A Late Antique Decorated Casket and Jewellery from the Roman Villa at Fordham, Essex

Abstract

During excavations of a Roman villa at Fordham, Essex, a remarkable series of decorated bone and antler veneer plaques were recovered from villa destruction deposits. They are datable to the later fourth or fifth centuries a.d. and probably once adorned a casket holding bathing equipment and jewellery. Spread through the three main rooms of the villa, fragments were recovered from at least 10 metres apart, so the object is likely already to have been broken when deposited. The plaques are decorated with ‘late antique’ style figural, zoomorphic, vegetal and architectural motifs on a cross-hatched background, with the best-preserved design probably relating to female bathing.

INTRODUCTION

Excavations on the site of a Roman villa near the village of Fordham, approximately six miles from Colchester, were undertaken by the Colchester Archaeological Group and the Fordham Local History Society. The excavation started in 2015, with the kind permission of the Woodland Trust and the local farmer Robert Chamley, and continued intermittently into 2022. On current evidence the villa was constructed in the fourth century (post a.d. 316) as a small three-roomed building with a large apse on the north-western aspect and with a detached bathhouse. The western room of the villa (Room B) had a hypocaust with a small ‘furnace room’ to the north. The detached bathhouse was some 14 m to the south. It was constructed in the ‘Reihentyp’ [row type] style and was tri-apsidal in form at the southern aspect (fig. 1). Both buildings were well constructed with approximately 600 mm wide hard-packed flint foundations. They had later been robbed for building material, probably for the nearby All Saints church, the earliest parts of which are largely of flint construction.

FIG. 1. General site plan of the Fordham villa complex.

The main subject of this paper is a series of 31 pieces of bone and antler veneer which, it will be argued, originally adorned a wooden casket.² Items of jewellery from the site that illustrate what may have been stored in the casket are also discussed.

THE VENEERS (fig. 6)

The pieces of veneer were found scattered in destruction deposits over at least a 10-metre horizon throughout the main building but were concentrated in the fill of a hypocaust room (Room B) on the western side. The first three fragments (fig. 2) were found in the apse area of Room D in 2015, marked by a red circle on the plan (fig. 3).³

FIG. 2. The first three fragments discovered. From room D.

FIG. 3. Distribution of the veneer fragments in the villa destruction deposits. Key: red = bone; black = antler.

When excavation resumed early in 2019 a further fourteen pieces were recovered, twelve from Room B and two from Room C. Fourteen more fragments have since been recovered from Room B, the floor and pilae stacks of which had been removed, leaving a large 500 mm deep space that had been backfilled with building debris and other material. We have assumed here that the pieces were originally glued onto a small casket, or perhaps more than one.⁴ No metal box-fittings were found in the destruction deposits, suggesting that the casket had a jointed wooden core.⁵ Given the wide distribution of the individual pieces of veneer, it would appear that by the time they were deposited the box(es) to which they had been attached had already broken and disintegrated.

The pieces are thin slivers of antler and bone, between 2.5 mm and 3.9 mm thick, sawn and filed into small plaques (see below).⁶ In shape they are roughly rectangular or trapezoidal and incised with figural, architectural, zoomorphic and vegetal motifs on a hatched background. Some are slightly curved, reflecting the material from which they have been cut. Of the 31 pieces, 24 (77 per cent) were made from red deer antler and just seven (23 per cent) from bone. They are catalogued in the online Supplement.

The utilisation of red deer antler is common in late Roman contexts, including for caskets and veneered or inlaid furniture,⁷ but the use of both bone and antler together on one object is unusual. The processes behind obtaining these raw materials were rather different⁸ and that two different sources were used might suggest that we are dealing with more than one, similarly decorated, object (see below). That we are dealing with a finished product rather than a workshop is clear. A small number of pieces of sawn red deer antler have been recovered from the excavations (although none in contexts associated with the veneer finds), but evidence from elsewhere shows that the presence of large quantities of waste is required to provide significant evidence for a workshop.⁹

