Last of the Westland Whigs

In the late 17th century, the 'Westland Whigs' were the radical descendants of earlier Covenanters who had defied the absolutist rule of Stuart kings in south west Scotland.

Monday, April 18, 2016

Vikings, Normans and Cumbrians

In 1998 I visited the Royal Blind School in Edinburgh where my son Callum was going to become a pupil. Before catching the train back home I popped into a bookshop and found ‘The Uses of Place-Names’ edited by Simon Taylor which had just been published. It was and is an utterly fascinating book, Since reading it I must have spent hundreds of hours poring over new and old maps, searching through historical texts, plodding across muddy fields, pushing through thickets of trees and navigating tracts of moorland in search of named places. I have also written thousands of words about what I found in these different explorations.

However, although I can find the places and their history over the past few hundred years, I do not have the deep linguistic knowledge of place name experts. I cannot tell if a particular place name is Old English or Old Norse, or if a word which looks Scots was originally Gaelic.

For example ‘Lockhart Hill’ in Balmaghie parish, Stewartry of Kirkcudbright looks as if it was named after someone called Lockhart who once owned it. But Sir Herbert Maxwell in ‘The Place Names of Galloway’  compares it to Barlockhart and Drumlockhart in Wigtownshire and suggest they all contain the Scottish Gaelic word ’lucairt’ which has various meanings including ’castle’. Drumlockhart overlooks Lochnaw castle in the Rhinns of Galloway and Barlockhart a medieval motte in Glenluce village.

‘Lucairt’ has its origins in the Irish ‘longphort, a name first given to Viking camps but later meaning ‘palace’ or other important building. There are no palaces or castles near Lockhart Hill, but it is near Glenlochar, another possible ’luchairt’. Here the origin may be the large Roman fort at Glenlochar- the remains of which were still substantial enough to be mistaken for a ruined abbey in the seventeenth century-  or it could refer to an actual Viking longphort in the vicinity.

In a previous post I have looked at this last possibility. A large part of the evidence supporting the Viking longphort hypothesis comes from Scandinavian place names plus two Viking graves in the surrounding area. Unfortunately, I have discovered a possible complication. There is a possibility that some Old Norse speakers from Cumberland may have been planted in Galloway in the twelfth century. These not so Old Norse speakers could then have given Scandinavian names to their new farms…

This possibility comes from connecting the following sources.

First source -Richard Oram’ The Lordship of Galloway’ (2000) from pages 67 and  194.

Probably in the late 1140s, Uhtred [son of Fergus of Galloway] was married to Gunnilda, daughter of Waltheof of Allerdale…Through this marriage, Uhtred became lord a small estate at Torpenhow in west Cumberland…Gunnilda of Allerdale brought Uhtred  in contact with a changing world across the Solway…In return it was this new Cumberland and Westmorland  society that provided Uhtred and his son Roland  with the body of colonists they introduced into their territories.

These ‘colonists’ would have been the builders of most of the 31 mottes in the central Stewartry discussed by Christopher Tabraham in ‘Norman Settlement in Galloway’ ( In ‘Studies in Scottish Antiquity’ edited by David Breeze, 1984)

The second source is David Parsons article ‘ On the Origins of ‘Hiberno-Norse Inversion Compounds’’ in the Journal of Scottish Names Studies Vol 5 (2011)

Amongst the earliest witnesses to the vernacular language [of northwest England] there a small number of 12th century inscriptions which suggest that Old Norse was alive at that date.  Although it is not impossible that the language here was a recent reintroduction from the Irish Sea area, it also seems possible that Norse might have remained  the locally dominant vernacular in parts of the region between the 10th and 12th centuries… [page 126]

Parsons then suggests [page 139] that if some of the Gall-Ghaidheil (Gaelic speaking Vikings who gave their name to Galloway) had pushed further southwards across the Solway they may have (re)learned Norse in the English north-west…

It is therefore possible that when Fergus of Galloway’s son Uhtred married his Cumbrian wife, this was consolidating a connection between Galloway and Cumberland that already existed. As a consequence, the motte-building colonists Uhtred and Roland (possibly even Fergus himself) introduced to Galloway may have been distant kin of Galloway’s Gall-Ghaideil rather than ‘Normans’.

This may seem very unlikely, but there is evidence which links some of Christopher Tabraham’s  ‘Norman’ mottes with  Scandinavian place names.

Gribdae farm, map reference NX 73 50, was first recorded in 1365 as ‘Gretby’ and the -by (byr) means it is included in all the various lists of Scandinavian place names in Galloway. The record of the place name in 1365 comes from ‘The Register of the Great Seal of Scotland Volume I, 1306-1425’ entry 206 on pages 63-4. The text is in Latin so I can only give the jist of it, but it says that Dugal McDowall grants Littlilgretby in his lands of Kyrassalda to John Trumpour (trumpeter?), Herald of Carrick.

