Gwynedd Archaeological
Trust Regional Historic Environment Record
Enclosure and Settlement, Glyn Farm, Llanbedrgoch
Primary Reference Number (PRN) : 5348 Trust : Gwynedd Community : Llanfair-Mathafarn-Eithaf Unitary authority : Ynys Mon NGR : SH51808115 Site Type (preferred type first) : EARLY MEDIEVAL ENCLOSURE
Summary : In the early 1990s, a number of significant Viking-age metal detector finds at Llanbedrgoch were discovered. A geophysical survey then revealed a large enclosure, which was subsequently excavated by Mark Redknap of the National Museums and Galleries of Wales.
The roughly oval enclosure lies on a sheltered south-facing slope near Red Wharf Bay, and has a freshwater spring rising within it. This spring produced finds which dated back as far as the Neolithic, including a crouched burial. There were also a number of finds demonstrating activity at the site in the 1st - 6th centuries AD.
The main focus of the site at Llanbedrgoch began in the 7th and 8th centuries AD. An earthen bank and ditch enclosed a settlement of wooden round houses and rectangular halls.
In the 9th century, the enclosure bank was rebuilt with a massive drystone wall. This may have been a demonstration of the settlement's ability to protect itself from Viking raids, although the defences have also been thought more reminiscent of walled sites in Ireland. The internal buildings revealed changes in building styles, and at least six late 9th-early 10th century buildings have been identified, suggesting a change to more long-houses and halls along the inner face of the wall, with even a paved road running north-south across the site, and possible side roads running from it. Llanbedrgoch was abandoned by the late 11th/ early 12th century.
This low-lying, fortified site was identified as a new type of settlement, possibly a trading centre, and produced the largest collection of early medieval finds from North Wales. These included leather-working implements, antler, and metals such as silver, copper alloy, lead and iron. The discovery of quernstones, grain and animal bones indicate farming in this area, whilst trading is suggested by the presence of hack-silver and lead weights.
Llanbedrgoch may have suffered during the Viking raids of the 10th century. The site is thought to represent an early aristocratic power centre, possibly even with royal assistance from Rhodri Mawr, while both the fortifications and the hack-silver suggest an administrative or economic role for the site.
Description : See PRN 5347. <1>
Geophysical results indicate a D-shaped enclosure, with high spots of activity within the ditch. Excavation has established that the ditch was approximately 2m in width and 1m deep. Its fill contained charcoal and the knob from a crucible of early medieval type, suggesting metalworking activity in the area. Calibrated c14 dates from the ditch fills are AD 450-770 and ad 760-1035. <1>
An enclosed Viking-age settlement situated on gently rising S-facing ground on well-drained carboniferous limestone, near to the modern sea-side resort of Benllech and only some 900m from the broad sandy beach of Red Wharf Bay.
The size and extent of the early medieval settlement is now known, and preliminary analysis suggests that it was well-established by AD 600, when the main activity appears to have been farming. The site during this phase encompassed an area about 80m across, and was enclosed by a ditch measuring 1.7-2m wide and 0.6-1.4m deep.
Two entrances have been located and excavated, which survive as undug causeways across the ditch between a gap in a bank of upcast. Within this pre-Viking-age enclosure there were at least two different forms of wooden structure: at least one small round house, and a large rectangular hall. During this phase dumping of midden material started within the enclosure at the bottom of the slope on the W side. This activity appears to have continued into the late C10th.
During the C9th the enclosure boundary was rebuilt as a defensive structure 2.2-2.3m wide. Most of this wall has been obliterated by late robbing and farming, but sections of the lower courses were found in the S half of the site. Such an ambitious undertaking may be interpreted as a response to both domestic concerns and Viking pressures, construction coinciding with the onset of Viking raids on north Wales from the 850s.
A large slab-lined sub-rectangular pool or reservoir, c.4.7m by 4m, once served a spring emanating from the limestone at one end of the lowest points of the enclosure. Lenses of silt and sand within the pool contained large quantities of animal bone, as well as a number of early medieval artefacts, including the end of a silver ingot (hack-silver), a copper-alloy pin with wrythen-decorated globular head, ironwork and a punch-decorated folded strip of tin. A silver penny of Edmund (939-946) from the upper silt suggests that it remained open well into the C10th.
The remains of a paved road made of limestone slabs and smaller stones between raised stone kerbs were identified to the N of the pool, suggesting a direct link between the pool and the buildings situated about 50m to the N.
The remains of at least five buildings have been found which belong to the C9th/C10th phase. Building 1 - c.10.5m by 5m, sunken floor and narrow, low walls of limestone blocks forming a revetment and a ground-level foundation sill to support a timber superstructure. Building 2 - Rectangular and sunken floored, but larger; 7.5 x 11m internal dimensions. Building 3 - Little remained except flagging and sunken floor. Building 4 - Constructed parallel to the inside of the enclosure wall on the W side of the site. Only a few courses of limestone drystone walling survived. There was at least one further construction phase within the enclosure, involving the erection of a large rectangular building, with a slightly raised central hearth, over the site of building 1.
