II. FINDS REPORTED UNDER THE PORTABLE ANTIQUITIES SCHEME

Citing this paper Please note that where the full-text provided on King's Research Portal is the Author Accepted Manuscript or Post-Print version this may differ from the final Published version. If citing, it is advised that you check and use the publisher's definitive version for pagination, volume/issue, and date of publication details. And where the final published version is provided on the Research Portal, if citing you are again advised to check the publisher's website for any subsequent corrections.


II. FINDS REPORTED UNDER THE PORTABLE ANTIQUITIES SCHEME
By JOHN PEARCE and SALLY WORRELL The Portable Antiquities Scheme (PAS) was established in 1997 as an initiative to record archaeological objects found by members of the general public. Following an initial pilot phase, it was extended to the whole of England and Wales from 2003. 1 Surveys of Roman period finds recorded by the PAS have been published in Britannia from 2004 onwards. This nineteenth annual report first briefly summarises the general character of Roman finds reported in 2021. Following a change of practice in 2019, we no longer present artefact numbers by county, since consistent regional differences in artefact frequencies recorded by the PAS were established in previous reports. The majority of this report therefore comprises the publication of significant individual artefact discoveries recorded by Finds Liaison Officers (FLOs). OVERVIEW In total 20,634 objects of Roman date were documented by the PAS in 2021. As a single record may sometimes document more than one item, typically in cases of coin hoards or assemblages of ceramics or building materials, the total number of objects documented exceeds the number of records (9548). The object and record numbers for 2021 are very similar to those for 2020. The totals for both years are much lower than those for previous years; in 2019, for example 29,571 objects of Roman date were documented within 18,088 records. The results for 2021 reveal the continuing substantial impact of the Covid-19 pandemic both on the discovery of objects and on the work of FLOs in documenting those discoveries. The relative frequency of object types and their spatial distribution across England and Wales closely resemble those attested in previous years. The largest numbers of finds records are again associated with counties in eastern and central southern England. 2 In Wales, north-west, north-east and south-west England, the number of artefacts documented is much smaller. The most significant factors in determining this distribution of reported artefacts remain as follows: circulation and deposition of metal objects in the past; land-use, including regional differences in historic and current agricultural practice; associated constraints on metal-detector use and its consequent uneven intensity. 3 Discoveries of coins and brooches as ever significantly outnumber those of other Roman artefacts. With the addition of the 6811 coins recorded in 2021, approximately 417,000 Roman coins have now been recorded in these annual summaries. This number reflects both the exemplify the diverse artistic traditions represented in portable metal objects from Roman Britain, illustrated here in terms of both style and motif. NORTHUMBERLAND (1) Corsenside (DUR-7BDE31) (FIG. 1). 10 A well-preserved copper-alloy harness mount with a dark greenish-brown patina. The mount is c. 71 mm long and weighs 37.8 g. It takes the form of a lyre-shaped frame whose joining arms meet two roundels. A circle is created within the frame by a link between the loops, within which is a scroll motif. This terminates in a single lobe where it touches the frame at one end and flares and bifurcates at the other, leaving a circular void at its centre. A stalk ending in a circular stud projects from this frame. The two roundels (one damaged) carry deep circumferential grooves. On the reverse of the frame are three short spikes cast with the object, one on either side of the loop, the other where the lyre curves meet. Rectangular loops at either end, one beneath the stalk, the other beneath the 'spectacle' roundels, appear to be soldered into place after separate casting.
Mounts of similar size, form and decoration have previously been documented by the PAS, including two as components of a hoard at Wood Dalling, Norfolk, as well as a further metal-detected example from the possible sanctuary site at Shillington, Bedfordshire. 11 These complement excavated examples from the Iron Age and Roman settlement at Dragonby, Lincs and the hillfort at Rainsborough, Northants. 12 The contexts and decoration suggest a date of later first century B.C.-A.D. first century for mounts of this type. The Corsenside mount is an outlier from the distribution which otherwise focuses on the East Midlands. Its openwork motif also differentiates it from other examples. The mounts' specific function has not been definitively established, in particular the purpose of the spikes or prongs which would not easily have enabled fixing to a strap. Dr Pauline Norris suggests that they were used as studded cheek-pieces to deter horses from turning their heads in motion, a possibility supported by the discovery of a pair in the Wood Dalling hoard. 13 CUMBRIA (2) Near Ulverston (LANCUM-58371E) (FIG. 2). 14 A copper-alloy amulet, well preserved with a consistent green patina, 90 mm long and weighing 64 g. It comprises two main elements, a terminal in boar's head form and a rod which projects from it. The boar's head is modelled in some detail, with ears raised on either side of a crested forehead, small eyes framed by concentric curved incisions rendering the musculature of the face, a worn snout with mouth slightly open and a tusk visible on the left side; further incisions mark bristles beneath the chin. A suspension loop with signs of wear projects beneath the bottom jaw. Raised semi-circular mouldings, incised as if they were further flaps of skin, mark the junction between the terminal and the projecting rod. The latter curves downwards from the back of the head.
