Edited by Chris Webster from contributions by Mick Aston, Bruce Eagles, David Evans, Keith Gardner, Moira and Brian Gittos, Teresa Hall, Bill Horner, Susan Pearce, Sam Turner, Howard Williams and Barbara Yorke
The South West of England, and in particular the three western counties of Cornwall, Devon and Somerset, has a long history of study of the Early Medieval period. This has concentrated on the perceived “gap” between the end of the Roman period and the influence of Anglo-Saxon culture; a gap of several hundred years in the west of the region. There has been less emphasis on the eastern parts of the region, perhaps as they are seen as peripheral to Anglo-Saxon studies focused on the east of England. The region identified as the kingdom of Dumnonia has received detailed treatment in most recent work on the subject, for example Pearce (1978; 2004), KR Dark (1994) and Somerset has been covered by Costen (1992) with recent work reviewed by Webster (2000b). The area that became Wessex has been discussed from a historical viewpoint by Yorke (1995) and its later archaeology by Hinton (1977; 1994a). Wiltshire has been reviewed by Eagles (2001), Gloucestershire is covered by Hooke (1985), Heighway (1987) and has also been the subject of two recent conferences (Ecclestone et al. 2004). There is a short review of the Dorset evidence by Hinton (1998) aimed at the non-specialist. The collection of papers edited by Aston and Lewis (1994) covers much of the region.
The Early Medieval archaeology of the region is internally diverse in terms of chronology, with the 5th and 6th centuries (Sub-Roman, Dark Ages, Early Anglo-Saxon) looking very different from the 7th to early 9th centuries (Middle Anglo-Saxon) and the late 9th to mid-11th centuries (Late Anglo-Saxon or Viking period). Moreover, the region has always been perceived, both in terms of landscape history, and in terms of Early Medieval political and ethnic geography, as two entities: one “British” (covering most of the region in the 5th century, and only Cornwall by the end of the period), and one “Anglo-Saxon” (focusing on the Old Sarum/Salisbury area from the later 5th century and covering much of the region by the 7th and 8th centuries). This is important, not only because it has influenced past research questions, but also because this ethnic division does describe (if not explain) a genuine distinction in the archaeological evidence in the earlier part of the period. Consequently, research questions have to deal less with a period, than with a highly complex sequence of different types of Early Medieval archaeology, shifting both chronologically and geographically in which issues of continuity and change from the Roman period, and the evolution of medieval society and landscape, frame an internally dynamic period.
Our understanding of the Early Medieval archaeology of the South West has been dependent upon a series of key excavations and, fortunately, many of the extensive campaigns of excavation, particularly from the 1960s and 70s, have recently been published, key sites such as: Tintagel island (C Thomas 1993; Harry and Morris 1997; Barrowman et al. 2007), Tintagel churchyard (Nowakowski and Thomas 1990; 1992), Trethurgy (Quinnell 2004a), Lundy (C Thomas 1991; C Thomas 1994, 163–182), Cannington cemetery (Rahtz et al. 2000), Cadbury Castle (Alcock 1995), Wells Cathedral (Rodwell 2001), Shepton Mallet (Leach and Evans 2001), Poundbury (Green 1987; Sparey-Green 1996; 2004), Cadbury Congresbury (Rahtz et al. 1992), Henley Wood, Congresbury (Watts and Leach 1996), Bath (Cunliffe and Davenport 1985), Frocester (Price 2000), Gloucester, St Oswald (Heighway and Bryant 1999), Uley (A Woodward and Leach 1993), Deerhurst (Rahtz and Watts 1997) and Trowbridge (Graham and Davies 1993).
Many of these reports contain discussions of themes within the period as well as accounts of the excavations. We are consequently in a much stronger position to assess and appraise future research questions than would have been possible only a decade ago.
The publication of the Shapwick report (Gerrard and Aston forthcoming) will also provide a wealth of evidence for rural settlement in Somerset but for the Early Medieval period, its main conclusion is that, despite 10 years of fieldwalking, geophysics and targeted excavation, the evidence for Early Medieval rural settlement is remarkably elusive.
Probably the largest challenge facing the archaeological study of this period is the lack of chronologically diagnostic artefacts (discussed for the later part of this period by Hinton 1994a, 35–37). There is very little pottery production (although evidence for this is increasing) until the very end of the period and imported wares appear during very limited phases and are geographically restricted. Without scientific dating it is extremely hard to identify sites to this period for excavation and many that have been excavated have been identified either by the presence of imported pottery or were located serendipitously. The increasing use of scientific dating is showing, however, that sites that might once have been thought to belong to another period, for example Trethurgy (Quinnell 2004a), Shepton Mallet (Leach and Evans 2001 and more recent unpublished work), or Hayes Farm (Simpson et al. 1989), have Early Medieval phases. In the eastern parts of the region there is material from burials but a similar lack of domestic material that would allow settlement sites to be located, although chaff-tempered ware has been identified during fieldwalking in the Avon valley (Light et al. 1995).
This paucity of evidence makes identification of settlement difficult and thus questions such as population, land-use or social hierarchy very difficult to approach.
Doubtless these issues also occur in earlier periods (possibly unrecognised) but the problems are complicated by the presence of historical sources of very varying degrees of reliability and hence the expectation that archaeology should answer very specific chronological and socio-political questions that involve precise dating. The uncritical use of these sources in the past led to a very simplistic view of the region based largely on the framework provided by the Anglo-Saxon Chronicle. However, it must be remembered that the Chronicle was being compiled in the 9th century for propaganda purposes, although it did make use of earlier written material, some of which has not survived in any other form. The early annals appear as a seamless whole, but conceal within them a point at which contemporary recordkeeping began, probably around the middle of the 7th century (Stenton 1926). The annal entries for the late 7th and early 8th centuries cluster around the north-west border of Wiltshire and it is possible that they were compiled at Malmesbury. Therefore entries from the late 7th century onwards carry much more authority than those for earlier periods. The earlier entries should not be dismissed altogether, but the dates for anything before 600 have been shown to be completely unreliable (Dumville 1985). The 577 entry for Dyrham is not without interest, however, and the reference to three cities and three leaders recalls the forms of Welsh triads (Sims-Williams 1983). It may preserve some authentic information, but the least reliable facet of it is the date attached and this should not be cited as a point around which Anglo-Saxon penetration of the region can be dated.
The historical evidence for the 5th and 6th centuries provided by Gildas (Winterbottom 1978), is traditionally associated with Glastonbury but this is not accepted by recent authorities who would see the association as a product of Glastonbury’s medieval pilgrimage “industry”. However, Higham (1991), KR Dark (1994, appendix 1) and Sharpe (2002, 108) have all argued that Gildas may well have written in the lowland zone of the region, with Higham and Dark favouring Dorset or Wiltshire while Sharpe prefers Somerset or Gloucestershire. KR Dark (1993) further argues for St Patrick originating in the South West and arguments along these lines have existed for some time (for instance, O’Rahilly 1942) and continue (Jelley 1999) but there is less acceptance of these claims among the academic community today, where the case for the North West of England is considered to be stronger.
There is also the potential of other Anglo-Saxon written sources to illuminate the nature of society in the Early Middle Ages. Bede has an important account (Ecclesiastical History ii, 2) of the meeting between Augustine and British bishops at Augustine’s Oak, and its aftermath. The law code of Ine has a series of clauses dealing with the Britons under his authority that may help us understand the nature of the AngloSaxon annexation of much of the region (Attenborough 1922). Another important source is the writings of Aldhelm of Malmesbury who seems to have been charged with bringing the British church under West Saxon control and enforcing conformity with “Roman” customs in the late 7th century (Lapidge and Herren 1979; Lapidge and Rosier 1985). Other sources, for example the medieval saints’ “lives” have been studied (for instance by Pearce 1978, appendix 3) but there are also the products of Arthurian studies, Celtic mythologies and other “New Age” beliefs to contend with. These last clearly have an important effect on the public’s perceptions of this period and need to be appreciated, engaged with and perhaps countered, if we are to move forward. As well as public perceptions of mythologies and the romance of the Early Medieval period, it is equally important to recognise that the Early Medieval period has a fundamental position in the popular imagination in terms of nation-building. Therefore, while archaeologically authoritative accounts must negotiate with many misconceptions over Early Medieval “races” and “ethnicities”, they must not forget that this debate is central to modern claims of identity. Nested and conflicting modern identities of British, English and Cornish have often seemed to be predicated upon Early Medieval racial distinctions between Celts and Saxons, and these continue to pervade our attempts to explain and communicate the Early Middle Ages.
In broad outline, discussion of the period can be divided into three chronological episodes, although the first two are often complicated by synchronicity across the region and questions of ethnicity.
There is continuing debate about the transitions that occurred around the end of Roman control of the provinces of Britannia with sharply contrasting views ranging from total systems collapse to the survival of a recognisably Roman administration, lacking only access to new forms of material culture. The date of this transition is also argued, with some seeing the start of Roman “decline” occurring during the 4th century whilst others see continuity for several centuries after the traditional date of 410.
Most of the debates over the timing and nature of the transformation of the Roman province (see for example KR Dark 1994 and Henig 2004 for views which stress continuity and Faulkner 2000; 2004 for a systems collapse model) have focused upon southeastern Britain. There are real challenges in applying models constructed in these regions to the west of Britain where the nature of Romano-British society may well have been substantially different, as the archaeological record seems to suggest. At the end of Roman administration it is likely that the South West occupied a large proportion of the province of Britannia Prima but the ways that the withdrawal of imperial control affected the region are poorly understood.
The second phase would traditionally be described as the Anglo-Saxon conquest and sees the gradual absorption of the southern parts of the region into the kingdom of Wessex, with control of the Hwicce of Gloucestershire taken by Mercia. The Hwicce were also probably originally a British kingdom; they and the people to the north, the Magonsaete appear to be Christian from at least c.660 (Sims-Williams 2000, 55–9, 75–9, 84; Bassett 2000). Central to this traditional view is the Chronicle date of 577 for the battle assumed to be located at Dyrham to the north of Bath when three “kings” were defeated and three cities captured. A more secure date is provided by charter evidence from Glastonbury and Malmesbury in the reign of Centwine (676–85) and the life of St Boniface which refers to an English abbot at Exeter in 680 (Yorke 1995, 60). Obviously, the eastern parts of the region had an Anglo-Saxon identity from an earlier date and Cornwall was never fully incorporated into the West Saxon kingdom.
