# Cadbury Castle

hillfort in Somerset, England, UK

South Cadbury and Sutton Montis, United Kingdom

Settlement · Fortification. British Bronze Age, British Iron Age, Romano-British period

## Story

• Cadbury Castle is a multivallate hillfort on Cadbury Hill in South Cadbury, Somerset, with a plateau of about 7.28 hectares inside an inner bank 1.2 kilometres long.
• Mesolithic flint and a small Neolithic settlement came before a Bronze Age return around 1000 BC, and a Yetholm-type shield was found in a ditch below the fort in 1997.
• From about 300 BC the hill held roundhouses, storage pits, metalworking and a complex south-western gate, and it was attacked in the first century AD.
• The defences were rebuilt from about 470, a hall 20 by 10 metres stood on the plateau, and a mint operated there from about 1010 to 1020.
• A tradition written down in 1542 linked the hill with King Arthur’s Camelot, and excavations from 1966 to 1970 traced the long sequence under that story.

## Visiting

Entry is free and the hillfort has open access all year, and boots or strong shoes are required where the paths are muddy.

## History

Cadbury Castle is a Bronze Age and Iron Age hillfort on Cadbury Hill in South Cadbury, Somerset, on the southern edge of the Somerset Levels and about 5 miles north-east of Yeovil. The limestone summit stands 153 metres above the sea. A plateau of 7.28 hectares is ringed by an inner bank 1.2 kilometres long, with four ramparts on the north-west and south and two surviving on the east.

The earliest finds are Mesolithic flints, more than 700 pieces recovered from the surface in about 1890, and a small unenclosed Neolithic settlement of pits, pottery and bones radiocarbon dated to 3500 and 3300 BC. Mary Harfield’s field-walking in the 1950s added Neolithic pottery and an early medieval sherd. A bank under the later defences is likely to be a lynchet from early ploughing.

After an apparent gap in the earlier Bronze Age, the hill was occupied again around 1000 BC. In 1997 a Yetholm-type shield, the first Bronze Age shield from an excavation in north-western Europe, was found in a ditch at Milsom’s Corner. It is too thin for combat, it had been stabbed, and a bone under the rim dates to about 940 BC, while the type is also placed between 1350 and 1150 BC.

From about 300 BC a stone enclosure stopped ploughing on the hilltop, roundhouses replaced rectangular buildings, and storage pits and metalworking filled the main years of the fort. The outer ramparts and a complex south-western gate belong to this phase. One reading of the banks ends with a final Iron Age repair in about AD 60.

The date of the Roman attack is disputed. One account places a violent taking around AD 43, Leslie Alcock placed the burning of the gate around AD 70, and Richard Tabor argues for AD 43 or 44, while the structural report puts fragmentary burnt human remains, weapons and dress fittings in the gateway late in the first century. Timber barracks, a small shrine and a possible masonry temple followed, and the gate burned again probably in the second century. There is no sign that this Roman use continued into the early medieval occupation.

From about 470 until some time after 580 the inner defences were rebuilt and a hall 20 by 10 metres stood on the plateau. The defended area was twice the size of any other known fort of the period, and pottery from the eastern Mediterranean reached the hill. A tradition first written down by John Leland in 1542 called it King Arthur’s Camelot. Geoffrey Ashe argued for a historical Arthur here, and that case has not been widely accepted.

Between 1010 and 1020 the hill held a temporary mint standing in for Ilchester, in the last years of Æthelred the Unready and the first of Cnut, and the top was refortified. An unfinished church foundation has been read as provision for the moneyers. In 1086 the hill appears to have lain in the barony of North Cadbury held by Turstin FitzRolf, and later use was mainly farming.

James Bennett dug here in 1890 and Harold St George Gray opened the south-western gate in 1913. Leslie Alcock’s excavations from 1966 to 1970, with a return in 1973, sampled about 6 per cent of the fort. Work since 1992 on the surrounding landscape has dated the pottery anew and has called the published Iron Age dating of the fort into question. The finds, including the shield, are in the museum at Taunton.

