NMGW-9BEE04: Iron Age hoard

Rights Holder: National Museum Wales
CC License: All Rights Reserved

Image use policy

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


Unique ID: NMGW-9BEE04

Object type certainty: Certain
Workflow status: Published Find published

Curator's Report

Description of object

Rein-ring or harness ring (Found in association with Terret (NMGW-9B2D52)

This is a substantial cast bronze ring with a circular cross-section. At one point, there is a slight casting imperfection, visible as a slight pitted thinning with a slightly bulbous or raised margin. This would not have compromised the structural strength of the piece. Slight thinning and smoothing of the internal surface on one side of the ring may represent a slight wear facet, created through pressure from straps or a bit-link. On one upper surface, there is a scrape, which appears to be modern. The ring has a grey green-brown patination, not unlike the terret. There are small patches of green corrosion deposits covering a small proportion of the ring surface.

On its own, and without follow-up metallurgical analysis, there is little which could be said to be stylistically or chronologically diagnostic about this artefact, beyond its cast technology and ancient patination. However, its association here with another piece of Iron Age horse equipment provides strong ground for comparison and parallel.

Cast bronze rings have been found as separate pieces in a number of hoards and deposits strongly associated with horse pieces such as terrets, strap-unions and bridle-bits: for example at Llyn Cerrig Bach (Anglesey), Seven Sisters (Neath Port Talbot), Polden Hill (Somerset), Stanwick (North Yorkshire) and Middlebie (Dumfrieshire) (Brailsford 1975, 232-3; Macgregor 1962, 49, Pls. III&IV, Cats. 92-4; 1976, Cats. 12-13; Davies & Spratling 1976, 127-8, Cat. 13; Jope 2000, Pl 278b; Macdonald 2007, 219, Cat. 9). Their function has been considered to be either a component of horse harness, or elements of bridle-bits.

Iron Age bridle-bits are composite objects, each comprising two or three bit-links and two rein-rings. Rein-rings may be made of iron, iron plated with bronze, bronze tubes or cast bronze. Simple cast bronze rings are typical of the two-link bridle-bits of the Polden Hills subtype (Macdonald 2007, 72-78 & Table 8) and not of three-link or derivative three-link forms. Comparison of ring diameters suggests that this ring, at 62-63mm, is within the lower end of the acceptable range for a rein-ring. For example, two bridle bits from Llanaber (Gwynedd) have ring diameters of 56-60mm (Ward Perkins 1939; Bowen & Gresham 1967, 175 & Fig. 74; Spratling 1972, Cat. 168), while those from the Polden Hill hoard range from 68-88mm (Brailsford 1975, 224-7). Though of an unusual and slightly different one-link type, the rein-rings on a bridle-bit from the Middlebie hoard, and four associated cast bronze rings, have diameters of 58-59mm and 55-58mm respectively (Macgregor 1976, Cats.11-13). The single cast bronze ring in the Seven Sisters hoard has a diameter of 63.5-65mm (Davies & Spratling 1976, 127-8, Cat. 13 & Fig. 4).

In conclusion, the ring can be identified with a reasonable degree of certainty as horse-associated and with good Iron Age parallels from hoards. Moreover, a scan of rein-ring dimensions on complete bridle-bits indicates that the ring could once have been a rein-ring on a bridle-bit of two-link form. These bridle bits of the Polden Hills subtype have been dated to the first century AD, with most well-dated associations belonging to the Campaigning period the Roman army in Wales and England of AD50-75, towards the end of the Iron Age (Macdonald 2007, 74).

External diameter 61.9mm-63.1mm, internal diameter 48.5-49.1mm, cross-section diameter 6.5-6.9mm, weight 47.6g

Metallurgical analysis

(MD) Analysis to ascertain the elemental compositions of the ring and terret (NMGW-9B2D52) was carried out using a CamScan MaXim 2040 analytical scanning electron microscope (SEM), with an Oxford Instruments Link Isis energy dispersive X-ray spectrometer (EDX).

As both artefacts had very fine surface patinas with no flaws, it was only possible to analyse corrosion products and not the sub-surface metal. Therefore findings are qualitative, rather than quantitative. The terret and ring had near identical compositions, containing copper and tin. No zinc or lead was present. The patination and type of surface corrosion products on both objects were also alike, implying not only similar compositions, but also similar burial environments.

An unleaded bronze composition is consistent with Iron Age objects (e.g. Northover 1984; Dungworth 1996). Zinc and brasses were introduced in Britain during the 1st century AD and became common during the Romano-British period; however they did not entirely replace unleaded tin bronzes.


