by Graham Robb
The second and third and fourth and fifth battles were on another river, which is called Dubglas and is in the Linnuis region.
Now, when I looked back at the second-century map, figures seemed to be moving across it. An invasion force landing on the west coast, in the bay which Ptolemy calls Vindogara, might, like the later Viking invaders, have followed the river valleys to Lindum and the Dubglas.
Suddenly, the map lit up with another lost connection. The Roman road between the Irish Sea and Lindum runs alongside a river which flows into the sea at Irvine. The river is now called the Irvine, but, ‘strictly speaking, its parent stream, on account of its length and the volume of water it carries’, is the river Glen.50 Of the three British rivers called Glen, this is the only one that could be said to have a mouth or an estuary (‘ostium’).
The first battle was at the mouth of the river named Glein.
There was no longer any doubt that these places had existed. Not only that, but they also appeared to have been listed in a logical, geographical order. This was a campaign which had actually taken place. The date of the campaign and the identity of the ‘leader of battles’ were still a mystery, but pieces of that Dark Age puzzle were beginning to fall into place like the tumblers of a lock.
The twelve battles of Arthur are usually thought to have taken place in the sixth century. The fact that the key had been supplied by a map of the second century seemed to be no more than a fortunate coincidence. I made a note of the discovery and filed it away. It had no obvious connection with the history of the Borders and the Debatable Land, and I was not keen to make it public. By then, my book on the ancient Celts had been published. One of its themes was the seductive power of insignificant coincidences. I had illustrated the point with a jokey reference to the defunct ‘Camelot’ amusement park near Wigan in Lancashire. This had generated garish articles in two national newspapers and a corresponding chorus of tweets and retweets: a ‘historian’ with my name and face was claiming to have discovered the fabled court of King Arthur just off the M6 motorway.
The newspapers reported this as though Camelot, that Hogwarts of the Middle Ages, might conceivably have existed. If I published the discovery now, it would be tantamount to saying, ‘I haven’t found Camelot, but I am on the trail of King Arthur – and, by the way, he’s Scottish!’ I decided to attribute the discovery of the battle sites to an anonymous reader and pursued the investigation.
*
It soon became clear that the map had more to say on the subject. It revealed one other battle site and confirmed the locations of two others, leaving only three unidentified sites. It also suggested a rational itinerary which might provide clues to the three remaining sites.
Those place names had not been conjured out of thin air by a Brittonic bard. The course of a forgotten war traced itself across the map, and the ghostly warriors of Arthurian legend began to look substantial, imbued with purpose and direction. The date of the battles was narrowed down to a period much closer to the map of Roman Britain than to the age of the Saxons. In fact, there was no sign at all of the Saxons whom the compiler of the Historia Brittonum supposed to have been the enemy: no Saxon invasion ever took place on the west coast of Scotland, and there are no Saxon settlements anywhere in the region.
If the battle list pre-dated the Saxons, who, then, had been the enemy? Piratical raids from the Irish Sea went on for centuries, but there was no full-scale invasion, and Vikings did not arrive in Irvine Bay until several decades after the Historia Brittonum was written. The text refers to a confederation of British kings fighting under a single leader. ‘King’, in Latin and in ancient Celtic, was the usual term for a tribal chief. This might suggest the Dark Age kingdoms which were established or restored at the end of Roman rule in the 400s, but the battles of this period seem to have pitted Briton against Briton and there is no trace of any pan-British alliance.
The historical chronometer has to be wound back to an even earlier age. The first two battle sites, and those to come, are Roman rather than British. Most – perhaps all – of the battles took place, not at Celtic hill forts, but at key military installations on the Roman road network. The original source seems to have had an unusually good grasp of the geography of a wide area: this is the most coherent sequence of place names in early British literature. Along with the alliance of kings under a single leader, this would imply a relatively stable political situation which was troubled for a time by an invasion sufficiently momentous to be commemorated in a poem.