The final phase, from the mid-9th century sees the effects of growing international contacts and trade, reflected in both the Viking raids and the growth of urbanism at towns such as Exeter and Gloucester. In the countryside, it is from the ninth century that we may have clear and unequivocal evidence for village nucleation in some parts of the South West as evidenced by the Shapwick Project (Gerrard and Aston forthcoming). There is also the development of a more central authority and more fully Medieval forms of ecclesiastical and secular settlement.
Examination of rural settlement is hampered by the problems outlined above in the identification of sites belonging to this period. Our knowledge of forms of settlement at the beginning of the period is also hampered by limited understanding of rural settlement in the later Roman period but there is a growing body of evidence for continuing occupation on Roman settlement sites. Small enclosed settlements, perhaps containing a few round (or in Cornwall, oval) houses are a feature of the western part of the area from the Bronze Age onwards and recent excavations have indicated settlement into the Early Medieval period at some. In Cornwall, Trethurgy (Quinnell 2004a), which began in the late pre-Roman Iron Age continues into the 6th century with the best published sequence covering the 1st millennium AD in the region.
Another significant settlement was excavated at Gwithian during the early 1950s, directed by Charles Thomas, which formed part of a series of related programmes of fieldwork which documented the long history of settlement and land-use in this unique coastal setting on the north Cornish coast. Stratified occupation horizons with stone and wooden structures, stone-lined hearths and settlement middens were found together with evidence for small-scale (industrial) iron working. The remains of a contemporary field system were also discovered. The site was rich with finds and is of particular significance as it produced a long stratified sequence of key local, post-Roman and Early Medieval ceramics as well as a unique metalwork assemblage. These are key ceramic assemblages for the post-Roman period and a recent programme of accelerator dating residues on selected Gwithian sherds show a continuous sequence of settlement which dates from the 5th to at least the 8th centuries cal AD. A further dating programme will continue during current work on the archive (Nowakowski 2004; pers. comm.).
Sites in Devon, such as Hayes Farm (Simpson et al. 1989) show activity on similar enclosed settlements and it appears likely that the Somerset site of Yarford, where a late prehistoric enclosure was succeeded by a small villa, will have a post-Roman phase (Wilkinson et al. 2004). Work at the old church site in Shapwick, Somerset, has shown that the church (later resited to the planned village) was associated with a manorial complex which overlay a Roman settlement. Whilst no diagnostic artefacts were recovered, radiocarbon dates indicate that one building was constructed around 710–720 cal AD (OxA-11461, 11474, 11475, 11873, 11874, 11930, 11931) and went out of use before 910–1060 cal AD (OxA-11932, 11933, GU-5898, 5899, Chris Gerrard pers. comm.; Gerrard and Aston forthcoming) suggesting that occupation of the site may have continued from the Roman through to the Medieval period. Interestingly the underlying Roman settlement was not one of the two villas identified in the parish, perhaps showing that it was not the Roman period elite who founded the Medieval manorial centre. There is other evidence for late reuse of Roman buildings in Shapwick; the small Roman building at Sladwick was partly demolished following a period of abandonment and used to shelter a hearth around 430–640 cal AD (SUERC2938; Chris Gerrard pers. comm.; Gerrard and Aston forthcoming). A similar picture may be seen at Worth Matravers where a single-roomed building, probably a barn, which contained a corn-drier was floored over late in its life and subsequently used for a single burial. Unfortunately there were no dates for these events but they are likely to be very late or post-Roman (Graham et al. 2002).
Recent fieldwork in northern Somerset has identified 8 enclosed settlements in the woodlands around Congresbury within 1km of the hillfort settlement of Cadbury Congresbury (Keith Gardner pers. comm.). Although these are as yet undated, further examination may provide a context for the occupation at the hillfort.
Excavations at the Roman villa at Frocester have shown significant post-Roman occupation of the site. The main villa building appears to have been partly destroyed by fire but the front corridor was then reoccupied. Post-holes suggested divisions of the room into an eastern end with a hearth, associated with chaff-tempered pottery, and a western end where further post-holes and wear to the floor suggested, to the excavator, a byre, and thus a layout similar to that of a Medieval longhouse (Price 2000, 115–6). Reynolds (2006) has pointed to the lack of a drain in the byre area which would argue against this. Three timber buildings appear to have been constructed within and around the courtyard of the villa, one with a radiocarbon date of 430–660 cal AD (CAR-1475 – the calibration curve suggests that a date after 530 is more likely). Unfortunately, the dating of the villa fire and the subsequent occupation is not precise enough to distinguish between continuity of occupation or reoccupation.
Further east, enclosed settlement appears to be absent and, without this visible (even when ploughedout) attribute, few settlements are known. In Wiltshire a settlement at Collingbourne Ducis (Pine 2001) was excavated in advance of development. Eight sunken-featured buildings were recorded together with a possible post-built structure. Calibrated radiocarbon dates suggest occupation in the 8th and 9th centuries (BM-3162–4) but one date centres on the later 6th century (BM-3165). A cemetery containing some 5th-century artefacts had previously been excavated 150m away (Gingell 1975/6). At Market Lavington another apparently contemporary settlement and cemetery of similar date have been excavated (Williams and Newman 2006).
In Gloucestershire finds are concentrated in the upper Thames valley, for example, Sherborne House, Lechlade, (Bateman et al. 2003), almost certainly due to the level of development there. In these areas, cemeteries are likely to be much easier to identify due to their diagnostic grave goods, whereas settlements of the same period are difficult to locate. However, by analogy with Anglo-Saxon settlements elsewhere in the upper Thames valley, we may be looking at a densely occupied landscape from the later 5th century with sizable dispersed (and perhaps shifting) settlements situated on the gravel terraces above the Thames and its tributaries. The Sherborne House site comprised sunken featured buildings and “halls” associated with boundary ditches dating from the 6th century onwards. There was evidence of shifts within the village and Reynolds (2006) notes both the early date of property boundaries within the settlement and also parallels with the buildings at Poundbury (Green 1987), suggesting a western tradition of construction.
There are also places where there is little archaeological knowledge but where there are suggestions that work might pay dividends. One such is the Exmoor village of Porlock, where early sculpture is known and whose raiding by the Vikings suggests a settlement of some importance. The church tower also appears to have a defensive aspect to it.
Rural settlement is also hard to identify in the later periods, until the 10th century when pottery begins to make a reappearance. The coastal settlement at Mawgan Porth falls into this period (Bruce-Mitford 1997) and finds of 10th-century pottery from rural sites are increasing in areas such as Somerset. The evidence from Shapwick (Gerrard and Aston forthcoming) suggests that by this date a previously very dispersed settlement pattern, which we are unable to identify because of the lack of diagnostic artefacts and its scattered and small-scale nature, was becoming more concentrated and is now hidden under many modern farms and villages. The situation further east is similarly difficult because of the lack of chronologically diagnostic artefacts with few patterns emerging (Hinton 1994a).
High-status rural sites are dealt with below but also need to be considered in the context of possible ecclesiastical origins and urban developments.
The contribution of placename studies to the understanding of rural settlement is particularly relevant to Cornwall where pre-English names survive in large numbers (Rose and Preston-Jones 1995). This has allowed more detailed work on settlement patterns than elsewhere in the region (see Turner 2006a for a summary). Placenames have also been put to good use on a smaller scale at Shapwick where the identification of field-names with habitative components has been used to reconstruct the pre-open field settlement pattern (Gerrard and Aston forthcoming).
The decline of urban settlement in the major Roman towns, such as Cirencester, Gloucester, Dorchester, Ilchester and Exeter is a well-known feature of the late 4th and 5th centuries but in most cases the evidence for the fate of the towns is poor. The enigmatic “dark earth” has been recorded at a number of these urban sites, for example Bath, Gloucester and Exeter, but the nature of this deposit continues to be debated. More work on the longevity of the Dorset Black Burnished Ware industry may help to provide dating evidence (Gerrard 2004b). The best evidence for continuing populations is perhaps the large organised cemeteries of oriented burials at such places as Poundbury (Farwell and Molleson 1993) or Ilchester (Leach 1994). Unfortunately as these sites are “Roman” they are rarely dated by radiocarbon, so it is usually not possible to say when they went out of use. Interestingly, the cemetery at Queenford Mill near Dorchester-on-Thames (Oxfordshire) has produced dates suggesting use into the 7th century AD (HAR-5324, 5325, 5350, 5351, Chambers 1987) which suggests that the dating of simliar sites in the South West might be fruitful. The evidence from Poundbury, however, is that that part of the cemetery was out of use and built-over sometime in the 5th century (Sparey-Green 1996). The cemetery at Kingsholm outside Gloucester does, however, see one possibly early 5th-century burial within what is believed to have been a Roman mausoleum (Hurst 1975). The body of a male aged 25–35 was accompanied by a silver belt-buckle, an iron knife and other silver buckles. Parallels for these items suggest a south-eastern European origin, in which case, a Late Roman date may be preferred (Hills and Hurst 1989). It is less clear what was going on within the town walls, although the Greyhound Yard site at Dorchester seems to show continuing use of buildings in the north of the excavated area, with chalk floors laid over mosaics, whilst the southern part was given over to cultivation (PJ Woodward et al. 1993). Other parts of the town appear to show a similar picture of less urban use, although with some (undated) activities continuing (Hinton 1998, 11). There is late occupation of the amphitheatre at Cirencester, perhaps suggestive of its use as a defended enclosure (Wacher 1976; Holbrook 1998), although the nature of the blocking of the entrance does not look particularly defensive.
At Exeter, the discovery of a post-Roman cemetery dating from perhaps 450 indicates continuing population but not necessarily urban life (Allan 1991, 29–35) as may the church of St Pancras which predates the medieval street pattern (Bidwell 1979). In Gloucester the excavation at St Mary de Lode suggests a (?continuing) ecclesiastical presence at least (Bryant and Heighway 2003), although Blair (2005, 31n) stresses that the first phase is, in fact, undated and could be late.
Recently good evidence has been found for the continuing occupation of the Roman roadside settlement at Shepton Mallet where some of the plots had been used as burial grounds for groups of burials (Leach and Evans 2001). More recent work on the site has produced archaeomagnetic dates of 285±80, 400±90, 515±45 and 515±65 from stonebuilt hearths within demolished buildings (Peter Leach, pers. comm.).
