An Early Anglo-Saxon Cemetery and Archaeological Survey at Breamore, Hampshire, 1999–2006

A Byzantine pail, datable to the sixth century AD, was discovered in 1999, in a field near the River Avon in Breamore, Hampshire. Subsequent fieldwork confirmed the presence there of an early Anglo-Saxon cemetery. In 2001, limited excavation located graves that were unusual, both for their accompanying goods and for the number of double and triple burials. This evidence suggests that Breamore was the location of a well-supplied ‘frontier’ community which may have had a relatively brief existence during the sixth century. It seems likely to have had strong connections with the Isle of Wight and Kent to the south and south-east, rather than with communities up-river to the north and north-east.


INTRODUCTION
In 1999, a metal-detectorist, Steve Bolger, showed various objects found in a field at The Shallows, Breamore, Hampshire (SU 1615 1730; Illus. 1), to one of us, Sally Worrell, then the Hampshire Finds Liaison Officer for the Portable Antiquities Scheme. They included a sixth-century Byzantine/East Mediterranean engraved copper-alloy pail, and suggested that the site was an important Anglo-Saxon cemetery. Geophysics were initially undertaken by Kate Clark, University of Southampton, and then by English Heritage. Test-pits excavated by Berkshire Archaeological Services and directed by Roy Entwhistle subsequently confirmed that human burial had taken place. Sally Worrell brought the site to the attention of Videotext Communications, producers of a television programme called Time Team, which undertook the excavation of several small trenches in late August 2001, and subsequently paid for artefact conservation and the production of a report collated by Kate Edwards for Wessex Archaeology, which was submitted to Hampshire County Council in 2005.
Subsequent fieldwork undertaken in Breamore during 2001 to 2006 was co-directed by David Hinton and Christopher Loveluck, and with geophysical survey by Kristian illus. 1 Location maps: modern county boundary shown medieval as well; much of the gravel and lower chalk slope was cultivated open field (Light et al. 1995, 90-5). The principal road, now the A338, runs north-south, generally clinging to the edge of the gravel terrace above the flood-plain, where the ground is flattest. Various roads and lanes lead westwards up to the chalk, but only one now goes eastwards, leading towards Shallows Farm, and crossing the Avon over a bridge alongside a mill (Illus. 2). Another route is now only a footpath, crossing the river at South Charford. The 'Cloven Way', a prehistoric/Roman trackway, crossed the river at Charford (Crawford, O. G. S. 1931).
Domesday Book records that Breamore was a royal estate in the late eleventh century, and this royal connection presumably accounts for the special nature of the church, which is not precisely dated but is generally agreed to be late Anglo-Saxon, illus. 2 Site plan and location of Berkshire Archaeological Services test-pits breamore early anglo-saxon cemetery with much of the fabric surviving. It is larger than a typical estate church, has vestiges of carvings of the highest quality (Rodwell and Rouse 1984;Tweddle, Biddle and Kjølbye-Biddle 1995, 251-6), and an inscriptionperhaps originally more than one which has a disputed meaning but seems to be a reference to the Genesis rainbow (Withers 1999, 121-4). The church may have been a 'minster' serving the north-west of Hampshire, bounded by Fordingbridge to the south, beyond which lay the well documented parochia of Twynham/Christchurch (Hase 1988, 47). The parish boundary, both now and in the nineteenth century, was also the boundary between Hampshire and Wiltshire, with the large parish of Downton, the Bishop of Winchester's estate in the Middle Ages, on the other side of the shire boundary. Domesday links Breamore with Rockbourne, and an origin of this royal estate as the territory of the big Roman villa there, taking in also the intervening Whitsbury, can be considered (Hinton 2013, 159-60); it is outside the area in which Eagles (2015) has recently argued that 'small shires' can be identified, but some sort of post-Roman sub-region is possible.
Ownership of the Breamore estate passed from the king to the de Redvers family in c. 1100. Soon afterwards, this family established an Augustinian priory in the river valley, where a low rise was presumably sufficient to preserve the buildings from flood-water (Hinton and Worrell 2016, S06). The priests of the minster probably became the founding members of the new house, as happened on a bigger scale, though without a change of site, at Twynham/Christchurch (Hase 1988, 49-58). The transfer meant that the existing church at Breamore remained as the parish church, and so was not subject to the alterations required over the next four hundred years by the canons at Twynham/Christchurch. It was presumably large enough for its secular congregation, despite general population growth, so that most of its Anglo-Saxon fabric survived intact.

BREAMORE IN THE EARLY MEDIEVAL PERIOD
Fieldwalking by the Avon Valley Archaeological Society in the 1980s located five Iron Age/Romano-British sites within Breamore parish. One, in fields bordering the west side of the main road, on the gravel terrace north of the marsh and south of a lane called North Street, produced several sherds of organic-tempered pottery that indicated usage, quite possibly continuously, into the early medieval period. A second, to the north, yielded a single such post-Roman sherd. Pottery of this sort was found at two other sites, without enough Roman material to suggest settlement in that period; one of them, at South Charford Farm, yielded a brooch probably of sixth-century date (Light et al. 1995, 81-4). Organic-tempered pottery is now seen as a fifth-century introduction; it used to be thought that it was brought in by Germanic immigrants, but has now been found at such sites as Frocester, Gloucestershire (Timby 2000, 137-8) in areas that were not influenced by them until later. The South Charford brooch, however, is a small-long type that would be regarded as Anglo-Saxon because of its date and distribution across southern and eastern England (Lucy 2000, 31-3, 134).
Two cemeteries have been investigated in the Avon Valley below Salisbury. About 7 km upstream of Breamore, at Charlton Plantation, is an Anglo-Saxon cemetery with material datable to the later fifth and sixth centuries (Davies 1984;Illus. 1). Downstream, however, the Bargates cemetery, at the mouth of the river outside Christchurch, probably did not begin before the later sixth century (Jarvis 1983, 127-8). Recent investigations to the south of Fordingbridge at Ibsley revealed traces of a small settlement, with a vestigial sunken-featured building, and a pit which has been radiocarbon dated to cal. AD 561-648 (Coles and Ford 2016, 30 and 80-2). A similar site was found two miles away at Hucklesbrook (Davies and Graham 1984).
Early Anglo-Saxon finds in Breamore parish, other than at The Shallows, include a copper-alloy mid-fifth/mid-sixth-century bridle-fitting with red glass insets, probably made in the west or central Mediterranean, though likely to have been channelled through the Franks, to whose culture is also attributed the making of a belt-mount inset with red garnets (Eagles and Ager 2004). These and other objects may indicate that at least one other cemetery exists in Breamore in addition to the one probably in South Charford, but none was located by geophysics or test-pitting.

THE SHALLOWS
The field in which the Byzantine/East Mediterranean pail (Br 1 below) was found is close to the River Avon, on a gravel promontory that is just high enough above waterlevel to avoid flooding, and from which surface water drains quickly. The pail had come from the western part of what was, at the time of discovery, a pasture field, in the area of a low oval mound, now measuring some 70 m by 60 m, which could still be observed despite much ploughing, rising to about 0.7 m above the surrounding land (Illus. 2). Other finds from the field shown to Sally Worrell by Steve Bolger included a coin-weight and fragments of two Anglo-Saxon brooches (Br 2-5 below). Unfortunately, disturbances in the field suggested activity by other detectorists, and some material may have been removed illicitly; nothing has subsequently appeared on the antiquities market that could certainly be attributed to this site, however.
Seven one-metre square test-pits were excavated in November 2000 by Berkshire Archaeological Services (Entwhistle 2001). The English Heritage geophysical survey had indicated a possible ditch on the south-west side of the mound, but the test-pits (Illus. 2, nos 4 and 6) sited to investigate this came straight down to underlying gravel and flint. Other test-pits showed that the mound was constructed largely of scraped-up topsoil, suggesting that no ditch had existed. This absence would explain why air-photography had not revealed it. Fourteen sherds of early Bronze Age Collared Urn found in Testpit 3, and other small pieces of pottery and flint, indicate that the mound was Bronze Age in date. Similarly ditchless mounds are known elsewhere in the area (Ashbee 1960;Ellison 1980), and the context record sheets from the subsequent excavation indicate that the mound material was siltier and contained less flint than the underlying layers; as upcast from a ditch would have contained more flint, this helps to confirm the absence of any such feature.
One test-pit (no. 2) was dug where Steve Bolger reported finding the pail; no other artefacts were found, but some small fragments of copper alloy may have derived from the pail, and some human bone confirmed the presence of a cemetery, although no grave was observed cut into the lower contexts. In none of the test-pits was it possible to see the edges of any features.
A test-pit to the north (no. 7), separated by a 0.5 m baulk from no. 2, located the badly preserved lower leg of a skeleton, probably supine and laid out south-north; a shield-boss (Sf 6), had been placed above it, with iron studs and a shield fitting (Sf 7-10). The rest of the skeleton was not revealed (subsequent excavation suggests that Test-pit 2 and Test-pit 7 located the same skeleton, which was given the number 128: see Trench 1 report, below). In view of the poor condition of the bones, it was decided to leave all in situ bodies unexcavated.
Test-pit 5 was dug 5 m to the north-west of 7. It revealed at least two supine inhumations laid out approximately south-south-east/north-north-west, and very close together. Like those in Test-pit 7, these skeletons were left in place (and were subsequently numbered as Skeletons 304 and 305: see Trench 3, below). The western of the two (later 304) had a shield-boss (Sf 1) over its chest, with a stud (Sf 2) about 0.2 m away; a knife (Sf 5), was close to its left elbow. The other had a spear-head (Sf 3) by the skull and a possible knife (Sf 4) by the pelvis. About 0.3 m to the north-west, right in the corner of the test-pit, was a third skull, overlying a lower right leg that suggested an inhumation aligned south-south-west/north-north-east; these could have been much disturbed remains of the same skeleton, or the skull could have come from a separate body. There were no artefacts in the very small area investigated around this skeletal material (which was not subsequently recorded by Time Team, and is dubbed Skeleton1/2 in this article).
Although only 5 m west of Test-pit 5, no burials or artefacts of Anglo-Saxon date were found in Test-pit 3. Likewise, away from the mound, nothing of Anglo-Saxon date was found in Test-pit 1. Furthermore, no human bone fragments were found in Test-pit 3's upper levels, although not only were there some in Test-pits 2, 5 and 7, near the centre of the mound, but also in 4 and 6, around its edges.
Drawing this evidence together, the test-pits confirmed the presence of a cemetery focusing upon the mound, and the condition of the bone suggested that even the more deeply buried skeletons were now close to the surface and eroding from the soil. The spread of bone through the upper contexts, and damage seen on the skull of Skeleton 1/2, indicated that considerable plough disturbance had already occurred, and would continue. The involvement of Videotext Communications allowed for the further excavation that was clearly desirable.
Geophysical survey always featured in Time Team programmes, and gradiometry was undetaken in the Shallows Farm field by GSB Prospection. Their results were consistent with the earlier work, in particular getting no indication that the mound had been surrounded by a ring-ditch (subsequent excavation did not locate a ditch, but the trenches were not placed where they might have revealed one). A 20-m diameter ringditch was outlined in the north-east part of the field, however, and traces of what may be others were noted (Hinton and Worrell 2016, S04, in the online supplemental data, http://dx.doi.org/10.1080/00665983.2017.1235151). Topographical survey was undertaken, and this also helped Time Team to decide where to locate its trenches. The programme involved filming archaeologists at work over a three-day period, in the hope of producing enough interesting results for 50 minutes of transmission each evening (shown as Series 9, Episode 13). A fourth, unfilmed, day was used for recording and back-filling, but in other respects the 'rules' were adhered to; despite the exigencies of the need to produce interesting programmes, the professionalism of the archaeologists involved ensured that the practical work of recovery was done to high standards, albeit that the restrictions of small trenches probably meant that less information was gained than might have been from open-area excavation. Context recording was maintained, though sometimes a little sketchily. Few 'still' photographs seem to have been taken.
At the time of the excavation, the adjacent field to the south was under crop and not available for investigation. Geophysical survey in it was subsequently undertaken by Kristian Strutt, University of Southampton (Hinton and Worrell 2016, S04, Figs 23-7), and a metal-detector group spent one Sunday prospecting in it, observed by Chris Gifford and David Hinton, but no discoveries were made.
After the excavation, Time Team funded conservation work, and the production of a grey literature report collated by Kate Edwards, which was submitted to Hampshire County Council in 2005. No further excavation has been possible, and (up to March 2015) no other finds have been recorded by the Portable Antiquities Scheme as coming from the cemetery. The reports that follow are mostly abbreviated from those submitted in 2005, with revisions and updates; all the contributors have had the opportunity to approve any changes. The reports on Br 1-Br 6 were written in 2014-6.
EXCAVATION BY TIME TEAM, 28-31 AUGUST, 2001 by KATE EDWARDS Five trenches were opened on the mound, and four others at various points in the field. Their locations were determined partly by the test-pit discoveries, and partly by geophysical results. (A summary table of the contexts is in Hinton and Worrell 2016, S01.) the mound (trenches 1, 3 and 5-7 ) All but one of the five trenches dug on the mound located inhumed skeletons, the exception being Trench 6 (Illus. 3). No trenches were excavated in the mound's centre, presumably because anomalies did not show up in the geophysics. Trenches 1, 3 and 6 were situated in the barrow's north-east quadrant, while 5 and 7 were located on its south side, nearer the mound's edge.
Graves had been dug into the make-up of the mound, their fills being hard sand and gravel silts, much compacted, and the excavators were not able to differentiate them from the undisturbed material around them; consequently, no grave dimensions are known. No graves were observed definitely to have penetrated to layers underlying the mound; in Trench 4 the natural geology was closely packed flints within compact silt. As the mound's height at the time that the burials were inserted is not known, the original depths of the graves cannot be established; those on the edges of the mound, where it would have been quite low, went further below OD than those in the centre, but a double burial in Trench 7 was only some 0.7 m lower at its base than those in Trench 1, so their relative depths were probably very little different.
Bone preservation was generally very poor; estimates of ages, heights and biological sex where determinable are taken from Roberts (this article).

