How Roman Britain Became England

How Britannia Became English

Educated British people are generally aware that for centuries up until about 400CE, England and Wales south of Hadrian’s Wall was Roman Britain, or Britannia, and that in the following centuries it fell into the hands of the Anglo-Saxons. What is much less well grasped is how that transition took place, or when.

The first key to a clearer understanding of that process is the political structure of Roman Britain in 400. Britain was divided into four distinct Provinces, and history took fundamentally different course in each of them over the quarter-millennium after the Roman legions evacuated in 409.

Each Province in turn was comprised of a number of civitates, districts for local governance, about the size of two counties. The seminal event of 409 was that the leaders of the civitates rebelled, demanding freedom from the burden of Roman taxation. The Roman Emperor Honorius could not spare troops to put down the rebellion, because the Empire was hard-pressed by Germanic invaders from across the Rhine frontier into Gaul (France and the Rhineland). So he withdrew his legions from Britain and told the civitates to defend themselves.

Twice afterwards, in 418 and in 429, Rome sent forces to Britain to aid Roman Britons fighting against raiders from beyond its borders, with some successful effect; but when aid was asked for again in 446, the Roman General Aetius ignored the request, again essentially because his forces were already stretched beyond capacity trying to defend Gaul.


Britannia Secunda

In two of the Provinces, Britannia Prima (Wales and the West), and Britannia Secunda (the North), the civitates did see to their own defences thoroughly and effectively, and with little or no contribution from foreigners. Military logistics of the fifth century were favourable for autonomous governance units the size of a civitas.

The four civitates of Britannia Secunda, plus tribal lands that had been divided from them by Hadrian’s Wall with military but not sociological logic,morphed over the following century into three kingdoms, all of which feature strongly in later history: Rheged (approximately Ayrshire, Dumfries-and-Galloway, Cumbria, and maybe Lancashire); Gododdin (Lothian, the Borders, Northumberland and Durham); and Elmet (North, West, and South Yorkshire, the Derwent valley of Derbyshire, and probably northern Nottinghamshire north of Sherwood). Elmet appears to have combined the small civitas of York with the much larger one surrounding it whose Roman city was Aldborough. All three kingdoms were ruled by successions of Roman British, i.e. Celtic, kings. Britain’s most famous bard Taliesin wrote poems in praise of two of these kings, Urien of Rheged and Gwallawg of Elmet, in the third quarter of the sixth century. Urien temporarily extended Rheged’s boundaries eastwards across north Yorkshire at that time.

Undoubtedly, for well over a century after the departure of the Roman legions, the territory that had been Britannia Secunda remained in the control of kings who were Celtic British by tribe and culture, influenced by a slowly fading Romanised past. But in approximately 547, the English warrior-king Ida disturbed the far north by occupying and fortifying Bamburgh and conquering a coastal part of the Gododdin kingdom as his Kingdom of Bernicia. This initiated a half-century of skirmish warfare, highlighted by the British defenders calling their English enemy leader in the 570s “Fflamdwyn”, “Fire-Setter”. As of 580, the English had been driven out of all except their coastal fortress.

The tide turned, though, and over the next fifteen years or so a new kingdom of Northumbria, formed by a union of Deira (East Yorkshire, see below) and Bernicia, conquered most of Britannia Secunda east of the Pennines. York fell in 581; Catraeth (Maiden Castle near Catterick), famously, probably in the 590s. Rheged remained independent until 633, and culturally, especially religiously, influenced its English-speaking neighbour either side of 620, when Rum was king of Rheged and Edwin of Northumbria. In 633, however, the male line of Rheged’s dynasty failed and the kingdom merged with Northumbria by royal marriage of Rheged’s daughter.


Britannia Prima

Looking next at Britannia Prima, the Roman Province south of the Mersey and west of a line from the Staffordshire Moorlands to the Hampshire Avon and Bournemouth, as in Secunda, the civitates morphed into kingdoms, and their considerable sense of British cultural unity did not translate into any political unity. On the contrary, small-scale inter-kingdom fighting became normal. It would be misleading to call western Britain “Roman Britain” any more by the mid-sixth century, these former civitates that were fully independent and squabbling kingdoms with no governance links beyond their borders and no Roman empire to link to anyway, but the term “sub-Roman” can be used with fair justification, for their elite were Romanised in cultural style, wrote (when they did) in Latin, and kept up strong trade links to Byzantium.

There is evidence of several areas near the western coast becoming conquered or heavily influenced by Irish intruders. There is none of any area that had been part of Britannia Prima coming under English control prior to 577. The Great Raid of 473 broke into the Cirencester civitas; in 490 King Arthur of Lindinis (mostly Somerset) won a famous victory over English invaders at the Battle of the Badonic Hill, i.e. the hill in the small Spa civitas of Bath; for the next eighty years, the kings of western Britain were at peace with their English neighbours.

But in 577 the agriculturally rich kingdom of the people known to the Romans as the Dobunni was lost at the Battle of Dyrham. Its area, covering all of Gloucestershire and Worcestershire east of the Severn, West Midlands County, and western parts of Oxfordshire and Warwickshire, plus the petty kingdom that had been the Spa civitas for Bath, became the English kingdom of the Hwicce (whose name is remembered by place-names such as Wychavon, Wychwood, and Bromwich). Very slowly, more territory came under English rule in the next one and a half centuries or so, with notable English battle victories at Chester in 611, Pengwern (?Shrewsbury; in ?634), Bradford-on-Avon in 652, Penselwood in south-east Somerset (?South Cadbury, a.k.a. Camelot) in 658, and Somerton in 733. However, the Westcountry west of the Parrett, western Herefordshire and Wales remained under the rule of British kings into at least the ninth century.


Flavia Caesariensis

By contrast, in the Province of Flavia Caesariensis, which comprised approximately East Yorkshire, Lincolnshire, Leicestershire, a little of eastern Warwickshire, and parts of Nottinghamshire and Derbyshire, such little evidence as there is is consistent with a wholly peaceful transition from Roman to English rule. There are not even signs of a short period of Celtic British independent kingships in-between, except in western Derbyshire where the recording in 676 of a semi-independent Mercian territory of Pecsaetna[i] attests to a territory around the Peak – a Peak district, the successor state to the small Roman Spa civitas of Buxton – that was not settled by English people.