## Photographs

- [Cadbury Castle from Corton Ridge, foreground detail lifted out of shadow — Joe D, modified by JimChampion](https://upload.wikimedia.org/wikipedia/commons/5/51/Somerset_cadbury_castle_modified.jpg)
- [w:Cadbury Castle — Joe D](https://upload.wikimedia.org/wikipedia/commons/1/10/050326_073_somerset_cadbury_castle.jpg)
- [A Yetholm type shield found at South Cadbury. Dating from the bronze age. Photograph was taken in the Somerset County Museum in Taunton on 29-Oct-05. Found by — Gaius Cornelius](https://upload.wikimedia.org/wikipedia/commons/3/3c/Yetholm_type_shield%2C_South_Cadbury_01.JPG)
- [Map of earthworks at Cadbury Castle, Somerset, England. — Editor - W.Page](https://upload.wikimedia.org/wikipedia/commons/3/37/Cadbury_Castle_Somerset_Map.jpg)
- [Cadbury Castle, Somerset. From the Neolithic Age (3000BC) to the early 11th century, the fortress of Cadbury was in turn a military stronghold, centre of trade — Paul Dixon](https://upload.wikimedia.org/wikipedia/commons/1/1d/Cadbury_Castle%2C_Somerset_-_geograph.org.uk_-_2717.jpg)
- [Cadbury Castle, Somerset. Taken from the car park at ST 632 253 this shows the eastern flanks of Cadbury Castle — Paul Dixon](https://upload.wikimedia.org/wikipedia/commons/c/cc/Cadbury_Castle%2C_Somerset_-_geograph.org.uk_-_2721.jpg)
- [Cadbury Castle defences. A closer look at the deep ditch surrounding the summit of Cadbury Castle, taken at the western edge of the castle. — Paul Dixon](https://upload.wikimedia.org/wikipedia/commons/b/b9/Cadbury_Castle_defences_-_geograph.org.uk_-_2720.jpg)
- [Cadbury Castle. The Macmillan Way follows this lane. — Patrick Mackie](https://upload.wikimedia.org/wikipedia/commons/4/4f/Cadbury_Castle_-_geograph.org.uk_-_108254.jpg)
- [Cadbury Castle from Parrock Hill Despite appearances and its name, Cadbury Castle is not a chocolate mountain. These days getting to it is easy on the A303 — Graham Horn](https://upload.wikimedia.org/wikipedia/commons/1/12/Cadbury_Castle_from_Parrock_Hill_-_geograph.org.uk_-_486359.jpg)
- [Ramparts of Cadbury Castle looking south Cadbury Castle is a natural place to defend, and virtually impregnable when these ramparts were built. Looking south — Graham Horn](https://upload.wikimedia.org/wikipedia/commons/8/8c/Ramparts_of_Cadbury_Castle_looking_south_-_geograph.org.uk_-_486304.jpg)

## Sources

- [Cadbury Castle, Somerset](https://en.wikipedia.org/wiki/Cadbury_Castle,_Somerset) — Wikipedia
- [Large multivallate hillfort and associated earthworks at South Cadbury](https://historicengland.org.uk/listing/the-list/list-entry/1011980) — Historic England
- [Iron Age hillfort, South Cadbury Castle](https://historicengland.org.uk/listing/heritage-gateway/results/record/7ae57b43-b498-445d-8bed-e1e4f9b8947c/54552) — Historic England
- [South Cadbury Shield](https://wshc.org.uk/south-cadbury-shield/) — Wiltshire and Swindon History Centre
- [Cadbury Castle, Somerset: the later prehistoric and early historic archaeology](https://archaeologydataservice.ac.uk/archiveDS/archiveDownload?t=arch-1416-1/dissemination/pdf/9781848021808_all_72dpi.pdf) — Archaeology Data Service
- [Cadbury Castle: focusing the landscape](https://www.somersetheritage.org.uk/downloads/publications/150years/HES_150_Years_Chapter_5.pdf) — Somerset Archaeological and Natural History Society
- [Prehistoric pottery and associated radiocarbon dating from the hinterland of South Cadbury](https://eprints.bournemouth.ac.uk/35330/1/Tabor%20%26Darvill%20%282020%20PSANHS%20South%20Cadbury%29.pdf) — Somerset Archaeological and Natural History Society
- [Land at Castle Farm, South Cadbury: archaeological trial trench evaluation](https://doi.org/10.5284/1129257) — Archaeology Data Service
- [Cadbury Castle](https://somersetbirding.org.uk/sites/cadbury-castle) — Somerset Ornithological Society
- [King Arthur](https://en.wikipedia.org/wiki/King_Arthur) — Wikipedia
- [Dumnonia](https://en.wikipedia.org/wiki/Dumnonia) — Wikipedia
- [Durotriges](https://en.wikipedia.org/wiki/Durotriges) — Wikipedia
- [Roman conquest of Britain](https://en.wikipedia.org/wiki/Roman_conquest_of_Britain) — Wikipedia
- [Leslie Alcock](https://en.wikipedia.org/wiki/Leslie_Alcock) — Wikipedia
- [Somerset cadbury castle modified.jpg](https://commons.wikimedia.org/wiki/File:Somerset_cadbury_castle_modified.jpg) — Wikimedia Commons
- [050326 073 somerset cadbury castle.jpg](https://commons.wikimedia.org/wiki/File:050326_073_somerset_cadbury_castle.jpg) — Wikimedia Commons
- [Yetholm type shield, South Cadbury 01.JPG](https://commons.wikimedia.org/wiki/File:Yetholm_type_shield,_South_Cadbury_01.JPG) — Wikimedia Commons
- [Cadbury Castle Somerset Map.jpg](https://commons.wikimedia.org/wiki/File:Cadbury_Castle_Somerset_Map.jpg) — Wikimedia Commons
- [Cadbury Castle, Somerset - geograph.org.uk - 2717.jpg](https://commons.wikimedia.org/wiki/File:Cadbury_Castle,_Somerset_-_geograph.org.uk_-_2717.jpg) — Wikimedia Commons
- [Cadbury Castle, Somerset - geograph.org.uk - 2721.jpg](https://commons.wikimedia.org/wiki/File:Cadbury_Castle,_Somerset_-_geograph.org.uk_-_2721.jpg) — Wikimedia Commons
- [Cadbury Castle defences - geograph.org.uk - 2720.jpg](https://commons.wikimedia.org/wiki/File:Cadbury_Castle_defences_-_geograph.org.uk_-_2720.jpg) — Wikimedia Commons
- [Cadbury Castle - geograph.org.uk - 108254.jpg](https://commons.wikimedia.org/wiki/File:Cadbury_Castle_-_geograph.org.uk_-_108254.jpg) — Wikimedia Commons
- [Cadbury Castle from Parrock Hill - geograph.org.uk - 486359.jpg](https://commons.wikimedia.org/wiki/File:Cadbury_Castle_from_Parrock_Hill_-_geograph.org.uk_-_486359.jpg) — Wikimedia Commons
- [Ramparts of Cadbury Castle looking south - geograph.org.uk - 486304.jpg](https://commons.wikimedia.org/wiki/File:Ramparts_of_Cadbury_Castle_looking_south_-_geograph.org.uk_-_486304.jpg) — Wikimedia Commons