The manufacturing techniques are very simple and required only the use of saw, knife or graver and file. As these were also the basic tools of woodworking, it is quite possible that the craftsman who made the veneer also made the wooden object to which it was attached. There are no fixing pegs or perforations on any of the Fordham pieces, so it seems clear that they must have been glued in place. File marks can be seen on the reverse of all the pieces (fig. 4). These had the dual function of both flattening the reverse (to ensure that all the pieces to be applied were of roughly similar thickness) and providing keying for the glue. The front of the sheets would also have been filed flat, then smoothed, before the decoration was added. The decoration is all knife-cut, including the background cross-hatching. It is possible that the veneer was finished with a wax polish, or that the pieces were originally painted or inlaid; however, microscopic examination of each piece has not revealed any residue of paint, staining or inlay.

FIG. 4. Examples of reverses of veneer plates showing working (not to scale).

There are three sets of two adjoining pieces.¹⁰ The first consists of a human figure (fig. 7.2 and 7.5; catalogue nos 7 and 8), which were found close together in a group of four pieces, all lying face downwards (fig. 5);¹¹ the other two pieces were not directly related to the figure (fig. 8.11).¹² The second consists of the head and neck of a zoomorph (fig. 8.12 and 8.14) which were found 2.84 m apart in separate rooms. They were also vertically separated by 0.5 m, with one (fig. 8.12) found in the backfill of the hypocaust of Room B and another (fig. 8.14) in room C. Finally, two more joining pieces of a human figure were both found in room B but 2.41 m apart (fig. 7.3 and fig. 9.29).

FIG. 5. Four pieces found close together in Room B.

FIG. 6. The veneers (2015–2020).

FIG. 7. Figural representations.

FIG. 8. Zoomorphic and alabastron designs.

FIG. 9. Other designs.

THE DECORATION

HUMAN OR DIVINE FIGURES

The best-preserved image is of a nearly complete figure, probably female, missing the left side of her head, her right leg and her feet (figs 7.1 and 10.1). Her wild hair and mask-like features are distinctive and dramatic. A segmented band around the base of her neck and a band above each hand either represent jewellery, perhaps a necklace and bracelets, or the hemline and cuffs of a garment. The former may be most likely as her legs are shown bare for their full length; this is most unusual for females, who are generally depicted fully clothed, apart from Venus and the nymphs, who are usually nude. In front of her loins, she holds what appears to be a long roll of cloth, perhaps a garment or a towel, which hangs down on each side. It is not certain whether she is standing or seated, no chair or bench is indicated, and the full length of her legs is shown. Linear marks on the roll probably represent folds in the fabric, although those on the sides resemble binding. A standing nymph on a third-century relief from High Rochester of Venus bathing holds a roll of cloth in front of her in exactly the same position, although it does not fall to either side as shown here.¹³

FIG. 10. Full figures (Scale 1:1).

There are at least two further examples of this figure, although only one (the above cited) shows her upper torso and head; the others simply depict the legs with the same drapery wrap (fig. 7.2, 7.4–5; fig. 10.2–3). Among the missing pieces from the overall design we must therefore assume more parts from at least two similar women, presumably displaying the same shock of hair on their heads. All three representations are on trapezoidal veneers. Although the most complete design has been made from a single piece of bone, another was manufactured from two adjoining pieces (fig. 10.3). To judge from the most complete figure and from the two joining pieces, it is likely that the other two figures were manufactured from at least four individual pieces. The reason for this appears to be that the largest piece was manufactured from a single piece of bone, probably from a scapula, while the smaller plates were all from antler.

Another figure appears to be that of a male wearing a belted tunic (fig. 10.4). It is formed by a piece showing a belted body (fig. 9.29) which joins to one depicting the lower edge of the garment above legs shown only from the knee downwards and with out-turned feet (fig. 7.3). The upper part of the body is missing, making it impossible to match this figure to any of the detached heads described below.