This object is similar in size and form to several artefacts which combine an animal head or protome with a short curving copper-alloy rod and have been identified as amulets for suspension from draught-animal harness. Others in boar's head form include examples from Great Canfield, Esch and Trier (boar or dog) also with a loop directly beneath the jaw. 15 Among the diverse other heads or protomes are the river god Achelous, felines, a human bust and a possible goat. 16 These are smaller equivalents to larger amulets fixed on the yoke in which boar's tusks or a metal equivalent were set into the back of the boar protome terminal. 17 The significance of formal differences within this group for the object's function is not clear. For example, the suspension loop is sometimes attached to the head (Trier, Esch, this example) and sometimes to the rod behind (Great Canfield). The rod itself usually curves upwards, tusk-like, from the back of the head but curves downwards in the present instance and takes a sinuous form in the case of Esch.  (FIG. 3). 18 A complete cast silver finger-ring with intaglio setting. At its maximum diameter it measures 26.6 mm and it weighs 6.4 g. The bezel stands proud above the hoop, which narrows from broad everted shoulders. The ring has affinities with Henig Type VIII / Guiraud Type III finger rings and dates therefore to the late second or third century A.D. 19 The shoulders are slightly concave or hunched like those of a ring from Wittering (Cambs). 20 The intaglio, light red in colour and slightly transparent, is most likely a carnelian. It carries an engraved image of a winged figure standing to the right, a sceptre diagonally across her body, seemingly with a long tunic (chiton). The attributes allow her to be identified as Nemesis; it may be that this figure like other images of this goddess makes the apotropaic gesture of spitting into a fold of her chiton, but damage makes it difficult to be certain in this case. 21 The engraving is closely paralleled by a find from Braughing (Herts) and has looser affinities to other representations of Nemesis from Britain on which other attributes appear, for example the olive branch and the wheel. 22 POWYS (4) Talgarth (NMGW-E6F3B0) (FIG. 4). 23 A cast copper-alloy janiform scabbard chape end, well preserved if worn with a green-brown patina and areas of visible iron corrosion. It is 39.8 mm long and weighs 38.5g. Corroded iron within what survives of the chape frame marks the remnant of the sword tip, more clearly visible in an x-ray. Its presence indicates that sword and scabbard were originally deposited together. 24 The chape end itself is modelled in the form of two opposed human heads, closely resembling one another, narrow and elongated with high foreheads and a The record also identifies the presence of traces of mineralised wood, likely derived from the scabbard, and notes the results of XRF analysis, indicating a tin-bronze alloy with some lead, consistent with pre-Roman copper-alloy composition. thin groove for a mouth set very low above a small chin. The raised moulding for the splayed nose almost mirrors the brows to which it links via a narrow bridge. The lentoid eyes are unevenly sized with the eyeball standing quite prominently in its recess. The hair-line is formed by a raised moulding which curves around the side of the face to extend below the chin, thickening and turning back on itself where it meets the frame to define recessed trumpet-like panels. At the tip of the chape behind the heads, mouldings define triangular panels which widen where they meet on the central axis of the fitting.
In its overall form the Talgarth find shares characteristics with northern British scabbard chapes defined by their bifurcating tip, in particular Stead's type 'h' which is exemplified by swords found at Melsonby and Flasby (N. Yorks). 25 These parallels would place the Talgarth sword, geographically an outlier from this small group, in the first century A.D. However, the anthropomorphic decoration is not paralleled in this group or indeed on other chapes. 'Anthropoid' hilts are documented for some short swords but these are not janiform, typically facing in one direction only. 26 Opposed heads in profile can occasionally be matched in other Late Iron Age metalwork, for example on the Marlborough bucket. 27 Janiform representations themselves are rare in Iron Age art in Britain, although are more frequently documented in continental Europe and in a Roman context. 28 Among these other janiform representations,  embody 'two contrastive states of being', through differences of age or expression, for example youth and age, wakefulness and sleep, life and death. 29 The key handle from Rhoose, Glamorgan, with its contrasting clean-shaven and bearded heads, is a case in point. 30 However, the Talgarth heads possess no such obvious distinction. Placed at the sword's tip, they might instead reference the taking of heads battle trophies, as is suggested for other disembodied heads on martial metalwork, for example on the Tal-y-Llyn plaques (Gwynnedd). 31 Alternatively they might reference or even enhance a watchful 'omnipercipience' on the part of the sword's wielder. 32 SWANSEA (5) Gower Caves (NMGW-0322B4) (FIG. 5) 33 A copper-alloy figurine of likely late Iron Age or early Roman date, found on the surface within Culver Hole Cave, in association with human skeletal material, having probably been disturbed by recent tidal activity. The corrosion and pitting is consistent with the marine environment. 34 It is 84 mm high and weighs 63.2 g. Cast in the round and near complete, the standing figure, with legs together and large feet, has short straight arms held against its sides, perhaps truncated by corrosion. With the exception of the toes, a slight groove to separate the legs and a thickening at the buttocks, little or no anatomical detail is visible on the body. The feet are pierced by a circular perforation, presumably to allow for attachment to a base. The head is significantly outsized in relation to the body, rising from a thick neck. Broader at the forehead, the face tapers to the chin with its features crudely rendered. A short broad ridge forms the nose, with grooves forming lentoid The object record notes that non-destructive XRF analysis of the copper alloy indicates a leaded bronze with trace levels of antimony.
II. FINDS REPORTED UNDER THE PORTABLE ANTIQUITIES SCHEME eyes angled downwards on either side and a brow ridge is perhaps modelled above. A straight narrow mouth spans the width of the face. A large circular crest extends from one side of the head to the other, forming a likely headdress.
