborough, Bottles- ford & Wilcot Horton, Allington, Stanton, Alton, Wilcot & Pewsey All Cannings Urchfont West Lavington, Tilshead, Shrewton, Winterbourne Stoke, Wilton & Salisbury West Lavington, Shrewton & Salisbury Melksham, Holt & Bradford Beckhampton, West Kennett, Marlborough & Swindon (Omnibus) Bratton Atworth & Melksham Rowde, Bromham, Sandy Lane, Heddington & Calne Beckhampton, Avebury, Upper Kennet villages, Wrough- ton & Swindon Orcheston, Shrewton & Berwick St. James Melksham, Lacock & Corsham Cherhill Lyneham, Hilmarton, Wootton Bassett & Swindon Netheravon, Fittleton, Enford, Amesbury & Salisbury Nearly half (17) of the Carriers listed above came from furthest distances within ten miles of Devizes; but note that a number came from as far afield as Swindon (20 miles direct, approx. 25 via Marlborough) and Salisbury (25-30 miles according to route). Twelve Carriers operated on other than market days, eleven for instance on Saturdays, the next busiest day; but on Wednesdays and Fridays only the Market Lavington Carrier, Draper, plied to and from Devizes. * On 29th June, 1865, R. Draper advertised in The Devizes Advertiser‘... that he will in future leave (Market) Lavington at 9.15 a.m., arriving at Devizes at 11, returning from the Elm Tree daily at 4 p.m.’ of market-day traffic for these three in the accompanying diagrams (Fics. 1-6). Of course, markets beyond the confines of Wiltshire could be similarly mapped, and only a daily analysis could show what effect they had on their neighbours. Perhaps the (Highworth) Swindon Swindon Swindon ae Swindon Lyneham Marlborough Bath Chippenham Avebury Geane Cherhill Compton Bassett Lacock Marlborough Corsham Wootton Rivers E Pewsey aie Atworth «<———_- ? DEVIZES All Cannings = : Holt ¢ 2 paoreee Bradford Pop. (1861) Marden Bottlesford Seend 6,700 Charlton Trowbridge wilsford Rushall Upavon Bratton Collingbournes cdington potteenc Tidworth TELAT EL LLLLT] Urchfont Upavon cnford West Lavington ie CN Little Cheverell Imber casterton Netthesaven Market | Lavington Tilshead Figheldean Chitterne Amesbury Heytesbury Shrewton ee Salisbury Berwick St. James Salisbury Fic. 1 Carriers from Devizes market (Thursdays), 1865. Each line represents one carrier. Based on J. and G. Harrod *y ) . . ~ oe ‘ and Company’s Postal and Commercial Directory of Wiltshire. 166 Lambourn Highworth South Marston Swindon Baydon Wanborough pene Wroughton Aldbourne Winterbourne Bassett Avebury Sa i pe a ieee af Mw i i : 24 § Lome a i 1 ~ “= : 7 ee fH = : a ” Rs, : = 1 “ \ — ‘ 1 ‘ Me ' r Pe 1 co INDEX TO VOLUME 66 Accessions: County Record Office, 201-02; Diocesan Record Office, 202; library, 200-01; museum, 196-200 Allington, 190-01 Ambrosius Aurelianus, 144 Anglo-Saxon: boundary, 191; brooch, 58, 178; cemeteries, 58, 179, 189; grave-goods, 189— go Ashton Keynes, 177 Atworth, 189 Barnston, Dr. John, 148 Bayly, Dr. Richard, 151 Boot pedigree, 185 Boscombe, Allington, 190-01 Bowerman, Paul, 58, 60, 193 Bronze Age: Deverel-Rimbury culture, 55; enclosures, 55, 82; knife, 82, 99; linear earthworks, 55; pottery, 47, 53, 56, 82; round barrows, 54, 76; settlements, 55, 82; spearhead, 177; sword, 196; ‘Wessex cul- ture’, 54 Brooch, iron penannular, 136 Brooks pedigree, 185 Buckland pedigree, 185 Burge pedigree, 185 Byrt pedigree, 185 Carriers in N. Wilts, 162-76 Carriers, Alexander, Geo and Wm., 164; Ashley, 165; Axford, 165; Bailey, 165; Basings, John, 164; Beaven, 165; Bedford, Rich., 164; Bristow, Sam, 171; Brown, 165; Bush, 165; Cowley, David, 164; Clark, Stephen, 171; Clark, Thos., 164; Clench, Chas., 164; Coleman, Edward, 171; Cook, Wm., 164; Cooke, 165; Cruse, 165; Davis, 165; Draper, R., 165; Dyke, Rich., 171, 173; Eatwell, 165; Fay, 165; Fowler, Sam., 164; Franklin, 165; Gale, Edward, 164; George, 164; George, Thos., 165; Gibbs, Chas., 164; Giddings, 163; Giles, 163; Goddard, 165; Hammond, Jeremiah, 164; Harding, 163; Hawkins, 165; Hayden, 165; Hillier, 163-4; Hooper, 163; James, 165; Jefferies, Robt., 164; Jordan, 165; Lapham, Thos., 163; Long, 165; Longstreet, Joseph, 163; Lucas, 165; Martin, 165; Maslen, 165; Matthews, 165; May, 163-4; Mead, 165; Miles, 163, 165; Mitchell, 163; Mizen, 165; Moody, 163; Mortimer, 165; Pepler, 165; Powell, 163; Purdue, 163; Reeve, 164; Rix, Sam., 163; Salmon (Selman), Rich., 163; Salter, 163; Sawyer, 165; Selwood, Isaac, 164; Sexton, 165; Simpkins, 165; Smart, Chas., 164; Smith, Chas., 163-4; Smith, R., 172; Spackman, Peter, 165; Spackman, Stephen, 163; Thomas (Calne), 165; Stockham, David, 165, 171; Stratton, James, 163; Stroud, Jeremiah, 164; Tackle, 165; Tanner, Chas and John, 164; Taylor, Fred, 171; Whiting, Isaac, 171 Carter pedigree, 185 Castlecombe pedigree, 185 Cathedral furnishings, 147-55 Ceawlin, 143-5 ‘Celtic’ fields, 63 Cerdicings, 144-5 Clay, R. C. C., obituary, 207 Clay pipes, 156-61 Coins: Iron Age, 197; Roman, 74, 86, 116, 177, 189-90, 198 Cow Down, Longbridge Deverill, 67-8 Cox pedigree, 185 Cricklade, 177; pedigrees in museum, 185-7 Cuss pedigree, 185 Cynegils, 145 Dartnell, G. E., 71 Davy, Captain, 151-3 de Abingdon pedigree, 185 de la Bere pedigree, 185 de la More pedigree, 186 Dennis pedigree, 185 Denny Sutton Hipend, Pewsey, 67 de Reviers pedigree, 186 de Toeni pedigree, 186 de Wik pedigree, 187 Drake, Dr. Richard, 151-3 Draycot Farm, Wilcot, 71 Dudley pedigree, 185 Dunch pedigree, 185 Dunn, Sir Philip, 60 Dunstanville pedigree, 185 Durocornovium (Wanborough), 188-9 Durrington Walls, 76-128 Earle, Bridget, 151; Dr. John, 151 Earle pedigree, 185 Easton Royal, 58 Eccles pedigree, 185 Ellesfield pedigree, 185 Ernle pedigree, 185 Fettiplace pedigree, 186 Fiennes, Celia, 149 Fluce pedigree, 186 Franklin pedigree, 186 Frome, George, 153 George pedigree, 186 Ghillo (Gillow), Edward, 153; Tom, 153 Giles pedigree, 186 Godby pedigree, 186 Gore pedigree, 186 Greenland, Richard, 158-9 Heathcote, Josiah Eyles, 182 Higford pedigree, 186 Hodges pedigree, 186 Huish, 71 Hungerford pedigree, 186 Hunt, Edward, 