The nature of the burial association must be considered further and here a potential issue is that the two artefacts were reported to have been found at depth (around 18-20 inches), yet dispersed 6-7 metres apart (around 20 feet). This could be taken to suggest that the two objects were not directly associated upon burial. However our knowledge of the ploughing history of this field, during Medieval and post-Medieval times is poor. Discovery at a depth of 45-50cm might preclude artefact displacement through anything other than deep ploughing, yet it is possible that the finder's estimate of the depth was slightly exaggerated, (as has tended to be the case in recent past investigations of hoard find-spots). Secondly our understanding of the possible displacement of archaeological deposits, since the burial of the metalwork, yet in the ancient past, is not good. In this instance, it was highly unlikely that the precise find-spots could be relocated and investigated through survey and limited archaeological investigation. Therefore, there is no practical way to demonstrate the point, one way or another.

An examination of the known archaeological sites in the area reveals a high density of potentially Iron Age and Romano-British settlements in the close vicinity of the find-spot. Most prominent is the hillfort and Romano-British settlement of Caer Dynnaf, located in fields immediately to the south of the B4270 road. This scheduled ancient monument, centring on SS 983 742 (SAM No. GM100; PRN No. 02444s) is a large earthwork, 3.8ha. in area and 300m long east to west, located on a pronounced overlooking ridge. There is a further annexe enclosing 1.1ha. on the eastern side. The monument has been surveyed and described, beside the other hillforts of Glamorgan (RCAHMW 1976, 40-1, Cat. 670 & Fig 20), although has only received limited excavation between 1965-7. Here, excavation of a small area of the lower terrace on the south side of the fort revealed a Romano-British building overlying Late Iron Age occupation evidence (Davies 1965; 1966; 1967). On current wider understanding, the hillfort was probably constructed during the Early to Middle Iron Age (650-150BC), seeing further phases of development and occupation during the Late Iron Age (150BC-AD47/78) and Romano-British period (AD47/78-410). However, within the fort, on the northern side, is an earlier Bronze Age ring-cairn centring on SS 98365 74323 (PRN No. 03812s). On the eastern side of the fort is a later Medieval castle centring on SS 9852 7427 (PRN No.00262s) and an adjacent Medieval building located at SS 9853 7428 (PRN No. 001915s). In a small field immediately north of the hillfort, are recorded a scatter of three Roman coins located at SS 983744 (PRN No. 02548s). These include two of early fourth century AD date, belonging to Licinius and Helena, and one of second century AD date belonging to Trajan or Hadrian (Guest & Wells, 2007, 166, Cat. 461).

Extending within the eastern side of the field where the artefacts were found and centring on SS 9829 7456, is a probable curvilinear enclosure, which has been identified from an aerial photograph, held at the Heritage Environment Record held by the Glamorgan-Gwent Archaeological Trust (PRN No. 03608.0s). In addition, on the northern side of the field and centring on SS 981 746, a series of cropmarks has also been identified from aerial photographs. These include a B-shaped enclosure, linear land boundary features and a heavily damaged building foundation (PRN No.03023s).These have not been dated or investigated, though on the basis of shape comparisons with many other better known settlement enclosures from Wales, it is likely that the curvilinear enclosure is Iron Age or Romano-British in date. Also noted within the same field is a Medieval coin (PAS Ref. NMGW-D3AE26) and a post-Medieval dam structure (PRN No. 03602s).

The close vicinity of settlement foci and enclosures, possibly occupied during the Late Iron Age, requires the possibility of the terret and rein ring as being located within extensive settlement scatters to be considered. If demonstrably so, they would not be viewed as being buried as a directly associated and a closed prehistoric base metal group, and would therefore not be considered as treasure within the current definition of the Treasure Act 1996 (DCMS 2002, 9-10, Para. 11). However, on current understanding, though there is evidence for possible Iron Age settlement foci within 100-200 metres of the find-spot, there is no convincing evidence for a settlement focus or dense artefact scatter directly overlying the reported find location of the terret and rein ring. Moreover, closed artefact associations can be located even within settlement foci and so even in these instances could be regarded as treasure. Therefore, a direct association (though slightly disturbed in antiquity or later), rather than a dispersed settlement scatter, must remain a reasonable likelihood.

In summary and despite continuing uncertainties, on a balance of probability and on the basis of current understanding, there are good grounds for suggesting that these artefacts were probably buried at the same time and in direct association. The similar patination and metallurgical composition, plausible contemporaneity and similar function of these finds within 6-7 metres of each other, represent a specific convergence of similarities, which are most convincingly explained via burial and direct association in the ground together as a closed association.

In identifying these two artefacts, a discussion of the wider parallels has been offered. These have shown the terret to be stylistically distinctive and diagnostic, while the rein-ring or harness-ring is less so on its own, though becomes more diagnostic by virtue of its close association with the terret and secondly, through comparative patination and metallurgical analysis. Chronologically, on the basis of many parallels, it can be suggested that these two artefacts were placed in the ground during the first century AD, probably within the Campaigning period of the Roman Army in Wales, between AD47-78. However, the terret has a native Iron Age ancestry spanning the Late Iron Age and earlier, while the rein-ring was probably made between AD1-75, either prior to the Roman invasion of Wales (from AD 47 onwards) or during the ensuing Campaigning period. Both artefacts have been shown to be found in frequent association with native or Celtic Iron Age styled metalwork, therefore both can reasonably be regarded to be of prehistoric manufacture. The style of the terret and the absence of zinc in the bronze support this.