Skeletons of five individuals were found on the W of the site, buried in the upper fill of the ditch immediately outside the defensive wall. The burials were laid in shallow graves and all aligned N-S. A sixth extended inhumation was discovered in 2001 in a different area of the site, within the enclosure and aligned E-W. <6>
Two skeletons found during excavation at Glyn in 1998. One extended inhumation almost overlying a crouched inhumation. (Longley & Richards, 2000)
The site was first identified through metal detecting by Archie Gillespie who identified a scatter of early medieval and Romano-British finds over three fields at Glyn Farm, Llanbedrgoch. A systematic programme of geophysical survey and excavation identified a large and relatively long-lived settlement dating to the early medieval and Viking period to the sixth century to the late-eleventh century AD (Figure 6.17; Redknap 1996; 1997; 1999; 2001a; 2001b; 2004). Occupation in the Roman period is demonstrated by radiocarbon dates from features belonging to an earlier settlement which span the mid-third to the mid-fifth century cal. AD, as well as pottery sherds, bronze brooches dating to the first and second centuries AD and a silver coin dating from c. 400AD, which were identified on the surface of the fields.
Early medieval settlement (c. 450-850 AD) The first major phase of occupation following the Romano-British settlement is represented by a group of timber buildings which were enclosed by a bank and ditch. The first phase of the enclosure is represented by a narrow ditch, up to 0.4m wide, which produced a date of cal. AD 90560 from the upper fill. This was replaced in the sixth or seventh century AD by a large D-shaped enclosure, which covered an area of c. 1ha and was defined by a substantial ditch up to 2m wide and 1.4m deep (Redknap 2004, 148). A wooden palisade in the boundary ditch was dated to cal. AD 620-860 (Beta-113403). The evidence suggests that the settlement was occupied between the mid-sixth century and the mid-ninth century. Finds associated with the ditch confirm this date for example, one of the ditch fills produced a crucible fragment of early medieval type, demonstrating that the settlement was also the focus for metalworking activities. An eighth-century copper-alloy dress pin also derived from a ditch fill.
The settlement contained a number of timber buildings, including a wattle-and-daub roundhouse. This had a diameter of c. 5m and was defined by a curvilinear gully and stakeholes. Material from the gully fill was dated to cal AD 680-890 (Beta-113392). The roundhouse was sealed beneath a large, rectangular timber hall which was partially overlapped and sealed by another timber hall (Redknap 1997, 94). A posthole associated with the first building produced a date of cal. AD 590-870 (Beta-113402) and it is suggested that occupation was focussed between the sixth and seventh centuries AD (Redknap 2004, 148). Finds dating to the eighth century were also recovered from within the enclosure, including an early medieval bone or antler pin. Evidence for artefact production is also apparent in the form of bone blanks and off-cuts and metal working residues (Redknap 1996, 81). During this phase, a dark-earth artefact-rich midden began to accumulate at the bottom of the enclosure on the western side. This produced large numbers of animal bones and a relatively large assemblage of finds, such as a seventh-century copper-alloy bird-headed pin (parallels exist in Yorkshire; Redknap 1997, 95). The artefactual evidence suggests the main phase of accumulation was initiated in the sixth century AD but this continued well into the tenth century AD (Redknap 2004, 149; see below). A stone-lined pond, present during the next phase of occupation, may also have been constructed during this phase.
Viking-age settlement (c. 8501000 AD) During the mid-ninth century AD, the enclosure boundary was enhanced through the construction of a monumental stone-faced wall, up to 2.3m wide, which was built inside the partially silted-up ditch (Redknap 1997, 95). The interior of the enclosure was reorganized through the construction of new rectangular buildings which were interlinked with trackways. The construction of this larger enclosure has been interpreted as a response to local and Viking pressures, with construction possibly coinciding with the onset of the Viking raids on north Wales from the 850s onwards (Redknap 1999, 61). The boundary may have also have displayed the success of the inhabitants who were closely involved in long-distance trading activities and artefact production. The settlement in the late-ninth and tenth centuries functioned as a centre for farming, craft production and trade (Redknap 1996, 82; 1999, 61).
The excavations focussed on five rectangular buildings, some activity areas defined by hearths, as well as a stone-lined pond. Sill beam methods of construction were utilized in three of the buildings, and they were rich in animal bones and artefacts, suggesting they were the focus for food consumption events, such as feasting. Building 1 was c. 10m by 5m and consisted of a sunken-floored building with low stone revetment walls and a ground-level foundation sill to support a timber frame. The building contained a stone-capped drain and a stone-lined rectangular central hearth which was dated to cal. AD 770-980 (Beta-101535). The hearth was surrounded on three sides by raised earthen areas, interpreted as low benches for sitting or sleeping (Redknap 1997, 94). The floor of the southern half was paved and may indicate a division in space. The adjacent Building 2 was constructed from stone walls which supported a sill beam, c. 7.5m wide by 11m long, and floor deposits have been dated to cal. AD 7051035 (Redknap 2004, 153). The building was larger than Building 1 and it has been interpreted as a hall or barn. Both buildings appear to be contemporary and are associated with a paved and metalled yard surface. Building 3 was very badly preserved but is indicated by stone flagging and a sunken floor (Redknap 2004, 153). A decorated tenth century copper-alloy buckle was recovered from one of the floor deposits. Buildings 4 and 5 are represented by fragmentary courses of walling on the eastern and western sides of the enclosure. During the final period of occupation, a large square shaped building was constructed over Building 1, sometime in the tenth century cal. AD. This building contained a raised central hearth, and possibly functioned as another hall.
