Roman Forts in Scotland

The Roman Empire's northern frontier

roman forts
The remaining earthworks at Ardoch fort in Perthshire | Image credit: © Neil Ritchie, editor

Scotland can boast of a great many Roman forts, though unfortunately few now have upstanding remains. Often all that can be seen is the profile of the flat fort platform. In some instances, however, such as Ardoch and Rough castle impressive elements of the earthwork defences remain and hint at Roman Scotland’s former glory.

Roman forts in Scotland, as elsewhere in the empire, are to be found in a wide variety of sizes. Each was built and modified where necessary to accommodate an appropriately sized garrison – Roman Scotland we must remember was primarily a military affair. The wide range of sizes included the massive fortress designed to accommodate a legion of some five and a half thousand men, the smaller forts designed to accommodate entire auxiliary units of a thousand or the more usual five hundred men – the workhorse of Roman Scotland, as well as smaller forts where manpower pressures required an auxiliary unit to be subdivided between several locations.

At the smaller end of the scale are the fortlets, small posts generally keeping watch and ward along Roman roads in exposed locations and river crossings as well as the diminutive signal or watch towers accommodating a squad of eight men operating out from their parent units. These towers could be sited individually on high ground close to a larger low lying fort to increase its long-range signalling capability or as one in a distinct series allowing speedy communications through the Roman army’s various tried and tested signalling devices.

Forts also provided the backbone of a frontier or “limes”, whether as a string of forts located at key strategic locations such as that Agricola had built between the Forth – Clyde isthmus in 80 AD or latterly in concert with an increasingly sophisticated range of other frontier devices as Sallustius Lucullus put in hand up to and beyond the Tay in the years following 84 AD and finally with the addition of a running barrier such as Hadrian`s Wall and the Antonine wall.

As ever, the Romans standardised, and while no two Roman forts are ever identical, each being a response to the site and garrison`s specific requirements, they are fairly atypical with key elements being the bedrock in their design. This often allows fairly accurate reconstructions of their form to be plotted after only fairly minor excavation. Usually, all that is required is to identify the location of certain key elements and the remaining parts can be filled in to give a tolerable picture of the fort’s layout. This was pioneered at Fendoch by Sir Ian Richmond in 1935.

The layout at Fendoch was for long seen as the classic picture of the completely excavated and fully understood Roman fort. The fact remains that Richmond did not excavate the whole fort but pioneered the art of successfully plotting the layout, especially of the internal buildings from more limited intrusive works on the remains. It is in this respect that the modern descendant of the minimal intrusion culture – the geotechnical ground-penetrating scan can offer a great deal in Scotland when resources are limited and excavation time consuming and expensive, as well as being unavoidably damaging to the remains themselves.

As many Roman fort sites in Scotland saw only one period of occupation these techniques can easily shed a great deal of information. Where a site was occupied for a long time or over several periods, each involving rebuilding such as Ardoch, Strageath and Newstead the picture is much more complicated and resort to the spade is unavoidable to understand the nature of the remains and the sequence of different forts on the same site.

The fort was set out from the very centre where the Principia would ultimately sit, the point being known as the “Groma” after the surveying tool which would be located at this point as the constituent elements of the fort were set out. Religious ceremony often attended this particular process, sacrifices may have been offered, augers consulted and votive offerings buried to propitiate the local spirits.

Most forts in Scotland had ramparts formed with turf blocks which enclosed the spoil cast up from the excavation of the defensive ditches. There were often two of these ditches but numbers vary. Sometimes blocks of cut clay were used in lieu of turf blocks where good turf was not readily available in sufficient quantity.

The base of the rampart could be seated in different manners, bearing either directly on the ground, or on a base of cobbles or a corduroy of branches. The outer face of the rampart was generally steeper than the gentler inner face. To prevent this encroaching within the fort the rampart slope was often cut back and retained to a low height by a low timber retaining wall. Timber logs, driven like piles also retained the earth at the openings in the defensive circuit left for the entrances, usually four though six can sometimes be encountered.

