The Mammoth Book of King Arthur

by Mike Ashley

  In order that we can see how this legend emerged I want to follow through all of the surviving ancient documents that cover this period, no matter how dubious.

  The principal documents are the De Excidio et Conquestu Britanniae (The Ruin of Britain) by Gildas, the Welsh Annals (Annales Cambriae), the Historia Brittonum, usually credited to Nennius, and the Anglo-Saxon Chronicle. There are also the genealogies, a few ancient poems such as the Y Goddodin, and the lives of the saints, none of which is contemporary and few of which are reliable. More reliable are ancient inscriptions on stones, but these have been subject to weathering and destruction.

  There may well have been more documents at some stage. When Geoffrey of Monmouth wrote his History of the Kings of Britain in the 1130s, he referred to a “certain very ancient book” which he had consulted. But such chronicles as may have been kept in the fifth or sixth centuries would mostly have been compiled and retained in monasteries, and these were subject to regular attack from the Vikings for over two hundred years, let alone the ravages of time and other dangers such as fire and flood. The library at Glastonbury Abbey was all but destroyed by fire in 1184 and one can but weep at what irreplaceable documents were lost.

  The Venerable Bede, regarded as the father of British history, was a dedicated researcher and may have had access to some of these lost documents, but he relied heavily on Gildas for his coverage of the fifth and sixth centuries and, like Gildas, makes no mention of Arthur. That may by itself seem significant, but Bede was not that interested in events before the arrival of St. Augustine, and would not have looked further into ancient British history. He was, however, the first to provide the name Vortigern. His primary research relates to later years, which means that England’s foremost historian of the Dark Ages can provide no help with the story of Arthur.

  One can live in hope that some long lost document may surface in an ancient archive, but until then we have to work with what we’ve got and hope that archaeology may help substantiate or further define the world in which the events took place. I shall look at each of these sources over the next few chapters, which will also help flesh out a chronology so we know where in time to place Arthur.

  Before doing that, though, it is worthwhile listing here the various pedigrees that survive in the ancient records. These are far from reliable – in fact at times they are wholly misleading – and they are almost impossible to date. But we will encounter many of these people as we travel through the other documents so it is worth acquainting ourselves with them here and trying to get at least a rough chronology. This will also show where the various individuals named Arthur or Artúir or Arthwyr appear.

  GENEALOGIES AND KING LISTS

  One of the key essentials to identifying Arthur is to place him in a specific period of time, along with his contemporaries. Without that we will get nowhere. In the next few chapters I will go through the various chronicles and see what timeline they suggest. Here, in order to acquaint ourselves with the names and territories that later emerged in Britain, I shall set out the various “royal” pedigrees and make some attempt to date them. Several authorities, not least Dr David Dumville, one of the undisputed experts on the Dark Ages, have demonstrated the difficulty in trying to get any chronology from the pedigrees for reasons I shall cover in a moment. So I start with a huge caveat that of all the sources covered in the next few chapters, these are amongst the most unreliable. But it seems to make sense to start with the data which is the least in focus and fine tune it as we go along.

  The British pedigrees and regnal lists are extensive and survive in a wide variety of ancient documents, though none contemporary with Arthur’s period. There are three major sources and many minor. The major ones are known as the Harleian MS. 3859, Jesus College MS. 20 and Bonedd y Saint. The Harleian manuscript is part of the text which also includes Nennius’s Historia Brittonum, but the oldest surviving copy with the genealogies dates from about 1100. The name Harleian comes from the original collection, now housed in the British Library, established by Robert Harley (1661–1724), first Earl of Oxford. The surviving copy of Jesus College MS. 20 (now in the Bodleian Library in Oxford) has been dated to around 1340, and was probably drawn from a copy completed about a century earlier. The Bonedd y Saint, or “Lineage of the Saints”, survives in many copies and versions, but the oldest dates from the end of the thirteenth century. This is held in the National Library of Wales and is known as Peniarth MS. 183. Both the Anglo-Saxon Chronicle and Nennius also provide a number of pedigrees.

