The Romano-British Settlement at Springhead; Excavation of Temple I, Site Cl

vthmlaQk (fyuntwnu THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD ; EXCAVATION OF TEMPLE I, SITE C 1. By W. S. PENN, B.SO. PAET I. GENERAL INTRODUCTION THE location of the Romano-British settlement at Springhead, Southfleet, Kent has been given previously (1). The position of the temple complex and other buildings in the settlement is given in Fig. 15 of the present report. As a result of work over the past two years, it is quite clear that an important temple complex exists at Springhead. It is known that the temenos, which is surrounded by a wall and occupies one enthe insula of the settlement, contains at least four structures. Temple I has the usual square celia surrounded by a portico or ambulatory, and in addition possesses many unusual features. Temple II has the celia wall replaced by a series of plinths in the form of a square to give a most unusual arrangement. By the side of Temple I is a smaller square or rectangular building, the purpose of which is at present unknown. In front of Temple I is a pedestal which supported a freestanding column surmounted by a Corinthian capital and possibly a statue. Thus the complex must have been quite impressive and may actually have extended further. Unfortunately the south-east corner of Temple I I is covered by a railway embankment, and there may be another building similarly covered. Temple I forms the subject of the present report as Site C 1. The pedestal and part of the temenos wall have been reported previously in Site B (2). In the next report it is proposed to describe Temple II as Site C 2 and the square or rectangular building as Site C 3. I t should be stressed that although the report on Temple I is complete in itself it is only tentative. It will not be possible to interpret the finds fully until all the buildings in the temple complex have been excavated. References will be found on p. 55. 1 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD SUMMARY OF WORK Basically Temple I has the normally accepted features of a Romano- Celtic temple: that is a square celia, surrounded by a square ambulatory or portico. In addition, it has a porch entrance on the east, a projecting store-room on the west and wings on each side of the porch, the wings being later additions. The temple has tessellated floors throughout, including several mosaics. Small finds were also of great interest. A well-preserved, uninscribed altar was found which is virtually unique, this also applying to part of a figurine of pseudo- Venus. They were both found in the celia of the temple, a unique occurrence in this country. Other small finds such as the votive bronze thumb and the seeds used as offerings are also of great interest. ARRANGEMENT OF REPORT The Report has been divided into five parts. It is hoped that this will facilitate a study of various aspects of an involved and detailed subject. Part I is concerned with general matters. Parts II, IV and V are technical reports on structural history, architectural remains and finds respectively. Part III is a more descriptive account of the features of the temple and should be of general interest. Usually individual dating is not given in the body of the report. Full dating evidence for the various strata and features is given in Table I. ACKNOWLEDGMENTS Our indebtedness to the farmers, Messrs. J. Bartholomew & Sons for allowing us to excavate on theh valuable land, grows each year. It is with regret that I must record the death in 1959 of Mr. P. Bartholomew, who was so kind to us. His son, Mr. J. Bartholomew has been interested in our work and we hope and trust that the association between us will continue for many years. Our sincere thanks are due to all those authorities who helped with advice or the study of small finds. Professor Richmond offered useful suggestions regarding the functions of various features of the temple. Mr. S. S. Frere, M.A., F.S.A., gave much valuable advice and criticism during the preparation of this report. Mr. Franlc Jenkins, F.S.A., has been kind enough to write a note on the possible cult of the temple. Special thanks are due to Mr. R. A. G. Carson, M.A., for dealing with the coins and to Mr. M. R. Hull, M.A., F.S.A., for dealing with the fibuke. Thanks are also due to Mr. J. W. Brailsford, F.S.A., for the identification of various small finds ; Miss D. Charlesworth, M.A., for 2 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD reporting on the glass ; Mrs. J. E. King for commenting on the animal bones ; Miss H. A. M. Maedonald for the petrographical report and Mr. F. L. Balfour-Browne for reporting on the charcoals. The author also wishes to thank Mr. E. Tilley for his considerable help in dealing with the small finds, both on the site and in the preparation of this report. Mr. W. Gee's drawings again attain theh usual excellence for which I am most grateful. Mr. P. Connolly again gave valuable assistance with the photography. Of the excavators, the author offers special thanks to Messrs. E. Tilley, W. Gee, G. Buries and F. Turner. Others who gave assistance were Messrs. P. Anderson ; P. Connolly ; D. Ford ; R. Ivell ; R. Chaplain ; A. Harrison ; M. Hubbard and Mr. and Mrs. Tidby. The work at Springhead is carried out by the Excavation Committee of the Gravesend Historical Society. The author is indebted to the President, Chahman and Council of the Society for theh help and encouragement during the years. We are all grateful for the support of the members of the Society and must also record our appreciation to the Carnegie (U.K.) Trust and the Council for British Archaeology for a grant to purchase certain equipment. Miss Taylor of the Journal of Roman Studies kindly loaned blocks for Figure 1 and Plate III A. PART II STRUCTURAL HISTORY OF TEMPLE PHASE Z The earliest masonry structure was built c. late first/early second century, and is described as Phase A below. However, the site was occupied from Claudian times, although unfortunately positive evidence of wooden structures was not found, due to the disturbances of subsequent building operations. The early period in the site's history has therefore been designated Phase Z to distinguish it clearly from the well-defined later structures. Abundant pottery, some brooches and coins, indicate a fahly intensive occupation during Phase Zl, of Claudian date. A heavy scatter of pebbles over the area subsequently occupied by the celia, and along the path to an eastern entrance, indicates that there may have been a wooden structure at the time. There were no pebbles beneath the portico (see Sections, Fig. 3 and 6) which would suggest a temple lacking such a feature, if indeed there was a temple at all during this period. In this connection it may be noted that occupation debris outside the area of the temple was so sparse as to indicate some special significance for the temple site. There was another occupation period, Phase Z2, during the Flavian period. The evidence to associate it with a temple is even less than with 3 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Phase Zl. There were no pebbles in the celia area, simply a charcoal scatter below the celia and portico (see Section, Fig. 3) with some sand under the east portico. However, there was much pottery ; the occupational debris of the period did occupy the same area as the previous phase and once again the area outside the temple produced very little. A summary of the situation indicates that there was some special significance attached to the area and that there may well have been wooden temples during Phases Zl and Z2. If there were temples on the site at this time, the path, portico and celia were all at the same level as the surrounding ground. There was no evidence of pre-Roman occupation, to indicate why the site was chosen. PHASE A Although there is doubt concerning the existence of temples during Phases Zl and Z2, there is none regarding the Phase A temple. This was a well built masonry structure with the usual features of Romano- Celtic temples and some unusual ones as well (see plan, Fig. 1). A trench was dug through the Flavian occupation stratum, and filled with loose flints as foundation stones. The flint walls were built on these, the level being indicated by mortar droppings at several points (see Sections, Figs. 3 and 4). The floors were subsequently made up with clay throughout the enthe structure, including the porch. All floors were at the same level, and all were about 5 in. above the outside ground level. A considerable amount of pottery and several coins (including two of Domitian) all sealed by the clay floors, dated the structure to the late first/early second century. A shght amount of plaster indicated that the walls were plastered externally and internally. I t seems rather incongruous that a well-made masonry temple should be provided with a clay floor. However, there is a precedent for this at Woodeaton (3), where the Period I temple is similarly furnished. It is also of similar date (Neronian or Flavian) and the celia of similar dimensions (16 ft. 4 in. by 18 ft. 11 in., externally, compared with 18 ft. 8 in. by 18 ft. 8 in. at Springhead). The structure is of normal plan insofar as the square celia surrounded by a square portico is concerned. The eastern entrance is also usual but masonry porch walls a t such an early date are unusual and are not to be found at Harlow (4), Colchester (5), Verulamium (6) and Woodeaton (3). Even more unusual is the small building projecting from, and of one build with, the west portico wall. Its significance will be discussed later but in the few temples where such a feature exists, it is usually a later addition, normally described as an annexe (see the temple at Frilford, Berks. (7) for example). 4 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD With the exception of the porch walls, which are 15 in. wide, all the walls were 21 m. wide. This is comparatively narrow, a featoe re cTed M P T ^ ^ ^ °f * * ^ It a ^^neeTZ • • » —- SCALE OF FfET STRONG ROOM PHASE A (LATE lst/EARLY 2""; PHASE B (ANTON/NE) PHASE C (LATE 2»<) TE/ S^E i t AJE I : > FLOOR (PHASE R) (PORCH MOSAIC -PHASE C° DESTROYED MOSAICS (PHASE D,4») m ANTAE 9HHi n 1 ' i . : i n • I i i WaBtth wm \\l: ML-' ] i ' VESTIBULE r. ; ••:••; s^* ^'"'-[ifi PEBBLE PATH ;:*? Mill ^ife «sr, sBSI ANTAE [Reproduced by courtesy of Miss M. V. Taylor and the ,. FIG. 1. Plan of Temple I, Site 01. I of Roman Studies There may have been a paved walk around the temple A layer of flints exists by the side of the porch (Fig. 2) and a layer of chalk at the west side of the temple (Fig. 5). THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD PHASE B Some significant and interestmg additions were made to the temple during Phase B, which is of early Antonine date. These changes were not made simultaneously and it has been possible to distinguish between Phases Bl, B2 and B3. It should be stressed, however, that the times which elapsed between the sub-phases cannot be determined and may have been days, weeks or years. The sub-phases may simply represent the sequence of building operations. However, Phase B as a whole is early Antonine and possibly A.D.150-160, although the dating evidence is not so clear as for earher periods. All the walls appear to have been rebuilt during the period, the building levels being best seen in the sections, Figs. 4 and 5. The porch was paved with an extremely hard cement during Phase Bl (see Fig. 2) at the same level as the surrounding ground. The vestibule and the remamder of the portico were levelled off with a light soil, at 2 in. above the porch entrance, and the celia was made up with soil and then floored with clay to raise it 9 in. above the porch entrance and outside ground level. A most interesting feature of Phase Bl, is the hearth in the centre of the west side of the celia (see Section, Fig. 3). Its extent was strictly limited to that part subsequently occupied by the apsidal suggestus and it must surely have had a ritual significance. It is paralleled by three superimposed hearths, this time in the centre of the celia, in the Period I, Woodeaton temple (3). The thick layer of charcoal was mixed with a few cockle and mussel shells. If the 6 in. of burning may be interpreted as of ritual origin, the problem of the length of Phase Bl arises. Was it simply the remains of one offering to sanctify the temple or was it used in this way for some time ? The thickness of the burnt layer may admit of several occasions of use, but the dating evidence is inconclusive. From the clay floor came a coin of Sabina and from the burnt layer two sherds of early Antonine pottery. During Phase B2, the cement entrance was retained, the corridor was provided with a clay floor, the suggestus was built and the extant mosaic floor was laid in the celia. This meant that the vestibule (corridor) floor was 7 in. higher than the porch floor and the celia floor 17 in. higher than the corridor floor. These differences in height clearly necessitated the provision of steps and the gaps left when these were later removed may be clearly seen against the east portico and celia walls (see Section, Fig. 3). This point is discussed further under Phase B3. Two plain red tesserse were sealed under the clay floor of the vestibule. Various explanations for this may be postulated and one possible reason is that the tesserse were left behind during the con- 6 KEY TO STRATA (ALL SECTIONS) SECTION, FIGURE 2 =3 T. DISTURBED PLOUGH LEVEL LIGHT SOIL CHALK H CLAY MORTAR.PLASTER OR OPUS SIG. MORTAR REINFORCED WITH PEBBLES PEBBLES TESSERAE PLASTERY SOIL DARK SOIL DARK SILTY CLAY BURNT EARTH.CHARCOAL ETC. BROKEN TILES SLAG SHERDS OF POTTERY OYSTER SHELLS NORTH SOUTH DISTURBED PLOUGH LEVEL 5S8J ikooQ. PORCH PORCH WALL WALL CREEMSANC OR TAB OP=SET £ WZ&Z iPPpM m mWWWm. OMi « « » B SECTION FIGURE 3 SCALE OF FEET \ \ 1 : NO P 3Q FEET WEST PORTICO C E L L A DISTURBED PLOUGH LEVEL — U — AI T AO ^ ^^ i "W- * ' ' .. I I I J-~ / J J -GROWN D LE V£ i_ V 6 STI B U LE PORCH — ^ —^J <££? ^ ^^==>^7 ^7i 5 = ^ •=== '—r i t. S3 (LTD iipGESTUs ALTAR WrtMfflw/g& ^2M ^ s s ^ ft^ff^ .(g< l l l l l l f 1 T SECTION, FIGURE 5 SCALE OF FEET EAST/WEST SECTION BREAK IN SECTION DISTURBED PLOUGH LEVEL SOUTH/NORTH SECTION m aD>an> ^S£ £ £ ^S wwmi^m^w§BM • • • • i i i SCALE OF FSET PIG. 4. Seotion through S.E. Corner of Temple. FIG. 5. Seotion through W. Corridor and Store-Room. [face ». 