II. FINDS REPORTED UNDER THE PORTABLE ANTIQUITIES SCHEME

The Portable Antiquities Scheme (PAS) was established in 1997 as an initiative to record archaeological objects found by members of the general public. Following a pilot phase, it was extended to all of England and Wales from 2003 onwards.¹ From 2004 surveys of Roman-period finds documented by the PAS have been published in Britannia. This twentieth annual report briefly summarises the general character of Roman finds reported in 2022 and publishes significant individual artefact discoveries recorded by Finds Liaison Officers (FLOs).²

OVERVIEW

In 2022 18,466 objects of Roman date were documented by the PAS. Since a single record may sometimes document more than one item, above all in cases of coin hoards as well as in assemblages of ceramics or building materials, the total number of objects as usual exceeds the number of records made in the year (13,658). These quantities resemble those for 2021. These totals continue to be lower than those recorded in the years preceding the COVID-19 related lockdowns which began in March 2020.³ The results for 2022 likely indicate continuing impact of the pandemic both on the discovery of objects, on their reporting and on the work of FLOs in documenting them. The spatial distribution of the finds reported in 2022 across England and Wales closely resembles the distributions attested in previous years. The highest numbers of objects were reported from counties in eastern and central southern England.⁴ In Wales, northern and western England, the numbers of artefacts documented and of records made are smaller: most of the unitary authorities where ten or fewer Roman-period objects were recorded in 2022 are located in these areas. The most important factors responsible for this distribution of reported artefacts remain as follows: the circulation and deposition of metal objects in the past; historic and current land-use, including agricultural practice, and the distribution of heavily urbanised areas (with the partial exception of London); constraints on metal-detecting linked to land-use and its consequent uneven intensity.⁵

Discoveries of coins and brooches as ever significantly outnumber those of other Roman artefacts. 12,881 coins were recorded in 9,608 records (of ‘coins’ and ‘coin hoards’), a somewhat smaller percentage (c. 70 per cent) of individual objects than most years. With the addition of the those recorded in 2022, approximately 430,000 Roman coins have now been recorded, as documented in these annual summaries. This number continues to reflect both the numerical preponderance of coins among objects discovered and the emphasis of the PAS on recording assemblages of Roman coins in toto.⁶ 36,900 brooches have been documented by the Scheme since 2003, including the 1,812 further examples recorded in 2022. This total comprises brooches with date spans across the late Iron Age–Roman transition as well as those dated within the Roman period proper.⁷ The distribution of brooch finds continues to follow well-established patterns, larger numbers being documented in eastern and in (some) southern counties than in northern and western England and Wales.⁸ As in previous years, this disparity is complemented by significant regional variability in the ratio of coin to brooch finds. In particular, in much of Wales and adjacent English counties, especially in the north-west Midlands, the ratio of brooches to coins continues to be significantly higher than the average.⁹

ARTEFACT DESCRIPTIONS

The following entries publish highlights among the past year's finds recorded by members of the PAS and Treasure Department at the British Museum.¹⁰ The objects are selected for their contribution to the study of the material and visual culture of Roman Britain, in particular where new discoveries extend understanding of object form, typological and decorative variability and/or distribution. They are arranged in approximate geographical sequence, from northern to south-western England. The reference number associated with each record is the unique identifier which can be used to consult individual object records on the PAS website: www.finds.org.uk. Some entries report objects also treated as Treasure cases and are therefore also designated with their Treasure number in the format of year (20XX) and reference number (TXX).¹¹

The 23 copper-alloy, silver and gold artefacts included in this report were made between the late first millennium b.c. and late Antiquity, and belong to various functional categories, including figurines representing divine beings and animals, vessel mounts, furniture fittings and a likely staff terminal, horse harness pieces and items for adorning the human body, including rings, brooches including likely imported examples, sometimes with a military connection, and a hair pin, and tools for modifying personal appearance, including a nail cleaner and small knife handle as well as a possible oil container in anthropomorphic form. These discoveries illustrate the diversity of style and motif documented in portable metal objects from Roman Britain.

NORTH YORKSHIRE

(1) Conistone with Kilnsey (SWYOR-A187CC) (fig. 1)¹² A copper-alloy Trompetenmuster brooch 34 mm long, 21.6 mm wide and weighing 7.2 g. The openwork brooch is well preserved with an uneven light green patina and some traces of white metal on the upper surface. It has a roughly drop-shaped overall form, combining three trumpet motifs forming two loops. One loop is formed by two trumpets joined at their wider flaring mouths; their other terminals connect to the third trumpet which curves back on itself, swelling into a wide mouth, to close the second loop. The pin, mostly preserved, extends along the long axis of the brooch, its hinge being housed in a double lug beneath the conjoining trumpet mouths. The catch plate, partly preserved, is located beneath the wide mouth of the third trumpet.

FIG. 1. Conistone with Kilnsey, brooch (No. 1). Scale 1:1 (photo: I. Whitehead). (© West Yorkshire Archaeology Advisory Service)

This brooch finds parallels with others formed by two or more conjoined trumpet motifs. One such was also documented in 2022 from Long Bennington (LIN-6267B8), Lincs, somewhat bigger (39 mm long) and formed by two trumpet motifs set back-to-back, curving onto themselves as the loops of a figure-of-eight. This brooch type has a strong connection with the Roman army. Most with known provenance are documented on or in the hinterland of German and Danubian frontier sites, the largest group being from Zugmantel.¹³ An exceptional variant, configuring multiple trumpets as a swimming duck, was excavated from a ditch fill of Severan date at Vindolanda.¹⁴ The brooch form strongly echoes the much more numerous Trompetenmuster harness mounts of mid-second- to early-third-century date, also associated with the Roman army in the same regions, again especially well attested at Zugmantel, as well as similar belt fittings.¹⁵ The Long Bennington brooch specifically resembles a particular mount from the Netherlands also based on double-ended trumpets.¹⁶

(2) Clapham cum Newby (SWYOR-8F0CFF) (fig. 2).¹⁷ A copper-alloy and white metal repoussé disc plate brooch, of mid-imperial date, 36.5 mm in diameter and weighing 10.2 g. The brooch comprises a flat circular plate, the surface of which carries fragments of an applied repoussé white metal disc, possibly silver. This disc carries some surviving traces of stamped decoration. At its centre is a triskele motif, comprising a central circle and satellite smaller roundels, all holding dot-formed rosettes. Concentric mouldings, both simple and beaded, surround the triskele and extend to the brooch edge. Beneath, one lug survives of the pair which housed the hinge of the now lost pin. The rivet which attached it is visible on the brooch's upper surface, as is that on the opposite side for the catchplate, the latter curled to hold the pin. Adjacent is a hole, perhaps once occupied by an earlier fastening. In the centre of the brooch back is a further copper-alloy rivet, flattened flush with the surface, likely attaching the applied decoration. A second brooch was found in close association (SWYOR-8EF89F), very similar in form and dimensions (36.3 mm diameter) and with traces of the curvilinear decoration on the applied disc. The lugs are better preserved and the flattened rivet near the centre of the reverse is more evident.