Trench 1 (Illus. 4)
This 4 m x 3 m north-south trench involved reopening Test-pits 2 and 7, which had been positioned where the pail had reportedly been found; an in situ human skeleton had been revealed but not excavated in Test-pit 7, and some human bone fragments in Test-pit 2.

SKELETON 121
Aligned approximately south−north (i.e. head at south end), less than a quarter of an extended, supine individual survived, the head facing east. The individual was aged between 17 and 25 years. Associated were two shield-bosses: Sf 1011 over the upper left leg, and Sf 1012 over the knees. A copper-alloy buckle (Sf 1010) was in the waist area, a wooden bucket (Sf 1007) on the west side of the head, and a spear-head (Sf 1006) much closer to it on the east side. The spear-head's point was found directed southwards, indicating a wooden shaft of which there was no trace parallel to the person's side. SKELETON 128 This skeleton is thought to be the one revealed in Test-pit 7, with a shield-boss and other ironwork (Sf 6-10).
Aligned south-west−north-east, the skeleton is recorded as supine, with crossed legs. So little bone survived that even its approximate age could not be determined. Two buckets (Sf 1008 and Sf 1013) were located around where the head should have been, in illus. 3 Time Team trench plan one of which (Sf 1008) an intact glass bowl (Sf 1008ii) was found subsequently during conservation. Three unstratified iron items, recorded as found by metal-detector users and presumably from the spoil-heaps, could have accompanied this burial: a spear-head (Sf 1002), a ferrule (Sf 1001) and a knife (Sf 1003).
Interpretation of this burial is inevitably difficult, as it had already been investigated in test-pits, and the absence of bone at the expected depth of a burial in Test-pit 2 illus. 4 Trench 1 plan breamore early anglo-saxon cemetery had been explained as an anomaly, possibly resulting from a crouched burial (Entwhistle 2001, 10). The positions of Sf 1008 and Sf 1013 indicate a normal fulllength burial, however, so it seems probable that in the cramped conditions of the test-pit, it had not been seen that the much-eroded upper part of the skeleton had in fact already been accidentally disturbed by detectorists. The Time Team photographs show that Test-pit 7 (not shown on their plan, Illus. 4, unlike Test-pit 2) had come down on to the legs. FEATURE 120 An east−west linear cut, F 120, filled with dark brown sand, contained a bucket (Sf 1009); a single long-bone was found outside the cut, but within it was no evidence of a skeleton, perhaps because of bone deterioration. The feature was probably a grave, the rest of which was outside the excavation area. It was the only cut feature other than post-holes observed during excavation. POSSIBLE POST-HOLES Several features, F 110-2, 116 etc., interpreted as holes for stakes or posts were halfsectioned. Being filled with dark brown, soft sandy material, they were observable, unlike grave cuts. They have no clear pattern, though a circle is possible. They were recorded as being later than the burials. FEATURE 123 An east−west feature labelled as a possible grave cut is shown on the site plan, but was not recorded in the context record sheets, and a note on the plan states that it was not excavated.

Trench 3 (Illus 5-8)
This trench, also 4 m x 3 m, included the location of Test-pit 5, in which two inhumations had been discovered; the skeletons had been left in situ, but ironwork (Sf 1-5) had been recovered. There is no mention in the Time Team records of the rediscovery of the third skull and the long bone reported in Test-pit 5.
The two inhumations were excavated as Skeletons 304 and 305. They were laid south−north alongside each other (Illus 5 and 7). No edges of a grave cut were seen, but the bodies must have been in a single grave, not two separate ones, as the left arm of 304 had fallen or been placed over the right side of 305, and their heads were at the same depth. No further grave-goods were found with either. The south-east corner of the trench was extended northwards and eastwards to complete the excavation of another inhumation, Skeleton 314 (Illus 6 and 8).

SKELETON 304
This was supine and was a child of between 9½ and 14½ years: too young for a sex determination, even if the bone had survived in a much better condition than it did. Less than a quarter of the skeleton could be recovered.

SKELETON 305
This skeleton was lying slightly on its left side, the head facing east. The age was between 17 and 25 years, but the sex could not be determined because of the condition of the bones.
Because of the circumstances, attribution of the ironwork specifically to one or other of these two burials is not possible. The test-pit record would seem to place Sf 1, a illus. 5 Trench 3 plan: Skeletons 304-6 (see also Illus 6 and 7) shield-boss, Sf 2, a rivet, and Sf 5, a knife, with the more westerly skeleton (i.e. the child, 304), although for someone aged less than fifteen to have been buried with a shield would be very unusual: Dickinson and Härke (1992, 68) recorded only one example in the 8 to 14 years category. Sf 3, a spear-head, and Sf 4, probably a knife, were thought to be with the skeleton later numbered 305 (Entwhistle 2001, 9). It seems more likely that the shield had been included because of the juvenile, 305, but had been placed to overlap both bodies. The test-pit report says that a metal-detector indicated other material, which was not lifted because to have done so would have involved disturbing the bones. Those bones were undisturbed when Time Team found them, which would imply that the detector had been tracing residues of the removed objects. SKELETON

306
Only parts of the legs and cranium survived of a child aged between 6 and 10 years to the south of the double burial. Like them, it was supine and aligned south−north; the head seemed to be facing slightly west. The distance from the legs of Skeletons 304 and 305 implies a separate grave. A large rounded pebble (Sf 3001) was found on top of the back of the jaw; this was photographed in situ, but the plan was drawn up after most of the head had been removed. The photograph suggests that the mouth was closed, in which case the pebble had been placed above the head, or had come to rest there by accident. The latter is less likely, as no naturally occurring stones like it were recorded elsewhere.
illus. 6 Trench 3 plan: Skeleton 314 (see also Illus. 8) SKELETON 314 (ILLUS 6 AND 8) This was supine, aligned east−west, the head facing slightly south. An iron shield-boss (Sf 3003) was above its head, and an iron spear-head (Sf 3004) was to its right, the point directed westwards, but again with no trace of a shaft or ferrule. An iron fragment (Sf 3006) might have been the remains of a knife-blade. The body was of someone aged between 17 and 25 years.
POST-HOLES Irregular holes were found penetrating into the ground below the mound. They were filled with red/brown fine silt like those in Trench 1, though the two trenches together do not give them greater coherence. They became visible at higher levels than the skeletons, and are not thought to have had anything to do with the cemetery.
Also recorded were ten patches underlying the levels at which Skeletons 304 and 305 lay, which were initially thought to be post-holes, but the excavators finally judged that they were all probably natural, caused by varying ground conditions.
Trench 5 (Illus 9-10) On the south-west side of the mound, this trench contained two inhumations, aligned approximately west−east, lying on their right sides. As in Trench 3, the bodies were very close together, including their flexed legs, which suggests that they must have been in the illus. 7 Time Team photograph of Trench 3 double burial, Skeletons 304 and 305, from the north same grave. The context record sheets do not state whether this trench was fully excavated, but the photographs suggest that there was empty space around the skeletons.

SKELETON 505
Up to three-quarters of the skeleton survived of an adult male, aged 35-45 years. An iron spear-head (Sf 5015) was by the right side of his head, an iron shield-boss (Sf 5012) over his left shoulder, and an iron knife-blade (Sf 5014) probably belonged with him.

SKELETON 506
Only about a quarter of this skeleton survived, so although an age of between 25 and 35 years could be suggested, sex and height could not be. An iron spear-head (Sf 5007) was behind the head. Above and between the knees was a small copper-alloy buckle (Sf 5013) and in the crook of the knees was a bucket (Sf 5011). Above the left leg was a shield-boss (Sf 5003); Sf 5004, Sf 5005, Sf 5016, Sf 5017 (not shown on plan) and Sf 5018 are probably studs from the edge of the shield-board. A copper-alloy buckle with insets (Sf 5010) was by the left hip.
illus. 8 Time Team photograph of Trench 3 Skeleton 314, from the east. Shield-boss Sf 3003 is shown overlying the head Described as unstratified but attributed to this trench are a spear-head (Sf 5001) and a knife (Sf 5002). Some Romano-British pottery, one Bronze Age and one late Iron Age/ early Romano-British sherd were also found.

Trench 6
This was the most northerly and also the smallest of the mound trenches, excavated because the geophysics gave a strong reading, which turned out to be because of iron-panning, not an artefact. No skeletal material was located either. Several worked flints in better condition than those found in the off-mound trenches were recovered (Harding, this article).
Below the topsoil was a 0.3 m thick layer of compact silt and gravel, below which the iron-panning was between 2 and 3 mm, on top of the old ground surface, here soft grey-brown silt with quantities of worked flint. The subsoil was also grey-brown silt, but with only a few pebbles.
Trench 7 (Illus. 11) The final trench on the mound was on the lower part of its south-south-east side; it was another that had to be extended to take in a complete burial. This trench contained a triple burialtwo adults and a very young child. No grave cut was seen; layer 706, outlined on the plan, was an underlying patch of sandy brown silt.

SKELETON 703
More than three-quarters of a supine, 25-35-year-old male survived, aligned approximately north−south, the head facing east; the arms were folded across the waist, with the left underneath the right. An iron spear-head (Sf 7005) was above the right shoulder (ferrule Sf 7010, below, could have been from the other end of the shaft), a shield-boss (Sf 7002) was just above the left knee, and a bucket (Sf 7011) consisting only of wooden staves was between the feet. Other items could not be so positively associated with one or other skeleton, but an iron lozenge-shaped plate (Sf 7006) was probably from the shield-board centred by boss Sf 7002, although very close to the leg of skeleton 704.

SKELETON 704
Equally well preserved was the skeleton of another male aged between 25 and 35 years. Also supine and aligned north−south, but with the head facing west, this body must illus. 10 Time Team photograph of Trench 5 double burial, Skeletons 505 and 506 (nearer camera), from the south. Bucket Sf 5011 is shown partly excavated prior to lifting en bloc; behind it is buckle Sf 5013; the garnet-set buckle Sf 5010 is by the hip of Skeleton 506 have been placed in the grave first, as the fingers of 703's right arm are recorded as slightly overlapping 704's right radius (this is not clear from the plan). A knife (Sf 7008) and a pair of copper-alloy tweezers (Sf 7009) are shown on the site plan as having been found below the lower left arm, but were not recorded on the context record sheets. A illus. 11 Trench 7 plan breamore early anglo-saxon cemetery spear-head (Sf 7001) is said in the site finds inventory to have been to the left of the pelvis, with an iron ferrule (Sf 7010) to the west of the ankles, but the ferrule looks as likely to have belonged with 703's spear-head, Sf 7005; the length would be appropriate, whereas if Sf 7001 really was beside 704's pelvis, the spear would have had a very short shaft if completeunless it had been broken. The head's length precludes interpretation as an arrow.

SKELETON 705
Fragments of bone above the legs of 703 and 704 were those of a very young child, aged c. 3 years. Recorded as overlying those bones were a shield-boss and grip (Sf 7007).
trenches off the mound (trenches 2, 4 and 8-9 ) Test-pit 1 (Illus. 2) and four Time Team trenches (Illus. 3) were excavated at points in the field where geophysics revealed anomalies. No further burials, or other traces of Anglo-Saxon usage, were located, but that does not mean that none remain to be found, especially to the south of the mound.

