The Mammoth Book of King Arthur

by Mike Ashley

  Connecting Stow with a fort called Guinnion is not straightforward. Skene and others simply based it on the fact that a Roman fort was known to be nearby, and that this must have been Guinnion (“White Fort”). The nearby Gala Water tumbles at high spate along the valley and is sometimes called the “White Strath” or Gwen-y-strad, though this seems rather convoluted. Alistair Moffat, in Arthur and the Lost Kingdoms, follows a more convincing route. He reminds us that gwen, or more properly gwyn, means not only “white” but also “holy”, in the sense of “pure”. Thus the name Stow may simply have been a Saxon translation of an earlier Celtic name. Some etymologists suggest that Wedale was originally Woe-dale, or “dale of woe”, remembering a Saxon defeat, whilst Moffat suggests the name derived from Guidh-dail, the Valley of Prayer, but had previously been the Holy Valley or Gwyn-dail, possibly corrupted into Gwyn-ion. This all seems rather tenuous to me.

  In fact as Gwynion the name is quite common in Wales. There are at least four noted hills, or crags, called Carreg Gwynion, near Pembroke, Rhayader, Rhos and in the Berwyn Mountains at Llanarmon. This last is the site of a well preserved Celtic hill fort, which seems more likely to have been the “fortress of Guinnion” than at Guidh-hail. This locale would better suit a battle between Welsh factions, but it is less than a day’s ride from Chester and cannot be ruled out. The fact that the name is fairly common in Wales suggests that at one time it was probably equally common across the rest of Britain. If so the name may have adapted to Wenbury or Winbury or Whitsbury, near Fordingbridge, a site I discuss later in relation to Cerdic’s battles.

  A more intriguing possibility is Wanborough, just outside Swindon. The name was once Wenbeorge, which is usually treated as “wenn beorge,” meaning the “place at the tumour-shaped mounds”, as wenn is Saxon for tumour. However, wen could as easily be derived from the Welsh gwyn for “white”, a theory strengthened by the fact that nearby are two sites, White Hill, renowned in Roman times for its pottery production, and Whitefield Hill, near the site of some ancient earthworks. The surrounding hillsides are covered by the many famous chalk carvings, such as the White Horse at Uffington. Wanborough was the site of the Roman fort Durocornovium, the Fort of the Cornovii, and this may well have been known locally as the White Fort. What adds to the intrigue of this site is that a little way to the south is Liddington Castle, one of the most favoured sites for the battle of Badon.

  There have been other suggestions, including Burgh Castle (Gariannonum) in Norfolk, Winchester (Caer Guinn), and the Wrekin (Caer Guricon) in Shropshire, but the only other one that has some merit is Binchester, near Durham. Here was the Roman fort of Vinovium, one of the earliest in Britain which was later refortified and remained in use until the early fifth century. The origin of the fort’s name is uncertain, most suggesting “the Way of the Wine” or similar. One suggestion is that it meant “pleasant spot”, which may link back to the Celtic word gwyn, which could have been Latinized to vin. The Celtic spelling of Vinovium is Uinnouion, becoming Gwinnouion. It was the largest fort in the north-east and held a contingent of Germanic soldiers. Its location must have been important to later settlers because they also buried their dead here, showing that it became a sustained community. At nearby Escomb is the oldest surviving Saxon church in England, and it may be that, once converted, the Saxons were drawn to what had long been a holy and venerated area, as Arthur’s battle in the name of the Virgin Mary might imply.

  9. The ninth battle was waged in the City of the Legion

  This ought to be reasonably straightforward but unfortunately isn’t. There were three main legionary towns, Caerleon, Chester and York. With the construction of Hadrian’s Wall, Carlisle also became a legionary town but not in the conventional sense, and, unless it was the capital of Valentia and its status changed, not at the end of the Roman period. Although Caerleon remained a legionary base, the legion was seldom there, beyond a skeletal force. It was involved in the construction of Hadrian’s Wall, but by the third and fourth centuries was assigned elsewhere, including Richborough in Kent. So whilst Caerleon can rightly claim the name City of the Legion, it was far less significant than either York or Chester.