8 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD no datable material in the stratum immediately underneath and therefore the mosaic cannot be positively dated. However, it is clearly later than early Antonine, since it overlies the Antonine porch floor, and seals a layer of dark soil which accumulated against the early Antonine portico wall. The cement base to the step is shghtly overlaid by the second pebble/flint porch floor, which has been dated by sherds of pottery, again including castor ware, to the late second century. It seems reasonable therefore to assume that the mosaic is of late second-century date. The tesserse and floor base of this mosaic are quite different from the early Antonine floors (see Part IV). The celia mosaic and possibly the portico floors may have been repaired about this period and also late in the third or early fourth century. A single sherd of pottery under the tesserae, on the east side of the celia by the south groove (see Fig. 8.7) may be of late secondcentury date or even later. This would indicate a repair at this time. In the cavity at the end of the north groove a coin dated c. A.D. 268-278 was found. This sherd of pottery and the coin could at first be thought to date the mosaic itself and not just a repair. Against this are the facts that (a) the tesserae where the dating evidence was found were very disturbed and the sherd and coin are easily explained as later intrusions ; (6) not one sherd of pottery or coin later than early Antonine was found below the solid foundation of the celia (or the other) floors and (c) the early Antonine structure (and therefore floor) is overlaid by a late second-century deposit. In the portico floors, mainly on the north side are some large patches of tesserae, containing a number of white instead to red tesserae. Presumably these were later repairs since the remainder of the floors were so well made. During Phase C, the outside ground level was raised, and the new porch floor was laid. The latter was about 3 in. above ground level, the vestibule mosaic about 6 in. above, and this reduced the celia level to only 11 in. above ground. There appeared to be no definite surround to the temple. A short wall projecting from the south portico wall will be noticed and this is actually of one build with the flanking wing walls. Its full import cannot be considered in the present report since it connects Temple I to Temple I I . It will be considered in the report on Temple II. PHASE D I t should be stressed that dating evidence for Phase D is sparse. However, a well-defined mortar layer (see Sections Fig. 4 and fi) sealed several coins of the second half of the third century, as well as a few sherds of Rhenish ware. It appears therefore that the walls (those which can at present be seen above floor level) were built in the 9 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD late third or early fourth century. These were no minor repairs and the quoins, for example, were well-constructed with tiles. This could then be a part of the building revival encouraged by Constantius and may be termed Phase Dl. SECTION, FIGURE 6 SOUTH NORTH DISTURBED PLOUGH LEVEL c ^ C O <2E2> c = a ^ <^C3 - ^JX ^ o<^ ^ r- m> ^ ^ G33 C22r*r ri H333 OS GE3 CT-TT3 &&g g g ° rrrTT-rrrrm/ / / / / n T-riininninniunTTTiHnnnfliniiinjm t ' LHAHJWJ I N FIG. 6. Section through N. Corridor. The section, Fig. 4, indicates that there were minor repairs later than those of Phase Dl, but there is no indication of date. The period of the repairs may be deemed Phase D2. A loose layer of tiles, including many tegulae, was used to repair the porch floor. This sealed a single coin of Carausius wliich means a repair in phase Dl or D2. Only two coins were found associated with all the porch floors, which clearly means that they had no ritual significance as at Frilford (7), where 78 were found. 10 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Other structural changes took place, presumably during one or other of the two phases. The northern flanking wall of the main doorway from the vestibule into the celia had been ripped out. It was replaced by a screen or door (see Part III). The mosaic floors sustained some considerable damage, presumably during Phase Dl or D2. If the damage had occurred earher, it would have undoubtedly been repaired. The vestibule mosaic and the tessellated floor in the projecting western room were completely removed, although the floor foundations were undamaged. The portico floors sustained minor damage at this time. Possible reasons for the destruction of the floors are considered in Part 3. The temple ceased to be used for sacred purposes sometime during the fourth century. Only three coins were found immediately over the portico (except on the west corridor) and celia floors, the latest coin being dated to A.D.335-341. The temple was thus open in the middle of the fourth century, but not necessarily in use. This may be termed Phase D3. Phase D4, although it cannot be precisely dated, has a terminus post quern of A.D.341. Over the temple floor of the west portico was a thick layer of charcoal, slag and partly worked iron fragments (see Section Fig. 3). This part of the building was clearly used as a blacksmith's shop (see Part III) when the temple had ceased to be used for rehgious purposes. Five coins, all of the Oonstantinian dynasty and not later than A.D.341, indicate that the blacksmith moved in some time after mid-way through the century, clearly soon after the temple had been deserted by the priests. Finally after a few years, the temple collapsed. The roof collapsed first, followed by the plaster off the walls, and then the flint and chalk walls themselves. This was normally the order in which the rubble was found. The altar was left as it toppled over which presumably indicates that there could not have been mnch activity anywhere on the site after this time. A summary of the constructional changes during the various phases, and of the levels of the different floors is given in Table 14. PART III THE FUNCTION OF THE TEMPLE AND ITS CONTENTS INTRODUCTION The extant temple comprises the porch with its mosaic and flanking wings ; the vestibule with its destroyed mosaic ; the portico connecting with the west projecting room and the celia with the apsidal suggestus. These features were the culmination of almost two centuries of building 11 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD from Phase B2 c. A.D.150 to Phase D2 sometime in the fourth century. Any explanation of these features will thus cover the function of the temple components from Phase B2 onwards. Moreover, the west projecting room was in existence in Phases A and B l so that its purpose during those phases will also be explained. The exceptional feature of the hearth in Phase Bl has already been discussed in Part II. The extant structural features of the temple and the positions of the small finds are given in Fig. I. General views of the temple are given in Plate I, A and B. THE PORCH WINGS (Antes). In the late second century, the porch walls were incorporated into flanking rectangular structures. The purpose of these structures is the first problem to be considered. The addition of an annexe at some time during the life of a temple was not uncommon. Rarely, however, has there been any evidence to explain their use. The evidence from Springhead, whilst not conclusive, does offer certain useful pointers. In appearance the temple at Frilford, Berkshire (7) is somewhat similar to Temple I at Springhead. There is a threefold division outside one end of the temple. However, at Springhead the outbuildings flank the entrance on the east, whereas the buildings at Frilford are on the west, opposite the entrance. The walls at Frilford were deeply founded, which suggested an extensive superstructure to the excavators, and the buildings were floored. Thus it appears that at Frilford, the added structures were functional. The same can probably be said for other annexes of this type which were not situated on the entrance side of the building. At Springhead there was no floor, and the walls were not deeply founded. They were also narrower than the main walls of the temple. Thus they could not have supported a very extensive superstructure, at least of stone, wliich throws doubt on their use as rooms. Moreover, there were no doorways, the walls having survived to a sufficient height to make this quite clear. At La Foret de Beaumont-le-Roger (4) a forebuilding projects 20 yards from the entrance side of the building and the suggestion was made that it supported a flight of steps. This was clearly not so at Springhead, but functionally the structures could have been similar in that the remainder of the forebuilding on each side of the steps, and the structures at Springhead on each side of the porch were in the nature of cheek-walls (sometimes called antae) of Italian style classical temples. There is a better example at Harlow, Essex (4). A forebuilding, of lighter construction than the remainder of the building extends 12 ft. 6 in. from the front. It had been artificially filled with earth, thus 12 PLATE I f A. Aerial View of Temple. B. Looking West from Porch. [fact p. 12 PLATE II A. The Suggestus. B. The Store (or Strong) Room. PLATE III [Reproduced by courtesy of Miss M. V. Taylor and the Journal of Roman Studies A. Celia Mosaic. \ B. Porch showing Step : Slots for Partition ; Drainage Channel and Mosaic. C. Poreh Mosaic. PLATE IV * f l A. East Corridor, showing position of B. West Corridor, showing holes in destroyed mosaic in vestibule. tessellated floor. C. South Corridor. D. North Corridor. PLATE V A. N.W. Quoin of Celia. B. Rectangular Stone Base in situ. t&* C. Altar and Base from front and right side. D. Altar and Base from front and left side. PLATE VI A. South Vestibule Wall B. North Groove in CeUa. C. Tiles in Make-up of Vestibule Floor. D. Clay Ramp in West Corridor. THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD « n| o / r r \) 32 33 34 35 36 ? ; V A 37 38 39 40 41 42 FIG. 7. Samian Ware ; Potter's Mark and Standard Forms. All illustrations %. 13 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD making a platform in front. It may well have supported steps, again giving the appearance of cheek-walls on either side. At Springhead the flanking structures (at 13 ft., almost the same length as at Harlow) were not solid filled, but contained the strata of periods later than the walls. The buildings must, therefore, have been open at various times during their history, almost certainly during rebuilding operations. Such an arrangement would not therefore have been in accord with their use as rooms or platforms. I t has already been observed that the walls were light, and it is likely that they never extended much above their present height. Some support for this view is given by the fact that the rubble from the main temple walls overlies the low walls of the porch structures whereas the main temple walls, which are higher, were too close to the surface to have any rubble remaining over them. The wings subsequently became filled with rubble from the superstructure including a large number of tiles. It is therefore suggested that the flanking buildings were low structures, hollow, but covered over with tiles. They would have had the appearance of the cheekwalls of Itahan type classical temples. Such an interpretation would explain the absence of floors (which existed everywhere else in the temple) the light walls and the occupation strata inside the walls. There is some difficulty in giving a name to these structures flanking the porch. They appear to have been purely ornamental like the cheekwalls on each side of the steps of certain temples—the Maison Carree at Nimes for example, but they are rather more than walls. They also have some of the features of antce, with the columns in antis. However, neither of the names is strictly accurate although the meaning of either could be extended to include the present type of structure. The construction of Temple II, to be described in the next report, offers considerable support for this theory. Aesthetically also, the theory has much to commend it. Walking into the temple between two high walls would have been oppressive and is not in accordance with normal Roman temple architecture. THE PORCH AND ITS MOSAIC At the end of the temple path, with its flanking wings, was a wooden porch, protecting the structures inside. The slots which held the wooden partitions are lined with tiles which may be seen in Plate I I I B, along the front of the porch mosaic. A sturdy threshold step was also placed just in front of the floor. The fine porch mosaic is a panel 26 in. by 54£ in. flanked by tile floors laid in a herringbone pattern. The decorative motif is simple with fine designs in colour which is primarily red but with some 14 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD relieving yellow, pink and blue tesserse (see Plate IIIo and Part IV for full details). The interesting feature is the drainage channel in the right flanking herringbone pattern floor. It is a structure, well built with tiles, passing through the wooden partition to the gravel path outside. The purpose of this channel is not easy to decide, but is of great interest. The possibilities may be divided into two broad classes, functional and ritual. The functional aspect arises because of the damage which water causes if left lying on tessellated floors. If the mosaic had been washed, or rain found its way through the porch, the water would have had to be drained off. A channel at Silchester has been explained in this way (9). The ritual explanation is concerned with libations which may have been made on entering the temple. The right hand side of the temple is the lucky side, and there may have been a small altar by the side of the channel which is on the right.1 Thus the choice remains between the use of the channel for draining off rain water and making libations. Logically, one would think that for draining off water in general, i.e. rain, the whole floor would have to slope gently towards the drain. This is certainly not so and the drain is designed for draining off a specific quantity of liquid, e.g. libations. On this basis, the drain would have been used for libations. THE VESTIBULE MOSAIC Immediately behind the porch, the portico has its continuity broken by two short projecting walls. The area between the two may be conveniently termed the vestibule, although it has some of the features of a pronaos. The principal features of the vestibule are the destroyed mosaic floor, the large stone base and a number of unusual iron objects. There is little doubt that a large mosaic once occupied most of the area of the vestibule. Fortunately there were a few white tesserse in the corners to show that the mosaic was contained in a panel 7 ft. by 9 ft. 2 ins. (Fig. 1). The area of the vestibule mosaic was thus about 64 sq. ft. compared with only 7 sq. ft., for the porch mosaic and 16 sq. ft., for the ceUa mosaic. Insufficient evidence exists to determine how the mosaic was destroyed, but a few observations may be made. It was not by post- Roman ploughing since the floor is protected by walls and no loose tesserae were found. It was almost certainly not destroyed by robbers removing a marble or other structure, since no fragments of any stone or any prepared foundations were found. Wear is a likely explanation since no doubt the loose tesserse would 1 I am indebted to Professor Richmond for this suggestion. 15 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD t TO 7 ^ ? ^ TCOCCOCCOOOt 9 8 5 IlllOUCtlO^ V Z^> 20 3 23 4 ^ v <> 24- I 25 19 -V II 1 12 S l3 \ W ) 15 6 ) Fio. 8. Coarse Ware and Mortaria. All £. 