FIG. 2. Clapham cum Newby, brooch (No. 2). Scale 1:1 (photo: I. Whitehead). (© West Yorkshire Archaeology Advisory Service)

The discovery in close proximity of a pair of near-identical brooches may indicate their derivation from a burial, though paired brooches are not often documented in funerary settings, especially in northern England where grave furnishing is typically limited. The plate brooch with repoussé triskele decoration is a widely documented form, dated by Mackreth to between a.d. 150 and 250.¹⁸ Similar brooches (eight) from the possible votive deposit at Piercebridge also carry elaborate triskeles, though these cover more of the surface, lack the rosettes and are edged by a single border.¹⁹ A few closer comparanda have been documented. In particular the notably better preserved brooch from distant Allithwaite by Grange over Sands, Cumbria (with silver appliqué) (LANCUM-D19609) reveals the full detail of the decoration, the rosettes being framed within a scroll which extends the brooch's surface, ending in trumpet terminals.²⁰ A brooch from nearby Ingleborough (LANCUM-58DE68) may be of similar type, but too little is preserved to state this with confidence. Mackreth and others propose a military association for brooches of this general type, in particular with reference to Severan refurbishment of northern garrisons.²¹ There is no nearby known fort in this case, though the findspot is not far from the hypothesised but unproven route from Skipton running west across the Pennines towards Settle and beyond.

WEST CHESHIRE

(3) Kingsley (LVPL-C7C9F6) (fig. 3).²² An incomplete copper-alloy figurine of a youthful satyr with strong affinities with the image of Bacchus, 115 mm high and weighing 154.1 g. The right arm and leg as well as left foot are missing but the figurine is otherwise well preserved, if worn, with its patina a smooth green, save at the break point on the right hip.

FIG. 3. Kingsley, figurine (No. 3). Scale 1:2 (photo: H. Beeton). (© National Museums Liverpool)

The figure is naturalistically rendered though wear has blurred surface detail. The head tilts to the left, encircled by a wreath tied at the nape of the neck, with clumps of fruit beneath the leaves which cover the ears, likely of ivy, the more common plant for garlands in Bacchic images. Clumps of hair frame the face, drawn up into a top knot above the temple, with shallow curving lines representing strands which radiate from the crown. The figure is naked except for a nebris (conventionally the skin of a deer or fawn) in the form of a quite thin band which wraps the body from right shoulder to left hip; at the latter point the hide's surface detail is better preserved, especially the fur rendered as short parallel incisions. The torso is well modelled, as is the musculature of the rounded buttocks and surviving leg. A small tail loops upwards from the base of the spine. The left arm extends to the figure's left from the shoulder, its hand now empty.

This figurine is similar in all but one respect to others identified as Bacchus, a god not often represented in this form in Britain or neighbouring provinces.²³ Its specific form and pose, of a smooth-cheeked youth standing naked save for a nebris with one arm lowered, the other extended, is paralleled in two groups of statuettes, one of an adult Bacchus, the other younger. Examples from Germany of the second group are especially similar, though they hold a drinking horn in the left hand while the Kingsley figure's hand once more likely held a thyrsus (staff).²⁴ A better preserved statuette from Ashby cum Oby, Norfolk (NMS-038224), shows a similar figure in a dancing pose, as does another figurine from Haynford.²⁵ Both may represent the god, especially the former, but they have previously been identified as satyrs. In the case of the Kingsley figurine the tail points more strongly towards identification as a satyr. The latter, often not dissimilar in form to the god himself, are attested in multiple roles within the Bacchic retinue, as dancers, flute players, wine carriers or pourers, sometimes propping up the drunken Bacchus, sometimes intoxicated themselves.²⁶ Although statuettes are uncommon, the proliferation of Bacchic motifs across Roman objects is widely attested by the PAS; satyrs join other hybrid creatures and party animals as single decorative motifs on mounts for furniture, vessels, and vehicles.²⁷

DENBIGHSHIRE

(4) Llandegla (LVPL-D780C9) (figs 4a and 4b).²⁸ A copper-alloy trumpet brooch dating to c. a.d. 75–175, near complete (lacking its pin) with a smooth dark-green patina. The brooch is 101.6 mm long and weighs 53.2 g. The long bow is D-shaped in section. It is divided in two parts by a large semi-circular waist moulding, itself framed on either side by three smaller fluted mouldings. Above this waist the bow widens into the trumpet head from which a wide circular head-plate extends (diameter 31.8 mm), ending in a small suspension loop. Set off-centre on the head-plate, near the loop, is a rosette, standing proud with a rising stem at its middle. To either side are eight stamped circles. On the reverse is the hinge for the missing pin. Below the waist the bow (with three longitudinal ribs) tapers to a rounded foot knob, separated by further fluted mouldings. The long triangular catchplate survives complete.

FIG. 4a. Llandegla, trumpet brooch (No. 4). Scale 1:2 (photo: H. Beeton). (© National Museums Liverpool)

FIG. 4b. Distribution of trumpet brooches (published examples) documented by the PAS (map by Edwin Wood, base map from Ordnance Survey Open Data).