A large slab-lined oval pool, c. 5m by 4.8m, with steps leading down to the water, was constructed in the southern half of the enclosure (Redknap 2001, 144). Lenses of silt and sand up to 0.3m deep within the feature produced large quantities of animal bone, as well as a number of early medieval artefacts, including the end of a silver ingot, a copper-alloy pin and a decorated folded strip of tin. On the basis of the dating evidence, Redknap suggests that the pond was constructed during the pre-Viking period of occupation, although it evidently remained open well into the tenth century AD (Redknap 2004, 150). The remains of a paved and cobbled road, up to 3m wide, were identified to the north of the pool, and this feature possibly led directly to the rectangular buildings in the northern half of the enclosure.
Five burial inhumations were also recovered from the upper fills of the enclosure ditch on the western side of the site. The burials were unusually orientated northsouth (rather than the Christian tradition of eastwest) and were situated within shallow graves. Some of the bodies displayed signs of injury and two of the individuals had had their wrists tied either infront or behind their backs, suggesting that their deaths were violent. It has been suggested that the individuals, who were possibly members of a family group and included children and adults, were the victim of Viking raiding activities which occurred sometime between cal. AD 790 and AD 1040 (Redknap 2004, 1545). A sixth extended inhumation was also discovered in a different area of the site. The circumstances of this burial were evidently different, and it has been interpreted as early-Christian due to the eastwest orientation of the body (Redknap 2004, 155).
The finds assemblage from the site was particularly rich and indicates the prosperous status of the settlement and some of its occupants. Finds of hack silver and lead weights reinforce the importance of Hiberno-Norse trading at the site, and large numbers of high-quality artefacts were also recovered, such as early-tenth century arm rings of the Bangor Hoard type, a range of decorative artefacts such as brooches, decorative plates and buckles, and tenth-century ringed pins (Redknap 2004, 15664). Metalworking residues, in the form of silver and copper-alloy droplets and casting sprues, and bone-working residues, such as antler cut-offs, reveal that artefact production was frequently performed at the site. The presence of iron tools, such as awls and toothed tools, also indicate that leather working was being carried out. Large numbers of rotary quern fragments were also recovered, and the large cattle bone assemblage and charred cereal remains reveal that the settlement held an important agricultural role within the community (Redknap 2001, 159) and was possibly the focus for conspicuous community feasts. This is further supported by the dark-earth midden deposits which continued to accumulate up to the tenth century AD to a depth of 0.80m. (Waddington, 2013)
Further excavations were conducted by the Department of Archaeology and Numismatics, National Museums and Galleries of Wales at Glyn, Llanbedrgoch (for summaries of previous work, see AW34, 58-60; AW 35. 58-9; AW36, 81-2).
Four areas were investigated. The excavation of Building 1, originally a longhouse measuring 6m wide by about 11.2m in length, was completed. Areas of raised natural within the house represented the positions of low benching for sleeping and sitting on, arranged along the walls around three sides of a central hearth. the hearth had a carefully-laid rectangular setting of curb stones, and had been relined with clay on several occasions. Post-holes for supports within the structure were found. Broad radiocarbon dates for the building (AD 780-980) were refined by archaeo-magnetic dating of the hearth by the Clark Laboratory, which established that it was last used between AD 890-970 (95% probability). A curvilinear gully associated with stake-holes and fired daub lay stratigraphically beneath Building 1 and formed a continuation of a gully found in 1995. It is thought to form part of a small round-house with a diameter of 5m (radiocarbon age from gully is cal AD 680-885 (to 2 sigma; Beta-113392).
The western end of the large sunken-floored hall or barn (Building 2) immediately to the east was represented by a short section of stone rubble within a shallow foundation trench. Radiocarbon dates of the deposits within the building have already been established that it is broadly contemporary with Building 1 (c AD 855-1000). The rectangular building was 8m wide by 12m long; a paved and metalled surface originally connected the two structures. A post-hole thought to form part of an aisled timber hall beneath Building 2 has now been radiocarbon dated: cal AD 620-800 ( to 2 sigma; Beta-113402).
A narrow trench (N) across the double ditch, revealed originally as a magnetometer anomaly in 1995 (at the eastern side of the enclosure), uncovered evidence for a small bank with stone-revetted inner face, and two narrow linear features thought to be palisade slots within the enclosure ditch which defined the settlement boundary (one with radiocarbon date of cal AD 645-800 to 2 sigma; Beta-113403). The bank had partially subsided into the upper fill of an earlier ditch (radiocarbon date for lower fill cal AD 555-665 to 2 sigma; Beta-113383), and consequently had survived later disturbance.
A third area of the south-west corner of the enclosure (trench O) was examined in order to establish a context for a significant cluster of ploughsoil finds discovered in 1996. A 'dark-earth' midden sequence was found to have built up against the limestone scarp to a depth of 0.8m , and contained large quantities of animal bone and a range of artefacts. The objects associated with these deposits were remarkably well-preserved. These included a copper-alloy mounts and bucket bindings. A southern section of enclosure ditch was located, and its fill contained a 7th-century copper-alloy bird-headed brooch, paralleled by Anglican brooches from Yorkshire, a reminder of pre-Viking contact between Anglesey and Northumbria.
