The Hanging and the Verses
Seven panels of wool on linen, each a little over a metre square, hang in the long gallery of a house near Ilchester in Somerset. They were bought as a set in 1873 by Edmund Hale Purse, a Yeovil solicitor with a subscription to the Society of Antiquaries and no training whatever, who catalogued them as seventh-century British work and never revised the entry. The verses below were copied by the same hand onto the flyleaf of an estate account book, from a source Purse identifies only as the ballad, and which nobody has found since.
Purse was wrong about the panels by roughly six hundred years and wrong about the ballad by rather more, and the ballad convicts itself in its first line. It is written in fourteeners rhymed ABAB, a measure most commonly found in English verse of the sixteenth and seventeenth centuries: the metre dates the poem as flatly as the pointed arches date panel III. The wool is dyed in madder red and woad blue in the manner of the Bayeux Tapestry, which is embroidery rather than tapestry, was worked in England in the 1070s, runs to nearly seventy metres and captions all fifty-eight of its scenes in legible Latin. The first of the Ilchester panels carries a border of lettering that spells nothing at all; the other six are bordered with beasts, wheel-crosses, zigzags and plain bands, and attempt no writing.
The larger difficulty is the subject. There was no siege of Camelot, because there was no Camelot. The name enters the written record in the 1170s, in Chrétien de Troyes’ Lancelot, the Knight of the Cart, where the court is named once, in passing, is not described, and does not appear in all the manuscripts. Everything the hanging shows — the fellowship, the hall, the relics, the enemy at the gate — postdates the war it depicts by six or seven hundred years, and almost all of it can be traced to two men writing for Norman patrons. The ballad is set out below as Purse copied it, one stanza to a panel, in the order the panels hang.
The mist lay low on Cadbury when first the horns were blown,
and out of it a thousand spears came standing like a wood,
with banners hung on ashen poles, their beasts in madder sewn,
and one great gilded disc above the place where Arthur stood.
Panel I: the muster. The crowned rider is presumably Arthur, though nothing on the panel says so, and the great gilt disc carried on a staff above him has no counterpart anywhere in the Arthurian material — it is not a shield, not a wheel-cross, and not explained.
He drew his line before the gate upon the rising ground,
with towers at his shoulder and the spires behind his head;
he turned his horse among the ranks and let them close him round,
and raised the blade above them all, and not a word was said.
Panel II: the line drawn up, the sword raised, spires at the King's back. Every shield in the block carries a device, which is heraldry; heraldry does not exist as a system until the seals of about 1135 to 1155, and the shields on the Bayeux Tapestry, seventy years before that, are painted but not yet armorial.
He called them to the pillared hall and stood them in a round,
that none might sit above the rest nor any go before;
Sir Lancelot with his own grief, Sir Gawain sword-renowned,
and Merlin at the King’s left hand, who had advised the war.
Panel III: the fellowship called together. The arches are pointed and the vaults ribbed, which is thirteenth-century Gothic; and in the one scene that ought to show the Round Table there is no table at all, only knights standing in a ring.
Then Arthur drew the water-sword the Lady gave to him,
and lifted it, and all the field went white as bone with light;
and round it in the empty air stood wards and sigils grim
that Merlin hung about the walls and would not name outright.
Panel IV: the sword lifted, with the ground behind it filled by roundels — interlace, wheel-crosses, one device with vanes like a mill. Nothing in the Arthurian material accounts for them; they are field-filler of the kind a metalworker puts round the boss of a brooch.
The gate stood open all that day and no man barred the way,
and horse met horse before the wall, and neither host would yield;
he drove the long pale sword through mail and helmet where it lay,
and nine hundred and sixty men were counted on the field.
Panel V: the fighting in front of the gate, which stands open and unbroken. There is no breach anywhere in the wall and no ladder anywhere in the picture. Behind the mêlée runs a curtain wall with crenellated flanking towers and a two-storey gatehouse — high medieval castle architecture doing duty for a refortified hill-fort of timber and earth. The stanza's nine hundred and sixty is Nennius's number for Badon, lifted whole.
The Saxons went off through the mist before the day was gone,
and left the water-meadow full of men they did not take;
and one girl held a dark horse still; the white one bore him on,
and no one spoke till evening fell, for any dead man’s sake.
Panel VI: the morning after, the spears still upright and the dead not yet lifted. The bird over the towers is the carrion bird that Welsh heroic verse feeds on every field it describes.
They set the masons on the wall before the frost came in,
and cut new oak, and raised the gate, and paid for it in gold;
and Arthur walked the river-road with all his household kin,
not knowing what the river knew, nor what the year would hold.
Panel VII: the rebuilding, felled timber stacked at the foot of the walls and the King out on the river road. Round mural towers in dressed ashlar are thirteenth-century work — the same century that built the castle at Tintagel, and for much the same reason.