More than 3,000 trumpet brooches are now documented by the PAS, distributed across Wales and England. The findspot of this example lies on the western margin of the type's main concentration in Midland and north-eastern England and East Anglia, not far from one of the rare documented brooch production sites at Prestatyn.²⁹ This brooch, the largest (by length) documented among these 3,000+ examples, and larger than almost all documented by Mackreth, has affinities with sub-types mainly found in the western part of this area. Its general form recalls Bayley and Butcher's group C and Mackreth's group 1.x.2, though the head-plate is larger than in other examples of this type and is responsible for the brooch's exceptional size.³⁰ The rosette on the Llandegla brooch's head-plate is also paralleled in a cluster of similar trumpet brooches, albeit sometimes simplified into a more wheel-like motif. These too are usually larger than average. They include brooches from nearby Gwernafield (WREX-302324, 85.2 mm) and Gresford (LVPL-68B8B4, 81 mm), as well as two from the Little Orme's Head hoard near Llandudno, respectively 90 mm and 99 mm long, one missing its rosette. More distant examples are documented to the west from Pentir, Gwynedd (GAT-BE46B4) and from Segontium fort on the Menai strait, the latter with the same stamped circle decoration around the margin of the headplate, and to the east from Ilam, Staffs (LVP-3056D6).³¹ The Gwernafield and Ilam brooches share a larger suspension loop. This represents a further sub-group within the noted regional variability of the trumpet type.³²

The rosette motif on the head plate or loop is shared with other trumpet brooches which stand out through their form or decoration. A part-preserved silver example from Ulgham, Northumberland (NCL-609BF6) carries a rosette motif on the bow adjacent to a square head plate characteristic of brooches labelled by Mackreth as ‘very large with extraordinary features’, including the ‘giants’ from Great Chesters and Perth, both over 170 mm long.³³ Another member of the latter category, found at Auldearn, Highland, also carries stylised rosette-like motifs on the head plate.³⁴ The diameter (10.1 mm) of a detached silver rosette recently documented at Chilcomb, Hants (BERK-B79E76) suggests that it too once belonged a very large trumpet brooch.³⁵

EAST YORKSHIRE

(5) Ellerker (LVPL-080ED2 / 2022T261) (fig. 5).³⁶ A complete sheet-gold oval ring setting depicting Leda and the swan in repoussé. It is 13.4 mm long and weighs 0.5 g. A raised border runs around the setting, within which Leda is seated, facing left, her left leg before her right, with Jupiter in swan form perched on her left leg. He faces Leda, his wings folded behind the body, neck arched and head extended forwards, almost touching Leda's. Beneath the pair is a partial stamped beaded border.

FIG. 5. Ellerker, ring setting (No. 5). Scale 2:1 (photo: H. Beeton). (© National Museums Liverpool)

Among precious metal ring settings the same subject also features on examples from Elms Farm, Essex (silver-gilt) and Great Walsingham (silver).³⁷ This is a further instance of the ‘immense succès’ of this mythological coupling across multiple Roman media, also rendered in other ring settings, for example glass and jasper intaglios from Pentney, Norfolk (NMS-49D8E7/ 2006T305) and Springhead.³⁸ The general use of precious metal settings to embellish finger rings is also documented in numerous similar-sized examples recorded by the PAS. Mythological subjects include Mercury, Vulcan, Ganymede and Jupiter, and Pan.³⁹ Dextrarum iunctiones (shaking hands motif) are also a widely documented motif in this medium from southern Britain and beyond.⁴⁰

WAKEFIELD

(6) Hemsworth (LIN-49BDD1/2022T537) (fig. 6).⁴¹ An incomplete silver-gilt bow brooch with affinities with the Polden Hill type, carrying extensive curvilinear decoration, 37.9 mm long, 32.7 mm wide and weighing 27.7 g. The brooch is decorated on the front and sides; the plain reverse is bordered by straight mouldings. At the head the brooch's wings are cylindrical, with the casing partly broken away on the reverse, revealing fragments of spring and axis bar. A hole allowing the axis bar to pivot survives on one wing cap. From the back of each cap, edged with mouldings, extends a small hook, one better preserved than the other. From the head the bow curves and then tapers with a ridged spine towards a now missing foot. The pin too is lost and a worn remnant of the catchplate survives beneath the bow.

FIG. 6. Hemsworth, brooch (No. 6). Scale 1:1 (photo: L. Brundle). (© Portable Antiquities Scheme)

The wing caps are decorated with leaf-framed fans, centred on the hole for the axis bar. From the caps extends curvilinear decoration which develops along the front and sides of the brooch, symmetrically arranged around its central axis. The sinuous motifs are created through mouldings which comprise a central ridge running between two channels. These sometimes run in tandem to shape larger forms along the front of the brooch. These larger forms flare and swell in higher relief as they converge at the centre of the bow and then taper and contract as they fall away at the sides, culminating in two opposed high relief S-scrolls before the bow's broken end. The decoration on the sides of the brooch stands in counterpoint to the front. Mouldings are shallower and narrower, with curves sometimes echoing the larger forms, but sometimes undercut to form cusps, punctuating the momentum of the decoration as it scrolls along the brooch.

In form this brooch's closest parallels lie with Polden Hill types, although it is difficult to assign to a particular category.⁴² It thus likely dates from the later first to later second century a.d. Examples in silver are uncommon, accounting for only five of c. 2,800 Polden Hill brooches recorded by the PAS, some of which echo the Hemsworth brooch's form.⁴³ Its decoration recalls that on other brooches with extensive curvilinear motifs, for example Aesica types, as exemplified in the eponymous brooch from Great Chesters and those from Backworth.⁴⁴

LINCOLNSHIRE

(7) Near Bassingham (LIN-AED713/ 2022T877) (fig. 7).⁴⁵ A gold finger ring, near complete, with a bezel set with a dark red intaglio. The external diameter of the small ring is 18.8 mm, with a bezel 11.6 mm long. The ring weighs 3.4 g. The hoop comprises three strands, separating to form openwork shoulders which carry opposed S-shaped filigree decoration with a gold pellet at both ends of each scroll. The bezel takes the form of an oval box-setting with a rope-work flange, likely soldered to a backplate formed by the hoop where it widened. The oval intaglio, perhaps re-used from an earlier ring, is set flush with the bezel and is incised with a Capricorn facing left, its details (bifurcated tail, horns) quite schematically rendered but clearly visible.⁴⁶

FIG. 7. Near Bassingham, finger ring (No. 7). Scale 3:2 (photo: L. Brundle). (© Portable Antiquities Scheme)

The form of the ring with its flaring shoulders, box bezel with rope-work collar and decoration with filigree and pellets suggests a late Roman date. The rings in the Thetford hoard, dated to the final decades of the fourth century or early fifth provide parallels for this example, as do other likely fourth-century gold rings in the British Museum collection, for instance from Richborough and from Odiham, Hampshire.⁴⁷ A cluster of examples from North Yorkshire show general similarities with this ring.⁴⁸ Other rings distributed over a wider area parallel its individual details, including filigree, pellets and elaborate openwork shoulders.⁴⁹ Collectively the substantial number of late Roman gold rings now documented by the PAS expands the repertoire of forms of this bright, often polychrome and highly textured personal ornament.

