LEFT The three Eildon Hills in the central Scottish Borders provided the Romans with the obvious name Trimontium for their largest fort, which was situated to dominate the ancient geopolitical landscape. The local Selgovae tribe also chose the summit of one of these hills for their huge 40-acre oppidum.

New light on Trimontium: building a museum against the odds

The recent renovation of Trimontium Museum in Melrose has given archaeologists the opportunity to re-examine the enduring mystery of what happened at the nearby Roman fortress of the same name – as John Reid explains.

Start

The large and well-defended fort of Trimontium lies in the heart of the central Scottish Borders, and for a time it was the most northerly outpost of Rome’s far-flung empire. It was also the setting for some macabre events not yet fully understood, and has preserved an astonishing array of finds from this period. Recent re-evaluation of some of these artefacts, coupled with state- of-the-art museum interpretation, have helped us reframe the story of the fort’s tempestuous past. But first, some background.

LEFT The three Eildon Hills in the central Scottish Borders provided the Romans with the obvious name Trimontium for their largest fort, which was situated to dominate the ancient geopolitical landscape. The local Selgovae tribe also chose the summit of one of these hills for their huge 40-acre oppidum.
The three Eildon Hills in the central Scottish Borders provided the Romans with the obvious name Trimontium for their largest fort, which was situated to dominate the ancient geopolitical landscape. The local Selgovae tribe also chose the summit of one of these hills for their huge 40-acre oppidum.

It now seems likely that considerable numbers of Roman troops, led by the imperial governor Quintus Petillius Cerialis, made forays into what is present-day Scotland in the early AD 70s, some three decades after the Claudian invasion had first reached Britain. History is created by those who write it down, however, and so, thanks to the great historian Tacitus, the glory for first driving Rome’s interests into the far north went to Gnaeus Julius Agricola (Tacitus’s father-in-law), who was governor in AD 77-84. Tacitus recorded that Agricola established strongpoints along the new invasion route, and we can be fairly sure that by the time the famous battle of Mons Graupius was fought somewhere north of Aberdeen in AD 83, his grip on the Scottish Lowlands was already tight.

One of the key power centres, which would have required close military control, was the heartland of the Selgovae, a people whose oppidum lay on the conspicuous summit of one of the three Eildon Hills near modern-day Melrose. The tribe’s massive 40-acre hillfort, which even today displays vestiges of hundreds of roundhouses, lay at the heart of an extensive web of subordinate hilltop settlements, many linked by miles of earthworks, cultivation terraces, and dykes. It therefore made good strategic sense for Agricola to place his premier lowland fort at the foot of the Eildon Hills and to construct a signal station on the summit.

Located near the modern-day hamlet of Newstead, the fort was large enough to accommodate a garrison of well over 1,000, and we know from surviving inscriptions that it was, at one point, home to a sizeable detachment of the XXth Legion and to a cohort of cavalry from Provence. A fascinating side story to the personnel of the fort was the presence of one Lucius Maximius Gaetulicus, a centurion of the XXth, who set up an altar at Trimontium. He went on to dedicate another at Great Chesters on Hadrian’s Wall, and finally commissioned an inscription at the legionary fortress of Novae on the Danube (in what today is Bulgaria), on which he is recorded as having served for 57 years. This might sound like something of a record, but in fact he is only the third-longest-serving Roman soldier that we currently know about.

Above This aerial photograph shows the fort site during the drought of 2018 (see CA 343 for more images of long-buried features that were brought to light elsewhere during this dry spell); streets and buildings are picked out by the parch-marks.
= This aerial photograph shows the fort site during the drought of 2018 (see CA 343 for more images of long-buried features that were brought to light elsewhere during this dry spell); streets and buildings are picked out by the parch-marks.

As for the fort’s name, a wooden writing tablet discovered at Carlisle and a milestone from just outside Edinburgh attest that it was known to the Romans as Trimontium. The same name also appears on Ptolemy’s 2nd-century map of the Roman world. Allowing for remodelling and repairs, the fort seems to have seen extended occupation during two main periods of activity. The first ran from Agricola’s invasion until sometime before the building of Hadrian’s Wall; the second extended from the time of Antoninus Pius’s reconquest of Scotland in AD 140 until the fort’s final destruction, probably in the great rebellion recorded during Commodus’s reign in AD 184. Even then, the derelict fort site was used as a mustering point by the legions for later campaigns in the 3rd and probably 4th centuries. Notably, this included that of the Emperor Septimius Severus himself when he campaigned in Scotland (AD 209-210) with his son Caracalla against the northern tribes.