Kyrssalda is now Kirkarswell in Rerrick parish. The mention of Kyrssalda in the 1365 text is the first record of the name. Both Maxwell (1930) and Daphne Brooke  (‘Northumbrian Settlements in Galloway and Carrick’, PSAS 1991) give the place name as ‘Kirk Oswald’- a chapel dedicated to the Northumbrian king and martyr St Oswald. However Andrew Symson writing in 1682 called Kirkcarswell, Kirkcastel.

However, as Brooke notes (page 305) the lands of Kirkcarswell  were the only part of Rerrick parish which did not belong to Dundrennan Abbey, founded by Fergus of Galloway in 1142.

Tabraham (‘Norman Settlements‘, 1986, page 112) includes Kirkcarswell motte - map ref NX 754 487 -  as a ‘definite’ motte plus bailey. Tabraham also notes that Kirkcarswell was not part of the Dundrennan lands and implies that  this means there was already in 1142 an estate with the Kirkcarswell motte as its ’caput’ or centre.

East Kirkcarswell motte

Unfortunately, Tabraham does not develop this possibility any further. Oram (2000, pages 224-5) mentions Kirkcarswell as an example of a ‘classic’ motte but goes on to date building of the Galloway mottes to 1160 and later, as the work of Uhtred and Roland/Lochlann. Tabraham also mentions a possible motte site at Bombie  map ref NX 715 505.  Bombie was Bundeby in 1296 so another Scandinavian place name. Tabraham describes the site at Bombie as probably a ‘ringwork’.

At Boreland of Anowth, map ref NX 548 550, there is a Cumbrian connection to what Tabraham describes as ’this fine motte-and-bailey’. Circa 1170 the church at Anwoth ‘was in the possession of David fitz Terri lord of Over Denton in Cumberland’, who is assumed to have been the builder of Boreland motte. However. Jack Scott [‘The Partition of a Kingdom: Strathclyde 1092-1153’, Transactions of the  Dumfriesshire and Galloway Natural History and Antiquarian Society, 1997, page 25] argues that the Anwoth motte may have been built 50 or so years earlier, before Fergus of Galloway came to power. [Oram, page 56 disagrees]

Boreland of Anwoth motte and bailey

If the Kirkcarswell (Gribdae), Bombie and Anwoth earthworks were built in the very early twelfth century by ‘Normanised’ Cumbrians who settled Norse speaking tenants on their lands, this could be the origin of some of the Scandinavian place / farm names.

Alternatively, later plantation of lands gifted by Fergus. Uhtred and/ or Roland/Lochlann to Cumbrians who also had still Norse speaking followers could have had a similar effect.

Or the Scandinavian place/ farm names could have been  created by Vikings….

Monday, April 11, 2016

A Viking longphort in Galloway?

The Dee-Ken inland naviagation

When the Vikings arrived in Ireland, they set up temporary encampments which the Irish called ‘longphorts’. One of these was long-lived and evolved into the city of Dublin, but most were not so enduring. Although western Galloway became part of a Dublin Viking kingdom, no longphorts have been found in Galloway, unless Herbert Maxwell, ‘Place Names of Galloway’ is correct and Long Fort farm near Auchenmalg Bay in Old Luce parish was a longphort. It is close to Sinniness which Maxwell derives from Old Norse sunan nes -southern headland.

Glenlochar = Gleann + luchairt? 

Excitingly, as part of his ongoing research into the Gaelic place names of Galloway and Carrick, Michael Ansell may have found one.  He suggests that the Gaelic place name Glenlochar in eastern Galloway might not be 'gleann luachair'- the rushy glen but  'gleann' + 'luchairt'. Luchairt is a Scottish Gaelic word derived from 'longphort', a Viking encampment. If Michael is correct, then Glenlochar represents ‘glen of the Viking encampment’.

A Viking longphort at Glenlochar means that the Vikings were accessing  25 miles of inland navigation on the Dee/ Ken river system - potentially from Kirkcudbright, where a Viking grave has been found, to the Boatpool of Dalry.

The location at Glenlochar is immediately interesting. In 1947, an aerial photograph revealed that that the site of what had previously been regarded as the site of an ‘abbey’ at Glenlochar was actually a Roman fort on the east bank of the Galloway river Dee. Altogether it is now known that there were two, possibly three, Roman forts at Glenlochar and at least five marching camps.

A mile downstream from Glenlochar is Threave castle, constructed in the late fourteenth century on an island site previously occupied by the Lords of Galloway in the twelfth century.  Between Threave and Glenlochar is the site of a tower house used as a summer residence by the Gordons of Kenmure castle. Andrew Symson [‘ A Large Description of Galloway’, written 1682] noted that ‘the Vicecount of Kenmuir may easily transport himself and his furniture by boat’ from his castle to this residence, a distance of 12 miles. In the later eighteenth century, barges were worked from Threave island to the Boatpool of Dalry, a distance of 15 miles.