Documentary evidence is only that of absence. Both British and English sources enthusiastically recorded battles – the English, only ones they won. Yet none are recorded in this Province. It is apparent from Bede, however, that these areas were all under English rule from long before the eighth century. East Yorkshire was the English kingdom of Deira since well before 547. Lindsey was an English kingdom since the fifth century. Leicestershire and English settlements close to the Trent were known as the Mierce (meaning Borderland, cognate with modern German maerchenland and English ‘marches’), later Mercia. Bede referred to North and South Mercia, divided by the Trent. Archaeological evidence from excavations near Hessle in East Yorkshire supports the view that the transition in Flavia was peaceful. The Roman city of Lincoln seems to have ceased to function within a few decades of 409.


Maxima Caesariensis

The story of the remaining Province, Maxima Caesariensis, the south and east of the island, is very different. Here, there was warfare from the mid-fifth century onwards. Here, there was destruction of Roman cities and slaughter in the streets, imprecisely and sorrowfully recorded on the British side by Gildas and occasionally gleefully and with variable reliability in the Anglo-Saxon Chronicle. Here, the course of history was affected by diplomatic marriages and savage treachery. English warriors were employed as mercenaries under Roman command, and they remained after 409, soon becoming dominant forces in some civitates – Winchester’s, probably, on the archaeological record; Kent, famously by treaty between the British ruler Vortigern and the English warrior-commander Hengist, in about 450; Essex and Sussex (which included most of Middlesex and of Surrey respectively), reported to have been as ransom for Vortigern after his capture, possibly in or soon after 470.

Nennius records that Vortigern came to power in 430.[ii] Power to what extent? The battles and political dealings attributed to him involve Essex, Kent, Sussex and Hampshire. De-facto Governor (hence Gildas’s title for him ‘tyrannus’, i.e. ‘non-legitimate ruler’) of MaximaCaesariensis Province would be the answer that best fits the geographical record. Whether he could exercise any influence in Britannia Prima can only be speculated.

Vortigern’s son Vortimer won a battle against Hengist at a location called by the British Set Thirbagail, and by the English Aylesford, which is on the Medway near Maidstone, recorded by both sides: in Nennius, chapter 44; and in the Anglo-Saxon Chronicle, for 455CE. The battle next recorded in chapter 44 of Nennius, at which the Saxons were defeated and “fled to their ships”, may likewise be the same as the one recorded in the Anglo-Saxon Chronicle for 465.

Alas for Britain, Vortimer then took his army across the channel to help the struggling Roman Empire in Gaul, where in 470 they were wiped out in Burgundy (c.f. comparable events in 1940), or anyway disappeared from the record. Vortigern regained power, but was displaced by Ambrosius in or before 479 following the disastrous Saxon raid across southern Britain in 479.

During the fifth century, the civitas of the Iceni, in Norfolk and Suffolk, and the one whose Roman city was Water Newton near Peterborough, which later became the kingdoms of the Gyrwas covering approximately Cambridgeshire and Holland, and perhaps Kesteven, also came within the arc of English rule from Christchurch to Beverley which Gildas refers to as “our enemies in the east”. One Roman source says Britain “had passed into the hands of the Saxons” by the early 440s. He may have meant these eastern civitates plus the Flavia Province.

Historians do not believe that the battles which the Anglo-Saxon Chronicle says were fought by the Gewissae (later known as the West Saxons and then Wessex) ever occurred. But the date for the founding of their kingdom, 495, may be genuine. It is consistent with archaeological evidence of late-fifth century English settlements at Abingdon and at Dorchester-on-Thames. It is well known that the first kings of the Gewissae, Cedric and Cynric, had British and part-Irish names, and it is possible that these men were British rulers of the Silchester civitas which chose from that time to rely on English warriors for its defence.

Thus it appears that by the start of the sixth century all of Maxima Caesariensis was under English rule, except for the large civitas whose Roman capital was St. Albans, which included Hertfordshire, Bedfordshire, Northamptonshire, Buckinghamshire and eastern Oxfordshire. This civitas, though ceasing to be governed from and maybe not controlling St. Albans, became the British kingdom of Calchfynydd. This name means ‘chalk hills’ and appears in English translation in the Mercian Tribal Hidage of 672 as the tributary kingdom of the ‘Ciltern-saetna’, the people who live in the Chilterns (the chalk hills). Calchfynydd remained British until it lost the Battle of Bedford in 571 and forfeited the fortresses of Aylesbury, Benson (a military site also today, the RAF base on the left bank of the Thames in mid-Oxfordshire), and Eynsham (NW of Oxford and close to the Calchfynydd-Dobunnic border). The remnant of Calchfynydd, basically Northamptonshire, may have come effectively under English conquest when they killed King Cadog at Bannaventa (the Roman town near Weedon Bec) in 580.


The Non-Uniform Process of Anglicisation

This brief survey shows that the Anglicisation of Britannia was neither a uniform process of conquest, nor a uniform peaceful transition. Rather, there were four distinct stories, substantially but not precisely coterminous with the four prior Roman provinces, each story radically different in time and in process from the others. It is these contrasts that are the primary cause of confusion about how and when Britannia became English and the running disputation as to whether the change was essentially a conquest or a peaceful transition.

In the centre-east of Britannia, one Roman Province and a couple of adjacent civitates became English-ruled within the first generation after the Roman legions departed in 409, without any recorded warfare – quite possibly by the voluntary choice of their first sub-Roman rulers. Most of another Province, in the south-east, became English through force of arms in the thrust and counter-thrust of warfare during the second half of the fifth century. A third Province, in the north, became English through political and military power over the years 581-633. Parts of the fourth, the western, Province plus an adjacent civitas from the south-east (Calchfynydd) were conquered piecemeal over nearly three centuries (from 571 to 838), but other parts (Wales and Cornwall) never became incorporated into Anglo-Saxon England.

Thus by 500 a considerable chunk of eastern and south-eastern Britain was under English rule, the area behind the coast from Bournemouth to Beverley, divided among a dozen young kingdoms; but the bulk of the country – all of the north, all of the west, and the centre as far east as Bedfordshire and Buckinghamshire – was comprised of sub-Roman Celtic-British kingdoms. This was still the geographical reality in 570.

During the next eleven years English kings conquered British territories equivalent to about nine counties, mostly in the Midlands. But in 582, even after these losses, all the regions west of the Pennines, the Severn and Selwood Forest remained under British control, and there was little further change in the following half-century. Only after the early 630s, with Powys’s loss of Pengwern, the defeat of Cadwallon of Gwynedd after his brief conquest of Northumbria, and the merger of Rheged into Northumbria, can it be said that English kings ruled over most of what was to become England.