The nature of Viking settlement in Wales remains one of the great mysteries of early-medieval archaeology. Fieldwork by the National Museums & Galleries of Wales is starting to change our perception of the subject, and has found the first indisputable remains of Viking-age settlement in Wales.
The enclosed settlement appears to have been operating from the 6th century, possibly earlier, an to have been at its peak in the 10th century, and there is evidence for the harvesting of crops, the keeping of livestock, the presence of artisans and craftsman, and contact with 'Scandinavian' merchants. The 7th-century bird-headed brooch is paralleled by Anglian brooches from Yorkshire (Sewerby and Uncleby), and reminds us of contact at this time between Anglesey and Northumbria: the overkingship of Northumbrian king Edwin of the Westerners' recorded in the Tribal Hidage, and Bede's record of Edwin's supremacy over the Britons, particularly Anglesey and Man. Cadwallon was responsible for the death of Edwin in 633, who by later tradition has been fostered by Cadwallon's father Cadfan (Catamanus Rex, king of Gwynedd on the Llangadwaladr inscribed stone, died 625). In 634 Cadwallon killed Osric, became ruler of Deira and controlled Northumbria for a year.
Finds unearthed from the Viking-age phase of site are staggering in quantity, already more than doubling the total number of 10th-century artefacts previously known in Wales. They reveal clues as to the everyday function of the Viking-age settlement, and offer a means (when considered in context) to assess the degree to which 'Viking' culture was assimilated into the society of North Wales. Some objects recovered from within the enclosure bear the unmistakable mark of the Hiberno-Norse style typical of the Irish-Sea area. Hacksilver (silver in the form of cut pieces of jewellery and ingot) recovered from within the enclosure shows that silver circulated on Anglesey as currency within the bullion economy of the Hiberno-Norse world. It appears likely that the site may have operated as a seasonal marketplace with itinerant craftsmen, perhaps with access to tradable commodities via the ruler of Gwynedd.
The Viking-age settlement appears to lie on an earlier native site with enclosure ditch, bank, timber buildings and metal-working hearths, showing a conformity to the pre-existing settlement pattern (as occurs at similar sites in Scotland). Of particular importance is the discovery of pre-Viking age wattle-and-daub round-house associated with a timber hall, reminiscent of the enclosure at Pant-y-saer just to the north of the site, with its complex sequence of huts and and occupation. It also recalls the famous description by Gerald of Wales, writing in the late 12th century, of the Welsh building 'dwellings of woven rods sufficient for a year's occupation, assembled with a minimum of labour and a modicum of expense.' The excavations at Llanbedrgoch have established that in the 7th and 8th century, wattle round-houses and substantial halls built principally of wood may have co-existed (Redknap 1997).
While the main thrust of the excavation campaign remains the early-medieval occupation of the site, a number of artefacts were found during the 1998 and 1999 seasons of excavation at Llanbedrgoch which shed further light on activity on the site during the Roman period. These include a late 1st-centuiry Colchester derivative bronze brooch, a 1st-century bow brooch, a late 1st to mid-2nd century bow brooch of 'Severn' type, and sherds of late Roman colour-coated ware.
Fieldwork in the surrounding area identified the site of a possible enclosed hut group situated on a limestone rise just above the 85m contour, to the west of the early-medieval closure (SH 511 825). A denarius of Diva Daustina I (c AD 141-161) was found immediately outside the enclosure bank. If the antiquity of this site is confirmed, apart from providing an interesting subject of comparison with a number of sites in the area, its period of occupation may be contemporary with the use of a number of erly Roman artefacts found in its vicinity., which include a 2nd-century zoomorphic brooch in the form of a duck. Further speculation is pointless without further site investigation
The early medieval occupation of the site is perhaps best considered in terms of of the progress made in each seperate season.
The 1998 Excavation The three-week season of excavation conducted in August concentrated on three areas: the eastern end of Building 2 (first identified in 1995, and dated to the 10th century), a transect across the early-medieval enclosure and ditch systems on the eastern side of the site, and the ditch on the west side of the enclosure (for summaries of previous work see AW 34, 58-60; 35, 58-59; 36, 81-82; 37, 94-96).
The main trench (trench R) was L-shaped and measured c 45m by 8m; an extension to the north (R1) uncovered the eastern end of Building 2. Evidence for smithing was found beneath the later infill of this building, which was pre-dated by a number of small features which had been cut by the stone footings of its eastern end. One pit was radiocarbon dated cal AD 405-670 (2 sigma; Beta-123198, 1510±80 BP), and a stone lined drain to the east of the building provided a radiocarbon date of cal AD 630-980 (2 sigma; Beta-123196, 1270±90 BP). A large flat-bottomed circular pit (diameter 2m) found to the south of the building was unusual in being lined with upright slabs of limestone in 'stave' fashion (Plate 1). Of certain function, the pit fill, which was rich in animal bone and charcoal,, has provided a radiocarbon date of cal AD 540-890 (2 sigma; Beta-123209, 1290±70 BP).
