Pen-y-Gwryd

The Roman army of the 1st century A.D. had become a force capable of rapid penetration into enemy territory. Usually, a field army composed of legionaries and auxiliaries split up into separate campaign groups called ‘vexillations’ these individual groups then sought to paralyze their opponent by a concerted attack on their territory from several fronts. These vexillations could be of numerous sizes, from very small half legion sized units, roughly three or four thousand strong, to massive three or four legion camps which could be anything from 40-160 acres and with equal numbers of attendant auxilia, these field armies could total 40.000. As can be seen from the second Gulf War, an invasion army, regardless of its quality, and standards of technology in the field kit, are subject to sudden and potentialy fatal attack from quite small groups of hostile countrymen, usually while traveling from fixed points, but also when the army has encamped for the night. The Roman army was familiar with the concept of a secure night camp, and these campaign groups regardless of their size, constructed a temporary camp, surrounded by fossatum et vallum - a bank and ditch, to afford the unit some security overnight. The Romans had learned their military craft from the Greeks, Spartans, and Macedonians, while previous to this they had fought wars against similar Italy based City States, (Rome had started life simply as a city) such as the Etruscans and the Samnites, by the 1st century A.D. they reached their peak of military perfection, utilizing the best, and adapting the worst features of all these peoples. The Greeks had been the greatest inspiration, they not only introduced to Rome the ‘phallanx’ type of offensive unit, but the organization of Rome’s officer class closely followed Greek type, and also the army units such as centuries and cohorts seem to have been based on contemporary Greek army units. The Greeks are remembered as very brave and tenacious soldiers, often literally fighting to the last man, but for some unknown reason they neglected to construct temporary camps around their armies. The concept of secure night camps was not a new one, many ancient armies were familiar with the concept, so it is no surprise that Rome was to adopt it. Rome’s Republican armies had not been endowed with very much common sense, defeats in these years being common, Peter Connolly claims the army had been slow to come to terms with the idea of employing scouts to check the safety of the ground ahead, it seems they had enough common sense to develop the temporary camp into something of an art form.

The Roman Marching Camp

It should be stressed from the outset that no Roman temporary field camp is definitive proof of military action. Obviously most are actual overnight stopping points in actual wars, but they can equally be camp sites in a peace time situation. Small camps have been used by very small units employed in road building, fort construction, or even simply for later safety while scouring a region for mineral assets. There is an even smaller class of camps, at 50 x 50 metres on average, these are ‘practice camps’ their small size easily discounts them being marching camps.

So, while on field maneouvers, either in war or peace, any Roman unit secured its camp site with a fortification. These range in size from tiny 3.4 acre camps, Craven Arms, to massive 165 acre camps in Scotland. The defence took the form of a rampart and ditch, either could be omitted depending on the terrain. In very rocky areas the ditch was left out, equally this would also happen if there was no threat, while in sandy areas a simple palisade of stakes would have been used. In a hostile situation a ditch was dug, the Roman army had a regulated size range of ditches, ‘fossa legitima’ or regular ditch. The smallest regular ditch was five feet wide by three feet deep, the maximum allowed was 17 feet wide, with a possible depth of ten feet, its base usually took the profile of a slot. The up-cast from this ditch was used to build a bank, or rampart, and was usually about four feet high, the front was lined with neat courses of turves. The Roman soldiers are famous for carrying pointed stakes with their marching packs, doubtless many more were carried in carts, or possibly on animals, these were used to form a fence, either interlocked together, or simply rammed into the top of the bank. The camp had breaks in the ramparts to form gates and are usually in similar places to gates in permanent forts. Though there was no actual doors, the rampart ends were shaped to form outworks which would hinder an enemy charge, and possibly had been fitted with more examples of the palisade stakes. The camps could be any shape, the available ground usually dictating this, but the familiar ‘playing card’ shape was always preferred if possible, though some Flavian types could be square. The playing card shape is more correctly referred to as ‘tertiary’, that is the smaller sides are exactly half that of the longer sides. The tertiary shape could also be rhomboidal, and this was due either to mistakes in the laying out phase, or simply to go with the lie of the land. The Roman army surveyors, ‘mensores’ or ‘gromatici’ would arrive at the camp site well in advance of the main army unit, presumably backed up with cavalry support, and using the ‘groma’, plotted the shape of the camp, flags were inserted in the ground at the main points of the camp. When the advance of the unit arrived it was set to work right away digging the ditch, in very hostile situations work would not commence until the whole vexillation had arrived, and half would be on guard duty while the remainder constructed the bank and ditch. We know all this information from genuine Roman sources, Polybius, who wrote in the 2nd century BC about Rome’s war with Carthage goes into some details about marching camps in Spain. Hyginus Gromaticus wrote probably in the 2nd century A.D. also goes into some details, while Vegetius, writing in the late 4th century A.D. in his ‘Epitome of Military Science’ rather dolefully comments on marching camps. Dolefully, because he comments basically ‘when we used to build marching camps we used to win wars, if we build them again maybe we can start winning again’ while the fall of the Roman Empire is not under discussion here, it goes someway to indicated the importance of the genre