  The pedigrees start with a contemporary descendant and work backwards through the generations. For instance, the first genealogy of the kingdom of Gwynedd listed in Harleian MS. 3859 begins:

  Uen map iguel map catell map Rotri map mermin map etthil . . .

  . . . “Owen son of Iguel son of Catell” and so on. For the purposes of this book I want to reverse them into the order we usually understand genealogies, reading from earliest to latest.

  Clearly these genealogies are so far removed in time from the Arthurian period that their accuracy is spurious. This is not simply because they may have been corrupted by scribal errors, but because there has been ample time for genealogies to have been fabricated. The primary reason for producing a pedigree is to identify a priority of descent, and thereby a claim to a title or land, and later rulers would have no compunction in having their scribes create a false genealogy. It is only by comparing the many hundreds of documents that survive that we can identify variances and attempt to correct them.

  The other problem is that all too few of these genealogies contain identifiable dates. The only way to create a chronology is by working backward or forward from known dates and for that reason I take many of the following lists beyond our period of 400–600AD in order to get a firm footing. But dating pedigrees has an inherent problem. There is a general rule of thumb that a generation covers 25 to 30 years. We can easily test that. In the Mammoth Book of British Kings & Queens, I list a pedigree for Queen Elizabeth II from Beli Mawr, the first known British king who lived in about 100BC. It consists of 73 generations covering approximately 2050 years, equalling 28 years per generation. However, it is a very approximate yardstick. We do not always know if a name in a genealogy is a first born or last born, and a man could father a son at any time from, say, age 15 to 65. It is quite easy to have a youngest son who is younger than his own nephew. Unless we have some corroborating dates it is easy to be out by an entire generation.

  In the following pedigrees I alternate generations by 25 and 30 years to keep the average to around 28. I use the term floruit to denote the period of an individual’s prime of life, from about age 20 to 50. The dates given for floruit therefore are not birth-death. Where any real dates are known I provide them. In some cases I list generations from brothers and because you have to allow ten or more years between a range of brothers I have extended the generational span accordingly from 30 to 40 years. Clearly all of this is very approximate but, if the pedigree itself is in any way accurate, it will give us a bearing on an individual at least to within thirty to forty years. Even so, some displacements in time do occur, which suggests corruption within the pedigree.

  I am indebted to the work of P.C. Bartrum who has collected and assimilated many of these surviving pedigrees in Early Welsh Genealogical Tracts (1966) and explored them further in A Welsh Classical Dictionary (1993). Without his work the following would have been extremely difficult. However I have not always followed the dates that Bartrum has assigned to individuals, preferring to follow my own logic as consistently as I can.

  Following the sequence I discussed above, starting with the kingdoms in the north, the following charts begin with the ancestry of Coel and Cunedda and work through the Men of the North to Wales, the south-west and finally Brittany. I also include the Saxon pedigrees, such as they are.

  This first table lists the two pedigrees in Harleian MS. 3859 that show the descent of both Cunedda and Coel through collateral lines from
Beli. The names given are first the Latin names, as per the pedigree, followed by their Celtic equivalent, as per Bartrum.

  Table 3.2 The ancestors of Cunedda and Coel

  The table shows the limitations of the 25–30 year average for each generation, especially when the starting point is also vague. If we assume that Amalech in the first column is a duplication of Aballac, then there are 15 generations to Coel and 17 to Cunedda, which gives us roughly 410 and 465 respectively. That gives a reasonable mid-life floruit for Coel, but the extra generation places Cunedda too late. Since Cunedda is supposed to have married Coel’s daughter, he must have lived a generation earlier, in the 420s. Although Coel’s dates seem about right, if the name listed as his father, Guotepauc, was really the title “Protector”, it would push him back a generation, making him too early.

  It shows that though the generation calculation may get you to roughly the right period, you need other data to fine tune it, albeit still approximately. The more we work through the pedigrees, the more the chronology will come into focus.

