HOARD

Unique ID: NMGW-9C0216

Object type certainty: Certain
Workflow status: Published Find validated and published by finds advisers

On 19th December 2007, Mr. Craig Mills discovered a hoard of two bronze bowls and a bronze wine strainer while metal detecting at Langstone, Newport. A complete wooden and bronze tankard was also discovered by the finder at the same time and within a few metres (estimated at 25 feet by the finder, but established through follow up survey and investigation as 12.8m) of the hoard. The discoveries were made in a field under pasture, which had not been ploughed within living memory, in a low-lying area prone to water-logging.

The finder promptly reported his discoveries to the National Roman Legion Museum Caerleon on 22nd December, leaving the artefacts there for safe-keeping and identification. Being very close to Christmas, there was no experienced curatorial presence on site, therefore the significance of these artefacts was not appreciated until early in the New Year. On 2nd January, one of the authors (Mark Lewis) immediately recognised their antiquity, their status as possible treasure and the fragility of the wooden tankard. The artefacts were transported to the National Museum Cardiff for the purpose of identification and reporting on 3rd January, while careful monitoring and assessment of the tankard was then initiated (Mark Lewis & Mary Davis).

The coroner was informed of the discovery by the finder on 4th January 2008 and a completed Treasure Receipt form was sent on 11th January 2008. At this time, Mark Lewis drafted a detailed description of the discovery, with the help of the finder and an accurate identification of the artefacts (Lewis, unpublished report). This information has been incorporated into the current report.

Given the fragility of the tankard and the emerging understanding that it did not constitute treasure, it was deemed desirable to acquire this organic artefact through private purchase. This would enable conservation assessment and treatment to commence without delay, therefore minimising deterioration of the object. The request was placed in a letter to the coroner on 4th February 2008 and permission to proceed with this course of action was received on 8th February.With the agreement of the finder and landowners, the tankard was acquired through private sale by Amgueddfa Cymru - National Museum Wales in May 2008 (Museum Accession Number 2008.15H).

The hoard & the tankard

1.The large bowl dimensions: External rim diameter 197mm; internal rim diameter 174mm; max. height as currently sitting 99.5mm; max. girth diameter 215-220mm; top rim width 10.2mm; max. rim width 14.0mm; internal rim height 25.8-26.3mm; external ring diameter 44.2-44.7mm. max. thickness ring 4.7-5.0mm; gap between terminals 2.0mm; height escutcheon c. 37.0 mm; max. width escutcheon 24.5mm; diameter rivet on interior surface 4.7mm Weight: 356.5gThis virtually complete copper alloy bowl has a shallow squat form with a rounded base. It has a low girth, upright neck and everted rim. The vessel has an internal rim diameter of 174mm and a depth of circa 99.5mm.The body and base of the bowl appears noticeably thinner than the rim. The everted rim is flat topped and turned down at the outer edge (max. depth 7.3mm). The rim top is decorated with four pairs of close set incised grooves circling the rim - two on the flat top surface and two on the rolled down outer margin. Between each pair of grooves a wavy line in low relief has been created using a punch, to create a stitch-work decoration effect. On the interior neck surface, two borders of incised grooves have been created, one at the top and the other at the base of the pronounced vertical neck. The upper border comprises four grooves, while the narrow band between the top and second groove has been decorated with dot-work from a fine punch, the dots spaced 0.6-1.0mm apart. The lower border comprises two grooves with no decoration between them.

At the centre of the base is a small perforation c. 1mm in diameter, created during the making of the bowl. This would have been sealed and plugged for use, though the plug has since decayed. On the interior bowl base surface, seven concentric circular grooves are visible with the aid of an X-ray plate. These are placed in three sets of two with a single inner ring of 29mm in diameter, with the perforation at their centre. The largest is approximately 120mm in diameter, and their spacing is irregular. These suggest the bowl was finished on a lathe. The bowl shows signs of damage caused by differential erosion and impact. A number of small void gashes, with irregular tear edges, are present around the girth of the bowl. There are also four dents in the vessel, three on the girth and lower body and one on the base. Two of these have areas of fresh metal exposed, suggesting slight damage during retrieval of the bowl by the finder. Further small holes are apparent around the girth and on the base. The vessel has a brown patination, with areas and patches of green surface corrosion.

The bowl has a cast copper alloy escutcheon, which has a perforation through its underside to accommodate a bronze ring. This provided the means to hang and store the bowl, rather than to act as an effective handle. The escutcheon is attached to the bowl by a circular shank perforating the lower upright rim. The end sits relatively flat with the internal surface of the vertical neck, projecting only 0.2mm and with no securing washer. The escutcheon was shaped very carefully so as to fit the curves of the bowl. The top lobe presses against the exterior edge of the rim, the reverse contours sitting snugly against the angle between the lower neck and the globular girth. The escutcheon is a complex decorative piece with a globular top lobe and a lower expanded lobe elaborated with two lateral wings. A central prominent horizontal moulding bisects the piece. The top lobe is decorated with a bi-concave sided recessed triangle with a flat base and forming an apex near the rim. This is in-filled and decorated with a red 'sealing wax' glass of Iron Age tradition, the zone bordered and defined with deep incised line-work parallel with the edges. The basal lobe has a narrower triangle recess, with three concave edges, and coming to the apex at the base. This is also filled with red glass. This is bordered and defined with incised line-work parallel with the edges. The lateral wings and central moulding are likewise decorated and bordered by incised line-work following the edges and contours. The ring is sub-rectangular in section, though sitting as a lozenge with angles at the centre of the top and rear surfaces and to either side. It is made of a rod, shaped as a near circle, with a gap of 2.0mm between the terminals, which are currently concealed beneath the escutcheon.

