After the death of Severus in AD 211 and the creation of the buffer zone by Caracalla, Roman historians were much less interested in southern Scotland. For most of the third century there was little trouble from the north; it seems to have been a period of consolidation. The Romanized Votadini quietly survived and prospered, despite a direct frontier with the warlike Selgovae. And on the Clyde the Damnonii held on to their fertile lands, despite pressure from the north and the Novantes and Selgovae in the south.
The Romans remained watchful. They reinforced five outpost forts north of Hadrian’s Wall: at High Rochester and Risingham on Dere Street in the east, and at Bewcastle, Birrens and Netherby in the west. Cohorts called exploratores were based at these camps. Netherby is described in a Roman road map as Castra Exploratorum. They were detachments of mounted scouts who patrolled the Southern Uplands alert for trouble and able to report it back quickly to the commander of the Wall. There was also a unit of spearmen, the Raeti Gaesati from the Tyrol, based at Risingham.
The medieval ruins of Jedburgh Abbey contain two memories of these frontier years. Trimmed down to serve as building stones, the old abbey Walls contain two Roman altars: one is for the Raeti Gaesati and the other for a cavalry cohort, the Vardulli, based at Rochester. There is a plentiful supply of good stone in Jedburgh and these altars will not have been brought any distance.³⁹ The meeting of the Jed Water and the River Teviot had some strategic significance and it looks as though patrols met there and maintained a presence permanent enough to see altars put up. Jedburgh lies on the fringe of the territory of the Votadini and it looks as though it was a point of regular contact with Rome in the third century. In such a peaceful time, few problems would have been reported to the scouts of their Roman allies.
The Votadini must have been nervous though. They had welcomed Agricola and become consistent clients of Rome throughout the occupation and beyond. Their neighbours had been equally consistently hostile and with the Brigantes and Novantae had inflicted defeat and damage on the greatest fighting machine that the world had ever seen. And yet not only did the Votadini maintain their boundaries after the Roman evacuation, they seem to have thrived. By the beginning of the fourth century they had extended their territory west and north of the Forth, holding a substantial part of Fife.
The only realistic explanation for this is a military one. No amount of Roman statecraft in the second century had subdued the hill tribes of the north; only force had worked and even then for limited periods. The Votadini clearly learned much from the Voconti and the Sarmatians: the skills they acquired on the parade ground and in the riding school at Trimontium had earned them vigorous independence from their warlike neighbours. They had become a dominating military power.
Some of the lessons were also logistical. It looks as though Trimontium was occupied in the early part of the third century by soldiers who had their families living with them. These were veterans who had decided to settle in the territory of the Votadini and it seems likely that the lighting skills of these men stiffened the tribal forces needed to check the Selgovae. They had to maintain the fort for the same reasons that Agricola first built it, as a point of containment at the mouth of Selgovan valleys. And also the Votadini would have wanted to be near their great sacred site at Eildon Hill North. The Military Road built with great labour was designed to deal with chariot-driving enemies approaching from the south-west. This was the Selgovan heartland of the Ettrick and Yarrow valleys. And at Carbantoritum or Carby Hill in Liddesdale, the Romans had had to deal with the charge of the chariots of the hill men much earlier.
The containment strategy of the Romans was extended further south to a place called Yeavering, near Wooler in north Northumberland. This was an important location. The Votadini maintained a large hill fort on Yeavering Bell that dominates the mouth of the College valley which winds up to the slopes of Cheviot itself and into the territory of the Brigantes; and also the Bowmont valley which offers a direct route through the foothills to the ridge country above Kelso and its crossing of the River Tweed. Yeavering was built in an ancient landscape of standing stones, abandoned villages and forts, and its name is old too: P-Celtic from Ad Gefrin, or Goathill. The bell looks east out over the flatlands to the North Sea to the island now called Lindisfarne but known to the P-Celts as Medcaut. It was the first landfall of Anglian pirates in the fifth century.
Archaeology shows decline at Trimontium and at some time in the third century the Votadini abandoned it, although there is evidence that the area to the west of Newstead village was inhabited by people who were still using Roman currency into the third and fourth centuries. Both it and Eildon Hill North depended on the fertile farmlands to the east for supply, and the fort was close to the border with the Selgovae. Prudence persuaded the Votadini to move their centre further downriver to somewhere smaller in scale, more easily defensible and a place where they could corral and pasture their warhorses in safety. As I hope to show, at some point in the third century they came to an old settlement on the Tweed where they built and fortified a stronghold perfectly suited to their military needs.
At the same time the Selgovae were changing. Names as ever suggest this, and a number of scraps of evidence, documentary, archaeological and toponymic, can be combined into a clear pattern. Two tribal groups had been a source of endless trouble to Britannia and its governors: the Caledonii and the Maeatae north of the Forth and Clyde, and the Selgovae, Novantes and Brigantes on either side of Hadrian’s Wall. In the third century it seems that these groups combined so that the hill country from the Highlands down to the Peak District bristled with enemies who became generally known to southern contemporary writers as the Picts.
Like Pretani this is held to be a name derived from body painting or body decoration. The Picts were ‘the painted people’. While this seems a little too easy a connection, there are few alternative derivations on offer. The incidence of the place-name prefix ‘Pit’ or ‘Pet’ broadly falls in with the northern areas of Scotland thought to be inhabited by the Picts, and there is a similarity between the terms – ‘Pit’ and ‘Pict’. But their language is lost to us. They were illiterate and left no inscriptions that can be safely deciphered. However, St Columba’s biographer Adomnan⁴⁰ remembers that the Q-Celtic-speaking saint needed an interpreter when he visited the court of the Pictish King Brude near Inverness sometime after 574. The likelihood is that they spoke a dialect of P-Celtic understandable to their southern neighbours, but not to Q-Celts.