## Detailed history

Cadbury Castle is a Bronze Age and Iron Age hillfort in the parish of South Cadbury and Sutton Montis, in the county of Somerset. It stands on the summit of Cadbury Hill, a limestone hill of lias stone on the southern edge of the Somerset Levels, with flat lowland to the north. The summit is 153 metres, or 500 feet, above the sea. The hill lies about 5 miles, or 8 kilometres, north-east of Yeovil. The plateau covers 7.28 hectares, about 18 acres, and one structural account describes the same fort as a seven-hectare enclosure. The inner bank that rings the plateau is 1.2 kilometres, or three-quarters of a mile, long. Four ramparts survive on the north-west and south. Only two remain on the east, and the missing lines there may have slipped down the slope. Entrances stood at the south-west, the north-east, and the east. The eastern entrance was either later or unfinished, and it was blocked at some stage, while the north-eastern and south-western entrances stayed in use through the hillfort’s history. The interior is dominated by a summit plateau on the west, with steep slopes to the west and south. Terraced banks and ditches run around the hill, and trees stand on the flanks. The neighbouring plateau of Sigwells overlooks the fort. Scrub and tree growth on the banks have been judged a risk to the earthworks. The hill is a scheduled monument.

The name uses the Old English ending byrig, a word for a fort or a town. Eilert Ekwall, the Swedish scholar of English place-names, derived the first part of Cadbury from the River Cam, as with the nearby villages of West Camel and Queen Camel. Other scholars have suggested a personal name, Cador or Cada. Cadbury Castle is one of three places in Somerset that carry the Cadbury name. The others are Cadbury Camp, near Tickenham, and Cadbury Hill, also called Cadbury-Congresbury so that the sites are not confused. The county name Somerset most likely comes from Old English Sumorsæte, short for the people living at or dependent on Somerton, and it is first recorded in the law code of Ine, king of Wessex from 688 to 726. The centre of the county is the Levels, a coastal plain and wetland. Mesolithic hunters used dry points in that wetland, and the Sweet Track, an engineered roadway across the Levels, dates from 3807 or 3806 BC. Numerous Iron Age hillforts stand in the county. Cadbury Castle and Ham Hill are among those later reused in the Early Middle Ages.

Hillforts, or the enclosed places that Leslie Alcock thought the better name, were built in Britain mainly in the Iron Age, with earlier examples in the Neolithic and a few in the later Bronze Age. About 3,300 hillforts or similar defended enclosures are known in Britain. Most cover less than a hectare, and most of the rest fall between that size and about 12 hectares, so a plateau of a little over 7 hectares is large without belonging to the rare class of enclosures over 80 hectares. They cluster in the south and south-west, the Welsh Marches, the west coast of Wales and Scotland, and the Scottish border hills. Excavators in the first half of the twentieth century dug defences on the assumption that these places were chiefly military. From 1960, attention turned to the interiors. Hillforts developed in the Late Bronze Age and the Early Iron Age, around the start of the first millennium BC. It has been argued that they were military sites raised against invasion from continental Europe, that invaders built them, or that they were a military answer to social tension caused by a rising population and pressure on farmland. The dominant view since the 1960s has been that the spread of iron changed society. Iron ore lay in different places from the tin and copper needed for bronze, trade shifted, old elites lost standing, and power passed to new people. Barry Cunliffe, whose own excavations at Danebury in Hampshire ran through many summers and who has argued that Atlantic Europe shaped Celtic culture before inland Hallstatt traditions did, still gives population a role. He has said that the forts provided defensive possibilities for the community when the stress of an increasing population burst out into open warfare, that he would not see them as having been built because there was a state of war, that they would be functional as defensive strongholds when there were tensions, that some were attacked and destroyed, and that war was not the only or even the most significant factor in their construction. Another reading treats them as seasonal meeting places for farming families. The British Iron Age, from about 800 BC until the Roman conquest of the first century AD, was the age in which hillfort building increased sharply, iron was used for tools and weapons, and some southern communities minted coins and kept links with Gaul.