2. The small bowl dimensions: External rim diameter 181.0mm; internal rim diameter 164.0mm; max. height as currently sitting 86.5mm; max. girth diameter 200-205mm; top rim width 8.0mm; max. rim width 9.4mm; internal rim height 19.8-20.1mm; external ring diameter 20.5-23.9mm; max. thickness ring 2.0mm; gap between terminals 4.0mm; length escutcheon 32.7mm; max. width escutcheon 20.5mm; diameter of escutcheon perforation c.4.0-5.0mm; external rivet casing (internal surface) 11.1mm; diameter rivet head (internal) 5.7-6.0mm Weight: 243.1g

This near complete copper alloy bowl has a shallow squat form with a rounded base. It has a low girth, upright neck and an everted rim. The vessel has an internal rim diameter of 164mm and a depth of circa 86mm. The body and base of the bowl appears notably thinner than the rim. The everted rim is flat-topped and turned down slightly at the outer edge (max. depth 2.5-3mm). The rim top is decorated with four incised grooves circling the rim, three on the flat top surface and one on the rolled down outer margin. Between the inner edge and the first groove and between the second and third grooves (working from interior to exterior) are incised lines, set at right angles to the grooves and spaced 0.5-0.8mm apart, giving a decorated ladder effect around the rim. In the interior neck surface are two borders of incised grooves, one at the top and the other at the base of the pronounced vertical neck. The execution is poor, leading to three grooves emerging in places and some divergence and convergence with the top and bottom edges of the neck.

At the centre of the base is a small perforation c. 1mm in diameter, once sealed and plugged. On the upper external surface of the body are feint concentric striations around the bowl. Together, these suggest the bowl was finished on a lathe. The bowl shows sign of damage caused by metal erosion while in the ground and impact. A long void gash, with in an out-turned irregular torn edges extends along 90 degrees of the girth circumference. A pronounced dent at its margin has caused a tear at the interface between the base of the upright rim and the body. Further small holes and dints are apparent elsewhere around the girth and in one or two places on the base. The vessel has a brown patination, with patches of green surface corrosion. Original bronze coloured areas are revealed particularly in places on the rim.

The bowl has a cast bronze escutcheon, which has a perforation through its underside to accommodate a small bronze ring. This provided a means to hang and store the bowl, rather than to act as an effective handle. The escutcheon has been displaced sideways from its original position. It is attached to the bowl by a circular shank perforating the lower upright rim and secured by a deep circular bronze washer (5-6mm thick). The escutcheon was shaped very carefully so as to fit the curves of the bowl and the edge of the rim. The top lobe originally pressed against the rim, a slight localised depression in the outer rim margin and a radial crack in the rim top indicating its original positioning. The escutcheon is a complex piece, comprising upper and lower curved lobes, with decorative and moulded lipping. The upper lobe is wider and more pronounced. The central section widens and has a convex arched form to accommodate the rear perforation for the ring-handle. A bi-concave 'toggle' motif is defined by projecting lips from the lobes and lateral inscribed margins. The small and insubstantial ring is made of a circular sectioned rod. It is currently oval in shape, the terminals having been pulled apart a distance of 2mm and the original circular shape lost during use or prior to burial.

3. The wine strainer dimensions: External rim diameter 178-179mm; internal rim diameter 130-131mm; width brim 24.5-25.0mm; width inner raised brim 10-11mm; height upturned brim edge 1.8-2.0mm; max. external diameter of body c. 140mm; height c. 113mm; diameter of perforated design in plan c. 95mm; diameter ring 29.2-30.8mm; thickness 1.5-1.6mm; separation between terminals 2.3mm; loop length 9.9mm; max. loop width 9.9mm; thickness 1.5-1.6mm; diameter of rivet on top surface 4.2mm. Weight 223.3g

This copper alloy strainer is made up of a round bottomed bowl or body with a wide circular flange. The strainer looks to have been finished on a lathe, although no clear concentric striations are visible on surfaces. The body has a diameter of 130-131mm at the rim, while the flange is 23.5-25mm wide, making the external diameter 178-179mm. The exterior margin of the flange is upturned and 2.0-2.5mm high. The interior band of 10-10.5mm is raised with convex curving form rolling downwards into the body, which then widens outwards to undercut the internal rim top. This has the effect of giving a rolled inner 'ledge' to the flange and defining the inner rim.

The base of the bowl is decorated with a perforation pattern occupying a circle approximately 100mm in diameter. The outer circles of perforations provide a border to the central circular motif. A triskele design with circular flourishes at the end of each limb is defined by the absence of perforations. At the centre of the triskele is a central perforation surrounded by a single circle of perforations c. 10-11mm in diameter. The surrounding spaces filled by perforations comprise three arched trumpet motifs, each occupying 120 degrees of the circular design. The interplay of positive and negative motifs means that the circular flourish at the end of each triskele limb defines the underside of a trumpet (facing outwards). However, the arch of the same trumpet forms the upper and middle extension of the next triskele limb, giving an endless circular flowing quality. The whole pattern is unified by a continuous outer circle of perforations, itself sitting within the two ring border. The spacing of perforations varies between 1.5-6mm from centre to centre, the perforations being 0.6-2mm in diameter. In one place, two perforations intersect.

The strainer has a simple looped bronze escutcheon attached to the underside of the flange, from which suspends a small bronze ring. One end of the loop of the escutcheon has an integral circular bronze shank which perforates the flange and sits flat with the flange upper surface. The escutcheon loop is 10.3mm wide and 6-7mm deep and has a concave exterior surface along its long axis. The ring is near circular and shaped from a rod, whose terminals are separated by a distance of 2.3mm.

The strainer is complete, though there are two radial cracks on the flange, one where the escutcheon shank perforates the flange. The outer flange edge is bent upwards at one point, through impact, and causing distortion to the interior rolled ledge. At another point on the upturned flange edge, there is a nick exposing bronze metal and probably caused during recent retrieval from the ground. There are one or two areas of scratches on the top and underside of the flange. The bowl retains perfect shape and has no damage. The strainer has a brown patination with areas of surface accretions on the upper flange and on the interior and exterior surfaces of the bowl.