The exceptions, lands outwith English control beyond the 630s, were Ergyng (the part of modern England south and west of the Severn), ruled by a British king known to the Mercians as Merewalh (which means ‘illustrious Welshman’), Lindinis (Somerset, western Wiltshire, northern Dorset and northern Devon), and most of Kernow (Cornwall and south Devon). Some of these areas remained under Celtic British rule for centuries; Cornwall into the tenth century.


Plenty of people publish on the internet, or in a few cases in books, their discoveries of aspects of the process by which Roman Britannia became a patchwork of English kingdoms, including me in earlier posts on this blog about King Arthur, Somerset, and Lindinis in the fifth and sixth centuries, and indeed through to the eighth century. But I trust and believe some readers will find it useful to have this summary of the process, of the overall picture of the change across three centuries, gathered here in one relatively brief narrative.



[i] In the Mercian Tribal Hidage. In that document, the ending –saetna was used for some petty kingdoms that were not part of Mercia but were scheduled for paying land taxes to Mercia, including Pecsaetna and Cilternsaetna. These areas were under English political influence but not areas of English immigrant settlement.

[ii] He transmits a record that from “Rufus and Rubelius” – obviously meaning the consuldom of Fufius and Rubellius in the first half of 29CE – to Stilicho was 373 years and from Stilicho to the start of Vortigern’s reign was 28 years, which in combination take the date to 430CE. He also names Valentinian and Theodosius as consuls that year, which is true of 430 (and also of 426 and 435).


Dindraithou = Dundry (Dundry hill in Somerset, near Bristol)

Dundry = Dindraithou

I recently visited Dundry. Dundry is just outside the city of Bristol, on the south side. The Dundry hillfort is an impressive defensive site, with a mighty ditch on its eastern side, deeper even today than a man, and extensive outlying walls beyond the inner perimeter. The flat land immediately to the south of the fort could easily be a paddock grazing ground for a king’s horses when he was in residence.

How little they know of South Cadbury who only South Cadbury know. The South Cadbury hillfort, (‘Cadbury-Camelot’) has been excavated, was evidently a major royal residence of the late fifth and early sixth centuries, and was indeed probably often used as a kaer (defended citadel, royal residence and feasting-hall) by King Arthur.

Dundry by contrast has never, to the best of my knowledge, been excavated. But it has an obvious strategic value: its views. Places on the Dundry plateau fairly near the hillfort[i] command a vista over the entire city of Bristol, across the estuarial valley of the River Avon westwards to the Severn shore, and to the north-east as far as the hills near Bath which I and others argue was ‘the Badonic region’ where Arthur’s most important battle success was won in 490[ii] and, I think, to Dyrham (site of the disastrous battle in 577 when the prosperous Circencester civitas was lost, along with the cities of Gloucester and Bath, their territories becoming the English-allied Kingdom of Hwicce). No enemy could have advanced on Lindinis from the north by land or water without being seen and heard afar by lookout warriors on the Dundry hill. A less impressive, but still militarily useful, overview of lands to the south – especially the south-west – is possible from a different point on the Dundry plateau.

Cador, the Governor of Lindinis in the mid-fifth century,[iii] had need to guard the northern frontier of Lindinis, primarily from the threat of Irish pirate raiders penetrating up the Bristol Channel and along the Avon, as they had devastatingly done in 398, and also from the possibility of Saxon adventurers intruding from the north-east. Dundry makes great sense as a base from which he would have done so.

Dindraithou, which I regard as the ancient name for Dundry, appears twice in the historical record. The more important reference is in the Life of St. Carantoc, [iv] which names it as the royal residence of both Cador and Arthur which the saint visited. The other reference is in the list of 33 cities given in the Historia Brittonum; the significance of that is that Dundry Dindraithou was a contemporary or remembered important British royal kaer at the (unknown, perhaps seventh century) time of compilation of that list. (Cadbury-Camelot, there called Penselwood, is also on the list.)

In my post on Cador (9-3-13), I mentioned several places which were fortified in the fifth century for which name and/or date evidence point to his being the commander responsible. Dundry should be added to that list. So too, its name suggests, should Cadbury Heath. Cadbury Heath is now a decidedly ordinary suburb of Bristol, but its location is above the Avon, just a mile to the north of the river. It is a little surprising to find the Cad- name on the north side of the Avon, since the river itself would have been the natural boundary between Lindinis and the Cirencester civitas. But Cador was a military man with a genius for strategic defensive planning, and it is easy to suppose that he would have wanted to ensure that his men could keep watch over the Avon water-transport route from both sides, and that he had the force to take command of Cadbury Heath – with or without the approval of his opposite number in charge of Cirencester, about whom nothing is known.


[i] On the road to East Dundry.

[ii] For the basis for identifying 490 as the battle’s date, see my post here of 22nd Sep 2013.

[iii] See my post here of 9th March 2013.

[iv] See my post here of 4th March 2013.

Map of King Arthur’s Lindinis (Somerset) Kingdom

This map is great – except that, the Google I can use obviously takes the line representing KingArthur’s Lindinis kingdom’s boundary along today’s roads rather than along hilltops where in reality it must have been. Also, accuracy is limited by the number of reference points I can use. In particular, the line from Beaminster (marked M on the map) to Iwerne Minster (marked N) should be along the N Dorset Downs watershed, not dipping southwards to Dorchester.

The Real King Arthur

This blog piece has taken five hours / fifty years to make. Five hours to compose – and, behind that, fifty years of study and love, visits and maps, intellect and intuition. I have been in pursuit of the real King Arthur since I first visited Glastonbury, and then South Cadbury, in the 1960s.

As a schoolboy in Taunton, I believed that he was one of ours – a man of what we knew as The Westcountry, a land that included Somerset, Devon, and Cornwall. Now I am confident. The pieces fit together. All the indicative evidence points in one direction. Arthur was King of Lindinis Civitas. This was northern Devon, most of Somerset, northern Dorset, and much of western Wiltshire.

Everything, every story or notice that has a geographical element and is maybe derived from a source in the fifth or sixth centuries, has a certain or probable Lindinis provenance. Modern writers placing Arthur in the north, Scotland, Lincolnshire, or the third century, base their thinking on no more than the odd piece of etymology (such as Camboglanna on Hadrian’s Wall as a claimant for Camlan, or Lincoln – Lindum in Latin – as a claimant for Linnuis), and bypass the inconvenient total absence of Arthur’s name from the Bonedd y gwyr gogledd (“Descent of the men of the north”), a document of genuine antiquity in which the names of the actual kings of the various regions of northern Britannia are given in several genealogical compilations; and any connecting of Arthur to Tintagel, Caerleon, Colchester or London is derivative of the fanciful imaginings of Geoffrey of Monmouth.

