The Ritual Shaft at Warbank, Keston

THE RITUAL SHAFT AT WA RB.ANK, KESTON By the late NAN0Y PIERCY Fox, B.A., .F.S.A. !NTRODUOTION It was decided in 1960 to re-excavate the site at Warbank, Keaton, of the foundations of the circular building known as the Romrun Temple, first excavated by T. Crofton Croker, F.S.A., and A. J. Kempe, F.S.A., in 1828.1 Work on the excavation commenced in the late autumn of 1960, and was continued in 1961, with the advice and help of R. F. Jessup, F.S.A. The work was made possible by the permission and co-operation of the Bromley Borough Council, the Ministry of Public Building and Works, and the owners of the adjoining land, E. J. Boosey, Esq., and A. Brooksban.k, Esq. It is intended to publish later the results of the re-excavation of the Temple. Towards the end of the season, a trial trench was found to intersect a well-cut circular rim in the solid chalk, 56 ft. north of the outside of the temple wall. Investigation by digging in 1962 revealed a deep circular pit which appeared to be a Ritual Shaft. SUmLIBY The Ritual Sha.ft or Bothros, 16 ft. deep and 11 ft. in diameter, was excavated in the natural chalk rock. It was divided into two sections by a projection 1 ft. wide, at a depth of 7 ft. 6 in., this projection becoming a flat step, 4 ft. wide, at one point in the western segment. The shaft was carefully made and smoothly finished to a depth of 12 ft. Below this level the sides sloped in towards the floor and were not so well finished. It was an impressive sight when excavated. The cremation of a lap-dog and a smaller dog had ta.ken place above the flat step in the adjacent area. A late Roman quarry had broken into the shaft near this point, destroying the area above and behind the step; but fortunately there remained, set on the 4 ft. wide step and extending backwards for 1 ft. 2 in., a foundation of clay, burnt in situ and having the appearance of tile. In the burnt clay were set blocks of very hard fawn-coloured chalk covered with more burnt clay, and in the debris of this were large, severely comminuted flints showing signs of intense heat. The quarry had destroyed the evidence necessary to reconstruct the site of the cremation, but one 1 Archcsologia, xxii. (1829). 184 ,. 00 Ol -----N sarcophagus removed Cl795 Foundations of tombs foundations of temple excavated 1828. 0 10 20 30 scale in feet Fro. 1. Relation of the Ritual Shaft to other structures at Warbank. Bothros or ritual shaft excavated 1962. trio trench. THE RITUAL SHAFT AT WA RBANK, KESTON clue suggested that this area may have been circular-perhaps a small circular pit opening off the shaft. When the cremations were completed, the remains of the dogs, mixed with finely comminuted charcoal, were placed in a pile on the floor of the shaft and resting against the side. An examination of the cremated animal bones from this pile by Dr. I. Cornwall, of the Institute of Archaeology, produced evidence that the remains were of a lap-dog and a smaller dog. Charcoal from the pile was examined by Professor G. W. Dimbleby at the Department of Forestry, Oxford, who identified the woods used as mainly hazel, with a little birch and hawthorn. He also found some carbonized organic matter not derived from wood. Seven sherds of a large redware vessel were placed on the pile of cremated bones and charcoal. These sherds were examined by Professor W. F. Grimes, C.B.E., F.S.A., but this redware cannot be dated precisely. The pile was covered with an extremely hard layer, I ft. 3 in. thick, lying evenly over the floor of the shaft. It was made up of chalk and a fine brown substance as yet unidentified; the surface was completely covered with a fine, soft brown powder, 3 in. deep (Appendix II), which thickened towards the sides of the shaft, suggesting that the sides had contributed to this layer and that the shaft had remained undisturbed for a considerable time. The shaft was then filled in, and a reddish daub was found clinging to the top 6 in. of the shaft where the top was undamaged, and in places extending a few inches over the rim. Numerous pieces of hard daub were found in the filling; they may be fragments of the cover; and a shallow key out in the rim of the shaft may have supported one end of a transverse beam. No soil from the surface found its way into the shaft, suggesting a tightly-fitting cover. In the late Roman period, quarrying for