In 296 the usurper Emperor Allectus removed part of the British garrison to support his ambitions in Europe. This was a signal to the Picts, now consolidating north of the Votadini and Damnonii, to assemble raiding parties. For the first time history records the appearance alongside them of a new people: the Scots of Dalriada. The combined force reached as far south as the great Roman fort at Chester, before the empire struck back and retrieved northern Britannia once more.
It seems likely that the Picts and Scots did not invade the province with any thought of conquest. They sought to destroy or weaken their hated oppressors, and most important to plunder their goods and livestock. They also avoided pitched battles with Roman legions and mostly waited for opportunities like Allectus’s withdrawal to diminish the garrison of Hadrian’s Wall and empty the forts which controlled and policed the roads of Britannia.
They also had to avoid the waxing kingdoms of the Damnonii on the Clyde and the Votadini in the east. I believe that they did this, and inflicted raids in the far south hundreds of miles from the Highlands of Scotland, by forming an alliance with those other implacable enemies of Rome, the Selgovae and the Brigantes. There are pieces of evidence that can be brought together to make a clearer picture.
Names first. Immediately to the south of Edinburgh lie the Pentland Hills. Rising steeply to 1,600 feet the range tails south-west into the heart of Selgovae territory. Pentland is derived from ‘Pictland’. It is the same name that the Norse called the barrier sea between the Orkneys and Caithness: ‘Pettaland-fjord’, the Pictland Firth, the Pentland Firth.⁴¹
There is also a good deal of archaeological evidence for Pictish influence and immigration to the Southern Uplands. On the Borthwick Water near Hawick there is a beautiful Pictish symbol stone of a sort generally only found far to the north. It is carved with the image of a fish and is not something that could have been imported, dropped or even moved. It must have been erected by a group of people who understood its beauty, a group of Pictish settlers.⁴²
In the Pentland Hills there is at Castlelaw Fort an example of Pictish domestic architecture. It is a souterrain or underground chamber, a place probably used for food or grain storage. There is another nearby at the village of Crichton and, most significantly, two examples south of the old Roman fort of Trimontium. One of them is built with stones from buildings in the settlement and must therefore postdate both Roman occupation and Votadinian control. They were probably built in the fourth century.
In addition there are ten forts of what is known as the nuclear design distributed throughout the Selgovan hills.⁴³ Again in the north the Picts built in exactly the same way all over their territory. One of the most elaborate of these is a stone-walled structure on Carby Hill. And even more characteristic, and expensive, is the Torwoodlee Broch just outside Galashiels. Brochs are towers (in shape not unlike a miniature cooling tower at a steelworks) of immense thickness which are far more typical of the Highlands and Hebrides, but clearly a powerful man found himself in the valley of the Gala Water and his way of showing prestige was to have a broch built. Since this was destroyed in the second century ad, probably by the Romans, it is unlikely that this man was a Pict, more likely a Selgovan chief who had had contact with Pictish allies, and had seen a northern broch. There is also evidence of two other brochs on the Gala Water, at Bow Castle and Crosslee. And there is a much larger stone-built hill fort at the Rink between Selkirk and Galashiels; about ten times the size of a broch, it uses the same dry-stone techniques and has a wall thickness of about thirteen feet. It is similar to the fort on Carby Hill in Roxburghshire. The sixteenth-century spelling of the name is Langrinck which is related to the derivation of the old county town of Lanark, which in turn is from lanerc for a cleared area. This place name has a distribution in Scotland almost exclusively in Pictish territory. Telling connections.
Two more notices are important. Ninian is the first named Christian of Scottish record, and he flourished between 360 and 432. Almost exactly 300 years later Bede wrote in his Historia Ecclesiastica that Ninian had converted the southern Picts long before Columba had preached to the northern Picts.⁴⁴ These two groups lived either side of the central Highland massif, explained Bede. ‘The southern Picts who live on this side of the mountains’ were the object of Ninian’s mission when he founded the white church of Candida Casa at Whithorn in Galloway. This must mean that Bede thought of the Novantae, the Selgovae and perhaps their allies the Brigantes as ‘the southern Picts’. If they were not, then Ninian founded his church a very long way from the people he hoped would populate it.
The other glimpse of the Picts in the Southern Uplands is on a Victorian map which insists that the Catrail, the defensive or demarcating earthwork separating the Selgovan hills from the Anglian flatlands, was properly called ‘The Picts Work Ditch’. Meaning that it was made by them.⁴⁵
There is a final word on this from Gildas, the sixth-century writer who bemoaned the state of his country in On the Fall of Britain.⁴⁶ He explains that the Picts, the greatest enemies of Britain, had penetrated right up to the Roman wall by taking over the lands of the native peoples. Since the Picts were already very close to the Antonine Wall, Gildas must mean Hadrian’s and the conquest of the lands of the Selgovae. Alliance or conquest, it is difficult to tell.
However, it is clear that for logistical reasons the Picts, the Selgovae and the Brigantes would have found it advantageous to combine. Not only were they more powerful, gaining access to the spine of Britain, from the Highlands to the Peak District, unhampered by the Damnonii or the Votadini, and the remaining Roman garrisons of Lancashire and Yorkshire, the Picts’ alliance also allowed them to raid far into the south of rich Britannia. The classical historian Ammianus reports Pictish war-bands attacking London. The Romans found them difficult, elusive opponents and later writers such as Bede and the Anglo-Saxon Chronicle compilers called them the scourge of Britain, its greatest enemies. Later I will show that Arthur fought a brilliant campaign against the Picts and on five occasions brought them down from their hills to battle and defeated them.