The camps had to be built with a series of constraints which had to be adhered to, basically the camp should not be overlooked, it should have a good supply of clean water, but these water courses must not run within the camp, the ground should be flat, with no marsh or minor hillocks, and at all camps the tactical superiority should be always sought. Right away it must be stressed that very few camps can satisfy all these criterior, and most will display at least one shortcoming as the surveyors had to do the best while working in unfavourable conditions. The general idea gained is that most camps display shrewd planning by the surveyors, and as will be seen below, at Pen-y-Gwryd.

The lowlands and the east coast of Scotland have the best preserved collection of marching camps, in what was the Roman Empire, and upwards of 200 are recorded as having existed in Scotland alone, The flimsy nature of the bank means that has long since been ploughed out, unless in marginal land, all that will survive is the cropmark of the buried ditch. The Scottish camps are of different types (chiefly their gates) and their names have gone on to refer to all similar camps throughout Britain. For example, the camp at Stracathro, with its rhomboidal shape and double clavicular gates, has given a name to all camps with this gate style, the camp was Flavian, incidentally. This refers to the outworks serving as barriers, some camps had a bank called a ‘titulum’ constructed out from, but parallel to the opening, while others had curved banks and ditches, both inside and outside of the rampart, these are called ‘clavicula’ or curves, and it is from were we get the name for the human collar bone.

How many troops did these camps hold? answer, we just don’t know, one of the largest Scottish camps, for instance Durno, at 144 acres, is likely to have held the famous Agricola’s entire field army and must have nearly held 40.000 in total, though the 165 acre camps are likely to date a century later to the campaigns of Severus and his son Caracalla. A camp the size of Blaen cwm Bach in South Wales is 62.5 acres and it is thought this could have housed two legions (15.000, approx) with equal amounts of auxilia. How big a camp did a single legion require? again open to conjecture, a legion would always be backed up by auxilia foot and horse, usually in similar numbers so a camp would have to be fairly large, it is thought 20 acres on average. Many Flavian camps in Scotland seem to indicate a single legion/auxilia, with possibly greater numbers of auxilia, Cleghorn - 46 acres, Oakwood - 40 acres, Pennymuir - 43 acres. Many camps in South Wales, also seem to indicate this, Pen-y-coed-Cae - 37.5 acres, Y Pigwn 1, and Coelbren being similar, while many Welsh camps are smaller, Y Pigwn 2, Walford, Bromfield, Ystradfelt all average 21-25 acres. These surely indicate a legionary vexillation (but not the entire legion) plus a generous allocation of auxilia. There is a size of camp averaging 14-16 acres both in Wales and Scotland, Scotland gives us camps at Ardoch and Dunblane both 14 acres, while in Wales we have Cwm-is-yr-Rhiw - 16.3 acres, Twyn-y-Briddallt - 16 acres, and Esgair Perfedd - 15.3 acres. These are likely to be accommodation for troops engaged in maintenance locally. The smaller camps such as Pen-y-Gwryd, Wales - 9.5 acres, and Pennymuir 2 - 9 acres are possibly likely to fit into this category. Finally tiny camps such as Woodhead, Scotland- 3.6 acres and Buckton, Wales -3.5 acres are not likely to have held more than twol cohorts 500-1000, engaged in peacetime practices. The progress of these camps makes interesting viewing, it is easy to trace the route of the invading army, often in twin prongs, eventually meeting up in one large camp at the conclusion of the campaign. The Ordnance Survey Map of Roman Britain is the best way to view the genre, the campaigns up the valleys of South Wales being particularly clear. North Wales does not display many camps, two can be viewed near Caer Gai fort at Bala, Llanfor, and one is also stood near Pen Llystyn fort near Brithdir, the most famous example in Gwynedd can be viewed at Pen-y-Gwryd near Capel Curig.

