Artistic obscurity: analysing Britain’s most elusive Roman sculptures

Research in the hinterland of Hadrian’s Wall has revealed remarkable examples of Romano-British artistry hidden in plain sight. Ian Haynes and Lindsay Allason-Jones explore some of the key examples identified by the Elusive Sculptures project, and how these carvings have evaded archaeological attention in the past.


What might a pigsty, a chimney breast, a rock garden, and a font have in common? An obvious answer would, of course, be that they can all be made of stone – but for a special few there is a particular claim to distinction. Over the last three years, members of the British Academy/Leverhulme Trust-funded Elusive Sculptures team have been looking beyond the obvious in their quest to locate and interpret Romano-British art preserved in unlikely – and, at times, somewhat precarious – places in the north of England, and important examples have been recorded at these sites and more.

This unprepossessing pigsty in Tebay, Cumbria, in fact preserves a Romano-British carving of Hercules (below) among its more mundane masonry – just one of many previously undocumented sculptures identified by the Elusive Sculptures project.

Students of Roman archaeology will be familiar with the great undertaking that is CSIR, Corpus Signorum Imperii Romani, an initiative – or, rather, series of initiatives – to document Roman sculpted stone around the empire. The range of material recorded spans everything from the delicate addition of carved decoration on otherwise simple stone panels, to larger-than-life-size depictions of deities. Britain is especially well served: to-date, ten CSIR volumes have been published and much of the province is covered. Just one remains to be completed: material from what might be collectively termed the hinterland of Hadrian’s Wall – the counties of Northumberland south of the Stanegate; Tyne and Wear (items not included in CSIR I.1); County Durham; Cumbria; modern Lancashire; and Derbyshire – is yet to join the list. However, after a final push conducted against a challenging COVID-19 backdrop, 65 of Britain’s most elusive sculptures have now been added to the 500 stones from the area that had already been catalogued for this last British volume.

Cutting-edge imaging technology has been vital to the project’s efforts. Here, a mother goddess identified at St Cuthbert’s Church in Dufton, Cumbria, is documented in models derived using an Artec LEO White Light Scanner (ABOVE) and Structure from Motion photography (BELOW). ‘Sometimes,’ the team notes, ‘we do not know which system will work best before we try them both.’

There are many surprising ways in which ancient sculpture can elude scholarly attention. Hiding out in the fabric of a sty usually guarded by aggressive pigs is but one of any number of options, but less dramatic routes to obscurity are also possible. Some examples of sculpture proudly shown in plain sight still manage to avoid academic interest, while others have even made their way to museum stores and displays where, accessioned or not, they are positioned in a manner that defies access or close inspection. Worked pieces have occasionally been interpreted as something much more recent in date or, conversely, have been identified as Roman when they emphatically are not. Churches have proven fine repositories of ancient art, but the project team has also searched far beyond the ecclesiastical, scaling chimney breasts, barn walls, and even rock faces in their quest to bring this rich assemblage to a wider audience.

Discoveries and rediscoveries

As with all good archaeological work, the process of investigation must begin well before researchers take to the field. In this instance, a decade of research by Lindsay Allason-Jones as she compiled CSIR I.11 had identified a series of likely suspects that required further investigation. Some items, documented in the 17th and 18th centuries and long since believed lost, were painstakingly tracked down. To offer but a few examples, within Co. Durham five stones, previously thought lost, have been relocated at Ebchester, while at Lanchester three reliefs of genii (protective spirits), two of which were presumed lost, have now been identified as being one and the same stone on display in Chesters Museum on Hadrian’s Wall. Meanwhile, in Cumbria, a Roman tombstone last seen in 1889 and thought vanished has been relocated at Old Carlisle, while at Beckfoot a small altar to a genius, which had been variously published as dedicated to Diana, Mithras, and Luna Lucifera, and whose whereabouts was unknown, was rediscovered and correctly attributed.

Sometimes accessibility is the reason that a piece of art has eluded academic attention: here, project team members assess how to get to an alleged example of Roman sculpture located between two farm buildings.

But there were new discoveries too. A previously undocumented fragment of a figural tombstone was found built into the fabric of Brougham Castle (Cumbria), and at a private house at Penrith a rock garden yielded a further three fragments of sculpture unknown to archaeologists, probably also from Brougham. Built into the parish church in Gosforth, which is more famous for its early medieval sculpture, a relief of a Celtic deity was lurking, ignored, in plain sight, and also in Cumbria, at Tebay, the aforementioned pigsty yielded a relief of Hercules.