Trench 2
A 44 m-long trench was cut using a machine, running partly down a steep slope towards the river, and locating the edge of one of its channels. Further up the slope the trench bisected two shallow ditches (F 202, F 204); seventeenth-to nineteenthcentury pottery, made locally at Verwood, was found in the former. Burnt flint was recovered from the river's edge, of the sort usually associated with prehistoric cooking, especially Bronze Age burnt mounds. Unworked wood of unknown species included a knobbly piece characteristic of coppicing; worked oak chips were also found, with facets indicating that a very sharp metal axe, more probably made of iron than of copper alloy, had been used on them. Active management of the river bank is indicated. (This trench was supervised by Francis Pryor, and the wood was scrutinized by Maisie Taylor.) Trench 4 This trench located four possible post-holes, three in an irregular line, the fourth containing mortar from a lath and plaster wall. Although the possibility of a building was entertained, the three in-line post-holes do not look much like the fourth, and the mortar in it need be no more than rubble packing for a stake, not an in situ structure. Fencing seems a likely alternative.

Trench 8
To the north of the mound, two anomalies suggested ditches, which were duly found, 4 m apart. Although their contents were sieved, no dating evidence came from them; they could have been either side of a ploughed-out trackway, unrelated to the mound. Tony Light has noted that a track is shown in this position on nineteenth-century maps.

Trench 9
This was excavated to investigate whether features revealed by geophysics were the remains of a barrow. A curvilinear ditch was located, but not a corresponding arc to its north that would have indicated a barrow or roundhouse. The fill included a piece of Bronze Age pottery, but also a wheel-thrown medieval sherd. A complex in the northeast part of the trench was thought to be natural, apart from a possible pit.

HUMAN REMAINS
by ALICE ROBERTS Some on-site investigation of the bones was undertaken during the Time Team excavation, but they were washed, examined, analysed and recorded subsequently in the Department of Anatomy, University of Bristol. The material was aged and sexed, insofar as its condition permitted, according to standard anthropological techniques (Brothwell 1981;van Beek 1983;Hillson 1996;Mays 1998) (Table 1). Biological sex was determined mainly on pelvic and cranial traits; stature estimates were calculated from long-bone measurements using the regression formulae developed by Trotter (1970). Only the two adult skeletons in Trench 7 survived well enough for metric traits to have been identified, had they been present.
skeleton 121 Tooth attrition indicated a juvenile aged between 17 and 25 years; enamel hypoplasia indicated a period of physiological stress in early childhood. As less than 25% of the skeleton survived, sex and stature could not be determined, and no pathology could be observed on the bones. skeleton 128 Less than 25% of this skeleton was preserved. Age, sex and stature could not be determined, and no pathology was observed.
skeleton 304 Tooth eruption and uncalcified third molar crown indicated a child of 12 years, +/-30 months. Less than 25% of this skeleton was preserved. Sex and stature could not be determined, and no pathology was observed.
skeleton 305 Tooth attrition indicated a juvenile aged between 17 and 25 years. As less than 25% of the skeleton survived, sex and stature could not be determined, and no pathology could be observed. breamore early anglo-saxon cemetery skeleton 306 Tooth eruption and wear indicated a child of 8 years, +/-24 months; enamel hypoplasia indicated a period of physiological stress in early childhood. Less than 25% of this skeleton was preserved. Sex and stature could not be determined, and no pathology was observed on the bones. skeleton 314 Tooth attrition indicated a juvenile aged between 17 and 25 years. The skull exhibited a metopic suture, present in 9% of modern populations (Williams, P. L. 1995). As less than 25% of the skeleton survived, sex and stature could not be determined, and no pathology could be observed on the bones. skeleton 505 Male, aged 35-45 years based on tooth attrition. There was one carious cavity in a molar, and another molar had been lost before death. The skull was thickened, but radiological examination excluded Paget's disease. 50-75% of this skeleton was preserved, but stature was not determined, and no pathology was recorded on the bones. skeleton 506 Tooth attrition indicated a young adult aged between 25 and 35 years. As less than 25% of the skeleton survived, sex and stature could not be determined, and no pathology could be observed.

skeleton 703
Male, aged 25-35 years based on tooth attrition. There was widespread calculus on the teeth, and an asymmetric wear pattern on the molars. There was a hole in the cranium, which may have been caused by an intracranial hydatid cyst. Significant expansion of the left tibia was confirmed as Paget's disease by radiological examination. More than 75% of this skeleton was preserved, but stature was not determined; there were no nonmetric traits.
skeleton 704 Male, aged 25-35 years, probably in the lower end of the range as the sacral bodies had not fused. More than 75% of this skeleton was preserved; stature was determined from the femur as about 1.785 m (5 ft 10¼ ins). There were no non-metric traits or pathology.
skeleton 705 Tooth eruption indicated a child of 3 years, +/-12 months. Less than 25% of this skeleton was preserved. Sex and stature could not be determined, and no non-metric traits, pathology or bony changes were observed on the bones.

ANGLO-SAXON GRAVE-GOODS AND OTHER ARTEFACTS
by SALLY WORRELL Most of the objects described in this section came from the inhumations described above, and have Sf numbers. A few others were found by metal-detection, and have Br numbers for ease of reference in this article. No Anglo-Saxon pottery was found in the trenches on the mound, and pottery of other periods from it has been included in the summary by Duncan Brown which follows. Objects in any grave are subject to movement after burial, as the body decomposes and the weight of the soil compacts the grave's contents; burrowing animals can transport objects over some distance; and the discovery of frog bones in some of the buckets is an indicator that, in dry weather, the soil at The Shallows cracked and created fissures. All these factors mean that objects in the graves may have been found in positions rather different from those in which they were originally placed. In particular, this means that, in the case of the double burials, a few objects cannot be definitively assigned to one or other body. This report was completed in 2004, but has been updated; the notes by the named specialists were written or revised for publication in 2014-5.

containers
The cemetery was remarkable for the number of containers found: the pail discovered by Steve Bolger, six small stave-built buckets, and a glass bowl contained inside one of them.  A skull recorded in Test-pit 5 was not located in Trench 3. The test-pit report also refers to a pelvis and lower right leg, possibly a different body. These are shown on the archive drawing in the corner of the test-pit and have been dubbed 'Skeleton 1/2' for present purposes, but are not included in the table above.
An incomplete pail (the term 'bucket' is confined in this report to the stave-built vessels discussed below, to avoid confusion) with nearly vertical sides and a swing handle has a surviving but detached base. The pail is made from hammered copper-alloy sheet with traces of white-metal coating on the outside; there are two semi-circular raised lugs on the rim; the swing handle has long, out-turned 'swan's neck' ends.
illus. 13 Byzantine pail Br 1, photogrammetry image James Miles. The complete photogrammetric record, which can be rotated, is available through Hinton and Worrell 2016, S02 Below the rim is an inscription in Greek uncial letters, rendered with annular punches. The beginning and end of the text are marked by a circular motif beneath one of the lugs. It has punched border lines. Beneath the inscription is a punched and incised frieze where three male figures, bushy-haired, two of them booted (the third is broken off below the knees), armed and wearing a chlamys (short cloak) but otherwise naked, confront two savage beasts. The background is decorated with sprigs, bushes and rosettes. The figure in the centre of the drawing, with shield thrust forward with his left arm and his right arm outstretched behind him and holding a discus, and the figure to his left, holding a long spear and (probably) a shield, now worn away, engage with a quadruped adversary. The animal's body is decorated with circles of punched dots, but its species cannot be identified; perhaps a bear or hyena is intended. The third figure, with a shield extended in front of him and a sword in his right hand behind, also challenges a fierce feline, with a handsome scrolling tail; its spots suggest identification as a leopard. As the tops of the figures' heads overrun the inscription, the latter was executed first.
The metal of the Breamore pail has not been analysed, unlike some other examples of this type of vessel which have been found to be brasses with high copper and zinc ratios, sufficiently hard to take finely detailed working, allowing for the friezes and inscriptions.
Although incomplete, with part of the lower body and base detached, the Breamore pail represents a significant piece of Byzantine art whose historical importance is illus. 14 Byzantine pail Br 1, section illus. 15 Byzantine pail Br 1, frieze and Greek inscription enhanced by its pathway to deposition in Hampshire. It fits into a small and closely related group of eleven copper-alloy vessels with punched and engraved decoration presented as a central frieze, and often, although not always, a punched inscription in Greek between the frieze and rim of the vessel. It is very likely that these pails were produced in the same or closely related workshops, perhaps located in Antioch, in the sixth century AD (Mango et al. 1989;Mango 1995).

The inscription is transcribed as follows:
Ὑγιένουσα χροῦ κυρὰ πολῦς σε χρόνυς κὲ καλῦς For the formally correct: Ὑγιαίνουσα χρῶ κυρὰ πολλοῖς σε χρόνοις και καλοῖς 'Use this in good health, lady, for many happy years' The spelling is not unusual in a Late Antique context, and the lettering would support a sixth-century date. It is also very similar to that on the other known pails with inscriptions, and the Greek text looks consistent with them. The formula πολλοῖς σε χρόνοις και καλοῖς is found on three others, and on a bronze pitcher discussed by Scholl (1994); he explains the obtrusive σε as for τε . . καί, 'both and'. It also seems illus. 16 Byzantine pail Br 1, enlarged photograph of the letter X, showing the circular punchmarks used to create it. Focus-stacked image produced by combining a series of original macro photographs by David Wheatley, University of Southampton possible that this phrasing provides metrical character to the text. On the pitcher and the other three pails the formula is preceded by ἐν, 'in'; the phrasing still makes sense, however, without the preposition. The hunt is one of four scenes depicted in the surviving friezes; others include Christian and mythological scenes as well as representations of animals. Seven pails are decorated with the hunt in a style paralleled in Late Antique hunting scenes on mosaic pavements, particularly in Antioch. The hunt scene on the Breamore pail provides a fine example of this subject and is generally close in style, composition and the forms and attributes of participants, to friezes on other pails, including another find with an English provenance from Bromeswell, Suffolk (Fern 2015, 176-80), but in the Breamore case, fewer hunters are represented than in other scenes (Mango 1995, pls I-II, nos 5-7). Of particular interest on the Breamore frieze is the apparent competence, vibrancy and realism in the depiction of the animals, especially the leopard, in contrast to the depictions of the warriors. The heroic nudity of the human participants supports Mango's (1989, 299) interpretation of similar scenes as showing mythological huntsmen; in a later article (1995) she suggests that a foundation myth related to Antioch may be represented. The inscription is also very similar to that found on several other vessels, both in content and position (Mango 1995, nos 2-5). Based on the similarities between the inscriptions found on these pails and similar texts on related vessels, Mango (1995, 279-81) proposed that they served as part of a washing kit, perhaps issued to soldiers in the late Roman army. The injunction to be healthy certainly makes sense in a bathing context; this inscription would perhaps suggest that the pail was one of a pair, the other given to a husband. Mango et al. (1989, 306-8) showed how Byzantine metalwork in general is found in important graves and contexts in England, a conclusion vindicated by the two eastern Mediterranean vessels subsequently excavated in the wood-lined chamber grave at Prittlewell, Essex (Hirst 2004). On an international basis, decorated Byzantine pails dating to the fifth/sixth century AD and of East Mediterranean manufacture, such as that from Breamore, are incredibly rare. Only eighteen decorated examples, produced in copperalloy, silver or glass, are known. Where a secure archaeological provenance has been recorded, their widespread distribution extends from Saudi Arabia to Turkey, Spain and England (Mango et al. 1989). Perhaps rather surprisingly, England is the country with the largest number of decorated pails, with only two others recorded: one from a richlyfurnished grave on Chessell Down, Isle of Wight, excavated in the nineteenth century (Arnold 1982, 27), and one from Bromeswell, Suffolk, found in plough-soil but now seen as likely to have come from a cemetery which was partly excavated in 2000-1 (Fern 2015, 176-80) on the same ridge, and to the north of, the famous 'princely' Sutton Hoo cemetery. Both those contexts further underline the significance of the Breamore pail.