The earliest historical source on Arthur pointing in part away from Lindinis is the list of twelve battles in the Historia Britonnum, an eighth- or ninth-century compilation; however, all the battles there listed which are identifiably far from Lindinis are also identifiably ones that were unconnected to King Arthur (Chester; Wallop; Armterid – now called Arthuret, north of Carlisle; etc): they were fought by other warriors at dates spread across the quarter of a millennium after the British overthrew their Roman governors.

It is a lost cause to attempt to convince academic historians that there even was a real King Arthur, let alone that anything specific can be said about him. Such historians allow as evidence only documents that are beyond doubt contemporary to the events they attest, or copied unaltered from ones that were so, and physical archaeology and artefacts. In an era when very few people could write, when the main method of culture transmission was oral storytelling – primarily by bards in poetic song – and when even the most imposing residences and strongholds of the social elite were built in wood, evidence that attains academics’ thresholds is virtually non-existent. It is not only for King Arthur that this type of evidence is lacking: even the reality of (piecemeal) conquest of England by Englishmen (the “Anglo-Saxons”) is questioned for lack of such ultra-hard proof.

But the indications for King Arthur are diverse and consistent. What happened has often come down to us as told in allegory and story, a culturally natural form for the fifth and sixth centuries, rather than as the dry facts beloved of modern scholarship.

The one dry source is the Annales Cambriae, which names him as a Christian and as the victor of the Battle of Badon – at a date that has to be corrected to 490 – and as being killed in 538 at the Battle of Camlan. The other sources, which I have discussed in previous posts here, are oral traditions later written down, saints’ Lives, and an eighth-century battle list that straddles the border between eulogy and record-keeping.

In probable chronological sequence, the life of King Arthur thereby conveyed is this:-

  • As a youth when Cador was Governor of Lindinis Civitas, Prince Arthur met St. Carantoc at the mouth of the Doniford Brook, near Watchet in Somerset. They travelled to see Cador in his kaer (citadel) at Dundry (Somerset). Cador chose Christianity to be the official religion of Lindinis. He granted St. Carantoc a land charter to build a church at the commercially significant town of Carhampton (Somerset).
  • Arthur as the new man in charge of Lindinis had a culturally essential duty to fight and win raids on his borders. He won one such at Glein (possibly Clannaborough, in the Lindinis part of Devon), and four more skirmishes on the hills above the River Divelish (near Ibberton, on the border of the Lindinis part of Dorset).
  • He won a battle on the Bristol Channel coast (Somerset); and then another that may have been near Beaminster (on the border of the Lindinis part of Dorset) or near Mere (near the border of the Lindinis part of Wiltshire).
  • He was given Divine protection from death by a Christian Mystery initiation at Beckery on what was then the Glastonbury peninsula (Somerset).
  • He defeated English aggressors at the stunningly successful Battle of the Badonic Hill near Bath (Somerset) in 490.
  • He ruled Lindinis for half a century and became known as King Arthur (rather than Roman-style Governor). During the long peacetime that followed his Badonic victory, his largest citadel which he used often to host feasts for kings of other British kingdoms, was the Cadbury-Camelot hillfort at South Cadbury (Somerset).
  • He was killed at the Battle of Camlan at Queen Camel (Somerset) in 538.
  • He was buried by monks of Glastonbury Abbey (Somerset), most likely at Nyland, the island in the marshes (now the Somerset Levels) held sacred as the Gate to the Otherworld, rather than by the Old Church at Glastonbury itself.

After his death he was remembered as a great warrior hero by Celtic bards of the later sixth century (Aneurin, Taliesin), and as their own past warrior hero by the bard who eulogised a battle lost in the eighth century at Langport (Somerset). He was remembered in song particularly by minstrels of Brittany (culturally descended from immigrants from the Westcountry).

From them, his fame entered mainstream European literature as the fictionalised hero of “Arthurian legend”. The real King Arthur, successful Brittonic warrior leader, commander of a hillfort in SE Somerset beside the River Cam later known as Camelot, Christian ruler of one of the ten former Roman civitates of Britannia Prima, Lindinis, a name after his time corrupted to Lyonesse…. became transformed into the wizard-guided idealised model English (!) king, born at Tintagel, and governing the whole island of Britain (and then some) with the help of the Knights of the Round Table – voided in the public imagination of all lifetime connection to his true home among the hills, coasts, forests, and extensive brackish marshes of Somerset.

Dating the Battle of Badon


The Date of the Battle of Badon Controversy

The The date of the battle of Badon has long been the subject of controversy among students of the period. The uncertainty is due to the ambiguity of Gildas’s words in his book De Excidio et Conquestu Britanniae. He refers in the context of Badon to “the 44th year with one month elapsed”, i.e. a period of 43 years and one month. But it is not linguistically clear whether the period was before the Battle of Badon – between a previous event mentioned in his book and the battle – or after it, between the battle and the date of writing.


How to Interpret Gildas

A major difficulty with interpreting this as the period between the battle and the time of writing is: how could Gildas while he was still writing have known to the month at what date his work would be finished? The alternative that the 43 years and one month ended with the battle (which of course did have an exact date) leads to the question: which of three events Gildas discussed in the preceding chapters of De Excidio he intended as the starting point. It could reasonably be any of: the letter to Aetius (Chapter 20); the council of Britain deciding to invite some Saxons to fight for Britain (Chapter 23); or the coming to power of Ambrosius (Chapter 25).

Gildas could reasonably have known the month and year of any of these. But the coming to power of Ambrosius should be ruled out, because Gildas indicates that it happened after and in consequence of the Great Raid of 473. This would require the date of Badon to be 516 or up to a few years later. A date as late as this is inconsistent with Gildas’s saying that the people who were in positions of responsibility during the warring that culminated at Badon, including kings, officials and priests, had all died by the time of writing. De Excidio was published during the lifetime of King Maelgwn (Mailcun in the older spelling used in the Annales Cambriae [AC]), so not later than 548CE.[1] A period of thirty years or so is not sufficient for this to be true. Fifty years at least would be necessary.

Also, Gildas records that he was born in the year of the battle of Badon. Hagiographical writings record that he gave or sent a bell to St. Bride. (It is presumed that he was a bell-maker.) If true, this has to have been before her death in 524, and therefore Gildas’s birth can hardly have been later than, at a squeeze, 510. There is also the tradition that he founded Rhuys Abbey in the 520s. This date is not as assured as that of St. Bride’s death – but even if the foundation date were in the 530s it would speak in favour of a date of Gildas’s birth earlier than 516.