On the east side of the enclosure, a partially paved early-medieval structure was found to overlie an inner enclosure ditch, and a layer associated with this paving has produced a radiocarbon date of cal AD 920-950 (2 sigma; Beta 123206, 1230±50 BP) has been dated cal AD 540-770 (2 sigma; Beta-123200, 1410±70 BP).
Trenches S and T in the south-western area of the enclosure examined 5ht/6th- to 10th-century midden deposits and re-evaluated the western enclosure perimeter identified in 1997. In Trench S, a massive enclosure wall 2.2m wide and orientated approximately north-south had been constructed on large limestone blocks with rubble infill. Traces of this wall had been discovered in 1997 within the uppermost filll of the inner enclosure ditch on the eastern side of the site. The remarkable preservation of the enclosure wall in Trench S was partly as a result of a degree of protection from ploughing provided by a sudden drop in the limestone scarp at this point. Similar subsidence had resulted in the preservation of the corner of a building constructed on limesetone footings (similar to Buildings 1 and 2: AW 35, 58-59), whose rear wall lay parallel to the inner face of the enclosure wall (Plate 3). A deposit immediately beneath the building produced a radiocarbon date of cal AD 690-990 (2 sigma; Beta-123201, 1190±60 BP).
The skeletons of two individuals were found in the upper fill of a ditch situated immediately outside the enclosure wall. The ditch, with a radiocarbon dae of cal AD 620-775 (2 sigma; Beta-123212, 1360±50 BP) for the fill immediately beneath these burials,, represents a continuation of the outer ditch found on the other side of the enclosure. Contrary to usual Christian practice, both burials were orientated approximately north-south rather than east-west. The adult was extended, with slightly flexed legs; the child was crouched, lying on its left side. A sample from one of the leg bones of the extended burial provided on AMS date of cal AD 770-970 (2 sigma; Beta-123213, 1460±70 BP).
Another paved surface, possibly the interior of a timber structure lying beneath later midden deposits, was found in Trench T, in the south-western area of the enclosure. The paved surface lay in the centre of a small ditched enclosure measuring some 8m across, with a more substantial ditch to the south. The radiocarbon date from the fill of the smaller ditch indicates that it belongs to an early phase i the development of the early medieval site: cal AD 630-780 (2 sigma; Beta-123197, 1350±50 BP). Deposits of charcoal and fired clay in this area, probably associated with metalworking activity, have produced a radiocarbon date of cal AD 530-780 (2 sigma; Beta-123202, 1410±80 BP).
The 1999 Excavation The main aims of the three-week season of excavation conducted in late August and September 1999 were to examine the area around the two burials identified in 1998, and to investigate a cropmark in the field to the south whcih suggested that the enclosure and ditch system extended further to the south.
Trench V measured c 17m by 10m; a further impressive length of enclosure wall, 2.2m wide, was exposed. The ditch outside the wall was c 1m deep and 1.2m wide at its base, broadening to c 3m wide at its highest surviving level. At least three phases of recutting were identified, and significant quantities of animal bone (including dog) were recovered from its upper fills.
Three burials wre discovered in the upper fill of this ditch, immediately to the norht of the two burials discovered in 1998 (Plate 4). One child or young adolescent and two adults had been buried without care or ceremony. As in the case of the 1998 burials, their orientation was approximately north-south or south-north rather than east-west. The child or young adolescent was found in the upper ditch fill, orientated with head to south and feet to north. An adult had been thrown directly on top of this body, this time with the head to north. To judge from the unusual positions of the arms, the adult's wrists may habe been tied behind the back.
All the bodies appear to have been casually dumped into shallow graves which utilised the depression representing the enclosure ditch, which had by now silted up. The circumstances of burial and lack of Christian orientation have led to speculation that the three adults and two children were victims of a Viking raid.
Within the enclosure, a shallow ditch, 1.4m wide, lay stratigraphically below midden deposits which are thought to pre-date the construction of the stone enclosure wall. A smaller ditch or gully, c 0.4m wide, lay perpedicular to the main axis of this feature, and to the west of the enclosure wall (Plate 5). An early feature, it may be associated with the 'enclosure' found in Trench T in 1998.
Trench W (c 20m by 21m) located on the southern-western corner of the enclosure in the adjacent field. This confirmed that the site was larger than first thought, being c 1 ha (2.2 acres) in size. Traces of a palisade gully and bank were found, predating the stone enclosure wall, which survived only as a stone spread c 4m wide. The enclosure ditch in this area was c 1.4m wide, and at least 1m deep (the water-table prevented complete excavation). A burnt mound (?bronze age) measuring c 2m by 1m, beneath the enclosure bank, had been truncated by this ditch.
Discussion The discovery of a substantial defensive wall around the site, and evidence for a fifth and sixth building had provided a new perspective on the appearance and spatial organisation. Early-medieval artefacts found during 1998-9 excavations include polychrome glass beads, fragments of decorated bone combs, a small bell with ring-and-dot decoration and a copper-alloy zoomorphic mount. A lead motif piece with punched decoration suggests that Hiberno-Norse armlets of broad band type may have been manufactured on the site during the 10th century.
The settlement at Llanbedrgoch would appear to represent a new type-site for 9th- to 10th-century Wales: a low-lying, fortified multifunctional centre. The preliminary sequence of site development during the early medieval period can now be correlated with an historical framework.