The earliest activity on Cadbury Hill is older than the ramparts. More than 700 pieces of Mesolithic flintwork, blades, microliths, flakes and cores, were recovered from the surface of the hillfort in about 1890. The earliest settlement so far recognised is a small unenclosed agricultural place, marked by pits and post holes with Neolithic pottery and flints. Bones from the site have been radiocarbon dated to 3500 and 3300 BC. A bank under the later Iron Age defences is likely to be a lynchet, a terrace left by early ploughing of the hilltop. Early and late Neolithic ritual pits may have marked the summit as special. A row of pits along the spine of the spur at Milsom’s Corner marked the way to and from the hill through woodland that was still mostly uncleared. Early Neolithic pottery in the Hembury style, abundant hazelnuts, a polished axe, and a quern that probably ground nuts as well as grain, belong to that early use. On one rampart cutting, an Early Neolithic occupation soil, through which pits had been dug, was covered by a slight field bank. A gully found in 1967 was thought to be part of a structure, but later digging did not confirm it. The nature of the Neolithic occupation remains uncertain, and the Neolithic and earlier evidence was not treated when the 1960s excavations were first published.

Mary Harfield, a self-taught archaeologist born on Shark Island in Australia in 1889, who trained as a nurse in London and later settled in Somerset, walked the fields around the hill in the 1950s. Her collection identified a range of ceramics, including Neolithic pottery and an early medieval sherd, among the Iron Age and Roman material that surface finds were expected to produce. Ralegh Radford used her work to argue that Neolithic occupation on the hill was significant. Her finds archive later went to the museum in Taunton. The presence of imported pottery from the Mediterranean, of about the fifth or sixth century AD, was a main reason for the decision to survey and excavate the hill in the later 1960s. Some of the diagnostic pottery of the early phases is credited to her collecting.

The later Bronze Age is the next clear occupation. Ovens of that period have been identified on the hill. Leslie Alcock’s reading of his own trenches was that, after an apparent gap in the earlier Bronze Age, the hill was lived on again in the centuries around 1000 BC and then continuously until at least the first century AD. The published structural account sets Early Cadbury at about 1000 to 300 BC, taking in ceramic assemblages 4, 5 and 6. In that span a cluster of buildings stood on a flat and relatively sheltered part of the plateau, with a small number of pits and a spread of occupation debris that included late Bronze Age metalwork. A few circular buildings and four-post and six-post rectangular buildings grew up along an east-west road, which probably ran into a hollow way climbing from the north-eastern entrance. Deposits that seem to come from metalworking, mixed with other debris, formed just north of that road and toward the centre of the excavated area. The industrial patch was cobbled from time to time and stayed in use through the Early and Middle phases. Before the first timber-revetted stone bank, the edge of the hill was marked by field lynchets, banks and fence lines. Any early gate in the south-western entrance has been lost to later erosion and rebuilding.

Below the fort, on a spur at Milsom’s Corner, the South Cadbury Environs Project found a bronze shield in June 1997, in a silt-filled ditch. It is a Yetholm-type shield, named from three shields found at Yetholm in southern Scotland. About twenty of the type are known from Britain and Ireland. This was the first Bronze Age shield recovered by excavation in north-western Europe, and the first of the type to come from an excavated context there. The shield is 665 millimetres across and made of sheet bronze only 0.6 millimetres thick, with a central boss and 25 rows of raised ribs and 25 rows of small bosses. Two tabs on the back took a strap. The hand grip was a bar of bronze hammered to sheet thickness in the middle and folded over, and metallurgical study of this grip and of six others of the same type found a core of pure tin. The metal is too thin for combat. Shields of this kind are read as objects of status, and they are usually found in wet places as ritual deposits. The South Cadbury shield is the exception: it lay in a dry deposit. It had been stabbed in three places, and a stake had been thrust through it. Whether the holes were made as a ritual killing at burial, or earlier, is not known. A bone directly under the rim gave a radiocarbon date of 940 plus or minus 110 calibrated years BC. One account places the manufacture of this form of shield in the period 1350 to 1150 BC, and another places the making in the twelfth century BC with deposition about two centuries later. The landscape study puts the deliberate deposition probably in the tenth century BC, perhaps associated with the ceramic phases called Cadbury 4 and Cadbury 5. Pollen in the ditch was poorly preserved. It did show lime and holly growing nearby. The shield was lifted in more than 75 kilograms of wet soil, kept damp for six and a half months, and conserved from January 1998. Only one face could be displayed. It is now shown at the museum in Taunton, with the other finds from the hill. A metalworking building and an enclosure about 2 kilometres to the south-east are roughly contemporary with the period in which such shields were made. On Sigwells, moulds for at least eight weapons of about that date lay in a pit.