  Table 3.3 introduces the second “Arthur” after Artorius. Because Coel’s descendants are so numerous, I have grouped them by generation, giving an idea of their territories. I have excluded Coel’s daughter Gwawl who married Cunedda, who appears in a later chart, and Coel’s third son Dydrwr, whose descendants are not known. Because each line includes older and younger sons, I’ve lengthened the prime-of-life “floruit” to forty years, and averaged the generation span to about thirty years.

  This table is a synthesis of several pedigrees, not all of which concur. Presenting them in a chronological form opens up even more queries. For instance, the few certain dates we know are the life of Kentigern and the fall of Ceredig, last king of Elmet, who was expelled by Edwin of Northumbria around 619/620. Ceredig is usually regarded as the son, or successor, of Gwallawg, who was involved in battles with the early kings of Bernicia (northern Northumbria and the southern territory of the Gododdin) in the 580s. But Gwallawg is recorded as the son of Llenauc, great-grandson of Coel, and thus could only have lived around the early 500s. It is possible that the sons were born in their father’s older years, but that raises the question of older sons more likely to succeed (or, if they were killed in battle, to be remembered in the poems). It suggests there may be a missing generation. The same applies to Dunod, who is always listed as a son of Pabo, yet the annals give his death as 590, suggesting either that he lived a very long time or that there is a generation missing.

  Table 3.3 The descendants of Coel

  The most confused genealogy belongs to the children of Mar (also called Mor) and Maeswig, grandsons of Coel. Indeed, Bartrum conjectures that Mar and Maeswig may have been the same person, as they feature commonly in the ancestry of their descendants. Mar’s son Einion is sometimes listed as a son of Arthwys, but we know that Einion’s son Rhun must have been contemporary with Maelgwyn Gwynedd (i.e. 500–540) because Rhun’s daughter married Maelgwyn’s son. Eliffer is sometimes listed as a son of Gwrwst, but the earlier pedigrees treat him as a son of Arthwys and this best suits the chronology.

  This table should not be set in stone. It is an approximation of descendants and chronology but it is unlikely to be out more than 25/30 years either way. It places Arthwys somewhere in Yorkshire in the period 470–510 which, as we will see, ties in with the probable dates of Arthur of Badon. It does not mean that he is the same as King Arthur, but it raises the question as to whether some activities attributed to Arthwys in now-lost ancient records were picked up by Nennius and Geoffrey. We have tentatively recognised a part of the jigsaw.

  These were not the only descendants of Coel and Cunedda, or “Men of the North”. There were also the British rulers of Alclud (Strathclyde), with their capital at Dumbarton. Only one of the ancient records lists their pedigree, so we have no corroboration. Some of the other Men of the North, who ruled amongst the Votadini at Din Eidyn (Edinburgh) also belong to this pedigree, through Dyfnwal Hen rather than Coel, so I have amalgamated all of them below. The only change I have made is that, in the pedigrees, the future rulers of Strathclyde (Neithon and Bili), are listed as descended from Dyfnwal’s son Gwyddno, but that is impossible according to the time scale. I believe this was a scribal error mistaking their descent from a later Gwyddno, descended from Garwynwyn. This is supported by the later Gwyddno having another son called Alpin who is recorded amongst the princes of Strathclyde.

  Table 3.4 The descendants of Ceretic of Strathclyde

  To complete the North, we need to match all of the above against the rulers of Dál Riata in Argyll, and the Pictish kings. The chronology of the kings of the Picts at this time is extremely confused and complicated, and is further aggravated by their kingship passing through the female line, making paternity difficult to track. The table below shows both sets of rulers as a list of kings, rather than a pedigree. This includes our third “Arthur”.

  Table 3.5 The rulers of Dál Riata and the Picts

  Picts

  Dál Riata

  Talorc (400–424)

  Drust (424–453)

  Talorc (453–457)

  Nechtan Morbet (457–468)

  Drest (468–498)

  Galanan (498–513)

  Fergus (498–501)

  Drest mac Drust (513–516 and 521–529)

  Domangart (501–507)

  Drest mac Girom (513–521 and 529–533)

  Comgall (507–538)

  Gartnait and Cailtram (533–541)

  Gabhran (538–558)

  Talorg (541–552)

  Drest (552–553)

  Cennalath (553–557)

  Brude (556–584)

  Conall (558–574)

  Gartnait (584–602)

  Aedan (574–608) and his son Artúir (c560–596)

  Nechtan/Neithon (602–621) (same as Neithon of Alclud in Table 3.4)

  We now move our attention to Wales. Table 3.6 lists the pedigree of the kings of Dyfed. They were descended from the Irish tribe of the Déisi, who were driven out of Leinster in the fourth century and settled in Demetia in south-west Wales, under Eochaid mac Artchorp.