Four other fragments depict heads of uncertain gender (fig. 11). Of the first two, only the lower part is preserved (figs 7.6–7 and 11.1–2). The most distinctive feature here is a long curl to either side of the chin, remarkably similar to the hairstyle of a figure on a veneer plaque from a pit on the site of a Romano-Celtic temple at Chelmsford (fig. 12.3). It too is incised and depicts the upper part of a human figure, probably but not certainly female, with the same curls on either side of the face, or perhaps, as Wickenden suggests, an elaborate head-dress.¹⁴ If this latter interpretation is correct, we might see the head-dress of the Fordham fragments as similarly indicative of status, whether secular or religious. The Chelmsford plaque was found in a pit from a robbing phase after the temple had gone out of use and was associated with a coin of Constans and late fourth-century pottery. The pit cut an oven that was itself dated to after c. a.d. 390.¹⁵ The similarity of the Chelmsford piece to those from Fordham is striking, as are the comparatively close proximity and dating of the two find spots. Although abraded, a piece of veneer from Piddington Roman villa in Northamptonshire is very similar in appearance to the Fordham and Chelmsford figures and may be considered alongside them here (fig. 12.4). It was, unfortunately, found on the spoil heap, although the excavator has suggested that it was probably derived from late Roman midden material.¹⁶

FIG. 11. Other heads (Scale 1:1).

FIG. 12. Fordham (1, 2), Chelmsford (3) and Piddington fragments (4) compared.

A further head has curly or wavy hair, perhaps with locks falling onto the left shoulder though this is uncertain (figs 7.8 and 11.3). The upper left side and lower right side of the coiffure are lost, and the fragment ends at the chin, while from the mouth issues a flattened oval object, perhaps a mirror shown in profile or more probably a wind instrument, perhaps a pipe. While the lock (if it is a lock) falling to the shoulder is suggestive of the head being that of a female it may nevertheless be male, as immature cupids – putti – are generally shown with short curly locks. Another head, of which even less remains, also appears to have curls on the top, although none are shown on the sides (figs 7.9 and 11.4), so also plausibly a cupid. The left hand lifts or supports an object with curved lower edge and central decoration, again perhaps a disc mirror or a percussion instrument, a tambourine, above the head. The right arm is missing but may also have been raised to support this object. Although earlier and carved in low relief, two objects may be pertinent in contrasting ways to these Fordham fragments and to their interpretation. First, a sarcophagus from Rome shows a parade of curly-headed putti playing musical instruments, including a tibia, panpipes and drum, the latter raised level with the head.¹⁷ The second, an ivory from Greenwich Park, London, depicts Victory holding above her head a shield decorated with a long-petalled flower.¹⁸ Another ivory plaque from Caerleon depicts a cupid supporting an object, albeit of a different kind, on its head.¹⁹

The simple linear style of the Fordham figures, although naïve, is in a Romano-British tradition. They can also be compared to a graffito sketch of the head of a woman on a piece of tile from Neatham, Hampshire, which is probably also of fourth-century date and on a recently discovered copper-alloy belt plate from Wiltshire.²⁰

ZOOMORPHS

There is a zoomorphic motif present on several pieces. One depicts the right profile head of a creature with open mouth and a mane or fin-like projection on its head and what appears to be a raised foreleg (figs 8.10 and 13.1). Another shows the left profile of a different type of creature, whose wide muzzle appears to be bound by a ribbon or perhaps a form of harness, as, for example, on mosaics from North Africa showing the triumph of Neptune and Amphitrite (figs 8.12 and 13.3). It does, however, join a piece that shows it to have a long sinuous neck, so it is not equine.²¹ A third piece depicts the heads of a pair of creatures similar to the first, their confronted muzzles touching (figs 8.11 and 13.2). One further item (fig. 9.24) may show the crowns of two conjoined heads and so be a further example of this design. Other pieces of veneer carved with sinuous forms may possibly belong with these creatures (fig. 8.13 and 8.15).

FIG. 13. Zoomorphic figures (scale 1:1).