43 Years after Aetius or Hengist

It is much more probable therefore that Gildas intended us to understand his period of 43 years and 1 month to begin with one of the other two seminal events he reports, the letter to Aetius or the invitation to Hengist. The letter to Aetius is the most probable, as it is the only one of the three events in writing, and therefore with an exact date on its face. It is figured to have been sent in 446 because that was the year of his third consulship, to which the letter refers.

In any case though, if the council of Britain was meant it makes little difference to the calculation for the date of Badon. The council cannot have been long after the letter, for response to the emergency prompting the request to Aetius for military aid was urgent. It makes logical sense to postulate that the council convened late in 446 or early in 447, the year when according to Mageoghagan’s Irish annals Hengist actually arrived with his English warriors in response to that council’s request. On either basis, the resulting calculation is that the Battle of Badon was fought in 489 or 490.


Badon in 490 CE Makes Sense

This date for Badon and therefore for Gildas’s birth fits the information about Gildas’s life excellently. It is also fully consistent with the archaeological record, which shows a break in the English penetration of Britain approximately comprising the first half of the sixth century; and with the Anglo-Saxon Chronicle which gives no battles other than in Wessex between 490 and 550. Its supposed battles in Wessex have been interpreted as an ‘origin myth’ for the Gewissae dynasty, rather than factual truth.


The Annales Cambriae Date for Badon can be Amended to 490

It can be reconciled with the AC on a straightforward hypothesis: that the original Christian source said the Battle of Badon was “490 years after the Incarnation of Christ”, and that an early copyist mis-transcribed that as “490 years after the Passion of Christ’. Scribes before the introduction of the AD calendar sometimes expressed dates anno passio, and 27 needs to be subtracted from the number thus given to give the corresponding CE year.[2] The AC date for the Battle of Badon on a +445 calibration is 517CE, and 517 minus 27 is 490.

490CE as the date of the Battle of Badon fits well with all the evidence. It does not require the AC compilers to have made a complex derivation of the date of Badon, such as by interpreting De Excidio and the Anglo-Saxon Chronicle while not using other dates also found in the same sources. It is compatible with the Anglo-Saxon Chronicle, Bede, the archaeological evidence, the floruit of Ambrosius, the Irish annals, De Excidio, the events of Gildas’s life, and the other entries in the AC.

The 27-year Passion/Incarnation mistake does nothing to diminish the likelihood of authenticity of the AC entry. On the contrary, it enhances it, for the error is much more probable with a source that was originally written before the invention of the AD calendar in 529.

[1] On the calibration of the AC to the CE calendar by adding 445 to the AC’s internal year ticker count, which I advocate for its pre-565 entries (see my previous post here of 21 September 2013).

[2] The Historia Brittonum demonstrates that such mistakes happened: Chapters 16 and 66 show a similar mistake in reverse, 405 and 400 years after the Incarnation, respectively, being written when “years after the Passion” should have been written.

Calibrating the Annales Cambriae

The Need to Calibrate the Annales Cambriae

The Annales Cambriae (AC) are a terse record of British historical events from the fifth to the tenth centuries CE, with dates. But the dates in it are only on its own year count. It (in the “A” text) numbers years from its own year 1 to its year 533, with no mention of the CE calendar or any other calendar used elsewhere. A user of the AC therefore has to choose a calibration of its dates to years CE.


Calibration should not be uniform

Some AC entries are for events also recorded elsewhere whose dates are beyond dispute, such as the birth of St. Columba, his move to Iona, and the death of Pope Gregory. No calibration can map all of these undisputed dates neatly and exactly over the AC, not even when allowance is made for some ancient sources starting the year on 25th March and others on 1st September. The calibration most often used is +444, i.e. AC year 1 is deemed to be 445CE, year 100 is 544, and so on down to AC year 533 = 977CE. This makes sense, for it is usually regarded as the best-fit calibration for the later centuries covered by the AC.


However, it may not be wise to make a uniform calibration across the whole of the AC. Every translator of and commentator that I have ever read, prior to Gough-Cooper in 2012, converts the AC to the CE calendar by adding the same number to every AC year. As I showed, though, in my previous post here (19th September 2013), the AC is a compilation from several sources. Its compilers could have made a calibration inconsistency when meshing their sources.


On the best fit for the early years, AC Year 1 = 446CE

The best fit calibration for the AC entries for events prior to the mid-560s CE is +445. From the later 560s to at least the 630s the best-fit calibration is +442. Later, it is indeed +444. Gough-Cooper, imaginatively, endeavours to straighten the kinks by deeming the AC to have erroneously repeated the four years corresponding to 558-561CE and, perhaps influenced by McCarthy’s finding of missing years in the Irish annals when he synchronised them, he reckons the AC to have missed years elsewhere. The effect of his thorough attempts to calibrate AC dates as closely as possible with externally known dates is that his AC calibration changes frequently, wandering between +441 (in the mid-seventh century) and +448 (in the mid-sixth century).[1]


As described in my previous post here, I interpret the linguistic and geographical elements of the AC as indicating that all of its entries for events before the 530s CE and a minority of those for the period 530 to 613 originated from one source, which I suggest was Glastonbury Abbey, and that the majority of the entries for the 530-613 period were derived from other, Welsh, originals.


Dates before 529CE could not have been contemporaneously described in the AD calendar (now often called “CE” for “Common Era”), as this calendar was only invented (by Dionysius Exiguus of Constantinople) in that year. They could have been recorded by the regnal year of the Roman Emperor – though this would have been inhibited after 474CE by the demise of the Western Roman Empire; or on the Anno Mundi calendar (counting dates from the supposed creation of the world); or on a Year of the Incarnation (birth of Jesus) calendar non-identical with the AD calendar; or on a Year of the Passion calendar which started from the death of Jesus, which that calendar dated to 27 years after his birth.


Which one, or ones, might have been used at Glastonbury is not knowable. What matters is that the information from Glastonbury that the AC compilers used, particularly for events prior to 529CE, would have been calibrated to AC years independently from the data they acquired from Welsh abbeys for events that Christian monks would have registered from the beginning against AD dates. I reckon therefore that the shift in calibration in the 560s is due to an inexact meshing by the AC compilers of the dates provided by their different sources.