Historical sources attest that town life seems to have become re-established by the late 9th century at Exeter and other places, although there is very little archaeological evidence for this. The work at Trowbridge (Graham and Davies 1993) showed a sequence of structures below the later castle, church and cemetery. These included sunken-featured and post-built buildings. In the absence of closely dateable pottery or other artefacts it was not possible to say whether occupation was continuous with the succeeding manorial enclosure and then castle. The status of the site is thus unclear. Other places such as Taunton and Ilchester are assumed to have some urban characteristics by the late 9th century because of the presence within them of mints but there has been no archaeological evidence from this early date with the exception of a radiocarbon date of 770–1150 cal AD from the cemetery beneath Taunton Castle (HAR-2674, Clements 1984, 28). There may be evidence of planned settlement from the 9th century at places such as Glastonbury and Iron Acton.
John Blair (2005) has recently argued that many of these urban places (including many of those mentioned in the Burghal Hideage) were associated with the presence of earlier minsters. The interaction between religious communities, royal estates, markets, defended sites and later urban status needs further work.
In a period with poor survival of material culture it is the settlements of the elite that are most visible with their imported “luxury” items and metalwork. In the early period these appear to take the form of refurbished hilltop sites such as Cadbury Congresbury (Rahtz et al. 1992) and Cadbury Castle (Alcock 1995), a phenomenon known from across western Britain. The possibility of a “hall” of Early Medieval date, tentatively reconstructed from the patterns of postholes at Cadbury Castle, may be further evidence of elite occupation and emphasises the challenge of identifying structures at sites of this nature. PostRoman phases have been reported from Crickley Hill (Dixon 1988a, 78) in two areas, one suggested as a defended elite centre, but in the absence of full publication these cannot be confirmed (Charles Parry pers. comm.). In Dorset, a similar refortification to that at Cadbury Castle has been suggested at Poundbury (Hinton 1998, 15) where excavations in 1939 (Richardson 1940) located a limestone wall constructed on the top of the Iron-Age ramparts and not associated with any refurbishment of the ditch. The excavator dated this to the latest Iron Age on the basis of a sherd of pottery from the ditch but in the absence of any indication of occupation of this date from the interior (RCHME 1970) a post-Roman date must be considered.
Tintagel continues to hold a special place in the interpretation of these sites. The interpretation of Tintagel island as a “celtic monastery” is almost universally out of fashion, but its role as a high-status residence or central place continues to hold sway, although debates continue about its size, status and duration of use (see for instance KR Dark 2000; C Thomas 1993). The recent work by the University of Glasgow (Harry and Morris 1997; Barrowman et al. 2007) has examined (and re-examined) the evidence from Radford’s excavations. New excavations on the lower terrace of site “C” identified three structural phases dated by radiocarbon. The earliest (395–460 cal AD) is associated with hearths, floors and stakeholes and the only pottery is Romano-British. In the second phase (415–535 cal AD) Romano-British pottery is associated with the arrival of Mediterranean imports and in the last phase (560–670 cal AD), the stone structures excavated by Radford are associated with large quantities of imported pottery. Reexcavation on the upper and middle terraces located an ephemeral building predating the stone structures located by Radford and similar to those found on the lower terrace. The deposits in these structures suggested that they may have been seasonally occupied. The stone building was confirmed to belong to this period and close to it was found a remarkably inscribed slate with two phases of lettering (Late Roman and post-Roman) and a collection of glass, possibly Spanish. Work also showed that the “Great Ditch”, by the lower ward of the castle on the mainland, had been in use (and its natural defensive qualities enhanced) in the immediate post-Roman centuries. These findings, together with 5th–6th century material from below the medieval lower ward of the castle itself have reinforced the sheer scale and significance of Tintagel island (Rachel Barrowman, pers. comm.).
Later high-status sites do not appear to be characterised by the presence of similar diagnostic artefacts and those that are known have been identified by chance or aerial photography. Perhaps the best known is Cheddar, excavated by Philip Rahtz (1979) and interpreted by him as a palace of the kings of Wessex. The site is adjacent to the sites of several Roman buildings, one at least with the character of a villa, and comprises a series of timber halls and associated structures. John Blair has questioned the royal origins of the site (Blair 1996; Blair 2005, 326–7) suggesting that the king’s hall was founded in the late-9th or early-10th century adjacent to a minster (with probable British origins) and grew at the church’s expense.
There is also the enigmatic site at Foxley to the west of Malmesbury where aerial photography, geophysical survey and limited excavation have shown what appears to be a hall (of several phases) with subsidiary buildings as well as a church in its own enclosure. Only one radiocarbon date, 420–770 cal AD (HAR-6216), has been published (Hinchliffe 1986), from charcoal in the wall trench of one of the ancillary buildings; an unpublished date from the one of the hall post-holes calibrates to AD 660–970 (HAR-8082). The earlier date range, as Eagles (2001) points out, spans both British and AngloSaxon control of the area. Blair, however, sees this site as a monastic cell of Malmesbury with a significant agricultural function (Blair 2005, 214–5).
In Gloucestershire (Reynolds 2006) a similar site was excavated at Kingsholm to the north of Gloucester where timber halls have been identified with a palace mentioned in the Anglo-Saxon Chronicle in the mid-10th century (Hurst et al. 1985) and another site was excavated at Holm Hill, Tewkesbury; a site which later became the residence of the Earls of Gloucester (Hannan 1997).
There is increasing evidence for the continuity of pottery manufacture, particularly from Cornwall. “Gwithian Style” pottery in gabbroic fabric was identified as long ago as 1958 as continuing Late Roman forms (C Thomas 1958) and the current work on the Gwithian archive will provide further information on this. There is also material from Trethurgy in a similar fabric which appears to continue a limited range of Late Roman forms into the 6th century (Quinnell 2004a, 111). There is one sherd of the later grass-marked pottery from Trethurgy which is petrologically identical to the earlier material, suggesting continuity of manufacture (David Williams in Quinnell 2004a, 126–7). Grassmarked pottery is elsewhere more firmly dated to the 8th century but the evidence for continuity of manufacture may require revision of this date (and/or that for the cessation of Late Roman forms). There is also the possibility of local breaks in pottery availability and use (Quinnell 2004a, 127). This detailed evidence is currently limited to Cornwall but there is some evidence for local wares elsewhere, such as at Mothecombe (Turner and Gerrard 2004) and Bantham (Silvester 1981; Griffith 1986; May and Weddell 2002; Bidwell and Reed in preparation) both associated with Mediterranean imports. To the north there are sherds of grass-marked pottery from Cannington (Rahtz et al. 2000) and at Cadbury Congresbury there also appeared to be continuing use of pottery made in an Iron Age tradition into the early 5th century (Rahtz et al. 1992). Shell-tempered ware is found in the north of the region, for example at Lower Woods, Hawkesbury, where it was found in layers which overlay a late 4th-century mosaic. The later period is equally poorly served, with little pottery before small quantities appear in the 10th century. Collections of this material have been published from Cheddar (Rahtz 1979) and, more recently, material from 1950s’excavations at Glastonbury Abbey has been re-examined (Kent 1997; 1998). The exception, to this aceramic picture is the presence of chaff-tempered wares in Gloucestershire and Wiltshire but these change little during their long date-range and are rarely found in non-excavation contexts such as fieldwalking.
The Glastonbury pottery was associated with glass furnaces which were dated to “before the middle of the 10th century” (Radford 1956, 69–72; Radford 1981); current thinking might push the date earlier into the 9th (Harden 1971) or even late 8th century (Justine Bailey quoted by Kent 1997). A full report on the excavation and finds is awaited.
An iron-working site at Blacklake Wood on Exmoor has recently produced a radiocarbon date of cal AD 415–650 and five sherds of pottery that appear, on the basis of granitic inclusions, to be products of the South West (Martin Gillard pers. comm.). The site appears to have been using surface ores and the waste products indicate a different technology from that used at Sherracombe Ford (also on Exmoor) in the Roman period. There was no evidence of smithing, implying that artefacts were being fabricated elsewhere. Other later dates are known from iron working sites such as Bywood Farm, Dunkeswell and North Hill Glider Field (6th–9th century) on the Blackdown Hills (Reed 1997; Horner 1998; Bill Horner pers. comm.) and at Burlescombe (7th–11th century, Reed et al. 2006). Iron working is known from further east, at Ramsbury (Haslam 1980), Gillingham (Heaton 1992) and Worgret (Hinton 1992a). The last site produced dendrochronological dates of between 664 and 709 for a structure, perhaps a watermill, filled with slag and furnace residues. If the structure at Worgret was a mill it would be only one of many mentioned in Domesday Book and it is, perhaps, surprising that more evidence for milling has not been located.
There is some limited evidence for quarrying: a charter from Bishop’s Lydeard (AD 899×909, Sawyer 1968, No. 380) refers to stangedelfe, the stone digging (although the boundary clauses may be later, Robert Dunning pers. comm.) and there is a similar reference from Pucklechurch (c.950).
There appears to be very little evidence for other activities, although wooden fish traps in Bridgwater Bay have recently produced 10th-century dendrochronological dates (Groves et al. 2004) and radiocarbon dates beginning in the 9th century (GU6002–6011, 6038, 6039, Richard Brunning pers. comm.). The context of these structures is, at present, not understood but it is possible that they are a feature of royal estates along the Somerset coast. There is also the enigmatic site at Duckpool in North Cornwall where industrial hearths have been found close to the beach which date from the Late Roman period and the 8–12th centuries. There was no evidence from the area excavated that occupation was continued between these two periods and the industrial processes involved are unknown (Ratcliffe 1995).
As mentioned above, assumptions about racial identity (more recently reformulated in terms of “ethnicity”) have had a central place in historical and archaeological research into the Early Medieval period. Archaeological evidence has long been used and abused in this debate, from Victorian studies of “celtic” crosses and “teutonic” graves. Questions of ethnicity evoke much current debate (see for example Lucy 2000; Hills 2003). The key debate has tended to focus on the nature of “Anglo-Saxon” ethnicity, often linked to the scale and character of Anglo-Saxon immigration.
Drawing upon the last 25 years of revision in Early Medieval history, ethnicity in the Early Middle Ages has been reformulated as a “situational construct” defined by political allegiance rather than biological or even familial ties. The implications of this argument for the assumption of “British” and “AngloSaxon” identities from the material evidence have been noted as cautionary, but have yet to be fully explored. At one extreme, conflating the issues of ethnicity and migration, some archaeologists have taken these arguments to mean that we can reject the supposed population movement: Bede’s “Adventus Saxonum”. However, this “immobilist” approach that emphasises “British” continuity over “Saxon” immigration also has its roots in Medieval biases and British nationalism (Härke 1998) and needs to be challenged as rigorously as traditional stereotypes of Germanic immigration and population replacement. Certainly we have evidence of novel forms of burial, settlement and material culture that can be associated with Germanic groups, but whether these are immigrants (Eagles 1994), or indigenous groups under the influence or direct control of invading elites (Higham 1992) remains open to debate. The South West may provide the best place to study such changes due to the apparent progression of Anglo-Saxon culture, both from east to west and through time.