The Temporary Roman Camp at Pen-y-Gwryd

Map drawn and designed by Kanovium Project, and is after RCAHMW Inventory to Caernvonshire Vol 2, Pen-y-Gwryd

This marching camp, can be viewed today, though with difficulty, it is sited on a low coll beneath the mountains Glyder Fach, and Moel Berfedd, north and west respectively, the cliffs of Moel Berfedd beetling down on the camp in particular. The camp is today covered by many modern features, the Pen-y-Gwryd Hotel, famous for accommodating the soon to be climbers of Everest, the A4086, and A498 roads, and the majority of the east rampart has been covered by a marsh following the creation of a man made lake, Llyn Lockwood. Other areas of the camp are no longer present due to man made disturbance or simply the natural weathering away of the remains, which even when first constructed can have been no more than four feet high. The RCAHMW Inventory to Caernarvonshire Vol 2, which appears to largely be the only available hard copy source of information on the site records its rampart as at best being three feet high, with a shallow ditch, which in total occupied fourteen feet. The rhomboidal camp occupies 9.5 acres, with sides of 240x 200 yards, which makes it neither square or tertiary. Only one exit gate is today visible, seen on a crest, east of the Beddgelert bound A498, there would have been others but are now obliterated by the Llanberis bound A4086, the hotel, and the Llyn Lockwood marsh. This gate, which featured neither titulum or claviculae, which is not surprising viewing the ground, is 60 feet east of the midpoint, and was ten feet wide. The site is at 900ft above sea level, south, the ground falls away very severely to become the Nant Gwynant Pass, south of the camp is a ‘pill box’ which is a gun battery probably dating from World War Two. The small square gun emplacement clearly indicates the strategic nature of the spot even in the 20th century. The siting of the camp was of great strategic importance to the Romans, it sits today astride three roads, and at the junction of the passes leading to Beddgelert, Capel Curig, and Llanberis. It is not implied that these roads existed then, but there would have been some kind of tracks, so it would severely have limited communication, and in a way it was kind of a ‘glen blocker’ a title used to describe forts in Scotland that deny access to the valleys and foothills. What is there to see today? The majority of the south rampart is still visible, though it is only less than two feet in height.

This image is looking north, taken from the A498 road, the crest centre, is crossed by the marching camp rampart, right of this, is the remaining entrance. The rampart is visible as a low bank, visiblity helped by the low sun in this winter photograph, it is seen just above the middle of the crest and eventually ends at a fence, the natural ground here clearly having been cut away by the modern road. This is possibly the strongest remaining feature of the marching camp, and it has been noted in road repair work, that if this cutting is disturbed then the turf layers of the rampart can clearly be viewed.

The south east corner of the rampart and ditch, here fairly well preserved, looking south toward Gallt y Wenallt, an outlier of Snowdon, and the Gwynant Valley. North of this the rampart is submerged by the man made lake.

Looking north east from the south east corner, the corner of the marching camp stands on a slight bluff, the rampart desends the slope to run into the lake Llyn Lockwood, above which, is the A4086 to Capel Curig.

The same viewpoint as the last two images, but this time looking west, the A4086 bound for Llanberis clings precariously to the slope of Moel Berfedd. The south west corner of the camp is in the ground beneath the road and is bounded, middle by the water courses. The camp bank is beneath the reeds centre and the nature of the bluff is emphasised by the amount the ground actually drops and then rises toward the road.

A closer view of the ground just described, this image is taken from the gun emplacement looking north west, the image is divided by the Llanberis road, but the south west rampart is clearly seen at the middle of the image to eventually join with the road far right. The plan shows the rampart has vanished at this point, but it would seem not so, even on this image it is clear to see the modern road runs above the very slight and scrubby rampart. The corner is visible, but its turn east is covered here by the ground, but the ditch of the camp has been reused as a modern field drain, for a tributary of the Afon Cynnyd.

