Located around 12 miles west of Hull, Brough-on-Humber is traditionally associated with Roman Petuaria, a settlement listed in Ptolemy’s 2nd-century AD Geographia. During the Iron Age, it lay within the territory of the Parisi, a people associated with the distinctive square barrow burials and chariot burials of the Arras Culture (see CA 363), and it has been suggested that the settlement became a civitas, or tribal capital, after the Roman conquest, although this interpretation is not uncontroversial. Nonetheless, Brough enjoys an indisputably prominent spot on what was the main Roman road to Eboracum (York, c.30 miles to the north), and on the Humber Estuary, which could have provided an easy route south towards Lincoln.
Evidence of Roman activity has been documented in the area since the 17th century, but it was another 300 years before any systematic excavations took place. The first, in 1933-1937, focused on the Burrs Playing Field (then known as Bozzes Field) in the centre of the modern town. It was directed by Philip Corder (later Secretary of the Society of Antiquaries of London), with the rather appropriately named Rev. Thomas Romans, and revealed traces of a sequence of forts, a stone wall with external towers, and various buildings. The highlight of their work, however, was a stone inscription (RIB 707; see www.romaninscriptionsofbritain.org/inscriptions/707) discovered by their foreman, Bertie Gott, and now held by Hull Museum. Its text recorded the presentation of a proscaenium, or stage for a theatre, by Marcus Ulpius Januarius, an aedile (an official responsible for public buildings and other civic duties) during the reign of Antoninus Pius, c.AD 140. The inscription is almost complete, but is missing a corner, cutting off the end of the name of the town receiving the stage. It begins ‘PETV –’; might we reconstruct this as Petuaria?
It would be another generation before Brough would see further archaeological work, but in 1958- 1961 John Wacher identified traces of early military activity dating to the Flavian period (AD 69-96), perhaps relating to the Roman conquest of this area of England c.AD 71. He also uncovered the foundations of an impressive stone gateway associated with a walled enclosure and other buildings. Wacher’s report, published by the Society of Antiquaries of London in 1969, remains the only monograph on this site, though this would not be the end of excavations there. Between 1977 and 1978, as modern Brough expanded, the East Riding Archaeological Society (ERAS) excavated on the northern fringes of the growing town, revealing substantial stone structures several hundred metres to the north-west of the Burrs, on the edge of a former tidal inlet of the River Humber – perhaps the remains of a harbour or landing stage of some kind.
In the early 1980s, ERAS went on to excavate on the opposite side of the road to the Burrs, uncovering further evidence of activity north of the walled settlement, while a resistivity survey on the Burrs itself, carried out in 1988 by Steve Jallands of Durham University, located a number of structures and a roadway. These findings would prove to be just a sample of what was preserved beneath the field, however, as in 2014 we carried out our own magnetometer survey, revealing that the Roman site was much more complex and extensive than was previously thought.
The anomalies were not as clear as might have been expected within a Roman town, but they attracted the interest of Martin Credland, then chair of the Elloughton cum Brough Town Council and Playing Fields Association, who was keen to find a new focus for promoting Brough’s heritage, as the town had long been associated with aircraft manufacturing, but was at that time facing the loss of its BAE Systems factory. To help celebrate Brough’s Roman story, a community heritage project called ‘Petuaria ReVisited’ was born, and thanks to funding from local businesses and the Centurion Club (through which interested members of the public provided donations), between November 2018 and April 2019 David Staveley was commissioned to carry out a GPR survey covering most of Burrs and associated gardens within the scheduled area. These investigations proved illuminating, revealing the changing outlines of the Roman structures in amazing detail.
Some of the features that we could see matched those excavated by Corder, particularly the line of the town defences and a number of rectilinear buildings, but the remains detected during our survey were far more impressive and extensive than hitherto supposed. They included a substantial road flanked by buildings, some of considerable size, which could be seen running roughly south-west to north-east across the whole length of the field, with side branches extending to the north. Immediately to the east of the road, resistivity survey by ERAS (also undertaken in November 2018) confirmed the presence of further buildings – but the most noteworthy part of the site lay in the south-western corner of the field. There, GPR had revealed a complex of features that appeared to be a series of rooms arranged around a courtyard, at the centre of which was a substantial D-shaped feature. Could this represent the footprint of the Roman theatre to which the famous inscription referred? To find out more, we needed to assess the buried archaeology through excavation.
