It would be difficult to imagine a figure that has been the subject of more controversy or conjecture than Arthur of Britain. Even his very existence is a matter of considerable dispute with views covering the entire spectrum from total denial to outright acceptance. While it is clear that the core of Arthurian Legend had become established by the early 9th century, it was the publication of the History of the British kings by Geoffrey of Monmouth about 1136AD which fixed Arthur as a historio-literary character. However, this and much subsequent writing on the subject is more often a commentary on the contemporary society than any examination of dark-age affairs and from these Arthur emerges as a figure with a mythical past, a huge present and an ambiguous future.

Even in an era of sparse literary sources, references to Arthur's historical presence are miniscule, being confined to two entries in the Welsh Annals and a longer battle listing passage in the Historia Britonnum. It is notable that neither Bede nor Gildas refer to him, and not unsurprisingly the Anglo-Saxon Chronicle is silent on the subject. Curiously, while a number of people were known as Arthur (or variants), no dynasty claimed him as an ancestor and his name does not appear in Walia Pura genealogies until Tudor times [1]. This paucity of material has led to different approaches to the elucidation of the Arthurian character. Studies of the meaning of the name have been undertaken, and in tandem, with the implicit acceptance that Arthur was a nom-de-guerre or nickname [2], attempts have been made to identify him with known historical figures [3]. In a recent on-line review, Green [4], drawing on the work of Padel and others, has promoted the more radical concept that Arthur was a pan-Brythonic folkloric character, a defender against the supernatural, who became historicised by the early 9th century with the deeds of past warriors, and who presumably supplanted them from later ken. One advantage claimed for this theory is that many if not all, ambiguities associated with the Arthurian persona are removed [5], and many partial theories can obtain a limited level of acceptance e.g. that of Riothamus.

This assay represents an attempt to reassess the "Historical Arthur" in the light of, but not overwhelmed by, recent research. It should be noted that much of the material is available on-line, as are excellent catalogues of the Arthurian battle sites and potential Arthurian candidates.

The Historical Evidence.

As noted previously, the extent evidence is very limited. The first and longest piece is the battle passage in the Historia Britonnum. This work was compiled about 830 AD by a monk who might or might not have been called Nennius, and which, like its author is the subject of some controversy. Some 35 manuscripts (or part) have survived [6], usually divided into two families- the Harleian and the Vatican, which itself dates from about 944 AD [7]. The content of different manuscripts varies somewhat with material appearing in some but not others. The extant Harleian MS dates from the 11th century [8].

By contrast, the Annales Cambriae were originally completed about 950 AD and are usually thought to be the work of the scriptorum of St.Davids [9]. Obviously, they do not represent a continuous record but were compiled from a number of subsidiary Easter annals or computi. It appears that the original database was a set of Irish Annals. Thus, between 453 and 595 AD the Harleian MS contains entries against 18 years, totalling 25 items. Of these, 18 are present in the Annals of Tigernach or Ulster. Later editions of the Annals exist, each adding more material; indeed the final MS appears to be heavily influenced by Geoffrey of Monmouth [10].

The Arthurian Setting

Archeology, and to a lesser extent, literary sources, have indicated that there was no unique Adventus Saxonum. Their invasive presence was certainly known by 428 AD or earlier, and 442 AD the Gallic Chronicle could report that a significant part of Britain was under Saxon control. The distribution of early pagan graves [11] indicates that in the first half of the 5th century, they had become established in an area roughly defined by Kent, Oxford and the Humber. However, by the end of the century, their boundaries had barely expanded except along the south coast although the density of graves had increased significantly. Such findings confirm the evidence of the Historia whose prevailing theme is one of a continuously expanding Saxon population. Now when Gildas wrote his diatribe about 540 AD, he selected five British rulers for particular vituperation. Significantly, the (very) approximate boundaries of their domains lay to the west of a line joining Southampton to Chester. Since Gildas makes it clear that at the time of writing, foreign wars had ceased, it would seem reasonable to suppose that this line represented the approximate boundary between the British and the Saxons, and given the decisiveness of Badon, that the frontier had not changed significantly since then. It is notable that the heavily wooded valleys and the scrubland of the midland plateau held as little attraction for the Saxons as they did for the British, and in fact, the triangle bordered by Chester, Lincoln and Gloucester had remained seriously underpopulated since Roman times or before [12]. Therefore it can be conjectured that by the end of 5th century, the Saxon peoples of central England were coming under increasing pressure. Inefficient agricultural practices coupled with a burgeoning population would require a continuous expansion into new lands. But to the south, lay other Saxon tribes; to the north, the land was harder to exploit and might even have been under some sort of coherent British control. To the west lay the forests of the midlands and the strategic Severn barrier, but beyond these were the fertile Welsh Marches. The literary evidence for the Saxon occupation is virtually nil and the Gildasian account, with its emphasis on the fire and slaughter of divine vengeance, is scarcely more instructive. Yet it is possible to discern within the Excidio, if only dimly, three phases of expansion: - the rebellion and occupation of Kent, an advance into lowland Britain and finally, the raiding and rapine which led to the resistance under Ambrosius.