The confronted animal heads and the similar creature shown in profile (fig. 13.1–2) are very similar to those portrayed on the late Roman horse and dolphin buckles first described by Hawkes and Dunning²² and more recently by Henry.²³ Confronted animal heads also appear on some late Roman double-sided combs²⁴ and may be seen portrayed on the late Roman mosaic from the temple at Lydney Park, Gloucestershire,²⁵ while a comb from Langton in Yorkshire shows two confronted dolphins with notched crests, their beaks joined by a short bar.²⁶ These animal heads have good parallels in the iconography of late Roman Britain. The third figure is probably a ketos, a mythological marine creature with long sinuous neck and fins (fig. 13.3). It is fairly closely paralleled on one of the panels of the Projecta Casket from the Esquiline Treasure, dated c. a.d. 330–370, where a nereid riding on a sea-creature is depicted on two of the side panels: a hippocamp on one side and a ketos on the other.²⁷ A further link may be provided by the shawl of the nereid riding the ketos, as it rises in an arc above her head, its central fold and slightly irregular width perhaps providing a match for the enigmatic fig. 8.13.

ALABASTRA AND VEGETAL SCROLLS

Five pieces (fig. 8.16–20), depict alabastra; an alabastron is a small vessel with a narrow neck and broader body which contained perfume or unguents and was part of the necessary equipment for bathing. They are all trapezoidal, but the designs, as well as the quite distinctive leaf and tongue decoration above, are sufficiently different from one another to render it unlikely that they were placed immediately adjacent to one another on one casket.

Three of the other pieces here are vegetal scrolls (fig. 9.21–23), which may have been intended to divide the figural scenes. However, if this was the case very few have survived.

OTHER IMAGES

One rectangular piece has a pedimented aedicula or doorway (fig. 9.27). Another shows part of a floral or forest scene, or perhaps the top of a red deer's head with pronged antlers²⁸ or even of a sea-stag, like the example depicted on a spoon in the Canterbury Treasure (fig. 9.26).²⁹ There are several other enigmatic pieces that are difficult to fit into a general scheme.

THE DESIGN

It is difficult to know how much of the original design of the casket remains. It seems very likely that many more fragments remain to be discovered, although we do not know, of course, what condition the casket was in when it became a part of the dump of rubbish placed within the ruined villa building. However, the fact that the fragments were scattered over at least 10 metres and occurred in 11 different contexts suggests that it was already broken when deposited. As we do not have all the pieces, perhaps even only a small fraction of them, especially if more than one casket is involved, reconstruction of the overall design is difficult. We may, however, make some observations based on the shapes of the veneer and broad suggestions as to how they might have fitted together, always presuming that the box was rectangular. Given that there are three female figures, one certainly and the others possibly ‘wild haired’, we might, for example, propose that there was one on each side panel and one is missing, unless the back of the box were plain. Alternatively, there may have been one on the lid and then one on either of the two longer, or shorter, sides. Just taking this simple example illustrates how difficult it is to determine not only the original design, but also how much of it we have remaining.

The veneers fall into two basic shapes, trapezoidal and rectangular. All use cross-hatching to emphasise the decoration.

Trapezoids: The largest of the of the trapezoidal pieces shows a female figure; others of this form are of different sizes and carry figural, geometric and floral designs.

Rectangles: The main rectangular pieces show zoomorphs but include at least one of the heads and a number of the more enigmatic pieces. The zoomorphic pieces are particularly instructive. One (fig. 8.11) has a plain area above the design; this is repeated on two other fragments (fig. 9.24–25). It would seem, therefore that the rectangular pieces might form a two tiered ‘frame’ around the casket.

We might assume, therefore, that we have a basic division of decoration and shape. While it is certain that we have just a part of (probably) one original design, it is still possible to imagine the rectangles forming a two-tiered border to the overall scheme with the trapezoids forming arcs or even entire circles around a central figural design.³⁰ If we examine the lengths of the two main decorative trapezoids (the alabastra and the vegetal pieces), we see that not only do they vary in size, but the dividing lines over the alabastra are all of different thickness, suggesting they are from different parts of the casket, or that the maker was not too concerned about matching the different pieces. Similarly, the design of each alabastron is slightly different, as is the floral decoration in the field above.