4 The tankard dimensions prior to conservation treatment: height 149-150mm; external rim diameter 158-160mm; internal rim diameter 150mm; circumference at mid height 523mm; base diameter 166mm; base plate thickness 8-9mm; offset between stave base and underside of base-plate 10-12mm; stave base thickness 11.2-11.5mm; thickness of chamfered stave bases 6.5-7.0mm; handle length 77-78mm; handle plate widths 45-48mm; metal binding strip widths: top 59-60mm; bottom 59-60mm; metal lip height (external): 5mm Weight: potentially variable, therefore not taken at this stage.

This is a complete wooden stave-built tankard with sheet copper alloy fittings and a cast copper alloy handle. The tankard has a near vertical profile with a simple rim and a flat base. The tankard is slightly wider at base than at the rim and the staves thicken from 5-6mm at rim to 11.2-11.5mm at the base. The vessel has an internal rim diameter of 150mm, an external height of 149-150mm and an internal depth of c. 130mm. The body is made of six staves of yew wood, which have been carved into a curved shaped section in plan view. Near to the base, a rectangular shaped horizontal recess has been cut into each interior surface, into which the circular shaped base plate has been slotted. This is approximately 7mm thick, rectangular shaped in side profile and probably also made of yew wood (Nayling pers. comm.). Close inspection of the grain structure suggests that the base plate was, in fact, made from two different pieces of wood. The lower surface of the base plates sit approximately 10-12mm above the base of the tankard staves. Inspection of the underside of the tankard reveals that the base edges of the staves have been carefully worked. The exterior edge has been smoothed and rounded slightly, while the interior edge has been angled and smoothed, to provide a thinned and finished base. The divided base plate, made from two pieces and the presence of knots in some of the staves, suggests that the wood selected for making the tankard was not of the highest quality, despite the care taken in shaping and constructing this complex artefact.

Two wide circular bands of sheet copper alloy have been fitted around the outer circumference of the tankard. There is no visible evidence of a join, indicating these bands or hoops were made to correct diameter, then temperature expanded to fix over and around the tankard, contracting tight around the staves upon cooling. In areas visible beneath the bands, the surface of the wood is blackened, probably created by the heat of the bands when fitted. The two bands are plain, with the exception of the definition of the top and bottom edges with an inscribed line, inset 0.1mm from the edges. The bands are separated in the middle by a gap of 4-7mm, with slight slippage evident since burial and retrieval. The lower band sits 11-12mm above the base, while the upper sits approximately 10mm below the top of the wooden rim. The lower band is complete and undamaged. However, the upper band has suffered through differential erosion and subsequent bending of edges of holes, particularly to either side of the handle of the tankard. A continuous, overturned copper alloy strip forms a rim lip around the rim of the tankard. This strip projects 3-4mm down the external surface and 4.0-4.5mm down the internal surface. This strip has a single incised line defining the external surface edge and inset 0.1-0.2mm from the edge. The strip has also been damaged through differential erosion of the outer face, with a number of small fragments being bent outwards and out of shape.

Three distinctive grooves are visible around the outer circumference of the tankard body, one between the rim lip and upper hoop, the second between the two hoops and the third around the base below the second hoop. The upper two are 7-8mm wide, while the lowest is slightly wider at 11-12mm. All are 0.2-0.3mm deep, yet seem to have been cut into the staves after the copper alloy bands had been fitted. It is possible that they were cut, to remove the heat blackened wood surface and to reveal the natural colour of the wood. The cast handle comprises a plain continuous C-shaped back (in profile), with T-shaped lateral attachment plates. The back is also C-shaped in cross-section with a prominent central concavity on the exterior surface. It is possible that this concavity was once filled with an organic material such as leather. The projecting margins are c. 2mm wide and have a central groove down the mid-axes. The lateral plates are rectangular with convex upper surfaces in the horizontal plane. Each lateral vertical margin is defined with a groove. The handle is attached to the body of the tankard by two pairs (one behind each attachment plate) of bronzerivets with shanks of 1.5-2mm in diameter, perforating the copper alloy hoops and wooden wall of the tankard. These have prominent widened and slightly domed heads (8mm in diameter) projecting above the interior surface of one of the staves. The handle is attached at a slight offset from vertical.

The tankard has suffered some cracking damage as a result of the drying out process since retrieval from a waterlogged context. This shows as a series of near vertical cracks on staves, slight contraction and pulling away of staves to reveal cracks between staves and slight curving and mis-shaping of stave tops and edges. One stave has been damaged through pressure or impact near the rim, causing a small break and an in-turning of stave tops and bending of the copper alloy rim lip. The two circular base plate pieces have contracted and distorted, to reveal a wide crack. A lead plug was inserted along this line of weakness, to seal an original hole. It is possible that internal stave and base joins may have been sealed with an organic sealant.

Dating and parallels

The two bowls may be identified as of the southern British 'Rose Ash' type or form, characterised by their shallow squat form with flat rim, vertical neck and bulging girth or belly with a rounded base (Macgregor 1976, 149-50, 168-70 & Map 20; Spratling 1972, 222-4 & Cats. 383-90; Raftery 1984, 222). These often show evidence of having been finished and polished on a lathe. They contrast with a contemporary northern British form of bowls, known as the 'Lamberton Moor' type, which tend to be smaller and have deeper, rounded forms and simpler rims (Macgregor 1976, 149-50; 168-70, Map 20 & Cats. 293-8; Raftery 1984, 214-23). Examples of this form occur in north Wales, including two examples from Llanenddwyn, one from Barmouth and one from Crib Goch, Snowdon, all from Gwynedd (RCAHMW 1921, 4 &103-4, Figs. 32-4; Savory 1976, 62, Cat. 32 & Fig. 38a).