Now, it would appear that this raiding would have been along the British/ Saxon frontier, and it is here that the where of the Arthurian setting should be sought i.e. to the west and north of the Humber, Chester and Southampton. The Saxons themselves would be the people later known as Angles. However, Gildas gives no dates for the events he reports and the extent of his contractions is such that it is difficult to anchor them to any time scale. Unfortunately the date of Badon given in the Annals is generally thought to be some 20 years too late and an oft quoted date for the unpleasantness at Wallop (437 AD) between Ambrosius and Vitalinus would suggest that hostilities began during the decade 460-470 AD, and this has led to the view, that the resistance began under Ambrosius and was continued by Arthur. Indeed the available translations of Gildas usually imply a passage of time between the beginning of the fight back and Badon. However a recent re-examination of the oldest Gildasian MS is said not to support this view and Badon "reads naturally as the victory that crowned the career of Ambrosius Aurelianus"[13]. Such an interpretation is consistent with a dating of Wallop based on the Saxon arrival in Kent i.e. about 465 AD. By contrast, a preliminary statement in the Historia simplifies the dating of the Arthurian campaign: - "in those times, the Saxons grew and thrived in Britain. On the death of Hengist, Octha, his son travelled down from the north to the kingdom of Kent and it is from him, that the Kentish kings are descended. Arthur fought against them in those days, together with the rulers of the British, but he himself was the overall commander"[14]. The death of Hengist is usually taken as 488 AD and although there are complications associated with the Kentish dynasty [15], this would seem to be the date which Nennius would use. Thus the campaign began about 485 AD and continued to 497 AD or thereabouts.

The Arthurian Campaign

The battles allegedly fought by Arthur are contained in a list that follows the sentence..."Arthur fought against them..." The passage has been much discussed and numerous scenarios have been proposed to account for the locations of the battles and even the campaign itself. The section is part of the final chapter of the Historia dealing with 5th and 6th century affairs. In one sense, it marks a serious stylistic break in the text in that when dealing with Vortigern or Patrick the author appears to be narrating, whereas from this section to the end of the Historia proper, he appears to be reporting.

Nevertheless, the list itself has been heavily and repeatedly criticised as a reliable historical document. In particular, the sites of two of the battles (Urbs Legions, Coit Celidon) are often said to duplicate later actions at Chester (613 AD) and Arfderydd (573 AD) [16] and others can be identified in early poetry (Tribruit, Breguoin). To complicate matters further, the list exists in two versions: - the Harleian and the Vatican which despite some minor variations in vocabulary, differ considerably in that the 11th battle of the Harleian is quoted as Agned, while in the Vatican MS it is given as Breguoin. These differences have given rise to a family of theories which suggest that in the past a list existed which contained both Agned and Breguoin, but which became corrupted by the addition of Badon which was judged too important to ignore [17]. Clearly, such allegations cannot increase confidence in the reliability of the material and so it is necessary to re-examine these locations.

The first point to consider is that the authors of the Harleian MS and the Recension were as unsure of the locations of the battles as is the modern reader. The Harleian text uses just three verbs to describe the sites - 'dicitur', it is said; 'vocatur' it is called; 'fuit' was. To these the Vatican MS adds 'nominatur in Brittanice', usually called in British. In both versions Coit Celidon, Urbs Legionis, Castell Guinnion and Badon are noted as fuit; Bassas and Tribruit as 'vocatur' and Glein as 'dicitur'. Agned and Dubglas are 'dicitur' in the Harleian while in the Vatican, Dubglas and Breguoin are given as 'nominatur'. Expressed in this way, the varying degrees of certainty are clearly demonstrated and it is almost certainly significant that the sites of the dicitur and vocatur groups are those which have engendered the most discussion. Comments on the individual sites follow and these should be examined in conjunction with the more detailed web-page.

Glein. This is often located in Northumberland or East Anglia. It is said to mean pure [18], but it should also be noted, that in Welsh it would be given as Afon Glein, and its estuary (where the battle actually took place) would be Aberglein.

Dubglas. The meaning of this word has been given variously as blue-black, black water, or even black sheltered water [19]. Here the problem lies in the large and widespread number of potential sites, which makes a specific identification meaningless in any primary sense. The helpful hint, ' in the region of Linnu' only translates as in the region of the lakes, and is of little real assistance. Indeed, it is quite possible that in a different climatic regime a lake region might have existed then, but not now. Interestingly, no identification with the Lake District appears to have been published. The identification with Lindsey stems from Geoffrey.

Bassas. The site is virtually unlocatable. However, Cambuslang in Scotland and Baschurch (the church of Bassa) in Shropshire have been proposed [20].

Coit Celidon. Opinion appears to be entrenched in regarding this site as referring to the forested area of southwest Scotland, but Roman Caledonia referred to Pictland, an area much further to the north. If Celidon and Caledon were identical, then there has been a misprint that has proved remarkably persistent. However, it is to be noted, with some trepidation, that it was Geoffrey who originally made this amendment and this tends to suggest that other sites should be considered. The name appears to be Welsh/Old English compound, Celli -wood, don-hill. Such etymology is not uncommon in stable frontier areas e.g. Bredon in Worcestershire, and in this context, a site within the extensive Midland forests should not be dismissed.