This figurine, lacking obvious gender attributes, is difficult to parallel closely among objects of Iron Age and Roman date. It shares some characteristics with other figurines from the Bristol Channel region, one excavated at the Henley Wood sanctuary near Yatton (N. Somerset), two found near the Severn river cliffs at Aust. 35 Like the Gower Cave figure, these too have schematised forms but with much more obvious female traits. The larger Aust figurine shares the crest across the head as well as a pin for attachment at the foot. It is tempting to see some affinity between these crests and the diverse head-pieces found in late Iron Age (and Roman) burials and hoards and depicted on some Iron Age anthropomorphic representations. These have been claimed as possible markers of priestly status and even as signifiers of the potential to gain esoteric knowledge through trances. 36 The place of discovery, a marine cave, is a liminal space suitable for such an experience. It is also the findspot of another Roman figurine, likely female (and perhaps carrying a child), as well as of other objects dated from the later prehistoric to early medieval periods. 37 NORTH YORKSHIRE (6) Skipton-on-Swale (SWYOR-ADC234) (FIG. 6). 38 A copper-alloy mount with smooth dark green patina, worn and perhaps broken away at its narrow ends. It is 52.7 mm long and weighs 9.1 g. The mount takes the form of a flat, rectangular sheet with a rivet at each end and a stepped edge on one long side. Within a straight-grooved frame, parallel to both long sides and one short side, it carries nine repeated motifs on its long axis, comprising a circle stamped slightly off-centre adjacent to an arrow-like double chevron. Beading runs parallel to one short and one long side, outside the grooved frame. A wider and deeper groove has been cut across one short edge, cutting across the beaded border, though the rivet is afterward driven through this groove. The decorative motifs have affinities to similar pairings of chevron and circles on late Roman belt fittings, for example in Nicolay's group E and Sommer's 'Sorte 1 Form C' (variant F). 39 The beaded edges are also shared by belt buckles of the same period. 40 Two smaller early medieval mounts documented by the PAS also have similarities of form and decoration. 41 EAST RIDING (7) Wetwang (NLM-214DFB) (FIG. 7). 42 A cast copper-alloy openwork mount, c. 49 mm in diameter, 6.8 mm thick and weighing 38.1 g. The mount or phalera takes the form of a triskele, with anti-clockwise curving legs which flare into trumpet motifs and then taper before expanding again to touch the rim. Decoration is incised on each leg, becoming more complex and expansive where they widen, with curving triangular zones delimited by hatching on the trumpet 'mouths' and on points of contact with the rim. At the centre are the corroded remains of an iron pin or rivet which held the fitting in place. This is at least the third example of a (likely) harness mount of this kind to be documented by the PAS, others being from Lewknor, Oxon, and Hanslope, Bucks, all with repeated trumpet-form or Trompetenmuster motifs. 43 (FIG. 8). 45 A small copper-alloy mount in zoomorphic form with a dark green patina with patches of corrosion, 43.5 mm long and weighing 22.2 g. It takes the form of a narrow strip with bevelled edges from which two studs project. From this extends (at an angle) a highly stylised zoomorphic neck and head. The thin neck, elliptical in cross-section, curves outwards and downwards, expanding where it meets the head. Grooves along the back of the neck render a crest. The head has three elements, a curving lower jaw and two facets for the face, wider at the ears and tapering to the tip of the muzzle, meeting at a marked angle. With the exception of a groove for the mouth and the tips of the ears, individual features are not rendered. The underside of the neck is quite worn.
The form of the object suggests that it may be a vessel mount, the wear perhaps indicating its role as a handle escutcheon. In detail, however, it is hard to find specific parallels, either for the tapering strip by which it was attached or to the zoomorphic terminal. Similar mounts have been documented by the PAS at Ipplepen, Devon (DEV-EA5E82) and at Berners Roding, Essex (ESS-CC7432). The former is more bird-like in form and detailing, but the latter is more boar-like in the modelling of the head. The 'dished' face of the Hampole mount recalls the profiles of boars as Late Iron Age figurines. 46 A date in the first centuries B.C. or A.D. is likely. (9) Burghwallis (2020T1014; SWYOR-756881) (FIG. 9.) 47 A well-preserved silver 'proto-hand' pin, with surface patinated black in places. The pin is 123.9 mm long and weighs 18.2 g. The pin head comprises two elements, a plain flat semi-circular plate or 'palm' and a row of three linked short cylinders or 'fingers' with one flat end and the other slightly rounded (for the two outer fingers). These 'fingers' project at right angles to the plate to which they connect at both ends. The head is parallel to the circular shank. The latter turns at right angles near its junction with the plate; at the other end it tapers to a point. The pin closely resembles other examples documented by the PAS in copper alloy and silver, some of which also carry curvilinear decoration and are inlaid with enamel and niello on the plate. 48 The Burghwallis and related finds illuminate the changing form of these large pins, used as likely cloak fasteners in the late Roman and early medieval periods. In terms of typological development they would appear to provide more examples of the 'intermediate' forms identified by Youngs which lie between the 'proto hand-pins' of the late Roman period and the much larger hand-pins with tubular 'fingers' of the seventh century A.D. 49 The pinheads in the precursor forms have a thinner plate joined to beaded decoration, for example as documented at Oldcroft (Glos) in a coin hoard (tpq of A.D. 359), in association with fourth-to fifth-century buildings at Caves Inn, Warws (albeit unstratified), as well as a metal-detected find at Welton-le-Wold. 50 The Burghwallis and similar pins have a thicker head with a set of 'fingers', albeit often fewer in number and shorter than the early medieval examples. In two cases from the West Country, also documented by the PAS, they carry scrolling decoration and inlay. Their dating is likely to lie in the period between the late fourth and sixth centuries A. II. FINDS REPORTED UNDER THE PORTABLE ANTIQUITIES SCHEME shoulders, and very worn, tapering to a rounded snout with little detail visible. The principal head is less worn, broader with a stubbier snout, mouth open and large eyes, better preserved on the right with oval rim and iris defined. The dog's rough coat is distinctively modelled by sets of parallel curving grooves visible across almost the entirety of its body, a shaggy texture occasionally paralleled among canine statuettes of Roman date. 53 Small-scale bronzes of Cerberus, the mythical guardian of the underworld, are occasionally attested. 