Subscribe now for full access and no adverts
Known to its inhabitants as ‘Isca’, the Roman fortress at Caerleon (near Newport in south-east Wales) was one of three permanent legionary bases established in Britain, together with Chester (Deva) and York (Eboracum). Probably founded c.AD 75, during campaigns against the local Silures tribe, the 50-acre (20ha) complex occupied a strategic crossing point on the River Usk, and was home to more than 5,000 soldiers from the Second Augustan Legion (Legio II Augusta, part of the original invasion force of AD 43). Today, two-thirds of the fortress lies beneath the modern town, while the remainder (to the west) is covered only by green fields. The site’s long history of archaeological interest – from 19th-century antiquarian explorations to more methodical investigations beginning with the Wheelers and the Caerleon Excavation Committee in the 1920s, as well as more recent insights arising from development-led digs – means that much of its layout has been gradually pieced together. This picture remains patchy, however, particularly in the western portion of the fortress’ interior, which has been less impacted by urban expansion and less intensively explored.
Important new details were added to this plan in 2006-2007, when Cardiff University and GeoArch carried out magnetometry and resistivity surveys within two previously unexcavated areas of the fortress, revealing the footprints of numerous buildings (the first surveys in what would become the Mapping Isca project). In Priory Field, which covers the south-eastern corner of Isca’s interior close to its western gate, the team identified a row of eight barrack blocks and a trio of 40m-long rectangles that (with hints of long internal walls that could have supported raised floors) were interpreted as granaries. While the outlines of these groups of structures (which tend to be fairly standardised across Roman military sites) made it easy to pin down their purpose, others were less clear-cut. Facing the granaries across a broad open yard was a large (60m by 64m), squarish structure comprising four ranges of rooms surrounding a central courtyard. On other military sites, such buildings have been variously interpreted as hospitals, workshops, and storehouses, and there are also parallels with the civilian horrea, or warehouses, that still stand in the port town of Ostia outside Rome.
A team from Cardiff University and UCL set out to learn more, and (after gaining Scheduled Monument consent from Cadw, which manages the site on behalf of the Welsh government), over the course of ten weeks split across 2008 and 2010 they opened a large trench over the courtyard building’s west (front) range and part of the yard on to which it faced. This was the first excavation inside the fortress where trenches had been positioned to answer specific research questions (instead of reflecting the location of housing developments), the first Caerleon dig to involve university students, and the first that was open to the public, attracting more than 7,000 visitors over the two seasons. CA 226 and 268 reported on the project’s progress at the time, but we are revisiting Caerleon in this issue as the excavation’s full findings and subsequent analysis have now been published in a new monograph by the investigation’s co-directors, Dr Peter Guest (then Cardiff University, now Vianova Archaeology) and Professor Andrew Gardner (UCL; see ‘Further reading’ below).
Excavating Isca
To summarise the ‘story so far’, in 2008 and 2010 a team of archaeologists, students, and local volunteers excavated around 70 per cent of the western range, revealing seven internal spaces behind the remains of a substantial façade. Although their walls and even the foundations had been heavily robbed (in multiple episodes spanning the late Roman period to the 19th century, later analysis established), enough remained to attest to this being a masonry building rather than a timber structure set on a few dampcourses of stone. The building’s paved entranceway was flanked by two small chambers: one, which had a hard-wearing floor of opus signinum, was interpreted as a possible guard room, while the other, measuring just 1.5m wide, is thought to represent the remains of a stairwell leading to an upper storey (traces of a portico that could have supported a balcony for a second floor were identified on the edge of the courtyard). The remaining four square rooms were distinctly utilitarian in decor with plain, whitewashed walls and simple, beaten-earth floors. The contents of these functional spaces quickly confirmed that the courtyard building had indeed served as a storehouse – but more on that subject later.
