On the far side of the Fairgreen stood the fair church of St James. Its bells rang for the last time in the middle of the seventeenth century; by 1649 there were only six communicants.¹⁵⁵ And in the nineteenth century the Duke of Roxburghe successfully resisted the temptation to fall in with fashion and retain the ruined aisle of St James which offered a romantic vista from the windows of Floors Castle. He decided that horse racing would look better and demolished the old church so that jockeys could enjoy an uninterrupted view of the next fence. Accordingly nothing at all of St James now survives except a tombstone, rescued from the cemetery and placed in Kelso Abbey. It and another offer a window into the workings of the old city.
‘Here lies Johanna Bulloc who died in the year of Our Lord 1371. Pray for her soul’ is inscribed on the recto side of one, while on the verso is carved a pair of sheep shears. The other stone is from the Franciscan Friary of St Peter which stood nearby the wych elm used in the Whipmen’s Ride, and which was also flattened. All that can be made out on the face is the word ‘mercer’ but the stone is interesting. It was quarried near Tournai in France out of the distinctive pierre bleue limestone.¹⁵⁶
The medieval wool trade underpinned Roxburgh’s prosperity. Like much that thrives in commerce, it was a simple business. Sheep farming was organized on a semi-industrial basis by the great monastic houses of the Borders. Melrose, Jedburgh, Dryburgh and Kelso all controlled vast sheep ranches in the Cheviot and Selgovan Hills, in places that kept their P-Celtic names and where, like in Cumberland, the shepherds probably still spoke a form of Welsh. Each ranch produced a carefully calculated crop of wool. There exists a record of this compiled by a Florentine merchant called Pegolotti in 1340.¹⁵⁷ At Roxburgh market, under the keen eye of royal burgh tax officers, the wool packs were sold to Italian and particularly to Flemish merchants. The importance of the latter is remembered in the high incidence of the surname Fleming around Kelso. They carted the fleeces along the Via Regis, the King’s Highway, which led to Berwick-upon-Tweed and ships bound for the cloth-making towns of the Rhine estuary. So that the wool could find its way easily to Roxburgh as well as from it, there was another Via Regis which led from Annandale up over Teviotdale and down to the old city.¹⁵⁸ Both of these were based on Roman roads. The regulations for Roxburgh’s busy market were set as a paradigm for other emergent towns; the Leges Burgorum were applied to Dumbarton and elsewhere in the Middle Ages.
Because no archaeology is allowed and no plan survives it is extremely difficult to visualize the place or get a sense of how big it was. However, a careful and patient reading of the land grants made inside the town walls offers some clues. Three streets are mentioned. The Headgate led out of the east gate of the castle and at some point changed its name to King’s Street, while Market Street ran crosswise from south to north. The arrangement was probably similar to Edinburgh where Castlehill runs out of the east gate of the castle before becoming the Lawnmarket and is then crossed by George IV Bridge. Not so steep-sided and smaller in scale than that, but similar.¹⁵⁹
Roxburgh had a long circuit of walls, possibly only a wooden stockade since Tweed and Teviot already offered protection, which was extended at least once to include a new town which had grown up in the early twelfth century. Two churches stood within. In addition to St James there was Holy Sepulchre which, documents suggest, may have been a rotunda church. To the south of the walls stood the Friary of St Peter, founded in 1231 and deriving its income from control of the ford of the River Teviot. Friars Cottage, standing now at the north side of Teviot Bridge, still has the name. A fourth church, possibly the oldest, stood inside the castle walls. This was St John the Evangelist and an important place that saw audiences with papal legates, treaties signed and royal baptism.
King David I MacMalcolm knew that a growing economy needed more than high levels of production and wide distribution. He needed to create his own standard of exchange and in 1135 he caused Scotland’s first coins of the Middle Ages to be struck at Roxburgh. Silver pennies were minted in sufficient number to allow numismatologists to discern the hands of five coin-makers working in the city.
An ability to count was also needed, and Scotland’s earliest schools come on record in 1251 when a man called Thomas is noted as rector of more than one. If there were two then it is likely that they were attached to Roxburgh’s municipal churches of Holy Sepulchre and St James.