54 This is the first documented example from Britain and its form suggests that it does not derive from the same source as other small bronze representations of the monster. The statuette might once have formed part of a group, although the dog's pose does not offer obvious pairings -Cerberus is usually seated rather than standing when paired with Serapis while Hercules is more commonly shown dragging a reluctant triple-headed hound. 55 Nonetheless the cult of Hercules provides the most plausible context for such a representation. Figurines of the god are a common find in southern Britain, with some examples also from Fenland sites to the south-east of the findspot, including two from the Cole-Ambrose collection, likely found at Ely. 56 NORTH LINCOLNSHIRE (11) Barnetby-le Wold (LIN-B96982) (FIG. 11). 57 A complete cast copper-alloy vessel escutcheon in figural form, with a dark green patina and some corrosion damage. The mount is 57.8 mm long (from the tip of the bull's nose to the top of the human head) and weighs 61.1 g. It takes the form of a human figure riding a composite animal formed of a bull and a marine or avian creature. The rider has a flat, thin body and very short legs, with limited rendering of anatomy. The flat round face carries schematically rendered detail, two circular eyes delineated by grooves, a triangular nose and narrow mouth. The hair is swept back in thick parallel grooves which converge at the rounded back of the head. Below the short neck the torso tapers slightly to the waist. Little is modelled of the figure's anatomy, save for two small breasts, suggesting that it may be female. The right arm is raised; on the back of the 'mitten' hand shallow incisions differentiate fingers. detail on the figure comprises the evidence for dress or adornment. Mouldings at the wrists may indicate bracelets, while a belt encircles the waist. On the front of the torso lines of punched decoration run across the shoulders, from neck to armpits and down the torso. This may be intended to render a fabric texture or indicate an item worn, a body chain perhaps.
The simply modelled bull's head is wedge-shaped, tapering slightly towards the nose. The thick horns curve upwards and backwards, the left being gripped by the rider. Large flaps for ears extend from either side of the head. Other features are schematically rendered: lentoid eyes rise from the muzzle, grooves delineate the nostrils and mouth. Fur is rendered by short incisions along the ears and between the horns. A slot cut deeply into the underside of the head enables attachment to a vessel rim. Beneath the rider the creature transforms into a curving zoomorphic head and neck. From a narrow crest housing the 'eye' (a perforation) the head splays into a broad 'bill.' The head and 'bill' are incised with short parallel grooves, schematically representing the texture of fur, feather or skin. Traces of iron corrosion and a broken off rivet are visible respectively on the upper and lower surfaces of the 'bill'. This creature may be intended to be a bird, but with occasional exceptions bird beaks are narrower and pointed in copper-alloy representations of this period. 58 Alternately this is a marine creature, perhaps a dolphin, with an exaggerated beak or rostrum and its back curving beneath the rider.
The multi-purpose decorative motif. 61 Composite figures are very rare. On a small copper-alloy object from Fivehead, Somerset, a rider also grips a bull's horn, but the object otherwise differs in size and form. Some other mounts share the principle of tripartite combinations, for example from Condover (Shropshire), the River Ribble, Thealby (Lincs) and Twyford (Leics), with human mask, bovine and avian elements, but these have little specific affinity with the mount under discussion. 62 Classical motifs and/or stories may have influenced its form. The abduction of Europa, for example, showing her riding Jupiter in the guise of a bull, is a common scene in Roman mosaics, including examples from Britain. 63 The same scene features in other portable media, including cameos and terra sigillata. 64 However Europa usually rides side-saddle and her mount is rarely hybridised. The image of Eros, right hand often raised and riding a dolphin in media ranging from mosaics to coins, figurines and oil lamps, supplies an alternative source of inspiration. 65 In Britain it famously features on a well-preserved mosaic at Fishbourne, but a coin of Tincomarus also adapts the scene from a Roman Republican type. 66 Whatever the source, this mount shows a playful adaptation of the motifs to the medium. The attachment points, the slot beneath the bull's jaw and the rivet in the beak of the bird/dolphin indicate that when connected to the vessel the human figure would not have stood vertically (as the photograph shows it). Assuming strongly sloping sides for the vessel, the bull's head would have projected outwards horizontally, while the avian/marine creature would have been plunging downwards towards the liquid in the vessel's interior and the human figure would have slanted at a steep angle leaning back with the exhilaration of the ride after breaking the surface. obscured by wear but are quite finely modelled, including some strands of hair swept over large ears on the otherwise bald head, snub nose, round eyes, a slot mouth and luxuriant curling moustache and beard. An angled narrow platform carries the head and the beard's symmetrically arranged curls, beneath which the mount expands, carrying further incised linear decoration. At the back of the head and at the bottom of what survives of the mount are short projections, the probable remnant of fixing studs.