Mysterious metalwork

After this, the fort’s location was lost to the effects of decay, stone-robbing and agricultural degradation, and remained lost for nearly 2,000 years – only resurfacing when the Victorians pushed a railway cutting through its southern annexe in the mid-19th century. It was not until pioneering excavations by a local lawyer, James Curle, in 1905-1910, though, that the true treasures of Trimontium were uncovered – and the real mystery began. What sets this site apart from the scores of other Roman forts known in Britain is the mass of material, literally tons of it, that had been concealed in more than 100 pits and wells scattered throughout the fort. Armour, parade helmets, face masks, weapons, tools, and implements: a mass of iron- and bronze-work of almost every description.

Although there are other examples in Scotland of the Roman army concealing ‘surplus’ metalwork such as cart-fittings or nails – almost certainly to deny them to local asset-strippers – at Trimontium things went much further. In what appears to have been an extraordinary act of destruction, the pits and wells of the fort were packed full not only of metalwork, but with organic items. These grisly caches included cadavers – carcasses of cavalry horses, lots of them, as well as cattle and human remains, some complete, others less so. Several of the pits also contained decapitated human skulls, one of which, Curle noted, was accompanied by the birch pole on which it had been displayed.

Above Eildon Hill North hillfort, once home to Selgovae tribe. Taken from the east, this photograph shows the 2km- long rampart circuit and the ‘dimples’ of scooped roundhouse platforms that cover the summit.
Eildon Hill North hillfort, once home to Selgovae tribe. Taken from the east, this photograph shows the 2km- long rampart circuit and the ‘dimples’ of scooped roundhouse platforms that cover the summit. IMAGE: National Museums Scotland.

To Curle, the man on the spot, the fort’s demise had all the hallmarks of a disaster – but of what sort? A cataclysmic attack on the fort or a strategic withdrawal in the face of overwhelming enemy opposition? Whatever the exact cause, Curle deduced from the dating evidence provided by coins and pottery, that in fact two catastrophes had overtaken the fort. The first, and probably most violent, was early in the 2nd century, and the final desertion came probably half a century later. Within 18 months of completing his five-year dig, Curle had published his findings in a landmark book, A Roman Frontier Post and its People – an extraordinary feat of scholarship, even by today’s standards.

Since Curle’s publication, though, less traumatic interpretations of the cornucopia of finds at Trimontium have been conjectured, varying from ‘the Romans tidying up’ to some bizarre form of ‘ritual’. The reality, however, is that there are literally dozens of other abandoned Roman forts throughout Britain, but none with Trimontium’s sinister compendium of artefacts. This has been recognised by popular writers from Arthur Conan Doyle, creator of Sherlock Holmes, who wrote a short ghost story about Curle’s dig, to Rosemary Sutcliff who set a chapter of The Eagle of the Ninth within the ruins of the fort.

Despite the mammoth volume of information that Curle crammed into his book just before the outbreak of the First World War, over a century later there remain many unanswered questions about the fort itself, those who populated it, and how they interacted with the indigenous peoples of the area. Since Curle’s day, though, a number of smaller investigations have been carried out to look into these matters – most recently by Bradford University in the 1990s, whose work confirmed, among other things, the nature of the fort’s amphitheatre, which is the most northerly example yet known. A number of desk-based investigations have been carried out, too, including a review of coins from the site that has shed further light on its likely occupation periods.

BELOW A jasper gemstone from a Roman signet ring showing the bust of Caracalla, son of emperor Septimius Severus. Both father and son campaigned for two years against the northern tribes, and one of their camps lay very close to Trimontium’s perimeter. The stone (which may have been a personal gift from the emperor) was found by local historian Walter Elliot, when field walking at the site.
A jasper gemstone from a Roman signet ring showing the bust of Caracalla, son of emperor Septimius Severus. Both father and son campaigned for two years against the northern tribes, and one of their camps lay very close to Trimontium’s perimeter. The stone (which may have been a personal gift from the emperor) was found by local historian Walter Elliot, when field walking at the site.