Buckets (Illus 17-23)
Six buckets were excavated and all but one had been built using wooden staves held in place on the outside by thin copper-alloy sheet hoops and uprights. The wood was mostly preserved by contact with the metal fittings, so the buckets are not complete, there being no surviving bases, for example. Sf 7011 was stave-built, but had no metal fittings. The wood was identified as far as possible to species by Jacqui Watson, then of English Heritage. In most cases the original number of staves could not be estimated: some had decayed away completely, others had split, making edge recognition difficult; and even the metal fittings had deteriorated beyond the point of recovery. The metal uprights were usually the only clue to a vessel's original height, except in the case of Sf 1007, where an internal groove in one stave showed where the bottom of the vessel would have been slotted in. The metal fittings were held in place by copper-alloy rivets; most of those at Breamore were solid, sometimes split at the ends to tighten them, but a few were tubular. In many if not all cases the heads had been soldered on. The handle mounts and the only surviving handle had all been made from cut sheet, not cast in individual moulds. If any had fishplates, strengthening the junction of the handle and the mount, they were not evident.
Seven samples from the fills of the buckets, including one from each of the six, were flotation-sieved by Rachel Ballantyne at the Pitt-Rivers Laboratory, University of Cambridge, but they were not waterlogged and the contents probably derived from post-burial growth of rootlets, worm action, and surface cracking of the soil in dry weather introducing seeds and wood fragments. An exception might be very small illus. 17 Bucket Sf 1007 (see also Illus. 18) animal bonesincluding some from frogs/toads which were found in Sf 7011but even those are likely to be intrusive. Both the flora and the frogs/toads indicate damp local conditions, as does a Planorbid snail within Sf 5011. A charred hazel-nut shell fragment in Sf 1009 might just possibly have been put in the bucket before burial, but the survival of only a single fragment rather than a quantity suggests otherwise.
The buckets from the Breamore excavation are included in the late Jean Cook's Corpus, edited by Birte Brugmann (Cook 2004), but in an interim form and could not be included in its discussions or illustrations. The catalogue descriptions that follow conform with those used in the Corpus, and use the typologies illustrated in it (Cook 2004, 38-9). illus. 18 Buckets Sf 1007 (see also Illus. 17) and Sf 1008 breamore early anglo-saxon cemetery Three of the staves of this bucket had survived, one having an internal U-shaped groove into which the base would have been slotted. Four broken hoops include border lines of repoussé dots, as do the six 'type c, simple' uprights that overlie them intermittently. They are held in place by copper-alloy rivets, those at the bottom having larger heads, at least one of which was knurled. Two bifurcated 'type a' handle-mounts are fixed by split-pin rivets over two of the uprights; one has a surviving extension rising above the top hoop. Although 'type a' mounts are plain, whereas 'type b' and 'type c' are animal-headed, the prominent rivets near the ends of this mount may have represented eyes. At least part of the top hoop was rolled over to create a tube, strengthened by at least one clip. A perforation at the end of the handle-mount allowed the handle to swivel on it, attached by a missing rivet. The handle is a cut flat sheet of 'type d'. It has a repair in the centre, a small strip of a different metal, identified during conservation by Margaret Brooks, then of English Heritage, as a ternary bronze, whereas the rest is a quaternary bronze containing some zinc; as the repair is noticeably brighter, the conservator suggested that it might have been added immediately before burial. The surface of the repair has scratches between the heads of the rivets holding it in place; a suggestion that these might be runes can be neither confirmed nor absolutely refuted. Three of the four hoops of this bucket are plain, the top of the fourth having repoussé dots along the lower edge, and inscribed overlapping, upright semi-circles in the field. There is a separate rim. Two uprights are preserved, broken off at their bottoms, and with bordering lines of repoussé dots and a central row of ring and dot illus. 19 Bucket Sf 1009 ornament. The top two rivets on each upright secure a triangular strengthener to the inside. No handle mounts or traces of handle were preserved, and only fragments of its staves. During conservation, this bucket was found to contain the glass bowl, Sf 1008ii.
Height of bucket (approx.): 115 mm; diameter of bucket at top: 122 mm. Conifer, probably yew wood (taxus sp.), staves, in poor condition. Found in Trench 1 in a cut feature, 120, in which no trace of a skeleton was observed.
This bucket has three plain hoops, a separate rim and a narrow internal hoop at the top, with one rim clip surviving. One hoop may have had a white metal coating (a short strip shown on one hoop in the drawing is part of an upright, slewed through 90°). The four uprights are plain, except for one with intersecting incised semi-circles; two can be classified as 'type c simple', two as 'type a extended' (i.e. extended above the rim and perforated for the handle)one of which has a broken perforation at its top. No handle survived. Six or seven staves were traceable. Only three hoops survive in part of this bucket, all decorated with bordering lines of repoussé dots. A separate rim and rim clips are preserved, but there was no handle Parts of two uprights classifiable as 'type a simple' survive, as do traces of a third; all three have bordering lines of repoussé dots and stamped with oblongs and Cs along the centre. Empty rivet-holes could indicate missing hoops or previous use, but more probably indicate that they were punched before assembly, and were superfluous. Two kidneyshaped handle-mounts each had two openwork circles and two openwork triangles, and bordering lines of repoussé dots; the extensions that would have been perforated for the handle are broken off. Parts of four staves survived. This bucket has four hoops, three of which are plain, the fourth and topmost having bordering lines of semi-circle and dot, each with a line of repoussé dots above. It has a separate rim, and six of an estimated original seven rim clips survive. Two narrow uprights join the lowest three hoops, each with bordering lines of repoussé dots. Overlying and riveted to the lower edge of the top hoop is a narrow copper-alloy band, decorated with a line of repoussé dots, to which had been riveted before attachment to the bucket a serrated strip imitating separately attached triangular appliqués. They are variously decorated with repoussé dots, illus. 21 Bucket Sf 1013 (see also Illus. 20) circles, semi-circles and bosses, and are riveted at their apexes to the staves. Two narrow perforated handle-mounts, one in line with an upright, one slightly to one side, extend above the rim. The handle-mounts are riveted to internal strengtheners. No handle was found. Eight staves were preserved in part. SF 7011 BUCKET (NOT ILLUS.) No measurements possible. Yew wood (taxus sp.) staves. Found between the feet of Skeleton 703.
Listed in Jacqui Watson's report, but without further details and no longer extant, this was presumably a stave-built bucket held together by organic materials such as twine or withies.

DISCUSSION OF BUCKETS
Including those from Breamore, sixty-two stave-built buckets with copper-alloy fittings were included in the corpus by the late Jean Cook. However, the Breamore examples were found too late for inclusion on the published distribution map (Cook 2004, 28: sixteen fragments had been added to the PAS database by March 2011, some perhaps from the same vessel). No other cemetery in southern England has produced so many buckets from so few graves (Table 3 for comparisons), although identifications based on metal fragments are not always certain, and any bound with twine or withies, like Breamore's Sf 7011, would usually not have survived well enough to be recognized, so they might have been more frequently buried than can be known. They are found in both cremation and inhumation graves: the large East Anglian cemetery at Spong Hill, Norfolk, had up to twenty-four buckets, all but one in cremation graves (Hills et al. 1984, 93;Hills and Lucy 2013, 87-90).
The discussion of Spong Hill usefully demonstrates that, although stave-built buckets with copper-alloy bindings may be fifth century, and that they run on into the seventh century, they peak in the first half of the sixth, with 40% up to its mid-point. Iron bindings, of which there were no traces at Breamore, seem to date from the later sixth century.
Most of the Breamore buckets were made of yew wood, as were most Anglo-Saxon ones when the species can be identified, that wood being a little less susceptible to decay than many others. Oak may not have been used for small buckets, although recorded for larger ones with iron fittings, so its non-use at Breamore is not significant. The use of pine, however, for the staves of Sf 1013 is important; it is not unique, as it has occasionally been recorded before, though not in Hampshire or the Isle of Wight. Although native to Britain, pine trees are not likely to have been growing in southern England in the Anglo-Saxon period (Hooke 2010, 275), so the wood, or even the whole bucket, must have travelled from some distance further north.
illus. 23 Bucket Sf 5011 (see also Illus. 22) As Jacqui Watson pointed out in her conservation report, the vessels were difficult to make; work by Earwood (1993) and Morris (2000) has shown that white cooperagethe craft that produced stave-built bucketsrequires skill and specialised tools. The staves have to be carefully chamfered for a tight fit; making and fitting the rivets would also have required expertise. The fittings of Sf 1013, with the pinewood staves, are distinctive for their kidney-shaped handlemounts, of which only four other examples are known. However, unlike the wood used to make Sf 1013, there is nothing northern about the distribution of the kidney-shaped mounts; for example, one was found in the large cemetery at Long Wittenham, Oxfordshire (Cook 2004, 37). The decoration on the Breamore buckets, such as it is, is not specific to the area, or to southern England: repoussé dots are ubiquitous, semi-circles less so, but are on one of the two from Blacknall Field, Pewsey, Wiltshire (Annable and Eagles 2010, fig. 46, 9). None of the Breamore buckets are obviously of better quality than any of the others, and none are as well decorated as some in Kent and elsewhere. None of the Breamore buckets have spangles or other additions. Sf 5011 is embellished with pendant triangles below the top hoop, but those are not of particularly good workmanship, their decoration being simple, and their manufacture perfunctory, as some had not been detached from the strip from which they were cut. The way the strip was riveted over the top hoop may be the only feature unique to the Breamore buckets.
At least one bucket, Sf 1007, was demonstrably old when buried, as its handle had been repaired, perhaps specifically for inclusion in the grave. This was the only handle recovered; being closest to the ground surface, and not directly protected by wood, handles may have been particularly vulnerable to corrosion and decay. Nevertheless, that the buckets were not necessarily in 'as new' condition when buried must be borne in mind.
Buckets are often in graves that suggest above-average ownership, such as having weaponry in them, as at Breamore. Although they are found with both males and females, and occasionally with infants and juveniles, those in Wessex have a distinct male bias (Stoodley 1999, 48). They are occasionally found in exceptional burials, such as the late seventh-century 'bed-burial' at Swallowcliffe Down, Wiltshire, although that one had hoops and a handle made of iron, not copper alloy (Speake 1989, 54-8). The only grave containing more than one bucket recorded in the corpus is no. 34 at Wakerley, Northamptonshire, with three (Cook 2004, 77); so if Skeleton 128 really had both Sf 1008 and Sf 1013 beside its head, it is exceptional, but the circumstances of the test-pitting and excavation must leave the association open to question.
Evidence of the use of stave-built buckets is sparse, and interpretation is not helped by their positions in graves, some being at the feet, some at the head, as at Breamore, but elsewhere some are near the waist. Whether they were usually buried with anything inside them is problematic; it is not even known if they would have been waterproof. Food contents could be indicated by one from Mucking, Essex, that was reported to have had nuts in it, and an animal bone was in one from Alton, Hampshire, but as that was found at the bucket's top it might have sunk in from overlying soil (Evison 1988, 78). So it was disappointing that the flotation of the contents of the Breamore buckets yielded no definite food remains or other residues, although it was important that the attempt was made. Sf 1008 is the first bucket to have been found to contain a glass bowl, Sf 1008ii, presumably alluding to drinking, and strengthens the assumption that the buckets were generally associated with feasting (Stoodley 1999, 33;Lee 2007, 74-5, 87-90), either so that the dead person could share in the funeral rites, or for that person to use up after their bodily death, or as an offering to spirits or gods in the next world. A single glass vessel was found: a bowl inside bucket Sf 1008, made of pale green glass, with narrow, marvered, opaque, faint white trails externally below the everted rim. The base, which has a slight kick to make it concave, has the scar left by the pontil when it was removed after blowing. The glass has many small bubbles and tiny impurities; decomposition of its surface in the ground shows up swirl-marks, and failure to achieve a completely homogeneous mix.
Although glass drinking vessels of various sorts were deposited in many graves in east Kent, and occasionally elsewhere (Evison 2000a, 55-6 for distribution maps), this is the only one known to have been found inside another container. It is classified as a bowl because the kick in the base allows it to stand on a flat surface, unlike the commoner rounded palm cups. It has a fairly close likeness to a bowl excavated in grave 1612 at Carisbrooke, Isle of Wight, for which three parallels from Kent were cited (Morris and Dickinson 2000, 96). Although glass vessels were probably made in Kent during the early Anglo-Saxon period, these bowls seem from their forms and their white trails more likely to have been imports from northern France or the Meuse valley (Werner 1957, fig. 28; Evison 2000b). An early sixth-century coin in the Carisbrooke grave is better dating evidence than anything usually available; although not new, the coin is unlikely to have been more than quarter of a century old when deposited, so AD 525-550 was indicated, which is consistent with the dating of Continental glass bowls. That grave also contained a stave-built bucket with copper-alloy fittings, a copper-alloy bowl and a drinkinghorn, which seem to emphasize a connection with feasting and drinking rites.

weapons and knives
As with the buckets, the large number of weapons, particularly the remains of shields, found in The Shallows cemetery was exceptional for Anglo-Saxon England in relation to the small number of excavated burials. Margaret Brooks reported that much of the iron was mineralized with surfaces that had spalled and blistered, though solid iron could be seen in a few core areas.