There are two more factors that we should consider. First, that part of the design might have been made of wood. It is known that bone or antler inlay were used together with wood in the late Roman period, as on the door at Hayton, East Yorkshire.³¹ Second, although no traces of any surface treatment were found (see above), we must consider that the whole may have been originally polished and painted,³² and that the veneer we see today does not represent the original appearance of the casket as conceived by the maker and seen by the owner.

THE SIGNIFICANCE OF THE MATERIALS

We have seen above that the pieces were manufactured from both bone and antler. This is unusual, and when we look at the distribution of bone and antler pieces a pattern emerges (fig. 3). The antler items are more tightly packed in Room B, whereas those of bone are spread over most of the width of the building. The distribution of the bone pieces is as follows (see also fig. 3):

• Room D: The most complete figure (fig. 7.1) and two of the vegetal trapezoids (fig. 9.22–23) are of bone. But also note that the three fragments of a similar figure (fig. 7.2 and 7.4–5; fig. 10.2–3) are manufactured from antler, as is the third vegetal trapezoid (fig. 9.21).

• Room C: One of the two faces with long curls (fig. 7.6).

• Room B: Two more heads (fig. 7.7–8).

• Outside Room B: The deer, sea-deer or vegetal element (fig. 9.26).

Apart from the last piece, which finds no parallel within the Fordham veneers, the other six pieces all fit within the overall decorative scheme suggested by the antler items, but the use of a different material plus the wider distribution might support the suggestion that originally there were two similarly decorated caskets rather than one.

SUMMARY

We have a very incomplete picture of the overall decoration. However, we can ascertain that the design consisted of circular or semi-circular design(s) formed by at least two groups of trapezoids (floral and alabastra), perhaps surrounding female figures. We have at least three examples of one figure, suggesting that they either formed a group or were placed on separate faces of the casket. In addition, there are rectangular designs consisting of two levels with a small blank area between them; these may have formed borders around the other designs. The zoomorphs are easily accommodated within this group, but there are several other pieces such as the architectural design that do not fit happily with what survives. There may have been two caskets sharing common themes, such as the floral design and the ‘wild haired female’, but this must remain speculative.

In the discussion that follows it is proposed first to discuss comparanda in Britain and elsewhere; then to suggest a possible theme, hazardous as it is with such incomplete evidence, and finally what may have been contained in the casket, including an account of two items worth publishing in their own right.

COMPARANDA

INCISED BONE AND ANTLER PLAQUES

Incised bone casings from wooden boxes or caskets are by no means uncommon in Roman Britain.³³ There are, for example, several late Roman wooden boxes covered with bone veneer with ring-and-dot and geometric decoration, compass- and graver-drawn styles of ornamentation that continued to be used on skeletal materials into the medieval period.³⁴ Such boxes are typified by examples from Richborough,³⁵ Winchester³⁶ and Hoxne.³⁷ The Winchester box was made to hold a double-sided comb with a matching design on the central panel. It had a sliding lid, as did another complete box from Heilbronn, Germany, which had a chi-rho on the lid and can be identified as Christian.³⁸ The veneer on such boxes was generally secured to the wooden core by small bone or copper-alloy rivets, perhaps alone or supplemented by glue. This use of all-over decorated veneer in the very late Roman period provides a background of decorated boxes against which the Fordham find may be considered.