Archaeological evidence from Wiltshire certainly seems to suggest significant “British” elements within the “Anglo-Saxon” population. Cemeteries such as Collingbourne Ducis and Market Lavington have produced burials which appear to have been wearing Roman-style dress, and female burials at Harnham appeared to favour bracelets and finger-rings, again a Roman fashion (Eagles 2001, 218). The rite of clothed inhumation itself, however, is a novel development that has parallels not only in other parts of England but also with Francia. One distinctive feature of Wiltshire is the low proportion of cremation burials, marking the burial rites of the region out from other parts of Wessex, and Anglo-Saxon England as a whole (Williams 2002). This might be evidence of greater “acculturation” between natives and immigrants, or alternately, it could reflect a distinctive local identity. More research is needed to ascertain whether this absence of cremation is a result of poor preservation or indeed is a reality of past mortuary practice.
Evidence from Gildas, Bede and Aldhelm suggests that, in the period from c.550 to the early 8th century, there was pretty uncompromising hostility that solidified the identity of both sides as “British” and “Anglo-Saxon” respectively, irrespective of the question of whether in reality many of the AngloSaxons of eastern Wessex were of British descent (Hines 2000). Later, the laws of Ine (688–725) show an “apartheid” between Anglo-Saxons and British; the latter were presumably people of the South West who had been recently incorporated into Wessex. They are described as wealas from which comes “Welsh”, but at the time of the laws the word probably had the connotation of “foreigners”. The British had rights and were assigned wergilds and oath values according to class, but these were worth less than those of their Anglo-Saxon counterparts.
The history of the church in western Wessex could support this. Evidence from Aldhelm’s writings suggests British clergy had to conform over matters such as the celebration of Easter and the form of the tonsure in order to continue in their positions. One could suggest that what lies behind these arrangements are the conditions on which the British submitted to the English – the British agreed to accept Saxon authority and in return had their legal (but second class) status assured. No doubt landowners also had to surrender much of their land to incomers like the family of the missionary Boniface who seem to have moved from Hampshire to Devon probably in 670s. It is possible that some of the villages in Devon, otherwise an area of dispersed settlement, were founded at this time. The battles as recorded in the Anglo-Saxon Chronicle may indicate that the Saxons advanced district by district (or by groups of districts). Dumnonia, for instance, does not seem to have fallen in one fell swoop, but to have suffered a gradual loss of its easternmost territories with boundaries inexorably moving further and further west. By the time Alfred (871–99) drew up his law code, everyone he ruled had become West Saxons – there is no separate provision for people of British descent. The prevalence of the Old English language as seen in placenames suggests that the former British had had to become “Anglo-Saxon” rather than that there was any compromise.
Archaeologically the “Anglo-Saxons” also appear distinct with a range and use of material culture that differed from that of areas to the west. For much of the region, however, the incorporation of the area into English kingdoms took place six or seven generations after the first appearance of Anglo-Saxon material culture in the east of England, and at a time when the use of distinctive items deposited in burials was ending. It is possibly that DNA studies may help but possibly more fruitful is the increasing use of stable isotope analysis to identify the diet, environment and place of origin from the human skeleton. Provisional strontium and oxygen isotope analyses of a range of Early Medieval cemeteries across England (Budd et al. 2004) suggest that first-generation immigrants from over the North Sea are likely to be recognised but also a range of local and regional migrations could be identified. Inevitably such studies are hindered by the poor bone preservation from much of the South West.
In Wiltshire the distribution of Anglo-Saxon material appears to show an expansion of Anglo-Saxon burial sites from the Avon valley (particularly around Salisbury) in the 5th century, across Wiltshire in the 6th, and into Dorset in the 7th. Eagles suggests that a clear area of Anglo-Saxon penetration into eastern Dorset beyond Bokerley Dyke sometime in the 6th century is visible in the distributions, possible halting at a refurbished Combs Ditch (Eagles 2004).
To the north there are large 6th-century cemeteries with Anglo-Saxon material which appear to follow the present county boundary of Gloucestershire (for instance that recently excavated at Butlers Field, Lechlade, Boyle et al. 1998) but to the west such burials are less common, although there are small numbers of secondary interments in barrows (Heighway 1984; 1987; Reynolds 2006). The small group from Bishop’s Cleeve stands out from this as it would appear to represent the burials of a small community (perhaps 13 people burying over 50 years) in an area where other communities were disposing of the dead without Germanic material culture (Holbrook 2000).
There is even less information on the other migrations reported in historical sources and suggested by placenames; movements of Irish, either direct or via South Wales, and the, presumed larger, movement to Brittany (Giot et al. 2003).
One group makes a large impact on the history of the 9th and 10th centuries: the Vikings. Raids are reported along many coasts following the killing of the king’s reeve at Portland (see Pearce 2004, fig. 111) but the archaeological evidence for raiding or settlement is extremely sparse. There are two (with a fragment of a third) hogback tombs from Cornwall (Langdon 1896) and the dedication of a church in Exeter to St Olaf may show the presence of Norse communities (Pearce 2004, 277). There are also hints of peaceful activities at Gloucester in 877 when the Vikings are reported to have “built booths in the streets” (Heighway 1984, 236). The re-interpretation (Gardner and Ternstrom 1997) of a 19th-century discovery of burials on Lundy as a Viking cemetery will need further work before it can be considered convincing. As in many other areas the chief legacy visible to us is the renaming of nautical features, but the mechanism by which this name replacement took place is extremely obscure (C Thomas 1997).
Much of what is known of this period comes from the scanty historical sources with a limited contribution from archaeology. Gildas only refers to the presence of kings in the western part of the region and it is possible that when he writes in ch. 27 that Reges habet Britannia, sed tyrannos; iudices habet, sed impios (Britain has kings, but they are tyrants; she has judges but they are wicked) he was contrasting two separate political systems with which he was familiar (Woolf 2003). All the kings that he mentions are associated with the least Romanised areas of western Britain and it could be argued that these areas had been left more under native control in the Roman period and so circumstances were more favourable for the emergence of kingdoms when Roman authority was removed altogether. In the more Romanised eastern areas there might have been a more gradual adaptation of local government to changing circumstances, and this may be reflected in the iudices (“magistrates”?) of Gildas and the three leaders (not necessarily “kings”) based in towns in the Chronicle annal for 577.
It could be argued that, in the west of Britain at least, the evidence of Gildas and St Patrick provides more support for gradual adaptation to changing circumstances following the withdrawal of imperial control. The case for a basic continuity of the British church and its structures from the 4th to 5th centuries and beyond has been cogently made (Sharpe 2002). That, however, need not mean that Christianity had made a substantial impact on all areas of the region before 400.
The presence of “elite” sites has been taken to indicate the presence of a strongly hierarchical society (together with Gildas’s “kings”) but the archaeological evidence for status is based solely on the presence of imported artefacts and, in some cases, the amount of labour needed to refurbish the defences. However, in the absence of other contemporary sites, it is not possible to show that the presence of imported pottery is a characteristic of status and the size of the defended area might suggest communal defence rather than an exclusive elite residence. The theoretical model employed to understand these sites has been based on a now-questioned and out-moded model employed for the pre-Roman Iron Age. Since the interpretation of hillforts has undergone a major shift from military defence and elite residence to communal and seasonal gathering places and ritual centres, we must equally open the debate with regard to postRoman sites.
As noted above there is historical evidence for the people of Dumnonia and the Hwicce but this is not well-reflected in the archaeological record. One pattern which may be evident is the distribution of enclosed settlements which does appear to concentrate within the area believed to be Dumnonia with its eastern border along the Quantock Hills. This may, however, be more of a reflection of the pastoral nature of the west of the region and a prevalence of cattle raiding. There do seem to have been distinctive settlement types in Cornwall during the Roman period (the courtyard house and the oval house); the latter tradition, at least, appears to continue after 400.
To the east the importance of Selwood as a boundary may be perpetuated in the two dioceses of Wessex: Winchester and Sherborne. The latter was established in c.705 for what seem to have been the predominantly “British” areas of Wessex where Aldhelm, the first bishop, had been labouring for some years to bring the church into conformity with Canterbury. The earlier presence and importance of Selwood is suggested by its British name of Coit Maur, Great Wood, but this needs to be treated with caution as it may only be its 9th-century name known to Asser. Eagles (2001, 214) has suggested that it formed the boundary of the Roman civitas of the Belgae.
Another boundary appears to have followed the Bristol Avon which divided the Hwicce from the people of Somerset and later divided Wessex from Mercia. This would appear to be the boundary that the West Wansdyke was built to defend. The boundary between Gloucestershire and Somerset uses the river itself and Reynolds and Langlands (in press) suggest that this move took place in the 10th century.
Eagles (2001; 2004) has argued that a part of the boundary between Hampshire and Dorset follows that of the Durotrigan territory and was defended by the Bokerley Dyke. He also suggests that further to the north the boundary lay to the east of the county boundary, perhaps through Teffont to the River Wylye, based on finds distributions and placenames (Eagles 2004).
The origins of the boundary between Somerset and Dorset may be based on the possible Late Roman division of the Durotriges into two cantons based on Dorchester and Ilchester. Costen (1992, 85–6) notes that the placenames Rimpton in Somerset and Ryme in Dorset come from the Old English for “edge” and are likely to have been named by the earliest English speakers in the area. He suggests that the boundary must have been laid out at around this time but there is no reason that it could not be earlier. It is unknown whether the Somerset/Devon border bears any relationship to that of the former Dumnonia which is often believed to have lain further north-east, closer to the River Parrett.