Shields (Illus 25-33)
Iron bosses and handle-grips are often all that survives of shields with wooden boards; sometimes iron mineralizes some of the wood, but that did not happen sufficiently at Breamore for timber species to be identified. Bosses sometimes have copper-alloy discs on their apexes (Dickinson and Härke 1992, 19), but none at Breamore had those extra embellishments; one, Sf 5002, may have had white-metal coated rivets at the edge of its shield-board (Sf 5004 and Sf 5005).
The terminology and typology used to describe the Breamore shields are taken from Dickinson and Härke (1992). Since the report was prepared in 2004, however, a new classification of shield-bosses has been proposed by Bayliss et al. (2013, 148-62), which is important for its dating implications, discussed below. Problems of applying it to the Breamore bosses include a degree of uncertainty over the apex types, as that on Sf 7002 breamore early anglo-saxon cemetery had broken off and the very top of Sf 1012 may have disappeared. The items listed include those found in the work carried out by Berkshire Archaeological Services (Sf 1-2, Sf 6-7, 10) before the Time Team programme. SF 1 SHIELD-BOSS (ILLUS. 25) Diameter: 192 mm. Found in Test-pit 5 by Berkshire Archaeological Services, above the chest of a skeleton that is probably the one subsequently excavated as Skeleton 304.
Test-pitting revealed a Group 1.1 (= Bayliss et al. 2013, SB3-b1) shield-boss with five flange-rivets and a Type Ib grip. Traces of mineralized wood and leather survived from the shield-board and presumably also from a carrying-strap. SF 2 FIVE STUDS (ILLUS. 25) Diameters: c. 29 mm; shank lengths: 6.8 mm to 8.1 mm. Found close to Sf 1.
Five studs were found, though only one was individually numbered. That was recorded as being about 200 mm from Sf 1. The studs would have been from the side of the board that faced the enemy (Dickinson and Härke 1992, 27). This stud was corroded on to a straight, broken, bar with a flaring terminal probably part of Sf 9. A straight, broken, bar, with flaring, perforated terminal was originally thought to be a knife, but it has a rectangular section and probably joined the bar now corroded to stud Sf 8, although not in its immediate proximity. It is probably a fitting from the shield-board. This is the only shield-boss attributed to Group 4 (and to Bayliss et al. 2013, SB1-b); it has five flange-rivets: a complete set, two with their hammered ends still on the inside. The boss was found with remains of a Type IIa handle-grip. Both the boss and the grip retained traces of mineralized wood. SF 3007 STUD (NOT ILLUS.) Diameter: 13 mm.
The site report prepared by Kate Edwards states that a stud with part of a copperalloy washer at the end of the shank was found above the head of Skeleton 314. It is not shown on the plan, and has not been traced, so is not illustrated here. It was probably not from the missing end of the grip of Sf 3003, as the extant rivet on that is wider. A copper-alloy 'washer' seems unlikely. Margaret Brooks in the conservation report noted that X-rays of both these studs indicated a white-metal coating, but that no grey layer, the usual indicator of such plating, illus. 29 Shield-boss Sf 3003 breamore early anglo-saxon cemetery was seen during removal of corrosion. Two other items, Sf 5016 and 5018, are shown on the plan at similar distances on the north side; although not described in the inventory, these may have been two more studs. The studs suggest an attachment close to the edge of the shield-board, such as a strengthening or repair batten, or some embellishment. Allowing for a margin between the studs and the edge and assuming that the board was flat not concave, a diameter of about 0.65 m for the shield is indicated, at the higher end of the 'medium' size group, though larger than would normally be expected before the seventh century (Dickinson and Härke 1992, 43-60). This reconstruction would mean that the shield was adjacent to bucket Sf 5011, but not covering it. This elongated lozenge-shaped plate, riveted at its ends and with a central nail, the head hammered over, was presumably attached to the shield-board of which boss Sf illus. 30 Shield-boss Sf 5003, with studs Sf 5004 and Sf 5005 7002 was the centre. If so, the board's diameter was at least 0.65 m, much the same as the estimate for Sf 5003-5.

DISCUSSION OF SHIELDS
Including the two bosses excavated by Berkshire Archaeological Services, remains of nine shields were found at Breamore. Because the grave of what became labelled Skeleton 128 was investigated initially during metal-detecting, then by the test-pitting, and finally by Time Team, details of Trench 1 are a little uncertain. Still, Skeleton 121 is clearly recorded as having had two shield-bosses interred with it. This is an exceptional practice for early Anglo-Saxon weapon-graves (Dickinson and Härke 1992, 63). The other weapon-burials had the typical practice of a single shield, though with the very high ratio of seven shields with seven juvenile/adult burials ( Table 3), assuming that Sf 1 was included in the grave because of the juvenile 305 rather than because of the child illus. 31 Shield-boss Sf 5012 304. In other words, from those graves excavated, no-one aged 17 years or over was interred without at least one shield.
The positions of the shields in the graves varied. Those represented by Sf 1011 and Sf 1012 were above the legs of Skeleton 121; also above legs were Sf 6, Sf 5003, Sf 7002 and Sf 7007, while Sf 3003 was over or by the head of Skeleton 314, and Sf 5012 was on the left shoulder of Skeleton 505. Only Sf 1, with Skeleton 304, seems to have been over the chest/ waist. This conforms to a regional trend identified by Dickinson and Härke (1992, 65-6), with 'Saxon' areas favouring the waist and upper legs, as at Breamore. The shield represented by boss Sf 7007, however, was over the lower legs of Skeleton 704, which is unusual, and it also covered the bones of a very young child, Skeleton 705, which may be unprecedented. Shields were almost never placed with children, even older ones (Dickinson and Härke 1992, 68-9), and the inclusion of the child under Sf 7007 must have been done deliberately and been a meaningful action. Nevertheless, the shield was probably included because of the older person rather than because of the child.
If the calculations of 0.65 m diameter for the shield-boards with Skeletons 506 and 703 are correct, they were a little above the median size for early Anglo-Saxon shields (Dickinson and Härke 1992, 44-8, 65, 68-9). Nothing otherwise seems in any way illus. 32 Shield-boss Sf 7002 and shield-plate Sf 7006 exceptional; no unusual types of shield-boss occurred. The lozenge-shaped fitting, Sf 7006, if fixed to the board of the shield that went with boss Sf 7002, is something with a Wessex focus, at Harnham Hill, Wiltshire, Droxford and Portway (both Hampshire) (Dickinson and Härke 1992, 27).
The four groups into which the Breamore shield-bosses fall in the typology proposed by Dickinson and Härke 1992 are dated by them to the fifth and sixth centuries. The classification by Bayliss et al. (2013) is consistent with this; in particular, though, three of the Breamore bosses are types within their SB3 classification, which 'falls on the last, sixth-century, peak of the probability distribution . . . in use for a fairly restricted period around the middle decades of the sixth century' (Bayliss et al. 2013, 247). The absence of any examples of their type SB4, current around AD 600, is also significant for dating, as is the absence of domed rivet heads, all the Breamore examples being discoid-headed. This is also consistent with the absence of groups dated to the seventh century by Dickinson and Härke (1992).

Spears (Illus 34-7)
The iron spear-heads have been classified according to the typology published by Swanton (1973). The implications of the revised classification proposed by Bayliss et al. (2013) are considered in the discussion below, but the condition of the ironwork has precluded its application to the catalogue entries. Most of the spear-heads have traces of mineralized wood but in no cases could they be identified to a specific tree species. The position in which this object was found strongly suggests that it was a knife, despite the mention of a socket. It is described as totally mineralized, with traces of wood in the socket, and is no longer extant. SF 5015 SPEAR-HEAD (ILLUS. 36) Length: 318 mm. Found on the right side of the head of Skeleton 505.
Unusually fitted with a binding-ring at the end of the socket, this Type I1 spear-head also had a transverse rivet half-way along it. SF 7001 SPEAR-HEAD (ILLUS. 37) Length: 386 mm. Position not recorded on context record sheets, nor on plans, though thought to have been with Skeleton 704. This is the third type H2 spear-head; it has two transverse rivets in the socket. SF 7005 SPEAR-HEAD (ILLUS. 37) Length: 294 mm. Found above the right shoulder of Skeleton 703.
The fourth and last of the Type H2 spear-heads that was found in the cemetery. Nine excavated spear-heads, six definitely from graves, is a high ratio in relation to the number of burials (Table 3). There were no recorded instances of burials with more than one spear. All the spear-heads are of types common in early Anglo-Saxon graves. The number of ferrules is uncertain, as the position of Sf 5014 suggests that it was not a ferrule as originally identified; on the other hand, Sf 3006, classified as a knife below, suggests that it was a ferrule. The revision to Swanton's (1973) Swanton (1973, 115) called 'corrugated' and thought imparted illus. 36 Spears: spear-head Sf 5007 (left); spear-head Sf 5015 (right) longitudinal strength; and there are none of the triangular or parallel-sided types (Class SP3). The absence of types indicative of the seventh century suggests that the cemetery did not remain in use after AD 600. This is consistent with the dating of a typology established for four East Anglian cemeteries, which would place the Breamore spear-heads in one of the three most common, fifth-/sixth-century, categories, including the 'corrugated' type into which Sf 1002 and Sf 5015 have been fitted (Penn and Brugmann 2007, 21).
Too little evidence survived for any correlation between the spear-heads and the skeletons to be made, except that none were definitely associated with anyone aged less than 17 years of age, and that, just as everyone aged 17 years or over seems to have had at least one shield, so too there are enough spears for every juvenile and adult to have had one. The longest at Breamore, Sf 7001, was not with the oldest male, but at 390 mm it is in the third longest tier of the East Anglian range, established from a total of about a hundred (Penn and Brugmann 2007, fig. 4.2).
Whether any features of the spear-heads signify social status should be considered: two, Sf 1002 and Sf 5015, had special treatment given to the socket ends, the former having four circumference grooves beaten into it, the latter seemingly having a detachable binding-ring. Swanton's (1973) corpus shows that such features are rare; he illustrated only one with a binding-ring at the end (Prittlewell, Essex, fig. 20d) or further up the shaft (Nassington, Northamptonshire, fig. 53e). A few have circumferential copper-alloy wire inlays, and one from Battersea seems to have a single groove ( fig. 43a). Closer to Breamore geographically, one from Grave 22, Blacknall Field, Pewsey, Wiltshire, has bands and inlay, and is from an inhumation with a sword in a scabbard with a finely crafted Style I mouth-band (Annable and Eagles 2010, 200-2). One from Charlton Plantation was unstratified (Davies 1984, no. 127, 135 and fig. 18). Also published since Swanton's corpus is the cemetery at Finglesham, Kent, in which one spear-head had a copper-alloy binding-ring, and two had circumferential grooves (Hawkes and Grainger 2006, 277, grave 133, no. 1.1;272, grave 117, no. 1.1;310, grave 204, no. 7.1), so perhaps these features are not as unusual as at first appears. The ferrules should also be considered, however, as even if their number was only two, they are not commonly found and are 'the sign of a well-finished spear' often in graves which also had swords (Swanton 1973, 7). The Breamore spears may therefore be considered to have been a little more special than the norm.

Knives (Illus. 38)
Knives were multi-functional objects, found as often with women as with men, but are included here with weapons to group iron items together. The classifications adopted are from Böhner (1958) and from the series proposed for Dover (Evison 1987, 113-6) where possible, but corrosion mostly limits finer discussion. SF  The thickness of this small fragment suggests that it is part of a knife-blade, although its position by the foot of a body might favour interpretation of it as a ferrule from the end of a spear, as knives were usually carried in sheaths suspended from belts. This is a Type D knife (also Evison 1987, Type 6) with a very pointed handle and traces of leather from a sheath on the blade.