Militarily the site was of further importance in later years, sited on a small bluff about 100 metres outside the marching camp is a well preserved gun battery, of very stout build, it appears to be scanning the Gwynant Valley and the whole site is completely overlooked by the cliffs of Moel Berfedd. So what can you say about the siting of the camp? It looks an odd place to build a camp, but at 9.5 acres it is only a small camp, and its rhomboidal shape is more likely to represent shrewd siting rather than a mistake in laying out. The south west corner is clearly avoiding the small brook of the Cynnyd, as equally the north rampart and north west corner also seem to be avoiding the small stream that runs east past the Pen-y-Gwryd Hotel and into Llyn Lockwood. We cannot comment on the ground now occupied by the lake, but it would seem the surveyor(s) responsible for this had a very difficult task on their hands, and sited the camp as clear of waterways as they could. So while that criteria was satisfied, there was also clean water available just outside the camp, probably on three sides. The camp was tragically overlooked as the above image shows, not only from Moel Berfedd, but also from the east. It would not have been easy for the army to erect their tents in rows, the bluff from which some of these photos were taken from would have taken up maybe 20% of the available ground space, and the south west angle drops and then rises.

This discussion of the actual availability of ground inside the camp brings us on to how big was the unit(s) using the camp? As we have seen a legion, 5.500 approx and equal auxilia need a camp of 15-25 acres, Pen-y-Gwryd is 9.5 acres, and adding up the awful nature of some of its ground, then this camp can only comfortably house less than half a legion and very few auxilia. The ground in the main would not be suitable for horses, it would be difficult to accommodate many, especially in the southern area of the camp. So it would be pointless conjecturing the size of the unit, but surely little more than two or three cohorts of legionaries, 1500, plus maybe the same of auxilia? If the camp was situated in lower ground it would be tempting to consider it a labour camp for legionaries building roads etc but not in such a location. The small size of the camp tells us something of Rome’s attitude at this time, it is believed the camp is Flavian, therefore around 78 A.D, they were confident enough to send quite small vexillations into the heart of the mountain fastness of Snowdonia. Was the confidence flawed? as at this time a cavalry brigade was routed by a Celtic warband in Merionedd, this is reported to us by Tacitus, his son in law was the governor Agricola, and he was quick to hit back at North Wales and almost totally annihilate the tribe known as the Ordovices. So if this marching camp represents a part of the final Agricolan campaign of 78 A.D is unknown. The map of known marching camps in North Wales is rather blank in comparison to South Wales or Scotland, so in consequence we cannot be sure which way the Romans had been going, possibly east in the direction of Bryn y Gefeiliau fort and the Conwy Valley. Possibly it was simply the overnight camp of a vexillation heading for Bangor or Anglesey? possibly there already being some form of track down the Llanberis Pass, on through Nant Peris and on to the coast facing Anglesey? There is a local Celtic connection, nearby on the slopes of Crib Goch, several miles north west of Pen-y-Gwryd, a high status item of Celtic art was discovered in Victorian times. It was a handle/applique from a large ceremonial bronze cauldron, of enameled bronze in the shape of a cat’s head. The destruction of a Roman cavalry unit already mentioned hastened Agricola (who had only just arrived as governor in London) to attack North Wales, he ended by hitting Anglesey hard and rather pointlessly, as it is reported some of the inhabitants had already considered themselves beaten fifteen years earlier by Suetonius Paullinus. In days gone by it was believed North Wales had been conquered from the large fort at Chester (Deva), it seems likely Agricola came in by the back door, and more likely used Wroxetter as the base for his final push into North Wales once all the legionaries and auxiliaries had been mustered for this final push. It is still likely that Deva took a prominent part in this operation, possibly being used by the Roman fleet, and finally the separate units meet up on the coast near Bangor.

The Written Sources
RCAHMW Inventory for the Ancient Monuments of Caernarvonshire Vol 2. 1955

The Romans In Scotland, Gordon.S. Maxwell, Edinburgh, 1989

The Roman Frontier in Wales, M.G Jarrett ed’ Cardiff, 1969.

Vegetius, Epitome of Military Science, Liverpool, 1989

Greece and Rome at War, Peter Connolly, Macdonald, 1981

Kanovium Project aims to produce interesting material about little known Roman sites in North Wales, check out its parent site at Kanovium Project if you want to know a bit more detail about our fascinating Roman heritage in North Wales.