By the middle of the second millennium BC, open swathes of ground from south-east to north-west on Sigwells were used for grazing, bounded by long ditches. Some of those ditches may have taken on a territorial role, marked by round barrows set over them. Another barrow on the narrow spur at Milsom’s Corner may have covered a flexed burial in a slatted, boat-like coffin, aligned on Glastonbury Tor, about 18 kilometres to the north-west. An enclosure straddling that spur cut through part of the burial. Its south and east ditches held a sequence of unusual bone deposits whose radiocarbon dates may run from the twelfth to the tenth centuries BC, overlapping dates tied to the late Bronze Age Cadbury 4 fabric. By then the hill had become a major focus in the landscape, and it remained one for several centuries before the ramparts were built. When those banks went up, a field system on Sigwells would have framed a view of the fort as a large yellow earthwork against green ground. West of the hill, ditches curved around a knoll and the foot of the spur to make a funnel into a corridor already more than 200 metres long before it climbed to the south-western gate. The extra distance emphasised the central place and the effort of approaching it.

Middle Cadbury, about 300 BC to AD 40 or 50, is the main period of occupation inside the fort, and it takes in ceramic assemblages 7 and 8. A stone enclosure with timber revetting was built around 300 BC, and ploughing on the hilltop stopped. Four-post and six-post rectangular buildings were gradually replaced by roundhouses. In the areas that were dug, roundhouses were built and rebuilt across the eastern plateau and beside the hollow way on the northern slope. They surely stood more widely, including in the quarry scoops left by construction of the inner bank. Most of the rock-cut storage pits belong to this period. Some seem tied to particular houses. Others were dug on the more exposed, and apparently unoccupied, western plateau. The industrial cobbling on the plateau continued, and parts of it were fenced off from the houses and the stores. The outer ramparts were built in this phase. At the south-western entrance a complex sequence of gates and guard chambers developed, and the rampart ends on either side of the roadway had to be revetted because use was wearing the road down. Ceramic assemblage 8 is a long-lived phase in the rampart sequence. It includes a major rearward extension of the stone and timber rampart and the start of a wall inserted into its back. Ovoid jars continue in quantity, straight-sided jars peak, proto-bead-rim jars are common, and the new forms include a bead-rimmed jar and the first bead-rimmed bowls, tempered with shell or oolitic inclusions. Decorated wares include saucepan pots and Glastonbury ware jars. Shell temper dominates, with a significant share of oolitic temper and a small presence of sandy wares.

A separate reading of the rampart sections, using the older numbered phases, starts with an undefended early Iron Age phase, Cadbury 5. Cadbury 6 was enclosed by rampart A, a timber-laced bank of clay and rubble. Rampart B, a stone bank revetted with timber and lias walling, defended Cadbury 7. Rampart C, much less well preserved, belongs with Cadbury 8. Cadbury 9A, in which Durotrigian material appears, overlies rampart C. Cadbury 9B lies on the tail of rampart D1 and was incorporated into D2. The last Iron Age phase, Cadbury 9C, belongs with the final refurbishing of rampart D3 in about AD 60. Roundhouses, storage pits, continuing metalworking, and the elaborate south-western gate all sit in this defended sequence. The intensity of occupation seems to have fallen toward the end of the last millennium BC, and it may already have been declining before the first century AD. In the later part of Middle Cadbury, carcasses of calves and cattle were buried in the part of the plateau that had been the industrial area. House plans are hard to recover. Medieval and later ploughing had cut most of the interior down to the bedrock, except where cobbled surfaces on the eastern plateau, and parts of the northern slope, protected stake lines and residues. Many hundreds of postholes could not be resolved into buildings, and few contained datable finds, so the plans that can be drawn are dated more by inference than by objects in the holes. Stake-walled roundhouses survived on two terraces of the northern slope. That kind of architecture would not show on a plough-eroded surface.