The modelling recalls images of Silenus, the balding satyr, with hints of grinning bestiality in the face. Small bronze appliques rendering Silenus and other satyrs are not uncommon as ornamental fittings on furniture, boxes, doors and tools. 68 However some of the typical features of these images, for example pointed ears, ivy or vine garlands encircling the head, or animal skin (nebris) across the shoulder, are not visible in this image. 69 Instead therefore the figure might have been intended to render a different type; a similar representation of a balding bearded male head from Augst is identified as a philosopher. 70 However, the frequency of Silenus representations among the repertoire of Bacchic images on portable objects may mean that ancient viewers were more likely to see him in this image, whatever the original intention. 71 BUCKINGHAMSHIRE (13) Buckinghamshire (DENO-2BAD49) (FIG. 13). 72 A copper-alloy harness brooch or 'horse brooch' in two pieces, a broad brooch plate and a narrow mount, both carrying curvilinear decoration with vitreous inlay, together forming a T-shape. The harness brooch is 178 mm long, 122 mm at its widest and weighs 169 g.
The flat brooch plate is tripartite in form, comprising two larger semi-circular projections symmetrically arranged around an opposed smaller semi-circle. It carries complex curvilinear decoration, symmetrically arranged around the short axis of the plate, with contrasting zones of reserved metal and panels recessed to contain red vitreous inlay of which traces remain. On the central smaller semi-circle a central thick loop of reserved metal frames a drop-shaped perforation. This loop is itself set within inlaid fan-shaped panels. The sides of the loop converge, extend and then bifurcate to form two 'eyebrow'-like projections from the larger semi-circles. On either side of this central axis the dominant motif is a kidney-shaped zone of reserved metal. Each 'kidney' is framed by smaller inlaid recesses of triangular, crescentic and trumpet form, connected by narrow lines near the plate's outer edge. Within each of these smaller recesses are further motifs in reserved metal, expanding into circles and commas from trumpets which taper from the plate's edges. On each side a circular perforation lies at the junction between the larger and smaller semi-circles. The plate reverse carries rectangular projections respectively to house the hinge for the pin (still containing the corroded remnant of the iron pivot) and a substantial catch-plate for its tip. Hinges

II. FINDS REPORTED UNDER THE PORTABLE ANTIQUITIES SCHEME
and catch-plate were likely attached later. From the reverse of the smaller semi-circle projects a lug which fits into a perforation into the narrow mount, so as to link the two components of the brooch.
This mount is spindle-shaped with a narrow curving terminal at one end and a broader curved terminal at the other, the form and symmetrical decoration of the latter echoing that of the brooch plate. The mount is scored with incisions that define its edges as far the 'step' where it joins the broader terminal. Incisions also frame two circular cells on the narrow terminal and converge from here and from the opposite corners to frame a circle of reserved metal in the centre of the mount, itself with two circular recesses to house further inlay. The mount is cracked adjacent to this central motif. On the broader terminal in reserved metal a central loop or lyre motif spawns scrolls which curl back on themselves to form circles, from the edges of which project opposed lobes. Within the circles this divides zones of recessed inlay into circles and a pelta, outside it creates other motifs, trumpets, fans and a central circle. On the reverse of the broader terminal is a rectangular strap loop, also seemingly attached later.
The form of the brooch is paralleled in a similar harness fitting in the Polden Hill hoard, the 'yoke piece' as Fox describes it, which is one of two elements serving to fix a horse blanket in place according to his reconstruction. 73  have likely passed through the strap loop on the mount while the pin on the plate fixed the brooch to the cloth. Towards the tail a corresponding brooch or 'rump piece', of which there were two examples in the Polden Hill hoard, provided an additional fixing point. On the Buckinghamshire mount the strap loop was perhaps wide enough for the crupper to pass through it, unlike the narrow loop on the Polden Hill example which required a more complex arrangement in Fox's reconstruction. The two-piece character (i.e. mount and plate) gave the 'yoke piece' brooch flexibility while the horse was in motion, enabling movement in the horizontal plane in the case of the Buckinghamshire brooch and in the vertical plane through the hinging of the Polden Hill example. The inhibited vertical movement between the elements in the Buckinghamshire case may account for the crack across the mount. 74 The contrast of glossy red inlay and bronze reserved metal on the Buckinghamshire harness brooch has clear affinities with decorated metalwork of the first centuries B.C. and A.D., especially with the Polden Hill horse gear dated to the mid-first century A.D. 75 Further recent hoard finds support a date during the Late Iron Age to conquest periods, including another enamelled example with curvilinear decoration. 