  York was not only the home of the VI Legion, it was the military capital of Britain, and the fortress remained permanently manned and strengthened throughout the Roman period. The headquarters of the Fort was such a major building that it remained in use well into the ninth century.

  Like Caerleon, Chester’s XX Valeria Victrix legion was often stationed elsewhere, especially during the third century, but Chester was refortified in the fourth century and remained so until the end of the Roman period. Nennius refers to the city in the singular, as urbe Legionis, “city of the Legion”, as if by the time of this battle all but one legion had left Britain. Gildas, in describing York, used the plural Legionum urbis cives, “the city of the Legions”. Unless this was too subtle for Nennius’s source we are evidently talking about a different place. The XX Valeria Victrix was the last legion to leave Britain.

  So whilst York was the major legionary fortress, and more likely to have been a focus for Anglo-British confrontation, Chester was the centre of the last legion, and is known to have been the site of a major Anglo-British battle in about 615. Both therefore have an equal case to argue. Which one Nennius meant can only be solved by identifying the other localities in the list.

  10. The tenth battle was waged on the shore of a river which is called Tribruit.

  Like Bassas, this river has almost defied analysis, and most authorities admit defeat. The early analysts, Skene and Glennie, considered the Celtic version of Tribruit, Tryfrwyd, a name which also appears in the poem Pa Gur (see Chapter 8), where it is spelled Trywruid. In a study of Scotland written in 1165, they found that the old British name for the Firth of Forth was Werid, derived from Gwruid, meaning “men of the forth”. The word “shore” is significant as it suggests more of a sea-shore than a river bank. The Celtic word is traeth and the word Tribruit or Trywruid may have originally been a combination of Traeth and Gwruid, with the “g” dropped, becoming Traewruid. It sounds plausible, albeit tortuous, and the site suggested is the Links of Forth between the river and the heights of Stirling Castle, a site better known for the Battle of Bannockburn in 1314.

  O.G.S. Crawford, the pioneer of aerial surveys in archaeology, also suspected that this battle was waged on the Forth, but further east at the Fords of Frew, between Gargunnuck and Kippen. This was one of only two safe crossing places on the Forth, used most notably by Bonnie Prince Charlie in 1745. Two other streams join the Forth at the point of the Fords, the Boquhan Burn to the west and the Goodie Water to the east. Crawford believed that the old name for the Frew was the Bruit, so that the Fords of Frew, marking the stretch of three streams, was the Tribruit. The site seems more likely than its explanation, as it was a key crossing point, regarded as the gateway between the Lowlands and the Highlands. It was doubtless a frontier for many engagements between the British and the Picts.

  Others who reject this derivation, including Kenneth Jackson whose essay “Once Again Arthur’s Battles” (1945) is considered one of the cornerstones of Arthurian research, do not necessarily reject the location. He notes that tryfrwyd could be used to describe something pierced or broken, and that Nennius’s location is not necessarily a river’s name but a description of a shoreline as “the broken place”. This could still refer to a ford, particularly one where the river is very shallow, leaving sandy and stony banks breaking the surface of the river.

  For another possible location, Barber and Pykitt looked to the story of Culhwch and Olwen, which tells of the hunt for the Irish Boar, the Twrch Trwyth. They suggest that this tale involves a play on words with the River Twrc (also called the Troggy) and trwyth as a variant of traeth. Twrch Trwyth therefore not only meant the Irish Boar, but the “shores of the Twrc”, becoming, over time, Try-Troit, and later Trywruid. They also believe that the story of the Irish Boar hunt is the retelling of a battle between th
e British and the Gewisse with the final decisive battle at the mouth of the river Twrc on the Severn estuary near Caerleon. If the Dulas Brook mentioned above also corresponds to one of the twelve battles, then these two sites may relate to Arthur’s conflict with the Twrch Trwyth.

  One other suggestion of interest is the River Ribble. The Ribble is known as the “roaring river”, a name adopted by the Roman fort at Ribchester, known as Bremetennacum. The Bre-prefix comes from the Celtic breffw, which means “to bellow”. The original river name may therefore have been something like Breffwrd, from which the “b” was eventually dropped, becoming Reffwrd. The Ribble meets the Douglas from the south and the Dow from the north at what is now called Hutton and Longton Sands. Because of this confluence of three rivers it has been conjectured that the locality may have been known as Trireffwrd, which could easily mutate into Tryfrwyd. With the Lune and Douglas in this vicinity, this might suggest a series of battles against the Irish marauders.