If wool paid the bills then the castle protected the profits of the wool trade at Roxburgh. It was the initial raison d’être for the whole set-up. Placed at the narrow neck before the wide haughland opens up to be safely bounded by the two rivers, the castle’s location impressed Henry VIII’s general, the Earl of Hertford: ‘It is one of the strongest seats of a fortress that I have ever seen.’ Very large, on a steep-sided oblong mound closely bounded by Teviot and Tweed, Roxburgh Castle commanded a wide view on all sides. It had seven substantial towers, a donjon or keep in its centre, extensive barracks and stabling, a well and a church. State events were regularly staged within the walls and so impregnable was it thought to be that political prisoners and hostages were imprisoned in its dungeons.
One final part of the medieval picture remains to be drawn. David I originally founded his Tironian abbey at Selkirk on the River Ettrick near the hamlet of Lindean. He was persuaded to uproot the French colonists and bring them downriver to Kelso ‘near Roxburgh’ as the documents have it. The kingdom was prospering, the economy expanding and David needed a civil service to help run it efficiently. Abbeys produced literate, numerate and educated men who owed everything to the king, whose probity and loyalty were unquestioning. But why did he not plant the new church by the town of Roxburgh, at his elbow?
The reasons are nowhere written down and seem to be implied in religious tradition and precedent rather than administrative logic. It is clear that there was a pre-existing church at Kelso dedicated to St Mary. Like Cuthbert’s Church at Old Melrose, it stood in a loop of the river, an allusion to the hermetic inclination of the early Celtic Church. Not an island like Iona or Lindisfarne cut off from the sinful temporal world, but almost. The comparison with Melrose is significant in another way. When David I brought Cistercian monks to build a new abbey in 1136, they moved it from the loop of the Tweed where the site may have been too restrictive. But they kept the dedication to St Mary. The Tironians did the same thing at Kelso. Remembering the ninth-century tradition of a pilgrimage shrine to the Virgin at Stow in Wedale and Arthur’s bearing her likeness on his shield at Guinnion, and at Badon, is there a whisper of a cult of St Mary in the Borders in the Dark Ages?
At all events David I lavished gifts on Kelso, their value incalculable in modern terms, and quickly made St Mary’s the wealthiest and most politically significant abbey in all Scotland.¹⁶⁰ Her abbots were statesmen, her priors preferred to great ecclesiastical office and the double cruciform design of the church made Kelso the largest and most impressive ecclesiastical foundation north of Durham.
The medieval glory of Kelso and the power of Roxburgh lasted for only two centuries. In 1296 Edward I of England turned his armies north and began 300 years of intermittent but brutal Border warfare. Roxburgh and Kelso lay too near the frontier to prosper; the castle was occupied by the English for nearly 100 years and the abbey was burned and looted repeatedly. By 1460 Roxburgh Castle was destroyed by the Scots in an artillery fusillade which killed King James II in a fearful accident, when a cannon burst near where he was standing. He had wanted to destroy Roxburgh so totally that should it fall into English hands again, it would be useless to them. The town began to decay in the fourteenth century, and by the mid seventeenth century the growing town of Kelso had taken over many of its functions as a market. In the eighteenth century Roxburgh’s buildings and churches were used as stables and cow byres and by 1800 there was nothing to see, no stone left standing on another save the ruined arches of St James and the Friary of St Peter.
But whatever the historical reason, and by an
y measure, the complete disappearance of Roxburgh is astonishing. Because there is nothing to see, no tour of the ruins available, no physical investigation allowed, this great city and all that happened within it has been ignored or underplayed by historians. No one goes there except local people looking for a good walk, or punters attending the point-to-point races twice a year. In every sense Roxburgh is a lost city.
Be that as it may, it is important to be clear about the connection to this present narrative. Context is the answer. When the city first comes on record as such a substantial and vibrant centre, it must have been a going concern for a considerable time before 1113. Its military significance is obscured by time but no one in the Middle Ages missed it and before that its virtues as a stronghold would have been recognized. Economically it is clear that the wool trade was operating in some volume before the monks arrived in the early 1100s with their tally sticks. Relative political stability followed Malcolm II MacMalcolm and King Owain of Strathclyde’s victory at Carham over the Northumbrians in 1018, but even before that the castle and town were in existence. The evidence needs a little background.