What the Sources Actually Say
Everything a fifth- or sixth-century Briton left us about the war against the Saxons would fit in a pamphlet. Gildas wrote De Excidio et Conquestu Britanniae — a sermon, not a chronicle, in three parts, the second of which simply abuses five named kings. He records the British victory at the siege of Mons Badonicus, says he was born in the year of it, and does not say who commanded. He does not mention Arthur. Nothing else names him for three centuries, unless Y Gododdin does: one line of that elegy praises a warrior called Gwawrddur by saying he was no Arthur, which would be the earliest reference of all, and which Jarman’s edition lists as a possible interpolation.
The twelve battles arrive with the Historia Brittonum, compiled around 829 — the mouth of the Glein, four more on the Dubglas in Linnuis, the Bassas, the wood of Celidon, the fort of Guinnion, the City of the Legion, the Tribruit, the mountain called Agned, and Badon last, where 960 men are said to have fallen in one charge to Arthur and to nobody else. That is a poem’s arithmetic in a historian’s clothes, and most of the places cannot be found on a map. The Annales Cambriae, written down about 1100 from a mid-tenth-century original, add two lines and no more: Badon in 516, and in 537 the strife of Camlann, where Arthur and Medraut fell. The annal does not say they fought each other. It does not say who won.
Then the invention begins, and it is possible to date it fairly precisely. Geoffrey of Monmouth finished the Historia Regum Britanniae about 1136 and supplied Uther, Tintagel, Caliburnus forged in Avalon, the continental campaign and the end — claiming throughout to be translating an ancient book in the British tongue that no one else has ever laid eyes on. The Round Table is Wace’s, added in the Roman de Brut in 1155 for the stated reason that Arthur’s barons would not accept a lower place than one another. Chrétien contributed Camelot and Lancelot in the 1170s. Even the painted shields the panels hand out to the fellowship are late: heraldry becomes a system somewhere in the seals of about 1135 to 1155, and the arms on the Bayeux Tapestry, seventy years earlier, are decoration that nobody inherits. Malory assembled the lot about 1470 and Caxton printed it at Westminster in 1485 in 21 books and 506 chapters, having first written a preface disagreeing with his own author about where Camelot was. Two copies of that printing survive, and Oskar Sommer reprinted it page for page in 1889, so the whole of it is free to read. Malory’s Arthur keeps a fifteenth-century court in fifteenth-century plate, which is exactly the anachronism the Ilchester panels commit in wool.
What Is Actually There
Two sites carry the weight of the archaeology, and neither of them produces an Arthur. Cadbury Castle in Somerset is a hill-fort of 7.28 hectares whose innermost Iron Age bank was refortified in the late fifth century — the largest fortification of its date known in post-Roman Britain — with a timber hall twenty metres by ten on the summit and eastern Mediterranean pottery in the soil. Occupation runs from about 470 to after 580. The antiquary John Leland wrote down the local claim that it was Camelot in 1542, and Leslie Alcock dug it from 1966 to 1970, reporting the seasons in the Antiquaries Journal under titles nobody remembers and then publishing the whole of it in 1972 under one nobody forgets: By South Cadbury Is That Camelot…. He took the new chair of archaeology at Glasgow the following year and spent the rest of his career trying to be known for something else.
Tintagel is stranger. The castle on the headland was put up by Richard, Earl of Cornwall, in 1233, deliberately old-fashioned in style, for no better reason than that Geoffrey had staged Arthur’s conception there a century earlier — a literary attraction built on a literary claim. Underneath it lies a real fifth- and sixth-century settlement of large, well-made buildings whose occupants drank eastern Mediterranean wine off North African and Gaulish tableware. The inscribed slate found in 1998 shows that Latin literacy survived the end of Rome in Cornwall. The name cut into it is Artognou.
The one Arthurian object anyone can go and look at is a fake, and an unusually honest one. The Round Table at Winchester is five and a half metres across and weighs 1.2 tonnes, and it hangs on the wall of the Great Hall built for Henry III between 1222 and 1235. Dendrochronology puts its timbers at 1250 to 1280: a genuine medieval table, made for a Round Table tournament in the reign of Edward I, by men playing at a fellowship a Norman poet had invented about a century earlier. The paint is later again, ordered by Henry VIII: twenty-four knights named around the rim, Galahad first and Mordred last, and Arthur not one of them. The crowned figure enthroned above them is widely said to wear Henry’s own face, which is a claim the table’s own record of the commission does not make.
As for the host in the ballad, the Saxon settlement has stopped looking like an invasion at all. Germanic speakers were already in Britain before Roman rule ended, mostly as soldiers; what followed looks less like conquest than a warrior elite whose language and dress the locals took up. The genetics, when they arrived, split the difference. A 2022 study of 460 early medieval genomes found people in eastern England carrying up to 76 per cent continental northern European ancestry, put the share in the modern English at 25 to 47 per cent, and found intermarriage everywhere it looked. Nobody withdrew into the mists. They married in.