Castell Guinnion. This is a fuit battle of which nothing is known and even the most enthusiastic interpreters of the list appear defeated by it. However the site is quite special, as it is the only battle in the list which has any appended description apart from Badon, whose entry in the Annals it closely resembles. It has been proposed that the origin of the list is a now-lost battle-naming poem [21]. In such works the intricacies of form and metre demand the extensive use of metaphor and it is argued that the term Castell Guinnion was used in this way. While Castell is now an accepted Welsh word (Castell Nedd - Neath) it is derived from the Latin castellum - a castle, fort, or figuratively, a refuge. In the entry it is inflected, which might indicate Latin usage or perhaps another problem arising from the passage of the text through Welsh and English hands in the manner of scuit/d. Yet, Guinnion appears to look suspiciously like a word akin to 'white people' which in context can only refer to the Saxons and thus can be translated as - stronghold of the whites. Coupled with a similarity to the Badon entry, it becomes very easy to associate the two with Guinnion being employed as a metaphor clearly understood by the original audience but lost to the transcriber. If this is so, and the two are identical, then the argument for one initial list cannot be sustained.

Urbs Legionis. This is another fuit battle, but one for which two strong candidates exist - Chester and Caerleon. As noted above, it has been regarded as a duplication of an action fought outside Chester in 613 AD when Aelfrith of Northumbria defeated a Welsh coalition. However, as his initial action was to massacre some 1150 monks [22], it would seem unlikely that such an act would have been easily forgotten, nor would any Welsh clerk regard it as a resounding victory. The readership for which the Historia was compiled might have been naïve, but it was neither stupid nor ignorant. Again, the problem stems from a confusion of Welsh and Latin usage so that Cair L[eg]ion and Urbs Legionis were often applied indiscriminately. But were they? Even Bede was reduced to over-elaboration when trying to describe the battle of Chester and he had to cite its location in Welsh, Anglo-Saxon and Latin. (Carlegion, Legicaster and Urbs Legionis) which suggests that the names were causing confusion even in his time. In this case, the insertion in the Vatican MS 'quae Brittanice Cairlion dicitur' does little to assist any specific identification. At the risk of diversion, it is possible to provide a partial resolution of the issue and add a further potential site. Bede recorded that the Welsh bishops met twice with Augustine. The first conference was 'at St. Augustine's Oak on the borders of the Hwicce and the West Saxons'. Ashe [23] has convincingly placed this at Aust and indeed the name is derived from Augustine [24]. But Bede does not give the site of the second meeting. However, the Annales Cambriae note a synod of Urbs Legionis in 601 AD. Whether this represented a second meeting with Augustine is not clear, but it is difficult to accept that any contemporary synod would be held at Chester. First, it is by no means certain that Chester even existed as a town at that time- the Roman town of Deva had been destroyed in 296 AD [25], the legions were transferred in 367 AD [26] and Niall of the nine hostages reputedly sacked the area in 395 AD [27]. By 893 AD it had ceased to exist since the ASC records a Danish army sheltering within its ruined walls. In any case, it would seem far more likely that any ecclesiastic conference would be held at Bangor Iscoed. Nor does it match the Augustinian mien that he should undertake such an arduous journey beyond the aegis of his sponsor, Aethelbert of Kent. Nevertheless, there is some evidence that Caerleon had survived to this date as a religious centre [28], although whether it represented much of a going concern in any urban sense remains moot. However, there is third possibility, but one noted with due caution. In an obscure secondary source, dating from 1830 [29], there is a note stating that a bishop of Llanbadarn attended a synod in 603 in Worcestershire. No further details are given, but it seems reasonable to assume that this refers to the synod of Urbs Legionis and as such, might be a more reasonable site for the confrontation with Augustine.

Tribruit. This battle is known from poetry and features Bedwyr fighting Gwrgi Gwlwyd on the shores of the river Tryfrwyd. Incidentally, Gwrgi appears in Triad 32 as a serial killer of Cymry. The poem has been dated to the 9th, 10th or 11th century's [30] and thus might post-date the Historia by a hundred years, which suggests that the site is being used as a literary vehicle. A Scottish location appears to be the most favoured.

Agned. A dicitur battle which Geoffrey identifies with Edinburgh although not in any context of a campaign. By contrast, Ashe has identified it with Anger in France. However, for this site, the most relevant question is from where did Geoffrey obtain his information.

Breguoin. Battle 11 of the Vatican list has caused no little discussion with proposed solutions including Leintwardine (Roman Bravonium) in Herefordshire through High Rochester (Roman Bremenium) to a poetic reference to a battle fought by Urien of Rheged in the Cells of Brewyn [31]. The insert, 'which we call Bregoin' might suggest that the scribe was English. The word appears to mean 'White Hill' and Brewyn can be derived with little difficulty but it might also be a figurative use in the sense of Castell Guinnion; indeed both contain the phrase 'in fugam versi'.

Badon. This is the final and decisive battle and the only one with any historical provenance. The action has been the subject of much controversy. The usually accepted date (ca 495/7AD) is at variance with that appearing in the Annals and in essence, despite numerous attempts to derive a credible location, no real consensus exists. Yet, in the light of the prevailing strategic situation (discussed above), it is difficult to consider any site other than the Welsh Marches, and there is indeed some literary testimony to support this although it must also be admitted that the evidence, based on glosses in later manuscripts should be treated with due caution.