One possibility, first raised in 1924, is that ‘Arthur’ was a Roman. A certain Lucius Artorius Castus commanded the Sixth Legion in Britain, which had its headquarters in York. An experienced and successful soldier, he had served in Syria and Judea, and later led two legions against the Armenians. The theory has been discredited by the tenuous speculations on which the chronologically deranged film King Arthur (2004) was based, yet the twelve battles do appear to have been fought between the Humber and the Firth of Clyde, which was precisely the area controlled by Artorius’s Sixth Legion.
Perhaps the name ‘Artorius’ stuck in the popular mind. Later, Celticized as ‘Art(h)ur’ or ‘Arto-rix’ (‘bear king’), the Roman name might have been attached to a home-grown British hero. But if Arthur’s army was a combined force of Romans and Britons, this would have been a strange campaign to celebrate in a Brittonic poem. Most of the battles would have been defeats, with the invaders pushing ever southwards through the lowlands, the borderlands and the Pennines until they reached the heart of Roman power at the ‘City of the Legion’.
The ‘Roman Arthur’ theory has proved compelling because it fits the traditional fable of the nation’s origins: the daggy, tartan-clad warriors of ancient Britain who skulked in smoky huts like people of the Stone Age, living on porridge, roots and beer, were given the gift of civilization by the Roman army. This view, which most Romans would have shared, has been repeatedly demolished by archaeology. The Celts had towns and roads, high-speed transport and well-managed farms. They used metal-working techniques which have yet to be reinvented. Several ancient texts refer to their meritocratic education system. Yet these sophisticated ancestors are still viewed with the same colonialist prejudice with which the metropolitan Scots and English of the Middle Ages regarded the barbarians of the borderlands.
The map of second-century Britain turns this tale on its head. This was not a Roman-led campaign of resistance to foreign invaders. The course of the war makes sense only if this was, as the battle list says, a British force united against a common enemy. That enemy was neither Saxon nor Celtic. The kings of Britain banded together – as Celtic ‘kings’ or chieftains often did in times of national emergency – in order to reconquer from the Roman usurpers the lands that had once been theirs.
*
A few weeks after finding the first battle sites on Ptolemy’s map, we set off from home, crossed the border into the Debatable Land and headed north. By then, I knew that the itinerary revealed by the map agreed with recorded history. The Roman chronicler Cassius Dio had described a major invasion of Britannia which took place in the early 180s.51 This invasion is consistent with the sequence of battles, and it was on a scale that might well have secured it a place in legend. According to Cassius Dio, it was ‘the biggest war’ fought anywhere in the Roman Empire during the reign of Emperor Commodus (AD 177–92). For the Romans, the invasion was a near catastrophe, but for the allied tribes, it would have been a formative event in the birth of a united British nation of the north.52
27
The Great Caledonian Invasion
Despite appearances, this is the end to the story of the Debatable Land: the discovery of its lost beginnings and the recognition of its lasting place in British history. The future of another United Kingdom was soon to be decided, and although physical geography in the twenty-first century seemed to have become detached from political history, there was a powerful sign of long-term forces at work in the fact that the paths of that
epic campaign of the second century led back to the borderlands.
*
In about AD 163, the Antonine Wall was abandoned and Hadrian’s Wall, eighty miles to the south, once again became the main line of defence. To control the troublesome northern tribes, several forts between the two walls retained garrisons. The borderlands then became a Roman buffer zone or limes, with command centres at Castledykes (Lindum) and Newstead protecting the main roads to north and south.
In the early 180s, ominous tidings reached the south. The unconquered tribes of Caledonia had ‘crossed the wall which separated them from the Roman legions’: ‘they proceeded to do much mischief and cut down a strategos [a general or military governor] together with his troops’. The account by Cassius Dio is corroborated by signs of destruction or rebuilding at various forts in northern England and the Scottish lowlands (fig. 14).