Glenlochar Roman forts, Old Greenlaw and Threave Castle 

Navigation downstream from Glenochar was blocked  by a ridge of rock  at Tongland near Kirkcudbright. However, since it is known that Vikings could portage their boats around such obstacles, the rocks at Tongland would not have been a major difficulty for them. That there were Vikings in Kirkcudbright is known from the presence of a Viking era grave found in 1888 in the town. There is also a cluster of Scandinavian settlement place names around Kirkcudbright. Another probable Viking burial was found in a cairn on Blackerne farm in the eighteenth century. Blackerne is 3 miles east of Glenlochar.

A Viking grave was found in 1756 on Blackerne farm.
Cockleathes, Halketleathes and Leathes contain the Scandinavian place name element 'hltha', barn. In 1557, Ernespie farm was recorded as 'Quesby'. -by is another Scandinavian place name element.

The most important element in the understanding of Galloway’s history is the land itself.  Soil quality and land utilisation maps show a large central area of poor quality rough grazing land. This stretches from the eastern shore of Loch Ryan in the west to Upper Nithsdale in the east. The Wigtownshire Rhinns are an area of good quality land. From Ballantrae along the Ayrshire coast lies another area of better quality land widening out until it joins with the better quality land of Upper Nithsdale in the Cumnock area. The area of better quality land then extends down towards Dumfries. Running north-west to south-east between Dumfries and Gatehouse of Fleet is broad strip of better quality land which also occupies the eastern half of the Wigtownshire Machars. North of Castle Douglas, an area of better quality land extends up the Dee/Ken valley to Dalry, but to the south underlying granite between Screel/ Bengairn and Criffel creates another area of rough grazing, apart from a small area of good quality land in Kirkbean and New Abbey parishes.

South-west Scotland land quality map

Unfortunately, the route of a hypothesised Roman road between Lower Nithsdale and Glenlochar has not been established. It probably followed the same ridge of higher ground taken by the mid-eighteenth century Old Military Road, but crossed the Urr in the vicinity of Old Bridge of Urr rather then the Haugh of Urr to reach Glenlochar. West of Glenlochar, a direct line to gatehouse of Fleet where there was a Roman fortlet is blocked by an area of higher ground rising from 400 feet up to 1200 feet at Bengray so a Roman route would probably have skirted this area.

On the course of the road were crannog sites at Lochrutton and Milton loch. The Glenlochar site would also have given the Romans control of the Dee/Ken valley. High status objects- the Balmaclellan mirror (found eight miles north of Glenlochar), Wheatcroft rein-ring (found 1 miles south of Glenlochar) the Torrs pony cap and Carlingwark cauldron (both found two miles south)- suggest the Dee/Ken valley was an important area and that the area around present day Castle Douglas was a centre of religious/political power.

However, by the sixth century, the Mote of Mark and Trysty’s Hill had become high status sites with Ardwall Island (visible from Trusty’s Hill) an important Christian religious site. But although Ardwall continued as a religious site until the eleventh century, both the Mote of Mark and Trusty’s Hill were destroyed in the later seventh century most likely by Northumbrian invaders.

The 'T' is Trusty's Hill and 'W' Whithorn.

While Whithorn became an important Northumbrian religious centre in Wigtownshire, no centre of Northumbrian power of similar high status replaced the Mote of Mark and Trusty’s Hill in the Stewartry.

In ‘The Northumbrian Settlements of Galloway and Carrick’ (PSAS, 1991), Daphne Brooke argued that there was a Northumbria ‘shire’ in the central Stewartry. This extended from the river Fleet in the west to the river Urr in the east and inland to Burned Island on Loch Ken, 6 miles north of Glenlochar. The Castle Douglas area (Kelton parish) would have been at the centre of this shire. The various estates and landholdings of this shire, having passed from native British to Northumbrian ownership then survived to become the ‘building blocks’ of the medieval Lordship of Galloway. Brooke used a list of lands forfeited to the Scottish Crown by the last Douglas lord of Galloway in 1455 to identify some of these key building blocks.

Unfortunately, Brooke misidentified Burned Island on Loch Ken as the ‘Arsbutil’ in the list of forfeited lands. However, also called Erysbutil and Irisbutil, Erthbutil and Arthbutil in the ‘Exchequer Rolls’ between 1456 and 1477, in 1566 the ’Exchequer Rolls’ list Yrisbutil  as ‘alias Orchardton’.  Orchardton is in Buittle parish and John Cairns built a distinctive round tower house there after being granted  the lands of ‘Arsbutil’ by James II in 1456.

Without Arsbutil in the Glenkens, Brooke’s argument for a Northumbrian shire in the Stewartry is weakened. It is also possible to question some of her other Northumbrian evidence, as I have done previously. [‘Gaelic in Galloway Part Two-Contraction’ TDGNHAS 2012 pp 65-6]

If the Northumbrians did not take any great interest in the Stewartry, there would have been little to stimulate its agricultural economy. Without the need to support high status settlements, subsistence farming would have sufficed to meet the needs of its inhabitants.  With no centres of wealth and population to match Whithorn, there would have been little in the district to attract Viking raiders. This would also have meant that there would have been little in the way of opposition to Viking settlers.