chalk broke into the shaft in the cremation area, and cut away the side above the step before quarrying was halted. The first material that found its way on to the soft powdery smface which had accumulated inside the shaft were pieces of the hard burnt clay covered with the white plaster seal, a fire-damaged flint, and chips off the distinctive fawn-coloured chalk blocks from the same area. The decision to fill the shaft must have been made at this time. Clean angular chalk, with a small admixture of sherds and bones, was tipped in. Domestic animal bones and sherds of the first to fourth centuries, of rubbish-pit type, came in the next tip. On top of this came a tip of cremation debris, thrown in from the cremation site above the step. It contained a quantity of charcoal, sampled by Professor Dimbleby and found to be hazel only. The main filling followed, some 7 ft. of clean angular chalk, tipped to above the level of the projection and the step. The top of this filling was rammed smooth and level, 186 trial trench 8 TURF LINE QUARRY FILLING C HALK FILLING PLAN Chalk quarry -10'0 l face of bothros 11' diameter. A floor-16' SECTION 0 HEAVY CLAY, CHALK SHERDS & BONES. ROUGHLY CUT D WHITE PLASTER SEAL SHARP AN GULAR CHALK Fili.ING C R EMATION WATTL E & DAl8 TOP {] POSSIBLE ORI GINAL SECllON &P􀁳AN oQ 0 I FINE BROWN POWDER -􀁴-. BONES & CHARCOAL 5 / Fra. 2. Plan 1111d Section of the Ritual S11aft. j A (/Ck"e p. 187 THE RITUAL SHA.FT AT WAR BANK, KESTON and a flat-topped cone of weathered chalk stood in the centre. The sweepings of the cremation area were deposited between the cone and the cremation area, and this deposit trailed back over the step. Chalk rubble was tipped over the sweepings, the filling then changing to heavy clay, full of domestic rubbish, late Roman in date. Samian sherds were absent; but New Forest wares were present, and third- to fourth-century sherds were abundant. The main filling of the shaft above the soft powdery layer is of one date, not earlier than the fourth century (Appendix I). The chalk quarry was filled in at the same time, the lower filling being of chalk and the upper part of the late heavy clay filling, which extended over the whole of the quarried area; and all damage done to the site by quarrying was made good superficially. DESORIPTION OF LE'VELS (Fig. 2) Level 1-Top Soil This contained domestic animal bones, sherds, broken tile, oyster shells, and modern debris. There were also pieces of the side of the pit, very hard, with traces of burnt daub. Level 2-Top layer of filling This had been thrown in on top of the lower chalk filling. It contained animal bones and teeth, mainly heavy fragments of jaw, broken tile, snail shells, sherds in large quantities, and clean daub bearing wattle marks. Level 3 This was at first thought to be an intrusive grave, but was later found to be a quarry which had damaged part of the rim of the shaft. It contained heavy clay, chalk, large pieces of clean, burnt daub with wattle marks, animal bones and sherds in large quantities. Level 4 A fine reddish layer full of tiny pieces of daub, burnt black, and pieces of charcoal, lying above a step, I ft. 2 in. wide, cut in the north face of the shaft. Level 6 A clean angular chalk filling, largely sterile; it contained a few fossils, four pieces of burnt daub, animal bones, two pieces of charcoal, a few grey sherds, and a sherd resembling Patch Grove ware. Level 6 Was found not to be a distinct level and is not recorded. 187 THE RIT{JAL SHAFT AT W.A RBANK, KESTON Level 7-0rematio'li debris A dark layer sloping from north to south across the shaft. It consisted of a layer of charcoal and burnt daub lying on a layer of burnt and unburnt animal bones, flints fractured by · intense heat, and numerous sherds. Level 8 A clean chalk filling containing a few animal bones and sherds lying above a layer of fine brown powder. Level 9 This consisted of hard chalk and tufa, the lumps having a brown matrix. At the north end, beneath the deposit of fine brown powder, were sherds of redware, with adherent chalk, lying immediately above the charcoal and bones of the cremated dogs, whlch were against the side and formed Level 10. RITUAL SIGNIFIOANCE OF THE LAP-Doo Statuettes depicting a matron seated on a throne or hlgh-baoked chair, and holding a small dog in her arms, are well known in N.E. Gaul. In the temple area at Trier a number of these .figures were found together by a small shrine, before the door of which stood a stone statue of a seated matron who holds a basket of fruit on her lap, with a dog seated on her side. On thirteen of the monuments found at the site of the temple at Nehallenia at Domburg in Holland, the goddess is shown with a dog seated at her side. In both cases the presence of the dog indicates the chthonian manifestations of the universal mother-goddess who played an important role in the religion of Romano-Gaulish people. The :figurines are distributed along the trading routes of the Rhlne and Mose!, and the goddess Nehallenia (protectress of seafaring merchants trading with Britain) had her sanctuary on the Isle of Waloheren. The goddess with the lap-dog had other emblems-a pomegranate, a small bowl or patera, and an ear of corn-but the essential characteristic is the lap-dog, which suggests that it was her main attribute, indicating her principal function.2 The goddess Ceres had agricultural connections, and by Roman times had become identified with her Greek predecessor Demeter. Demeter was the goddess of corn, and more broadly of the growth and death of vegetation. There is a connection between the dog and agriculture, especially in those cults concerned with the growth of corn h1te the Robigalia. The Cerialia {19th April) was followed by the 2 Arohceologia Oantiana, lx.x (1956). 188 PLATE 1 .\ Lc,·cl ,-Cr·ein,it ion debris. The point of th,· ranging polo indicat,•s c-1-omu􀂱ion dehl'is. 􀂲 ft. ahov<> this is tho step cut in tho north foco of tho shuft, wi1 h tho lakr quat'ry ahovo agu.i11. 13. Yiew of the ::>haft. Remo,·ini, 1lre filling lwlo11· Le,·1•1 7. [/fire p. 1, ... THE RITUAL SHAFT AT WARBANK, KESTON Rohigalia (25th April) when a red dog was offered in sacrifice to Rohigus, the god of rust (mildew) in wheat. A large dog as the companion of an underworld deity, or as a denizen of the world of the dead, is very familiar. Cerberus was the guardian of the Gates of Hell, and is depicted by the side of the Lord of the Underworld. The Greeks regarded Hecate, leader in the wild chase for the souls of men, as having a dog-like form. In Greek iconography she is depicted with a canine companion; and the main constituent of the so-called Hecate supper was the flesh of a dog. A leaping dog is a-ssociated with Teuta.tes on the Gunderstrup cauldron; a dog accompanies the dieu au maillet identified with Sucellus, and ultimately with Dispater, from whom the Gauls claimed descent. The dog played a part in hydrotherapeutic cults at spring-heads, and in the Celtic water cults whose sanctuaries were sited at wells, spring-heads, and beside streams. The dog had healing properties at Epidaurus and Rome, where inscriptions testify to miraculous cures by the animal. Dogs were kept in the sacred precincts of the temple of Aesculapius for that purpose; one miraculous cure there was the restoration of sight to a blind child who had been licked by the sacred dogs. Representations of dogs as cult-objects were found at the temple ofNodens at Lydney.8 The goddess of the lap-dog has a link with Diana. Diana's lunarshaped symbol is worn as a.n amulet by the matrons of the Rhineland; the goddesses who hold lap-dogs a.lso wore them. Diana was the name by which the Romans knew the goddess Artemis; but the manybreasted Artemis of Ephesus was a very different goddess from the chaste huntress of the European Greeks. She was the great Asiatic Mother-Goddess, and was only gradually identified with the deity brought in by the Ionian colonists. Representations of Diana from classical sources show the goddess with a hunting-dog; when the Diana of the Ephesians is depicted with one, two, or three small lions or lion-cubs seated on her arms, playfully clinging to them, a link with the goddess of the lap-dog being suggested-particularly where the goddess is shown with one animal on each arm. Montfaucon illustrates seven representations of the goddess where she is characterized by an animal 01· animals on her arms, in addition to the distinctive tower head-dress which distinguished her from Isis, whose head-dress was in the form of a lotus. All the animals depicted are small lions or Hon-cubs, but in relation to the figure of the goddess, they are of lap-dog size. The series of illustrations demonstrates how easily the lion symbol could degenerate into the lap-dog symbol. The small animal on the arm symbol appears to be reserved for the great Mother-Goddess, 8 Exoo.vo.tions in Lydney Park, Glouoestershire-S. of A. Research Report, No. ix (1932), R. E. M. a.nd T. V. Wheeler. 