Rose Ash type bowls are recognised as being Late Iron Age and Native forms, with researchers generally ascribing a date of manufacture between 50 BC-AD 50, with the circulation and burial of bowls and bowl escutcheons continuing through into the later first century AD (Watson 1949, 56-7; Fox 1961, 194-5; Megaw 1963, 36; Stead 1967; Spratling 1972, 223-4; Macgregor 1976, 149-50; Raftery 1984, 223; Joy 2008, 165, 322-4). The most complete and closest parallels for the Langstone bowls are the examples from south west England, from Rose Ash (Devon), Youlton (Cornwall) and two examples from a rich Late Iron Age female burial at Birdlip (Gloucestershire) (Bellows 1881; Smith 1926; Fox 1961; Spratling 1972, Cats. 385 & 400, Figs. 178 & 186; Staelens 1982; Jope 2000, Pls. 244-7). The Rose Ash and Youlton bowls share exactly the same form and have very similar internal rim diameters of 195mm and 187mm respectively. They share the same rim decorative techniques: the Rose Ash bowl with wavy line relief decoration on the rim top and overturned edge, the Youlton bowl with grooves on the vertical inner neck, transverse tool marks between two pairs of grooves on the rim top and wavy line decoration in relief on the overturned rim edge. Both examples have decorative escutcheons to take hanging rings, a cast ring being present on the Rose Ash bowl but the ring missing on Youlton. Lipping and lobes are similarly evident. Both perforate the neck wall and are secured by bronze washers, as on the small Langstone bowl.

The Birdlip burial presents some of the best dating evidence for Rose Ash bowls. This female burial found in 1879 was accompanied by two bowls, a decorated bronze mirror, a silver gilt brooch, a bracelet, a bucket rim, vessel plate covers, brass rings, a bronze knife handle (with a head shaped as a female deer), and amber, marble and jet beads from a necklace (Bellows 1881; Jope 2000, Pls. 244-7). The bowls vary slightly from the Langstone ones in form, having slightly deeper and rounder profiles and shorter necks. One is larger, with an internal rim diameter of 215mm and the second is small with an internal rim diameter of 105mm. The larger has lathe-turned grooves and a square hole for an escutcheon on the neck. The smaller has a round topped rim with turned grooves and incised decoration. The brooch and the decorated mirror are most diagnostic, in dating terms. The brooch is an example of a beaked bow brooch, a form which has generally been dated to between 50 BC-50 AD (e.g. Olivier 1996, 231-2; Sealey 2006, 16). However, this highly decorated example, with lipping, punch decoration and pseudo-stitch wavy line decoration is stylistically more at home towards the end of this date range in the early to mid first century AD (Staelens 1982, 23; Joy 2008, 323-4). The mirror is one of a distinctive group of large examples from western Britain, with similar decorative motifs, which are buried late, between AD 40 -100 (Joy 2008,162-8). It is likely therefore that this mirror was made during the early first century AD and that the burial dates to the period AD 40-70. This would suggest that the Birdlip bowls were made around AD 25-60, remaining in use and circulation for some time before burial.

A further significant hoard or grave group is that found within the hillfort at Bulbury (or Belbury) in Dorset in 1881. Unfortunately, the description of the burial context is not precisely known, but the sizeable group of bronze and iron artefacts included a fragmentary decorated mirror, a tankard handle and bindings, a sword, a sheet fragment decorated with wavy line relief (like the bowls), two bronze fittings in the form of bulls and a wide array of ironwork (Cunliffe 1972). The tankard handle has been dated on stylistic grounds to between 25 BC-AD 25 (Corcoran 1952, 90), however it's perceived similarly to mirror handles, which have been re-dated in the light of new discoveries, could now extend the range into the period AD 25-50. Little of the mirror survives, but an early first century AD date is likely (Cunliffe 1972, 305). The sword is of Type IVA (Piggott 1950) or Group H (Stead 2006, Cat 235), finding closest parallel with examples from Hod Hill, Fairley Heath and Waddon Hill (Stead 2006, 75, 202, Cats. 247-51 & Fig. 107) and dating to the middle of the first century AD. Taken together, the group has been dated as buried between AD 20-50 (Cunliffe 1972, 305). Correspondingly, the bowls were also made and deposited between AD 1-50.

Similar decorated escutcheons from Rose Ash bowls have been discovered at Bagendon (Gloucestershire), Hod Hill (Dorset) and Ham Hill (Dorset). The Bagendon example, found at the site of a nucleated Late Iron Age settlement or oppidum, was found sealing a ditch dating to the period AD 35-45, and given the abandonment of the site in the early Flavian period, a date of AD 40-60 for burial is clear (Clifford 1961, 182-3, Fig. 36.4; Spratling 1972, Cat. 383). This escutcheon has multiple lipping very similar to that seen on the small bowl from Langstone. Two examples have been found at Hod Hill, an Iron Age hillfort with an early Roman fort superimposed. The first was a casual unstratified find, but the second was located in the foundation trench of a barrack block to the early fort, the time of burial securely dated to AD 43-55 (Smith 1926, 281-2, Fig. 7; Richmond 1968, 114, 117-9, Fig 57.28; Spratling 1972, Cats. 387-8 & Fig. 177). A less well dated example was found in the ditch of an Iron Age hillfort at Ham Hill, though the excavation seemingly produced artefacts of a Late Iron Age and early Roman date (St. George Gray 1924, 113, Pl. XIII, E7). Recently, a winged example with an inset of red glass decoration has been reported as a metal detector find from Llantilio Crosseny (Monmouthshire) (Lodwick 2005). Though not independently dated through stratigraphic evidence, it is closely similar to the securely dated example from Hod Hill.

Stylistically, the prevalence of beaks, ears, lips and wing mouldings on escutcheons from Rose Ash bowls echo a wider development of the mid first century AD, seen upon Native/Iron Age horse and chariot equipment. Group V-VII lipped and winged terrets and Group III two-link bits, in particular illustrate this trend (Spratling 1972, 28-30, 99-101, Cats. 35-63, 167-79). A number of diagnostic hoard groups, such as from Melsonby (North Yorkshire), Polden Hills (Somerset) and Santon (Norfolk), include examples and these were all buried between AD 43-75 (Macgregor 1962; Spratling 1972, 304-9; Brailsford 1975, 234; Fitts et al 1999, 48; Hutcheson 2004, 28). The Santon hoard also contained a lid, probably belonging to a Rose Ash type bowl (Macgregor 1976, 149,168-9; Spratling 1972, Cat. 398). The beak and wing mouldings on the silver-gilt Birdlip brooch could also be seen as a further expression of this broadly contemporary style.