DISCUSSION OF KNIVES
Even if Sf 3006 was a ferrule, the total of knives would probably still be seven, as Sf 5014 may not be a ferrule (above); knives were common objects, so, unlike the shields and buckets, there is nothing particularly significant about their number at Breamore. A more recent typology than Böhner's suggests that these knives probably all date to the sixth century, and are within Type A of that classification; the perforation in the tang of Sf 1003 does not imply a scale-tang fitting, but fixed a whittle tang of the usual kind (Riddler 2012, 127).

personal accessories
Buckles (Illus 39-45) Three buckles were excavated, their frames all of copper alloy, as were two of their pins, a third being of iron; one had an elaborate plate. A fourth buckle frame, Br 2, also of copper alloy, was reported by metal-detectorist Steve Bolger before 2001. The buckles from The Shallows were found too late for inclusion in the corpus volume by Marzinzik (2003); her terminology is used, except that 'frame' is preferred to 'loop' and 'pin' to 'tongue'; these are the terms used both by Egan for London (Egan and Pritchard 1991, 50-3) and Goodall for Winchester (1990, 526 A single frame, D-shaped buckle with shield-shaped pin and D-shaped two-part damaged plate (Typegroup 1.11b-2.12) was found with one of the young adults. The upper plate of the frame is attached by four small rivets to a back plate; strap leather or textile would have passed between them. The plate does not directly join the frame, as there is no provision for a hinge. The upper plate has cloisonné cellwork, forming a slightly raised oval within an outer border. The oval has a central strip of three rectangular cells, with four others shaped to fit the curve. The central rectangle has a blue glass inset, flanked by two red insets, one now broken in half; the red insets are either garnets or glass, but have not been analysed. Two of the other four cells hold red insets, one has a half-broken inset, and the fourth is empty. The outer zone now has three blue glass insets, two on the straight side. There are sixteen red insets, small and somewhat irregular in shape; two cells are void. A white substance visible within the voids may be calcium carbonate, used as a fixative; like most copper-alloy cloisonné, the cells did not have reflecting foils between the glue and the insets.
Sf 5010 is an exceptional discovery. As the plate was not directly attached to the frame, they may not originally have belonged together, though there is nothing incongruous about the match, unlike the cellwork plaque with an iron frame and pin at Pewsey (Annable and Eagles 2010, 227-8). Two similar buckles from Bifrons, Kent, were not attached to frames, and could have been from boxes (Hawkes 2000, 13-16). The D shape of the plate has not been reported previously in England, those with cellwork being rectangular, or kidney-or heart-shaped; most of those were found in Kent, the exceptions being Abingdon, Oxfordshire, and Pewsey and Petersfinger, both illus. 40 Buckle plate and frame Sf 5010 Wiltshire, the first heart-shaped, the other two rectangular, and all from male graves. Also exceptional is the unexciting pattern made by the cells, as others contain curvilinear elements and larger insets. They are paralleled on the Continent in graves dated to the second half of the fifth/first third of the sixth centuries, and the Breamore buckle is almost certainly an import, but could have come from anywhere in a long arc from Ibiza to the middle Rhine, reaching as far east as Hungary (Marzinzik 2003, 39-41). A remarkable local analogy is the bridle fitting found elsewhere in Breamore, which also had coloured insets and is attributed to a mid-fifth-/mid sixth-century western or central Mediterranean origin (Eagles and Ager 2004, 89). SF 5013 BUCKLE FRAME AND PIN (ILLUS. 39) Width: 21 mm. Found between the knees of Skeleton 506. This single frame, oval (Typegroup I.II) buckle was found with the same skeleton, 506, as Sf 5010. It has an iron pin, on which a patch of leather survived. Two incised lines are present on one end of the pin-bar.
The frame may have lost its original pin, as its combination with an iron pin looks unsuitable, although is not uncommon (Marzinzik 2003, 23). Small oval copper-alloy frames are frequently found in early Anglo-Saxon inhumation graves; Marzinzik identified those of 20 mm width or less as a little more likely to date after the mid-sixth century, but Sf 5013 is not distinctive enough for a close date to be suggested, which is true also of Sf 1010; after the seventh century, the number made in copper alloy seems to fall off, however (Hinton 1996, 7-8;Ottaway and Rogers 2002, 22-3).
BR 2 FRAME (ILLUS 39 AND 41) Width: 30 mm. Found during metal-detecting before the discovery of the pail, from or very close to the surface of the mound but not in a grave.
An oval (Typegroup I.2/I.5) frame with faceted sides and flat top and bottom cannot be closely dated, as its pin-bar is broken off and its pin is missing. At 20.1 g, it seems heavier than most buckle frames; the metal has not been analysed, but superficially looks dull in colour and may have a high lead content. Fine striations in places suggest filing down after casting. Patches of iron corrosion may derive from gravel staining not from adjacent artefacts. Despite being undiagnostic, Br 2 is probably from the Anglo-Saxon period and from the cemetery.

Tweezers
The only other metal object from the graves was a pair of tweezers. SF 7009 TWEEZERS (ILLUS. 42) Length: 77 mm. Found with Skeleton 704; the plan shows the location as below the left lower arm, though the context report says below the pelvis.
A pair of plain copper-alloy tweezers, with a suspension ring, was probably a toilet implement. Tweezers are found both in settlements and in men's and women's inhumation graves dating to the Anglo-Saxon period elsewhere in Hampshire (Hinton 1996, 46). In other areas they are often found attached to the same ring as implements that can be identified as ear-scoops and the like, so were for personal grooming. Combs and toilet implements suggest concern for appearance, in death as well as in life (Williams, H. 2007, 83-9).

Pebble
Sf 3001 Pebble (Not Illus.) The context record sheet describes how a 'rounded pebble' (Sf 3001) was found 'where the lower jaw of [Skeleton 306, an child] had been and may have been deliberately placed there.' This pebble does not seem to have been retained but is shown in the site photographs, and was obviously sufficiently distinctive to attract the excavators' attention. Such things are either rarely encountered, or are not noticed, but one of us, many years ago, excavated an Anglo-Saxon grave in Oxfordshire that contained a younger child and a distinctive, helical fossil 'that could conceivably have been deliberately included as one of the child's toys' (Hinton 1973, 121). A more prosaic interpretation would see such things as a means of soothing tooth-ache, but amulets might also be considered (Crawford 1999, 79-80), not necessarily to help the child's spirit, but to stop that spirit from issuing bad thoughts towards the living. Part of the foot-plate of a miniature square-headed brooch that had an old break across the bow lacks a terminal because of more recent damage. Enough survives for Style 1 animal ornament to be recognized. The artefact was analysed at the British Museum and found to be approximately 88% silver, with niello fillings in places (Portable Antiquities Scheme 2001, 26-7).
This type of brooch occurs mainly in Kent, the Isle of Wight and Saxon-settled areas in northern France; they were generally worn in pairs and can be dated to about AD 530/40-560/70 (Parfitt and Brugmann 1997, 37-9 and fig. 23). The use of silver is notable; a recent distribution map has only five sixth-century silver objects in southern England and the south Midlands, apart from in Kent and on the Isle of Wight A gilt copper-alloy small square-headed brooch has a rectangular head-plate, its top corners broken, containing a frame enclosing a circle and lines. It has an undecorated bow, and an expanded foot, also damaged. The foot has a poorly defined lozengeshaped panel and decoration. The catch-plate and lugs on the back show traces of iron corrosion, presumably from the pin. The surviving gilding shows signs of wear, having rubbed away from raised areas.
This type of brooch originated in Kent, but during the early to mid-sixth century the form was used in adult female-gendered costume throughout most of southern England. A slightly larger example from Kent, in silver, was recently dated to the early sixth century (Crummy 2015, 47), and one in gilt copper-alloy was excavated at Barrow Clumps, Wiltshire (Osgood and Andrews 2015, 34).
A cast circular copper-alloy disc with central perforation has on each side a circumferential line of small punched dots, and five larger punched dots around the perforation on each side.
This object is most similar to weight 11 from Gilton, Kent, Grave 66 (Scull 1990, fig. 2), one of the few weight-sets to have been recovered from a grave. When found individually, such things can be impossible to identify reliably (Hinton 2000, illus. 44 Brooch Br 4 63-6 and 111-2); in this case, the punched dots enabled the identification to be made. At 9.04 g, the disc is within a reasonable tolerance of seven times a unit of 1.3 g, based on Merovingian gold coinage, but is a little closer to six times the slightly heavier Byzantine unit of 1.5 g, both of which seem from surviving weightsets to have been known in Anglo-Saxon England. Most belong to the later sixth and seventh centuries, and are evidence of extensive trade and exchange systems. They have been found in both male and female graves (Scull 1990, 184-6 for associations; 187-97 for metrology).
MOUNT by JANE KERSHAW Br 6 Mount (PAS FASW-E5F-84-2; British Museum MLA 3001) (Illus. 42). Length: 31 mm; width: 17.5 mm; thickness: 1.5 mm. Found by Steve Bolger in the same field as the cemetery at Shallows Farm, but at some distance from it. This is a broken cast copper-alloy mount with two rivet-holes in the extant, concave end, with a copper-alloy rivet surviving in one. There is interlace ornament in the panel.
The contoured lines in the panel appear to be part of a stylized beast, with spiral hips, the last of which creates the inverse V-shape at the surviving end and terminates in a thin leg with an angled foot. Another leg emerges from a hip near the break: either the fore-leg of the same creature or the back leg of a separate, interweaving one, with pellets in-between. Although the rivets could suggest a strap-end, the drawing does not illus. 45 Weight Br 5, side 1 illus. 46 Weight Br 5, side 2 indicate a split end, and the straight sides indicate a mount such as those from Uppåkra, Sweden (Hardh 2010, 258, fig. 14).
The sharply angled limbs resemble Borre-style animals, but the ribbon shape of the main body suggests transition to the Jellinge style, so a date from the late ninth to the mid-tenth century is suggested (Kershaw 2013, 144). The design is not exactly paralleled, and suggests a misunderstanding of these Scandinavian styles, during their adaptation in England. The continuous looping does not form complete knots or volutes, for instance, but towards the break the interlace pattern is more symmetrical, the effect perhaps intended. Brooches from Sweden are in the same tradition, as is a strap-end from Wharram Percy, Yorkshire (Kershaw 2013, 174). Closer to Breamore geographically, and also outside the Danelaw, is a copperalloy strap-end found recently at Mudford, Somerset. Wessex does not have many such Anglo-Scandinavian objects (Kershaw 2016), so although later in date than the cemetery at Shallows Farm, the mount may show that the River Avon retained importance.
POTTERY by DUNCAN H. BROWN An assemblage of forty-nine sherds was recovered from sixteen contexts excavated by Time Team. No Anglo-Saxon pottery was found on or in the mound. Not inspected for this report were fourteen sherds thought to be from an Early Bronze Age Collared Urn found in their Test-pit 3 by Berkshire Archaeological Services (Entwhistle 2001, 8). Unstratified material (consisting of one fragment of white refined earthenware) and three fragments of fired clay are also excluded from this discussion. The pottery was recorded in 2002 and quantified by sherd count, weight in grams, rim percentage and maximum vessel count. Table 2 lists the contexts, quantities and types of pottery recovered. Topsoil or subsoil layers contained most of the assemblage, much of which is post-medieval in date. There is nineteenth-/twentieth-century flower pot and Verwood type ware (Algar et al. 1979), with seventeenth-century glazed redware. Prehistoric, Romano-British and Anglo-Saxon pottery is also present. The Anglo-Saxon organic-tempered ware may be fifth or sixth century in date. Most of the pottery associated with the mound is prehistoric, although it is too fragmentary and abraded to be associated with any specific activity and is most probably all re-deposited. The linear feature in Trench 8 produced two sherds of Romano-British pottery, too small for closer dating. The linear feature in Trench 9 produced a small sherd of Bronze Age pottery and a piece of thirteenth-century Laverstock-type coarseware with a characteristic scored external surface (Musty et al, 1969, 99-107).
Overall, the small sherd size and abraded condition suggests that none of these finds were from primary deposits but the assemblage does reflect activity from the late Bronze Age to the early modern periods. Flakes and broken flakes were found in most test-pits and trenches, as was burnt flint; all was locally derived. The quantity found below the mound, in Trench 6 and Test-pit 3, and its absence from other interventions which also reached the old ground surface, may indicate that the mound had enhanced a slight natural dome. Trench 7 reached the same overall depth as Trench 6, but without a flint scatter being found; the former may therefore have located a shallow scoop from which mound material had been scraped.
The range of flint within the mound was similar to that from below it, but in different proportions, notably a concentration of cores in Trench 5. No worked flint was recorded off the mound in Trenches 8 and 9; flint in Trench 2 showed clear signs of damage by water-rolling.
The waste flakes were generally squat in plan, with no evidence of deliberate blade production, and are therefore indicative of a late Neolithic or early Bronze Age date. A chisel arrowhead within the mound is likely to be late Neolithic. This is consistent with evidence from both the mouth and the up-stream parts of the River Avon, with woodland clearance east of Cranborne Chase well advanced (Gardiner 1984). The burnt flint concentration could be from one of the burnt mounds that are well known in the valley.
CONCLUSION by DAVID A. HINTON activity at the shallows Sundry flint tools and burnt flints indicate late Neolithic and Bronze Age activity continuing in the Breamore area after its initial Neolithic clearance. A concentration of burnt flint near the river in Trench 2 is probably the residues of a burnt mound; fieldwalking in the 1980s by the Avon Valley Archaeological Society found similar evidence in other fields close to the river (Light et al. 1995, 72-5). The wood chips indicate active management of the river's banks after the Bronze Age.
No absolute dating evidence for the construction of the mound at The Shallows was discovered, but the flints and Collared Urn sherds are consistent in suggesting an Early Bronze Age date. The mound may have enhanced a slight natural feature. It seems not to have had a surrounding ditch, though excavation did not test this fully. If discolourations observed in Trench 3 really were post-holes, and were contemporary with the mound, they might represent some sort of enclosure around it, such as a fence or postring. The other post-holes could be of the same date, or could be the same date as the cemetery, or post-date it. They do not cohere with its graves in any obvious manner, however, and had different fills.
The Iron Age and Roman sherds suggest minor activity, such as ploughing, which is consistent with the earlier fieldwalking data. The Anglo-Saxon sherd might indicate domestic activity somewhere nearby, though could have come from a plough-disturbed grave or from a cemetery ritual. Absence of any later medieval sherds, other than the single one from Laverstock, suggests that the meadow was used for grazing and hay after the cemetery went out of use, though the isolated Anglo-Scandinavian mount Br 6 (Kershaw, above) might possibly indicate an occasional pre-Conquest visitor, other than a farmer. The post-medieval Verwood pottery may reflect another ploughing episode, before the modern intensive agricultural regime.
the anglo-saxon cemetery Some years have passed since the excavation of the cemetery in 2001 and the submission of the excellent grey literature reports in 2005, and inevitably much work has been published subsequently on the early Anglo-Saxon period which could not benefit from discussion of the site and the artefacts from it.