158; Flower, 156; Jeffry, 156- 61; John, 156; Thomas, 156-9; William, 156 Hyde, Dr. Alexander, 148; Edward, earl of Clarendon, 148; Dr. James, 148; Sir Robert, 148 Ice-houses, 181-4 Idmiston, 190 Iron Age: animal bones, 70, 83, 123; coin, 197; enclosures, 58, 61-71, 76, 190; iron slag, 71; loom weight ?, 69; pottery, 67, 83, 100, 197; quern, 69; settlements, 58, 76, 82, 190 Jeffreys, George, Lord Chief Justice, 151 Jenner pedigree, 186 Kembel pedigree, 186 Kingshill, Cricklade, 177 Kinnier pedigree, 186 Larkhill, Amesbury, 76 Latton, 177 Laurie pedigree, 186 Lawrence pedigree, 186 Little pedigree, 186 Low, George, 153 210 Lushill pedigree, 185 Lydiard Park, Lydiard Tregoze, 181 Mancroft pedigree, 186 Medieval: kilns, 179; miscellaneous objects, 198; pottery, 181, 1g1, 198; tiles, 180 Mells, Edward, 159 Mesolithic: chronology, 48; flints, 197 Morgan pedigree, 186 Naish Hill, Lacock, 179 Naish, Thomas, 153 Neeld pedigree, 186 Neolithic: antler and bone, 81, 98, 122-3; axes, 196-7; causewayed camps, 50; flints, 52, 97, 100, 197; flint mines, 52, 76; henges, 51; land usage, 49; long barrows, 48-9, 51; pottery, 50, 53, 76, 81, 94-7, 197; settle- ments, 49, 78-82 Newnton, North, 58 Nicholas pedigree, 186 Northern pedigree, 186 Nott pedigree, 186 Ockwell pedigree, 186 Packer pedigree, 186 Palaeolithic: hand-axe, 197 Parham pedigree, 186 Peto, Lt.-Col. Sir (James) Michael, obituary, 207 Pewsey, 58; Black Patch Field, 58, 189-90; Vale of, 50, 58 Pewsey Hill, enclosure, 61-71 Pierce, Thomas, dean, 147, 151, 153 Pitt pedigree, 186 Pleydell pedigree, 186 pollen analysis, 136-40 Porton Down, Idmiston, 190 Poulton pedigree, 186 Radiocarbon dating, 47-57, 81 Railway, military, 184-5 Reviews: Two Anglo-Saxon Cemeteries at Winnall, Winchester, Hampshire, by A. L. Meaney and S. C. Hawkes, 205; Art of the European Iron Age (A Study of the Elusive Image), by J. V. S. Megaw, 206; Discovering Regional Archaeology: Wessex, by L. Grinsell and J. Dyer, 206 Roman roads, 142, 177, 188 Romano-British: animal bones, 87, 123; bracelets, 74, 85, 118; bronze artefacts, 118; brooches, 189, 197; burials, 87, 125-7; corn- drying kilns, ovens, 83, 87; enclosures, 83; glass bead, 120; iron artefacts, 118; miscel- laneous objects, 118-22, 189, 197-8; pottery, 67-8, 73-4, 76, 100-16, 130, 132, 135, 177; 189, 197-8; querns, 87; river port, 177; settlements, 58, 67, 71-4, 83-94, 177, 188-9; stone artefacts, 120; tombstone, 177; villas, 189-90 St. John family, 181 St. Omer pedigree, 186 Salisbury cathedral, 147-55 Salmon, William, 184; William Wroughton, 184 Sancroft, William, archbishop, 147, 150 Savernake Forest, 134, 141 Selleck, Dr. John, 149 Selman pedigree, 186 Servington pedigree, 186 Sheldon, Gilbert, archbishop, 150 Shottesbroke pedigree, 186 Snail analysis, 140 Southbroom, Devizes, 181 Stanchester, Alton, 74 Swanton, Neil, 60 Sykes, Rev. W. S., 177 Townsend pedigree, 186 Trenchard pedigree, 186 Wakefield pedigree, 186 Walrond (Waleran) pedigree, 186 Walton pedigree, 