(8) Long Bennington (LIN-2B92B3) (fig. 8).⁵⁰ A near-complete and well-preserved if worn copper-alloy hairpin. The shaft and terminal combined measure 106.3 mm. The pin weighs 18.9 g. The terminal takes the form of a double-handled vessel or cantharus, closed at the top, short handles between its neck and shoulders and a foot formed by one of two circular mouldings which separate terminal from shaft. The shaft's sides are slightly faceted, the parts closer to the terminal carrying repeated close-set diagonal lines. The shaft tapers to a point, slightly broken. With its terminal in object form this hairpin joins a group represented by a limited number of other examples (Group 18C, ‘object-headed pins’) in Cool's classification of hairpins from Roman Britain.⁵¹ More widely in the empire pins with a cantharus terminal are documented in other materials, bone, gold, and jet, although not commonly in copper alloy.⁵² In Britain the form is best documented in jet hairpins from burials in late Roman York, especially from the Railway Station cemetery where one pair was found fixed in a well-preserved head of hair in a plaster burial.⁵³ While the cantharus motif lends itself to adaptation as a terminal, like the frequent conical and knobbed types, it may have Bacchic associations, given its origins in the double-handed drinking cup linked to the symposium and Dionysus. Such symbolism was more consciously activated, perhaps, in funerary contexts like those from York.⁵⁴

FIG. 8. Long Bennington, hairpin (No. 8). Scale 2:3 (photo: L. Brundle). (© Portable Antiquities Scheme)

(9) Empingham (LIN-86E933) (fig. 9).⁵⁵ A small copper-alloy knife handle with a bird's head terminal, quite well-preserved, with a dark-brown patina and some patches of active corrosion. The handle is 33.7 mm long and weighs 15 g. It is approximately rectangular in form and cross-section. The stylised bird's head is represented by a short curved beak, eye and tripartite crest. On each side the eye is formed by a large circular recess, perhaps originally inlaid with glass or enamel. The eye is framed by an incised line which scrolls outwards to extend across the handle, separating two curvilinear zones, one decorated with basket-weave hatching (one side more systematically than the other), the other plain. Corroded remnants of the iron blade/tang are visible on both the narrow sides of the handle. Traces of a likely rivet can also be seen on both sides.

FIG. 9. Empingham, knife handle (No. 9). Scale 1:1 (photo: L. Brundle). (© Portable Antiquities Scheme)

The stylised animal form, incised curvilinear decoration and cross-hatching place the handle in the decorative tradition of the late Iron Age, finding parallels in mirror decoration, for example.⁵⁶ As a decorated knife handle the object is hard to parallel among documented late Iron Age artefacts, and belongs broadly to the much larger corpus of Roman-period small knives of fixed- and folding-blade form. No specific parallel to the decoration of the Empingham handle is known among these but small knife handles with stylised avian and canine terminals are documented in both excavated and PAS examples.⁵⁷

NORTHAMPTONSHIRE

(10) Whilton (ESS-2E252D) (fig. 10).⁵⁸ A copper-alloy sub-rectangular mount, cast in the form of a bearded and horned male face, its details quite finely modelled, largely intact if worn and disfigured by extensive corrosion. The mount is 47 mm wide, 50.3 mm high and weighs 85.8 g. The beard is rendered as thick clumps and the hair as sinuous locks which rise from the forehead, mirroring the spiralling ridged horns on either side of the face. The latter is mask-like, with a prominent brow ridge, worn nose and round eyes with carefully rendered lids and pupils. The mouth is a narrow slot obscured like other adjacent detail by corrosion. The reverse of the mount is concave, with the cavity to the rear containing a projection (rivet?) set within whiter metal (lead?) and a mass of iron corrosion.

FIG. 10. Whilton, Jupiter Ammon mount (No. 10). Scale 2:3 (photo: L. Brundle). (© Portable Antiquities Scheme)

The face is recognisable as that of Jupiter Ammon, from its distinctive combination of Jupiter's bearded features with Ammon's ram horns. This is the first documented example from Britain of a mount of mid-imperial date (c. a.d. 100–300) noted across the Roman world, including a large group of 12 near-identical mounts from Banasa, Morocco, probably from a chest.⁵⁹ The Banasa mounts are a little larger, with a face set in higher relief on a disc 7.4 cm in diameter, but the mask itself has the same diameter as the Whilton face. However, the face is modelled with greater restraint in the Banasa mounts and most other examples, in contrast to the fuller hair and larger horns of the Whilton piece. Given the size and form of the Whilton mount, it too is likely to be a furniture decoration. The motif has an imperial pedigree, being adopted into Augustan imagery from Hellenistic ruler portraiture, most famously in the marble clipei from the Forum of Augustus in Rome and related monuments.⁶⁰ It was copied in images linked to imperial authority, by client kings like Cunobelinus on their coins or on military dona.⁶¹ It then proliferated onto diverse media, including other portable objects, for example, lamps, rings, intaglios and button-and-loop fasteners, becoming a general decorative and possibly apotropaic motif.⁶²

(11) Titchmarsh (NARC-E3BCDD) (fig. 11).⁶³ A complete copper-alloy strap mount, well preserved with a smooth green-brown patina and traces of enamel. It is 71.8 mm long and weighs 46.1 g. It comprises an openwork loop and a rectangular plate, connected by a right-angled collar.