Brooke (1991) listed 28 Scandinavian settlement place names in the Stewartry of Kirkcubdright. Of these, eight are existing or former (Galtway) parish names. The rest, a form Hestan Island, are farm names. One of these, Sypland near Kirkcudbright, was first recorded in 1210. It is now divided into Little Sypland and Meikle Sypeland- which are two miles apart. The original farm was therefore very large. Almorness, a three mile long peninsula, was first recorded as one farm in 1376. It was still one farm in 1456 but had been divided into six farms by the eighteenth century.

Brooke Scandinavian Place Names 1991

Mapping the full extent of Viking settlement is difficult. As examples, Brooke includes Southwick in her list of Northumbrian place names and her list of Scandinavian places names. Herbert Maxwell (‘Place Names of Galloway’) identifies Senwick  on the Dee estuary as  Scandinavian but Brooke as Northumbrian. Maxwell gives the river Fleet as either Scandinavian or Anglo-Saxon (Northumbrian) while Brooke includes it in her Northumbrian list.

It may be significant that Brooke was carrying out the research that produced ‘Northumbrian Settlements in Galloway and Carrick’ while the 1984-1991 ‘Whithorn Dig’ was producing evidence for the importance of Northumbrian Whithorn. The place name evidence for the western Machars shows Scandinavian settlements -Eggerness, Sorbie, Bysbie, Arbrack, Kidsdale, Physgill, Appleby and Ravenstone- in an arc around the Northumbrian settlements of Whithorn, Oughton, Broughton and Pouton.

This evidence has led to the suggestion that Northumbrian power in the Machars was strong enough to employ Viking settlers as a protective shield around  Whithorn. [e.g. Brooke ‘Wild Men and Holy Place, 1994, p.65]

But in the Stewartry, there was no equivalent to Whithorn. If there was, however, an inland/upriver Viking longphort at Glenlochar- close to earlier (late Iron Age/ Roman) and later ( medieval lordship of Galloway) centres of power, then the balance of probability between Northumbrian or Scandinavian settlement place names shifts towards Scandinavian origins.

In the 1930s, the Dee/Ken river system (and also the river Doon in Ayrshire) was altered by the Galloway Hyrdro-electric Scheme. On the lower Dee a barrage was constructed at Glenlochar and a dam at Tongland. It is therefore necessary to rely on pre 1930s Ordnance Survey maps to survey the lower Dee. These older maps show that even before the barrage was constructed at Glenlochar, the Dee was much wider above Glenlochar than it was below. Between Tongland and Old Bridge of Dee, the river was much narrower, in places less than 100 feet wide.

The pre 1930s maps also show four fords between Old Bridge of Dee and Tongland as well as three ‘forts’ and two ‘moats’ close to the river. There is a definite motte at Kirkcormack and also a chapel site. Kirkcormack was a pre-Reformation parish, now part of Kelton parish. Although the motte at Kirkcormack is post-Viking, the farms of Milnthird and Netherthird which were in Kirkcormack parish are included in Brooke’s list of Scandinavian settlements.

Assuming that the Vikings used the river Dee to reach the Glenlochar longphort, the need for portage at Tongland and to navigate five miles of narrow and sometimes shallow river from Tongland to Old Bridge of Dee would have been a strategic weakness. This suggests that either the lands along the lower Dee were already settled by Vikings or that they were soon after the Glenlochar longphort was established.

That what became the parishes of Tongland and Balmaghie on the west bank of the lower Dee and Kirkcormack and Kelton on the east bank formed a connected unit of territory is suggested by evidence from the twelfth century.

The ‘Saints in Scottish Place Names ‘ database  (and other sources) links Kirkcormack to other parish and chapel sites on the lower Dee.

Between 1172 and 1174 King William grants to Holyrood Abbey the churches or chapels in Galloway (in Galweia) in which Iona Abbey has proprietary right (que ad ius abbatie de Hij Columchille pertinent), with all teinds and other benefits (beneficiis), viz the church of Kirchecormach [medieval parish, now in Kelton], and ‘ecclesiam Sancti Andree [= Balmaghie] and that of Balencros [Barncrosh, in Tongland] and that of Cheletun [Kelton]. RRS ii no. 141.

Significantly, although he suggests that the four churches were gifted to Iona Abbey by Fergus in the earlier twelfth century, Richard Oram [‘The Lordship of Galloway’, 2000, p. 10 ] notes that

Dedications to Columban saints, such as Colman and Buittle, Urr and Colomonell in Carrick, Aid mac Bricc at Kirkmabreck, Bride at Blaiket, Cormac at Kirkcormack and Cummene at Kirkcolm, display a striking correlation to the main zones of Scandinavian settlement…particularly to the zone around Kirkcudbright.

Two miles south-east of Kirkcormack there is Kirkbride farm.There is another Kirkbride in Anwoth parish and Kirklebride in Kirkpatrick Durham parish. Near Dalbettie, what is now the Little Kirgunzeon Lane was called the Pollchilbride- St Bride’s chapel stream- in the twelfth century. However, unlike Cormac and Aid mac Bricc, none of the chapels dedicated to Bride became parishes. On the other hand, they may indicated other areas of Viking settlement.