Location and Population
The cemetery's location within a prehistoric mound is not unusual (Semple 2013, 15-16); the practice may signify claims to territorial rights through association with an ancestral past. In the Breamore area, however, there were other pre-existing monuments that were perhaps still visible in the Anglo-Saxon period and could have been chosen instead, such as the barrows in the fields north of Bullcroft, on the edge of the terrace overlooking the flood-plain of the River Avon, an orthodox location for an Anglo-Saxon cemetery (see Hinton and Worrell 2016, S05). The Shallows, however, was exceptionally placed in a river-bend. The most famous of all cemeteries, Sutton Hoo, is always said to be close to the River Deben, although the ships there had to be dragged some 700 m up to it. The seventh-century barrow at Taplow, Buckinghamshire, is much closer to a river, the Thames, seemingly at a spot where its rich and powerful occupant could still dominate passage up and down the waterway. Near Taplow, in Cookham parish, Berkshire, are barrows at Cock Marsh, in one of which was excavated in the late nineteenth century 'a platycephalous man with his dog and various articles', according to the report to the Society of Antiquaries (quoted by Meaney 1964, 45). Later reports call the 'articles' a shield-boss, a knife and an urn, so the burial was probably Anglo-Saxon. The barrows are very close to a bend in the River Thames, though a wider one than at The Shallows. The site would be a close parallel if it were a cemetery and not an isolated burial.
Reverting to the Breamore area, the grave in the Bargates cemetery, near the mouth of the Avon at Christchurch (Illus. 1) closest to the flood-plain, was only about 40 m from it (Jarvis 1983, fig. 55), which is more or less the same as the distance from the graves at The Shallows to the flood-plain there. Upstream, the nearest cemetery is at Charlton Plantation, on the edge of the terrace, over 300 m from the Avon (Illus. 1; Davies 1984, fig. 1), which is much more typical of early Anglo-Saxon cemetery locations. The Shallows is on a promontory just high enough not to flood; unless Cock Marsh was one, no other known early Anglo-Saxon cemetery has a river on three sides.
The full extent of the cemetery was not established, as the geophysics is not definitive, and only a single small trench, 6, was dug on the mound that did not find graves. That was the most northerly trench, however; if it could be used as representative, the graves might all have been on the southern side of the mound, but not extending southwards off it, assuming that the metal-detected objects Br 2-5 had been plough-dragged from their original positions. The trenches are less able to suggest the cemetery's extent eastwards and westwards; but even if there were graves only in the spaces between Trenches 1, 3, 5 and 7 (which together contained at least eleven), at the same density, there could have been more than fifty. That would be comparable in known number to Charlton Plantation, where forty-two graves extended over 50 m in one direction. The irregular grouping at Breamore is more comparable to Bargates, but there the thirty graves were scattered across an area approximately 40 m x 35 m. As Trench 7 suggests fairly wide spacing between graves, The Shallows may actually have had quite a low total, in the twenties or thirties.
Although only eleven skeletons were definitely located, there may have been two others in Test-pit 5/Trench 3 and another in Trench 1, if cut F120, which contained a bucket, was an unexcavated grave. Even using a total of fourteen, to have what are fairly clearly two double burials (Skeletons 304 and 305; 503 and 506) and one triple burial (703, 704 and 705) is a remarkably high proportion (Table 3). Furthermore, none of the skeletons that could be sexed biologically were shown to be female, and none of the excavated grave-goods were of a type normally gender-linked to females. During the Time Team programme, it was suggested that Skeletons 305 and 506 might be females, but subsequent analysis after cleaning did not substantiate this (Roberts, above). The objects lifted during test-pitting cannot be attributed with certainty to skeletons subsequently fully excavated, but both 305 and 506 seem to have had shields and spears with them; weapon-sets like those are normally associated with males, there being only a single example of a skeleton buried with a shield and spear that has been reliably identified as probably female on osteological grounds (McKinley 2003, 108; graves probably of females with spears are known, but are very rare). Even some of the grave-goods other than weapons, such as the buckets, have a slight bias to male associations, at least within Wessex.
The metal-detected brooches would normally be taken as female items, and might be used to argue for a more balanced gender ratio at Breamore, but it should be borne in mind that female graves tend to have more items in them of the sort found by metaldetection, so two brooch fragments is still actually quite a low number. The three children, however, show that the cemetery may have been more family-based than the excavated graves and artefacts otherwise suggest; ratios of children to juveniles and adults in Anglo-Saxon cemeteries vary widely, and Breamore is not exceptional in its proportion of those aged 3 years to 16 years, or in the absence of neonates and infants (Sayer 2014; but by taking infancy up to 5 years, and childhood from 6 to 12 years, his figures are not precisely comparable to Breamore's. See Table 3 for local comparisons).