The development of smaller territorial units, such as estates, parishes, hundreds and the like, has been studied in several areas but much has depended on the availability of historical evidence, principally charters, to define the earliest recorded units. A good review has been provided by Hooke (1994) which draws attention to the potential of defining British territorial units in the far west as well as the better recorded estates in the Wessex heartlands. The use of Iron Age hillforts, some reoccupied in the Early Medieval period, to define Early Medieval territories has a long history, and continues to hold potential (Burrow 1981; Rahtz et al. 1992). The extent to which this is possible might be testable through the work of the Cadbury Castle environs project (Davey 2004). Costen (1994) has examined some estates in detail in Somerset and adjoining counties and there have been detailed studies of particular estates in the same area by Corcos (2002). There remains considerable potential for integrating landscape, placename, charter evidence and archaeological research to understand the development and evolution of territories and estates in the Early Medieval South West (see Reynolds 1999). For instance, the place-name research of Aliki Pantos on assembly places across England (Pantos 2004) combined with an ongoing project to examine Early Medieval assembly sites archaeologically by Sarah Semple (2004), contain the potential to heighten our understanding of some of the most important, yet archaeologically understudied, categories of “central place” in the landscape as does the research of Andrew Reynolds concerning execution places and burials in Wiltshire.
The identification and analysis of the structure of royal estates also has great potential for the understanding of many aspects of the period. Most minster churches appear to have been founded on these estates and they also appear to have a bearing on the location of mints and of other settlements, for example burhs, which often appear to lie on their edges, as at Axbridge on the edge of the Cheddar estate. It is possible that while the minster lay on the royal estate, it and the royal “villa” were not always on the same site. The site at Foxley (Hinchliffe 1986) may represent such a royal site with a small chapel, the minster lying at Malmesbury, although Blair (2005, 213–4) would see it as a subsidiary monastic cell. Most of such sites will, however, have developed into modern settlements and thus be hard to discover. In particular, minsters seem to develop into towns, often with planned royal additions. These sites may also be identified by later royal attempts to regularise (and thus profit from) the “customary” market.
Two episodes of external trade are visible in the archaeological record for the early part of the period (see Pearce 2004, table 3 for a useful overview). The earlier, which seems to centre on the period 475–550, is characterised by pottery from the eastern Mediterranean with a smaller component of North African origin (which appears to be slightly later, Campbell 1996). The vast majority of this material (originally defined as Tintagel A and B wares) in Britain comes from Cornwall, Devon, Somerset and South Wales with the largest amount from Tintagel (C Thomas 1993). There are many sites in Cornwall that have produced a few sherds (C Thomas 1981b) and more recent finds such as at St Michael’s Mount (Herring 2000). In the last few years, sites along the south coast of Devon have begun to produce larger amounts of this material, for instance, Bantham (Silvester 1981; Griffith 1986; May and Weddell 2002; Bidwell and Reed in preparation) and Mothecombe, where there appear to be houses to one side of the beach with an area of hearths on the other (Sam Turner, pers. comm.). Similar occupation was discovered at Wembury along the coast, but without the imported pottery; it has been dated by radiocarbon to cal AD 420–600 (Wk-13086–8, Reed 2005). It is possible that Tintagel and sites on the south coast such as St Michael’s Mount (Herring 2000) or Bantham (Silvester 1981; Griffith 1986; May and Weddell 2002; Bidwell and Reed in preparation) were fulfilling the same function as the so-called “productive” sites (Pestell and Ulmschneider 2003) known from eastern England and may have been recognised trading centres like the later wics.
The other sites producing quantities of this material are Cadbury Congresbury (Rahtz et al. 1992) and Cadbury Castle (Alcock 1995), neither of which lies on the coast and both are assumed to have been high-status sites which consumed the contents of the amphorae and used the tablewares. Tintagel island appears to have combined both roles. Small quantities of this material turn up on other sites in the region, such as Trethurgy (Quinnell 2004a), Lundy (McBride in Allan and Blaylock 2005, 88) and Carhampton (McCrone 1995) which do not appear to be particularly high-status, although the last two may be monastic. It is likely that more remains unrecognised in excavation archives but the distribution does appear to exclude Dorset and Gloucestershire, a fact which cannot be explained by the presence of AngloSaxons at such an early date. Corfe Castle has been suspected of having a pre-Norman phase (Hinton 1994b; 1998) and the collections from here might be worth re-examining.
The later episode comprises material dated to the 6th and 7th centuries. Most of this is E-Ware which dates towards the later part of the range; the earlier D-Ware is extremely rare. This material is believed to derive from western France and is much more widespread than A and B wares in the British Isles but it does not tend to be found on hillforts such as Cadbury Congresbury and Cadbury Castle which has led to the suggestion that these were abandoned before this date. It is, however, mostly absent from Somerset in any case which again emphasises the problems when only “high-status” sites are known.
Glass also appears to form part of this later continental trade as it has a similar distribution and is stratified with D-Ware and E-ware at Whithorn in southern Scotland (Campbell 1996; 2000; Hill 1997). The only site with significant amounts in the South West is Cadbury Congresbury. There is some indication of the movement of items from English areas, although debate continues as to whether this was in the form of raw glass for local manufacture (Whitehouse 2003) or as complete vessels (Campbell 1996, 93). Certainly some complete vessels are found such as the jar from Pagans Hill (Evison in Rahtz and Watts 1989, 341– 5) and metal objects of English origin were found at Cadbury Castle (Alcock 1995).
Consideration of trade patterns in the later periods is similarly hampered by the almost complete absence of material culture until the 10th century (Hinton 1994a).
The spread of Christianity is one of the areas where much research has focused in the past, stemming from a long tradition of interest in this topic by the Medieval church and later historians. Burial is also one of the few areas where there is a comparatively large amount of data. The religious affiliations of the population at the beginning of the period are obviously deeply connected to those of the Late Roman period and there has been much argument on the extent to which Christianity had spread in urban and rural populations by AD 400 (C Thomas 1981a; Petts 2003). The numerous saints’ “lives” which survive (mostly from much later than the period) paint a picture of missionary activity from South Wales but Gildas appears to be writing in a Christian society, more concerned with heresy than paganism.
Unfortunately the most direct evidence for Late or post-Roman Christianity, the Shepton Mallet amulet (Johns in Leach and Evans 2001, 257–260), appears to be a modern forgery; its authenticity has been in doubt for some years, on both art-historical (Martin Henig, pers. comm.) and scientific (Johns in Leach and Evans 2001, 260) grounds. More recent scientific tests have confirmed that the silver alloy contains too few contaminants to be pre-Industrial Revolution (Stephen Minnitt pers. comm.). The presence at Shepton Mallet of separate groups of oriented and north-south aligned graves, however, still suggests the presence of a Christian (and a pagan) community. This leaves the best evidence at Poundbury (Farwell and Molleson 1993), where Late Roman mausolea contained Christian imagery, and the well-known Christian mosaics at villa sites in Dorset. In addition to these there is now a suggestion that the recently discovered villa at Bradford-on-Avon had a baptistery and possibly a later cemetery. There are also pagan temple sites that appear to have been re-used as Christian churches, for example at Uley (A Woodward and Leach 1993) where a wooden basilica was built over the site of the pagan temple. This was replaced around 600 by a smaller stone building which then had an apse added to the north-east end. These are both interpreted as chapels but the careful burial of the unweathered head of Mercury from the Roman temple near the junction with the apse suggests a complex picture. Similar associations of Roman temples with later, oriented, structures are known at Brean Down (ApSimon 1965) and Lamyatt Beacon (Leech 1986) where they were also associated with east-west burials.
There is no sign of any surviving Roman church in eastern Wessex at the time of the appointment of Birinus as bishop in 635 (Hase 1994) but in the west there is both historical and archaeological evidence. Gildas believed that the church of his day had evolved from its Roman antecedents without a break and the presence of seven bishops and other churchmen at the meeting with Augustine in c.601 shows a wellorganised structure. Eagles (2003) has suggested that Augustine may have travelled along the Thames and that the meeting place, known to Bede as Augustine’s Oak lay close to the source of the river, which burial evidence suggests is close to the limit of AngloSaxon settlement at the time. If this location, or other locations in the South Gloucestershire area, is correct and taken with the origin (from Bangor-is-yCoed near Chester) of some of the British representatives it is possible that that the province of Britannia Prima still operated for ecclesiastical purposes at that time. Gildas, indeed, may have been writing specifically for this province. The locations of the other bishoprics depend in part on the fate of the civitas capitals (presumably the centres of Roman sees).
Monasticism was already coming into western Britain when Gildas wrote De Excidio (a date much debated but probably in the first half of the 6th century) and by the time of his later writings was more fully established. These later writings suggest that we should not expect all monasteries in the west to follow the same degree of asceticism – Gildas did not approve of the extreme asceticism associated with St David. Studies throughout the insular world suggest that monasterium and associated words could be applied to any religious community and did not necessarily have connotations of celibacy and the following of a rule.
The meeting with Augustine and the writings of Aldhelm suggest that by the 7th century the British church was seen as very conservative and adhering to customs of Easter calculation and tonsuring that were prevalent when Christianity had been introduced into western Britain in the later Roman period. The British views on Easter were seen as bordering on the heretical and it is maintained in some quarters that the heresy of Pelagianism (which concerned the doctrine of free will) was rife in the British church (Herren and Brown 2002) but, if this were the case, one might have expected reference to it in Anglo-Saxon sources.
Many sites have been identified by archaeologists and historians as those of early churches or monastic settlements. In Cornwall, although Tintagel island is now suggested to be a secular elite site, there is continued interest in the origins of place names incorporating the element *lann (enclosure) and showing circular form (Olson 1989; PrestonJones 1992; Pearce 2004, 136–48). David Petts has argued convincingly, however, that the circular churchyards are probably later as he can find no evidence for enclosed cemeteries in Britain before the 8th century (Petts 2002). Several Cornish churchyards have produced cist graves (Preston-Jones 1984) as has, possibly, Street in Somerset (Bridgers 2003; Calder 2004). Street (“Lantokai”) may well be the original focus for the religious sites of Glastonbury with the others, most famously the Tor (Rahtz 1970; Rahtz and Watts 2003), originating as outlying hermitages. These hermitages do not appear to have been approved of by the Roman church and were regularised in some way to become the chapels which survived into the Medieval period (Aston 2004), for example St Michael’s on the Tor or St Bridget’s at Beckery (Rahtz and Hirst 1974). Recent work on the causeway linking Street to Glastonbury has produced radiocarbon determinations suggesting an 8th-century date (Richard Brunning pers. comm.), probably associated with the growth of the Anglo-Saxon monastery at Glastonbury itself.
There are also other locations in Somerset that have been suggested as early monastic sites, such as Carhampton (McCrone 1995), Banwell, Congresbury (Oakes and Costen 2003), Kewstoke (Calder 2004) and St Decumans at Watchet (Calder 2004). In Dorset the excavated settlement at Poundbury has been proposed as monastic, based primarily on the continued use of the Roman stone-built mausolea as churches (Sparey-Green 1996; 2004).