A +445 calibration for the early years of the AC makes it possible to make sense of a peculiar feature of the A text. The AC marked out years from its year 1 to year 533 because the Roman church’s method for the calculation of the date of Easter repeats on a 532-year cycle. So when any year is arbitrarily defined as Year 1, Year 533 is the first year of the next cycle. The last 23 years of the AC’s 533 are blank. Modern scholars say this was because these years were still in the future when the AC were completed: a fully rational cause, given the compilers’ motivation for marking out 533 years. But the first eight years are also without entries. This makes no rational sense at all: why not start by choosing for year 1 a year with an event in it, and add eight more blank future years at the time of presenting the AC to King Rhodri?


There was however a very important event in 446CE which could have inspired the AC compilers to choose it as the year 1 of the AC. It was the year in which according to St. Gildas, whom all agree is a very reliable source, the British warlord Vortigern invited English warriors led by Hengist to be mercenaries to help him defeat Irish and Scottish forces that were attacking Roman Britain. He gave Hengist sovereignty over the Isle of Thanet as his army base. Vortigern turned to Hengist for support because Aetius, the de-facto commander of Roman forces in the near Continent to whom he appealed first, was too pressed to be able to spare any military aid for Britain. Vortigern’s decision was retrospectively regarded as momentous because it was perceived as the first, or the most world-changing, of a chain of events that led to English rulers being kings over the entire arc of coastal Britain from East Yorkshire round to east Hampshire within half a century, and over most of what had been Roman Britannia by 633.


On the +445 calibration, Vortigern’s invitation to Hengist indeed happened in AC year 1. It is so recorded in the B text of the AC, with an entry that translates as “the coming of the English of Horsa and Hengist in the time of King Vortigern”. The C text has almost identical words.


It is not in the A text. But it is quite possible that it was there originally, for the corner of the manuscript where it could have been is torn off. If we interpolate that there was originally an entry for Vortigern and Hengist, and an intended +445 calibration to the CE calendar, the selection of the starting point of the AC becomes coherent.


Although a +445 calibration is the best fit for the early years of the AC, it does not result in all the entries being correctly dated. Several dates differ from ones identified elsewhere by one or two years (see Table). The variations may be due to such factors as the differences in when a new year was deemed to start, delays in reporting news or divergent interpretations of the meaning of phrases such as “December of last year”, and slippery miscopyings of Roman numerals, as well as to arithmetic errors in conversion of source data given in non-AD calendars.


Such discrepancies are not terribly important; and, besides, there is no guarantee that it is always the AC that is wrong and the English or Irish other source that is correct.


However, there are three entries for years prior to 613CE where the AC date is substantially different from that given elsewhere, and is wrong: the birth of St. Bride, the Battle of Badon, and the death of St. David. I can offer no explanation for the error on St. David, other than to presume a scribe’s mistake.[2] The date of Badon requires a substantial commentary which will be the subject of a separate post. As for the birth of St. Bride, as I explained in my post of 17th July 2013 (under the subheading “St. Columba and St. David”), a credible basis for this 16 year error in her year of birth is confusion by subsequent recordkeepers at Glastonbury Abbey between two visits she made, 16 years apart, to that holy shrine.


These mistakes apart, the AC shows a sufficiently close similarity to other sources where it can be corroborated to justify regarding it as a good source of dates – give or take a couple of years – of the events that it records where no such corroboration is possible, including those such as the Battles of Camlan and of Armterid whose existence is attested elsewhere but not their date. And I have sought to show that the AC dates should be translated to the CE calendar by adding 445 to the AC year count ticker number from the AC’s year 1 to AC118, and by adding 442 to the AC year count ticker number from the year of its next event record, AC126 until the mid-seventh century.[3]




Schedule of AC Events


Intended year CE as calibrated

Translation of entry

Actual date where one can be determined or estimated other than from the AC


Easter changed to Sunday by Pope Leo, Bishop of Rome



St. Bride was born



St. Patrick went to God



Death of Bishop Benignus



Bishop Ebur died in Christ



Battle of Badon in which Arthur carried the cross of our lord Jesus Christ for 3 days and 3 nights and the Britons were the victors



St. Columcille was born



Death of St. Bride



Strife of Camlann in which Arthur and Medraut were killed



There was mortality in Britain and in Ireland



Death of Ciaran



Great mortality in which Mailcun, King of Gwynedd died. 



Gabran son of Dungart died



Columcillae went out into Britain


568 *

Gildas died



Battle of Armterid



Brendan Byror died



Guurci and Peretur died



Battle against the Isle of Man



Dispositio for Danielus of the Bangors



Conversion of Constantinus to the Lord



Columcille died



King Dunaut died


593 (596**)

Augustinus [and] Mellitus converted the Angles to Christ

596 (601***)


Synod of Chester



Gregorius died in Christ



Bishop Dauid of Menevia



Dispositio for Bishop Cinauc



Aidan son of Gabran died



Death of Conthigirn



and [i.e. death] of Bishop Dibric



Battle of Chester where Selim of the sons of Cinan was killed and Iacob of the sons of Beli died




* Perhaps January 569. The Welsh sources for the AC may have used a 25th March New Year.


** 596 on a +445 calibration, as might be appropriate for this entry if the AC compilers took the data from a Glastonbury source. The gap in the manuscript between this entry and the other two events entered for AC year 151 could be due to those data having come from different sources.


*** According to Bede, Mellitus arrived four or five years after Augustine.

[1] Gough-Cooper ( ) gives respect to every blank year marked by “an” (for annus) in the manuscript. I regard the decade count in Roman numerals there as the reliable indicator of the intended years of events it records, to be given precedence over adding the empty an’s where these two methods conflict.


Gough-Cooper assumes that the “baptism of King Edwin” recorded in the AC is the same event as his baptism recorded by Bede. However, a different baptism of Edwin, by Rum, is recorded in the Historia Brittonum. An explanation that reconciles both the sources is that Edwin was baptised by Rum but, because Rum (a British prince of Rheged) was following the date-of-Easter custom upheld by the Celtic churches, the Roman church did not recognise that baptism as valid and Edwin allowed himself to be re-baptised by the Roman Bishop Paulinus. Bede vigorously favoured the Roman stance in the date-of-Easter controversy and therefore reported only Edwin’s baptism by Paulinus; the AC Northumbrian entries, by contrast, were written by a supporter of the Celtic side in that dispute – perhaps indeed Rum himself. Gough-Cooper also appears to have ignored McCarthy’s evidence that the correct date for St. Columba’s death is 593 and not the customarily believed 597.