76 The juxtaposition of red and bronze also made for a strong decorative contrast with the fabric and leather once connected by the brooch. Close up, the harness brooch engaged its viewers through echoes and contrasts in form and decoration across its two components. The decoration on the mount's broader terminal resembles that on the brooch plate in its layout but gives much greater space to inlaid areas; by contrast expanses of reserved metal on the brooch plate are islanded by red inlay frames. On the latter these reserved metal zones in turn isolate the circles of light admitted by the perforations at their corners. Abstract motifs potentially convert to anthropomorphic forms when the brooch is turned 'upside-down' in relation to the way it is printed here. With its well-preserved decoration, this is a significant addition to the small corpus of related objects of this kind documented across southern England. 77 LONDON (14) Dowgate (LON-9B6D66) (FIG. 14). 78 A complete red jasper intaglio carved with a bust, 13.3 mm long, 11.2 mm wide, 2.9 mm thick and weighing 0.6 g. The stone is cut as an oval and is II. FINDS REPORTED UNDER THE PORTABLE ANTIQUITIES SCHEME bevelled, the engraved face being wider than the reverse. The left-facing bust shows a young male whose hair falls in thick clumps to the nape of the neck, likely to be the god Apollo. 79 Martin Henig writes: 'This lovely red jasper can be dated from its style to around the very end of the second century AD. It is a full, nude bust of a youth with his hair bound with a fillet and long locks down the nape of his neck. These locks suggest that this is to be interpreted as a bust of Apollo, despite the lack of an attribute such as a laurel, for example as on an earlier full bust version from Xanten. 80 A similar example is documented in the British Museum collection, but with no laurel. 81 The corkscrew locks are paralleled on another image from the British Museum collection and on an onyx intaglio from Carnuntum depicting a shallow bust of the god. 82 The style, certainly of the physiognomy, can be compared to a red jasper found at Birdoswald which shows a bust of a youth, though the hair of this example is in a different style and may actually represent the young Geta.' 83 HAMPSHIRE (15) Winslade (HAMP-47B5C) (FIG. 15). 84 A cast copper-alloy lock component in the form of a duck, 48 mm high, weighing 21.6 g. The duck's body is highly stylised. Its head has large round eyes inlaid with enamel and a tapering beak. From the high neck the body widens, with a low back, and then rises and narrows again at the tail. Wings and tail are demarcated by ridges along the sides and back of the body. The wing plumage is rendered by incised lines. The body is separated from the flat base by a small groove around its circumference. Beneath this base projects a flat rectangular shank with opposed cutaways near its tip. in connection with other elements of the lock plate. 86 The strong stylisation of the Winslade duck is paralleled in finds from the Magdalensberg and Mont Beuvray. 87 The Winslade bird provides a further example of Iron Age and Roman water fowl images documented by the PAS, its stylisation perhaps making a date in the earlier part of the first century A.D. more likely. 88 (16) Houghton (SUSS-BF337C) (FIG. 16). 89 An incomplete copper-alloy openwork plate with a mid-green patina. The fitting is 66.9 mm long, 50.4 mm wide including figures, 4.8 mm thick and weighs 37.8 g. The rectangular plate carries openwork decoration, organised with rough symmetry around a central 'strut' across the short axis. On either side of this strut are three longer parallel voids with multiple lobes. Along both long edges are smaller sub-rectangular and sub-circular apertures. The original openwork configuration is slightly obscured by the loss of some connecting metal. Along the edges and in the corners are holes which are likely to have housed rivets for fixing the mount in place Along one side there are three smaller circular holes, with slight traces of corrosion on the back of the fitting. On the other long edge the holes are less regular in form. Two contain the corroded in situ remains of rivets. The angled edges on the shorter sides of the frame are incised with cross-hatching, while the angled edge on one long side carries some incised diagonal grooves; more may be obscured by the corrosion in the adjacent rivet holes. The other long edge carries three small schematically rendered animals seen in profile, perhaps horses. The reverse is plain. This mount has some general affinities to the motto openwork mounts known from baldrics and belts, for example those inscribed with invocations to Jupiter or more general utere felix exhortation, although in this case no obvious lettering can be detected and it lacks the hinges found on these baldric plates, while it is too wide to be part of a belt fitting of similar mid-imperial date. 90 It is possible instead that it was a furniture mount: the plates decorating a  (FIG. 17). 92 A charming copper-alloy figurine of a seated baby boy, well preserved (save for its missing right hand) if worn with a dark patina, 47.1 mm high and weighing 40.9 g. The boy's short arms are outstretched, the left holding a ball or a similar object. The short legs are bent as if the figure is seated, with upper legs slightly splayed but feet closer together. A hole in the rounded left buttock may have served to enable attachment to a base. The body, outsized in relation to the limbs, carries little detail save for a small penis. The sloping shoulders connect directly to the head. The facial features are schematically modelled, with large eyes within triangular orbits, a broad and very worn nose, short fat lips and rounded cheeks and a cap of hair with parallel incised grooves rendering its individual strands.