In 617, after the Gododdin rout at Catterick in 600, Edwin united the Angles of Deira in Yorkshire and Bernicia in Northumberland into the army of Northumbria. Twenty years later they captured Edinburgh and overran the Tweed basin which they held until Carham. In that time the Angles conferred many new place-names; some of them recall actual events such as at Pallin’s Burn where St Paulinus baptized the heathen aristocracy of Northumbria.¹⁶¹ That happened in 627 and since Pallin’s Burn is only twelve miles from Roxburgh, it shows how close the Angles had encroached only a century after Arthur’s death.
As in modern political takeover, only the previous elite suffered immediately and most P-Celtic peasants were left on the land they cultivated. The incoming ruling class needed to eat just as the outgoing did and it would have been foolish to drive food producers away. Instead the Anglian chiefs seized places that offered control, and one of them made straight for the strongest fort in the Tweed basin and gave it and the settlement around it his name.¹⁶² He was an aristocrat called Hroc. Like hrosa for ‘stallion’, hroc is an early animal name and it means ‘rook’. The old P-Celtic name died and soon the place became known first as Hroc’s Burh then Rokesburg and later Roxburgh. It is impossible to date this event but given the importance of the site and the dateable taking of Edinburgh in 638, Hroc must have made himself master of Roxburgh early on in the Anglian occupation of the Tweed basin.
As ever, names of places whisper secrets. When the Anglians conferred new names they were very often after people. This was in contrast to P-Celtic names which mostly described geography and sometimes function. Just as Dodin’s Tun replaced Tref yr Llin near Edinburgh in an earlier example, Hroc stamped himself on a place whose former name described what happened there. The ghost name of Roxburgh never completely faded and in several disparate places it is also called Marchidun, Merchidun or the Marchmound.¹⁶³ In lazy ignorance I had believed that this was a reference to its proximity to the marches between England and Scotland, rather as in the aristocratic title the Earl of March acquired from his ancestors’ role as marcher lords.
Taking the earlier form Marchidun, the last element is familiar as a P-Celtic word for a fort. There are many ‘duns’ in Celtic Britain and I had missed a simple reference. It led me to open dictionaries to find the first element. March or Marc has nothing to do with frontiers. In both P- and Q-Celtic it has a very specific meaning: marc is a ‘cavalry horse, a charger, a steed’. Marchog is a ‘cavalry warrior’, marcslaugh is a ‘war-band of cavalry warriors’.¹⁶⁴ There are many historical references. But one crystal thing was clear: Marchidun means ‘cavalry fort’. And it meant that before Hroc made Roxburgh his place in the early seventh century.
At first I found the name puzzling. Forgetting the old town and the haughland, I wondered about the practicality of keeping horses enclosed in a place so steep-sided and confining where feed and water would have had to be handled in large quantities up the castle hill. But as soon as I looked at an aerial photograph everything fell into place. Roxburgh, or Marchidun, is a perfect place to base a large cavalry force. The logistics tell the story by themselves.
When armed and mounted, Arthur’s cavalry was a fearsome, almost irresistible force if massed in numbers on the right ground, choosing its moment to strike. When dismounted they were very vulnerable indeed, much more than their infantry opponents. If a foot soldier is surprised all he has to do is reach for his spear and shield and he is ready to defend himself. If a cavalry warrior is threatened with sudden attack he has first to arm himself, then catch his horse which may be some distance away, then he has to get a saddle and bridle on it, and then at last he can mount and be ready to fight. That in turn dictated much about the organization of horseback warfare.¹⁶⁵ First and most vital a secure, defensible, difficult-to-access base was needed to house dismounted warriors overnight and also, near at hand, another safe, probably fenced area was needed which had pasture and water for horses. Both of these places, finally, needed good all-round vision to allow early warning of attack. Few sites like that exist.