[2] A note in the C text of the AC hints at the possibility that the original AC miscalculated this date. Marginalia on this manuscript in a different hand from the entry can be translated “in the year after the Incarnation of the Lord DC xlv”. 587CE is on my calibration Year 145 of the AC, i.e. Year Cxlv. It could be, therefore, that the person responsible for the insert in the C text margin had in their hand or their head both the correct date for St. David’s death and the year ticker structure of the A text, and – correctly – wrote C xlv as their calculation of the year of the AC that St. David’s death should be recorded in, but wasn’t. That a calculation was going on is shown by the writing in this marginalia of numeral letters that were crossed out. (The words “in the year after the Incarnation of the Lord” are of course out of place here.)

[3] With the exception of the conversion of the Angles to Christ.


Annales Cambriae: real evidence?

Arthur in the Annales Cambriae

Did King Arthur exist? Two references to him in a source known as the “Annales Cambriae” affirm that he did. The significance of them, however, is controversial. There is evidence that the Annales Cambriae (AC) were written as such in 953CE. The AC is a compilation of brief notes of historical events over the previous half-millennium, drawn from earlier sources. It does not date them to the CE calendar; rather, it uses its own, unique, counting from its year 1 (which I will endeavour to show here in another post is 446CE) to its year 533.

The first Arthurian entry is dated to AC year 72. In translation from the Latin, it says: “The battle of Badon, in which Arthur carried the cross of our Lord Jesus Christ for three days and three nights on his shoulders, and the Britons were the victors.” The other entry is dated to AC year 93. In translation from its mix of Welsh and Latin, it says: “The strife of Camlann in which Arthur and Medraut were slain. And there was mortality in Britain and in Ireland.”

Some historians, most notably David Dumville, assert that King Arthur never ruled anywhere. They say that his purported existence is the result of a story-making process of historicisation of a mythical figure. They dismiss the AC references as a mere consequence of that story-making. The necessary implication is that the Arthur references were added to the sources from which the AC was compiled several centuries after the dates they are marked against.

The difficulty with this is that it cannot be disproved. Manuscripts were often amended by copyists and their originals sometimes lost. Indeed, a clear example of this is the manuscript of the AC that historians call the “B” text (in contrast to the extant copy of the 953CE version regarded as pure, which they call the “A” text). The B text contains most but not all of the words of the A text – set against a different, and also unique, year count – plus a substantial number of additional entries. The B text’s first entry with Arthur – the Badon entry – contains most of the same words as the A text. However, it omits “the Britons were the victors” and it adds “king” before Arthur. The second Arthur entry – the Camlan record – is more substantially different. It says “the famous [inclitus] Arthur, king of the Britons and the traitor Modred” were killed. The differences undoubtedly indicate that the copyist who created the B text was influenced in his choice of words by Geoffrey of Monmouth’s twelfth century book “the history of the kings of the Britons”, a book now known to be a fabulous tale of British history rather than a historical account. The copyist probably thought he was improving the AC (as indeed with some entries he was), innocently unaware of the gay abandon with which Geoffrey invented stories and reshuffled names, dates, and places to formulate his creative masterpiece.

However, there is nothing textual to hint at any comparable copyist’s addition to the original sources of the A text. The Badon entry is in pure grammatical Latin just like the ones that precede it. The Camlan entry is in a mix of Welsh and Latin similar to some of the entries that follow it.

Nor is there any other evidence to suggest that the Arthur entries in the A text might be corrupt.

Significantly, there is nothing to honour Arthur above other district kings named in later sixth-century AC entries such as Maelgwn, Gabran, and Peretur. The AC entries are in this regard fully consistent with the “Arthur, king of Lindinis” hypothesis, provided that the battles of Badon and Camlann were fought in or near to the Lindinis territory. In earlier posts, I have given reasons for believing that indeed both were so located, Badon being probably Bath or possibly in NE Wiltshire – in either case, just outside Lindinis; and identifying Camlan as Queen Camel in SE Somerset.


Sources of the Annales Cambriae

Historians have customarily given little credibility to the AC as evidence on the history of the fifth and sixth centuries, because its pre-953 origins cannot be traced. What appears to have happened, though, is that around 950 King Rhodri commissioned some monks to compile a British historical record as a counterforce to the Anglo-Saxon Chronicle published half a century earlier, to demonstrate that the British, not just the English, had an honourable and long history. The monks gathered the data from several different monasteries which had kept records.

Both the language and the content of the AC change from one time period to another. These changes are signposts for changes of sources. The first eight entries, those for before the 530s CE, are all in grammatically correct classical Latin with British and Irish names Latinized, with case endings.

One of these entries is about the Battle of Badon. The other seven all concern matters of interest only to Christians. Six of them record the births and deaths of Christians famous in Ireland; the seventh records a papal decision about the dating of Easter.

The second period runs from the 530s until the mid-610s. There is an important linguistic change. Many of the entries in this phase show distinct Welsh traits: Welsh names of people and places without Latin endings, such as ‘mailcun’ and ‘armterid’; Welsh words, such as ‘gueith’ (battle) and ‘map’ (son of); Welsh variants of Irish names such as ‘gabran’ and ‘byror’.

The content also changes. Content concerning Britain rather than, or occasionally along with, Ireland predominates; and political events such as battles and deaths of kings are recorded as well as matters only of interest to Christians.

The change is not total. Some of the entries for the second period have the same style and content focus as those for the first period. This makes sense if the second period entries were drawn from two or more sources, one of which was the same source as the first-period entries and the other(s) were new sources.

The language, and more especially the content, change again after 613. In particular, the content becomes politico-military rather than Christian, and it includes several events in the English kingdom of Northumbria as well as ones in and affecting Welsh kingdoms. Clearly the AC compilers had another additional new source for this period.


Source of Earliest AC Entries: Glastonbury Abbey

It is striking that six of the eight entries in the first period concern people with links to Glastonbury Abbey. Glastonbury is the only abbey that existed during most of that period in the regions covered by the AC (now Wales, Herefordshire, and the south-west peninsula). Also, it still existed in the 950s, albeit under English control. It is therefore the most credible source for monks working in the 950s to have obtained pre-530 historical information from. [1]

The book by L.S. Lewis, ‘Glastonbury – Her Saints’, includes chapters with these titles:






These five people are the focus of six of the eight first-period AC entries; a seventh, the decision to keep Easter on Sundays, has no personal or geographical provenance; the only missing link is Bishop Ebur. According to William of Malmesbury, Sts. Patrick and Benignus were abbots of Glastonbury, and Sts. Bride and Columba were among its famous visitors. King Arthur is well known to have supposedly been buried there. Even Bishop Ebur could well have been in Glastonbury’s awareness, via its connection with St. Bride, for he founded a monastery called Beg Erin on an island in Wexford harbour that St. Bride visited before coming to Glastonbury. [3]


Later AC Entries from Welsh Abbeys

The change in the second period of the AC, from the 530s to 613, is that more than half the entries contain Welsh words, such as gueith (strife) and map (son of), or Welsh forms of names of persons and places, such as Gabran and Armterid, or have Welsh geographical provenance.