Having established the broad and secure Iron Age pedigree of the Rose Ash bowls, one or two emerging elements of Romanising form and technique implicit on the Langstone, Rose Ash and Youlton bowls need to be noted. The pronounced vertical form of the interior necks on these examples appear cognisant of the repertoire of Roman pans, platters and dishes witnessed in early Roman military contexts or as early imports, even if these bowls retained round bottomed forms. In addition, the circular grooves on the interior base of the large Langstone bowl could be seen to echo the turned concentric mouldings on the bases of early Roman trulleii (e.g. Eggers 1966). This can be taken as tentative evidence to suggest the manufacture of these western bowls late within the range of 50 BC-AD 50 so far established in the published literature for these bowls. A date of manufacture around AD 25-60 can therefore be tentatively suggested, seeing their deposition slightly later circa AD 50-75. This would be consistent with the time of early campaigning in these regions by the Roman Army.

The wine strainer may be identified as one of a small group of strainers with broad flanges and no handles (Tomalin 1989). They have bowls with perforated designs, while a range of decorative techniques are used on the flanges. By virtue of their association, to date, with trulleii, dippers, oar-handled strainers and shallow bowls of Roman form, they have been dated to the late first and early second centuries AD (AD 50-150) (Tomalin 1989; Curle 1931, 306-9; Wainwright 1967, 87-8). Examples occur within vessel hoards at Coygan Cave (Carmarthenshire), Manorbier (Pembrokeshire) and Helmsdale (Sutherland), while single finds have been made at Thorpe (Surrey) and Marston Moor (Yorkshire).

The direct association of the Langstone example with two bowls of Iron Age and Native form provides important new insight into their early dating. The fitting of this example with a ring-handle, in common with the bowls, and the compatibility of the strainer with both vessels in dimensional terms, provides persuasive evidence that this example was made and used as part of a Native / Iron Age set, at around the time of campaigning of the Roman Army in Wales (between AD 47-78). The selection of a triskele motif in the perforation pattern confirms a Native, rather than Roman preference. On these grounds, a date of manufacture between AD 40-60 and subsequent burial between AD 40-75 is proposed.

Looking at the decorative styles and associations of the other strainers, it is possible to find supporting evidence for continuing Native expression, suggesting their transitional currency as vessels between Iron Age and Roman stylistic worlds. The Coygan Cave strainer, with a rectilinear running scroll motif and four roundels was associated with a Roman trulleus. However, later attached to the base of this vessel was an openwork base-plate with a triskele motif. A date between AD 60-100 would be consistent with the appearance of such a vessel in west Wales and its subsequent modification to suit Native tastes. Stylistically, both this motif and that on the Langstone strainer resemble triskeles on the shield boss and openwork discs in the Late Iron Age Tal-y-Llyn hoard (Savory 1964; 1966), which has recently been argued as dating between AD 50-75 (Davis & Gwilt 2008, 171). Similar triskeles have been found on a bronze disc from a burial on Lambay Island (Co. Dublin) and a bronze plate from Moel Hiraddug hillfort (Flintshire), both dating to a similar time (Hemp 1928; Raftery 1984, 282-4 & Fig. 139; Savory 1976, Cat. 23, Pl. II).

Two strainers were recently discovered with the Manorbier hoard (Pembrokeshire) also including a trulleus, two dippers, two flat vessels and contained within an iron rimmed cauldron (Chapman pers. comm.). The strainers have a very similar Romanising design, involving concentric circular scrolls made up of linked S-shaped motifs and bounded by circular perforated borders. However, on one of the strainers, a repeat border of in-filled geometric triangles is located on the upper flange on the inner edge. This technique is viewed on the Tal-y-Llyn shield fittings and has been observed on Native decorated horse equipment and bronze collars dating to between AD 50-100 (Davis & Gwilt 2008, 170). The Roman trulleus, like the Coygan Cave example, has an openwork decorative base-plate attached with a central knobbed rivet. The complex symmetrical design has both Native and Romanising attributes. The four openwork plates decorated with triskele motifs in the Tal-y-Llyn hoard also recall this technique (Savory 1964, 463-8 & Pl. VI). Similar knobbed rivets are found on one of the Rose Ash bowls from Bulbury and the Trawsfynydd tankard, dating to AD 20-50 and AD 50-100 respectively (Cunliffe 1972, 298 & Pl. LVII; Corcoran 1952, 97 & Pl. XII). Overall, a date of AD 60-100 can be suggested for the assembling of the Manorbier hoard (Chapman pers. comm.).

The Thorpe strainer has a triple roundel perforation motif, echoing the triplism on the triskele motifs and in addition shows the same geometric infilled triangle decoration on the inner and outer borders of its upper flange surface. This suggests a continuing Native taste and, by comparison, would tend to suggest a date of manufacture, use and burial between AD 50-100. The Helmsdale (Sutherland) and Marston Moor (Yorkshire) strainers from northern Britain have a more strongly Romanised style, with rosette perforation designs and far more heavily decorated flanges (incised and punched on the Marston Moor example and impressed work on the Helmsdale example). While a simple linear progression from Native to Roman styles cannot be assumed (significant overlap in tastes probably existed), it is possible that these strainers are slightly later in date than the others - in the range AD 80-120. This would be consistent with the later campaigning of the Roman Army in northern Britain, beginning in Scotland, with the Agricolan Campaign during the 80s AD.