Despite the difficulties it is clearly important to try to identify early church sites, and the type of church involved, as they appear to have had a significant influence on the landscape and societies in their areas (Turner 2003; 2005; 2006a).
Many of the known ecclesiastical sites in the west continued as important places in the AngloSaxon church suggesting that the organisation inplace there was one familiar to Anglo-Saxon rulers and churchmen. Both structures appear to have comprised royally-held central places, accompanied by the more important churches, surrounded by a dispersed settlement. The main difference appears to be a greater closeness between secular and religious in the Saxon kingdom which may explain occurrences such as the movement of church sites over small but significant distances, as may have occurred at Sherborne (Hase 1994, note 30). The old site then seems to have become a chapel dependant on the new minster. There may also have been political and doctrinal reasons for this as as discussed below.
Various ecclesiastical sources indicate a rapid colonisation by Anglo-Saxon churchmen and women in the wake of the Anglo-Saxon take-over of the South West. It is apparent that many of Boniface’s correspondents lived in the western part of Wessex, but unfortunately we do not know the location of all of their communities (Yorke 1998). Some places only known to us as parish churches may have been significant male/female religious communities with high standards of Latin culture even though little is known of their material culture.
The extent to which John Blair’s minster theory can be universally applied to the South West requires further exploration (Blair 1994). Key archaeological questions relate to the identification of the plans and extent of Early Medieval monasteries, most of which are poorly understood. The combination of focused archaeological research, combined with topographical and place-name analysis, used by Hall and Whyman (1996) at Ripon, must provide a template for future studies of Early Medieval ecclesiastical sites in the South West. As well as the monasteries, we also need to understand the development of the daughter and manor churches. There seems to be a very large number of churches which, from their siting, are likely to have started life as manorial chapels or churches and this, too, needs to be understood.
Hase (1994) discussed the location of the early “mother” churches and suggested evidence for their location on sites close to water but above the floodplain in contrast to other parts of the country where Roman sites and other earlier fortifications seem to have been preferred. More recent work (for example, Blair 2005; Gittos 2002) has suggested that the situation is more complex; the South West had fewer towns and stone-walled forts (which appear elsewhere to have been favoured over villas for church sites) than other regions. Other factors, such as the Christianisation of pagan sites, for example, at Uley (Hase 1994, 48), Knowlton (the church has 12th-century features but is otherwise hard to date, RCHME 1975, 111) and Bath (Cunliffe 1986; Davenport 2002), which may also have contributed to the apparent preference for sites close to water, will also have been important. The location of Malmesbury within an Iron-Age hillfort may be due to its Irish origins (Blair 2005, 190). The decision, in many cases, is likely to have lain with the donor of the land. This is an area of continuing debate where the South West may provide much of the evidence.
The differences may be explained by the existence of a pre-existing Christian population in western Wessex so that sites were located close to existing British centres as, for instance, with the siting of the seat of the new diocese of Selwood close to Lanprobi (believed to be Sherborne) rather than at Ilchester or Dorchester (Hall 2005). Sherborne is one of a number of places where earlier British monastic centres appear to have been replaced by new minsters, built in large rectangular enclosures that still survive in the modern street-patterns of towns and villages today. This may be the result of a deliberate suppression of British ideas following the Synod of Whitby and the arrival of the reforming archbishop Theodore in 668. There was a hardening of attitudes towards the British church which was increasingly seen as heretical in its views on the date of Easter and also in its monasticism which, with its emphasis on individual asceticism, was seen to be contrary to the views on grace propounded by St Augustine of Hippo. These views may have led to the replacement of “remote” British sites by minsters providing pastoral care, on new accessible sites and based on rectangular plans to emphasise their “Roman-ness” and therefore orthodoxy. In Dorset and the adjacent areas of Somerset, a system of these minsters seems to have been created under Ine (688–725), possibly by Aldhelm whose writings show a strong concern for orthodoxy. Similar beliefs may have prompted the start of stone church building, seen by Bede as “in the Roman manner” (Hall 2000; 2004; forthcoming) and, perhaps, a preference for new locations, uncontaminated by the past, which appears to have been the case at places such as Yeovil (Gittos and Gittos 2004b).
The small size and large number of these early sites prevented them from becoming too powerful and wealthy and, again, may have been a deliberate policy of royal control. This system began to break down towards the end of the period under the pressures of the increasing wealth of some churches based on endowments, the pastoral needs of a dispersed population and the establishment of churches by powerful laymen at their own residences. These developments, together with the monastic reforms initiated by Dunstan at Glastonbury from about 940, led to the creation of the medieval parish system.
Glastonbury itself has been studied extensively by both historians (see Carley 1988; Abrams and Carley 1991) and archaeologists (Rahtz and Watts 2003 and references therein) but there has been less archaeological work at the other reformed monasteries of the 10th century listed by Hill (1984): Abbotsbury, Athelney (unpublished Time Team excavations), Bath (Davenport 2002, 57–60), Buckfast, Cerne, Cranborne, Deerhurst (Rahtz and Watts 1997), Exeter, Gloucester St Peter’s, Horton, Malmesbury (Haslam 1984, 111–7), Muchelney, Tavistock, Westbury-onTrym (unpublished excavations in 1968 revealed 10thcentury timber-buildings, burials and a jetty, Bob Jones pers. comm. – see Ponsford 1968; 1981) and Winchcombe (Mullin 2006). The nunneries are similarly poorly known archaeologically.
The Shapwick project has shown that a stonebuilt church had seen several phases of alteration (and acquired burial rights) before a burial dated to 890–1150 cal AD (SUERC-2937; Chris Gerrard pers. comm.; Gerrard and Aston forthcoming).
The evidence surviving in standing church fabrics is poorly understood over most of the South West. The Taylors’ survey (Taylor and Taylor 1965) is now over 40 years old and more recent work, both published and unpublished, suggests that there is a far greater survival rate than they supposed. Important investigations have been conducted on some notable buildings, for example, Deerhurst (Rahtz and Watts 1997) and Bradford-on-Avon (Hinton 2001), raising many questions about their complex history and design while evidence of others has been gathered through excavation: Gloucester St Oswald’s (Heighway and Bryant 1999), Cirencester Abbey (PDC Brown and Evans 1998), Bath Abbey (summarised in Davenport 2002) – both of these were constructed within areas of standing Roman masonry, Wells (Rodwell 2001, but see Blair 2004 for the origins), Cheddar (Rahtz 1979; Blair 1996), Glastonbury and Tor (Rahtz and Watts 2003), Muchelney Abbey(Taylor and Taylor 1965, 451– 3), Potterne (Davey 1964) and Exeter (Allan 1991).
Of parish churches, in addition to the published work on East Coker (Gittos and Gittos 1991) and Exeter St Martin (Blaylock and Westcott 1989), many retain features highly suggestive of an Anglo-Saxon origin or building tradition. There is documentary and substantive evidence of many other churches in existence in Anglo-Saxon times which include Shaftesbury (Keynes 1999), Crediton (Orme 1980), Ramsbury (Taylor and Taylor 1965, with its collection of 9th–10th century carved stones, Cramp 2006, 228–34) Malmesbury (Pugh and Crittall 1956, 210–31), Wareham (Hinton and Webster 1987), Milborne Port (considered a late 11th-century rebuilding by Blair 2005 and Gem 1988 but it could have been an updating of the existing church which, in any case, must have existed), Bitton (Ellacombe 1878; Taylor and Taylor 1965) and Britford (Chambers 1958–60; RCHME 1987, 113–5). The areas around Bath, Cirencester and Sherborne show particular concentrations of Anglo-Saxon physical evidence which are unlikely to be the result of differential survival.
Although there have been some previous local studies (for example Foster 1987), the South West’s carved stone has recently been thoroughly catalogued and reviewed (Cramp 2006). The material is more substantial than might be thought at first and includes highlights such as the standing crosses of the far west, the font at Melbury Bubb (Cramp 2006, 104– 6), the recently discovered sculpture from Congresbury (Figure 10.1; Oakes and Costen 2003), the fine and important collection at Ramsbury (Cramp 2006, 228–34) and elements incorporated into standing buildings, such as at Keynsham (Cottle and Lowe in Lowe 1987, 103–6).
Burial rites have been used by archaeologists as a source of cultural, religious, social and ideological evidence for contemporary society. The burial tradition of the sub-Roman South West from the 5th century comprises both large and small cemeteries with oriented burials at what appear to us, as we know so little of the settlement pattern, to be remote locations (Petts 2004). These presumably have some relationship to the large Late Roman cemeteries such as Poundbury (Farwell and Molleson 1993). There are also indications of other continuing Late Roman mortuary practices, such as the burial with hobnails at Shepton Mallet dated to cal AD 430–680 (GU-5293, Leach and Evans 2001, 45). The burials within most of these post-Roman cemeteries, however, are usually simple interments with little evidence for coffins but there is also a tradition of stone “lining” to graves. The most developed of these is the cist grave as recorded, for example, from 5th and 6th century contexts at Tintagel churchyard (Nowakowski and Thomas 1990; 1992) where the cist graves were lined and roofed with large slates forming a complete stone burial chamber. A similar picture is seen at Ulwell in Dorset, which dated to the 7th century and contained 13 cist graves (some with lids), 13 stone-lined graves (some with only head and foot stones) and 27 simple graves (Cox 1988b). Similar structures form a small component of the large cemetery at Cannington (Rahtz et al. 2000) but with no indication here of roofing slabs.
Other burials at Cannington contained smaller amounts of stone, appearing on excavation as a line of stones along the sides of the grave and these are also known at cemeteries with no true cist graves such as the cemetery on the site of the Late Roman temple at Henley Wood, which had 30 burials with stone lining together with 26 simple graves (Watts and Leach 1996). The small cemetery at the foot of Brean Down (Bell 1990, 73–83), which may be associated with an oriented building replacing a Roman temple (ApSimon 1965) on the Down above, contained graves lined with irregular boulders and produced 3 radiocarbon dates spanning 340–900 cal AD (HAR-8548, 8549, Birm-246). A similar cemetery partly investigated at Wembdon Hill (Langdon 1986; Croft and Woods 1987; Croft 1988) again revealed graves with partial stone linings and radiocarbon dates from the 7th to the 10th centuries (GU-5149–5151). Yet again, some cemeteries appear to consist entirely of simple graves such as Lamyatt Beacon, where in a very similar way to Brean Down, a Roman temple was replaced by an oriented building and accompanied by 16 burials (Leech 1986). Such cemeteries continue to be located, usually as a result of work on other sites, or as a result of development such as the cemetery of oriented inhumations at Tolpuddle (Hearne and Birbeck 1999, 55–62, 150–161, 226–230, 246–148), with radiocarbon dates centring on the 7th century (OxA-8297–8300, 8320).