While much of the masonry had been plundered for later recycling, the site’s stratigraphy was particularly well preserved, and this, together with post-excavation analysis of coins and pottery, has allowed Peter and Andrew to reconstruct a detailed chronology for the building’s use. It appears to have been built c.AD 90-100, apparently on a plot that had remained vacant since the fortress’s foundation some 15-25 years before, as no sign of timber predecessors or other hints of previous occupation were identified on this spot. During the construction process, the ground had been levelled with layers of refuse including broken pottery and the bones of animals that had presumably been consumed by Isca’s inhabitants. These latter finds, comprising the remains of cows, pigs, and sheep/goats, shed surprising light on how the early garrison had been provisioned. Some 37 bones were sampled for isotope analysis, which revealed that, while most of the animals had been reared locally as might be expected, a significant minority (at least 25 per cent) had come from much further afield. As we touched on in CA 334, and as is examined in greater detail in the new monograph, one group was found to have been raised on the chalklands of southern or central England, while another produced results compatible with much older geology found in Devon and Cornwall – indicating that the fortress community had relied, at least initially, on wide-ranging supply networks to feed its thousands-strong population.
The courtyard building appears to have remained in use for around 250 years, undergoing minor structural alterations and repairs (including resurfacing the entranceway) over this long period, but it seems to have been abandoned and allowed to fall into ruin in the first half of the 4th century. The central part of the range may have been deliberately demolished, as few facing stones or large pieces of roof tile were found in this area, hinting at a somewhat methodical dismantling and recycling of any useful materials. The rooms towards the end of the row, however, contained quantities of broken tegulae and imbrices, suggesting that these spaces had deteriorated until their roof fell in, again c.AD 300/350. The resulting debris had sealed the floors of Rooms 2 and 7, and beneath this was a wealth of evidence revealing that it was not only the building that had been seemingly abandoned, but its contents.
Roman recycling?
In Room 2, the floor was littered with fragments of iron and copper-alloy. Although somewhat scattered by the impact of the roof collapsing, their distribution suggests that they may represent objects stored on free-standing shelves around the edges of the room. Archaeological conservators from Amgueddfa Cymru-National Museum Wales lifted 38 soil blocks from this area – some containing single items, some containing masses of metal – and these were carefully transported back to the lab in Cardiff so that they could be X-rayed and the painstaking process of micro-excavation could begin. There, under carefully controlled conditions, contents of the blocks were gently teased from the surrounding soil. Much of the metal was very oxidised, distorting its shape beyond recognition by the naked eye, but X-ray imaging revealed that many of the clumps of rust disguised plates and rivets from dismantled military equipment. This included at least one set of lorica segmentata, the characteristic plate armour worn by legionaries in the 1st to 3rd centuries. Fairly complete examples have been found at Newstead fort in the Scottish Borders and Corbridge near Hadrian’s Wall, revealing that the kit’s style had evolved over time. The pieces from Caerleon represent the simpler Newstead style, but the cuirass (or cuirasses) from which they came appears to have been stripped of its copper-alloy fittings, which were not found, suggesting equipment that had been dismantled for repair or recycling.
Two more soil blocks produced around 110-120 thin copper-alloy scales from a different kind of armour: the flexible lorica squamata, which comprised thousands of such scales sewn on to a fabric backing. Some of the scales found in Room 2 still had the copper ties that would have fixed them to this material, and this area of the storeroom also yielded a larger rectangular piece from a breastplate and a thin scrap of sheet decorated with an image of a Capricorn – one of the emblems of the Legio II Augusta – delicately picked out in tiny, closely spaced punch marks. It was not only equipment for humans that had been kept in this space, however. A series of metal studs and pins are decorative elements from a chamfron, a leather mask used to protect the head and face of a horse. Leather sheets from such items have been found at Newstead, Vindolanda, and Carlisle; all of these had been stripped of their ornaments, but holes in the leather preserve the designs that they had once formed. At Caerleon, we have the opposite situation: the leather backing had decayed, but the metalwork survived in a jumbled mass. Among this was a leaf-shaped plaque with an image of a head wearing a Phrygian cap. Comparisons to finds from Vindolanda indicate that this was from a horse’s ear-guard, but its partner is missing, as are other larger elements such as the domed eye-guards associated with chamfrons, again suggesting an incomplete item.