There is also an Arkland in Anwoth parish which is included in Brooke’s list of Scandinavian places names. In Kirkpatrick Durham there is another Arkland which is adjacent to Kirklebride. In the former Kirkcormack parish there is a third Arkland, now divided into Over, High and Low Arkland. In her 2003 PhD thesis E A Grant [ Scandinavian Place-Names in Northern Britain as Evidence for Language Contact and Interaction ]  following Maxwell, suggested that Areeming in Kirkpatrick Durham, which is adjacent to Kirklebride and Arkland is similar to the Wigtownshire Airiehemming. Grant suggested both contain the Old Norse personal name Hemingr + Gaelic airigh.

This raises the possibility that Aireland (originally in Gelston parish, now Kelton) and which shared grazing in common with Arkland (Kirkcormack) in the seventeenth century may also have a connection to same period of settlement as the block of Scandinavian named farms to the south and west.

Sketch map of possible Viking settlements by density of place names.

Although some Irish longphort had inland locations beside rivers, most were coastal. [See Clare Downham ]

The most obvious locations for longphorts in the Stewartry would have been around the tidal limits of the Cree, Fleet, Dee, Urr or Nith estuaries. Since a Viking grave was found at Kirkcudbright on the Dee, Kirkcudbright would seem the more obvious location for a longphort rather than  Glenlochar 12 miles inland.

One possibility for choosing an inland location was security. In 902, the Vikings were driven out of Dublin by the Irish. If there already was Viking settlement around Kirkcudbright, one of the leaders of the Dublin Viking community may have chosen Glenlochar as a safe haven to rebuild and regroup before attempting to retake Dublin- which was achieved in 917 by Sitric Cáech (Sigtryggr) and Ragnall ua Ímair (Røgnvaldr).

Since Ragnall/ Røgnvaldr may have ruled over an area of southern Scotland and/or the Isle of Man, the Glenlochar longphort may have been his location between 902 and 917. But what was he doing there during those years?

Possibly he was exploiting a local resource- timber. Seventeenth century records show that there were still woodlands at Cumston near Kirkcudbright, near New Galloway and above Dalry where Alexander Gordon of Earlston sold his woodlands for 23 000 merks (£640 sterling) in 1691. In the tenth century, the Dee/Ken river system would have been well wooded. A longphort at Glenlochar would have given its occupier access to this timber which could then have been used to build a handy fleet of Viking longships in preparation for a return to Dublin. (As an aside, in 1204, Alan of Galloway offered to provide King John of England with a fleet of 100 ships when John was at war with Phillip II of France. Also see Stockarton discussed below)

Alternatively, if the Glenlochar longphort was established earlier, it may have been used to control trade and communication routes through the Stewartry. Discovered in 1912 and dated to 875, a Viking era metalwork hoard from Talnotry, 17 miles north west of Glenlochar, implies that what was to become a medieval pilgrimage route between Edinburgh and Whithorn was already in existence.  This east-west route crossed a north-south route into Ayrshire at the head of Loch Ken. A group of tenth century cross-slabs near Carsphairn lie on this route, possibly as markers.

The Roman east-west route crossed the Dee at Glenlochar and another east-west route crossed marshlands which stretch from the Dee to Gelston on a ridge of high ground beside Carlingwark loch two miles to the south.

However, since the only high status location any of these routes gave access to at the time was Whithorn, control over them would have provided slim pickings for the Glenlochar Vikings.

On the other hand, the Talnotry hoard could be an indication that travel between the Firth of Clyde and the Solway Firth by land through the Southern Uplands was believed to be safer than by sea through the North Channel. A longphort at the head of Loch Ken, possibly near the site of Kenmure Castle would have facilitated control of both east-west and north-south routes.

Even if the north-south overland route was only of minor importance for the Vikings compared to the sea-lanes, it may have been one of the routes along which Gaelic spread into Galloway. The most obvious route for the spread of Gaelic into Galloway is the short hop across the North Channel but at the beginning of the tenth century the Wigtownshire coast was still being described as the ‘Saxon shore’ in Irish annals. [Both Alex Woolf ‘Pictland to Alba’ 2007 and Thomas Clancy ‘The Gall-Ghàidheil and Galloway‘, JSNS 2008 note this.]

Greater Galloway

It is possible that there was already a Viking presence at Whithorn by then since the last Northumbrian coins found at Whithorn date to around 880. But by the time Echmarcach mac Ragnaill, a former king of Viking Dublin died while on pilgrimage to Rome in 1064, the Saxon shore had acquired a new name. The territory Echmarcach mac Ragnaill ruled over at the time of his death was called ‘na Renna’- the Rhinns..