Elsewhere, in the case 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 a Wales context. H.M. Coroner for Powys judged firstly that manufacture was determinative over use and burial. Therefore, an artefact of Roman manufacture, though used by native Iron Age societies and buried in direct association with an Iron Age artefact between AD47-78, did not constitute a 'prehistoric' association and therefore could not be deemed as treasure. Secondly, and most germane to this case, it was accepted that in Wales, artefacts of native or Celtic Iron Age style continued to be manufactured and used between AD47-78 - the period of Campaigning of the Roman Army in Wales, prior to final conquest and occupation in AD78. Two or more associated base metal artefacts of native of Celtic style made and buried between AD47-78 would therefore constitute a 'prehistoric' association and be deemed treasure, within the current wording of the Treasure Act 1996. To return to this case, it is argued that the terret and rein-ring may both be regarded as native or Celtic Iron Age in manufacture and style. Being manufactured between 150BC-AD75 (terret) and AD1-75 (rein-ring) and probably buried together between AD50-75, they may be considered a 'prehistoric' base metal association.


Earlier Description

Cast bronze ring with a circular cross-section (external diameter 61.9-63.1mm, ring diameter 6.5-6.9mm, weight 47.6g).

This rein-ring was found 6-7 metres from the Group I simple terret, but was deemed, on a balance of probability, to be associated when buried. Semi-quantitative metallurgical analysis of both indicated a similar tin bronze composition, with no lead or zinc present. Simple terrets were rein guides for chariots or wagons and examples have elsewhere been found in contexts dating between the Middle Iron Age and early Romano-British periods (400BC-AD150). However, many have been found in contexts spanning the Late Iron Age and Campaigning period of the Roman army (150BC-AD75). The rein-ring has a similar diameter and form as the rings on two-link bridle bits of the Polden Hills sub-type, which date to the first century AD. These associated artefacts are therefore considered to be of Late Iron Age manufacture, though probably buried during the Campaigning period of the Roman army in Wales (AD50-75).

Subsequent actions

Current location of find: National Museum of Wales
Subsequent action after recording: Donated to museum after being declared Treasure


Broad period: IRON AGE
Subperiod from: Late
Period from: IRON AGE
Date from: Circa AD 50
Date to: Circa AD 75

Dimensions and weight

Quantity: 1
Weight: 47.6 g
Diameter: 63.1 mm

Discovery dates

Date(s) of discovery: Thursday 1st February 2007

Personal details

Found by: This information is restricted for your login.
Recorded by: Mr Mark Lodwick
Identified by: Mr Mark Lodwick

Other reference numbers

Other reference: Treasure (Wales): 2007.6 & Found with PAS no NMGW-9B2D52

Materials and construction

Primary material: Copper alloy
Manufacture method: Cast
Completeness: Complete

Spatial metadata

Region: Wales (European Region)
County or Unitary authority: the Vale of Glamorgan (Unitary Authority)
District: the Vale of Glamorgan (Unitary Authority)
Parish or ward: Cowbridge with Llanblethian (Community)

Spatial coordinates

4 Figure: SS9874
Four figure Latitude: 51.45575985
Four figure longitude: -3.46943322
1:25K map: SS9874
1:10K map: SS97SE
Grid reference source: From a paper map
Unmasked grid reference accurate to a 100 metre square.

Discovery metadata

Method of discovery: Metal detector
Current location: National Museum of Wales
General landuse: Cultivated land

References cited

Similar objects

Find number: NMGW-9B2D52
Broadperiod: IRON AGE
CORONER'S REPORT Description of object Terret (Found in association with Rein-ring (NMGW-9BEE04) This is a large cast bronze terret,…
Workflow: Awaiting validationFind awaiting validation

Find number: WAW-714386
Object type: BRIDLE BIT
Broadperiod: ROMAN
A fragment of cast copper alloy bridle bit of derivative three-link form, dating to the Late Iron Age or Early Roman period. The fragment is …
Workflow: Awaiting validationFind awaiting validation

Find number: KENT-F7A1ED
Object type: ASSEMBLAGE
Broadperiod: POST MEDIEVAL
1. Harness ring A plain copper alloy ring, circular in cross-section, most likely associated with horse harness. Probably originally circula…
Workflow: Awaiting validationFind awaiting validation

Timeline of associated dates

Audit data

Recording Institution: NMGW
Created: 10 years ago
Updated: 5 years ago

Other formats: this page is available as qrcode json xml geojson pdf rdf representations.