Despite the schematisation, this is an expressive model of a seated boy of young age, of a baby able to sit but not yet walk (though the seated pose, with bent legs implying a chair, is not appropriate for the age). It finds a number of parallels both from Britain and neighbouring provinces of children of a similar age, seated, sometimes with objects or animals. 93 Although the portrait has elements of local stylisation, above all in the hair, it therefore sits within a wider genre tradition of representing younger children, often reproducing on a small scale images known in larger media. 94 There are examples from Britain, for instance of the well-known scene of the boy (or Cupid) playing or struggling with a goose. One is documented as a metal-detected find; another, not previously identified as such, was deposited in the Lexden grave. 95 The holding of a ball is also a sporadically attested gesture of children in Roman commemorative art. 96 The seated pose also echoes that of divinities, for example of Cupid and Bacchus. 97 SUFFOLK (18) South-west Suffolk (SF-61DC76) (FIG. 18). 98 Four fragments of the same enamelled copper-alloy object, a possible chain-linked belt or chain, one comprising a single link, the others multiple joined links. The fragments were discovered in close proximity to one another and clearly comprise elements of a single, larger assemblage. 99 They are in good condition, with a brown patina, although the enamelled decoration is poorly preserved. The majority of the links comprise rectangular or lobed lozenge plates, c. 32-36 mm long, 14-19 mm wide, varying from 2.2 to 3.3 mm in thickness, and with projecting loops in the same plane as the plate or set at right angles to it. One link is annular (38.3 mm by 22 mm), one is a much narrower rectangle (39 mm x 7.5 mm). Among the group comprising four links, one takes the form of a plain box (10 mm deep excluding the handle fitting), with decorated lid, and another is a thin rectangular link (0.7 mm thick) without enamelled decoration and with likely remains of solder on one face. The links are mainly connected by the loops, though the box lid also incorporates a swivelling lug which enables the attachment of a ring handle. With one exception each link carries on one surface champlevé panels for enamelled decoration: the surviving colours include red, blue, white and a yellow or yellowish-green, typically in single colour panels, sometimes subdivided by thinner stripes of a different colour. Occasionally the arrangement is more complex, on the narrow rectangular link and (probably) on the annular link.
In the absence of direct parallels the function of this object is uncertain. The repeating elements recall the enamelled mounts in similar colours on some horse-harness and military belts. 100 However, the latter differ in their attachment to leather strapping, evidence for which is largely absent here; in this case only one part (the upper face of the plain link) shows evidence for a  Identified and recorded by Anna Booth and Jude Plouviez. We thank Jude Plouviez for making her research on the assemblage available to us. Details of individual fragments and further discussion are available in the PAS database entry. 99 The four fragments of this object were found in March 2020 in close proximity, with a maximum gap of about 18 m between them, by two detectorists carrying out a systematic survey on a Roman site with evidence for a villa-type building. 100  likely connection to another material. Enamelled fittings on harness and military belts also differ in terms of decoration, often carrying chequerboard design of greater complexity than the champlevé enamel in this case. The more complex configuration of coloured decoration on the narrow link and probably also on the annular link, seemingly of several colours in a single cell, echoes some brooch decoration of mid-imperial date. 101 The original size or shape of the complete object is not clear; the four fragments could be restored as a continuous length (over c. 300 mm), although other reconstructions are possible, for example joining at an angle to the ring handle on the box. As a continuous length it might be part of a belt with affinities to the metal-chain belts of second-and first-century B.C. date documented in continental Europe, 130-145 cm long when surviving complete. An enamelled belt, originally 120 mm in length, of possible first-century B.C. date from a hillfort at Thuin, Hainaut, Belgium, offers one of the closest comparanda among these. 102 Other possible comparanda include hinged sections from a likely belt from Brettenham, Norfolk and four fragments from a deposit of priestly 'regalia' near Bury St Edmunds, Suffolk. A single link from Gedgrave, Suffolk has the closest direct similarity to the elements documented above. 103 Perhaps, then, this object is part of a belt or chain suspended from the body or head, a mid-imperial date suggested from the style of the enamelling.
WEST SUSSEX (19) Coombes (SUSS-EBC181) (FIG. 19). 104 A complete copper-alloy cosmetic set comprising a nail cleaner, tweezers and 'toilet spoon', joined by a wire loop, with dark green patina and weighing 9 g. The nail cleaner, 46.1 mm long, has a suspension loop ending in a narrow neck and sloping shoulders. It tapers to a 'step' c. two-thirds along its shaft, marked by grooves, and terminates in a bifid tip. The plain tweezers, 47.4 mm long, are formed from a single flat strip of metal, folded in half so as to create the two arms, hooped to enable attachment to the suspension loop and with wider tips which curve inwards to meet. The toilet spoon, 52.1 mm long, has a suspension loop like that of the nail cleaner, a similar narrow neck and sloping shoulders, grooved decoration on neck and shaft and a shallow bowl at the end.
Individual components of cosmetic sets are not uncommon finds, both in excavation and as documented through the PAS, though for the latter only nail cleaners and ear scoops can be confidently dated to the later Iron Age and Roman periods. 105 More than 600 nail cleaners, for example, have been documented by the PAS. 106 However this is one of a very small number of complete or near-complete sets recorded by the PAS, among which are examples on plain wire loops, rivetted together, or suspended from an enamelled shackle and bar. 107 The purpose of individual instruments, in particular the nail cleaner and ear scoop, continues to be debated; their names are a matter of archaeological convention rather than a definitively established functional label. They likely belong to the repertoire of specialised tools for modifying bodily appearance of late Iron Age and Roman date in Britain. 108 A medical use has long been argued for the tweezers and ear scoops, for cleaning wounds and removing foreign material from them, for example, and has been more recently advocated for 'nail-cleaners'. 109 The  (FIG. 20). 111 A complete knife, the iron blade being quite well preserved despite some pitting, the handle with a dark green patina. The knife is 230 mm long and weighs 103.9 g. It has a slim plain cast copper-alloy handle (81.4 mm long), rectangular in cross-section, with an openwork suspension loop with scrolling terminals. From the suspension loop the handle curves and narrows towards the marked step at the junction with the blade. Its good preservation conceals any sign of the tang within. The blade (148.2 mm long) has a blunt straight edge and a curving sharp edge which narrows to the point.