  Table 3.6 The rulers of Dyfed

  This is a rare example where there is both a Welsh pedigree and an independent Irish one. The latter, from the Book of Uí Maine, is listed in the first column, as reprinted by Bartrum from a twelfth century document held in the Bodleian Library (MS. Rawlinson B.502). The second column is the Welsh version from Jesus College MS.20. The Welsh list is dubious for the first five generations where at some stage a different pedigree has been fused on to Tryphun to create a descent from Magnus Maximus. I have placed those names in brackets but they are best ignored. From Tryphun on the two pedigrees agree. This pedigree is important because the third Arthur is our first “real” Arthur.

  The chronology looks reliable. It allows for Eochaid to settle in Demetia in the mid to late fourth century, which fits in with known events. It allows Vortipor to be an old man at the time of Gildas (the above would give Vortipor’s life-span as 470–540), and it terminates at the known dates assigned to Rhain. Allowing for an error of maybe no more than 20 years, we can fix Arthur of Dyfed firmly in the late sixth century.

  The pedigree of the rulers of Gwent and Glywysing (Table 3.7), which includes our fourth “Arthur”, is both complicated and confusing. Unlike in Gwynedd (Table 3.8), where a strong hereditary kingship became established early on, in Gwent this proved harder to do. Leslie Alcock, who undertook a major archaeological study at Dinas Powys in Glamorgan, has suggested that because Gwent and Glamorgan had been strongly Romanized, Gwent clung more tenaciously to the Roman way of life and no single hereditary kingship emerged for some time. Instead, there were competing administrators and governors, no doubt many from the old Silurian nobility, all of whom sought overall authority but few of whom achieved it. When chroniclers tried to piece this back together two or three centuries later the key records were lost. The position is not helped by Gwent incorporating thre
e or four small kingdoms, which began independently and at various times merged or regained independence. Gwent and Glywysing were the two main kingdoms. Part of Glywysing was originally called Cernyw and became Gwynllwg. In later years when Glywysing merged with Gwent it was called Morgannwg. To the east of Gwent was Ergyng, which later became a sub-kingdom of Gwent.

  The following table depicts all of these parallel and sometimes overlapping dynasties, and tries to rectify some of the obvious errors in the old genealogies. For instance, the Jesus College manuscript shows a descent from Caradog Vreichfras, placing him so far back as to be contemporary with the Emperor Constantine. Yet other sources we will encounter show him as a companion of King Arthur. A study of the pedigree shows that two recurring names (Meurig and Erb) have become repeated, conflating two pedigrees into one and doubling the span of time.

  Table 3.7 The rulers of Gwent and Glywysing

  Dates for some of the reigns are more reliable by the eighth century, and the death of Ffernfael ap Ithel is recorded as 775 in the Welsh Annals. We also know that Meurig’s father, Tewdrig, died after the battle of Tintern when he was already of an advanced age. That battle has been variously dated between 577 and 630, with around 626 being the most likely. However, Morgan ap Athrwys is believed to have died in 665, which is too early for his position in the chart. We know that many of these kings lived to an advanced age, even the later ones not listed here. Hywel ap Rhys died in around 885, well into his eighties; Tewdrig ap Llywarch was also into his eighties. So we may find a 25–30 generation span insufficient in this instance. However, that makes it even more difficult to count back from Tewdrig, as it would push Owain Finddu, son of Magnus Maximus, back too far. Meurig’s mother Enhinti is identified as either the daughter or sister of Urien of Rheged, so I have placed him in the mid sixth century, even though he was probably of the same generation as Cadog.