Given the lack of weaponry or other military items such as soldiers’ belt fittings, this room does not have the feel of an armoury. Indeed, much of the equipment described above would have been very old at the time of the roof collapse, and its dismantled state and missing parts indicate that the armour was far from battle-ready. Instead, the team suggest that this room might have been used to store ‘special’ items that were only brought out on particular occasions, or to keep antiquated armour that was in the process of being recycled.
As well as military equipment, the rooms appear to have been used to store the personal belongings of legionaries and their families, perhaps keeping them safe while soldiers were away from the fortress on policing, construction, or campaign duties. Brooches (including a colourful six-pointed flower and three shaped like fish), finger rings, glass beads, bone and metal pins, fragments of bracelets (including a shale example that is so small as to suggest a very young wearer), and part of a copper-alloy mirror speak of the presence of men, women, and children. There were also decorative pieces of metalwork (including mounts depicting a lion, a cat, and a helmeted deity) from furniture and boxes or caskets. The handful of locks and keys recovered by the team could have helped to secure some of these containers, while sacks or parcels may have been tagged with the lead labels that were found in multiple places. Cut from thin lead sheet with a hole punched in one corner, their cursive writing was not always easy to decipher but usually included a name in its genitive (possessive) form, as if to indicate ‘property of [name]’. From these, the names of some of the men associated with the fortress come back to us across the span of centuries: Candidus, Aurelius Severus, Iulius Iucundus, and Marinianus.
Post-Roman occupation
Caerleon’s later history is not as well-understood as its legionary life. It is thought that Legio II Augusta departed c.AD 300, but rubbish in the fortress’ baths and barracks indicates that some occupation continued into the 4th century. Perhaps a small detachment of soldiers had stayed behind for a period; similar activity in the extra-mural settlements suggests that this was not simply a case of the civilian population moving inside the fortifications.
Above & below: A brooch depicting a six-pointed flower and another representing a fish.
The finds from Priory Field add to this picture, as two new masonry buildings were built among the partially demolished ruins of the storehouse, using its substantial front wall (which must have been still standing to some degree) as a spine. Building 1 is the easiest to reconstruct: a three-roomed rectangle measuring 11.75m by 3.7m, it had been built outside the storehouse, against its façade. One of its walls had toppled outwards, leaving a spread of 14-15 courses attesting that it was built entirely from stone and may have stood up to 3.5m tall – and the reason for this collapse was immediately apparent. Despite its sturdy walls, Building 1 was very superficially constructed, with no mortar to secure its masonry and no foundations: it had been built directly on to the cobbled surface of the outer yard. Building 2, meanwhile, is represented only by a single, short stretch of masonry sitting on top of a stub of wall that had once divided the west range’s Rooms 2 and 3.
As for when these structures were built, the associated material culture looks characteristically late Roman, but radiocarbon dating tells a different story: charred grain from a stone-lined pit in Building 1 has produced results clustered around AD 500. Evidence of 5th-/6th-century occupation has been identified at other Roman sites, such as Birdoswald and Wroxeter, but there the new structures appear to have been timber. Caerleon’s masonry buildings seem less like a break with tradition, Peter and Andrew note, and the apparent continued use of culturally Roman items is intriguing. ‘It is clear that, in Caerleon at least, terms like “late Roman” and “post-Roman” are not mutually exclusive and probably should be considered as political, rather than chronological or cultural labels’, the monograph notes. ‘Although Caerleon had changed very dramatically from 300 to 500, it seems to have remained a place where a Roman way-of-life continued long after the demise of Britannia.’
Further reading:
Peter Guest and Andrew Gardner (2025) Excavations in the Roman Legionary Fortress at Caerleon: The Priory Field Store Building, 2007-2010 (Archaeopress, ISBN 978-1803276908, £50; open access ebook: DOI 10.32028/9781803276908).
See https://vianovaarchaeology.com/roman-caerleon for more on recent research at Caerleon, including the excavations at Priory Field.
All images: Vianova Archaeology & UCL, unless otherwise stated