Stage 1: Development of low-lying coastal settlement, engaged in agriculture. Timber buildings, within a bank or ditch. Stage 2: Increasing diversification. Evidence for craft production (eg bronze smithing) during the 9th century. Stage 3: Fortification of trading settlement in response to hostile Viking pressure during the 9th century. Possibility of Viking attack and temporary take-over. Stage 4: Renewed development of the site's potential as a trading post and multifunctional centre during the late 9th and early 10th century with the opening up of long-distance maritime commerce. ?Stage 5: The intergration of Scandinavians and native Welsh, and possibly even Norse residence. The burials are particularly intriguing. All five individuals may have been interred at the same time, and no attempt appears to have been made to commemorate them or mark the location of their graves. A number of lines of inquiry are being pursued in order to establish their ages and cause of death, the probable date of death, and whether any family links can be established between them. Were they the unfortunate victims of violence? If so, who were the perpetrators? Vikings wre not the only aggressors; Mercians destroyed Degannwy in 823 and were campaigning in north Wales for much of the 9th century (the English killed Rhodri Mawr in 873). Nevertheless, in view of Llanbedrgoch's prominent location, Viking attack and take-over of the site, albeit for a short period are distinct possibilities.
In 2000 it is planned to investigate the deposits within the spring which lies inside the enclosure.(Redknap, 1999)
A three-week season of excavation took place in August under the direction of Dr M Redknap, and supervised by Evan Chapman, Mark Lodwick and Heather Jackson. The main aims of the 2000 season were to conduct a preliminary examination within the early medieval enclosure of the feature known as the 'spring' and to confirm the assumed eastern alignment of the enclosure ditch and wall. The latter appeared to be indicated by a patch of unploughable rough ground and some partially visible stones (though to represent blocks from the lower courses of the wall, dislodged by the plough).
Trench Y (c 24m by 14m) was located over the western half of the enclosure spring. Prior to excavation, this 'spring' feature had the form of a steep-sided depression measuring c 30m by 15m. The top of the silts within the depression, which were waterlogged and on occasions below water, lay at a depth of c 1m below the surrounding ground level. The removal of thick vegetation from the western half of the spring revealed previously concealed irregular contours, which subsequent excavation estabished to be the product of a number of distinct phases of activity, in particular the construction of a post-medieval slab-lined well, and the more recent excavation and dumping of material to create much of the present hollow, in order to form a larger watering hole for cattle.
A post-medieval layer of compacted rubble, associated with the post-medieval spring outfall, lay stratigraphically over an earlier slab-lined rectangular pool or 'reservoir' of early medieval date. This pool was roughly oval in plan, and measured c 4.7m long by 4m wide. Early medieval deposits within it survived to a depth of c 0.3m, and were sealed directly by post-medieval layers associated with the later well. The early-medieval contexts comprised complex lenses of grey silt and sand, containing large quantities of animal bone, as well as a number of diagnostic early-medieval artefacts. These included a silver penny of Edmund (AD 936-46) from the upper early-medieval horizon, and the end of a silver ingot (hack-silver), a copper-alloy pin with wrythen-decorated gobular head (similar to exampkes in late 9th- and early 10th-century deposits in York), ironwork and a punch-decorated folded strip of tin from other horizons. Two simple drystone steps had been carefully constructed against slabs lining the northern side of the pool, and would have provided easy access to water.
Excavation of the pool was hampered by continued flooding of the trench with spring water; despite waterlogged conditions and careful wet-sieving of the excavated fills, no organic artefacts were recovered from these silts. About 50% of the early medieval pool deposits were excavated, the remainder being left for the 2001 season.
The remains of a paved street 3m wide, carefully constructed with raised kerbs, were found to the north of the early-medieval pool. Black earth from a layer immediately above the street surface produced a copper-alloy ringed pin of the 10th-century type.
A small extensions wa made to the west of the trench to investiage an area of ifred clay, This represented a large circular oven (c 0.9m in diameter), whose walls were constructed on limestone embedded in clay. The oven lay adjacent to an alignment of limestone blocks whcih resembled the supports for a wall sill. The oven would appear to be of early-medieval date, and will be further investigated in 2001.
An L-shaped trench (Trench Z) was located over the assumed line of the eastern defences. A short length of the enclosure wall (c 2.3m wide) was exposed, identical in charcater to streches uncovered in previous seasons (AW 39, 60, pls 4-5). The outer face of the wall was more complrete thatn the inner face, which had been damaged by ploughing. The ditch outside the wall was c 1m deep and c 3m wide at its highest surviving level, narrowing to 1m wide at its base (bedrock). At least three phases of recutting were identified, as seen in 1999 in the ditch of the west side of the enclosure. Significant quantities of animal bone were recovered from its upper fills.
Within the enclosure wall, a complex sequence of intercutting ditches (some c 0.5m deep) and pits were found beneath a layer of black earth, some pre-dating the 9th- to 10th-century phase on the site. A shallow gully (0.4m wide), orientated east-west, lay stratigraphically below the stone enclosure wall, and at least one of the shallow ditchs (c 1m wide) turned to form a small enclosed area which may have been similar in function and date to the 'enclosure', contained what appeared to be collapsed dry-stone walling, roughly parallel with the enclosure wall. Some features in Trench Z may be Romano-British or earlier in date, and associated with pre-enclosure features outside the ditch.