189 THE RITUAL SHAFT AT WARBANK, KESTON and may be taken to represent a particular attribute or function. A Syrian goddess represented as Diana of the Ephesians, with a small animal on each arm, is the only other classical example known to the writer. Normally the god or goddess holds an object in the hand.4 AOKNOWLEDGMENTS Thanks are due to Messrs. J. C. Evans, Clarkson, Hills, and G. and Len Smith, who did the heavy work required to excavate the shaft; to R. F. Jessup, F.S.A., for advice and visiting the site; to Dr. I. Cornwall for examining the bones; to Professor G. W. Dimbleby for examining the charcoal; to Professor W. F. Grimes, C.B.E., F.S.A., and P. J. Tester, F.S.A., for examining the sherds of pottery; and to J. N. Carreok for reporting on the brown powder. REFERENOES Frank Jenkins, 'The Role of the Dog in Romano-Gaulish Religion', Latomus T., xvi (1957). Frank Jenkins, 'Nameless or N ehallenia', Archaiologia Oantiana, Ixx (1956). Montfaucon, 1719. Tome I. Premiere partie. Plates xciii and xciv, Diana of the Ephesians. Plate V, Syrian goddess. N. Piercy Fox, 'Warbank, Keston: A Romano-British Site', Arclueologia Oantiana, lxix (1955). A. J. Kempe, Archreologia, xxii (1829). I. Noel Hume, 'Ritual Burials on the Upchurch Marshes', Arclueologia Oantiana, lxx (1956). R. E. M. and T. V. Wheeler, 'Excavations in Lydney Park, Gloucestershire', Society of Antiquaries Research Report, No. IX, 1932. APPENDIX I THE CoA.RSE PoT1.'ERY By P. J. TESTER, F.S.A. THE stratified pottery recovered from Levels 2 to 9 comprises for the most part typical Romano-British sherds of common domestic vessels such as cooking-pots and bowls. Its main interest lies i n the dating evidence which it provides for its context, and the following summary is concerned mainly with this consideration. There are a few pieces of diagnostic value from Levels 2, 3 and 7, and so far as the sherds from 4, 5 and 8 show any distinguishing features they appear to be contemporary with the rest. ' Montfauoon, Tome I (1719). 190 THE RITUAL SHAFT AT WARBANK, KESTON Allowing for the inclusion of a few possibly earlier rubbish-sherds, the entire assemblage bears a distinctly fourth-century character. Hardly anything is likely to be earlier than the third century and there is a significant lack of samian in the material passed to me for examination. Levels 2, 3 and 7 contain examples of the common flanged bowls with straight sides, as figured in Rwhborough I, pl. XXIX, Nos. 121-22.6 This is one of the pottery forms most frequently represented in late Romano-British contexts in Kent. At Colchester it is said not to occur before the middle of the third century,6 a conclusion which my own observations elsewhere tend to support. Level 3 contained part of a cooking pot with outbent moulded rim and regular horizontal rilling on the body, matched closely by Richborough 336, and there dated to the fourth century. With it was a red bowl imitating samian form 38, for which good parallels would be Rwhborough 109-11, all dated to the late-third or fourth century. In Level 9 were found several joining fragments of a fairly large vessel of red-tile clay. Unfortunately the rim is missing and the sherds are insufficient in number to supply much information as to the general shape of the body. I am unable to make any useful suggestion as to its age. APPENDIX II LEVEL 9. THE BROWN POWDER By J. N. CARREOK A SOLUTION of the brown material, in particular, and to a much leas extent the white, in dilute hydrochloric acid, produced an insoluble brown residue, very marked in the brown material. This residue can, from the analysis, only be ascribed to the Manganese or Phosphatic content. The apparent Manganese content, so high in the brown material, may have originated from the manganese hydroxide present in very small quantities in the local chalk, or perhaps that laid down in a water-logged situation by certain bacteria or algae, but evidence is entirely lacking. Museum Section, Department of Geology, Queen Mary College, University of London. ' J. P. Bushe-Fox, First Report on the Excavation of the Roman Fort at Richborough, Kent (1926). • M. R. Hull, .Roman Ool.chester (1958), 288, no. 305A. 191 18B