16 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD have been removed. Destruction by iconoclasts is also possible, if the mosaic had a pagan motif. Some shght support is given to this theory by the fact that the mosaic in the store-room was also destroyed and wear would be less likely in this case. THE LARGE STONE BASE A large stone slab was found lying in the vestibule, near one of the projecting walls (Plate VB). The stone is 24| in. by 2 1 | in. by 5£ in. in size. Three of the undersides are chamfered, the remaining side being vertical. The upper part of the stone contains four holes each If in. deep, 1| in. square at the bottom, widening to 2£ in. at the top. Two of the holes still contain iron rods cemented in with lead. The stone had clearly been fixed in a cavity in the ground, with the vertical side against a wall. The chamfered edges fitted into a similarly shaped cavity and the edges bore extensive traces of a hard white cement. The top was quite rough and dirty, whilst the underside was well trimmed and clean. The position in which the stone was found was not its original one, and it had clearly been re-used. When it had faUen from its position in the vestibule there was already a thin layer of soil over the floor and a flint had fallen from the wall. There was also a 1 in. line of red paint along one edge of the underside of the stone. This could only have been put on if the stone had been subsequently incorporated into a wall. Since the stone was re-used the questions arise, where did it come from and what was it used for ? It is certain that it had not been let into the floors of Temple I, at least in those places where there were tessellated floors. It is possible that it could have stood outside the east portico wall to form the base of a votive stone as at the Large Temple, St. Helena's School, Colchester (5), but this is not yet known. It could have belonged to Temple I I but this is yet to be determined. As far as purpose is concerned, it is clearly a base. The base of a statue is indicated at first sight, the iron bars being simply used for keying another stone to the base, but it is difficult to visualize anything with four legs placed so symmetrically. This is unlikely, however, since the upper surface was well worn and dirty, and the lead cements were not quite flush with the surface. The iron bars must have supported something, probably well above the base and it must have been heavy, if the weight of the base is anything to judge by. THE IRON FITTINGS Some unusual iron objects were found in the vestibule (see Fig. 11) and near the celia wall. They were mixed with the rubble of the walls 17 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD and roofs, well above the floor level and only a few inches below the top of the wall. This suggests that they were part of the timber structure of the roof and fittings of the door into the celia. There were a few nails of common type ; an X-shaped object (Fig. 11.3) ; a large dome-headed rivet through a " U " shaped staple (Fig. 11.9) ; a number of large split staples (Fig. 11.7) and a number of large rivets, one exceptionally large (Figs. 11.6 and 11.11). Detailed descriptions are given in Part V, but clearly they were all used on woodwork of considerable dimensions. Some were undoubtedly from the lintels of a door into the celia (see later under Celia), but others must have been from beams in the roof structure. THE CORRIDORS AND BLACKSMITH'S SHOP A view of each corridor is given in Plate IV. The only point of interest in the north and north-east corridors is the repairs made with pink and white tesserse. A roughly made hinge was found on the floor of the north-east corridor. Associated with it was a ring on a large spht staple. There certainly was not a door in this part of the temple so that the hinge must have come from some structure in the roof or on the wall. A wooden shutter to let in the light is a possibility. A large nail was driven through the plaster and into the portico wall of the south-east corridor. If this was the nail used to attach a votive offering to the wall, the object had gone. The interesting feature of the south corridor is the large gap in the wall and floor at the east end (Plate IVo). It appears that in late Roman times, an attempt had been made to dig a pit. Perhaps the solidarity of the floor deterred those concerned. The pit was not completed and the loose tesserse from the floor were left behind. The vestibule occupies much of the east corridor and had a mosaic floor. The rest of the corridor floors are plain red tesserse, very well preserved. All the corridor walls were plastered with line designs (see Part IV) but there were no quarter-round mouldings or skirtings. In the west corridor and the west ends of the north and south corridors, the floor was covered with a thick layer of blackened earth and charcoal, mixed with much iron slag, and partly made iron objects. This feature also extended into the west projecting room. In the west corridor were a number of shallow holes (Plate IVB) which however were deep enough to have been the post holes of some wood structure. No doubt the-walls of the temple helped support the structure so that the holes of any posts need only be shallow. Most holes were about 6 in. diameter and. 6 in. deep. One was 1 ft. 1 in. square and 10 in. deep. Clearly the post-holes were associated with the occupant who worked iron in the corridor, since the holes were not found elsewhere. 18 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD SaiJs 8 FIG. 9. The Brooches. All A. 19 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD The area occupied by the blacksmith—for such he must have been— was indicated very clearly. Not only did the charcoal finish at well defined points, but the plaster had fallen from the walls, although outside the area of the shop it had survived on all walls. The blacksmith must have had some type of furnace, but it had not survived, at least not completely. In the north half of the west corridor was a clay ramp and a hole in the celia wall (Plate VID). I t may have been associated with a furnace since drops of solidified iron were found on the tesserse by the side of it. The clay sealed a coin dated to A.D.335-341. Near the south-west corner of the celia was a rough triangular hearth made of tiles, chalk and flints. It showed considerable signs of burning and there was much charcoal. Twenty three samples of the charcoal were examined and were identified as ash, hazel and beech. The burnt layer contained much slag and many fragments of iron. One, was clearly a knife, one may have been a sickle and there were a number of nails and one or two rivets. The last two were no doubt from the roof. From the evidence it is clear that a blacksmith used the west corridor as his shop some time during the second half of the fourth century (Part II). The wall plaster and flints and chalk from the walls and roofing tiles had collapsed over his rubble. THE WEST PROJECTING ROOM (THE STORE OR STRONG ROOM) This room had been built originally during Phase A of the temple and was retained in use until the last, even by the blacksmith. It was of one build with the portico walls and was therefore not an annexe like those attached to so many other temples. An annexe of similar size to the room at Springhead was attached to the north wall of the temple at Frilford (7). The use to which the room was put, presents a difficult problem, to which however there are a few clues. First the walls are about 1 in.-2 in. thicker than the remainder of the temple walls. Second, the courses of bonding tiles were carried right round the walls and right through the walls. In the rest of the temple only the quoins had bonding tiles on the outsides, although the porch wings had tiles along the complete length, but only on the outside. Since the walls of the room were plastered, the bonding tiles were not decorative, but to give extra stability. Both of these features indicate a store room, although there were no signs of a door into it (Plate IIB). THE ENTRANCE TO THE CELLA The entrance from the vestibule to the cella was originally symmetrical. The full length of the wall may be seen on the south side 20 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD of the entrance. For some reason, part of the wall on the north was destroyed. I t appears that the two symmetrical curved grooves in the cella projecting from the wall (Fig. 1 and Plate VIB) were associated with the entrance. They were not formed by a door scraping over the surface, since all the tesserse were well formed. The floor foundation is most solid and there can be little doubt that the grooves were made deliberately. One suggestion is that the doors (if any) were pivoted in sockets and the grooves allowed the doors to be fitted into place in the sockets.1 Unfortunately there were no signs of iron or stone sockets, although there were cavities about 3 in. diameter and 3 in. deep.- Neither were there any post holes for door jambs, although the jambs could have been supported by the lintels. It is possible that there were no doors but simply screens and the grooves were associated with them in some way. Whatever the situation during the earher phases of the temple, it is certain that there was a door during the last phase of its existence. Part of the north entrance wall was destroyed, and replaced presumably by a wooden screen, since the grooves left by it may be seen in the cement surfacing of the wall (Fig. 1). Two pairs of substantial hinges were found just in front of the screen, which presumably indicates a fairly substantial door into the cella. Again there was no provision for post holes for jambs so that the door posts must have been supported by lintels. Along the north wall of the cella were a number of spht staples and along the south wall a number of large nails. They were virtually in a line and gave the impression that they had come from a fallen beam. These could have been the door jambs. THE CELLA MOSAIC The tessellated floor of the cella was red everywhere except for the small square mosaic just in front of the entrance. There are some interesting features and aspects of this mosaic. It is almost four feet square, including the surround of four rows of cream tesserae (Plate I I IA) . The colour scheme of the mosaic is described in Part IV. At first sight, the pattern appears to be symmetrical. Close examination, however, reveals dehberate inconsistencies. For example, along the north the first row of diamonds is blue/black. On the south edge the first row is red. Whether this was due to a whim of the architect or whether it had some rehgious significance is difficult to say, but it is interesting to note that the dimensions of the temple are not quite symmetrical. For example, the north corridor has a length of 33 ft. 3 in. and the south corridor 33 ft. 7 in. 1 I am indebted to Professor Richmond for this suggestion. 21 6 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD 1 I 10 II o 8 FIG. 10. The Bronze Objeots. All • 22 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD The position of the mosaic undoubtedly had some significance. I t was placed strategically in front of the altar and cult statue on its suggestus, and was probably where the priest had to stand during some important part of the ritual. Such mosaics have been found in the same position at Great Chesterford, Essex (10) and Harfleur, France (4). Further significance was given to these mosaics in some temples. In their centre are the remains of pedestals on which stood a statue or altar. There is a square socket at Harfleur and a circular pattern in the mosaic at Great Chesterford. There was no such provision at Springhead. However, at Springhead there was a mark on the mosaic where the tesserse has worn less than elsewhere, indicating that something had stood on the mosaic at the protected point. The mark was not in the centre of the panel and in fact was quite asymmetrical, although roughly rectangular 15 in. by 10 in. THE SEEDS AND INCENSE CUP. In the report on Site A (1) reference was made to a cluster of carbonized seed of Atriplex patula in the bakery. It was concluded that these seeds had been specially collected for some purpose, probably to grow as fodder, but no association with bread was implied. A further cluster of carbonized seeds was found on the floor of the cella, which again were identified as those of Atriplex patula. They had clearly fallen from a small incense cup (Fig. 8.17) nearby. It appears therefore that the seeds could well have been burnt as a type of offering. A possible reason for this may be given. Further evidence has been found regarding Atriplex, which definitely associated it with bread. I t was found, for example, in winnowing pits at Silchester (11) and Newstead (12) for the Roman period ; Itford Hill for the Late Bronze Age (13) and Meare, Somerset for the Early Iron Age (13). Some carbonized buns from Glastonbury contained Atriplex (13). Since Atriplex is not normally a weed associated with wheat, and since the seeds in the bakery were collected specially, it could be that Atriplex was used in bread production. Now it was common practice in Ancient Rome to make offerings of specially made bread. Associated with this was the sprinkling of grains of salt on the sacrificial fire (14). When Atriplex is burnt, all varieties contain a high proportion of salt, in some cases as much as 40 per cent. It may be possible, therefore, that Atriplex seeds were sprinkled into the fire, as associated with both bread and salt. (Frank Jenkins reaches a similar conclusion but by a different approach—see Appendix 1.) Another point is that none of the incense bowls found at Springhead snowed signs of burning. It appears that nothmg was 23 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD burnt in them, but they were just receptacles used for sprinkling the offerings on to the fire. The one found in the cella was quite small and clearly was used simply for holding the seeds. THE ALTAR A fine altar and its base were found lying in the cella (Plate Vo and D). The base, which was incomplete, was lying near the hollow of the suggestus, with the altar near the centre of the cella. The altar is uninscribed, although the front panel is well faced. There were no traces of plaster wliich could have covered the altar and have been inscribed. The back of the altar is quite rough and obviously meant to be hidden. In view of the fact that the base was found near the hollow of the suggestus, it seems likely that the altar stood in the hollow. The other interesting feature of the altar is the tripod arrangement on top. Three equidistant holes were made outside the focus, and three iron rods were cemented into them with lead. A fragment of iron was also embedded in the tessellated floor in front of the top of the altar, probably the remams of the object which had been supported on the tripod. To be embedded in the floor so deeply, it could not have have been rusted and indicated that the altar was toppled over when in a good state of preservation. I t is more likely that the object on top was a brazier in which the sacred fire would be kindled and the offerings made. The suggestion that the focus was used for such a purpose is difficult to believe in view of its small size. The use of a brazier either on the altar or elsewhere seems much more likely. THE SUGGESTUS The apsidal or hollow-fronted suggestus is unusual in this country, but more frequent in Northern Gaul for example the Jupiter-tempel in the temple area at Trier (4). The temple at Caerwent had an apse (not a suggestus) in a similar position, but projecting backwards into the portico (15). The width of the suggestus, not including the projecting sides, is 3 ft. 3 in. This can only mean that something stood on it, presumably the cult statue. Votive offerings (see later) may also have been placed here. It is possible that the projecting sides supported pillars and that there was a roof over all to make a shrine. The suggestus was probably not much higher than at present, certainly not higher than the altar placed in front of it (Plate 11A). That the spot occupied by the suggestus was considered to be sacred can hardly be doubted. Apart from the significance of the suggestus itself, it was built over the ritual hearth. 