The people of this part of southern Britain in the late Iron Age are identified as the Durotriges, a Celtic grouping whose lands covered modern Dorset, south Wiltshire, south Somerset, and Devon east of the River Axe, with a possible extension to the western half of the Isle of Wight argued from a hoard found at Shalfleet in 2009. Their name is known from Ptolemy and from two inscriptions on Hadrian’s Wall recording the civitas of the Durotriges of Lindinis. One breakdown of the name reads it as fort-dwellers, which would fit a region of hillforts, though many of those forts seem to have gone out of use by 100 BC, long before the Roman army arrived. Another derivation takes the first element from a word for water and the second from a word for kings. Barry Cunliffe placed them on the cultural periphery around a southern core. They issued coins, simple and uninscribed, so no ruler’s name is known from the currency. The coinage was progressively debased, which has been read as economic contraction and growing isolation. Their cross-Channel trade, strong in the first half of the first century BC, when the potter’s wheel arrived, dried up in the decades before the conquest, and the outlet was at Hengistbury Head. Analysis of their pottery has been taken to show production increasingly centred on Poole Harbour. Thin sections of Durotrigian pottery from Cadbury itself show that some sherds match compositions from the Wareham and Poole Harbour area, while others that look the same to the eye come from a different, unknown source. Burial was by inhumation, often crouched in a shallow oval grave, with a last ritual meal. A young woman’s grave at Langton Herring, with a decorated mirror, has a radiocarbon date between AD 25 and 53. High-status goods found mostly in women’s graves, and a study of 57 genomes published in 2025 that found settlements organised on maternal lines with incoming unrelated men, are the grounds on which a relatively high status for women, and a matrilocal pattern, have been argued. Cadbury Castle is one of the places listed as a Durotrigian settlement, with Maiden Castle, Ham Hill, Hod Hill, Badbury Rings, Hambledon Hill and others. After the conquest their administrative centres were Durnovaria, modern Dorchester, described as the probable original capital, and Lindinis, modern Ilchester, whose earlier status is unknown and was raised by that role. To the west lay the Dumnonii. To the east lay the Belgae.

The Roman conquest of most of Britain began in earnest in AD 43 under the emperor Claudius and was largely complete in the south by AD 87. Julius Caesar’s expeditions in 55 and 54 BC had left a century of tribute, hostages and client kings without occupation. By AD 40 the Catuvellauni under Caratacus had overrun the Atrebates and driven out Verica, a king recognised by Rome. That exile was the pretext. Aulus Plautius led the invasion. The only legion directly attested in the force was the Second Augustan Legion, commanded by the future emperor Vespasian, who was born in AD 9, served as that legion’s legate in the British campaign, and much later, from 69 to 79, ruled as emperor and began the Flavian dynasty. The main landing is usually placed at Richborough, though a landing toward the Solent has also been argued. The army defeated British forces at the Medway and, by AD 47, held the country south-east of the Fosse Way. Suetonius records battles between tribes of southern Britain and the Second Augustan Legion under Vespasian. It was long thought that the Durotriges resisted strongly in AD 43 or 44. It is not known whether they, in particular, fought those battles. At Maiden Castle, a cemetery in the eastern gate was once read as a Roman assault and massacre. Later excavation and survey have made that reading unlikely. By AD 70 the Durotriges were already being drawn into the province.

At Cadbury the accounts of the attack do not agree, and both the dates and the readings have to stand. One account says the fort was taken violently around AD 43, and that the defences were slighted again later in the first century after a Roman army barracks was built on the hilltop. Leslie Alcock placed the severe violence at the south-western gate, with weaponry and burning, around AD 70. Richard Tabor argues for the initial invasion, AD 43 or 44. Another statement is that the hill was a place of vigorous resistance by the Durotriges and the Dobunni to the Second Augustan Legion under Vespasian. The structural publication of the prehistoric and Roman deposits places a destruction in the gateway late in the first century AD. With it is a spread of fragmentary and partly burnt human remains, weaponry and dress fittings, and the indications are of Roman military action against the people in the fort. A trail of human remains along the western approach, and extensive burning of the gate and the corridor leading to it, belong with that destruction. After it, a roadway was laid through the gate. The passage was refurbished in four phases, and a group of Roman timber barracks was built in the interior. The gate was finally destroyed by fire, probably in the second century. One building of the middle to late first century is taken to be a small timber shrine on the plateau, and there are indications of a Roman masonry building, possibly a temple. A separate description speaks of a possible sequence of small rectangular temples or shrines. There is no indication of later Roman activity, and so no continuity from that occupation into the early medieval reuse of the hill. Alcock’s own summary was that the Roman evidence was scant apart from a barracks block of the latter first century. Late Cadbury in the ceramic scheme runs from about AD 40 or 50 to about AD 400, covering assemblages 9 and 10, but the structures do not show the hill occupied as a fort through that whole span. Work since 1994 on pottery and radiocarbon dates in the landscape around the fort has called into question the dating of Early and Late Iron Age pottery in the final publication of the hillfort excavations, and it may require key events in that publication to be interpreted again. The published sequence is therefore one account, not a settled chronology.

The digging that produced this sequence has its own history. The first known excavations were in the late nineteenth century, by James Bennett, a local clergyman, in 1890, and little was written down. Harold St George Gray, born Harold Gray at Lichfield in 1872, trained under Augustus Pitt Rivers, became curator of the museum in Taunton and stayed until 1949, and dug at Arbor Low, the Glastonbury and Meare lake villages, Maumbury Rings, Avebury and Windmill Hill. In 1913 he cut trenches at the south-western entrance of Cadbury Castle and located a stone-built passage. His record of fieldwork also includes an excavation in 1922 at Cadbury Camp, the separate hillfort near Tickenham, and that season should not be folded into the South Cadbury gate. The 1913 passage was hoped, when the gate was reopened in the 1960s, to be the latest, early eleventh-century entrance, and to guide the diggers into the deposits. It proved of little help. The gate was the most complex Iron Age and later structure on the site.