Rhinns is a Gaelic word, not a Scandinavia one, however. The most likely  explanation for this is that while leading Vikings in Ireland continued to speak Old Norse, they would also have spoken Gaelic and most of their  followers would have been Gaelic speakers. As a result the main group of Irish Vikings who settled in and took control of Wigtownshire during the tenth century would have been Gaelic speakers. The Viking/ Gaelic Kingdom of the Rhinns included the Whithorn and the Wigtownshire Machars, but it does not seem to have included the Stewartry or extended north through the Moors into Carrick and south Ayrshire.

Yet by the time Fergus emerges in the 1120s as  ruler of a kingdom which included the Stewartry west of the Urr, Wigtownshire and Carrick, the whole territory had become Gaelic speaking. Significantly, Fergus’ kingdom was called Galloway, not the Rhinns.

Galloway takes its name from the Gall-Ghàidheil, but confusingly the Gall-Ghàidheil are first mentioned in mid ninth century Irish records. The  Gall-Ghàidheil then disappear between 857 and 1034 when the death of Suibne mac Cinaeda, king of the Gall-Ghàidheil is recorded. What the record does not reveal is where Suibne ruled, but presumably not in the same territory as Echmarcach mac Ragnaill.

Thomas Clancy [2008] has suggested that Bute was in the territory of the Gall-Ghàidheil at the beginning of the tenth century. Two hundred years later, Scottish records use ‘Galloway’ to describe a territory which stretched from Renfrewshire down through Ayrshire and  into western Dumfriesshire. This was a much larger area than the territory Fergus controlled. Through the twelfth century this greater Galloway was gradually absorbed into the kingdom of Scotland.

In 1160, Scots King Malcolm IV,  conquered Fergus’ kingdom but as the Lords of Galloway, Fergus’ descendants continued to hold sawy over the region.. When he died in 1234, Fergus’ great-grandson Alan was described as ‘king of the Gall-Ghàidheil’ in the Annals of Ulster. When Alan’s great grandson Edward Balliol made a bid for the Scottish Crown in 1332, his strongest support came from Galloway where he was described as the region’s ‘special lord’.

After Edward Balliol’s death in 1365, Archibald ‘the Grim’ Douglas re-established the Lordship of Galloway. The Douglas Lordship of Galloway survived  until 1455, when all the Douglas lands were  forfeited to the Scottish Crown.  In 1456, the Abbot of Dundrennan Abbey produced  a list of all these lands. The survival in this list of lands with 750 year old Northumbrian names led Daphne Brooke to propose that these lands, including Arsbutil (discussed above), had been the building blocks of Fergus of Galloway’s kingdom.

To return to the Gall-Ghàidheil. If the early tenth century territory of the Gall-Ghàidheil lay around Bute  and was then extended south along the Clyde coast, the existence of a Dublin controlled Kingdom of the Rhinns would have been an obstacle to this expansion. But if the Kingdom of the Rhinns did not extend into Carrick to the north or the Stewartry and Nithsdale to east, these areas would have been available for settlement in the later tenth or early eleventh centuries.

However, if there had still been a strong Viking presence along the Dee/ Ken river system, as there still was in Wigtownshire, this would have been an obstacle to the Gall-Ghàidheil. Alternatively, if the main phase of activity associated with the Glenlochar longphort  was in the early tenth century, by the later tenth century and early eleventh century, the Vikings settlers remaining in the lower Dee valley may have formed a defensive alliance with the incoming Gall-Ghàidheil to prevent their territory becoming absorbed into the Kingdom of the Rhinns.

This possibility would fit with Richard Oram’s suggestion that the original nucleus of Fergus of Galloway’s kingdom was centred on Kirkcudbright and the lower Dee valley. Although there is no historical evidence to link Fergus with Loch Fergus above Kirkcudbright, the Palace Isle and Stable Isle on the now drained loch may have been his original caput. Christopher Tabraham described the site as of ‘undoubtedly major importance’ in his study ‘Norman Settlement in Galloway’ (1984). Loch Fergus is one mile from the Scandinavian Bombie and Meikle Sypland and only half a mile from the Viking grave site in Kirkcudbright.

Scandinavian place names orange, Gaelic blue, Scots red

Summary and Conclusion

Although not discussed above, the discovery of a major Viking Hoard in Galloway in 2014 must lead to a re-assessment of the importance of Galloway during the Viking era. But if there was a stronger Viking presence in Galloway than previously suspected, the ripples from the Galloway Viking Hoard will affect our understanding of the preceding Northumbrian period and the later origins of Fergus of Galloway’s kingdom.

If we take Michael Ansell’s Glenlochar/ Glenluchairt ‘longphort’ suggestion as pointing to a significant inland Viking settlement in the Stewartry, then the presence of  Scandinavian parish, farm and place names between the Fleet and the Urr, with a concentration in Rerrick, Kirkcudbright and Kelton parishes (which include the medieval parishes of Galtway, Kirkcormack and Gelston) becomes significant. What is difficult to see in this area is something similar to the pattern found in Wigtownshire, where an arc of Scandinavian farm and place names encircles a core of Northumbrian farm names around Whithorn.