Discussion While the post-medieval spring-head of Llanbedrgoch has never faded from local memmory, its early-medieval precursor has provided a diagnostic form of structure for this period. It is rare to identify such a prolonged sequene of activity around a spring. At Llanbedrgoch a reservoir appears to have been constructed, to judge from the intial spot-dating, during the 9th or 10th century at the latest. It would seem likely, however, that the spring was also used in the preceding periods. No clear evidence has yet been found for the deposition of votive items, or for the reverence of the site as a sacred spot, although it is significant that a late neolithic or early bronze-age burial was placed c 40m to the east of the spring and some 10m north of the assumed course of its outfall stream. It is well recognised that many watery places have in the past held special religious significance, some as 'liminal' places where 'the present world' links with the Underworld. the significance of Roman artefacts from the course of the outfall stream, and of the early-medieval finds from the pool silts will be reviewed following further excavation. A wide range of motives may prompt the throwing of objecs into such features (in 10th-century Christian Wales,, perhaps, 'good luck' rather than votive), while some objects may be from secondary contexts.
The paved street was contemporary, and provided a direct, easy link between the pool and Building 1 to the north. These features provide an imporatnt key tothe spatial organisation of the site during the Viking period. The street appears to have formed a north-south throughfare across the site. Moreover, the nature of its construction suggests that similar areas of limestone paving discovered in 1998 may also be represent arterial buildings, as previosuly thought. One such route may have run to a building range inside the wall on the eastern side of the enclosure. A number of new factors point to a settlement enclosure with the capacity in the 10th century to house a significant number of people: (1) the well-constructed lanes and streets within the enclosure; (2) the prospect of more buildings set parallel to the enclosure wall; (3) the discovery of a substantial oven (rather than customary open hearth), probably for bread production.
Early-medieval finds include waste globules of leaded bronze, lead spindle whorls, hack-silver, the copper-alloy terminal from an 8th century brooch with sockey for stud and recess for applied foil decoration, fragments of ringed pin, offcuts of worked antler, fragments of rotary quernstone, iron buckles, tools and knife blades. (Redknap, 2000).
A number of prehistoric features were found during the 2000 season of excavation. Trench Z revealed two small pits with dark silty fills containing late neolithic or early bronze-age pottery, flint, burnt hazelnut shell and bone; one of these contained in addition a rubbing stone a a partially-perforated pebble.
The crouched burial of a young person was found in an oval pit, measuring 1.2m by 0.7m, situated to the east (outside) of the early medieval enclosure ditch. Three snall slabs of limestone lined the east side of the pit (maximum lengths 114mm, 215mm, 240mm), with the two slightly larger limestone slabs acting as capstones (that are the southern end measuring 460mm in length; the other placed across the central area), in imitation of a cist. The burial lay on its left side, with the head to the north (ie facing east). No grave goods were found. The fill was a sticky clay, derived from upcast of the pit.
Most of the prehistoric pottery recovered from Trench Z was coarse and thick-walled, and included sherds with impressed thumb-nail, stabbed or scored line decoration. When the dates of the burial and those pits and slots possessing prehistoric characteristics have been confirmed by radiocarbon dating, it will be possible to assess the relationship of the burial to contemporary features. It is possible that the spring provided fresh water during this period, though this is as yet unproven. (Redknap, 2000).
A number of post-medieval layers were encoutered in Trench Y, most of them being associated with a sunken stone-lined boc (settling tank) for fresh water and the foundations for a well chamber above. The well chamber was 3.6m square in plan, with wide drystone footings of varying width (1.3m to 1.44m) which probably provided an inner ledge with rough steps down to the water's edge. The footings of the wall on the south side incorporated at least one fragment of early medieval quernstone. Above foundation level, narrower walls forming part of the superstructure (now missing) appear to have been bonded with mortar. Fragments of slate and corrugated iron were recovered from the backill of the well area and may derive from former roofing materials.
The post-medieval spring- and well-chamber may have been constructed during the 17th century, in association with the rebuilding of the Glyn farm and other estate work by William Bold and subsequent owners. This is also suggested by the fieldname Cae yr ffynnon wen, which goes back to 1637. The spring-head appears to have been operating into the 19th century. The revival of a well and water supply only metres from the early-medieval reservoir emphasises the continued recognition of the natural advantages of the site. Fortunately the excavation of a larger watering hole, probably by bucket and drag-line some time before the 1940s, appears only to have truncated the north wall of the post-medieval well chamber and some of the higher early-medieval deposits to the north and east. After the clearance of black mud and vegetation which clogged up the spring-head, a constant slow flow of clean feseh water once again filled both the post-medieval well and the excavated portion of the early-medieval reservoir. (Redknap, 2000).
A final three-week season of excavation took place in August and September under the direction of Dr M Redknap, and supervised by Evan Chapman,, Mark Lodwick, David Stevens and Brian Milton. The main aims of the 2001 season were to complete the examination of the early medieval 'srping' (AW 40, 105) and to confirm the assumed existence of a second entrance on the northern side of the enclosure, as indicated by the results of the 1994 geophysical survey.