  11. The eleventh battle was fought on the mountain which is called Agned.

  This battle has the added confusion that there is an alternative entry. A later version of the Nennius manuscript calls this site Breguoin, and other manuscripts have other spellings, including Bregnion and Bregomion. One manuscript even combines the two as Agned Catbregomion, implying that the two names mean the same site: that is, the battle (Cat) of Bregomion on Mount Agned. Just as with Dubglas and Linnuis, bringing the two together is not easy, though the fact that Breguion and Agned are such rare names means that if we can identify the two we would almost certainly have a unique site.

  Mount Agned is referred to by Geoffrey of Monmouth, though not in connection with Arthur’s battles. He says that the British ruler Ebrauc founded the cities of Kaerebrauc, Alclud and Mynydd Agned. John of Fordun, in his fourteenth-century Scotichronicon, states that Agned was an old name for Edinburgh. Edinburgh was usually called Eidyn and the fortress there, on top of what is now called Arthur’s Seat, was Din Eidyn. It would require some philological contortions to convert Eidyn into Agned, and even more to make it convincing. August Hunt suggests the two are a play on words. Eidyn could relate to the Greek eidon, meaning “to behold or envision”, similar to the Latin agnitio, which means “recognition or understanding.” Both would suggest that Din Eidyn might have been called Mount Agned because both could mean the “Mount of Understanding”. It’s a romantic notion, but one not even hinted at in folklore.

  Geoffrey of Monmouth says that Mount Agned was known as “The Castle of the Maidens”. This relates to the Picts. The right to kingship passed through the female line, and, according to legend, the Picts kept all the eligible royal maidens in the Castle for their security and education. However, Eidyn was a British stronghold and though it was briefly captured by the Picts, it was never settled by them. The Picts had several strongholds along the Forth in the area of the Manau, especially at Myot Hill to the west of Camelon. Their territory was at Stirling and one might imagine the rock of Stirling Castle being an equally suitable site for a “Castle of the Maidens”. However, in the Middle Ages, the Castle of Maidens was always believed to refer to Edinburgh, regardless of any historical accuracy.

  Geoffrey also called Mount Agned the Dolorous Mountain, which may be appropriate as one understanding of the word agned is that it is related to the Welsh ochenaid meaning “sigh”. This makes an interesting connection to the possibility of Wedale as the eighth battle at Guinnion, since the Eildon Hills near Wedale are referred to as the Dolorous Mountains in the Arthurian romance Fergus of Galloway. This may link with the location of Breguoin, which has been shown to derive from Bremenium, the Roman fort at High Rochester, a day’s ride to the south, perhaps suggesting a continuation of the first battle.

  The philologist Alfred Anscombe demonstrated that had Breguoin been spelled Breguein, it would have derived from Bravonium, the Roman fort at Leintwardine, in Herefordshire. This is just west of Ludlow and within hurling distance of the Clun Forest, one of the candidates for the Celidon battle. Bravonium appears in one Roman itinerary as Branogenium. Kenneth Jackson has suggested that Branogenus means “born of the raven”, but it can equally mean “born of the king” or, taking genus in its more general sense, “people of the king.” This could suggest that the original Celtic site of Bravonium/Branogenium was a hill fort occupied by royalty or descendants of royalty.

  Linda Malcor has shown that Bremetennacum, the name of the fort at Ribchester which has the same prefix as Bremenium, would also adapt to Breguoin. Bremetennacum was the fort at which Lucius Artorius Castus was based, and would be a natural candidate during the period when the Picts were attacking the forts south of the wall.

  One other intriguing suggestion arose from a marginal note in an old translation of Nennius found by Joseph Ritson in 1825, suggesting that the site was Cathbregyon, in Somerset. Barber and Pykitt identified this as Catbrain in Bristol, now at one end of Filton airfield. It has also been suggested that this might be Cadbury, where the Saxon Caddesbyrig was derived from the Celtic Cat-bregyon. Cadbury has long been associated with Arthur, as a possible site for Camelot, but not with this battle. However, the name is derived from the personal name Cada, not from Cat for battle.