In some cases, there are tantalising ambiguities. Whether some of the items identified are truly new discoveries, or more properly rediscoveries, is open to debate. A torso, probably belonging to a Venus statue, built into a water feature in a private garden at Ravenglass (Cumbria) might conceivably be the one recorded by Hutchinson in 1794 – the lost Venus of Maryport – though the possibility that the two are distinct pieces must be allowed. Hutchinson’s published image suggests that rather more of his example’s arms may have survived, but also omits elements of drapery and hair tresses preserved on the torso from the pond.

This enigmatic torso, thought to have once been part of a statue of Venus, had been built into an arch associated with a pond in a private garden at Ravenglass in Cumbria.

What all this underscores is that tracking the pieces down is only one part of the process: it is then necessary to look at them very closely indeed, and for stones that have often withstood centuries of exposure to the elements, the very latest in imaging technology can make all the difference. Over the years, the Newcastle University team has experimented with an array of different techniques to record 3D surfaces with maximum precision. Take the Written Rock of Gelt, for example. Located near Brampton, 5.5km from Hadrian’s Wall, this Roman quarry wall is covered in graffiti left by 3rd-century soldiers who used its sandstone to repair the frontier fortifications. Intrepid investigations at this remarkable site, undertaken for Historic England and reported on in CA 351 (the project also won the 2020 CA Award for Rescue Project of the Year), depended on the use of SfM (Structure from Motion) photography.

Intrepid archaeology: Jon Allison is shown recording 3rd-century graffiti on the Written Rock of Gelt, a Roman quarry wall near Brampton.

How does this technique work? Simply put, SfM uses photographs taken from multiple overlapping angles which can, when processed through appropriate software, be brought together to generate a 3D image. Such an approach can be very effective, but the use of White Light Scanners offers an even more accurate system for object modelling. In the course of the Elusive Sculptures project, the team used both SfM and Artec series White Light Scanners to capture detail potentially hard to read with the naked eye, and to make the resulting images availably digitally for future researchers.

A warrior god identified at St Mary and All Saints Church in Whalley, Lancashire. PHOTO: Jon Allison/Newcastle University.

Key finds

Among the star pieces identified by the project were, perhaps unsurprisingly, those objects that had attracted little or no academic attention, but which had been cared for and displayed in churches over the centuries. Of these, three examples warrant further attention. The first is an example of a mural-crowned genius, once the guardian of a Roman settlement, which had been hiding unremarked since 1912 on a window ledge in the parish church at Bridekirk in Cumbria, a fine building better known for its remarkable 12th-century font complete with runic inscription. The second is a warrior god resplendent in his detailed 3rd-century militaria, who today oversees a play area for young parishioners at Whalley in Lancashire. For an especially striking example of notoriety co-existing with underestimated importance, however, we must surely note a Roman altar in the church of St John the Evangelist at The Lund (Lancashire). The records of the parish attest that in 1688, one Matt Hall, Churchwarden of Kirkham, installed ‘a scandalous trough’ as a font, an action that raised local outrage. Staring down the nave are three respectable Matres (mother goddesses), a familiar sight in Romano-British iconography, even if not otherwise associated with Christian baptism. More striking, however, is the detail on both sides, which are decorated with dancing figures. The dancers and Matres seen here are a juxtaposition unparalleled in our knowledge of Romano-Celtic art, and are startling even today; certainly, they were unlikely to have been to the taste of a 17th-century congregation. Do we catch here, in a Lancashire parish church, a unique glimpse of the rituals used to venerate those powerful figures in the hinterland of Hadrian’s Wall 1,800 years ago?

Recording the Scandalous Trough of Lund using an Artec LEO White Light Scanner.

Our quest for elusive sculptures in the hinterland of Hadrian’s Wall is now complete. It has been a marvellous journey, full of surprises. Along the way, we have been helped by an array of extraordinarily generous and welcoming people, and multiple tip-offs. The whole experience has been tremendously enjoyable, but it has also raised important questions about just how many other examples might still be hiding out in Britain and, indeed, across the Roman Empire, waiting to be rediscovered.

A digital model showing more detail of the Trough.
Ian Haynes and Lindsay Allason-Jones would like to thank their fellow travellers and accomplices Alex Turner and Jon Allison, and the many people who helped throughout this investigation. Particular thanks go to the British Academy and Leverhulme Trust for the small grant that made this work possible (SRG18R1\180588). Corpus Signorum Imperii Romani (Volume I, Fascicule 11: Sculpture from the Hinterland of Hadrian’s Wall) will be published shortly.