A museum against the odds

It was around the same time, some 30 years ago, that the Trimontium Trust was formed by a committed band of history enthusiasts dedicated to improving public understanding of the Romans in Scotland. This group has helped curate and lead guided walks to the site, and, in addition to an impressive educational output, the Trust was able to found and operate a small independent museum in Melrose that displays some of the fabulous artefacts found at Newstead. As the museum reached its 25th birthday, however, it was beginning to show its age, and so the decision was taken to upgrade the building by creating additional display space and by adopting the latest interpretation methods to bring Trimontium’s dramatic story to the attention of a new generation of visitors.

ABOVE Just some of the diverse finds from the Trimontium pits, including a cavalry face mask, swords, shoes, and centurions’ medals (phalerae).
Just some of the diverse finds from the Trimontium pits, including a cavalry face mask, swords, shoes, and centurions’ medals (phalerae). IMAGE: National Museums Scotland.

Raising the required funds – almost £2 million – from a standing start was not going to be easy, however, particularly when set against the backdrop of the financial crash that had dented the confidence of many charities to give to small independents. Add to this the political/financial/supply chain turbulence associated with Brexit, and you can see why it took considerable conviction from the board of trustees to push the project forward in the current climate. Thankfully, a network of support from the National Lottery Heritage Fund, Scottish Borders Council, Scottish Government, Museums Galleries Scotland, and many others fell into place all at the right time. Simultaneously our museum partners – National Museum Scotland and Live Borders – stepped up to the plate with generous expert help and a fabulous wealth of previously unseen artefacts to swell the displays.

BELOW Many of the site’s pits and wells were packed with enigmatic finds, including quantities of human and animal bone. LEFT shows two of the many horse skulls from Pit 17; the animals appear to have been healthy and there is no evidence that they were butchered for food. RIGHT is a human skull found in one of the wells, displayed alongside stages of facial reconstruction one of the Trimontium Trust’s recent projects.
Many of the site’s pits and wells were packed with enigmatic finds, including quantities of human and animal bone. ABOVE shows two of the many horse skulls from Pit 17; the animals appear to have been healthy and there is no evidence that they were butchered for food. IMAGES: National Museums Scotland.
ABOVE is a human skull found in one of the wells, displayed alongside stages of facial reconstruction one of the Trimontium Trust’s recent projects.

Just as the ‘green button’ to start the physical process of renovation had been pressed in December 2019, there emerged a hideous new menace that threatened to completely derail all elements of the finely balanced project. COVID-19 rapidly became the big story, producing massive knock-on effects on builders, supplies, access to stores, lockdown of personnel, and 101 other stumbling blocks that the new and evolving restrictions quickly generated. Amazingly, thanks to the resilience and professionalism of all involved, particularly our stellar project team, we were able to push on. Masked to the eyeballs, everyone from carpenters and display-case manufacturers to mount-makers and museum curators played their part, and by the late summer of 2021 the museum was taking shape. Final installation was complete by August of that year, and although only the tail-end of the season remained, the museum had a ‘soft’ opening to great acclaim… only three months late, after a gestation of seven years!

ABOVE The oval scoop of Trimontium’s amphitheatre, which had a wooden superstructure. It lies just to the side of the fort’s parade ground.
The oval scoop of Trimontium’s amphitheatre, which had a wooden superstructure. It lies just to the side of the fort’s parade ground.

New technology and the future

Apart from the inherent friendliness of small museums, much of their charm comes from their ability to tell a concise story that is pitched at the right level for their audience, and to do all this within a relatively small space. Some of these aims can now be aided by modern technology, and in Trimontium’s case, where little remains to be seen above the surface of the arable fields at Newstead, 3D visualisation plays a significant role in telling the site’s story. At the heart of the display lies an exceptionally detailed virtual model, created by Simon Fleming, that acts as the core for flyarounds, walkthroughs, and cutaways. Added to this are copious other audiovisuals, which – together with cunning modern methods of mounting and display – all help to capture the visitor’s imagination.

Rob Longworth, curator of the Trimontium Museum, surveys the main museum room at the start of the renovation in early 2020. Although the group of eight piles found beneath the old floor look like hypocaust pillars, they were in fact supports for a Victorian billiard table that stood here over a hundred years ago.
The completed museum interior.

In a progressively cash-strapped economy, with less disposable wealth to go round, all museums face a difficult future, not least those small independents that have to raise most of their own finance. On the bright side, though, there appears to have been a dramatic upsurge of popular interest in, and support for, heritage projects, perhaps partly driven by the existential introspection caused by a pandemic of COVID-19’s magnitude. Whatever the reason, though, the museum-going general public seem to have embraced Britain’s past with gusto, which is a cheering thought as we look ahead.