FIG. 11. Titchmarsh, strap mount (No. 11). Scale 2:3 (photo: E. Cox). (© Northamptonshire County Council)

The loop frames a face-like configuration of two roundels and an openwork mouth-like pelta. The strip forming the loop itself widens with distance from the collar and on its expanded zone carries enamel inlay. This starts on each side with a triangular motif, followed by five rectangular cells, inlaid with yellow enamel. These frame three roundels, similarly inlaid. The roundels within the loop are similar but larger, their channels being inlaid with red enamel, their inner recesses with yellow. The collar expands where it joins the plate, being thickened on the front with a strip-like moulding with a rectangular strap slide projecting from the back. Below a further row of three similar roundels, also inlaid with red and yellow enamel, is framed by two incised lines.

The mount has strong affinities with lyre-shaped strap mounts dated from the later first century b.c. to first century a.d., most being documented in the East Midlands and East Anglia.⁶⁴ An association with horse harness is likely. The prongs which project from other examples have suggested their use as cheek pieces to discourage horses in harness from turning their heads and thus to ensure greater control of vehicle direction, a possibility supported by the discovery of a mirror-image pair in the Wood Dalling hoard, Norfolk.⁶⁵ In several respects, however, the Titchmarsh example stands out among these mounts, expanding their formal and decorative repertoire. It lacks the studs or prongs characteristic of most, its openwork loop is circular rather than ‘lyre-form’ and it places the decorated roundels within the loop, creating the distinctive face-like appearance noted above.

NORFOLK

(12) Wood Dalling (NMS-AE7F1C) (fig. 12).⁶⁶ A copper-alloy figurine of Mars, c. 74 mm high, worn with a mid-green patina, the left hand partly and the right hand wholly lost to damage. Both legs are bent, the left placed forward, giving the appearance of a figure in motion. The left arm is held downwards and closer to the body, with the hand likely once resting on a now lost object. The right arm extends outwards, once perhaps holding a spear. On the slightly outsize head the eyes are large, with recessed pupils; the nose and mouth are almost lost to wear. The figure is clean-shaven and the hair is hidden by a very worn crested helmet. The stocky figure is clad in armour, most readily identifiable by the pteryx formed by a single row of fabric strips beneath a belt rendered as a narrow moulding with diagonal incisions. Grooves on the upper arm represent the tunic edge beneath what must be a cuirassed torso. There is limited modelling of the leg musculature and no visible detail indicating greaves or footwear.

FIG. 12. Wood Dalling, figurine (No. 12). Scale 2:3 (photo: A. Marsden). (© Norfolk County Council)

The armour allows the figure to be identified as Mars, one of 18 examples of figurines of this god in standing form now recorded by the PAS, alongside 24 documented from other sources.⁶⁷ The armoured version of Mars is best known as the Mars Ultor type, frequent among PAS examples.⁶⁸ The Wood Dalling figurine exemplifies one plain variant of this type, with an undecorated cuirass, a single row of lappets (strips) in the pteryx and clean-shaven features in place of the full beard.⁶⁹ The stance of the god is a little distinctive, seemingly almost moving rather than in the characteristic contrapposto pose and he lacks the greaves seen on most examples. With a similar example from London it nonetheless illustrates the presence of this widely documented variant in Britain.⁷⁰

ESSEX

(13) Near Southend (ESS-3DB994) (fig. 13).⁷¹ A near complete copper-alloy vessel 8.5 cm high in the form of the bust of a youthful bearded man emerging from a waisted calyx. Two integrally cast suspension loops rise from the top of the head on either side of an aperture. The handle once secured by these is lost. A lug survives from the now lost hinged lid.⁷² The bust lacks visible detail, but the head is carefully modelled with sub-Saharan African features. The hair's curls are formed from parallel strips incised with transverse lines, arranged in opposed pairs. One set of strips radiates from the aperture at the top of the head, extending halfway down it on back and sides and to the temples in front. The strips’ angled ends interlock with those of a lower row of similar strips, bringing the hair to the neck. The facial hair is rendered by clumps of curling hair on the lower jaw, a short moustache and a beard (not well preserved) forming opposed corkscrew curls on the chin. The worn face is sensitively modelled, smooth-cheeked, with a broad nose and quite deepset eyes with heavy lids.

FIG. 13. Near Southend, anthropomorphic vessel (No. 13). Scale 2:3 (photo: C. Rangel de Lima). (© Portable Antiquities Scheme)

Anthropomorphic vessels of this form are conventionally designated as balsamaria, hypothesised to have contained either aromatic resins, perfumes or perfumed oil, incenses, or even spices, etc. In the absence of a watertight stopper their storage of a dry substance may be more likely than that of a liquid, although the recurrent funerary association with strigils may support the possibility that they contained oils.⁷³ Almost 300 were identified in a survey completed almost 30 years ago. This type in the form of a black youth is one of the commonest, documented in more than 100 examples, some as in this case with stylised facial hair, others without.⁷⁴ Whatever the finesse of the modelling, this individual's features show somatic stereotypes widely used in figural art on Roman artefacts to identify individuals as an ‘Aethiopian’ or Aethiops, i.e. an inhabitant of sub-Saharan Africa, especially regions towards the Red Sea, as well as of regions beyond it to the Indian Ocean.⁷⁵ Such images are thus ‘figures of appropriation’, their range of reference conditioned by the setting of the image or its context of use. In this case, therefore, they might signify the distant origins of the vessel's contents, or perhaps served as proxy images for the enslaved individuals whose labour and looks were likely exploited in the grooming routines in which those contents were used.⁷⁶ The clustering of findspots of this and other copper-alloy balsamaria (in anthropomorphic form or with relief decoration) on the Thames estuary points to the imported status of these vessels, thought to be made in an as yet unidentified western provincial setting.⁷⁷

HERTFORDSHIRE

(14) Puckeridge (BH-747C63) (fig. 14).⁷⁸ A fragmentary copper-alloy object in the form of a horned human head, 33.8 mm long, 19.6 mm wide and weighing 11.2 g. Its curvature suggests that this may be a finger-ring bezel, with part of the shoulder preserved, although the break points are set slightly asymmetrically in relation to the face. Patination of these suggests breakage in antiquity. Large eyes dominate the broad face, delineated by a thicker moulding. Wear has removed most of the nose and mouth. The slightly curved horns project forwards over the face, tapering towards their tip, turning inwards, perhaps part of a headdress. Short diagonal incisions on the moulding framing the face may indicate hair, fur or fabric.