Artefacts and Dating
The dating of the Shallows Farm cemetery can only be provisional, since it is likely that there are other graves there that might yield further evidence. The artefacts from the graves, and found by metal-detecting, all point to the sixth century, to which the manufacture of the Byzantine pail Br 1 is attributed. The very similar vessel found at Chessel Down on the Isle of Wight excavated in 1855 was in a reasonably wellrecorded grave that contained three square-headed brooches recently given a daterange of 525-70 (Hines 1997, 230). Re-evaluation of the 'Bromeswell bucket', the third of this type of object from England, has shown that it probably derived from the cemetery that was partly excavated in 2000-1 at Tranmer House, Suffolk, near Sutton Hoo, and that dates to the sixth century, possibly just extending to the beginning of the seventh (Fern 2015, 181-91). The glass bowl Sf 108ii is similar to one from Carisbrooke in a grave that has its dating helped by a coin, possibly of some age when buried, but pointing to 525-50 for deposition. The other Breamore containers are the stave-built buckets, which are not attributed very close dates, but at least do not include any that are iron-bound, which is a trait of the seventh century.
The Breamore ironwork also points towards the sixth century. Recent typologies place the shield-bosses in that century, quite possibly in its first two thirds, and the knives also seem sixth-century; the spear-heads are less diagnostic, but none are a seventh-century type. This runs counter to a recent synthesis that attributed some of the Breamore ironwork to post-570, but gave no supporting evidence (Harrington and Welch 2014, 132). The Continental buckle Sf 5010 has a suggested range of c. 450 to c. 530, the metal-detected miniature square-headed brooch Br 3 has been dated from c. 530 to c. 570 by Barry Ager, and Br 4 is probably also sixth-century. The weight, Br 5, seems related to the metrology of the gold coins that circulated in England mainly in the sixth and seventh centuries. The earliest weight set with a date attributed to it was found at Watchfield (Oxon), for which c. 525-75 was suggested; it was in a cemetery with nothing definitively ascribed to the seventh century (Scull 1992, 252).
Objects were not necessarily new when buried: bucket Sf 1007 had a repaired handle, buckle Sf 5013 had an iron pin that may have been a replacement, buckle Sf 5010 was probably a 'marriage' of two separate elements, and the pail may have taken some time to reach Breamore from the Mediterranean. However, the deterioration of most of the ironwork hinders further discussion. If the imported buckle Sf 5010 had taken some time to reach Breamore, or was made towards the end of its date-range, the whole assemblage could be fitted into the first two thirds of the sixth century, and indeed could all have gone into the ground in its second quarter.
Because of the circumstances of discovery and excavation, associations are not precise, but the ratio of shields and spears to burials is considerably above the local norm, as Table 3 shows; whether any burials with shields did not have a spear as well cannot be certain, but seems unlikely. None of the people aged over 17 years can be shown not to have had weapons, which is again exceptional for early Anglo-Saxon cemeteries, just as it is for one of the dead to have had two shields (Skeleton 121 with Sf 1011 and Sf 1012: Dickinson and Härke 1992, 63; but Trench 1 was much disturbed and the record may not be entirely reliable). The shields do not seem to have been outstanding individually, only one even having a decorative attachment (Sf 7002 and Sf 7006). The spears and ferrules hint at something beyond the norm, but no grave contained the most prestigious Anglo-Saxon weapon, the sword. As it has been argued that swords tended to be with men of higher stature as well as of higher status (e.g. Härke 1990, 38-40;but note Lucy 2000, 74), then it could be significant that the height  128 is included in the 'males sexed by grave-goods' column. The spear-heads include those recorded as unstratified, but excludes ferrules. One bucket, Sf 1009, was from a cut feature, not necessarily a grave.
Bargates: as some grave-cuts did not contain identifiable body stains, the number of inhumations cannot be certain; no grave contained more than one shield-boss or spear-head. One cremation was probably of a female, and an unstratified bead also indicates female presence. Blacknall Field: adults at this site were taken as from 20 years of age, juveniles as aged 14-19 years, four of which were identified as female, seven as male, with children from 3 to 13 years. The column for spear-heads includes one unstratified and one that may be an arrow-head. Charlton Plantation: a large part of this site was destroyed by roadworks without record; many more graves were clearly in that area. The age range for juveniles was 15-25 years, and the total for children includes two aged 0-5 years. Both those recorded infants/neonates were found during roadworks before excavation. The two swords, one shield-boss and four spear-heads were not excavated from graves, but were salvaged from workmen; other goods were almost certainly removed and not handed in. The buckets total excludes a piece of yew wood not certainly a stave. of the only measurable skeleton, 703, was about two inches above the national average, but that is not exceptional (Egging Dinwiddy 2011, 99-100).
Despite the absence of a sword, some of The Shallows grave-goods are exceptional: not only the Byzantine pail Br 1, but also buckle Sf 5010, the silver brooch Br 3 and the weight Br 5. To have so many stave-built buckets is one of the outstanding features ( Table 3 for comparisons and contrasts with other nearby cemeteries); furthermore, one of the six, Sf 1013, used 'imported' wood, pine. One grave in Trench 1 may have had two buckets, Sf 1008 and Sf 1013, but as with the double-shield burial in that muchdisturbed trench, absolute certainty is impossible. On the other hand, no doubt exists about the discovery inside Sf 1008 of a glass vessel, Sf 1008ii, which seems unique, and emphasizes the role that these containers may have had in feasting, which was an important social activity (Lee 2007). The Byzantine pail may have had a similar use in England, even if its original function was for washing (Pearce et al. above). No-one in Breamore would have known what the Greek inscription on it meant, though the letters might have been regarded as possessing special power, like the runes added to the Chessel Down vessel. The hunting scene would also have needed explanation to a north European audience; the animals were probably too exotic for craftsmen to use as inspiration.
The Byzantine pail is only the third with an inscription and hunting scene found in England. Harris (2003, 165-70 and 179-80) pointed out that there are no examples of them in Francia; they do not have a 'down-the-line' distribution via Italy and southern Gaul, or along the Danube-Rhine, which suggests that they were not routinely traded. One possibility is that they arrived by sea directly from the Mediterranean into western Britain, like pottery (Campbell, E. 2007, 126-32), and were then taken eastwards as loot or tribute. Perhaps more likely is that they were in some way associated with the sixthcentury diplomatic contacts that probably saw at least a few people from England at the Byzantine court, who could have brought them back as gifts (Campbell, J. 2000, 77-9). Objects arriving in England by this means would have spent less time travelling than items coming by 'down-the-line' exchange, so could have been buried closer to the time of their manufacture than some other imports. Deposition dates as early as c. 550 for the Breamore and Chessel Down pails may therefore be quite acceptable. Certainly pails seem to have had special status in sixth-century England; the Chessel Down example was in a grave that contained not only the three square-headed brooches, but also gold braid, a gold ring and other objects showing that the burial was of a woman who had been exceptionally well provided for (Arnold 1982, 26-8 and 66). The 'Bromeswell bucket' is not from a grave, but the Tranmer House cemetery is very close to, and perhaps the predecessor of, the great Sutton Hoo complex. A Byzantine pail without incised decoration sold in 2011 was said to have been found 'in the early 1990s near Ipswich' (TimeLine Auctions 24/6/11, lot 669), a description that could be applied to Tranmer House and Sutton Hoo, and may be another example of the demand for east Mediterranean imports.
Byzantine vessels, incised or plain, are much less frequently found than 'Coptic' bowls or Celtic hanging-bowls. That many of those have been found in graves of the highest status, as at Sutton Hoo and Prittlewell, in the seventh century may be an indication that the supply of the Byzantine white-metal coated and inscribed pails had ceased by then, justifying Harris's argument that they arrived through different exchange mechanisms. That also suggests that what seems to have been a 'Coptic' vessel, recorded as found on the coast near modern Bournemouthnot otherwise known as an Anglo-Saxon landing-placein 1805, and, if correctly illustrated, likely to have been buried or discarded in the late sixth or seventh century (Eagles and Ager 2004, 92), did not arrive by the same means or at the same time as the pail at Breamore.
The weight, Br 5, is another object of a type not previously found in central southern England. The punch-marks, which are now indistinct, have been recognized on some of the other weights, including the reused Roman coins that feature in most sets, but the perimeter dots are more unusual, apparently occurring in England only on one of the sixth-century weights from Watchfield, Oxfordshire (Scull 1990, fig. 5, no. 15). However, the circumferences of higher-value Roman coins usually had them, and they reduced the chances of clipping or shaving down the edges. The Breamore weight is only the third English find to have a central perforation, perhaps surprisingly, as to have carried weights on a string would seem a sensible precaution. The other two are both from the long-running cemetery at Castledyke South, North Lincolnshire: no. 5, weighing 8.25 g, and no. 9, a thin disc weighing only 0.7 g (Scull 1990, fig. 6). If put together, those two would weigh 8.95 g, only slightly less than the Breamore weight's 9.04 g, which could indicate that 9.1 g was also the target weight at Castledyke. 9.04 g is a slightly better fit with the Byzantine weight standard based on a unit of 1.52 g than with the Merovingian 1.32 g (6 x 1.52 = 9.12, 7 x 1.32 = 9.24). Scull has shown that the weight-sets found in England allowed for gold coins of both systems to be checked, and that some individual weights in Anglo-Saxon graves were copied from Byzantine ones, even if not actually imported (Scull 1990, 197). The Breamore weight therefore shows familiarity with gold coins and the likelihood that someone expected to handle them at some time, if not actually in Breamore itself. It is less likely that it was a Byzantine import into western Britain, as the pail might possibly have been; unlike the incised vessels, weight-sets are found predominantly in Kent, which had strong links with Francia, where most sixth-century gold coins found in Anglo-Saxon England were minted.
Although its place of manufacture cannot be pin-pointed, the copper-alloy buckle Sf 5010 was certainly brought into England from somewhere in the area dominated by the Franks (Marzinzik, above). Not found at The Shallows, but elsewhere in Breamore, a copper-alloy mid fifth-/mid sixth-century bridle-fitting with red glass insets was probably made in the west or central Mediterranean area and was therefore also probably channelled through the Franks, and a garnet-inlaid belt-mount was probably made by one of them (Eagles and Ager 2004). The bucket Sf 1013 may have had a more northern origin, because of the pine-wood used for its staves, but the glass bowl Sf 1008ii argues for connections with the Isle of Wight, and beyond there with Kent, both sixth-century power-houses with access to fine imports, the latter also manufacturing high-quality ornaments like the small square-headed brooch, Br 4. Other Anglo-Saxon objects reported from Breamore and the Charfords, such as a saucer brooch and two button brooches, are of types which have wider distributions and would not be unexpected in any cemetery in southern England (e.g. Suzuki 2008, 104, 107), any more than would a pierced fourth-century coin, quite possibly from a necklace (these metal-detected finds were recorded by Chris Gifford, to whom we are grateful for this and other help). Those finds are late fifth-or sixth-century, and all are unlike the fifthcentury material now being identified in eastern Hampshire (Stoodley 2011). The western part of the modern county seems initially to have had a different Anglo-Saxon settlement trajectory.
Interpreting the cemetery Its location, its artefacts and its multiple burials all suggest that The Shallows cemetery was not an ordinary one. The number of weapons from it might favour the view that it was a 'warrior graveyard' associated with the honoured dead slain in a particular battle. Against taking this to extremes is the presence of children, one and probably two too young to have borne weapons. Furthermore, the metal-detected brooches suggest women's presence. Weapon injuries were not noted on any of the bones, though the poor survival conditions of those could mean that battle damage would not be identifiable. A battle is indeed recorded as having taken place near-by: at more or less the same time as the cemetery was being investigated, work on place-names was reinstating the claim to historical veracity of the King Natanleod named in the Anglo-Saxon Chronicles under the year 508. For long it had been accepted that he was a fiction created to explain the place-name Netley, 'wet wood' (Coates 1989, 121-2), and that no such person had existed. Breeze, however, has shown that the place-name is likely to be an uncomprehendingly devised compound, from 'Natan l eod', or Prince Natan, with l eod being Old English for 'prince' in poetry (though merely 'man' in prose), and Natan being comparable to British names like Nechtan or Pictish Naiton, and one remembered by his Anglo-Saxon enemies (Breeze 2000). They were led by Cerdic and Cynric, and the place of their victory was recorded as Cerdices ford, which Aethelward's late tenth-century version of the Chronicles says was on the River Avon, and would therefore seem to be Charford, a place-name now perpetuated by North and South Charford (see Introduction). There are the two Domesday manors called Cerdeford, a name which could derive from OE ceart, 'rough ground', though seems more likely to be a contraction of Cerdices.
This seemingly straightforward picture is clouded by a 'doubling' of the annal, as the success of Cerdic and Cynric was repeated under the year 519, but without mention of 'Natanleod' and with the earliest manuscripts calling the place of victory Cerdices leah. Repetitions and name variations underline that early dates given to the annals are late constructs giving artificial precision to myths and traditions (Swanton trans. and ed. 2000, 14-17 for the annals; Yorke 1995, 32-3). So battle(s) recorded in the Chronicles may derive from an origin story, and even if the 508 and 519 dates should not be taken as precise, they are not inconsistent with sixth-century archaeological dating. Cerdic and Cynric may not have been involved in anything that happened in the area, however, for the origins of the Gewisse family probably lay in the Upper Thames valley, not on the south coast (Yorke 1995, 33-6). Furthermore, Cerdic's name has a British not a Germanic origin, so a straightforward conflict between incoming Anglo-Saxons and established Britons cannot be taken for granted. In any case, that a battle took place in the Breamore area, and was remembered as an important one, with the Wessex kingdom's founders later credited with the victory, is not implausible.
An alternative is that a battle never happened, but that the cemetery was remembered as a place where a special group of people had been buried. As time passed, actual knowledge of them faded, and their origin needed explanation; a battle was the stuff of stories, told to explain why The Shallows cemetery was still thought of as different. Cerdic may never have been in the area, but as it became accepted that his success had led to the foundation of the West Saxon kingdom, so people wanted to associate themselves with him, rather than with anyone with a less resonant name. Perhaps it was expedient for the Avon settlers to expunge the name of their own founder-hero, whose resonance had finally failed to protect them against Gewissan dominance, and to give the victory over Natan to Cerdic and Cynric.
Another model that cannot be absolutely precluded is that the relatively narrow timespan, argued above on the evidence of the archaeology, is actually an overestimate, and that this was a 'crisis cemetery', caused by an outbreak of disease, for instance. However, would people buried because of some horrible mortality have been given such grave-goods? Perhaps, if those goods did not reflect status in life, but rather a need to appease the evil spirits that had entered the dead people, by honouring them in their burials. If the perceived imbalance between adult males and females was a real one, a crisis that affected the former more than the latter, and also carried off children, must be envisaged. In such circumstances, multiple burials might be more likely, although it might also be expected that the graves would have been both closer together and smaller, to reduce the time and effort in digging them, with bodies flexed to be fitted in. The variation in burial alignments does not favour a 'crisis', either, because although few early Anglo-Saxon cemeteries show neat rows, burial within a really short timescale might have produced more uniformity as well as closer spacing.
An interesting argument has been put forward that the Tranmer House cemetery in Suffolk from which it is likely that the Bromeswell Byzantine 'bucket' derived might be classified as having been for 'warrior-retainer communities or even household troops in "royal" service' (Fern 2015, 210). Like Breamore, a high ratio of weapon-graves was found, though with cremation as the rite more favoured for women's bodies, of which no trace was found at The Shallows. The proximity of Tranmer House to Rendlesham, a royal centre by the early seventh century, has a possible parallel to Breamore, if there is any validity in seeing the eleventh-century royal estate recorded in Domesday Book as having a Roman origin; if the big villa at Rockbourne was a power-base, its role could have passed to the hill-fort at Whitsbury where post-Roman work on the defences has been found (Ellison and Rahtz 1987, 72-5 and 80) and on to Breamore, where a royal house could have existed west of the church (Hinton and Worrell 2006, S04). What the Suffolk 'retainer' model lacks is demonstration that there was a contemporary cemetery for the 'royal' family, as Tranmer House only overlaps chronologically with the Sutton Hoo graves marginally, if at all; a contemporary 'royal' cemetery with markedly different practices and grave-goods has not been recognized. That also applies to Breamore, except insofar as a few graves on the Isle of Wight, such as the ones from Chessel Down and Carisbrooke previously cited, had an even more eye-catching range of objects with them.
Breamore may be better viewed as the cemetery of a 'frontier community' (Härke 2011, 14), which would explain the high number of weapons as well as the women and breamore early anglo-saxon cemetery children's presence. Such a group might have felt the need to express themselves in their members' deaths through ostentatious display, both of armaments and of goods acquired from far away and used for feasting. Fifty miles to the west, at Hardown Hill, Dorset, a group of spear-heads, one unique in having an inlaid symbol, and a copper-alloy brooch, might represent something similar, although Austin (2014, 64-6) has recently shown that they were probably deposited as a hoard rather than as grave-goods, perhaps to await recycling rather than as an act of ritual. How so many Anglo-Saxon objects reached so far west, and remained as a group, remains hard to understand in terms of friendly contacts between incoming and indigenous cultures, and watchful antagonism might still be predicated.
A 'frontier community' may have needed to be especially careful of itself, but still have had fruitful contacts with other groups and peoples. Eagles and Ager (2004, 92-3) argued that the modern boundary between Hampshire and Wiltshire, which is here the north boundary of North Charford, was a 'frontier' between Jutes to the south and the Wilsaete people to the north, before both were incorporated into the burgeoning West Saxon kingdom. They stressed that the River Avon could have been a navigation route, if only for simple log-boats, though it should be noted that the scale replica of the seventh-century Sutton Hoo boat built by the late Edwin Gifford was able to sail and be rowed quite effectively on the modern river channel. Eagles and Ager stressed trade and exchange, explaining why a number of Frankish objects have been identified to the north in Wiltshire: the Breamore weight, Br 5, is strong supporting evidence for this model.
A 'frontier community' interpretation of The Shallows encourages a speculation that developments to the north, with Charlton Plantation showing people with affiliations looking northwards rather than southwards, made the lower Avon valley increasingly uncomfortable, and its settlers withdrew. This would be consistent with their originating on the Isle of Wight, for the island that was so rich in sixth-century grave-goods became much less so in seventh-century gold (Leahy's 2010 map has only two gold objects from the island, despite the metal's increased usage generally); its colonists in Breamore could not obtain the goods that they needed to be able to offer their neighbours. Bede, writing in the early eighth century, recorded how people of Jutish origin were on Wight and the coast opposite, but was more specific only about the 'Meonware', the folk of the River Meon, east of Southampton Water (Colgrave and Mynors 1969, 50-1, 372-3), a later reference to the New Forest being Ytene as the sole evidence of their presence on the west side (Yorke 1995, 37-9). A possible scenario is therefore that 'Jutes' won control of the Avon valley against the British Natan's people early in the sixth century, but lost it to Gewissans within a generation or two, the first indication that 'Jutish' power was on the wane. Whether the Bargates cemetery represents their last, impoverished stand, or the first appearance of 'West Saxon' people in the area, remains a moot point. However, although Anglo-Saxon expansion has never been seen as a single onward-marching process, because of recorded set-backs such as the British victory at the Battle of Mons Badonicus, what has not been so often considered is that different Anglo-Saxon culture-groups may have been suspicious of, and sometimes hostile to, each other, well before the documented power-struggles between the kingdoms of the seventh and eighth centuries. The Shallows may therefore be offering a new understanding of sixth-century conditions.