Straight to the point: Investigating Roman swords and extensive settlements in the Cotswolds

This article is from Current Archaeology issue 426

Subscribe now for full access and no adverts

In CA 405, we reported on the rare discovery of a pair of Roman cavalry swords, or spathae, which were found during a metal-detecting rally near Willersey in the north Cotswolds in 2023. The blades were promptly reported to the Portable Antiquities Scheme, and their finder, Glenn Manning, and the landowner subsequently donated them to the Corinium Museum in Cirencester. Since then, the swords have undergone expert analysis and conservation, as well as X-ray imaging at Historic England’s Fort Cumberland science facilities (see box below) – and this year archaeologists returned to the findspot to learn more about the context of this unusual find. The investigations, funded by Historic England and directed jointly with Cotswold Archaeology, revealed that the spathae appear to have been buried within the grounds of a substantial building, possibly even a corridor villa, and also shed light on an area of extensive Iron Age settlement that preceded this activity.

The first clues to this rich archaeological landscape emerged during geophysical surveys by Historic England in November 2023, focusing on two fields around the location where the swords had been found. A previous study of aerial photos had not indicated the presence of buried remains, said Helen Winton, Head of Historic England’s Archaeological Investigation Team, but magnetometry and resistivity told a very different story. These showed clear traces of significant prehistoric and Romano-British remains concealed beneath a surface shaped by medieval ridge-and-furrow cultivation.

Within the northern field, part of a large, rectangular enclosure could be seen, as well as smaller, sub-circular enclosures and what appeared to be a ‘ladder settlement’. The rest of the rectangular enclosure extended into the southern field – and, in its northwest corner, earth resistance survey detected the possible corner of a masonry building. The same field also contained a smaller, double-ditched, sub-rectangular enclosure about 70m (230ft) south of the main enclosure. Most significantly, however, earth resistance detected part of a substantial building overlying the southeastern corner of the enclosure ditch, which has been provisionally interpreted as a potential Romano-British corridor villa, though more fieldwork is required to confirm this. It appeared to have a narrow central range with wings flanking each end – and another possible wall could be seen running along the southern edge of the rectangular enclosure.

Cutting-edge research

Spathae take their name from a word that roughly translates to ‘stirrer’, possibly because their blades were significantly longer than those of the gladii used by Roman infantrymen. They were used primarily from horseback, both by cavalrymen and by civilian travellers who wanted additional protection during their journeys, but during the 2nd century spathae were also adopted by foot soldiers (who perhaps appreciated the extra reach), eventually completely replacing gladii as their main weapon.

The examples from Willersey are thought to have been still in their scabbards at the time of burial: while the wood has not survived well, their copper-alloy chapes and other fittings were present in the ground. Since their excavation, they have been sent to Historic England’s laboratories at Fort Cumberland, near Portsmouth, to be X-rayed. The images (below) revealed intriguing details concealed beneath corrosion, showing that the iron blades had been made in very different ways. One was plain in design, but the other had been pattern-welded, suggesting that it was a much-more-expensive and higher-status item.

A handful of single spathae have been found elsewhere in Britain, but there is only one other known example where two such blades had been buried together. Discovered by Canterbury Archaeological Trust in 1977, during excavations at Canterbury Castle, they had a dramatic story to tell. Two men were found lying face-down in a pit, with the swords seemingly thrown in on top of them (CA 62). The twisted position of their bodies and the haphazard arrangement of their limbs suggests that they had been thrown into the pit very roughly, and the grave itself was irregular in shape, possibly hastily dug. The burial was dated to the later 2nd century and, significantly, it had taken place within the walls of the settlement, something that was strictly taboo in Roman culture – which perhaps indicated a clandestine act. While no sign of violence was observed on the skeletons themselves, this unusual burial has been interpreted as evidence of a double murder.

Here, too, the swords were still in their scabbards, and were found alongside bronze and iron belt fittings characteristic of military items. In the British Museum’s recent Legion exhibition, the burial featured in a section highlighting the hazards faced by Roman soldiers who were tasked with policing occupied territories (CA 409). Might these have been military men who were murdered while they were on duty, after which their weapons were buried with their bodies to hide the evidence – or tossed into the grave as a gesture of contempt for the regime that they represented?

Iron Age activity

In order to shed more light on these intriguing anomalies, and to explore the burial circumstances of the swords in more detail, in January and February of this year a team of professional archaeologists and volunteers, as well as Historic England trainees and members of their net-zero team, travelled to the site in order to carry out targeted excavations. Directed by Peter ‘Buzz’ Busby of Cotswold Archaeology and Ian Barnes of Historic England, the team persisted through very challenging weather conditions and soon relocated the spot where the swords had been discovered, inside the large rectangular enclosure that had been identified during the geophysical surveys.