There are also graves that are marked out as, in some way, special such as the proposed “cella memoria” on Lundy, believed to be the burial place of a “saint” – Charles Thomas (1991; 1994, 171) suggests St Nectan before his translation to Hartland – and graves enclosed by square ditches at Kenn (Weddell 2000) and Stoneage Barton (Webster and Brunning 2004). Very similar enclosures are also known at Poundbury (Farwell and Molleson 1993), where they lie beyond the main burial area, and at several other places in Britain, in Roman and post-Roman contexts. It has been suggested that the square-ditched enclosures are secular elite burials consciously copying Roman mausoleum forms (Webster and Brunning 2004).
There may also be more unusual forms of burial; a prone burial of a young woman in water-logged conditions by the Wiltshire Avon near Lake produced a radiocarbon date of 400–610 cal AD (GU-4921). It was suggested that the unusual location of the burial by the river may have been conditioned by the use of the river as a boundary at the time (McKinley 2003). Single burials, with or without unusual rites, may be common but without scientific dating are unlikely to be assigned to this period.
Another unusual burial was excavated at Filton in 2005 (Cullen et al. forthcoming). A woman had been buried after the removal of both lower legs, the feet being placed below the knees. The bones of the missing right leg were crossed over the upper leg bones but those of the left leg were missing. The grave was of full size and was surrounded, at a slght distance, by clusters of others, some intercutting to suggest that closeness was important. Other graves in the cemetery were arranged in spaced rows. The mutilation of the body has echoes of some late Roman burials but a radiocarbon date of cal AD 555–655 (Wk-17495) suggests that this was one of the later graves in the cemetery which appears to date from cal AD 400–650 (Wk-17495–17498).
A further feature of the western parts of the area is the presence of inscribed stones, known in greater numbers in Wales and Ireland. These are usually believed to be grave markers although few (Lundy with 4 and St Just) have been found in a cemetery (Okasha 1993; C Thomas 1994). The majority may have marked isolated graves and/or functioned as territorial markers but given the poor survival of bone in the area it is unlikely that burials will be located. The identification of stones in association with burials and other archaeological features is clearly of considerable importance to situate and contextualise this important source of evidence; their inscriptions could then be understood in context (Handley 1998; Howlett 1998). An exception to the westerly distribution of these stones is the group from Wareham where five stones have been found in and around Lady St Mary Church (RCHME 1970, 308, 310–12, pl 105, 106; Hinton 1992b; Higgett 2006) The presence of these stones, so far east of the main distribution, contrasts interestingly with the lack of imported pottery from Dorset, as does the re-use of Roman masonry for the carvings with the undressed stones of the west.
The earliest Anglo-Saxon burials appear to lie in, and to the east of, the valley of the Wiltshire Avon. The exception to this is the site at Market Lavington some 10km to the west which may represent the first phase of expansion (Eagles 2001). These cemeteries appear to provide for the burial of small groups occupying dispersed settlements, similar to cemeteries at Roman rural sites but in contrast to the larger cemeteries found adjacent to Late Roman towns and elsewhere further west. A further characteristic of these, and later, cemeteries is their association with earlier burial sites, often barrows, and there is also a tradition of the insertion of secondary burials into existing monuments (Meaney 1964; Williams 1997).
An intriguing site was excavated at Monkton Deverill where a water pipeline revealed a cemetery of unaccompanied oriented burials, one within a ring ditch. Some of the graves had stone lining (dressed stones, presumably from a Roman building) and similar stones were found in the top of the ring ditch. A secondary burial had been placed within the ring ditch; a crouched inhumation with an iron knife of 7th-century date which provided the only dating for the cemetery (Rawlings 1995). Although interpreted as Anglo-Saxon, most of the graves would appear to demonstrate “British” traits (see above) with only the intrusive burial showing any Germanic features. Another site showing features from both cultures is at Portesham (Valentin 2003). Here Durotrigian crouched burials, possibly associated with a circular, non-domestic building were succeeded by oriented burials with two radio-carbon dates spanning cal AD 640–870 (Beta-167358, 167359). At least four burials were in a large pit and one single burial had an iron saw by the foot (although this could be residual). Another burial, however, was accompanied by bird bones – a known “Anglo-Saxon” trait (Lucy 2000, 90–4, 112–13). There was organic-tempered pottery and the Roman pottery included some very late Black Burnished Ware forms. The site was later crossed by a large ditch containing Saxo-Norman pottery in its fills that might have formed the enclosure of the minster that Hall (2000, 19–20, 72–3) has suggested was sited here.
In the 7th century, “Anglo-Saxon” burials are found more widely as the influence of Wessex spread westwards. There is an increase in the wealth of material deposited in the graves which form part of a pattern seen more widely in England than the more locally distinctive burials of the 5th and 6th centuries. There is a continued emphasis on association with, and reuse, of old burial sites, together with the construction of new barrows in some places (Williams 1999). The burials also occupy prominent locations in relation to routes, later hundred boundaries, and situations with all-round views.
This “final-phase” of furnished burials may incorporate the last of pagan burial rites but also, following the conversion, furnished burials continuing in a Christian context. The latter include some exceptionally rich barrow-burials (notably of females) at Roundway Down (Meaney 1964) and Swallowcliffe Down (c.700, Speake 1989). The radiocarbon dating of such burials to ascertain their precise location in relation to the conversion process, and their investigation to understand to what extent they related to the contemporary and evolving Early Medieval landscape, both need to be considered. Some of the outlying furnished burial sites are among the most intriguing and should be investigated further to ascertain the nature of the burial rites employed and whether they are the same as burial sites further east. Some of the Dorset and Somerset sites, for instance, are poorly understood and require further investigation under modern archaeological conditions to address these questions.
There is then a significant gap in our knowledge until churchyard burial becomes the norm, perhaps at minsters by the 8th century and proprietary churches by the the 10th. As discussed by numerous authors, there remains a debate over how burial rites developed in the 8th, 9th and 10th centuries, hindered by the lack of dating evidence earlier provided by the furnished burials. John Blair (1994), Dawn Hadley (2002) and Andrew Reynolds (2002) are among those to suggest that burials continued to be located in the landscape away from churches and sometimes away from contemporary settlements. Prehistoric burial mounds may well have continued to feature as favoured locations (Williams 1997) although increasingly selected mounds took on demonic and dangerous associations (Semple 1998; 2002). The further investigation of placenames for prehistoric monuments attributed mythical or supernatural associations, charter references to execution sites connected to mounds and other earthworks, and evidence for later Anglo-Saxon execution cemeteries (Reynolds 2002), combine to give us a better impression of the sacred and political geography of later Anglo-Saxon England.
Evidence of high-status churchyard burials has been identified at ecclesiastical centres such as Gloucester (Heighway and Bryant 1999) where chests and charcoal burials are elements of a Christian elite repertoire of burial rites. However, our understanding of the development of parish churches in the South West has not developed to compliment the work undertaken at Raunds (Boddington 1996) from the East Midlands. The work undertaken at Shapwick Old Church (Gerrard and Aston forthcoming) may help to redress this balance although few burials were excavated. There are also smaller scale works on churches in use which add small but cumulative pieces of evidence, for example the burials laid on charred planks at Pulham (Claire Pinder pers. comm.). Few, however, of these churchyard sites have been available for large-scale excavation due to their disturbance by later burials and continuing use.
The historical sources give a strong emphasis to the presence of warfare during the period and the nature of defensive structures is such that they often survive. Battlefields, at this early period, are extremely hard to locate but that has not stopped a great deal of effort being spent, often in circular arguments, relating the location of battles to the supposed “frontier” between Britons and Saxons at any particular date. The historical sources become more informative later but that does not always allow more accurate location.
The excavations at Cadbury Castle (Alcock 1995) and Cadbury Congresbury (Rahtz et al. 1992), together with Ian Burrow’s (1981) work have shown that several Somerset hillforts were reoccupied in the 5th and 6th centuries and there are hints that this is also the case in other areas (for example around Bristol, Bob Jones pers. comm.). Poundbury hillfort outside Dorchester may also have been refortified. Radiocarbon dates have recently also been obtained from a ploughed-out hillfort at Raddon suggesting occupation, and possible refurbishment in the 5th–7th centuries (Gent and Quinnell 1999) and there is imported pottery from High Peak near Sidmouth (Pollard 1966). The work at Cadbury Castle involved the construction of about 1100m of timber-laced stone rampart on top of the earlier ramparts, a significant expenditure in terms of stone, timber and man-power. No other site exhibits this degree of refurbishment, although there was new rampart construction at Cadbury Congresbury which divided the original enclosure into two halves, and recent work at Brent Knoll (Papworth 2004) has supported the suggestion made by Ian Burrow that the ramparts were heightened there in the Roman period or later.
The other large construction that may belong to this period is the Wansdyke, whose western section runs south of the River Avon between Bristol and Bath and whose eastern section runs for 15km along the ridge to the south of the River Kennet. It is not clear, however, what the relationship between these two lengths of rampart is, although they appear to be joined by the course of a Roman road. The East Wansdyke has recently been discussed by Fowler (2001) following new fieldwork in Overton and Fyfield parishes. Examination of the earthworks appears to show that it was abandoned whilst under construction, suggesting the lifting of the threat from the north that it was intended to counter. Fowler draws attention to the similarities with the Roman walls in northern Britain, particularly in the provision of numerous gateways (although these are not all proved to be original), and also with the Late Roman (or early post-Roman) Bokerley Dyke in Dorset. He favours a late 5th-century date for construction in the context of Anglo-Saxon expansion from the Thames valley and discounts Reynolds’s suggestion (Reynolds 1999, 85; Reynolds and Langlands in press) that the dyke was a later boundary between Wessex and Mercia, believing that the name implies that its origins had been forgotten before it received an English name. Several recent excavations have shown that the West Wansdyke was carefully constructed to a uniform plan but have failed to produce conclusive dating evidence (Keith Gardner pers. comm.). The extent of the monument to the west continues to arouse discussion with Gardner (1998) reasserting a continuation to Stokeleigh Camp on the basis of fieldwork and medieval documentary evidence.
