Bryn y Gefeiliau Roman Fort, Capel Curig, Snowdonia

© Kanovium Project

The above image is of the North Western corner of the Roman fort on the banks of the Llugwy river, which is now between the modern Snowdonia tourist villages of Betws y Coed and Capel Curig.  This fort really has no name, either ancient or modern.  It was excavated in 1920-2 by J.P Hall, and I think for convenience he named the fort ‘Caer Llugwy’ in English this would mean Fort Llugwy, such a name has long since believed to be erroneous, and the fort is named after the farm to the South, Bryn y Gefeiliau, it could equally be called Gorddiniog, after the house to the North, over the small river Llugwy.  Today the site is as it was during Roman times, sited in a broad bend of the river, J.P Hall comments on a flooding problem of his trenches, and this would likely to have been a concern during it’s occupation as the river is only less than a metre below the fort platform.  Today the site is not a good example of a Roman fort, ploughing, stone robbing, and even the actual soil from the fort even being removed because it was rich in nitrates due to the occupation and was used as fertilizer for fields in the last couple of centuries.  As you drive toward the site from Betws y Coed, past the Swallow Falls Hotel the walls flanking the road have been constructed from Roman stone from this site, Hall comments in the 1920’s that this was in living memory.  I have also heard it quoted that most of the houses between the fort and Betws have been built with robbed stone.  While this is sad to someone like me, who would wish the remains hereabouts were of a less robbed nature, you cannot blame people in the past for doing this, unaware of the importance (plus ancient sites would have been more prominent several centuries ago, therefore viewed with even more ‘familiar contempt’) it goes somewhat to explain the denudation seen at Caerhun fort, even so it seems in the last 200 years.

R.J.A Wilson, A Guide to the Roman Remains in Britain 2nd edition Constable 1980

fails to even give the fort a gazetteer listing, it being confined to his appendix section,  which indicates stone building remains present, therefore giving it some importance in view of the sparse remains of some of Roman Britain  Wilson describes the remains as follows, I quote.


Prominent mounds with some stonework exposed ; in annexe on West side of a fort (invisible).I feel the site is worth a visit, not just to see this building, but if just to experience a Roman fort site totally untouched by modern development (a barn stands in the North Eastern corner but hardly detracts from the robbed and ploughed splendour).  Firstly the stone remains in the ‘annex’ mentioned above stand a metre above the meadow, many rooms being visible, and are just seen left above, now within a tiny ash wood, they are unusual in North Wales and still surprise me.  At the start of my interest in Roman history I believed they must represent a post Medieval farmstead, but a check out in  M.G. Garrett’s Roman Frontier in Wales proved they were an extant Roman building.  It is presumed that the visibleness of this courtyard building, three ranges and a courtyard facing the river, clear to be seen, must be due to excavation by J.P. Hall, a picture in the Caer Llugwy site report shows in 1920-2 there was no trees, probably prior to this there would simply have been a hump, if very prominent, bump, in the meadow.  The tree growth does not appear planned and shows what perhaps these areas would look like without farming.

Read about this fort in Hall’s site report, a rather thin affair in comparison to Baillie Reynold’s Kanovium, and Wheeler’s Segontium, (the trenches were only of a trial nature) but interesting all the same, but the finds report is pauce to say the least.

J.P. Hall, Caer Llugwy : Excavavation of the Roman Fort between Capel Curig and Bettws y Coed, Manchester 1923.

Hall was followed by

W. Bezzant Lowe - The Heart of Northern Wales, Llanfairfechan 1927, but The Roman Frontier in Wales  2nd Edition 1969 fails to specify the results, and also by V.E. Nash Williams for the 1st edition of The Roman Frontier in Wales.  Some trial trenches had been also dug in the building during 1965 by John Ellis Jones, this showed that some floors had been slate slabs, some had been of concrete, some of cobbles and some simply of beaten clay.  Like the bath building at Caerhun this building has clearly had some prodding during the 20th century, ‘stone remains attracting men with spades’. 

During 2002, Gwynedd Archaeological Trust completed a Geophysics survey of not just the Vicus, but of the entire area of the fort rampart, I was lucky enough to view the initial lap top computer print outs and while I am not at liberty to quote their findings until they themselves publish the work during 2003-4, I can at least say that while Hall was right about the strange positioning of the South gate, he was wrong about the fort and it’s relation to the annexe, which as such never existed, so it seems rather pointless to relate rampart size etc.  But until G.A.T publish their findings, I will relate what M.G. Jarrett, or as it happens John Ellis Jones, he being the actual contributor to this section of the 1969 book say on the knowledge available on this fort at the end of the 60’s.

A low lying site, encircled by low hills’, in fact ‘overlooked’ would be a better description, the comfort of the valley bottom sacrificing the security factor.  In my opinion the site is not a comfortable one, the view west to Moel Siabod is suberb, and must have been enjoyed by the Roman garrison, but the site is very cold and damp being next to the Llugwy river, during winter the sun never reaches the site, and a frost remains on the ground all day.