A group of veneers from Lant Street, Southwark, were part of a box placed at the feet of a fourth-century female inhumation. One piece shows a draped half-length female figure, while other pieces suggest that she was standing within an aedicula, an indication of divine status. Copper-alloy rivets found in the grave were presumably also part of the box, which had probably originally been designed to hold jewellery or cosmetics, although no contents survived. The veneer here was not an example of ‘all-over’ decoration as on the Fordham pieces,³⁹ since it served to decorate just one end or side of the casket. A further example of veneer recovered from a closed context is from Lullingstone Roman villa, Kent, dated to c. a.d. 300. Here a group of veneers and a small central disc depicting a head, identified by the excavator as of Medusa, had been placed on top of a coffin lid.⁴⁰ Although not certainly from a casket, such is the most likely identification. Both the Southwark and Lullingstone examples have a central design, which, to judge from the number of veneers,⁴¹ appears to have adorned just one side of a box. They are also linked by the use of small veneers, some triangular, some rectangular with simple linear decoration. Similar pieces have been found at several sites throughout Roman Britain, suggesting that veneer-decorated caskets were common in the fourth century.⁴² The rather different shapes of the pieces from the Fordham casket emphasise the unique design.

There are a small number of other isolated examples of veneer plaques depicting figures which are likely to have formed parts of larger designs on caskets. One from a late fourth-century context at Great Casterton villa depicts the upper part of a woman shown full face and wearing what may be a crown.⁴³ The image was taken by Toynbee⁴⁴ to be that of a city goddess. The plaque had been cut around the edge of the figure and may have been inlaid into furniture or glued to a box, as suggested by Liversidge.⁴⁵ The incised bust from excavations at Wroxeter is similar but not so fine; it may depict a male, although no attribute proclaims his identity.⁴⁶ A nude female figure from the Temple of Nodens at Lydney Park, Gloucestershire, was almost certainly a medical votive rather than a depiction of the goddess Venus, as there appears to be a cist, or perhaps a representation of her womb, on her belly.⁴⁷

More sophisticated use of incised figure work is found on a fourth-century bone plaque from Egypt in Merseyside Museums; it depicts a winged youth holding a hare, probably derived from a personification of the month of October, as shown in the Roman Calendar of a.d. 354.⁴⁸ A pair of incised plaques from one end of a bridal casket from Saqqara in the Coptic Museum, Cairo, depicts two female attendants to a bride,⁴⁹ and this may have some relevance to the Fordham find, if the casket was a marriage gift, the figural decoration alluding to the ceremonial life of an élite woman (see below).

OTHER CASKETS

The Fordham casket may have been a jewellery box such as the one found in 1889 in the second- or third-century interment of a young bride, Crepereia Tryphaena, on the site of the Palace of Justice in Rome.⁵⁰ The deceased wore gold earrings, a necklace, finger-ring and brooch, but the burial is best known for its life-like articulated ivory doll, which had its own casket containing miniature jewellery, mirrors and comb. For our present discussion, this casket is important as it demonstrates that veneers could be applied to the outside of boxes that did not have flat, flush, faces. The casket is composed of sheets of ivory that are plain, apart from at the front, which is embellished with three veneer rosettes, each of which is composed of eight rhomboidal strips of bone.⁵¹ It recalls the subject of one of the best-known painted panels from the ceiling of Constantine's palace at Trier (c. a.d. 320) which depicts a woman, a princess or a personification, holding a jewel box from which she extracts a necklace of pearls.⁵² Of course, there is no way of telling the material of which this casket, suitable for an Empress, would have been made, but at that level it was surely constructed of ivory.

Classicising work carved in relief or in the round from ivory and bone was regularly employed for caskets. There are many examples from Egypt, often figuring Dionysos, Aphrodite and Eros and nereids,⁵³ but there are very few ivories from Britain, or indeed from Western Europe. Especially noteworthy, therefore, is the corner and foot of a small box excavated at Hill Farm, Gestingthorpe, Essex.⁵⁴ Its outer face side bears in relief the figure of an infant Bacchus holding a thyrsus, while below the ground-line on which he stands is a feline paw. Slots cut into the back and base of the piece would have taken the side and the base panels.⁵⁵ An ivory plaque from Leicester portraying a crouched animal -headed figure has similar slots to the Gestingthorpe piece and must originally have come from a similar type of casket.⁵⁶ Two ivory plaques from the fortress of legio II Augusta at Caerleon, Monmouthshire, were attached by rivets to a backing, and may also have been attached to a wooden casket; one plaque is in the form of a tragic mask and the other is a draped woman, perhaps a maenad, placing a basket of fruit upon the head of a boy.⁵⁷ However, a plaque depicting Cupid holding a basket of fruit, also in relief, from the Saxon Shore fort at Dover, Kent, is riveted to the remains of a knife blade; behind it is a plain ivory panel.⁵⁸ Finally, although it is in a very fragmentary state, an ivory pyxis of roughly the same date as the Fordham material was included in the Hoxne Treasure; it includes the figure of a satyr, and was probably Bacchic in its iconography.⁵⁹ It may have served, like the Fordham casket, as a possible receptacle for jewellery, although Johns points out that ‘objects of carved ivory were highly esteemed in antiquity for their own sake’.⁶⁰