FIG. 14. Puckeridge, ring fragment (No. 14). Scale 1:1 (photo: M. Fittock). (© Portable Antiquities Scheme)

It is not clear if this is a hybrid creature or human. Representations of possibly horned human figures are uncommon in late Iron Age and early Roman art and lack a close resemblance to this object. For instance, the pelta-form crests which spring from the centre of the forehead on bucket mounts, for example from burials at Baldock or Welwyn, are very different.⁷⁹ The anthropomorphic vessel fitting from Boughton Aluph, Kent, and rein ring fitting from Aldborough have horn-like projections sprouting from either side of the head, as does a distinctive bucket mount from the Wallingford area (BERK-783763), but the resemblance is otherwise very limited.⁸⁰ Similarly there is little more than a passing echo of other horned figures, such as the famous antlered beings on the Gundestrup cauldron or in stone relief form on the Pilier des Nautes.⁸¹

WEST BERKSHIRE

(15) Streatley (SUR-C884C4) (fig. 15).⁸² A copper-alloy component of a strap fastener for horse harness, 34.3 mm long and 15.5 mm wide. This comprises a loop with a frame of keyhole-shape, with lobes from top and sides, the former in trefoil form, the latter curling back on themselves. This is attached to a collar which carries a strip of silver inlay, flanked by moulded ridges. On the other side of the collar projects a small loop which was part of a hinged connection to the main body of the fastener, now lost, which would have been fixed to the strap by studs. The keyhole aperture forms the ‘female’ part of a ‘bar-and-keyhole’ strap fastener used on horse harness. A T-shaped projection on an opposing strap terminal would have fitted through the keyhole slot.⁸³

FIG. 15. Streatley, strap fastener (No. 15). Scale 1:1 (photo: S. Maslin). (© Portable Antiquities Scheme)

The survival of many strap fittings of this type from military sites allows it to be dated to the first century a.d., with examples surviving from contexts early in the century onwards.⁸⁴ For instance, a complete fastener survived among the famous silvered harness fittings from Xanten.⁸⁵ The presence of inlay is quite widely documented. Further military-associated examples are also known from Camerton and Caerleon.⁸⁶ Several others have been recorded by the PAS from central and southern England, usually also of the keyhole element. As in this case these have sometimes become detached at the hinge from the main body of the fastener. These extend the range of findspots beyond known garrison contexts, one (Farnborough, W. Berks) lying only c. 10 km to the west of the Streatley fastener.⁸⁷ Among these the strap fastener from Owslebury deserves particular mention for its preservation and context. It preserved the bar which connected to the strap as well as the keyhole-shaped fitting and traces of foliate decoration on a silvered surface. Its findspot lies within a landscape with a large body of metal-detected finds of late Iron Age and very early Roman date but so far lacking any evidence for a garrison site.⁸⁸

KENT

(16) Chilham (KENT-E3CD00) (fig. 16).⁸⁹ Part of the left side of a Jupiter figurine, modelled with a high degree of naturalism with a pale-green patina. The object is 84 mm long and weighs 71.3 g. The fragment includes the clothing draped over the left shoulder of the figure and hanging down its side as well as the left arm, visible from the elbow. The upper arm is held close to the side while the forearm turns forward and outward, the wrist turned towards the body and the hand closed on a now lost object. The clothing is draped over the left shoulder and forearm, its falling folds being exceptionally well modelled, dropping at their furthest down the side to at least the calf. At the side the curves are longer and gentler; at front and back, where the cloak is hitched higher, the folds end in shorter zig-zags.

FIG. 16. Chilham, figurine fragment (No. 16). Scale 2:3 (photo: L. Rogerson). (© Portable Antiquities Scheme)

Despite the absence of the rest of the statuette, the figure to which this draped limb belongs can be identified with some confidence. At first glance the most likely candidates are Jupiter and Mercury, both of whom are sometimes represented in standing pose with a cloak suspended from their left shoulder and arm. In both cases the cloaks can be either more voluminous or more exiguous, suspended either from the arm from shoulder to elbow, or only from the shoulder.⁹⁰ In general, however, the arrangement of the folds corresponds very closely to the widely distributed statuette type showing Jupiter with thunderbolt in his right hand and sceptre in left.⁹¹ The pose of the Chilham fragment also helps to associate it with the same Jupiter type, in which the god's left arm is often raised to allow the hand to grasp a spear or sceptre high up its shaft; for statuettes of Mercury, cloaked or uncloaked, the left arm and hand are never raised in this way.⁹² The separate casting of this element of arm and cloak is not uncommon, being documented in other very similar fragments of statuettes as well as complete examples, the latter being especially well illustrated in figurines from Zollfeld (Kärnten), Auvernier (Neuchâtel) and Brigetio (Szőny).⁹³ The most striking characteristic of the Chilham fragment is the sophistication of hanging drapery, expressed in the layering and folding of the cloth. Comparison with a fragment of drapery from St Albans, also from the left side, well realised but without the same complexity, serves to highlight its quality, rarely seen in bronzes from Britain.⁹⁴ This elegance, also characteristic of many part-statuettes of this type noted above, would conventionally prompt its identification as a likely import.⁹⁵

SURREY

(17) Limpsfield (SUR-5770C2) (fig. 17).⁹⁶ A copper-alloy bracelet segment, with some traces of gilding, heavily worn and broken at the edges, probably from a hinged-bezel type bracelet. The fragment is 28.8 mm long and weighs 5.9 g. On both sides are double-lugged hinges with perforations for an axis bar, serving to connect this segment to the bracelet proper. Both hinges are decorated with oblique incision. The slightly convex disc which forms most of the fragment carries worn concentric decoration, the circular mouldings being separated by narrow channels. Some mouldings are thinner and plainer, others are thicker with ropework decoration formed by incision set in alternate directions. At the centre is a circular insert in a different material framing a central recess for a now lost setting.