Evidence of longevity
Scheduled Monument Consent was granted by Historic England to open a 25m by 3m trench over the D-shaped anomaly and adjacent features, and on 21 August 2020 our 16-day excavation began (funded by local businesses, the Royal Archaeological Institute, and the Roman Society). As soon as the topsoil was stripped away, we could see why the initial magnetometer survey had produced such ‘blurry’ images: a layer of crushed limestone, which, we learned, had been deposited on the instruction of the Department of the Environment in order to protect the underlying archaeology when the Burrs was converted from farmland to a playing field in 1971-1972. The GPR had been able to penetrate below this layer much more effectively than the magnetometry.
At the northern end of the trench, we encountered our first Roman feature: a hearth that contained a burnt coin dating to around AD 330 (another similar coin was found adjacent to it). The hearth lay in the centre of a floor surface that was surrounded on three sides by slots that had once contained walls, and these matched the footprint of one of the rooms in the courtyard building that had been revealed through GPR. This structure seemed to have been remodelled several times, but there were still plentiful clues to indicate how it may have looked. Beneath the rubble covering the remains of the southernmost room was a layer of very fragmented painted wall plaster, which had fallen face-down onto the floor surface – its back still bore indentations of the wattle work that it had covered. The best-preserved section was lifted, revealing a pattern of red and black lines on a cream background, and the recovery of other colourful fragments including blue and purple suggests that this was a high-status building – an interpretation that was bolstered by the discovery of ceramic and stone roof tiles, as well as box flue tiles from a hypocaust.
Whatever the structure’s status, however, it had been comprehensively and methodically demolished, with its building materials extensively robbed for recycling elsewhere. Beyond the southernmost robber trench, we found a thick spread of rubble stretching for around 4m, and in this location only a single block of facing stone from one of the walls remained. Underneath the rubble, in the centre of the trench, was a regular layer of stones running at an angle – an exciting development, as this feature coincided with the rear of the D-shaped anomaly in the GPR plots. The later-3rd- and 4th-century coins and pottery associated with it, however, make it unlikely that this line relates to a mid-2nd-century theatre, but as we investigated beneath it we found underlying stony layers provisionally interpreted as successive courtyard surfaces. Finds from these layers included roof tiles and iron nails, large quantities of oyster shell and plentiful butchered animal bone, and a well-preserved copper-alloy buckle with some decayed leather still adhering to it.
Over at the southern end of our excavation, we found large pieces of disturbed Roman stone, as well as echoes of Corder’s excavation, in the form of a narrow trench cutting through the plough soil. This area yielded pieces of Crambeck painted parchment ware pottery, one of the last fine wares to be manufactured in Roman Britain, as well as a crossbow-type brooch and coins. They add to a compelling picture; other artefacts excavated in 2020 – a ring-and-dot decorated bone needle case, a copper-alloy bangle and fragments of another made of shale, and jet beads – match late Roman assemblages found elsewhere in Roman Britain. Here, though, they are particularly important as evidence for the longevity of activity in Brough. Together, these objects provide evidence of continued Roman activity in Brough into the late 4th and probably 5th centuries AD, longer than previously thought, and contradicting previous conclusions about the end of Brough presented by Wacher, who suggested that it had faded out of use by the early 4th century.
Finds like these are also useful for testing the hypothesis that the walled enclosure with its external towers was some form of naval base controlling the River Humber and access to York, and possibly serving as part of the chain of coastal defences known as the Saxon Shore forts. Although the bulk of the coins we excavated date to the mid-4th century, there was an interesting cluster from the reigns of the usurper emperors Carausius (r.286-293) and Allectus (r.293-296). Wacher had at one stage argued that Brough may have become a naval base, and there may be a connection here, as Carausius had risen to power through his command of the Roman fleet in the English Channel.
The crossbow brooch, buckle, and a possible piece of scale armour hint at some kind of military association continuing long after the construction of the early forts. Perhaps the most remarkable find, however, was a simple oyster shell. This initially unprepossessing object held an enigmatic secret inside: its underside had been scored with a chequerboard and diamond pattern and, depending on its orientation, either a IX or XI. The meaning of these markings remains obscure, but it represents an intriguing link to the site’s Roman occupants.