THE THEME OF THE DESIGN

The placement of the various elements, human/divine figures, a marine element, an aedicule shrine and perhaps a stag, is hard to ascertain (see above), and, as suggested above, there may have been more than one box. The alabastra were evidently disposed in a circle and perhaps served as a decoration on the lid, much like the Heilbronn casket with a chi-rho on its lid.⁶¹ That casket also has part-circles on the sides, some intersecting, which point to alternative ways of arranging the trapezoidal pieces on the Fordham box.

An obvious starting point for determining a unifying theme is the complete standing figure with a shock of hair, who holds a roll of cloth in front of her. Venus was, of course, a popular subject for late Roman mosaics in Britain. These illustrate several features of our veneer. The Venus mosaic from Rudston, Yorkshire, depicts the goddess with hair flowing freely on both sides;⁶² a bust of Venus with long hair wearing a necklace around her nude shoulders on a mosaic from Bignor, West Sussex, shows Venus with a nimbus around her head;⁶³ and the Virgilian mosaic from Low Ham, Somerset, portrays Venus twice, in one case holding her mantle behind her and in the other bedecked with jewellery against her nude body.⁶⁴ Tempting as it would have been if we just had a single figure, we have to reject this explanation here, while maintaining the likelihood that there was a close connection with the goddess, as these figures in all likelihood represent her servants, nymphs or nereids, attending their mistress's toilet, and the inclusions of a male with belted tunic and sea-beasts would be in keeping with this theme. Such figures are portrayed on the sides of the much larger silver Projecta casket from the Esquiline Treasure, as well as on a cylindrical casket (no. 14) in the Sevso Treasure.⁶⁵ In addition to the High Rochester stone relief mentioned above, an example from Britain of a female figure holding a roll of cloth is a wooden statuette of probable third-century date from a temple in Winchester. Originally interpreted by Anne Ross as a statuette of the goddess Epona, this has recently been reinterpreted as a votary, presumably in this case about to present the garment to a god or goddess.⁶⁶ That the theme of the casket centred on bathing, an important aspect of daily life, is strengthened by the recognition that amongst the decorative elements on the pieces of veneer are alabastra, the little vessels that would have contained the oils and perfumes essential for bathing.

Apart from purely decorative items, there remain the other human heads. Those with the corkscrew curls or head-dress, so similar to those on the Chelmsford carving, bear a resemblance to the ornamental pendants hanging from the sides of the head of a stylised figure on a late Roman plaque, probably from a casket, found near a Christian basilica at Caričin Grad in Serbia.⁶⁷ These are ultimately dependent on the head ornaments of late Roman Emperors and Empresses. As a married couple are portrayed on the lid of the Projecta casket,⁶⁸ it may be that one of the Fordham heads with curly or waved hair, contemporary with early fourth-century hairstyles, may represent the master of the estate (the Dominus) while the other, more credibly female, may be that of his wife (the Domina).