FIG. 17. Limpsfield, bracelet fragment (No. 17). Scale 3:2 (photo: S. Maslin). (© Surrey County Council)

No exact parallels have been found but surviving whole and part bracelets with hinged medallions or central segments hint at the original form of bracelets of this type. A silver armlet identified as an armilla from the Boudican-period Fenwick hoard at Colchester has a hinged medallion as its focus, while a round medallion set with an intaglio figuring Jupiter is the focal element of a mid-imperial silver bracelet from the Dalton area, Cumbria.⁹⁷ Various examples of whole and part preserved thin bracelets with enamelled decoration from Gaul exemplify more modest examples in copper alloy.⁹⁸ Gold and silvered examples are documented elsewhere.⁹⁹

HAMPSHIRE

(18) Mortimer (BERK-AE0EAD / 2021T458) (fig. 18).¹⁰⁰ A sheet-gold disc 22.2 mm in diameter, 0.7 mm thick and weighing 0.7 g, carrying a repoussé portrait image tightly framed within a beaded border, with a second circle of larger beading (damaged) forming the outer border. The male portrait head faces right, with a laurel wreath with ribbons at the nape of the neck. The hairstyle and facial features resemble those of Julio-Claudian emperors.

FIG. 18. Mortimer (No. 18). Scale 1:1 (photo: P. Smither). (© Portable Antiquities Scheme)

The object is difficult to identify confidently both in terms of medium and subject. A small number of metal sheet images made as impressions from Roman imperial coinage are documented for Britain.¹⁰¹ However, size and detailing in this case indicate that it was not made from a coin rubbing, although it is likely inspired by a coin image. It has some affinities with other small-scale imperial portraits associated with portable objects. One such group, showing specifically the Julio-Claudian house and found entirely in the north-west provinces, comprises the decorative discs used as belt and scabbard fittings which emphasise the emperor's martial qualities by including attributes like the laurel wreath, armour, etc., also often drawing on coin types for their inspiration.¹⁰² The other is a more miscellaneous group of imperial figures, distributed over a longer period, which show the emperor's profile image on pendants, rings and so on.¹⁰³ Examples from Britain include a ring from Clifton Down, Bristol, with a portrait of Nero as the device on the bezel¹⁰⁴ and a very low relief portrait of Faustina Maior set in a possible belt fitting from Colchester.¹⁰⁵ As for the subject, the portrait most closely resembles Augustus and Tiberius but is hard to identify with confidence. Whatever its specific character, the disc provides a further example of small-scale representations of the Julio-Claudian emperors circulating in Britain.¹⁰⁶

WILTSHIRE

(19) Tisbury (OXON-DA8EA4) (fig. 19).¹⁰⁷ A complete decorated nail cleaner made from copper-alloy sheet, well preserved with a brown-green patina. It is 58.5 mm long, and weighs 1.9 g. From the suspension loop the nail cleaner narrows before swelling towards the centre, then narrowing again towards the forked terminal, cut by a short triangular slot. The nail cleaner carries incised decoration on all sides from top to tip. On one of the wider surfaces this includes alternating cross-hatching and sets of parallel transverse incised lines, on the other transverse incised lines only. The narrow sides also have approximately matching similar alternation of motifs.

FIG. 19. Tisbury, nail cleaner (No. 19). Scale 1:1 (photo: E. Caswell). (© Portable Antiquities Scheme)

The Tisbury nail cleaner is not easily assigned to one specific group. Its closest affinities are with the nail cleaners from nearby Wanborough characterised by Eckardt and Crummy as the ‘long exuberantly decorated’ group. It also the echoes the decoration on the ‘Late Roman tubular nail cleaner’ group, a widely distributed group with stratified examples from late fourth- and fifth-century a.d. contexts.¹⁰⁸ Along with further new finds from 2022, for example a very well preserved nail cleaner of the ‘early Roman Bone Disc’ type from South Marston, nr Swindon (HAMP-5112A2), it adds to the typological and decorative diversity of the more than 600 examples documented by the PAS.¹⁰⁹ The identification of these objects of later Iron Age and Roman date as nail cleaners is conventional, rather than definitively established, but they are a key part of the repertoire of instruments created in this period for bodily modification and/or possible medical purposes.

GLOUCESTERSHIRE

(20) Dursley (GLO-7A3DE9) (fig. 20).¹¹⁰ A copper-alloy convex circular mount bearing the head of Medusa, generally well preserved with a dark green patina but with some patches of active corrosion, modelled in detail. The mount is 73 mm in diameter and weighs 162 g. The gorgon's staring face stands in high relief. The well-shaped eyes are framed by mouldings for eyelids; sheet silver is applied to the eyes, with a central perforation exposing the copper alloy beneath as the pupil. Corrosion has removed much of the nose and mouth. Thick rope-like locks, modelled in high relief, sweep back from the face at the top and surround it at the sides. In places these serpentine locks turn in across the face where they are sometimes broken off. Over the temples corrosion and damage make the arrangement harder to unravel, but there appears to be a central parting, though it is hard to identify traces of wings. Around the edge of the mount are five small holes for attachment. The reverse is concave and plain but carries concentric lines and a small central hollow from its making on a lathe.

FIG. 20. Dursley, Medusa mount (No. 20). Scale 2:3 (photo: K. Adams). (© Bristol City Council)

As an apotropaic and general decorative motif the Medusa face is widely attested on portable objects, including those documented by the PAS.¹¹¹ The closest comparanda are other mounts which in principle can be distinguished by size and method of attachment, phalerae being wider and lighter and generally attached to their backing by pinning or stitching, harness mounts by their studs or loops for attachment and furniture mounts with evidence for rivets or solder.¹¹² Large roundels also occur on other objects, for example as medallions in handled pans.¹¹³ In practice, size is the main attribute for differentiation, with phalerae in particular being distinctly larger than others.¹¹⁴

The size of the other 16 examples documented by the PAS suggests that they too belong to the mount category, their largest dimensions typically lying between 30 and 45 mm; they are also flatter, though few have evidence for fixing. There is lead packing within the cavity of the mount from South Ferriby, North Lincs (NLM-C8D017) and traces of rivets from Stockbury, Kent (KENT-036B4C), Acton, Cheshire East (LVPL-3A0412) and Naburn, York (NLM-E131A5). A single attachment hole is pierced on the edge of the mount from Halton, Merseyside (LVPL-1DBDD5). Embellishment as here usually involves the application of other metals though a rare example from Cynwyl Gaeo, Carmarthenshire (SUR-1354F7) frames the gorgon's face with enamelled decoration. Only the furniture mount from Beeston with Bittering, Norfolk (NMS-650F10), approaches the size of the Dursley piece, 50 mm in diameter, also in high relief and with traces of turning.¹¹⁵