These gates were termed the Porta Praetoria, the Porta Decumana, the Porta Principalis Sinistra and the Porta Principalis Dextra. The latter two were the “main gates” and their orientation is usually used to refer to the direction in which the fort “faced”. As can be seen from the typical schematic the rectangular fort faced along its longer axis, though in Scotland particularly on smaller Antonine wall forts this trend was often reversed, no doubt on account of incorporating the works on the running barrier while refining features from the earlier Hadrians wall which may have been deemed in need of improvement.

“Ascensi” steps or ladders gave access to the top of the rampart from the fort interior. The rampart was crowned with a walkway, usually timber and a timber breastwork which was crenellated. Towers, usually three storeys in height were located at the curved corners of the fort and equally spaced around the perimeter. They were paired at the gates, a device designed to strengthen the weakest part of any defensive circuit.

roman forts

Stone defences, though rarer than earth and timber works were also used in Scotland at the legionary fortresses of Inchtuthil and Carpow as well as important forts such as Balmuildy and Castle Cary. The walls were faced in roughly worked local stone, the masons often giving the stone a light broached finish. These faces contained a rubble core, the whole wall being bound together with lime. The top of the wall was crenellated and the gate towers, when constructed in stone, could be extremely imposing structures. Towers could be found with a pitched roof, a practical consideration given the Scottish climate or a flat roof with a fighting wallhead.

Immediately inside the defensive circuit ran the Via Sagularis. This maintained a clear space back from the defences and also allowed access to utilitarian features located at the back of the rampart such as bread ovens, latrines and on occasion small stores and workshops. The soldiers barracks were normally the first buildings to be encountered in from the Via Sagularis. These barracks commonly took the form of a series of ten contubernium (eight men) dormitories making up a Century (eighty men), with the more spacious accommodation at the end nearest the rampart allotted to the Centurion. We can envisage some of the space in the latter suite of rooms performing the equivalent of the Century’s administrative office.

In Scotland, each contuberniums quarters was generally subdivided into a front and back room, probably delineating between a sleeping area and space for the storage of personal equipment and arms. Running along the front of the barracks was a covered veranda, allowing relief from the barracks which had limited ventilation and daylighting. Later barracks along Hadrian’s Wall saw the contuberniums dormitories giving way to terraced chalets for the soldiers and their families, civilians now being brought inside the defences as a result of the chaos on the late Hadrianic frontier with the northern tribes raiding with impunity.

The Contubernium and Century numbers given above are based on accepted current understanding. However, this is not entirely satisfactory and merits some further debate. Why we may ask ourselves would a “Century” – in Latin meaning one hundred be comprised of only eighty men? Though outwith current thinking it is nonetheless possible that the Century did indeed comprise one hundred men at full strength and that each contubernium comprised ten men. Accommodation which was fairly communal in barracks or campaign tent only required to accommodate eight at any given time as the remaining two will have been by rota on sentry duty.

The heart of the fort was the headquarters building – The Principia. The size of the Principia varied depending on the size and importance of each fort. Its features, however, were fairly consistent though and comprised a large and lofty hall, with a range of small offices behind, usually five. To the front of the hall was an enclosed courtyard, around the perimeter of which small rooms, possibly arms stores were commonly located. Fronting these and running around the courtyard was another covered veranda. A well was usually located within the courtyard and artefacts brought out of the well at Bar Hill on the Antonine Wall have included finely carved column capitals. This illustrates that the Principia was often a finely embellished building.

The garrison would have mustered for pay parades and inspections within the hall, often termed a “cross hall” as it sits longitudinally facing onto the courtyard. A raised pedestal was located at one end from which the commanding officer, a Cohort Prefect in most auxiliary forts could address his troops. Visible at the back of the cross-hall was the “Sacellum”, usually the central of the five offices of the back range. This was where the unit’s standards were kept, on display but under constant guard. The units pay and savings were kept here – on account of the guard – and sometimes these were set into small subterranean basements beneath the sacellum or the neighbouring offices. The remaining four rooms were most likely used for the unit’s headquarters administrative purposes, the Roman Army being famously bureaucratic.

Located next to the Principia and fronting the Via Principalis was the Forts Commanding Officers House – The Praetorium. These varied in size according to the size of the fort and the social status of its inhabitants, records from Vindolanda make it clear that families usually accompanied the commanding officer. They were grand luxurious dwellings based on the Mediterranean courtyard style of house. These face inwards to a small landscaped and unroofed open courtyard where privacy from the fort’s garrison could be achieved.

Also usually located on the Via Principalis was the forts granary – the Horrea. In some cases, there may have been more than one. The granary was intended to allow sufficient quantity of foodstuffs to be stockpiled to last the garrison a considerable time in the event of erratic crisply, a very real problem which seems to have failed quite spectacularly in Scotland following the Antonine occupation in the late 2nd century AD.

The atypical granary was set up off the ground on a grid of short pillars. This deterred pests and allowed air to circulate below the floor to prevent the food spoiling, an important consideration when grain is stored for any length of time. The remains of the lower external walls generally reveal large stone butts indicating that the external walls were reinforced at these points where loads from a fairly substantial roof were transferred from the rafters to the discontinuous wall below. This was necessarily discontinuous as timber louvres were probably incorporated in the wall to allow air to freely circulate within the granary. Various foodstuffs will have been kept within the granary. Grain will have been deposited in large bunkers while amphora holding various liquids will have been carefully stacked upright. Meatstuffs will have been hung.

Fronting the granary on the Via Principalis will have been a raised timber stage where goods would have been offloaded from carts at the correct level allowing them to be dragged or hauled inside the granary with the minimum of delay, greatly speeding up the process of delivering bulk supplies. The main pitched roof probably oversailed this loading dock. Protecting a forts supply of provisions would have been the highest priority of the garrison commander following the maintenance of secure defences, the goods arriving in carts would doubtless have been protected in transit under oiled canvas tarpaulins.

Where space allowed, storage buildings, workshops and stables were accommodated. The precise use to which the remains of some of these structures were put to originally is difficult to identify in many forts. Large versions of these utilitarian structures, however, as well as hospitals, were located in the huge legionary fortresses where they are readily identifiable on the basis of their recognisable architectural layout.

Lastly, bread ovens and latrines were built into the rear of the rampart. As a feature, it is commonly the ovens which the Romans would readily replace as required which can assist in identifying the separate phases of occupation of a multi-period fort site. The latrines would discharge their contents out into the fort ditch and were likely fed with water from drainage channels formed in the forts main streets. We should not imagine the forts ditch – or ditches – being anything other than unpleasant places!

Additional space could be achieved by attaching “annexes” to the main fort. These were often as large as the fort itself and were also defended by a ditch and rampart. Bathhouses were commonly located within the annexe. Speculation surrounds the function of the annexe as they rarely offer up the remains of any buildings within them. We can safely speculate however that the annexe provided a secure compound where small convoys in transit could be temporarily accommodated as well as allowing ancillary activities to be undertaken that the tightly planned fort did not have sufficient space for.

Forts which were occupied for periods of only short duration may have stood in isolation, though it is inconceivable that a fort will have had no ready access to the Roman road network, a natural harbourage or navigable waterway where supplies could be brought in regularly and reinforcements brought up in a time of need. Many forts saw only relatively short periods of use, but where forts located in prime strategic locations continued to thrive or were reoccupied in successive Roman occupations a civilian face sometimes appears. The “Vicus”, a settlement outside the fort, generally sprang up along the road leading to a fort. Vicus are not well known in Scotland though some certainly existed. A large example is known at Inveresk, an inscription from Carriden interestingly relates that the vicus there gained some self-governing rights while remains at Happrew near Peebles show that a vicus could spring up even in seemingly remote locations.

Article first published May 2008 (romanscotland.org.uk).

Further reading:

Andrew Tibbs, Beyond the Empire: A Guide to the Roman Remains in Scotland, (Robert Hale Ltd, 2019)