Fixed-blade knives of Roman date are less numerous as PAS finds than the distinctive folding-handled knives in figural form, the absence of a typologically specific handle often making it difficult confidently to attribute date. Several handles which share the same distinctive suspension loop have nonetheless been recorded. One, a well-preserved handle from Thurnham, Kent, carries a distinctive scale pattern, shared with first-to early second-century A.D. knives from London and elsewhere. 112 Other finds have hollow slots in the handle which once likely contained an insert in organic material (wood, bone), also with affinities to knives from London of similar date. 113 These are all, however, smaller than the Lenham knife and can be grouped with knives identified as sharpeners for quill or reed pens. 114 The elegance of Lenham's curving handle recalls that of knives from Mashbury, Essex (ESS-04F9E1) or Ladderfel, Gwynedd (LVPL-D420F6), but these too are smaller. It is not easy to parallel the knife among the types documented from Roman Britain. While the suspension loop gives the handle some affinity with knives of Manning's type 1, handle and blade taper more than those of knives in this group and the blade is longer. 115 In terms of overall form and function, the Lenham knife has closer affinities with larger knives used in processes requiring greater application of force and with larger guards protecting against slippage. 116 WILTSHIRE incisions, arranged chevron-like, render individual strands of both hair and beard in schematic form. Much of the detail of the lentoid eyes has been lost within the hollow formed by the long straight nose and brow ridge. The moustache frames a narrow mouth. The body slopes and widens from the shoulders but no features are visible beneath the cloak, rendered in stylised form as oblique folds which meet at the front of the body. 118 At the front the feet are just visible on a flat rectangular base. The back is flat, save for a seat with short legs which projects from low on the figure, indicating a seated pose. The base is slightly concave and carries faint traces of a white substance, perhaps solder. It remains uncertain if this was intended to be free-standing piece or as a component of another object.
The figure is hard to parallel directly. The bearded features, cloak and seated pose recall in a generic way the representations of philosophers (i.e. philosophers proper, poets, other thinkers, teachers etc.) and portraits modelled on them, more particularly perhaps the fuller heads of hair and groomed beards of Epicurean images. 119 Such an identification is not unthinkable in Britain and the wider north-west provinces, noting, for example, the mosaic representations of philosophers in busts (Cologne) or as seated figures (Autun; Brading). 120 Seated philosophers also occasionally survive as statuettes; a silver figure from a third-century house context in Augst, for example, shows a seated and cloaked figure with a papyrus scroll. 121 The bearded male figures which form the handles of some folding knives, again swathed in a cloak and with legs crossed, have also been identified as philosophers. 122 Similarly portraits on intaglios and gems are also occasionally documented. 123 However these comparisons should not be pressed since other representations large and small differ from the Marlborough find. Allowing for stylisation, any resemblance to philosophers' images is at best generic rather than specific to individuals. Philosophers' cloaks are also typically worn over a naked torso or a tunic, their folds are not arranged symmetrically or in as all-encompassing a way and such figures typically gesture more expressively than the Wiltshire figurine. Perhaps instead the representation echoes that of bearded male divinities, but this figure lacks any attribute or gesture that allows confident linkage to a specific deity. The cloaked deities known as the genii cucullati are clean-shaven and hooded. 124 The hair and beard recall the stylised images in bronze of Antonine emperors, and while this might hint at a date for the figure, its dress does not otherwise align it with imperial representations. 125 CORNWALL (22) St Ives (CORN-32866D) (FIG. 22). 126 An incomplete copper-alloy anthropomorphic mount in the form of a hollow female bust, heavily worn with a dark green patina, 45.7 mm long and weighing 52.9 g. The heavy wear impedes the description of individual features of face and hair. The eyes are rendered by two long downward angled grooves, with shallower grooves above suggesting eyebrows (or brow ridge). They frame a broad triangular nose, below which is a short thin slot for a mouth. Little is visible of the ears. A row of curls frames the top of the head while shallow grooves render the individual strands pulled into the bun at the back of the head, this being the only well-preserved part of the coiffure. The bust is schematically rendered. The back of the bust and the base have circular perforations. This is a furniture mount, placed at the top of the legs (three or four in number) of a folding tripod, a widely documented portable Roman furnishing. The perforation at the base allowed for connection to the leg. From the back of the mount an L-shaped hook in either copper-alloy or iron would once have projected to enable the suspension of a shallow basin or tray. 127 Like some similar mounts recently documented by the PAS, the St Ives piece is small compared to many surviving examples which are more typically 7-8 cm or more in height. 128 Several other mounts of this type have been documented in Britain, including a significant number through the PAS, although this is a distant westerly outlier. 129 Some are specifically identifiable as the god Bacchus or his followers. 130 Others are youths, including possible representations of Antinous. 131 The British examples resemble in general those from (and identified as likely made in) Gaul and Germany documented by Klatt and are likely to be produced in the same period, namely the second-third centuries A.D. 132 However, like the St Just mount, several others from Britain have characteristics which are not paralleled in this wider corpus, including the stylised rendering of facial features, especially of eyes and hair, and the incised and stamped surface decoration for rendering clothing. 133