Leslie Alcock directed the main campaign. Born in 1925, he served in the Royal Gurkha Rifles, read history at Oxford, worked briefly as Mortimer Wheeler’s deputy at Mohenjo-daro and as the first director of the Archaeological Survey of Pakistan, and then taught at Cardiff, where he excavated Dinas Powys before South Cadbury. From 1966 to 1970, with a return in 1973 to cut back the face of an existing section, he sampled about 6 per cent of the 7.28 hectare fort for the Camelot Research Committee. A soil-conductivity survey covered a quarter of the interior by April 1967 and was used to place trenches. Open-area excavation was preferred to the box method then associated with Wheeler. Geophysical survey was still unusual. The work examined the inner bank, the south-western gate, the plateau, and a strip of the northern slope. Alcock’s book of 1972 was a large interim account. He published the early medieval evidence himself in 1995. The prehistoric and Roman evidence was published in 2000. From 1994 until his death in 2006 he was patron of the landscape project around the hill. The earthworks were surveyed in 1993 by staff of the Royal Commission on the Historical Monuments of England, as a contribution to that final publication. Visitors to the 1960s dig were shown a plastic skeleton set in the same spot each afternoon, with a bucket for contributions to a welfare fund for the diggers. The scale of the publicity tied Alcock’s name, in the public mind, to Arthur.

After the Roman administration withdrew, the hill is thought to have been in use from about 470 until some time after 580. The innermost Iron Age defences were refurbished. A great hall, 20 by 10 metres, or 66 by 33 feet, was revealed on the plateau. The defended area was twice the size of any other known fort of the period, and the refortification of the latter fifth century has been described as unparalleled in Britain at that time. Sherds of pottery from the eastern Mediterranean show trade reaching far beyond Somerset. It has seemed probable that this was the chief fort of a major Brythonic ruler, with his family, his warband, his servants and his horses. The kingdom in which that ruler would have sat was Dumnonia, a Brythonic realm of the sixth and seventh centuries in the farther south-west. Its pre-Roman predecessor, the Dumnonii, had been centred in what is now Devon and had also held Cornwall and part of Somerset, perhaps as far east as the River Parrett, to judge by where Dobunnic and Durotrigian coins fall. In the Roman period a provincial boundary separated the country governed from Exeter from the country governed from Dorchester and Ilchester. After Rome, Dumnonia covered Cornwall, Devon and parts of west Somerset, and its eastern edge shifted west as the West Saxon kingdom expanded. The people spoke a Brythonic dialect ancestral to Cornish and Breton. Tin from the south-west had been exported since antiquity, and an apparent surge of late fifth-century Mediterranean imports across the region has been linked to a trade in metals toward the eastern empire. Imported pottery of that age has been found on many sites in the region, and the eastern Mediterranean sherds at Cadbury belong in that pattern. Christianity persisted in the kingdom, with late Roman cemeteries continuing and important churches at Bodmin, Glastonbury and Exeter. Gildas, writing in the sixth century, used a punning form of the regional name to rebuke a ruler called Constantine.

King Arthur is the name that later writing attached to this hill. He is a legendary king of Britain and a central figure of the medieval Matter of Britain. In Welsh sources he is a leader of the post-Roman Britons against the Anglo-Saxons in the late fifth and early sixth centuries. He first appears in the Historia Brittonum and the Annales Cambriae, texts of about three centuries after the time in which he is supposed to have lived, and most historians of the period do not treat him as a historical person. The Historia Brittonum, a ninth-century compilation, lists twelve battles ending at Badon, where he is said to have killed 960 men alone. The Annales Cambriae date Badon to 516–518 and Camlann, where Arthur and Medraut both die, to 537–539. Those entries were probably added in the tenth century and the Badon notice likely derives from the earlier list. Gildas mentions Badon and does not mention Arthur. Bede, the Anglo-Saxon Chronicle, and surviving manuscripts written between 400 and 820 do not name him. One historian’s judgement is that there may well have been a historical Arthur, but that the historian can as yet say nothing of value about him. Another judgement is that there is no historical evidence about Arthur and that he should be left out of the histories. Geoffrey of Monmouth’s twelfth-century history made Arthur a king who defeated the Saxons, and it supplied Uther, Merlin, Guinevere, Excalibur, a conception at Tintagel, the battle with Mordred at Camlann, and rest in Avalon. Chrétien de Troyes, in the 1170s, added Lancelot and began Arthurian romance. Camelot, Arthur’s later capital, is absent from the early material. It first appears in Chrétien’s Lancelot, as a court held after the king had left Caerleon, and most scholars regard it as fictional. Arguments for a real Camelot have run since the fifteenth century.