BELOW Mount-maker Richard West arranges some fragile elements of the Ruberslaw Hoard of Roman bronze-ware within the new displays.
Mount-maker Richard West arranges some fragile elements of the Ruberslaw Hoard of Roman bronze-ware within the new displays.

Now that the museum has reopened, the Trust can turn its focus to further unwrapping the mysteries of what happened at Trimontium, with a planned raft of new initiatives to explore the archaeological evidence. Already under way are archaeometallurgy projects and cold-case laboratory-based post-excavation ventures including aDNA analysis, radiocarbon dating, and stable isotopic studies of the human remains, to learn more about these individuals and where they came from. These will be combined with renewed geophysical survey and aerial mapping of the Iron Age landscape, which will complement previous fieldwork – and with the fort’s stormy past, there is no danger of running out of questions or material. Finally, as part of the original vision for the museum, 2022 will see the development of its new Community Archaeology Centre, packed with the latest archaeological visualisation technology to help transport us back to when two cultures collided, leaving behind amazing traces of a story not yet fully written.

ABOVE This virtual model shows how the fort might have looked in the middle of the 2nd century AD. It was surrounded by an extensive fortified vicus or civilian settlement though recent geophysical survey has shown that there was little sign of occupation outside the defended annexes.
This virtual model shows how the fort might have looked in the middle of the 2nd century AD. It was surrounded by an extensive fortified vicus or civilian settlement though recent geophysical survey has shown that there was little sign of occupation outside the defended annexes.

There have already been exciting developments: ground-penetrating radar survey (by AOC Archaeology Group, using equipment loaned by Sygma) has just been completed over the central range of buildings, aiming to relocate trenches opened during the 1910 excavation, and to pin down the site of key structures such as the principia and bathhouse associated with the fort. Geophysical survey has also revealed the presence of what is thought to be a gyrus, a circular compound used for horse training. Facilities like these are rarely seen in Roman Britain – though a comparable example with a near-identical diameter has been found at the Lunt, near Coventry – and this discovery might suggest that horse breeding or supply was a key function of the fort at Trimontium.

left Recent magnetometry survey of the fort has revealed considerable internal detail these features are enhanced by the degree of in situ burning, traces of which can also be seen from the air (INSET).
Recent magnetometry survey of the fort has revealed considerable internal detail these features are enhanced by the degree of in situ burning, traces of which can also be seen from the air (BELOW).

Meanwhile, magnetometry survey has demonstrated very little evidence of occupation outside the heavily defended annexe compounds, suggesting that – unlike civil settlements in ‘safer’ parts of the province – we might understand Trimontium as more akin to the Roman equivalent of Camp Bastion than a friendly trading post. That this was not a ‘safe’ area of Britannia is also suggested by the results of recent drone surveillance, which has documented extensive traces of carbon-staining of the plough soil of the fort – signs of a pretty massive conflagration on the site. We had previously only done aerial surveillance when the crops were in place, obscuring these distinctive marks (which are visible in wet, newly ploughed soil) – but, purely by happy chance, they were initially spotted by Fraser Hunter, National Museums Scotland’s principal curator of Iron Age and Roman archaeology, as he glanced out of the window during a commercial flight to Germany that had just taken off from Edinburgh Airport. The images we have since recorded using the drone tie in with similar observations at other Lowland forts, such as Cappuck, Glenlochar, and Birrens, possibly suggesting that Trimontium’s destruction was part of a wider event.

LEFT Traces of what is thought to have been a gyrus, used for training horses; segments on the scale represent 10m.
Traces of what is thought to have been a gyrus, used for training horses; segments on the scale represent 10m. IMAGE: German Archaeological Institute.

This wider picture is reflected, too, in coins from the site, which have been independently reassessed by the numismatist Richard Brickstock. It appears that the first occupation of the fort was longer than generally accepted – unlike previous assessments, which placed Trimontium’s abandonment in AD 105, it now appears that its garrison managed to hold the fort until just before the building of Hadrian’s Wall (c.117-120). These new dates put the site in the firing line of a big rebellion early in Hadrian’s reign that recorded the losses of lots of Roman troops (including the famous IXth Legion). As fieldwork progresses, we hope that the story of the fort and its violent end will evolve further. Watch this space for further updates on our findings.

Further information
For the latest details about Trimontium, see www.trimontium.co.uk.