Close analysis shows that these entries are linked to one family, that of King Pabo Post Prydein (‘Pabo sturdy defender of Britain’). Pabo, or Peibio, appears to have been a king of Ergyng (part of Herefordshire plus the Forest of Dean). Among the people featured in second-period AC entries with Welsh markers, Dunaut was Pabo’s son. Dibric and Deiniol, who was Dunaut’s son, were cousins, both being grandsons of Pabo. Bishop Dibric founded Hentland Abbey in Ergyng. Bishop Cinauc was Dibric’s probable successor there; both bishops still have a church named after them near Hentland. According to the Bonedd y Gwyr Gogledd  (‘Descent of the Men of the North’) [4] Guurci and Peretur were cousins of Dunaut with a common grandfather Arthwys. These two warriors were among the victors at the Battle of Armterid, which accounts for this geographically unexpected battle being recorded in this part of the AC. In the mid-sixth century, St. Deiniol founded the abbeys of the two Bangors – Bangor-is-Coed, which apparently was abandoned after the Battle of Chester in 611, and Bangor Fawr, which later became Bangor Cathedral. King Maelgwn was the royal benefactor who granted the land for the founding of Bangor Fawr.

It is therefore a reasonable inference that a habit of recording brief historical notes was initiated by Sts Deiniol and Dibric at the three abbeys they founded. Records from Hentland were carried by fleeing monks to Llandaff Cathedral during later wartime, and would, we can presume, have been available there to the compilers of the AC. They would also have drawn on records at Bangor Fawr, perhaps including ones from Bangor-is-Coed likewise transferred there after 611.

There are specific word-forms in a few of these second-period AC entries that support the view that the original source from which the compilers copied them was contemporary or nearly so to the events they record. Mailcun is one: the sixth- and perhaps seventh-century form of this king of Gwynedd’s name, rather than the form ‘Maelgwn’ used in the ninth and later centuries. [5]

The –mail name-component also occurs elsewhere in the sixth century: Anglo-Saxon Chronicle entries for Brittonic kings killed by Saxons at the Battle of Dyrham in 577 include Farinmail and Commail. The form of their names can be taken as sixth century, as it will surely have been ossified in transmission by English writers between 577 and the writing of the Anglo-Saxon Chronicle, for these copyists would not have been influenced to update it by the natural modernising tendencies of Welsh-speakers and writers using their own language.

Similarly, the name-forms Selim (killed at the Battle of Chester) and Dibric belong to Selim and Dibric’s lifetime; the bishop’s name was later written Dyfrig while the king’s shifted to Selyf. (The mid-word letter b always became f in later orthography, for example with dybris becoming dyfrys[6] The earlier b and the later f both represent the sound for which a v is used in English. Likewise, m also became f.)


Conclusion: The Annales Cambriae Deserve Respect

So the words of the AC entries for the period before 613CE do deserve respect. It is realistically probable, though of course not provable, that they were copied from contemporary or near-contemporary sources, those from before the 530s, and also the later ones of similar language and provenance, being from Glastonbury Abbey, and those after 530 with Welsh language or provenance being from one of the abbeys founded in or soon after the 530s by St. Deiniol and St. Dibric. The calibration of the AC and the accuracy of its dates is a separate topic which I shall address here in another post.


[1] The conventional assumption is that the AC relied for its earliest years on Irish annals. For example, D.P.McCarthy, in his Chronological Synchronisation of the Irish Annals (Dublin 2005), says, giving no evidence, “it is clear that AC has taken these entries from an Irish source”.
Seven of the AC’s eight pre-530CE entries are for events also mentioned in those annals. Of these, five report births and deaths of Irish Christians (Bride, Benignus, Ebur and Columba), a sixth reports the death of Patrick, the Christian leader whose primary work was in Ireland, and the seventh records a church decision to always celebrate Easter on a Sunday, which mattered to all Christians. However, there are several problems with supposing an Irish source. Firstly, the words in the text of the AC do not match those of any of the several extant Irish annals. In itself this is not disproof of the hypothesis, as the existence of lost Irish annals (such ass the Liber Cuanach) is certain; but it does cast doubt on it. Secondly, it does not explain the one entry for an event that is not recorded in Irish annals, that for the Battle of Badon. The entry shows no obvious stylistic difference from the other seven, except for being twice as long as the longest of them. Thirdly, one of the entries notes the death of a bishop Ebur. This man’s name is spelt Iubair or Ibar in Irish annals; Ebur occurs only in the AC. It can be figured to be a Brittonic variant, like the variants of Irish names in the second AC period such as ‘Byror’ for ‘Birr’, rather than a word that would have been in any text copied from an Irish source. Fourthly, if these entries were copied from Irish annals, why were only these few selected? The Irish annals were all much more copious in their entries than the AC. The Annals of Ulster, for example, record the deaths of nearly forty people during the years corresponding to the first 84 years of the AC. The Annals of Tigernach, which only start half way through the period, and the Chronicon Scotorum, each record more than two dozen. The AC by contrast, records four.

[2] L.S. Lewis, Glastonbury: Her Saints (Orpington, 1985), p. ix. Chapter 6 is titled ‘St. Indract’. Indract lived centuries later than the people among whom Lewis places him and is not mentioned in the AC. Chapters 9 and 11 are titled, respectively, ‘St. David’ and ‘St. Gildas’. Both these Christians are mentioned in the second period of the AC.

[3] ‘The Monastery of St Ibhair [Ebur], Begerin, Wexford – Saint Brìghde, and other Celtic saints, sailed from here to [Glastonbury].’ P. David, Bride’s Mound.

[4] The Bonedd is part of the manuscript Peniarth 45, fos.291v-292r; the Harleian genealogies are preserved in Harl.3859 written in about 1100CE.

[5] ‘The Welsh name Mailcun, later Maelgwn’ – Sir E. Anwyl. Quoted in Taliesin, section 27, by Sir J. Morris-Jones. Y Cymmrodor (London, 1918), vol 28.

[6] J.E.C. Williams, The Poems of Taliesin (Dublin, 1987), p. 21, top line; and p. 56, ‘dybris’.