If, as Daphne Brooke proposed,  Kelton was at the centre of  a Northumbrian shire, a centre of secular Northumbrian power similar to the religious centre of Northumbrian power at Whithorn, a similar pattern of place name evidence might be expected.

An alternative hypothesis is that after  the destruction of the Mote of Mark and Trusty’s Hill sites, the Northumbrians did not establish any equivalent centres of secular power  in the Stewartry but did take over existing chapel sites and/or created some of their own. However, none of the Northumbrian religious sites in the Stewartry were on the scale of Whithorn in Wigtownshire or Hoddom in Dumfriesshire.

Without the economic stimulus of a major religious or secular site in the Stewartry, the districts economy would have been based on small scale subsistence farming on patches of better quality land concentrated on drier land in the river valleys and interspersed by woodlands and wetlands.

In this scenario, it would have been Norse speaking Vikings and then the Gaelic speaking Gall-Ghàidheil rather than the Northumbrians who created the large farms and estates which - at least in the Stewartry- later formed the building blocks of what was to become Fergus’ kingdom and survived to be recorded by the Abbot of Dundrennan in 1456.

Quite how the Galloway Viking Hoard will fit into this picture remains to be seen. Much will depend on its dating and the duration of the settlement it is associated with.

Appendix One - Extended list of possible Scandinavian place names in Stewartry of Kirkcudbright. Based on Brooke [1991] plus additions marked *

 Almorness- peninsula. Was one medieval farm, now several farms in Buittle parish.
Arbigland- farm in Kirkbean parish. Very high quality farmland.
*Areeming- farm in  Kirkpatrick Durham parish, airigh +Hemingr
*Airieland- farm in Gelston now Kelton parish.
Arkland -farm in Anwoth parish.
*Arkland- farm in Kirkcormack now Kelton parish.
*Arkland- farm in Kirkpatrick Durham parish.
Bagbie- farm in Kirkmabreck parish.
*Balmaghie (Iona pre-1172)- Balmaghie parish.
Bareness- farm in Colvend and Southwick parish. May not be Scandinavian but from Gaelic easa -waterfall since adjacent to Southwick mill and not a headland.
*Barncrosh (Iona pre-1172 )- farm in Tongland parish.
*Blackerene (probable Viking burial)- farm in Crossmichael parish. Adjacent to Leathes in Buittle parish.
Bombie- farm in Kirkcudbright parish.
Borgue- parish, now includes medieval Kirkandrews and Senwick parishes.
Borness- farm in Borgue parish.
Cockleaths- former farm, now just a house in Buittle parish.
*Ernespie- farm in Crossmichael parish. Recorded in 1557 as ‘Quesby’. Ernespie is adjacent to Blackerne (Viking burial) and *Leathes (Lathys, 1330) in Buittle parish.
Fairgirth- farm in Colvend and Southwick parish.
*Finniness- farm in Balmaghie parish. No waterfall nearby, so less likely to be Gaelic easa. Land extends into broad section of river Dee so possible Scandinavian ‘nes‘.
Fleet- river. Boundary between Anwoth and Girthon parishes.
Gaitgill- farm in Borgue parish.
Galtway- farm in medieval Galtway parish, now part of Kirkcudbright parish.
Galtway- medieval parish, now part of Kirkcudbright parish.
Gelston- medieval parish, now part of Kelton parish.
Girstingwood- farm in Rerrick parish.
*Glenlochar- may be Gaelic Gleann Luchairt- longphort. Place name location on both side of river Dee in Balmaghie and Crossmichael parishes. On Crossmichael side, also site of Roman forts and marching camps.
Gribdae- farm in Kirkcudbright parish.
*Grobdale- two farms, one in Girthon parish, other in Balmaghie parish. Grobdale Lane (watercourse) is parish boundary.
Halketleathes- farm in Buitle parish.
Hestan Island- Rerrick parish. Maxwell suggests from Scandinavian  hestum ey- horse island. Accessible at low tide from White Horse (White Port) Bay on which is adjacent to Horse Isles and Horse Isle Bay which are all on Almorness peninsula.
Kelton (Iona pre-1172)- parish, now includes Kirkcormack and Gelston.
*Kirkcormack (Iona pre-1172) - medieval parish, now part of Kelton parish.
Kirkcudbright (Viking burial)- parish, now includes medieval Galtway and Dunrod parishes.
*Kirkbride- farm in Anwoth parish.
*Kirkbride- farm in Kirkcudbright parish.
*Kirklebride- farm in Kirkpatrick Durham parish, adjacent to Areeming and Arkland farms.
Kirkdale- medieval parish, now part of Kirkmabreck parish.
Kirkmabreck- parish, now includes Kirkdale parish.
*Leathes- farm in Buittle parish. Recorded as ‘Lathys’ in 1330, with Cockleathes and Halketleathes as later sub-divisions. Adjacent to Blakerne (Viking burial). From ‘hlatha’ (Maxwell, 1930)
Mabie- farm in Troqueer parish, now Mabie Forest.
Milnthird- farm in Kirkcormack, now Kelton parish.
*Netherlaw- farm in Rerrick parish, recorded as ‘Nethrelathe’ in 1306. (‘hlatha’)
Netherthird- farm in Kirkcormack now Kelton parish.
*Newlaw- farm in Rerrick parish, recorded as ‘Newelathe’ in 1306. (‘hlatha’
*Overlaw- farm in Rerrick parish, recorded as ‘Ourelathe’ in 1306.
Rerrick- parish. (‘hlatha’
Southwick- medieval parish, now part of Colvend and Southwick parish.
Southerness- now a village in Kirkbean parish.
Sypland- now Meikle and Little Sypland in Kirkcudbrigth parish. The two farms are two miles apart.
Tongland- parish.