Their investigations also revealed illuminating information about the prehistoric settlement that had been hinted at during this previous work. The sub-circular ditches proved to represent the remains of a number of large roundhouses, up to 18.5m in diameter, whose surrounding gullies produced quantities of pottery placing the structures in the early to middle Iron Age (it is hoped that post-excavation analysis, which is ongoing, will refine this date further). As some of their outlines are intercutting, the roundhouses are thought to reflect multiple phases of activity rather than a single large settlement. The rectangular enclosure, meanwhile, is thought to have been a still-later addition, as its assumed inner bank would have covered some of the ring gullies.

Possibly dug during the late Iron Age or the very early Roman period, the rectangular enclosure covers an area of 68m by 55m. Its surrounding ditch was described by Buzz as ‘magnificent’, measuring 5m wide and more than 1.5m deep (the team hit the water table before reaching its base). The fills of this boundary produced a number of enigmatic finds, including the bones of a human arm and hand. As the fingers were still in their correct anatomical position, this limb appears to have entered the ground when it was still fleshy and articulated – and, slightly further along the ditch, the team found another collection of finger bones hinting at the presence of another hand. It is not yet clear whether these finds represent the scattered deposition of a dismembered body or the burial of selected portions of different individuals, Buzz said, but the remains are more in keeping with Iron Age traditions than Roman rites.

An intact individual was discovered, however, in Trench 2, located a short distance to the north-east. There, the team found the articulated skeleton of an adult who had been laid on their back with their knees drawn up. Unusually, this person had been buried with an iron ring around their right upper arm, and a complete pony skull was found in an adjacent pit. The burial lies just within the rectangular enclosure, but could alternatively be associated with a roundhouse that butts up against the ditch, perhaps placed within the dwelling as some kind of closing gesture.

Roman remains

What about the Roman features? Trench 4 was opened to investigate the possible masonry corner over the north-western corner of the rectangular enclosure, and this did prove to be a clear, L-shaped portion of a building, albeit one that had been heavily robbed out, leaving behind a trench full of limestone rubble. The main wall’s construction-cut could be seen continuing beyond this, and another, somewhat slighter wall appeared to lead off from the robbed corner at a right angle, seemingly aligned with the ‘wall’ that had been noted along the southern edge of the rectangular enclosure and the walls of the possible villa in the survey results.

What kind of structures might these walls have belonged to? The most illuminating finds came from Trench 2, which had been opened at the top of the rectangular enclosure to examine the possible ladder settlement. Within this area, the team found a large, shallow pit that was packed full of Roman building materials, including ceramic roof tiles, box-flue tiles, and quantities of brightly painted wall plaster. ‘This was not a common dwelling, but something very upmarket,’ Buzz commented.

Whether the structure to the east of the rectangular enclosure can be identified as a villa remains an open question – the ‘corridor’ building spotted in the survey results has not been excavated – but the demolition materials recovered nearby speak of a high-status construction of considerable size. The ceramic roof tiles, which include both curved and flat forms, would have formed a heavy roof that needed the support of substantial walls. The painted plaster also hints at wealthy inhabitants, while the box-flue tiles testify to the presence of a hypocaust somewhere nearby, possibly serving a bathhouse, which would be a key aspect of any villa complex.

Further trenches opened on the final day of the project revealed the remains of further limestone structures – possibly ancillary buildings – to the north of the masonry corner. As for the slender wall running from the robbed-out section and perhaps edging the rectangular enclosure, Buzz suggests that the ditched space had later been replaced by a walled area – possibly a garden or an internal courtyard to the west of the possible villa – that was partly built over its predecessor, while still preserving its line. It was within this walled space that the swords had been buried.

Re-excavation of the findspot confirmed that the cut in which they had been placed was only small, and as there was also no sign of human remains, the blades are not thought to have been associated with a grave. Instead, Buzz offered a number of possible scenarios that could explain their burial. The first might see a wealthy family who had fallen on hard times burying their most prized possessions (the swords were found together with a copper-alloy bowl) in the garden to hide them from debt collectors, with the intention of retrieving them in happier times. Alternatively, mindful of the prolific Anglo-Saxon settlement evidence known from the surrounding area, a later date of burial could have been linked to differently turbulent times. Finally, given the apparent continuity from, or reuse of, the Iron Age enclosure, Buzz wonders if the swords’ owners may have made some kind of ancestral association with the site. As post-excavation analysis continues and more precise dating evidence emerges, there may be further clues to suggest when – and why – such prestigious swords were committed to the ground in this corner of a Cotswold field.