Trench AA (c30m by 16m) investigated the area between Trench Y (opened in 2000 over the western half of the enclosure sprinf) and Trench T (opened in 1998; see AW 39, 59). Vegetation was removed from that part of the spring which had been kept open since 2000; the remaining pool deposits, in the northern half, were excavated and retained for wet sieving. Previous descritpios of the pool's plan and dimensions (given in 2000 as c 4.7m long and 4m wide) can now be revised. The pool is now known to be quadrangular in plan, with drystone wall sides, measuring c 5m by 4.8m internally. Early medieval deposits within it survived to a depth of c 0.3m to 0.4m, and were sealed directly by compacted stony post-medieval layers associated with the later well. The early medieval contexts comprised complex lenses of grey silt and sandy gravel at the bottom, overlain by thicker depostis of sand and peat gravel. These layers contained large quantities of animal bone, and a number of early medieval artefacts, including two fragmentary decorated glass beads and a strand of gold filigree wire. Along the southern side of the pool, a drystone platform 0.8m to 1.0m wide had been constructed, parallel to the northern side, and opposite stone steps described in the previous report (AW 40, 106).
Excavation of the pool was hampered by continual flooding of the tench with spring water. Despite thse waterlogged conditons and careful wet-sieving of the excavated fills, no organic artefacts have yet been idenitified from the silts, although small fragments of wood have been recovered. Behind the stone pool wall, on its north-west side wre found the remains of a wooden roundhouse stake of uncertain date (but proably early medieval or earlier) which had been driven vertically into grey silt and peat deposits.
The circular oven first noted in 2000, wsa found to lie within a sunken-floored building whose walls, though poorly preserved, had limestone footinfs. A flat slab of limestone beneath the oven acted as capping for a stone-lined drain which ensured that the oven stayed dry whe in use. A narrow slot curved with a projected radius of 4m to 5m around the northern half of the structure, and is thought to have been a drip gully. Samples of fired clay wr taken from the clay matrux of two linings of the oven for archaemagentic dating by Dr Peter Rauxloh, Museum of London Archaeological Srvices, but evaluation by Paul Linfrod, Englis Heritage Centre for Archaeology, establised that the clay particles were insufficiently highly fired to align their magnetic domains and provide a relaible date. Dating will consequently depend on radiocarbon dating of the abundant charcoal depostis founf within and around the oven. Other features to the west of the 'bakery' include post-holes and pits, and a sequence of inercuttin curvilinear ditches which are thoug to belong t the pre-Viking phase of the site.
An articulated human inhumation (the seventh burial fro the site) was found in the south-west corner of the trench, within the 9th- to 10th-century enclosure wall. The body had been carefully laid out, and orientated east-west without grave goods, with the head to the east. Stratigraphical evidence suggest that it is early medieval in date.
Trench AB, c 4.6m by 3.5m, investigated a pronounced geophysical anomaly identified in 1994 outside the enclosure, on its eastern side. This proved to be a large, irregularly-shaped pit (3.6m by 2.8m) possibly for the extraction of clay. The fill contained a little animal bone and a few artefacts: a fragment of decorated copper-alloy sheet, and the end of a lead-tin alloy ingot.
A third Trench (AC) was located over the assumed entrance through the northern defences. No trace of the stone defensive wall was found, all traces having been removed by ploughing. The ditch outside the wall was c 1m deep, c 1.8m to 2m wide at its highest surviving level, and the entrance causeway was 2.1m wide. Pre-dating this was a smaller ditch, 0.4m wide at its highest survivng level, which ran inside the eastern edge of the enclosure perimeter and probably defined an earlier enclosed area.
Discussion The silts from the spring were wet sieved on site, for further detailed examination in the museum, in the light of discovery of beads and gold filigree. Trench AA overlapped with areas examined in 1998 (Trench T) and 2000 (Trench Y) in which a number of hullies and ditche had been partially exposed but incompletly understood. Previous radiocarbon dates confirm that some of these features date to the 6th to 8th centuries. It is now clear that some of these features either enclosed or drained small areas, while others relate activity in the bakery and spring area. It has been confirmed that ther was a second entrance on the northern side of the site, and that the pre-Viking period ditched enclosure had a entrance at the same point. The oven, located within a building, implies the presence of a bakery. the complete plan of the early medieval spring, though to have been silting up in the mid 10th century, has been recovered. Early medieval finds included a lead spindle whorl, a copper-alloy pin with a globular head decorated with an incised herringbone pattern and an incised cross at the top of the head, fragments of ringed pins, fragments of whetstones and quernstones, and ironwork (including severeal needels). Prior to the excavation, a metal detector search of the ploughsoil produced two smaller copper-alloy bells, and a small plain equal-arm lead cross with perforation for suspension from a necklace.
This season completed the programme of eight years of excavation by the National Museums & Galleries of Wales on this highly productive site. (Redknap, 2001).
Events : 43306 : A Geophysical Survey Around Find Spots of Viking Artefacts (year : 1994) 43307 : Llanbedrgoch: a Viking Site on Anglesey? (year : 2000) 43784 : Glyn, Llanbedrgoch (year : 2004) 40360 : Early Medieval Burial in Gwynedd (year : 2000) 44557 : Early Celtic Societies in North Wales (year : 2010) 40568 : Early Medieval Burial and Ecclesiastical Sites 2001-2002 (year : 2002)