  12. The twelfth battle was on Mount Badon in which there fell in one day 960 men from one charge by Arthur; and no one struck them down except Arthur himself, and in all the wars he emerged as victor.

  This is the one battle for which we have irrefutable historical support because Gildas mentioned it, and said that it happened in the year of his birth. We have already tentatively dated it to some time in the 490s, probably between 493 and 497. We have also seen that both the original battle poem and the Welsh Annals connect Badon with Arthur. Unfortunately Gildas does not tell us who the commander was at Badon, or where it was fought. Why did he need to? In his day, the audience for his De Excidio would have remembered Badon or have heard of it, and the location was well known.

  The fact remains that Badon was the decisive battle which forced back the Saxons, resulting in a period of comparative peace in Britain. Whoever was the victor at Badon became the Arthur of legend. Tying Badon into the landscape is thus vital in helping identify Arthur.

  Gildas would have referred to the place by its British name, and Badon or Baddon is British for “bath”. Nennius had the same view. In his historical miscellany is a reference to “the Hot Lake, where the baths of Badon are, in the country of the Hwicce.” (§67). The passage then describes what are clearly Roman baths, and the obvious assumption is that the reference is to Bath itself.

  The territory of the Hwicce centred upon Worcester and its main town Winchcombe, but included Gloucester and Cirencester, and thus were the lands opened up to the Saxons following the battle of Dyrham in 577. Bath was tucked in at the southern end of the Hwicce, and the southern boundary follows the River Avon to the coast and along part of the Wansdyke defensive earthwork. However, the Hwicce people must have roamed because their name is remembered as far afield as Whiston (formerly Hwiccingtune), east of Northampton.

  Bath continued to be known as the city of the “Hot Baths” long after the Roman period and into the Saxon. It was called aet Badum in the foundation charter for Bath Abbey in 676AD. Just when the original of Nennius’s list of wonders was compiled is not known. The baths would have to have survived in sufficiently useable condition to be regarded as a “wonder”. Barry Cunliffe, in his excavations at Bath in the 1980s, confirmed that whilst there was considerable stone-robbing in the post Roman period, efforts were made to maintain other buildings well into the fifth and even sixth centuries. The likelihood is that Bath was still a functioning British city at the time of the Battle of Dyrham, and the baths there must still have been held in awe.

  This would seem to prove that Badon must be Bath, but we need to be cautious. To begin with, Gildas does not refer specifically to the town of Bath. His phrase is obsessionis Badonici month, “the siege of Mount Badon” or “Badon Hill”. Bath isn’t on
a hill – quite the contrary. The area of the hot spring which fed the baths was originally in a marshy valley. Bath is, however, surrounded by hills, and most authorities assume that Gildas meant one of those, but which one?

  All other references to Badon date from several centuries later, and they give us three other pieces of dubious information. Firstly, that Arthur fought at Badon carrying the Cross of Jesus on his shoulder, or shield. This presumably refers to an emblem, unless it is meant figuratively, in that Arthur is defending a church, or fighting in the name of Christ, as at Guinnion. Secondly, Nennius tells us that the siege lasted for three days and nights. Thirdly, Nennius also states that Arthur killed 960 of the enemy. This does not mean Arthur killed them single-handedly, but that he led the charge that resulted in such wholesale, and doubtless exaggerated, slaughter.

  If Badon Hill is one of the hills surrounding Bath, it could be one of several. The most favoured are either Banner down Hill or Little Solsbury Hill, both at Batheaston. Analysis by John Morris, for instance, suggests that the Saxon forces were almost certainly infantry, perhaps no more than a thousand strong, whilst the British forces were probably cavalry. Little Solsbury Hill has a major hill fort on its plateau-like summit, which was certainly large enough to house a cavalry unit, but a difficult site for a cavalry charge. It is unlikely that the Saxons would take the hill fort and then be besieged by the British from below. However, the Saxons could have been hemmed in on Bannerdown Hill, less than a mile to the east, which has no hill fort.