Back to Brough
When we returned to the site last year (excavating 6-24 July), work continued on the main trench, removing the substantial wall core that we had excavated in 2020 to reveal a sand levelling layer. A further stone and clay foundation for a substantial structure lay underneath and slightly to the west of this, with two large facing stones of local limestone surviving – it is clear that, at some stage, this building too had been systematically robbed. In the centre of the trench, beneath the floor surface where we had found the collapsed section of painted wall plaster, we discovered clear signs of burning, including a dense rectangular patch of charcoal, which might represent the remains of a wooden panel. This burning might be contemporary with similar traces encountered by Corder in 1936; then, he described finding the remains of a building that had been destroyed in a large fire in the later 3rd century, as well as a sequence of subsequent buildings that had been constructed directly above the footprint of their predecessor.
Immediately to the south of the courtyard building’s southernmost ‘robber’ trench, we uncovered one of the more enigmatic finds from our latest season: a dense spread of Elland flagstone roof tile, some of which had slumped into the top of what turned out to be an ovoid pit measuring 2m by 1.5m. It was over a metre deep, and had been methodically packed with layers of oyster shell, large pieces of pottery, and roof tiles, before being carefully sealed with another layer of the last material. Dating to the 3rd century, this feature has been interpreted as possibly representing some form of structured deposition – that is, not a random rubbish dump, but a carefully curated collection that might have had some kind of ceremonial significance, perhaps marking the construction of the courtyard building itself.
Most of the artefacts recovered from the main trench in 2021 dated from the 3rd century, with some 4th-century material in the upper layers, including a decorated copper alloy bangle, a jet bead, and a pair of copper-alloy tweezers. The lower contexts contained pottery and other items from the 1st and 2nd centuries AD, however, including a well-preserved copper-alloy item known as a fob dangler – possibly part of a horse harness, which still had a portion of iron chain attached to it.
This most recent excavation also saw us open three further trenches in the north-eastern and north-western corners of the playing field, just outside the scheduled area, as well as another in a garden to the south. We were hoping to investigate the town defences, but the north-east trench immediately proved problematic when we encountered an unexpected 20th-century brick structure. It was not recorded on Ordnance Survey maps, and its purpose was not immediately apparent, though some suggested it could have been a culvert directing water into a feature locally known as ‘Roman Drain’ – in fact, a Roman ditch, one of the only elements of the settlement still visible above ground. Enquiries amongst Brough residents yielded another interpretation, though: it might be part of a Second World War air raid shelter, as an elderly resident had a recollection of German prisoners of war being marched from a nearby camp to demolish it. We did eventually find Roman levels, however – including, most importantly, the edge of a ditch that may correspond with a fort ditch excavated by Corder in 1935. Finds there included a sherd of a decorated Samian bowl and a piece of a rusticated jar, typical of the later 1st and early 2nd century AD, which supports an early date for this feature.
In the north-western corner of the Burrs Playing Field, Roman deposits were much closer to the surface, and although they were heavily disturbed by tree roots, we did find the remains of a stone wall documented by Corder, as well as the top of what he described as a clay and turf rampart. We also found traces of one of Corder’s trenches – and of one of his excavators, who had left behind a metal six-inch ruler in the fill. In our trench to the south, there was further disturbance in the form of a sewage pipe and a land drain, which had almost certainly been placed within a substantial Roman ditch. This area also yielded Roman pottery including a sherd of decorated colour-coated ceramic, most probably from the Rhineland, and greywares from the extensive Holme-on-Spalding Moor production centre that has been identified c.15 miles from Brough.
Above all, throughout 2020 and 2021 it has become clear that much of what was found in our main trench is strikingly similar to the structures and features that were identified in Corder’s ‘Site V’, just a short distance to the east. The remains that he uncovered form the continuation of the eastern range of the courtyard building that we were investigating, and although we have not yet found conclusive proof that the D-shaped anomaly was a theatre, we have reached contexts that may be of later 2nd-century date, and the search continues.
As our 2020-2021 excavations were conducted during the time of the Covid-19 pandemic, a major aim was to involve the community of Brough and its wider region in discovering their local heritage with a view to enhancing wellbeing.
Much of the recruitment of volunteers was done through the Services Archaeology and Heritage Association, by Diarmaid Walshe, with great success. Particular emphasis was placed on involving ‘Blue Light’ and NHS professionals, as well as serving and ex-military personnel. In our most recent excavation, a total of 128 volunteers excluding staff actively participated, and for 87 per cent of these it was the first time they had taken part in a dig. Over 52 per cent either lived or worked within a 3km (2-mile) radius of Brough, and the majority of the remainder came from Hull and the East Riding of Yorkshire more generally.
ALL IMAGES: P Halkon unless otherwise stated.