Appendix Two-  Stewartry parishes with area in square miles, arable potential (based on New Statistical Account), number of Gaelic farm names, medieval mottes (from Tabraham, 1986) and possible Scandinavian place names.

The need to be aware of parishes with medieval (Norman style) mottes is important since Middle English/Early Scots speakers settled around the mottes in the twelfth century may have introduced place names which have Old English or Scandinavian roots, thus creating confusion. On the other hand the mottes-representing grants of land by the medieval Lords of Galloway to non-Gaelic speakers - also correlate with areas of land suitable for arable farming so would have been attractive to pre-Norman Northumbrians and Vikings as well. Parishes which have medieval mottes but no Scandinavian places names are highlighted #. These are Balmaclellan, Dalry, New Abbey, Parton, Troqueer, Twynholm and Urr.

Mottes from Tabraham 1986 plus nearby Gaelic farms and Scandinavian place names

 Anwoth- 16 square miles 50 % arable, 19 Gaelic farm names + medieval motte, possible Viking place name
#Balmaclellan- 37 square  miles, 19% arable, 38 Gaelic farm names + medieval motte
Balmaghie-34 square miles, 29% arable, 29 Gaelic farm names, no medieval motte, but has Threave castle, possible Viking place name
Borgue-20 square miles, 65% arable, 22 Gaelic farm names + 3 medieval mottes, possible Viking place names
Buittle- 19 square miles, 32% arable, 22 Gaelic farm names, no medieval motte but has Buittle castle, possible Viking place name
Carsphairn- 88 square miles, 2% arable, 33 Gaelic farm names, no medieval motte, 10th century cross slabs, possibly Viking route
Colvend and Southwick- 37 square miles, arable ?, 32 Gaelic farm names + medieval motte, possible Viking place name
Crossmichael- 15 square miles, 80% arable, 27 Gaelic farm names, no medieval motte. Parish belonged to Lincluden Abbey/church, fortified site at Old Greenlaw, possible Viking grave.
#Dalry- 52 square miles, 19% arable, 35 Gaelic farm names + 2 medieval mottes
Girthon- 24 square miles, 25% arable, 27 Gaelic farm names, + medieval motte, possible Viking place name
Irongray -23 square miles, 52% arable, 17 Gaelic farm names, no medieval motte.
Kells- 74 square miles, 3% arable, 43 Gaelic farm names, no medieval motte, unless under Kenmure Castle.
Kelton- 18 square miles, 28% arable, 10 Gaelic farm names + 2 medieval mottes (Gelston and Kirkcormack), possible Viking place names
Kirkbean-17 square miles, 44% arable, 5 Gaelic farm names, no medieval motte, possible Viking place name
Kirkcudbright- 23 square miles, 23% arable, 12 Gaelic farm names, no medieval motte, but fortified site at loch Fergus and had later castle (Castledykes), possible Viking place names plus grave
Kirkgunzeon- 12 square miles, 42% arable, 17 Gaelic farms, no medieval motte.
Kirkmabreck-36 square miles, 22% arable, 25 Gaelic farm names + medieval motte, possible Viking place name
Kirkpatrick Durham- 31 square miles, 42% arable, 26 Gaelic farm names, no medieval motte, possible Viking place name
Lochrutton- 10 square miles, 80% arable, 10 Gaelic farm names, no medieval motte but crannog site occupied in 12th century
Minnigaff- 137 square miles, 7% arable, 83 medieval farm names, no medieval motte, possible Viking route (Talnotry Hoard)
#New Abbey- 17 square miles, 35% arable, 16 Gaelic farm names + medieval motte, Abbey
#Parton- 27 square miles, 4 % arable, 23 Gaelic farm names + two medieval mottes
Rerrick- 32 square miles, 63% arable, 28 Gaelic farms + medieval motte, Abbey, possible Viking place names
Terregles- 8 square miles, 88% arable, 4 Gaelic farm names, no medieval motte
Tongland- 10 square miles, 20% arable, 11 Gaelic farm names, no medieval motte, Abbey, possible Viking place name
#Troqueer- 9 square miles, 89% arable, 13 Gaelic farms names  + medieval motte, Abbey
#Twynholm- 17 square miles, 65% arable, 10 Gaelic farm names + 2 medieval mottes
#Urr- 47 square miles, 83% arable, 14 Gaelic farm names + one massive motte, Motte of Urr