In the Wonders of Britain, the author states.... "the third wonder is the hot lake where the baths of Badon are (in quo balnea sunt Badonis) in the country of the Hwicce...The fourth wonder is the salt springs found there, from which salt is boiled wherewith various foods can be salted: they are not near the sea, but emerge from the ground". Now, at first sight, the passage would seem to refer to Bath. But on closer examination this interpretation may be questioned. Firstly, there can be some considerable doubt if Bath was ever 'in the country of the Hwicce', peripherally, perhaps, but hardly within; secondly, at the time Bath might have been known as Akmancaster [32], and thirdly and most importantly the waters of Bath are not saline at all. However, the area around Droitwich - definitely in the country of the Hwicce was renowned as a salt producing area since ancient times [33] and has also received much publicity as a spa. Since salt represented virtually the only means of preserving meat during the hard winter months, it is difficult to overestimate its importance to the contemporary economy. In fact, the Saxons began to exploit the salt deposits of the area as soon as they had secured it [34]. In this sense, the confusion of Bath and Droitwich appears incredible especially in the context of a passage that concentrates on the Severn valley/Gwent/Powys region.

The second gloss appears in the Cambridge Manuscript of Gildas where Badon is described as ' que prope ostium Sabrinum' i.e. not far from the mouth of Severn [35]. In such circumstances, this would seem to apply more closely to Droitwich (or an area to the south) than Bath which would presumably be described as 'close to the Severn Sea'.

Wherever Badon occurred, the battle was decisive. The pressure on the Saxons was maintained and by 539AD, Procopius could report a significant repatriation of Angles [36]. Nevertheless while the strategic aspects of the engagement are known or can be guessed, the tactical position is much less clear. Here the importance of the Castell Guinnion-Badon link cannot be underestimated as it strongly suggests the investment of a Saxon stronghold, perhaps their headquarters. It might be considered that its fall deprived the invaders of their leadership and as importantly their stock of weaponry.

Arthur or Arthurs

Over the years, there has been considerable speculation concerning not only the whereabouts of the Arthurian battles but also their very existence. This exercise of 'find the battle' has been pursued vigorously since the days of Geoffrey. And it has been noted that given sufficient ingenuity and linguistic manipulation it is possible to place the battles in just about any locality [37]. The historical reality of the battles is less easy to judge and their very obscurity has been used as an argument to bolster their existence [38]. Further, the presence of Tribruit in poetry has been employed to suggest that the battle was fictional. Yet, no clear date for the poem has been given with quoted references vary from the 9th to the 11th century. Arguments have been proposed to suggest the poem 'Cat Godau' as the source of Coit Celidon [39]. However the poem appears to this (highly) untrained eye as containing the key elements, battle, trees and an appeal to Arthur, which taken together, can hardly be regarded as compelling evidence.

However, if the battle list is a concoction, it is a remarkably deficient one. Given that the author wished to complete a full complement of twelve, he would have had a large catalogue of battles both known or obscure to choose from, and in this context, the necessity to include the four skirmishes along the Dubglas would appear to suggest reality rather than fiction. But what is curious is that, while the engagements are specifically noted as being fought against the Saxons, analysis of the piece would place the sites in two general locations - Wales and the North [40]. In this context, discussion whether the battles are real or not need not be strictly relevant since the salient point to emerge is that they define the areas where an Arthurian tradition became established at an early date. In turn this leads to the view that either two Arthurs existed, probably fighting during different periods, or less plausibly, that a mobile Arthur existed who was active in Wales but who also claimed an extensive reputation in the North. Of the latter candidates, only Arthwys of Elmet and Arthwys of the Pennines need be considered. Yet it seems unlikely that Elmet might extend its power over so large a range, (although it is not impossible that Arthwys was responsible for the traditions of the Lindsey or Fenland sites), however in the case of a Pennine Arthur, an argument can be constructed.

In his preface, Gildas notes that he does not intend to consider the deeds of great warriors, yet he goes on to name Ambrosius Aurelianus as providing the leadership of the resistance. It is thus possible to infer that he regarded Ambrosius as the organiser, but not necessarily the executor, of the fight-back, and it might be argued (just) that a mercenary band was recruited by Ambrosius from the North having Arthwys as its leader. Since it is not improbable that he fought Saxons in Northumbria and was also involved in internal wranglings, a leader could emerge unnamed by Gildas with both Welsh and northern connections and whose pedigree would not appear in any Welsh source. However this theory is weakened if not compromised, by the timing of the terminal Saxon invasion of the north and also the point that a decade is a long time to employ a mercenary band - although again, it might be possible to explain the Glein/Dubglas connection and indeed the question of a Dux Bellorum fighting with the British leaders makes much sense in this context.