‘The station comprised a small squarish fort and an annexe on its West side’ the fort is neither square or smallish as it happens, but certainly has the round angled corners as described, and they were correct in stating that the fort’s ploughed out ramparts are just about visible.  The East rampart is still fairly noticeable, as is the gate midway along its course, other areas of rampart, in particular the West, are more spread, but look to be of a fairly substantial nature.  The ‘square’ was found to be 131 by 120 metres wide, which enclosed an area inside the defenses of 3.9 acres or 1.57 hectares, which and an area of ‘annexe’ 131 by 91 metres or 430 by 300 feet.  This was defended by two ditches (one with a midrib like Kanovium?) tested on the North and East sides and measured 4.25m wide and 2 metres deep and 3.3 wide and 1.3 metres deep, which giving a total of 30 feet of ditch, roughly is very similar to Kanovium, as is the 9-10 feet depth of the inner ditch.  Hall’s initial verdict of the rampart was that it comprised two parallel walls with a rubble cavity (indeed unusual) the 1969 book describes it as wrong.  It in fact goes to describe an earth and rubble rampart resting on a bed of cobbles, which like Kanovium was of differing width at different points, it fails to explain the presence of a definite stone wall, which obviously represents a cutting back of the turf bank and the laying of a stone wall up against it, just as at Kanovium, possibly around A.D 140.

The fort was believed to have been founded in A.D. 90, during the reign of Emperor Domitian, G.A.T’s Geophysics has proved that terminology used on the fort, through most of the 20th century, ‘small’, ‘minor’, ‘unimportant’ etc is clearly wrong, it was a normal Roman fort.  Also called a ‘minor mining station’ by many, it  is also inappropriate as all Roman forts were mining stations.  Much evidence of metal smelting has been located by the various small scale digs on the site, but most forts are similar in this respect.  Bryn y Gefeiliau actually translates (difficultly) as ‘the hill of smithies’ it may or may not be an actual folk memory of the Romans polluting the mountain air, rather like goblins in some ‘sword and sorcery fable’, but mines are to be seen today to the north west of the Pont ty Hyll (the ugly house bridge) so mining is likely have continued at the area long after the Roman withdrawl.

The fort was believed to have only had a serious occupation from A.D 90 - A.D 120, the levels being exceedingly thin, a couple of inches of clay representing most Roman floors.  Why was it not laid down during A.D. 78 as with Kanovium? possibly the late date represents the pulling back of the Scotland garrison and a reconsolidation of the North Wales frontier, which by A.D 90 (discounting the mining station theory) must still not have been stable.  The pottery evidence takes the occupation down to the end of the 2nd century, there was no coin evidence, no coins had been found, any local people got one? contact me, finds where sparse, though a complete Spanish amphora was found in the stone building, which has been interpreted as a ‘Mansio’, an official posting station complete with a bath-house which may take the occupation of the site into the early 3rd century.  Not a really interesting site, but worth half an hour, permission from farm of same name, pick a day of snow or low sun to get any idea of the forts dimensions.

The aforementioned large scale stripping of the site has ensured little is seen today, but why didn’t they rob the stone building?  Richard Fenton in his 1804-13 Tours of Wales records here was to be seen  ‘tiled floors upon pillars’ the slate slabbed hypocaust remains clearly illustrating the bath section of the stone building still being extant in the early 19th century.  None of this remains today, and it had clearly vanished by the 1920’s only just over 100 years later.  A strange thing was that when I first visited the site in 1993 it was still possible to view scattered slate roofing tiles, each with a single round punched hole for the nail, around the building remains, this was what led me to believe the remains represented a post medieval farm house, such was the upstanding condition of the remains.  According to Wirral Historian, Eric Foulkes if these roof tiles are indeed 2nd or 3rd century Roman examples then they are without a doubt the oldest recorded Roman slate roof tiles in the world.  While the slate floor slabbing is undoubtably of this date, we have to look to examples found at Segontium Roman fort in a 4th century context.  This use of Gwynedd slate (and Denbighshire slate at Kanovium Roman site) is the first recorded use of an industry that would from early Victorian times roof the world.  I have heard it postulated that site should be completely re - excavated and that the stone remains should be cleared and consolidated.  Local authorities agreed if half the funds could be raised by public donation, this never happened which is a shame because the building would probably be better preserved than the buildings on display at Segontium, and may even present a picture to be seen of buildings at forts such as Housesteads and Vindolanda on Hadrian’s Wall.  View the fort on a tour including Caerhun, the marching camp at Pen y Gwryd, the fort at Segontium, and  if you really are addicted to this period, the Roman crossing point into Mona of Suetonius Paulinus (Tacitus and his story of black cloaked furies and Druids scaring the Roman force of two legions and auxiliaries almost into apoplexy) which is now the area beyond the brake lining factory East of Caernarvon, though such a large force would have been spread over several miles of Menia Strait.

A Selection of Finds from the Fort at Bangor Museum
 Samian Ware

© Kanovium Project

Glass Melon Beads

© Kanovium Project

An Unguent Pot

© Kanovium Project

Samian Dr37 Bowl, Domitianic, Recovered 1993, Farming Activity

© Kanovium Project, Private Collection