Fear of invasion spread far beyond the frontier. In southern Britannia, a massive building programme was launched. Between the mid-180s and the mid-190s, many towns, both large and small, which until then had been open and undefended, began to surround themselves with earthworks. Nowhere else in the Roman Empire was there such a rush to protect the urban population. The crossing of the wall (whether Hadrian’s or the Antonine) and the killing of a Roman general and his troops were alarming enough to induce a sense of panic, and there is palpable evidence of the relief that was felt at the end of the emergency in the coins which were struck in 184 and 185 to celebrate a victory in Britannia.
*
More than a thousand years separate the Border reivers from the confederate army of British kings, but the geography of the region is largely unchanged and the pastoral, cattle-raiding society of the Celts would have been recognizable to a medieval warden. The Caledonian warriors were ‘very fond of plundering’, says Cassius Dio. ‘Consequently, they choose their boldest men as rulers.’ They ride ‘small, swift horses’ and ‘can endure hunger and cold and any kind of hardship’.
Cassius Dio probably owed his information to a Roman Robert Carey who had seen his enemy disappear into the trackless mosses:
They plunge into the swamps and exist there for many days with only their heads above water, and in the forests they support themselves upon bark and roots, and for all emergencies they prepare a certain kind of food, the eating of a small portion of which, the size of a bean, prevents them from feeling either hunger or thirst.
These hardy tribes of the northlands, equipped with nimble ponies and the Iron Age equivalent of the energy bar, either marched south from the Antonine Wall or sailed down the coast from the Firth of Clyde. After an opposed landing at the mouth of the Glen in Irvine Bay – which was once the largest port in western Scotland – another four battles were fought along the Douglas in the region of Lindum (Castledykes). Here, the road connecting the Antonine Wall with Hadrian’s Wall met the road which crossed Britain from the Irish Sea to the North Sea. For several days, weeks or even months, this vital crossroads of the northern British network became the theatre of war. After the fifth battle, the invaders finally broke through into the Southern Uplands where the roads led south into the heart of Roman Britain, towards the place which the battle poem calls ‘the City of the Legion’.
The sixth battle, on the river ‘Bassas’ – somewhere on the route from Castledykes to the border – was probably at the fort of Tassies Height (Coria, on Ptolemy’s map) along the shallow river Annan, which flows to the Solway Firth. The seventh was in a place we now know well: Celidon Wood, where the bard called Myrddin took refuge after the Battle of Arfderydd in the parish of Arthuret (here).53
Remnants of that wood still exist along the steep banks of the Liddel where most of North Cumbria’s ancient woodland grows, preserved from browsing deer and wood-cutting humans by its inaccessibility. On a modern map, the patches of old broad-leaf forest on the Debatable Land boundary from Netherby and Carwinley to Penton Wood look like the shredded cloak of someone fleeing through the thickets towards the Kielder Forest. They are still a refuge for several threatened species of plant and animal.
This would therefore be the earliest known of the many battles that were fought around the Debatable Land or, as it was then, the buffer zone of the Damnonii, the Selgovae and the Votadini. A victory in this crucial area would have opened the way to the south. If ever ‘Arthur’ saw the twin knolls of Arthuret, it would have been after his seventh battle. The wood which covers the knolls is known locally, but not on maps, as Crow Wood. The people of Longtown who played in Crow Wood as children like to think of it as the grave of King Arthur. On a foggy day, when the lower slopes are draped with shining white sheets of plastic mulch, the site has an almost ceremonial air.
*
Invasions tend to follow the same routes from one century to the next. Like the Jacobite army in 1745, the Caledonian invaders headed for Carlisle, where an inscription of the AD 180s, unearthed on the site of Bonnie Prince Charlie’s headquarters in English Street, gives thanks to the deified emperor Hercules-Commodus for ensuring the safety of the garrison when it was attacked by a barbarian horde. The inscription does not explicitly commemorate a victory: the ‘barbarians’ may simply have passed through that notoriously pregnable city. The battle list shows that they had a more distant goal: beyond Shap and the forbidding and beautiful gorge of the river Lune, where the drystone walls and teetering sheep give an exact measure of the fells’ steepness, lay the fort of Guinnion.