The Langstone tankard is a native from of drinking vessel used for communal drinking of beer and cider and especially prevalent across western Britain (Webster 1975; Gwilt 2007, 313-4). The lack of decorative fittings and the plain cast handle would ally it with examples of Class V tankards (Corcoran 1952, 93). These include complete examples from Pentuan (Cornwall) and Shapwick (Somerset), with additional tankard handles only from Seven Sisters (Neath- Port Talbot), Segontium (Gwynedd), Newstead (Roxburghshire), Greenhill (Dorset), Bartlow (Essex), Catterick (North Yorkshire) and Okstrow (Orkney) (Corcoran 1952, 96-101; Macgregor 1976, 148-9, Cats 288, 290-1; Allason-Jones 1993, 178-9, Cat. 103, Fig 10.8). An overview of their burial contexts provides a range of dates spanning the mid first to fourth centuries AD, therefore is only of limited chronological insight for their manufacture and use.

Recently, the handle and bronze fittings of a tankard was found near Brackley (Northamptonshire). It had three copper alloy bands around the exterior surface of the vessel, similar to the two surrounding the Langstone tankard, though the handle was decorated rather than plain. Importantly, it was directly associated with Claudio-Neronian pottery dating to between AD 41-68 (Joy, unpublished manuscript). The Langstone tankard handle appeared to be unparalleled in precise form in the published tankard literature. However, an exact parallel, previously identified as 'a handle or ?fitting from a chest', has been found at the Roman fort at Loughor (Swansea) (Lloyd-Morgan 1997, 251-2, Fig. 95 & Cat. 60). Unfortunately, the dating of the burial context, a slump fill into the top of a well is confused (Marvell & Owen-John 1997, 69-70). A range between AD 115/120-260 is indicated, though most likely within the early to mid second century AD (Chapman pers comm.). How long the handle had been in use and circulation before burial is unknown, however it indicates tankards with these handle forms were probably being manufactured during the late first or second centuries AD.

It is not possible to indicate whether the tankard was a contemporary deposit with the Late Iron Age hoard, or a later Romano-British burial event. However, the drinking association of these two deposits, wine with the bowls and strainer and beer or cider with the tankard suggests this place had similar symbolic and social significances. The two were certainly placed within the same peaty bog context, though this may have been long-lived. Unfortunately, as yew is such a slow-growing tree species, radiocarbon dating of the wood is unlikely to reveal an accurate date for the making and use of the tankard, since the wood was probably formed hundreds of years prior.

Other discoveries in the vicinity of the find-spot

1-3 Are all from the same field as the hoard.

1. Terret Dimensions: external diameter (side to side) 43.1mm; external diameter (bar to top of ring) 39.7mm; internal diameter 30.1mm and 29.6mm; surviving bar length c. 9.0mm; surviving bar width 2.7mm; surviving bar thickness 1.4mm; loop diameter 5.5-7.7mm Weight: 25.5g.

This is a small cast bronze terret, once used as a harness loop on a horse-drawn vehicle, through which reins were passed. In form, it comprises an oval shaped ring with an attachment bar on one longer side. The bar, set horizontally has been heavily eroded and thinned, having broken, with a gap of 2.5mm. The margins of the bar are defined by narrow and heavily eroded collars. The terret has been forcibly bent laterally, the terret loop no longer sitting flat. Two prominent wear facets are visible, one on each side face and there is also slight wearing on each interior loop surface. The surfaces have an eroded patina, brown in colour with areas of green patching.

This may be identified as a Group I simple terret (Spratling 1972; Macdonald 2007). Around 60 examples are known in the published literature and many more have been recently reported via the Portable Antiquities Scheme across England (Macdonald 2007, 12; Worrell 2007, 377-80, Fig. 2 & Table 3). A distinctively British artefact type, they have been dated to between the Middle Iron Age and the early Romano-British periods, being found in contexts dating to between 400BC-AD150. However in southern England and Wales, many examples span the Late Iron Age and Roman Campaigning periods between 150BC-AD75 (Spratling 1972, 25-6; Macdonald 2007, 13-16). They are of an Iron Age or Native style, rather than a Roman military or Classical style.

2. Brooch Dimensions: surviving length 30.4mm; maximum width 12.3mm, surviving weight 10.4g. A cast copper-alloy trumpet brooch fragment, broken at the lug for the spring (missing) and just beneath the waist knop, so that the entire catch-plate end of the brooch is missing. The elliptical trumpet is small (12.3mm wide and 15.2mm long) and plain. The waist-knop is exaggerated, being almost as wide as the trumpet head (11.0mm diameter and 5.6mm in width), but is apparently plain.

The brooch fragment is heavily corroded and exhibits losses due to bronze disease. This brooch may be identified as one of a class of trumpet brooches of Romano-British date and belonging to the later first and early second century AD. These are a common occurrence on early Romano-British military and civilian sites in south Wales (e.g. Webster 1981, 172-5, Cats. 21-5; 1995, 83-7, Cats. 52-6; Lloyd-Morgan 1997, 237-40, Cats12, 13 & 18).

3. Coin (NMGWPA No. 2008.9) Weight: 1.652g. A corroded nummus of Constantine II, Trier mint, Providentia Caess. Dates to the early fourth century AD (c. AD 325-30).

Artefacts 4-10 are known from the close vicinity of the hoard site, but were not found in the same field.

4. Decorative mount Dimensions: height 30mm; width 23mm; weight 15.0g A small copper-alloy bust of a young, clean-shaven, male deity with long hair framing the face beneath a pileus or topknot was discovered and reported in 2006 (Lewis, unpublished artefact record). It was found by a metal-detectorist on low-lying waterlogged land sitting in the same basin as the bowl and strainer hoard. It is located approximately 250m to the south west of the hoard site and approximately 150m to the east of Ford Farm Villa. Copper alloy corrosion and loss has obscured some detail. The torso is naked and emanates from an acanthus-like arrangement of foliage. It is thought once to have been attached to a vessel or item of furniture. The depiction may represent a young Bacchus, or possibly Amour or Eros, and while its classical style and Romano-British date is assured, a precise date between AD50-410 is currently difficult to define. A possible religious association and watery place burial provides interesting parallel with the hoard and tankard, while the possible drinking association of Bacchus with a vessel could be significant, in relation to the wider connotations of this locality.