(21) Cotswolds (GLO-452F33) (fig. 21).¹¹⁶ A copper-alloy terminal in owl form, 68 mm tall in its entirety, quite worn but with some surface detail surviving. The owl stands on a hollow waisted pedestal that expands into a broader flange (41 mm in diameter). The bird stands upright, its body tilted a little towards its left. A pronounced grooved bill and above all the large eyes dominate the face – inlay has been lost from the eye cavity which perforates the head. Above the eyes some traces of feathering, rendered as vertical grooves, survive on the pronounced eyebrows. On the stocky body slight traces of feathering are also visible as repeated semi-circles below the neck. The wings are folded on the bird's back with the tail projecting to the rear; traces of feathering are just visible but detail is mainly lost to wear. A circular hollow above the right leg may be an artefact of the casting process. The cross-hatching which renders feathering is much better preserved on the stout legs. Both feet are visibly divided into the three forward claws.

FIG. 21. Cotswolds, owl terminal (No. 21). Scale 2:3 (photo: K. Adams). (© Bristol City Council)

As an attribute of Minerva, owls are sporadically documented in small-scale form, for example as plate brooches, occasionally as pin heads as well as mounts and figurines.¹¹⁷ The closest parallel in this case lies with the owl from Willingham Fen, argued from the association with copper-alloy binding and terminals to have been a possible mount for a sceptre.¹¹⁸ Examples from Oedenburg (Bas Rhin) and Chester convey an impression of the original appearance of such figurines when the inlay (respectively glass and enamel) survives in situ in the eye cavity.¹¹⁹

(22) Wickwar (GLO-0ED7EB) (fig. 22).¹²⁰ A copper-alloy figurine in the form of a standing boar, 50 mm long. The animal's feet are attached to a disc with a hole in the centre, perhaps allowing for its attachment. The body is oval in cross-section with limited modelling of anatomy or surface detail. The most prominent feature is the crest-like mane which runs the length of the back from ears to tail. The limited modelling of the head includes raised ears, large eyes with channels framing cavities once likely inlaid, a short snout with two recesses for nostrils and a hint of a slot mouth with short worn tusks to either side. The face is quite short compared to boar figurines, more characteristic of a domesticated pig.¹²¹ Apart from the short tail and testicles hanging between the back legs there is limited modelling of limbs and torso.

FIG. 22. Wickwar, boar figurine (No. 22). Scale 2:3 (photo: K. Adams). (© Bristol City Council)

Finds documented by the PAS have contributed 15 further examples to the corpus of boar figurines of Iron Age and Roman date.¹²² In the absence of stratigraphic evidence the precise date of individual examples is hard to establish.¹²³ These representations are almost all highly schematised, emphasising the mane which characterises the adult male wild boar, long snout and raised ears, as well as short legs, exemplified in a figurine from Great Sturton, Lincs; occasionally hatching indicates the coat, for instance on figurines from Colchester and East Herts.¹²⁴ Boars in protome form emphasise the same features, especially the height of the mane.¹²⁵ While the Wickwar boar has clear affinities with this wider group it also has unusual details. The sex of the boar is otherwise only occasionally indicated by modelling of the genitals, as in this case. With one exception, from Birch, Essex (ESS-C5CB81), it is the only boar figurine with evident signs of inlay in the eyes. It is also difficult to parallel the attachment on the base, unless this is a casting artefact.¹²⁶ The naturalism of the Lexden tumulus figurine, persuasively identified by Foster as an import, remains exceptional in this corpus.¹²⁷

DEVON

(23) Sparkwell (DEV-6D331A) (fig. 23).¹²⁸ An enamelled copper-alloy brooch of probable second-century a.d. date, largely complete save for its pin, of hybrid form. It measures 68 mm long and weighs 47.3 g. Broad wings house the casing for the hinge. Traces of the latter are visible within the casing in the form of iron corrosion product. The pin is also visible at the end of the wing. The wings carry three recesses on either side of the bow, their inlay mostly lost but likely enamel. Some traces of ropework mouldings are visible around the cavities. The triangular head-plate carries a crescent recess with enamelling in alternating colours (red and blue, the blue having mostly decayed), above which is a circular setting. On both sides there is a loop at the junction between head plate and wings and crescent projections closer to the headplate's tip. The bow has a marked L-profile, flaring trumpet-like where it meets the headplate. Ropework moulding is visible at the junction of bow and wings. The main part of the bow expands to carry tripartite plate-like decoration. In the centre of the bow is a large disc with concentrically arranged channels separated by reserved metal and set with enamelled squares in alternating colours, yellow and red in one, red and blue in another. At the centre is a poorly preserved setting seemingly with further white and blue enamel. Between this disc and the head is a smaller disc, with a channel framing a central enamelled setting and crescent projections on either side. At the foot, separated by a broad waist, is a further, larger, disc, also flanked by crescent projections and again with a channel framing a central enamelled setting. On the reverse a medial ridge expands to form a deep catchplate.

FIG. 23. Sparkwell, plate brooch (No. 23). Scale 1:1 (photo: R. Webley). (© Somerset County Council)

This brooch is hard to classify or parallel and appears experimental in form, combining elements of two brooch types. In its general form it can partly be described as a developed T-shaped brooch (in Butcher and Bayley's classification), with a T-form lent by the wide wings complemented with enamelled decoration, a type documented more commonly in south-west England.¹²⁹ However the expansion of the bow decoration is much more exuberant than is typical for brooches of this type. Instead its spirit is strongly paralleled in enamelled plate brooches of continental type, especially the tripartite type with central disc decorated with coloured squares set in a circular arrangement and smaller elements at head and foot, formed by double crescents. The focus of the 100+ examples known lies in eastern and northern France, Switzerland and the Rhineland.¹³⁰ Continental plate brooches of this and other types are very rare in the south-west peninsula, with the major exception of the votive assemblage at Nor'nour, Scilly, among which is one of these.¹³¹ Another bow brooch from Nor'nour also offers a partial parallel, with its echo of the same decorative scheme, with a central disc with an enamelled rosette and smaller studs placed at the head and foot of the bow.¹³² Wherever it was made, the Sparkwell brooch fits a wider preference for brooches with polychrome decoration in coastal zones in south-west England.¹³³