More convincingly, Ambrosius is often identified with Arthur [41], but herein lies a problem because as little is known of Ambrosius as Arthur himself. Despite a recent comprehensive survey by Veprauskas, the hard Ambrosian biography remains obscure. A son of a family perhaps having strong Imperial connections, Ambrosius led the resistance to Saxon invasion, seemingly in the N.E. Wales area [42]. Yet, again, he founded no dynasty, but his degenerate grandchildren were known to Gildas. There is evidence that he was militarily active against Vitalinus about 460-5 AD [43]. If it is accepted that his appearances in the Historia and Bede are derivative, then the only reliable source for his existence is Gildas, and Gildas is by no means a simple source where names are concerned. For instance, of his big bad five, three are given pseudonyms, and earlier, Vortigern is called the Tyrannus Superbus. Thus it is possible to question whether Ambrosius is actually his real name or the product of rollicking Gildasian wit. But it is evident that Ambrosius was a man of considerable power and Gildas was sufficiently informed of the family to castigate his descendents. While no direct evidence exists to support the conjecture, there was one leader present at the relevant time and place: Cadell Ddyrnllwg, in Welsh - shining hilt, an epithet by no means totally inconsistent with that of Ambrosius - divine golden one. In fact Cadell is a derivative of the Roman family name Catellius [44] and to stretch matters much further, the prefect of the Votadini (a tribe known through Cunedda to have translocated to Wales) was once Cattelius Decius (Catel ap Decion of the Harleian Gododdin genealogy) [45]. To speculate further, it has been suggested that the Aurelius Caninus of Gildas was in fact Cyngen, the grandson of Cadell [46].

But what of the Northern Arthur? Again, here a candidate exists in Urien Rheged. He certainly fought successfully against the Saxons and his assassination at the instigation of Morcant might lie behind the story of betrayal and death embodied in the Camlan legend. The Arthurian connection is extended by the tradition that Arthur would stay with Urien when in the north. In the absence of further information, Urien is as good a candidate as any.

Camlan

In many ways Camlan is the typical Arthurian battle and its noble and tragic overtones of betrayal and loss seem far more appropriate than any squalid skirmish on a wet Welsh hillside. The date of Camlan (given as 537 in the Annals) has often been questioned, and like that of Badon it is often regarded as some twenty years too late i.e. about 517AD. But even this date would appear to be too late to apply to Ambrosius. Now there is a body of opinion that regards the early indigenous dates in the Annales as much later insertions; thus a date of any kind might be irrelevant or conversely inapplicable to Arthur. Certainly, the early history of Powys seems bespattered with dynastic dislocation. However, it should also be noted, that the Welsh Camlans are to the west of Powys territory, within the hegemony of Maelgwn Gwynedd, who did indeed suffer from family troubles - with his uncle (he killed him) and his own wife and nephew (he killed them too). Interestingly the two dates are consistent with both the conflicts. However it is also notable that the Irish annals note the famine (not a plague)- 'perditio panis' in 537AD, which might suggest that universal strife in such a year might well be remembered. Indeed, the death of Ambrosius in civil war does not seem to be the sort of issue to pass unmentioned by Gildas.

As noted earlier, the murder of Urien appears to have been incorporated into the tradition and it is curious that while this is noted in some detail in the Historia, by the time of Annals the battle is fully associated with Arthur.

The Annals, the Historia and Amendment of the Record.

In many ways, the Annals are a strange document. The traces of layering derived from the original Irish database are clearly evident. While the incorporation of indigenous material is often said to continue only from 613 AD [51], there is some evidence to suggest that a small body of specifically northern material was included from 573 (Arfderydd)

These entries make no mention of Urien at all which by itself is strange if they were later concoctions although this might be explicable if the Annals were compiled in a Menevia where his fame was largely unknown. Only three early British items are present: Badon (516), Camlan (537) and the Yellow Plague (547). This last entry would appear to be a simple amendment of the extant Irish item to accommodate the strong local tradition that Maelgwn had succumbed in the visitation [52]. But intriguingly, the others are both Arthurian. If Camlan is rejected as being at best a confusion with the affairs of Maelgwn, the Badon entry is also historically uncertain not only in terms of its date but also in terms of its content. It reads:

"516. The battle of Badon in which Arthur carried the Cross of our Lord Jesus Christ on his shoulders for three days and nights and the British were the victors".

Compared with the other entries, this item is anomalous. It is longer than most and contains the explicatory gloss "in quo..." although it is not clear whether this extends to cover the final clause 'the British...'. The alleged scuit/d confusion is repeated, and the very fact that that it was necessary to emphasise the British victory suggests that either the entry was prepared so long after the battle that the identity of the winners had been forgotten, or, more likely, that the scribe was not himself Welsh. Alternatively, it might reflect an entry in a now-lost Irish Annal, but this would be even more anomalous since the Irish entries in the Annales Cambriae can be traced without exception. Nevertheless it should be considered that the item has been modified from the 950AD ' original'.

Evidence of further tampering may be found in the Historia. It is important to remember here that the extant early MS of both the Historia and the Annals are not necessarily exact copies of the originals. Therefore in the case of the religious gloss inserted into the Castell Guinnion and Badon entries, there is no certainty which one represents the initial modification of the existing text. In this case the point that the scuit/d confusion could only occur once validly defines the dependence of one on the other, but not their order. However, it is also possible to argue that the Latin is correct and that a painted tunic might be the source of the tradition. In any case the different iconography need only reflect varying religious preference.