186 Wanborough (Durocornovium), 188-9 Wansdyke, 129-46, 194 Ward, Seth, bishop, 147 Watkins pedigree, 186 Wells pedigree, 186 Wilcot, 177 Wilsford Hill, Wilsford (N.), 63 Wilton, Grafton, 58, 178 Wilton (S.), 191 Winter pedigree, 187 Woodbridge Inn, North Newnton, 58 Writh (Wriothesley) pedigree, 187 Young, G. M., 73-4 way ol aihey SY Leda0d. ah VE prelate y tay bewtnuial Wit Wselatt, 18 CUNT sling reach i oa | yd Td iy goes Taal Nia sel t vitae ih Es | 7 Bees 7 Fe Syd eed aoe AP, E _ F - — a 5, iy Ae eure ye ‘ Se ‘ 4 a = a ore = ma) Fen : ; a i 4 => a ” i - ' - 2 r : “ Ls, i 7 = y = He i te i = ~ x ~ NR ia ores Loita PLATE I a. Aerial view of the excavations from the east. b. The corn-drying kiln. Photographs, P. Sandiford DURRINGTON WALLS, 1970 PLATE II a. The corn-drying kiln. b. Infant burial in Pit 215. DURRINGTON WALLS, TQ7O0 Photographs, P. Sandiford The Magazine The Magazine is at present issued once a year. It is issued free to members of the Society. Contributions, editorial correspondence and books for review should be sent to the Editor at The Museum, 41 Long Street, Devizes. Back numbers of Magazines can be obtained from the Honorary Librarian, with the other publications listed below. Notes for the guidance of contributors will be found on pp. 207-8 of Volume 60 (1965). Publications to be obtained from the Librarian, The Museum, Devizes A GuiwE CATALOGUE OF THE NEOLITHIC AND BRoNnzE AGE COoLLeEcTIoNs AT DEvIzEs Museum, by F. K. Annable and D. D. A. Simpson. 1964. Post free, £1.40. A BIBLIOGRAPHY OF THE GREAT STONE MONUMENTS OF WILTSHIRE: STONEHENGE AND AvEBURY, by W. J. Harrison. No. LXX XIX (1901) of W.A.M. gtp. WILTsHIRE INQuISITIONES Post Mortem: Henry III, Epwarp I anp Epwarp II. 7op. Dirro. Epwarp III. 7op. Devizes BorouGH ANNALS, EXTRACTS FROM THE CORPORATION Recorps, by B. H. Cunnington. Vol. II, 1792 to 1835. g1p. (Vol. I is out of print.) WILTSHIRE Birps, by L. G. Peirson. 29p. Supplement. 1op. THE MACROLEPIDOPTERA OF WILTSHIRE, by Baron de Worms. Pp. xv, 177. 1962. £1.31. (Members £1.06.) PARLIAMENTARY HisTORY OF CRICKLADE, by W. B. Crouch. 1961. 66p. A HisTory OF SAVERNAKE Forest, by The Marquess of Ailesbury. 1962. 6gp. The Society’s Museum and Library, Long Street, Devizes The Curator is Mr. F. K. Annable; the Honorary Librarian, Mr. R. E. Sandell. All members of the Society are asked to give an annual subscription towards the upkeep of the Museum and Library collections. The Museum contains many objects of great local interest; and the Library a rich collection of books, articles and notes about the history of Wiltshire. Subscriptions should be sent to The Treasurer, The Museum, Devizes. Old printed material and photographs of Wiltshire buildings or other objects of interest will be welcomed by the Librarian at the Museum. The repository for records, e.g. old deeds, maps, plans, etc., is the Wiltshire Record Office, County Hall, Trowbridge. aollars) & in return for rk each year, with Be and Ma M. Ae oe Ail ae