Local tradition, first written down by the antiquary John Leland in 1542, holds that Cadbury Castle was King Arthur’s Camelot. Leland’s itinerary, compiled in the sixteenth century, became a foundation for English local history. There is no evidence that the name Camelot was equated with Cadbury Hill before the sixteenth century, when the tradition is likely to have been invented. William Stukeley, the physician and clergyman who measured Stonehenge and Avebury and in 1723 drew the hill, captioned his engraving “Prospect of Camalet Castle”. Geoffrey Ashe, a cultural historian of the Arthurian legend and co-founder, with Ralegh Radford, and secretary of the Camelot Research Committee, argued that the hill was the base of a historical Arthur. His opinion has not been widely accepted by students of the period. He also argued that the career of the fifth-century commander Riothamus underlies a major part of Geoffrey of Monmouth’s Arthur. What he did claim from the digging is narrower and better supported by the trenches: Leland had picked a hill that really was refortified in the latter fifth century, in works that could not have been guessed from the earthworks alone, and that were then unparalleled in Britain. Ashe said there must have been a tradition about the hill and its powerful overlord, handed down from those centuries. A military reading that does not require Arthur by name is also on record. The hill could have served refugees and the south-western Britons of Dumnonia against attack from the east. If the tradition that Arthur was conceived at Tintagel were followed, he might have been a prince of Dumnonia using Cadbury as a stronghold on an eastern frontier. That chain is tradition and argument. The archaeology shows a large refortified hill, a hall, imported pottery and a date in the later fifth and sixth centuries. It does not name the ruler.

Between 1010 and 1020 the hill was occupied again, as a temporary mint standing in for the mint at Ilchester. The royal mint of Æthelred the Unready at the place called Cadanbyrig, in 1010, is likely, on documents and the place-name, to have been in the neighbourhood of South Cadbury. Æthelred II, king of the English from 978 to 1013 and again from 1014 until his death on 23 April 1016, ruled through a renewal of Viking attacks. Raids that were small in the 980s became large from the 990s. Tribute was paid to make armies leave. In December 1013 Swein Forkbeard of Denmark conquered England and Æthelred went to Normandy. He returned in 1014. The coin evidence from Cadbury spans the end of his reign and the beginning of the reign of Cnut, about 1010 to 1020. Excavation showed that the hilltop was refortified in this period and that it still had strategic weight. A foundation trench for a probable cruciform church was found, apparently never finished, and it has been read as provision for moneyers moved onto the hill for security in the early eleventh century. Later activity on the hill was mainly farming.

In 1086, according to the record of holdings made after the Norman conquest, Cadbury Castle appears to have been part of the feudal barony of North Cadbury held by Turstin FitzRolf. Turstin, also called Turstin le Blanc, was a companion of William of Normandy at Hastings in 1066, from Bec-de-Mortagne in the Pays de Caux, and the Domesday record shows him as the first holder of that extensive barony, with manors in Somerset and in nearby counties. Some small-scale fortification of the hill may also have taken place in the thirteenth century. Ploughing in the medieval period and after is what reduced the interior deposits.

Since 1992 the South Cadbury Environs Project has studied the fort and the country out to 8 kilometres. Local volunteers from the South East Somerset Archaeological and Historical Society and the Yeovil Archaeology and Local History Society worked with archaeologists from the universities of Bristol, Glasgow, Birmingham and Oxford. The work was funded by the Leverhulme Trust and the Arts and Humanities Research Council. Methods included gradiometry, electrical resistivity tomography, sampling of the disturbed ploughsoil, test pits, deeper excavation, and mapping in a geographic information system. The shield was one result. Another is a set of papers by Richard Tabor, with Timothy Darvill and others, on prehistoric pottery and radiocarbon dates from the hinterland. Fieldwork from 1994 to 2007 produced pottery from secure contexts, several of them dated by more than one consistent radiocarbon determination, from the Early Neolithic to the Late Bronze Age and onward to the Late Iron Age. Those dates refine Alcock’s ceramic sequence and, as already noted, put the published Iron Age dating of the fort itself in question. Petrology and chemistry of the fabrics have been used to trace raw materials and changing outside contacts, from Cornwall to central southern Britain. Pottery, being heavy, is treated as a poorer traveller than stone or metal unless the clay or the potter moved. A trial-trench evaluation in October 2024 on land at Castle Farm, South Cadbury, cut two trenches, each 10 metres long and 1.6 metres wide, on the footprint of a proposed slurry store. No archaeological features or deposits were present. A single fragment of green bottle glass of the late nineteenth or early twentieth century came from an overlying soil. The ancient sequence remains the one recovered on the hill and in the wider study area, not in that small footprint. The finds from the fort, from the late Bronze Age through the early Iron Age and including the shield, are displayed in the museum at Taunton.

Canonical page: https://getsitesee.com/places/cadbury-castle