What is clear is that the pieces have been modified. Further it can also be argued that the amendments were made at different periods and for different motives. For instance it might be suggested that the initial sentence of the battle passage...."And Arthur fought against them..." represented an original part of the text which conflated the deeds of two figures separated in both time and space. But such a passage, together with the Badon description (killing 960 Saxons in one charge) and to an extent, the Camlan gloss, are all aimed at the definition of Arthurian martial prowess. However, the other pieces appear to be of a different stamp: the portrayal of Arthur in an overtly religious setting would not seem to be the aggrandisement of Arthurian stature but the enhancement of the religious ideal using the warrior to demonstrate the power of the church over the pagan Saxons. As such, this evidence need not be either true or relevant to the elucidation of the Arthurian character. Indeed a similar motive occurs in the early Saint's Lives where he is usually depicted as being in conflict with the Church. However, here the situation is somewhat more complicated as the anti-Arthur genre appears to have begun during the 11th century at Llancarfan [53], where Gildas, traditionally no friend of Arthur's, had been Abbot,

Epilogue - Who was Arthur?

The historical evidence outlined above is at best ambiguous, at worst invalid, in judging whether Arthur was an indubitable historical figure. Yet, to the monks who compiled the Historia and the Annals there was no such uncertainly. As was noted above, if the glosses and interpolations are removed, the only remaining coherent historical text is the battle list and this can also be questioned severely. Again, it has been suggested that the bi-directional bias of the list can be explained by the presence of two Arthurian figures. Such a stance might imply that the list originally represented a series of victories against the Saxons which a scribe, being ignorant of the course of events, attributed to Arthur alone. Indeed, it is possible to argue that the Dux Bellorum fighting with the leaders of the British implies an acceptance that a composite figure is being described. However this view is weakened by the temporal definition - "in those days".

It is time to consider Arthur as a folkloric character, the virtual Arthur, the defender against the supernatural; however this view can be questioned on its own terms. For instance, why does nearly all the historical material emphasise the conflict with Saxons. This of course might reflect the dread that they instilled among the British, but Scots and Picts caused as much havoc within the isle. In this context, the necessity of Arthur's 'death' is anomalous (no death notice was published for Hengist); it might reflect a severe historicisation of the figure, but if the events surrounding Camlan can be clearly seen as 9th century concoction, the parallel with the assassination of Urien, perhaps back-dated by 50 years or more, seems compulsive. Thirdly, it must be appreciated that the contemporary perception of the supernatural was very different. The Celts regarded the otherworld as a definite but distinct region of existence [54]; thus in the Mabinogion, characters drift in and out of reality without difficulty. Bede could suspend sober historical narrative to describe miracles and later Giraldus could relate anecdotes concerning incubi, and in neither case with any apparent textual strain. Such a mind-frame could certainly accommodate a historicisation but would be much more receptive to an extension of a real figure into the supernatural, and such exaggeration need only be limited by the credulity of the audience. Indeed, given the willing suspension of belief, this need not apply at all, and all manner of tales could radiate from a central figure of sufficient charisma. Ashe has consistently employed a wild-west analogy for this process, but a more appropriate example, and one closer to the parent tradition, can be found in the work of Max Boyce. In an era when Welsh rugby strutted across the world's playing fields, Boyce's depiction of the players imbued them with near supernatural powers, and while his songs and poems were usually (but not always) light-hearted, they reflected the Anglo-Welsh rivalry in a manner closely akin to the tales of Arthur. But his heroes were real, and in this respect, there is a clear example of the transition of the heroic historical to the mythical [55].

But in an other sense, the either /or of historical to mythical versus mythical to historicisation need not apply as both processes can proceed simultaneously. It is clear that by the 9th century Arthur was regarded as a real figure and the compilers of the Historia and Annals were assiduous in placing him into what they believed to be the correct historical context while also making a clear distinction between the historical and the mythical. The Historia demonstrates this explicitly. The battle passage is presented as history, but the obviously apocryphal pieces concerning Amir's grave and Cabal's footprint are relegated to the section of British wonders. And this is the crux of the problem. The 'historical' matter can be seen as seriously deficient. Yet, with thought, this need not be a problem if it is assumed that both the historical and mythical material is regarded as two sides of the same coin. The compilers were presented with a mass of subject matter; the more credible they regarded as history, the more outlandish as folklore. But, as is stated above, both can stem from the perception of real figures whose deeds were transmitted through folk history but who left no accurate contemporary record.

The salient point is that in virtually all the historical material Arthur is a warrior fighting Saxons. The paranatural character does not, and there is no early reference joining them in this way. Even the poetic references in Y Gododdin or Gereint use him purely as a metaphor, and it is thus easy to suggest that the basis of the figure was a warrior whose fame became established from successfully contesting the Saxon invasion.