THE DATE OF THE CASKET

As we have seen, the veneer derives from villa destruction deposits. Amongst the series of coins from destruction deposits are 27 of the 44 Theodosian coins of a.d. 388+ from the site. The deposits in which the veneer was found were sealed in places by an area of burning which had some copper and iron slag in it. This was cut by an inhumation radiocarbon dated to a.d. 431–618 (95.4 per cent probability) with a date in the second half of the sixth or early seventh century most likely.⁶⁹ The grave backfill also contained some copper and iron slag, presumably derived from the deposits mentioned above. The stratigraphic evidence therefore places the deposits in which the veneer were found between the very end of the fourth century and the mid-fifth to early seventh centuries. The comparisons we have discussed above place the designs on the veneer very firmly in the later fourth century or the very early fifth. However, as the casket had been broken before the veneer fragments were deposited, the construction date for the casket could be earlier in the fourth century.

The majority of the Fordham finds so far recovered are later Roman, but there is one object indicative of an even later date. Found in the robbed-out furnace room of the bath-house to the south of the villa building (fig. 1), and not necessarily contemporary with the casket, was a small antler pendant (fig. 14) decorated with triangles of ring and dots between three bands of incised lines; it measures 45 mm in height and 12 mm in diameter at the base. Such pendants are found throughout northern Europe and are generally ascribed a fourth- to seventh-century date. They are not frequently found in Britain where they typically have an early Anglo-Saxon association.⁷⁰ This could support the continuation of occupation on the site into the fifth or later centuries, although it should not be forgotten that the basic shape has Roman precedents in the ‘Hercules club’ pendants or earrings, such as those found in the late fourth-century Thetford Treasure.⁷¹

FIG. 14. Pendant from the Fordham villa destruction deposits (scale 1:1).

POSSIBLE CASKET CONTENTS

An élite woman's casket is likely to have contained jewellery and/or cosmetics. There are numerous female-gendered dress accessories from the site, including part of a bracelet and a gold finger-ring (fig. 15), both probably contemporary with the veneers although not found in the same contexts. The bracelet consists of two strands of copper-alloy wire and one of silver twisted together, a regular fourth-century type, of which there are several examples from Hill Farm, Gestingthorpe.⁷² The ring is exceptional as a villa find and is precisely the sort of item which might have been kept in such a casket, as its material marks it out as an object of real value. Found in 2016, it weighs just 1.37 g and has a diameter of 21 mm; the band is 1.0 mm wide and 1.42 mm thick. It too dates from the very late third or the fourth century and is just the sort of item a (putative) female owner would have kept in her casket. The decoration consists of V-shaped alternating notches, giving the effect of a continuous angular wave. No other gold examples of the type are known from Britain, although comparison may be made with an unprovenanced gold ring in the Koch collection, which is dated to the third century and is similarly plain and decorated with notches around its circumference.⁷³ The late Roman love of broken surfaces allowed the gold to shimmer, as is also the case with the ‘crinkum-crankum’ bracelets in the Thetford and Hoxne Treasures, although it must be stressed that these broken surfaces in the precious arts are a widespread Late Roman phenomenon.⁷⁴ The style of decoration is found on copper-alloy finger-rings, such as an example from a burial dated to later than a.d. 360 in the Butt Road cemetery at Colchester.⁷⁵ The design is more common on late Roman insular bracelets, again of copper alloy, from Poundbury, Dorchester, Dorset, Richborough, and in Essex at Hill Farm, Gestingthorpe, and Chelmsford.⁷⁶ In conclusion, it appears that the Fordham ring was almost certainly of British manufacture, perhaps made by a goldsmith fairly locally, either in Colchester or London, and could very likely have belonged to the wife of the owner of the Fordham villa.

FIG. 15. The gold ring (scale 2:1).

CONCLUSION

The Fordham veneers are a significant addition to our knowledge of late antique art in Roman Britain. Their decorative elements are in most cases unique but find broad parallels in other fourth- and fifth-century objects. The casket (or perhaps caskets) sits within a wider context of decorated boxes dating from the fourth century, and we believe that it belongs relatively late in the sequence, having been destroyed in the later fourth or early fifth century. While it is of insular manufacture, the themes it depicts are classical in nature. It is likely to have originally been a box belonging to an elite female, holding bathing equipment or jewellery, the latter possibly including a gold ring found on the site.