Large ramparted hill forts often enclose a wide area where both men and horses can be safe but very often water has to be brought up or existing wells are too piddling to provide the 400 or so gallons that 300 horses need every day. Good for overnight but useless for much longer, hill forts that were large enough to accommodate the animals often had long perimeters which were impossible to defend against concerted attack, particularly under cover of darkness. Something Venutius and his Brigantine and Selgovan warriors discovered to their fatal cost at Stanwick 300 years before.
Roxburgh is a perfect place for a cavalry fort. The castle mount is very steep, and shows signs of having been made steeper on the northern side where what looks like a dry moat may have been a Dark Ages ditch; the modern road makes it very difficult to visualize. And its perimeter could be easily defended by 200 to 300 well-organized soldiers. In any case, sentries could see for up to two miles on every side. On the south the River Teviot comes very close to the foot of the mound, while on the north the Tweed is only 150 yards away. By itself the castle mount is a superb place for a stronghold.
But sitting at the narrow neck of a huge area of haughland bounded by two large rivers, it has no parallel with any other site in Britain. There are upwards of 400 acres of pasture with drinking water all around. Not only do the Tweed and Teviot keep horses in, they are too deep, except in one place, for men to cross without boat or bridge. At the narrow neck at the east end of the castle mound, where the Headgate would have started, there is an interesting archaeological detail. Laid bare by the roots of an old chestnut tree are the first few courses of a massive stone wall. The Tweed is only 150 yards to the north and that point would have been the prime place to throw a defensive wall across the neck of the haugh. Any attacker would have had to risk a sort of crossfire from running the gauntlet of 200 yards of the castle ramparts before they reached a wall built at that point.
The haughland is so large that with careful management several cuts of hay could have been taken to provide winter feed when the campaigning season was over. And there is also enough room for a herd of brood mares to live all the year round, supplying replacement mounts as they were needed. And by the rivers there are tracts of flat ground for breaking and schooling and for getting both horse and rider fit for the summer.
My contention is that Arthur used this uniquely suitable place for his base, but also he did so using Roman training techniques mixed with Celtic horse lore augmented by what seeped into Borders horse culture from the training schools of Trimontium. But there is a much more secure and recent Roman military connection to be traced to Roxburgh.
After the Barbarian Conspiracy of 367, Count Theodosius set prefects over the P-Celtic kingdoms. He cannot have expected these men to impose their authority without some sort of military backing and since b
y the end of the fourth century most Roman frontier troops were mounted, it is probable that they commanded troops of cavalry. Paternus Pesrut, the man with the red cloak, was given authority over the southern Gododdin in the Tweed basin. My belief is that he did not consider reoccupying Trimontium because he needed a legion to do it and all he had was an ala or wing of 300 cavalry (the same number of Gododdin warriors who rode to their deaths at Catterick). For reasons of logistics I think he built a fort on the castle mount at Roxburgh and ran his horses on the haughland. If only military common sense and toponymy was behind this claim then it might be less than substantial. But the choice of Roxburgh as base for Paternus Pesrut is underpinned by a remarkable and recent find of Roman coins.
Marchidun circa AD 500
Directly opposite Roxburgh Castle on the south bank of the Teviot lies the old Springwood Estate. It used to belong to the Douglas family who had built a beautiful country house, now sadly demolished. The grounds are now occupied by a caravan site and a retirement village, while the field bordering the Teviot is often ploughed and sown for barley or hay. Field walkers and a very learned and persistent local historian have found more than 300 Roman coins in Springwood and fifty-four of these date from the time of the Emperor Valentinian of 364 to 378, the man for whom Theodosius renamed the province Valentia. The coins are not gold or silver or objects of any sort of intrinsic value. Most are bronze and measure half an inch across like five pence pieces or cents. Neither treasure, nor booty, they are the small change of a money economy and they arrived at Roxburgh in the pouches of Roman soldiers. And, more, they were found not in the castle mount or on the haughland where archaeology is forbidden but on a site separated by a wide river from the centre of Roxburgh.
Arthur and the Lost Kingdoms Page 24