Here the evidence of the name is instructive [56]. The terms, guardian (or defender) of the (Great) Bear, ploughman, or even fierce (bear) man are all loaded with symbolism. Yet attempts to identify a real character who might have had the term applied as a nickname or nom de guerre have proved unsatisfactory and indeed the existence of a doubly located tradition does not suggest uniqueness. Yet if the mythic figure of Arcturus was a distinct feature of the British mind-frame, it is possible to define a sequence of events which might account for the evidence. Thus initially real leaders became identified with the eponymous mythic figure probably in a metaphorical context. Because of the power and charisma of the warriors, a folk culture grew around them and this supplanted the original Arcturan cult, which eventually disappeared from popular ken [57]. The popularity of the new cult was able to admit both the pseudo-historical and mythical elements and thus remained tied to the Arthurian persona or its perception. Therefore the evidence so far adduced need not be applied to the virtual figure but can be explained by the use of the name in a specific sense, not unlike the Saxon Bretwalda. It was not a post or position but a post-hoc status akin to a Dark Age job-description: a man who killed large numbers of Saxons and as important, comprehensively beat them in battle. Now, while the fragmented nature of the post-roman fiefdoms obviously assisted their assimilation by the Saxons, it is also true that any leader who could rise above local rivalries and forge a coalition (a Dux Bellorum fighting with the British leaders) would be in a strong position to inflict severe losses on the invaders. However what has not been perceived is that such a leader would require political and diplomatic skills at least equal to his martial expertise. Consequently the enduring fame of such Arthurs might be based on a broader assessment of their leadership abilities than merely a talent for slaughter.

With its non-Christian overtones, such a status would be unknown or ignored by Gildas. It would not be a candidate as a baptismal name in Wales until its connotations have been long forgotten. Indeed, no such candidate would have been available for a thousand years; it could have applied to Ambrosius/Cadell or Urien but not Gwawrddur who in the Gododdin,

'Glutted black ravens on the walls of the fort
But he was no Arthur'

Although he was actively and enthusiastically reducing the Saxon population, there is no evidence that his skills were anything other than martial and in any case, he would eventually lose the battle and thus could never be regarded as an Arthur.

Was Arthur a real figure? It is argued that he was both. It is clear that the 'Historical Evidence" is deficient, but it can also be argued that the Battle List itself might represent a core of tradition concerning the existence of real warriors who fought and beat the Saxons and were themselves the real Arthurs. Around these figures legends grew which became accepted as history or legend depending on their credibility. But in another and broader context, the question might be considered irrelevant, as the symbolic Arthur is as real as any conjectured, and the question of his existence, at the interface of fact, fiction and myth might seem more often to reflect the stance of the enquirer.

Books.

1. Alcock, L. Arthur's Britain, Penguin 1971.
2. Ashe, G. King Arthurs Avalon. Fontana 1973.
3. Ashe, G. (Ed) The Quest for Arthurs Britain.
4. Barber, C. More mysterious Wales, Paladin 1986.
5. Bede, History of the English Church and People, trans. L. Sherley-Price, Penguin 1968.
6. Chadwick, N. The Celts, Penguin 1971.
7. Delany, F. The Celts, Grafton, 1989.
8. Dudley, D. Roman Society, Penguin, 1970.
9. Gastineau, H. South Wales Illustrated, originally published Scott, 1830, reprinted EP 1976.
10. Geoffrey of Monmouth, History of the Kings of Britain, trans. Lewis Thorpe, Penguin 1966
11. Gerald of Wales, The Journey through Wales/ The Description of Wales, trans. Lewis Thorpe, Penguin, 1978.
12. The Mabinogion, trans. Ganz, J. Penguin, 1976.
13. Mills, D. A Popular dictionary of English place Names, Parragon, 1996.
14. Morris, J. The Age of Arthur, Phillimore, 1977
15. Nennius and The Welsh Annals, trans. Morris, Phillimore 1980.
16. Phillips, G and Keatman, M. King Arthur, The True Story, Arrow, 1993.
17. Richmond, I. Roman Britain, Penguin 1963.
18. Whitelock, D. The Beginnings of British Society, Penguin, 1991.
19. Wills, C. Plagues, Flamingo, 1996.
20. Zinsser, H. Rats, Lice and History, Bantam 1960.

On-Line

1. Ashe, G. Origins of the Arthurian legend
2. Ashe, G. A Quest for Arthur.
3. ASC - The Anglo Saxon Chronicle.
4. Brynjulfson, S, Idle Pursuits
5. Cad Godau.
6. Davies, J. Gwarnant
7. Davis, M. King Arthur and Cuneglasus.
8. Ford, D. Early British Kingdoms. See in particular Arthur's Battles and Arthur the King.
9. Gildas, De Exidio et Conquestu Britanniae.
10. Godesky, J. Who was King Arthur.
11. Green, T. The Historicity and Historicisation of Arthur.
12. Griffen, T. Athur's Name.
13. Harleian Genealogies.
14. Historia Britonnum.
15. Hunt, A. The Road from Avalon.
16. Sims, D. Gildas the Monk and Maelgwn the Murderer.
17. Sims, D. Troubles with Vortigerns.
18. Veprauskas, M. The generations of Ambrosius.
19. Wiseman, H. On the Destruction of Britain.
20. Y Gododdin.
21. Annales Cambriae.
22. Annals of the Four Masters.
23. Annals of Tigernach.
24. Annals of Ulster.

Notes and References.