The Great Caledonian Invasion. The uninterrupted lines show the probable route; the broken lines show the possible route. For more detail, see figs 14 and 15.
‘Guinnion’ is a later form of ‘Vinnovium’, which is shown on Ptolemy’s map of northern England. Previously assumed to be Binchester on the other side of England, it can now be identified as Lancaster, where vestiges of a Roman fort have been found on Castle Hill. The geographical logic of the battle list is obvious: Lancaster lies on Roman roads to Chester and to York, and, like several of the battle sites, it was an inland port. A century before, Agricola had advanced up the west coast with the fleet on the Irish Sea supporting the ground forces. The British kings might have used a similar strategy. The first eight battles – from Irvine Bay to Lancaster – suggest a mass attack on the western seaboard, while the later battle sites are consistent with a parallel or subsequent invasion from the east.
After the victory at Guinnion, which the author of the Historia Brittonum attributes anachronistically to the power of the Virgin Mary, the hypothetical route divides. The ninth battle took place ‘in the city of the Legion’. This must be either Chester (labelled ‘Legio XX’ on Ptolemy’s map) – which was no more than a ‘rearward works establishment’ in the 180s – or, more likely, York (labelled ‘Legio VI’), which by then was the main base of Roman power in the north. For the British kings, the legionary fortress at York would have been the more significant target. The final battles were fought north of the Humber, and so ‘the Legion’ would naturally have referred to the Sixth, which operated all over the north, from Manchester to the Antonine Wall, and which, at that time, was stationed in York under the command of Lucius Artorius Castus.
A spectacular Roman road runs through the Aire Gap and over the moors to the city of York. By following this route, the Britons would have remained on familiar, upland territory. Along that road of terraced climbs and hair-raising descents, two forts, Ribchester and Ilkley – both shown on Ptolemy’s map – were rebuilt in the late second century and may have been wrecked in the invasion.
The tenth battle, at ‘Tribruit’, is a mystery but the search can be narrowed to a tidal estuary on the North Sea coast – probably Tweedmouth near the eastern end of the future Anglo-Scottish border or South Shields at the terminus of Hadrian’s Wall. The site of the eleventh battle is less mysterious: Bregion or Bregomion is the fort of Bremenium, which stood at a junction on Dere Street. Its location is confirmed by Ptolemy’s map: it lies off the main Newcastle road below the border pass of Carter Bar, beyond t
he rotting hulk of a reconstructed Iron Age house and a sign marked ‘No Access to Military Vehicles or Troops’.
Bremenium was one of the forts in the frontier zone north of Hadrian’s Wall which retained a Roman garrison after the abandonment of the Antonine Wall. Its ruins are as evocative of Border reivers as of Romans. In the sixteenth century, a bastle was built out of the crumbled walls as a defence against the rampaging Elliots of Liddesdale. But some of the stones were left in situ and several burial mounds and monumental tombs have been found nearby. The cemetery at Bremenium remained in use from the early second to the early fourth centuries and was apparently reserved for officers. Perhaps it was there that the general mentioned by Cassius Dio was ‘cut down together with his troops’.
*
By the time the British rebel army reached Bremenium, it would have covered more than four hundred miles. It had ravaged northern Britannia, caused panic in the south and was being talked about in Rome. The British kings and ‘Arthur’ now found themselves on the principal frontier of the Roman province and the only permanently defended border in the far north of the empire.
Arthurian scholars have long suspected that the twelfth and last battle in the Historia Brittonum list was borrowed from another source. The famous Battle of Mount Badon was fought in an entirely different time and place: in the south of England, in the sixth century and against the Saxons. As a British victory, it provided the ninth-century historian with a gratifying conclusion, but, as we know, second-century Britain was not reconquered from the Romans by a barbarian horde. Another ‘last battle’ is recorded in the tenth-century Welsh Annals: the Battle of Camlann, at which Arthur died. Many scholars believe that this unhappy episode was the original ending. The itinerary suggested by Ptolemy’s map supports this theory, as does the fact that ‘Camlann’ would have fitted the rhyme scheme of the battle list.