5-8 Coins

Four late Roman coins were reported from the same field and locality as the decorative mount (4 above). They were found by the same metal-detectorist at the same time as the mount.

5. House of Constantine, nummus, Trier mint, Beata Tranquillitas type, dated c. AD 321-3.

6. Constantine I, nummus, Trier mint, Sarmatia Devicta, dated c. Ad 323-4.

7. House of Constantine, nummus, Trier mint, Gloria Exercitus (2 standards), dated c. 330-5.

8. Irregular nummus, Urbs Roma/wolf and twins, dated c. 330-5+.

9. Terret
An elaborate Romano-British terret with a bar beneath the loop and surrounded by a skirt with knobbed projections was discovered in 2003 by a metal detectorist on low-lying waterlogged land some 629m to the west-north-west of the hoard find spot (Lodwick 2003a). This sits within the same peat basin as the hoard. Similar examples from Housesteads (Northumberland), Poltross Burn (Cumberland) and Melandra Castle (Derbyshire) suggest a date of manufacture during the early second century AD (Spratling 1972, Cat. 98; Macgregor 1976, 46, Cats. 85 & 99).

10. Enamelled seal box
A complete lozenge shaped Romano-British seal box was discovered in 2003 by a metal detectorist some 255m to the north of the hoard find-spot on slightly raised land at the dryland margin of the low-lying basin (Lodwick 2003b). This was decorated with a whirligig motif and recesses, filled with enamel. It has been dated to between AD 100-300.

Archaeological context and discussion

The hoard and tankard find-spot is located towards the north-eastern margin of a low-lying basin at 14-20m OD. This has a long axis from west-north-west to east-south-east of approximately 1.1km and a width of approximately 400-500m from the north-north-east to south-south-west. The basin is fed by three water-courses from the north, east and west and water currently flows via water drainage channels into the Monk's Ditch and onto the Gwent Levels to the sea. A series of springs also feed these header water courses. The alluvial deposits contained in the basin are likely to be river or lake formed (fluvial or lacustrine), although little is known of the detailed environmental history of this locality during the Pleistocene and Holocene. However, more is known of the sequences in the neighbouring Lower Nedern Valley to the east and on the Gwent Levels to the south (Nayling & Caseldine 1997; Bell et al 2000). The surrounding solid geology comprises a Triassic and Jurassic sequence of marls and lias mudstones and clays to the south and west, with older Devonian Brownstones (Old Red Sandstone) to the north and east (Welch & Trotter 1961, 46-7, 116-7; BGS 1981). The tentative evidence revealed through the sequence of deposits in the two test-pits excavated, suggest phases of water-born clays and stream channels, interspersed with periods of peat bog formation, possibly under an area of standing water.

At the centre of this basin is an ovoid shaped island of First Terrace Gravel Deposit (BGS 1981; Williams unpublished report, 7, Fig. 3). Sitting on this slightly raised island, at the centre of this boggy basin is a known Romano-British site, known as Ford Farm Villa. This is a recently scheduled ancient monument (SAM No. MM298; NPRN 310034; PRN 07836g), discovered and explored by metal-detectorists in 2001. The centre of this site is located some 350m to the west-south-west of the Langstone hoard find-spot. Unfortunately, the site was damaged prior to scheduling by JCB trenching conducted by the metal-detectorists, as a way of producing artefact finds. At this time, two mosaic floors were uncovered within two of six trenches opened, while the others all contained structural evidence of Roman activity. Many artefacts were discovered, yet few were reported for identification. The pottery and coins viewed are predominantly of 3rd or 4th century AD date, however the discovery of a trumpet brooch and samian pottery also suggest early Romano-British phases of 1st and 2nd century AD date (Williams unpublished report 19-21 & 61).

A contour and geophysical survey was commissioned by Cadw over part of the site and recently undertaken by Daryl Williams. This appears to confirm the existence of a Romanised building of some size (30 by 10m) and potentially a winged villa (Williams unpublished report, 61). However, an extensive complex of activity and buildings is suggested, potentially of different phases of use and development. Although some of this activity would appear to be late Romano-British in date (AD 200-400), it is entirely likely that a Late Iron Age to early Romano-British site, and contemporary with the hoard and tankard, preceded it. The location of this site at the centre of a bog is unusual and raises interesting questions about how it was approached. It is possible that this 'place' had particular and very specific liminal associations, possibly relating to religious devotion (a temple or native shrine?) or bathing (a bathhouse?) in addition to domestic settlement functions. Within a broader settlement context, the site sits only 4.5-5km to the east-south-east of the Roman legionary fortress of Isca at Caerleon, constructed around AD 74-5 and 8-8.5km east-north-east of the Roman civitas capital Venta Silurum at Caerwent, which developed from around AD 120. Such strategic factors could have had significant bearing upon the development of the Romano-British phases of this site.

The burial context of the Langstone hoard and tankard at the margins of an area of peaty and boggy land is worthy of further comment. The evidence suggests careful placing, either in small pits or within surface standing water, the bowls and strainer placed together from within the bog, rather than thrown from the edge. Such a watery context is closely echoed by the very similar wetland burial contexts of the Youlton and Rose Ash bowls in Cornwall and Devon (Smith 1926, 280; Fox 1961, 188-90 & 192). The account of the discovery of the Rose Ash bowl indicates that a post was seen near to the bowl, raising the possibility that it may have hung from its ring-handle on the post out in the bog. No evidence of a nearby post was found at Langstone, though it remains one possibility that the vessels and strainer could have been suspended from their ring-handles and used at this place, prior to incorporation into the bog. The Youlton bowl was found in peaty bog land near to a low platform of hard ground (Fox 1961, 192), raising questions about the possible relationship of the Langstone artefacts with the nearby Iron Age / Romano-British site on the central gravel island, some 350 metres away.