1. Bartum. 1965. Quoted in Green.
2. Eg Griffen or Davis.
3. Eg Godesky, Ford.
4. Green. This essay, regularly being updated, represents one of the most interesting sources of material on line. While his conclusions on the reality of a Historical Arthur are essentially negative, it is essential reading and recommended to all Arthurian students.
5. Green. However, it must also be considered that the abolition of speed limits would drastically reduce the number of speeding offences.
6. Morris, Introduction to Nennius.
7. Surviving copies date from 976 AD or later.
8. Alcock.
9. Alcock
10. See photo reproduced in Tolstoy. MS attached to cover off Domesday Book P.R.O. In this context, it should be noted that the Annals of Ulster contain an entry for 467AD which records the death of Uther Pendragon and the succession of Arthur.
11. Maps by Chadwick Hawkes, given in Morris.
12. Dudley
13. Padel, quoted in Green
14. Text from Morris, translation by Author
15 It is now generally accepted that Hengist and Horsa were mythical characters. Yet it is notable that the ASC explicitly states that Horsa died in battle (Aylesford) in 455AD and that Hengist and Aesc then ruled together. In context this would suggest that the Saxons suffered a major defeat and it would appear that from then until 488AD, Aesc was the prominent Saxon leader in Kent. However in that year, the ASC notes that Aesc took the kingdom alone. The death of Hengist is not mentioned at all. Bede explains this apparent discrepancy by reporting that Aesc was actually the clan name and that there were a father and son, Oeric and Octha respectively. By contrast, Nennius gives the succession as Hengist - Octha - Ossa (Aesc). This is consistent with the earlier statement that Octha was a son of Hengist invited with cousin Ebissa to settle in the north. Given the similarity between Ebissa and Alusa, a known early northern invader, the piece smacks of allegory, and thus the reference to Octha travelling south at the beginning of the battle piece would seem to be an effort to maintain consistency. However it might argued that if Nennius were correct and he was repeating a tradition that on the death of Aesc someone had come from the North to claim the Kent, then the succession might be neither as tidy nor simple as might be thought. Certainly, the Historia is emphatic in stating that it is from Octha that the Kentish Kings were descended. On this point, it has been stated that the dates given in the ASC are often arbitrary. Thus the date for the accession of Octha and indeed the resistance need not be as certain as the text implies.
16 Green
17 Brynjulfson, Hunt.
18 Alcock
19 The reference here is to the Dulais stream that runs into the Wye near Hay (Gastineau). Apart from being in the right strategic area, the stream has an associated legend of running red with blood after a battle (Barber). But is this a later addition. This motif is quite common, and a similar legend is borne by a stream running from the Wenallt near Rhiwbina, Cardiff.
20 The derivation of Bassas is from Mills, the identification from Phillips and Keatman. It does occur that Cambuslang seems closer to Camlan than Bassas.
21 See Alcock or Green.
22 The numbers of dead vary, Bede gives 1150, ASC, 150; as both agree that only 50 escaped it seems likely that the ASC is repeating a misprint. It should be noted that Bangor is said to have housed some 2400 monks, and that after the disaster the survivors fled to Bardsey, leaving the Monastery to disappear without trace. It is interesting to note that by making no adverse comment about this, Bede appears to be a model of political correctness.
23 Ashe
24 Mills
25 Dudley
26 Dudley
27 Richmond, Dudley.
28 See Caerleon, this Site.
29 Gastineau.
30 Triad 32 reads in part... One of the three men who performed the three Unfortunate Assassinations...and Diffydell, son of Dysgyfdawd who slew Gwrgi Garalwyd. That Gwrgi used to make a corpse of one of the Cymri every day and two on Saturday so as not to kill on the Sunday. The dating here would seem to be a problem.
31 See Alcock or Ashe for an easily accessible discussion of Breguoin
32 Mills
33 Richmond.
34 Whitelock
35 The only reference known to the Author is in 'Our Man in Camelot' (A. Price, Futura. 1982). The provenance of the MS is given by Price but the quotation is also used in an Arthurnet discussion.
36 Ashe, or Wiseman.
37 Green
38 Alcock, Ashe.
39 Green.
40 There is little evidence in the list to associate him with the West Country, or even the midlands.
41 Alcock, Ashe etc.
42 Wiseman
43 See Sims. Troubles with Vortigerns.
44 Morris
45 Morris
46 Ford.
47 Historia Britonnum, Urien has been a favourite candidate in an Arthurian role for sometime.
48 Quoted in Green
49 Gildas, but see Sims
50 Annals of Ulster, Tigernach.
51 Quoted in Green
52 There has been some debate (see Alcock) as to whether Maelgwn actually died in this particular plague. However to deny this is to misunderstand the magnitude of this visitation. Procopius in De Bello Persico (Zinsser) noted that the plague of Justinian originated in Arabia, spread to Egypt about 540AD and by 542AD was ravaging Byzantium. While it rumbled on in Europe until 590AD and its derivatives for even longer, its major effects were felt earlier and it has been estimated that a half of the European population died before it finally came to an end (Wills). It seems unlikely that the death of Maelgwn in such an event would be mistaken.
53 Alcock
54 Various authors make this point, see for instance Chadwick.
55 Perhaps the most apposite example is the tale of "Steve Mostyn Evans, the world's first bionic No. 8"; Boyce exactly reproduces the sentiment of Y Gododdin. ".... Is he not the greatest ever? - Yes I said he's very good, but a Mervyn Davies - Never.
56 See for instance, Griffen.
57 A good instance of this mechanism is found in the terms, Tommy Atkins, Uncle Sam, Right Charlie, Boycott etc. where the original people have been long forgotten but the terms have survived.