The southern part of Offa’s Dyke was identified in Gloucestershire by Sir Cyril Fox (1955) but more recent work (Hill and Worthington 2004) has suggested that the earthworks he identified as the Dyke are discontinuous and of unknown date. There are problems with this reinterpretation, such as a 14th-century reference to “Offediche” (Herbert 1996, 249), but the Gloucestershire sections certainly seem to be separated from the dyke further north by a gap of 60km.
Later, during the time of the Viking attacks, the system of burhs was established to provide protected locations. Some of these, such as Wareham, retain impressive earthwork remains but at several the identification of the documented site is uncertain and at others, like Axbridge or Wilton, there is no evidence of the boundary today. Such sites are often classified as urban but work at Wareham appears to show that much of the large interior was not occupied until later in the medieval period. Unfortunately the important excavations at Lydford, by Peter Addyman in the 1960s, remain unpublished despite recent attempts to achieve this. There has been recent work by the Time Team at the Alfredian fort at Athelney (not yet published) which has given some idea of the range of activities there and the finding of a sherd of an imported Mediterranean amphora (Hollinrake in Bagwell and Webster 2005, 171) has supported previous work (the discovery of a 6thcentury bank, Reed 2002) suggesting that the site was important before Alfred. Andrew Reynolds has recently reassessed the evidence from Avebury and, controversially, suggested that a planned town or burh was established there, to the west of the henge (Reynolds 2001a;b) associated with other defensive sites on Silbury Hill and at Yatesbury (Reynolds 2000). There are also other sites, such as Totnes which does not appear in the Burghal Hidage but which is known from archaeology and numismatics (Dyer and Allan 2004a; SW Brown 1999), and Oldaport (Rainbird 1998) which appears to be completely undocumented but which has produced a radiocarbon date of 810–1030 cal AD (NZA-17401) from the mortared stone wall which survives at the site (Rainbird and Druce 2004).
Lab. Ref. | 14C age BP | Cal AD | Site | Context | Reference |
---|---|---|---|---|---|
Beta-167358 | 1320±40 | 640 – 780 | Portesham | Human burial | Valentin (2003) |
Beta-167359 | 1260±40 | 660 – 870 | Portesham | Human burial | Valentin (2003) |
Birm-246 | 1300±80 | 600 – 900 | Brean Down sandcliff | Human burial | Rahtz (1977) |
BM-3162 | 1160±50 | 710 – 990 | Collingbourne Ducis | Sunken floored building 104 | Pine (2001); Ambers and Bowman (2003) |
BM-3163 | 1245±50 | 660 – 890 | Collingbourne Ducis | Sunken floored building 103 | Pine (2001); Ambers and Bowman (2003) |
BM-3164 | 1210±50 | 670 – 950 | Collingbourne Ducis | Sunken floored building 102 | Pine (2001); Ambers and Bowman (2003) |
BM-3165 | 1460±50 | 430 – 670 | Collingbourne Ducis | Sunken floored building 101 | Pine (2001); Ambers and Bowman (2003) |
CAR-1475 | 1475±60 | 430 – 660 | Frocester | Ox skull on floor of building E | Price (2000) |
GU-2710 | 1790±50 | 120–390 | Aller Farm, Devon | Peat | Hatton and Caseldine (1991) |
GU-4921 | 1560±50 | 400 – 610 | Lake | Plank from “grave” | McKinley (2003) |
GU-5149 | 1300±90 | 580 – 970 | Wembdon Hill | Human burial | Bob Croft pers. comm. |
GU-5150 | 1240±70 | 650 – 970 | Wembdon Hill | Human burial | Bob Croft pers. comm. |
GU-5151 | 1060±90 | 770 – 1180 | Wembdon Hill | Human burial | Bob Croft pers. comm. |
GU-5293 | 1450±70 | 430 – 680 | Shepton Mallet | Human burial | Leach and Evans (2001) |
GU-5898 | 940±50 | 1010 – 1220 | Shapwick Church Field | Animal bone | Gerrard and Aston (forthcoming) |
GU-5899 | 1050±50 | 880 – 1160 | Shapwick Church Field | Animal bone | Gerrard and Aston (forthcoming) |
GU-6002 | 1090±50 | 810 – 1030 | Bridgwater Bay | Fish weir 307 | Richard Brunning pers. comm. |
GU-6003 | 1150±50 | 720 – 1000 | Bridgwater Bay | Fish weir 307 | Richard Brunning pers. comm. |
GU-6004 | 1150±60 | 710 – 1020 | Bridgwater Bay | Fish weir 309 | Richard Brunning pers. comm. |
GU-6005 | 1170±50 | 710 – 990 | Bridgwater Bay | Fish weir 309 | Richard Brunning pers. comm. |
GU-6006 | 430±50 | 1400 – 1640 | Bridgwater Bay | Fish weir 202 | Richard Brunning pers. comm. |
GU-6007 | 340±50 | 1450 – 1650 | Bridgwater Bay | Fish weir 202 | Richard Brunning pers. comm. |
GU-6008 | 960±50 | 980 – 1190 | Bridgwater Bay | Fish weir 306 | Richard Brunning pers. comm. |
GU-6009 | 1050±50 | 880 – 1160 | Bridgwater Bay | Fish weir 306 | Richard Brunning pers. comm. |
GU-6010 | 1060±50 | 870 – 1150 | Bridgwater Bay | Fish weir 204 | Richard Brunning pers. comm. |
GU-6011 | 1160±70 | 690 – 1020 | Bridgwater Bay | Fish weir 204 | Richard Brunning pers. comm. |
GU-6038 | 1050±50 | 880 – 1160 | Bridgwater Bay | Fish weir 205 | Richard Brunning pers. comm. |
GU-6039 | 940±50 | 1010 – 1220 | Bridgwater Bay | Fish weir 205 | Richard Brunning pers. comm. |
HAR-2674 | 1090±70 | 770 – 1150 | Taunton Castle | Human burial | Clements (1984) |
HAR-5324 | 1430±70 | 430 – 770 | Queenford Mill (Oxon) | Human burial | Chambers (1987) |
HAR-5325 | 1480±70 | 420 – 660 | Queenford Mill (Oxon) | Human burial | Chambers (1987) |
HAR-5350 | 1550±70 | 380 – 650 | Queenford Mill (Oxon) | Human burial | Chambers (1987) |
HAR-5351 | 1550±80 | 340 – 650 | Queenford Mill (Oxon) | Human burial | Chambers (1987) |
HAR-6216 | 1430±80 | 420 – 770 | Foxley | Charcoal from wall trench | Hinchliffe (1986) |
HAR-8082 | 1220±70 | 660 – 970 | Foxley | Charcoal from hall post-hole | Radiocarbon, 32 (1990), 189–90 |
HAR-8548 | 1550±80 | 340 – 650 | Brean Down sandcliff | Human burial | Bell (1990) |
HAR-8549 | 1430±70 | 430 – 770 | Brean Down sandcliff | Human burial | Bell (1990) |
NZA-17401 | 1098±45 | 810 – 1030 | Oldaport | Hazel charcoal from wall mortar | Rainbird and Druce (2004) |
OxA-8297 | 1380±35 | 590 – 690 | Tolpuddle Ball | Human burial | Hearne and Birbeck (1999) |
OxA-8298 | 1440±35 | 560 – 660 | Tolpuddle Ball | Human burial | Hearne and Birbeck (1999) |
OxA-8299 | 1660±35 | 250 – 540 | Tolpuddle Ball | Human burial | Hearne and Birbeck (1999) |
OxA-8300 | 1450±30 | 560 – 655 | Tolpuddle Ball | Human burial | Hearne and Birbeck (1999) |
OxA-8320 | 1470±35 | 530 – 650 | Tolpuddle Ball | Human burial | Hearne and Birbeck (1999) |
OxA-11461 | 1156±32 | 770 – 980 | Shapwick Church Field | Charcoal from building posthole | Gerrard and Aston (forthcoming) |
OxA-11474 | 1251±32 | 670 – 870 | Shapwick Church Field | Charcoal from building posthole | Gerrard and Aston (forthcoming) |
OxA-11475 | 1277±31 | 660 – 810 | Shapwick Church Field | Duplicate of OxA-11474 | Gerrard and Aston (forthcoming) |
OxA-11873 | 1189±30 | 720 – 950 | Shapwick Church Field | Charcoal from building posthole | Gerrard and Aston (forthcoming) |
OxA-11874 | 1196±30 | 710 – 940 | Shapwick Church Field | Charcoal from building posthole | Gerrard and Aston (forthcoming) |
OxA-11930 | 1277±27 | 660 – 780 | Shapwick Church Field | Charcoal from building posthole | Gerrard and Aston (forthcoming) |
OxA-11931 | 1301±26 | 660 – 780 | Shapwick Church Field | Charcoal from building posthole | Gerrard and Aston (forthcoming) |
OxA-11932 | 880±24 | 1040 – 1220 | Shapwick Church Field | Charcoal from building posthole | Gerrard and Aston (forthcoming) |
OxA-11933 | 942±25 | 1020 – 1160 | Shapwick Church Field | Charcoal from building posthole | Gerrard and Aston (forthcoming) |
SUERC-2937 | 1025±35 | 890 – 1150 | Shapwick Old Church | Human burial | Gerrard and Aston (forthcoming) |
SUERC-2938 | 1510±35 | 430 – 640 | Sladwick, Shapwick | Bone | Gerrard and Aston (forthcoming) |
Wk-13086 | 1552±45 | 410 – 610 | Wembury Bay | Burnt pit fill | Reed (2005) |
Wk-13087 | 1635±53 | 250 – 550 | Wembury Bay | Burnt gully fill | Reed (2005) |
Wk-13088 | 1510±44 | 430 – 640 | Wembury Bay | Burnt gully fill | Reed (2005) |
Wk-17495 | 1451±32 | 555 – 655 | Filton | Human burial | Cullen et al. (forthcoming) |
Wk-17496 | 1491±30 | 460 – 650 | Filton | Human burial | Cullen et al. (forthcoming) |
Wk-17497 | 1515±32 | 430 – 620 | Filton | Human burial | Cullen et al. (forthcoming) |
Wk-17498 | 1571±31 | 410 – 560 | Filton | Human burial | Cullen et al. (forthcoming) |
We are grateful to those who provided comments on earlier drafts and to Martin Watts of Cotswold Archaeology for information about the Filton site in advance of publication.