A votive burial (and possibly also a votive use) for the Langstone bowls, strainer and tankard in the bog is suggested and this accords well with strong wider evidence during later prehistory, and across northern and western Europe, for the deposition of complete and prestige metalwork within rivers, lakes and bogs (e.g. Fox 1946; Fitzpatrick 1984; Wait 1985, 15-50; Bradley 1990; Pryor 2001; Field & Parker-Pearson 2003). It is possible that the plain terret, trumpet brooch, decorative mount and coins also found within the close vicinity of the hoard, represent smaller and continuing votive deposits into the bog during the Late Iron Age and Romano-British periods. The terret and brooch also seem to offer supporting evidence for a focus of activity preceding the later Romano-British phases witnessed at Ford Farm villa.

Chronologically, it has been established that the bowls were probably made between AD 25-60 and buried between AD 50-75. Culturally, the bowls are Iron Age and Native in style and inspiration, rather than Roman. The strainer is decorated with a Native triskele perforation design and was made with a ring-handle to work as part of a set with the bowls. Though possibly made slightly later than the bowls, between AD 40-60, it was in common use and buried in direct association with the bowls between AD 50-75. The tankard could have been made and buried at the same time, or later during the Romano-British period. The Roman army continued to actively campaign against the peoples of Wales until AD 78, when occupation of the entire geography was secured. During this time, Iron Age or Native metalwork styles continued in manufacture and use beside the new Classical Roman styles.

Elsewhere, in the cases of an Iron Age toggle associated with a bell of Roman manufacture, discovered at Maescar, near Sennybridge (Treasure case 05.6 for Wales), a carefully considered precedent was established for the determination and interpretation of 'prehistoric' in Wales. H.M. Coroner for Powys accepted that artefacts of Native or Celtic Iron Age (i.e. prehistoric) style continued to be manufactured and used between AD 47-78. This was the period of campaigning of the Roman army in Wales with Native Iron Age peoples, prior to the consolidated occupation of the whole of Wales in AD 78. Two or more associated base metal artefacts of native or Celtic style, made and buried between AD 47-78 would therefore constitute a 'prehistoric' association and be deemed treasure, within the current wording of the Treasure Act 1996. This principle was also upheld by H.M. Coroner for Cardiff and The Vale of Glamorgan more recently, in the case of a terret and rein-ring from Cowbridge (Treasure Case 07.6 for Wales).

Opinion and recommendation

The manufacture of the two bowls is clearly of Celtic Iron Age style, the date of manufacture spanning the Late Iron Age and Roman Campaigning period (AD 25-60). The strainer, also with Native style and use attributes, was possibly manufactured slightly later, though also spanning the immediate pre-Campaigning and early Roman Campaigning period of AD 40-60. As all three were used and buried at the same time and in direct association, it is our opinion that they can be viewed as constituting a 'prehistoric' base metal association in relation to the reviewed definition of treasure Treasure Act 1996 and the Treasure (Designation) Order 2002). The tankard, though buried in the same general locality or 'place', was not buried in direct association with the hoard, being 12.8m distant from it. There is also continuing uncertainty about its dating, which may be Romano-British, rather than prehistoric.

Subsequent actions

Subsequent action after recording: Undergoing further examination at a museum

Treasure details

Treasure case tracking number: 2007T24

Chronology

Broad period: IRON AGE
Subperiod from: Late
Period from: IRON AGE [scope notes | view all attributed records]
Period to: IRON AGE [scope notes | view all attributed records]
Date from: Circa AD 25
Date to: Circa AD 75

Dimensions and weight

Quantity: 4

Personal details

Recorded by: Adam Gwilt - [ view all attributed records]
Identified by: Mr Mark Lodwick - [view all attributed records]

Other reference numbers

Other reference: Treasure (Wales): 2007.24. Associated PAS references: NMGW-7D0AA5, NMGW-4C6212, NMGW-6A53C4
Treasure case number: 2007T24

Materials and construction

Primary material: Copper alloy [scope notes | view all attributed records]
Completeness: Complete [scope notes | view all attributed records]

A resized image of Langstone_07.24a

Image use policy

Our images can be used under a CC BY-SA licence (unless stated otherwise).

Spatial metadata

Region: Wales
County: Newport
District: Newport
Parish: Ringland

Spatial coordinates

4 Figure: ST3889
Four figure Latitude: 51.596417 Four figure longitude: -2.896475
1:25K map: ST3889
1:10K map: ST38NE
WOEID: 27041
Grid reference source: From a paper map
Unmasked grid reference accurate to a 100 metre square.

Discovery metadata

Method of discovery: Metal detector [scope notes]

References cited

Similar objects

Find number: LON-AAED92
Object type: TANKARD
Broadperiod: POST MEDIEVAL
A Post Medieval wood tankard, probably dating 16th - early 18th century. This tankard is of stave construction and the handle has been carved…
Workflow: Awaiting validationFind waiting to be validated

Find number: YORYM-8565C5
Object type: MOUNT
Broadperiod: IRON AGE
A complete cast copper alloy stave built vessel mount of late Iron Age date - circa 120 BC - AD 80. The object has a tubular fitting at one e…
Workflow: Awaiting validationFind waiting to be validated

Find number: WMID-B145A3
Object type: VESSEL
Broadperiod: EARLY MEDIEVAL
An incomplete copper-alloy hanging bowl escutcheon, dating from the Early-Medieval period. The escutcheon consists of a broadly flat, sub-ova…
Workflow: PublishedFind validated and published by finds advisers

Report a mistake

Comment on this artefact's record

Your comment will be checked before being posted. We also check for spam using akismet.

Enter your comments:

* This will not be displayed to the public.

* Not compulsory

The following HTML tags can be used - a,p,ul,li,em,strong,br,img,a - and paragraphs will be automatically created

Audit data

Created: Friday 23rd January 2009
Updated: Monday 22nd April 2013

This page is available in: qrcode json xml geojson rdf representations.

Social Bookmarking: