Roman writing on the wall

22 mins read

Recording inscriptions at a Hadrian’s Wall quarry

In the depths of a Cumbrian wood, intrepid archaeologists have been abseiling down the wall of a Roman quarry to document eroding inscriptions left by 3rd-century soldiers tasked with harvesting the sandstone to help repair Hadrian’s Wall. Carly Hilts visited the site and spoke to Jon Allison and Mike Collins to find out more.

Researchers using ropes to access eroding Roman inscriptions on an ancient quarry wall in Cumbria
Intrepid archaeology: researchers led by Jon Allison have been using ropes to access eroding Roman inscriptions on an ancient quarry wall in Cumbria. [Image: Jon Allison, courtesy of Historic England]

Standing on the bank of the River Gelt, surrounded on all sides by dense trees and the sound of birdsong and rushing water, it is all too easy to forget the modern world. But while today the patch of Cumbrian woodland that we had come to visit feels evocatively isolated, 1,800 years ago the rust-red rockface before us would have been a hive of activity, bustling with workmen and ringing with the rhythmic sound of their picks, hammers, and chisels.

This once-busy quarry is located just 5.5km from Hadrian’s Wall, near the small market town of Brampton. The river-cut sandstone gorge in which it lies was a vital source of stone to help fortify the Roman frontier, and this site was just one of several works – known as officinae – operating in the area. But this quarry, known since the 16th century as the ‘Written Rock of Gelt’, is special.

Like many major industries in the Roman Empire (see CA 336), Cumbrian stone extraction was administered by the army, and at a handful of these quarries, inscriptions survive to record the names and affiliations of some of the men who laboured there. The most significant collection of such carvings is found on the Written Rock of Gelt – not only do they identify some of the legions that were assigned to the quarry and the men who oversaw their efforts, but the inscriptions also include rare dating evidence for these activities, and a number of sculptured drawings.

William Camden's illustration of the Written Rock of Gelt
The Written Rock of Gelt was visited by antiquarians including William Camden and John Collingwood Bruce. Both published images of the inscriptions that were visible at the time, though some artistic licence was taken – the face featured in Camden’s illustration is far more detailed than the actual carving. [Image: National Central Library of Florence]

The site’s significance was reported as early as 1599, when the antiquarians William Camden and Julius Cotton visited the site and Camden subsequently included his observations about the inscriptions in his 1607 edition of Britannia. Intense academic interest in the 18th and 19th centuries followed (including from John Collingwood Bruce, who featured the site in his 1851 book, The Roman Wall – he described the Rock of Gelt as ‘the most remarkable of this class of Antiquities,’ adding that ‘the beauty of the surrounding scenery will give additional zest to the ramble’), greatly enhancing this picture; by examining the drawings that these visitors produced, we can trace the evolution of their understanding. Yet these images also highlight how fragile the site is. Exposed as it is to the elements, the soft sandstone of the rockface is being constantly eroded by the weather and dripping water, with clear consequences for the archaeology: some of the inscriptions that stand out clearly in antiquarian sketches are today difficult to make out, while one has vanished altogether.

John Collingwood Bruce's inscription of the Written Rock of Gelt
John Collingwood Bruce’s inscription of the Written Rock of Gelt [Image: Getty Research Institute]

In order to preserve the content of these eroding inscriptions – and the invaluable insights into the area’s Roman past that they provide – Newcastle University archaeologists led by Jon Allison have been working with cutting-edge technology to document the surviving inscriptions in a project funded by Historic England. The team is using structure-from-motion (SFM) photogrammetry to create a 3D record of the inscriptions that will endure long after the original carvings have worn away – and which will be freely available online at when work is complete.

Reaching the inscriptions themselves was no small task, though – the carved portion of the rockface stands 10m above the riverbank, and a footpath that used to allow closer access collapsed in the 1980s. Instead, Jon and his team descended on ropes from the top of the quarry to get close enough to the markings to examine them in detail. Thanks to these adventurous efforts the researchers have documented the known inscriptions in unprecedented detail – revealing that antiquarian interpretations of some of them were not particularly accurate – as well as rediscovering a number of carvings that had long been thought lost to erosion, and even identifying several inscriptions and shallow relief sculptures that had never been recorded before.


Compared to the formal, carefully incised inscriptions seen on many Roman public buildings and monuments, the early 3rd-century writings on the Rock of Gelt are decidedly more irregular; pecked roughly into the rockface using stoneworkers’ tools, some of their letters are also formed in unusual ways. Yet their contents, which we will explore here, are intriguing. Those that were known before the current project began are recorded in Roman Inscriptions of Britain, the key body of information on Romano-British epigraphy (see the ‘Further reading’ box at the end of this article), and in the following discussion we will refer to them using their RIB number. What, then, can we learn from these fragmentary texts?

As most of the Rock of Gelt inscriptions survive only in part, their meanings are somewhat obscure, but there are a number that appear to represent male names – probably individual soldiers associated with the quarry. Two of these were earlier discoveries long thought lost, but the recent work has located them once more, not eroded but hidden beneath overgrown vegetation. One (RIB 1013), a deeply incised series of letters first recorded in the late 18th century but invisible since 1944, spells out the letters IULIN, possibly part of ‘Julianus’, while another (RIB 1015), initially noted in the 1940s, more clearly reads EPPIUS M.

Other markings are more enigmatic, such as the pair of numerals, IX X (RIB 1007), that were also rediscovered beneath greenery during the present research project. Isolated numbers frequently appear on building stones where they are sometimes interpreted as construction guides, Jon notes, but they are rarely seen in a quarry context. He wonders if they might be batch numbers, marking out areas of stone destined for specific destinations or drop-off points.

Close-by, a jumble of superimposed numbers are harder to interpret, while another clutch of letters and numbers (RIB 1011) are particularly intriguing for their unusual appearance. Initially discovered in the 19th century, today this inscription is heavily eroded but a possible text of TI . IIV III has been reconstructed by the team (a slight variant on the reading given in Roman Inscriptions of Britain, where earlier observers had interpreted the first letter as an S – something that closer inspection of the rock face has revised, Jon said). TI is a familiar abbreviation in Roman epigraphy, where it stands for the name ‘Tiberius’, but what does IIV signify? Jon wonders if it might be an incorrect spelling of III (following the same logic that IV represents 4), but the correct rendering of this numeral immediately afterwards would make this a curious error.


The above markings are intriguing but, for now, largely difficult to interpret. There are, though, inscriptions that are much more intact, and which make for very interesting reading. One of these is known as RIB 1014. It is badly eroded and some letters have disappeared even between an earlier inspection by Jon in 2018 and the beginning of the current project this year, but its text appears to read: IUL. PECULIARIS VIIXILATIO LEG XX VV. The ‘II’ of ‘VIIXILATIO’ is a vulgar form of the letter E, Jon says, revealing this to be a vexillatio, or detachment, of a legion commanded by a man called Julianus Peculiaris. Specifically, these are troops from the 20th Legion Valeria Victrix – a legion known to have participated in the construction of Hadrian’s Wall (a century and a half earlier, its members were also key to putting down the Boudican rebellion of AD 60-61), making it doubly exciting to find signs of their presence in a quarry associated with maintaining the monument. A new line of text associated with this inscription has also been discovered during the current work; it contains the word DEO (‘god’), but its meaning is as yet unclear.

A modern group of visitors crossing the River Gelt to see the inscriptions
A rather more modern group of visitors braved the River Gelt crossing to see the inscriptions for themselves in March. CA tagged along. [Image: C Hilts]

These were not the only soldiers who worked at the Gelt Woods quarry during the 3rd century, however; RIB 1008, which was first noted by Camden in 1599, bears witness to the labours of a detachment of the 2nd Legion Augusta: VEX LIIG [note again the II for ‘E’ in this abbreviation of legio] AVG OF APR[…] SUB AGRICOLA OPTIONE. The name of the legion’s centurion responsible for the officina (‘working face’, abbreviated to OF in the inscription) is mostly lost, though it appears to have begun with ‘Apr’ – however, the name of his optio, or second-in-command, Agricola, is still easily readable. It is this latter individual who was most likely overseeing the soldiers’ work at the quarry. And while there is no direct dating evidence to conclusively link this text to a rudimentary face carved directly above it, it is tempting, Mike Collins (Inspector of Ancient Monuments for Historic England on Hadrian’s Wall) suggests, to interpret the sketch as possibly representing a cheeky caricature of their boss drawn by the troops.

‘Quarrying huge amounts of stone and moving them out of the gorge to Hadrian’s Wall was a massive logistical undertaking, and finds like this add a human dimension to the bigger industrial picture,’ he said. ‘People will be people, and the idea that amongst the hard, dangerous work of quarrying someone paused to caricature their boss has an appeal – it helps us empathise with the soldiers and think about the inscriptions not only as important historical records but as artefacts created by people.’


The Agricola inscription is carved along one side of a prow-shaped protrusion of rock that sticks prominently from the quarry wall, and on its other face is another inscription recording the name of a different officer in charge of the officina. This area of the rock face is otherwise completely devoid of toolmarks, Jon says, and the two inscriptions appear to have been deliberately sited in order to display their words clearly for those walking along the far riverbank.

The second inscription, RIB 1009, is a remarkable historical source as it provides unusually precise clues to help narrow down when the Gelt quarry was in operation. Its text reads: APRO ET MAXIMO [with an unusual +-shaped X] CONSVLIBVS OFICINA MERCATI, or ‘In the consulship of Aper and Maximus, the working-face of Mercatius’. The appearance of these two consuls allows us to pin down the presence of Mercatius’ troops to a single year, AD 207 – making the Rock of Gelt the only quarry in the Roman Empire that can be dated so exactly. This episode of quarrying, at least, can therefore be placed within the reign of Septimius Severus, just before he began his campaigns in Scotland (AD 208-211) – a fascinating insight into efforts to reinforce the fortifications before his armies advanced north. The informative inscription is not the last we hear of Mercatius – a fuller version of his name also appears 30cm to the east, where we find the words MERCATIUS FERNI.

Aside from written inscriptions, the rock is also home to a number of images, including a carefully sculptured phallus (a symbol of good luck or protection against evil during the Roman period – see CA 315) which Jon spotted this year. Meanwhile, the ‘caricature’ mentioned above is not the only human face to grace the quarry wall; a short distance above it is a much more naturalistic outline, identified by Jon last year, which shows a male head in profile almost like a bust or a portrait on a coin. Facing to the left and executed in (rather worn) shallow relief, the lines of the individual’s forehead, nose, and chin can be clearly seen even when looking up from the riverbank metres below.

A bust engraved on the quarry wall
Preserved in almost pristine condition, this bust is accompanied by an inscription that might spell the end of its subject’s name. [Image:
Jon Allison, courtesy of Historic England]

Another even clearer bust lies a short distance to the right, identified during this year’s work on the site. It is sculptured into a rocky projection that sticks out at 45 degrees to the rest of the quarry wall, meaning that it has been shielded from the water that has run over, and eroded, many of the other inscriptions. Instead, this carving looks almost pristine – though, despite its brand-new appearance, the team is confident that it too is Roman. The outline is associated with an inscription that was initially interpreted as the name GAIUS, though Jon now suspects that the first two letters are in fact more likely to be an O and an unusually wide N. There are plenty of male Roman names that end in -onius, though one candidate that Jon finds particularly tempting is ‘Apollonius’, whose name is preserved in an inscription recorded by Collingwood Bruce at another Cumbrian quarry site in Combcrag Wood, just 350m south of Hadrian’s Wall.


As work continues to document the inscriptions, the aims of the project are two-fold: to create a lasting record of the eroding archaeology, and to make the remarkable quarry carvings more accessible to the general public – particularly now that they are so difficult and dangerous to see up close.

‘This site is a very important source of quarry inscriptions, because of their quality, their quantity, and the human aspect of their content,’ said Mike Collins. ‘But they are gradually eroding – some of the inscriptions that were very clear in the 18th century are now nearly unreadable, and we owe it to future generations to create as full a record as possible and to make it widely available, particularly to the many local people who know and love these woods and their Roman inscriptions.’

Jon added: ‘The Written Rock of Gelt is prominent amongst the other inscribed Roman quarries of Hadrian’s Wall frontier because not only can it be tightly dated; it also provides a wealth of information about people who laboured there and the military units to which they belonged. Camden’s 1607 edition of Britannia might not have been the first documented account, however. The text of RIB 2238, OFICIVM ROMANORVM, provides a post-Roman caption to the Written Rock; it is translated as ‘the working place of the Romans’. Early commentators, including Camden, noted that the font of the letters M looked to be Norman in style, but we cannot examine the evidence for ourselves because the inscription was chiselled out and a date of AD 1821 added.’

A phallic engraving
In the Roman period, phallic imagery was associated with good luck. [Image:
Jon Allison, courtesy of Historic England]

Jon would be delighted to hear from potential volunteers to conduct structure-from-motion photogrammetry in other Roman quarries and for geophysical survey along Hadrian’s Wall. He can be contacted at [email protected]
To explore the new 3D models of the inscriptions, visit

Roman Inscriptions of Britain vol. III, by R S O Tomlin, R P Wright, and M W C Hassall, Oxbow Books, £70, ISBN 978-1842173688; see also

This article appeared in Issue 351 of CA. Click here to subscribe.

Leave a Reply

Your email address will not be published.