24 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD psa H Q 10 es£!is^&&3^&i33, FIG. 11. The Iron Objects. Alii. 25 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD THE CULT OF TEMPLE I A detailed consideration of the cult of Temple I has kindly been made by Frank Jenkins. His views, for which I am most grateful, are given in full in Appendix 1. No remains of the cult statue were found to correspond with the fragments of fife size statuary at Silchester (16). However, several fragments of an unusUal clay figurine of Venus were found in the cella, near the suggestus on which it may have well stood. A full account of this figurine is given in Appendix 1 by Frank Jenkins. The name Vagniacse is significant. An altar found in Cologne has a carving under the inscription, on which is an altar with a tripod. The inscription is to the goddess Vagdervercustia1 and the derivation of this name and that of Vagniacse, may well be associated with springs. A votive offering found in the cella was a small lead axe or mattock. In one of the porch wings was found a bronze thumb, probably connected with a healing cult, and just outside the temple was found a bronze tack, driven through a fragment of silver sheet, possibly part of a votive plaque. A part of a gilted bronze letter A was found in the gap between the suggestus and the cella wall. THE SUPERSTRUCTURE The only certainty about the superstructure is that all the roofs were well tiled. Many hundreds of broken and a few complete tiles were found in the cella, portico, porch-wings and store-room. Their only interest were the imprints of a few animals, a child's foot and a hobnailed shoe. Except for the porch wings and the store-room, all the walls were of the same thickness of 21 in. This is of some significance and may afford a clue regarding the superstructure of such temples. Several reconstructions of concentric-sqaa,Te temples have been made, including the one at Harlow (4) and the temple in Insula XXXV at Silchester (16). Most of the temples of this type have been visualized with a comparatively high central tower, containing clerestory windows, with a pent-roof over the portico, supported on dwarf columns on the portico walls. For this arrangement, much thicker cella walls than portico walls are likely to be necessary. In Table 15 dimensions for eight temples are given and only in two cases is there a difference in thickness. Again the postulation of an open portico must be open to considerable doubt. The tessellated floors would not have survived many winters thus exposed. In any case, such an arrangement is not the best for our climate. 1 I am indebted to Frank Jenkins, F.S.A. for this information. 26 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD I t is difficult to say whether the walls were of wood or stone. The latter is perhaps to be preferred in view of the scarcity of nails in the rubble of the building. Not more than forty were found in the entire building and the only large staples were found in the cella, associated with the cella door. Even the tiles were generally cemented in place on the roof. Only a few were found with nail holes, but many had mortar on the flanges of the tegulse. Evidence such as the above seems to point to solid portico walls, not very high. The cella walls might have been solid or supported pillars, the latter being quite likely and it certainly was the case in Temple II. With this arrangement, the cella, corridors, porch and store room may have had separate gabled roofs. Goodchild and Kirk have postulated a similar arrangement at Woodeaton where the portico walls were thicker than the cella walls. The outer walls in the octagonal temple at Chew Stoke actually had buttresses. The further suggestion that part of the cella might have been open to the sky is unlikely at Springhead. The floors would have been damaged by the rain and there were no means of foaming it off. The above arrangement does raise some problems with regard to lighting, which however are not insurmountable. No lamps were found so perhaps this method can be discounted. The temple entrance faced east, and would be ht by the sun at certain times and for this method of hghting, columns would be expected around the cella walls. Skylights are also a possibility and an example of this has already been mentioned. PART IV ARCHITECTURAL DETAILS INTRODUCTION Probably no other temple of the type found in the United Kingdom was as well preserved as Temple I at Springhead. It has been possible to obtain dating evidence for the various phases of its construction and obtain some insight into the purpose for which the various parts of the building were used. I t seems worth while to record in detail the dimensions and construction of the temple since opportunity for such detailed analysis so rarely arises. In part IV these details are provided and discussed where appropriate. FOUNDATIONS The foundations for Phase A temple were made by cutting a trench about 12 in. deep and the width of the wall, and filling with loose flints. The flints were comparatively small, about 3 in. across. 27 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD f > t » D FIG. 12. Miscellaneous Objects. All £. 28 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD In subsequent reconstructions the previous walls were used as foundations. Normally one was built exactly on top of the other, except in the porch where there was an offset (see Fig. 2). In subsequent additions, the wings (antse) for example, the same procedure was adopted as for Phase A, except for the suggestus and projecting vestibule walls where the masonry was laid on to a thin mortar or flint layer. Such insecure foundations allowed the features concerned to subside, the suggestus considerably. In subsiding, it damaged the associated tessellated floor. THE WALLS (a) General The extant walls were constructed from any available stone, usually flint. However, the centre parts frequently had a chalk rubble composition. Here and there a little sandstone was used. As far as the upper walls are concerned, some evidence survived. The extant walls vary in height from 12 in. to 16 in. above the tessellated floors. From the amount of plaster recovered, the cella and outer walls were at least 12 in. to 18 in. higher than this originally. The rubble from the upper walls, which was considerable, was composed of large flints and chalk blocks. Many of the chalk blocks were of quite large size (16 in. by 8 in. by 5 in.) and had been carefully shaped. There were no traces of daub and surprisingly few nails. The evidence such as it is, does not indicate timber walls. It has been shown in Part III that the outer wahs were possibly solid, the inner walls possibly supporting pillars. There is nothing in the evidence against this theory, although there is nothing positive to support it: Bonding tiles were not generally used in the main waUs. One course was found in the store-room about 12 in. above the tessellated floor. The tiles are 1£ in. thick and extend right through the thickness of the walls. Similar bonding tiles were used for the wing walls, these being at floor level inside and out. They projected about 1£ in. from the outside of the wall and 3 in. from the inside of the wall. The former was a convenient limit to the plaster and the latter was covered over with a very hard cement. All walls were made in layers. The imperfections of the surface of the first flint layer would be filled with cement and smoothed down. A second layer of- flints would be placed on top, filled with cement and so on. (6) Quoins All important quoins were strengthened with bonding tiles. These were the outer corners of the cella, outer corners of the temple, store- 29 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD room, wings and the walls flanldng the entrance of the store-room (see Plate VA). The usual arrangement is illustrated in Fig. 13.4. There were probably two groups of two tiles, although usually only the impression of the upper tile remains. The tiles were bonded with opus signinum, although the main body of the wahs was bonded with ordinary yellow mortar. (c) The Suggestus and Vestibule Walls The large suggestus was adequately reinforced throughout with bonding tiles and opus signinum. Its construction is illustrated in Fig. 13.1. Apart from the tiles, the other building materials were flint on the outside, well faced, and flint and chalk rubble in the centre. The vestibule walls were rather crudely made, mainly with chalk blocks and rubble. They were, however, reinforced with bonding tiles (Fig. 13.2). It is well to remember, in examining these structures, that they were all covered with painted plaster. THE FLOORS (a) The Porch With the exception of the wings wliich had no made up floors, all parts of the building had well-made floors. The entrance through the porch had two superimposed floors, both essentially consisting of a gravel surface and flint base (Section, Fig. 2). The porch mosaic is quite differently constructed to the other tessellated floors. The base is a layer of flints, followed by a layer of mortar in which the tesserse are set (Section, Fig. 3). The tesserse are rather crude, on average about 2 in. deep and \ in. to 1 in. square. They are made of coloured tiles, although the blue effect is obtained by the use of overbaked tiles. The mosaic itself has a simple geometric design of a central diamond with a circle on each side. The colour arrangement of the diamond is as follows : the single centre tesserse is red ; the first inner diamond is yellow ; the second diamond is blue ; the third diamond is yellow ; the fourth and fifth diamonds are red ; the sixth diamond is pink ; the seventh diamond is yellow and the outer diamond is red. The arrangement of the two circles is as foUows : there is an outer circle of | in. square red tesserse ; the first circle inwards also has red tesserse but is 1 in. wide ; the second circle in has \ in. wide yellow tesserse and the third circle in has | in. red tesserse. This circle is divided into quadrants by a cross of red tessera?. In each of the centre angles of the cross are four blue 1£ in. square tesserae, the remainder of the quadrants being filled with yellow tesserse. The diameter of the 30 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD THE SUOSESTUS SOUTH VESTIBULE WALL PLAN THE ARRANGEMENT OF TILES IS IDENTICAL ALL ABOUND THE STRUCTURE J FLINTS FRONT ELEVATJON CHALK FLINT FOUNDATION FRONT ELEVATtQN SIDE EiEVATIOH PLAN AND SECTION OF SW QUOIN QP CELLA T H E PLASTER IN THE CELLA CtUA FLOP* f SIDE ELEVATION ICALI |N MIT. SIDE ELEVATION rcAtt IM ttir FIG. 13. 1-4, Architectural Details. All \. 31 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD outer circle is 15J in. and the width of the diamond, apex to apex, 17£ in. The patterns were surrounded by red tesserse of varying size. The whole mosaic panel was only 54£ in. by 18 in. and was surrounded by herringbone pattern floors. The tiles used for the latter floors were about 4 in. deep. The drainage channel at the right hand side of the pattern was 3£ in. wide at the west end, 5 in. wide at the east end and. about 1 in. deep. (6) The Vestibule, Corridors and Store-Room The floors of the vestibule, corridors and store-room are all much the same construction. The tesserae, all made from tiles, are about \ in. to | in. cubes. They are cemented together, and also to a J in. layer of opus signinum, with a hard white cement. Under the opus signinum is 2 | in. mortar, reinforced with pebbles, followed by 3 in. small flints and 6 in. rammed chalk (Section Fig. 3). The composite structure was very solid indeed and there had been no sinking of the floor at any point. There is no doubt that the vestibule once contained a mosaic. First parts of a double line of white tesserse survived across the corridor to show that the mosaic had been separated from the plain red tessellated floor. Parts of the corners of the mosaic also survived, particularly in the N.E. corner. Here there was a corner consisting of a three row tesserse band coloured white ; inside this a two row tesserse band coloured black followed by traces of another white band (see Fig. 1). No traces of the tessellated floor in the store-room survived, although the base of the floor was untouched and was as elsewhere. Only one or two tesserse remained in the rubble found above the floor foundation, and on this evidence alone, and the importance of the room, a mosaic panel may perhaps be inferred. (c) The CeUa As far as the cella floor is concerned, the tesserse are as those in the corridors, but the base is different. The tesserse are cemented, as before, to a | in. layer of opus signinum with a white mortar. This is followed by 2\ in. yellow mortar, reinforced with pebbles and 10 in. small flints (Fig. 3). This .floor subsided badly in the centre, there being a difference of 6 in. between the edges and the centre. The mosaic in the cella was a geometric design of triangles and diamonds. The panel was contained in a four deep tesserse band coloured white. The outer dimensions of the band were 3 ft. 11 in. (W.-E.) by 3 ft. 1\ in. (N.-S.). Immediately inside the band was a row of triangles. Those on the north and south sides consisted of an outer triangle of a single row of white tesserae, with an inner single 32 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD triangular shaped tessera coloured cream. The triangles on the west and east were similar except that the outer row of tesserse was black and the inner triangle red. The colours of the rows of diamonds reading from the north were black, white, cream, red, black, white, cream, red, black, white, cream, red, black, white, cream, red, black, white and red. It will be noted that the start and finish of the rows is not symmetrical. The colours across the rows were constant and the black had an element of blue about it. THE PLASTEB (a) Outside the Outer Walls All the walls, at all periods, were plastered externally and internally, with the exception of the insides of the walls of the wings. Very few traces of plaster from the earlier periods remained so that no attempt at reconstruction is possible. Neither is much evidence available for the plaster used outside the walls, at any period. What there is indicates that at aU periods a fairly plain pink colouring was used. (6) Inside the Outer Walls Internally a reasonable amount of plaster survived. Much was in situ to a height of about 12 in., particularly on the insides of the outer walls of the ambulatory, except that on these walls it had been completely stripped off for the extent of the blacksmith's shop. Where the plaster had survived on the walls, it had normally been lifted off (anciently) by frosts and the gap thus created, filled with soil. The following is the type of arrangement for the inside walls of the porch and the ambulatory walls. There was no quarter-round moulding or skirting but a simple plastered wall. First came 17 in. plain pink, except that at 4 ft. intervals there were three vertical bands each \ in. wide and 4£ in. apart. They may have meant to represent fluted columns, although no traces of a capital painting were found. Above this was a 1 in. yellow band (with a horizontal streak of red through it at intervals) ; 3 in. red band ; 4 in. cream band (with horizontal and vertical thin red lines on it to form irregular rectangles) ; 1 in. gold band ; 3 in. red band ; \ in. buff band ; 3 in. green band ; 1 in. orange band and a 3 in. cream band. All the upper bands were liable to have different coloured streaks in them, diagonal lines and bands. There seemed to be no system to the designs. The arrangement of the bands above, except for the bottom 12 in. was deduced from reconstructed plaster fallen from the walls. The total height of wall obtained is thus 36\ in. which is clearly a minimum value. Undoubtedly much plaster was lost. 33 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD It may be presumed that the plaster on the inside walls of the west corridor was similar to that in the other corridors. Some was indeed found over the burnt layer which confirms this. However, the amount of plaster from the walls in the store-room was not great. That surviving indicated quite a different colour scheme to that in the corridors. The colours seemed more vivid and instead of horizontal bands of colour being broken up by thin bands of other colours, the thick bands were at all angles including horizontal. There were no ISECTION n. =1 2 ' O ' SECTION FIGURE 6 13 Fio. 14. Position of Sections. signs of animals of any type, although a fragment or two seemed to indicate a leaf. (c) Outside the Cella Walls There was very httle plaster in situ on these walls, and for the north and south walls very Httle even on the floors. Whether this indicates a low wall is impossible to say. The arrangement was quite different to that on the outer walls. The lower pink band was only 1 | in. wide. Above this was a most 34 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD complex design which varied from foot to foot along the length of the wall. It seems as though the painter had a completely free hand and made up varied designs as he went along. A description of the 30 in. of plaster in situ will best explain the situation. Above the 1 | in. band was a 4 in. band. The first 10 in. of this (along its length) had a pink background followed by 8 in. of a cream background. Over both colours was a criss-cross of orange lines. At 18 in. along, there was a 1 in. vertical pink band ; at 19 in. to 21 in. a cream vertical band ; at 21 in. to 22 in. a vertical fawn band and from 22 in. to 30 in. the plaster was white, foUowed by a thin vertical streak of red. Above this complex 4 in. band was a £ in. gold band, followed by £ m. pink and by \ in. red, but only for 18 in. of the 30 in. The rest was a plain pink band. (d) Outside the West Cella Wall The west waU was quite different and the amount of plaster considerable. If the amount of plaster is any indication, the west ceUa wah was higher than the others. The colours and patterns on the plaster were quite like the ceUa to be described next. Again there was no quarter-round moulding. The colour sequence was as follows : nearest the floor was a 3 in. grey band foUowed by \ in. red ; 1 in. cream ; \ in. red ; 3 in. cream ; \ in. red ; 3 in. cream ; £ in. red ; 1 in. cream ; 2 in. red ; 6£ in. cream and Z\ in. red. Many of the lower bands were probably plain, although the possibility of a few vertical and obhque lines cannot be excluded. Certainly the 6 | in. cream band was decorated with a variety of designs. One was a gold festoon with a 4 in. radius, with a tangential \ in. wide green hne about 6 in. long. A 1 in. vertical red stripe was also found. Above the levels quoted above (24 in.) insufficient plaster survived to make a reconstruction. However, it is quite clear that the decoration became even more fantastic. There were the usual horizontal bands of colour. The bands were decorated with a variety of designs, apparently varying considerably aU over the temple. Here there were festoons and half festoons in aU colours ; obhque lines of aU colours ; criss-cross patterns and " hanging baUoons " or circles of colour connected to the line above with a streak of colour. (e) The inside Cella Walls If the colours and designs of the outside west ceUa waU were unusual, those of the inner cella waUs were even more so. There was nothmg dismal and depressing about the cult associated with this Roman temple. First, there was a quarter-round moulding at the junction of aU waUs and floors, including the suggestus. This moulding had a radius 35 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD of 3 in. Above this, there was 25 in. of plaster to form a skirting (Fig. 13.3). Above 25 in. the plaster sloped inwards at an angle to the horizontal of 40 degrees for a vertical depth of about 1 in. There was then a normal covering of plaster on the waU, which probably did not extend upwards more than 12 in. This gives a total plaster height (and possibly waU) of 42 in. The skirting had comparatively simple colour designs. The quarter-round moulding was painted red. There was then a sequence of bands as foUows : off-white |- in. ; red I in. ; off-white \ in. ; green 2 in. ; yeUow \ in. ; off-white \ in. ; dark-red \ in. ; off-white | in. ; dark-red \ in. ; coral pink 1 in. ; off-white 1 in. ; coral-pink 1 in. ; dark-red \ in. ; off-white \ in. ; dark-red \ in. ; off-white \ in. ; dark-red \ in. ; yellow 7 in. ; off-white 6 in. and 1 in. dark-red. The slope of the skirting was also dark-red. The broad yeUow band had a series of vague brush marks in red along its complete length. For the plaster above the skirting there was again a great variety of designs and vivid colours. These were simUar to those described for outside of the west-ceUa waU. In addition there was a series of large blobs of colours, in white, yeUow and red. Very httle plaster was found near the suggestus and no skirting plaster. As far as the evidence goes, its treatment was simUar to that of the waUs of the ceUa. The vestibule waUs were treated in plaster. The stonework of these waUs has a sloping and a straight portion, but the plaster converted this to a stepped-off effect. The colours were similar to those on the outside of the ceUa waUs. The whole effect must have been startling and gay. The painting was also of a very high standard. The colours were laid on evenly, and there was no smudging where additional lines and colours were apphed. The edges of aU the bands were sharp and straight. THE ROOE The roof was tUed with the usual tegulse and imbrices. The tUes were found in great numbers, some complete. Very few marks were found on them, but the impressions of a sandal, a child's bare foot, one or two animals, the usual parabolic curves and a few finger and thumb prints were found. The method of securing the tUes to the wooden structure of the roof was quite clear. Only a very few tUes with nail holes were found and presumably these were the bottom ones around the edges. The others were buUt up from these firmly secured tUes, flange to flange. The flanges were then covered HberaUy with mortar and the imbrex then placed over, there being sufficient mortar to fill the curve. 36 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD PLAN OF ROMANO-BRITISH SETTLEMENT SPRINGHEAD (ONETREE HELD) / PRESUMED FOURTH CENTURY OCCUPATION AREA BAKERY SITE A / V SITE E / MANY IRREGULAR ,-' x / CROP MARKS AT / / P*^ i i i i i n u I I I I I I I I i i 1! H WELL WITH SURROUND SITE D ^°«o sr^s*' HEDC EMBANKMENT OF MODERN WATLING STREET (A2) EXCAVATED CROP MARKS PRESUMED a |Q 50 30 4Q 60 SCALE OP FEET Fia. 16. Plan of Site at Springhead. PAST V THE FINDS IHTEODTJOTION There were the usual number of finds from aU periods with a considerable amount of pottery from the early periods. The foUowing is an index of aU the finds made. Table or Appendix Table 2. Table 3. Table 4. Objects The Coins. Decorated Samian Ware. Mortaria. 37 7 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Table or Appendix Table 5. Table 6. Table 7. Table 8. Table 9. Table 10. Table 11. Table 12. Table 13. Appendix 1. Appendix 2 Appendix 3. Objects Plain Samian Ware. Pottery, Standard Forms. Coarse Ware. The Fibulae. The Bronze Objects. The Iron Objects. The Glass. MisceUaneous Objects. The Bones. The Pseudo-Venus. A Note on the Flanged Sherd. Petrographical Report. GENERAL COMMENTS Most of the fibulse found were fairly common. However, attention is drawn to the fine gUt brooch of fourth-century date (Brooch No. 1, Table 8, Fig. 9.1). Some interesting iron objects were found on the temple floor. Attention is particularly drawn to Objects Nos. 4, 5, 6, 7, 8, 10 and 11, Table 10. POTTERY Of the unusual types of pottery found, special attention is drawn to the incense cups (Sherds Nos. 17-21) Table 7 and Figs. 8.17 to 8.21. In the Site B (2) report, a distinction was made between pottery in general and certam common standard forms. Forms Sl-31 were previously covered and a further eleven standard forms have been added in the present report to give S.32-42 (Table 6). AU these forms are common and fairly closely dated. CULT AND VOTIVE OBJECTS The two most important of these objects, to which special attention is drawn, are the altar and figurine of Venus, The fact that they were found in situ in the ceUa gives added importance to their discovery. The note by Frank Jenkins on the figurine is of great interest and importance and the author is very much indebted to him for his help. Other objects which may be noted include the thumb (object No. 13, Table 9) ; part of a letter A (Object No. 24, Table 9) and the seeds of Atriplex patula (Object No. 9, Table 12). 38 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD TABLE 1 DATING OF STRATA Phase Datable Object Zl. Coin Coin Brooch Decorated Samian Plain Samian Coarse Ware Coarse Ware Standard Forms of Coarse Ware. (See Par t 5) Coarse Ware Note : This stratum may Z2. Brooch Decorated Samian Coarse Ware Coarse Ware Coarse Ware Coarse Ware Coarse Ware Standard Forms Note : This stratum may A. Coin Coin Coin Coin Brooch Brooch Brooch Brooch Decorated Samian Plain Samian Plain Samian Plain Samian Coarse Ware Coarse Ware Standard Forms Standard Forms Standard Forms Standard Forms Standard Forms Standard Forms Standard Forms Note : This stratum may B. Coin Coin Coin Date Claudian Claudian 1st half 1st cent. C. A.D.50 Claudian Claudian Claudian C. A.D.50 Claudian be dated Claudian. Mid 1st cent. Flavian Claudian 0. A.D.50 A.D. 50-75 A.D. 50-70 C. A.D. 60 A.D. 50-80 A.D. 60-80 A.D. 60 A.D. 60-75 A.D. 60-75 A.D. 60-75 A.D. 60-75 C. A.D. 70 be dated Flavian. Domitian Domitian Vespasian Agrippa 1st century 1st century 1st century 1st century Last quarter 1st century 1st century 1st century 1st century A.D. 70-90 A.D. 50-75 A.D. 60-80 Late 1st century Late 1st century A.D. 75-90 A.D. 75-90 A.D. 75-90 A.D. 75-90 be dated late lst/early Hadrian Vespasian Vespasian Reference Table 2, 36 Table 2, 37 Table 8, 5 Table 3, 4 Table 5, 4 Table 7, 2 Table 7, 9 S.28 6 sherds of Fig. 10.16 Site B Report Arch. Cant., Vol. LXXI. Table 8, 3 Table 3, 2 Table 7, 3 Table 7, 4 Table 7, 10 Table 7, 14 Table 7, 16 4 of S.5 2 of S.13 1 of S.30 2 of S.32 2 of S.33 4 of S.34 1 of S.35 1 of S.36 Table 2, 12 Table 2, 14 Table 2, 16 Table 2, 7 Table 8, 4 Table 8, 6 Table 8, 7 Table 8, 8 Table 3, 5 Table 5, 6 Table 5, 8 Table 5, 11 Table 7, 8 Table 7, 15 1 of S.13 4 of S.20 6 of S.23 1 of S.39 4 of S.40 1 of S.41 2 of S.42 2nd century. Table 2, 1 Table 2, 15 Table 2, 20 39 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD vase Datable Object Coin Brooch Decorated Samian Plain Samian Plain Samian Plain Samian Plain Samian Plain Samian Plain Samian Plain Samian Mortarium Coarse Ware Coarse Ware Coarse Ware Coarse Ware Coarse Ware Standard Forms Standard Forms Standard Forms Standard Forms Standard Forms Note : This stratum C. Coin Brooch Decorated Samian Plain Samian Plain Samian Potter's Mark Mortarium Coarse Ware Coarse Ware Coarse Ware Coarse Ware Standard Forms Standard Forms Standard Forms Standard Forms Standard Forms Note : This stratum Dl Coin and Coin D2 Coin Coin Brooch Plain Samian Plain Samian Mortarium Mortarium Coarse Ware Standard Forms Date Sabina 1st century c. A.D. 140-150 Antonine Late 1st century Antonine 1st century 1st century 1st century Antonine Late lst/early 2nd century C. A.D. 50 o. A.D. 80-120 Late 1st century Late 1st century Late 1st century Late lst/early 2nd century Late lst/early 2nd century Hadrianic Antonine A.D. 75-90 may be dated early Antonine Antoninus Pius 2nd century C. A.D. 170 C. A.D. 175 Antonine o. A.D. 140-190 Late 2nd century Mid 1st century First half, 2nd century Late 1st century Early 2nd century Antonine Antonine Late 2nd century A.D. 75-90 A.D. 75-90 Reference Table 2, 34 Table 8, 9 Table 3, 1 Table 5, 5 Table 5, 7 Table 5, 9 Table 6, 10 Table 5, 12 Table 5, 13 Table 5, 15 Table 4, 5 Table 7, 1 Table 7, 5 Table 7, 12 Table 7, 19 Table 7, 20 1 of S.19 1 of S.26 1 of S.15 2 of S.6 2 of S.42 Table 2, 17 Table 8, 2 Table 3, 3 Table 5, 1 Table 5, 14 Table 5 Table 4, 1 Table 7, 6 Table 7, 11 Table 7, 18 Table 7, 21 2 of S.6 1 of S.22 8 of S.10 1 of S.40 1 of S.42 may be dated late 2nd century. Claudius II Carausius Carausius Postumus Probably 4th century C. A.D. 160 C. A.D. 175 Late 3rd century Late 2nd/4th century A.D. 160-190 Antonine Table 2, 5 Table 2, 10 Table 2, 11 Table 2, 21 Table 8, 1 Table 5, 2 Table 5, 3 Table 4, 2 Table 4, 3 and 4 Table 7, 13 1 of S.6 40 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Phase D3. Datable Object Date Reference Standard Forms Late 2nd century 3 of S. 10 Standard Forms 4th century 6 of S.3 Note : Probable date—1st half of 4th century. Coin Coin Standard Forms Standard Forms Standard Forms Note : Probable date- 4th century c. 335-341 4th century 4th century 4th century -2nd half of 4th century. Table 2, 18 Table 2, 25 4 of S.4 3 of S.7 1 of S.3 D4. Coin Coin Coin Coin Standard Forms Standard Forms Standard Forms Constantius II Constantius II c. 335-341 c. 330-335 4th century 4th century 4th century Table 2, 2 Table 2, 28 Table 2, 30 Table 2, 33 10 of S.4 5 of S.7 4 of S.3 Note : Date—2nd half of 4th century. TABLE 2 THE COINS (Reported on by R Coin No. Position 1 Figure 2, Stratum B3. In porch. 2 1 in. above floor, blacksmith's shop. 3 Under rubble in southern of two flanking -wings. Under tile repair of latest porch floor. Figure 6, Stratum Dl. Under latest mortar layer, outside west wall. Unstratified. Figure 2, Stratum A. Northern of two flanking wings. Under rubble in southern of two flanking wings. In rubble in southern of two flanking wings. A. G. Carson, Esq., M.A.) Description Hadrian 1 AE (Sestertius). RIC. II, p. 414, No. 581 (b) Variety. Constantius II. 3 AE. Q. HK. 89. (17). House of Valentinian I (uncertain attribution). 3 AE. Emperor standing right, hand on head of kneeling captive, cf. RIC. IX, p. 119, No. 15. Carausius. 3 AE. (Antoninianus). RIC. V.2, p. 536, No. 893. Claudius II. 3 AE. (Antoninianus). RIC. V. i, p. 233, No. 259. Arcadius or Honorius. 4 AE. Obv. Illegible. Bust diademed right. Rev. (SAL VS REIPVBLICAE). Victory advancing left, dragging captive by the hair. cf. RIC. IX, p. 133, No. 64 c. Agrippa. 2 AE (As). RIC. I, p. 108, Tiberius, 32. Prototype of cf. Tetricus II. c. A.D. 270-4. Obv. Radiate head right. Rev. Spes. Augg type crudely drawn, inscription illegible. Imitation Claudius II. 3 AE. Obv. Illegible. Head, radiate right. Rev. (CONSE) CRATIO, Eagle facing, wings outspread, head turned right. cf. RIC. V. i, p. 234, No. 266. 41 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Com No. Position 10 Figure 4, Stratum D2. In southern of two flanking wings, under latest mortar layer. 11 Figure 5, Stratum Dl. Under latest mortar layer outside west wall. 12 Figure 2, Stratum A. 13 Under rubble, immediately south of temple. 14 Figure 2, Stratum A. 15 Figure 4, Stratum Bl. In flint floor in southern of two flanking wings. 16 Figure 3, Stratum A. In vestibule. 17 Figure 4, Stratum C. In make-up of southern of two flanking wings. 18 2 in. above floor in vestibule. 19 Unstratified. 20 Figure 4, Stratum Bl. 21 Figure 4, Stratum Dl. 22 In rubble of south corridor. 23 Under rubble of north corridor. 24 Under rubble of north corridor. 25 On floor of north corridor. Description Carausius. 3 AE (Antoninianus). RIC. V.2, p. 534, No. 855. Carausius. 3 AE (Antoninianus). RIC. V.2, p. 534, No. 857 c. Domitian 2 AE (Dupondius). RIC. I I , p. 206, No. 417. Valentinian I, 3 AE. cf. RIC. IX, p. 44, Lug. No. 12. Domitian. 2 AE (As). RIC. II, p. 196, No. 335. Vespasian. 2 AE (As). RIC. II, p. 104, No. 763. Vespasian. 2 AE (As). Obv. IMP CAESAR VESPASIAN AVG COSH II. Head laureate right. . Rev. AEQVITAS AVGVSTI. In field S.C. Aequitas standing left, holding in right hand balance, in left sceptre. Antonius Pius. 2 AE (As). RIC. Il l , p. 98, No. 533. 4th century. Mis-struck 3 AE. Arcadius. 4 AE. Obv. (DN)ARCADIUS (PF AVG), Bust diademed and draped, right. Rev. (VICTORIA AVGGG) Victory, advancing left, holding wreath and palm branch, cf. RIC. IX, p. 52, Trier 44 (d). Vespasian. 2 AE (As). RIC. II, p. 75, No. 503. Postumus. 3 AE (Antoninianus). RIC. V.2, p. 341, No. 54. Valens. 3 AE. RIC. IX, p. 97, Aquileia, No. 18 (a). Imitation Tetricus II. 3 AE. Obv. (CPE)TETRICUS CAES. Youthful bust, radiate and draped, right. Rev. P A (XAV)G Pax standing left, with olive branch and vertical sceptre, cf. RIC. V.2, p. 422, No. 248. Imitation Claudius II. c. A.D. 270. Obv. DIVO CLAV(DIO) Bust, radiate, right. Rev.CONSE(CRATIO)Eagle,faoing,head, right. . cf. RIC. VI, p. 234, No. 266. Imitation diademed head. Minim weight 3-75 grains, c. A.D. 335-41. Obv. Head right. Rev. 1 Gloria Exercitus. One standard. 42 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Coin No. Position 26 In rubble of north corridor. 27 In rubble over Blacksmith's shop. 28 In burnt layer, blacksmith's shop. 29 In rubble over blacksmith's shop. 30 In clay ramp in blacksmith's shop. 31 In rubble over north corridor. 32 Under row of tiles in vestibule. 33 In burnt layer in store room. 34 Figure 3, Stratum Bl. 35 In rubble over store room. 36 Figure 3, Stratum Zl. 37 Figure 3, Stratum Zl. 38 In rubble above store-room. 39 Under the tile fall in the northern of two flanking wings. Description Domitian. 2 AE (As). RIC. II, Vesp., p. 108, No. 791 (a). Crispus. 3 AE. Obv. (CRIS)PVS NOB C. Bust helmeted and cuirassed right. Rev. BEATA TRANQUILLITAS. On altar, inscribed VOTIS XX, a globe ; above three stars. Mint mark, PTRV Trier, cf. K.213 (18). Constantius II. 3 AE. HK.187. Constantine I. 3 AE. K.383. Imitation. Minim, weight 3-5 grains, c. 335-341. Obv. Head right. Rev. Type of Gloria Exercitus—one standard. Constantine I. 3 AE. K.318. Augustus. 2 AE (As). RIC. I, Aug. 360. Imitation Constantinopolis issue. Minim, weight 9 grains, c. A.D. 330-335. Obv. Helmeted head, left. Rev. Wolf and twins. Poor representation. H.K.59. Sabina. Imitation As. c. A.D. 117-138. Obv. SABINA AUGUSTA. Bust, diademed, right, with long plait at back. Rev. S.C. Inscription off flan. 1 VENERI GENETRICI Venus standing left. cf. RIC. I I , p. 479, No. 1050. Valentinian I. 3 AE. CK.311. Imitation Claudius I 2 AE (As). Obv. (TI) CLAV DIVS CAESAR AVG (PM TR P IMP P P) Head, bare left. Rev. Minerva advancing right, hurling javelin and holding shield. In field, Q p cf. RIC. I, p. 129, No. 66. Imitation Claudius I. 2 AE (As). As for coin No. 90. Carausius (Irregular). 3 AE (Antoninianus). Obv. IMP CARAVSI(VS . . . . ) . Bust, radiate right. Rev. Illegible. Pax type 1 cf. RIC. V.2, p. 535, N. 880A. Claudius I I . 3 AS (Antoninianus). RIC. V.i, p. 219, No. 102. 43 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD TABLE 3 DECORATED SAMIAN WARE Form Figure No. No. Position Drag. 37 7.1 Figure 2, Stratum B In porch. Drag. 37 7.2 Drag. 37 7.3 Description (a) Ovolo. The ovolo is a simple double " U ", with a simple tongue with circle. It could have been produced any time during the 1st and 2nd centuries. (6) A bead row below the ovolo, again could be Flavian or 2nd century, but perhaps its association with the simple ovolo and tongue indicate a 2nd- rather than a lst-century date. (c) The pot is not particularly large and if size can be taken as any criterion then it should be placed in the early 2nd century. (d) The continuous scroll, with the concavities occupied by large medallions characteristic of the Antonine period, more precisely probably mid 2nd century. Some of the concavities may have leaves, joined by stalks to the main scroll. The treatment is very reminiscent of CINNAMUS of Lezoux. (e) The scroll is very similar to Wroxeter III, p. 37, No. 11, with leaves and medallions in concavities dated 90-130. Date : This pot is early Antonine, say A.D. 140-150. Near Figure 4, There is nothing inconsistent with Stratum Z2. a Flavian date : (i) the continuous chevron ; (ii) the arrow heads ; (iii) the medallion, possibly in continuous scroll; (iv) the wavy line demarcation motif. Date : Flavian. Figure 4, (a) The hare running to right is of Stratum C. little help. O. No. 2094 which it most closely resembles is dated late Antonine, whilst its next closest counterpart O. No. 2078 is dated Nero/Vespasian. (6) The bead rows dividing the metopes are primarily a 2ndcentury feature. (e) The divided metope and large medallions, with detached festoon, is very characteristic of Antonine period. (d) Similar to an Antonine bowl Newstead II. (Newstead, PIXL. 14). Date : c. A.D. 170. 44 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Sherd No. 4 Form No. Drag. 29 Figure No. Position 7.4 Near Figure 4, Stratum Zl. Drag. 37 7.5 Figure 3, Stratum A. Description The features of this sherd may indicate pre-Flavian. (i) The tongue has a rosette terminal (although this persists into 2nd century) ; (ii) there is no demarcating motif under the ovolo ; (iii) there is a slight trace of a continuous scroll. Date : c. A.D. 50. (a) The continuous scroll with leaf patterns indicate 1st century. (6) The moniliform binding with three beads indicates a date not later than Domitian. Date : Last quarter 1st century. Sherd Figure No. No. 1 8.22 8.23 8.24 Position Figure 2, Stratum C. Near Figure 4, Stratum D. Figure 4, Stratum D. 8.24 Figure 2, Stratum D. 8.25 Figure 5, Stratum B. TABLE 4 MORTARIA Description Roll Ha rd buff ware. bead rim. Cf. Wroxeter I , No. 98. Late 2nd century. Hard buff ware. Flanged type. Cf. Wroxeter I , Nos. 134 and 138. Late 3rd century ? Hard buff ware. Vertical rim type. Cf. Wroxeter I, Nos. 218-238. Late 2nd century to end of 4th. As sherd 3. Hard buff ware. Flanged, turned under at end. Cf. Wroxeter I, Nos. 14 and 46 and Collingwood No. 4. Late lstjearly 2nd century. TABLE 5 PLAIN SAMIAN WARE AND POTTER'S MARK Sherd Form No. No. 1 Curie 15. 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Position Figure 4, Stratum C. Drag. 33. Figure 5, Stratum D. Drag. 31. Figure 2, Stratum D. Drag. 27. Figure 3, Stratum Zl. Drag. 18/31. Figure 3, Stratum Bl. Drag. 18. Figure 2, Stratum A. Drag. 27. Figure 6, Stratum B. 18. Figure 6, Stratum A. 31. Figure 2, Stratum B. 27. Figure 2, Stratum B. 18. Figure 2, Stratum A. 18. Figure 4, Stratum B. 27. Near Figure 4, Stratum B, 36. Figure 4, Stratum C. 36. Figure 3, Stratum B. Drag. Drag. Drag. Drag. Drag. Drag. Drag. Date The Pudding Pan Rock variation No. 14. (O. & P . PI. LVI). C. A.D. 175. c. A.D. 150 (Footstand). c. A.D. 175. Part of base. Claudian. Antonine. 1st century A.D. Late 1st century A.D. 1st century A.D. Antonine. 1st century A.D. 1st century A.D. 1st century A.D. 1st century A.D. Antonine. Antonine. 45 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Note . Standard Forih No. S.32. S.33 S.34 S.35 S.36 S.37 S.38 S.39 S.40 S.41 S.42 No. 7.32 7.33 7.34 7.35 POTTER'S MARK AIISTIV=AESTIVTM. Figure 7.6. Form Drag. 31. Aestivus, a Lezoux potter dated c. 140-190. Numerous stamps recorded from Pudding Pan Rock. See also Richborough II, p. 72, No. 67 and Wroxeter II, p. 31, No. 134. (Figure 2, Stratum C). TABLE 6 COARSE POTTERY, STANDARD FORMS These forms continue the series given in Table 9, Site B report, positions are not given and will be found under Table 1. " Figure Date Description A.D. 60-75 Cooking pot of hard gritty ware. Incurving rim typical of 1st century. Cf. S.20 and 23 and Joyden's Wood, Arch. Cant. LXVIII, p. 176, Nos. 8-12. A.D. 60-75 Similar to S.32. A.D. 60-75 Similar to S.32. A.D. 60-75 Fumed grey ware. Plate with curved side, shghtly beaded rim. Cf. Richborough I, No. 86. 1st century. 1st century. 7.36 c. A.D. 70 ' Carinated beaker. Cf. No. 291 (A.D. 80-120). 7.37 A.D. 90 Ovoid cooking pot, with a simple everted . rim. Hard pink fabric. Cf. Camulodunum p. 269. (A.D. 10-48.) 7.38 A.D. 75-90 Broad rim, hard gritty. Shght development (increased flattening of top of rim) of sherd from Lullingstone. (Arch. Cant., LXVI, p . 29, No. 121.) Dated 60-80. 7.39 A.D. 75-90 Ja r possibly carinated. Cf. Canterbury, Arch. Cant., LXVIII, p. 117, No. 77. (1st half 2nd A.D. or earlier.) 7.40 A.D. 75-90 Olla with bead rim recessed on inside. One groove in shoulder. Similar Richborough I I I , No. 245. Date : 80-120. 7.41 A.D. 75-90 Development of S.40, but with cordons on shoulder. 7.42 A.D. 75-90 Bowl, grey ware, with overhanging rim. c. A.D. 75-90. Find Similar Drag. 18, Richborough III, Sherd No. 1 Figure No. 8.1 8.2 8.3 TABLE 7 MISCELLANEOUS COARSE WARE (Including Incense Cups Nos. 17-21) Position Near Figure 4, Stratum Bl. Figure 3, Stratum Zl. Near Figure 4, Stratum Z2. Black furrowed ware. Cf. Richborough II, PL XXIX, No. 1. Date : c. A.D. 50. Narrow-necked vase, w i th pronounced cordon on neck. Hard grey fabric. Cf. Richborough I, No. 4. Date : Claudian. Imitation of Drag. 24/25. Flat half round moulding above and below rouletted rim. Hard black ware. Date : Claudian. 46 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Sherd No. 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Figure No. Position 8.4 Figure 4, Stratum Z2. 8.5 8.6 8.7 8.8 8.9 8.10 8.11 8.12 8.13 8.14 8.15 8.16 8.17 8.18 8.19 8.20 8.21 Figure 3, Stratum B3. Figure 4, Stratum C. Immediately under tesserse in cella. Figure 6, Stratum A. Figure 2, Stratum Zl. Figure 2, Stratum Z2. Figure 4, Stratum C. Figure 3, Stratum B2. Figure 2, Stratum D. Figure 3, Stratum Z2. Figure 3, Stratum A. Figure 3, Stratum Z2. Found on N. wall of cella. Figure 4, Stratum C. Figure 2, Stratum B. (Antse). Figure 3, Stratum B2. Figure 4, Stratum C. Description Bead rim olla. Pink fabric fumed grey. Cf. Richborough III, No. 243. Date : c. A.D. 50. Olla with outbent moulded rim. Cf. Richborough III, Nos. 264 and 269. Date : c. A.D. 80-120. Neck of flagon with three-ringed neck. Brick red fabric, cream slip. Date : Mid 1st century A.D. See Appendix 2. Jug with five ringed neck. Pinky-buff ware. Cf. Richborough IV, No. 374. Date : A.D. 70-90. Bowl. Dirty grey ware. Wide rim. Cf. Richborough III, No. 276. Claudian. Date : Claudian. Beaker with moulded overhanging rim. Hard pale grey. Cf. Richborough III, No. 260. c. 59-75. Date : A.D. 50-75. Carinated jar—grey ware. Cf. Canterbury, Arch. Cant., LXVIII, p. 117, No. 77. Date : First half 2nd century A.D. or earlier. As Sherd No. 11 but grey centre with thick baked red coat, and then red glaze outside. Date : Late 1st century A.D. Grey ware, traces orange paint. Cf. Verulamium p. 184, No. 17. Date : A.D. 160-190. Black ware, with outer black glaze. Level lip. Cf. Canterbury, Vol. LXIII. p. 99, No. 12 dated Claudian. Date : A.D. 50-70. A typical mid lst-century rim, black with traces brown slip. Cf. Canterbury, LXIII, p. 101, No. 24 and Richborough IV, No. 381 (70-85). Date : A.D. 50-75 Rather similar to Sherd No. 15 but grooves under rim. Hard grey ware. Date : c. A.D. 60. Small incense cup. Similar to No. 48 but smaller. Verulamium p. 193, No. 48. Grey centre but very light orange surface. Date : Marly 2nd century A.D. Part incense cup. Similar to No. 30. Richborough I, p. 95. Date : Late 1st century A.D. Similar to Sherd No. 18. Similar to Sherd No. 17. Similar to Sherd No. 17. 47 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD TABLE 8 THE FIBULAE (Reported on by M. R. Hull, Esq., M.A., F.S.A.) Brooch No. Position 1 Figure 4, Stratum D2, outside temple. Figure No. 9.1 Figure 2, Stratum C, in northern of two antse. Figure 3, Stratum Z2, in cella. Figure 2, Stratum A, in northern of two antse. Figure 4, Stratum Zl, in southern of two antoa. Description Fine oval brooch, gilt, with glass stone mounted in centre. Belongs to a wellknown group which is attributed to the 4th century but not, as yet, more closely dated. The stone is conical, in a tallish setting, which is surrounded by the flat plate which bears two troughs bordered by sharp-edged walls ; the bottoms of the troughs are decorated with repeated impressions of a metal punch, in this case making an S-shaped mark. The inner trough is bordered by a low, corded fillet. The spring is of very few turns held by a single lug. Cf. British Museum," Antiquities of Roman Britain ", Fig. 11, No. 36. 9.2 Flat enamelled disc brooch, with small central hole for a stud, now missing. The enamel was in two bands, the outer of small squares of white and royal blue mosaic, the inner of blue and red mosaic A small loop at the top is broken. Polychrome work, perhaps from the Villa Anthee in Belgium, and usually ascribed to the 2nd century. Cf. Verulamium Report, Fig. 46, No. 36. Wroxeter Report III, p. 25, No. 11. Newstead Report, PI. LXXXIX and many others. — Fragment of bronze pennannular brooch with rolled over terminal forming small loop. No decoration. Common Roman type of the mid 1st century. 9.3 An early, large and fine example of the second group of derivatives from the Colchester brooch ; the type does not occur at Camulodunum, so is not before c. A.D. 65 at Colchester. It is common in Kent, where the earliest examples might possibly yet be found to be rather earlier than at Colchester. This brooch has no axial bar through the spring. Cf. Verulamium Report, Fig. 44, No. 22. 9.4 A very fine example of the Langton Down brooch, of the smaller, B, size, tinned, and of the very square-headed oonstruotion which is oommon around Trier in Germany. The very flat bow bears a simple reeding, the central member of which is an ill-executed knurled rib. Date : First half 1st century. 48 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Brooch Figure No. Position No. 6 Figure 4, Stratum A, 9.5 just outside southern of two antse. Figure 3, Stratum A, 9.6 in vestibule. Figure 4, Stratum A, just outside southern of two antse. Figure 3, Stratum B, 9.8 in cella. Description This brooch is of the same group as No. 4, but is of the normal small and povertystricken type ; the groove on the head, with notching or hatching on its edges, is typical. These brooches are probably not earlier than Flavian, and it is doubtful whether they extend into the 2nd century. They are very common in Kent. Many parallels could be quoted. As brooch No. 6. 9.7 As brooch No. 6. Bronze brooch of the same type as Nos. 6, 7 and 8 with similar groove on the head and hatching either side of its edges. The catch plate is pierced with a single large hole. Object No. Position 1 On cella floor. Figure No. — 10.1 TABLE 9 THE BRONZE OBJECTS (Reported on by J. W. Brailsford, Esq., F.S.A., and E. W. Tilley, Esq.) Description Bronze edging. Probably belonging to a sheath. Length 1£ in., width -J in. Cf. Curie, Newstead 1911,pi. XXXV, 1-7. Foot-stand. Common Roman type, with iron securing rivet. Cf. Lydney Park Report, Fig. 20, No. 102. Ring. Small. Bronze Object. The central portion is folded lengthways while the ends have been hammered out to form spatulas. Length 1£ in. Ring. Diameter $ in. No decoration. — Tack. Hollow-domed head. Pin. Part of Spherical Head. Incised orossed lines in head and three lines round neck. Chatelaine. Tweezers, nail-cleaner and part of an ear-scoop on a ring. Tack. Flat oiroular head for ornamentation of leather. Cf. Curie, Newstead, 1911, pi. XXV, 1-27. Teiret. Part of. Of Iron Age type. Cf. " Later Prehistoric Antiquities of the British Isles ", pi. X,.No. 2. . 2 Figure 4, Stratum C, just outside southern of antse. 3 As Object No. 2. 4 Latest Roman ground level, just outside S.W. corner of temple. 5 Figure 4, Stratum C, just outside southern of two antse. € Figure 4, Stratum D4, in southern of two antse. 7 Figure 3, Stratum Bl, in cella. •8 Figure 3, Stratum Bl, in vestibule. 9 Figure 2, Stratum B, in northern of two antse. 10 In rubble over north corridor floor. 10.2 10.3 49 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Object Figui No. Position No. 11 In rubble over cella 10.4 floor. 12 Figure 2, Stratum B, 10.5 in southern of two antse. 13 At latest ground level, 10.6 just outside south wall of temple. 14 Figure 2, Stratum C, — in porch. 15 Figure 6, Stratum B, 10.7 southern of two antse. 16 At latest Roman level, 10.8 outside S.E. corner of temple. 17 Figure 4, Stratum Bl, outside of two antse. 18 Just outside south wall — of temple. 19 Plough level. 10.9 20 In rubble above south •— corridor. 21 Just over floor of south — corridor. 22 In strong-room on —- destroyed floor. 23 In rubble of south 10.10 corridor. Description Bracelet. Fragment of, with snake's-head terminal. The body is decorated with incised diagonal lines. Cf. Verulamium Report, Fig. 46, No. 44, where a similar bracelet occurred in a late third century deposit. Thumb. Clearly shows nail and is bent at joint. A votive offering ? Corroded iron at the end of the thumb. Armlet. Bronze wire armlet of very simple type. One end is bent round to form a loop into which the other end slid. Two narrow coils of four turns of bronze wire are bound round it to form the only decoration. A similar example in silver is illustrated in Wheeler " London in Roman Times ", pi. XL, Fig. 4. Bezel. Iron ring and set with glass. Pin. Long, with plain, flat, circular head ; three grooves below the neck. Ring. Small bronze finger-ring. Thin flat hoop, the shoulders broadening to a raised ridge on either side of a circular bezel in which is set what appears to be a small piece of mother-of-pearl surrounded by white enamel. The shoulders are incised with diagonal lines. Cf. Richborough IV, pi. XXXV, No. 103 where it is dated to the 3rd or 4th century. " A degraded example of such types as B.M. 540, 544 or 1416 (cf. Henkel 1328, etc.)." References are given to the BM. Catalogue of Finger Rings, Greek, Etruscan and Roman by F. H. Marshall, 1907. Ring. Fluted. Diameter •$ in. Cf. Richborough II, pi. XV, Fig. 1, No. 7. Tack through silver. SmaU bronze tack which has been driven through thin silver sheet, a fragment of which still adheres around the head. Handle. Bronze object which is probably handle. It is of roughly circular section for most of its length except for the ends which have been squared off for about half an inch. These ends are decorated with incised crossed lines on two sides, and a single diagonal line on each of the other two sides. Overall length 3f in. Pin. Roughly made with square shank and spherical head. Stud. Hollow domed, filled internally with lead. Nail. Square shank and plain spherical head. Chisel. 3 • 1 in. long. SO THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Object No. Position 24 In gap between suggestus and west cella wall. Figure No. 10.11 25 Figure 3, Stratum Bl , — in cella. 26 Figure 3, Stratum A, — in cella. •ton Letter. Fragment of gilt bronze which appears to be the r ight leg and horizontal bar of a letter " A ". It is cut out of thin sheet bronze, pierced with nail holes, and raised to a central ridge. If i t were complete, the height would be about Z\ in. Several sites have produced similar letters, e.g. Lydney Park (Excavation of Lydney Park, Report of Soc. of Antiquaries, IX, pi. XXIV. p. 102, No. 8). Pfunz and Schierenhof on the German Limes (see O.R.L., Lief. 14, Taf. xiv, and Lief, 7, Taf. ii). etc. Strip. Pierced at one end with a small hole. Sheet. Thin, 3.9 in. long, 0'9 in. wide. One edge straight, the other roughly scalloped. TABLE 10 IRON OBJECTS (Reported on by E. W. Tilley, Esq.) Object No. 1 2 3 4 5 6 Position In rubble over vestibule floor. In rubble of southern of two antse. Found on tessellated floor in blacksmith's shop. In rubble, just over vestibule floor. From cella floor. Found in a line along Figure No. — 11.1 11.2 11.3 11.4 11.5 11 cella floor parallel to north wall. Iron floor of vestibule. 11.6 8 From cella floor near staples (Object No. 6). 9 In rubble above north corridor floor. 10 On vestibule floor. From floor, N.E. corridor. 11.7 11.8 11.9 11.10 Nails. Two Description large dome-headed nails. Diameter of head If in., length 2J in. Stylus. Incomplete. Length 3f in. Knife. Length 5-2 in. Width of blade 0-8 in. at its widest part. The tang thickens from 0 • 1 in. to 0 • 5 in. X-shaped Object. Small hole at intersection. Arms are turned at right angles t o the general plane of the object. Hinges. Each has four iron securing nails. Staples. Four, T-shaped. Similar objects were found at Newstead (pi. LXVI I , 1-3) where it has been suggested that they were used as clamps for " fastening tiles and tubes to the walls ". Rivet. Large with hollow-domed head and securing plate. Staples. Split types . Cf. Curie, Newstead, 1911, pi. LXVII, 10-13. Key. T-shaped iron slide-key with looped handle. Cf. Wheeler " London in Roman Times ", pi. XXXA, No. 1. Rivet. Large dome-headed rivet through a U shaped staple. Possibly used as a shackle. Ring and Staple. A ring on a large split staple.- Cf. Curie. Newstead, 1911, pi. LXVII, 12. 51 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD Object No. 12 13 14 15 16 Position In rubble over N.E. corridor floor. In latest Roman level, southern of two antse. In rubble just over vestibule floor. From floor of blacksmith's shop. From floor of blacksmith's shop. Figure No. — 11.11 11.12 Description Hinge. Roughly made. Cf. Curie, Newstead, 1911, pi. LXXXIII, Figs. 8 and 12. Shears. Blade and part of handle of pair of shears. Cf. Camulodunum, p. 343, Fig. 7. Bolt. Heavy iron bolt with thick circular plate at each end. Sickle. Badly corroded iron fragment which may be part of sickle. Horseshoe. Badly corroded iron fragment which may be part of horseshoe. TABLE 11 THE GLASS (Reported on by Miss D. Charlesworth, M.A.) Object Figure No. Position No. 1 Figure 2, Stratum B, — northern of two antse. Figure 4, Stratum B, 12.2 southern of two antse. Figure 4, Stratum C, — southern of two antse. Figure 4, Stratum A, 12.3 just south of temple. Figure 4, Stratum A, — in southern of two antse. Figure 2, Stratum D, — in porch. Figure 2, Stratum D, — in porch. Description Counter. Fragment from a green, square bottle, which has been roughly chipped to disc shape for use as a gaming counter, 0 • 3 in. thick. Associated with Antonine pottery. Bead. Oblate of blue glass. Fragment of green-glass incised with fernleaf design. Bead. Blue, ribbed, melon-shaped, glasspaste bead of common pattern (portion only). Wroxeter I, pi. X, Fig. 2, No. 22. Richborough II, pi. XIV, Fig. 2, No. 6. Handle. Fragment from the handle of a late 1st or 2nd century flagon, a type made in the Seine/Rhine area and imported in large quantities into this country, e.g. Richborough (4th Report No. 367 in pit 122 with late 1st and mid 2nd-century material and No. 368 in Area XVI c. 76-90). Bexhill and Faversham (Guide to the Antiquities of Roman Britain, British Museum, 1951, pi. XII, 10 and 11). Associated with late lst/early 2nd-century pottery. Base. Fragment of a base of square bluegreen bottle. Associated with 4th century rubble. Fragment. From shoulder, clear green square bottle. 52 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD TABLE 12 MISCELLANEOUS OBJECTS Object No. 1 Position Figure 2, Stratum D, in northern of two antse. Figure 4, Stratum D, in southern of two antse. On cella floor. 6 In rubble of north corridor. 7 On cella floor. 8 On cella floor. 9 On cella floor. No. 12.5 12.4 Figure 4, Stratum A, — south of southern of two antse. In rubble of south — corridor. 12.1 Bone pin. Description Biconical head. Pewter Leaf. Gilt-covered ivy leaf, pierced with two holes. One of these retained part of a bronze rivet. Axe-head. Miniature lead axe-head. Votive offering. Leather-thong. Short fragment with knot. Tegula. Two tiles both with clear impression of hob-nailed sole of a calceus. Cf. Wheeler " London in Roman Times ", pi. XLI I I , No. 3. Tegula. Impressions of a child's bare foot. " Pseudo- Venus." See par t 3, and Appendix 1. Altar. See Part 3 and Appendix 3. Seeds. See Part 3. TABLE 13 THE BONES (Reported on by Miss J . E. King.) Fragment No. 1 2 3 4 5 6 7 8 9 10 11 12 13 Position Figure 2, Stratum Bl, Porch. I n rubble, cella. Figure 4, Stratum C, southern of two antse. I n rubble, north corridor. Ox, proximal phalange. I n rubble, Blacksmith's Ox, right side lower jaw Description Dog, incomplete humerus. Ox, middle phalange. Ox, middle phalange. shop. I n rubble, Blacksmith's shop. I n rubble, Blacksmith's shop. Figure 3, Stratum A, I n vestibule. Figure 2, Stratum D, in northern of two antse. Figure 3, Stratum A, in cella. I n rubble, northern of two antse. On cella floor. Figure 6, Stratum Bl , just outside south of temple. Ox, left side lower jaw. Sheep, left side lower jaw. Sheep, horn core. Horse, upper molar. Ox, upper molar. Tarso metatarsus fragment of game bird, possibly large domestic fowl, gallus gallus. Red deer antler fragment. Fragment of furoulum of game bird, probably domestic fowl. 63 8 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD TABLE 14 STRUCTURAL SEQUENCES AND FLOOR LEVELS Phase Structural Sequence Floor Level with Respect to outside Ground Level Zl Z2 A B l B2 B3 C Dl D2 D3 D4 Wooden building ? Claudian Wooden building ? Flavian First stone building and porch walls. Clay floors and entrance. Mortar approach path; earth floor in corridor ; clay floor in cella ritual hearth (Early Antonine). Mortar approach path ; clay floor in corridor and step to cella ; suggestus built, and tessellated floor in cella. Flint/pebble approach path ; vestibule projecting walls added and tessellated floors in corridor. Porch flanking wings (antse) added; porch mosaic and new porch path. (Late 2nd century). Walls of temple rebuilt. (Late 3rd to middle of 4th century A.D.). Minor repairs to walls ; porch path repaired; vestibule and storeroom mosaics destroyed (?) Temple ceased to be used for religious purposes, (c. A.D. 350-360). Temple used by Blacksmith as his shop and then deserted. Porch Same Same 5 in. above Same Same 7 in. above 3 in. above 2 in. above Same Same Same Vestibule and Corridor Same Same 5 in. above 2 in. above 7 in. above 19 in. above 6 in. above 2 in. above Same Same Same Cella Same Same 5 in. above 9 in. above 24 in. above 24 in. above 11 in. above 7 in. above 6 in. above 5 in. above 5 in. TABLE 15 WIDTHS OF TEMPLE WALLS Site Cella Wall External Wall 1. Springhead, Temple I. (Present report). 21 in. 21 in. 2. Richborough, III, p. 34, Temple 2. 38 in.-48 in. 39 in.-47 in. 3. Maiden Castle. 23 in. 23 in. 4. Harlow(4) c. 36 in. o. 36 in. 6. Worth(4) 48in.-52in. 39in.-40in. 6. Frilford(7) 30 in.-36 in. 30 in. -36 in. 7. Colchester 42 in. 24 in. 8. Richborough, III, p. 34, Temple 1. 34 in.-38 in. 34 in.-38 in. 54 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD REFERENCES (1) W. S. PENN, Arch. Cant., LXXI, 53, 1957. (2) W. S. PENN, Arch. Cant., LXXII, 77, 1958. (3) R. GOODOHILD and J. R. KIRK, Oxoniensia, XIX, 15, 1954. (4) R. E. M. WHEELER, Antiq. Journ., VIII, 300, 1928. (5) M. R. HULL, Roman Colchester (The Large Temple, St. Helena's School), p . 224, 1958. (6) R. E. M. and T. V. WHEELER, Verulamium, p. 131, 1936. (7) Oxoniensia, IV, 27. (8) R. E. M. and T. V. WHEELER, Lydney Park, Fig. 3, 1932. (9) G. C. BOON, Roman Silchester, p. 155, 1957. (10) R. C. NEVILLE, Sepulchra Exposita, p. 89, 1848. (11) Archceologia, Vol. 57, p . 252. (12) CURLE, Newstead, p. 359. (13) H. HELBAEK, Proceedings of the Prehistoric Society, New Series, Vol. XVIII, pt. 2, p . 194 (1952). (14) OVID, Fasti II, 5-29. (15) Archceologia, Vol. 62, p . 4. (16) Archceologia, Vol. 61, p . 207 and p . 215. (17) P. V. HILL and J. P. 0. KENT, " Bronze Coinage of the House of Constantine, A.D. 324-346 ". Spinks Numismatic Circular, 1955-56. (18) J. P. C. KENT, Numismatic Chronicle, 1957, pp. 16 ff. APPENDIX 1 SOME POSSIBLE INDICATIONS OP THE CULT by F. Jenkins, F.S.A. The clay figurine (now represented by five fragments which join) was found resting on the floor of the ceUa, a fact of great interest for it is the first time that a clay figurine of any type has definitely occurred in a Romano-Celtic temple in Britain. This apparent rarity is very remarkable, for when we examine the wealth of smaU objects which have been found in simUar temples in Gaul and Western Germany, it is clear that the clay figurines of diverse types, including that represented by the example to be described, are very numerous. The votive character of these figurines is weU attested, and in view of its surroundings, it is almost certam that the one found at Vagniacae was used for the same purpose. The figurine is now incomplete but fortunately enough has survived so that it can be reconstructed and the type identified with some degree of accuracy. It is of a standing female, supporting in her left hand a robe which has slipped down over the hips, to leave the upper part of the body nude. A bracelet is worn on the left arm just above 55 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD the elbow. The head and right side of the figure are now lost, but it is almost certain that the right arm was bent upwards so that a tress of hair, like the one to be seen on the left shoulder, was held in the right hand. OriginaUy the figure was mounted upon a smaU base made in the form of a hollow cube. This reconstruction is based on a complete figurine found at the site of a temple at Horperath.1 Figurines precisely of the type just described, seem to be characteristic of the Rhineland figurine industry which had its main centre at Cologne, where it was estabhshed at the end of the 1st century A.D., and reached its peak period of production about fifty years later.2 Having studied the products of that industry at first hand, the present writer considers that the clay and style of the figurine from Vagniacae leave httle doubt that it was made in the Cologne or allied workshops and entered Britain either by way of trade, or as the personal possession of a traveUer from that region. Having progressed thus far we must now identify the personage which the figurine represents. The art-type whence the present example was copied, is that of classical Venus, but there are reasons for tliinking that the deity in this case is not the Roman goddess of love, but a native deity in Roman guise. This problem has been discussed elsewhere by the present writer,3 therefore in order to avoid undue repetition, it wiU suffice if attention is drawn to a few of the salient points which may aid us in our quest for knowledge of the cult once practised at Vagniacse. If the deity was originaUy a native goddess then it foUows that although she appears in the guise of Venus, she may not have been primarily concerned with amorous pursuits. It is more hkely that she was regarded by her worshippers as a personification of the great earthmother so widely venerated in manifold forms and under various titles by people of Gauhsh extraction. As a mother-goddess she would have been an aU-purpose deity who had numerous functions chief of wliich was to protect aU hving things, including the crops. Perhaps the goddess at Vagniacae was of this kind, but before we reach any conclusions it may be useful if we examine further evidence from the temple. The comparative scarcity of cult-objects from the temple is noteworthy, considering the wealth of material derived from the continental sites. The altar, which undoubtedly stood in the cella in front of the suggestus, being uninscribed, provides no clue to guide us, but one object at least, seems to have served a ritual purpose. This is a socalled " incense-cup ", a smaU pottery vessel of a type which is characteristic of temple furniture, but what is even more significant 1 See topographical index No. 1. 2 See topographical index Nos. 9-11 inclusive. 3 F. Jenkins, " The Cult of the Pseudo-Venus in Kent ", Arch. Cant., LXXII. 56 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD is that it contained a quantity of seeds of Atriplex patula (Common Orache). Now this herb to-day is found on cultivated and waste ground, therefore it is curious that its seeds should have been selected in preference to seed-corn, which obviously would have been a more appropriate offering to the kind of deity we have in mind, viz. a goddess concerned with the growth of the corn. Hence as the seeds do not belong to the cereal famUy, there must be another explanation for thek presence in a seemingly votive context, and the answer may he in the works of classical writers. Of these Varro provides us with some interestmg information. He teUs us that at the Vinalia rustica the gardeners kept a hohday in honour of Venus,1 and gives her the title procuratio hortorum, that is, " keeper of the gardens ".2 That the dealers in pot-herbs and vegetables celebrated her festival on the dedication day of her temple at Rome, certainly seems significant, and must indicate that she presided over thek welfare by protecting thek crops. The meaning of her name seems to provide us with another link in the chain of evidence, for it ought to signify " beautiful appearance". This, as Rose points out, was not of women but that of a piece of well-tilled ground, in fact a vegetable garden.3 From this it is but a short step to conclude that prior to her assimUation with Greek Aphrodite, Venus must have originated as a goddess of vegetation. Can it be that the deity represented by the clay figurine, was of that kind ? The presence of the seeds certainly suggests that she was. If then we assume for the moment that the goddess has been correctly identified, we can progress a stage further along this line of inquiry. As the seeds seem to be the key to the problem it is possible that there is further evidence relating to them wliich may aid us in our search for clues. Atriplexum, the more ancient form of the name of the plant, occurs in the works of a number of classical writers,4 where it is specified as a kitchen vegetable. In his previous report, Mr. Penn has akeady cited several instances where species of Atriplex have been used for various purposes, and the present writer agrees with his suggestion that it seems not unreasonable to suppose that Atriplex p. was cultivated in this country by the Romans, either as fodder or for human consumption. 5 If this is true, then either use does not affect the present argument, for the success of the annual crop would have been vitaUy important to the grower. Hence, it may weU be that in order to secure the next crop, an offering was made to the deity, the first fruits of the 1 Varro, De Lingua Latino, 6, 20. 2 Varro, De Re Rustica 1, 1, 6. 8 H. J. Rose, Ancient Roman Religion, p. 92 ff. 1 For a number of references see Lewis and Short, Latin Dictionary (Oxford) 1945 edition. 5 W. S. Penn, Arch. Cant., LXXI, p. 68. 57 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD plant upon which his hvelihood depended, being the orthodox gift in this case. In pondering upon this fascinating problem, the present writer has thought of another possible use for the plant. Perhaps it was valued for its supposed medicinal quahties, a not unhkely possibUity, for from very early times, herbal remedies have occupied an important place in the sphere of rustic medicine. Is it too far-fetched to presume that the oracular utterances which may have formed part of the temple ritual, prescribed an infusion of the leaves of the herb Atriplex to cure a votary who had invoked the deity to relieve his sufferings ? We cannot be sure, but we do know that healing of diseases has a prominent place in cult-ritual in many of the Romano-Celtic temples in Gaul and Western Germany. In the present writer's view, it is not beyond the bounds of probabihty that the priest who officiated at Vagniacae included the role of local medicine-man in his temple duties. In the assemblage of ex vetoes found in the Gauhsh temples there are many objects which clearly indicate that sick persons resorted to these shrines in order to invoke the gods to reheve thek sufferings. UsuaUy the ex voto took the form of a model of the affected part of the body to be cured, such as genital organs, breasts, limbs, fingers and thumbs. It is therefore interestmg to know that a bronze model of a bent thumb has turned up at Vagniacae.1 It is complete in itself and is not part of something larger such as a cult statue. In fact it is identical to those found in the Gaulish temples where the one time presence of healing cults is well-attested. In view of this we may regard the present example as an offering deposited by a votary, who either hoped that in so doing he would be cured, or that he had been cured of his affliction, and wished to express his thanks in this manner. Now we must return to the subject of the figurines to see if there is any evidence that they were associated with the cults of healing. I t is significant that in Gaul and Germany many examples have occurred in this context. One example, closely similar to that found at Vagniacoz, came from the temple of Lenus-Mars at Trier, a god whose main function as a healer is proved by epigraphic evidence.2 A feature of the temple precinct was a sacred spring. Now the Gauls tended to buUd thek temples on the sites of springs, which they believed, had healing quahties as weU as being the source of hfe gushing forth from the earth personified as the universal mother-goddess. At Vagniacoz we know from the excavations that the temple was buUt on marshy ground which may mean that a spring determined 1 The literature relating to ex, votoea associated with healing oults is enormous and the temples are numerous. Of these the temple of Sequana, the goddess of the source of the Seine, has yielded a wide variety of objects of this type. a For bibliographical references see topographical index No. 6. 58 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD its position. It also stands at the head of a creek where the fresh water flows out into the Thames. The modern place-name is also highly significant for it is hardly necessary to point out that it is Springhead. In an envkonment such as this, the goddess represented in the guise of Venus in Romano-Gaulish inconography would be at home. One feature of the temple which seems to be unique in this country is the suggestus. In its present state we have no means of knowing its original height, but as the altar stood in front of it, it is reasonable to suppose that it formed a fitting background and probably supported some land of shrine buUt above altar level. If this is true, then it may have been hke a Lararium found at Pompeii which housed the statuettes of various deities.1 Perhaps it was in a shrine of that kind the clay figurine of Venus was placed. That this conjecture may be correct is indicated by the remains of a simUar structure built against the north waU of the cella in the shrine of the Xulsigise, a group of mothergoddesses, which stood near the temple of Lenus-Mars at Trier.2 Here there is httle doubt that the various ex voto statuettes were once placed on it to be seen by the worshipper as he crossed the threshold. A feature which at present remains something of an enigma is a burnt layer which was found at a lower level under the place where the suggestus was finally erected. Perhaps it represents the sacrificial fire lit during the ceremony when the hnes of the temple were defined, but this is pure conjecture, and until further evidence is obtained it is wiser to suspend judgment as to the precise nature of this feature. In reviewing the evidence obtained from this temple it is clear that we are dealing with a highly complex problem. In attempting to solve it, several avenues of research have been explored, so now it is time to summarize what evidence we possess. Fkstly, it is certain that a goddess akin to Venus, or that deity herself, was venerated at Vagniacoz. We have seen that there may have been several aspects of her cult, namely, fertility, the care of the crops and healing, coupled with that of the sacred springs. Our aim has not been to be too dogmatic in our conclusions but to set forth the evidence in an impartial manner to form the basis for future research into this fascinating problem. A TOPOGRAPHICAL INDEX OE THE GEOGRAPHICAL DISTRIBUTION OP CLAY FIGURINES OE THE SPRINGHEAD (VAGNIAO^S) TYPE, IN GERMANY (Appendix 1) (1) Horperath. Bonner Jahrbucher 143/144, S. 398 ff., Taf. 73, Abbs. 1 and 2. 1 C. Bailey, The Legacy of Rome (Oxford), p. 241, illustration. 2 For reference see topographical index No. 6. 59 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD (2) Gusenburg. F. Hettner, Drei Tempelbezirke im Trevererlande, S. 89, Taf. XIII, No. 64. (3) Dhroneeken. F. Hettner, ibid, S. 59, Taf. XII, 52. (4) Trier. S. Loeschcke, Der Tempelbezirk im Altbachtale zu Trier, Heft I, S. 61, Taf. 24, No. 1. (5) Trier. S. Loeschcke, Die Erforschung des Tempelbezirkes im Albachtale zu Trier, Abb. 16. (6) Trier. E. Gose, Der Tempelbezirk des Lenus-Mars in Trier, S. 55, Abb. 68, No. 7. All the above examples were found on the sites of temples. Those which foUow are stray finds where the precise nature of the site was not determined. (7) Bonn. Rheinischen Landesmuseums Bonn, No. A1199. (8) Bonn. Rhein. L.M. Bonn, No. U 240. (9) Stockstadt. F. Fremersdorf, Erzeugnisse Kolner Manufacturen ... in Saalburg Jahrbuck IX (1939) S. 11, Taf. 6.1. (10) Zugmantel. F. Fremersdorf, ibid, S. 11, Taf. 6.2. (11) Cologne. F. Fremersdorf, ibid, S. 11, Taf. 6.3. APPENDIX 2 A NOTE ON THE FLANGED SHERD (SHERD No. 7, TABLE 7) This flanged bowl presents a difficult problem. It was found under the tessera at the east edge of the cella. The mosaic is definitely Antonine, and since at first sight this sherd would not appear out of place in a fourth-century context, it must receive close study. The hp projects very slightly above the flange and is therefore different from the feature of S.42 weU dated A.D. 75-90. It is very simUar to Wroxeter I, p. 71, No. 16, except that there are no painted patterns and the clay is not cream or drab. No. 14 with hard grey sandy clay, also with a flange, is also dated 80-120. These examples from Wroxeter at least show that a flanged bowl does not necessarUy belong to the fourth century. If mortaria are examined for evidence, a low Hp at Richborough (Richborough IV, No. 498) has a pre-Flavian date. As the Up becomes higher (Richborough IV, No. 504) the date becomes later (80-120) and with a high Up the date reaches c. A.D. 276 (Richborough IV. No. 516). This tendency is reflected in Samian mortaria with Ritterling Type 12 and Curie Type 11 with a low lip having a date no later than the late 1st century, and Drag. 38, with a very high hp dated towards the end of 2nd century. 60 THE ROMANO-BRITISH SETTLEMENT AT SPRINGHEAD When the true flanged bowls are reached with indifferent dull fabrics (and usuaUy black paint) they are aU given a fourth-century date. Compare Nos. 121 and 122 Richborough I and Nos. 41-43 Lydney Park. The present example has a very low hp above the flange, which might place it at the beginning of the 2nd century. However, it does not possess the hghter forms of fabric of these early types and therefore perhaps should be placed later. The end of the second century might be a good compromise but this deduction must remain conjectural. I t does raise the point however that excavators are possibly too eager to place the flanged bowl, krespective of variations, in the fourth century. (The present example has a dull brown fabric and is fumed black.) I t may be noted that this sherd was found at the end of the south groove in the ceUa. In the north groove, a coin dated c. A.D. 268-278, was found. This could therefore mean that the sherd could be late thkd or early fourth century. As a general rule, however, it seems that the higher the flange the later the date. APPENDIX 3 PETROGRAPHICAL REPORT Report by Miss H. A. M. Maedonald No. Object and Position Description 1 Porch Step. Limestone. SimUar to a specimen of hmestone from the Grinstead clay at East Grinstead, Sussex. 2 The Altar and its Base Millstone grit. 3 The Rectangular Base from Oolitic Limestone. SimUar to a the Vestibule. specimen in the museum collection from Laynton, Oxfordshke. 61